diff options
Diffstat (limited to 'arch/s390/kernel')
108 files changed, 34033 insertions, 0 deletions
diff --git a/arch/s390/kernel/.gitignore b/arch/s390/kernel/.gitignore new file mode 100644 index 0000000000..bbb90f92d0 --- /dev/null +++ b/arch/s390/kernel/.gitignore @@ -0,0 +1,2 @@ +# SPDX-License-Identifier: GPL-2.0-only +vmlinux.lds diff --git a/arch/s390/kernel/Makefile b/arch/s390/kernel/Makefile new file mode 100644 index 0000000000..0df2b88cc0 --- /dev/null +++ b/arch/s390/kernel/Makefile @@ -0,0 +1,85 @@ +# SPDX-License-Identifier: GPL-2.0 +# +# Makefile for the linux kernel. +# + +ifdef CONFIG_FUNCTION_TRACER + +# Do not trace tracer code +CFLAGS_REMOVE_ftrace.o = $(CC_FLAGS_FTRACE) + +# Do not trace early setup code +CFLAGS_REMOVE_early.o = $(CC_FLAGS_FTRACE) +CFLAGS_REMOVE_rethook.o = $(CC_FLAGS_FTRACE) + +endif + +GCOV_PROFILE_early.o := n +KCOV_INSTRUMENT_early.o := n +UBSAN_SANITIZE_early.o := n +KASAN_SANITIZE_ipl.o := n +KASAN_SANITIZE_machine_kexec.o := n + +# +# Passing null pointers is ok for smp code, since we access the lowcore here. +# +CFLAGS_smp.o := -Wno-nonnull + +# +# Disable tailcall optimizations for stack / callchain walking functions +# since this might generate broken code when accessing register 15 and +# passing its content to other functions. +# +CFLAGS_stacktrace.o += -fno-optimize-sibling-calls +CFLAGS_dumpstack.o += -fno-optimize-sibling-calls +CFLAGS_unwind_bc.o += -fno-optimize-sibling-calls + +obj-y := head64.o traps.o time.o process.o earlypgm.o early.o setup.o idle.o vtime.o +obj-y += processor.o syscall.o ptrace.o signal.o cpcmd.o ebcdic.o nmi.o +obj-y += debug.o irq.o ipl.o dis.o diag.o vdso.o cpufeature.o +obj-y += sysinfo.o lgr.o os_info.o +obj-y += runtime_instr.o cache.o fpu.o dumpstack.o guarded_storage.o sthyi.o +obj-y += entry.o reipl.o kdebugfs.o alternative.o +obj-y += nospec-branch.o ipl_vmparm.o machine_kexec_reloc.o unwind_bc.o +obj-y += smp.o text_amode31.o stacktrace.o abs_lowcore.o + +extra-y += vmlinux.lds + +obj-$(CONFIG_SYSFS) += nospec-sysfs.o +CFLAGS_REMOVE_nospec-branch.o += $(CC_FLAGS_EXPOLINE) + +obj-$(CONFIG_MODULES) += module.o +obj-$(CONFIG_SCHED_TOPOLOGY) += topology.o +obj-$(CONFIG_NUMA) += numa.o +obj-$(CONFIG_AUDIT) += audit.o +compat-obj-$(CONFIG_AUDIT) += compat_audit.o +obj-$(CONFIG_COMPAT) += compat_linux.o compat_signal.o +obj-$(CONFIG_COMPAT) += $(compat-obj-y) +obj-$(CONFIG_EARLY_PRINTK) += early_printk.o +obj-$(CONFIG_KPROBES) += kprobes.o +obj-$(CONFIG_KPROBES) += kprobes_insn_page.o +obj-$(CONFIG_KPROBES) += mcount.o +obj-$(CONFIG_RETHOOK) += rethook.o +obj-$(CONFIG_FUNCTION_TRACER) += ftrace.o +obj-$(CONFIG_FUNCTION_TRACER) += mcount.o +obj-$(CONFIG_CRASH_DUMP) += crash_dump.o +obj-$(CONFIG_KEXEC_CORE) += machine_kexec.o relocate_kernel.o +obj-$(CONFIG_UPROBES) += uprobes.o +obj-$(CONFIG_JUMP_LABEL) += jump_label.o + +obj-$(CONFIG_KEXEC_FILE) += machine_kexec_file.o kexec_image.o +obj-$(CONFIG_KEXEC_FILE) += kexec_elf.o +obj-$(CONFIG_CERT_STORE) += cert_store.o +obj-$(CONFIG_IMA_SECURE_AND_OR_TRUSTED_BOOT) += ima_arch.o + +obj-$(CONFIG_PERF_EVENTS) += perf_event.o +obj-$(CONFIG_PERF_EVENTS) += perf_cpum_cf.o perf_cpum_sf.o +obj-$(CONFIG_PERF_EVENTS) += perf_cpum_cf_events.o perf_regs.o +obj-$(CONFIG_PERF_EVENTS) += perf_pai_crypto.o perf_pai_ext.o + +obj-$(CONFIG_TRACEPOINTS) += trace.o +obj-$(findstring y, $(CONFIG_PROTECTED_VIRTUALIZATION_GUEST) $(CONFIG_PGSTE)) += uv.o + +# vdso +obj-y += vdso64/ +obj-$(CONFIG_COMPAT) += vdso32/ diff --git a/arch/s390/kernel/abs_lowcore.c b/arch/s390/kernel/abs_lowcore.c new file mode 100644 index 0000000000..f9efc54ec4 --- /dev/null +++ b/arch/s390/kernel/abs_lowcore.c @@ -0,0 +1,46 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <linux/pgtable.h> +#include <asm/abs_lowcore.h> + +unsigned long __bootdata_preserved(__abs_lowcore); + +int abs_lowcore_map(int cpu, struct lowcore *lc, bool alloc) +{ + unsigned long addr = __abs_lowcore + (cpu * sizeof(struct lowcore)); + unsigned long phys = __pa(lc); + int rc, i; + + for (i = 0; i < LC_PAGES; i++) { + rc = __vmem_map_4k_page(addr, phys, PAGE_KERNEL, alloc); + if (rc) { + /* + * Do not unmap allocated page tables in case the + * allocation was not requested. In such a case the + * request is expected coming from an atomic context, + * while the unmap attempt might sleep. + */ + if (alloc) { + for (--i; i >= 0; i--) { + addr -= PAGE_SIZE; + vmem_unmap_4k_page(addr); + } + } + return rc; + } + addr += PAGE_SIZE; + phys += PAGE_SIZE; + } + return 0; +} + +void abs_lowcore_unmap(int cpu) +{ + unsigned long addr = __abs_lowcore + (cpu * sizeof(struct lowcore)); + int i; + + for (i = 0; i < LC_PAGES; i++) { + vmem_unmap_4k_page(addr); + addr += PAGE_SIZE; + } +} diff --git a/arch/s390/kernel/alternative.c b/arch/s390/kernel/alternative.c new file mode 100644 index 0000000000..e7bca29f9c --- /dev/null +++ b/arch/s390/kernel/alternative.c @@ -0,0 +1,75 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/module.h> +#include <linux/cpu.h> +#include <linux/smp.h> +#include <asm/text-patching.h> +#include <asm/alternative.h> +#include <asm/facility.h> +#include <asm/nospec-branch.h> + +static int __initdata_or_module alt_instr_disabled; + +static int __init disable_alternative_instructions(char *str) +{ + alt_instr_disabled = 1; + return 0; +} + +early_param("noaltinstr", disable_alternative_instructions); + +static void __init_or_module __apply_alternatives(struct alt_instr *start, + struct alt_instr *end) +{ + struct alt_instr *a; + u8 *instr, *replacement; + + /* + * The scan order should be from start to end. A later scanned + * alternative code can overwrite previously scanned alternative code. + */ + for (a = start; a < end; a++) { + instr = (u8 *)&a->instr_offset + a->instr_offset; + replacement = (u8 *)&a->repl_offset + a->repl_offset; + + if (!__test_facility(a->facility, alt_stfle_fac_list)) + continue; + + if (unlikely(a->instrlen % 2)) { + WARN_ONCE(1, "cpu alternatives instructions length is " + "odd, skipping patching\n"); + continue; + } + + s390_kernel_write(instr, replacement, a->instrlen); + } +} + +void __init_or_module apply_alternatives(struct alt_instr *start, + struct alt_instr *end) +{ + if (!alt_instr_disabled) + __apply_alternatives(start, end); +} + +extern struct alt_instr __alt_instructions[], __alt_instructions_end[]; +void __init apply_alternative_instructions(void) +{ + apply_alternatives(__alt_instructions, __alt_instructions_end); +} + +static void do_sync_core(void *info) +{ + sync_core(); +} + +void text_poke_sync(void) +{ + on_each_cpu(do_sync_core, NULL, 1); +} + +void text_poke_sync_lock(void) +{ + cpus_read_lock(); + text_poke_sync(); + cpus_read_unlock(); +} diff --git a/arch/s390/kernel/asm-offsets.c b/arch/s390/kernel/asm-offsets.c new file mode 100644 index 0000000000..fa5f6885c7 --- /dev/null +++ b/arch/s390/kernel/asm-offsets.c @@ -0,0 +1,190 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Generate definitions needed by assembly language modules. + * This code generates raw asm output which is post-processed to extract + * and format the required data. + */ + +#define ASM_OFFSETS_C + +#include <linux/kbuild.h> +#include <linux/kvm_host.h> +#include <linux/sched.h> +#include <linux/purgatory.h> +#include <linux/pgtable.h> +#include <linux/ftrace.h> +#include <asm/idle.h> +#include <asm/gmap.h> +#include <asm/stacktrace.h> + +int main(void) +{ + /* task struct offsets */ + OFFSET(__TASK_stack, task_struct, stack); + OFFSET(__TASK_thread, task_struct, thread); + OFFSET(__TASK_pid, task_struct, pid); + BLANK(); + /* thread struct offsets */ + OFFSET(__THREAD_ksp, thread_struct, ksp); + BLANK(); + /* thread info offsets */ + OFFSET(__TI_flags, task_struct, thread_info.flags); + BLANK(); + /* pt_regs offsets */ + OFFSET(__PT_PSW, pt_regs, psw); + OFFSET(__PT_GPRS, pt_regs, gprs); + OFFSET(__PT_R0, pt_regs, gprs[0]); + OFFSET(__PT_R1, pt_regs, gprs[1]); + OFFSET(__PT_R2, pt_regs, gprs[2]); + OFFSET(__PT_R3, pt_regs, gprs[3]); + OFFSET(__PT_R4, pt_regs, gprs[4]); + OFFSET(__PT_R5, pt_regs, gprs[5]); + OFFSET(__PT_R6, pt_regs, gprs[6]); + OFFSET(__PT_R7, pt_regs, gprs[7]); + OFFSET(__PT_R8, pt_regs, gprs[8]); + OFFSET(__PT_R9, pt_regs, gprs[9]); + OFFSET(__PT_R10, pt_regs, gprs[10]); + OFFSET(__PT_R11, pt_regs, gprs[11]); + OFFSET(__PT_R12, pt_regs, gprs[12]); + OFFSET(__PT_R13, pt_regs, gprs[13]); + OFFSET(__PT_R14, pt_regs, gprs[14]); + OFFSET(__PT_R15, pt_regs, gprs[15]); + OFFSET(__PT_ORIG_GPR2, pt_regs, orig_gpr2); + OFFSET(__PT_FLAGS, pt_regs, flags); + OFFSET(__PT_CR1, pt_regs, cr1); + OFFSET(__PT_LAST_BREAK, pt_regs, last_break); + DEFINE(__PT_SIZE, sizeof(struct pt_regs)); + BLANK(); + /* stack_frame offsets */ + OFFSET(__SF_BACKCHAIN, stack_frame, back_chain); + OFFSET(__SF_GPRS, stack_frame, gprs); + OFFSET(__SF_EMPTY, stack_frame, empty[0]); + OFFSET(__SF_SIE_CONTROL, stack_frame, sie_control_block); + OFFSET(__SF_SIE_SAVEAREA, stack_frame, sie_savearea); + OFFSET(__SF_SIE_REASON, stack_frame, sie_reason); + OFFSET(__SF_SIE_FLAGS, stack_frame, sie_flags); + OFFSET(__SF_SIE_CONTROL_PHYS, stack_frame, sie_control_block_phys); + DEFINE(STACK_FRAME_OVERHEAD, sizeof(struct stack_frame)); + BLANK(); + /* idle data offsets */ + OFFSET(__CLOCK_IDLE_ENTER, s390_idle_data, clock_idle_enter); + OFFSET(__TIMER_IDLE_ENTER, s390_idle_data, timer_idle_enter); + OFFSET(__MT_CYCLES_ENTER, s390_idle_data, mt_cycles_enter); + BLANK(); + /* hardware defined lowcore locations 0x000 - 0x1ff */ + OFFSET(__LC_EXT_PARAMS, lowcore, ext_params); + OFFSET(__LC_EXT_CPU_ADDR, lowcore, ext_cpu_addr); + OFFSET(__LC_EXT_INT_CODE, lowcore, ext_int_code); + OFFSET(__LC_PGM_ILC, lowcore, pgm_ilc); + OFFSET(__LC_PGM_INT_CODE, lowcore, pgm_code); + OFFSET(__LC_DATA_EXC_CODE, lowcore, data_exc_code); + OFFSET(__LC_MON_CLASS_NR, lowcore, mon_class_num); + OFFSET(__LC_PER_CODE, lowcore, per_code); + OFFSET(__LC_PER_ATMID, lowcore, per_atmid); + OFFSET(__LC_PER_ADDRESS, lowcore, per_address); + OFFSET(__LC_EXC_ACCESS_ID, lowcore, exc_access_id); + OFFSET(__LC_PER_ACCESS_ID, lowcore, per_access_id); + OFFSET(__LC_OP_ACCESS_ID, lowcore, op_access_id); + OFFSET(__LC_AR_MODE_ID, lowcore, ar_mode_id); + OFFSET(__LC_TRANS_EXC_CODE, lowcore, trans_exc_code); + OFFSET(__LC_MON_CODE, lowcore, monitor_code); + OFFSET(__LC_SUBCHANNEL_ID, lowcore, subchannel_id); + OFFSET(__LC_SUBCHANNEL_NR, lowcore, subchannel_nr); + OFFSET(__LC_IO_INT_PARM, lowcore, io_int_parm); + OFFSET(__LC_IO_INT_WORD, lowcore, io_int_word); + OFFSET(__LC_MCCK_CODE, lowcore, mcck_interruption_code); + OFFSET(__LC_EXT_DAMAGE_CODE, lowcore, external_damage_code); + OFFSET(__LC_MCCK_FAIL_STOR_ADDR, lowcore, failing_storage_address); + OFFSET(__LC_PGM_LAST_BREAK, lowcore, pgm_last_break); + OFFSET(__LC_RETURN_LPSWE, lowcore, return_lpswe); + OFFSET(__LC_RETURN_MCCK_LPSWE, lowcore, return_mcck_lpswe); + OFFSET(__LC_RST_OLD_PSW, lowcore, restart_old_psw); + OFFSET(__LC_EXT_OLD_PSW, lowcore, external_old_psw); + OFFSET(__LC_SVC_OLD_PSW, lowcore, svc_old_psw); + OFFSET(__LC_PGM_OLD_PSW, lowcore, program_old_psw); + OFFSET(__LC_MCK_OLD_PSW, lowcore, mcck_old_psw); + OFFSET(__LC_IO_OLD_PSW, lowcore, io_old_psw); + OFFSET(__LC_RST_NEW_PSW, lowcore, restart_psw); + OFFSET(__LC_EXT_NEW_PSW, lowcore, external_new_psw); + OFFSET(__LC_SVC_NEW_PSW, lowcore, svc_new_psw); + OFFSET(__LC_PGM_NEW_PSW, lowcore, program_new_psw); + OFFSET(__LC_MCK_NEW_PSW, lowcore, mcck_new_psw); + OFFSET(__LC_IO_NEW_PSW, lowcore, io_new_psw); + /* software defined lowcore locations 0x200 - 0xdff*/ + OFFSET(__LC_SAVE_AREA_SYNC, lowcore, save_area_sync); + OFFSET(__LC_SAVE_AREA_ASYNC, lowcore, save_area_async); + OFFSET(__LC_SAVE_AREA_RESTART, lowcore, save_area_restart); + OFFSET(__LC_CPU_FLAGS, lowcore, cpu_flags); + OFFSET(__LC_RETURN_PSW, lowcore, return_psw); + OFFSET(__LC_RETURN_MCCK_PSW, lowcore, return_mcck_psw); + OFFSET(__LC_SYS_ENTER_TIMER, lowcore, sys_enter_timer); + OFFSET(__LC_MCCK_ENTER_TIMER, lowcore, mcck_enter_timer); + OFFSET(__LC_EXIT_TIMER, lowcore, exit_timer); + OFFSET(__LC_LAST_UPDATE_TIMER, lowcore, last_update_timer); + OFFSET(__LC_LAST_UPDATE_CLOCK, lowcore, last_update_clock); + OFFSET(__LC_INT_CLOCK, lowcore, int_clock); + OFFSET(__LC_BOOT_CLOCK, lowcore, boot_clock); + OFFSET(__LC_CURRENT, lowcore, current_task); + OFFSET(__LC_KERNEL_STACK, lowcore, kernel_stack); + OFFSET(__LC_ASYNC_STACK, lowcore, async_stack); + OFFSET(__LC_NODAT_STACK, lowcore, nodat_stack); + OFFSET(__LC_RESTART_STACK, lowcore, restart_stack); + OFFSET(__LC_MCCK_STACK, lowcore, mcck_stack); + OFFSET(__LC_RESTART_FN, lowcore, restart_fn); + OFFSET(__LC_RESTART_DATA, lowcore, restart_data); + OFFSET(__LC_RESTART_SOURCE, lowcore, restart_source); + OFFSET(__LC_RESTART_FLAGS, lowcore, restart_flags); + OFFSET(__LC_KERNEL_ASCE, lowcore, kernel_asce); + OFFSET(__LC_USER_ASCE, lowcore, user_asce); + OFFSET(__LC_LPP, lowcore, lpp); + OFFSET(__LC_CURRENT_PID, lowcore, current_pid); + OFFSET(__LC_GMAP, lowcore, gmap); + OFFSET(__LC_LAST_BREAK, lowcore, last_break); + /* software defined ABI-relevant lowcore locations 0xe00 - 0xe20 */ + OFFSET(__LC_DUMP_REIPL, lowcore, ipib); + OFFSET(__LC_VMCORE_INFO, lowcore, vmcore_info); + OFFSET(__LC_OS_INFO, lowcore, os_info); + /* hardware defined lowcore locations 0x1000 - 0x18ff */ + OFFSET(__LC_MCESAD, lowcore, mcesad); + OFFSET(__LC_EXT_PARAMS2, lowcore, ext_params2); + OFFSET(__LC_FPREGS_SAVE_AREA, lowcore, floating_pt_save_area); + OFFSET(__LC_GPREGS_SAVE_AREA, lowcore, gpregs_save_area); + OFFSET(__LC_PSW_SAVE_AREA, lowcore, psw_save_area); + OFFSET(__LC_PREFIX_SAVE_AREA, lowcore, prefixreg_save_area); + OFFSET(__LC_FP_CREG_SAVE_AREA, lowcore, fpt_creg_save_area); + OFFSET(__LC_TOD_PROGREG_SAVE_AREA, lowcore, tod_progreg_save_area); + OFFSET(__LC_CPU_TIMER_SAVE_AREA, lowcore, cpu_timer_save_area); + OFFSET(__LC_CLOCK_COMP_SAVE_AREA, lowcore, clock_comp_save_area); + OFFSET(__LC_LAST_BREAK_SAVE_AREA, lowcore, last_break_save_area); + OFFSET(__LC_AREGS_SAVE_AREA, lowcore, access_regs_save_area); + OFFSET(__LC_CREGS_SAVE_AREA, lowcore, cregs_save_area); + OFFSET(__LC_PGM_TDB, lowcore, pgm_tdb); + BLANK(); + /* gmap/sie offsets */ + OFFSET(__GMAP_ASCE, gmap, asce); + OFFSET(__SIE_PROG0C, kvm_s390_sie_block, prog0c); + OFFSET(__SIE_PROG20, kvm_s390_sie_block, prog20); + /* kexec_sha_region */ + OFFSET(__KEXEC_SHA_REGION_START, kexec_sha_region, start); + OFFSET(__KEXEC_SHA_REGION_LEN, kexec_sha_region, len); + DEFINE(__KEXEC_SHA_REGION_SIZE, sizeof(struct kexec_sha_region)); + /* sizeof kernel parameter area */ + DEFINE(__PARMAREA_SIZE, sizeof(struct parmarea)); + /* kernel parameter area offsets */ + DEFINE(IPL_DEVICE, PARMAREA + offsetof(struct parmarea, ipl_device)); + DEFINE(INITRD_START, PARMAREA + offsetof(struct parmarea, initrd_start)); + DEFINE(INITRD_SIZE, PARMAREA + offsetof(struct parmarea, initrd_size)); + DEFINE(OLDMEM_BASE, PARMAREA + offsetof(struct parmarea, oldmem_base)); + DEFINE(OLDMEM_SIZE, PARMAREA + offsetof(struct parmarea, oldmem_size)); + DEFINE(COMMAND_LINE, PARMAREA + offsetof(struct parmarea, command_line)); + DEFINE(MAX_COMMAND_LINE_SIZE, PARMAREA + offsetof(struct parmarea, max_command_line_size)); +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + /* function graph return value tracing */ + OFFSET(__FGRAPH_RET_GPR2, fgraph_ret_regs, gpr2); + OFFSET(__FGRAPH_RET_FP, fgraph_ret_regs, fp); + DEFINE(__FGRAPH_RET_SIZE, sizeof(struct fgraph_ret_regs)); +#endif + OFFSET(__FTRACE_REGS_PT_REGS, ftrace_regs, regs); + DEFINE(__FTRACE_REGS_SIZE, sizeof(struct ftrace_regs)); + return 0; +} diff --git a/arch/s390/kernel/audit.c b/arch/s390/kernel/audit.c new file mode 100644 index 0000000000..02051a596b --- /dev/null +++ b/arch/s390/kernel/audit.c @@ -0,0 +1,81 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/init.h> +#include <linux/types.h> +#include <linux/audit.h> +#include <asm/unistd.h> +#include "audit.h" + +static unsigned dir_class[] = { +#include <asm-generic/audit_dir_write.h> +~0U +}; + +static unsigned read_class[] = { +#include <asm-generic/audit_read.h> +~0U +}; + +static unsigned write_class[] = { +#include <asm-generic/audit_write.h> +~0U +}; + +static unsigned chattr_class[] = { +#include <asm-generic/audit_change_attr.h> +~0U +}; + +static unsigned signal_class[] = { +#include <asm-generic/audit_signal.h> +~0U +}; + +int audit_classify_arch(int arch) +{ +#ifdef CONFIG_COMPAT + if (arch == AUDIT_ARCH_S390) + return 1; +#endif + return 0; +} + +int audit_classify_syscall(int abi, unsigned syscall) +{ +#ifdef CONFIG_COMPAT + if (abi == AUDIT_ARCH_S390) + return s390_classify_syscall(syscall); +#endif + switch(syscall) { + case __NR_open: + return AUDITSC_OPEN; + case __NR_openat: + return AUDITSC_OPENAT; + case __NR_socketcall: + return AUDITSC_SOCKETCALL; + case __NR_execve: + return AUDITSC_EXECVE; + case __NR_openat2: + return AUDITSC_OPENAT2; + default: + return AUDITSC_NATIVE; + } +} + +static int __init audit_classes_init(void) +{ +#ifdef CONFIG_COMPAT + audit_register_class(AUDIT_CLASS_WRITE_32, s390_write_class); + audit_register_class(AUDIT_CLASS_READ_32, s390_read_class); + audit_register_class(AUDIT_CLASS_DIR_WRITE_32, s390_dir_class); + audit_register_class(AUDIT_CLASS_CHATTR_32, s390_chattr_class); + audit_register_class(AUDIT_CLASS_SIGNAL_32, s390_signal_class); +#endif + audit_register_class(AUDIT_CLASS_WRITE, write_class); + audit_register_class(AUDIT_CLASS_READ, read_class); + audit_register_class(AUDIT_CLASS_DIR_WRITE, dir_class); + audit_register_class(AUDIT_CLASS_CHATTR, chattr_class); + audit_register_class(AUDIT_CLASS_SIGNAL, signal_class); + return 0; +} + +__initcall(audit_classes_init); diff --git a/arch/s390/kernel/audit.h b/arch/s390/kernel/audit.h new file mode 100644 index 0000000000..4d4b596412 --- /dev/null +++ b/arch/s390/kernel/audit.h @@ -0,0 +1,16 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __ARCH_S390_KERNEL_AUDIT_H +#define __ARCH_S390_KERNEL_AUDIT_H + +#include <linux/types.h> + +#ifdef CONFIG_COMPAT +extern int s390_classify_syscall(unsigned); +extern __u32 s390_dir_class[]; +extern __u32 s390_write_class[]; +extern __u32 s390_read_class[]; +extern __u32 s390_chattr_class[]; +extern __u32 s390_signal_class[]; +#endif /* CONFIG_COMPAT */ + +#endif /* __ARCH_S390_KERNEL_AUDIT_H */ diff --git a/arch/s390/kernel/cache.c b/arch/s390/kernel/cache.c new file mode 100644 index 0000000000..56254fa06f --- /dev/null +++ b/arch/s390/kernel/cache.c @@ -0,0 +1,170 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Extract CPU cache information and expose them via sysfs. + * + * Copyright IBM Corp. 2012 + */ + +#include <linux/seq_file.h> +#include <linux/cpu.h> +#include <linux/cacheinfo.h> +#include <asm/facility.h> + +enum { + CACHE_SCOPE_NOTEXISTS, + CACHE_SCOPE_PRIVATE, + CACHE_SCOPE_SHARED, + CACHE_SCOPE_RESERVED, +}; + +enum { + CTYPE_SEPARATE, + CTYPE_DATA, + CTYPE_INSTRUCTION, + CTYPE_UNIFIED, +}; + +enum { + EXTRACT_TOPOLOGY, + EXTRACT_LINE_SIZE, + EXTRACT_SIZE, + EXTRACT_ASSOCIATIVITY, +}; + +enum { + CACHE_TI_UNIFIED = 0, + CACHE_TI_DATA = 0, + CACHE_TI_INSTRUCTION, +}; + +struct cache_info { + unsigned char : 4; + unsigned char scope : 2; + unsigned char type : 2; +}; + +#define CACHE_MAX_LEVEL 8 +union cache_topology { + struct cache_info ci[CACHE_MAX_LEVEL]; + unsigned long raw; +}; + +static const char * const cache_type_string[] = { + "", + "Instruction", + "Data", + "", + "Unified", +}; + +static const enum cache_type cache_type_map[] = { + [CTYPE_SEPARATE] = CACHE_TYPE_SEPARATE, + [CTYPE_DATA] = CACHE_TYPE_DATA, + [CTYPE_INSTRUCTION] = CACHE_TYPE_INST, + [CTYPE_UNIFIED] = CACHE_TYPE_UNIFIED, +}; + +void show_cacheinfo(struct seq_file *m) +{ + struct cpu_cacheinfo *this_cpu_ci; + struct cacheinfo *cache; + int idx; + + this_cpu_ci = get_cpu_cacheinfo(cpumask_any(cpu_online_mask)); + for (idx = 0; idx < this_cpu_ci->num_leaves; idx++) { + cache = this_cpu_ci->info_list + idx; + seq_printf(m, "cache%-11d: ", idx); + seq_printf(m, "level=%d ", cache->level); + seq_printf(m, "type=%s ", cache_type_string[cache->type]); + seq_printf(m, "scope=%s ", + cache->disable_sysfs ? "Shared" : "Private"); + seq_printf(m, "size=%dK ", cache->size >> 10); + seq_printf(m, "line_size=%u ", cache->coherency_line_size); + seq_printf(m, "associativity=%d", cache->ways_of_associativity); + seq_puts(m, "\n"); + } +} + +static inline enum cache_type get_cache_type(struct cache_info *ci, int level) +{ + if (level >= CACHE_MAX_LEVEL) + return CACHE_TYPE_NOCACHE; + ci += level; + if (ci->scope != CACHE_SCOPE_SHARED && ci->scope != CACHE_SCOPE_PRIVATE) + return CACHE_TYPE_NOCACHE; + return cache_type_map[ci->type]; +} + +static inline unsigned long ecag(int ai, int li, int ti) +{ + return __ecag(ECAG_CACHE_ATTRIBUTE, ai << 4 | li << 1 | ti); +} + +static void ci_leaf_init(struct cacheinfo *this_leaf, int private, + enum cache_type type, unsigned int level, int cpu) +{ + int ti, num_sets; + + if (type == CACHE_TYPE_INST) + ti = CACHE_TI_INSTRUCTION; + else + ti = CACHE_TI_UNIFIED; + this_leaf->level = level + 1; + this_leaf->type = type; + this_leaf->coherency_line_size = ecag(EXTRACT_LINE_SIZE, level, ti); + this_leaf->ways_of_associativity = ecag(EXTRACT_ASSOCIATIVITY, level, ti); + this_leaf->size = ecag(EXTRACT_SIZE, level, ti); + num_sets = this_leaf->size / this_leaf->coherency_line_size; + num_sets /= this_leaf->ways_of_associativity; + this_leaf->number_of_sets = num_sets; + cpumask_set_cpu(cpu, &this_leaf->shared_cpu_map); + if (!private) + this_leaf->disable_sysfs = true; +} + +int init_cache_level(unsigned int cpu) +{ + struct cpu_cacheinfo *this_cpu_ci = get_cpu_cacheinfo(cpu); + unsigned int level = 0, leaves = 0; + union cache_topology ct; + enum cache_type ctype; + + if (!this_cpu_ci) + return -EINVAL; + ct.raw = ecag(EXTRACT_TOPOLOGY, 0, 0); + do { + ctype = get_cache_type(&ct.ci[0], level); + if (ctype == CACHE_TYPE_NOCACHE) + break; + /* Separate instruction and data caches */ + leaves += (ctype == CACHE_TYPE_SEPARATE) ? 2 : 1; + } while (++level < CACHE_MAX_LEVEL); + this_cpu_ci->num_levels = level; + this_cpu_ci->num_leaves = leaves; + return 0; +} + +int populate_cache_leaves(unsigned int cpu) +{ + struct cpu_cacheinfo *this_cpu_ci = get_cpu_cacheinfo(cpu); + struct cacheinfo *this_leaf = this_cpu_ci->info_list; + unsigned int level, idx, pvt; + union cache_topology ct; + enum cache_type ctype; + + ct.raw = ecag(EXTRACT_TOPOLOGY, 0, 0); + for (idx = 0, level = 0; level < this_cpu_ci->num_levels && + idx < this_cpu_ci->num_leaves; idx++, level++) { + if (!this_leaf) + return -EINVAL; + pvt = (ct.ci[level].scope == CACHE_SCOPE_PRIVATE) ? 1 : 0; + ctype = get_cache_type(&ct.ci[0], level); + if (ctype == CACHE_TYPE_SEPARATE) { + ci_leaf_init(this_leaf++, pvt, CACHE_TYPE_DATA, level, cpu); + ci_leaf_init(this_leaf++, pvt, CACHE_TYPE_INST, level, cpu); + } else { + ci_leaf_init(this_leaf++, pvt, ctype, level, cpu); + } + } + return 0; +} diff --git a/arch/s390/kernel/cert_store.c b/arch/s390/kernel/cert_store.c new file mode 100644 index 0000000000..554447768b --- /dev/null +++ b/arch/s390/kernel/cert_store.c @@ -0,0 +1,812 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * DIAG 0x320 support and certificate store handling + * + * Copyright IBM Corp. 2023 + * Author(s): Anastasia Eskova <anastasia.eskova@ibm.com> + */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#include <linux/delay.h> +#include <linux/device.h> +#include <linux/fs.h> +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/key-type.h> +#include <linux/key.h> +#include <linux/keyctl.h> +#include <linux/kobject.h> +#include <linux/module.h> +#include <linux/seq_file.h> +#include <linux/slab.h> +#include <linux/sysfs.h> +#include <crypto/sha2.h> +#include <keys/user-type.h> +#include <asm/debug.h> +#include <asm/diag.h> +#include <asm/ebcdic.h> +#include <asm/sclp.h> + +#define DIAG_MAX_RETRIES 10 + +#define VCE_FLAGS_VALID_MASK 0x80 + +#define ISM_LEN_DWORDS 4 +#define VCSSB_LEN_BYTES 128 +#define VCSSB_LEN_NO_CERTS 4 +#define VCB_LEN_NO_CERTS 64 +#define VC_NAME_LEN_BYTES 64 + +#define CERT_STORE_KEY_TYPE_NAME "cert_store_key" +#define CERT_STORE_KEYRING_NAME "cert_store" + +static debug_info_t *cert_store_dbf; +static debug_info_t *cert_store_hexdump; + +#define pr_dbf_msg(fmt, ...) \ + debug_sprintf_event(cert_store_dbf, 3, fmt "\n", ## __VA_ARGS__) + +enum diag320_subcode { + DIAG320_SUBCODES = 0, + DIAG320_STORAGE = 1, + DIAG320_CERT_BLOCK = 2, +}; + +enum diag320_rc { + DIAG320_RC_OK = 0x0001, + DIAG320_RC_CS_NOMATCH = 0x0306, +}; + +/* Verification Certificates Store Support Block (VCSSB). */ +struct vcssb { + u32 vcssb_length; + u8 pad_0x04[3]; + u8 version; + u8 pad_0x08[8]; + u32 cs_token; + u8 pad_0x14[12]; + u16 total_vc_index_count; + u16 max_vc_index_count; + u8 pad_0x24[28]; + u32 max_vce_length; + u32 max_vcxe_length; + u8 pad_0x48[8]; + u32 max_single_vcb_length; + u32 total_vcb_length; + u32 max_single_vcxb_length; + u32 total_vcxb_length; + u8 pad_0x60[32]; +} __packed __aligned(8); + +/* Verification Certificate Entry (VCE) Header. */ +struct vce_header { + u32 vce_length; + u8 flags; + u8 key_type; + u16 vc_index; + u8 vc_name[VC_NAME_LEN_BYTES]; /* EBCDIC */ + u8 vc_format; + u8 pad_0x49; + u16 key_id_length; + u8 pad_0x4c; + u8 vc_hash_type; + u16 vc_hash_length; + u8 pad_0x50[4]; + u32 vc_length; + u8 pad_0x58[8]; + u16 vc_hash_offset; + u16 vc_offset; + u8 pad_0x64[28]; +} __packed __aligned(4); + +/* Verification Certificate Block (VCB) Header. */ +struct vcb_header { + u32 vcb_input_length; + u8 pad_0x04[4]; + u16 first_vc_index; + u16 last_vc_index; + u32 pad_0x0c; + u32 cs_token; + u8 pad_0x14[12]; + u32 vcb_output_length; + u8 pad_0x24[3]; + u8 version; + u16 stored_vc_count; + u16 remaining_vc_count; + u8 pad_0x2c[20]; +} __packed __aligned(4); + +/* Verification Certificate Block (VCB). */ +struct vcb { + struct vcb_header vcb_hdr; + u8 vcb_buf[]; +} __packed __aligned(4); + +/* Verification Certificate Entry (VCE). */ +struct vce { + struct vce_header vce_hdr; + u8 cert_data_buf[]; +} __packed __aligned(4); + +static void cert_store_key_describe(const struct key *key, struct seq_file *m) +{ + char ascii[VC_NAME_LEN_BYTES + 1]; + + /* + * First 64 bytes of the key description is key name in EBCDIC CP 500. + * Convert it to ASCII for displaying in /proc/keys. + */ + strscpy(ascii, key->description, sizeof(ascii)); + EBCASC_500(ascii, VC_NAME_LEN_BYTES); + seq_puts(m, ascii); + + seq_puts(m, &key->description[VC_NAME_LEN_BYTES]); + if (key_is_positive(key)) + seq_printf(m, ": %u", key->datalen); +} + +/* + * Certificate store key type takes over properties of + * user key but cannot be updated. + */ +static struct key_type key_type_cert_store_key = { + .name = CERT_STORE_KEY_TYPE_NAME, + .preparse = user_preparse, + .free_preparse = user_free_preparse, + .instantiate = generic_key_instantiate, + .revoke = user_revoke, + .destroy = user_destroy, + .describe = cert_store_key_describe, + .read = user_read, +}; + +/* Logging functions. */ +static void pr_dbf_vcb(const struct vcb *b) +{ + pr_dbf_msg("VCB Header:"); + pr_dbf_msg("vcb_input_length: %d", b->vcb_hdr.vcb_input_length); + pr_dbf_msg("first_vc_index: %d", b->vcb_hdr.first_vc_index); + pr_dbf_msg("last_vc_index: %d", b->vcb_hdr.last_vc_index); + pr_dbf_msg("cs_token: %d", b->vcb_hdr.cs_token); + pr_dbf_msg("vcb_output_length: %d", b->vcb_hdr.vcb_output_length); + pr_dbf_msg("version: %d", b->vcb_hdr.version); + pr_dbf_msg("stored_vc_count: %d", b->vcb_hdr.stored_vc_count); + pr_dbf_msg("remaining_vc_count: %d", b->vcb_hdr.remaining_vc_count); +} + +static void pr_dbf_vce(const struct vce *e) +{ + unsigned char vc_name[VC_NAME_LEN_BYTES + 1]; + char log_string[VC_NAME_LEN_BYTES + 40]; + + pr_dbf_msg("VCE Header:"); + pr_dbf_msg("vce_hdr.vce_length: %d", e->vce_hdr.vce_length); + pr_dbf_msg("vce_hdr.flags: %d", e->vce_hdr.flags); + pr_dbf_msg("vce_hdr.key_type: %d", e->vce_hdr.key_type); + pr_dbf_msg("vce_hdr.vc_index: %d", e->vce_hdr.vc_index); + pr_dbf_msg("vce_hdr.vc_format: %d", e->vce_hdr.vc_format); + pr_dbf_msg("vce_hdr.key_id_length: %d", e->vce_hdr.key_id_length); + pr_dbf_msg("vce_hdr.vc_hash_type: %d", e->vce_hdr.vc_hash_type); + pr_dbf_msg("vce_hdr.vc_hash_length: %d", e->vce_hdr.vc_hash_length); + pr_dbf_msg("vce_hdr.vc_hash_offset: %d", e->vce_hdr.vc_hash_offset); + pr_dbf_msg("vce_hdr.vc_length: %d", e->vce_hdr.vc_length); + pr_dbf_msg("vce_hdr.vc_offset: %d", e->vce_hdr.vc_offset); + + /* Certificate name in ASCII. */ + memcpy(vc_name, e->vce_hdr.vc_name, VC_NAME_LEN_BYTES); + EBCASC_500(vc_name, VC_NAME_LEN_BYTES); + vc_name[VC_NAME_LEN_BYTES] = '\0'; + + snprintf(log_string, sizeof(log_string), + "index: %d vce_hdr.vc_name (ASCII): %s", + e->vce_hdr.vc_index, vc_name); + debug_text_event(cert_store_hexdump, 3, log_string); + + /* Certificate data. */ + debug_text_event(cert_store_hexdump, 3, "VCE: Certificate data start"); + debug_event(cert_store_hexdump, 3, (u8 *)e->cert_data_buf, 128); + debug_text_event(cert_store_hexdump, 3, "VCE: Certificate data end"); + debug_event(cert_store_hexdump, 3, + (u8 *)e->cert_data_buf + e->vce_hdr.vce_length - 128, 128); +} + +static void pr_dbf_vcssb(const struct vcssb *s) +{ + debug_text_event(cert_store_hexdump, 3, "DIAG320 Subcode1"); + debug_event(cert_store_hexdump, 3, (u8 *)s, VCSSB_LEN_BYTES); + + pr_dbf_msg("VCSSB:"); + pr_dbf_msg("vcssb_length: %u", s->vcssb_length); + pr_dbf_msg("version: %u", s->version); + pr_dbf_msg("cs_token: %u", s->cs_token); + pr_dbf_msg("total_vc_index_count: %u", s->total_vc_index_count); + pr_dbf_msg("max_vc_index_count: %u", s->max_vc_index_count); + pr_dbf_msg("max_vce_length: %u", s->max_vce_length); + pr_dbf_msg("max_vcxe_length: %u", s->max_vce_length); + pr_dbf_msg("max_single_vcb_length: %u", s->max_single_vcb_length); + pr_dbf_msg("total_vcb_length: %u", s->total_vcb_length); + pr_dbf_msg("max_single_vcxb_length: %u", s->max_single_vcxb_length); + pr_dbf_msg("total_vcxb_length: %u", s->total_vcxb_length); +} + +static int __diag320(unsigned long subcode, void *addr) +{ + union register_pair rp = { .even = (unsigned long)addr, }; + + asm volatile( + " diag %[rp],%[subcode],0x320\n" + "0: nopr %%r7\n" + EX_TABLE(0b, 0b) + : [rp] "+d" (rp.pair) + : [subcode] "d" (subcode) + : "cc", "memory"); + + return rp.odd; +} + +static int diag320(unsigned long subcode, void *addr) +{ + diag_stat_inc(DIAG_STAT_X320); + + return __diag320(subcode, addr); +} + +/* + * Calculate SHA256 hash of the VCE certificate and compare it to hash stored in + * VCE. Return -EINVAL if hashes don't match. + */ +static int check_certificate_hash(const struct vce *vce) +{ + u8 hash[SHA256_DIGEST_SIZE]; + u16 vc_hash_length; + u8 *vce_hash; + + vce_hash = (u8 *)vce + vce->vce_hdr.vc_hash_offset; + vc_hash_length = vce->vce_hdr.vc_hash_length; + sha256((u8 *)vce + vce->vce_hdr.vc_offset, vce->vce_hdr.vc_length, hash); + if (memcmp(vce_hash, hash, vc_hash_length) == 0) + return 0; + + pr_dbf_msg("SHA256 hash of received certificate does not match"); + debug_text_event(cert_store_hexdump, 3, "VCE hash:"); + debug_event(cert_store_hexdump, 3, vce_hash, SHA256_DIGEST_SIZE); + debug_text_event(cert_store_hexdump, 3, "Calculated hash:"); + debug_event(cert_store_hexdump, 3, hash, SHA256_DIGEST_SIZE); + + return -EINVAL; +} + +static int check_certificate_valid(const struct vce *vce) +{ + if (!(vce->vce_hdr.flags & VCE_FLAGS_VALID_MASK)) { + pr_dbf_msg("Certificate entry is invalid"); + return -EINVAL; + } + if (vce->vce_hdr.vc_format != 1) { + pr_dbf_msg("Certificate format is not supported"); + return -EINVAL; + } + if (vce->vce_hdr.vc_hash_type != 1) { + pr_dbf_msg("Hash type is not supported"); + return -EINVAL; + } + + return check_certificate_hash(vce); +} + +static struct key *get_user_session_keyring(void) +{ + key_ref_t us_keyring_ref; + + us_keyring_ref = lookup_user_key(KEY_SPEC_USER_SESSION_KEYRING, + KEY_LOOKUP_CREATE, KEY_NEED_LINK); + if (IS_ERR(us_keyring_ref)) { + pr_dbf_msg("Couldn't get user session keyring: %ld", + PTR_ERR(us_keyring_ref)); + return ERR_PTR(-ENOKEY); + } + key_ref_put(us_keyring_ref); + return key_ref_to_ptr(us_keyring_ref); +} + +/* Invalidate all keys from cert_store keyring. */ +static int invalidate_keyring_keys(struct key *keyring) +{ + unsigned long num_keys, key_index; + size_t keyring_payload_len; + key_serial_t *key_array; + struct key *current_key; + int rc; + + keyring_payload_len = key_type_keyring.read(keyring, NULL, 0); + num_keys = keyring_payload_len / sizeof(key_serial_t); + key_array = kcalloc(num_keys, sizeof(key_serial_t), GFP_KERNEL); + if (!key_array) + return -ENOMEM; + + rc = key_type_keyring.read(keyring, (char *)key_array, keyring_payload_len); + if (rc != keyring_payload_len) { + pr_dbf_msg("Couldn't read keyring payload"); + goto out; + } + + for (key_index = 0; key_index < num_keys; key_index++) { + current_key = key_lookup(key_array[key_index]); + pr_dbf_msg("Invalidating key %08x", current_key->serial); + + key_invalidate(current_key); + key_put(current_key); + rc = key_unlink(keyring, current_key); + if (rc) { + pr_dbf_msg("Couldn't unlink key %08x: %d", current_key->serial, rc); + break; + } + } +out: + kfree(key_array); + return rc; +} + +static struct key *find_cs_keyring(void) +{ + key_ref_t cs_keyring_ref; + struct key *cs_keyring; + + cs_keyring_ref = keyring_search(make_key_ref(get_user_session_keyring(), true), + &key_type_keyring, CERT_STORE_KEYRING_NAME, + false); + if (!IS_ERR(cs_keyring_ref)) { + cs_keyring = key_ref_to_ptr(cs_keyring_ref); + key_ref_put(cs_keyring_ref); + goto found; + } + /* Search default locations: thread, process, session keyrings */ + cs_keyring = request_key(&key_type_keyring, CERT_STORE_KEYRING_NAME, NULL); + if (IS_ERR(cs_keyring)) + return NULL; + key_put(cs_keyring); +found: + return cs_keyring; +} + +static void cleanup_cs_keys(void) +{ + struct key *cs_keyring; + + cs_keyring = find_cs_keyring(); + if (!cs_keyring) + return; + + pr_dbf_msg("Found cert_store keyring. Purging..."); + /* + * Remove cert_store_key_type in case invalidation + * of old cert_store keys failed (= severe error). + */ + if (invalidate_keyring_keys(cs_keyring)) + unregister_key_type(&key_type_cert_store_key); + + keyring_clear(cs_keyring); + key_invalidate(cs_keyring); + key_put(cs_keyring); + key_unlink(get_user_session_keyring(), cs_keyring); +} + +static struct key *create_cs_keyring(void) +{ + static struct key *cs_keyring; + + /* Cleanup previous cs_keyring and all associated keys if any. */ + cleanup_cs_keys(); + cs_keyring = keyring_alloc(CERT_STORE_KEYRING_NAME, GLOBAL_ROOT_UID, + GLOBAL_ROOT_GID, current_cred(), + (KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ, + KEY_ALLOC_NOT_IN_QUOTA | KEY_ALLOC_SET_KEEP, + NULL, get_user_session_keyring()); + if (IS_ERR(cs_keyring)) { + pr_dbf_msg("Can't allocate cert_store keyring"); + return NULL; + } + + pr_dbf_msg("Successfully allocated cert_store keyring: %08x", cs_keyring->serial); + + /* + * In case a previous clean-up ran into an + * error and unregistered key type. + */ + register_key_type(&key_type_cert_store_key); + + return cs_keyring; +} + +/* + * Allocate memory and create key description in format + * [key name in EBCDIC]:[VCE index]:[CS token]. + * Return a pointer to key description or NULL if memory + * allocation failed. Memory should be freed by caller. + */ +static char *get_key_description(struct vcssb *vcssb, const struct vce *vce) +{ + size_t len, name_len; + u32 cs_token; + char *desc; + + cs_token = vcssb->cs_token; + /* Description string contains "%64s:%05u:%010u\0". */ + name_len = sizeof(vce->vce_hdr.vc_name); + len = name_len + 1 + 5 + 1 + 10 + 1; + desc = kmalloc(len, GFP_KERNEL); + if (!desc) + return NULL; + + memcpy(desc, vce->vce_hdr.vc_name, name_len); + snprintf(desc + name_len, len - name_len, ":%05u:%010u", + vce->vce_hdr.vc_index, cs_token); + + return desc; +} + +/* + * Create a key of type "cert_store_key" using the data from VCE for key + * payload and key description. Link the key to "cert_store" keyring. + */ +static int create_key_from_vce(struct vcssb *vcssb, struct vce *vce, + struct key *keyring) +{ + key_ref_t newkey; + char *desc; + int rc; + + desc = get_key_description(vcssb, vce); + if (!desc) + return -ENOMEM; + + newkey = key_create_or_update( + make_key_ref(keyring, true), CERT_STORE_KEY_TYPE_NAME, + desc, (u8 *)vce + vce->vce_hdr.vc_offset, + vce->vce_hdr.vc_length, + (KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ, + KEY_ALLOC_NOT_IN_QUOTA); + + rc = PTR_ERR_OR_ZERO(newkey); + if (rc) { + pr_dbf_msg("Couldn't create a key from Certificate Entry (%d)", rc); + rc = -ENOKEY; + goto out; + } + + key_ref_put(newkey); +out: + kfree(desc); + return rc; +} + +/* Get Verification Certificate Storage Size block with DIAG320 subcode2. */ +static int get_vcssb(struct vcssb *vcssb) +{ + int diag320_rc; + + memset(vcssb, 0, sizeof(*vcssb)); + vcssb->vcssb_length = VCSSB_LEN_BYTES; + diag320_rc = diag320(DIAG320_STORAGE, vcssb); + pr_dbf_vcssb(vcssb); + + if (diag320_rc != DIAG320_RC_OK) { + pr_dbf_msg("Diag 320 Subcode 1 returned bad RC: %04x", diag320_rc); + return -EIO; + } + if (vcssb->vcssb_length == VCSSB_LEN_NO_CERTS) { + pr_dbf_msg("No certificates available for current configuration"); + return -ENOKEY; + } + + return 0; +} + +static u32 get_4k_mult_vcb_size(struct vcssb *vcssb) +{ + return round_up(vcssb->max_single_vcb_length, PAGE_SIZE); +} + +/* Fill input fields of single-entry VCB that will be read by LPAR. */ +static void fill_vcb_input(struct vcssb *vcssb, struct vcb *vcb, u16 index) +{ + memset(vcb, 0, sizeof(*vcb)); + vcb->vcb_hdr.vcb_input_length = get_4k_mult_vcb_size(vcssb); + vcb->vcb_hdr.cs_token = vcssb->cs_token; + + /* Request single entry. */ + vcb->vcb_hdr.first_vc_index = index; + vcb->vcb_hdr.last_vc_index = index; +} + +static void extract_vce_from_sevcb(struct vcb *vcb, struct vce *vce) +{ + struct vce *extracted_vce; + + extracted_vce = (struct vce *)vcb->vcb_buf; + memcpy(vce, vcb->vcb_buf, extracted_vce->vce_hdr.vce_length); + pr_dbf_vce(vce); +} + +static int get_sevcb(struct vcssb *vcssb, u16 index, struct vcb *vcb) +{ + int rc, diag320_rc; + + fill_vcb_input(vcssb, vcb, index); + + diag320_rc = diag320(DIAG320_CERT_BLOCK, vcb); + pr_dbf_msg("Diag 320 Subcode2 RC %2x", diag320_rc); + pr_dbf_vcb(vcb); + + switch (diag320_rc) { + case DIAG320_RC_OK: + rc = 0; + if (vcb->vcb_hdr.vcb_output_length == VCB_LEN_NO_CERTS) { + pr_dbf_msg("No certificate entry for index %u", index); + rc = -ENOKEY; + } else if (vcb->vcb_hdr.remaining_vc_count != 0) { + /* Retry on insufficient space. */ + pr_dbf_msg("Couldn't get all requested certificates"); + rc = -EAGAIN; + } + break; + case DIAG320_RC_CS_NOMATCH: + pr_dbf_msg("Certificate Store token mismatch"); + rc = -EAGAIN; + break; + default: + pr_dbf_msg("Diag 320 Subcode2 returned bad rc (0x%4x)", diag320_rc); + rc = -EINVAL; + break; + } + + return rc; +} + +/* + * Allocate memory for single-entry VCB, get VCB via DIAG320 subcode 2 call, + * extract VCE and create a key from its' certificate. + */ +static int create_key_from_sevcb(struct vcssb *vcssb, u16 index, + struct key *keyring) +{ + struct vcb *vcb; + struct vce *vce; + int rc; + + rc = -ENOMEM; + vcb = vmalloc(get_4k_mult_vcb_size(vcssb)); + vce = vmalloc(vcssb->max_single_vcb_length - sizeof(vcb->vcb_hdr)); + if (!vcb || !vce) + goto out; + + rc = get_sevcb(vcssb, index, vcb); + if (rc) + goto out; + + extract_vce_from_sevcb(vcb, vce); + rc = check_certificate_valid(vce); + if (rc) + goto out; + + rc = create_key_from_vce(vcssb, vce, keyring); + if (rc) + goto out; + + pr_dbf_msg("Successfully created key from Certificate Entry %d", index); +out: + vfree(vce); + vfree(vcb); + return rc; +} + +/* + * Request a single-entry VCB for each VCE available for the partition. + * Create a key from it and link it to cert_store keyring. If no keys + * could be created (i.e. VCEs were invalid) return -ENOKEY. + */ +static int add_certificates_to_keyring(struct vcssb *vcssb, struct key *keyring) +{ + int rc, index, count, added; + + count = 0; + added = 0; + /* Certificate Store entries indices start with 1 and have no gaps. */ + for (index = 1; index < vcssb->total_vc_index_count + 1; index++) { + pr_dbf_msg("Creating key from VCE %u", index); + rc = create_key_from_sevcb(vcssb, index, keyring); + count++; + + if (rc == -EAGAIN) + return rc; + + if (rc) + pr_dbf_msg("Creating key from VCE %u failed (%d)", index, rc); + else + added++; + } + + if (added == 0) { + pr_dbf_msg("Processed %d entries. No keys created", count); + return -ENOKEY; + } + + pr_info("Added %d of %d keys to cert_store keyring", added, count); + + /* + * Do not allow to link more keys to certificate store keyring after all + * the VCEs were processed. + */ + rc = keyring_restrict(make_key_ref(keyring, true), NULL, NULL); + if (rc) + pr_dbf_msg("Failed to set restriction to cert_store keyring (%d)", rc); + + return 0; +} + +/* + * Check which DIAG320 subcodes are installed. + * Return -ENOENT if subcodes 1 or 2 are not available. + */ +static int query_diag320_subcodes(void) +{ + unsigned long ism[ISM_LEN_DWORDS]; + int rc; + + rc = diag320(0, ism); + if (rc != DIAG320_RC_OK) { + pr_dbf_msg("DIAG320 subcode query returned %04x", rc); + return -ENOENT; + } + + debug_text_event(cert_store_hexdump, 3, "DIAG320 Subcode 0"); + debug_event(cert_store_hexdump, 3, ism, sizeof(ism)); + + if (!test_bit_inv(1, ism) || !test_bit_inv(2, ism)) { + pr_dbf_msg("Not all required DIAG320 subcodes are installed"); + return -ENOENT; + } + + return 0; +} + +/* + * Check if Certificate Store is supported by the firmware and DIAG320 subcodes + * 1 and 2 are installed. Create cert_store keyring and link all certificates + * available for the current partition to it as "cert_store_key" type + * keys. On refresh or error invalidate cert_store keyring and destroy + * all keys of "cert_store_key" type. + */ +static int fill_cs_keyring(void) +{ + struct key *cs_keyring; + struct vcssb *vcssb; + int rc; + + rc = -ENOMEM; + vcssb = kmalloc(VCSSB_LEN_BYTES, GFP_KERNEL); + if (!vcssb) + goto cleanup_keys; + + rc = -ENOENT; + if (!sclp.has_diag320) { + pr_dbf_msg("Certificate Store is not supported"); + goto cleanup_keys; + } + + rc = query_diag320_subcodes(); + if (rc) + goto cleanup_keys; + + rc = get_vcssb(vcssb); + if (rc) + goto cleanup_keys; + + rc = -ENOMEM; + cs_keyring = create_cs_keyring(); + if (!cs_keyring) + goto cleanup_keys; + + rc = add_certificates_to_keyring(vcssb, cs_keyring); + if (rc) + goto cleanup_cs_keyring; + + goto out; + +cleanup_cs_keyring: + key_put(cs_keyring); +cleanup_keys: + cleanup_cs_keys(); +out: + kfree(vcssb); + return rc; +} + +static DEFINE_MUTEX(cs_refresh_lock); +static int cs_status_val = -1; + +static ssize_t cs_status_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + if (cs_status_val == -1) + return sysfs_emit(buf, "uninitialized\n"); + else if (cs_status_val == 0) + return sysfs_emit(buf, "ok\n"); + + return sysfs_emit(buf, "failed (%d)\n", cs_status_val); +} + +static struct kobj_attribute cs_status_attr = __ATTR_RO(cs_status); + +static ssize_t refresh_store(struct kobject *kobj, struct kobj_attribute *attr, + const char *buf, size_t count) +{ + int rc, retries; + + pr_dbf_msg("Refresh certificate store information requested"); + rc = mutex_lock_interruptible(&cs_refresh_lock); + if (rc) + return rc; + + for (retries = 0; retries < DIAG_MAX_RETRIES; retries++) { + /* Request certificates from certificate store. */ + rc = fill_cs_keyring(); + if (rc) + pr_dbf_msg("Failed to refresh certificate store information (%d)", rc); + if (rc != -EAGAIN) + break; + } + cs_status_val = rc; + mutex_unlock(&cs_refresh_lock); + + return rc ?: count; +} + +static struct kobj_attribute refresh_attr = __ATTR_WO(refresh); + +static const struct attribute *cert_store_attrs[] __initconst = { + &cs_status_attr.attr, + &refresh_attr.attr, + NULL, +}; + +static struct kobject *cert_store_kobj; + +static int __init cert_store_init(void) +{ + int rc = -ENOMEM; + + cert_store_dbf = debug_register("cert_store_msg", 10, 1, 64); + if (!cert_store_dbf) + goto cleanup_dbf; + + cert_store_hexdump = debug_register("cert_store_hexdump", 3, 1, 128); + if (!cert_store_hexdump) + goto cleanup_dbf; + + debug_register_view(cert_store_hexdump, &debug_hex_ascii_view); + debug_register_view(cert_store_dbf, &debug_sprintf_view); + + /* Create directory /sys/firmware/cert_store. */ + cert_store_kobj = kobject_create_and_add("cert_store", firmware_kobj); + if (!cert_store_kobj) + goto cleanup_dbf; + + rc = sysfs_create_files(cert_store_kobj, cert_store_attrs); + if (rc) + goto cleanup_kobj; + + register_key_type(&key_type_cert_store_key); + + return rc; + +cleanup_kobj: + kobject_put(cert_store_kobj); +cleanup_dbf: + debug_unregister(cert_store_dbf); + debug_unregister(cert_store_hexdump); + + return rc; +} +device_initcall(cert_store_init); diff --git a/arch/s390/kernel/compat_audit.c b/arch/s390/kernel/compat_audit.c new file mode 100644 index 0000000000..a7c46e8310 --- /dev/null +++ b/arch/s390/kernel/compat_audit.c @@ -0,0 +1,48 @@ +// SPDX-License-Identifier: GPL-2.0 +#undef __s390x__ +#include <linux/audit_arch.h> +#include <asm/unistd.h> +#include "audit.h" + +unsigned s390_dir_class[] = { +#include <asm-generic/audit_dir_write.h> +~0U +}; + +unsigned s390_chattr_class[] = { +#include <asm-generic/audit_change_attr.h> +~0U +}; + +unsigned s390_write_class[] = { +#include <asm-generic/audit_write.h> +~0U +}; + +unsigned s390_read_class[] = { +#include <asm-generic/audit_read.h> +~0U +}; + +unsigned s390_signal_class[] = { +#include <asm-generic/audit_signal.h> +~0U +}; + +int s390_classify_syscall(unsigned syscall) +{ + switch(syscall) { + case __NR_open: + return AUDITSC_OPEN; + case __NR_openat: + return AUDITSC_OPENAT; + case __NR_socketcall: + return AUDITSC_SOCKETCALL; + case __NR_execve: + return AUDITSC_EXECVE; + case __NR_openat2: + return AUDITSC_OPENAT2; + default: + return AUDITSC_COMPAT; + } +} diff --git a/arch/s390/kernel/compat_linux.c b/arch/s390/kernel/compat_linux.c new file mode 100644 index 0000000000..f9d418d1b6 --- /dev/null +++ b/arch/s390/kernel/compat_linux.c @@ -0,0 +1,289 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * S390 version + * Copyright IBM Corp. 2000 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), + * Gerhard Tonn (ton@de.ibm.com) + * Thomas Spatzier (tspat@de.ibm.com) + * + * Conversion between 31bit and 64bit native syscalls. + * + * Heavily inspired by the 32-bit Sparc compat code which is + * Copyright (C) 1997,1998 Jakub Jelinek (jj@sunsite.mff.cuni.cz) + * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu) + * + */ + + +#include <linux/kernel.h> +#include <linux/sched.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/file.h> +#include <linux/signal.h> +#include <linux/resource.h> +#include <linux/times.h> +#include <linux/smp.h> +#include <linux/sem.h> +#include <linux/msg.h> +#include <linux/shm.h> +#include <linux/uio.h> +#include <linux/quota.h> +#include <linux/poll.h> +#include <linux/personality.h> +#include <linux/stat.h> +#include <linux/filter.h> +#include <linux/highmem.h> +#include <linux/mman.h> +#include <linux/ipv6.h> +#include <linux/in.h> +#include <linux/icmpv6.h> +#include <linux/syscalls.h> +#include <linux/sysctl.h> +#include <linux/binfmts.h> +#include <linux/capability.h> +#include <linux/compat.h> +#include <linux/vfs.h> +#include <linux/ptrace.h> +#include <linux/fadvise.h> +#include <linux/ipc.h> +#include <linux/slab.h> + +#include <asm/types.h> +#include <linux/uaccess.h> + +#include <net/scm.h> +#include <net/sock.h> + +#include "compat_linux.h" + +#ifdef CONFIG_SYSVIPC +COMPAT_SYSCALL_DEFINE5(s390_ipc, uint, call, int, first, compat_ulong_t, second, + compat_ulong_t, third, compat_uptr_t, ptr) +{ + if (call >> 16) /* hack for backward compatibility */ + return -EINVAL; + return compat_ksys_ipc(call, first, second, third, ptr, third); +} +#endif + +COMPAT_SYSCALL_DEFINE3(s390_truncate64, const char __user *, path, u32, high, u32, low) +{ + return ksys_truncate(path, (unsigned long)high << 32 | low); +} + +COMPAT_SYSCALL_DEFINE3(s390_ftruncate64, unsigned int, fd, u32, high, u32, low) +{ + return ksys_ftruncate(fd, (unsigned long)high << 32 | low); +} + +COMPAT_SYSCALL_DEFINE5(s390_pread64, unsigned int, fd, char __user *, ubuf, + compat_size_t, count, u32, high, u32, low) +{ + if ((compat_ssize_t) count < 0) + return -EINVAL; + return ksys_pread64(fd, ubuf, count, (unsigned long)high << 32 | low); +} + +COMPAT_SYSCALL_DEFINE5(s390_pwrite64, unsigned int, fd, const char __user *, ubuf, + compat_size_t, count, u32, high, u32, low) +{ + if ((compat_ssize_t) count < 0) + return -EINVAL; + return ksys_pwrite64(fd, ubuf, count, (unsigned long)high << 32 | low); +} + +COMPAT_SYSCALL_DEFINE4(s390_readahead, int, fd, u32, high, u32, low, s32, count) +{ + return ksys_readahead(fd, (unsigned long)high << 32 | low, count); +} + +struct stat64_emu31 { + unsigned long long st_dev; + unsigned int __pad1; +#define STAT64_HAS_BROKEN_ST_INO 1 + u32 __st_ino; + unsigned int st_mode; + unsigned int st_nlink; + u32 st_uid; + u32 st_gid; + unsigned long long st_rdev; + unsigned int __pad3; + long st_size; + u32 st_blksize; + unsigned char __pad4[4]; + u32 __pad5; /* future possible st_blocks high bits */ + u32 st_blocks; /* Number 512-byte blocks allocated. */ + u32 st_atime; + u32 __pad6; + u32 st_mtime; + u32 __pad7; + u32 st_ctime; + u32 __pad8; /* will be high 32 bits of ctime someday */ + unsigned long st_ino; +}; + +static int cp_stat64(struct stat64_emu31 __user *ubuf, struct kstat *stat) +{ + struct stat64_emu31 tmp; + + memset(&tmp, 0, sizeof(tmp)); + + tmp.st_dev = huge_encode_dev(stat->dev); + tmp.st_ino = stat->ino; + tmp.__st_ino = (u32)stat->ino; + tmp.st_mode = stat->mode; + tmp.st_nlink = (unsigned int)stat->nlink; + tmp.st_uid = from_kuid_munged(current_user_ns(), stat->uid); + tmp.st_gid = from_kgid_munged(current_user_ns(), stat->gid); + tmp.st_rdev = huge_encode_dev(stat->rdev); + tmp.st_size = stat->size; + tmp.st_blksize = (u32)stat->blksize; + tmp.st_blocks = (u32)stat->blocks; + tmp.st_atime = (u32)stat->atime.tv_sec; + tmp.st_mtime = (u32)stat->mtime.tv_sec; + tmp.st_ctime = (u32)stat->ctime.tv_sec; + + return copy_to_user(ubuf,&tmp,sizeof(tmp)) ? -EFAULT : 0; +} + +COMPAT_SYSCALL_DEFINE2(s390_stat64, const char __user *, filename, struct stat64_emu31 __user *, statbuf) +{ + struct kstat stat; + int ret = vfs_stat(filename, &stat); + if (!ret) + ret = cp_stat64(statbuf, &stat); + return ret; +} + +COMPAT_SYSCALL_DEFINE2(s390_lstat64, const char __user *, filename, struct stat64_emu31 __user *, statbuf) +{ + struct kstat stat; + int ret = vfs_lstat(filename, &stat); + if (!ret) + ret = cp_stat64(statbuf, &stat); + return ret; +} + +COMPAT_SYSCALL_DEFINE2(s390_fstat64, unsigned int, fd, struct stat64_emu31 __user *, statbuf) +{ + struct kstat stat; + int ret = vfs_fstat(fd, &stat); + if (!ret) + ret = cp_stat64(statbuf, &stat); + return ret; +} + +COMPAT_SYSCALL_DEFINE4(s390_fstatat64, unsigned int, dfd, const char __user *, filename, + struct stat64_emu31 __user *, statbuf, int, flag) +{ + struct kstat stat; + int error; + + error = vfs_fstatat(dfd, filename, &stat, flag); + if (error) + return error; + return cp_stat64(statbuf, &stat); +} + +/* + * Linux/i386 didn't use to be able to handle more than + * 4 system call parameters, so these system calls used a memory + * block for parameter passing.. + */ + +struct mmap_arg_struct_emu31 { + compat_ulong_t addr; + compat_ulong_t len; + compat_ulong_t prot; + compat_ulong_t flags; + compat_ulong_t fd; + compat_ulong_t offset; +}; + +COMPAT_SYSCALL_DEFINE1(s390_old_mmap, struct mmap_arg_struct_emu31 __user *, arg) +{ + struct mmap_arg_struct_emu31 a; + + if (copy_from_user(&a, arg, sizeof(a))) + return -EFAULT; + if (a.offset & ~PAGE_MASK) + return -EINVAL; + return ksys_mmap_pgoff(a.addr, a.len, a.prot, a.flags, a.fd, + a.offset >> PAGE_SHIFT); +} + +COMPAT_SYSCALL_DEFINE1(s390_mmap2, struct mmap_arg_struct_emu31 __user *, arg) +{ + struct mmap_arg_struct_emu31 a; + + if (copy_from_user(&a, arg, sizeof(a))) + return -EFAULT; + return ksys_mmap_pgoff(a.addr, a.len, a.prot, a.flags, a.fd, a.offset); +} + +COMPAT_SYSCALL_DEFINE3(s390_read, unsigned int, fd, char __user *, buf, compat_size_t, count) +{ + if ((compat_ssize_t) count < 0) + return -EINVAL; + + return ksys_read(fd, buf, count); +} + +COMPAT_SYSCALL_DEFINE3(s390_write, unsigned int, fd, const char __user *, buf, compat_size_t, count) +{ + if ((compat_ssize_t) count < 0) + return -EINVAL; + + return ksys_write(fd, buf, count); +} + +/* + * 31 bit emulation wrapper functions for sys_fadvise64/fadvise64_64. + * These need to rewrite the advise values for POSIX_FADV_{DONTNEED,NOREUSE} + * because the 31 bit values differ from the 64 bit values. + */ + +COMPAT_SYSCALL_DEFINE5(s390_fadvise64, int, fd, u32, high, u32, low, compat_size_t, len, int, advise) +{ + if (advise == 4) + advise = POSIX_FADV_DONTNEED; + else if (advise == 5) + advise = POSIX_FADV_NOREUSE; + return ksys_fadvise64_64(fd, (unsigned long)high << 32 | low, len, + advise); +} + +struct fadvise64_64_args { + int fd; + long long offset; + long long len; + int advice; +}; + +COMPAT_SYSCALL_DEFINE1(s390_fadvise64_64, struct fadvise64_64_args __user *, args) +{ + struct fadvise64_64_args a; + + if ( copy_from_user(&a, args, sizeof(a)) ) + return -EFAULT; + if (a.advice == 4) + a.advice = POSIX_FADV_DONTNEED; + else if (a.advice == 5) + a.advice = POSIX_FADV_NOREUSE; + return ksys_fadvise64_64(a.fd, a.offset, a.len, a.advice); +} + +COMPAT_SYSCALL_DEFINE6(s390_sync_file_range, int, fd, u32, offhigh, u32, offlow, + u32, nhigh, u32, nlow, unsigned int, flags) +{ + return ksys_sync_file_range(fd, ((loff_t)offhigh << 32) + offlow, + ((u64)nhigh << 32) + nlow, flags); +} + +COMPAT_SYSCALL_DEFINE6(s390_fallocate, int, fd, int, mode, u32, offhigh, u32, offlow, + u32, lenhigh, u32, lenlow) +{ + return ksys_fallocate(fd, mode, ((loff_t)offhigh << 32) + offlow, + ((u64)lenhigh << 32) + lenlow); +} diff --git a/arch/s390/kernel/compat_linux.h b/arch/s390/kernel/compat_linux.h new file mode 100644 index 0000000000..ef23739b27 --- /dev/null +++ b/arch/s390/kernel/compat_linux.h @@ -0,0 +1,101 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _ASM_S390X_S390_H +#define _ASM_S390X_S390_H + +#include <linux/compat.h> +#include <linux/socket.h> +#include <linux/syscalls.h> +#include <asm/ptrace.h> + +/* + * Macro that masks the high order bit of a 32 bit pointer and + * converts it to a 64 bit pointer. + */ +#define A(__x) ((unsigned long)((__x) & 0x7FFFFFFFUL)) +#define AA(__x) ((unsigned long)(__x)) + +/* Now 32bit compatibility types */ +struct ipc_kludge_32 { + __u32 msgp; /* pointer */ + __s32 msgtyp; +}; + +/* asm/sigcontext.h */ +typedef union { + __u64 d; + __u32 f; +} freg_t32; + +typedef struct { + unsigned int fpc; + unsigned int pad; + freg_t32 fprs[__NUM_FPRS]; +} _s390_fp_regs32; + +typedef struct { + psw_t32 psw; + __u32 gprs[__NUM_GPRS]; + __u32 acrs[__NUM_ACRS]; +} _s390_regs_common32; + +typedef struct { + _s390_regs_common32 regs; + _s390_fp_regs32 fpregs; +} _sigregs32; + +typedef struct { + __u32 gprs_high[__NUM_GPRS]; + __u64 vxrs_low[__NUM_VXRS_LOW]; + __vector128 vxrs_high[__NUM_VXRS_HIGH]; + __u8 __reserved[128]; +} _sigregs_ext32; + +#define _SIGCONTEXT_NSIG32 64 +#define _SIGCONTEXT_NSIG_BPW32 32 +#define __SIGNAL_FRAMESIZE32 96 +#define _SIGMASK_COPY_SIZE32 (sizeof(u32) * 2) + +struct sigcontext32 { + __u32 oldmask[_COMPAT_NSIG_WORDS]; + __u32 sregs; /* pointer */ +}; + +/* asm/signal.h */ + +/* asm/ucontext.h */ +struct ucontext32 { + __u32 uc_flags; + __u32 uc_link; /* pointer */ + compat_stack_t uc_stack; + _sigregs32 uc_mcontext; + compat_sigset_t uc_sigmask; + /* Allow for uc_sigmask growth. Glibc uses a 1024-bit sigset_t. */ + unsigned char __unused[128 - sizeof(compat_sigset_t)]; + _sigregs_ext32 uc_mcontext_ext; +}; + +struct stat64_emu31; +struct mmap_arg_struct_emu31; +struct fadvise64_64_args; + +long compat_sys_s390_truncate64(const char __user *path, u32 high, u32 low); +long compat_sys_s390_ftruncate64(unsigned int fd, u32 high, u32 low); +long compat_sys_s390_pread64(unsigned int fd, char __user *ubuf, compat_size_t count, u32 high, u32 low); +long compat_sys_s390_pwrite64(unsigned int fd, const char __user *ubuf, compat_size_t count, u32 high, u32 low); +long compat_sys_s390_readahead(int fd, u32 high, u32 low, s32 count); +long compat_sys_s390_stat64(const char __user *filename, struct stat64_emu31 __user *statbuf); +long compat_sys_s390_lstat64(const char __user *filename, struct stat64_emu31 __user *statbuf); +long compat_sys_s390_fstat64(unsigned int fd, struct stat64_emu31 __user *statbuf); +long compat_sys_s390_fstatat64(unsigned int dfd, const char __user *filename, struct stat64_emu31 __user *statbuf, int flag); +long compat_sys_s390_old_mmap(struct mmap_arg_struct_emu31 __user *arg); +long compat_sys_s390_mmap2(struct mmap_arg_struct_emu31 __user *arg); +long compat_sys_s390_read(unsigned int fd, char __user *buf, compat_size_t count); +long compat_sys_s390_write(unsigned int fd, const char __user *buf, compat_size_t count); +long compat_sys_s390_fadvise64(int fd, u32 high, u32 low, compat_size_t len, int advise); +long compat_sys_s390_fadvise64_64(struct fadvise64_64_args __user *args); +long compat_sys_s390_sync_file_range(int fd, u32 offhigh, u32 offlow, u32 nhigh, u32 nlow, unsigned int flags); +long compat_sys_s390_fallocate(int fd, int mode, u32 offhigh, u32 offlow, u32 lenhigh, u32 lenlow); +long compat_sys_sigreturn(void); +long compat_sys_rt_sigreturn(void); + +#endif /* _ASM_S390X_S390_H */ diff --git a/arch/s390/kernel/compat_ptrace.h b/arch/s390/kernel/compat_ptrace.h new file mode 100644 index 0000000000..3c400fc7e9 --- /dev/null +++ b/arch/s390/kernel/compat_ptrace.h @@ -0,0 +1,64 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _PTRACE32_H +#define _PTRACE32_H + +#include <asm/ptrace.h> /* needed for NUM_CR_WORDS */ +#include "compat_linux.h" /* needed for psw_compat_t */ + +struct compat_per_struct_kernel { + __u32 cr9; /* PER control bits */ + __u32 cr10; /* PER starting address */ + __u32 cr11; /* PER ending address */ + __u32 bits; /* Obsolete software bits */ + __u32 starting_addr; /* User specified start address */ + __u32 ending_addr; /* User specified end address */ + __u16 perc_atmid; /* PER trap ATMID */ + __u32 address; /* PER trap instruction address */ + __u8 access_id; /* PER trap access identification */ +}; + +struct compat_user_regs_struct +{ + psw_compat_t psw; + u32 gprs[NUM_GPRS]; + u32 acrs[NUM_ACRS]; + u32 orig_gpr2; + /* nb: there's a 4-byte hole here */ + s390_fp_regs fp_regs; + /* + * These per registers are in here so that gdb can modify them + * itself as there is no "official" ptrace interface for hardware + * watchpoints. This is the way intel does it. + */ + struct compat_per_struct_kernel per_info; + u32 ieee_instruction_pointer; /* obsolete, always 0 */ +}; + +struct compat_user { + /* We start with the registers, to mimic the way that "memory" + is returned from the ptrace(3,...) function. */ + struct compat_user_regs_struct regs; + /* The rest of this junk is to help gdb figure out what goes where */ + u32 u_tsize; /* Text segment size (pages). */ + u32 u_dsize; /* Data segment size (pages). */ + u32 u_ssize; /* Stack segment size (pages). */ + u32 start_code; /* Starting virtual address of text. */ + u32 start_stack; /* Starting virtual address of stack area. + This is actually the bottom of the stack, + the top of the stack is always found in the + esp register. */ + s32 signal; /* Signal that caused the core dump. */ + u32 u_ar0; /* Used by gdb to help find the values for */ + /* the registers. */ + u32 magic; /* To uniquely identify a core file */ + char u_comm[32]; /* User command that was responsible */ +}; + +typedef struct +{ + __u32 len; + __u32 kernel_addr; + __u32 process_addr; +} compat_ptrace_area; + +#endif /* _PTRACE32_H */ diff --git a/arch/s390/kernel/compat_signal.c b/arch/s390/kernel/compat_signal.c new file mode 100644 index 0000000000..cecedd01d4 --- /dev/null +++ b/arch/s390/kernel/compat_signal.c @@ -0,0 +1,422 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2000, 2006 + * Author(s): Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com) + * Gerhard Tonn (ton@de.ibm.com) + * + * Copyright (C) 1991, 1992 Linus Torvalds + * + * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson + */ + +#include <linux/compat.h> +#include <linux/sched.h> +#include <linux/sched/task_stack.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/kernel.h> +#include <linux/signal.h> +#include <linux/errno.h> +#include <linux/wait.h> +#include <linux/ptrace.h> +#include <linux/unistd.h> +#include <linux/stddef.h> +#include <linux/tty.h> +#include <linux/personality.h> +#include <linux/binfmts.h> +#include <asm/ucontext.h> +#include <linux/uaccess.h> +#include <asm/lowcore.h> +#include <asm/switch_to.h> +#include <asm/vdso.h> +#include "compat_linux.h" +#include "compat_ptrace.h" +#include "entry.h" + +typedef struct +{ + __u8 callee_used_stack[__SIGNAL_FRAMESIZE32]; + struct sigcontext32 sc; + _sigregs32 sregs; + int signo; + _sigregs_ext32 sregs_ext; + __u16 svc_insn; /* Offset of svc_insn is NOT fixed! */ +} sigframe32; + +typedef struct +{ + __u8 callee_used_stack[__SIGNAL_FRAMESIZE32]; + __u16 svc_insn; + compat_siginfo_t info; + struct ucontext32 uc; +} rt_sigframe32; + +/* Store registers needed to create the signal frame */ +static void store_sigregs(void) +{ + save_access_regs(current->thread.acrs); + save_fpu_regs(); +} + +/* Load registers after signal return */ +static void load_sigregs(void) +{ + restore_access_regs(current->thread.acrs); +} + +static int save_sigregs32(struct pt_regs *regs, _sigregs32 __user *sregs) +{ + _sigregs32 user_sregs; + int i; + + user_sregs.regs.psw.mask = (__u32)(regs->psw.mask >> 32); + user_sregs.regs.psw.mask &= PSW32_MASK_USER | PSW32_MASK_RI; + user_sregs.regs.psw.mask |= PSW32_USER_BITS; + user_sregs.regs.psw.addr = (__u32) regs->psw.addr | + (__u32)(regs->psw.mask & PSW_MASK_BA); + for (i = 0; i < NUM_GPRS; i++) + user_sregs.regs.gprs[i] = (__u32) regs->gprs[i]; + memcpy(&user_sregs.regs.acrs, current->thread.acrs, + sizeof(user_sregs.regs.acrs)); + fpregs_store((_s390_fp_regs *) &user_sregs.fpregs, ¤t->thread.fpu); + if (__copy_to_user(sregs, &user_sregs, sizeof(_sigregs32))) + return -EFAULT; + return 0; +} + +static int restore_sigregs32(struct pt_regs *regs,_sigregs32 __user *sregs) +{ + _sigregs32 user_sregs; + int i; + + /* Always make any pending restarted system call return -EINTR */ + current->restart_block.fn = do_no_restart_syscall; + + if (__copy_from_user(&user_sregs, &sregs->regs, sizeof(user_sregs))) + return -EFAULT; + + if (!is_ri_task(current) && (user_sregs.regs.psw.mask & PSW32_MASK_RI)) + return -EINVAL; + + /* Test the floating-point-control word. */ + if (test_fp_ctl(user_sregs.fpregs.fpc)) + return -EINVAL; + + /* Use regs->psw.mask instead of PSW_USER_BITS to preserve PER bit. */ + regs->psw.mask = (regs->psw.mask & ~(PSW_MASK_USER | PSW_MASK_RI)) | + (__u64)(user_sregs.regs.psw.mask & PSW32_MASK_USER) << 32 | + (__u64)(user_sregs.regs.psw.mask & PSW32_MASK_RI) << 32 | + (__u64)(user_sregs.regs.psw.addr & PSW32_ADDR_AMODE); + /* Check for invalid user address space control. */ + if ((regs->psw.mask & PSW_MASK_ASC) == PSW_ASC_HOME) + regs->psw.mask = PSW_ASC_PRIMARY | + (regs->psw.mask & ~PSW_MASK_ASC); + regs->psw.addr = (__u64)(user_sregs.regs.psw.addr & PSW32_ADDR_INSN); + for (i = 0; i < NUM_GPRS; i++) + regs->gprs[i] = (__u64) user_sregs.regs.gprs[i]; + memcpy(¤t->thread.acrs, &user_sregs.regs.acrs, + sizeof(current->thread.acrs)); + fpregs_load((_s390_fp_regs *) &user_sregs.fpregs, ¤t->thread.fpu); + + clear_pt_regs_flag(regs, PIF_SYSCALL); /* No longer in a system call */ + return 0; +} + +static int save_sigregs_ext32(struct pt_regs *regs, + _sigregs_ext32 __user *sregs_ext) +{ + __u32 gprs_high[NUM_GPRS]; + __u64 vxrs[__NUM_VXRS_LOW]; + int i; + + /* Save high gprs to signal stack */ + for (i = 0; i < NUM_GPRS; i++) + gprs_high[i] = regs->gprs[i] >> 32; + if (__copy_to_user(&sregs_ext->gprs_high, &gprs_high, + sizeof(sregs_ext->gprs_high))) + return -EFAULT; + + /* Save vector registers to signal stack */ + if (MACHINE_HAS_VX) { + for (i = 0; i < __NUM_VXRS_LOW; i++) + vxrs[i] = current->thread.fpu.vxrs[i].low; + if (__copy_to_user(&sregs_ext->vxrs_low, vxrs, + sizeof(sregs_ext->vxrs_low)) || + __copy_to_user(&sregs_ext->vxrs_high, + current->thread.fpu.vxrs + __NUM_VXRS_LOW, + sizeof(sregs_ext->vxrs_high))) + return -EFAULT; + } + return 0; +} + +static int restore_sigregs_ext32(struct pt_regs *regs, + _sigregs_ext32 __user *sregs_ext) +{ + __u32 gprs_high[NUM_GPRS]; + __u64 vxrs[__NUM_VXRS_LOW]; + int i; + + /* Restore high gprs from signal stack */ + if (__copy_from_user(&gprs_high, &sregs_ext->gprs_high, + sizeof(sregs_ext->gprs_high))) + return -EFAULT; + for (i = 0; i < NUM_GPRS; i++) + *(__u32 *)®s->gprs[i] = gprs_high[i]; + + /* Restore vector registers from signal stack */ + if (MACHINE_HAS_VX) { + if (__copy_from_user(vxrs, &sregs_ext->vxrs_low, + sizeof(sregs_ext->vxrs_low)) || + __copy_from_user(current->thread.fpu.vxrs + __NUM_VXRS_LOW, + &sregs_ext->vxrs_high, + sizeof(sregs_ext->vxrs_high))) + return -EFAULT; + for (i = 0; i < __NUM_VXRS_LOW; i++) + current->thread.fpu.vxrs[i].low = vxrs[i]; + } + return 0; +} + +COMPAT_SYSCALL_DEFINE0(sigreturn) +{ + struct pt_regs *regs = task_pt_regs(current); + sigframe32 __user *frame = (sigframe32 __user *)regs->gprs[15]; + sigset_t set; + + if (get_compat_sigset(&set, (compat_sigset_t __user *)frame->sc.oldmask)) + goto badframe; + set_current_blocked(&set); + save_fpu_regs(); + if (restore_sigregs32(regs, &frame->sregs)) + goto badframe; + if (restore_sigregs_ext32(regs, &frame->sregs_ext)) + goto badframe; + load_sigregs(); + return regs->gprs[2]; +badframe: + force_sig(SIGSEGV); + return 0; +} + +COMPAT_SYSCALL_DEFINE0(rt_sigreturn) +{ + struct pt_regs *regs = task_pt_regs(current); + rt_sigframe32 __user *frame = (rt_sigframe32 __user *)regs->gprs[15]; + sigset_t set; + + if (get_compat_sigset(&set, &frame->uc.uc_sigmask)) + goto badframe; + set_current_blocked(&set); + if (compat_restore_altstack(&frame->uc.uc_stack)) + goto badframe; + save_fpu_regs(); + if (restore_sigregs32(regs, &frame->uc.uc_mcontext)) + goto badframe; + if (restore_sigregs_ext32(regs, &frame->uc.uc_mcontext_ext)) + goto badframe; + load_sigregs(); + return regs->gprs[2]; +badframe: + force_sig(SIGSEGV); + return 0; +} + +/* + * Set up a signal frame. + */ + + +/* + * Determine which stack to use.. + */ +static inline void __user * +get_sigframe(struct k_sigaction *ka, struct pt_regs * regs, size_t frame_size) +{ + unsigned long sp; + + /* Default to using normal stack */ + sp = (unsigned long) A(regs->gprs[15]); + + /* Overflow on alternate signal stack gives SIGSEGV. */ + if (on_sig_stack(sp) && !on_sig_stack((sp - frame_size) & -8UL)) + return (void __user *) -1UL; + + /* This is the X/Open sanctioned signal stack switching. */ + if (ka->sa.sa_flags & SA_ONSTACK) { + if (! sas_ss_flags(sp)) + sp = current->sas_ss_sp + current->sas_ss_size; + } + + return (void __user *)((sp - frame_size) & -8ul); +} + +static int setup_frame32(struct ksignal *ksig, sigset_t *set, + struct pt_regs *regs) +{ + int sig = ksig->sig; + sigframe32 __user *frame; + unsigned long restorer; + size_t frame_size; + + /* + * gprs_high are always present for 31-bit compat tasks. + * The space for vector registers is only allocated if + * the machine supports it + */ + frame_size = sizeof(*frame) - sizeof(frame->sregs_ext.__reserved); + if (!MACHINE_HAS_VX) + frame_size -= sizeof(frame->sregs_ext.vxrs_low) + + sizeof(frame->sregs_ext.vxrs_high); + frame = get_sigframe(&ksig->ka, regs, frame_size); + if (frame == (void __user *) -1UL) + return -EFAULT; + + /* Set up backchain. */ + if (__put_user(regs->gprs[15], (unsigned int __user *) frame)) + return -EFAULT; + + /* Create struct sigcontext32 on the signal stack */ + if (put_compat_sigset((compat_sigset_t __user *)frame->sc.oldmask, + set, sizeof(compat_sigset_t))) + return -EFAULT; + if (__put_user(ptr_to_compat(&frame->sregs), &frame->sc.sregs)) + return -EFAULT; + + /* Store registers needed to create the signal frame */ + store_sigregs(); + + /* Create _sigregs32 on the signal stack */ + if (save_sigregs32(regs, &frame->sregs)) + return -EFAULT; + + /* Place signal number on stack to allow backtrace from handler. */ + if (__put_user(regs->gprs[2], (int __force __user *) &frame->signo)) + return -EFAULT; + + /* Create _sigregs_ext32 on the signal stack */ + if (save_sigregs_ext32(regs, &frame->sregs_ext)) + return -EFAULT; + + /* Set up to return from userspace. If provided, use a stub + already in userspace. */ + if (ksig->ka.sa.sa_flags & SA_RESTORER) { + restorer = (unsigned long __force) + ksig->ka.sa.sa_restorer | PSW32_ADDR_AMODE; + } else { + restorer = VDSO32_SYMBOL(current, sigreturn); + } + + /* Set up registers for signal handler */ + regs->gprs[14] = restorer; + regs->gprs[15] = (__force __u64) frame; + /* Force 31 bit amode and default user address space control. */ + regs->psw.mask = PSW_MASK_BA | + (PSW_USER_BITS & PSW_MASK_ASC) | + (regs->psw.mask & ~PSW_MASK_ASC); + regs->psw.addr = (__force __u64) ksig->ka.sa.sa_handler; + + regs->gprs[2] = sig; + regs->gprs[3] = (__force __u64) &frame->sc; + + /* We forgot to include these in the sigcontext. + To avoid breaking binary compatibility, they are passed as args. */ + if (sig == SIGSEGV || sig == SIGBUS || sig == SIGILL || + sig == SIGTRAP || sig == SIGFPE) { + /* set extra registers only for synchronous signals */ + regs->gprs[4] = regs->int_code & 127; + regs->gprs[5] = regs->int_parm_long; + regs->gprs[6] = current->thread.last_break; + } + + return 0; +} + +static int setup_rt_frame32(struct ksignal *ksig, sigset_t *set, + struct pt_regs *regs) +{ + rt_sigframe32 __user *frame; + unsigned long restorer; + size_t frame_size; + u32 uc_flags; + + frame_size = sizeof(*frame) - + sizeof(frame->uc.uc_mcontext_ext.__reserved); + /* + * gprs_high are always present for 31-bit compat tasks. + * The space for vector registers is only allocated if + * the machine supports it + */ + uc_flags = UC_GPRS_HIGH; + if (MACHINE_HAS_VX) { + uc_flags |= UC_VXRS; + } else + frame_size -= sizeof(frame->uc.uc_mcontext_ext.vxrs_low) + + sizeof(frame->uc.uc_mcontext_ext.vxrs_high); + frame = get_sigframe(&ksig->ka, regs, frame_size); + if (frame == (void __user *) -1UL) + return -EFAULT; + + /* Set up backchain. */ + if (__put_user(regs->gprs[15], (unsigned int __force __user *) frame)) + return -EFAULT; + + /* Set up to return from userspace. If provided, use a stub + already in userspace. */ + if (ksig->ka.sa.sa_flags & SA_RESTORER) { + restorer = (unsigned long __force) + ksig->ka.sa.sa_restorer | PSW32_ADDR_AMODE; + } else { + restorer = VDSO32_SYMBOL(current, rt_sigreturn); + } + + /* Create siginfo on the signal stack */ + if (copy_siginfo_to_user32(&frame->info, &ksig->info)) + return -EFAULT; + + /* Store registers needed to create the signal frame */ + store_sigregs(); + + /* Create ucontext on the signal stack. */ + if (__put_user(uc_flags, &frame->uc.uc_flags) || + __put_user(0, &frame->uc.uc_link) || + __compat_save_altstack(&frame->uc.uc_stack, regs->gprs[15]) || + save_sigregs32(regs, &frame->uc.uc_mcontext) || + put_compat_sigset(&frame->uc.uc_sigmask, set, sizeof(compat_sigset_t)) || + save_sigregs_ext32(regs, &frame->uc.uc_mcontext_ext)) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->gprs[14] = restorer; + regs->gprs[15] = (__force __u64) frame; + /* Force 31 bit amode and default user address space control. */ + regs->psw.mask = PSW_MASK_BA | + (PSW_USER_BITS & PSW_MASK_ASC) | + (regs->psw.mask & ~PSW_MASK_ASC); + regs->psw.addr = (__u64 __force) ksig->ka.sa.sa_handler; + + regs->gprs[2] = ksig->sig; + regs->gprs[3] = (__force __u64) &frame->info; + regs->gprs[4] = (__force __u64) &frame->uc; + regs->gprs[5] = current->thread.last_break; + return 0; +} + +/* + * OK, we're invoking a handler + */ + +void handle_signal32(struct ksignal *ksig, sigset_t *oldset, + struct pt_regs *regs) +{ + int ret; + + /* Set up the stack frame */ + if (ksig->ka.sa.sa_flags & SA_SIGINFO) + ret = setup_rt_frame32(ksig, oldset, regs); + else + ret = setup_frame32(ksig, oldset, regs); + + signal_setup_done(ret, ksig, test_thread_flag(TIF_SINGLE_STEP)); +} + diff --git a/arch/s390/kernel/cpcmd.c b/arch/s390/kernel/cpcmd.c new file mode 100644 index 0000000000..b210a29d3e --- /dev/null +++ b/arch/s390/kernel/cpcmd.c @@ -0,0 +1,114 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * S390 version + * Copyright IBM Corp. 1999, 2007 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), + * Christian Borntraeger (cborntra@de.ibm.com), + */ + +#define KMSG_COMPONENT "cpcmd" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/kernel.h> +#include <linux/export.h> +#include <linux/slab.h> +#include <linux/spinlock.h> +#include <linux/stddef.h> +#include <linux/string.h> +#include <linux/mm.h> +#include <linux/io.h> +#include <asm/diag.h> +#include <asm/ebcdic.h> +#include <asm/cpcmd.h> + +static DEFINE_SPINLOCK(cpcmd_lock); +static char cpcmd_buf[241]; + +static int diag8_noresponse(int cmdlen) +{ + asm volatile( + " diag %[rx],%[ry],0x8\n" + : [ry] "+&d" (cmdlen) + : [rx] "d" (__pa(cpcmd_buf)) + : "cc"); + return cmdlen; +} + +static int diag8_response(int cmdlen, char *response, int *rlen) +{ + union register_pair rx, ry; + int cc; + + rx.even = __pa(cpcmd_buf); + rx.odd = __pa(response); + ry.even = cmdlen | 0x40000000L; + ry.odd = *rlen; + asm volatile( + " diag %[rx],%[ry],0x8\n" + " ipm %[cc]\n" + " srl %[cc],28\n" + : [cc] "=&d" (cc), [ry] "+&d" (ry.pair) + : [rx] "d" (rx.pair) + : "cc"); + if (cc) + *rlen += ry.odd; + else + *rlen = ry.odd; + return ry.even; +} + +/* + * __cpcmd has some restrictions over cpcmd + * - __cpcmd is unlocked and therefore not SMP-safe + */ +int __cpcmd(const char *cmd, char *response, int rlen, int *response_code) +{ + int cmdlen; + int rc; + int response_len; + + cmdlen = strlen(cmd); + BUG_ON(cmdlen > 240); + memcpy(cpcmd_buf, cmd, cmdlen); + ASCEBC(cpcmd_buf, cmdlen); + + diag_stat_inc(DIAG_STAT_X008); + if (response) { + memset(response, 0, rlen); + response_len = rlen; + rc = diag8_response(cmdlen, response, &rlen); + EBCASC(response, response_len); + } else { + rc = diag8_noresponse(cmdlen); + } + if (response_code) + *response_code = rc; + return rlen; +} +EXPORT_SYMBOL(__cpcmd); + +int cpcmd(const char *cmd, char *response, int rlen, int *response_code) +{ + unsigned long flags; + char *lowbuf; + int len; + + if (is_vmalloc_or_module_addr(response)) { + lowbuf = kmalloc(rlen, GFP_KERNEL); + if (!lowbuf) { + pr_warn("The cpcmd kernel function failed to allocate a response buffer\n"); + return -ENOMEM; + } + spin_lock_irqsave(&cpcmd_lock, flags); + len = __cpcmd(cmd, lowbuf, rlen, response_code); + spin_unlock_irqrestore(&cpcmd_lock, flags); + memcpy(response, lowbuf, rlen); + kfree(lowbuf); + } else { + spin_lock_irqsave(&cpcmd_lock, flags); + len = __cpcmd(cmd, response, rlen, response_code); + spin_unlock_irqrestore(&cpcmd_lock, flags); + } + return len; +} +EXPORT_SYMBOL(cpcmd); diff --git a/arch/s390/kernel/cpufeature.c b/arch/s390/kernel/cpufeature.c new file mode 100644 index 0000000000..1b2ae42a0c --- /dev/null +++ b/arch/s390/kernel/cpufeature.c @@ -0,0 +1,46 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2022 + */ + +#include <linux/cpufeature.h> +#include <linux/bug.h> +#include <asm/elf.h> + +enum { + TYPE_HWCAP, + TYPE_FACILITY, +}; + +struct s390_cpu_feature { + unsigned int type : 4; + unsigned int num : 28; +}; + +static struct s390_cpu_feature s390_cpu_features[MAX_CPU_FEATURES] = { + [S390_CPU_FEATURE_MSA] = {.type = TYPE_HWCAP, .num = HWCAP_NR_MSA}, + [S390_CPU_FEATURE_VXRS] = {.type = TYPE_HWCAP, .num = HWCAP_NR_VXRS}, + [S390_CPU_FEATURE_UV] = {.type = TYPE_FACILITY, .num = 158}, +}; + +/* + * cpu_have_feature - Test CPU features on module initialization + */ +int cpu_have_feature(unsigned int num) +{ + struct s390_cpu_feature *feature; + + if (WARN_ON_ONCE(num >= MAX_CPU_FEATURES)) + return 0; + feature = &s390_cpu_features[num]; + switch (feature->type) { + case TYPE_HWCAP: + return !!(elf_hwcap & BIT(feature->num)); + case TYPE_FACILITY: + return test_facility(feature->num); + default: + WARN_ON_ONCE(1); + return 0; + } +} +EXPORT_SYMBOL(cpu_have_feature); diff --git a/arch/s390/kernel/crash_dump.c b/arch/s390/kernel/crash_dump.c new file mode 100644 index 0000000000..7af69948b2 --- /dev/null +++ b/arch/s390/kernel/crash_dump.c @@ -0,0 +1,652 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * S390 kdump implementation + * + * Copyright IBM Corp. 2011 + * Author(s): Michael Holzheu <holzheu@linux.vnet.ibm.com> + */ + +#include <linux/crash_dump.h> +#include <asm/lowcore.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/mm.h> +#include <linux/gfp.h> +#include <linux/slab.h> +#include <linux/memblock.h> +#include <linux/elf.h> +#include <linux/uio.h> +#include <asm/asm-offsets.h> +#include <asm/os_info.h> +#include <asm/elf.h> +#include <asm/ipl.h> +#include <asm/sclp.h> +#include <asm/maccess.h> + +#define PTR_ADD(x, y) (((char *) (x)) + ((unsigned long) (y))) +#define PTR_SUB(x, y) (((char *) (x)) - ((unsigned long) (y))) +#define PTR_DIFF(x, y) ((unsigned long)(((char *) (x)) - ((unsigned long) (y)))) + +static struct memblock_region oldmem_region; + +static struct memblock_type oldmem_type = { + .cnt = 1, + .max = 1, + .total_size = 0, + .regions = &oldmem_region, + .name = "oldmem", +}; + +struct save_area { + struct list_head list; + u64 psw[2]; + u64 ctrs[16]; + u64 gprs[16]; + u32 acrs[16]; + u64 fprs[16]; + u32 fpc; + u32 prefix; + u32 todpreg; + u64 timer; + u64 todcmp; + u64 vxrs_low[16]; + __vector128 vxrs_high[16]; +}; + +static LIST_HEAD(dump_save_areas); + +/* + * Allocate a save area + */ +struct save_area * __init save_area_alloc(bool is_boot_cpu) +{ + struct save_area *sa; + + sa = memblock_alloc(sizeof(*sa), 8); + if (!sa) + return NULL; + + if (is_boot_cpu) + list_add(&sa->list, &dump_save_areas); + else + list_add_tail(&sa->list, &dump_save_areas); + return sa; +} + +/* + * Return the address of the save area for the boot CPU + */ +struct save_area * __init save_area_boot_cpu(void) +{ + return list_first_entry_or_null(&dump_save_areas, struct save_area, list); +} + +/* + * Copy CPU registers into the save area + */ +void __init save_area_add_regs(struct save_area *sa, void *regs) +{ + struct lowcore *lc; + + lc = (struct lowcore *)(regs - __LC_FPREGS_SAVE_AREA); + memcpy(&sa->psw, &lc->psw_save_area, sizeof(sa->psw)); + memcpy(&sa->ctrs, &lc->cregs_save_area, sizeof(sa->ctrs)); + memcpy(&sa->gprs, &lc->gpregs_save_area, sizeof(sa->gprs)); + memcpy(&sa->acrs, &lc->access_regs_save_area, sizeof(sa->acrs)); + memcpy(&sa->fprs, &lc->floating_pt_save_area, sizeof(sa->fprs)); + memcpy(&sa->fpc, &lc->fpt_creg_save_area, sizeof(sa->fpc)); + memcpy(&sa->prefix, &lc->prefixreg_save_area, sizeof(sa->prefix)); + memcpy(&sa->todpreg, &lc->tod_progreg_save_area, sizeof(sa->todpreg)); + memcpy(&sa->timer, &lc->cpu_timer_save_area, sizeof(sa->timer)); + memcpy(&sa->todcmp, &lc->clock_comp_save_area, sizeof(sa->todcmp)); +} + +/* + * Copy vector registers into the save area + */ +void __init save_area_add_vxrs(struct save_area *sa, __vector128 *vxrs) +{ + int i; + + /* Copy lower halves of vector registers 0-15 */ + for (i = 0; i < 16; i++) + sa->vxrs_low[i] = vxrs[i].low; + /* Copy vector registers 16-31 */ + memcpy(sa->vxrs_high, vxrs + 16, 16 * sizeof(__vector128)); +} + +static size_t copy_oldmem_iter(struct iov_iter *iter, unsigned long src, size_t count) +{ + size_t len, copied, res = 0; + + while (count) { + if (!oldmem_data.start && src < sclp.hsa_size) { + /* Copy from zfcp/nvme dump HSA area */ + len = min(count, sclp.hsa_size - src); + copied = memcpy_hsa_iter(iter, src, len); + } else { + /* Check for swapped kdump oldmem areas */ + if (oldmem_data.start && src - oldmem_data.start < oldmem_data.size) { + src -= oldmem_data.start; + len = min(count, oldmem_data.size - src); + } else if (oldmem_data.start && src < oldmem_data.size) { + len = min(count, oldmem_data.size - src); + src += oldmem_data.start; + } else { + len = count; + } + copied = memcpy_real_iter(iter, src, len); + } + count -= copied; + src += copied; + res += copied; + if (copied < len) + break; + } + return res; +} + +int copy_oldmem_kernel(void *dst, unsigned long src, size_t count) +{ + struct iov_iter iter; + struct kvec kvec; + + kvec.iov_base = dst; + kvec.iov_len = count; + iov_iter_kvec(&iter, ITER_DEST, &kvec, 1, count); + if (copy_oldmem_iter(&iter, src, count) < count) + return -EFAULT; + return 0; +} + +/* + * Copy one page from "oldmem" + */ +ssize_t copy_oldmem_page(struct iov_iter *iter, unsigned long pfn, size_t csize, + unsigned long offset) +{ + unsigned long src; + + src = pfn_to_phys(pfn) + offset; + return copy_oldmem_iter(iter, src, csize); +} + +/* + * Remap "oldmem" for kdump + * + * For the kdump reserved memory this functions performs a swap operation: + * [0 - OLDMEM_SIZE] is mapped to [OLDMEM_BASE - OLDMEM_BASE + OLDMEM_SIZE] + */ +static int remap_oldmem_pfn_range_kdump(struct vm_area_struct *vma, + unsigned long from, unsigned long pfn, + unsigned long size, pgprot_t prot) +{ + unsigned long size_old; + int rc; + + if (pfn < oldmem_data.size >> PAGE_SHIFT) { + size_old = min(size, oldmem_data.size - (pfn << PAGE_SHIFT)); + rc = remap_pfn_range(vma, from, + pfn + (oldmem_data.start >> PAGE_SHIFT), + size_old, prot); + if (rc || size == size_old) + return rc; + size -= size_old; + from += size_old; + pfn += size_old >> PAGE_SHIFT; + } + return remap_pfn_range(vma, from, pfn, size, prot); +} + +/* + * Remap "oldmem" for zfcp/nvme dump + * + * We only map available memory above HSA size. Memory below HSA size + * is read on demand using the copy_oldmem_page() function. + */ +static int remap_oldmem_pfn_range_zfcpdump(struct vm_area_struct *vma, + unsigned long from, + unsigned long pfn, + unsigned long size, pgprot_t prot) +{ + unsigned long hsa_end = sclp.hsa_size; + unsigned long size_hsa; + + if (pfn < hsa_end >> PAGE_SHIFT) { + size_hsa = min(size, hsa_end - (pfn << PAGE_SHIFT)); + if (size == size_hsa) + return 0; + size -= size_hsa; + from += size_hsa; + pfn += size_hsa >> PAGE_SHIFT; + } + return remap_pfn_range(vma, from, pfn, size, prot); +} + +/* + * Remap "oldmem" for kdump or zfcp/nvme dump + */ +int remap_oldmem_pfn_range(struct vm_area_struct *vma, unsigned long from, + unsigned long pfn, unsigned long size, pgprot_t prot) +{ + if (oldmem_data.start) + return remap_oldmem_pfn_range_kdump(vma, from, pfn, size, prot); + else + return remap_oldmem_pfn_range_zfcpdump(vma, from, pfn, size, + prot); +} + +static const char *nt_name(Elf64_Word type) +{ + const char *name = "LINUX"; + + if (type == NT_PRPSINFO || type == NT_PRSTATUS || type == NT_PRFPREG) + name = KEXEC_CORE_NOTE_NAME; + return name; +} + +/* + * Initialize ELF note + */ +static void *nt_init_name(void *buf, Elf64_Word type, void *desc, int d_len, + const char *name) +{ + Elf64_Nhdr *note; + u64 len; + + note = (Elf64_Nhdr *)buf; + note->n_namesz = strlen(name) + 1; + note->n_descsz = d_len; + note->n_type = type; + len = sizeof(Elf64_Nhdr); + + memcpy(buf + len, name, note->n_namesz); + len = roundup(len + note->n_namesz, 4); + + memcpy(buf + len, desc, note->n_descsz); + len = roundup(len + note->n_descsz, 4); + + return PTR_ADD(buf, len); +} + +static inline void *nt_init(void *buf, Elf64_Word type, void *desc, int d_len) +{ + return nt_init_name(buf, type, desc, d_len, nt_name(type)); +} + +/* + * Calculate the size of ELF note + */ +static size_t nt_size_name(int d_len, const char *name) +{ + size_t size; + + size = sizeof(Elf64_Nhdr); + size += roundup(strlen(name) + 1, 4); + size += roundup(d_len, 4); + + return size; +} + +static inline size_t nt_size(Elf64_Word type, int d_len) +{ + return nt_size_name(d_len, nt_name(type)); +} + +/* + * Fill ELF notes for one CPU with save area registers + */ +static void *fill_cpu_elf_notes(void *ptr, int cpu, struct save_area *sa) +{ + struct elf_prstatus nt_prstatus; + elf_fpregset_t nt_fpregset; + + /* Prepare prstatus note */ + memset(&nt_prstatus, 0, sizeof(nt_prstatus)); + memcpy(&nt_prstatus.pr_reg.gprs, sa->gprs, sizeof(sa->gprs)); + memcpy(&nt_prstatus.pr_reg.psw, sa->psw, sizeof(sa->psw)); + memcpy(&nt_prstatus.pr_reg.acrs, sa->acrs, sizeof(sa->acrs)); + nt_prstatus.common.pr_pid = cpu; + /* Prepare fpregset (floating point) note */ + memset(&nt_fpregset, 0, sizeof(nt_fpregset)); + memcpy(&nt_fpregset.fpc, &sa->fpc, sizeof(sa->fpc)); + memcpy(&nt_fpregset.fprs, &sa->fprs, sizeof(sa->fprs)); + /* Create ELF notes for the CPU */ + ptr = nt_init(ptr, NT_PRSTATUS, &nt_prstatus, sizeof(nt_prstatus)); + ptr = nt_init(ptr, NT_PRFPREG, &nt_fpregset, sizeof(nt_fpregset)); + ptr = nt_init(ptr, NT_S390_TIMER, &sa->timer, sizeof(sa->timer)); + ptr = nt_init(ptr, NT_S390_TODCMP, &sa->todcmp, sizeof(sa->todcmp)); + ptr = nt_init(ptr, NT_S390_TODPREG, &sa->todpreg, sizeof(sa->todpreg)); + ptr = nt_init(ptr, NT_S390_CTRS, &sa->ctrs, sizeof(sa->ctrs)); + ptr = nt_init(ptr, NT_S390_PREFIX, &sa->prefix, sizeof(sa->prefix)); + if (MACHINE_HAS_VX) { + ptr = nt_init(ptr, NT_S390_VXRS_HIGH, + &sa->vxrs_high, sizeof(sa->vxrs_high)); + ptr = nt_init(ptr, NT_S390_VXRS_LOW, + &sa->vxrs_low, sizeof(sa->vxrs_low)); + } + return ptr; +} + +/* + * Calculate size of ELF notes per cpu + */ +static size_t get_cpu_elf_notes_size(void) +{ + struct save_area *sa = NULL; + size_t size; + + size = nt_size(NT_PRSTATUS, sizeof(struct elf_prstatus)); + size += nt_size(NT_PRFPREG, sizeof(elf_fpregset_t)); + size += nt_size(NT_S390_TIMER, sizeof(sa->timer)); + size += nt_size(NT_S390_TODCMP, sizeof(sa->todcmp)); + size += nt_size(NT_S390_TODPREG, sizeof(sa->todpreg)); + size += nt_size(NT_S390_CTRS, sizeof(sa->ctrs)); + size += nt_size(NT_S390_PREFIX, sizeof(sa->prefix)); + if (MACHINE_HAS_VX) { + size += nt_size(NT_S390_VXRS_HIGH, sizeof(sa->vxrs_high)); + size += nt_size(NT_S390_VXRS_LOW, sizeof(sa->vxrs_low)); + } + + return size; +} + +/* + * Initialize prpsinfo note (new kernel) + */ +static void *nt_prpsinfo(void *ptr) +{ + struct elf_prpsinfo prpsinfo; + + memset(&prpsinfo, 0, sizeof(prpsinfo)); + prpsinfo.pr_sname = 'R'; + strcpy(prpsinfo.pr_fname, "vmlinux"); + return nt_init(ptr, NT_PRPSINFO, &prpsinfo, sizeof(prpsinfo)); +} + +/* + * Get vmcoreinfo using lowcore->vmcore_info (new kernel) + */ +static void *get_vmcoreinfo_old(unsigned long *size) +{ + char nt_name[11], *vmcoreinfo; + unsigned long addr; + Elf64_Nhdr note; + + if (copy_oldmem_kernel(&addr, __LC_VMCORE_INFO, sizeof(addr))) + return NULL; + memset(nt_name, 0, sizeof(nt_name)); + if (copy_oldmem_kernel(¬e, addr, sizeof(note))) + return NULL; + if (copy_oldmem_kernel(nt_name, addr + sizeof(note), + sizeof(nt_name) - 1)) + return NULL; + if (strcmp(nt_name, VMCOREINFO_NOTE_NAME) != 0) + return NULL; + vmcoreinfo = kzalloc(note.n_descsz, GFP_KERNEL); + if (!vmcoreinfo) + return NULL; + if (copy_oldmem_kernel(vmcoreinfo, addr + 24, note.n_descsz)) { + kfree(vmcoreinfo); + return NULL; + } + *size = note.n_descsz; + return vmcoreinfo; +} + +/* + * Initialize vmcoreinfo note (new kernel) + */ +static void *nt_vmcoreinfo(void *ptr) +{ + const char *name = VMCOREINFO_NOTE_NAME; + unsigned long size; + void *vmcoreinfo; + + vmcoreinfo = os_info_old_entry(OS_INFO_VMCOREINFO, &size); + if (vmcoreinfo) + return nt_init_name(ptr, 0, vmcoreinfo, size, name); + + vmcoreinfo = get_vmcoreinfo_old(&size); + if (!vmcoreinfo) + return ptr; + ptr = nt_init_name(ptr, 0, vmcoreinfo, size, name); + kfree(vmcoreinfo); + return ptr; +} + +static size_t nt_vmcoreinfo_size(void) +{ + const char *name = VMCOREINFO_NOTE_NAME; + unsigned long size; + void *vmcoreinfo; + + vmcoreinfo = os_info_old_entry(OS_INFO_VMCOREINFO, &size); + if (vmcoreinfo) + return nt_size_name(size, name); + + vmcoreinfo = get_vmcoreinfo_old(&size); + if (!vmcoreinfo) + return 0; + + kfree(vmcoreinfo); + return nt_size_name(size, name); +} + +/* + * Initialize final note (needed for /proc/vmcore code) + */ +static void *nt_final(void *ptr) +{ + Elf64_Nhdr *note; + + note = (Elf64_Nhdr *) ptr; + note->n_namesz = 0; + note->n_descsz = 0; + note->n_type = 0; + return PTR_ADD(ptr, sizeof(Elf64_Nhdr)); +} + +/* + * Initialize ELF header (new kernel) + */ +static void *ehdr_init(Elf64_Ehdr *ehdr, int mem_chunk_cnt) +{ + memset(ehdr, 0, sizeof(*ehdr)); + memcpy(ehdr->e_ident, ELFMAG, SELFMAG); + ehdr->e_ident[EI_CLASS] = ELFCLASS64; + ehdr->e_ident[EI_DATA] = ELFDATA2MSB; + ehdr->e_ident[EI_VERSION] = EV_CURRENT; + memset(ehdr->e_ident + EI_PAD, 0, EI_NIDENT - EI_PAD); + ehdr->e_type = ET_CORE; + ehdr->e_machine = EM_S390; + ehdr->e_version = EV_CURRENT; + ehdr->e_phoff = sizeof(Elf64_Ehdr); + ehdr->e_ehsize = sizeof(Elf64_Ehdr); + ehdr->e_phentsize = sizeof(Elf64_Phdr); + ehdr->e_phnum = mem_chunk_cnt + 1; + return ehdr + 1; +} + +/* + * Return CPU count for ELF header (new kernel) + */ +static int get_cpu_cnt(void) +{ + struct save_area *sa; + int cpus = 0; + + list_for_each_entry(sa, &dump_save_areas, list) + if (sa->prefix != 0) + cpus++; + return cpus; +} + +/* + * Return memory chunk count for ELF header (new kernel) + */ +static int get_mem_chunk_cnt(void) +{ + int cnt = 0; + u64 idx; + + for_each_physmem_range(idx, &oldmem_type, NULL, NULL) + cnt++; + return cnt; +} + +/* + * Initialize ELF loads (new kernel) + */ +static void loads_init(Elf64_Phdr *phdr, u64 loads_offset) +{ + phys_addr_t start, end; + u64 idx; + + for_each_physmem_range(idx, &oldmem_type, &start, &end) { + phdr->p_filesz = end - start; + phdr->p_type = PT_LOAD; + phdr->p_offset = start; + phdr->p_vaddr = start; + phdr->p_paddr = start; + phdr->p_memsz = end - start; + phdr->p_flags = PF_R | PF_W | PF_X; + phdr->p_align = PAGE_SIZE; + phdr++; + } +} + +/* + * Initialize notes (new kernel) + */ +static void *notes_init(Elf64_Phdr *phdr, void *ptr, u64 notes_offset) +{ + struct save_area *sa; + void *ptr_start = ptr; + int cpu; + + ptr = nt_prpsinfo(ptr); + + cpu = 1; + list_for_each_entry(sa, &dump_save_areas, list) + if (sa->prefix != 0) + ptr = fill_cpu_elf_notes(ptr, cpu++, sa); + ptr = nt_vmcoreinfo(ptr); + ptr = nt_final(ptr); + memset(phdr, 0, sizeof(*phdr)); + phdr->p_type = PT_NOTE; + phdr->p_offset = notes_offset; + phdr->p_filesz = (unsigned long) PTR_SUB(ptr, ptr_start); + phdr->p_memsz = phdr->p_filesz; + return ptr; +} + +static size_t get_elfcorehdr_size(int mem_chunk_cnt) +{ + size_t size; + + size = sizeof(Elf64_Ehdr); + /* PT_NOTES */ + size += sizeof(Elf64_Phdr); + /* nt_prpsinfo */ + size += nt_size(NT_PRPSINFO, sizeof(struct elf_prpsinfo)); + /* regsets */ + size += get_cpu_cnt() * get_cpu_elf_notes_size(); + /* nt_vmcoreinfo */ + size += nt_vmcoreinfo_size(); + /* nt_final */ + size += sizeof(Elf64_Nhdr); + /* PT_LOADS */ + size += mem_chunk_cnt * sizeof(Elf64_Phdr); + + return size; +} + +/* + * Create ELF core header (new kernel) + */ +int elfcorehdr_alloc(unsigned long long *addr, unsigned long long *size) +{ + Elf64_Phdr *phdr_notes, *phdr_loads; + size_t alloc_size; + int mem_chunk_cnt; + void *ptr, *hdr; + u64 hdr_off; + + /* If we are not in kdump or zfcp/nvme dump mode return */ + if (!oldmem_data.start && !is_ipl_type_dump()) + return 0; + /* If we cannot get HSA size for zfcp/nvme dump return error */ + if (is_ipl_type_dump() && !sclp.hsa_size) + return -ENODEV; + + /* For kdump, exclude previous crashkernel memory */ + if (oldmem_data.start) { + oldmem_region.base = oldmem_data.start; + oldmem_region.size = oldmem_data.size; + oldmem_type.total_size = oldmem_data.size; + } + + mem_chunk_cnt = get_mem_chunk_cnt(); + + alloc_size = get_elfcorehdr_size(mem_chunk_cnt); + + hdr = kzalloc(alloc_size, GFP_KERNEL); + + /* Without elfcorehdr /proc/vmcore cannot be created. Thus creating + * a dump with this crash kernel will fail. Panic now to allow other + * dump mechanisms to take over. + */ + if (!hdr) + panic("s390 kdump allocating elfcorehdr failed"); + + /* Init elf header */ + ptr = ehdr_init(hdr, mem_chunk_cnt); + /* Init program headers */ + phdr_notes = ptr; + ptr = PTR_ADD(ptr, sizeof(Elf64_Phdr)); + phdr_loads = ptr; + ptr = PTR_ADD(ptr, sizeof(Elf64_Phdr) * mem_chunk_cnt); + /* Init notes */ + hdr_off = PTR_DIFF(ptr, hdr); + ptr = notes_init(phdr_notes, ptr, ((unsigned long) hdr) + hdr_off); + /* Init loads */ + hdr_off = PTR_DIFF(ptr, hdr); + loads_init(phdr_loads, hdr_off); + *addr = (unsigned long long) hdr; + *size = (unsigned long long) hdr_off; + BUG_ON(elfcorehdr_size > alloc_size); + return 0; +} + +/* + * Free ELF core header (new kernel) + */ +void elfcorehdr_free(unsigned long long addr) +{ + kfree((void *)(unsigned long)addr); +} + +/* + * Read from ELF header + */ +ssize_t elfcorehdr_read(char *buf, size_t count, u64 *ppos) +{ + void *src = (void *)(unsigned long)*ppos; + + memcpy(buf, src, count); + *ppos += count; + return count; +} + +/* + * Read from ELF notes data + */ +ssize_t elfcorehdr_read_notes(char *buf, size_t count, u64 *ppos) +{ + void *src = (void *)(unsigned long)*ppos; + + memcpy(buf, src, count); + *ppos += count; + return count; +} diff --git a/arch/s390/kernel/debug.c b/arch/s390/kernel/debug.c new file mode 100644 index 0000000000..a85e0c3e70 --- /dev/null +++ b/arch/s390/kernel/debug.c @@ -0,0 +1,1574 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * S/390 debug facility + * + * Copyright IBM Corp. 1999, 2020 + * + * Author(s): Michael Holzheu (holzheu@de.ibm.com), + * Holger Smolinski (Holger.Smolinski@de.ibm.com) + * + * Bugreports to: <Linux390@de.ibm.com> + */ + +#define KMSG_COMPONENT "s390dbf" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/stddef.h> +#include <linux/kernel.h> +#include <linux/errno.h> +#include <linux/slab.h> +#include <linux/ctype.h> +#include <linux/string.h> +#include <linux/sysctl.h> +#include <linux/uaccess.h> +#include <linux/export.h> +#include <linux/init.h> +#include <linux/fs.h> +#include <linux/minmax.h> +#include <linux/debugfs.h> + +#include <asm/debug.h> + +#define DEBUG_PROLOG_ENTRY -1 + +#define ALL_AREAS 0 /* copy all debug areas */ +#define NO_AREAS 1 /* copy no debug areas */ + +/* typedefs */ + +typedef struct file_private_info { + loff_t offset; /* offset of last read in file */ + int act_area; /* number of last formated area */ + int act_page; /* act page in given area */ + int act_entry; /* last formated entry (offset */ + /* relative to beginning of last */ + /* formated page) */ + size_t act_entry_offset; /* up to this offset we copied */ + /* in last read the last formated */ + /* entry to userland */ + char temp_buf[2048]; /* buffer for output */ + debug_info_t *debug_info_org; /* original debug information */ + debug_info_t *debug_info_snap; /* snapshot of debug information */ + struct debug_view *view; /* used view of debug info */ +} file_private_info_t; + +typedef struct { + char *string; + /* + * This assumes that all args are converted into longs + * on L/390 this is the case for all types of parameter + * except of floats, and long long (32 bit) + * + */ + long args[]; +} debug_sprintf_entry_t; + +/* internal function prototyes */ + +static int debug_init(void); +static ssize_t debug_output(struct file *file, char __user *user_buf, + size_t user_len, loff_t *offset); +static ssize_t debug_input(struct file *file, const char __user *user_buf, + size_t user_len, loff_t *offset); +static int debug_open(struct inode *inode, struct file *file); +static int debug_close(struct inode *inode, struct file *file); +static debug_info_t *debug_info_create(const char *name, int pages_per_area, + int nr_areas, int buf_size, umode_t mode); +static void debug_info_get(debug_info_t *); +static void debug_info_put(debug_info_t *); +static int debug_prolog_level_fn(debug_info_t *id, + struct debug_view *view, char *out_buf); +static int debug_input_level_fn(debug_info_t *id, struct debug_view *view, + struct file *file, const char __user *user_buf, + size_t user_buf_size, loff_t *offset); +static int debug_prolog_pages_fn(debug_info_t *id, + struct debug_view *view, char *out_buf); +static int debug_input_pages_fn(debug_info_t *id, struct debug_view *view, + struct file *file, const char __user *user_buf, + size_t user_buf_size, loff_t *offset); +static int debug_input_flush_fn(debug_info_t *id, struct debug_view *view, + struct file *file, const char __user *user_buf, + size_t user_buf_size, loff_t *offset); +static int debug_hex_ascii_format_fn(debug_info_t *id, struct debug_view *view, + char *out_buf, const char *in_buf); +static int debug_sprintf_format_fn(debug_info_t *id, struct debug_view *view, + char *out_buf, const char *inbuf); +static void debug_areas_swap(debug_info_t *a, debug_info_t *b); +static void debug_events_append(debug_info_t *dest, debug_info_t *src); + +/* globals */ + +struct debug_view debug_hex_ascii_view = { + "hex_ascii", + NULL, + &debug_dflt_header_fn, + &debug_hex_ascii_format_fn, + NULL, + NULL +}; +EXPORT_SYMBOL(debug_hex_ascii_view); + +static struct debug_view debug_level_view = { + "level", + &debug_prolog_level_fn, + NULL, + NULL, + &debug_input_level_fn, + NULL +}; + +static struct debug_view debug_pages_view = { + "pages", + &debug_prolog_pages_fn, + NULL, + NULL, + &debug_input_pages_fn, + NULL +}; + +static struct debug_view debug_flush_view = { + "flush", + NULL, + NULL, + NULL, + &debug_input_flush_fn, + NULL +}; + +struct debug_view debug_sprintf_view = { + "sprintf", + NULL, + &debug_dflt_header_fn, + &debug_sprintf_format_fn, + NULL, + NULL +}; +EXPORT_SYMBOL(debug_sprintf_view); + +/* used by dump analysis tools to determine version of debug feature */ +static unsigned int __used debug_feature_version = __DEBUG_FEATURE_VERSION; + +/* static globals */ + +static debug_info_t *debug_area_first; +static debug_info_t *debug_area_last; +static DEFINE_MUTEX(debug_mutex); + +static int initialized; +static int debug_critical; + +static const struct file_operations debug_file_ops = { + .owner = THIS_MODULE, + .read = debug_output, + .write = debug_input, + .open = debug_open, + .release = debug_close, + .llseek = no_llseek, +}; + +static struct dentry *debug_debugfs_root_entry; + +/* functions */ + +/* + * debug_areas_alloc + * - Debug areas are implemented as a threedimensonal array: + * areas[areanumber][pagenumber][pageoffset] + */ + +static debug_entry_t ***debug_areas_alloc(int pages_per_area, int nr_areas) +{ + debug_entry_t ***areas; + int i, j; + + areas = kmalloc_array(nr_areas, sizeof(debug_entry_t **), GFP_KERNEL); + if (!areas) + goto fail_malloc_areas; + for (i = 0; i < nr_areas; i++) { + /* GFP_NOWARN to avoid user triggerable WARN, we handle fails */ + areas[i] = kmalloc_array(pages_per_area, + sizeof(debug_entry_t *), + GFP_KERNEL | __GFP_NOWARN); + if (!areas[i]) + goto fail_malloc_areas2; + for (j = 0; j < pages_per_area; j++) { + areas[i][j] = kzalloc(PAGE_SIZE, GFP_KERNEL); + if (!areas[i][j]) { + for (j--; j >= 0 ; j--) + kfree(areas[i][j]); + kfree(areas[i]); + goto fail_malloc_areas2; + } + } + } + return areas; + +fail_malloc_areas2: + for (i--; i >= 0; i--) { + for (j = 0; j < pages_per_area; j++) + kfree(areas[i][j]); + kfree(areas[i]); + } + kfree(areas); +fail_malloc_areas: + return NULL; +} + +/* + * debug_info_alloc + * - alloc new debug-info + */ +static debug_info_t *debug_info_alloc(const char *name, int pages_per_area, + int nr_areas, int buf_size, int level, + int mode) +{ + debug_info_t *rc; + + /* alloc everything */ + rc = kmalloc(sizeof(debug_info_t), GFP_KERNEL); + if (!rc) + goto fail_malloc_rc; + rc->active_entries = kcalloc(nr_areas, sizeof(int), GFP_KERNEL); + if (!rc->active_entries) + goto fail_malloc_active_entries; + rc->active_pages = kcalloc(nr_areas, sizeof(int), GFP_KERNEL); + if (!rc->active_pages) + goto fail_malloc_active_pages; + if ((mode == ALL_AREAS) && (pages_per_area != 0)) { + rc->areas = debug_areas_alloc(pages_per_area, nr_areas); + if (!rc->areas) + goto fail_malloc_areas; + } else { + rc->areas = NULL; + } + + /* initialize members */ + spin_lock_init(&rc->lock); + rc->pages_per_area = pages_per_area; + rc->nr_areas = nr_areas; + rc->active_area = 0; + rc->level = level; + rc->buf_size = buf_size; + rc->entry_size = sizeof(debug_entry_t) + buf_size; + strscpy(rc->name, name, sizeof(rc->name)); + memset(rc->views, 0, DEBUG_MAX_VIEWS * sizeof(struct debug_view *)); + memset(rc->debugfs_entries, 0, DEBUG_MAX_VIEWS * sizeof(struct dentry *)); + refcount_set(&(rc->ref_count), 0); + + return rc; + +fail_malloc_areas: + kfree(rc->active_pages); +fail_malloc_active_pages: + kfree(rc->active_entries); +fail_malloc_active_entries: + kfree(rc); +fail_malloc_rc: + return NULL; +} + +/* + * debug_areas_free + * - free all debug areas + */ +static void debug_areas_free(debug_info_t *db_info) +{ + int i, j; + + if (!db_info->areas) + return; + for (i = 0; i < db_info->nr_areas; i++) { + for (j = 0; j < db_info->pages_per_area; j++) + kfree(db_info->areas[i][j]); + kfree(db_info->areas[i]); + } + kfree(db_info->areas); + db_info->areas = NULL; +} + +/* + * debug_info_free + * - free memory debug-info + */ +static void debug_info_free(debug_info_t *db_info) +{ + debug_areas_free(db_info); + kfree(db_info->active_entries); + kfree(db_info->active_pages); + kfree(db_info); +} + +/* + * debug_info_create + * - create new debug-info + */ + +static debug_info_t *debug_info_create(const char *name, int pages_per_area, + int nr_areas, int buf_size, umode_t mode) +{ + debug_info_t *rc; + + rc = debug_info_alloc(name, pages_per_area, nr_areas, buf_size, + DEBUG_DEFAULT_LEVEL, ALL_AREAS); + if (!rc) + goto out; + + rc->mode = mode & ~S_IFMT; + refcount_set(&rc->ref_count, 1); +out: + return rc; +} + +/* + * debug_info_copy + * - copy debug-info + */ +static debug_info_t *debug_info_copy(debug_info_t *in, int mode) +{ + unsigned long flags; + debug_info_t *rc; + int i, j; + + /* get a consistent copy of the debug areas */ + do { + rc = debug_info_alloc(in->name, in->pages_per_area, + in->nr_areas, in->buf_size, in->level, mode); + spin_lock_irqsave(&in->lock, flags); + if (!rc) + goto out; + /* has something changed in the meantime ? */ + if ((rc->pages_per_area == in->pages_per_area) && + (rc->nr_areas == in->nr_areas)) { + break; + } + spin_unlock_irqrestore(&in->lock, flags); + debug_info_free(rc); + } while (1); + + if (mode == NO_AREAS) + goto out; + + for (i = 0; i < in->nr_areas; i++) { + for (j = 0; j < in->pages_per_area; j++) + memcpy(rc->areas[i][j], in->areas[i][j], PAGE_SIZE); + } +out: + spin_unlock_irqrestore(&in->lock, flags); + return rc; +} + +/* + * debug_info_get + * - increments reference count for debug-info + */ +static void debug_info_get(debug_info_t *db_info) +{ + if (db_info) + refcount_inc(&db_info->ref_count); +} + +/* + * debug_info_put: + * - decreases reference count for debug-info and frees it if necessary + */ +static void debug_info_put(debug_info_t *db_info) +{ + if (!db_info) + return; + if (refcount_dec_and_test(&db_info->ref_count)) + debug_info_free(db_info); +} + +/* + * debug_format_entry: + * - format one debug entry and return size of formated data + */ +static int debug_format_entry(file_private_info_t *p_info) +{ + debug_info_t *id_snap = p_info->debug_info_snap; + struct debug_view *view = p_info->view; + debug_entry_t *act_entry; + size_t len = 0; + + if (p_info->act_entry == DEBUG_PROLOG_ENTRY) { + /* print prolog */ + if (view->prolog_proc) + len += view->prolog_proc(id_snap, view, p_info->temp_buf); + goto out; + } + if (!id_snap->areas) /* this is true, if we have a prolog only view */ + goto out; /* or if 'pages_per_area' is 0 */ + act_entry = (debug_entry_t *) ((char *)id_snap->areas[p_info->act_area] + [p_info->act_page] + p_info->act_entry); + + if (act_entry->clock == 0LL) + goto out; /* empty entry */ + if (view->header_proc) + len += view->header_proc(id_snap, view, p_info->act_area, + act_entry, p_info->temp_buf + len); + if (view->format_proc) + len += view->format_proc(id_snap, view, p_info->temp_buf + len, + DEBUG_DATA(act_entry)); +out: + return len; +} + +/* + * debug_next_entry: + * - goto next entry in p_info + */ +static inline int debug_next_entry(file_private_info_t *p_info) +{ + debug_info_t *id; + + id = p_info->debug_info_snap; + if (p_info->act_entry == DEBUG_PROLOG_ENTRY) { + p_info->act_entry = 0; + p_info->act_page = 0; + goto out; + } + if (!id->areas) + return 1; + p_info->act_entry += id->entry_size; + /* switch to next page, if we reached the end of the page */ + if (p_info->act_entry > (PAGE_SIZE - id->entry_size)) { + /* next page */ + p_info->act_entry = 0; + p_info->act_page += 1; + if ((p_info->act_page % id->pages_per_area) == 0) { + /* next area */ + p_info->act_area++; + p_info->act_page = 0; + } + if (p_info->act_area >= id->nr_areas) + return 1; + } +out: + return 0; +} + +/* + * debug_output: + * - called for user read() + * - copies formated debug entries to the user buffer + */ +static ssize_t debug_output(struct file *file, /* file descriptor */ + char __user *user_buf, /* user buffer */ + size_t len, /* length of buffer */ + loff_t *offset) /* offset in the file */ +{ + size_t count = 0; + size_t entry_offset; + file_private_info_t *p_info; + + p_info = (file_private_info_t *) file->private_data; + if (*offset != p_info->offset) + return -EPIPE; + if (p_info->act_area >= p_info->debug_info_snap->nr_areas) + return 0; + entry_offset = p_info->act_entry_offset; + while (count < len) { + int formatted_line_residue; + int formatted_line_size; + int user_buf_residue; + size_t copy_size; + + formatted_line_size = debug_format_entry(p_info); + formatted_line_residue = formatted_line_size - entry_offset; + user_buf_residue = len-count; + copy_size = min(user_buf_residue, formatted_line_residue); + if (copy_size) { + if (copy_to_user(user_buf + count, p_info->temp_buf + + entry_offset, copy_size)) + return -EFAULT; + count += copy_size; + entry_offset += copy_size; + } + if (copy_size == formatted_line_residue) { + entry_offset = 0; + if (debug_next_entry(p_info)) + goto out; + } + } +out: + p_info->offset = *offset + count; + p_info->act_entry_offset = entry_offset; + *offset = p_info->offset; + return count; +} + +/* + * debug_input: + * - called for user write() + * - calls input function of view + */ +static ssize_t debug_input(struct file *file, const char __user *user_buf, + size_t length, loff_t *offset) +{ + file_private_info_t *p_info; + int rc = 0; + + mutex_lock(&debug_mutex); + p_info = ((file_private_info_t *) file->private_data); + if (p_info->view->input_proc) { + rc = p_info->view->input_proc(p_info->debug_info_org, + p_info->view, file, user_buf, + length, offset); + } else { + rc = -EPERM; + } + mutex_unlock(&debug_mutex); + return rc; /* number of input characters */ +} + +/* + * debug_open: + * - called for user open() + * - copies formated output to private_data area of the file + * handle + */ +static int debug_open(struct inode *inode, struct file *file) +{ + debug_info_t *debug_info, *debug_info_snapshot; + file_private_info_t *p_info; + int i, rc = 0; + + mutex_lock(&debug_mutex); + debug_info = file_inode(file)->i_private; + /* find debug view */ + for (i = 0; i < DEBUG_MAX_VIEWS; i++) { + if (!debug_info->views[i]) + continue; + else if (debug_info->debugfs_entries[i] == file->f_path.dentry) + goto found; /* found view ! */ + } + /* no entry found */ + rc = -EINVAL; + goto out; + +found: + + /* Make snapshot of current debug areas to get it consistent. */ + /* To copy all the areas is only needed, if we have a view which */ + /* formats the debug areas. */ + + if (!debug_info->views[i]->format_proc && !debug_info->views[i]->header_proc) + debug_info_snapshot = debug_info_copy(debug_info, NO_AREAS); + else + debug_info_snapshot = debug_info_copy(debug_info, ALL_AREAS); + + if (!debug_info_snapshot) { + rc = -ENOMEM; + goto out; + } + p_info = kmalloc(sizeof(file_private_info_t), GFP_KERNEL); + if (!p_info) { + debug_info_free(debug_info_snapshot); + rc = -ENOMEM; + goto out; + } + p_info->offset = 0; + p_info->debug_info_snap = debug_info_snapshot; + p_info->debug_info_org = debug_info; + p_info->view = debug_info->views[i]; + p_info->act_area = 0; + p_info->act_page = 0; + p_info->act_entry = DEBUG_PROLOG_ENTRY; + p_info->act_entry_offset = 0; + file->private_data = p_info; + debug_info_get(debug_info); + nonseekable_open(inode, file); +out: + mutex_unlock(&debug_mutex); + return rc; +} + +/* + * debug_close: + * - called for user close() + * - deletes private_data area of the file handle + */ +static int debug_close(struct inode *inode, struct file *file) +{ + file_private_info_t *p_info; + + p_info = (file_private_info_t *) file->private_data; + if (p_info->debug_info_snap) + debug_info_free(p_info->debug_info_snap); + debug_info_put(p_info->debug_info_org); + kfree(file->private_data); + return 0; /* success */ +} + +/* Create debugfs entries and add to internal list. */ +static void _debug_register(debug_info_t *id) +{ + /* create root directory */ + id->debugfs_root_entry = debugfs_create_dir(id->name, + debug_debugfs_root_entry); + + /* append new element to linked list */ + if (!debug_area_first) { + /* first element in list */ + debug_area_first = id; + id->prev = NULL; + } else { + /* append element to end of list */ + debug_area_last->next = id; + id->prev = debug_area_last; + } + debug_area_last = id; + id->next = NULL; + + debug_register_view(id, &debug_level_view); + debug_register_view(id, &debug_flush_view); + debug_register_view(id, &debug_pages_view); +} + +/** + * debug_register_mode() - creates and initializes debug area. + * + * @name: Name of debug log (e.g. used for debugfs entry) + * @pages_per_area: Number of pages, which will be allocated per area + * @nr_areas: Number of debug areas + * @buf_size: Size of data area in each debug entry + * @mode: File mode for debugfs files. E.g. S_IRWXUGO + * @uid: User ID for debugfs files. Currently only 0 is supported. + * @gid: Group ID for debugfs files. Currently only 0 is supported. + * + * Return: + * - Handle for generated debug area + * - %NULL if register failed + * + * Allocates memory for a debug log. + * Must not be called within an interrupt handler. + */ +debug_info_t *debug_register_mode(const char *name, int pages_per_area, + int nr_areas, int buf_size, umode_t mode, + uid_t uid, gid_t gid) +{ + debug_info_t *rc = NULL; + + /* Since debugfs currently does not support uid/gid other than root, */ + /* we do not allow gid/uid != 0 until we get support for that. */ + if ((uid != 0) || (gid != 0)) + pr_warn("Root becomes the owner of all s390dbf files in sysfs\n"); + BUG_ON(!initialized); + + /* create new debug_info */ + rc = debug_info_create(name, pages_per_area, nr_areas, buf_size, mode); + if (rc) { + mutex_lock(&debug_mutex); + _debug_register(rc); + mutex_unlock(&debug_mutex); + } else { + pr_err("Registering debug feature %s failed\n", name); + } + return rc; +} +EXPORT_SYMBOL(debug_register_mode); + +/** + * debug_register() - creates and initializes debug area with default file mode. + * + * @name: Name of debug log (e.g. used for debugfs entry) + * @pages_per_area: Number of pages, which will be allocated per area + * @nr_areas: Number of debug areas + * @buf_size: Size of data area in each debug entry + * + * Return: + * - Handle for generated debug area + * - %NULL if register failed + * + * Allocates memory for a debug log. + * The debugfs file mode access permissions are read and write for user. + * Must not be called within an interrupt handler. + */ +debug_info_t *debug_register(const char *name, int pages_per_area, + int nr_areas, int buf_size) +{ + return debug_register_mode(name, pages_per_area, nr_areas, buf_size, + S_IRUSR | S_IWUSR, 0, 0); +} +EXPORT_SYMBOL(debug_register); + +/** + * debug_register_static() - registers a static debug area + * + * @id: Handle for static debug area + * @pages_per_area: Number of pages per area + * @nr_areas: Number of debug areas + * + * Register debug_info_t defined using DEFINE_STATIC_DEBUG_INFO. + * + * Note: This function is called automatically via an initcall generated by + * DEFINE_STATIC_DEBUG_INFO. + */ +void debug_register_static(debug_info_t *id, int pages_per_area, int nr_areas) +{ + unsigned long flags; + debug_info_t *copy; + + if (!initialized) { + pr_err("Tried to register debug feature %s too early\n", + id->name); + return; + } + + copy = debug_info_alloc("", pages_per_area, nr_areas, id->buf_size, + id->level, ALL_AREAS); + if (!copy) { + pr_err("Registering debug feature %s failed\n", id->name); + + /* Clear pointers to prevent tracing into released initdata. */ + spin_lock_irqsave(&id->lock, flags); + id->areas = NULL; + id->active_pages = NULL; + id->active_entries = NULL; + spin_unlock_irqrestore(&id->lock, flags); + + return; + } + + /* Replace static trace area with dynamic copy. */ + spin_lock_irqsave(&id->lock, flags); + debug_events_append(copy, id); + debug_areas_swap(id, copy); + spin_unlock_irqrestore(&id->lock, flags); + + /* Clear pointers to initdata and discard copy. */ + copy->areas = NULL; + copy->active_pages = NULL; + copy->active_entries = NULL; + debug_info_free(copy); + + mutex_lock(&debug_mutex); + _debug_register(id); + mutex_unlock(&debug_mutex); +} + +/* Remove debugfs entries and remove from internal list. */ +static void _debug_unregister(debug_info_t *id) +{ + int i; + + for (i = 0; i < DEBUG_MAX_VIEWS; i++) { + if (!id->views[i]) + continue; + debugfs_remove(id->debugfs_entries[i]); + } + debugfs_remove(id->debugfs_root_entry); + if (id == debug_area_first) + debug_area_first = id->next; + if (id == debug_area_last) + debug_area_last = id->prev; + if (id->prev) + id->prev->next = id->next; + if (id->next) + id->next->prev = id->prev; +} + +/** + * debug_unregister() - give back debug area. + * + * @id: handle for debug log + * + * Return: + * none + */ +void debug_unregister(debug_info_t *id) +{ + if (!id) + return; + mutex_lock(&debug_mutex); + _debug_unregister(id); + mutex_unlock(&debug_mutex); + + debug_info_put(id); +} +EXPORT_SYMBOL(debug_unregister); + +/* + * debug_set_size: + * - set area size (number of pages) and number of areas + */ +static int debug_set_size(debug_info_t *id, int nr_areas, int pages_per_area) +{ + debug_info_t *new_id; + unsigned long flags; + + if (!id || (nr_areas <= 0) || (pages_per_area < 0)) + return -EINVAL; + + new_id = debug_info_alloc("", pages_per_area, nr_areas, id->buf_size, + id->level, ALL_AREAS); + if (!new_id) { + pr_info("Allocating memory for %i pages failed\n", + pages_per_area); + return -ENOMEM; + } + + spin_lock_irqsave(&id->lock, flags); + debug_events_append(new_id, id); + debug_areas_swap(new_id, id); + debug_info_free(new_id); + spin_unlock_irqrestore(&id->lock, flags); + pr_info("%s: set new size (%i pages)\n", id->name, pages_per_area); + + return 0; +} + +/** + * debug_set_level() - Sets new actual debug level if new_level is valid. + * + * @id: handle for debug log + * @new_level: new debug level + * + * Return: + * none + */ +void debug_set_level(debug_info_t *id, int new_level) +{ + unsigned long flags; + + if (!id) + return; + + if (new_level == DEBUG_OFF_LEVEL) { + pr_info("%s: switched off\n", id->name); + } else if ((new_level > DEBUG_MAX_LEVEL) || (new_level < 0)) { + pr_info("%s: level %i is out of range (%i - %i)\n", + id->name, new_level, 0, DEBUG_MAX_LEVEL); + return; + } + + spin_lock_irqsave(&id->lock, flags); + id->level = new_level; + spin_unlock_irqrestore(&id->lock, flags); +} +EXPORT_SYMBOL(debug_set_level); + +/* + * proceed_active_entry: + * - set active entry to next in the ring buffer + */ +static inline void proceed_active_entry(debug_info_t *id) +{ + if ((id->active_entries[id->active_area] += id->entry_size) + > (PAGE_SIZE - id->entry_size)) { + id->active_entries[id->active_area] = 0; + id->active_pages[id->active_area] = + (id->active_pages[id->active_area] + 1) % + id->pages_per_area; + } +} + +/* + * proceed_active_area: + * - set active area to next in the ring buffer + */ +static inline void proceed_active_area(debug_info_t *id) +{ + id->active_area++; + id->active_area = id->active_area % id->nr_areas; +} + +/* + * get_active_entry: + */ +static inline debug_entry_t *get_active_entry(debug_info_t *id) +{ + return (debug_entry_t *) (((char *) id->areas[id->active_area] + [id->active_pages[id->active_area]]) + + id->active_entries[id->active_area]); +} + +/* Swap debug areas of a and b. */ +static void debug_areas_swap(debug_info_t *a, debug_info_t *b) +{ + swap(a->nr_areas, b->nr_areas); + swap(a->pages_per_area, b->pages_per_area); + swap(a->areas, b->areas); + swap(a->active_area, b->active_area); + swap(a->active_pages, b->active_pages); + swap(a->active_entries, b->active_entries); +} + +/* Append all debug events in active area from source to destination log. */ +static void debug_events_append(debug_info_t *dest, debug_info_t *src) +{ + debug_entry_t *from, *to, *last; + + if (!src->areas || !dest->areas) + return; + + /* Loop over all entries in src, starting with oldest. */ + from = get_active_entry(src); + last = from; + do { + if (from->clock != 0LL) { + to = get_active_entry(dest); + memset(to, 0, dest->entry_size); + memcpy(to, from, min(src->entry_size, + dest->entry_size)); + proceed_active_entry(dest); + } + + proceed_active_entry(src); + from = get_active_entry(src); + } while (from != last); +} + +/* + * debug_finish_entry: + * - set timestamp, caller address, cpu number etc. + */ + +static inline void debug_finish_entry(debug_info_t *id, debug_entry_t *active, + int level, int exception) +{ + unsigned long timestamp; + union tod_clock clk; + + store_tod_clock_ext(&clk); + timestamp = clk.us; + timestamp -= TOD_UNIX_EPOCH >> 12; + active->clock = timestamp; + active->cpu = smp_processor_id(); + active->caller = __builtin_return_address(0); + active->exception = exception; + active->level = level; + proceed_active_entry(id); + if (exception) + proceed_active_area(id); +} + +static int debug_stoppable = 1; +static int debug_active = 1; + +#define CTL_S390DBF_STOPPABLE 5678 +#define CTL_S390DBF_ACTIVE 5679 + +/* + * proc handler for the running debug_active sysctl + * always allow read, allow write only if debug_stoppable is set or + * if debug_active is already off + */ +static int s390dbf_procactive(struct ctl_table *table, int write, + void *buffer, size_t *lenp, loff_t *ppos) +{ + if (!write || debug_stoppable || !debug_active) + return proc_dointvec(table, write, buffer, lenp, ppos); + else + return 0; +} + +static struct ctl_table s390dbf_table[] = { + { + .procname = "debug_stoppable", + .data = &debug_stoppable, + .maxlen = sizeof(int), + .mode = S_IRUGO | S_IWUSR, + .proc_handler = proc_dointvec, + }, + { + .procname = "debug_active", + .data = &debug_active, + .maxlen = sizeof(int), + .mode = S_IRUGO | S_IWUSR, + .proc_handler = s390dbf_procactive, + }, + { } +}; + +static struct ctl_table_header *s390dbf_sysctl_header; + +/** + * debug_stop_all() - stops the debug feature if stopping is allowed. + * + * Return: + * - none + * + * Currently used in case of a kernel oops. + */ +void debug_stop_all(void) +{ + if (debug_stoppable) + debug_active = 0; +} +EXPORT_SYMBOL(debug_stop_all); + +/** + * debug_set_critical() - event/exception functions try lock instead of spin. + * + * Return: + * - none + * + * Currently used in case of stopping all CPUs but the current one. + * Once in this state, functions to write a debug entry for an + * event or exception no longer spin on the debug area lock, + * but only try to get it and fail if they do not get the lock. + */ +void debug_set_critical(void) +{ + debug_critical = 1; +} + +/* + * debug_event_common: + * - write debug entry with given size + */ +debug_entry_t *debug_event_common(debug_info_t *id, int level, const void *buf, + int len) +{ + debug_entry_t *active; + unsigned long flags; + + if (!debug_active || !id->areas) + return NULL; + if (debug_critical) { + if (!spin_trylock_irqsave(&id->lock, flags)) + return NULL; + } else { + spin_lock_irqsave(&id->lock, flags); + } + do { + active = get_active_entry(id); + memcpy(DEBUG_DATA(active), buf, min(len, id->buf_size)); + if (len < id->buf_size) + memset((DEBUG_DATA(active)) + len, 0, id->buf_size - len); + debug_finish_entry(id, active, level, 0); + len -= id->buf_size; + buf += id->buf_size; + } while (len > 0); + + spin_unlock_irqrestore(&id->lock, flags); + return active; +} +EXPORT_SYMBOL(debug_event_common); + +/* + * debug_exception_common: + * - write debug entry with given size and switch to next debug area + */ +debug_entry_t *debug_exception_common(debug_info_t *id, int level, + const void *buf, int len) +{ + debug_entry_t *active; + unsigned long flags; + + if (!debug_active || !id->areas) + return NULL; + if (debug_critical) { + if (!spin_trylock_irqsave(&id->lock, flags)) + return NULL; + } else { + spin_lock_irqsave(&id->lock, flags); + } + do { + active = get_active_entry(id); + memcpy(DEBUG_DATA(active), buf, min(len, id->buf_size)); + if (len < id->buf_size) + memset((DEBUG_DATA(active)) + len, 0, id->buf_size - len); + debug_finish_entry(id, active, level, len <= id->buf_size); + len -= id->buf_size; + buf += id->buf_size; + } while (len > 0); + + spin_unlock_irqrestore(&id->lock, flags); + return active; +} +EXPORT_SYMBOL(debug_exception_common); + +/* + * counts arguments in format string for sprintf view + */ +static inline int debug_count_numargs(char *string) +{ + int numargs = 0; + + while (*string) { + if (*string++ == '%') + numargs++; + } + return numargs; +} + +/* + * debug_sprintf_event: + */ +debug_entry_t *__debug_sprintf_event(debug_info_t *id, int level, char *string, ...) +{ + debug_sprintf_entry_t *curr_event; + debug_entry_t *active; + unsigned long flags; + int numargs, idx; + va_list ap; + + if (!debug_active || !id->areas) + return NULL; + numargs = debug_count_numargs(string); + + if (debug_critical) { + if (!spin_trylock_irqsave(&id->lock, flags)) + return NULL; + } else { + spin_lock_irqsave(&id->lock, flags); + } + active = get_active_entry(id); + curr_event = (debug_sprintf_entry_t *) DEBUG_DATA(active); + va_start(ap, string); + curr_event->string = string; + for (idx = 0; idx < min(numargs, (int)(id->buf_size / sizeof(long)) - 1); idx++) + curr_event->args[idx] = va_arg(ap, long); + va_end(ap); + debug_finish_entry(id, active, level, 0); + spin_unlock_irqrestore(&id->lock, flags); + + return active; +} +EXPORT_SYMBOL(__debug_sprintf_event); + +/* + * debug_sprintf_exception: + */ +debug_entry_t *__debug_sprintf_exception(debug_info_t *id, int level, char *string, ...) +{ + debug_sprintf_entry_t *curr_event; + debug_entry_t *active; + unsigned long flags; + int numargs, idx; + va_list ap; + + if (!debug_active || !id->areas) + return NULL; + + numargs = debug_count_numargs(string); + + if (debug_critical) { + if (!spin_trylock_irqsave(&id->lock, flags)) + return NULL; + } else { + spin_lock_irqsave(&id->lock, flags); + } + active = get_active_entry(id); + curr_event = (debug_sprintf_entry_t *)DEBUG_DATA(active); + va_start(ap, string); + curr_event->string = string; + for (idx = 0; idx < min(numargs, (int)(id->buf_size / sizeof(long)) - 1); idx++) + curr_event->args[idx] = va_arg(ap, long); + va_end(ap); + debug_finish_entry(id, active, level, 1); + spin_unlock_irqrestore(&id->lock, flags); + + return active; +} +EXPORT_SYMBOL(__debug_sprintf_exception); + +/** + * debug_register_view() - registers new debug view and creates debugfs + * dir entry + * + * @id: handle for debug log + * @view: pointer to debug view struct + * + * Return: + * - 0 : ok + * - < 0: Error + */ +int debug_register_view(debug_info_t *id, struct debug_view *view) +{ + unsigned long flags; + struct dentry *pde; + umode_t mode; + int rc = 0; + int i; + + if (!id) + goto out; + mode = (id->mode | S_IFREG) & ~S_IXUGO; + if (!(view->prolog_proc || view->format_proc || view->header_proc)) + mode &= ~(S_IRUSR | S_IRGRP | S_IROTH); + if (!view->input_proc) + mode &= ~(S_IWUSR | S_IWGRP | S_IWOTH); + pde = debugfs_create_file(view->name, mode, id->debugfs_root_entry, + id, &debug_file_ops); + spin_lock_irqsave(&id->lock, flags); + for (i = 0; i < DEBUG_MAX_VIEWS; i++) { + if (!id->views[i]) + break; + } + if (i == DEBUG_MAX_VIEWS) { + rc = -1; + } else { + id->views[i] = view; + id->debugfs_entries[i] = pde; + } + spin_unlock_irqrestore(&id->lock, flags); + if (rc) { + pr_err("Registering view %s/%s would exceed the maximum " + "number of views %i\n", id->name, view->name, i); + debugfs_remove(pde); + } +out: + return rc; +} +EXPORT_SYMBOL(debug_register_view); + +/** + * debug_unregister_view() - unregisters debug view and removes debugfs + * dir entry + * + * @id: handle for debug log + * @view: pointer to debug view struct + * + * Return: + * - 0 : ok + * - < 0: Error + */ +int debug_unregister_view(debug_info_t *id, struct debug_view *view) +{ + struct dentry *dentry = NULL; + unsigned long flags; + int i, rc = 0; + + if (!id) + goto out; + spin_lock_irqsave(&id->lock, flags); + for (i = 0; i < DEBUG_MAX_VIEWS; i++) { + if (id->views[i] == view) + break; + } + if (i == DEBUG_MAX_VIEWS) { + rc = -1; + } else { + dentry = id->debugfs_entries[i]; + id->views[i] = NULL; + id->debugfs_entries[i] = NULL; + } + spin_unlock_irqrestore(&id->lock, flags); + debugfs_remove(dentry); +out: + return rc; +} +EXPORT_SYMBOL(debug_unregister_view); + +static inline char *debug_get_user_string(const char __user *user_buf, + size_t user_len) +{ + char *buffer; + + buffer = kmalloc(user_len + 1, GFP_KERNEL); + if (!buffer) + return ERR_PTR(-ENOMEM); + if (copy_from_user(buffer, user_buf, user_len) != 0) { + kfree(buffer); + return ERR_PTR(-EFAULT); + } + /* got the string, now strip linefeed. */ + if (buffer[user_len - 1] == '\n') + buffer[user_len - 1] = 0; + else + buffer[user_len] = 0; + return buffer; +} + +static inline int debug_get_uint(char *buf) +{ + int rc; + + buf = skip_spaces(buf); + rc = simple_strtoul(buf, &buf, 10); + if (*buf) + rc = -EINVAL; + return rc; +} + +/* + * functions for debug-views + *********************************** +*/ + +/* + * prints out actual debug level + */ + +static int debug_prolog_pages_fn(debug_info_t *id, struct debug_view *view, + char *out_buf) +{ + return sprintf(out_buf, "%i\n", id->pages_per_area); +} + +/* + * reads new size (number of pages per debug area) + */ + +static int debug_input_pages_fn(debug_info_t *id, struct debug_view *view, + struct file *file, const char __user *user_buf, + size_t user_len, loff_t *offset) +{ + int rc, new_pages; + char *str; + + if (user_len > 0x10000) + user_len = 0x10000; + if (*offset != 0) { + rc = -EPIPE; + goto out; + } + str = debug_get_user_string(user_buf, user_len); + if (IS_ERR(str)) { + rc = PTR_ERR(str); + goto out; + } + new_pages = debug_get_uint(str); + if (new_pages < 0) { + rc = -EINVAL; + goto free_str; + } + rc = debug_set_size(id, id->nr_areas, new_pages); + if (rc != 0) { + rc = -EINVAL; + goto free_str; + } + rc = user_len; +free_str: + kfree(str); +out: + *offset += user_len; + return rc; /* number of input characters */ +} + +/* + * prints out actual debug level + */ +static int debug_prolog_level_fn(debug_info_t *id, struct debug_view *view, + char *out_buf) +{ + int rc = 0; + + if (id->level == DEBUG_OFF_LEVEL) + rc = sprintf(out_buf, "-\n"); + else + rc = sprintf(out_buf, "%i\n", id->level); + return rc; +} + +/* + * reads new debug level + */ +static int debug_input_level_fn(debug_info_t *id, struct debug_view *view, + struct file *file, const char __user *user_buf, + size_t user_len, loff_t *offset) +{ + int rc, new_level; + char *str; + + if (user_len > 0x10000) + user_len = 0x10000; + if (*offset != 0) { + rc = -EPIPE; + goto out; + } + str = debug_get_user_string(user_buf, user_len); + if (IS_ERR(str)) { + rc = PTR_ERR(str); + goto out; + } + if (str[0] == '-') { + debug_set_level(id, DEBUG_OFF_LEVEL); + rc = user_len; + goto free_str; + } else { + new_level = debug_get_uint(str); + } + if (new_level < 0) { + pr_warn("%s is not a valid level for a debug feature\n", str); + rc = -EINVAL; + } else { + debug_set_level(id, new_level); + rc = user_len; + } +free_str: + kfree(str); +out: + *offset += user_len; + return rc; /* number of input characters */ +} + +/* + * flushes debug areas + */ +static void debug_flush(debug_info_t *id, int area) +{ + unsigned long flags; + int i, j; + + if (!id || !id->areas) + return; + spin_lock_irqsave(&id->lock, flags); + if (area == DEBUG_FLUSH_ALL) { + id->active_area = 0; + memset(id->active_entries, 0, id->nr_areas * sizeof(int)); + for (i = 0; i < id->nr_areas; i++) { + id->active_pages[i] = 0; + for (j = 0; j < id->pages_per_area; j++) + memset(id->areas[i][j], 0, PAGE_SIZE); + } + } else if (area >= 0 && area < id->nr_areas) { + id->active_entries[area] = 0; + id->active_pages[area] = 0; + for (i = 0; i < id->pages_per_area; i++) + memset(id->areas[area][i], 0, PAGE_SIZE); + } + spin_unlock_irqrestore(&id->lock, flags); +} + +/* + * view function: flushes debug areas + */ +static int debug_input_flush_fn(debug_info_t *id, struct debug_view *view, + struct file *file, const char __user *user_buf, + size_t user_len, loff_t *offset) +{ + char input_buf[1]; + int rc = user_len; + + if (user_len > 0x10000) + user_len = 0x10000; + if (*offset != 0) { + rc = -EPIPE; + goto out; + } + if (copy_from_user(input_buf, user_buf, 1)) { + rc = -EFAULT; + goto out; + } + if (input_buf[0] == '-') { + debug_flush(id, DEBUG_FLUSH_ALL); + goto out; + } + if (isdigit(input_buf[0])) { + int area = ((int) input_buf[0] - (int) '0'); + + debug_flush(id, area); + goto out; + } + + pr_info("Flushing debug data failed because %c is not a valid " + "area\n", input_buf[0]); + +out: + *offset += user_len; + return rc; /* number of input characters */ +} + +/* + * prints debug data in hex/ascii format + */ +static int debug_hex_ascii_format_fn(debug_info_t *id, struct debug_view *view, + char *out_buf, const char *in_buf) +{ + int i, rc = 0; + + for (i = 0; i < id->buf_size; i++) + rc += sprintf(out_buf + rc, "%02x ", ((unsigned char *) in_buf)[i]); + rc += sprintf(out_buf + rc, "| "); + for (i = 0; i < id->buf_size; i++) { + unsigned char c = in_buf[i]; + + if (isascii(c) && isprint(c)) + rc += sprintf(out_buf + rc, "%c", c); + else + rc += sprintf(out_buf + rc, "."); + } + rc += sprintf(out_buf + rc, "\n"); + return rc; +} + +/* + * prints header for debug entry + */ +int debug_dflt_header_fn(debug_info_t *id, struct debug_view *view, + int area, debug_entry_t *entry, char *out_buf) +{ + unsigned long sec, usec; + unsigned long caller; + unsigned int level; + char *except_str; + int rc = 0; + + level = entry->level; + sec = entry->clock; + usec = do_div(sec, USEC_PER_SEC); + + if (entry->exception) + except_str = "*"; + else + except_str = "-"; + caller = (unsigned long) entry->caller; + rc += sprintf(out_buf, "%02i %011ld:%06lu %1u %1s %04u %px ", + area, sec, usec, level, except_str, + entry->cpu, (void *)caller); + return rc; +} +EXPORT_SYMBOL(debug_dflt_header_fn); + +/* + * prints debug data sprintf-formated: + * debug_sprinf_event/exception calls must be used together with this view + */ + +#define DEBUG_SPRINTF_MAX_ARGS 10 + +static int debug_sprintf_format_fn(debug_info_t *id, struct debug_view *view, + char *out_buf, const char *inbuf) +{ + debug_sprintf_entry_t *curr_event = (debug_sprintf_entry_t *)inbuf; + int num_longs, num_used_args = 0, i, rc = 0; + int index[DEBUG_SPRINTF_MAX_ARGS]; + + /* count of longs fit into one entry */ + num_longs = id->buf_size / sizeof(long); + + if (num_longs < 1) + goto out; /* bufsize of entry too small */ + if (num_longs == 1) { + /* no args, we use only the string */ + strcpy(out_buf, curr_event->string); + rc = strlen(curr_event->string); + goto out; + } + + /* number of arguments used for sprintf (without the format string) */ + num_used_args = min(DEBUG_SPRINTF_MAX_ARGS, (num_longs - 1)); + + memset(index, 0, DEBUG_SPRINTF_MAX_ARGS * sizeof(int)); + + for (i = 0; i < num_used_args; i++) + index[i] = i; + + rc = sprintf(out_buf, curr_event->string, curr_event->args[index[0]], + curr_event->args[index[1]], curr_event->args[index[2]], + curr_event->args[index[3]], curr_event->args[index[4]], + curr_event->args[index[5]], curr_event->args[index[6]], + curr_event->args[index[7]], curr_event->args[index[8]], + curr_event->args[index[9]]); +out: + return rc; +} + +/* + * debug_init: + * - is called exactly once to initialize the debug feature + */ +static int __init debug_init(void) +{ + s390dbf_sysctl_header = register_sysctl("s390dbf", s390dbf_table); + mutex_lock(&debug_mutex); + debug_debugfs_root_entry = debugfs_create_dir(DEBUG_DIR_ROOT, NULL); + initialized = 1; + mutex_unlock(&debug_mutex); + return 0; +} +postcore_initcall(debug_init); diff --git a/arch/s390/kernel/diag.c b/arch/s390/kernel/diag.c new file mode 100644 index 0000000000..f9f06cd8fc --- /dev/null +++ b/arch/s390/kernel/diag.c @@ -0,0 +1,269 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Implementation of s390 diagnose codes + * + * Copyright IBM Corp. 2007 + * Author(s): Michael Holzheu <holzheu@de.ibm.com> + */ + +#include <linux/export.h> +#include <linux/init.h> +#include <linux/cpu.h> +#include <linux/seq_file.h> +#include <linux/debugfs.h> +#include <linux/vmalloc.h> +#include <asm/asm-extable.h> +#include <asm/diag.h> +#include <asm/trace/diag.h> +#include <asm/sections.h> +#include "entry.h" + +struct diag_stat { + unsigned int counter[NR_DIAG_STAT]; +}; + +static DEFINE_PER_CPU(struct diag_stat, diag_stat); + +struct diag_desc { + int code; + char *name; +}; + +static const struct diag_desc diag_map[NR_DIAG_STAT] = { + [DIAG_STAT_X008] = { .code = 0x008, .name = "Console Function" }, + [DIAG_STAT_X00C] = { .code = 0x00c, .name = "Pseudo Timer" }, + [DIAG_STAT_X010] = { .code = 0x010, .name = "Release Pages" }, + [DIAG_STAT_X014] = { .code = 0x014, .name = "Spool File Services" }, + [DIAG_STAT_X044] = { .code = 0x044, .name = "Voluntary Timeslice End" }, + [DIAG_STAT_X064] = { .code = 0x064, .name = "NSS Manipulation" }, + [DIAG_STAT_X08C] = { .code = 0x08c, .name = "Access 3270 Display Device Information" }, + [DIAG_STAT_X09C] = { .code = 0x09c, .name = "Relinquish Timeslice" }, + [DIAG_STAT_X0DC] = { .code = 0x0dc, .name = "Appldata Control" }, + [DIAG_STAT_X204] = { .code = 0x204, .name = "Logical-CPU Utilization" }, + [DIAG_STAT_X210] = { .code = 0x210, .name = "Device Information" }, + [DIAG_STAT_X224] = { .code = 0x224, .name = "EBCDIC-Name Table" }, + [DIAG_STAT_X250] = { .code = 0x250, .name = "Block I/O" }, + [DIAG_STAT_X258] = { .code = 0x258, .name = "Page-Reference Services" }, + [DIAG_STAT_X26C] = { .code = 0x26c, .name = "Certain System Information" }, + [DIAG_STAT_X288] = { .code = 0x288, .name = "Time Bomb" }, + [DIAG_STAT_X2C4] = { .code = 0x2c4, .name = "FTP Services" }, + [DIAG_STAT_X2FC] = { .code = 0x2fc, .name = "Guest Performance Data" }, + [DIAG_STAT_X304] = { .code = 0x304, .name = "Partition-Resource Service" }, + [DIAG_STAT_X308] = { .code = 0x308, .name = "List-Directed IPL" }, + [DIAG_STAT_X318] = { .code = 0x318, .name = "CP Name and Version Codes" }, + [DIAG_STAT_X320] = { .code = 0x320, .name = "Certificate Store" }, + [DIAG_STAT_X500] = { .code = 0x500, .name = "Virtio Service" }, +}; + +struct diag_ops __amode31_ref diag_amode31_ops = { + .diag210 = _diag210_amode31, + .diag26c = _diag26c_amode31, + .diag14 = _diag14_amode31, + .diag0c = _diag0c_amode31, + .diag8c = _diag8c_amode31, + .diag308_reset = _diag308_reset_amode31 +}; + +static struct diag210 _diag210_tmp_amode31 __section(".amode31.data"); +struct diag210 __amode31_ref *__diag210_tmp_amode31 = &_diag210_tmp_amode31; + +static struct diag8c _diag8c_tmp_amode31 __section(".amode31.data"); +static struct diag8c __amode31_ref *__diag8c_tmp_amode31 = &_diag8c_tmp_amode31; + +static int show_diag_stat(struct seq_file *m, void *v) +{ + struct diag_stat *stat; + unsigned long n = (unsigned long) v - 1; + int cpu, prec, tmp; + + cpus_read_lock(); + if (n == 0) { + seq_puts(m, " "); + + for_each_online_cpu(cpu) { + prec = 10; + for (tmp = 10; cpu >= tmp; tmp *= 10) + prec--; + seq_printf(m, "%*s%d", prec, "CPU", cpu); + } + seq_putc(m, '\n'); + } else if (n <= NR_DIAG_STAT) { + seq_printf(m, "diag %03x:", diag_map[n-1].code); + for_each_online_cpu(cpu) { + stat = &per_cpu(diag_stat, cpu); + seq_printf(m, " %10u", stat->counter[n-1]); + } + seq_printf(m, " %s\n", diag_map[n-1].name); + } + cpus_read_unlock(); + return 0; +} + +static void *show_diag_stat_start(struct seq_file *m, loff_t *pos) +{ + return *pos <= NR_DIAG_STAT ? (void *)((unsigned long) *pos + 1) : NULL; +} + +static void *show_diag_stat_next(struct seq_file *m, void *v, loff_t *pos) +{ + ++*pos; + return show_diag_stat_start(m, pos); +} + +static void show_diag_stat_stop(struct seq_file *m, void *v) +{ +} + +static const struct seq_operations show_diag_stat_sops = { + .start = show_diag_stat_start, + .next = show_diag_stat_next, + .stop = show_diag_stat_stop, + .show = show_diag_stat, +}; + +DEFINE_SEQ_ATTRIBUTE(show_diag_stat); + +static int __init show_diag_stat_init(void) +{ + debugfs_create_file("diag_stat", 0400, NULL, NULL, + &show_diag_stat_fops); + return 0; +} + +device_initcall(show_diag_stat_init); + +void diag_stat_inc(enum diag_stat_enum nr) +{ + this_cpu_inc(diag_stat.counter[nr]); + trace_s390_diagnose(diag_map[nr].code); +} +EXPORT_SYMBOL(diag_stat_inc); + +void notrace diag_stat_inc_norecursion(enum diag_stat_enum nr) +{ + this_cpu_inc(diag_stat.counter[nr]); + trace_s390_diagnose_norecursion(diag_map[nr].code); +} +EXPORT_SYMBOL(diag_stat_inc_norecursion); + +/* + * Diagnose 14: Input spool file manipulation + */ +int diag14(unsigned long rx, unsigned long ry1, unsigned long subcode) +{ + diag_stat_inc(DIAG_STAT_X014); + return diag_amode31_ops.diag14(rx, ry1, subcode); +} +EXPORT_SYMBOL(diag14); + +static inline int __diag204(unsigned long *subcode, unsigned long size, void *addr) +{ + union register_pair rp = { .even = *subcode, .odd = size }; + + asm volatile( + " diag %[addr],%[rp],0x204\n" + "0: nopr %%r7\n" + EX_TABLE(0b,0b) + : [rp] "+&d" (rp.pair) : [addr] "d" (addr) : "memory"); + *subcode = rp.even; + return rp.odd; +} + +/** + * diag204() - Issue diagnose 204 call. + * @subcode: Subcode of diagnose 204 to be executed. + * @size: Size of area in pages which @area points to, if given. + * @addr: Vmalloc'ed memory area where the result is written to. + * + * Execute diagnose 204 with the given subcode and write the result to the + * memory area specified with @addr. For subcodes which do not write a + * result to memory both @size and @addr must be zero. If @addr is + * specified it must be page aligned and must have been allocated with + * vmalloc(). Conversion to real / physical addresses will be handled by + * this function if required. + */ +int diag204(unsigned long subcode, unsigned long size, void *addr) +{ + if (addr) { + if (WARN_ON_ONCE(!is_vmalloc_addr(addr))) + return -1; + if (WARN_ON_ONCE(!IS_ALIGNED((unsigned long)addr, PAGE_SIZE))) + return -1; + } + if ((subcode & DIAG204_SUBCODE_MASK) == DIAG204_SUBC_STIB4) + addr = (void *)pfn_to_phys(vmalloc_to_pfn(addr)); + diag_stat_inc(DIAG_STAT_X204); + size = __diag204(&subcode, size, addr); + if (subcode) + return -1; + return size; +} +EXPORT_SYMBOL(diag204); + +/* + * Diagnose 210: Get information about a virtual device + */ +int diag210(struct diag210 *addr) +{ + static DEFINE_SPINLOCK(diag210_lock); + unsigned long flags; + int ccode; + + spin_lock_irqsave(&diag210_lock, flags); + *__diag210_tmp_amode31 = *addr; + + diag_stat_inc(DIAG_STAT_X210); + ccode = diag_amode31_ops.diag210(__diag210_tmp_amode31); + + *addr = *__diag210_tmp_amode31; + spin_unlock_irqrestore(&diag210_lock, flags); + + return ccode; +} +EXPORT_SYMBOL(diag210); + +/* + * Diagnose 8C: Access 3270 Display Device Information + */ +int diag8c(struct diag8c *addr, struct ccw_dev_id *devno) +{ + static DEFINE_SPINLOCK(diag8c_lock); + unsigned long flags; + int ccode; + + spin_lock_irqsave(&diag8c_lock, flags); + + diag_stat_inc(DIAG_STAT_X08C); + ccode = diag_amode31_ops.diag8c(__diag8c_tmp_amode31, devno, sizeof(*addr)); + + *addr = *__diag8c_tmp_amode31; + spin_unlock_irqrestore(&diag8c_lock, flags); + + return ccode; +} +EXPORT_SYMBOL(diag8c); + +int diag224(void *ptr) +{ + int rc = -EOPNOTSUPP; + + diag_stat_inc(DIAG_STAT_X224); + asm volatile( + " diag %1,%2,0x224\n" + "0: lhi %0,0x0\n" + "1:\n" + EX_TABLE(0b,1b) + : "+d" (rc) :"d" (0), "d" (ptr) : "memory"); + return rc; +} +EXPORT_SYMBOL(diag224); + +/* + * Diagnose 26C: Access Certain System Information + */ +int diag26c(void *req, void *resp, enum diag26c_sc subcode) +{ + diag_stat_inc(DIAG_STAT_X26C); + return diag_amode31_ops.diag26c(req, resp, subcode); +} +EXPORT_SYMBOL(diag26c); diff --git a/arch/s390/kernel/dis.c b/arch/s390/kernel/dis.c new file mode 100644 index 0000000000..89dc826a8d --- /dev/null +++ b/arch/s390/kernel/dis.c @@ -0,0 +1,590 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Disassemble s390 instructions. + * + * Copyright IBM Corp. 2007 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), + */ + +#include <linux/sched.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <linux/errno.h> +#include <linux/ptrace.h> +#include <linux/timer.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/init.h> +#include <linux/interrupt.h> +#include <linux/delay.h> +#include <linux/export.h> +#include <linux/kallsyms.h> +#include <linux/reboot.h> +#include <linux/kprobes.h> +#include <linux/kdebug.h> +#include <linux/uaccess.h> +#include <linux/atomic.h> +#include <linux/io.h> +#include <asm/dis.h> +#include <asm/cpcmd.h> +#include <asm/lowcore.h> +#include <asm/debug.h> +#include <asm/irq.h> + +/* Type of operand */ +#define OPERAND_GPR 0x1 /* Operand printed as %rx */ +#define OPERAND_FPR 0x2 /* Operand printed as %fx */ +#define OPERAND_AR 0x4 /* Operand printed as %ax */ +#define OPERAND_CR 0x8 /* Operand printed as %cx */ +#define OPERAND_VR 0x10 /* Operand printed as %vx */ +#define OPERAND_DISP 0x20 /* Operand printed as displacement */ +#define OPERAND_BASE 0x40 /* Operand printed as base register */ +#define OPERAND_INDEX 0x80 /* Operand printed as index register */ +#define OPERAND_PCREL 0x100 /* Operand printed as pc-relative symbol */ +#define OPERAND_SIGNED 0x200 /* Operand printed as signed value */ +#define OPERAND_LENGTH 0x400 /* Operand printed as length (+1) */ + +struct s390_operand { + unsigned char bits; /* The number of bits in the operand. */ + unsigned char shift; /* The number of bits to shift. */ + unsigned short flags; /* One bit syntax flags. */ +}; + +struct s390_insn { + union { + const char name[5]; + struct { + unsigned char zero; + unsigned int offset; + } __packed; + }; + unsigned char opfrag; + unsigned char format; +}; + +struct s390_opcode_offset { + unsigned char opcode; + unsigned char mask; + unsigned char byte; + unsigned short offset; + unsigned short count; +} __packed; + +enum { + UNUSED, + A_8, /* Access reg. starting at position 8 */ + A_12, /* Access reg. starting at position 12 */ + A_24, /* Access reg. starting at position 24 */ + A_28, /* Access reg. starting at position 28 */ + B_16, /* Base register starting at position 16 */ + B_32, /* Base register starting at position 32 */ + C_8, /* Control reg. starting at position 8 */ + C_12, /* Control reg. starting at position 12 */ + D20_20, /* 20 bit displacement starting at 20 */ + D_20, /* Displacement starting at position 20 */ + D_36, /* Displacement starting at position 36 */ + F_8, /* FPR starting at position 8 */ + F_12, /* FPR starting at position 12 */ + F_16, /* FPR starting at position 16 */ + F_24, /* FPR starting at position 24 */ + F_28, /* FPR starting at position 28 */ + F_32, /* FPR starting at position 32 */ + I8_8, /* 8 bit signed value starting at 8 */ + I8_32, /* 8 bit signed value starting at 32 */ + I16_16, /* 16 bit signed value starting at 16 */ + I16_32, /* 16 bit signed value starting at 32 */ + I32_16, /* 32 bit signed value starting at 16 */ + J12_12, /* 12 bit PC relative offset at 12 */ + J16_16, /* 16 bit PC relative offset at 16 */ + J16_32, /* 16 bit PC relative offset at 32 */ + J24_24, /* 24 bit PC relative offset at 24 */ + J32_16, /* 32 bit PC relative offset at 16 */ + L4_8, /* 4 bit length starting at position 8 */ + L4_12, /* 4 bit length starting at position 12 */ + L8_8, /* 8 bit length starting at position 8 */ + R_8, /* GPR starting at position 8 */ + R_12, /* GPR starting at position 12 */ + R_16, /* GPR starting at position 16 */ + R_24, /* GPR starting at position 24 */ + R_28, /* GPR starting at position 28 */ + U4_8, /* 4 bit unsigned value starting at 8 */ + U4_12, /* 4 bit unsigned value starting at 12 */ + U4_16, /* 4 bit unsigned value starting at 16 */ + U4_20, /* 4 bit unsigned value starting at 20 */ + U4_24, /* 4 bit unsigned value starting at 24 */ + U4_28, /* 4 bit unsigned value starting at 28 */ + U4_32, /* 4 bit unsigned value starting at 32 */ + U4_36, /* 4 bit unsigned value starting at 36 */ + U8_8, /* 8 bit unsigned value starting at 8 */ + U8_16, /* 8 bit unsigned value starting at 16 */ + U8_24, /* 8 bit unsigned value starting at 24 */ + U8_28, /* 8 bit unsigned value starting at 28 */ + U8_32, /* 8 bit unsigned value starting at 32 */ + U12_16, /* 12 bit unsigned value starting at 16 */ + U16_16, /* 16 bit unsigned value starting at 16 */ + U16_32, /* 16 bit unsigned value starting at 32 */ + U32_16, /* 32 bit unsigned value starting at 16 */ + VX_12, /* Vector index register starting at position 12 */ + V_8, /* Vector reg. starting at position 8 */ + V_12, /* Vector reg. starting at position 12 */ + V_16, /* Vector reg. starting at position 16 */ + V_32, /* Vector reg. starting at position 32 */ + X_12, /* Index register starting at position 12 */ +}; + +static const struct s390_operand operands[] = { + [UNUSED] = { 0, 0, 0 }, + [A_8] = { 4, 8, OPERAND_AR }, + [A_12] = { 4, 12, OPERAND_AR }, + [A_24] = { 4, 24, OPERAND_AR }, + [A_28] = { 4, 28, OPERAND_AR }, + [B_16] = { 4, 16, OPERAND_BASE | OPERAND_GPR }, + [B_32] = { 4, 32, OPERAND_BASE | OPERAND_GPR }, + [C_8] = { 4, 8, OPERAND_CR }, + [C_12] = { 4, 12, OPERAND_CR }, + [D20_20] = { 20, 20, OPERAND_DISP | OPERAND_SIGNED }, + [D_20] = { 12, 20, OPERAND_DISP }, + [D_36] = { 12, 36, OPERAND_DISP }, + [F_8] = { 4, 8, OPERAND_FPR }, + [F_12] = { 4, 12, OPERAND_FPR }, + [F_16] = { 4, 16, OPERAND_FPR }, + [F_24] = { 4, 24, OPERAND_FPR }, + [F_28] = { 4, 28, OPERAND_FPR }, + [F_32] = { 4, 32, OPERAND_FPR }, + [I8_8] = { 8, 8, OPERAND_SIGNED }, + [I8_32] = { 8, 32, OPERAND_SIGNED }, + [I16_16] = { 16, 16, OPERAND_SIGNED }, + [I16_32] = { 16, 32, OPERAND_SIGNED }, + [I32_16] = { 32, 16, OPERAND_SIGNED }, + [J12_12] = { 12, 12, OPERAND_PCREL }, + [J16_16] = { 16, 16, OPERAND_PCREL }, + [J16_32] = { 16, 32, OPERAND_PCREL }, + [J24_24] = { 24, 24, OPERAND_PCREL }, + [J32_16] = { 32, 16, OPERAND_PCREL }, + [L4_8] = { 4, 8, OPERAND_LENGTH }, + [L4_12] = { 4, 12, OPERAND_LENGTH }, + [L8_8] = { 8, 8, OPERAND_LENGTH }, + [R_8] = { 4, 8, OPERAND_GPR }, + [R_12] = { 4, 12, OPERAND_GPR }, + [R_16] = { 4, 16, OPERAND_GPR }, + [R_24] = { 4, 24, OPERAND_GPR }, + [R_28] = { 4, 28, OPERAND_GPR }, + [U4_8] = { 4, 8, 0 }, + [U4_12] = { 4, 12, 0 }, + [U4_16] = { 4, 16, 0 }, + [U4_20] = { 4, 20, 0 }, + [U4_24] = { 4, 24, 0 }, + [U4_28] = { 4, 28, 0 }, + [U4_32] = { 4, 32, 0 }, + [U4_36] = { 4, 36, 0 }, + [U8_8] = { 8, 8, 0 }, + [U8_16] = { 8, 16, 0 }, + [U8_24] = { 8, 24, 0 }, + [U8_28] = { 8, 28, 0 }, + [U8_32] = { 8, 32, 0 }, + [U12_16] = { 12, 16, 0 }, + [U16_16] = { 16, 16, 0 }, + [U16_32] = { 16, 32, 0 }, + [U32_16] = { 32, 16, 0 }, + [VX_12] = { 4, 12, OPERAND_INDEX | OPERAND_VR }, + [V_8] = { 4, 8, OPERAND_VR }, + [V_12] = { 4, 12, OPERAND_VR }, + [V_16] = { 4, 16, OPERAND_VR }, + [V_32] = { 4, 32, OPERAND_VR }, + [X_12] = { 4, 12, OPERAND_INDEX | OPERAND_GPR }, +}; + +static const unsigned char formats[][6] = { + [INSTR_E] = { 0, 0, 0, 0, 0, 0 }, + [INSTR_IE_UU] = { U4_24, U4_28, 0, 0, 0, 0 }, + [INSTR_MII_UPP] = { U4_8, J12_12, J24_24 }, + [INSTR_RIE_R0IU] = { R_8, I16_16, U4_32, 0, 0, 0 }, + [INSTR_RIE_R0UU] = { R_8, U16_16, U4_32, 0, 0, 0 }, + [INSTR_RIE_RRI0] = { R_8, R_12, I16_16, 0, 0, 0 }, + [INSTR_RIE_RRP] = { R_8, R_12, J16_16, 0, 0, 0 }, + [INSTR_RIE_RRPU] = { R_8, R_12, U4_32, J16_16, 0, 0 }, + [INSTR_RIE_RRUUU] = { R_8, R_12, U8_16, U8_24, U8_32, 0 }, + [INSTR_RIE_RUI0] = { R_8, I16_16, U4_12, 0, 0, 0 }, + [INSTR_RIE_RUPI] = { R_8, I8_32, U4_12, J16_16, 0, 0 }, + [INSTR_RIE_RUPU] = { R_8, U8_32, U4_12, J16_16, 0, 0 }, + [INSTR_RIL_RI] = { R_8, I32_16, 0, 0, 0, 0 }, + [INSTR_RIL_RP] = { R_8, J32_16, 0, 0, 0, 0 }, + [INSTR_RIL_RU] = { R_8, U32_16, 0, 0, 0, 0 }, + [INSTR_RIL_UP] = { U4_8, J32_16, 0, 0, 0, 0 }, + [INSTR_RIS_RURDI] = { R_8, I8_32, U4_12, D_20, B_16, 0 }, + [INSTR_RIS_RURDU] = { R_8, U8_32, U4_12, D_20, B_16, 0 }, + [INSTR_RI_RI] = { R_8, I16_16, 0, 0, 0, 0 }, + [INSTR_RI_RP] = { R_8, J16_16, 0, 0, 0, 0 }, + [INSTR_RI_RU] = { R_8, U16_16, 0, 0, 0, 0 }, + [INSTR_RI_UP] = { U4_8, J16_16, 0, 0, 0, 0 }, + [INSTR_RRE_00] = { 0, 0, 0, 0, 0, 0 }, + [INSTR_RRE_AA] = { A_24, A_28, 0, 0, 0, 0 }, + [INSTR_RRE_AR] = { A_24, R_28, 0, 0, 0, 0 }, + [INSTR_RRE_F0] = { F_24, 0, 0, 0, 0, 0 }, + [INSTR_RRE_FF] = { F_24, F_28, 0, 0, 0, 0 }, + [INSTR_RRE_FR] = { F_24, R_28, 0, 0, 0, 0 }, + [INSTR_RRE_R0] = { R_24, 0, 0, 0, 0, 0 }, + [INSTR_RRE_RA] = { R_24, A_28, 0, 0, 0, 0 }, + [INSTR_RRE_RF] = { R_24, F_28, 0, 0, 0, 0 }, + [INSTR_RRE_RR] = { R_24, R_28, 0, 0, 0, 0 }, + [INSTR_RRF_0UFF] = { F_24, F_28, U4_20, 0, 0, 0 }, + [INSTR_RRF_0URF] = { R_24, F_28, U4_20, 0, 0, 0 }, + [INSTR_RRF_F0FF] = { F_16, F_24, F_28, 0, 0, 0 }, + [INSTR_RRF_F0FF2] = { F_24, F_16, F_28, 0, 0, 0 }, + [INSTR_RRF_F0FR] = { F_24, F_16, R_28, 0, 0, 0 }, + [INSTR_RRF_FFRU] = { F_24, F_16, R_28, U4_20, 0, 0 }, + [INSTR_RRF_FUFF] = { F_24, F_16, F_28, U4_20, 0, 0 }, + [INSTR_RRF_FUFF2] = { F_24, F_28, F_16, U4_20, 0, 0 }, + [INSTR_RRF_R0RR] = { R_24, R_16, R_28, 0, 0, 0 }, + [INSTR_RRF_R0RR2] = { R_24, R_28, R_16, 0, 0, 0 }, + [INSTR_RRF_RURR] = { R_24, R_28, R_16, U4_20, 0, 0 }, + [INSTR_RRF_RURR2] = { R_24, R_16, R_28, U4_20, 0, 0 }, + [INSTR_RRF_U0FF] = { F_24, U4_16, F_28, 0, 0, 0 }, + [INSTR_RRF_U0RF] = { R_24, U4_16, F_28, 0, 0, 0 }, + [INSTR_RRF_U0RR] = { R_24, R_28, U4_16, 0, 0, 0 }, + [INSTR_RRF_URR] = { R_24, R_28, U8_16, 0, 0, 0 }, + [INSTR_RRF_UUFF] = { F_24, U4_16, F_28, U4_20, 0, 0 }, + [INSTR_RRF_UUFR] = { F_24, U4_16, R_28, U4_20, 0, 0 }, + [INSTR_RRF_UURF] = { R_24, U4_16, F_28, U4_20, 0, 0 }, + [INSTR_RRS_RRRDU] = { R_8, R_12, U4_32, D_20, B_16 }, + [INSTR_RR_FF] = { F_8, F_12, 0, 0, 0, 0 }, + [INSTR_RR_R0] = { R_8, 0, 0, 0, 0, 0 }, + [INSTR_RR_RR] = { R_8, R_12, 0, 0, 0, 0 }, + [INSTR_RR_U0] = { U8_8, 0, 0, 0, 0, 0 }, + [INSTR_RR_UR] = { U4_8, R_12, 0, 0, 0, 0 }, + [INSTR_RSI_RRP] = { R_8, R_12, J16_16, 0, 0, 0 }, + [INSTR_RSL_LRDFU] = { F_32, D_20, L8_8, B_16, U4_36, 0 }, + [INSTR_RSL_R0RD] = { D_20, L4_8, B_16, 0, 0, 0 }, + [INSTR_RSY_AARD] = { A_8, A_12, D20_20, B_16, 0, 0 }, + [INSTR_RSY_CCRD] = { C_8, C_12, D20_20, B_16, 0, 0 }, + [INSTR_RSY_RDRU] = { R_8, D20_20, B_16, U4_12, 0, 0 }, + [INSTR_RSY_RRRD] = { R_8, R_12, D20_20, B_16, 0, 0 }, + [INSTR_RSY_RURD] = { R_8, U4_12, D20_20, B_16, 0, 0 }, + [INSTR_RSY_RURD2] = { R_8, D20_20, B_16, U4_12, 0, 0 }, + [INSTR_RS_AARD] = { A_8, A_12, D_20, B_16, 0, 0 }, + [INSTR_RS_CCRD] = { C_8, C_12, D_20, B_16, 0, 0 }, + [INSTR_RS_R0RD] = { R_8, D_20, B_16, 0, 0, 0 }, + [INSTR_RS_RRRD] = { R_8, R_12, D_20, B_16, 0, 0 }, + [INSTR_RS_RURD] = { R_8, U4_12, D_20, B_16, 0, 0 }, + [INSTR_RXE_FRRD] = { F_8, D_20, X_12, B_16, 0, 0 }, + [INSTR_RXE_RRRDU] = { R_8, D_20, X_12, B_16, U4_32, 0 }, + [INSTR_RXF_FRRDF] = { F_32, F_8, D_20, X_12, B_16, 0 }, + [INSTR_RXY_FRRD] = { F_8, D20_20, X_12, B_16, 0, 0 }, + [INSTR_RXY_RRRD] = { R_8, D20_20, X_12, B_16, 0, 0 }, + [INSTR_RXY_URRD] = { U4_8, D20_20, X_12, B_16, 0, 0 }, + [INSTR_RX_FRRD] = { F_8, D_20, X_12, B_16, 0, 0 }, + [INSTR_RX_RRRD] = { R_8, D_20, X_12, B_16, 0, 0 }, + [INSTR_RX_URRD] = { U4_8, D_20, X_12, B_16, 0, 0 }, + [INSTR_SIL_RDI] = { D_20, B_16, I16_32, 0, 0, 0 }, + [INSTR_SIL_RDU] = { D_20, B_16, U16_32, 0, 0, 0 }, + [INSTR_SIY_IRD] = { D20_20, B_16, I8_8, 0, 0, 0 }, + [INSTR_SIY_RD] = { D20_20, B_16, 0, 0, 0, 0 }, + [INSTR_SIY_URD] = { D20_20, B_16, U8_8, 0, 0, 0 }, + [INSTR_SI_RD] = { D_20, B_16, 0, 0, 0, 0 }, + [INSTR_SI_URD] = { D_20, B_16, U8_8, 0, 0, 0 }, + [INSTR_SMI_U0RDP] = { U4_8, J16_32, D_20, B_16, 0, 0 }, + [INSTR_SSE_RDRD] = { D_20, B_16, D_36, B_32, 0, 0 }, + [INSTR_SSF_RRDRD] = { D_20, B_16, D_36, B_32, R_8, 0 }, + [INSTR_SSF_RRDRD2] = { R_8, D_20, B_16, D_36, B_32, 0 }, + [INSTR_SS_L0RDRD] = { D_20, L8_8, B_16, D_36, B_32, 0 }, + [INSTR_SS_L2RDRD] = { D_20, B_16, D_36, L8_8, B_32, 0 }, + [INSTR_SS_LIRDRD] = { D_20, L4_8, B_16, D_36, B_32, U4_12 }, + [INSTR_SS_LLRDRD] = { D_20, L4_8, B_16, D_36, L4_12, B_32 }, + [INSTR_SS_RRRDRD] = { D_20, R_8, B_16, D_36, B_32, R_12 }, + [INSTR_SS_RRRDRD2] = { R_8, D_20, B_16, R_12, D_36, B_32 }, + [INSTR_SS_RRRDRD3] = { R_8, R_12, D_20, B_16, D_36, B_32 }, + [INSTR_S_00] = { 0, 0, 0, 0, 0, 0 }, + [INSTR_S_RD] = { D_20, B_16, 0, 0, 0, 0 }, + [INSTR_VRI_V0IU] = { V_8, I16_16, U4_32, 0, 0, 0 }, + [INSTR_VRI_V0U] = { V_8, U16_16, 0, 0, 0, 0 }, + [INSTR_VRI_V0UU2] = { V_8, U16_16, U4_32, 0, 0, 0 }, + [INSTR_VRI_V0UUU] = { V_8, U8_16, U8_24, U4_32, 0, 0 }, + [INSTR_VRI_VR0UU] = { V_8, R_12, U8_28, U4_24, 0, 0 }, + [INSTR_VRI_VVUU] = { V_8, V_12, U16_16, U4_32, 0, 0 }, + [INSTR_VRI_VVUUU] = { V_8, V_12, U12_16, U4_32, U4_28, 0 }, + [INSTR_VRI_VVUUU2] = { V_8, V_12, U8_28, U8_16, U4_24, 0 }, + [INSTR_VRI_VVV0U] = { V_8, V_12, V_16, U8_24, 0, 0 }, + [INSTR_VRI_VVV0UU] = { V_8, V_12, V_16, U8_24, U4_32, 0 }, + [INSTR_VRI_VVV0UU2] = { V_8, V_12, V_16, U8_28, U4_24, 0 }, + [INSTR_VRR_0V] = { V_12, 0, 0, 0, 0, 0 }, + [INSTR_VRR_0VV0U] = { V_12, V_16, U4_24, 0, 0, 0 }, + [INSTR_VRR_RV0UU] = { R_8, V_12, U4_24, U4_28, 0, 0 }, + [INSTR_VRR_VRR] = { V_8, R_12, R_16, 0, 0, 0 }, + [INSTR_VRR_VV] = { V_8, V_12, 0, 0, 0, 0 }, + [INSTR_VRR_VV0U] = { V_8, V_12, U4_32, 0, 0, 0 }, + [INSTR_VRR_VV0U0U] = { V_8, V_12, U4_32, U4_24, 0, 0 }, + [INSTR_VRR_VV0U2] = { V_8, V_12, U4_24, 0, 0, 0 }, + [INSTR_VRR_VV0UU2] = { V_8, V_12, U4_32, U4_28, 0, 0 }, + [INSTR_VRR_VV0UUU] = { V_8, V_12, U4_32, U4_28, U4_24, 0 }, + [INSTR_VRR_VVV] = { V_8, V_12, V_16, 0, 0, 0 }, + [INSTR_VRR_VVV0U] = { V_8, V_12, V_16, U4_32, 0, 0 }, + [INSTR_VRR_VVV0U0] = { V_8, V_12, V_16, U4_24, 0, 0 }, + [INSTR_VRR_VVV0U0U] = { V_8, V_12, V_16, U4_32, U4_24, 0 }, + [INSTR_VRR_VVV0UU] = { V_8, V_12, V_16, U4_32, U4_28, 0 }, + [INSTR_VRR_VVV0UUU] = { V_8, V_12, V_16, U4_32, U4_28, U4_24 }, + [INSTR_VRR_VVV0V] = { V_8, V_12, V_16, V_32, 0, 0 }, + [INSTR_VRR_VVVU0UV] = { V_8, V_12, V_16, V_32, U4_28, U4_20 }, + [INSTR_VRR_VVVU0V] = { V_8, V_12, V_16, V_32, U4_20, 0 }, + [INSTR_VRR_VVVUU0V] = { V_8, V_12, V_16, V_32, U4_20, U4_24 }, + [INSTR_VRS_RRDV] = { V_32, R_12, D_20, B_16, 0, 0 }, + [INSTR_VRS_RVRDU] = { R_8, V_12, D_20, B_16, U4_32, 0 }, + [INSTR_VRS_VRRD] = { V_8, R_12, D_20, B_16, 0, 0 }, + [INSTR_VRS_VRRDU] = { V_8, R_12, D_20, B_16, U4_32, 0 }, + [INSTR_VRS_VVRDU] = { V_8, V_12, D_20, B_16, U4_32, 0 }, + [INSTR_VRV_VVXRDU] = { V_8, D_20, VX_12, B_16, U4_32, 0 }, + [INSTR_VRX_VRRDU] = { V_8, D_20, X_12, B_16, U4_32, 0 }, + [INSTR_VRX_VV] = { V_8, V_12, 0, 0, 0, 0 }, + [INSTR_VSI_URDV] = { V_32, D_20, B_16, U8_8, 0, 0 }, +}; + +static char long_insn_name[][7] = LONG_INSN_INITIALIZER; +static struct s390_insn opcode[] = OPCODE_TABLE_INITIALIZER; +static struct s390_opcode_offset opcode_offset[] = OPCODE_OFFSET_INITIALIZER; + +/* Extracts an operand value from an instruction. */ +static unsigned int extract_operand(unsigned char *code, + const struct s390_operand *operand) +{ + unsigned char *cp; + unsigned int val; + int bits; + + /* Extract fragments of the operand byte for byte. */ + cp = code + operand->shift / 8; + bits = (operand->shift & 7) + operand->bits; + val = 0; + do { + val <<= 8; + val |= (unsigned int) *cp++; + bits -= 8; + } while (bits > 0); + val >>= -bits; + val &= ((1U << (operand->bits - 1)) << 1) - 1; + + /* Check for special long displacement case. */ + if (operand->bits == 20 && operand->shift == 20) + val = (val & 0xff) << 12 | (val & 0xfff00) >> 8; + + /* Check for register extensions bits for vector registers. */ + if (operand->flags & OPERAND_VR) { + if (operand->shift == 8) + val |= (code[4] & 8) << 1; + else if (operand->shift == 12) + val |= (code[4] & 4) << 2; + else if (operand->shift == 16) + val |= (code[4] & 2) << 3; + else if (operand->shift == 32) + val |= (code[4] & 1) << 4; + } + + /* Sign extend value if the operand is signed or pc relative. */ + if ((operand->flags & (OPERAND_SIGNED | OPERAND_PCREL)) && + (val & (1U << (operand->bits - 1)))) + val |= (-1U << (operand->bits - 1)) << 1; + + /* Double value if the operand is pc relative. */ + if (operand->flags & OPERAND_PCREL) + val <<= 1; + + /* Length x in an instructions has real length x + 1. */ + if (operand->flags & OPERAND_LENGTH) + val++; + return val; +} + +struct s390_insn *find_insn(unsigned char *code) +{ + struct s390_opcode_offset *entry; + struct s390_insn *insn; + unsigned char opfrag; + int i; + + /* Search the opcode offset table to find an entry which + * matches the beginning of the opcode. If there is no match + * the last entry will be used, which is the default entry for + * unknown instructions as well as 1-byte opcode instructions. + */ + for (i = 0; i < ARRAY_SIZE(opcode_offset); i++) { + entry = &opcode_offset[i]; + if (entry->opcode == code[0]) + break; + } + + opfrag = *(code + entry->byte) & entry->mask; + + insn = &opcode[entry->offset]; + for (i = 0; i < entry->count; i++) { + if (insn->opfrag == opfrag) + return insn; + insn++; + } + return NULL; +} + +static int print_insn(char *buffer, unsigned char *code, unsigned long addr) +{ + struct s390_insn *insn; + const unsigned char *ops; + const struct s390_operand *operand; + unsigned int value; + char separator; + char *ptr; + int i; + + ptr = buffer; + insn = find_insn(code); + if (insn) { + if (insn->zero == 0) + ptr += sprintf(ptr, "%.7s\t", + long_insn_name[insn->offset]); + else + ptr += sprintf(ptr, "%.5s\t", insn->name); + /* Extract the operands. */ + separator = 0; + for (ops = formats[insn->format], i = 0; + *ops != 0 && i < 6; ops++, i++) { + operand = operands + *ops; + value = extract_operand(code, operand); + if ((operand->flags & OPERAND_INDEX) && value == 0) + continue; + if ((operand->flags & OPERAND_BASE) && + value == 0 && separator == '(') { + separator = ','; + continue; + } + if (separator) + ptr += sprintf(ptr, "%c", separator); + if (operand->flags & OPERAND_GPR) + ptr += sprintf(ptr, "%%r%i", value); + else if (operand->flags & OPERAND_FPR) + ptr += sprintf(ptr, "%%f%i", value); + else if (operand->flags & OPERAND_AR) + ptr += sprintf(ptr, "%%a%i", value); + else if (operand->flags & OPERAND_CR) + ptr += sprintf(ptr, "%%c%i", value); + else if (operand->flags & OPERAND_VR) + ptr += sprintf(ptr, "%%v%i", value); + else if (operand->flags & OPERAND_PCREL) { + void *pcrel = (void *)((int)value + addr); + + ptr += sprintf(ptr, "%px", pcrel); + } else if (operand->flags & OPERAND_SIGNED) + ptr += sprintf(ptr, "%i", value); + else + ptr += sprintf(ptr, "%u", value); + if (operand->flags & OPERAND_DISP) + separator = '('; + else if (operand->flags & OPERAND_BASE) { + ptr += sprintf(ptr, ")"); + separator = ','; + } else + separator = ','; + } + } else + ptr += sprintf(ptr, "unknown"); + return (int) (ptr - buffer); +} + +static int copy_from_regs(struct pt_regs *regs, void *dst, void *src, int len) +{ + if (user_mode(regs)) { + if (copy_from_user(dst, (char __user *)src, len)) + return -EFAULT; + } else { + if (copy_from_kernel_nofault(dst, src, len)) + return -EFAULT; + } + return 0; +} + +void show_code(struct pt_regs *regs) +{ + char *mode = user_mode(regs) ? "User" : "Krnl"; + unsigned char code[64]; + char buffer[128], *ptr; + unsigned long addr; + int start, end, opsize, hops, i; + + /* Get a snapshot of the 64 bytes surrounding the fault address. */ + for (start = 32; start && regs->psw.addr >= 34 - start; start -= 2) { + addr = regs->psw.addr - 34 + start; + if (copy_from_regs(regs, code + start - 2, (void *)addr, 2)) + break; + } + for (end = 32; end < 64; end += 2) { + addr = regs->psw.addr + end - 32; + if (copy_from_regs(regs, code + end, (void *)addr, 2)) + break; + } + /* Code snapshot usable ? */ + if ((regs->psw.addr & 1) || start >= end) { + printk("%s Code: Bad PSW.\n", mode); + return; + } + /* Find a starting point for the disassembly. */ + while (start < 32) { + for (i = 0, hops = 0; start + i < 32 && hops < 3; hops++) { + if (!find_insn(code + start + i)) + break; + i += insn_length(code[start + i]); + } + if (start + i == 32) + /* Looks good, sequence ends at PSW. */ + break; + start += 2; + } + /* Decode the instructions. */ + ptr = buffer; + ptr += sprintf(ptr, "%s Code:", mode); + hops = 0; + while (start < end && hops < 8) { + opsize = insn_length(code[start]); + if (start + opsize == 32) + *ptr++ = '#'; + else if (start == 32) + *ptr++ = '>'; + else + *ptr++ = ' '; + addr = regs->psw.addr + start - 32; + ptr += sprintf(ptr, "%px: ", (void *)addr); + if (start + opsize >= end) + break; + for (i = 0; i < opsize; i++) + ptr += sprintf(ptr, "%02x", code[start + i]); + *ptr++ = '\t'; + if (i < 6) + *ptr++ = '\t'; + ptr += print_insn(ptr, code + start, addr); + start += opsize; + pr_cont("%s", buffer); + ptr = buffer; + ptr += sprintf(ptr, "\n "); + hops++; + } + pr_cont("\n"); +} + +void print_fn_code(unsigned char *code, unsigned long len) +{ + char buffer[128], *ptr; + int opsize, i; + + while (len) { + ptr = buffer; + opsize = insn_length(*code); + if (opsize > len) + break; + ptr += sprintf(ptr, "%px: ", code); + for (i = 0; i < opsize; i++) + ptr += sprintf(ptr, "%02x", code[i]); + *ptr++ = '\t'; + if (i < 4) + *ptr++ = '\t'; + ptr += print_insn(ptr, code, (unsigned long) code); + *ptr++ = '\n'; + *ptr++ = 0; + printk("%s", buffer); + code += opsize; + len -= opsize; + } +} diff --git a/arch/s390/kernel/dumpstack.c b/arch/s390/kernel/dumpstack.c new file mode 100644 index 0000000000..d2012635b0 --- /dev/null +++ b/arch/s390/kernel/dumpstack.c @@ -0,0 +1,224 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Stack dumping functions + * + * Copyright IBM Corp. 1999, 2013 + */ + +#include <linux/kallsyms.h> +#include <linux/hardirq.h> +#include <linux/kprobes.h> +#include <linux/utsname.h> +#include <linux/export.h> +#include <linux/kdebug.h> +#include <linux/ptrace.h> +#include <linux/mm.h> +#include <linux/module.h> +#include <linux/sched.h> +#include <linux/sched/debug.h> +#include <linux/sched/task_stack.h> +#include <asm/processor.h> +#include <asm/debug.h> +#include <asm/dis.h> +#include <asm/ipl.h> +#include <asm/unwind.h> + +const char *stack_type_name(enum stack_type type) +{ + switch (type) { + case STACK_TYPE_TASK: + return "task"; + case STACK_TYPE_IRQ: + return "irq"; + case STACK_TYPE_NODAT: + return "nodat"; + case STACK_TYPE_RESTART: + return "restart"; + default: + return "unknown"; + } +} +EXPORT_SYMBOL_GPL(stack_type_name); + +static inline bool in_stack(unsigned long sp, struct stack_info *info, + enum stack_type type, unsigned long stack) +{ + if (sp < stack || sp >= stack + THREAD_SIZE) + return false; + info->type = type; + info->begin = stack; + info->end = stack + THREAD_SIZE; + return true; +} + +static bool in_task_stack(unsigned long sp, struct task_struct *task, + struct stack_info *info) +{ + unsigned long stack = (unsigned long)task_stack_page(task); + + return in_stack(sp, info, STACK_TYPE_TASK, stack); +} + +static bool in_irq_stack(unsigned long sp, struct stack_info *info) +{ + unsigned long stack = S390_lowcore.async_stack - STACK_INIT_OFFSET; + + return in_stack(sp, info, STACK_TYPE_IRQ, stack); +} + +static bool in_nodat_stack(unsigned long sp, struct stack_info *info) +{ + unsigned long stack = S390_lowcore.nodat_stack - STACK_INIT_OFFSET; + + return in_stack(sp, info, STACK_TYPE_NODAT, stack); +} + +static bool in_mcck_stack(unsigned long sp, struct stack_info *info) +{ + unsigned long stack = S390_lowcore.mcck_stack - STACK_INIT_OFFSET; + + return in_stack(sp, info, STACK_TYPE_MCCK, stack); +} + +static bool in_restart_stack(unsigned long sp, struct stack_info *info) +{ + unsigned long stack = S390_lowcore.restart_stack - STACK_INIT_OFFSET; + + return in_stack(sp, info, STACK_TYPE_RESTART, stack); +} + +int get_stack_info(unsigned long sp, struct task_struct *task, + struct stack_info *info, unsigned long *visit_mask) +{ + if (!sp) + goto unknown; + + /* Sanity check: ABI requires SP to be aligned 8 bytes. */ + if (sp & 0x7) + goto unknown; + + /* Check per-task stack */ + if (in_task_stack(sp, task, info)) + goto recursion_check; + + if (task != current) + goto unknown; + + /* Check per-cpu stacks */ + if (!in_irq_stack(sp, info) && + !in_nodat_stack(sp, info) && + !in_restart_stack(sp, info) && + !in_mcck_stack(sp, info)) + goto unknown; + +recursion_check: + /* + * Make sure we don't iterate through any given stack more than once. + * If it comes up a second time then there's something wrong going on: + * just break out and report an unknown stack type. + */ + if (*visit_mask & (1UL << info->type)) + goto unknown; + *visit_mask |= 1UL << info->type; + return 0; +unknown: + info->type = STACK_TYPE_UNKNOWN; + return -EINVAL; +} + +void show_stack(struct task_struct *task, unsigned long *stack, + const char *loglvl) +{ + struct unwind_state state; + + printk("%sCall Trace:\n", loglvl); + unwind_for_each_frame(&state, task, NULL, (unsigned long) stack) + printk(state.reliable ? "%s [<%016lx>] %pSR \n" : + "%s([<%016lx>] %pSR)\n", + loglvl, state.ip, (void *) state.ip); + debug_show_held_locks(task ? : current); +} + +static void show_last_breaking_event(struct pt_regs *regs) +{ + printk("Last Breaking-Event-Address:\n"); + printk(" [<%016lx>] ", regs->last_break); + if (user_mode(regs)) { + print_vma_addr(KERN_CONT, regs->last_break); + pr_cont("\n"); + } else { + pr_cont("%pSR\n", (void *)regs->last_break); + } +} + +void show_registers(struct pt_regs *regs) +{ + struct psw_bits *psw = &psw_bits(regs->psw); + char *mode; + + mode = user_mode(regs) ? "User" : "Krnl"; + printk("%s PSW : %px %px", mode, (void *)regs->psw.mask, (void *)regs->psw.addr); + if (!user_mode(regs)) + pr_cont(" (%pSR)", (void *)regs->psw.addr); + pr_cont("\n"); + printk(" R:%x T:%x IO:%x EX:%x Key:%x M:%x W:%x " + "P:%x AS:%x CC:%x PM:%x", psw->per, psw->dat, psw->io, psw->ext, + psw->key, psw->mcheck, psw->wait, psw->pstate, psw->as, psw->cc, psw->pm); + pr_cont(" RI:%x EA:%x\n", psw->ri, psw->eaba); + printk("%s GPRS: %016lx %016lx %016lx %016lx\n", mode, + regs->gprs[0], regs->gprs[1], regs->gprs[2], regs->gprs[3]); + printk(" %016lx %016lx %016lx %016lx\n", + regs->gprs[4], regs->gprs[5], regs->gprs[6], regs->gprs[7]); + printk(" %016lx %016lx %016lx %016lx\n", + regs->gprs[8], regs->gprs[9], regs->gprs[10], regs->gprs[11]); + printk(" %016lx %016lx %016lx %016lx\n", + regs->gprs[12], regs->gprs[13], regs->gprs[14], regs->gprs[15]); + show_code(regs); +} + +void show_regs(struct pt_regs *regs) +{ + show_regs_print_info(KERN_DEFAULT); + show_registers(regs); + /* Show stack backtrace if pt_regs is from kernel mode */ + if (!user_mode(regs)) + show_stack(NULL, (unsigned long *) regs->gprs[15], KERN_DEFAULT); + show_last_breaking_event(regs); +} + +static DEFINE_SPINLOCK(die_lock); + +void __noreturn die(struct pt_regs *regs, const char *str) +{ + static int die_counter; + + oops_enter(); + lgr_info_log(); + debug_stop_all(); + console_verbose(); + spin_lock_irq(&die_lock); + bust_spinlocks(1); + printk("%s: %04x ilc:%d [#%d] ", str, regs->int_code & 0xffff, + regs->int_code >> 17, ++die_counter); +#ifdef CONFIG_PREEMPT + pr_cont("PREEMPT "); +#elif defined(CONFIG_PREEMPT_RT) + pr_cont("PREEMPT_RT "); +#endif + pr_cont("SMP "); + if (debug_pagealloc_enabled()) + pr_cont("DEBUG_PAGEALLOC"); + pr_cont("\n"); + notify_die(DIE_OOPS, str, regs, 0, regs->int_code & 0xffff, SIGSEGV); + print_modules(); + show_regs(regs); + bust_spinlocks(0); + add_taint(TAINT_DIE, LOCKDEP_NOW_UNRELIABLE); + spin_unlock_irq(&die_lock); + if (in_interrupt()) + panic("Fatal exception in interrupt"); + if (panic_on_oops) + panic("Fatal exception: panic_on_oops"); + oops_exit(); + make_task_dead(SIGSEGV); +} diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c new file mode 100644 index 0000000000..442ce0489e --- /dev/null +++ b/arch/s390/kernel/early.c @@ -0,0 +1,320 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2007, 2009 + * Author(s): Hongjie Yang <hongjie@us.ibm.com>, + */ + +#define KMSG_COMPONENT "setup" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/compiler.h> +#include <linux/init.h> +#include <linux/errno.h> +#include <linux/string.h> +#include <linux/ctype.h> +#include <linux/lockdep.h> +#include <linux/extable.h> +#include <linux/pfn.h> +#include <linux/uaccess.h> +#include <linux/kernel.h> +#include <asm/asm-extable.h> +#include <linux/memblock.h> +#include <asm/diag.h> +#include <asm/ebcdic.h> +#include <asm/ipl.h> +#include <asm/lowcore.h> +#include <asm/processor.h> +#include <asm/sections.h> +#include <asm/setup.h> +#include <asm/sysinfo.h> +#include <asm/cpcmd.h> +#include <asm/sclp.h> +#include <asm/facility.h> +#include <asm/boot_data.h> +#include <asm/switch_to.h> +#include "entry.h" + +#define decompressor_handled_param(param) \ +static int __init ignore_decompressor_param_##param(char *s) \ +{ \ + return 0; \ +} \ +early_param(#param, ignore_decompressor_param_##param) + +decompressor_handled_param(mem); +decompressor_handled_param(vmalloc); +decompressor_handled_param(dfltcc); +decompressor_handled_param(facilities); +decompressor_handled_param(nokaslr); +#if IS_ENABLED(CONFIG_KVM) +decompressor_handled_param(prot_virt); +#endif + +static void __init kasan_early_init(void) +{ +#ifdef CONFIG_KASAN + init_task.kasan_depth = 0; + sclp_early_printk("KernelAddressSanitizer initialized\n"); +#endif +} + +static void __init reset_tod_clock(void) +{ + union tod_clock clk; + + if (store_tod_clock_ext_cc(&clk) == 0) + return; + /* TOD clock not running. Set the clock to Unix Epoch. */ + if (set_tod_clock(TOD_UNIX_EPOCH) || store_tod_clock_ext_cc(&clk)) + disabled_wait(); + + memset(&tod_clock_base, 0, sizeof(tod_clock_base)); + tod_clock_base.tod = TOD_UNIX_EPOCH; + S390_lowcore.last_update_clock = TOD_UNIX_EPOCH; +} + +/* + * Initialize storage key for kernel pages + */ +static noinline __init void init_kernel_storage_key(void) +{ +#if PAGE_DEFAULT_KEY + unsigned long end_pfn, init_pfn; + + end_pfn = PFN_UP(__pa(_end)); + + for (init_pfn = 0 ; init_pfn < end_pfn; init_pfn++) + page_set_storage_key(init_pfn << PAGE_SHIFT, + PAGE_DEFAULT_KEY, 0); +#endif +} + +static __initdata char sysinfo_page[PAGE_SIZE] __aligned(PAGE_SIZE); + +static noinline __init void detect_machine_type(void) +{ + struct sysinfo_3_2_2 *vmms = (struct sysinfo_3_2_2 *)&sysinfo_page; + + /* Check current-configuration-level */ + if (stsi(NULL, 0, 0, 0) <= 2) { + S390_lowcore.machine_flags |= MACHINE_FLAG_LPAR; + return; + } + /* Get virtual-machine cpu information. */ + if (stsi(vmms, 3, 2, 2) || !vmms->count) + return; + + /* Detect known hypervisors */ + if (!memcmp(vmms->vm[0].cpi, "\xd2\xe5\xd4", 3)) + S390_lowcore.machine_flags |= MACHINE_FLAG_KVM; + else if (!memcmp(vmms->vm[0].cpi, "\xa9\x61\xe5\xd4", 4)) + S390_lowcore.machine_flags |= MACHINE_FLAG_VM; +} + +/* Remove leading, trailing and double whitespace. */ +static inline void strim_all(char *str) +{ + char *s; + + s = strim(str); + if (s != str) + memmove(str, s, strlen(s)); + while (*str) { + if (!isspace(*str++)) + continue; + if (isspace(*str)) { + s = skip_spaces(str); + memmove(str, s, strlen(s) + 1); + } + } +} + +static noinline __init void setup_arch_string(void) +{ + struct sysinfo_1_1_1 *mach = (struct sysinfo_1_1_1 *)&sysinfo_page; + struct sysinfo_3_2_2 *vm = (struct sysinfo_3_2_2 *)&sysinfo_page; + char mstr[80], hvstr[17]; + + if (stsi(mach, 1, 1, 1)) + return; + EBCASC(mach->manufacturer, sizeof(mach->manufacturer)); + EBCASC(mach->type, sizeof(mach->type)); + EBCASC(mach->model, sizeof(mach->model)); + EBCASC(mach->model_capacity, sizeof(mach->model_capacity)); + sprintf(mstr, "%-16.16s %-4.4s %-16.16s %-16.16s", + mach->manufacturer, mach->type, + mach->model, mach->model_capacity); + strim_all(mstr); + if (stsi(vm, 3, 2, 2) == 0 && vm->count) { + EBCASC(vm->vm[0].cpi, sizeof(vm->vm[0].cpi)); + sprintf(hvstr, "%-16.16s", vm->vm[0].cpi); + strim_all(hvstr); + } else { + sprintf(hvstr, "%s", + MACHINE_IS_LPAR ? "LPAR" : + MACHINE_IS_VM ? "z/VM" : + MACHINE_IS_KVM ? "KVM" : "unknown"); + } + dump_stack_set_arch_desc("%s (%s)", mstr, hvstr); +} + +static __init void setup_topology(void) +{ + int max_mnest; + + if (!test_facility(11)) + return; + S390_lowcore.machine_flags |= MACHINE_FLAG_TOPOLOGY; + for (max_mnest = 6; max_mnest > 1; max_mnest--) { + if (stsi(&sysinfo_page, 15, 1, max_mnest) == 0) + break; + } + topology_max_mnest = max_mnest; +} + +void __do_early_pgm_check(struct pt_regs *regs) +{ + if (!fixup_exception(regs)) + disabled_wait(); +} + +static noinline __init void setup_lowcore_early(void) +{ + psw_t psw; + + psw.addr = (unsigned long)early_pgm_check_handler; + psw.mask = PSW_KERNEL_BITS; + S390_lowcore.program_new_psw = psw; + S390_lowcore.preempt_count = INIT_PREEMPT_COUNT; +} + +static noinline __init void setup_facility_list(void) +{ + memcpy(alt_stfle_fac_list, stfle_fac_list, sizeof(alt_stfle_fac_list)); + if (!IS_ENABLED(CONFIG_KERNEL_NOBP)) + __clear_facility(82, alt_stfle_fac_list); +} + +static __init void detect_diag9c(void) +{ + unsigned int cpu_address; + int rc; + + cpu_address = stap(); + diag_stat_inc(DIAG_STAT_X09C); + asm volatile( + " diag %2,0,0x9c\n" + "0: la %0,0\n" + "1:\n" + EX_TABLE(0b,1b) + : "=d" (rc) : "0" (-EOPNOTSUPP), "d" (cpu_address) : "cc"); + if (!rc) + S390_lowcore.machine_flags |= MACHINE_FLAG_DIAG9C; +} + +static __init void detect_machine_facilities(void) +{ + if (test_facility(8)) { + S390_lowcore.machine_flags |= MACHINE_FLAG_EDAT1; + __ctl_set_bit(0, 23); + } + if (test_facility(78)) + S390_lowcore.machine_flags |= MACHINE_FLAG_EDAT2; + if (test_facility(3)) + S390_lowcore.machine_flags |= MACHINE_FLAG_IDTE; + if (test_facility(50) && test_facility(73)) { + S390_lowcore.machine_flags |= MACHINE_FLAG_TE; + __ctl_set_bit(0, 55); + } + if (test_facility(51)) + S390_lowcore.machine_flags |= MACHINE_FLAG_TLB_LC; + if (test_facility(129)) { + S390_lowcore.machine_flags |= MACHINE_FLAG_VX; + __ctl_set_bit(0, 17); + } + if (test_facility(130)) + S390_lowcore.machine_flags |= MACHINE_FLAG_NX; + if (test_facility(133)) + S390_lowcore.machine_flags |= MACHINE_FLAG_GS; + if (test_facility(139) && (tod_clock_base.tod >> 63)) { + /* Enabled signed clock comparator comparisons */ + S390_lowcore.machine_flags |= MACHINE_FLAG_SCC; + clock_comparator_max = -1ULL >> 1; + __ctl_set_bit(0, 53); + } + if (IS_ENABLED(CONFIG_PCI) && test_facility(153)) { + S390_lowcore.machine_flags |= MACHINE_FLAG_PCI_MIO; + /* the control bit is set during PCI initialization */ + } + if (test_facility(194)) + S390_lowcore.machine_flags |= MACHINE_FLAG_RDP; +} + +static inline void save_vector_registers(void) +{ +#ifdef CONFIG_CRASH_DUMP + if (test_facility(129)) + save_vx_regs(boot_cpu_vector_save_area); +#endif +} + +static inline void setup_control_registers(void) +{ + unsigned long reg; + + __ctl_store(reg, 0, 0); + reg |= CR0_LOW_ADDRESS_PROTECTION; + reg |= CR0_EMERGENCY_SIGNAL_SUBMASK; + reg |= CR0_EXTERNAL_CALL_SUBMASK; + __ctl_load(reg, 0, 0); +} + +static inline void setup_access_registers(void) +{ + unsigned int acrs[NUM_ACRS] = { 0 }; + + restore_access_regs(acrs); +} + +static int __init disable_vector_extension(char *str) +{ + S390_lowcore.machine_flags &= ~MACHINE_FLAG_VX; + __ctl_clear_bit(0, 17); + return 0; +} +early_param("novx", disable_vector_extension); + +char __bootdata(early_command_line)[COMMAND_LINE_SIZE]; +static void __init setup_boot_command_line(void) +{ + /* copy arch command line */ + strscpy(boot_command_line, early_command_line, COMMAND_LINE_SIZE); +} + +static void __init sort_amode31_extable(void) +{ + sort_extable(__start_amode31_ex_table, __stop_amode31_ex_table); +} + +void __init startup_init(void) +{ + kasan_early_init(); + reset_tod_clock(); + time_early_init(); + init_kernel_storage_key(); + lockdep_off(); + sort_amode31_extable(); + setup_lowcore_early(); + setup_facility_list(); + detect_machine_type(); + setup_arch_string(); + setup_boot_command_line(); + detect_diag9c(); + detect_machine_facilities(); + save_vector_registers(); + setup_topology(); + sclp_early_detect(); + setup_control_registers(); + setup_access_registers(); + lockdep_on(); +} diff --git a/arch/s390/kernel/early_printk.c b/arch/s390/kernel/early_printk.c new file mode 100644 index 0000000000..d9d53f4400 --- /dev/null +++ b/arch/s390/kernel/early_printk.c @@ -0,0 +1,36 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2017 + */ + +#include <linux/console.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include <asm/sclp.h> + +static void sclp_early_write(struct console *con, const char *s, unsigned int len) +{ + __sclp_early_printk(s, len); +} + +static struct console sclp_early_console = { + .name = "earlysclp", + .write = sclp_early_write, + .flags = CON_PRINTBUFFER | CON_BOOT, + .index = -1, +}; + +static int __init setup_early_printk(char *buf) +{ + if (early_console) + return 0; + /* Accept only "earlyprintk" and "earlyprintk=sclp" */ + if (buf && !str_has_prefix(buf, "sclp")) + return 0; + if (!sclp.has_linemode && !sclp.has_vt220) + return 0; + early_console = &sclp_early_console; + register_console(early_console); + return 0; +} +early_param("earlyprintk", setup_early_printk); diff --git a/arch/s390/kernel/earlypgm.S b/arch/s390/kernel/earlypgm.S new file mode 100644 index 0000000000..c634871f0d --- /dev/null +++ b/arch/s390/kernel/earlypgm.S @@ -0,0 +1,23 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright IBM Corp. 2006, 2007 + * Author(s): Michael Holzheu <holzheu@de.ibm.com> + */ + +#include <linux/linkage.h> +#include <asm/asm-offsets.h> + +SYM_CODE_START(early_pgm_check_handler) + stmg %r8,%r15,__LC_SAVE_AREA_SYNC + aghi %r15,-(STACK_FRAME_OVERHEAD+__PT_SIZE) + la %r11,STACK_FRAME_OVERHEAD(%r15) + xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) + stmg %r0,%r7,__PT_R0(%r11) + mvc __PT_PSW(16,%r11),__LC_PGM_OLD_PSW + mvc __PT_R8(64,%r11),__LC_SAVE_AREA_SYNC + lgr %r2,%r11 + brasl %r14,__do_early_pgm_check + mvc __LC_RETURN_PSW(16),STACK_FRAME_OVERHEAD+__PT_PSW(%r15) + lmg %r0,%r15,STACK_FRAME_OVERHEAD+__PT_R0(%r15) + lpswe __LC_RETURN_PSW +SYM_CODE_END(early_pgm_check_handler) diff --git a/arch/s390/kernel/ebcdic.c b/arch/s390/kernel/ebcdic.c new file mode 100644 index 0000000000..0e51fa5372 --- /dev/null +++ b/arch/s390/kernel/ebcdic.c @@ -0,0 +1,401 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * EBCDIC -> ASCII, ASCII -> EBCDIC, + * upper to lower case (EBCDIC) conversion tables. + * + * S390 version + * Copyright IBM Corp. 1999 + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> + * Martin Peschke <peschke@fh-brandenburg.de> + */ + +#include <linux/types.h> +#include <linux/export.h> +#include <asm/ebcdic.h> + +/* + * ASCII (IBM PC 437) -> EBCDIC 037 + */ +__u8 _ascebc[256] = +{ + /*00 NUL SOH STX ETX EOT ENQ ACK BEL */ + 0x00, 0x01, 0x02, 0x03, 0x37, 0x2D, 0x2E, 0x2F, + /*08 BS HT LF VT FF CR SO SI */ + /* ->NL */ + 0x16, 0x05, 0x15, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, + /*10 DLE DC1 DC2 DC3 DC4 NAK SYN ETB */ + 0x10, 0x11, 0x12, 0x13, 0x3C, 0x3D, 0x32, 0x26, + /*18 CAN EM SUB ESC FS GS RS US */ + /* ->IGS ->IRS ->IUS */ + 0x18, 0x19, 0x3F, 0x27, 0x22, 0x1D, 0x1E, 0x1F, + /*20 SP ! " # $ % & ' */ + 0x40, 0x5A, 0x7F, 0x7B, 0x5B, 0x6C, 0x50, 0x7D, + /*28 ( ) * + , - . / */ + 0x4D, 0x5D, 0x5C, 0x4E, 0x6B, 0x60, 0x4B, 0x61, + /*30 0 1 2 3 4 5 6 7 */ + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, + /*38 8 9 : ; < = > ? */ + 0xF8, 0xF9, 0x7A, 0x5E, 0x4C, 0x7E, 0x6E, 0x6F, + /*40 @ A B C D E F G */ + 0x7C, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, + /*48 H I J K L M N O */ + 0xC8, 0xC9, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, + /*50 P Q R S T U V W */ + 0xD7, 0xD8, 0xD9, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, + /*58 X Y Z [ \ ] ^ _ */ + 0xE7, 0xE8, 0xE9, 0xBA, 0xE0, 0xBB, 0xB0, 0x6D, + /*60 ` a b c d e f g */ + 0x79, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, + /*68 h i j k l m n o */ + 0x88, 0x89, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, + /*70 p q r s t u v w */ + 0x97, 0x98, 0x99, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, + /*78 x y z { | } ~ DL */ + 0xA7, 0xA8, 0xA9, 0xC0, 0x4F, 0xD0, 0xA1, 0x07, + /*80*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*88*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*90*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*98*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*A0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*A8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*B0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*B8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*C0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*C8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*D0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*D8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*E0 sz */ + 0x3F, 0x59, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*E8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*F0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*F8*/ + 0x90, 0x3F, 0x3F, 0x3F, 0x3F, 0xEA, 0x3F, 0xFF +}; + +/* + * EBCDIC 037 -> ASCII (IBM PC 437) + */ +__u8 _ebcasc[256] = +{ + /* 0x00 NUL SOH STX ETX *SEL HT *RNL DEL */ + 0x00, 0x01, 0x02, 0x03, 0x07, 0x09, 0x07, 0x7F, + /* 0x08 -GE -SPS -RPT VT FF CR SO SI */ + 0x07, 0x07, 0x07, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, + /* 0x10 DLE DC1 DC2 DC3 -RES -NL BS -POC + -ENP ->LF */ + 0x10, 0x11, 0x12, 0x13, 0x07, 0x0A, 0x08, 0x07, + /* 0x18 CAN EM -UBS -CU1 -IFS -IGS -IRS -ITB + -IUS */ + 0x18, 0x19, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, + /* 0x20 -DS -SOS FS -WUS -BYP LF ETB ESC + -INP */ + 0x07, 0x07, 0x1C, 0x07, 0x07, 0x0A, 0x17, 0x1B, + /* 0x28 -SA -SFE -SM -CSP -MFA ENQ ACK BEL + -SW */ + 0x07, 0x07, 0x07, 0x07, 0x07, 0x05, 0x06, 0x07, + /* 0x30 ---- ---- SYN -IR -PP -TRN -NBS EOT */ + 0x07, 0x07, 0x16, 0x07, 0x07, 0x07, 0x07, 0x04, + /* 0x38 -SBS -IT -RFF -CU3 DC4 NAK ---- SUB */ + 0x07, 0x07, 0x07, 0x07, 0x14, 0x15, 0x07, 0x1A, + /* 0x40 SP RSP ä ---- */ + 0x20, 0xFF, 0x83, 0x84, 0x85, 0xA0, 0x07, 0x86, + /* 0x48 . < ( + | */ + 0x87, 0xA4, 0x9B, 0x2E, 0x3C, 0x28, 0x2B, 0x7C, + /* 0x50 & ---- */ + 0x26, 0x82, 0x88, 0x89, 0x8A, 0xA1, 0x8C, 0x07, + /* 0x58 ß ! $ * ) ; */ + 0x8D, 0xE1, 0x21, 0x24, 0x2A, 0x29, 0x3B, 0xAA, + /* 0x60 - / ---- Ä ---- ---- ---- */ + 0x2D, 0x2F, 0x07, 0x8E, 0x07, 0x07, 0x07, 0x8F, + /* 0x68 ---- , % _ > ? */ + 0x80, 0xA5, 0x07, 0x2C, 0x25, 0x5F, 0x3E, 0x3F, + /* 0x70 ---- ---- ---- ---- ---- ---- ---- */ + 0x07, 0x90, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, + /* 0x78 * ` : # @ ' = " */ + 0x70, 0x60, 0x3A, 0x23, 0x40, 0x27, 0x3D, 0x22, + /* 0x80 * a b c d e f g */ + 0x07, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + /* 0x88 h i ---- ---- ---- */ + 0x68, 0x69, 0xAE, 0xAF, 0x07, 0x07, 0x07, 0xF1, + /* 0x90 ° j k l m n o p */ + 0xF8, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, + /* 0x98 q r ---- ---- */ + 0x71, 0x72, 0xA6, 0xA7, 0x91, 0x07, 0x92, 0x07, + /* 0xA0 ~ s t u v w x */ + 0xE6, 0x7E, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, + /* 0xA8 y z ---- ---- ---- ---- */ + 0x79, 0x7A, 0xAD, 0xAB, 0x07, 0x07, 0x07, 0x07, + /* 0xB0 ^ ---- § ---- */ + 0x5E, 0x9C, 0x9D, 0xFA, 0x07, 0x07, 0x07, 0xAC, + /* 0xB8 ---- [ ] ---- ---- ---- ---- */ + 0xAB, 0x07, 0x5B, 0x5D, 0x07, 0x07, 0x07, 0x07, + /* 0xC0 { A B C D E F G */ + 0x7B, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, + /* 0xC8 H I ---- ö ---- */ + 0x48, 0x49, 0x07, 0x93, 0x94, 0x95, 0xA2, 0x07, + /* 0xD0 } J K L M N O P */ + 0x7D, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50, + /* 0xD8 Q R ---- ü */ + 0x51, 0x52, 0x07, 0x96, 0x81, 0x97, 0xA3, 0x98, + /* 0xE0 \ S T U V W X */ + 0x5C, 0xF6, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, + /* 0xE8 Y Z ---- Ö ---- ---- ---- */ + 0x59, 0x5A, 0xFD, 0x07, 0x99, 0x07, 0x07, 0x07, + /* 0xF0 0 1 2 3 4 5 6 7 */ + 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, + /* 0xF8 8 9 ---- ---- Ãœ ---- ---- ---- */ + 0x38, 0x39, 0x07, 0x07, 0x9A, 0x07, 0x07, 0x07 +}; + + +/* + * ASCII (IBM PC 437) -> EBCDIC 500 + */ +__u8 _ascebc_500[256] = +{ + /*00 NUL SOH STX ETX EOT ENQ ACK BEL */ + 0x00, 0x01, 0x02, 0x03, 0x37, 0x2D, 0x2E, 0x2F, + /*08 BS HT LF VT FF CR SO SI */ + /* ->NL */ + 0x16, 0x05, 0x15, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, + /*10 DLE DC1 DC2 DC3 DC4 NAK SYN ETB */ + 0x10, 0x11, 0x12, 0x13, 0x3C, 0x3D, 0x32, 0x26, + /*18 CAN EM SUB ESC FS GS RS US */ + /* ->IGS ->IRS ->IUS */ + 0x18, 0x19, 0x3F, 0x27, 0x22, 0x1D, 0x1E, 0x1F, + /*20 SP ! " # $ % & ' */ + 0x40, 0x4F, 0x7F, 0x7B, 0x5B, 0x6C, 0x50, 0x7D, + /*28 ( ) * + , - . / */ + 0x4D, 0x5D, 0x5C, 0x4E, 0x6B, 0x60, 0x4B, 0x61, + /*30 0 1 2 3 4 5 6 7 */ + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, + /*38 8 9 : ; < = > ? */ + 0xF8, 0xF9, 0x7A, 0x5E, 0x4C, 0x7E, 0x6E, 0x6F, + /*40 @ A B C D E F G */ + 0x7C, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, + /*48 H I J K L M N O */ + 0xC8, 0xC9, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, + /*50 P Q R S T U V W */ + 0xD7, 0xD8, 0xD9, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, + /*58 X Y Z [ \ ] ^ _ */ + 0xE7, 0xE8, 0xE9, 0x4A, 0xE0, 0x5A, 0x5F, 0x6D, + /*60 ` a b c d e f g */ + 0x79, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, + /*68 h i j k l m n o */ + 0x88, 0x89, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, + /*70 p q r s t u v w */ + 0x97, 0x98, 0x99, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, + /*78 x y z { | } ~ DL */ + 0xA7, 0xA8, 0xA9, 0xC0, 0xBB, 0xD0, 0xA1, 0x07, + /*80*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*88*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*90*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*98*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*A0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*A8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*B0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*B8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*C0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*C8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*D0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*D8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*E0 sz */ + 0x3F, 0x59, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*E8*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*F0*/ + 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, + /*F8*/ + 0x90, 0x3F, 0x3F, 0x3F, 0x3F, 0xEA, 0x3F, 0xFF +}; + +/* + * EBCDIC 500 -> ASCII (IBM PC 437) + */ +__u8 _ebcasc_500[256] = +{ + /* 0x00 NUL SOH STX ETX *SEL HT *RNL DEL */ + 0x00, 0x01, 0x02, 0x03, 0x07, 0x09, 0x07, 0x7F, + /* 0x08 -GE -SPS -RPT VT FF CR SO SI */ + 0x07, 0x07, 0x07, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, + /* 0x10 DLE DC1 DC2 DC3 -RES -NL BS -POC + -ENP ->LF */ + 0x10, 0x11, 0x12, 0x13, 0x07, 0x0A, 0x08, 0x07, + /* 0x18 CAN EM -UBS -CU1 -IFS -IGS -IRS -ITB + -IUS */ + 0x18, 0x19, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, + /* 0x20 -DS -SOS FS -WUS -BYP LF ETB ESC + -INP */ + 0x07, 0x07, 0x1C, 0x07, 0x07, 0x0A, 0x17, 0x1B, + /* 0x28 -SA -SFE -SM -CSP -MFA ENQ ACK BEL + -SW */ + 0x07, 0x07, 0x07, 0x07, 0x07, 0x05, 0x06, 0x07, + /* 0x30 ---- ---- SYN -IR -PP -TRN -NBS EOT */ + 0x07, 0x07, 0x16, 0x07, 0x07, 0x07, 0x07, 0x04, + /* 0x38 -SBS -IT -RFF -CU3 DC4 NAK ---- SUB */ + 0x07, 0x07, 0x07, 0x07, 0x14, 0x15, 0x07, 0x1A, + /* 0x40 SP RSP ä ---- */ + 0x20, 0xFF, 0x83, 0x84, 0x85, 0xA0, 0x07, 0x86, + /* 0x48 [ . < ( + ! */ + 0x87, 0xA4, 0x5B, 0x2E, 0x3C, 0x28, 0x2B, 0x21, + /* 0x50 & ---- */ + 0x26, 0x82, 0x88, 0x89, 0x8A, 0xA1, 0x8C, 0x07, + /* 0x58 ß ] $ * ) ; ^ */ + 0x8D, 0xE1, 0x5D, 0x24, 0x2A, 0x29, 0x3B, 0x5E, + /* 0x60 - / ---- Ä ---- ---- ---- */ + 0x2D, 0x2F, 0x07, 0x8E, 0x07, 0x07, 0x07, 0x8F, + /* 0x68 ---- , % _ > ? */ + 0x80, 0xA5, 0x07, 0x2C, 0x25, 0x5F, 0x3E, 0x3F, + /* 0x70 ---- ---- ---- ---- ---- ---- ---- */ + 0x07, 0x90, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, + /* 0x78 * ` : # @ ' = " */ + 0x70, 0x60, 0x3A, 0x23, 0x40, 0x27, 0x3D, 0x22, + /* 0x80 * a b c d e f g */ + 0x07, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + /* 0x88 h i ---- ---- ---- */ + 0x68, 0x69, 0xAE, 0xAF, 0x07, 0x07, 0x07, 0xF1, + /* 0x90 ° j k l m n o p */ + 0xF8, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, + /* 0x98 q r ---- ---- */ + 0x71, 0x72, 0xA6, 0xA7, 0x91, 0x07, 0x92, 0x07, + /* 0xA0 ~ s t u v w x */ + 0xE6, 0x7E, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, + /* 0xA8 y z ---- ---- ---- ---- */ + 0x79, 0x7A, 0xAD, 0xAB, 0x07, 0x07, 0x07, 0x07, + /* 0xB0 ---- § ---- */ + 0x9B, 0x9C, 0x9D, 0xFA, 0x07, 0x07, 0x07, 0xAC, + /* 0xB8 ---- | ---- ---- ---- ---- */ + 0xAB, 0x07, 0xAA, 0x7C, 0x07, 0x07, 0x07, 0x07, + /* 0xC0 { A B C D E F G */ + 0x7B, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, + /* 0xC8 H I ---- ö ---- */ + 0x48, 0x49, 0x07, 0x93, 0x94, 0x95, 0xA2, 0x07, + /* 0xD0 } J K L M N O P */ + 0x7D, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50, + /* 0xD8 Q R ---- ü */ + 0x51, 0x52, 0x07, 0x96, 0x81, 0x97, 0xA3, 0x98, + /* 0xE0 \ S T U V W X */ + 0x5C, 0xF6, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, + /* 0xE8 Y Z ---- Ö ---- ---- ---- */ + 0x59, 0x5A, 0xFD, 0x07, 0x99, 0x07, 0x07, 0x07, + /* 0xF0 0 1 2 3 4 5 6 7 */ + 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, + /* 0xF8 8 9 ---- ---- Ãœ ---- ---- ---- */ + 0x38, 0x39, 0x07, 0x07, 0x9A, 0x07, 0x07, 0x07 +}; + + +/* + * EBCDIC 037/500 conversion table: + * from upper to lower case + */ +__u8 _ebc_tolower[256] = +{ + 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, + 0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, + 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, + 0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F, + 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, + 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x2D, 0x2E, 0x2F, + 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, + 0x38, 0x39, 0x3A, 0x3B, 0x3C, 0x3D, 0x3E, 0x3F, + 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, + 0x48, 0x49, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, + 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, + 0x58, 0x59, 0x5A, 0x5B, 0x5C, 0x5D, 0x5E, 0x5F, + 0x60, 0x61, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, + 0x48, 0x49, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, + 0x58, 0x79, 0x7A, 0x7B, 0x7C, 0x7D, 0x7E, 0x7F, + 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, + 0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x8D, 0x8E, 0x8F, + 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, + 0x98, 0x99, 0x9A, 0x9B, 0x9C, 0x9D, 0x9C, 0x9F, + 0xA0, 0xA1, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, + 0xA8, 0xA9, 0xAA, 0xAB, 0x8C, 0x8D, 0x8E, 0xAF, + 0xB0, 0xB1, 0xB2, 0xB3, 0xB4, 0xB5, 0xB6, 0xB7, + 0xB8, 0xB9, 0xBA, 0xBB, 0xBC, 0xBD, 0xBE, 0xBF, + 0xC0, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, + 0x88, 0x89, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, + 0xD0, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, + 0x98, 0x99, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, + 0xE0, 0xE1, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, + 0xA8, 0xA9, 0xEA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, + 0xF8, 0xF9, 0xFA, 0xDB, 0xDC, 0xDD, 0xDE, 0xFF +}; + + +/* + * EBCDIC 037/500 conversion table: + * from lower to upper case + */ +__u8 _ebc_toupper[256] = +{ + 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, + 0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, + 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, + 0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F, + 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, + 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x2D, 0x2E, 0x2F, + 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, + 0x38, 0x39, 0x3A, 0x3B, 0x3C, 0x3D, 0x3E, 0x3F, + 0x40, 0x41, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, + 0x50, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x59, 0x5A, 0x5B, 0x5C, 0x5D, 0x5E, 0x5F, + 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x7B, 0x7C, 0x7D, 0x7E, 0x7F, + 0x80, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, + 0xC8, 0xC9, 0x8A, 0x8B, 0xAC, 0xAD, 0xAE, 0x8F, + 0x90, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, + 0xD8, 0xD9, 0x9A, 0x9B, 0x9E, 0x9D, 0x9E, 0x9F, + 0xA0, 0xA1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xAA, 0xAB, 0xAC, 0xAD, 0xAE, 0xAF, + 0xB0, 0xB1, 0xB2, 0xB3, 0xB4, 0xB5, 0xB6, 0xB7, + 0xB8, 0xB9, 0xBA, 0xBB, 0xBC, 0xBD, 0xBE, 0xBF, + 0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, + 0xC8, 0xC9, 0xCA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, + 0xD8, 0xD9, 0xDA, 0xFB, 0xFC, 0xFD, 0xFE, 0xDF, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, + 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF +}; + +EXPORT_SYMBOL(_ascebc_500); +EXPORT_SYMBOL(_ebcasc_500); +EXPORT_SYMBOL(_ascebc); +EXPORT_SYMBOL(_ebcasc); +EXPORT_SYMBOL(_ebc_tolower); +EXPORT_SYMBOL(_ebc_toupper); + diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S new file mode 100644 index 0000000000..49a11f6dd7 --- /dev/null +++ b/arch/s390/kernel/entry.S @@ -0,0 +1,669 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * S390 low-level entry points. + * + * Copyright IBM Corp. 1999, 2012 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), + * Hartmut Penner (hp@de.ibm.com), + * Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com), + */ + +#include <linux/export.h> +#include <linux/init.h> +#include <linux/linkage.h> +#include <asm/asm-extable.h> +#include <asm/alternative-asm.h> +#include <asm/processor.h> +#include <asm/cache.h> +#include <asm/dwarf.h> +#include <asm/errno.h> +#include <asm/ptrace.h> +#include <asm/thread_info.h> +#include <asm/asm-offsets.h> +#include <asm/unistd.h> +#include <asm/page.h> +#include <asm/sigp.h> +#include <asm/irq.h> +#include <asm/vx-insn.h> +#include <asm/setup.h> +#include <asm/nmi.h> +#include <asm/nospec-insn.h> + +_LPP_OFFSET = __LC_LPP + + .macro STBEAR address + ALTERNATIVE "nop", ".insn s,0xb2010000,\address", 193 + .endm + + .macro LBEAR address + ALTERNATIVE "nop", ".insn s,0xb2000000,\address", 193 + .endm + + .macro LPSWEY address,lpswe + ALTERNATIVE "b \lpswe; nopr", ".insn siy,0xeb0000000071,\address,0", 193 + .endm + + .macro MBEAR reg + ALTERNATIVE "brcl 0,0", __stringify(mvc __PT_LAST_BREAK(8,\reg),__LC_LAST_BREAK), 193 + .endm + + .macro CHECK_STACK savearea +#ifdef CONFIG_CHECK_STACK + tml %r15,THREAD_SIZE - CONFIG_STACK_GUARD + lghi %r14,\savearea + jz stack_overflow +#endif + .endm + + .macro CHECK_VMAP_STACK savearea,oklabel +#ifdef CONFIG_VMAP_STACK + lgr %r14,%r15 + nill %r14,0x10000 - THREAD_SIZE + oill %r14,STACK_INIT_OFFSET + clg %r14,__LC_KERNEL_STACK + je \oklabel + clg %r14,__LC_ASYNC_STACK + je \oklabel + clg %r14,__LC_MCCK_STACK + je \oklabel + clg %r14,__LC_NODAT_STACK + je \oklabel + clg %r14,__LC_RESTART_STACK + je \oklabel + lghi %r14,\savearea + j stack_overflow +#else + j \oklabel +#endif + .endm + + /* + * The TSTMSK macro generates a test-under-mask instruction by + * calculating the memory offset for the specified mask value. + * Mask value can be any constant. The macro shifts the mask + * value to calculate the memory offset for the test-under-mask + * instruction. + */ + .macro TSTMSK addr, mask, size=8, bytepos=0 + .if (\bytepos < \size) && (\mask >> 8) + .if (\mask & 0xff) + .error "Mask exceeds byte boundary" + .endif + TSTMSK \addr, "(\mask >> 8)", \size, "(\bytepos + 1)" + .exitm + .endif + .ifeq \mask + .error "Mask must not be zero" + .endif + off = \size - \bytepos - 1 + tm off+\addr, \mask + .endm + + .macro BPOFF + ALTERNATIVE "nop", ".insn rrf,0xb2e80000,0,0,12,0", 82 + .endm + + .macro BPON + ALTERNATIVE "nop", ".insn rrf,0xb2e80000,0,0,13,0", 82 + .endm + + .macro BPENTER tif_ptr,tif_mask + ALTERNATIVE "TSTMSK \tif_ptr,\tif_mask; jz .+8; .insn rrf,0xb2e80000,0,0,13,0", \ + "j .+12; nop; nop", 82 + .endm + + .macro BPEXIT tif_ptr,tif_mask + TSTMSK \tif_ptr,\tif_mask + ALTERNATIVE "jz .+8; .insn rrf,0xb2e80000,0,0,12,0", \ + "jnz .+8; .insn rrf,0xb2e80000,0,0,13,0", 82 + .endm + +#if IS_ENABLED(CONFIG_KVM) + /* + * The OUTSIDE macro jumps to the provided label in case the value + * in the provided register is outside of the provided range. The + * macro is useful for checking whether a PSW stored in a register + * pair points inside or outside of a block of instructions. + * @reg: register to check + * @start: start of the range + * @end: end of the range + * @outside_label: jump here if @reg is outside of [@start..@end) + */ + .macro OUTSIDE reg,start,end,outside_label + lgr %r14,\reg + larl %r13,\start + slgr %r14,%r13 + clgfrl %r14,.Lrange_size\@ + jhe \outside_label + .section .rodata, "a" + .balign 4 +.Lrange_size\@: + .long \end - \start + .previous + .endm + + .macro SIEEXIT + lg %r9,__SF_SIE_CONTROL(%r15) # get control block pointer + ni __SIE_PROG0C+3(%r9),0xfe # no longer in SIE + lctlg %c1,%c1,__LC_KERNEL_ASCE # load primary asce + larl %r9,sie_exit # skip forward to sie_exit + .endm +#endif + + .macro STACKLEAK_ERASE +#ifdef CONFIG_GCC_PLUGIN_STACKLEAK + brasl %r14,stackleak_erase_on_task_stack +#endif + .endm + + GEN_BR_THUNK %r14 + + .section .kprobes.text, "ax" +.Ldummy: + /* + * The following nop exists only in order to avoid that the next + * symbol starts at the beginning of the kprobes text section. + * In that case there would be several symbols at the same address. + * E.g. objdump would take an arbitrary symbol when disassembling + * the code. + * With the added nop in between this cannot happen. + */ + nop 0 + +/* + * Scheduler resume function, called by switch_to + * gpr2 = (task_struct *) prev + * gpr3 = (task_struct *) next + * Returns: + * gpr2 = prev + */ +SYM_FUNC_START(__switch_to) + stmg %r6,%r15,__SF_GPRS(%r15) # store gprs of prev task + lghi %r4,__TASK_stack + lghi %r1,__TASK_thread + llill %r5,STACK_INIT_OFFSET + stg %r15,__THREAD_ksp(%r1,%r2) # store kernel stack of prev + lg %r15,0(%r4,%r3) # start of kernel stack of next + agr %r15,%r5 # end of kernel stack of next + stg %r3,__LC_CURRENT # store task struct of next + stg %r15,__LC_KERNEL_STACK # store end of kernel stack + lg %r15,__THREAD_ksp(%r1,%r3) # load kernel stack of next + aghi %r3,__TASK_pid + mvc __LC_CURRENT_PID(4,%r0),0(%r3) # store pid of next + lmg %r6,%r15,__SF_GPRS(%r15) # load gprs of next task + ALTERNATIVE "nop", "lpp _LPP_OFFSET", 40 + BR_EX %r14 +SYM_FUNC_END(__switch_to) + +#if IS_ENABLED(CONFIG_KVM) +/* + * __sie64a calling convention: + * %r2 pointer to sie control block phys + * %r3 pointer to sie control block virt + * %r4 guest register save area + */ +SYM_FUNC_START(__sie64a) + stmg %r6,%r14,__SF_GPRS(%r15) # save kernel registers + lg %r12,__LC_CURRENT + stg %r2,__SF_SIE_CONTROL_PHYS(%r15) # save sie block physical.. + stg %r3,__SF_SIE_CONTROL(%r15) # ...and virtual addresses + stg %r4,__SF_SIE_SAVEAREA(%r15) # save guest register save area + xc __SF_SIE_REASON(8,%r15),__SF_SIE_REASON(%r15) # reason code = 0 + mvc __SF_SIE_FLAGS(8,%r15),__TI_flags(%r12) # copy thread flags + lmg %r0,%r13,0(%r4) # load guest gprs 0-13 + lg %r14,__LC_GMAP # get gmap pointer + ltgr %r14,%r14 + jz .Lsie_gmap + lctlg %c1,%c1,__GMAP_ASCE(%r14) # load primary asce +.Lsie_gmap: + lg %r14,__SF_SIE_CONTROL(%r15) # get control block pointer + oi __SIE_PROG0C+3(%r14),1 # we are going into SIE now + tm __SIE_PROG20+3(%r14),3 # last exit... + jnz .Lsie_skip + TSTMSK __LC_CPU_FLAGS,_CIF_FPU + jo .Lsie_skip # exit if fp/vx regs changed + lg %r14,__SF_SIE_CONTROL_PHYS(%r15) # get sie block phys addr + BPEXIT __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST +.Lsie_entry: + sie 0(%r14) +# Let the next instruction be NOP to avoid triggering a machine check +# and handling it in a guest as result of the instruction execution. + nopr 7 +.Lsie_leave: + BPOFF + BPENTER __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST +.Lsie_skip: + lg %r14,__SF_SIE_CONTROL(%r15) # get control block pointer + ni __SIE_PROG0C+3(%r14),0xfe # no longer in SIE + lctlg %c1,%c1,__LC_KERNEL_ASCE # load primary asce +.Lsie_done: +# some program checks are suppressing. C code (e.g. do_protection_exception) +# will rewind the PSW by the ILC, which is often 4 bytes in case of SIE. There +# are some corner cases (e.g. runtime instrumentation) where ILC is unpredictable. +# Other instructions between __sie64a and .Lsie_done should not cause program +# interrupts. So lets use 3 nops as a landing pad for all possible rewinds. +.Lrewind_pad6: + nopr 7 +.Lrewind_pad4: + nopr 7 +.Lrewind_pad2: + nopr 7 +SYM_INNER_LABEL(sie_exit, SYM_L_GLOBAL) + lg %r14,__SF_SIE_SAVEAREA(%r15) # load guest register save area + stmg %r0,%r13,0(%r14) # save guest gprs 0-13 + xgr %r0,%r0 # clear guest registers to + xgr %r1,%r1 # prevent speculative use + xgr %r3,%r3 + xgr %r4,%r4 + xgr %r5,%r5 + lmg %r6,%r14,__SF_GPRS(%r15) # restore kernel registers + lg %r2,__SF_SIE_REASON(%r15) # return exit reason code + BR_EX %r14 +.Lsie_fault: + lghi %r14,-EFAULT + stg %r14,__SF_SIE_REASON(%r15) # set exit reason code + j sie_exit + + EX_TABLE(.Lrewind_pad6,.Lsie_fault) + EX_TABLE(.Lrewind_pad4,.Lsie_fault) + EX_TABLE(.Lrewind_pad2,.Lsie_fault) + EX_TABLE(sie_exit,.Lsie_fault) +SYM_FUNC_END(__sie64a) +EXPORT_SYMBOL(__sie64a) +EXPORT_SYMBOL(sie_exit) +#endif + +/* + * SVC interrupt handler routine. System calls are synchronous events and + * are entered with interrupts disabled. + */ + +SYM_CODE_START(system_call) + stpt __LC_SYS_ENTER_TIMER + stmg %r8,%r15,__LC_SAVE_AREA_SYNC + BPOFF + lghi %r14,0 +.Lsysc_per: + STBEAR __LC_LAST_BREAK + lctlg %c1,%c1,__LC_KERNEL_ASCE + lg %r15,__LC_KERNEL_STACK + xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) + stmg %r0,%r7,STACK_FRAME_OVERHEAD+__PT_R0(%r15) + # clear user controlled register to prevent speculative use + xgr %r0,%r0 + xgr %r1,%r1 + xgr %r4,%r4 + xgr %r5,%r5 + xgr %r6,%r6 + xgr %r7,%r7 + xgr %r8,%r8 + xgr %r9,%r9 + xgr %r10,%r10 + xgr %r11,%r11 + la %r2,STACK_FRAME_OVERHEAD(%r15) # pointer to pt_regs + mvc __PT_R8(64,%r2),__LC_SAVE_AREA_SYNC + MBEAR %r2 + lgr %r3,%r14 + brasl %r14,__do_syscall + STACKLEAK_ERASE + lctlg %c1,%c1,__LC_USER_ASCE + mvc __LC_RETURN_PSW(16),STACK_FRAME_OVERHEAD+__PT_PSW(%r15) + BPON + LBEAR STACK_FRAME_OVERHEAD+__PT_LAST_BREAK(%r15) + lmg %r0,%r15,STACK_FRAME_OVERHEAD+__PT_R0(%r15) + stpt __LC_EXIT_TIMER + LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE +SYM_CODE_END(system_call) + +# +# a new process exits the kernel with ret_from_fork +# +SYM_CODE_START(ret_from_fork) + lgr %r3,%r11 + brasl %r14,__ret_from_fork + STACKLEAK_ERASE + lctlg %c1,%c1,__LC_USER_ASCE + mvc __LC_RETURN_PSW(16),STACK_FRAME_OVERHEAD+__PT_PSW(%r15) + BPON + LBEAR STACK_FRAME_OVERHEAD+__PT_LAST_BREAK(%r15) + lmg %r0,%r15,STACK_FRAME_OVERHEAD+__PT_R0(%r15) + stpt __LC_EXIT_TIMER + LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE +SYM_CODE_END(ret_from_fork) + +/* + * Program check handler routine + */ + +SYM_CODE_START(pgm_check_handler) + stpt __LC_SYS_ENTER_TIMER + BPOFF + stmg %r8,%r15,__LC_SAVE_AREA_SYNC + lghi %r10,0 + lmg %r8,%r9,__LC_PGM_OLD_PSW + tmhh %r8,0x0001 # coming from user space? + jno .Lpgm_skip_asce + lctlg %c1,%c1,__LC_KERNEL_ASCE + j 3f # -> fault in user space +.Lpgm_skip_asce: +#if IS_ENABLED(CONFIG_KVM) + # cleanup critical section for program checks in __sie64a + OUTSIDE %r9,.Lsie_gmap,.Lsie_done,1f + BPENTER __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST + SIEEXIT + lghi %r10,_PIF_GUEST_FAULT +#endif +1: tmhh %r8,0x4000 # PER bit set in old PSW ? + jnz 2f # -> enabled, can't be a double fault + tm __LC_PGM_ILC+3,0x80 # check for per exception + jnz .Lpgm_svcper # -> single stepped svc +2: CHECK_STACK __LC_SAVE_AREA_SYNC + aghi %r15,-(STACK_FRAME_OVERHEAD + __PT_SIZE) + # CHECK_VMAP_STACK branches to stack_overflow or 4f + CHECK_VMAP_STACK __LC_SAVE_AREA_SYNC,4f +3: lg %r15,__LC_KERNEL_STACK +4: la %r11,STACK_FRAME_OVERHEAD(%r15) + stg %r10,__PT_FLAGS(%r11) + xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) + stmg %r0,%r7,__PT_R0(%r11) + mvc __PT_R8(64,%r11),__LC_SAVE_AREA_SYNC + mvc __PT_LAST_BREAK(8,%r11),__LC_PGM_LAST_BREAK + stmg %r8,%r9,__PT_PSW(%r11) + + # clear user controlled registers to prevent speculative use + xgr %r0,%r0 + xgr %r1,%r1 + xgr %r3,%r3 + xgr %r4,%r4 + xgr %r5,%r5 + xgr %r6,%r6 + xgr %r7,%r7 + lgr %r2,%r11 + brasl %r14,__do_pgm_check + tmhh %r8,0x0001 # returning to user space? + jno .Lpgm_exit_kernel + STACKLEAK_ERASE + lctlg %c1,%c1,__LC_USER_ASCE + BPON + stpt __LC_EXIT_TIMER +.Lpgm_exit_kernel: + mvc __LC_RETURN_PSW(16),STACK_FRAME_OVERHEAD+__PT_PSW(%r15) + LBEAR STACK_FRAME_OVERHEAD+__PT_LAST_BREAK(%r15) + lmg %r0,%r15,STACK_FRAME_OVERHEAD+__PT_R0(%r15) + LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE + +# +# single stepped system call +# +.Lpgm_svcper: + mvc __LC_RETURN_PSW(8),__LC_SVC_NEW_PSW + larl %r14,.Lsysc_per + stg %r14,__LC_RETURN_PSW+8 + lghi %r14,1 + LBEAR __LC_PGM_LAST_BREAK + LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE # branch to .Lsysc_per +SYM_CODE_END(pgm_check_handler) + +/* + * Interrupt handler macro used for external and IO interrupts. + */ +.macro INT_HANDLER name,lc_old_psw,handler +SYM_CODE_START(\name) + stckf __LC_INT_CLOCK + stpt __LC_SYS_ENTER_TIMER + STBEAR __LC_LAST_BREAK + BPOFF + stmg %r8,%r15,__LC_SAVE_AREA_ASYNC + lmg %r8,%r9,\lc_old_psw + tmhh %r8,0x0001 # interrupting from user ? + jnz 1f +#if IS_ENABLED(CONFIG_KVM) + OUTSIDE %r9,.Lsie_gmap,.Lsie_done,0f + BPENTER __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST + SIEEXIT +#endif +0: CHECK_STACK __LC_SAVE_AREA_ASYNC + aghi %r15,-(STACK_FRAME_OVERHEAD + __PT_SIZE) + j 2f +1: lctlg %c1,%c1,__LC_KERNEL_ASCE + lg %r15,__LC_KERNEL_STACK +2: xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) + la %r11,STACK_FRAME_OVERHEAD(%r15) + stmg %r0,%r7,__PT_R0(%r11) + # clear user controlled registers to prevent speculative use + xgr %r0,%r0 + xgr %r1,%r1 + xgr %r3,%r3 + xgr %r4,%r4 + xgr %r5,%r5 + xgr %r6,%r6 + xgr %r7,%r7 + xgr %r10,%r10 + xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11) + mvc __PT_R8(64,%r11),__LC_SAVE_AREA_ASYNC + MBEAR %r11 + stmg %r8,%r9,__PT_PSW(%r11) + lgr %r2,%r11 # pass pointer to pt_regs + brasl %r14,\handler + mvc __LC_RETURN_PSW(16),__PT_PSW(%r11) + tmhh %r8,0x0001 # returning to user ? + jno 2f + STACKLEAK_ERASE + lctlg %c1,%c1,__LC_USER_ASCE + BPON + stpt __LC_EXIT_TIMER +2: LBEAR __PT_LAST_BREAK(%r11) + lmg %r0,%r15,__PT_R0(%r11) + LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE +SYM_CODE_END(\name) +.endm + +INT_HANDLER ext_int_handler,__LC_EXT_OLD_PSW,do_ext_irq +INT_HANDLER io_int_handler,__LC_IO_OLD_PSW,do_io_irq + +/* + * Load idle PSW. + */ +SYM_FUNC_START(psw_idle) + stg %r14,(__SF_GPRS+8*8)(%r15) + stg %r3,__SF_EMPTY(%r15) + larl %r1,psw_idle_exit + stg %r1,__SF_EMPTY+8(%r15) + larl %r1,smp_cpu_mtid + llgf %r1,0(%r1) + ltgr %r1,%r1 + jz .Lpsw_idle_stcctm + .insn rsy,0xeb0000000017,%r1,5,__MT_CYCLES_ENTER(%r2) +.Lpsw_idle_stcctm: + oi __LC_CPU_FLAGS+7,_CIF_ENABLED_WAIT + BPON + stckf __CLOCK_IDLE_ENTER(%r2) + stpt __TIMER_IDLE_ENTER(%r2) + lpswe __SF_EMPTY(%r15) +SYM_INNER_LABEL(psw_idle_exit, SYM_L_GLOBAL) + BR_EX %r14 +SYM_FUNC_END(psw_idle) + +/* + * Machine check handler routines + */ +SYM_CODE_START(mcck_int_handler) + BPOFF + la %r1,4095 # validate r1 + spt __LC_CPU_TIMER_SAVE_AREA-4095(%r1) # validate cpu timer + LBEAR __LC_LAST_BREAK_SAVE_AREA-4095(%r1) # validate bear + lmg %r0,%r15,__LC_GPREGS_SAVE_AREA # validate gprs + lmg %r8,%r9,__LC_MCK_OLD_PSW + TSTMSK __LC_MCCK_CODE,MCCK_CODE_SYSTEM_DAMAGE + jo .Lmcck_panic # yes -> rest of mcck code invalid + TSTMSK __LC_MCCK_CODE,MCCK_CODE_CR_VALID + jno .Lmcck_panic # control registers invalid -> panic + lctlg %c0,%c15,__LC_CREGS_SAVE_AREA # validate ctl regs + ptlb + lghi %r14,__LC_CPU_TIMER_SAVE_AREA + mvc __LC_MCCK_ENTER_TIMER(8),0(%r14) + TSTMSK __LC_MCCK_CODE,MCCK_CODE_CPU_TIMER_VALID + jo 3f + la %r14,__LC_SYS_ENTER_TIMER + clc 0(8,%r14),__LC_EXIT_TIMER + jl 1f + la %r14,__LC_EXIT_TIMER +1: clc 0(8,%r14),__LC_LAST_UPDATE_TIMER + jl 2f + la %r14,__LC_LAST_UPDATE_TIMER +2: spt 0(%r14) + mvc __LC_MCCK_ENTER_TIMER(8),0(%r14) +3: TSTMSK __LC_MCCK_CODE,MCCK_CODE_PSW_MWP_VALID + jno .Lmcck_panic + tmhh %r8,0x0001 # interrupting from user ? + jnz .Lmcck_user + TSTMSK __LC_MCCK_CODE,MCCK_CODE_PSW_IA_VALID + jno .Lmcck_panic +#if IS_ENABLED(CONFIG_KVM) + OUTSIDE %r9,.Lsie_gmap,.Lsie_done,.Lmcck_user + OUTSIDE %r9,.Lsie_entry,.Lsie_leave,4f + oi __LC_CPU_FLAGS+7, _CIF_MCCK_GUEST +4: BPENTER __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST + SIEEXIT +#endif +.Lmcck_user: + lg %r15,__LC_MCCK_STACK + la %r11,STACK_FRAME_OVERHEAD(%r15) + stctg %c1,%c1,__PT_CR1(%r11) + lctlg %c1,%c1,__LC_KERNEL_ASCE + xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) + lghi %r14,__LC_GPREGS_SAVE_AREA+64 + stmg %r0,%r7,__PT_R0(%r11) + # clear user controlled registers to prevent speculative use + xgr %r0,%r0 + xgr %r1,%r1 + xgr %r3,%r3 + xgr %r4,%r4 + xgr %r5,%r5 + xgr %r6,%r6 + xgr %r7,%r7 + xgr %r10,%r10 + mvc __PT_R8(64,%r11),0(%r14) + stmg %r8,%r9,__PT_PSW(%r11) + xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11) + xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) + lgr %r2,%r11 # pass pointer to pt_regs + brasl %r14,s390_do_machine_check + lctlg %c1,%c1,__PT_CR1(%r11) + lmg %r0,%r10,__PT_R0(%r11) + mvc __LC_RETURN_MCCK_PSW(16),__PT_PSW(%r11) # move return PSW + tm __LC_RETURN_MCCK_PSW+1,0x01 # returning to user ? + jno 0f + BPON + stpt __LC_EXIT_TIMER +0: ALTERNATIVE "nop", __stringify(lghi %r12,__LC_LAST_BREAK_SAVE_AREA),193 + LBEAR 0(%r12) + lmg %r11,%r15,__PT_R11(%r11) + LPSWEY __LC_RETURN_MCCK_PSW,__LC_RETURN_MCCK_LPSWE + +.Lmcck_panic: + /* + * Iterate over all possible CPU addresses in the range 0..0xffff + * and stop each CPU using signal processor. Use compare and swap + * to allow just one CPU-stopper and prevent concurrent CPUs from + * stopping each other while leaving the others running. + */ + lhi %r5,0 + lhi %r6,1 + larl %r7,stop_lock + cs %r5,%r6,0(%r7) # single CPU-stopper only + jnz 4f + larl %r7,this_cpu + stap 0(%r7) # this CPU address + lh %r4,0(%r7) + nilh %r4,0 + lhi %r0,1 + sll %r0,16 # CPU counter + lhi %r3,0 # next CPU address +0: cr %r3,%r4 + je 2f +1: sigp %r1,%r3,SIGP_STOP # stop next CPU + brc SIGP_CC_BUSY,1b +2: ahi %r3,1 + brct %r0,0b +3: sigp %r1,%r4,SIGP_STOP # stop this CPU + brc SIGP_CC_BUSY,3b +4: j 4b +SYM_CODE_END(mcck_int_handler) + +SYM_CODE_START(restart_int_handler) + ALTERNATIVE "nop", "lpp _LPP_OFFSET", 40 + stg %r15,__LC_SAVE_AREA_RESTART + TSTMSK __LC_RESTART_FLAGS,RESTART_FLAG_CTLREGS,4 + jz 0f + lctlg %c0,%c15,__LC_CREGS_SAVE_AREA +0: larl %r15,daton_psw + lpswe 0(%r15) # turn dat on, keep irqs off +.Ldaton: + lg %r15,__LC_RESTART_STACK + xc STACK_FRAME_OVERHEAD(__PT_SIZE,%r15),STACK_FRAME_OVERHEAD(%r15) + stmg %r0,%r14,STACK_FRAME_OVERHEAD+__PT_R0(%r15) + mvc STACK_FRAME_OVERHEAD+__PT_R15(8,%r15),__LC_SAVE_AREA_RESTART + mvc STACK_FRAME_OVERHEAD+__PT_PSW(16,%r15),__LC_RST_OLD_PSW + xc 0(STACK_FRAME_OVERHEAD,%r15),0(%r15) + lg %r1,__LC_RESTART_FN # load fn, parm & source cpu + lg %r2,__LC_RESTART_DATA + lgf %r3,__LC_RESTART_SOURCE + ltgr %r3,%r3 # test source cpu address + jm 1f # negative -> skip source stop +0: sigp %r4,%r3,SIGP_SENSE # sigp sense to source cpu + brc 10,0b # wait for status stored +1: basr %r14,%r1 # call function + stap __SF_EMPTY(%r15) # store cpu address + llgh %r3,__SF_EMPTY(%r15) +2: sigp %r4,%r3,SIGP_STOP # sigp stop to current cpu + brc 2,2b +3: j 3b +SYM_CODE_END(restart_int_handler) + + .section .kprobes.text, "ax" + +#if defined(CONFIG_CHECK_STACK) || defined(CONFIG_VMAP_STACK) +/* + * The synchronous or the asynchronous stack overflowed. We are dead. + * No need to properly save the registers, we are going to panic anyway. + * Setup a pt_regs so that show_trace can provide a good call trace. + */ +SYM_CODE_START(stack_overflow) + lg %r15,__LC_NODAT_STACK # change to panic stack + la %r11,STACK_FRAME_OVERHEAD(%r15) + stmg %r0,%r7,__PT_R0(%r11) + stmg %r8,%r9,__PT_PSW(%r11) + mvc __PT_R8(64,%r11),0(%r14) + stg %r10,__PT_ORIG_GPR2(%r11) # store last break to orig_gpr2 + xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) + lgr %r2,%r11 # pass pointer to pt_regs + jg kernel_stack_overflow +SYM_CODE_END(stack_overflow) +#endif + + .section .data, "aw" + .balign 4 +SYM_DATA_LOCAL(stop_lock, .long 0) +SYM_DATA_LOCAL(this_cpu, .short 0) + .balign 8 +SYM_DATA_START_LOCAL(daton_psw) + .quad PSW_KERNEL_BITS + .quad .Ldaton +SYM_DATA_END(daton_psw) + + .section .rodata, "a" +#define SYSCALL(esame,emu) .quad __s390x_ ## esame +SYM_DATA_START(sys_call_table) +#include "asm/syscall_table.h" +SYM_DATA_END(sys_call_table) +#undef SYSCALL + +#ifdef CONFIG_COMPAT + +#define SYSCALL(esame,emu) .quad __s390_ ## emu +SYM_DATA_START(sys_call_table_emu) +#include "asm/syscall_table.h" +SYM_DATA_END(sys_call_table_emu) +#undef SYSCALL +#endif diff --git a/arch/s390/kernel/entry.h b/arch/s390/kernel/entry.h new file mode 100644 index 0000000000..9f41853f36 --- /dev/null +++ b/arch/s390/kernel/entry.h @@ -0,0 +1,75 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _ENTRY_H +#define _ENTRY_H + +#include <linux/percpu.h> +#include <linux/types.h> +#include <linux/signal.h> +#include <asm/extable.h> +#include <asm/ptrace.h> +#include <asm/idle.h> + +extern void *restart_stack; + +void system_call(void); +void pgm_check_handler(void); +void ext_int_handler(void); +void io_int_handler(void); +void mcck_int_handler(void); +void restart_int_handler(void); +void early_pgm_check_handler(void); + +void __ret_from_fork(struct task_struct *prev, struct pt_regs *regs); +void __do_pgm_check(struct pt_regs *regs); +void __do_syscall(struct pt_regs *regs, int per_trap); +void __do_early_pgm_check(struct pt_regs *regs); + +void do_protection_exception(struct pt_regs *regs); +void do_dat_exception(struct pt_regs *regs); +void do_secure_storage_access(struct pt_regs *regs); +void do_non_secure_storage_access(struct pt_regs *regs); +void do_secure_storage_violation(struct pt_regs *regs); +void do_report_trap(struct pt_regs *regs, int si_signo, int si_code, char *str); +void kernel_stack_overflow(struct pt_regs * regs); +void handle_signal32(struct ksignal *ksig, sigset_t *oldset, + struct pt_regs *regs); + +void do_io_irq(struct pt_regs *regs); +void do_ext_irq(struct pt_regs *regs); +void do_restart(void *arg); +void __init startup_init(void); +void die(struct pt_regs *regs, const char *str); +int setup_profiling_timer(unsigned int multiplier); +unsigned long prepare_ftrace_return(unsigned long parent, unsigned long sp, unsigned long ip); + +struct s390_mmap_arg_struct; +struct fadvise64_64_args; +struct old_sigaction; + +long sys_rt_sigreturn(void); +long sys_sigreturn(void); + +long sys_s390_personality(unsigned int personality); +long sys_s390_runtime_instr(int command, int signum); +long sys_s390_guarded_storage(int command, struct gs_cb __user *); +long sys_s390_pci_mmio_write(unsigned long, const void __user *, size_t); +long sys_s390_pci_mmio_read(unsigned long, void __user *, size_t); +long sys_s390_sthyi(unsigned long function_code, void __user *buffer, u64 __user *return_code, unsigned long flags); + +DECLARE_PER_CPU(u64, mt_cycles[8]); + +unsigned long stack_alloc(void); +void stack_free(unsigned long stack); + +extern char kprobes_insn_page[]; + +extern char _samode31[], _eamode31[]; +extern char _stext_amode31[], _etext_amode31[]; +extern struct exception_table_entry _start_amode31_ex_table[]; +extern struct exception_table_entry _stop_amode31_ex_table[]; + +#define __amode31_data __section(".amode31.data") +#define __amode31_ref __section(".amode31.refs") +extern long _start_amode31_refs[], _end_amode31_refs[]; + +#endif /* _ENTRY_H */ diff --git a/arch/s390/kernel/fpu.c b/arch/s390/kernel/fpu.c new file mode 100644 index 0000000000..4666b29ac8 --- /dev/null +++ b/arch/s390/kernel/fpu.c @@ -0,0 +1,264 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * In-kernel vector facility support functions + * + * Copyright IBM Corp. 2015 + * Author(s): Hendrik Brueckner <brueckner@linux.vnet.ibm.com> + */ +#include <linux/kernel.h> +#include <linux/cpu.h> +#include <linux/sched.h> +#include <asm/fpu/types.h> +#include <asm/fpu/api.h> +#include <asm/vx-insn.h> + +void __kernel_fpu_begin(struct kernel_fpu *state, u32 flags) +{ + /* + * Limit the save to the FPU/vector registers already + * in use by the previous context + */ + flags &= state->mask; + + if (flags & KERNEL_FPC) + /* Save floating point control */ + asm volatile("stfpc %0" : "=Q" (state->fpc)); + + if (!MACHINE_HAS_VX) { + if (flags & KERNEL_VXR_V0V7) { + /* Save floating-point registers */ + asm volatile("std 0,%0" : "=Q" (state->fprs[0])); + asm volatile("std 1,%0" : "=Q" (state->fprs[1])); + asm volatile("std 2,%0" : "=Q" (state->fprs[2])); + asm volatile("std 3,%0" : "=Q" (state->fprs[3])); + asm volatile("std 4,%0" : "=Q" (state->fprs[4])); + asm volatile("std 5,%0" : "=Q" (state->fprs[5])); + asm volatile("std 6,%0" : "=Q" (state->fprs[6])); + asm volatile("std 7,%0" : "=Q" (state->fprs[7])); + asm volatile("std 8,%0" : "=Q" (state->fprs[8])); + asm volatile("std 9,%0" : "=Q" (state->fprs[9])); + asm volatile("std 10,%0" : "=Q" (state->fprs[10])); + asm volatile("std 11,%0" : "=Q" (state->fprs[11])); + asm volatile("std 12,%0" : "=Q" (state->fprs[12])); + asm volatile("std 13,%0" : "=Q" (state->fprs[13])); + asm volatile("std 14,%0" : "=Q" (state->fprs[14])); + asm volatile("std 15,%0" : "=Q" (state->fprs[15])); + } + return; + } + + /* Test and save vector registers */ + asm volatile ( + /* + * Test if any vector register must be saved and, if so, + * test if all register can be saved. + */ + " la 1,%[vxrs]\n" /* load save area */ + " tmll %[m],30\n" /* KERNEL_VXR */ + " jz 7f\n" /* no work -> done */ + " jo 5f\n" /* -> save V0..V31 */ + /* + * Test for special case KERNEL_FPU_MID only. In this + * case a vstm V8..V23 is the best instruction + */ + " chi %[m],12\n" /* KERNEL_VXR_MID */ + " jne 0f\n" /* -> save V8..V23 */ + " VSTM 8,23,128,1\n" /* vstm %v8,%v23,128(%r1) */ + " j 7f\n" + /* Test and save the first half of 16 vector registers */ + "0: tmll %[m],6\n" /* KERNEL_VXR_LOW */ + " jz 3f\n" /* -> KERNEL_VXR_HIGH */ + " jo 2f\n" /* 11 -> save V0..V15 */ + " brc 2,1f\n" /* 10 -> save V8..V15 */ + " VSTM 0,7,0,1\n" /* vstm %v0,%v7,0(%r1) */ + " j 3f\n" + "1: VSTM 8,15,128,1\n" /* vstm %v8,%v15,128(%r1) */ + " j 3f\n" + "2: VSTM 0,15,0,1\n" /* vstm %v0,%v15,0(%r1) */ + /* Test and save the second half of 16 vector registers */ + "3: tmll %[m],24\n" /* KERNEL_VXR_HIGH */ + " jz 7f\n" + " jo 6f\n" /* 11 -> save V16..V31 */ + " brc 2,4f\n" /* 10 -> save V24..V31 */ + " VSTM 16,23,256,1\n" /* vstm %v16,%v23,256(%r1) */ + " j 7f\n" + "4: VSTM 24,31,384,1\n" /* vstm %v24,%v31,384(%r1) */ + " j 7f\n" + "5: VSTM 0,15,0,1\n" /* vstm %v0,%v15,0(%r1) */ + "6: VSTM 16,31,256,1\n" /* vstm %v16,%v31,256(%r1) */ + "7:" + : [vxrs] "=Q" (*(struct vx_array *) &state->vxrs) + : [m] "d" (flags) + : "1", "cc"); +} +EXPORT_SYMBOL(__kernel_fpu_begin); + +void __kernel_fpu_end(struct kernel_fpu *state, u32 flags) +{ + /* + * Limit the restore to the FPU/vector registers of the + * previous context that have been overwritte by the + * current context + */ + flags &= state->mask; + + if (flags & KERNEL_FPC) + /* Restore floating-point controls */ + asm volatile("lfpc %0" : : "Q" (state->fpc)); + + if (!MACHINE_HAS_VX) { + if (flags & KERNEL_VXR_V0V7) { + /* Restore floating-point registers */ + asm volatile("ld 0,%0" : : "Q" (state->fprs[0])); + asm volatile("ld 1,%0" : : "Q" (state->fprs[1])); + asm volatile("ld 2,%0" : : "Q" (state->fprs[2])); + asm volatile("ld 3,%0" : : "Q" (state->fprs[3])); + asm volatile("ld 4,%0" : : "Q" (state->fprs[4])); + asm volatile("ld 5,%0" : : "Q" (state->fprs[5])); + asm volatile("ld 6,%0" : : "Q" (state->fprs[6])); + asm volatile("ld 7,%0" : : "Q" (state->fprs[7])); + asm volatile("ld 8,%0" : : "Q" (state->fprs[8])); + asm volatile("ld 9,%0" : : "Q" (state->fprs[9])); + asm volatile("ld 10,%0" : : "Q" (state->fprs[10])); + asm volatile("ld 11,%0" : : "Q" (state->fprs[11])); + asm volatile("ld 12,%0" : : "Q" (state->fprs[12])); + asm volatile("ld 13,%0" : : "Q" (state->fprs[13])); + asm volatile("ld 14,%0" : : "Q" (state->fprs[14])); + asm volatile("ld 15,%0" : : "Q" (state->fprs[15])); + } + return; + } + + /* Test and restore (load) vector registers */ + asm volatile ( + /* + * Test if any vector register must be loaded and, if so, + * test if all registers can be loaded at once. + */ + " la 1,%[vxrs]\n" /* load restore area */ + " tmll %[m],30\n" /* KERNEL_VXR */ + " jz 7f\n" /* no work -> done */ + " jo 5f\n" /* -> restore V0..V31 */ + /* + * Test for special case KERNEL_FPU_MID only. In this + * case a vlm V8..V23 is the best instruction + */ + " chi %[m],12\n" /* KERNEL_VXR_MID */ + " jne 0f\n" /* -> restore V8..V23 */ + " VLM 8,23,128,1\n" /* vlm %v8,%v23,128(%r1) */ + " j 7f\n" + /* Test and restore the first half of 16 vector registers */ + "0: tmll %[m],6\n" /* KERNEL_VXR_LOW */ + " jz 3f\n" /* -> KERNEL_VXR_HIGH */ + " jo 2f\n" /* 11 -> restore V0..V15 */ + " brc 2,1f\n" /* 10 -> restore V8..V15 */ + " VLM 0,7,0,1\n" /* vlm %v0,%v7,0(%r1) */ + " j 3f\n" + "1: VLM 8,15,128,1\n" /* vlm %v8,%v15,128(%r1) */ + " j 3f\n" + "2: VLM 0,15,0,1\n" /* vlm %v0,%v15,0(%r1) */ + /* Test and restore the second half of 16 vector registers */ + "3: tmll %[m],24\n" /* KERNEL_VXR_HIGH */ + " jz 7f\n" + " jo 6f\n" /* 11 -> restore V16..V31 */ + " brc 2,4f\n" /* 10 -> restore V24..V31 */ + " VLM 16,23,256,1\n" /* vlm %v16,%v23,256(%r1) */ + " j 7f\n" + "4: VLM 24,31,384,1\n" /* vlm %v24,%v31,384(%r1) */ + " j 7f\n" + "5: VLM 0,15,0,1\n" /* vlm %v0,%v15,0(%r1) */ + "6: VLM 16,31,256,1\n" /* vlm %v16,%v31,256(%r1) */ + "7:" + : [vxrs] "=Q" (*(struct vx_array *) &state->vxrs) + : [m] "d" (flags) + : "1", "cc"); +} +EXPORT_SYMBOL(__kernel_fpu_end); + +void __load_fpu_regs(void) +{ + struct fpu *state = ¤t->thread.fpu; + unsigned long *regs = current->thread.fpu.regs; + + asm volatile("lfpc %0" : : "Q" (state->fpc)); + if (likely(MACHINE_HAS_VX)) { + asm volatile("lgr 1,%0\n" + "VLM 0,15,0,1\n" + "VLM 16,31,256,1\n" + : + : "d" (regs) + : "1", "cc", "memory"); + } else { + asm volatile("ld 0,%0" : : "Q" (regs[0])); + asm volatile("ld 1,%0" : : "Q" (regs[1])); + asm volatile("ld 2,%0" : : "Q" (regs[2])); + asm volatile("ld 3,%0" : : "Q" (regs[3])); + asm volatile("ld 4,%0" : : "Q" (regs[4])); + asm volatile("ld 5,%0" : : "Q" (regs[5])); + asm volatile("ld 6,%0" : : "Q" (regs[6])); + asm volatile("ld 7,%0" : : "Q" (regs[7])); + asm volatile("ld 8,%0" : : "Q" (regs[8])); + asm volatile("ld 9,%0" : : "Q" (regs[9])); + asm volatile("ld 10,%0" : : "Q" (regs[10])); + asm volatile("ld 11,%0" : : "Q" (regs[11])); + asm volatile("ld 12,%0" : : "Q" (regs[12])); + asm volatile("ld 13,%0" : : "Q" (regs[13])); + asm volatile("ld 14,%0" : : "Q" (regs[14])); + asm volatile("ld 15,%0" : : "Q" (regs[15])); + } + clear_cpu_flag(CIF_FPU); +} +EXPORT_SYMBOL(__load_fpu_regs); + +void load_fpu_regs(void) +{ + raw_local_irq_disable(); + __load_fpu_regs(); + raw_local_irq_enable(); +} +EXPORT_SYMBOL(load_fpu_regs); + +void save_fpu_regs(void) +{ + unsigned long flags, *regs; + struct fpu *state; + + local_irq_save(flags); + + if (test_cpu_flag(CIF_FPU)) + goto out; + + state = ¤t->thread.fpu; + regs = current->thread.fpu.regs; + + asm volatile("stfpc %0" : "=Q" (state->fpc)); + if (likely(MACHINE_HAS_VX)) { + asm volatile("lgr 1,%0\n" + "VSTM 0,15,0,1\n" + "VSTM 16,31,256,1\n" + : + : "d" (regs) + : "1", "cc", "memory"); + } else { + asm volatile("std 0,%0" : "=Q" (regs[0])); + asm volatile("std 1,%0" : "=Q" (regs[1])); + asm volatile("std 2,%0" : "=Q" (regs[2])); + asm volatile("std 3,%0" : "=Q" (regs[3])); + asm volatile("std 4,%0" : "=Q" (regs[4])); + asm volatile("std 5,%0" : "=Q" (regs[5])); + asm volatile("std 6,%0" : "=Q" (regs[6])); + asm volatile("std 7,%0" : "=Q" (regs[7])); + asm volatile("std 8,%0" : "=Q" (regs[8])); + asm volatile("std 9,%0" : "=Q" (regs[9])); + asm volatile("std 10,%0" : "=Q" (regs[10])); + asm volatile("std 11,%0" : "=Q" (regs[11])); + asm volatile("std 12,%0" : "=Q" (regs[12])); + asm volatile("std 13,%0" : "=Q" (regs[13])); + asm volatile("std 14,%0" : "=Q" (regs[14])); + asm volatile("std 15,%0" : "=Q" (regs[15])); + } + set_cpu_flag(CIF_FPU); +out: + local_irq_restore(flags); +} +EXPORT_SYMBOL(save_fpu_regs); diff --git a/arch/s390/kernel/ftrace.c b/arch/s390/kernel/ftrace.c new file mode 100644 index 0000000000..c46381ea04 --- /dev/null +++ b/arch/s390/kernel/ftrace.c @@ -0,0 +1,340 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Dynamic function tracer architecture backend. + * + * Copyright IBM Corp. 2009,2014 + * + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> + */ + +#include <linux/moduleloader.h> +#include <linux/hardirq.h> +#include <linux/uaccess.h> +#include <linux/ftrace.h> +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/kprobes.h> +#include <trace/syscall.h> +#include <asm/asm-offsets.h> +#include <asm/text-patching.h> +#include <asm/cacheflush.h> +#include <asm/ftrace.lds.h> +#include <asm/nospec-branch.h> +#include <asm/set_memory.h> +#include "entry.h" +#include "ftrace.h" + +/* + * To generate function prologue either gcc's hotpatch feature (since gcc 4.8) + * or a combination of -pg -mrecord-mcount -mnop-mcount -mfentry flags + * (since gcc 9 / clang 10) is used. + * In both cases the original and also the disabled function prologue contains + * only a single six byte instruction and looks like this: + * > brcl 0,0 # offset 0 + * To enable ftrace the code gets patched like above and afterwards looks + * like this: + * > brasl %r0,ftrace_caller # offset 0 + * + * The instruction will be patched by ftrace_make_call / ftrace_make_nop. + * The ftrace function gets called with a non-standard C function call ABI + * where r0 contains the return address. It is also expected that the called + * function only clobbers r0 and r1, but restores r2-r15. + * For module code we can't directly jump to ftrace caller, but need a + * trampoline (ftrace_plt), which clobbers also r1. + */ + +void *ftrace_func __read_mostly = ftrace_stub; +struct ftrace_insn { + u16 opc; + s32 disp; +} __packed; + +#ifdef CONFIG_MODULES +static char *ftrace_plt; +#endif /* CONFIG_MODULES */ + +static const char *ftrace_shared_hotpatch_trampoline(const char **end) +{ + const char *tstart, *tend; + + tstart = ftrace_shared_hotpatch_trampoline_br; + tend = ftrace_shared_hotpatch_trampoline_br_end; +#ifdef CONFIG_EXPOLINE + if (!nospec_disable) { + tstart = ftrace_shared_hotpatch_trampoline_exrl; + tend = ftrace_shared_hotpatch_trampoline_exrl_end; + } +#endif /* CONFIG_EXPOLINE */ + if (end) + *end = tend; + return tstart; +} + +bool ftrace_need_init_nop(void) +{ + return true; +} + +int ftrace_init_nop(struct module *mod, struct dyn_ftrace *rec) +{ + static struct ftrace_hotpatch_trampoline *next_vmlinux_trampoline = + __ftrace_hotpatch_trampolines_start; + static const char orig[6] = { 0xc0, 0x04, 0x00, 0x00, 0x00, 0x00 }; + static struct ftrace_hotpatch_trampoline *trampoline; + struct ftrace_hotpatch_trampoline **next_trampoline; + struct ftrace_hotpatch_trampoline *trampolines_end; + struct ftrace_hotpatch_trampoline tmp; + struct ftrace_insn *insn; + const char *shared; + s32 disp; + + BUILD_BUG_ON(sizeof(struct ftrace_hotpatch_trampoline) != + SIZEOF_FTRACE_HOTPATCH_TRAMPOLINE); + + next_trampoline = &next_vmlinux_trampoline; + trampolines_end = __ftrace_hotpatch_trampolines_end; + shared = ftrace_shared_hotpatch_trampoline(NULL); +#ifdef CONFIG_MODULES + if (mod) { + next_trampoline = &mod->arch.next_trampoline; + trampolines_end = mod->arch.trampolines_end; + shared = ftrace_plt; + } +#endif + + if (WARN_ON_ONCE(*next_trampoline >= trampolines_end)) + return -ENOMEM; + trampoline = (*next_trampoline)++; + + /* Check for the compiler-generated fentry nop (brcl 0, .). */ + if (WARN_ON_ONCE(memcmp((const void *)rec->ip, &orig, sizeof(orig)))) + return -EINVAL; + + /* Generate the trampoline. */ + tmp.brasl_opc = 0xc015; /* brasl %r1, shared */ + tmp.brasl_disp = (shared - (const char *)&trampoline->brasl_opc) / 2; + tmp.interceptor = FTRACE_ADDR; + tmp.rest_of_intercepted_function = rec->ip + sizeof(struct ftrace_insn); + s390_kernel_write(trampoline, &tmp, sizeof(tmp)); + + /* Generate a jump to the trampoline. */ + disp = ((char *)trampoline - (char *)rec->ip) / 2; + insn = (struct ftrace_insn *)rec->ip; + s390_kernel_write(&insn->disp, &disp, sizeof(disp)); + + return 0; +} + +static struct ftrace_hotpatch_trampoline *ftrace_get_trampoline(struct dyn_ftrace *rec) +{ + struct ftrace_hotpatch_trampoline *trampoline; + struct ftrace_insn insn; + s64 disp; + u16 opc; + + if (copy_from_kernel_nofault(&insn, (void *)rec->ip, sizeof(insn))) + return ERR_PTR(-EFAULT); + disp = (s64)insn.disp * 2; + trampoline = (void *)(rec->ip + disp); + if (get_kernel_nofault(opc, &trampoline->brasl_opc)) + return ERR_PTR(-EFAULT); + if (opc != 0xc015) + return ERR_PTR(-EINVAL); + return trampoline; +} + +int ftrace_modify_call(struct dyn_ftrace *rec, unsigned long old_addr, + unsigned long addr) +{ + struct ftrace_hotpatch_trampoline *trampoline; + u64 old; + + trampoline = ftrace_get_trampoline(rec); + if (IS_ERR(trampoline)) + return PTR_ERR(trampoline); + if (get_kernel_nofault(old, &trampoline->interceptor)) + return -EFAULT; + if (old != old_addr) + return -EINVAL; + s390_kernel_write(&trampoline->interceptor, &addr, sizeof(addr)); + return 0; +} + +static int ftrace_patch_branch_mask(void *addr, u16 expected, bool enable) +{ + u16 old; + u8 op; + + if (get_kernel_nofault(old, addr)) + return -EFAULT; + if (old != expected) + return -EINVAL; + /* set mask field to all ones or zeroes */ + op = enable ? 0xf4 : 0x04; + s390_kernel_write((char *)addr + 1, &op, sizeof(op)); + return 0; +} + +int ftrace_make_nop(struct module *mod, struct dyn_ftrace *rec, + unsigned long addr) +{ + /* Expect brcl 0xf,... */ + return ftrace_patch_branch_mask((void *)rec->ip, 0xc0f4, false); +} + +int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr) +{ + struct ftrace_hotpatch_trampoline *trampoline; + + trampoline = ftrace_get_trampoline(rec); + if (IS_ERR(trampoline)) + return PTR_ERR(trampoline); + s390_kernel_write(&trampoline->interceptor, &addr, sizeof(addr)); + /* Expect brcl 0x0,... */ + return ftrace_patch_branch_mask((void *)rec->ip, 0xc004, true); +} + +int ftrace_update_ftrace_func(ftrace_func_t func) +{ + ftrace_func = func; + return 0; +} + +void arch_ftrace_update_code(int command) +{ + ftrace_modify_all_code(command); +} + +void ftrace_arch_code_modify_post_process(void) +{ + /* + * Flush any pre-fetched instructions on all + * CPUs to make the new code visible. + */ + text_poke_sync_lock(); +} + +#ifdef CONFIG_MODULES + +static int __init ftrace_plt_init(void) +{ + const char *start, *end; + + ftrace_plt = module_alloc(PAGE_SIZE); + if (!ftrace_plt) + panic("cannot allocate ftrace plt\n"); + + start = ftrace_shared_hotpatch_trampoline(&end); + memcpy(ftrace_plt, start, end - start); + set_memory_rox((unsigned long)ftrace_plt, 1); + return 0; +} +device_initcall(ftrace_plt_init); + +#endif /* CONFIG_MODULES */ + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +/* + * Hook the return address and push it in the stack of return addresses + * in current thread info. + */ +unsigned long prepare_ftrace_return(unsigned long ra, unsigned long sp, + unsigned long ip) +{ + if (unlikely(ftrace_graph_is_dead())) + goto out; + if (unlikely(atomic_read(¤t->tracing_graph_pause))) + goto out; + ip -= MCOUNT_INSN_SIZE; + if (!function_graph_enter(ra, ip, 0, (void *) sp)) + ra = (unsigned long) return_to_handler; +out: + return ra; +} +NOKPROBE_SYMBOL(prepare_ftrace_return); + +/* + * Patch the kernel code at ftrace_graph_caller location. The instruction + * there is branch relative on condition. To enable the ftrace graph code + * block, we simply patch the mask field of the instruction to zero and + * turn the instruction into a nop. + * To disable the ftrace graph code the mask field will be patched to + * all ones, which turns the instruction into an unconditional branch. + */ +int ftrace_enable_ftrace_graph_caller(void) +{ + int rc; + + /* Expect brc 0xf,... */ + rc = ftrace_patch_branch_mask(ftrace_graph_caller, 0xa7f4, false); + if (rc) + return rc; + text_poke_sync_lock(); + return 0; +} + +int ftrace_disable_ftrace_graph_caller(void) +{ + int rc; + + /* Expect brc 0x0,... */ + rc = ftrace_patch_branch_mask(ftrace_graph_caller, 0xa704, true); + if (rc) + return rc; + text_poke_sync_lock(); + return 0; +} + +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ + +#ifdef CONFIG_KPROBES_ON_FTRACE +void kprobe_ftrace_handler(unsigned long ip, unsigned long parent_ip, + struct ftrace_ops *ops, struct ftrace_regs *fregs) +{ + struct kprobe_ctlblk *kcb; + struct pt_regs *regs; + struct kprobe *p; + int bit; + + bit = ftrace_test_recursion_trylock(ip, parent_ip); + if (bit < 0) + return; + + regs = ftrace_get_regs(fregs); + p = get_kprobe((kprobe_opcode_t *)ip); + if (!regs || unlikely(!p) || kprobe_disabled(p)) + goto out; + + if (kprobe_running()) { + kprobes_inc_nmissed_count(p); + goto out; + } + + __this_cpu_write(current_kprobe, p); + + kcb = get_kprobe_ctlblk(); + kcb->kprobe_status = KPROBE_HIT_ACTIVE; + + instruction_pointer_set(regs, ip); + + if (!p->pre_handler || !p->pre_handler(p, regs)) { + + instruction_pointer_set(regs, ip + MCOUNT_INSN_SIZE); + + if (unlikely(p->post_handler)) { + kcb->kprobe_status = KPROBE_HIT_SSDONE; + p->post_handler(p, regs, 0); + } + } + __this_cpu_write(current_kprobe, NULL); +out: + ftrace_test_recursion_unlock(bit); +} +NOKPROBE_SYMBOL(kprobe_ftrace_handler); + +int arch_prepare_kprobe_ftrace(struct kprobe *p) +{ + p->ainsn.insn = NULL; + return 0; +} +#endif diff --git a/arch/s390/kernel/ftrace.h b/arch/s390/kernel/ftrace.h new file mode 100644 index 0000000000..7f75a96164 --- /dev/null +++ b/arch/s390/kernel/ftrace.h @@ -0,0 +1,24 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _FTRACE_H +#define _FTRACE_H + +#include <asm/types.h> + +struct ftrace_hotpatch_trampoline { + u16 brasl_opc; + s32 brasl_disp; + s16: 16; + u64 rest_of_intercepted_function; + u64 interceptor; +} __packed; + +extern struct ftrace_hotpatch_trampoline __ftrace_hotpatch_trampolines_start[]; +extern struct ftrace_hotpatch_trampoline __ftrace_hotpatch_trampolines_end[]; +extern const char ftrace_shared_hotpatch_trampoline_br[]; +extern const char ftrace_shared_hotpatch_trampoline_br_end[]; +extern const char ftrace_shared_hotpatch_trampoline_exrl[]; +extern const char ftrace_shared_hotpatch_trampoline_exrl_end[]; +extern const char ftrace_plt_template[]; +extern const char ftrace_plt_template_end[]; + +#endif /* _FTRACE_H */ diff --git a/arch/s390/kernel/guarded_storage.c b/arch/s390/kernel/guarded_storage.c new file mode 100644 index 0000000000..d14dd1c2e5 --- /dev/null +++ b/arch/s390/kernel/guarded_storage.c @@ -0,0 +1,128 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2016 + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> + */ + +#include <linux/kernel.h> +#include <linux/syscalls.h> +#include <linux/signal.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <asm/guarded_storage.h> +#include "entry.h" + +void guarded_storage_release(struct task_struct *tsk) +{ + kfree(tsk->thread.gs_cb); + kfree(tsk->thread.gs_bc_cb); +} + +static int gs_enable(void) +{ + struct gs_cb *gs_cb; + + if (!current->thread.gs_cb) { + gs_cb = kzalloc(sizeof(*gs_cb), GFP_KERNEL); + if (!gs_cb) + return -ENOMEM; + gs_cb->gsd = 25; + preempt_disable(); + __ctl_set_bit(2, 4); + load_gs_cb(gs_cb); + current->thread.gs_cb = gs_cb; + preempt_enable(); + } + return 0; +} + +static int gs_disable(void) +{ + if (current->thread.gs_cb) { + preempt_disable(); + kfree(current->thread.gs_cb); + current->thread.gs_cb = NULL; + __ctl_clear_bit(2, 4); + preempt_enable(); + } + return 0; +} + +static int gs_set_bc_cb(struct gs_cb __user *u_gs_cb) +{ + struct gs_cb *gs_cb; + + gs_cb = current->thread.gs_bc_cb; + if (!gs_cb) { + gs_cb = kzalloc(sizeof(*gs_cb), GFP_KERNEL); + if (!gs_cb) + return -ENOMEM; + current->thread.gs_bc_cb = gs_cb; + } + if (copy_from_user(gs_cb, u_gs_cb, sizeof(*gs_cb))) + return -EFAULT; + return 0; +} + +static int gs_clear_bc_cb(void) +{ + struct gs_cb *gs_cb; + + gs_cb = current->thread.gs_bc_cb; + current->thread.gs_bc_cb = NULL; + kfree(gs_cb); + return 0; +} + +void gs_load_bc_cb(struct pt_regs *regs) +{ + struct gs_cb *gs_cb; + + preempt_disable(); + clear_thread_flag(TIF_GUARDED_STORAGE); + gs_cb = current->thread.gs_bc_cb; + if (gs_cb) { + kfree(current->thread.gs_cb); + current->thread.gs_bc_cb = NULL; + __ctl_set_bit(2, 4); + load_gs_cb(gs_cb); + current->thread.gs_cb = gs_cb; + } + preempt_enable(); +} + +static int gs_broadcast(void) +{ + struct task_struct *sibling; + + read_lock(&tasklist_lock); + for_each_thread(current, sibling) { + if (!sibling->thread.gs_bc_cb) + continue; + if (test_and_set_tsk_thread_flag(sibling, TIF_GUARDED_STORAGE)) + kick_process(sibling); + } + read_unlock(&tasklist_lock); + return 0; +} + +SYSCALL_DEFINE2(s390_guarded_storage, int, command, + struct gs_cb __user *, gs_cb) +{ + if (!MACHINE_HAS_GS) + return -EOPNOTSUPP; + switch (command) { + case GS_ENABLE: + return gs_enable(); + case GS_DISABLE: + return gs_disable(); + case GS_SET_BC_CB: + return gs_set_bc_cb(gs_cb); + case GS_CLEAR_BC_CB: + return gs_clear_bc_cb(); + case GS_BROADCAST: + return gs_broadcast(); + default: + return -EINVAL; + } +} diff --git a/arch/s390/kernel/head64.S b/arch/s390/kernel/head64.S new file mode 100644 index 0000000000..45413b04ef --- /dev/null +++ b/arch/s390/kernel/head64.S @@ -0,0 +1,40 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright IBM Corp. 1999, 2010 + * + * Author(s): Hartmut Penner <hp@de.ibm.com> + * Martin Schwidefsky <schwidefsky@de.ibm.com> + * Rob van der Heij <rvdhei@iae.nl> + * + */ + +#include <linux/init.h> +#include <linux/linkage.h> +#include <asm/asm-offsets.h> +#include <asm/thread_info.h> +#include <asm/page.h> +#include <asm/ptrace.h> + +__HEAD +SYM_CODE_START(startup_continue) + larl %r1,tod_clock_base + mvc 0(16,%r1),__LC_BOOT_CLOCK +# +# Setup stack +# + larl %r14,init_task + stg %r14,__LC_CURRENT + larl %r15,init_thread_union+STACK_INIT_OFFSET + stg %r15,__LC_KERNEL_STACK + brasl %r14,sclp_early_adjust_va # allow sclp_early_printk + brasl %r14,startup_init # s390 specific early init + brasl %r14,start_kernel # common init code +# +# We returned from start_kernel ?!? PANIK +# + basr %r13,0 + lpswe dw_psw-.(%r13) # load disabled wait psw +SYM_CODE_END(startup_continue) + + .balign 16 +SYM_DATA_LOCAL(dw_psw, .quad 0x0002000180000000,0x0000000000000000) diff --git a/arch/s390/kernel/idle.c b/arch/s390/kernel/idle.c new file mode 100644 index 0000000000..e7239aaf42 --- /dev/null +++ b/arch/s390/kernel/idle.c @@ -0,0 +1,94 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Idle functions for s390. + * + * Copyright IBM Corp. 2014 + * + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> + */ + +#include <linux/kernel.h> +#include <linux/kernel_stat.h> +#include <linux/notifier.h> +#include <linux/init.h> +#include <linux/cpu.h> +#include <trace/events/power.h> +#include <asm/cpu_mf.h> +#include <asm/cputime.h> +#include <asm/nmi.h> +#include <asm/smp.h> +#include "entry.h" + +static DEFINE_PER_CPU(struct s390_idle_data, s390_idle); + +void account_idle_time_irq(void) +{ + struct s390_idle_data *idle = this_cpu_ptr(&s390_idle); + unsigned long idle_time; + u64 cycles_new[8]; + int i; + + if (smp_cpu_mtid) { + stcctm(MT_DIAG, smp_cpu_mtid, cycles_new); + for (i = 0; i < smp_cpu_mtid; i++) + this_cpu_add(mt_cycles[i], cycles_new[i] - idle->mt_cycles_enter[i]); + } + + idle_time = S390_lowcore.int_clock - idle->clock_idle_enter; + + S390_lowcore.steal_timer += idle->clock_idle_enter - S390_lowcore.last_update_clock; + S390_lowcore.last_update_clock = S390_lowcore.int_clock; + + S390_lowcore.system_timer += S390_lowcore.last_update_timer - idle->timer_idle_enter; + S390_lowcore.last_update_timer = S390_lowcore.sys_enter_timer; + + /* Account time spent with enabled wait psw loaded as idle time. */ + WRITE_ONCE(idle->idle_time, READ_ONCE(idle->idle_time) + idle_time); + WRITE_ONCE(idle->idle_count, READ_ONCE(idle->idle_count) + 1); + account_idle_time(cputime_to_nsecs(idle_time)); +} + +void noinstr arch_cpu_idle(void) +{ + struct s390_idle_data *idle = this_cpu_ptr(&s390_idle); + unsigned long psw_mask; + + /* Wait for external, I/O or machine check interrupt. */ + psw_mask = PSW_KERNEL_BITS | PSW_MASK_WAIT | + PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK; + clear_cpu_flag(CIF_NOHZ_DELAY); + + /* psw_idle() returns with interrupts disabled. */ + psw_idle(idle, psw_mask); +} + +static ssize_t show_idle_count(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct s390_idle_data *idle = &per_cpu(s390_idle, dev->id); + + return sysfs_emit(buf, "%lu\n", READ_ONCE(idle->idle_count)); +} +DEVICE_ATTR(idle_count, 0444, show_idle_count, NULL); + +static ssize_t show_idle_time(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct s390_idle_data *idle = &per_cpu(s390_idle, dev->id); + + return sysfs_emit(buf, "%lu\n", READ_ONCE(idle->idle_time) >> 12); +} +DEVICE_ATTR(idle_time_us, 0444, show_idle_time, NULL); + +void arch_cpu_idle_enter(void) +{ +} + +void arch_cpu_idle_exit(void) +{ +} + +void __noreturn arch_cpu_idle_dead(void) +{ + cpu_die(); +} diff --git a/arch/s390/kernel/ima_arch.c b/arch/s390/kernel/ima_arch.c new file mode 100644 index 0000000000..f3c3e6e1c5 --- /dev/null +++ b/arch/s390/kernel/ima_arch.c @@ -0,0 +1,14 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <linux/ima.h> +#include <asm/boot_data.h> + +bool arch_ima_get_secureboot(void) +{ + return ipl_secure_flag; +} + +const char * const *arch_get_ima_policy(void) +{ + return NULL; +} diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c new file mode 100644 index 0000000000..8d0b95c173 --- /dev/null +++ b/arch/s390/kernel/ipl.c @@ -0,0 +1,2522 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * ipl/reipl/dump support for Linux on s390. + * + * Copyright IBM Corp. 2005, 2012 + * Author(s): Michael Holzheu <holzheu@de.ibm.com> + * Volker Sameske <sameske@de.ibm.com> + */ + +#include <linux/types.h> +#include <linux/export.h> +#include <linux/init.h> +#include <linux/device.h> +#include <linux/delay.h> +#include <linux/kstrtox.h> +#include <linux/panic_notifier.h> +#include <linux/reboot.h> +#include <linux/ctype.h> +#include <linux/fs.h> +#include <linux/gfp.h> +#include <linux/crash_dump.h> +#include <linux/debug_locks.h> +#include <asm/asm-extable.h> +#include <asm/diag.h> +#include <asm/ipl.h> +#include <asm/smp.h> +#include <asm/setup.h> +#include <asm/cpcmd.h> +#include <asm/ebcdic.h> +#include <asm/sclp.h> +#include <asm/checksum.h> +#include <asm/debug.h> +#include <asm/abs_lowcore.h> +#include <asm/os_info.h> +#include <asm/sections.h> +#include <asm/boot_data.h> +#include "entry.h" + +#define IPL_PARM_BLOCK_VERSION 0 + +#define IPL_UNKNOWN_STR "unknown" +#define IPL_CCW_STR "ccw" +#define IPL_ECKD_STR "eckd" +#define IPL_ECKD_DUMP_STR "eckd_dump" +#define IPL_FCP_STR "fcp" +#define IPL_FCP_DUMP_STR "fcp_dump" +#define IPL_NVME_STR "nvme" +#define IPL_NVME_DUMP_STR "nvme_dump" +#define IPL_NSS_STR "nss" + +#define DUMP_CCW_STR "ccw" +#define DUMP_ECKD_STR "eckd" +#define DUMP_FCP_STR "fcp" +#define DUMP_NVME_STR "nvme" +#define DUMP_NONE_STR "none" + +/* + * Four shutdown trigger types are supported: + * - panic + * - halt + * - power off + * - reipl + * - restart + */ +#define ON_PANIC_STR "on_panic" +#define ON_HALT_STR "on_halt" +#define ON_POFF_STR "on_poff" +#define ON_REIPL_STR "on_reboot" +#define ON_RESTART_STR "on_restart" + +struct shutdown_action; +struct shutdown_trigger { + char *name; + struct shutdown_action *action; +}; + +/* + * The following shutdown action types are supported: + */ +#define SHUTDOWN_ACTION_IPL_STR "ipl" +#define SHUTDOWN_ACTION_REIPL_STR "reipl" +#define SHUTDOWN_ACTION_DUMP_STR "dump" +#define SHUTDOWN_ACTION_VMCMD_STR "vmcmd" +#define SHUTDOWN_ACTION_STOP_STR "stop" +#define SHUTDOWN_ACTION_DUMP_REIPL_STR "dump_reipl" + +struct shutdown_action { + char *name; + void (*fn) (struct shutdown_trigger *trigger); + int (*init) (void); + int init_rc; +}; + +static char *ipl_type_str(enum ipl_type type) +{ + switch (type) { + case IPL_TYPE_CCW: + return IPL_CCW_STR; + case IPL_TYPE_ECKD: + return IPL_ECKD_STR; + case IPL_TYPE_ECKD_DUMP: + return IPL_ECKD_DUMP_STR; + case IPL_TYPE_FCP: + return IPL_FCP_STR; + case IPL_TYPE_FCP_DUMP: + return IPL_FCP_DUMP_STR; + case IPL_TYPE_NSS: + return IPL_NSS_STR; + case IPL_TYPE_NVME: + return IPL_NVME_STR; + case IPL_TYPE_NVME_DUMP: + return IPL_NVME_DUMP_STR; + case IPL_TYPE_UNKNOWN: + default: + return IPL_UNKNOWN_STR; + } +} + +enum dump_type { + DUMP_TYPE_NONE = 1, + DUMP_TYPE_CCW = 2, + DUMP_TYPE_FCP = 4, + DUMP_TYPE_NVME = 8, + DUMP_TYPE_ECKD = 16, +}; + +static char *dump_type_str(enum dump_type type) +{ + switch (type) { + case DUMP_TYPE_NONE: + return DUMP_NONE_STR; + case DUMP_TYPE_CCW: + return DUMP_CCW_STR; + case DUMP_TYPE_ECKD: + return DUMP_ECKD_STR; + case DUMP_TYPE_FCP: + return DUMP_FCP_STR; + case DUMP_TYPE_NVME: + return DUMP_NVME_STR; + default: + return NULL; + } +} + +int __bootdata_preserved(ipl_block_valid); +struct ipl_parameter_block __bootdata_preserved(ipl_block); +int __bootdata_preserved(ipl_secure_flag); + +unsigned long __bootdata_preserved(ipl_cert_list_addr); +unsigned long __bootdata_preserved(ipl_cert_list_size); + +unsigned long __bootdata(early_ipl_comp_list_addr); +unsigned long __bootdata(early_ipl_comp_list_size); + +static int reipl_capabilities = IPL_TYPE_UNKNOWN; + +static enum ipl_type reipl_type = IPL_TYPE_UNKNOWN; +static struct ipl_parameter_block *reipl_block_fcp; +static struct ipl_parameter_block *reipl_block_nvme; +static struct ipl_parameter_block *reipl_block_ccw; +static struct ipl_parameter_block *reipl_block_eckd; +static struct ipl_parameter_block *reipl_block_nss; +static struct ipl_parameter_block *reipl_block_actual; + +static int dump_capabilities = DUMP_TYPE_NONE; +static enum dump_type dump_type = DUMP_TYPE_NONE; +static struct ipl_parameter_block *dump_block_fcp; +static struct ipl_parameter_block *dump_block_nvme; +static struct ipl_parameter_block *dump_block_ccw; +static struct ipl_parameter_block *dump_block_eckd; + +static struct sclp_ipl_info sclp_ipl_info; + +static bool reipl_nvme_clear; +static bool reipl_fcp_clear; +static bool reipl_ccw_clear; +static bool reipl_eckd_clear; + +static unsigned long os_info_flags; + +static inline int __diag308(unsigned long subcode, unsigned long addr) +{ + union register_pair r1; + + r1.even = addr; + r1.odd = 0; + asm volatile( + " diag %[r1],%[subcode],0x308\n" + "0: nopr %%r7\n" + EX_TABLE(0b,0b) + : [r1] "+&d" (r1.pair) + : [subcode] "d" (subcode) + : "cc", "memory"); + return r1.odd; +} + +int diag308(unsigned long subcode, void *addr) +{ + diag_stat_inc(DIAG_STAT_X308); + return __diag308(subcode, addr ? virt_to_phys(addr) : 0); +} +EXPORT_SYMBOL_GPL(diag308); + +/* SYSFS */ + +#define IPL_ATTR_SHOW_FN(_prefix, _name, _format, args...) \ +static ssize_t sys_##_prefix##_##_name##_show(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + char *page) \ +{ \ + return scnprintf(page, PAGE_SIZE, _format, ##args); \ +} + +#define IPL_ATTR_CCW_STORE_FN(_prefix, _name, _ipl_blk) \ +static ssize_t sys_##_prefix##_##_name##_store(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + const char *buf, size_t len) \ +{ \ + unsigned long long ssid, devno; \ + \ + if (sscanf(buf, "0.%llx.%llx\n", &ssid, &devno) != 2) \ + return -EINVAL; \ + \ + if (ssid > __MAX_SSID || devno > __MAX_SUBCHANNEL) \ + return -EINVAL; \ + \ + _ipl_blk.ssid = ssid; \ + _ipl_blk.devno = devno; \ + return len; \ +} + +#define DEFINE_IPL_CCW_ATTR_RW(_prefix, _name, _ipl_blk) \ +IPL_ATTR_SHOW_FN(_prefix, _name, "0.%x.%04x\n", \ + _ipl_blk.ssid, _ipl_blk.devno); \ +IPL_ATTR_CCW_STORE_FN(_prefix, _name, _ipl_blk); \ +static struct kobj_attribute sys_##_prefix##_##_name##_attr = \ + __ATTR(_name, 0644, \ + sys_##_prefix##_##_name##_show, \ + sys_##_prefix##_##_name##_store) \ + +#define DEFINE_IPL_ATTR_RO(_prefix, _name, _format, _value) \ +IPL_ATTR_SHOW_FN(_prefix, _name, _format, _value) \ +static struct kobj_attribute sys_##_prefix##_##_name##_attr = \ + __ATTR(_name, 0444, sys_##_prefix##_##_name##_show, NULL) + +#define DEFINE_IPL_ATTR_RW(_prefix, _name, _fmt_out, _fmt_in, _value) \ +IPL_ATTR_SHOW_FN(_prefix, _name, _fmt_out, (unsigned long long) _value) \ +static ssize_t sys_##_prefix##_##_name##_store(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + const char *buf, size_t len) \ +{ \ + unsigned long long value; \ + if (sscanf(buf, _fmt_in, &value) != 1) \ + return -EINVAL; \ + _value = value; \ + return len; \ +} \ +static struct kobj_attribute sys_##_prefix##_##_name##_attr = \ + __ATTR(_name, 0644, \ + sys_##_prefix##_##_name##_show, \ + sys_##_prefix##_##_name##_store) + +#define DEFINE_IPL_ATTR_STR_RW(_prefix, _name, _fmt_out, _fmt_in, _value)\ +IPL_ATTR_SHOW_FN(_prefix, _name, _fmt_out, _value) \ +static ssize_t sys_##_prefix##_##_name##_store(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + const char *buf, size_t len) \ +{ \ + strscpy(_value, buf, sizeof(_value)); \ + strim(_value); \ + return len; \ +} \ +static struct kobj_attribute sys_##_prefix##_##_name##_attr = \ + __ATTR(_name, 0644, \ + sys_##_prefix##_##_name##_show, \ + sys_##_prefix##_##_name##_store) + +/* + * ipl section + */ + +static __init enum ipl_type get_ipl_type(void) +{ + if (!ipl_block_valid) + return IPL_TYPE_UNKNOWN; + + switch (ipl_block.pb0_hdr.pbt) { + case IPL_PBT_CCW: + return IPL_TYPE_CCW; + case IPL_PBT_FCP: + if (ipl_block.fcp.opt == IPL_PB0_FCP_OPT_DUMP) + return IPL_TYPE_FCP_DUMP; + else + return IPL_TYPE_FCP; + case IPL_PBT_NVME: + if (ipl_block.nvme.opt == IPL_PB0_NVME_OPT_DUMP) + return IPL_TYPE_NVME_DUMP; + else + return IPL_TYPE_NVME; + case IPL_PBT_ECKD: + if (ipl_block.eckd.opt == IPL_PB0_ECKD_OPT_DUMP) + return IPL_TYPE_ECKD_DUMP; + else + return IPL_TYPE_ECKD; + } + return IPL_TYPE_UNKNOWN; +} + +struct ipl_info ipl_info; +EXPORT_SYMBOL_GPL(ipl_info); + +static ssize_t ipl_type_show(struct kobject *kobj, struct kobj_attribute *attr, + char *page) +{ + return sprintf(page, "%s\n", ipl_type_str(ipl_info.type)); +} + +static struct kobj_attribute sys_ipl_type_attr = __ATTR_RO(ipl_type); + +static ssize_t ipl_secure_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%i\n", !!ipl_secure_flag); +} + +static struct kobj_attribute sys_ipl_secure_attr = + __ATTR(secure, 0444, ipl_secure_show, NULL); + +static ssize_t ipl_has_secure_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%i\n", !!sclp.has_sipl); +} + +static struct kobj_attribute sys_ipl_has_secure_attr = + __ATTR(has_secure, 0444, ipl_has_secure_show, NULL); + +static ssize_t ipl_vm_parm_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + char parm[DIAG308_VMPARM_SIZE + 1] = {}; + + if (ipl_block_valid && (ipl_block.pb0_hdr.pbt == IPL_PBT_CCW)) + ipl_block_get_ascii_vmparm(parm, sizeof(parm), &ipl_block); + return sprintf(page, "%s\n", parm); +} + +static struct kobj_attribute sys_ipl_vm_parm_attr = + __ATTR(parm, 0444, ipl_vm_parm_show, NULL); + +static ssize_t sys_ipl_device_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + switch (ipl_info.type) { + case IPL_TYPE_CCW: + return sprintf(page, "0.%x.%04x\n", ipl_block.ccw.ssid, + ipl_block.ccw.devno); + case IPL_TYPE_ECKD: + case IPL_TYPE_ECKD_DUMP: + return sprintf(page, "0.%x.%04x\n", ipl_block.eckd.ssid, + ipl_block.eckd.devno); + case IPL_TYPE_FCP: + case IPL_TYPE_FCP_DUMP: + return sprintf(page, "0.0.%04x\n", ipl_block.fcp.devno); + case IPL_TYPE_NVME: + case IPL_TYPE_NVME_DUMP: + return sprintf(page, "%08ux\n", ipl_block.nvme.fid); + default: + return 0; + } +} + +static struct kobj_attribute sys_ipl_device_attr = + __ATTR(device, 0444, sys_ipl_device_show, NULL); + +static ssize_t ipl_parameter_read(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, char *buf, + loff_t off, size_t count) +{ + return memory_read_from_buffer(buf, count, &off, &ipl_block, + ipl_block.hdr.len); +} +static struct bin_attribute ipl_parameter_attr = + __BIN_ATTR(binary_parameter, 0444, ipl_parameter_read, NULL, + PAGE_SIZE); + +static ssize_t ipl_scp_data_read(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, char *buf, + loff_t off, size_t count) +{ + unsigned int size = ipl_block.fcp.scp_data_len; + void *scp_data = &ipl_block.fcp.scp_data; + + return memory_read_from_buffer(buf, count, &off, scp_data, size); +} + +static ssize_t ipl_nvme_scp_data_read(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, char *buf, + loff_t off, size_t count) +{ + unsigned int size = ipl_block.nvme.scp_data_len; + void *scp_data = &ipl_block.nvme.scp_data; + + return memory_read_from_buffer(buf, count, &off, scp_data, size); +} + +static ssize_t ipl_eckd_scp_data_read(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, char *buf, + loff_t off, size_t count) +{ + unsigned int size = ipl_block.eckd.scp_data_len; + void *scp_data = &ipl_block.eckd.scp_data; + + return memory_read_from_buffer(buf, count, &off, scp_data, size); +} + +static struct bin_attribute ipl_scp_data_attr = + __BIN_ATTR(scp_data, 0444, ipl_scp_data_read, NULL, PAGE_SIZE); + +static struct bin_attribute ipl_nvme_scp_data_attr = + __BIN_ATTR(scp_data, 0444, ipl_nvme_scp_data_read, NULL, PAGE_SIZE); + +static struct bin_attribute ipl_eckd_scp_data_attr = + __BIN_ATTR(scp_data, 0444, ipl_eckd_scp_data_read, NULL, PAGE_SIZE); + +static struct bin_attribute *ipl_fcp_bin_attrs[] = { + &ipl_parameter_attr, + &ipl_scp_data_attr, + NULL, +}; + +static struct bin_attribute *ipl_nvme_bin_attrs[] = { + &ipl_parameter_attr, + &ipl_nvme_scp_data_attr, + NULL, +}; + +static struct bin_attribute *ipl_eckd_bin_attrs[] = { + &ipl_parameter_attr, + &ipl_eckd_scp_data_attr, + NULL, +}; + +/* FCP ipl device attributes */ + +DEFINE_IPL_ATTR_RO(ipl_fcp, wwpn, "0x%016llx\n", + (unsigned long long)ipl_block.fcp.wwpn); +DEFINE_IPL_ATTR_RO(ipl_fcp, lun, "0x%016llx\n", + (unsigned long long)ipl_block.fcp.lun); +DEFINE_IPL_ATTR_RO(ipl_fcp, bootprog, "%lld\n", + (unsigned long long)ipl_block.fcp.bootprog); +DEFINE_IPL_ATTR_RO(ipl_fcp, br_lba, "%lld\n", + (unsigned long long)ipl_block.fcp.br_lba); + +/* NVMe ipl device attributes */ +DEFINE_IPL_ATTR_RO(ipl_nvme, fid, "0x%08llx\n", + (unsigned long long)ipl_block.nvme.fid); +DEFINE_IPL_ATTR_RO(ipl_nvme, nsid, "0x%08llx\n", + (unsigned long long)ipl_block.nvme.nsid); +DEFINE_IPL_ATTR_RO(ipl_nvme, bootprog, "%lld\n", + (unsigned long long)ipl_block.nvme.bootprog); +DEFINE_IPL_ATTR_RO(ipl_nvme, br_lba, "%lld\n", + (unsigned long long)ipl_block.nvme.br_lba); + +/* ECKD ipl device attributes */ +DEFINE_IPL_ATTR_RO(ipl_eckd, bootprog, "%lld\n", + (unsigned long long)ipl_block.eckd.bootprog); + +#define IPL_ATTR_BR_CHR_SHOW_FN(_name, _ipb) \ +static ssize_t eckd_##_name##_br_chr_show(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + char *buf) \ +{ \ + struct ipl_pb0_eckd *ipb = &(_ipb); \ + \ + if (!ipb->br_chr.cyl && \ + !ipb->br_chr.head && \ + !ipb->br_chr.record) \ + return sprintf(buf, "auto\n"); \ + \ + return sprintf(buf, "0x%x,0x%x,0x%x\n", \ + ipb->br_chr.cyl, \ + ipb->br_chr.head, \ + ipb->br_chr.record); \ +} + +#define IPL_ATTR_BR_CHR_STORE_FN(_name, _ipb) \ +static ssize_t eckd_##_name##_br_chr_store(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + const char *buf, size_t len) \ +{ \ + struct ipl_pb0_eckd *ipb = &(_ipb); \ + unsigned long args[3] = { 0 }; \ + char *p, *p1, *tmp = NULL; \ + int i, rc; \ + \ + if (!strncmp(buf, "auto", 4)) \ + goto out; \ + \ + tmp = kstrdup(buf, GFP_KERNEL); \ + p = tmp; \ + for (i = 0; i < 3; i++) { \ + p1 = strsep(&p, ", "); \ + if (!p1) { \ + rc = -EINVAL; \ + goto err; \ + } \ + rc = kstrtoul(p1, 0, args + i); \ + if (rc) \ + goto err; \ + } \ + \ + rc = -EINVAL; \ + if (i != 3) \ + goto err; \ + \ + if ((args[0] || args[1]) && !args[2]) \ + goto err; \ + \ + if (args[0] > UINT_MAX || args[1] > 255 || args[2] > 255) \ + goto err; \ + \ +out: \ + ipb->br_chr.cyl = args[0]; \ + ipb->br_chr.head = args[1]; \ + ipb->br_chr.record = args[2]; \ + rc = len; \ +err: \ + kfree(tmp); \ + return rc; \ +} + +IPL_ATTR_BR_CHR_SHOW_FN(ipl, ipl_block.eckd); +static struct kobj_attribute sys_ipl_eckd_br_chr_attr = + __ATTR(br_chr, 0644, eckd_ipl_br_chr_show, NULL); + +IPL_ATTR_BR_CHR_SHOW_FN(reipl, reipl_block_eckd->eckd); +IPL_ATTR_BR_CHR_STORE_FN(reipl, reipl_block_eckd->eckd); + +static struct kobj_attribute sys_reipl_eckd_br_chr_attr = + __ATTR(br_chr, 0644, eckd_reipl_br_chr_show, eckd_reipl_br_chr_store); + +static ssize_t ipl_ccw_loadparm_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + char loadparm[LOADPARM_LEN + 1] = {}; + + if (!sclp_ipl_info.is_valid) + return sprintf(page, "#unknown#\n"); + memcpy(loadparm, &sclp_ipl_info.loadparm, LOADPARM_LEN); + EBCASC(loadparm, LOADPARM_LEN); + strim(loadparm); + return sprintf(page, "%s\n", loadparm); +} + +static struct kobj_attribute sys_ipl_ccw_loadparm_attr = + __ATTR(loadparm, 0444, ipl_ccw_loadparm_show, NULL); + +static struct attribute *ipl_fcp_attrs[] = { + &sys_ipl_device_attr.attr, + &sys_ipl_fcp_wwpn_attr.attr, + &sys_ipl_fcp_lun_attr.attr, + &sys_ipl_fcp_bootprog_attr.attr, + &sys_ipl_fcp_br_lba_attr.attr, + &sys_ipl_ccw_loadparm_attr.attr, + NULL, +}; + +static struct attribute_group ipl_fcp_attr_group = { + .attrs = ipl_fcp_attrs, + .bin_attrs = ipl_fcp_bin_attrs, +}; + +static struct attribute *ipl_nvme_attrs[] = { + &sys_ipl_nvme_fid_attr.attr, + &sys_ipl_nvme_nsid_attr.attr, + &sys_ipl_nvme_bootprog_attr.attr, + &sys_ipl_nvme_br_lba_attr.attr, + &sys_ipl_ccw_loadparm_attr.attr, + NULL, +}; + +static struct attribute_group ipl_nvme_attr_group = { + .attrs = ipl_nvme_attrs, + .bin_attrs = ipl_nvme_bin_attrs, +}; + +static struct attribute *ipl_eckd_attrs[] = { + &sys_ipl_eckd_bootprog_attr.attr, + &sys_ipl_eckd_br_chr_attr.attr, + &sys_ipl_ccw_loadparm_attr.attr, + &sys_ipl_device_attr.attr, + NULL, +}; + +static struct attribute_group ipl_eckd_attr_group = { + .attrs = ipl_eckd_attrs, + .bin_attrs = ipl_eckd_bin_attrs, +}; + +/* CCW ipl device attributes */ + +static struct attribute *ipl_ccw_attrs_vm[] = { + &sys_ipl_device_attr.attr, + &sys_ipl_ccw_loadparm_attr.attr, + &sys_ipl_vm_parm_attr.attr, + NULL, +}; + +static struct attribute *ipl_ccw_attrs_lpar[] = { + &sys_ipl_device_attr.attr, + &sys_ipl_ccw_loadparm_attr.attr, + NULL, +}; + +static struct attribute_group ipl_ccw_attr_group_vm = { + .attrs = ipl_ccw_attrs_vm, +}; + +static struct attribute_group ipl_ccw_attr_group_lpar = { + .attrs = ipl_ccw_attrs_lpar +}; + +static struct attribute *ipl_common_attrs[] = { + &sys_ipl_type_attr.attr, + &sys_ipl_secure_attr.attr, + &sys_ipl_has_secure_attr.attr, + NULL, +}; + +static struct attribute_group ipl_common_attr_group = { + .attrs = ipl_common_attrs, +}; + +static struct kset *ipl_kset; + +static void __ipl_run(void *unused) +{ + diag308(DIAG308_LOAD_CLEAR, NULL); +} + +static void ipl_run(struct shutdown_trigger *trigger) +{ + smp_call_ipl_cpu(__ipl_run, NULL); +} + +static int __init ipl_init(void) +{ + int rc; + + ipl_kset = kset_create_and_add("ipl", NULL, firmware_kobj); + if (!ipl_kset) { + rc = -ENOMEM; + goto out; + } + rc = sysfs_create_group(&ipl_kset->kobj, &ipl_common_attr_group); + if (rc) + goto out; + switch (ipl_info.type) { + case IPL_TYPE_CCW: + if (MACHINE_IS_VM) + rc = sysfs_create_group(&ipl_kset->kobj, + &ipl_ccw_attr_group_vm); + else + rc = sysfs_create_group(&ipl_kset->kobj, + &ipl_ccw_attr_group_lpar); + break; + case IPL_TYPE_ECKD: + case IPL_TYPE_ECKD_DUMP: + rc = sysfs_create_group(&ipl_kset->kobj, &ipl_eckd_attr_group); + break; + case IPL_TYPE_FCP: + case IPL_TYPE_FCP_DUMP: + rc = sysfs_create_group(&ipl_kset->kobj, &ipl_fcp_attr_group); + break; + case IPL_TYPE_NVME: + case IPL_TYPE_NVME_DUMP: + rc = sysfs_create_group(&ipl_kset->kobj, &ipl_nvme_attr_group); + break; + default: + break; + } +out: + if (rc) + panic("ipl_init failed: rc = %i\n", rc); + + return 0; +} + +static struct shutdown_action __refdata ipl_action = { + .name = SHUTDOWN_ACTION_IPL_STR, + .fn = ipl_run, + .init = ipl_init, +}; + +/* + * reipl shutdown action: Reboot Linux on shutdown. + */ + +/* VM IPL PARM attributes */ +static ssize_t reipl_generic_vmparm_show(struct ipl_parameter_block *ipb, + char *page) +{ + char vmparm[DIAG308_VMPARM_SIZE + 1] = {}; + + ipl_block_get_ascii_vmparm(vmparm, sizeof(vmparm), ipb); + return sprintf(page, "%s\n", vmparm); +} + +static ssize_t reipl_generic_vmparm_store(struct ipl_parameter_block *ipb, + size_t vmparm_max, + const char *buf, size_t len) +{ + int i, ip_len; + + /* ignore trailing newline */ + ip_len = len; + if ((len > 0) && (buf[len - 1] == '\n')) + ip_len--; + + if (ip_len > vmparm_max) + return -EINVAL; + + /* parm is used to store kernel options, check for common chars */ + for (i = 0; i < ip_len; i++) + if (!(isalnum(buf[i]) || isascii(buf[i]) || isprint(buf[i]))) + return -EINVAL; + + memset(ipb->ccw.vm_parm, 0, DIAG308_VMPARM_SIZE); + ipb->ccw.vm_parm_len = ip_len; + if (ip_len > 0) { + ipb->ccw.vm_flags |= IPL_PB0_CCW_VM_FLAG_VP; + memcpy(ipb->ccw.vm_parm, buf, ip_len); + ASCEBC(ipb->ccw.vm_parm, ip_len); + } else { + ipb->ccw.vm_flags &= ~IPL_PB0_CCW_VM_FLAG_VP; + } + + return len; +} + +/* NSS wrapper */ +static ssize_t reipl_nss_vmparm_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return reipl_generic_vmparm_show(reipl_block_nss, page); +} + +static ssize_t reipl_nss_vmparm_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + return reipl_generic_vmparm_store(reipl_block_nss, 56, buf, len); +} + +/* CCW wrapper */ +static ssize_t reipl_ccw_vmparm_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return reipl_generic_vmparm_show(reipl_block_ccw, page); +} + +static ssize_t reipl_ccw_vmparm_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + return reipl_generic_vmparm_store(reipl_block_ccw, 64, buf, len); +} + +static struct kobj_attribute sys_reipl_nss_vmparm_attr = + __ATTR(parm, 0644, reipl_nss_vmparm_show, + reipl_nss_vmparm_store); +static struct kobj_attribute sys_reipl_ccw_vmparm_attr = + __ATTR(parm, 0644, reipl_ccw_vmparm_show, + reipl_ccw_vmparm_store); + +/* FCP reipl device attributes */ + +static ssize_t reipl_fcp_scpdata_read(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, + char *buf, loff_t off, size_t count) +{ + size_t size = reipl_block_fcp->fcp.scp_data_len; + void *scp_data = reipl_block_fcp->fcp.scp_data; + + return memory_read_from_buffer(buf, count, &off, scp_data, size); +} + +static ssize_t reipl_fcp_scpdata_write(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, + char *buf, loff_t off, size_t count) +{ + size_t scpdata_len = count; + size_t padding; + + + if (off) + return -EINVAL; + + memcpy(reipl_block_fcp->fcp.scp_data, buf, count); + if (scpdata_len % 8) { + padding = 8 - (scpdata_len % 8); + memset(reipl_block_fcp->fcp.scp_data + scpdata_len, + 0, padding); + scpdata_len += padding; + } + + reipl_block_fcp->hdr.len = IPL_BP_FCP_LEN + scpdata_len; + reipl_block_fcp->fcp.len = IPL_BP0_FCP_LEN + scpdata_len; + reipl_block_fcp->fcp.scp_data_len = scpdata_len; + + return count; +} +static struct bin_attribute sys_reipl_fcp_scp_data_attr = + __BIN_ATTR(scp_data, 0644, reipl_fcp_scpdata_read, + reipl_fcp_scpdata_write, DIAG308_SCPDATA_SIZE); + +static struct bin_attribute *reipl_fcp_bin_attrs[] = { + &sys_reipl_fcp_scp_data_attr, + NULL, +}; + +DEFINE_IPL_ATTR_RW(reipl_fcp, wwpn, "0x%016llx\n", "%llx\n", + reipl_block_fcp->fcp.wwpn); +DEFINE_IPL_ATTR_RW(reipl_fcp, lun, "0x%016llx\n", "%llx\n", + reipl_block_fcp->fcp.lun); +DEFINE_IPL_ATTR_RW(reipl_fcp, bootprog, "%lld\n", "%lld\n", + reipl_block_fcp->fcp.bootprog); +DEFINE_IPL_ATTR_RW(reipl_fcp, br_lba, "%lld\n", "%lld\n", + reipl_block_fcp->fcp.br_lba); +DEFINE_IPL_ATTR_RW(reipl_fcp, device, "0.0.%04llx\n", "0.0.%llx\n", + reipl_block_fcp->fcp.devno); + +static void reipl_get_ascii_loadparm(char *loadparm, + struct ipl_parameter_block *ibp) +{ + memcpy(loadparm, ibp->common.loadparm, LOADPARM_LEN); + EBCASC(loadparm, LOADPARM_LEN); + loadparm[LOADPARM_LEN] = 0; + strim(loadparm); +} + +static ssize_t reipl_generic_loadparm_show(struct ipl_parameter_block *ipb, + char *page) +{ + char buf[LOADPARM_LEN + 1]; + + reipl_get_ascii_loadparm(buf, ipb); + return sprintf(page, "%s\n", buf); +} + +static ssize_t reipl_generic_loadparm_store(struct ipl_parameter_block *ipb, + const char *buf, size_t len) +{ + int i, lp_len; + + /* ignore trailing newline */ + lp_len = len; + if ((len > 0) && (buf[len - 1] == '\n')) + lp_len--; + /* loadparm can have max 8 characters and must not start with a blank */ + if ((lp_len > LOADPARM_LEN) || ((lp_len > 0) && (buf[0] == ' '))) + return -EINVAL; + /* loadparm can only contain "a-z,A-Z,0-9,SP,." */ + for (i = 0; i < lp_len; i++) { + if (isalpha(buf[i]) || isdigit(buf[i]) || (buf[i] == ' ') || + (buf[i] == '.')) + continue; + return -EINVAL; + } + /* initialize loadparm with blanks */ + memset(ipb->common.loadparm, ' ', LOADPARM_LEN); + /* copy and convert to ebcdic */ + memcpy(ipb->common.loadparm, buf, lp_len); + ASCEBC(ipb->common.loadparm, LOADPARM_LEN); + ipb->common.flags |= IPL_PB0_FLAG_LOADPARM; + return len; +} + +#define DEFINE_GENERIC_LOADPARM(name) \ +static ssize_t reipl_##name##_loadparm_show(struct kobject *kobj, \ + struct kobj_attribute *attr, char *page) \ +{ \ + return reipl_generic_loadparm_show(reipl_block_##name, page); \ +} \ +static ssize_t reipl_##name##_loadparm_store(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + const char *buf, size_t len) \ +{ \ + return reipl_generic_loadparm_store(reipl_block_##name, buf, len); \ +} \ +static struct kobj_attribute sys_reipl_##name##_loadparm_attr = \ + __ATTR(loadparm, 0644, reipl_##name##_loadparm_show, \ + reipl_##name##_loadparm_store) + +DEFINE_GENERIC_LOADPARM(fcp); +DEFINE_GENERIC_LOADPARM(nvme); +DEFINE_GENERIC_LOADPARM(ccw); +DEFINE_GENERIC_LOADPARM(nss); +DEFINE_GENERIC_LOADPARM(eckd); + +static ssize_t reipl_fcp_clear_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%u\n", reipl_fcp_clear); +} + +static ssize_t reipl_fcp_clear_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + if (kstrtobool(buf, &reipl_fcp_clear) < 0) + return -EINVAL; + return len; +} + +static struct attribute *reipl_fcp_attrs[] = { + &sys_reipl_fcp_device_attr.attr, + &sys_reipl_fcp_wwpn_attr.attr, + &sys_reipl_fcp_lun_attr.attr, + &sys_reipl_fcp_bootprog_attr.attr, + &sys_reipl_fcp_br_lba_attr.attr, + &sys_reipl_fcp_loadparm_attr.attr, + NULL, +}; + +static struct attribute_group reipl_fcp_attr_group = { + .attrs = reipl_fcp_attrs, + .bin_attrs = reipl_fcp_bin_attrs, +}; + +static struct kobj_attribute sys_reipl_fcp_clear_attr = + __ATTR(clear, 0644, reipl_fcp_clear_show, reipl_fcp_clear_store); + +/* NVME reipl device attributes */ + +static ssize_t reipl_nvme_scpdata_read(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, + char *buf, loff_t off, size_t count) +{ + size_t size = reipl_block_nvme->nvme.scp_data_len; + void *scp_data = reipl_block_nvme->nvme.scp_data; + + return memory_read_from_buffer(buf, count, &off, scp_data, size); +} + +static ssize_t reipl_nvme_scpdata_write(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, + char *buf, loff_t off, size_t count) +{ + size_t scpdata_len = count; + size_t padding; + + if (off) + return -EINVAL; + + memcpy(reipl_block_nvme->nvme.scp_data, buf, count); + if (scpdata_len % 8) { + padding = 8 - (scpdata_len % 8); + memset(reipl_block_nvme->nvme.scp_data + scpdata_len, + 0, padding); + scpdata_len += padding; + } + + reipl_block_nvme->hdr.len = IPL_BP_FCP_LEN + scpdata_len; + reipl_block_nvme->nvme.len = IPL_BP0_FCP_LEN + scpdata_len; + reipl_block_nvme->nvme.scp_data_len = scpdata_len; + + return count; +} + +static struct bin_attribute sys_reipl_nvme_scp_data_attr = + __BIN_ATTR(scp_data, 0644, reipl_nvme_scpdata_read, + reipl_nvme_scpdata_write, DIAG308_SCPDATA_SIZE); + +static struct bin_attribute *reipl_nvme_bin_attrs[] = { + &sys_reipl_nvme_scp_data_attr, + NULL, +}; + +DEFINE_IPL_ATTR_RW(reipl_nvme, fid, "0x%08llx\n", "%llx\n", + reipl_block_nvme->nvme.fid); +DEFINE_IPL_ATTR_RW(reipl_nvme, nsid, "0x%08llx\n", "%llx\n", + reipl_block_nvme->nvme.nsid); +DEFINE_IPL_ATTR_RW(reipl_nvme, bootprog, "%lld\n", "%lld\n", + reipl_block_nvme->nvme.bootprog); +DEFINE_IPL_ATTR_RW(reipl_nvme, br_lba, "%lld\n", "%lld\n", + reipl_block_nvme->nvme.br_lba); + +static struct attribute *reipl_nvme_attrs[] = { + &sys_reipl_nvme_fid_attr.attr, + &sys_reipl_nvme_nsid_attr.attr, + &sys_reipl_nvme_bootprog_attr.attr, + &sys_reipl_nvme_br_lba_attr.attr, + &sys_reipl_nvme_loadparm_attr.attr, + NULL, +}; + +static struct attribute_group reipl_nvme_attr_group = { + .attrs = reipl_nvme_attrs, + .bin_attrs = reipl_nvme_bin_attrs +}; + +static ssize_t reipl_nvme_clear_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%u\n", reipl_nvme_clear); +} + +static ssize_t reipl_nvme_clear_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + if (kstrtobool(buf, &reipl_nvme_clear) < 0) + return -EINVAL; + return len; +} + +static struct kobj_attribute sys_reipl_nvme_clear_attr = + __ATTR(clear, 0644, reipl_nvme_clear_show, reipl_nvme_clear_store); + +/* CCW reipl device attributes */ +DEFINE_IPL_CCW_ATTR_RW(reipl_ccw, device, reipl_block_ccw->ccw); + +static ssize_t reipl_ccw_clear_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%u\n", reipl_ccw_clear); +} + +static ssize_t reipl_ccw_clear_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + if (kstrtobool(buf, &reipl_ccw_clear) < 0) + return -EINVAL; + return len; +} + +static struct kobj_attribute sys_reipl_ccw_clear_attr = + __ATTR(clear, 0644, reipl_ccw_clear_show, reipl_ccw_clear_store); + +static struct attribute *reipl_ccw_attrs_vm[] = { + &sys_reipl_ccw_device_attr.attr, + &sys_reipl_ccw_loadparm_attr.attr, + &sys_reipl_ccw_vmparm_attr.attr, + &sys_reipl_ccw_clear_attr.attr, + NULL, +}; + +static struct attribute *reipl_ccw_attrs_lpar[] = { + &sys_reipl_ccw_device_attr.attr, + &sys_reipl_ccw_loadparm_attr.attr, + &sys_reipl_ccw_clear_attr.attr, + NULL, +}; + +static struct attribute_group reipl_ccw_attr_group_vm = { + .name = IPL_CCW_STR, + .attrs = reipl_ccw_attrs_vm, +}; + +static struct attribute_group reipl_ccw_attr_group_lpar = { + .name = IPL_CCW_STR, + .attrs = reipl_ccw_attrs_lpar, +}; + +/* ECKD reipl device attributes */ + +static ssize_t reipl_eckd_scpdata_read(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, + char *buf, loff_t off, size_t count) +{ + size_t size = reipl_block_eckd->eckd.scp_data_len; + void *scp_data = reipl_block_eckd->eckd.scp_data; + + return memory_read_from_buffer(buf, count, &off, scp_data, size); +} + +static ssize_t reipl_eckd_scpdata_write(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, + char *buf, loff_t off, size_t count) +{ + size_t scpdata_len = count; + size_t padding; + + if (off) + return -EINVAL; + + memcpy(reipl_block_eckd->eckd.scp_data, buf, count); + if (scpdata_len % 8) { + padding = 8 - (scpdata_len % 8); + memset(reipl_block_eckd->eckd.scp_data + scpdata_len, + 0, padding); + scpdata_len += padding; + } + + reipl_block_eckd->hdr.len = IPL_BP_ECKD_LEN + scpdata_len; + reipl_block_eckd->eckd.len = IPL_BP0_ECKD_LEN + scpdata_len; + reipl_block_eckd->eckd.scp_data_len = scpdata_len; + + return count; +} + +static struct bin_attribute sys_reipl_eckd_scp_data_attr = + __BIN_ATTR(scp_data, 0644, reipl_eckd_scpdata_read, + reipl_eckd_scpdata_write, DIAG308_SCPDATA_SIZE); + +static struct bin_attribute *reipl_eckd_bin_attrs[] = { + &sys_reipl_eckd_scp_data_attr, + NULL, +}; + +DEFINE_IPL_CCW_ATTR_RW(reipl_eckd, device, reipl_block_eckd->eckd); +DEFINE_IPL_ATTR_RW(reipl_eckd, bootprog, "%lld\n", "%lld\n", + reipl_block_eckd->eckd.bootprog); + +static struct attribute *reipl_eckd_attrs[] = { + &sys_reipl_eckd_device_attr.attr, + &sys_reipl_eckd_bootprog_attr.attr, + &sys_reipl_eckd_br_chr_attr.attr, + &sys_reipl_eckd_loadparm_attr.attr, + NULL, +}; + +static struct attribute_group reipl_eckd_attr_group = { + .attrs = reipl_eckd_attrs, + .bin_attrs = reipl_eckd_bin_attrs +}; + +static ssize_t reipl_eckd_clear_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%u\n", reipl_eckd_clear); +} + +static ssize_t reipl_eckd_clear_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + if (kstrtobool(buf, &reipl_eckd_clear) < 0) + return -EINVAL; + return len; +} + +static struct kobj_attribute sys_reipl_eckd_clear_attr = + __ATTR(clear, 0644, reipl_eckd_clear_show, reipl_eckd_clear_store); + +/* NSS reipl device attributes */ +static void reipl_get_ascii_nss_name(char *dst, + struct ipl_parameter_block *ipb) +{ + memcpy(dst, ipb->ccw.nss_name, NSS_NAME_SIZE); + EBCASC(dst, NSS_NAME_SIZE); + dst[NSS_NAME_SIZE] = 0; +} + +static ssize_t reipl_nss_name_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + char nss_name[NSS_NAME_SIZE + 1] = {}; + + reipl_get_ascii_nss_name(nss_name, reipl_block_nss); + return sprintf(page, "%s\n", nss_name); +} + +static ssize_t reipl_nss_name_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + int nss_len; + + /* ignore trailing newline */ + nss_len = len; + if ((len > 0) && (buf[len - 1] == '\n')) + nss_len--; + + if (nss_len > NSS_NAME_SIZE) + return -EINVAL; + + memset(reipl_block_nss->ccw.nss_name, 0x40, NSS_NAME_SIZE); + if (nss_len > 0) { + reipl_block_nss->ccw.vm_flags |= IPL_PB0_CCW_VM_FLAG_NSS; + memcpy(reipl_block_nss->ccw.nss_name, buf, nss_len); + ASCEBC(reipl_block_nss->ccw.nss_name, nss_len); + EBC_TOUPPER(reipl_block_nss->ccw.nss_name, nss_len); + } else { + reipl_block_nss->ccw.vm_flags &= ~IPL_PB0_CCW_VM_FLAG_NSS; + } + + return len; +} + +static struct kobj_attribute sys_reipl_nss_name_attr = + __ATTR(name, 0644, reipl_nss_name_show, + reipl_nss_name_store); + +static struct attribute *reipl_nss_attrs[] = { + &sys_reipl_nss_name_attr.attr, + &sys_reipl_nss_loadparm_attr.attr, + &sys_reipl_nss_vmparm_attr.attr, + NULL, +}; + +static struct attribute_group reipl_nss_attr_group = { + .name = IPL_NSS_STR, + .attrs = reipl_nss_attrs, +}; + +void set_os_info_reipl_block(void) +{ + os_info_entry_add(OS_INFO_REIPL_BLOCK, reipl_block_actual, + reipl_block_actual->hdr.len); +} + +/* reipl type */ + +static int reipl_set_type(enum ipl_type type) +{ + if (!(reipl_capabilities & type)) + return -EINVAL; + + switch(type) { + case IPL_TYPE_CCW: + reipl_block_actual = reipl_block_ccw; + break; + case IPL_TYPE_ECKD: + reipl_block_actual = reipl_block_eckd; + break; + case IPL_TYPE_FCP: + reipl_block_actual = reipl_block_fcp; + break; + case IPL_TYPE_NVME: + reipl_block_actual = reipl_block_nvme; + break; + case IPL_TYPE_NSS: + reipl_block_actual = reipl_block_nss; + break; + default: + break; + } + reipl_type = type; + return 0; +} + +static ssize_t reipl_type_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%s\n", ipl_type_str(reipl_type)); +} + +static ssize_t reipl_type_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + int rc = -EINVAL; + + if (strncmp(buf, IPL_CCW_STR, strlen(IPL_CCW_STR)) == 0) + rc = reipl_set_type(IPL_TYPE_CCW); + else if (strncmp(buf, IPL_ECKD_STR, strlen(IPL_ECKD_STR)) == 0) + rc = reipl_set_type(IPL_TYPE_ECKD); + else if (strncmp(buf, IPL_FCP_STR, strlen(IPL_FCP_STR)) == 0) + rc = reipl_set_type(IPL_TYPE_FCP); + else if (strncmp(buf, IPL_NVME_STR, strlen(IPL_NVME_STR)) == 0) + rc = reipl_set_type(IPL_TYPE_NVME); + else if (strncmp(buf, IPL_NSS_STR, strlen(IPL_NSS_STR)) == 0) + rc = reipl_set_type(IPL_TYPE_NSS); + return (rc != 0) ? rc : len; +} + +static struct kobj_attribute reipl_type_attr = + __ATTR(reipl_type, 0644, reipl_type_show, reipl_type_store); + +static struct kset *reipl_kset; +static struct kset *reipl_fcp_kset; +static struct kset *reipl_nvme_kset; +static struct kset *reipl_eckd_kset; + +static void __reipl_run(void *unused) +{ + switch (reipl_type) { + case IPL_TYPE_CCW: + diag308(DIAG308_SET, reipl_block_ccw); + if (reipl_ccw_clear) + diag308(DIAG308_LOAD_CLEAR, NULL); + else + diag308(DIAG308_LOAD_NORMAL_DUMP, NULL); + break; + case IPL_TYPE_ECKD: + diag308(DIAG308_SET, reipl_block_eckd); + if (reipl_eckd_clear) + diag308(DIAG308_LOAD_CLEAR, NULL); + else + diag308(DIAG308_LOAD_NORMAL, NULL); + break; + case IPL_TYPE_FCP: + diag308(DIAG308_SET, reipl_block_fcp); + if (reipl_fcp_clear) + diag308(DIAG308_LOAD_CLEAR, NULL); + else + diag308(DIAG308_LOAD_NORMAL, NULL); + break; + case IPL_TYPE_NVME: + diag308(DIAG308_SET, reipl_block_nvme); + if (reipl_nvme_clear) + diag308(DIAG308_LOAD_CLEAR, NULL); + else + diag308(DIAG308_LOAD_NORMAL, NULL); + break; + case IPL_TYPE_NSS: + diag308(DIAG308_SET, reipl_block_nss); + diag308(DIAG308_LOAD_CLEAR, NULL); + break; + case IPL_TYPE_UNKNOWN: + diag308(DIAG308_LOAD_CLEAR, NULL); + break; + case IPL_TYPE_FCP_DUMP: + case IPL_TYPE_NVME_DUMP: + case IPL_TYPE_ECKD_DUMP: + break; + } + disabled_wait(); +} + +static void reipl_run(struct shutdown_trigger *trigger) +{ + smp_call_ipl_cpu(__reipl_run, NULL); +} + +static void reipl_block_ccw_init(struct ipl_parameter_block *ipb) +{ + ipb->hdr.len = IPL_BP_CCW_LEN; + ipb->hdr.version = IPL_PARM_BLOCK_VERSION; + ipb->pb0_hdr.len = IPL_BP0_CCW_LEN; + ipb->pb0_hdr.pbt = IPL_PBT_CCW; +} + +static void reipl_block_ccw_fill_parms(struct ipl_parameter_block *ipb) +{ + /* LOADPARM */ + /* check if read scp info worked and set loadparm */ + if (sclp_ipl_info.is_valid) + memcpy(ipb->ccw.loadparm, &sclp_ipl_info.loadparm, LOADPARM_LEN); + else + /* read scp info failed: set empty loadparm (EBCDIC blanks) */ + memset(ipb->ccw.loadparm, 0x40, LOADPARM_LEN); + ipb->ccw.flags = IPL_PB0_FLAG_LOADPARM; + + /* VM PARM */ + if (MACHINE_IS_VM && ipl_block_valid && + (ipl_block.ccw.vm_flags & IPL_PB0_CCW_VM_FLAG_VP)) { + + ipb->ccw.vm_flags |= IPL_PB0_CCW_VM_FLAG_VP; + ipb->ccw.vm_parm_len = ipl_block.ccw.vm_parm_len; + memcpy(ipb->ccw.vm_parm, + ipl_block.ccw.vm_parm, DIAG308_VMPARM_SIZE); + } +} + +static int __init reipl_nss_init(void) +{ + int rc; + + if (!MACHINE_IS_VM) + return 0; + + reipl_block_nss = (void *) get_zeroed_page(GFP_KERNEL); + if (!reipl_block_nss) + return -ENOMEM; + + rc = sysfs_create_group(&reipl_kset->kobj, &reipl_nss_attr_group); + if (rc) + return rc; + + reipl_block_ccw_init(reipl_block_nss); + reipl_capabilities |= IPL_TYPE_NSS; + return 0; +} + +static int __init reipl_ccw_init(void) +{ + int rc; + + reipl_block_ccw = (void *) get_zeroed_page(GFP_KERNEL); + if (!reipl_block_ccw) + return -ENOMEM; + + rc = sysfs_create_group(&reipl_kset->kobj, + MACHINE_IS_VM ? &reipl_ccw_attr_group_vm + : &reipl_ccw_attr_group_lpar); + if (rc) + return rc; + + reipl_block_ccw_init(reipl_block_ccw); + if (ipl_info.type == IPL_TYPE_CCW) { + reipl_block_ccw->ccw.ssid = ipl_block.ccw.ssid; + reipl_block_ccw->ccw.devno = ipl_block.ccw.devno; + reipl_block_ccw_fill_parms(reipl_block_ccw); + } + + reipl_capabilities |= IPL_TYPE_CCW; + return 0; +} + +static int __init reipl_fcp_init(void) +{ + int rc; + + reipl_block_fcp = (void *) get_zeroed_page(GFP_KERNEL); + if (!reipl_block_fcp) + return -ENOMEM; + + /* sysfs: create fcp kset for mixing attr group and bin attrs */ + reipl_fcp_kset = kset_create_and_add(IPL_FCP_STR, NULL, + &reipl_kset->kobj); + if (!reipl_fcp_kset) { + free_page((unsigned long) reipl_block_fcp); + return -ENOMEM; + } + + rc = sysfs_create_group(&reipl_fcp_kset->kobj, &reipl_fcp_attr_group); + if (rc) + goto out1; + + if (test_facility(141)) { + rc = sysfs_create_file(&reipl_fcp_kset->kobj, + &sys_reipl_fcp_clear_attr.attr); + if (rc) + goto out2; + } else { + reipl_fcp_clear = true; + } + + if (ipl_info.type == IPL_TYPE_FCP) { + memcpy(reipl_block_fcp, &ipl_block, sizeof(ipl_block)); + /* + * Fix loadparm: There are systems where the (SCSI) LOADPARM + * is invalid in the SCSI IPL parameter block, so take it + * always from sclp_ipl_info. + */ + memcpy(reipl_block_fcp->fcp.loadparm, sclp_ipl_info.loadparm, + LOADPARM_LEN); + } else { + reipl_block_fcp->hdr.len = IPL_BP_FCP_LEN; + reipl_block_fcp->hdr.version = IPL_PARM_BLOCK_VERSION; + reipl_block_fcp->fcp.len = IPL_BP0_FCP_LEN; + reipl_block_fcp->fcp.pbt = IPL_PBT_FCP; + reipl_block_fcp->fcp.opt = IPL_PB0_FCP_OPT_IPL; + } + reipl_capabilities |= IPL_TYPE_FCP; + return 0; + +out2: + sysfs_remove_group(&reipl_fcp_kset->kobj, &reipl_fcp_attr_group); +out1: + kset_unregister(reipl_fcp_kset); + free_page((unsigned long) reipl_block_fcp); + return rc; +} + +static int __init reipl_nvme_init(void) +{ + int rc; + + reipl_block_nvme = (void *) get_zeroed_page(GFP_KERNEL); + if (!reipl_block_nvme) + return -ENOMEM; + + /* sysfs: create kset for mixing attr group and bin attrs */ + reipl_nvme_kset = kset_create_and_add(IPL_NVME_STR, NULL, + &reipl_kset->kobj); + if (!reipl_nvme_kset) { + free_page((unsigned long) reipl_block_nvme); + return -ENOMEM; + } + + rc = sysfs_create_group(&reipl_nvme_kset->kobj, &reipl_nvme_attr_group); + if (rc) + goto out1; + + if (test_facility(141)) { + rc = sysfs_create_file(&reipl_nvme_kset->kobj, + &sys_reipl_nvme_clear_attr.attr); + if (rc) + goto out2; + } else { + reipl_nvme_clear = true; + } + + if (ipl_info.type == IPL_TYPE_NVME) { + memcpy(reipl_block_nvme, &ipl_block, sizeof(ipl_block)); + /* + * Fix loadparm: There are systems where the (SCSI) LOADPARM + * is invalid in the IPL parameter block, so take it + * always from sclp_ipl_info. + */ + memcpy(reipl_block_nvme->nvme.loadparm, sclp_ipl_info.loadparm, + LOADPARM_LEN); + } else { + reipl_block_nvme->hdr.len = IPL_BP_NVME_LEN; + reipl_block_nvme->hdr.version = IPL_PARM_BLOCK_VERSION; + reipl_block_nvme->nvme.len = IPL_BP0_NVME_LEN; + reipl_block_nvme->nvme.pbt = IPL_PBT_NVME; + reipl_block_nvme->nvme.opt = IPL_PB0_NVME_OPT_IPL; + } + reipl_capabilities |= IPL_TYPE_NVME; + return 0; + +out2: + sysfs_remove_group(&reipl_nvme_kset->kobj, &reipl_nvme_attr_group); +out1: + kset_unregister(reipl_nvme_kset); + free_page((unsigned long) reipl_block_nvme); + return rc; +} + +static int __init reipl_eckd_init(void) +{ + int rc; + + if (!sclp.has_sipl_eckd) + return 0; + + reipl_block_eckd = (void *)get_zeroed_page(GFP_KERNEL); + if (!reipl_block_eckd) + return -ENOMEM; + + /* sysfs: create kset for mixing attr group and bin attrs */ + reipl_eckd_kset = kset_create_and_add(IPL_ECKD_STR, NULL, + &reipl_kset->kobj); + if (!reipl_eckd_kset) { + free_page((unsigned long)reipl_block_eckd); + return -ENOMEM; + } + + rc = sysfs_create_group(&reipl_eckd_kset->kobj, &reipl_eckd_attr_group); + if (rc) + goto out1; + + if (test_facility(141)) { + rc = sysfs_create_file(&reipl_eckd_kset->kobj, + &sys_reipl_eckd_clear_attr.attr); + if (rc) + goto out2; + } else { + reipl_eckd_clear = true; + } + + if (ipl_info.type == IPL_TYPE_ECKD) { + memcpy(reipl_block_eckd, &ipl_block, sizeof(ipl_block)); + } else { + reipl_block_eckd->hdr.len = IPL_BP_ECKD_LEN; + reipl_block_eckd->hdr.version = IPL_PARM_BLOCK_VERSION; + reipl_block_eckd->eckd.len = IPL_BP0_ECKD_LEN; + reipl_block_eckd->eckd.pbt = IPL_PBT_ECKD; + reipl_block_eckd->eckd.opt = IPL_PB0_ECKD_OPT_IPL; + } + reipl_capabilities |= IPL_TYPE_ECKD; + return 0; + +out2: + sysfs_remove_group(&reipl_eckd_kset->kobj, &reipl_eckd_attr_group); +out1: + kset_unregister(reipl_eckd_kset); + free_page((unsigned long)reipl_block_eckd); + return rc; +} + +static int __init reipl_type_init(void) +{ + enum ipl_type reipl_type = ipl_info.type; + struct ipl_parameter_block *reipl_block; + unsigned long size; + + reipl_block = os_info_old_entry(OS_INFO_REIPL_BLOCK, &size); + if (!reipl_block) + goto out; + /* + * If we have an OS info reipl block, this will be used + */ + if (reipl_block->pb0_hdr.pbt == IPL_PBT_FCP) { + memcpy(reipl_block_fcp, reipl_block, size); + reipl_type = IPL_TYPE_FCP; + } else if (reipl_block->pb0_hdr.pbt == IPL_PBT_NVME) { + memcpy(reipl_block_nvme, reipl_block, size); + reipl_type = IPL_TYPE_NVME; + } else if (reipl_block->pb0_hdr.pbt == IPL_PBT_CCW) { + memcpy(reipl_block_ccw, reipl_block, size); + reipl_type = IPL_TYPE_CCW; + } else if (reipl_block->pb0_hdr.pbt == IPL_PBT_ECKD) { + memcpy(reipl_block_eckd, reipl_block, size); + reipl_type = IPL_TYPE_ECKD; + } +out: + return reipl_set_type(reipl_type); +} + +static int __init reipl_init(void) +{ + int rc; + + reipl_kset = kset_create_and_add("reipl", NULL, firmware_kobj); + if (!reipl_kset) + return -ENOMEM; + rc = sysfs_create_file(&reipl_kset->kobj, &reipl_type_attr.attr); + if (rc) { + kset_unregister(reipl_kset); + return rc; + } + rc = reipl_ccw_init(); + if (rc) + return rc; + rc = reipl_eckd_init(); + if (rc) + return rc; + rc = reipl_fcp_init(); + if (rc) + return rc; + rc = reipl_nvme_init(); + if (rc) + return rc; + rc = reipl_nss_init(); + if (rc) + return rc; + return reipl_type_init(); +} + +static struct shutdown_action __refdata reipl_action = { + .name = SHUTDOWN_ACTION_REIPL_STR, + .fn = reipl_run, + .init = reipl_init, +}; + +/* + * dump shutdown action: Dump Linux on shutdown. + */ + +/* FCP dump device attributes */ + +DEFINE_IPL_ATTR_RW(dump_fcp, wwpn, "0x%016llx\n", "%llx\n", + dump_block_fcp->fcp.wwpn); +DEFINE_IPL_ATTR_RW(dump_fcp, lun, "0x%016llx\n", "%llx\n", + dump_block_fcp->fcp.lun); +DEFINE_IPL_ATTR_RW(dump_fcp, bootprog, "%lld\n", "%lld\n", + dump_block_fcp->fcp.bootprog); +DEFINE_IPL_ATTR_RW(dump_fcp, br_lba, "%lld\n", "%lld\n", + dump_block_fcp->fcp.br_lba); +DEFINE_IPL_ATTR_RW(dump_fcp, device, "0.0.%04llx\n", "0.0.%llx\n", + dump_block_fcp->fcp.devno); + +static struct attribute *dump_fcp_attrs[] = { + &sys_dump_fcp_device_attr.attr, + &sys_dump_fcp_wwpn_attr.attr, + &sys_dump_fcp_lun_attr.attr, + &sys_dump_fcp_bootprog_attr.attr, + &sys_dump_fcp_br_lba_attr.attr, + NULL, +}; + +static struct attribute_group dump_fcp_attr_group = { + .name = IPL_FCP_STR, + .attrs = dump_fcp_attrs, +}; + +/* NVME dump device attributes */ +DEFINE_IPL_ATTR_RW(dump_nvme, fid, "0x%08llx\n", "%llx\n", + dump_block_nvme->nvme.fid); +DEFINE_IPL_ATTR_RW(dump_nvme, nsid, "0x%08llx\n", "%llx\n", + dump_block_nvme->nvme.nsid); +DEFINE_IPL_ATTR_RW(dump_nvme, bootprog, "%lld\n", "%llx\n", + dump_block_nvme->nvme.bootprog); +DEFINE_IPL_ATTR_RW(dump_nvme, br_lba, "%lld\n", "%llx\n", + dump_block_nvme->nvme.br_lba); + +static struct attribute *dump_nvme_attrs[] = { + &sys_dump_nvme_fid_attr.attr, + &sys_dump_nvme_nsid_attr.attr, + &sys_dump_nvme_bootprog_attr.attr, + &sys_dump_nvme_br_lba_attr.attr, + NULL, +}; + +static struct attribute_group dump_nvme_attr_group = { + .name = IPL_NVME_STR, + .attrs = dump_nvme_attrs, +}; + +/* ECKD dump device attributes */ +DEFINE_IPL_CCW_ATTR_RW(dump_eckd, device, dump_block_eckd->eckd); +DEFINE_IPL_ATTR_RW(dump_eckd, bootprog, "%lld\n", "%llx\n", + dump_block_eckd->eckd.bootprog); + +IPL_ATTR_BR_CHR_SHOW_FN(dump, dump_block_eckd->eckd); +IPL_ATTR_BR_CHR_STORE_FN(dump, dump_block_eckd->eckd); + +static struct kobj_attribute sys_dump_eckd_br_chr_attr = + __ATTR(br_chr, 0644, eckd_dump_br_chr_show, eckd_dump_br_chr_store); + +static struct attribute *dump_eckd_attrs[] = { + &sys_dump_eckd_device_attr.attr, + &sys_dump_eckd_bootprog_attr.attr, + &sys_dump_eckd_br_chr_attr.attr, + NULL, +}; + +static struct attribute_group dump_eckd_attr_group = { + .name = IPL_ECKD_STR, + .attrs = dump_eckd_attrs, +}; + +/* CCW dump device attributes */ +DEFINE_IPL_CCW_ATTR_RW(dump_ccw, device, dump_block_ccw->ccw); + +static struct attribute *dump_ccw_attrs[] = { + &sys_dump_ccw_device_attr.attr, + NULL, +}; + +static struct attribute_group dump_ccw_attr_group = { + .name = IPL_CCW_STR, + .attrs = dump_ccw_attrs, +}; + +/* dump type */ + +static int dump_set_type(enum dump_type type) +{ + if (!(dump_capabilities & type)) + return -EINVAL; + dump_type = type; + return 0; +} + +static ssize_t dump_type_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%s\n", dump_type_str(dump_type)); +} + +static ssize_t dump_type_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + int rc = -EINVAL; + + if (strncmp(buf, DUMP_NONE_STR, strlen(DUMP_NONE_STR)) == 0) + rc = dump_set_type(DUMP_TYPE_NONE); + else if (strncmp(buf, DUMP_CCW_STR, strlen(DUMP_CCW_STR)) == 0) + rc = dump_set_type(DUMP_TYPE_CCW); + else if (strncmp(buf, DUMP_ECKD_STR, strlen(DUMP_ECKD_STR)) == 0) + rc = dump_set_type(DUMP_TYPE_ECKD); + else if (strncmp(buf, DUMP_FCP_STR, strlen(DUMP_FCP_STR)) == 0) + rc = dump_set_type(DUMP_TYPE_FCP); + else if (strncmp(buf, DUMP_NVME_STR, strlen(DUMP_NVME_STR)) == 0) + rc = dump_set_type(DUMP_TYPE_NVME); + return (rc != 0) ? rc : len; +} + +static struct kobj_attribute dump_type_attr = + __ATTR(dump_type, 0644, dump_type_show, dump_type_store); + +static struct kset *dump_kset; + +static void diag308_dump(void *dump_block) +{ + diag308(DIAG308_SET, dump_block); + while (1) { + if (diag308(DIAG308_LOAD_NORMAL_DUMP, NULL) != 0x302) + break; + udelay(USEC_PER_SEC); + } +} + +static void __dump_run(void *unused) +{ + switch (dump_type) { + case DUMP_TYPE_CCW: + diag308_dump(dump_block_ccw); + break; + case DUMP_TYPE_ECKD: + diag308_dump(dump_block_eckd); + break; + case DUMP_TYPE_FCP: + diag308_dump(dump_block_fcp); + break; + case DUMP_TYPE_NVME: + diag308_dump(dump_block_nvme); + break; + default: + break; + } +} + +static void dump_run(struct shutdown_trigger *trigger) +{ + if (dump_type == DUMP_TYPE_NONE) + return; + smp_send_stop(); + smp_call_ipl_cpu(__dump_run, NULL); +} + +static int __init dump_ccw_init(void) +{ + int rc; + + dump_block_ccw = (void *) get_zeroed_page(GFP_KERNEL); + if (!dump_block_ccw) + return -ENOMEM; + rc = sysfs_create_group(&dump_kset->kobj, &dump_ccw_attr_group); + if (rc) { + free_page((unsigned long)dump_block_ccw); + return rc; + } + dump_block_ccw->hdr.len = IPL_BP_CCW_LEN; + dump_block_ccw->hdr.version = IPL_PARM_BLOCK_VERSION; + dump_block_ccw->ccw.len = IPL_BP0_CCW_LEN; + dump_block_ccw->ccw.pbt = IPL_PBT_CCW; + dump_capabilities |= DUMP_TYPE_CCW; + return 0; +} + +static int __init dump_fcp_init(void) +{ + int rc; + + if (!sclp_ipl_info.has_dump) + return 0; /* LDIPL DUMP is not installed */ + dump_block_fcp = (void *) get_zeroed_page(GFP_KERNEL); + if (!dump_block_fcp) + return -ENOMEM; + rc = sysfs_create_group(&dump_kset->kobj, &dump_fcp_attr_group); + if (rc) { + free_page((unsigned long)dump_block_fcp); + return rc; + } + dump_block_fcp->hdr.len = IPL_BP_FCP_LEN; + dump_block_fcp->hdr.version = IPL_PARM_BLOCK_VERSION; + dump_block_fcp->fcp.len = IPL_BP0_FCP_LEN; + dump_block_fcp->fcp.pbt = IPL_PBT_FCP; + dump_block_fcp->fcp.opt = IPL_PB0_FCP_OPT_DUMP; + dump_capabilities |= DUMP_TYPE_FCP; + return 0; +} + +static int __init dump_nvme_init(void) +{ + int rc; + + if (!sclp_ipl_info.has_dump) + return 0; /* LDIPL DUMP is not installed */ + dump_block_nvme = (void *) get_zeroed_page(GFP_KERNEL); + if (!dump_block_nvme) + return -ENOMEM; + rc = sysfs_create_group(&dump_kset->kobj, &dump_nvme_attr_group); + if (rc) { + free_page((unsigned long)dump_block_nvme); + return rc; + } + dump_block_nvme->hdr.len = IPL_BP_NVME_LEN; + dump_block_nvme->hdr.version = IPL_PARM_BLOCK_VERSION; + dump_block_nvme->fcp.len = IPL_BP0_NVME_LEN; + dump_block_nvme->fcp.pbt = IPL_PBT_NVME; + dump_block_nvme->fcp.opt = IPL_PB0_NVME_OPT_DUMP; + dump_capabilities |= DUMP_TYPE_NVME; + return 0; +} + +static int __init dump_eckd_init(void) +{ + int rc; + + if (!sclp_ipl_info.has_dump || !sclp.has_sipl_eckd) + return 0; /* LDIPL DUMP is not installed */ + dump_block_eckd = (void *)get_zeroed_page(GFP_KERNEL); + if (!dump_block_eckd) + return -ENOMEM; + rc = sysfs_create_group(&dump_kset->kobj, &dump_eckd_attr_group); + if (rc) { + free_page((unsigned long)dump_block_eckd); + return rc; + } + dump_block_eckd->hdr.len = IPL_BP_ECKD_LEN; + dump_block_eckd->hdr.version = IPL_PARM_BLOCK_VERSION; + dump_block_eckd->eckd.len = IPL_BP0_ECKD_LEN; + dump_block_eckd->eckd.pbt = IPL_PBT_ECKD; + dump_block_eckd->eckd.opt = IPL_PB0_ECKD_OPT_DUMP; + dump_capabilities |= DUMP_TYPE_ECKD; + return 0; +} + +static int __init dump_init(void) +{ + int rc; + + dump_kset = kset_create_and_add("dump", NULL, firmware_kobj); + if (!dump_kset) + return -ENOMEM; + rc = sysfs_create_file(&dump_kset->kobj, &dump_type_attr.attr); + if (rc) { + kset_unregister(dump_kset); + return rc; + } + rc = dump_ccw_init(); + if (rc) + return rc; + rc = dump_eckd_init(); + if (rc) + return rc; + rc = dump_fcp_init(); + if (rc) + return rc; + rc = dump_nvme_init(); + if (rc) + return rc; + dump_set_type(DUMP_TYPE_NONE); + return 0; +} + +static struct shutdown_action __refdata dump_action = { + .name = SHUTDOWN_ACTION_DUMP_STR, + .fn = dump_run, + .init = dump_init, +}; + +static void dump_reipl_run(struct shutdown_trigger *trigger) +{ + struct lowcore *abs_lc; + unsigned int csum; + + /* + * Set REIPL_CLEAR flag in os_info flags entry indicating + * 'clear' sysfs attribute has been set on the panicked system + * for specified reipl type. + * Always set for IPL_TYPE_NSS and IPL_TYPE_UNKNOWN. + */ + if ((reipl_type == IPL_TYPE_CCW && reipl_ccw_clear) || + (reipl_type == IPL_TYPE_ECKD && reipl_eckd_clear) || + (reipl_type == IPL_TYPE_FCP && reipl_fcp_clear) || + (reipl_type == IPL_TYPE_NVME && reipl_nvme_clear) || + reipl_type == IPL_TYPE_NSS || + reipl_type == IPL_TYPE_UNKNOWN) + os_info_flags |= OS_INFO_FLAG_REIPL_CLEAR; + os_info_entry_add(OS_INFO_FLAGS_ENTRY, &os_info_flags, sizeof(os_info_flags)); + csum = (__force unsigned int) + csum_partial(reipl_block_actual, reipl_block_actual->hdr.len, 0); + abs_lc = get_abs_lowcore(); + abs_lc->ipib = __pa(reipl_block_actual); + abs_lc->ipib_checksum = csum; + put_abs_lowcore(abs_lc); + dump_run(trigger); +} + +static struct shutdown_action __refdata dump_reipl_action = { + .name = SHUTDOWN_ACTION_DUMP_REIPL_STR, + .fn = dump_reipl_run, +}; + +/* + * vmcmd shutdown action: Trigger vm command on shutdown. + */ + +static char vmcmd_on_reboot[128]; +static char vmcmd_on_panic[128]; +static char vmcmd_on_halt[128]; +static char vmcmd_on_poff[128]; +static char vmcmd_on_restart[128]; + +DEFINE_IPL_ATTR_STR_RW(vmcmd, on_reboot, "%s\n", "%s\n", vmcmd_on_reboot); +DEFINE_IPL_ATTR_STR_RW(vmcmd, on_panic, "%s\n", "%s\n", vmcmd_on_panic); +DEFINE_IPL_ATTR_STR_RW(vmcmd, on_halt, "%s\n", "%s\n", vmcmd_on_halt); +DEFINE_IPL_ATTR_STR_RW(vmcmd, on_poff, "%s\n", "%s\n", vmcmd_on_poff); +DEFINE_IPL_ATTR_STR_RW(vmcmd, on_restart, "%s\n", "%s\n", vmcmd_on_restart); + +static struct attribute *vmcmd_attrs[] = { + &sys_vmcmd_on_reboot_attr.attr, + &sys_vmcmd_on_panic_attr.attr, + &sys_vmcmd_on_halt_attr.attr, + &sys_vmcmd_on_poff_attr.attr, + &sys_vmcmd_on_restart_attr.attr, + NULL, +}; + +static struct attribute_group vmcmd_attr_group = { + .attrs = vmcmd_attrs, +}; + +static struct kset *vmcmd_kset; + +static void vmcmd_run(struct shutdown_trigger *trigger) +{ + char *cmd; + + if (strcmp(trigger->name, ON_REIPL_STR) == 0) + cmd = vmcmd_on_reboot; + else if (strcmp(trigger->name, ON_PANIC_STR) == 0) + cmd = vmcmd_on_panic; + else if (strcmp(trigger->name, ON_HALT_STR) == 0) + cmd = vmcmd_on_halt; + else if (strcmp(trigger->name, ON_POFF_STR) == 0) + cmd = vmcmd_on_poff; + else if (strcmp(trigger->name, ON_RESTART_STR) == 0) + cmd = vmcmd_on_restart; + else + return; + + if (strlen(cmd) == 0) + return; + __cpcmd(cmd, NULL, 0, NULL); +} + +static int vmcmd_init(void) +{ + if (!MACHINE_IS_VM) + return -EOPNOTSUPP; + vmcmd_kset = kset_create_and_add("vmcmd", NULL, firmware_kobj); + if (!vmcmd_kset) + return -ENOMEM; + return sysfs_create_group(&vmcmd_kset->kobj, &vmcmd_attr_group); +} + +static struct shutdown_action vmcmd_action = {SHUTDOWN_ACTION_VMCMD_STR, + vmcmd_run, vmcmd_init}; + +/* + * stop shutdown action: Stop Linux on shutdown. + */ + +static void stop_run(struct shutdown_trigger *trigger) +{ + if (strcmp(trigger->name, ON_PANIC_STR) == 0 || + strcmp(trigger->name, ON_RESTART_STR) == 0) + disabled_wait(); + smp_stop_cpu(); +} + +static struct shutdown_action stop_action = {SHUTDOWN_ACTION_STOP_STR, + stop_run, NULL}; + +/* action list */ + +static struct shutdown_action *shutdown_actions_list[] = { + &ipl_action, &reipl_action, &dump_reipl_action, &dump_action, + &vmcmd_action, &stop_action}; +#define SHUTDOWN_ACTIONS_COUNT (sizeof(shutdown_actions_list) / sizeof(void *)) + +/* + * Trigger section + */ + +static struct kset *shutdown_actions_kset; + +static int set_trigger(const char *buf, struct shutdown_trigger *trigger, + size_t len) +{ + int i; + + for (i = 0; i < SHUTDOWN_ACTIONS_COUNT; i++) { + if (sysfs_streq(buf, shutdown_actions_list[i]->name)) { + if (shutdown_actions_list[i]->init_rc) { + return shutdown_actions_list[i]->init_rc; + } else { + trigger->action = shutdown_actions_list[i]; + return len; + } + } + } + return -EINVAL; +} + +/* on reipl */ + +static struct shutdown_trigger on_reboot_trigger = {ON_REIPL_STR, + &reipl_action}; + +static ssize_t on_reboot_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%s\n", on_reboot_trigger.action->name); +} + +static ssize_t on_reboot_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + return set_trigger(buf, &on_reboot_trigger, len); +} +static struct kobj_attribute on_reboot_attr = __ATTR_RW(on_reboot); + +static void do_machine_restart(char *__unused) +{ + smp_send_stop(); + on_reboot_trigger.action->fn(&on_reboot_trigger); + reipl_run(NULL); +} +void (*_machine_restart)(char *command) = do_machine_restart; + +/* on panic */ + +static struct shutdown_trigger on_panic_trigger = {ON_PANIC_STR, &stop_action}; + +static ssize_t on_panic_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%s\n", on_panic_trigger.action->name); +} + +static ssize_t on_panic_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + return set_trigger(buf, &on_panic_trigger, len); +} +static struct kobj_attribute on_panic_attr = __ATTR_RW(on_panic); + +static void do_panic(void) +{ + lgr_info_log(); + on_panic_trigger.action->fn(&on_panic_trigger); + stop_run(&on_panic_trigger); +} + +/* on restart */ + +static struct shutdown_trigger on_restart_trigger = {ON_RESTART_STR, + &stop_action}; + +static ssize_t on_restart_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%s\n", on_restart_trigger.action->name); +} + +static ssize_t on_restart_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + return set_trigger(buf, &on_restart_trigger, len); +} +static struct kobj_attribute on_restart_attr = __ATTR_RW(on_restart); + +static void __do_restart(void *ignore) +{ + smp_send_stop(); +#ifdef CONFIG_CRASH_DUMP + crash_kexec(NULL); +#endif + on_restart_trigger.action->fn(&on_restart_trigger); + stop_run(&on_restart_trigger); +} + +void do_restart(void *arg) +{ + tracing_off(); + debug_locks_off(); + lgr_info_log(); + smp_call_online_cpu(__do_restart, arg); +} + +/* on halt */ + +static struct shutdown_trigger on_halt_trigger = {ON_HALT_STR, &stop_action}; + +static ssize_t on_halt_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%s\n", on_halt_trigger.action->name); +} + +static ssize_t on_halt_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + return set_trigger(buf, &on_halt_trigger, len); +} +static struct kobj_attribute on_halt_attr = __ATTR_RW(on_halt); + +static void do_machine_halt(void) +{ + smp_send_stop(); + on_halt_trigger.action->fn(&on_halt_trigger); + stop_run(&on_halt_trigger); +} +void (*_machine_halt)(void) = do_machine_halt; + +/* on power off */ + +static struct shutdown_trigger on_poff_trigger = {ON_POFF_STR, &stop_action}; + +static ssize_t on_poff_show(struct kobject *kobj, + struct kobj_attribute *attr, char *page) +{ + return sprintf(page, "%s\n", on_poff_trigger.action->name); +} + +static ssize_t on_poff_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t len) +{ + return set_trigger(buf, &on_poff_trigger, len); +} +static struct kobj_attribute on_poff_attr = __ATTR_RW(on_poff); + +static void do_machine_power_off(void) +{ + smp_send_stop(); + on_poff_trigger.action->fn(&on_poff_trigger); + stop_run(&on_poff_trigger); +} +void (*_machine_power_off)(void) = do_machine_power_off; + +static struct attribute *shutdown_action_attrs[] = { + &on_restart_attr.attr, + &on_reboot_attr.attr, + &on_panic_attr.attr, + &on_halt_attr.attr, + &on_poff_attr.attr, + NULL, +}; + +static struct attribute_group shutdown_action_attr_group = { + .attrs = shutdown_action_attrs, +}; + +static void __init shutdown_triggers_init(void) +{ + shutdown_actions_kset = kset_create_and_add("shutdown_actions", NULL, + firmware_kobj); + if (!shutdown_actions_kset) + goto fail; + if (sysfs_create_group(&shutdown_actions_kset->kobj, + &shutdown_action_attr_group)) + goto fail; + return; +fail: + panic("shutdown_triggers_init failed\n"); +} + +static void __init shutdown_actions_init(void) +{ + int i; + + for (i = 0; i < SHUTDOWN_ACTIONS_COUNT; i++) { + if (!shutdown_actions_list[i]->init) + continue; + shutdown_actions_list[i]->init_rc = + shutdown_actions_list[i]->init(); + } +} + +static int __init s390_ipl_init(void) +{ + char str[8] = {0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40}; + + sclp_early_get_ipl_info(&sclp_ipl_info); + /* + * Fix loadparm: There are systems where the (SCSI) LOADPARM + * returned by read SCP info is invalid (contains EBCDIC blanks) + * when the system has been booted via diag308. In that case we use + * the value from diag308, if available. + * + * There are also systems where diag308 store does not work in + * case the system is booted from HMC. Fortunately in this case + * READ SCP info provides the correct value. + */ + if (memcmp(sclp_ipl_info.loadparm, str, sizeof(str)) == 0 && ipl_block_valid) + memcpy(sclp_ipl_info.loadparm, ipl_block.ccw.loadparm, LOADPARM_LEN); + shutdown_actions_init(); + shutdown_triggers_init(); + return 0; +} + +__initcall(s390_ipl_init); + +static void __init strncpy_skip_quote(char *dst, char *src, int n) +{ + int sx, dx; + + dx = 0; + for (sx = 0; src[sx] != 0; sx++) { + if (src[sx] == '"') + continue; + dst[dx++] = src[sx]; + if (dx >= n) + break; + } +} + +static int __init vmcmd_on_reboot_setup(char *str) +{ + if (!MACHINE_IS_VM) + return 1; + strncpy_skip_quote(vmcmd_on_reboot, str, 127); + vmcmd_on_reboot[127] = 0; + on_reboot_trigger.action = &vmcmd_action; + return 1; +} +__setup("vmreboot=", vmcmd_on_reboot_setup); + +static int __init vmcmd_on_panic_setup(char *str) +{ + if (!MACHINE_IS_VM) + return 1; + strncpy_skip_quote(vmcmd_on_panic, str, 127); + vmcmd_on_panic[127] = 0; + on_panic_trigger.action = &vmcmd_action; + return 1; +} +__setup("vmpanic=", vmcmd_on_panic_setup); + +static int __init vmcmd_on_halt_setup(char *str) +{ + if (!MACHINE_IS_VM) + return 1; + strncpy_skip_quote(vmcmd_on_halt, str, 127); + vmcmd_on_halt[127] = 0; + on_halt_trigger.action = &vmcmd_action; + return 1; +} +__setup("vmhalt=", vmcmd_on_halt_setup); + +static int __init vmcmd_on_poff_setup(char *str) +{ + if (!MACHINE_IS_VM) + return 1; + strncpy_skip_quote(vmcmd_on_poff, str, 127); + vmcmd_on_poff[127] = 0; + on_poff_trigger.action = &vmcmd_action; + return 1; +} +__setup("vmpoff=", vmcmd_on_poff_setup); + +static int on_panic_notify(struct notifier_block *self, + unsigned long event, void *data) +{ + do_panic(); + return NOTIFY_OK; +} + +static struct notifier_block on_panic_nb = { + .notifier_call = on_panic_notify, + .priority = INT_MIN, +}; + +void __init setup_ipl(void) +{ + BUILD_BUG_ON(sizeof(struct ipl_parameter_block) != PAGE_SIZE); + + ipl_info.type = get_ipl_type(); + switch (ipl_info.type) { + case IPL_TYPE_CCW: + ipl_info.data.ccw.dev_id.ssid = ipl_block.ccw.ssid; + ipl_info.data.ccw.dev_id.devno = ipl_block.ccw.devno; + break; + case IPL_TYPE_ECKD: + case IPL_TYPE_ECKD_DUMP: + ipl_info.data.eckd.dev_id.ssid = ipl_block.eckd.ssid; + ipl_info.data.eckd.dev_id.devno = ipl_block.eckd.devno; + break; + case IPL_TYPE_FCP: + case IPL_TYPE_FCP_DUMP: + ipl_info.data.fcp.dev_id.ssid = 0; + ipl_info.data.fcp.dev_id.devno = ipl_block.fcp.devno; + ipl_info.data.fcp.wwpn = ipl_block.fcp.wwpn; + ipl_info.data.fcp.lun = ipl_block.fcp.lun; + break; + case IPL_TYPE_NVME: + case IPL_TYPE_NVME_DUMP: + ipl_info.data.nvme.fid = ipl_block.nvme.fid; + ipl_info.data.nvme.nsid = ipl_block.nvme.nsid; + break; + case IPL_TYPE_NSS: + case IPL_TYPE_UNKNOWN: + /* We have no info to copy */ + break; + } + atomic_notifier_chain_register(&panic_notifier_list, &on_panic_nb); +} + +void s390_reset_system(void) +{ + /* Disable prefixing */ + set_prefix(0); + + /* Disable lowcore protection */ + __ctl_clear_bit(0, 28); + diag_amode31_ops.diag308_reset(); +} + +#ifdef CONFIG_KEXEC_FILE + +int ipl_report_add_component(struct ipl_report *report, struct kexec_buf *kbuf, + unsigned char flags, unsigned short cert) +{ + struct ipl_report_component *comp; + + comp = vzalloc(sizeof(*comp)); + if (!comp) + return -ENOMEM; + list_add_tail(&comp->list, &report->components); + + comp->entry.addr = kbuf->mem; + comp->entry.len = kbuf->memsz; + comp->entry.flags = flags; + comp->entry.certificate_index = cert; + + report->size += sizeof(comp->entry); + + return 0; +} + +int ipl_report_add_certificate(struct ipl_report *report, void *key, + unsigned long addr, unsigned long len) +{ + struct ipl_report_certificate *cert; + + cert = vzalloc(sizeof(*cert)); + if (!cert) + return -ENOMEM; + list_add_tail(&cert->list, &report->certificates); + + cert->entry.addr = addr; + cert->entry.len = len; + cert->key = key; + + report->size += sizeof(cert->entry); + report->size += cert->entry.len; + + return 0; +} + +struct ipl_report *ipl_report_init(struct ipl_parameter_block *ipib) +{ + struct ipl_report *report; + + report = vzalloc(sizeof(*report)); + if (!report) + return ERR_PTR(-ENOMEM); + + report->ipib = ipib; + INIT_LIST_HEAD(&report->components); + INIT_LIST_HEAD(&report->certificates); + + report->size = ALIGN(ipib->hdr.len, 8); + report->size += sizeof(struct ipl_rl_hdr); + report->size += sizeof(struct ipl_rb_components); + report->size += sizeof(struct ipl_rb_certificates); + + return report; +} + +void *ipl_report_finish(struct ipl_report *report) +{ + struct ipl_report_certificate *cert; + struct ipl_report_component *comp; + struct ipl_rb_certificates *certs; + struct ipl_parameter_block *ipib; + struct ipl_rb_components *comps; + struct ipl_rl_hdr *rl_hdr; + void *buf, *ptr; + + buf = vzalloc(report->size); + if (!buf) + goto out; + ptr = buf; + + memcpy(ptr, report->ipib, report->ipib->hdr.len); + ipib = ptr; + if (ipl_secure_flag) + ipib->hdr.flags |= IPL_PL_FLAG_SIPL; + ipib->hdr.flags |= IPL_PL_FLAG_IPLSR; + ptr += report->ipib->hdr.len; + ptr = PTR_ALIGN(ptr, 8); + + rl_hdr = ptr; + ptr += sizeof(*rl_hdr); + + comps = ptr; + comps->rbt = IPL_RBT_COMPONENTS; + ptr += sizeof(*comps); + list_for_each_entry(comp, &report->components, list) { + memcpy(ptr, &comp->entry, sizeof(comp->entry)); + ptr += sizeof(comp->entry); + } + comps->len = ptr - (void *)comps; + + certs = ptr; + certs->rbt = IPL_RBT_CERTIFICATES; + ptr += sizeof(*certs); + list_for_each_entry(cert, &report->certificates, list) { + memcpy(ptr, &cert->entry, sizeof(cert->entry)); + ptr += sizeof(cert->entry); + } + certs->len = ptr - (void *)certs; + rl_hdr->len = ptr - (void *)rl_hdr; + + list_for_each_entry(cert, &report->certificates, list) { + memcpy(ptr, cert->key, cert->entry.len); + ptr += cert->entry.len; + } + + BUG_ON(ptr > buf + report->size); +out: + return buf; +} + +int ipl_report_free(struct ipl_report *report) +{ + struct ipl_report_component *comp, *ncomp; + struct ipl_report_certificate *cert, *ncert; + + list_for_each_entry_safe(comp, ncomp, &report->components, list) + vfree(comp); + + list_for_each_entry_safe(cert, ncert, &report->certificates, list) + vfree(cert); + + vfree(report); + + return 0; +} + +#endif diff --git a/arch/s390/kernel/ipl_vmparm.c b/arch/s390/kernel/ipl_vmparm.c new file mode 100644 index 0000000000..b5245fadcf --- /dev/null +++ b/arch/s390/kernel/ipl_vmparm.c @@ -0,0 +1,38 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/minmax.h> +#include <linux/string.h> +#include <asm/ebcdic.h> +#include <asm/ipl.h> + +/* VM IPL PARM routines */ +size_t ipl_block_get_ascii_vmparm(char *dest, size_t size, + const struct ipl_parameter_block *ipb) +{ + int i; + size_t len; + char has_lowercase = 0; + + len = 0; + if ((ipb->ccw.vm_flags & IPL_PB0_CCW_VM_FLAG_VP) && + (ipb->ccw.vm_parm_len > 0)) { + + len = min_t(size_t, size - 1, ipb->ccw.vm_parm_len); + memcpy(dest, ipb->ccw.vm_parm, len); + /* If at least one character is lowercase, we assume mixed + * case; otherwise we convert everything to lowercase. + */ + for (i = 0; i < len; i++) + if ((dest[i] > 0x80 && dest[i] < 0x8a) || /* a-i */ + (dest[i] > 0x90 && dest[i] < 0x9a) || /* j-r */ + (dest[i] > 0xa1 && dest[i] < 0xaa)) { /* s-z */ + has_lowercase = 1; + break; + } + if (!has_lowercase) + EBC_TOLOWER(dest, len); + EBCASC(dest, len); + } + dest[len] = 0; + + return len; +} diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c new file mode 100644 index 0000000000..b020ff17d2 --- /dev/null +++ b/arch/s390/kernel/irq.c @@ -0,0 +1,402 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2004, 2011 + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>, + * Holger Smolinski <Holger.Smolinski@de.ibm.com>, + * Thomas Spatzier <tspat@de.ibm.com>, + * + * This file contains interrupt related functions. + */ + +#include <linux/kernel_stat.h> +#include <linux/interrupt.h> +#include <linux/seq_file.h> +#include <linux/proc_fs.h> +#include <linux/profile.h> +#include <linux/export.h> +#include <linux/kernel.h> +#include <linux/ftrace.h> +#include <linux/errno.h> +#include <linux/slab.h> +#include <linux/init.h> +#include <linux/cpu.h> +#include <linux/irq.h> +#include <linux/entry-common.h> +#include <asm/irq_regs.h> +#include <asm/cputime.h> +#include <asm/lowcore.h> +#include <asm/irq.h> +#include <asm/hw_irq.h> +#include <asm/stacktrace.h> +#include <asm/softirq_stack.h> +#include "entry.h" + +DEFINE_PER_CPU_SHARED_ALIGNED(struct irq_stat, irq_stat); +EXPORT_PER_CPU_SYMBOL_GPL(irq_stat); + +struct irq_class { + int irq; + char *name; + char *desc; +}; + +/* + * The list of "main" irq classes on s390. This is the list of interrupts + * that appear both in /proc/stat ("intr" line) and /proc/interrupts. + * Historically only external and I/O interrupts have been part of /proc/stat. + * We can't add the split external and I/O sub classes since the first field + * in the "intr" line in /proc/stat is supposed to be the sum of all other + * fields. + * Since the external and I/O interrupt fields are already sums we would end + * up with having a sum which accounts each interrupt twice. + */ +static const struct irq_class irqclass_main_desc[NR_IRQS_BASE] = { + {.irq = EXT_INTERRUPT, .name = "EXT"}, + {.irq = IO_INTERRUPT, .name = "I/O"}, + {.irq = THIN_INTERRUPT, .name = "AIO"}, +}; + +/* + * The list of split external and I/O interrupts that appear only in + * /proc/interrupts. + * In addition this list contains non external / I/O events like NMIs. + */ +static const struct irq_class irqclass_sub_desc[] = { + {.irq = IRQEXT_CLK, .name = "CLK", .desc = "[EXT] Clock Comparator"}, + {.irq = IRQEXT_EXC, .name = "EXC", .desc = "[EXT] External Call"}, + {.irq = IRQEXT_EMS, .name = "EMS", .desc = "[EXT] Emergency Signal"}, + {.irq = IRQEXT_TMR, .name = "TMR", .desc = "[EXT] CPU Timer"}, + {.irq = IRQEXT_TLA, .name = "TAL", .desc = "[EXT] Timing Alert"}, + {.irq = IRQEXT_PFL, .name = "PFL", .desc = "[EXT] Pseudo Page Fault"}, + {.irq = IRQEXT_DSD, .name = "DSD", .desc = "[EXT] DASD Diag"}, + {.irq = IRQEXT_VRT, .name = "VRT", .desc = "[EXT] Virtio"}, + {.irq = IRQEXT_SCP, .name = "SCP", .desc = "[EXT] Service Call"}, + {.irq = IRQEXT_IUC, .name = "IUC", .desc = "[EXT] IUCV"}, + {.irq = IRQEXT_CMS, .name = "CMS", .desc = "[EXT] CPU-Measurement: Sampling"}, + {.irq = IRQEXT_CMC, .name = "CMC", .desc = "[EXT] CPU-Measurement: Counter"}, + {.irq = IRQEXT_FTP, .name = "FTP", .desc = "[EXT] HMC FTP Service"}, + {.irq = IRQIO_CIO, .name = "CIO", .desc = "[I/O] Common I/O Layer Interrupt"}, + {.irq = IRQIO_DAS, .name = "DAS", .desc = "[I/O] DASD"}, + {.irq = IRQIO_C15, .name = "C15", .desc = "[I/O] 3215"}, + {.irq = IRQIO_C70, .name = "C70", .desc = "[I/O] 3270"}, + {.irq = IRQIO_TAP, .name = "TAP", .desc = "[I/O] Tape"}, + {.irq = IRQIO_VMR, .name = "VMR", .desc = "[I/O] Unit Record Devices"}, + {.irq = IRQIO_LCS, .name = "LCS", .desc = "[I/O] LCS"}, + {.irq = IRQIO_CTC, .name = "CTC", .desc = "[I/O] CTC"}, + {.irq = IRQIO_ADM, .name = "ADM", .desc = "[I/O] EADM Subchannel"}, + {.irq = IRQIO_CSC, .name = "CSC", .desc = "[I/O] CHSC Subchannel"}, + {.irq = IRQIO_VIR, .name = "VIR", .desc = "[I/O] Virtual I/O Devices"}, + {.irq = IRQIO_QAI, .name = "QAI", .desc = "[AIO] QDIO Adapter Interrupt"}, + {.irq = IRQIO_APB, .name = "APB", .desc = "[AIO] AP Bus"}, + {.irq = IRQIO_PCF, .name = "PCF", .desc = "[AIO] PCI Floating Interrupt"}, + {.irq = IRQIO_PCD, .name = "PCD", .desc = "[AIO] PCI Directed Interrupt"}, + {.irq = IRQIO_MSI, .name = "MSI", .desc = "[AIO] MSI Interrupt"}, + {.irq = IRQIO_VAI, .name = "VAI", .desc = "[AIO] Virtual I/O Devices AI"}, + {.irq = IRQIO_GAL, .name = "GAL", .desc = "[AIO] GIB Alert"}, + {.irq = NMI_NMI, .name = "NMI", .desc = "[NMI] Machine Check"}, + {.irq = CPU_RST, .name = "RST", .desc = "[CPU] CPU Restart"}, +}; + +static void do_IRQ(struct pt_regs *regs, int irq) +{ + if (tod_after_eq(S390_lowcore.int_clock, + S390_lowcore.clock_comparator)) + /* Serve timer interrupts first. */ + clock_comparator_work(); + generic_handle_irq(irq); +} + +static int on_async_stack(void) +{ + unsigned long frame = current_frame_address(); + + return ((S390_lowcore.async_stack ^ frame) & ~(THREAD_SIZE - 1)) == 0; +} + +static void do_irq_async(struct pt_regs *regs, int irq) +{ + if (on_async_stack()) { + do_IRQ(regs, irq); + } else { + call_on_stack(2, S390_lowcore.async_stack, void, do_IRQ, + struct pt_regs *, regs, int, irq); + } +} + +static int irq_pending(struct pt_regs *regs) +{ + int cc; + + asm volatile("tpi 0\n" + "ipm %0" : "=d" (cc) : : "cc"); + return cc >> 28; +} + +void noinstr do_io_irq(struct pt_regs *regs) +{ + irqentry_state_t state = irqentry_enter(regs); + struct pt_regs *old_regs = set_irq_regs(regs); + bool from_idle; + + irq_enter_rcu(); + + if (user_mode(regs)) { + update_timer_sys(); + if (static_branch_likely(&cpu_has_bear)) + current->thread.last_break = regs->last_break; + } + + from_idle = test_and_clear_cpu_flag(CIF_ENABLED_WAIT); + if (from_idle) + account_idle_time_irq(); + + do { + regs->tpi_info = S390_lowcore.tpi_info; + if (S390_lowcore.tpi_info.adapter_IO) + do_irq_async(regs, THIN_INTERRUPT); + else + do_irq_async(regs, IO_INTERRUPT); + } while (MACHINE_IS_LPAR && irq_pending(regs)); + + irq_exit_rcu(); + + set_irq_regs(old_regs); + irqentry_exit(regs, state); + + if (from_idle) + regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT); +} + +void noinstr do_ext_irq(struct pt_regs *regs) +{ + irqentry_state_t state = irqentry_enter(regs); + struct pt_regs *old_regs = set_irq_regs(regs); + bool from_idle; + + irq_enter_rcu(); + + if (user_mode(regs)) { + update_timer_sys(); + if (static_branch_likely(&cpu_has_bear)) + current->thread.last_break = regs->last_break; + } + + regs->int_code = S390_lowcore.ext_int_code_addr; + regs->int_parm = S390_lowcore.ext_params; + regs->int_parm_long = S390_lowcore.ext_params2; + + from_idle = test_and_clear_cpu_flag(CIF_ENABLED_WAIT); + if (from_idle) + account_idle_time_irq(); + + do_irq_async(regs, EXT_INTERRUPT); + + irq_exit_rcu(); + set_irq_regs(old_regs); + irqentry_exit(regs, state); + + if (from_idle) + regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT); +} + +static void show_msi_interrupt(struct seq_file *p, int irq) +{ + struct irq_desc *desc; + unsigned long flags; + int cpu; + + rcu_read_lock(); + desc = irq_to_desc(irq); + if (!desc) + goto out; + + raw_spin_lock_irqsave(&desc->lock, flags); + seq_printf(p, "%3d: ", irq); + for_each_online_cpu(cpu) + seq_printf(p, "%10u ", irq_desc_kstat_cpu(desc, cpu)); + + if (desc->irq_data.chip) + seq_printf(p, " %8s", desc->irq_data.chip->name); + + if (desc->action) + seq_printf(p, " %s", desc->action->name); + + seq_putc(p, '\n'); + raw_spin_unlock_irqrestore(&desc->lock, flags); +out: + rcu_read_unlock(); +} + +/* + * show_interrupts is needed by /proc/interrupts. + */ +int show_interrupts(struct seq_file *p, void *v) +{ + int index = *(loff_t *) v; + int cpu, irq; + + cpus_read_lock(); + if (index == 0) { + seq_puts(p, " "); + for_each_online_cpu(cpu) + seq_printf(p, "CPU%-8d", cpu); + seq_putc(p, '\n'); + } + if (index < NR_IRQS_BASE) { + seq_printf(p, "%s: ", irqclass_main_desc[index].name); + irq = irqclass_main_desc[index].irq; + for_each_online_cpu(cpu) + seq_printf(p, "%10u ", kstat_irqs_cpu(irq, cpu)); + seq_putc(p, '\n'); + goto out; + } + if (index < nr_irqs) { + show_msi_interrupt(p, index); + goto out; + } + for (index = 0; index < NR_ARCH_IRQS; index++) { + seq_printf(p, "%s: ", irqclass_sub_desc[index].name); + irq = irqclass_sub_desc[index].irq; + for_each_online_cpu(cpu) + seq_printf(p, "%10u ", + per_cpu(irq_stat, cpu).irqs[irq]); + if (irqclass_sub_desc[index].desc) + seq_printf(p, " %s", irqclass_sub_desc[index].desc); + seq_putc(p, '\n'); + } +out: + cpus_read_unlock(); + return 0; +} + +unsigned int arch_dynirq_lower_bound(unsigned int from) +{ + return from < NR_IRQS_BASE ? NR_IRQS_BASE : from; +} + +/* + * ext_int_hash[index] is the list head for all external interrupts that hash + * to this index. + */ +static struct hlist_head ext_int_hash[32] ____cacheline_aligned; + +struct ext_int_info { + ext_int_handler_t handler; + struct hlist_node entry; + struct rcu_head rcu; + u16 code; +}; + +/* ext_int_hash_lock protects the handler lists for external interrupts */ +static DEFINE_SPINLOCK(ext_int_hash_lock); + +static inline int ext_hash(u16 code) +{ + BUILD_BUG_ON(!is_power_of_2(ARRAY_SIZE(ext_int_hash))); + + return (code + (code >> 9)) & (ARRAY_SIZE(ext_int_hash) - 1); +} + +int register_external_irq(u16 code, ext_int_handler_t handler) +{ + struct ext_int_info *p; + unsigned long flags; + int index; + + p = kmalloc(sizeof(*p), GFP_ATOMIC); + if (!p) + return -ENOMEM; + p->code = code; + p->handler = handler; + index = ext_hash(code); + + spin_lock_irqsave(&ext_int_hash_lock, flags); + hlist_add_head_rcu(&p->entry, &ext_int_hash[index]); + spin_unlock_irqrestore(&ext_int_hash_lock, flags); + return 0; +} +EXPORT_SYMBOL(register_external_irq); + +int unregister_external_irq(u16 code, ext_int_handler_t handler) +{ + struct ext_int_info *p; + unsigned long flags; + int index = ext_hash(code); + + spin_lock_irqsave(&ext_int_hash_lock, flags); + hlist_for_each_entry_rcu(p, &ext_int_hash[index], entry) { + if (p->code == code && p->handler == handler) { + hlist_del_rcu(&p->entry); + kfree_rcu(p, rcu); + } + } + spin_unlock_irqrestore(&ext_int_hash_lock, flags); + return 0; +} +EXPORT_SYMBOL(unregister_external_irq); + +static irqreturn_t do_ext_interrupt(int irq, void *dummy) +{ + struct pt_regs *regs = get_irq_regs(); + struct ext_code ext_code; + struct ext_int_info *p; + int index; + + ext_code.int_code = regs->int_code; + if (ext_code.code != EXT_IRQ_CLK_COMP) + set_cpu_flag(CIF_NOHZ_DELAY); + + index = ext_hash(ext_code.code); + rcu_read_lock(); + hlist_for_each_entry_rcu(p, &ext_int_hash[index], entry) { + if (unlikely(p->code != ext_code.code)) + continue; + p->handler(ext_code, regs->int_parm, regs->int_parm_long); + } + rcu_read_unlock(); + return IRQ_HANDLED; +} + +static void __init init_ext_interrupts(void) +{ + int idx; + + for (idx = 0; idx < ARRAY_SIZE(ext_int_hash); idx++) + INIT_HLIST_HEAD(&ext_int_hash[idx]); + + irq_set_chip_and_handler(EXT_INTERRUPT, + &dummy_irq_chip, handle_percpu_irq); + if (request_irq(EXT_INTERRUPT, do_ext_interrupt, 0, "EXT", NULL)) + panic("Failed to register EXT interrupt\n"); +} + +void __init init_IRQ(void) +{ + BUILD_BUG_ON(ARRAY_SIZE(irqclass_sub_desc) != NR_ARCH_IRQS); + init_cio_interrupts(); + init_airq_interrupts(); + init_ext_interrupts(); +} + +static DEFINE_SPINLOCK(irq_subclass_lock); +static unsigned char irq_subclass_refcount[64]; + +void irq_subclass_register(enum irq_subclass subclass) +{ + spin_lock(&irq_subclass_lock); + if (!irq_subclass_refcount[subclass]) + ctl_set_bit(0, subclass); + irq_subclass_refcount[subclass]++; + spin_unlock(&irq_subclass_lock); +} +EXPORT_SYMBOL(irq_subclass_register); + +void irq_subclass_unregister(enum irq_subclass subclass) +{ + spin_lock(&irq_subclass_lock); + irq_subclass_refcount[subclass]--; + if (!irq_subclass_refcount[subclass]) + ctl_clear_bit(0, subclass); + spin_unlock(&irq_subclass_lock); +} +EXPORT_SYMBOL(irq_subclass_unregister); diff --git a/arch/s390/kernel/jump_label.c b/arch/s390/kernel/jump_label.c new file mode 100644 index 0000000000..e808bb8bc0 --- /dev/null +++ b/arch/s390/kernel/jump_label.c @@ -0,0 +1,82 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Jump label s390 support + * + * Copyright IBM Corp. 2011 + * Author(s): Jan Glauber <jang@linux.vnet.ibm.com> + */ +#include <linux/uaccess.h> +#include <linux/jump_label.h> +#include <linux/module.h> +#include <asm/text-patching.h> +#include <asm/ipl.h> + +struct insn { + u16 opcode; + s32 offset; +} __packed; + +static void jump_label_make_nop(struct jump_entry *entry, struct insn *insn) +{ + /* brcl 0,offset */ + insn->opcode = 0xc004; + insn->offset = (jump_entry_target(entry) - jump_entry_code(entry)) >> 1; +} + +static void jump_label_make_branch(struct jump_entry *entry, struct insn *insn) +{ + /* brcl 15,offset */ + insn->opcode = 0xc0f4; + insn->offset = (jump_entry_target(entry) - jump_entry_code(entry)) >> 1; +} + +static void jump_label_bug(struct jump_entry *entry, struct insn *expected, + struct insn *new) +{ + unsigned char *ipc = (unsigned char *)jump_entry_code(entry); + unsigned char *ipe = (unsigned char *)expected; + unsigned char *ipn = (unsigned char *)new; + + pr_emerg("Jump label code mismatch at %pS [%px]\n", ipc, ipc); + pr_emerg("Found: %6ph\n", ipc); + pr_emerg("Expected: %6ph\n", ipe); + pr_emerg("New: %6ph\n", ipn); + panic("Corrupted kernel text"); +} + +static void jump_label_transform(struct jump_entry *entry, + enum jump_label_type type) +{ + void *code = (void *)jump_entry_code(entry); + struct insn old, new; + + if (type == JUMP_LABEL_JMP) { + jump_label_make_nop(entry, &old); + jump_label_make_branch(entry, &new); + } else { + jump_label_make_branch(entry, &old); + jump_label_make_nop(entry, &new); + } + if (memcmp(code, &old, sizeof(old))) + jump_label_bug(entry, &old, &new); + s390_kernel_write(code, &new, sizeof(new)); +} + +void arch_jump_label_transform(struct jump_entry *entry, + enum jump_label_type type) +{ + jump_label_transform(entry, type); + text_poke_sync(); +} + +bool arch_jump_label_transform_queue(struct jump_entry *entry, + enum jump_label_type type) +{ + jump_label_transform(entry, type); + return true; +} + +void arch_jump_label_transform_apply(void) +{ + text_poke_sync(); +} diff --git a/arch/s390/kernel/kdebugfs.c b/arch/s390/kernel/kdebugfs.c new file mode 100644 index 0000000000..33130c7daf --- /dev/null +++ b/arch/s390/kernel/kdebugfs.c @@ -0,0 +1,14 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/debugfs.h> +#include <linux/export.h> +#include <linux/init.h> + +struct dentry *arch_debugfs_dir; +EXPORT_SYMBOL(arch_debugfs_dir); + +static int __init arch_kdebugfs_init(void) +{ + arch_debugfs_dir = debugfs_create_dir("s390", NULL); + return 0; +} +postcore_initcall(arch_kdebugfs_init); diff --git a/arch/s390/kernel/kexec_elf.c b/arch/s390/kernel/kexec_elf.c new file mode 100644 index 0000000000..9da6fa30c4 --- /dev/null +++ b/arch/s390/kernel/kexec_elf.c @@ -0,0 +1,136 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * ELF loader for kexec_file_load system call. + * + * Copyright IBM Corp. 2018 + * + * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com> + */ + +#include <linux/errno.h> +#include <linux/kernel.h> +#include <linux/kexec.h> +#include <asm/ipl.h> +#include <asm/setup.h> + +static int kexec_file_add_kernel_elf(struct kimage *image, + struct s390_load_data *data) +{ + struct kexec_buf buf; + const Elf_Ehdr *ehdr; + const Elf_Phdr *phdr; + Elf_Addr entry; + void *kernel; + int i, ret; + + kernel = image->kernel_buf; + ehdr = (Elf_Ehdr *)kernel; + buf.image = image; + if (image->type == KEXEC_TYPE_CRASH) + entry = STARTUP_KDUMP_OFFSET; + else + entry = ehdr->e_entry; + + phdr = (void *)ehdr + ehdr->e_phoff; + for (i = 0; i < ehdr->e_phnum; i++, phdr++) { + if (phdr->p_type != PT_LOAD) + continue; + + buf.buffer = kernel + phdr->p_offset; + buf.bufsz = phdr->p_filesz; + + buf.mem = ALIGN(phdr->p_paddr, phdr->p_align); + if (image->type == KEXEC_TYPE_CRASH) + buf.mem += crashk_res.start; + buf.memsz = phdr->p_memsz; + data->memsz = ALIGN(data->memsz, phdr->p_align) + buf.memsz; + + if (entry - phdr->p_paddr < phdr->p_memsz) { + data->kernel_buf = buf.buffer; + data->kernel_mem = buf.mem; + data->parm = buf.buffer + PARMAREA; + } + + ipl_report_add_component(data->report, &buf, + IPL_RB_COMPONENT_FLAG_SIGNED | + IPL_RB_COMPONENT_FLAG_VERIFIED, + IPL_RB_CERT_UNKNOWN); + ret = kexec_add_buffer(&buf); + if (ret) + return ret; + } + + return data->memsz ? 0 : -EINVAL; +} + +static void *s390_elf_load(struct kimage *image, + char *kernel, unsigned long kernel_len, + char *initrd, unsigned long initrd_len, + char *cmdline, unsigned long cmdline_len) +{ + const Elf_Ehdr *ehdr; + const Elf_Phdr *phdr; + size_t size; + int i; + + /* image->fobs->probe already checked for valid ELF magic number. */ + ehdr = (Elf_Ehdr *)kernel; + + if (ehdr->e_type != ET_EXEC || + ehdr->e_ident[EI_CLASS] != ELFCLASS64 || + !elf_check_arch(ehdr)) + return ERR_PTR(-EINVAL); + + if (!ehdr->e_phnum || ehdr->e_phentsize != sizeof(Elf_Phdr)) + return ERR_PTR(-EINVAL); + + size = ehdr->e_ehsize + ehdr->e_phoff; + size += ehdr->e_phentsize * ehdr->e_phnum; + if (size > kernel_len) + return ERR_PTR(-EINVAL); + + phdr = (void *)ehdr + ehdr->e_phoff; + size = ALIGN(size, phdr->p_align); + for (i = 0; i < ehdr->e_phnum; i++, phdr++) { + if (phdr->p_type == PT_INTERP) + return ERR_PTR(-EINVAL); + + if (phdr->p_offset > kernel_len) + return ERR_PTR(-EINVAL); + + size += ALIGN(phdr->p_filesz, phdr->p_align); + } + + if (size > kernel_len) + return ERR_PTR(-EINVAL); + + return kexec_file_add_components(image, kexec_file_add_kernel_elf); +} + +static int s390_elf_probe(const char *buf, unsigned long len) +{ + const Elf_Ehdr *ehdr; + + if (len < sizeof(Elf_Ehdr)) + return -ENOEXEC; + + ehdr = (Elf_Ehdr *)buf; + + /* Only check the ELF magic number here and do proper validity check + * in the loader. Any check here that fails would send the erroneous + * ELF file to the image loader that does not care what it gets. + * (Most likely) causing behavior not intended by the user. + */ + if (memcmp(ehdr->e_ident, ELFMAG, SELFMAG) != 0) + return -ENOEXEC; + + return 0; +} + +const struct kexec_file_ops s390_kexec_elf_ops = { + .probe = s390_elf_probe, + .load = s390_elf_load, +#ifdef CONFIG_KEXEC_SIG + .verify_sig = s390_verify_sig, +#endif /* CONFIG_KEXEC_SIG */ +}; diff --git a/arch/s390/kernel/kexec_image.c b/arch/s390/kernel/kexec_image.c new file mode 100644 index 0000000000..af23eff577 --- /dev/null +++ b/arch/s390/kernel/kexec_image.c @@ -0,0 +1,65 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Image loader for kexec_file_load system call. + * + * Copyright IBM Corp. 2018 + * + * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com> + */ + +#include <linux/errno.h> +#include <linux/kernel.h> +#include <linux/kexec.h> +#include <asm/ipl.h> +#include <asm/setup.h> + +static int kexec_file_add_kernel_image(struct kimage *image, + struct s390_load_data *data) +{ + struct kexec_buf buf; + + buf.image = image; + + buf.buffer = image->kernel_buf; + buf.bufsz = image->kernel_buf_len; + + buf.mem = 0; + if (image->type == KEXEC_TYPE_CRASH) + buf.mem += crashk_res.start; + buf.memsz = buf.bufsz; + + data->kernel_buf = image->kernel_buf; + data->kernel_mem = buf.mem; + data->parm = image->kernel_buf + PARMAREA; + data->memsz += buf.memsz; + + ipl_report_add_component(data->report, &buf, + IPL_RB_COMPONENT_FLAG_SIGNED | + IPL_RB_COMPONENT_FLAG_VERIFIED, + IPL_RB_CERT_UNKNOWN); + return kexec_add_buffer(&buf); +} + +static void *s390_image_load(struct kimage *image, + char *kernel, unsigned long kernel_len, + char *initrd, unsigned long initrd_len, + char *cmdline, unsigned long cmdline_len) +{ + return kexec_file_add_components(image, kexec_file_add_kernel_image); +} + +static int s390_image_probe(const char *buf, unsigned long len) +{ + /* Can't reliably tell if an image is valid. Therefore give the + * user whatever he wants. + */ + return 0; +} + +const struct kexec_file_ops s390_kexec_image_ops = { + .probe = s390_image_probe, + .load = s390_image_load, +#ifdef CONFIG_KEXEC_SIG + .verify_sig = s390_verify_sig, +#endif /* CONFIG_KEXEC_SIG */ +}; diff --git a/arch/s390/kernel/kprobes.c b/arch/s390/kernel/kprobes.c new file mode 100644 index 0000000000..d4b863ed0a --- /dev/null +++ b/arch/s390/kernel/kprobes.c @@ -0,0 +1,525 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * Kernel Probes (KProbes) + * + * Copyright IBM Corp. 2002, 2006 + * + * s390 port, used ppc64 as template. Mike Grundy <grundym@us.ibm.com> + */ + +#define pr_fmt(fmt) "kprobes: " fmt + +#include <linux/moduleloader.h> +#include <linux/kprobes.h> +#include <linux/ptrace.h> +#include <linux/preempt.h> +#include <linux/stop_machine.h> +#include <linux/kdebug.h> +#include <linux/uaccess.h> +#include <linux/extable.h> +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/hardirq.h> +#include <linux/ftrace.h> +#include <asm/set_memory.h> +#include <asm/sections.h> +#include <asm/dis.h> +#include "kprobes.h" +#include "entry.h" + +DEFINE_PER_CPU(struct kprobe *, current_kprobe); +DEFINE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk); + +struct kretprobe_blackpoint kretprobe_blacklist[] = { }; + +static int insn_page_in_use; + +void *alloc_insn_page(void) +{ + void *page; + + page = module_alloc(PAGE_SIZE); + if (!page) + return NULL; + set_memory_rox((unsigned long)page, 1); + return page; +} + +static void *alloc_s390_insn_page(void) +{ + if (xchg(&insn_page_in_use, 1) == 1) + return NULL; + return &kprobes_insn_page; +} + +static void free_s390_insn_page(void *page) +{ + xchg(&insn_page_in_use, 0); +} + +struct kprobe_insn_cache kprobe_s390_insn_slots = { + .mutex = __MUTEX_INITIALIZER(kprobe_s390_insn_slots.mutex), + .alloc = alloc_s390_insn_page, + .free = free_s390_insn_page, + .pages = LIST_HEAD_INIT(kprobe_s390_insn_slots.pages), + .insn_size = MAX_INSN_SIZE, +}; + +static void copy_instruction(struct kprobe *p) +{ + kprobe_opcode_t insn[MAX_INSN_SIZE]; + s64 disp, new_disp; + u64 addr, new_addr; + unsigned int len; + + len = insn_length(*p->addr >> 8); + memcpy(&insn, p->addr, len); + p->opcode = insn[0]; + if (probe_is_insn_relative_long(&insn[0])) { + /* + * For pc-relative instructions in RIL-b or RIL-c format patch + * the RI2 displacement field. We have already made sure that + * the insn slot for the patched instruction is within the same + * 2GB area as the original instruction (either kernel image or + * module area). Therefore the new displacement will always fit. + */ + disp = *(s32 *)&insn[1]; + addr = (u64)(unsigned long)p->addr; + new_addr = (u64)(unsigned long)p->ainsn.insn; + new_disp = ((addr + (disp * 2)) - new_addr) / 2; + *(s32 *)&insn[1] = new_disp; + } + s390_kernel_write(p->ainsn.insn, &insn, len); +} +NOKPROBE_SYMBOL(copy_instruction); + +static int s390_get_insn_slot(struct kprobe *p) +{ + /* + * Get an insn slot that is within the same 2GB area like the original + * instruction. That way instructions with a 32bit signed displacement + * field can be patched and executed within the insn slot. + */ + p->ainsn.insn = NULL; + if (is_kernel((unsigned long)p->addr)) + p->ainsn.insn = get_s390_insn_slot(); + else if (is_module_addr(p->addr)) + p->ainsn.insn = get_insn_slot(); + return p->ainsn.insn ? 0 : -ENOMEM; +} +NOKPROBE_SYMBOL(s390_get_insn_slot); + +static void s390_free_insn_slot(struct kprobe *p) +{ + if (!p->ainsn.insn) + return; + if (is_kernel((unsigned long)p->addr)) + free_s390_insn_slot(p->ainsn.insn, 0); + else + free_insn_slot(p->ainsn.insn, 0); + p->ainsn.insn = NULL; +} +NOKPROBE_SYMBOL(s390_free_insn_slot); + +/* Check if paddr is at an instruction boundary */ +static bool can_probe(unsigned long paddr) +{ + unsigned long addr, offset = 0; + kprobe_opcode_t insn; + struct kprobe *kp; + + if (paddr & 0x01) + return false; + + if (!kallsyms_lookup_size_offset(paddr, NULL, &offset)) + return false; + + /* Decode instructions */ + addr = paddr - offset; + while (addr < paddr) { + if (copy_from_kernel_nofault(&insn, (void *)addr, sizeof(insn))) + return false; + + if (insn >> 8 == 0) { + if (insn != BREAKPOINT_INSTRUCTION) { + /* + * Note that QEMU inserts opcode 0x0000 to implement + * software breakpoints for guests. Since the size of + * the original instruction is unknown, stop following + * instructions and prevent setting a kprobe. + */ + return false; + } + /* + * Check if the instruction has been modified by another + * kprobe, in which case the original instruction is + * decoded. + */ + kp = get_kprobe((void *)addr); + if (!kp) { + /* not a kprobe */ + return false; + } + insn = kp->opcode; + } + addr += insn_length(insn >> 8); + } + return addr == paddr; +} + +int arch_prepare_kprobe(struct kprobe *p) +{ + if (!can_probe((unsigned long)p->addr)) + return -EINVAL; + /* Make sure the probe isn't going on a difficult instruction */ + if (probe_is_prohibited_opcode(p->addr)) + return -EINVAL; + if (s390_get_insn_slot(p)) + return -ENOMEM; + copy_instruction(p); + return 0; +} +NOKPROBE_SYMBOL(arch_prepare_kprobe); + +struct swap_insn_args { + struct kprobe *p; + unsigned int arm_kprobe : 1; +}; + +static int swap_instruction(void *data) +{ + struct swap_insn_args *args = data; + struct kprobe *p = args->p; + u16 opc; + + opc = args->arm_kprobe ? BREAKPOINT_INSTRUCTION : p->opcode; + s390_kernel_write(p->addr, &opc, sizeof(opc)); + return 0; +} +NOKPROBE_SYMBOL(swap_instruction); + +void arch_arm_kprobe(struct kprobe *p) +{ + struct swap_insn_args args = {.p = p, .arm_kprobe = 1}; + + stop_machine_cpuslocked(swap_instruction, &args, NULL); +} +NOKPROBE_SYMBOL(arch_arm_kprobe); + +void arch_disarm_kprobe(struct kprobe *p) +{ + struct swap_insn_args args = {.p = p, .arm_kprobe = 0}; + + stop_machine_cpuslocked(swap_instruction, &args, NULL); +} +NOKPROBE_SYMBOL(arch_disarm_kprobe); + +void arch_remove_kprobe(struct kprobe *p) +{ + s390_free_insn_slot(p); +} +NOKPROBE_SYMBOL(arch_remove_kprobe); + +static void enable_singlestep(struct kprobe_ctlblk *kcb, + struct pt_regs *regs, + unsigned long ip) +{ + struct per_regs per_kprobe; + + /* Set up the PER control registers %cr9-%cr11 */ + per_kprobe.control = PER_EVENT_IFETCH; + per_kprobe.start = ip; + per_kprobe.end = ip; + + /* Save control regs and psw mask */ + __ctl_store(kcb->kprobe_saved_ctl, 9, 11); + kcb->kprobe_saved_imask = regs->psw.mask & + (PSW_MASK_PER | PSW_MASK_IO | PSW_MASK_EXT); + + /* Set PER control regs, turns on single step for the given address */ + __ctl_load(per_kprobe, 9, 11); + regs->psw.mask |= PSW_MASK_PER; + regs->psw.mask &= ~(PSW_MASK_IO | PSW_MASK_EXT); + regs->psw.addr = ip; +} +NOKPROBE_SYMBOL(enable_singlestep); + +static void disable_singlestep(struct kprobe_ctlblk *kcb, + struct pt_regs *regs, + unsigned long ip) +{ + /* Restore control regs and psw mask, set new psw address */ + __ctl_load(kcb->kprobe_saved_ctl, 9, 11); + regs->psw.mask &= ~PSW_MASK_PER; + regs->psw.mask |= kcb->kprobe_saved_imask; + regs->psw.addr = ip; +} +NOKPROBE_SYMBOL(disable_singlestep); + +/* + * Activate a kprobe by storing its pointer to current_kprobe. The + * previous kprobe is stored in kcb->prev_kprobe. A stack of up to + * two kprobes can be active, see KPROBE_REENTER. + */ +static void push_kprobe(struct kprobe_ctlblk *kcb, struct kprobe *p) +{ + kcb->prev_kprobe.kp = __this_cpu_read(current_kprobe); + kcb->prev_kprobe.status = kcb->kprobe_status; + __this_cpu_write(current_kprobe, p); +} +NOKPROBE_SYMBOL(push_kprobe); + +/* + * Deactivate a kprobe by backing up to the previous state. If the + * current state is KPROBE_REENTER prev_kprobe.kp will be non-NULL, + * for any other state prev_kprobe.kp will be NULL. + */ +static void pop_kprobe(struct kprobe_ctlblk *kcb) +{ + __this_cpu_write(current_kprobe, kcb->prev_kprobe.kp); + kcb->kprobe_status = kcb->prev_kprobe.status; + kcb->prev_kprobe.kp = NULL; +} +NOKPROBE_SYMBOL(pop_kprobe); + +static void kprobe_reenter_check(struct kprobe_ctlblk *kcb, struct kprobe *p) +{ + switch (kcb->kprobe_status) { + case KPROBE_HIT_SSDONE: + case KPROBE_HIT_ACTIVE: + kprobes_inc_nmissed_count(p); + break; + case KPROBE_HIT_SS: + case KPROBE_REENTER: + default: + /* + * A kprobe on the code path to single step an instruction + * is a BUG. The code path resides in the .kprobes.text + * section and is executed with interrupts disabled. + */ + pr_err("Failed to recover from reentered kprobes.\n"); + dump_kprobe(p); + BUG(); + } +} +NOKPROBE_SYMBOL(kprobe_reenter_check); + +static int kprobe_handler(struct pt_regs *regs) +{ + struct kprobe_ctlblk *kcb; + struct kprobe *p; + + /* + * We want to disable preemption for the entire duration of kprobe + * processing. That includes the calls to the pre/post handlers + * and single stepping the kprobe instruction. + */ + preempt_disable(); + kcb = get_kprobe_ctlblk(); + p = get_kprobe((void *)(regs->psw.addr - 2)); + + if (p) { + if (kprobe_running()) { + /* + * We have hit a kprobe while another is still + * active. This can happen in the pre and post + * handler. Single step the instruction of the + * new probe but do not call any handler function + * of this secondary kprobe. + * push_kprobe and pop_kprobe saves and restores + * the currently active kprobe. + */ + kprobe_reenter_check(kcb, p); + push_kprobe(kcb, p); + kcb->kprobe_status = KPROBE_REENTER; + } else { + /* + * If we have no pre-handler or it returned 0, we + * continue with single stepping. If we have a + * pre-handler and it returned non-zero, it prepped + * for changing execution path, so get out doing + * nothing more here. + */ + push_kprobe(kcb, p); + kcb->kprobe_status = KPROBE_HIT_ACTIVE; + if (p->pre_handler && p->pre_handler(p, regs)) { + pop_kprobe(kcb); + preempt_enable_no_resched(); + return 1; + } + kcb->kprobe_status = KPROBE_HIT_SS; + } + enable_singlestep(kcb, regs, (unsigned long) p->ainsn.insn); + return 1; + } /* else: + * No kprobe at this address and no active kprobe. The trap has + * not been caused by a kprobe breakpoint. The race of breakpoint + * vs. kprobe remove does not exist because on s390 as we use + * stop_machine to arm/disarm the breakpoints. + */ + preempt_enable_no_resched(); + return 0; +} +NOKPROBE_SYMBOL(kprobe_handler); + +/* + * Called after single-stepping. p->addr is the address of the + * instruction whose first byte has been replaced by the "breakpoint" + * instruction. To avoid the SMP problems that can occur when we + * temporarily put back the original opcode to single-step, we + * single-stepped a copy of the instruction. The address of this + * copy is p->ainsn.insn. + */ +static void resume_execution(struct kprobe *p, struct pt_regs *regs) +{ + struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + unsigned long ip = regs->psw.addr; + int fixup = probe_get_fixup_type(p->ainsn.insn); + + if (fixup & FIXUP_PSW_NORMAL) + ip += (unsigned long) p->addr - (unsigned long) p->ainsn.insn; + + if (fixup & FIXUP_BRANCH_NOT_TAKEN) { + int ilen = insn_length(p->ainsn.insn[0] >> 8); + if (ip - (unsigned long) p->ainsn.insn == ilen) + ip = (unsigned long) p->addr + ilen; + } + + if (fixup & FIXUP_RETURN_REGISTER) { + int reg = (p->ainsn.insn[0] & 0xf0) >> 4; + regs->gprs[reg] += (unsigned long) p->addr - + (unsigned long) p->ainsn.insn; + } + + disable_singlestep(kcb, regs, ip); +} +NOKPROBE_SYMBOL(resume_execution); + +static int post_kprobe_handler(struct pt_regs *regs) +{ + struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + struct kprobe *p = kprobe_running(); + + if (!p) + return 0; + + resume_execution(p, regs); + if (kcb->kprobe_status != KPROBE_REENTER && p->post_handler) { + kcb->kprobe_status = KPROBE_HIT_SSDONE; + p->post_handler(p, regs, 0); + } + pop_kprobe(kcb); + preempt_enable_no_resched(); + + /* + * if somebody else is singlestepping across a probe point, psw mask + * will have PER set, in which case, continue the remaining processing + * of do_single_step, as if this is not a probe hit. + */ + if (regs->psw.mask & PSW_MASK_PER) + return 0; + + return 1; +} +NOKPROBE_SYMBOL(post_kprobe_handler); + +static int kprobe_trap_handler(struct pt_regs *regs, int trapnr) +{ + struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + struct kprobe *p = kprobe_running(); + + switch(kcb->kprobe_status) { + case KPROBE_HIT_SS: + case KPROBE_REENTER: + /* + * We are here because the instruction being single + * stepped caused a page fault. We reset the current + * kprobe and the nip points back to the probe address + * and allow the page fault handler to continue as a + * normal page fault. + */ + disable_singlestep(kcb, regs, (unsigned long) p->addr); + pop_kprobe(kcb); + preempt_enable_no_resched(); + break; + case KPROBE_HIT_ACTIVE: + case KPROBE_HIT_SSDONE: + /* + * In case the user-specified fault handler returned + * zero, try to fix up. + */ + if (fixup_exception(regs)) + return 1; + /* + * fixup_exception() could not handle it, + * Let do_page_fault() fix it. + */ + break; + default: + break; + } + return 0; +} +NOKPROBE_SYMBOL(kprobe_trap_handler); + +int kprobe_fault_handler(struct pt_regs *regs, int trapnr) +{ + int ret; + + if (regs->psw.mask & (PSW_MASK_IO | PSW_MASK_EXT)) + local_irq_disable(); + ret = kprobe_trap_handler(regs, trapnr); + if (regs->psw.mask & (PSW_MASK_IO | PSW_MASK_EXT)) + local_irq_restore(regs->psw.mask & ~PSW_MASK_PER); + return ret; +} +NOKPROBE_SYMBOL(kprobe_fault_handler); + +/* + * Wrapper routine to for handling exceptions. + */ +int kprobe_exceptions_notify(struct notifier_block *self, + unsigned long val, void *data) +{ + struct die_args *args = (struct die_args *) data; + struct pt_regs *regs = args->regs; + int ret = NOTIFY_DONE; + + if (regs->psw.mask & (PSW_MASK_IO | PSW_MASK_EXT)) + local_irq_disable(); + + switch (val) { + case DIE_BPT: + if (kprobe_handler(regs)) + ret = NOTIFY_STOP; + break; + case DIE_SSTEP: + if (post_kprobe_handler(regs)) + ret = NOTIFY_STOP; + break; + case DIE_TRAP: + if (!preemptible() && kprobe_running() && + kprobe_trap_handler(regs, args->trapnr)) + ret = NOTIFY_STOP; + break; + default: + break; + } + + if (regs->psw.mask & (PSW_MASK_IO | PSW_MASK_EXT)) + local_irq_restore(regs->psw.mask & ~PSW_MASK_PER); + + return ret; +} +NOKPROBE_SYMBOL(kprobe_exceptions_notify); + +int __init arch_init_kprobes(void) +{ + return 0; +} + +int arch_trampoline_kprobe(struct kprobe *p) +{ + return 0; +} +NOKPROBE_SYMBOL(arch_trampoline_kprobe); diff --git a/arch/s390/kernel/kprobes.h b/arch/s390/kernel/kprobes.h new file mode 100644 index 0000000000..dc3ed5098e --- /dev/null +++ b/arch/s390/kernel/kprobes.h @@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +#ifndef _ARCH_S390_KPROBES_H +#define _ARCH_S390_KPROBES_H + +#include <linux/kprobes.h> + +DEFINE_INSN_CACHE_OPS(s390_insn); + +#endif diff --git a/arch/s390/kernel/kprobes_insn_page.S b/arch/s390/kernel/kprobes_insn_page.S new file mode 100644 index 0000000000..0fe4d725e9 --- /dev/null +++ b/arch/s390/kernel/kprobes_insn_page.S @@ -0,0 +1,22 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#include <linux/linkage.h> + +/* + * insn_page is a special 4k aligned dummy function for kprobes. + * It will contain all kprobed instructions that are out-of-line executed. + * The page must be within the kernel image to guarantee that the + * out-of-line instructions are within 2GB distance of their original + * location. Using a dummy function ensures that the insn_page is within + * the text section of the kernel and mapped read-only/executable from + * the beginning on, thus avoiding to split large mappings if the page + * would be in the data section instead. + */ + .section .kprobes.text, "ax" + .balign 4096 +SYM_CODE_START(kprobes_insn_page) + .rept 2048 + .word 0x07fe + .endr +SYM_CODE_END(kprobes_insn_page) + .previous diff --git a/arch/s390/kernel/lgr.c b/arch/s390/kernel/lgr.c new file mode 100644 index 0000000000..6652e54cf3 --- /dev/null +++ b/arch/s390/kernel/lgr.c @@ -0,0 +1,187 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Linux Guest Relocation (LGR) detection + * + * Copyright IBM Corp. 2012 + * Author(s): Michael Holzheu <holzheu@linux.vnet.ibm.com> + */ + +#include <linux/init.h> +#include <linux/export.h> +#include <linux/timer.h> +#include <linux/slab.h> +#include <asm/facility.h> +#include <asm/sysinfo.h> +#include <asm/ebcdic.h> +#include <asm/debug.h> +#include <asm/ipl.h> + +#define LGR_TIMER_INTERVAL_SECS (30 * 60) +#define VM_LEVEL_MAX 2 /* Maximum is 8, but we only record two levels */ + +/* + * LGR info: Contains stfle and stsi data + */ +struct lgr_info { + /* Bit field with facility information: 4 DWORDs are stored */ + u64 stfle_fac_list[4]; + /* Level of system (1 = CEC, 2 = LPAR, 3 = z/VM */ + u32 level; + /* Level 1: CEC info (stsi 1.1.1) */ + char manufacturer[16]; + char type[4]; + char sequence[16]; + char plant[4]; + char model[16]; + /* Level 2: LPAR info (stsi 2.2.2) */ + u16 lpar_number; + char name[8]; + /* Level 3: VM info (stsi 3.2.2) */ + u8 vm_count; + struct { + char name[8]; + char cpi[16]; + } vm[VM_LEVEL_MAX]; +} __packed __aligned(8); + +/* + * LGR globals + */ +static char lgr_page[PAGE_SIZE] __aligned(PAGE_SIZE); +static struct lgr_info lgr_info_last; +static struct lgr_info lgr_info_cur; +static struct debug_info *lgr_dbf; + +/* + * Copy buffer and then convert it to ASCII + */ +static void cpascii(char *dst, char *src, int size) +{ + memcpy(dst, src, size); + EBCASC(dst, size); +} + +/* + * Fill LGR info with 1.1.1 stsi data + */ +static void lgr_stsi_1_1_1(struct lgr_info *lgr_info) +{ + struct sysinfo_1_1_1 *si = (void *) lgr_page; + + if (stsi(si, 1, 1, 1)) + return; + cpascii(lgr_info->manufacturer, si->manufacturer, + sizeof(si->manufacturer)); + cpascii(lgr_info->type, si->type, sizeof(si->type)); + cpascii(lgr_info->model, si->model, sizeof(si->model)); + cpascii(lgr_info->sequence, si->sequence, sizeof(si->sequence)); + cpascii(lgr_info->plant, si->plant, sizeof(si->plant)); +} + +/* + * Fill LGR info with 2.2.2 stsi data + */ +static void lgr_stsi_2_2_2(struct lgr_info *lgr_info) +{ + struct sysinfo_2_2_2 *si = (void *) lgr_page; + + if (stsi(si, 2, 2, 2)) + return; + cpascii(lgr_info->name, si->name, sizeof(si->name)); + lgr_info->lpar_number = si->lpar_number; +} + +/* + * Fill LGR info with 3.2.2 stsi data + */ +static void lgr_stsi_3_2_2(struct lgr_info *lgr_info) +{ + struct sysinfo_3_2_2 *si = (void *) lgr_page; + int i; + + if (stsi(si, 3, 2, 2)) + return; + for (i = 0; i < min_t(u8, si->count, VM_LEVEL_MAX); i++) { + cpascii(lgr_info->vm[i].name, si->vm[i].name, + sizeof(si->vm[i].name)); + cpascii(lgr_info->vm[i].cpi, si->vm[i].cpi, + sizeof(si->vm[i].cpi)); + } + lgr_info->vm_count = si->count; +} + +/* + * Fill LGR info with current data + */ +static void lgr_info_get(struct lgr_info *lgr_info) +{ + int level; + + memset(lgr_info, 0, sizeof(*lgr_info)); + stfle(lgr_info->stfle_fac_list, ARRAY_SIZE(lgr_info->stfle_fac_list)); + level = stsi(NULL, 0, 0, 0); + lgr_info->level = level; + if (level >= 1) + lgr_stsi_1_1_1(lgr_info); + if (level >= 2) + lgr_stsi_2_2_2(lgr_info); + if (level >= 3) + lgr_stsi_3_2_2(lgr_info); +} + +/* + * Check if LGR info has changed and if yes log new LGR info to s390dbf + */ +void lgr_info_log(void) +{ + static DEFINE_SPINLOCK(lgr_info_lock); + unsigned long flags; + + if (!spin_trylock_irqsave(&lgr_info_lock, flags)) + return; + lgr_info_get(&lgr_info_cur); + if (memcmp(&lgr_info_last, &lgr_info_cur, sizeof(lgr_info_cur)) != 0) { + debug_event(lgr_dbf, 1, &lgr_info_cur, sizeof(lgr_info_cur)); + lgr_info_last = lgr_info_cur; + } + spin_unlock_irqrestore(&lgr_info_lock, flags); +} +EXPORT_SYMBOL_GPL(lgr_info_log); + +static void lgr_timer_set(void); + +/* + * LGR timer callback + */ +static void lgr_timer_fn(struct timer_list *unused) +{ + lgr_info_log(); + lgr_timer_set(); +} + +static struct timer_list lgr_timer; + +/* + * Setup next LGR timer + */ +static void lgr_timer_set(void) +{ + mod_timer(&lgr_timer, jiffies + msecs_to_jiffies(LGR_TIMER_INTERVAL_SECS * MSEC_PER_SEC)); +} + +/* + * Initialize LGR: Add s390dbf, write initial lgr_info and setup timer + */ +static int __init lgr_init(void) +{ + lgr_dbf = debug_register("lgr", 1, 1, sizeof(struct lgr_info)); + if (!lgr_dbf) + return -ENOMEM; + debug_register_view(lgr_dbf, &debug_hex_ascii_view); + lgr_info_get(&lgr_info_last); + debug_event(lgr_dbf, 1, &lgr_info_last, sizeof(lgr_info_last)); + timer_setup(&lgr_timer, lgr_timer_fn, TIMER_DEFERRABLE); + lgr_timer_set(); + return 0; +} +device_initcall(lgr_init); diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c new file mode 100644 index 0000000000..ce65fc0167 --- /dev/null +++ b/arch/s390/kernel/machine_kexec.c @@ -0,0 +1,286 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2005, 2011 + * + * Author(s): Rolf Adelsberger, + * Michael Holzheu <holzheu@linux.vnet.ibm.com> + */ + +#include <linux/device.h> +#include <linux/mm.h> +#include <linux/kexec.h> +#include <linux/delay.h> +#include <linux/reboot.h> +#include <linux/ftrace.h> +#include <linux/debug_locks.h> +#include <asm/pfault.h> +#include <asm/cio.h> +#include <asm/setup.h> +#include <asm/smp.h> +#include <asm/ipl.h> +#include <asm/diag.h> +#include <asm/elf.h> +#include <asm/asm-offsets.h> +#include <asm/cacheflush.h> +#include <asm/abs_lowcore.h> +#include <asm/os_info.h> +#include <asm/set_memory.h> +#include <asm/stacktrace.h> +#include <asm/switch_to.h> +#include <asm/nmi.h> +#include <asm/sclp.h> + +typedef void (*relocate_kernel_t)(unsigned long, unsigned long, unsigned long); +typedef int (*purgatory_t)(int); + +extern const unsigned char relocate_kernel[]; +extern const unsigned long long relocate_kernel_len; + +#ifdef CONFIG_CRASH_DUMP + +/* + * Reset the system, copy boot CPU registers to absolute zero, + * and jump to the kdump image + */ +static void __do_machine_kdump(void *data) +{ + struct kimage *image = data; + purgatory_t purgatory; + unsigned long prefix; + + purgatory = (purgatory_t)image->start; + + /* store_status() saved the prefix register to lowcore */ + prefix = (unsigned long) S390_lowcore.prefixreg_save_area; + + /* Now do the reset */ + s390_reset_system(); + + /* + * Copy dump CPU store status info to absolute zero. + * This need to be done *after* s390_reset_system set the + * prefix register of this CPU to zero + */ + memcpy(absolute_pointer(__LC_FPREGS_SAVE_AREA), + phys_to_virt(prefix + __LC_FPREGS_SAVE_AREA), 512); + + call_nodat(1, int, purgatory, int, 1); + + /* Die if kdump returns */ + disabled_wait(); +} + +/* + * Start kdump: create a LGR log entry, store status of all CPUs and + * branch to __do_machine_kdump. + */ +static noinline void __machine_kdump(void *image) +{ + struct mcesa *mcesa; + union ctlreg2 cr2_old, cr2_new; + int this_cpu, cpu; + + lgr_info_log(); + /* Get status of the other CPUs */ + this_cpu = smp_find_processor_id(stap()); + for_each_online_cpu(cpu) { + if (cpu == this_cpu) + continue; + if (smp_store_status(cpu)) + continue; + } + /* Store status of the boot CPU */ + mcesa = __va(S390_lowcore.mcesad & MCESA_ORIGIN_MASK); + if (MACHINE_HAS_VX) + save_vx_regs((__vector128 *) mcesa->vector_save_area); + if (MACHINE_HAS_GS) { + __ctl_store(cr2_old.val, 2, 2); + cr2_new = cr2_old; + cr2_new.gse = 1; + __ctl_load(cr2_new.val, 2, 2); + save_gs_cb((struct gs_cb *) mcesa->guarded_storage_save_area); + __ctl_load(cr2_old.val, 2, 2); + } + /* + * To create a good backchain for this CPU in the dump store_status + * is passed the address of a function. The address is saved into + * the PSW save area of the boot CPU and the function is invoked as + * a tail call of store_status. The backchain in the dump will look + * like this: + * restart_int_handler -> __machine_kexec -> __do_machine_kdump + * The call to store_status() will not return. + */ + store_status(__do_machine_kdump, image); +} + +#endif /* CONFIG_CRASH_DUMP */ + +/* + * Check if kdump checksums are valid: We call purgatory with parameter "0" + */ +static bool kdump_csum_valid(struct kimage *image) +{ +#ifdef CONFIG_CRASH_DUMP + purgatory_t purgatory = (purgatory_t)image->start; + int rc; + + rc = call_nodat(1, int, purgatory, int, 0); + return rc == 0; +#else + return false; +#endif +} + +#ifdef CONFIG_CRASH_DUMP + +void crash_free_reserved_phys_range(unsigned long begin, unsigned long end) +{ + unsigned long addr, size; + + for (addr = begin; addr < end; addr += PAGE_SIZE) + free_reserved_page(pfn_to_page(addr >> PAGE_SHIFT)); + size = begin - crashk_res.start; + if (size) + os_info_crashkernel_add(crashk_res.start, size); + else + os_info_crashkernel_add(0, 0); +} + +static void crash_protect_pages(int protect) +{ + unsigned long size; + + if (!crashk_res.end) + return; + size = resource_size(&crashk_res); + if (protect) + set_memory_ro(crashk_res.start, size >> PAGE_SHIFT); + else + set_memory_rw(crashk_res.start, size >> PAGE_SHIFT); +} + +void arch_kexec_protect_crashkres(void) +{ + crash_protect_pages(1); +} + +void arch_kexec_unprotect_crashkres(void) +{ + crash_protect_pages(0); +} + +#endif + +/* + * Give back memory to hypervisor before new kdump is loaded + */ +static int machine_kexec_prepare_kdump(void) +{ +#ifdef CONFIG_CRASH_DUMP + if (MACHINE_IS_VM) + diag10_range(PFN_DOWN(crashk_res.start), + PFN_DOWN(crashk_res.end - crashk_res.start + 1)); + return 0; +#else + return -EINVAL; +#endif +} + +int machine_kexec_prepare(struct kimage *image) +{ + void *reboot_code_buffer; + + if (image->type == KEXEC_TYPE_CRASH) + return machine_kexec_prepare_kdump(); + + /* We don't support anything but the default image type for now. */ + if (image->type != KEXEC_TYPE_DEFAULT) + return -EINVAL; + + /* Get the destination where the assembler code should be copied to.*/ + reboot_code_buffer = page_to_virt(image->control_code_page); + + /* Then copy it */ + memcpy(reboot_code_buffer, relocate_kernel, relocate_kernel_len); + return 0; +} + +void machine_kexec_cleanup(struct kimage *image) +{ +} + +void arch_crash_save_vmcoreinfo(void) +{ + struct lowcore *abs_lc; + + VMCOREINFO_SYMBOL(lowcore_ptr); + VMCOREINFO_SYMBOL(high_memory); + VMCOREINFO_LENGTH(lowcore_ptr, NR_CPUS); + vmcoreinfo_append_str("SAMODE31=%lx\n", (unsigned long)__samode31); + vmcoreinfo_append_str("EAMODE31=%lx\n", (unsigned long)__eamode31); + vmcoreinfo_append_str("KERNELOFFSET=%lx\n", kaslr_offset()); + abs_lc = get_abs_lowcore(); + abs_lc->vmcore_info = paddr_vmcoreinfo_note(); + put_abs_lowcore(abs_lc); +} + +void machine_shutdown(void) +{ +} + +void machine_crash_shutdown(struct pt_regs *regs) +{ + set_os_info_reipl_block(); +} + +/* + * Do normal kexec + */ +static void __do_machine_kexec(void *data) +{ + unsigned long data_mover, entry, diag308_subcode; + struct kimage *image = data; + + data_mover = page_to_phys(image->control_code_page); + entry = virt_to_phys(&image->head); + diag308_subcode = DIAG308_CLEAR_RESET; + if (sclp.has_iplcc) + diag308_subcode |= DIAG308_FLAG_EI; + s390_reset_system(); + + call_nodat(3, void, (relocate_kernel_t)data_mover, + unsigned long, entry, + unsigned long, image->start, + unsigned long, diag308_subcode); + + /* Die if kexec returns */ + disabled_wait(); +} + +/* + * Reset system and call either kdump or normal kexec + */ +static void __machine_kexec(void *data) +{ + pfault_fini(); + tracing_off(); + debug_locks_off(); +#ifdef CONFIG_CRASH_DUMP + if (((struct kimage *) data)->type == KEXEC_TYPE_CRASH) + __machine_kdump(data); +#endif + __do_machine_kexec(data); +} + +/* + * Do either kdump or normal kexec. In case of kdump we first ask + * purgatory, if kdump checksums are valid. + */ +void machine_kexec(struct kimage *image) +{ + if (image->type == KEXEC_TYPE_CRASH && !kdump_csum_valid(image)) + return; + tracer_disable(); + smp_send_stop(); + smp_call_ipl_cpu(__machine_kexec, image); +} diff --git a/arch/s390/kernel/machine_kexec_file.c b/arch/s390/kernel/machine_kexec_file.c new file mode 100644 index 0000000000..8d207b82d9 --- /dev/null +++ b/arch/s390/kernel/machine_kexec_file.c @@ -0,0 +1,379 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * s390 code for kexec_file_load system call + * + * Copyright IBM Corp. 2018 + * + * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com> + */ + +#define pr_fmt(fmt) "kexec: " fmt + +#include <linux/elf.h> +#include <linux/errno.h> +#include <linux/kexec.h> +#include <linux/module_signature.h> +#include <linux/verification.h> +#include <linux/vmalloc.h> +#include <asm/boot_data.h> +#include <asm/ipl.h> +#include <asm/setup.h> + +const struct kexec_file_ops * const kexec_file_loaders[] = { + &s390_kexec_elf_ops, + &s390_kexec_image_ops, + NULL, +}; + +#ifdef CONFIG_KEXEC_SIG +int s390_verify_sig(const char *kernel, unsigned long kernel_len) +{ + const unsigned long marker_len = sizeof(MODULE_SIG_STRING) - 1; + struct module_signature *ms; + unsigned long sig_len; + int ret; + + /* Skip signature verification when not secure IPLed. */ + if (!ipl_secure_flag) + return 0; + + if (marker_len > kernel_len) + return -EKEYREJECTED; + + if (memcmp(kernel + kernel_len - marker_len, MODULE_SIG_STRING, + marker_len)) + return -EKEYREJECTED; + kernel_len -= marker_len; + + ms = (void *)kernel + kernel_len - sizeof(*ms); + kernel_len -= sizeof(*ms); + + sig_len = be32_to_cpu(ms->sig_len); + if (sig_len >= kernel_len) + return -EKEYREJECTED; + kernel_len -= sig_len; + + if (ms->id_type != PKEY_ID_PKCS7) + return -EKEYREJECTED; + + if (ms->algo != 0 || + ms->hash != 0 || + ms->signer_len != 0 || + ms->key_id_len != 0 || + ms->__pad[0] != 0 || + ms->__pad[1] != 0 || + ms->__pad[2] != 0) { + return -EBADMSG; + } + + ret = verify_pkcs7_signature(kernel, kernel_len, + kernel + kernel_len, sig_len, + VERIFY_USE_SECONDARY_KEYRING, + VERIFYING_MODULE_SIGNATURE, + NULL, NULL); + if (ret == -ENOKEY && IS_ENABLED(CONFIG_INTEGRITY_PLATFORM_KEYRING)) + ret = verify_pkcs7_signature(kernel, kernel_len, + kernel + kernel_len, sig_len, + VERIFY_USE_PLATFORM_KEYRING, + VERIFYING_MODULE_SIGNATURE, + NULL, NULL); + return ret; +} +#endif /* CONFIG_KEXEC_SIG */ + +static int kexec_file_update_purgatory(struct kimage *image, + struct s390_load_data *data) +{ + u64 entry, type; + int ret; + + if (image->type == KEXEC_TYPE_CRASH) { + entry = STARTUP_KDUMP_OFFSET; + type = KEXEC_TYPE_CRASH; + } else { + entry = STARTUP_NORMAL_OFFSET; + type = KEXEC_TYPE_DEFAULT; + } + + ret = kexec_purgatory_get_set_symbol(image, "kernel_entry", &entry, + sizeof(entry), false); + if (ret) + return ret; + + ret = kexec_purgatory_get_set_symbol(image, "kernel_type", &type, + sizeof(type), false); + if (ret) + return ret; + + if (image->type == KEXEC_TYPE_CRASH) { + u64 crash_size; + + ret = kexec_purgatory_get_set_symbol(image, "crash_start", + &crashk_res.start, + sizeof(crashk_res.start), + false); + if (ret) + return ret; + + crash_size = crashk_res.end - crashk_res.start + 1; + ret = kexec_purgatory_get_set_symbol(image, "crash_size", + &crash_size, + sizeof(crash_size), + false); + } + return ret; +} + +static int kexec_file_add_purgatory(struct kimage *image, + struct s390_load_data *data) +{ + struct kexec_buf buf; + int ret; + + buf.image = image; + + data->memsz = ALIGN(data->memsz, PAGE_SIZE); + buf.mem = data->memsz; + if (image->type == KEXEC_TYPE_CRASH) + buf.mem += crashk_res.start; + + ret = kexec_load_purgatory(image, &buf); + if (ret) + return ret; + data->memsz += buf.memsz; + + return kexec_file_update_purgatory(image, data); +} + +static int kexec_file_add_initrd(struct kimage *image, + struct s390_load_data *data) +{ + struct kexec_buf buf; + int ret; + + buf.image = image; + + buf.buffer = image->initrd_buf; + buf.bufsz = image->initrd_buf_len; + + data->memsz = ALIGN(data->memsz, PAGE_SIZE); + buf.mem = data->memsz; + if (image->type == KEXEC_TYPE_CRASH) + buf.mem += crashk_res.start; + buf.memsz = buf.bufsz; + + data->parm->initrd_start = data->memsz; + data->parm->initrd_size = buf.memsz; + data->memsz += buf.memsz; + + ret = kexec_add_buffer(&buf); + if (ret) + return ret; + + return ipl_report_add_component(data->report, &buf, 0, 0); +} + +static int kexec_file_add_ipl_report(struct kimage *image, + struct s390_load_data *data) +{ + __u32 *lc_ipl_parmblock_ptr; + unsigned int len, ncerts; + struct kexec_buf buf; + unsigned long addr; + void *ptr, *end; + int ret; + + buf.image = image; + + data->memsz = ALIGN(data->memsz, PAGE_SIZE); + buf.mem = data->memsz; + + ptr = __va(ipl_cert_list_addr); + end = ptr + ipl_cert_list_size; + ncerts = 0; + while (ptr < end) { + ncerts++; + len = *(unsigned int *)ptr; + ptr += sizeof(len); + ptr += len; + } + + addr = data->memsz + data->report->size; + addr += ncerts * sizeof(struct ipl_rb_certificate_entry); + ptr = __va(ipl_cert_list_addr); + while (ptr < end) { + len = *(unsigned int *)ptr; + ptr += sizeof(len); + ipl_report_add_certificate(data->report, ptr, addr, len); + addr += len; + ptr += len; + } + + ret = -ENOMEM; + buf.buffer = ipl_report_finish(data->report); + if (!buf.buffer) + goto out; + buf.bufsz = data->report->size; + buf.memsz = buf.bufsz; + image->arch.ipl_buf = buf.buffer; + + data->memsz += buf.memsz; + + lc_ipl_parmblock_ptr = + data->kernel_buf + offsetof(struct lowcore, ipl_parmblock_ptr); + *lc_ipl_parmblock_ptr = (__u32)buf.mem; + + if (image->type == KEXEC_TYPE_CRASH) + buf.mem += crashk_res.start; + + ret = kexec_add_buffer(&buf); +out: + return ret; +} + +void *kexec_file_add_components(struct kimage *image, + int (*add_kernel)(struct kimage *image, + struct s390_load_data *data)) +{ + unsigned long max_command_line_size = LEGACY_COMMAND_LINE_SIZE; + struct s390_load_data data = {0}; + unsigned long minsize; + int ret; + + data.report = ipl_report_init(&ipl_block); + if (IS_ERR(data.report)) + return data.report; + + ret = add_kernel(image, &data); + if (ret) + goto out; + + ret = -EINVAL; + minsize = PARMAREA + offsetof(struct parmarea, command_line); + if (image->kernel_buf_len < minsize) + goto out; + + if (data.parm->max_command_line_size) + max_command_line_size = data.parm->max_command_line_size; + + if (minsize + max_command_line_size < minsize) + goto out; + + if (image->kernel_buf_len < minsize + max_command_line_size) + goto out; + + if (image->cmdline_buf_len >= max_command_line_size) + goto out; + + memcpy(data.parm->command_line, image->cmdline_buf, + image->cmdline_buf_len); + + if (image->type == KEXEC_TYPE_CRASH) { + data.parm->oldmem_base = crashk_res.start; + data.parm->oldmem_size = crashk_res.end - crashk_res.start + 1; + } + + if (image->initrd_buf) { + ret = kexec_file_add_initrd(image, &data); + if (ret) + goto out; + } + + ret = kexec_file_add_purgatory(image, &data); + if (ret) + goto out; + + if (data.kernel_mem == 0) { + unsigned long restart_psw = 0x0008000080000000UL; + restart_psw += image->start; + memcpy(data.kernel_buf, &restart_psw, sizeof(restart_psw)); + image->start = 0; + } + + ret = kexec_file_add_ipl_report(image, &data); +out: + ipl_report_free(data.report); + return ERR_PTR(ret); +} + +int arch_kexec_apply_relocations_add(struct purgatory_info *pi, + Elf_Shdr *section, + const Elf_Shdr *relsec, + const Elf_Shdr *symtab) +{ + const char *strtab, *name, *shstrtab; + const Elf_Shdr *sechdrs; + Elf_Rela *relas; + int i, r_type; + int ret; + + /* String & section header string table */ + sechdrs = (void *)pi->ehdr + pi->ehdr->e_shoff; + strtab = (char *)pi->ehdr + sechdrs[symtab->sh_link].sh_offset; + shstrtab = (char *)pi->ehdr + sechdrs[pi->ehdr->e_shstrndx].sh_offset; + + relas = (void *)pi->ehdr + relsec->sh_offset; + + for (i = 0; i < relsec->sh_size / sizeof(*relas); i++) { + const Elf_Sym *sym; /* symbol to relocate */ + unsigned long addr; /* final location after relocation */ + unsigned long val; /* relocated symbol value */ + void *loc; /* tmp location to modify */ + + sym = (void *)pi->ehdr + symtab->sh_offset; + sym += ELF64_R_SYM(relas[i].r_info); + + if (sym->st_name) + name = strtab + sym->st_name; + else + name = shstrtab + sechdrs[sym->st_shndx].sh_name; + + if (sym->st_shndx == SHN_UNDEF) { + pr_err("Undefined symbol: %s\n", name); + return -ENOEXEC; + } + + if (sym->st_shndx == SHN_COMMON) { + pr_err("symbol '%s' in common section\n", name); + return -ENOEXEC; + } + + if (sym->st_shndx >= pi->ehdr->e_shnum && + sym->st_shndx != SHN_ABS) { + pr_err("Invalid section %d for symbol %s\n", + sym->st_shndx, name); + return -ENOEXEC; + } + + loc = pi->purgatory_buf; + loc += section->sh_offset; + loc += relas[i].r_offset; + + val = sym->st_value; + if (sym->st_shndx != SHN_ABS) + val += pi->sechdrs[sym->st_shndx].sh_addr; + val += relas[i].r_addend; + + addr = section->sh_addr + relas[i].r_offset; + + r_type = ELF64_R_TYPE(relas[i].r_info); + + if (r_type == R_390_PLT32DBL) + r_type = R_390_PC32DBL; + + ret = arch_kexec_do_relocs(r_type, loc, val, addr); + if (ret) { + pr_err("Unknown rela relocation: %d\n", r_type); + return -ENOEXEC; + } + } + return 0; +} + +int arch_kimage_file_post_load_cleanup(struct kimage *image) +{ + vfree(image->arch.ipl_buf); + image->arch.ipl_buf = NULL; + + return kexec_image_post_load_cleanup_default(image); +} diff --git a/arch/s390/kernel/machine_kexec_reloc.c b/arch/s390/kernel/machine_kexec_reloc.c new file mode 100644 index 0000000000..b7182cec48 --- /dev/null +++ b/arch/s390/kernel/machine_kexec_reloc.c @@ -0,0 +1,56 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/elf.h> +#include <asm/kexec.h> + +int arch_kexec_do_relocs(int r_type, void *loc, unsigned long val, + unsigned long addr) +{ + switch (r_type) { + case R_390_NONE: + break; + case R_390_8: /* Direct 8 bit. */ + *(u8 *)loc = val; + break; + case R_390_12: /* Direct 12 bit. */ + *(u16 *)loc &= 0xf000; + *(u16 *)loc |= val & 0xfff; + break; + case R_390_16: /* Direct 16 bit. */ + *(u16 *)loc = val; + break; + case R_390_20: /* Direct 20 bit. */ + *(u32 *)loc &= 0xf00000ff; + *(u32 *)loc |= (val & 0xfff) << 16; /* DL */ + *(u32 *)loc |= (val & 0xff000) >> 4; /* DH */ + break; + case R_390_32: /* Direct 32 bit. */ + *(u32 *)loc = val; + break; + case R_390_64: /* Direct 64 bit. */ + case R_390_GLOB_DAT: + case R_390_JMP_SLOT: + *(u64 *)loc = val; + break; + case R_390_PC16: /* PC relative 16 bit. */ + *(u16 *)loc = (val - addr); + break; + case R_390_PC16DBL: /* PC relative 16 bit shifted by 1. */ + *(u16 *)loc = (val - addr) >> 1; + break; + case R_390_PC32DBL: /* PC relative 32 bit shifted by 1. */ + *(u32 *)loc = (val - addr) >> 1; + break; + case R_390_PC32: /* PC relative 32 bit. */ + *(u32 *)loc = (val - addr); + break; + case R_390_PC64: /* PC relative 64 bit. */ + *(u64 *)loc = (val - addr); + break; + case R_390_RELATIVE: + *(unsigned long *) loc = val; + break; + default: + return 1; + } + return 0; +} diff --git a/arch/s390/kernel/mcount.S b/arch/s390/kernel/mcount.S new file mode 100644 index 0000000000..ae4d4fd9af --- /dev/null +++ b/arch/s390/kernel/mcount.S @@ -0,0 +1,195 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright IBM Corp. 2008, 2009 + * + */ + +#include <linux/linkage.h> +#include <asm/asm-offsets.h> +#include <asm/ftrace.h> +#include <asm/nospec-insn.h> +#include <asm/ptrace.h> + +#define STACK_FRAME_SIZE_PTREGS (STACK_FRAME_OVERHEAD + __PT_SIZE) +#define STACK_PTREGS (STACK_FRAME_OVERHEAD) +#define STACK_PTREGS_GPRS (STACK_PTREGS + __PT_GPRS) +#define STACK_PTREGS_PSW (STACK_PTREGS + __PT_PSW) + +#define STACK_FRAME_SIZE_FREGS (STACK_FRAME_OVERHEAD + __FTRACE_REGS_SIZE) +#define STACK_FREGS (STACK_FRAME_OVERHEAD) +#define STACK_FREGS_PTREGS (STACK_FRAME_OVERHEAD + __FTRACE_REGS_PT_REGS) +#define STACK_FREGS_PTREGS_GPRS (STACK_FREGS_PTREGS + __PT_GPRS) +#define STACK_FREGS_PTREGS_PSW (STACK_FREGS_PTREGS + __PT_PSW) +#define STACK_FREGS_PTREGS_ORIG_GPR2 (STACK_FREGS_PTREGS + __PT_ORIG_GPR2) +#define STACK_FREGS_PTREGS_FLAGS (STACK_FREGS_PTREGS + __PT_FLAGS) + +/* packed stack: allocate just enough for r14, r15 and backchain */ +#define TRACED_FUNC_FRAME_SIZE 24 + +#ifdef CONFIG_FUNCTION_TRACER + + GEN_BR_THUNK %r1 + GEN_BR_THUNK %r14 + + .section .kprobes.text, "ax" + +SYM_FUNC_START(ftrace_stub) + BR_EX %r14 +SYM_FUNC_END(ftrace_stub) + +SYM_CODE_START(ftrace_stub_direct_tramp) + lgr %r1, %r0 + BR_EX %r1 +SYM_CODE_END(ftrace_stub_direct_tramp) + + .macro ftrace_regs_entry, allregs=0 + stg %r14,(__SF_GPRS+8*8)(%r15) # save traced function caller + + .if \allregs == 1 + # save psw mask + # don't put any instructions clobbering CC before this point + epsw %r1,%r14 + risbg %r14,%r1,0,31,32 + .endif + + lgr %r1,%r15 + # allocate stack frame for ftrace_caller to contain traced function + aghi %r15,-TRACED_FUNC_FRAME_SIZE + stg %r1,__SF_BACKCHAIN(%r15) + stg %r0,(__SF_GPRS+8*8)(%r15) + stg %r15,(__SF_GPRS+9*8)(%r15) + # allocate ftrace_regs and stack frame for ftrace_trace_function + aghi %r15,-STACK_FRAME_SIZE_FREGS + stg %r1,(STACK_FREGS_PTREGS_GPRS+15*8)(%r15) + xc STACK_FREGS_PTREGS_ORIG_GPR2(8,%r15),STACK_FREGS_PTREGS_ORIG_GPR2(%r15) + + .if \allregs == 1 + stg %r14,(STACK_FREGS_PTREGS_PSW)(%r15) + mvghi STACK_FREGS_PTREGS_FLAGS(%r15),_PIF_FTRACE_FULL_REGS + .else + xc STACK_FREGS_PTREGS_FLAGS(8,%r15),STACK_FREGS_PTREGS_FLAGS(%r15) + .endif + + lg %r14,(__SF_GPRS+8*8)(%r1) # restore original return address + aghi %r1,-TRACED_FUNC_FRAME_SIZE + stg %r1,__SF_BACKCHAIN(%r15) + stg %r0,(STACK_FREGS_PTREGS_PSW+8)(%r15) + stmg %r2,%r14,(STACK_FREGS_PTREGS_GPRS+2*8)(%r15) + .endm + +SYM_CODE_START(ftrace_regs_caller) + ftrace_regs_entry 1 + j ftrace_common +SYM_CODE_END(ftrace_regs_caller) + +SYM_CODE_START(ftrace_caller) + ftrace_regs_entry 0 + j ftrace_common +SYM_CODE_END(ftrace_caller) + +SYM_CODE_START(ftrace_common) +#ifdef CONFIG_HAVE_MARCH_Z196_FEATURES + aghik %r2,%r0,-MCOUNT_INSN_SIZE + lgrl %r4,function_trace_op + lgrl %r1,ftrace_func +#else + lgr %r2,%r0 + aghi %r2,-MCOUNT_INSN_SIZE + larl %r4,function_trace_op + lg %r4,0(%r4) + larl %r1,ftrace_func + lg %r1,0(%r1) +#endif + lgr %r3,%r14 + la %r5,STACK_FREGS(%r15) + BASR_EX %r14,%r1 +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +# The j instruction gets runtime patched to a nop instruction. +# See ftrace_enable_ftrace_graph_caller. +SYM_INNER_LABEL(ftrace_graph_caller, SYM_L_GLOBAL) + j .Lftrace_graph_caller_end + lmg %r2,%r3,(STACK_FREGS_PTREGS_GPRS+14*8)(%r15) + lg %r4,(STACK_FREGS_PTREGS_PSW+8)(%r15) + brasl %r14,prepare_ftrace_return + stg %r2,(STACK_FREGS_PTREGS_GPRS+14*8)(%r15) +.Lftrace_graph_caller_end: +#endif + lg %r0,(STACK_FREGS_PTREGS_PSW+8)(%r15) +#ifdef CONFIG_HAVE_MARCH_Z196_FEATURES + ltg %r1,STACK_FREGS_PTREGS_ORIG_GPR2(%r15) + locgrz %r1,%r0 +#else + lg %r1,STACK_FREGS_PTREGS_ORIG_GPR2(%r15) + ltgr %r1,%r1 + jnz 0f + lgr %r1,%r0 +#endif +0: lmg %r2,%r15,(STACK_FREGS_PTREGS_GPRS+2*8)(%r15) + BR_EX %r1 +SYM_CODE_END(ftrace_common) + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + +SYM_FUNC_START(return_to_handler) + stmg %r2,%r5,32(%r15) + lgr %r1,%r15 + aghi %r15,-(STACK_FRAME_OVERHEAD+__FGRAPH_RET_SIZE) + stg %r1,__SF_BACKCHAIN(%r15) + la %r3,STACK_FRAME_OVERHEAD(%r15) + stg %r1,__FGRAPH_RET_FP(%r3) + stg %r2,__FGRAPH_RET_GPR2(%r3) + lgr %r2,%r3 + brasl %r14,ftrace_return_to_handler + aghi %r15,STACK_FRAME_OVERHEAD+__FGRAPH_RET_SIZE + lgr %r14,%r2 + lmg %r2,%r5,32(%r15) + BR_EX %r14 +SYM_FUNC_END(return_to_handler) + +#endif +#endif /* CONFIG_FUNCTION_TRACER */ + +SYM_CODE_START(ftrace_shared_hotpatch_trampoline_br) + lmg %r0,%r1,2(%r1) + br %r1 +SYM_INNER_LABEL(ftrace_shared_hotpatch_trampoline_br_end, SYM_L_GLOBAL) +SYM_CODE_END(ftrace_shared_hotpatch_trampoline_br) + +#ifdef CONFIG_EXPOLINE +SYM_CODE_START(ftrace_shared_hotpatch_trampoline_exrl) + lmg %r0,%r1,2(%r1) + exrl %r0,0f + j . +0: br %r1 +SYM_INNER_LABEL(ftrace_shared_hotpatch_trampoline_exrl_end, SYM_L_GLOBAL) +SYM_CODE_END(ftrace_shared_hotpatch_trampoline_exrl) +#endif /* CONFIG_EXPOLINE */ + +#ifdef CONFIG_RETHOOK + +SYM_CODE_START(arch_rethook_trampoline) + stg %r14,(__SF_GPRS+8*8)(%r15) + lay %r15,-STACK_FRAME_SIZE_PTREGS(%r15) + stmg %r0,%r14,STACK_PTREGS_GPRS(%r15) + + # store original stack pointer in backchain and pt_regs + lay %r7,STACK_FRAME_SIZE_PTREGS(%r15) + stg %r7,__SF_BACKCHAIN(%r15) + stg %r7,STACK_PTREGS_GPRS+(15*8)(%r15) + + # store full psw + epsw %r2,%r3 + risbg %r3,%r2,0,31,32 + stg %r3,STACK_PTREGS_PSW(%r15) + larl %r1,arch_rethook_trampoline + stg %r1,STACK_PTREGS_PSW+8(%r15) + + lay %r2,STACK_PTREGS(%r15) + brasl %r14,arch_rethook_trampoline_callback + + mvc __SF_EMPTY(16,%r7),STACK_PTREGS_PSW(%r15) + lmg %r0,%r15,STACK_PTREGS_GPRS(%r15) + lpswe __SF_EMPTY(%r15) +SYM_CODE_END(arch_rethook_trampoline) + +#endif /* CONFIG_RETHOOK */ diff --git a/arch/s390/kernel/module.c b/arch/s390/kernel/module.c new file mode 100644 index 0000000000..42215f9404 --- /dev/null +++ b/arch/s390/kernel/module.c @@ -0,0 +1,576 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * Kernel module help for s390. + * + * S390 version + * Copyright IBM Corp. 2002, 2003 + * Author(s): Arnd Bergmann (arndb@de.ibm.com) + * Martin Schwidefsky (schwidefsky@de.ibm.com) + * + * based on i386 version + * Copyright (C) 2001 Rusty Russell. + */ +#include <linux/module.h> +#include <linux/elf.h> +#include <linux/vmalloc.h> +#include <linux/fs.h> +#include <linux/ftrace.h> +#include <linux/string.h> +#include <linux/kernel.h> +#include <linux/kasan.h> +#include <linux/moduleloader.h> +#include <linux/bug.h> +#include <linux/memory.h> +#include <asm/alternative.h> +#include <asm/nospec-branch.h> +#include <asm/facility.h> +#include <asm/ftrace.lds.h> +#include <asm/set_memory.h> +#include <asm/setup.h> + +#if 0 +#define DEBUGP printk +#else +#define DEBUGP(fmt , ...) +#endif + +#define PLT_ENTRY_SIZE 22 + +static unsigned long get_module_load_offset(void) +{ + static DEFINE_MUTEX(module_kaslr_mutex); + static unsigned long module_load_offset; + + if (!kaslr_enabled()) + return 0; + /* + * Calculate the module_load_offset the first time this code + * is called. Once calculated it stays the same until reboot. + */ + mutex_lock(&module_kaslr_mutex); + if (!module_load_offset) + module_load_offset = get_random_u32_inclusive(1, 1024) * PAGE_SIZE; + mutex_unlock(&module_kaslr_mutex); + return module_load_offset; +} + +void *module_alloc(unsigned long size) +{ + gfp_t gfp_mask = GFP_KERNEL; + void *p; + + if (PAGE_ALIGN(size) > MODULES_LEN) + return NULL; + p = __vmalloc_node_range(size, MODULE_ALIGN, + MODULES_VADDR + get_module_load_offset(), + MODULES_END, gfp_mask, PAGE_KERNEL, + VM_FLUSH_RESET_PERMS | VM_DEFER_KMEMLEAK, + NUMA_NO_NODE, __builtin_return_address(0)); + if (p && (kasan_alloc_module_shadow(p, size, gfp_mask) < 0)) { + vfree(p); + return NULL; + } + return p; +} + +#ifdef CONFIG_FUNCTION_TRACER +void module_arch_cleanup(struct module *mod) +{ + module_memfree(mod->arch.trampolines_start); +} +#endif + +void module_arch_freeing_init(struct module *mod) +{ + if (is_livepatch_module(mod) && + mod->state == MODULE_STATE_LIVE) + return; + + vfree(mod->arch.syminfo); + mod->arch.syminfo = NULL; +} + +static void check_rela(Elf_Rela *rela, struct module *me) +{ + struct mod_arch_syminfo *info; + + info = me->arch.syminfo + ELF_R_SYM (rela->r_info); + switch (ELF_R_TYPE (rela->r_info)) { + case R_390_GOT12: /* 12 bit GOT offset. */ + case R_390_GOT16: /* 16 bit GOT offset. */ + case R_390_GOT20: /* 20 bit GOT offset. */ + case R_390_GOT32: /* 32 bit GOT offset. */ + case R_390_GOT64: /* 64 bit GOT offset. */ + case R_390_GOTENT: /* 32 bit PC rel. to GOT entry shifted by 1. */ + case R_390_GOTPLT12: /* 12 bit offset to jump slot. */ + case R_390_GOTPLT16: /* 16 bit offset to jump slot. */ + case R_390_GOTPLT20: /* 20 bit offset to jump slot. */ + case R_390_GOTPLT32: /* 32 bit offset to jump slot. */ + case R_390_GOTPLT64: /* 64 bit offset to jump slot. */ + case R_390_GOTPLTENT: /* 32 bit rel. offset to jump slot >> 1. */ + if (info->got_offset == -1UL) { + info->got_offset = me->arch.got_size; + me->arch.got_size += sizeof(void*); + } + break; + case R_390_PLT16DBL: /* 16 bit PC rel. PLT shifted by 1. */ + case R_390_PLT32DBL: /* 32 bit PC rel. PLT shifted by 1. */ + case R_390_PLT32: /* 32 bit PC relative PLT address. */ + case R_390_PLT64: /* 64 bit PC relative PLT address. */ + case R_390_PLTOFF16: /* 16 bit offset from GOT to PLT. */ + case R_390_PLTOFF32: /* 32 bit offset from GOT to PLT. */ + case R_390_PLTOFF64: /* 16 bit offset from GOT to PLT. */ + if (info->plt_offset == -1UL) { + info->plt_offset = me->arch.plt_size; + me->arch.plt_size += PLT_ENTRY_SIZE; + } + break; + case R_390_COPY: + case R_390_GLOB_DAT: + case R_390_JMP_SLOT: + case R_390_RELATIVE: + /* Only needed if we want to support loading of + modules linked with -shared. */ + break; + } +} + +/* + * Account for GOT and PLT relocations. We can't add sections for + * got and plt but we can increase the core module size. + */ +int module_frob_arch_sections(Elf_Ehdr *hdr, Elf_Shdr *sechdrs, + char *secstrings, struct module *me) +{ + Elf_Shdr *symtab; + Elf_Sym *symbols; + Elf_Rela *rela; + char *strings; + int nrela, i, j; + struct module_memory *mod_mem; + + /* Find symbol table and string table. */ + symtab = NULL; + for (i = 0; i < hdr->e_shnum; i++) + switch (sechdrs[i].sh_type) { + case SHT_SYMTAB: + symtab = sechdrs + i; + break; + } + if (!symtab) { + printk(KERN_ERR "module %s: no symbol table\n", me->name); + return -ENOEXEC; + } + + /* Allocate one syminfo structure per symbol. */ + me->arch.nsyms = symtab->sh_size / sizeof(Elf_Sym); + me->arch.syminfo = vmalloc(array_size(sizeof(struct mod_arch_syminfo), + me->arch.nsyms)); + if (!me->arch.syminfo) + return -ENOMEM; + symbols = (void *) hdr + symtab->sh_offset; + strings = (void *) hdr + sechdrs[symtab->sh_link].sh_offset; + for (i = 0; i < me->arch.nsyms; i++) { + if (symbols[i].st_shndx == SHN_UNDEF && + strcmp(strings + symbols[i].st_name, + "_GLOBAL_OFFSET_TABLE_") == 0) + /* "Define" it as absolute. */ + symbols[i].st_shndx = SHN_ABS; + me->arch.syminfo[i].got_offset = -1UL; + me->arch.syminfo[i].plt_offset = -1UL; + me->arch.syminfo[i].got_initialized = 0; + me->arch.syminfo[i].plt_initialized = 0; + } + + /* Search for got/plt relocations. */ + me->arch.got_size = me->arch.plt_size = 0; + for (i = 0; i < hdr->e_shnum; i++) { + if (sechdrs[i].sh_type != SHT_RELA) + continue; + nrela = sechdrs[i].sh_size / sizeof(Elf_Rela); + rela = (void *) hdr + sechdrs[i].sh_offset; + for (j = 0; j < nrela; j++) + check_rela(rela + j, me); + } + + /* Increase core size by size of got & plt and set start + offsets for got and plt. */ + mod_mem = &me->mem[MOD_TEXT]; + mod_mem->size = ALIGN(mod_mem->size, 4); + me->arch.got_offset = mod_mem->size; + mod_mem->size += me->arch.got_size; + me->arch.plt_offset = mod_mem->size; + if (me->arch.plt_size) { + if (IS_ENABLED(CONFIG_EXPOLINE) && !nospec_disable) + me->arch.plt_size += PLT_ENTRY_SIZE; + mod_mem->size += me->arch.plt_size; + } + return 0; +} + +static int apply_rela_bits(Elf_Addr loc, Elf_Addr val, + int sign, int bits, int shift, + void *(*write)(void *dest, const void *src, size_t len)) +{ + unsigned long umax; + long min, max; + void *dest = (void *)loc; + + if (val & ((1UL << shift) - 1)) + return -ENOEXEC; + if (sign) { + val = (Elf_Addr)(((long) val) >> shift); + min = -(1L << (bits - 1)); + max = (1L << (bits - 1)) - 1; + if ((long) val < min || (long) val > max) + return -ENOEXEC; + } else { + val >>= shift; + umax = ((1UL << (bits - 1)) << 1) - 1; + if ((unsigned long) val > umax) + return -ENOEXEC; + } + + if (bits == 8) { + unsigned char tmp = val; + write(dest, &tmp, 1); + } else if (bits == 12) { + unsigned short tmp = (val & 0xfff) | + (*(unsigned short *) loc & 0xf000); + write(dest, &tmp, 2); + } else if (bits == 16) { + unsigned short tmp = val; + write(dest, &tmp, 2); + } else if (bits == 20) { + unsigned int tmp = (val & 0xfff) << 16 | + (val & 0xff000) >> 4 | (*(unsigned int *) loc & 0xf00000ff); + write(dest, &tmp, 4); + } else if (bits == 32) { + unsigned int tmp = val; + write(dest, &tmp, 4); + } else if (bits == 64) { + unsigned long tmp = val; + write(dest, &tmp, 8); + } + return 0; +} + +static int apply_rela(Elf_Rela *rela, Elf_Addr base, Elf_Sym *symtab, + const char *strtab, struct module *me, + void *(*write)(void *dest, const void *src, size_t len)) +{ + struct mod_arch_syminfo *info; + Elf_Addr loc, val; + int r_type, r_sym; + int rc = -ENOEXEC; + + /* This is where to make the change */ + loc = base + rela->r_offset; + /* This is the symbol it is referring to. Note that all + undefined symbols have been resolved. */ + r_sym = ELF_R_SYM(rela->r_info); + r_type = ELF_R_TYPE(rela->r_info); + info = me->arch.syminfo + r_sym; + val = symtab[r_sym].st_value; + + switch (r_type) { + case R_390_NONE: /* No relocation. */ + rc = 0; + break; + case R_390_8: /* Direct 8 bit. */ + case R_390_12: /* Direct 12 bit. */ + case R_390_16: /* Direct 16 bit. */ + case R_390_20: /* Direct 20 bit. */ + case R_390_32: /* Direct 32 bit. */ + case R_390_64: /* Direct 64 bit. */ + val += rela->r_addend; + if (r_type == R_390_8) + rc = apply_rela_bits(loc, val, 0, 8, 0, write); + else if (r_type == R_390_12) + rc = apply_rela_bits(loc, val, 0, 12, 0, write); + else if (r_type == R_390_16) + rc = apply_rela_bits(loc, val, 0, 16, 0, write); + else if (r_type == R_390_20) + rc = apply_rela_bits(loc, val, 1, 20, 0, write); + else if (r_type == R_390_32) + rc = apply_rela_bits(loc, val, 0, 32, 0, write); + else if (r_type == R_390_64) + rc = apply_rela_bits(loc, val, 0, 64, 0, write); + break; + case R_390_PC16: /* PC relative 16 bit. */ + case R_390_PC16DBL: /* PC relative 16 bit shifted by 1. */ + case R_390_PC32DBL: /* PC relative 32 bit shifted by 1. */ + case R_390_PC32: /* PC relative 32 bit. */ + case R_390_PC64: /* PC relative 64 bit. */ + val += rela->r_addend - loc; + if (r_type == R_390_PC16) + rc = apply_rela_bits(loc, val, 1, 16, 0, write); + else if (r_type == R_390_PC16DBL) + rc = apply_rela_bits(loc, val, 1, 16, 1, write); + else if (r_type == R_390_PC32DBL) + rc = apply_rela_bits(loc, val, 1, 32, 1, write); + else if (r_type == R_390_PC32) + rc = apply_rela_bits(loc, val, 1, 32, 0, write); + else if (r_type == R_390_PC64) + rc = apply_rela_bits(loc, val, 1, 64, 0, write); + break; + case R_390_GOT12: /* 12 bit GOT offset. */ + case R_390_GOT16: /* 16 bit GOT offset. */ + case R_390_GOT20: /* 20 bit GOT offset. */ + case R_390_GOT32: /* 32 bit GOT offset. */ + case R_390_GOT64: /* 64 bit GOT offset. */ + case R_390_GOTENT: /* 32 bit PC rel. to GOT entry shifted by 1. */ + case R_390_GOTPLT12: /* 12 bit offset to jump slot. */ + case R_390_GOTPLT20: /* 20 bit offset to jump slot. */ + case R_390_GOTPLT16: /* 16 bit offset to jump slot. */ + case R_390_GOTPLT32: /* 32 bit offset to jump slot. */ + case R_390_GOTPLT64: /* 64 bit offset to jump slot. */ + case R_390_GOTPLTENT: /* 32 bit rel. offset to jump slot >> 1. */ + if (info->got_initialized == 0) { + Elf_Addr *gotent = me->mem[MOD_TEXT].base + + me->arch.got_offset + + info->got_offset; + + write(gotent, &val, sizeof(*gotent)); + info->got_initialized = 1; + } + val = info->got_offset + rela->r_addend; + if (r_type == R_390_GOT12 || + r_type == R_390_GOTPLT12) + rc = apply_rela_bits(loc, val, 0, 12, 0, write); + else if (r_type == R_390_GOT16 || + r_type == R_390_GOTPLT16) + rc = apply_rela_bits(loc, val, 0, 16, 0, write); + else if (r_type == R_390_GOT20 || + r_type == R_390_GOTPLT20) + rc = apply_rela_bits(loc, val, 1, 20, 0, write); + else if (r_type == R_390_GOT32 || + r_type == R_390_GOTPLT32) + rc = apply_rela_bits(loc, val, 0, 32, 0, write); + else if (r_type == R_390_GOT64 || + r_type == R_390_GOTPLT64) + rc = apply_rela_bits(loc, val, 0, 64, 0, write); + else if (r_type == R_390_GOTENT || + r_type == R_390_GOTPLTENT) { + val += (Elf_Addr)me->mem[MOD_TEXT].base + + me->arch.got_offset - loc; + rc = apply_rela_bits(loc, val, 1, 32, 1, write); + } + break; + case R_390_PLT16DBL: /* 16 bit PC rel. PLT shifted by 1. */ + case R_390_PLT32DBL: /* 32 bit PC rel. PLT shifted by 1. */ + case R_390_PLT32: /* 32 bit PC relative PLT address. */ + case R_390_PLT64: /* 64 bit PC relative PLT address. */ + case R_390_PLTOFF16: /* 16 bit offset from GOT to PLT. */ + case R_390_PLTOFF32: /* 32 bit offset from GOT to PLT. */ + case R_390_PLTOFF64: /* 16 bit offset from GOT to PLT. */ + if (info->plt_initialized == 0) { + unsigned char insn[PLT_ENTRY_SIZE]; + char *plt_base; + char *ip; + + plt_base = me->mem[MOD_TEXT].base + me->arch.plt_offset; + ip = plt_base + info->plt_offset; + *(int *)insn = 0x0d10e310; /* basr 1,0 */ + *(int *)&insn[4] = 0x100c0004; /* lg 1,12(1) */ + if (IS_ENABLED(CONFIG_EXPOLINE) && !nospec_disable) { + char *jump_r1; + + jump_r1 = plt_base + me->arch.plt_size - + PLT_ENTRY_SIZE; + /* brcl 0xf,__jump_r1 */ + *(short *)&insn[8] = 0xc0f4; + *(int *)&insn[10] = (jump_r1 - (ip + 8)) / 2; + } else { + *(int *)&insn[8] = 0x07f10000; /* br %r1 */ + } + *(long *)&insn[14] = val; + + write(ip, insn, sizeof(insn)); + info->plt_initialized = 1; + } + if (r_type == R_390_PLTOFF16 || + r_type == R_390_PLTOFF32 || + r_type == R_390_PLTOFF64) + val = me->arch.plt_offset - me->arch.got_offset + + info->plt_offset + rela->r_addend; + else { + if (!((r_type == R_390_PLT16DBL && + val - loc + 0xffffUL < 0x1ffffeUL) || + (r_type == R_390_PLT32DBL && + val - loc + 0xffffffffULL < 0x1fffffffeULL))) + val = (Elf_Addr) me->mem[MOD_TEXT].base + + me->arch.plt_offset + + info->plt_offset; + val += rela->r_addend - loc; + } + if (r_type == R_390_PLT16DBL) + rc = apply_rela_bits(loc, val, 1, 16, 1, write); + else if (r_type == R_390_PLTOFF16) + rc = apply_rela_bits(loc, val, 0, 16, 0, write); + else if (r_type == R_390_PLT32DBL) + rc = apply_rela_bits(loc, val, 1, 32, 1, write); + else if (r_type == R_390_PLT32 || + r_type == R_390_PLTOFF32) + rc = apply_rela_bits(loc, val, 0, 32, 0, write); + else if (r_type == R_390_PLT64 || + r_type == R_390_PLTOFF64) + rc = apply_rela_bits(loc, val, 0, 64, 0, write); + break; + case R_390_GOTOFF16: /* 16 bit offset to GOT. */ + case R_390_GOTOFF32: /* 32 bit offset to GOT. */ + case R_390_GOTOFF64: /* 64 bit offset to GOT. */ + val = val + rela->r_addend - + ((Elf_Addr) me->mem[MOD_TEXT].base + me->arch.got_offset); + if (r_type == R_390_GOTOFF16) + rc = apply_rela_bits(loc, val, 0, 16, 0, write); + else if (r_type == R_390_GOTOFF32) + rc = apply_rela_bits(loc, val, 0, 32, 0, write); + else if (r_type == R_390_GOTOFF64) + rc = apply_rela_bits(loc, val, 0, 64, 0, write); + break; + case R_390_GOTPC: /* 32 bit PC relative offset to GOT. */ + case R_390_GOTPCDBL: /* 32 bit PC rel. off. to GOT shifted by 1. */ + val = (Elf_Addr) me->mem[MOD_TEXT].base + me->arch.got_offset + + rela->r_addend - loc; + if (r_type == R_390_GOTPC) + rc = apply_rela_bits(loc, val, 1, 32, 0, write); + else if (r_type == R_390_GOTPCDBL) + rc = apply_rela_bits(loc, val, 1, 32, 1, write); + break; + case R_390_COPY: + case R_390_GLOB_DAT: /* Create GOT entry. */ + case R_390_JMP_SLOT: /* Create PLT entry. */ + case R_390_RELATIVE: /* Adjust by program base. */ + /* Only needed if we want to support loading of + modules linked with -shared. */ + return -ENOEXEC; + default: + printk(KERN_ERR "module %s: unknown relocation: %u\n", + me->name, r_type); + return -ENOEXEC; + } + if (rc) { + printk(KERN_ERR "module %s: relocation error for symbol %s " + "(r_type %i, value 0x%lx)\n", + me->name, strtab + symtab[r_sym].st_name, + r_type, (unsigned long) val); + return rc; + } + return 0; +} + +static int __apply_relocate_add(Elf_Shdr *sechdrs, const char *strtab, + unsigned int symindex, unsigned int relsec, + struct module *me, + void *(*write)(void *dest, const void *src, size_t len)) +{ + Elf_Addr base; + Elf_Sym *symtab; + Elf_Rela *rela; + unsigned long i, n; + int rc; + + DEBUGP("Applying relocate section %u to %u\n", + relsec, sechdrs[relsec].sh_info); + base = sechdrs[sechdrs[relsec].sh_info].sh_addr; + symtab = (Elf_Sym *) sechdrs[symindex].sh_addr; + rela = (Elf_Rela *) sechdrs[relsec].sh_addr; + n = sechdrs[relsec].sh_size / sizeof(Elf_Rela); + + for (i = 0; i < n; i++, rela++) { + rc = apply_rela(rela, base, symtab, strtab, me, write); + if (rc) + return rc; + } + return 0; +} + +int apply_relocate_add(Elf_Shdr *sechdrs, const char *strtab, + unsigned int symindex, unsigned int relsec, + struct module *me) +{ + bool early = me->state == MODULE_STATE_UNFORMED; + void *(*write)(void *, const void *, size_t) = memcpy; + + if (!early) + write = s390_kernel_write; + + return __apply_relocate_add(sechdrs, strtab, symindex, relsec, me, + write); +} + +#ifdef CONFIG_FUNCTION_TRACER +static int module_alloc_ftrace_hotpatch_trampolines(struct module *me, + const Elf_Shdr *s) +{ + char *start, *end; + int numpages; + size_t size; + + size = FTRACE_HOTPATCH_TRAMPOLINES_SIZE(s->sh_size); + numpages = DIV_ROUND_UP(size, PAGE_SIZE); + start = module_alloc(numpages * PAGE_SIZE); + if (!start) + return -ENOMEM; + set_memory_rox((unsigned long)start, numpages); + end = start + size; + + me->arch.trampolines_start = (struct ftrace_hotpatch_trampoline *)start; + me->arch.trampolines_end = (struct ftrace_hotpatch_trampoline *)end; + me->arch.next_trampoline = me->arch.trampolines_start; + + return 0; +} +#endif /* CONFIG_FUNCTION_TRACER */ + +int module_finalize(const Elf_Ehdr *hdr, + const Elf_Shdr *sechdrs, + struct module *me) +{ + const Elf_Shdr *s; + char *secstrings, *secname; + void *aseg; +#ifdef CONFIG_FUNCTION_TRACER + int ret; +#endif + + if (IS_ENABLED(CONFIG_EXPOLINE) && + !nospec_disable && me->arch.plt_size) { + unsigned int *ij; + + ij = me->mem[MOD_TEXT].base + me->arch.plt_offset + + me->arch.plt_size - PLT_ENTRY_SIZE; + ij[0] = 0xc6000000; /* exrl %r0,.+10 */ + ij[1] = 0x0005a7f4; /* j . */ + ij[2] = 0x000007f1; /* br %r1 */ + } + + secstrings = (void *)hdr + sechdrs[hdr->e_shstrndx].sh_offset; + for (s = sechdrs; s < sechdrs + hdr->e_shnum; s++) { + aseg = (void *) s->sh_addr; + secname = secstrings + s->sh_name; + + if (!strcmp(".altinstructions", secname)) + /* patch .altinstructions */ + apply_alternatives(aseg, aseg + s->sh_size); + + if (IS_ENABLED(CONFIG_EXPOLINE) && + (str_has_prefix(secname, ".s390_indirect"))) + nospec_revert(aseg, aseg + s->sh_size); + + if (IS_ENABLED(CONFIG_EXPOLINE) && + (str_has_prefix(secname, ".s390_return"))) + nospec_revert(aseg, aseg + s->sh_size); + +#ifdef CONFIG_FUNCTION_TRACER + if (!strcmp(FTRACE_CALLSITE_SECTION, secname)) { + ret = module_alloc_ftrace_hotpatch_trampolines(me, s); + if (ret < 0) + return ret; + } +#endif /* CONFIG_FUNCTION_TRACER */ + } + + return 0; +} diff --git a/arch/s390/kernel/nmi.c b/arch/s390/kernel/nmi.c new file mode 100644 index 0000000000..38ec048752 --- /dev/null +++ b/arch/s390/kernel/nmi.c @@ -0,0 +1,513 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Machine check handler + * + * Copyright IBM Corp. 2000, 2009 + * Author(s): Ingo Adlung <adlung@de.ibm.com>, + * Martin Schwidefsky <schwidefsky@de.ibm.com>, + * Cornelia Huck <cornelia.huck@de.ibm.com>, + */ + +#include <linux/kernel_stat.h> +#include <linux/init.h> +#include <linux/errno.h> +#include <linux/entry-common.h> +#include <linux/hardirq.h> +#include <linux/log2.h> +#include <linux/kprobes.h> +#include <linux/kmemleak.h> +#include <linux/time.h> +#include <linux/module.h> +#include <linux/sched/signal.h> +#include <linux/kvm_host.h> +#include <linux/export.h> +#include <asm/lowcore.h> +#include <asm/smp.h> +#include <asm/stp.h> +#include <asm/cputime.h> +#include <asm/nmi.h> +#include <asm/crw.h> +#include <asm/switch_to.h> +#include <asm/ctl_reg.h> +#include <asm/asm-offsets.h> +#include <asm/pai.h> +#include <asm/vx-insn.h> + +struct mcck_struct { + unsigned int kill_task : 1; + unsigned int channel_report : 1; + unsigned int warning : 1; + unsigned int stp_queue : 1; + unsigned long mcck_code; +}; + +static DEFINE_PER_CPU(struct mcck_struct, cpu_mcck); + +static inline int nmi_needs_mcesa(void) +{ + return MACHINE_HAS_VX || MACHINE_HAS_GS; +} + +/* + * The initial machine check extended save area for the boot CPU. + * It will be replaced on the boot CPU reinit with an allocated + * structure. The structure is required for machine check happening + * early in the boot process. + */ +static struct mcesa boot_mcesa __aligned(MCESA_MAX_SIZE); + +void __init nmi_alloc_mcesa_early(u64 *mcesad) +{ + if (!nmi_needs_mcesa()) + return; + *mcesad = __pa(&boot_mcesa); + if (MACHINE_HAS_GS) + *mcesad |= ilog2(MCESA_MAX_SIZE); +} + +int nmi_alloc_mcesa(u64 *mcesad) +{ + unsigned long size; + void *origin; + + *mcesad = 0; + if (!nmi_needs_mcesa()) + return 0; + size = MACHINE_HAS_GS ? MCESA_MAX_SIZE : MCESA_MIN_SIZE; + origin = kmalloc(size, GFP_KERNEL); + if (!origin) + return -ENOMEM; + /* The pointer is stored with mcesa_bits ORed in */ + kmemleak_not_leak(origin); + *mcesad = __pa(origin); + if (MACHINE_HAS_GS) + *mcesad |= ilog2(MCESA_MAX_SIZE); + return 0; +} + +void nmi_free_mcesa(u64 *mcesad) +{ + if (!nmi_needs_mcesa()) + return; + kfree(__va(*mcesad & MCESA_ORIGIN_MASK)); +} + +static __always_inline char *nmi_puts(char *dest, const char *src) +{ + while (*src) + *dest++ = *src++; + *dest = 0; + return dest; +} + +static __always_inline char *u64_to_hex(char *dest, u64 val) +{ + int i, num; + + for (i = 1; i <= 16; i++) { + num = (val >> (64 - 4 * i)) & 0xf; + if (num >= 10) + *dest++ = 'A' + num - 10; + else + *dest++ = '0' + num; + } + *dest = 0; + return dest; +} + +static notrace void s390_handle_damage(void) +{ + union ctlreg0 cr0, cr0_new; + char message[100]; + psw_t psw_save; + char *ptr; + + smp_emergency_stop(); + diag_amode31_ops.diag308_reset(); + ptr = nmi_puts(message, "System stopped due to unrecoverable machine check, code: 0x"); + u64_to_hex(ptr, S390_lowcore.mcck_interruption_code); + + /* + * Disable low address protection and make machine check new PSW a + * disabled wait PSW. Any additional machine check cannot be handled. + */ + __ctl_store(cr0.val, 0, 0); + cr0_new = cr0; + cr0_new.lap = 0; + __ctl_load(cr0_new.val, 0, 0); + psw_save = S390_lowcore.mcck_new_psw; + psw_bits(S390_lowcore.mcck_new_psw).io = 0; + psw_bits(S390_lowcore.mcck_new_psw).ext = 0; + psw_bits(S390_lowcore.mcck_new_psw).wait = 1; + sclp_emergency_printk(message); + + /* + * Restore machine check new PSW and control register 0 to original + * values. This makes possible system dump analysis easier. + */ + S390_lowcore.mcck_new_psw = psw_save; + __ctl_load(cr0.val, 0, 0); + disabled_wait(); + while (1); +} +NOKPROBE_SYMBOL(s390_handle_damage); + +/* + * Main machine check handler function. Will be called with interrupts disabled + * and machine checks enabled. + */ +void s390_handle_mcck(void) +{ + struct mcck_struct mcck; + + /* + * Disable machine checks and get the current state of accumulated + * machine checks. Afterwards delete the old state and enable machine + * checks again. + */ + local_mcck_disable(); + mcck = *this_cpu_ptr(&cpu_mcck); + memset(this_cpu_ptr(&cpu_mcck), 0, sizeof(mcck)); + local_mcck_enable(); + + if (mcck.channel_report) + crw_handle_channel_report(); + /* + * A warning may remain for a prolonged period on the bare iron. + * (actually until the machine is powered off, or the problem is gone) + * So we just stop listening for the WARNING MCH and avoid continuously + * being interrupted. One caveat is however, that we must do this per + * processor and cannot use the smp version of ctl_clear_bit(). + * On VM we only get one interrupt per virtally presented machinecheck. + * Though one suffices, we may get one interrupt per (virtual) cpu. + */ + if (mcck.warning) { /* WARNING pending ? */ + static int mchchk_wng_posted = 0; + + /* Use single cpu clear, as we cannot handle smp here. */ + __ctl_clear_bit(14, 24); /* Disable WARNING MCH */ + if (xchg(&mchchk_wng_posted, 1) == 0) + kill_cad_pid(SIGPWR, 1); + } + if (mcck.stp_queue) + stp_queue_work(); + if (mcck.kill_task) { + printk(KERN_EMERG "mcck: Terminating task because of machine " + "malfunction (code 0x%016lx).\n", mcck.mcck_code); + printk(KERN_EMERG "mcck: task: %s, pid: %d.\n", + current->comm, current->pid); + if (is_global_init(current)) + panic("mcck: Attempting to kill init!\n"); + do_send_sig_info(SIGKILL, SEND_SIG_PRIV, current, PIDTYPE_PID); + } +} + +/* + * returns 0 if register contents could be validated + * returns 1 otherwise + */ +static int notrace s390_validate_registers(union mci mci) +{ + struct mcesa *mcesa; + void *fpt_save_area; + union ctlreg2 cr2; + int kill_task; + u64 zero; + + kill_task = 0; + zero = 0; + + if (!mci.gr || !mci.fp) + kill_task = 1; + fpt_save_area = &S390_lowcore.floating_pt_save_area; + if (!mci.fc) { + kill_task = 1; + asm volatile( + " lfpc %0\n" + : + : "Q" (zero)); + } else { + asm volatile( + " lfpc %0\n" + : + : "Q" (S390_lowcore.fpt_creg_save_area)); + } + + mcesa = __va(S390_lowcore.mcesad & MCESA_ORIGIN_MASK); + if (!MACHINE_HAS_VX) { + /* Validate floating point registers */ + asm volatile( + " ld 0,0(%0)\n" + " ld 1,8(%0)\n" + " ld 2,16(%0)\n" + " ld 3,24(%0)\n" + " ld 4,32(%0)\n" + " ld 5,40(%0)\n" + " ld 6,48(%0)\n" + " ld 7,56(%0)\n" + " ld 8,64(%0)\n" + " ld 9,72(%0)\n" + " ld 10,80(%0)\n" + " ld 11,88(%0)\n" + " ld 12,96(%0)\n" + " ld 13,104(%0)\n" + " ld 14,112(%0)\n" + " ld 15,120(%0)\n" + : + : "a" (fpt_save_area) + : "memory"); + } else { + /* Validate vector registers */ + union ctlreg0 cr0; + + /* + * The vector validity must only be checked if not running a + * KVM guest. For KVM guests the machine check is forwarded by + * KVM and it is the responsibility of the guest to take + * appropriate actions. The host vector or FPU values have been + * saved by KVM and will be restored by KVM. + */ + if (!mci.vr && !test_cpu_flag(CIF_MCCK_GUEST)) + kill_task = 1; + cr0.val = S390_lowcore.cregs_save_area[0]; + cr0.afp = cr0.vx = 1; + __ctl_load(cr0.val, 0, 0); + asm volatile( + " la 1,%0\n" + " VLM 0,15,0,1\n" + " VLM 16,31,256,1\n" + : + : "Q" (*(struct vx_array *)mcesa->vector_save_area) + : "1"); + __ctl_load(S390_lowcore.cregs_save_area[0], 0, 0); + } + /* Validate access registers */ + asm volatile( + " lam 0,15,0(%0)\n" + : + : "a" (&S390_lowcore.access_regs_save_area) + : "memory"); + if (!mci.ar) + kill_task = 1; + /* Validate guarded storage registers */ + cr2.val = S390_lowcore.cregs_save_area[2]; + if (cr2.gse) { + if (!mci.gs) { + /* + * 2 cases: + * - machine check in kernel or userspace + * - machine check while running SIE (KVM guest) + * For kernel or userspace the userspace values of + * guarded storage control can not be recreated, the + * process must be terminated. + * For SIE the guest values of guarded storage can not + * be recreated. This is either due to a bug or due to + * GS being disabled in the guest. The guest will be + * notified by KVM code and the guests machine check + * handling must take care of this. The host values + * are saved by KVM and are not affected. + */ + if (!test_cpu_flag(CIF_MCCK_GUEST)) + kill_task = 1; + } else { + load_gs_cb((struct gs_cb *)mcesa->guarded_storage_save_area); + } + } + /* + * The getcpu vdso syscall reads CPU number from the programmable + * field of the TOD clock. Disregard the TOD programmable register + * validity bit and load the CPU number into the TOD programmable + * field unconditionally. + */ + set_tod_programmable_field(raw_smp_processor_id()); + /* Validate clock comparator register */ + set_clock_comparator(S390_lowcore.clock_comparator); + + if (!mci.ms || !mci.pm || !mci.ia) + kill_task = 1; + + return kill_task; +} +NOKPROBE_SYMBOL(s390_validate_registers); + +/* + * Backup the guest's machine check info to its description block + */ +static void notrace s390_backup_mcck_info(struct pt_regs *regs) +{ + struct mcck_volatile_info *mcck_backup; + struct sie_page *sie_page; + + /* r14 contains the sie block, which was set in sie64a */ + struct kvm_s390_sie_block *sie_block = phys_to_virt(regs->gprs[14]); + + if (sie_block == NULL) + /* Something's seriously wrong, stop system. */ + s390_handle_damage(); + + sie_page = container_of(sie_block, struct sie_page, sie_block); + mcck_backup = &sie_page->mcck_info; + mcck_backup->mcic = S390_lowcore.mcck_interruption_code & + ~(MCCK_CODE_CP | MCCK_CODE_EXT_DAMAGE); + mcck_backup->ext_damage_code = S390_lowcore.external_damage_code; + mcck_backup->failing_storage_address + = S390_lowcore.failing_storage_address; +} +NOKPROBE_SYMBOL(s390_backup_mcck_info); + +#define MAX_IPD_COUNT 29 +#define MAX_IPD_TIME (5 * 60 * USEC_PER_SEC) /* 5 minutes */ + +#define ED_STP_ISLAND 6 /* External damage STP island check */ +#define ED_STP_SYNC 7 /* External damage STP sync check */ + +#define MCCK_CODE_NO_GUEST (MCCK_CODE_CP | MCCK_CODE_EXT_DAMAGE) + +/* + * machine check handler. + */ +void notrace s390_do_machine_check(struct pt_regs *regs) +{ + static int ipd_count; + static DEFINE_SPINLOCK(ipd_lock); + static unsigned long long last_ipd; + struct mcck_struct *mcck; + unsigned long long tmp; + irqentry_state_t irq_state; + union mci mci; + unsigned long mcck_dam_code; + int mcck_pending = 0; + + irq_state = irqentry_nmi_enter(regs); + + if (user_mode(regs)) + update_timer_mcck(); + inc_irq_stat(NMI_NMI); + mci.val = S390_lowcore.mcck_interruption_code; + mcck = this_cpu_ptr(&cpu_mcck); + + /* + * Reinject the instruction processing damages' machine checks + * including Delayed Access Exception into the guest + * instead of damaging the host if they happen in the guest. + */ + if (mci.pd && !test_cpu_flag(CIF_MCCK_GUEST)) { + if (mci.b) { + /* Processing backup -> verify if we can survive this */ + u64 z_mcic, o_mcic, t_mcic; + z_mcic = (1ULL<<63 | 1ULL<<59 | 1ULL<<29); + o_mcic = (1ULL<<43 | 1ULL<<42 | 1ULL<<41 | 1ULL<<40 | + 1ULL<<36 | 1ULL<<35 | 1ULL<<34 | 1ULL<<32 | + 1ULL<<30 | 1ULL<<21 | 1ULL<<20 | 1ULL<<17 | + 1ULL<<16); + t_mcic = mci.val; + + if (((t_mcic & z_mcic) != 0) || + ((t_mcic & o_mcic) != o_mcic)) { + s390_handle_damage(); + } + + /* + * Nullifying exigent condition, therefore we might + * retry this instruction. + */ + spin_lock(&ipd_lock); + tmp = get_tod_clock(); + if (((tmp - last_ipd) >> 12) < MAX_IPD_TIME) + ipd_count++; + else + ipd_count = 1; + last_ipd = tmp; + if (ipd_count == MAX_IPD_COUNT) + s390_handle_damage(); + spin_unlock(&ipd_lock); + } else { + /* Processing damage -> stopping machine */ + s390_handle_damage(); + } + } + if (s390_validate_registers(mci)) { + if (!user_mode(regs)) + s390_handle_damage(); + /* + * Couldn't restore all register contents for the + * user space process -> mark task for termination. + */ + mcck->kill_task = 1; + mcck->mcck_code = mci.val; + mcck_pending = 1; + } + + /* + * Backup the machine check's info if it happens when the guest + * is running. + */ + if (test_cpu_flag(CIF_MCCK_GUEST)) + s390_backup_mcck_info(regs); + + if (mci.cd) { + /* Timing facility damage */ + s390_handle_damage(); + } + if (mci.ed && mci.ec) { + /* External damage */ + if (S390_lowcore.external_damage_code & (1U << ED_STP_SYNC)) + mcck->stp_queue |= stp_sync_check(); + if (S390_lowcore.external_damage_code & (1U << ED_STP_ISLAND)) + mcck->stp_queue |= stp_island_check(); + mcck_pending = 1; + } + /* + * Reinject storage related machine checks into the guest if they + * happen when the guest is running. + */ + if (!test_cpu_flag(CIF_MCCK_GUEST)) { + /* Storage error uncorrected */ + if (mci.se) + s390_handle_damage(); + /* Storage key-error uncorrected */ + if (mci.ke) + s390_handle_damage(); + /* Storage degradation */ + if (mci.ds && mci.fa) + s390_handle_damage(); + } + if (mci.cp) { + /* Channel report word pending */ + mcck->channel_report = 1; + mcck_pending = 1; + } + if (mci.w) { + /* Warning pending */ + mcck->warning = 1; + mcck_pending = 1; + } + + /* + * If there are only Channel Report Pending and External Damage + * machine checks, they will not be reinjected into the guest + * because they refer to host conditions only. + */ + mcck_dam_code = (mci.val & MCIC_SUBCLASS_MASK); + if (test_cpu_flag(CIF_MCCK_GUEST) && + (mcck_dam_code & MCCK_CODE_NO_GUEST) != mcck_dam_code) { + /* Set exit reason code for host's later handling */ + *((long *)(regs->gprs[15] + __SF_SIE_REASON)) = -EINTR; + } + clear_cpu_flag(CIF_MCCK_GUEST); + + if (mcck_pending) + schedule_mcck_handler(); + + irqentry_nmi_exit(regs, irq_state); +} +NOKPROBE_SYMBOL(s390_do_machine_check); + +static int __init machine_check_init(void) +{ + ctl_set_bit(14, 25); /* enable external damage MCH */ + ctl_set_bit(14, 27); /* enable system recovery MCH */ + ctl_set_bit(14, 24); /* enable warning MCH */ + return 0; +} +early_initcall(machine_check_init); diff --git a/arch/s390/kernel/nospec-branch.c b/arch/s390/kernel/nospec-branch.c new file mode 100644 index 0000000000..d1b16d83e4 --- /dev/null +++ b/arch/s390/kernel/nospec-branch.c @@ -0,0 +1,156 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/module.h> +#include <linux/device.h> +#include <linux/cpu.h> +#include <asm/nospec-branch.h> + +static int __init nobp_setup_early(char *str) +{ + bool enabled; + int rc; + + rc = kstrtobool(str, &enabled); + if (rc) + return rc; + if (enabled && test_facility(82)) { + /* + * The user explicitly requested nobp=1, enable it and + * disable the expoline support. + */ + __set_facility(82, alt_stfle_fac_list); + if (IS_ENABLED(CONFIG_EXPOLINE)) + nospec_disable = 1; + } else { + __clear_facility(82, alt_stfle_fac_list); + } + return 0; +} +early_param("nobp", nobp_setup_early); + +static int __init nospec_setup_early(char *str) +{ + __clear_facility(82, alt_stfle_fac_list); + return 0; +} +early_param("nospec", nospec_setup_early); + +static int __init nospec_report(void) +{ + if (test_facility(156)) + pr_info("Spectre V2 mitigation: etokens\n"); + if (nospec_uses_trampoline()) + pr_info("Spectre V2 mitigation: execute trampolines\n"); + if (__test_facility(82, alt_stfle_fac_list)) + pr_info("Spectre V2 mitigation: limited branch prediction\n"); + return 0; +} +arch_initcall(nospec_report); + +#ifdef CONFIG_EXPOLINE + +int nospec_disable = IS_ENABLED(CONFIG_EXPOLINE_OFF); + +static int __init nospectre_v2_setup_early(char *str) +{ + nospec_disable = 1; + return 0; +} +early_param("nospectre_v2", nospectre_v2_setup_early); + +void __init nospec_auto_detect(void) +{ + if (test_facility(156) || cpu_mitigations_off()) { + /* + * The machine supports etokens. + * Disable expolines and disable nobp. + */ + if (__is_defined(CC_USING_EXPOLINE)) + nospec_disable = 1; + __clear_facility(82, alt_stfle_fac_list); + } else if (__is_defined(CC_USING_EXPOLINE)) { + /* + * The kernel has been compiled with expolines. + * Keep expolines enabled and disable nobp. + */ + nospec_disable = 0; + __clear_facility(82, alt_stfle_fac_list); + } + /* + * If the kernel has not been compiled with expolines the + * nobp setting decides what is done, this depends on the + * CONFIG_KERNEL_NP option and the nobp/nospec parameters. + */ +} + +static int __init spectre_v2_setup_early(char *str) +{ + if (str && !strncmp(str, "on", 2)) { + nospec_disable = 0; + __clear_facility(82, alt_stfle_fac_list); + } + if (str && !strncmp(str, "off", 3)) + nospec_disable = 1; + if (str && !strncmp(str, "auto", 4)) + nospec_auto_detect(); + return 0; +} +early_param("spectre_v2", spectre_v2_setup_early); + +static void __init_or_module __nospec_revert(s32 *start, s32 *end) +{ + enum { BRCL_EXPOLINE, BRASL_EXPOLINE } type; + static const u8 branch[] = { 0x47, 0x00, 0x07, 0x00 }; + u8 *instr, *thunk, *br; + u8 insnbuf[6]; + s32 *epo; + + /* Second part of the instruction replace is always a nop */ + memcpy(insnbuf + 2, branch, sizeof(branch)); + for (epo = start; epo < end; epo++) { + instr = (u8 *) epo + *epo; + if (instr[0] == 0xc0 && (instr[1] & 0x0f) == 0x04) + type = BRCL_EXPOLINE; /* brcl instruction */ + else if (instr[0] == 0xc0 && (instr[1] & 0x0f) == 0x05) + type = BRASL_EXPOLINE; /* brasl instruction */ + else + continue; + thunk = instr + (*(int *)(instr + 2)) * 2; + if (thunk[0] == 0xc6 && thunk[1] == 0x00) + /* exrl %r0,<target-br> */ + br = thunk + (*(int *)(thunk + 2)) * 2; + else + continue; + if (br[0] != 0x07 || (br[1] & 0xf0) != 0xf0) + continue; + switch (type) { + case BRCL_EXPOLINE: + /* brcl to thunk, replace with br + nop */ + insnbuf[0] = br[0]; + insnbuf[1] = (instr[1] & 0xf0) | (br[1] & 0x0f); + break; + case BRASL_EXPOLINE: + /* brasl to thunk, replace with basr + nop */ + insnbuf[0] = 0x0d; + insnbuf[1] = (instr[1] & 0xf0) | (br[1] & 0x0f); + break; + } + + s390_kernel_write(instr, insnbuf, 6); + } +} + +void __init_or_module nospec_revert(s32 *start, s32 *end) +{ + if (nospec_disable) + __nospec_revert(start, end); +} + +extern s32 __nospec_call_start[], __nospec_call_end[]; +extern s32 __nospec_return_start[], __nospec_return_end[]; +void __init nospec_init_branches(void) +{ + nospec_revert(__nospec_call_start, __nospec_call_end); + nospec_revert(__nospec_return_start, __nospec_return_end); +} + +#endif /* CONFIG_EXPOLINE */ diff --git a/arch/s390/kernel/nospec-sysfs.c b/arch/s390/kernel/nospec-sysfs.c new file mode 100644 index 0000000000..52d4353188 --- /dev/null +++ b/arch/s390/kernel/nospec-sysfs.c @@ -0,0 +1,23 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/device.h> +#include <linux/cpu.h> +#include <asm/facility.h> +#include <asm/nospec-branch.h> + +ssize_t cpu_show_spectre_v1(struct device *dev, + struct device_attribute *attr, char *buf) +{ + return sprintf(buf, "Mitigation: __user pointer sanitization\n"); +} + +ssize_t cpu_show_spectre_v2(struct device *dev, + struct device_attribute *attr, char *buf) +{ + if (test_facility(156)) + return sprintf(buf, "Mitigation: etokens\n"); + if (nospec_uses_trampoline()) + return sprintf(buf, "Mitigation: execute trampolines\n"); + if (__test_facility(82, alt_stfle_fac_list)) + return sprintf(buf, "Mitigation: limited branch prediction\n"); + return sprintf(buf, "Vulnerable\n"); +} diff --git a/arch/s390/kernel/numa.c b/arch/s390/kernel/numa.c new file mode 100644 index 0000000000..23ab9f02f2 --- /dev/null +++ b/arch/s390/kernel/numa.c @@ -0,0 +1,35 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * NUMA support for s390 + * + * Implement NUMA core code. + * + * Copyright IBM Corp. 2015 + */ + +#include <linux/kernel.h> +#include <linux/mmzone.h> +#include <linux/cpumask.h> +#include <linux/memblock.h> +#include <linux/node.h> +#include <asm/numa.h> + +struct pglist_data *node_data[MAX_NUMNODES]; +EXPORT_SYMBOL(node_data); + +void __init numa_setup(void) +{ + int nid; + + nodes_clear(node_possible_map); + node_set(0, node_possible_map); + node_set_online(0); + for (nid = 0; nid < MAX_NUMNODES; nid++) { + NODE_DATA(nid) = memblock_alloc(sizeof(pg_data_t), 8); + if (!NODE_DATA(nid)) + panic("%s: Failed to allocate %zu bytes align=0x%x\n", + __func__, sizeof(pg_data_t), 8); + } + NODE_DATA(0)->node_spanned_pages = memblock_end_of_DRAM() >> PAGE_SHIFT; + NODE_DATA(0)->node_id = 0; +} diff --git a/arch/s390/kernel/os_info.c b/arch/s390/kernel/os_info.c new file mode 100644 index 0000000000..6e1824141b --- /dev/null +++ b/arch/s390/kernel/os_info.c @@ -0,0 +1,173 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * OS info memory interface + * + * Copyright IBM Corp. 2012 + * Author(s): Michael Holzheu <holzheu@linux.vnet.ibm.com> + */ + +#define KMSG_COMPONENT "os_info" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/crash_dump.h> +#include <linux/kernel.h> +#include <linux/slab.h> +#include <asm/checksum.h> +#include <asm/abs_lowcore.h> +#include <asm/os_info.h> +#include <asm/maccess.h> +#include <asm/asm-offsets.h> + +/* + * OS info structure has to be page aligned + */ +static struct os_info os_info __page_aligned_data; + +/* + * Compute checksum over OS info structure + */ +u32 os_info_csum(struct os_info *os_info) +{ + int size = sizeof(*os_info) - offsetof(struct os_info, version_major); + return (__force u32)csum_partial(&os_info->version_major, size, 0); +} + +/* + * Add crashkernel info to OS info and update checksum + */ +void os_info_crashkernel_add(unsigned long base, unsigned long size) +{ + os_info.crashkernel_addr = (u64)(unsigned long)base; + os_info.crashkernel_size = (u64)(unsigned long)size; + os_info.csum = os_info_csum(&os_info); +} + +/* + * Add OS info entry and update checksum + */ +void os_info_entry_add(int nr, void *ptr, u64 size) +{ + os_info.entry[nr].addr = __pa(ptr); + os_info.entry[nr].size = size; + os_info.entry[nr].csum = (__force u32)csum_partial(ptr, size, 0); + os_info.csum = os_info_csum(&os_info); +} + +/* + * Initialize OS info structure and set lowcore pointer + */ +void __init os_info_init(void) +{ + struct lowcore *abs_lc; + + os_info.version_major = OS_INFO_VERSION_MAJOR; + os_info.version_minor = OS_INFO_VERSION_MINOR; + os_info.magic = OS_INFO_MAGIC; + os_info.csum = os_info_csum(&os_info); + abs_lc = get_abs_lowcore(); + abs_lc->os_info = __pa(&os_info); + put_abs_lowcore(abs_lc); +} + +#ifdef CONFIG_CRASH_DUMP + +static struct os_info *os_info_old; + +/* + * Allocate and copy OS info entry from oldmem + */ +static void os_info_old_alloc(int nr, int align) +{ + unsigned long addr, size = 0; + char *buf, *buf_align, *msg; + u32 csum; + + addr = os_info_old->entry[nr].addr; + if (!addr) { + msg = "not available"; + goto fail; + } + size = os_info_old->entry[nr].size; + buf = kmalloc(size + align - 1, GFP_KERNEL); + if (!buf) { + msg = "alloc failed"; + goto fail; + } + buf_align = PTR_ALIGN(buf, align); + if (copy_oldmem_kernel(buf_align, addr, size)) { + msg = "copy failed"; + goto fail_free; + } + csum = (__force u32)csum_partial(buf_align, size, 0); + if (csum != os_info_old->entry[nr].csum) { + msg = "checksum failed"; + goto fail_free; + } + os_info_old->entry[nr].addr = (u64)(unsigned long)buf_align; + msg = "copied"; + goto out; +fail_free: + kfree(buf); +fail: + os_info_old->entry[nr].addr = 0; +out: + pr_info("entry %i: %s (addr=0x%lx size=%lu)\n", + nr, msg, addr, size); +} + +/* + * Initialize os info and os info entries from oldmem + */ +static void os_info_old_init(void) +{ + static int os_info_init; + unsigned long addr; + + if (os_info_init) + return; + if (!oldmem_data.start) + goto fail; + if (copy_oldmem_kernel(&addr, __LC_OS_INFO, sizeof(addr))) + goto fail; + if (addr == 0 || addr % PAGE_SIZE) + goto fail; + os_info_old = kzalloc(sizeof(*os_info_old), GFP_KERNEL); + if (!os_info_old) + goto fail; + if (copy_oldmem_kernel(os_info_old, addr, sizeof(*os_info_old))) + goto fail_free; + if (os_info_old->magic != OS_INFO_MAGIC) + goto fail_free; + if (os_info_old->csum != os_info_csum(os_info_old)) + goto fail_free; + if (os_info_old->version_major > OS_INFO_VERSION_MAJOR) + goto fail_free; + os_info_old_alloc(OS_INFO_VMCOREINFO, 1); + os_info_old_alloc(OS_INFO_REIPL_BLOCK, 1); + pr_info("crashkernel: addr=0x%lx size=%lu\n", + (unsigned long) os_info_old->crashkernel_addr, + (unsigned long) os_info_old->crashkernel_size); + os_info_init = 1; + return; +fail_free: + kfree(os_info_old); +fail: + os_info_init = 1; + os_info_old = NULL; +} + +/* + * Return pointer to os infor entry and its size + */ +void *os_info_old_entry(int nr, unsigned long *size) +{ + os_info_old_init(); + + if (!os_info_old) + return NULL; + if (!os_info_old->entry[nr].addr) + return NULL; + *size = (unsigned long) os_info_old->entry[nr].size; + return (void *)(unsigned long)os_info_old->entry[nr].addr; +} +#endif diff --git a/arch/s390/kernel/perf_cpum_cf.c b/arch/s390/kernel/perf_cpum_cf.c new file mode 100644 index 0000000000..850c11ea63 --- /dev/null +++ b/arch/s390/kernel/perf_cpum_cf.c @@ -0,0 +1,1950 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Performance event support for s390x - CPU-measurement Counter Facility + * + * Copyright IBM Corp. 2012, 2023 + * Author(s): Hendrik Brueckner <brueckner@linux.ibm.com> + * Thomas Richter <tmricht@linux.ibm.com> + */ +#define KMSG_COMPONENT "cpum_cf" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/kernel.h> +#include <linux/kernel_stat.h> +#include <linux/percpu.h> +#include <linux/notifier.h> +#include <linux/init.h> +#include <linux/export.h> +#include <linux/miscdevice.h> +#include <linux/perf_event.h> + +#include <asm/cpu_mf.h> +#include <asm/hwctrset.h> +#include <asm/debug.h> + +enum cpumf_ctr_set { + CPUMF_CTR_SET_BASIC = 0, /* Basic Counter Set */ + CPUMF_CTR_SET_USER = 1, /* Problem-State Counter Set */ + CPUMF_CTR_SET_CRYPTO = 2, /* Crypto-Activity Counter Set */ + CPUMF_CTR_SET_EXT = 3, /* Extended Counter Set */ + CPUMF_CTR_SET_MT_DIAG = 4, /* MT-diagnostic Counter Set */ + + /* Maximum number of counter sets */ + CPUMF_CTR_SET_MAX, +}; + +#define CPUMF_LCCTL_ENABLE_SHIFT 16 +#define CPUMF_LCCTL_ACTCTL_SHIFT 0 + +static inline void ctr_set_enable(u64 *state, u64 ctrsets) +{ + *state |= ctrsets << CPUMF_LCCTL_ENABLE_SHIFT; +} + +static inline void ctr_set_disable(u64 *state, u64 ctrsets) +{ + *state &= ~(ctrsets << CPUMF_LCCTL_ENABLE_SHIFT); +} + +static inline void ctr_set_start(u64 *state, u64 ctrsets) +{ + *state |= ctrsets << CPUMF_LCCTL_ACTCTL_SHIFT; +} + +static inline void ctr_set_stop(u64 *state, u64 ctrsets) +{ + *state &= ~(ctrsets << CPUMF_LCCTL_ACTCTL_SHIFT); +} + +static inline int ctr_stcctm(enum cpumf_ctr_set set, u64 range, u64 *dest) +{ + switch (set) { + case CPUMF_CTR_SET_BASIC: + return stcctm(BASIC, range, dest); + case CPUMF_CTR_SET_USER: + return stcctm(PROBLEM_STATE, range, dest); + case CPUMF_CTR_SET_CRYPTO: + return stcctm(CRYPTO_ACTIVITY, range, dest); + case CPUMF_CTR_SET_EXT: + return stcctm(EXTENDED, range, dest); + case CPUMF_CTR_SET_MT_DIAG: + return stcctm(MT_DIAG_CLEARING, range, dest); + case CPUMF_CTR_SET_MAX: + return 3; + } + return 3; +} + +struct cpu_cf_events { + refcount_t refcnt; /* Reference count */ + atomic_t ctr_set[CPUMF_CTR_SET_MAX]; + u64 state; /* For perf_event_open SVC */ + u64 dev_state; /* For /dev/hwctr */ + unsigned int flags; + size_t used; /* Bytes used in data */ + size_t usedss; /* Bytes used in start/stop */ + unsigned char start[PAGE_SIZE]; /* Counter set at event add */ + unsigned char stop[PAGE_SIZE]; /* Counter set at event delete */ + unsigned char data[PAGE_SIZE]; /* Counter set at /dev/hwctr */ + unsigned int sets; /* # Counter set saved in memory */ +}; + +static unsigned int cfdiag_cpu_speed; /* CPU speed for CF_DIAG trailer */ +static debug_info_t *cf_dbg; + +/* + * The CPU Measurement query counter information instruction contains + * information which varies per machine generation, but is constant and + * does not change when running on a particular machine, such as counter + * first and second version number. This is needed to determine the size + * of counter sets. Extract this information at device driver initialization. + */ +static struct cpumf_ctr_info cpumf_ctr_info; + +struct cpu_cf_ptr { + struct cpu_cf_events *cpucf; +}; + +static struct cpu_cf_root { /* Anchor to per CPU data */ + refcount_t refcnt; /* Overall active events */ + struct cpu_cf_ptr __percpu *cfptr; +} cpu_cf_root; + +/* + * Serialize event initialization and event removal. Both are called from + * user space in task context with perf_event_open() and close() + * system calls. + * + * This mutex serializes functions cpum_cf_alloc_cpu() called at event + * initialization via cpumf_pmu_event_init() and function cpum_cf_free_cpu() + * called at event removal via call back function hw_perf_event_destroy() + * when the event is deleted. They are serialized to enforce correct + * bookkeeping of pointer and reference counts anchored by + * struct cpu_cf_root and the access to cpu_cf_root::refcnt and the + * per CPU pointers stored in cpu_cf_root::cfptr. + */ +static DEFINE_MUTEX(pmc_reserve_mutex); + +/* + * Get pointer to per-cpu structure. + * + * Function get_cpu_cfhw() is called from + * - cfset_copy_all(): This function is protected by cpus_read_lock(), so + * CPU hot plug remove can not happen. Event removal requires a close() + * first. + * + * Function this_cpu_cfhw() is called from perf common code functions: + * - pmu_{en|dis}able(), pmu_{add|del}()and pmu_{start|stop}(): + * All functions execute with interrupts disabled on that particular CPU. + * - cfset_ioctl_{on|off}, cfset_cpu_read(): see comment cfset_copy_all(). + * + * Therefore it is safe to access the CPU specific pointer to the event. + */ +static struct cpu_cf_events *get_cpu_cfhw(int cpu) +{ + struct cpu_cf_ptr __percpu *p = cpu_cf_root.cfptr; + + if (p) { + struct cpu_cf_ptr *q = per_cpu_ptr(p, cpu); + + return q->cpucf; + } + return NULL; +} + +static struct cpu_cf_events *this_cpu_cfhw(void) +{ + return get_cpu_cfhw(smp_processor_id()); +} + +/* Disable counter sets on dedicated CPU */ +static void cpum_cf_reset_cpu(void *flags) +{ + lcctl(0); +} + +/* Free per CPU data when the last event is removed. */ +static void cpum_cf_free_root(void) +{ + if (!refcount_dec_and_test(&cpu_cf_root.refcnt)) + return; + free_percpu(cpu_cf_root.cfptr); + cpu_cf_root.cfptr = NULL; + irq_subclass_unregister(IRQ_SUBCLASS_MEASUREMENT_ALERT); + on_each_cpu(cpum_cf_reset_cpu, NULL, 1); + debug_sprintf_event(cf_dbg, 4, "%s root.refcnt %u cfptr %d\n", + __func__, refcount_read(&cpu_cf_root.refcnt), + !cpu_cf_root.cfptr); +} + +/* + * On initialization of first event also allocate per CPU data dynamically. + * Start with an array of pointers, the array size is the maximum number of + * CPUs possible, which might be larger than the number of CPUs currently + * online. + */ +static int cpum_cf_alloc_root(void) +{ + int rc = 0; + + if (refcount_inc_not_zero(&cpu_cf_root.refcnt)) + return rc; + + /* The memory is already zeroed. */ + cpu_cf_root.cfptr = alloc_percpu(struct cpu_cf_ptr); + if (cpu_cf_root.cfptr) { + refcount_set(&cpu_cf_root.refcnt, 1); + on_each_cpu(cpum_cf_reset_cpu, NULL, 1); + irq_subclass_register(IRQ_SUBCLASS_MEASUREMENT_ALERT); + } else { + rc = -ENOMEM; + } + + return rc; +} + +/* Free CPU counter data structure for a PMU */ +static void cpum_cf_free_cpu(int cpu) +{ + struct cpu_cf_events *cpuhw; + struct cpu_cf_ptr *p; + + mutex_lock(&pmc_reserve_mutex); + /* + * When invoked via CPU hotplug handler, there might be no events + * installed or that particular CPU might not have an + * event installed. This anchor pointer can be NULL! + */ + if (!cpu_cf_root.cfptr) + goto out; + p = per_cpu_ptr(cpu_cf_root.cfptr, cpu); + cpuhw = p->cpucf; + /* + * Might be zero when called from CPU hotplug handler and no event + * installed on that CPU, but on different CPUs. + */ + if (!cpuhw) + goto out; + + if (refcount_dec_and_test(&cpuhw->refcnt)) { + kfree(cpuhw); + p->cpucf = NULL; + } + cpum_cf_free_root(); +out: + mutex_unlock(&pmc_reserve_mutex); +} + +/* Allocate CPU counter data structure for a PMU. Called under mutex lock. */ +static int cpum_cf_alloc_cpu(int cpu) +{ + struct cpu_cf_events *cpuhw; + struct cpu_cf_ptr *p; + int rc; + + mutex_lock(&pmc_reserve_mutex); + rc = cpum_cf_alloc_root(); + if (rc) + goto unlock; + p = per_cpu_ptr(cpu_cf_root.cfptr, cpu); + cpuhw = p->cpucf; + + if (!cpuhw) { + cpuhw = kzalloc(sizeof(*cpuhw), GFP_KERNEL); + if (cpuhw) { + p->cpucf = cpuhw; + refcount_set(&cpuhw->refcnt, 1); + } else { + rc = -ENOMEM; + } + } else { + refcount_inc(&cpuhw->refcnt); + } + if (rc) { + /* + * Error in allocation of event, decrement anchor. Since + * cpu_cf_event in not created, its destroy() function is not + * invoked. Adjust the reference counter for the anchor. + */ + cpum_cf_free_root(); + } +unlock: + mutex_unlock(&pmc_reserve_mutex); + return rc; +} + +/* + * Create/delete per CPU data structures for /dev/hwctr interface and events + * created by perf_event_open(). + * If cpu is -1, track task on all available CPUs. This requires + * allocation of hardware data structures for all CPUs. This setup handles + * perf_event_open() with task context and /dev/hwctr interface. + * If cpu is non-zero install event on this CPU only. This setup handles + * perf_event_open() with CPU context. + */ +static int cpum_cf_alloc(int cpu) +{ + cpumask_var_t mask; + int rc; + + if (cpu == -1) { + if (!zalloc_cpumask_var(&mask, GFP_KERNEL)) + return -ENOMEM; + for_each_online_cpu(cpu) { + rc = cpum_cf_alloc_cpu(cpu); + if (rc) { + for_each_cpu(cpu, mask) + cpum_cf_free_cpu(cpu); + break; + } + cpumask_set_cpu(cpu, mask); + } + free_cpumask_var(mask); + } else { + rc = cpum_cf_alloc_cpu(cpu); + } + return rc; +} + +static void cpum_cf_free(int cpu) +{ + if (cpu == -1) { + for_each_online_cpu(cpu) + cpum_cf_free_cpu(cpu); + } else { + cpum_cf_free_cpu(cpu); + } +} + +#define CF_DIAG_CTRSET_DEF 0xfeef /* Counter set header mark */ + /* interval in seconds */ + +/* Counter sets are stored as data stream in a page sized memory buffer and + * exported to user space via raw data attached to the event sample data. + * Each counter set starts with an eight byte header consisting of: + * - a two byte eye catcher (0xfeef) + * - a one byte counter set number + * - a two byte counter set size (indicates the number of counters in this set) + * - a three byte reserved value (must be zero) to make the header the same + * size as a counter value. + * All counter values are eight byte in size. + * + * All counter sets are followed by a 64 byte trailer. + * The trailer consists of a: + * - flag field indicating valid fields when corresponding bit set + * - the counter facility first and second version number + * - the CPU speed if nonzero + * - the time stamp the counter sets have been collected + * - the time of day (TOD) base value + * - the machine type. + * + * The counter sets are saved when the process is prepared to be executed on a + * CPU and saved again when the process is going to be removed from a CPU. + * The difference of both counter sets are calculated and stored in the event + * sample data area. + */ +struct cf_ctrset_entry { /* CPU-M CF counter set entry (8 byte) */ + unsigned int def:16; /* 0-15 Data Entry Format */ + unsigned int set:16; /* 16-31 Counter set identifier */ + unsigned int ctr:16; /* 32-47 Number of stored counters */ + unsigned int res1:16; /* 48-63 Reserved */ +}; + +struct cf_trailer_entry { /* CPU-M CF_DIAG trailer (64 byte) */ + /* 0 - 7 */ + union { + struct { + unsigned int clock_base:1; /* TOD clock base set */ + unsigned int speed:1; /* CPU speed set */ + /* Measurement alerts */ + unsigned int mtda:1; /* Loss of MT ctr. data alert */ + unsigned int caca:1; /* Counter auth. change alert */ + unsigned int lcda:1; /* Loss of counter data alert */ + }; + unsigned long flags; /* 0-63 All indicators */ + }; + /* 8 - 15 */ + unsigned int cfvn:16; /* 64-79 Ctr First Version */ + unsigned int csvn:16; /* 80-95 Ctr Second Version */ + unsigned int cpu_speed:32; /* 96-127 CPU speed */ + /* 16 - 23 */ + unsigned long timestamp; /* 128-191 Timestamp (TOD) */ + /* 24 - 55 */ + union { + struct { + unsigned long progusage1; + unsigned long progusage2; + unsigned long progusage3; + unsigned long tod_base; + }; + unsigned long progusage[4]; + }; + /* 56 - 63 */ + unsigned int mach_type:16; /* Machine type */ + unsigned int res1:16; /* Reserved */ + unsigned int res2:32; /* Reserved */ +}; + +/* Create the trailer data at the end of a page. */ +static void cfdiag_trailer(struct cf_trailer_entry *te) +{ + struct cpuid cpuid; + + te->cfvn = cpumf_ctr_info.cfvn; /* Counter version numbers */ + te->csvn = cpumf_ctr_info.csvn; + + get_cpu_id(&cpuid); /* Machine type */ + te->mach_type = cpuid.machine; + te->cpu_speed = cfdiag_cpu_speed; + if (te->cpu_speed) + te->speed = 1; + te->clock_base = 1; /* Save clock base */ + te->tod_base = tod_clock_base.tod; + te->timestamp = get_tod_clock_fast(); +} + +/* + * The number of counters per counter set varies between machine generations, + * but is constant when running on a particular machine generation. + * Determine each counter set size at device driver initialization and + * retrieve it later. + */ +static size_t cpumf_ctr_setsizes[CPUMF_CTR_SET_MAX]; +static void cpum_cf_make_setsize(enum cpumf_ctr_set ctrset) +{ + size_t ctrset_size = 0; + + switch (ctrset) { + case CPUMF_CTR_SET_BASIC: + if (cpumf_ctr_info.cfvn >= 1) + ctrset_size = 6; + break; + case CPUMF_CTR_SET_USER: + if (cpumf_ctr_info.cfvn == 1) + ctrset_size = 6; + else if (cpumf_ctr_info.cfvn >= 3) + ctrset_size = 2; + break; + case CPUMF_CTR_SET_CRYPTO: + if (cpumf_ctr_info.csvn >= 1 && cpumf_ctr_info.csvn <= 5) + ctrset_size = 16; + else if (cpumf_ctr_info.csvn == 6 || cpumf_ctr_info.csvn == 7) + ctrset_size = 20; + break; + case CPUMF_CTR_SET_EXT: + if (cpumf_ctr_info.csvn == 1) + ctrset_size = 32; + else if (cpumf_ctr_info.csvn == 2) + ctrset_size = 48; + else if (cpumf_ctr_info.csvn >= 3 && cpumf_ctr_info.csvn <= 5) + ctrset_size = 128; + else if (cpumf_ctr_info.csvn == 6 || cpumf_ctr_info.csvn == 7) + ctrset_size = 160; + break; + case CPUMF_CTR_SET_MT_DIAG: + if (cpumf_ctr_info.csvn > 3) + ctrset_size = 48; + break; + case CPUMF_CTR_SET_MAX: + break; + } + cpumf_ctr_setsizes[ctrset] = ctrset_size; +} + +/* + * Return the maximum possible counter set size (in number of 8 byte counters) + * depending on type and model number. + */ +static size_t cpum_cf_read_setsize(enum cpumf_ctr_set ctrset) +{ + return cpumf_ctr_setsizes[ctrset]; +} + +/* Read a counter set. The counter set number determines the counter set and + * the CPUM-CF first and second version number determine the number of + * available counters in each counter set. + * Each counter set starts with header containing the counter set number and + * the number of eight byte counters. + * + * The functions returns the number of bytes occupied by this counter set + * including the header. + * If there is no counter in the counter set, this counter set is useless and + * zero is returned on this case. + * + * Note that the counter sets may not be enabled or active and the stcctm + * instruction might return error 3. Depending on error_ok value this is ok, + * for example when called from cpumf_pmu_start() call back function. + */ +static size_t cfdiag_getctrset(struct cf_ctrset_entry *ctrdata, int ctrset, + size_t room, bool error_ok) +{ + size_t ctrset_size, need = 0; + int rc = 3; /* Assume write failure */ + + ctrdata->def = CF_DIAG_CTRSET_DEF; + ctrdata->set = ctrset; + ctrdata->res1 = 0; + ctrset_size = cpum_cf_read_setsize(ctrset); + + if (ctrset_size) { /* Save data */ + need = ctrset_size * sizeof(u64) + sizeof(*ctrdata); + if (need <= room) { + rc = ctr_stcctm(ctrset, ctrset_size, + (u64 *)(ctrdata + 1)); + } + if (rc != 3 || error_ok) + ctrdata->ctr = ctrset_size; + else + need = 0; + } + + return need; +} + +static const u64 cpumf_ctr_ctl[CPUMF_CTR_SET_MAX] = { + [CPUMF_CTR_SET_BASIC] = 0x02, + [CPUMF_CTR_SET_USER] = 0x04, + [CPUMF_CTR_SET_CRYPTO] = 0x08, + [CPUMF_CTR_SET_EXT] = 0x01, + [CPUMF_CTR_SET_MT_DIAG] = 0x20, +}; + +/* Read out all counter sets and save them in the provided data buffer. + * The last 64 byte host an artificial trailer entry. + */ +static size_t cfdiag_getctr(void *data, size_t sz, unsigned long auth, + bool error_ok) +{ + struct cf_trailer_entry *trailer; + size_t offset = 0, done; + int i; + + memset(data, 0, sz); + sz -= sizeof(*trailer); /* Always room for trailer */ + for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) { + struct cf_ctrset_entry *ctrdata = data + offset; + + if (!(auth & cpumf_ctr_ctl[i])) + continue; /* Counter set not authorized */ + + done = cfdiag_getctrset(ctrdata, i, sz - offset, error_ok); + offset += done; + } + trailer = data + offset; + cfdiag_trailer(trailer); + return offset + sizeof(*trailer); +} + +/* Calculate the difference for each counter in a counter set. */ +static void cfdiag_diffctrset(u64 *pstart, u64 *pstop, int counters) +{ + for (; --counters >= 0; ++pstart, ++pstop) + if (*pstop >= *pstart) + *pstop -= *pstart; + else + *pstop = *pstart - *pstop + 1; +} + +/* Scan the counter sets and calculate the difference of each counter + * in each set. The result is the increment of each counter during the + * period the counter set has been activated. + * + * Return true on success. + */ +static int cfdiag_diffctr(struct cpu_cf_events *cpuhw, unsigned long auth) +{ + struct cf_trailer_entry *trailer_start, *trailer_stop; + struct cf_ctrset_entry *ctrstart, *ctrstop; + size_t offset = 0; + + auth &= (1 << CPUMF_LCCTL_ENABLE_SHIFT) - 1; + do { + ctrstart = (struct cf_ctrset_entry *)(cpuhw->start + offset); + ctrstop = (struct cf_ctrset_entry *)(cpuhw->stop + offset); + + if (memcmp(ctrstop, ctrstart, sizeof(*ctrstop))) { + pr_err_once("cpum_cf_diag counter set compare error " + "in set %i\n", ctrstart->set); + return 0; + } + auth &= ~cpumf_ctr_ctl[ctrstart->set]; + if (ctrstart->def == CF_DIAG_CTRSET_DEF) { + cfdiag_diffctrset((u64 *)(ctrstart + 1), + (u64 *)(ctrstop + 1), ctrstart->ctr); + offset += ctrstart->ctr * sizeof(u64) + + sizeof(*ctrstart); + } + } while (ctrstart->def && auth); + + /* Save time_stamp from start of event in stop's trailer */ + trailer_start = (struct cf_trailer_entry *)(cpuhw->start + offset); + trailer_stop = (struct cf_trailer_entry *)(cpuhw->stop + offset); + trailer_stop->progusage[0] = trailer_start->timestamp; + + return 1; +} + +static enum cpumf_ctr_set get_counter_set(u64 event) +{ + int set = CPUMF_CTR_SET_MAX; + + if (event < 32) + set = CPUMF_CTR_SET_BASIC; + else if (event < 64) + set = CPUMF_CTR_SET_USER; + else if (event < 128) + set = CPUMF_CTR_SET_CRYPTO; + else if (event < 288) + set = CPUMF_CTR_SET_EXT; + else if (event >= 448 && event < 496) + set = CPUMF_CTR_SET_MT_DIAG; + + return set; +} + +static int validate_ctr_version(const u64 config, enum cpumf_ctr_set set) +{ + u16 mtdiag_ctl; + int err = 0; + + /* check required version for counter sets */ + switch (set) { + case CPUMF_CTR_SET_BASIC: + case CPUMF_CTR_SET_USER: + if (cpumf_ctr_info.cfvn < 1) + err = -EOPNOTSUPP; + break; + case CPUMF_CTR_SET_CRYPTO: + if ((cpumf_ctr_info.csvn >= 1 && cpumf_ctr_info.csvn <= 5 && + config > 79) || (cpumf_ctr_info.csvn >= 6 && config > 83)) + err = -EOPNOTSUPP; + break; + case CPUMF_CTR_SET_EXT: + if (cpumf_ctr_info.csvn < 1) + err = -EOPNOTSUPP; + if ((cpumf_ctr_info.csvn == 1 && config > 159) || + (cpumf_ctr_info.csvn == 2 && config > 175) || + (cpumf_ctr_info.csvn >= 3 && cpumf_ctr_info.csvn <= 5 && + config > 255) || + (cpumf_ctr_info.csvn >= 6 && config > 287)) + err = -EOPNOTSUPP; + break; + case CPUMF_CTR_SET_MT_DIAG: + if (cpumf_ctr_info.csvn <= 3) + err = -EOPNOTSUPP; + /* + * MT-diagnostic counters are read-only. The counter set + * is automatically enabled and activated on all CPUs with + * multithreading (SMT). Deactivation of multithreading + * also disables the counter set. State changes are ignored + * by lcctl(). Because Linux controls SMT enablement through + * a kernel parameter only, the counter set is either disabled + * or enabled and active. + * + * Thus, the counters can only be used if SMT is on and the + * counter set is enabled and active. + */ + mtdiag_ctl = cpumf_ctr_ctl[CPUMF_CTR_SET_MT_DIAG]; + if (!((cpumf_ctr_info.auth_ctl & mtdiag_ctl) && + (cpumf_ctr_info.enable_ctl & mtdiag_ctl) && + (cpumf_ctr_info.act_ctl & mtdiag_ctl))) + err = -EOPNOTSUPP; + break; + case CPUMF_CTR_SET_MAX: + err = -EOPNOTSUPP; + } + + return err; +} + +/* + * Change the CPUMF state to active. + * Enable and activate the CPU-counter sets according + * to the per-cpu control state. + */ +static void cpumf_pmu_enable(struct pmu *pmu) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + int err; + + if (!cpuhw || (cpuhw->flags & PMU_F_ENABLED)) + return; + + err = lcctl(cpuhw->state | cpuhw->dev_state); + if (err) + pr_err("Enabling the performance measuring unit failed with rc=%x\n", err); + else + cpuhw->flags |= PMU_F_ENABLED; +} + +/* + * Change the CPUMF state to inactive. + * Disable and enable (inactive) the CPU-counter sets according + * to the per-cpu control state. + */ +static void cpumf_pmu_disable(struct pmu *pmu) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + u64 inactive; + int err; + + if (!cpuhw || !(cpuhw->flags & PMU_F_ENABLED)) + return; + + inactive = cpuhw->state & ~((1 << CPUMF_LCCTL_ENABLE_SHIFT) - 1); + inactive |= cpuhw->dev_state; + err = lcctl(inactive); + if (err) + pr_err("Disabling the performance measuring unit failed with rc=%x\n", err); + else + cpuhw->flags &= ~PMU_F_ENABLED; +} + +/* Release the PMU if event is the last perf event */ +static void hw_perf_event_destroy(struct perf_event *event) +{ + cpum_cf_free(event->cpu); +} + +/* CPUMF <-> perf event mappings for kernel+userspace (basic set) */ +static const int cpumf_generic_events_basic[] = { + [PERF_COUNT_HW_CPU_CYCLES] = 0, + [PERF_COUNT_HW_INSTRUCTIONS] = 1, + [PERF_COUNT_HW_CACHE_REFERENCES] = -1, + [PERF_COUNT_HW_CACHE_MISSES] = -1, + [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = -1, + [PERF_COUNT_HW_BRANCH_MISSES] = -1, + [PERF_COUNT_HW_BUS_CYCLES] = -1, +}; +/* CPUMF <-> perf event mappings for userspace (problem-state set) */ +static const int cpumf_generic_events_user[] = { + [PERF_COUNT_HW_CPU_CYCLES] = 32, + [PERF_COUNT_HW_INSTRUCTIONS] = 33, + [PERF_COUNT_HW_CACHE_REFERENCES] = -1, + [PERF_COUNT_HW_CACHE_MISSES] = -1, + [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = -1, + [PERF_COUNT_HW_BRANCH_MISSES] = -1, + [PERF_COUNT_HW_BUS_CYCLES] = -1, +}; + +static int is_userspace_event(u64 ev) +{ + return cpumf_generic_events_user[PERF_COUNT_HW_CPU_CYCLES] == ev || + cpumf_generic_events_user[PERF_COUNT_HW_INSTRUCTIONS] == ev; +} + +static int __hw_perf_event_init(struct perf_event *event, unsigned int type) +{ + struct perf_event_attr *attr = &event->attr; + struct hw_perf_event *hwc = &event->hw; + enum cpumf_ctr_set set; + u64 ev; + + switch (type) { + case PERF_TYPE_RAW: + /* Raw events are used to access counters directly, + * hence do not permit excludes */ + if (attr->exclude_kernel || attr->exclude_user || + attr->exclude_hv) + return -EOPNOTSUPP; + ev = attr->config; + break; + + case PERF_TYPE_HARDWARE: + if (is_sampling_event(event)) /* No sampling support */ + return -ENOENT; + ev = attr->config; + if (!attr->exclude_user && attr->exclude_kernel) { + /* + * Count user space (problem-state) only + * Handle events 32 and 33 as 0:u and 1:u + */ + if (!is_userspace_event(ev)) { + if (ev >= ARRAY_SIZE(cpumf_generic_events_user)) + return -EOPNOTSUPP; + ev = cpumf_generic_events_user[ev]; + } + } else if (!attr->exclude_kernel && attr->exclude_user) { + /* No support for kernel space counters only */ + return -EOPNOTSUPP; + } else { + /* Count user and kernel space, incl. events 32 + 33 */ + if (!is_userspace_event(ev)) { + if (ev >= ARRAY_SIZE(cpumf_generic_events_basic)) + return -EOPNOTSUPP; + ev = cpumf_generic_events_basic[ev]; + } + } + break; + + default: + return -ENOENT; + } + + if (ev == -1) + return -ENOENT; + + if (ev > PERF_CPUM_CF_MAX_CTR) + return -ENOENT; + + /* Obtain the counter set to which the specified counter belongs */ + set = get_counter_set(ev); + switch (set) { + case CPUMF_CTR_SET_BASIC: + case CPUMF_CTR_SET_USER: + case CPUMF_CTR_SET_CRYPTO: + case CPUMF_CTR_SET_EXT: + case CPUMF_CTR_SET_MT_DIAG: + /* + * Use the hardware perf event structure to store the + * counter number in the 'config' member and the counter + * set number in the 'config_base' as bit mask. + * It is later used to enable/disable the counter(s). + */ + hwc->config = ev; + hwc->config_base = cpumf_ctr_ctl[set]; + break; + case CPUMF_CTR_SET_MAX: + /* The counter could not be associated to a counter set */ + return -EINVAL; + } + + /* Initialize for using the CPU-measurement counter facility */ + if (cpum_cf_alloc(event->cpu)) + return -ENOMEM; + event->destroy = hw_perf_event_destroy; + + /* + * Finally, validate version and authorization of the counter set. + * If the particular CPU counter set is not authorized, + * return with -ENOENT in order to fall back to other + * PMUs that might suffice the event request. + */ + if (!(hwc->config_base & cpumf_ctr_info.auth_ctl)) + return -ENOENT; + return validate_ctr_version(hwc->config, set); +} + +/* Events CPU_CYLCES and INSTRUCTIONS can be submitted with two different + * attribute::type values: + * - PERF_TYPE_HARDWARE: + * - pmu->type: + * Handle both type of invocations identical. They address the same hardware. + * The result is different when event modifiers exclude_kernel and/or + * exclude_user are also set. + */ +static int cpumf_pmu_event_type(struct perf_event *event) +{ + u64 ev = event->attr.config; + + if (cpumf_generic_events_basic[PERF_COUNT_HW_CPU_CYCLES] == ev || + cpumf_generic_events_basic[PERF_COUNT_HW_INSTRUCTIONS] == ev || + cpumf_generic_events_user[PERF_COUNT_HW_CPU_CYCLES] == ev || + cpumf_generic_events_user[PERF_COUNT_HW_INSTRUCTIONS] == ev) + return PERF_TYPE_HARDWARE; + return PERF_TYPE_RAW; +} + +static int cpumf_pmu_event_init(struct perf_event *event) +{ + unsigned int type = event->attr.type; + int err; + + if (type == PERF_TYPE_HARDWARE || type == PERF_TYPE_RAW) + err = __hw_perf_event_init(event, type); + else if (event->pmu->type == type) + /* Registered as unknown PMU */ + err = __hw_perf_event_init(event, cpumf_pmu_event_type(event)); + else + return -ENOENT; + + if (unlikely(err) && event->destroy) + event->destroy(event); + + return err; +} + +static int hw_perf_event_reset(struct perf_event *event) +{ + u64 prev, new; + int err; + + do { + prev = local64_read(&event->hw.prev_count); + err = ecctr(event->hw.config, &new); + if (err) { + if (err != 3) + break; + /* The counter is not (yet) available. This + * might happen if the counter set to which + * this counter belongs is in the disabled + * state. + */ + new = 0; + } + } while (local64_cmpxchg(&event->hw.prev_count, prev, new) != prev); + + return err; +} + +static void hw_perf_event_update(struct perf_event *event) +{ + u64 prev, new, delta; + int err; + + do { + prev = local64_read(&event->hw.prev_count); + err = ecctr(event->hw.config, &new); + if (err) + return; + } while (local64_cmpxchg(&event->hw.prev_count, prev, new) != prev); + + delta = (prev <= new) ? new - prev + : (-1ULL - prev) + new + 1; /* overflow */ + local64_add(delta, &event->count); +} + +static void cpumf_pmu_read(struct perf_event *event) +{ + if (event->hw.state & PERF_HES_STOPPED) + return; + + hw_perf_event_update(event); +} + +static void cpumf_pmu_start(struct perf_event *event, int flags) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + struct hw_perf_event *hwc = &event->hw; + int i; + + if (!(hwc->state & PERF_HES_STOPPED)) + return; + + hwc->state = 0; + + /* (Re-)enable and activate the counter set */ + ctr_set_enable(&cpuhw->state, hwc->config_base); + ctr_set_start(&cpuhw->state, hwc->config_base); + + /* The counter set to which this counter belongs can be already active. + * Because all counters in a set are active, the event->hw.prev_count + * needs to be synchronized. At this point, the counter set can be in + * the inactive or disabled state. + */ + if (hwc->config == PERF_EVENT_CPUM_CF_DIAG) { + cpuhw->usedss = cfdiag_getctr(cpuhw->start, + sizeof(cpuhw->start), + hwc->config_base, true); + } else { + hw_perf_event_reset(event); + } + + /* Increment refcount for counter sets */ + for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) + if ((hwc->config_base & cpumf_ctr_ctl[i])) + atomic_inc(&cpuhw->ctr_set[i]); +} + +/* Create perf event sample with the counter sets as raw data. The sample + * is then pushed to the event subsystem and the function checks for + * possible event overflows. If an event overflow occurs, the PMU is + * stopped. + * + * Return non-zero if an event overflow occurred. + */ +static int cfdiag_push_sample(struct perf_event *event, + struct cpu_cf_events *cpuhw) +{ + struct perf_sample_data data; + struct perf_raw_record raw; + struct pt_regs regs; + int overflow; + + /* Setup perf sample */ + perf_sample_data_init(&data, 0, event->hw.last_period); + memset(®s, 0, sizeof(regs)); + memset(&raw, 0, sizeof(raw)); + + if (event->attr.sample_type & PERF_SAMPLE_CPU) + data.cpu_entry.cpu = event->cpu; + if (event->attr.sample_type & PERF_SAMPLE_RAW) { + raw.frag.size = cpuhw->usedss; + raw.frag.data = cpuhw->stop; + perf_sample_save_raw_data(&data, &raw); + } + + overflow = perf_event_overflow(event, &data, ®s); + if (overflow) + event->pmu->stop(event, 0); + + perf_event_update_userpage(event); + return overflow; +} + +static void cpumf_pmu_stop(struct perf_event *event, int flags) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + struct hw_perf_event *hwc = &event->hw; + int i; + + if (!(hwc->state & PERF_HES_STOPPED)) { + /* Decrement reference count for this counter set and if this + * is the last used counter in the set, clear activation + * control and set the counter set state to inactive. + */ + for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) { + if (!(hwc->config_base & cpumf_ctr_ctl[i])) + continue; + if (!atomic_dec_return(&cpuhw->ctr_set[i])) + ctr_set_stop(&cpuhw->state, cpumf_ctr_ctl[i]); + } + hwc->state |= PERF_HES_STOPPED; + } + + if ((flags & PERF_EF_UPDATE) && !(hwc->state & PERF_HES_UPTODATE)) { + if (hwc->config == PERF_EVENT_CPUM_CF_DIAG) { + local64_inc(&event->count); + cpuhw->usedss = cfdiag_getctr(cpuhw->stop, + sizeof(cpuhw->stop), + event->hw.config_base, + false); + if (cfdiag_diffctr(cpuhw, event->hw.config_base)) + cfdiag_push_sample(event, cpuhw); + } else { + hw_perf_event_update(event); + } + hwc->state |= PERF_HES_UPTODATE; + } +} + +static int cpumf_pmu_add(struct perf_event *event, int flags) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + + ctr_set_enable(&cpuhw->state, event->hw.config_base); + event->hw.state = PERF_HES_UPTODATE | PERF_HES_STOPPED; + + if (flags & PERF_EF_START) + cpumf_pmu_start(event, PERF_EF_RELOAD); + + return 0; +} + +static void cpumf_pmu_del(struct perf_event *event, int flags) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + int i; + + cpumf_pmu_stop(event, PERF_EF_UPDATE); + + /* Check if any counter in the counter set is still used. If not used, + * change the counter set to the disabled state. This also clears the + * content of all counters in the set. + * + * When a new perf event has been added but not yet started, this can + * clear enable control and resets all counters in a set. Therefore, + * cpumf_pmu_start() always has to reenable a counter set. + */ + for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) + if (!atomic_read(&cpuhw->ctr_set[i])) + ctr_set_disable(&cpuhw->state, cpumf_ctr_ctl[i]); +} + +/* Performance monitoring unit for s390x */ +static struct pmu cpumf_pmu = { + .task_ctx_nr = perf_sw_context, + .capabilities = PERF_PMU_CAP_NO_INTERRUPT, + .pmu_enable = cpumf_pmu_enable, + .pmu_disable = cpumf_pmu_disable, + .event_init = cpumf_pmu_event_init, + .add = cpumf_pmu_add, + .del = cpumf_pmu_del, + .start = cpumf_pmu_start, + .stop = cpumf_pmu_stop, + .read = cpumf_pmu_read, +}; + +static struct cfset_session { /* CPUs and counter set bit mask */ + struct list_head head; /* Head of list of active processes */ +} cfset_session = { + .head = LIST_HEAD_INIT(cfset_session.head) +}; + +static refcount_t cfset_opencnt = REFCOUNT_INIT(0); /* Access count */ +/* + * Synchronize access to device /dev/hwc. This mutex protects against + * concurrent access to functions cfset_open() and cfset_release(). + * Same for CPU hotplug add and remove events triggering + * cpum_cf_online_cpu() and cpum_cf_offline_cpu(). + * It also serializes concurrent device ioctl access from multiple + * processes accessing /dev/hwc. + * + * The mutex protects concurrent access to the /dev/hwctr session management + * struct cfset_session and reference counting variable cfset_opencnt. + */ +static DEFINE_MUTEX(cfset_ctrset_mutex); + +/* + * CPU hotplug handles only /dev/hwctr device. + * For perf_event_open() the CPU hotplug handling is done on kernel common + * code: + * - CPU add: Nothing is done since a file descriptor can not be created + * and returned to the user. + * - CPU delete: Handled by common code via pmu_disable(), pmu_stop() and + * pmu_delete(). The event itself is removed when the file descriptor is + * closed. + */ +static int cfset_online_cpu(unsigned int cpu); + +static int cpum_cf_online_cpu(unsigned int cpu) +{ + int rc = 0; + + /* + * Ignore notification for perf_event_open(). + * Handle only /dev/hwctr device sessions. + */ + mutex_lock(&cfset_ctrset_mutex); + if (refcount_read(&cfset_opencnt)) { + rc = cpum_cf_alloc_cpu(cpu); + if (!rc) + cfset_online_cpu(cpu); + } + mutex_unlock(&cfset_ctrset_mutex); + return rc; +} + +static int cfset_offline_cpu(unsigned int cpu); + +static int cpum_cf_offline_cpu(unsigned int cpu) +{ + /* + * During task exit processing of grouped perf events triggered by CPU + * hotplug processing, pmu_disable() is called as part of perf context + * removal process. Therefore do not trigger event removal now for + * perf_event_open() created events. Perf common code triggers event + * destruction when the event file descriptor is closed. + * + * Handle only /dev/hwctr device sessions. + */ + mutex_lock(&cfset_ctrset_mutex); + if (refcount_read(&cfset_opencnt)) { + cfset_offline_cpu(cpu); + cpum_cf_free_cpu(cpu); + } + mutex_unlock(&cfset_ctrset_mutex); + return 0; +} + +/* Return true if store counter set multiple instruction is available */ +static inline int stccm_avail(void) +{ + return test_facility(142); +} + +/* CPU-measurement alerts for the counter facility */ +static void cpumf_measurement_alert(struct ext_code ext_code, + unsigned int alert, unsigned long unused) +{ + struct cpu_cf_events *cpuhw; + + if (!(alert & CPU_MF_INT_CF_MASK)) + return; + + inc_irq_stat(IRQEXT_CMC); + + /* + * Measurement alerts are shared and might happen when the PMU + * is not reserved. Ignore these alerts in this case. + */ + cpuhw = this_cpu_cfhw(); + if (!cpuhw) + return; + + /* counter authorization change alert */ + if (alert & CPU_MF_INT_CF_CACA) + qctri(&cpumf_ctr_info); + + /* loss of counter data alert */ + if (alert & CPU_MF_INT_CF_LCDA) + pr_err("CPU[%i] Counter data was lost\n", smp_processor_id()); + + /* loss of MT counter data alert */ + if (alert & CPU_MF_INT_CF_MTDA) + pr_warn("CPU[%i] MT counter data was lost\n", + smp_processor_id()); +} + +static int cfset_init(void); +static int __init cpumf_pmu_init(void) +{ + int rc; + + /* Extract counter measurement facility information */ + if (!cpum_cf_avail() || qctri(&cpumf_ctr_info)) + return -ENODEV; + + /* Determine and store counter set sizes for later reference */ + for (rc = CPUMF_CTR_SET_BASIC; rc < CPUMF_CTR_SET_MAX; ++rc) + cpum_cf_make_setsize(rc); + + /* + * Clear bit 15 of cr0 to unauthorize problem-state to + * extract measurement counters + */ + ctl_clear_bit(0, 48); + + /* register handler for measurement-alert interruptions */ + rc = register_external_irq(EXT_IRQ_MEASURE_ALERT, + cpumf_measurement_alert); + if (rc) { + pr_err("Registering for CPU-measurement alerts failed with rc=%i\n", rc); + return rc; + } + + /* Setup s390dbf facility */ + cf_dbg = debug_register(KMSG_COMPONENT, 2, 1, 128); + if (!cf_dbg) { + pr_err("Registration of s390dbf(cpum_cf) failed\n"); + rc = -ENOMEM; + goto out1; + } + debug_register_view(cf_dbg, &debug_sprintf_view); + + cpumf_pmu.attr_groups = cpumf_cf_event_group(); + rc = perf_pmu_register(&cpumf_pmu, "cpum_cf", -1); + if (rc) { + pr_err("Registering the cpum_cf PMU failed with rc=%i\n", rc); + goto out2; + } else if (stccm_avail()) { /* Setup counter set device */ + cfset_init(); + } + + rc = cpuhp_setup_state(CPUHP_AP_PERF_S390_CF_ONLINE, + "perf/s390/cf:online", + cpum_cf_online_cpu, cpum_cf_offline_cpu); + return rc; + +out2: + debug_unregister_view(cf_dbg, &debug_sprintf_view); + debug_unregister(cf_dbg); +out1: + unregister_external_irq(EXT_IRQ_MEASURE_ALERT, cpumf_measurement_alert); + return rc; +} + +/* Support for the CPU Measurement Facility counter set extraction using + * device /dev/hwctr. This allows user space programs to extract complete + * counter set via normal file operations. + */ + +struct cfset_call_on_cpu_parm { /* Parm struct for smp_call_on_cpu */ + unsigned int sets; /* Counter set bit mask */ + atomic_t cpus_ack; /* # CPUs successfully executed func */ +}; + +struct cfset_request { /* CPUs and counter set bit mask */ + unsigned long ctrset; /* Bit mask of counter set to read */ + cpumask_t mask; /* CPU mask to read from */ + struct list_head node; /* Chain to cfset_session.head */ +}; + +static void cfset_session_init(void) +{ + INIT_LIST_HEAD(&cfset_session.head); +} + +/* Remove current request from global bookkeeping. Maintain a counter set bit + * mask on a per CPU basis. + * Done in process context under mutex protection. + */ +static void cfset_session_del(struct cfset_request *p) +{ + list_del(&p->node); +} + +/* Add current request to global bookkeeping. Maintain a counter set bit mask + * on a per CPU basis. + * Done in process context under mutex protection. + */ +static void cfset_session_add(struct cfset_request *p) +{ + list_add(&p->node, &cfset_session.head); +} + +/* The /dev/hwctr device access uses PMU_F_IN_USE to mark the device access + * path is currently used. + * The cpu_cf_events::dev_state is used to denote counter sets in use by this + * interface. It is always or'ed in. If this interface is not active, its + * value is zero and no additional counter sets will be included. + * + * The cpu_cf_events::state is used by the perf_event_open SVC and remains + * unchanged. + * + * perf_pmu_enable() and perf_pmu_enable() and its call backs + * cpumf_pmu_enable() and cpumf_pmu_disable() are called by the + * performance measurement subsystem to enable per process + * CPU Measurement counter facility. + * The XXX_enable() and XXX_disable functions are used to turn off + * x86 performance monitoring interrupt (PMI) during scheduling. + * s390 uses these calls to temporarily stop and resume the active CPU + * counters sets during scheduling. + * + * We do allow concurrent access of perf_event_open() SVC and /dev/hwctr + * device access. The perf_event_open() SVC interface makes a lot of effort + * to only run the counters while the calling process is actively scheduled + * to run. + * When /dev/hwctr interface is also used at the same time, the counter sets + * will keep running, even when the process is scheduled off a CPU. + * However this is not a problem and does not lead to wrong counter values + * for the perf_event_open() SVC. The current counter value will be recorded + * during schedule-in. At schedule-out time the current counter value is + * extracted again and the delta is calculated and added to the event. + */ +/* Stop all counter sets via ioctl interface */ +static void cfset_ioctl_off(void *parm) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + struct cfset_call_on_cpu_parm *p = parm; + int rc; + + /* Check if any counter set used by /dev/hwctr */ + for (rc = CPUMF_CTR_SET_BASIC; rc < CPUMF_CTR_SET_MAX; ++rc) + if ((p->sets & cpumf_ctr_ctl[rc])) { + if (!atomic_dec_return(&cpuhw->ctr_set[rc])) { + ctr_set_disable(&cpuhw->dev_state, + cpumf_ctr_ctl[rc]); + ctr_set_stop(&cpuhw->dev_state, + cpumf_ctr_ctl[rc]); + } + } + /* Keep perf_event_open counter sets */ + rc = lcctl(cpuhw->dev_state | cpuhw->state); + if (rc) + pr_err("Counter set stop %#llx of /dev/%s failed rc=%i\n", + cpuhw->state, S390_HWCTR_DEVICE, rc); + if (!cpuhw->dev_state) + cpuhw->flags &= ~PMU_F_IN_USE; +} + +/* Start counter sets on particular CPU */ +static void cfset_ioctl_on(void *parm) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + struct cfset_call_on_cpu_parm *p = parm; + int rc; + + cpuhw->flags |= PMU_F_IN_USE; + ctr_set_enable(&cpuhw->dev_state, p->sets); + ctr_set_start(&cpuhw->dev_state, p->sets); + for (rc = CPUMF_CTR_SET_BASIC; rc < CPUMF_CTR_SET_MAX; ++rc) + if ((p->sets & cpumf_ctr_ctl[rc])) + atomic_inc(&cpuhw->ctr_set[rc]); + rc = lcctl(cpuhw->dev_state | cpuhw->state); /* Start counter sets */ + if (!rc) + atomic_inc(&p->cpus_ack); + else + pr_err("Counter set start %#llx of /dev/%s failed rc=%i\n", + cpuhw->dev_state | cpuhw->state, S390_HWCTR_DEVICE, rc); +} + +static void cfset_release_cpu(void *p) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + int rc; + + cpuhw->dev_state = 0; + rc = lcctl(cpuhw->state); /* Keep perf_event_open counter sets */ + if (rc) + pr_err("Counter set release %#llx of /dev/%s failed rc=%i\n", + cpuhw->state, S390_HWCTR_DEVICE, rc); +} + +/* This modifies the process CPU mask to adopt it to the currently online + * CPUs. Offline CPUs can not be addresses. This call terminates the access + * and is usually followed by close() or a new iotcl(..., START, ...) which + * creates a new request structure. + */ +static void cfset_all_stop(struct cfset_request *req) +{ + struct cfset_call_on_cpu_parm p = { + .sets = req->ctrset, + }; + + cpumask_and(&req->mask, &req->mask, cpu_online_mask); + on_each_cpu_mask(&req->mask, cfset_ioctl_off, &p, 1); +} + +/* Release function is also called when application gets terminated without + * doing a proper ioctl(..., S390_HWCTR_STOP, ...) command. + */ +static int cfset_release(struct inode *inode, struct file *file) +{ + mutex_lock(&cfset_ctrset_mutex); + /* Open followed by close/exit has no private_data */ + if (file->private_data) { + cfset_all_stop(file->private_data); + cfset_session_del(file->private_data); + kfree(file->private_data); + file->private_data = NULL; + } + if (refcount_dec_and_test(&cfset_opencnt)) { /* Last close */ + on_each_cpu(cfset_release_cpu, NULL, 1); + cpum_cf_free(-1); + } + mutex_unlock(&cfset_ctrset_mutex); + return 0; +} + +/* + * Open via /dev/hwctr device. Allocate all per CPU resources on the first + * open of the device. The last close releases all per CPU resources. + * Parallel perf_event_open system calls also use per CPU resources. + * These invocations are handled via reference counting on the per CPU data + * structures. + */ +static int cfset_open(struct inode *inode, struct file *file) +{ + int rc = 0; + + if (!perfmon_capable()) + return -EPERM; + file->private_data = NULL; + + mutex_lock(&cfset_ctrset_mutex); + if (!refcount_inc_not_zero(&cfset_opencnt)) { /* First open */ + rc = cpum_cf_alloc(-1); + if (!rc) { + cfset_session_init(); + refcount_set(&cfset_opencnt, 1); + } + } + mutex_unlock(&cfset_ctrset_mutex); + + /* nonseekable_open() never fails */ + return rc ?: nonseekable_open(inode, file); +} + +static int cfset_all_start(struct cfset_request *req) +{ + struct cfset_call_on_cpu_parm p = { + .sets = req->ctrset, + .cpus_ack = ATOMIC_INIT(0), + }; + cpumask_var_t mask; + int rc = 0; + + if (!alloc_cpumask_var(&mask, GFP_KERNEL)) + return -ENOMEM; + cpumask_and(mask, &req->mask, cpu_online_mask); + on_each_cpu_mask(mask, cfset_ioctl_on, &p, 1); + if (atomic_read(&p.cpus_ack) != cpumask_weight(mask)) { + on_each_cpu_mask(mask, cfset_ioctl_off, &p, 1); + rc = -EIO; + } + free_cpumask_var(mask); + return rc; +} + +/* Return the maximum required space for all possible CPUs in case one + * CPU will be onlined during the START, READ, STOP cycles. + * To find out the size of the counter sets, any one CPU will do. They + * all have the same counter sets. + */ +static size_t cfset_needspace(unsigned int sets) +{ + size_t bytes = 0; + int i; + + for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) { + if (!(sets & cpumf_ctr_ctl[i])) + continue; + bytes += cpum_cf_read_setsize(i) * sizeof(u64) + + sizeof(((struct s390_ctrset_setdata *)0)->set) + + sizeof(((struct s390_ctrset_setdata *)0)->no_cnts); + } + bytes = sizeof(((struct s390_ctrset_read *)0)->no_cpus) + nr_cpu_ids * + (bytes + sizeof(((struct s390_ctrset_cpudata *)0)->cpu_nr) + + sizeof(((struct s390_ctrset_cpudata *)0)->no_sets)); + return bytes; +} + +static int cfset_all_copy(unsigned long arg, cpumask_t *mask) +{ + struct s390_ctrset_read __user *ctrset_read; + unsigned int cpu, cpus, rc = 0; + void __user *uptr; + + ctrset_read = (struct s390_ctrset_read __user *)arg; + uptr = ctrset_read->data; + for_each_cpu(cpu, mask) { + struct cpu_cf_events *cpuhw = get_cpu_cfhw(cpu); + struct s390_ctrset_cpudata __user *ctrset_cpudata; + + ctrset_cpudata = uptr; + rc = put_user(cpu, &ctrset_cpudata->cpu_nr); + rc |= put_user(cpuhw->sets, &ctrset_cpudata->no_sets); + rc |= copy_to_user(ctrset_cpudata->data, cpuhw->data, + cpuhw->used); + if (rc) { + rc = -EFAULT; + goto out; + } + uptr += sizeof(struct s390_ctrset_cpudata) + cpuhw->used; + cond_resched(); + } + cpus = cpumask_weight(mask); + if (put_user(cpus, &ctrset_read->no_cpus)) + rc = -EFAULT; +out: + return rc; +} + +static size_t cfset_cpuset_read(struct s390_ctrset_setdata *p, int ctrset, + int ctrset_size, size_t room) +{ + size_t need = 0; + int rc = -1; + + need = sizeof(*p) + sizeof(u64) * ctrset_size; + if (need <= room) { + p->set = cpumf_ctr_ctl[ctrset]; + p->no_cnts = ctrset_size; + rc = ctr_stcctm(ctrset, ctrset_size, (u64 *)p->cv); + if (rc == 3) /* Nothing stored */ + need = 0; + } + return need; +} + +/* Read all counter sets. */ +static void cfset_cpu_read(void *parm) +{ + struct cpu_cf_events *cpuhw = this_cpu_cfhw(); + struct cfset_call_on_cpu_parm *p = parm; + int set, set_size; + size_t space; + + /* No data saved yet */ + cpuhw->used = 0; + cpuhw->sets = 0; + memset(cpuhw->data, 0, sizeof(cpuhw->data)); + + /* Scan the counter sets */ + for (set = CPUMF_CTR_SET_BASIC; set < CPUMF_CTR_SET_MAX; ++set) { + struct s390_ctrset_setdata *sp = (void *)cpuhw->data + + cpuhw->used; + + if (!(p->sets & cpumf_ctr_ctl[set])) + continue; /* Counter set not in list */ + set_size = cpum_cf_read_setsize(set); + space = sizeof(cpuhw->data) - cpuhw->used; + space = cfset_cpuset_read(sp, set, set_size, space); + if (space) { + cpuhw->used += space; + cpuhw->sets += 1; + } + } +} + +static int cfset_all_read(unsigned long arg, struct cfset_request *req) +{ + struct cfset_call_on_cpu_parm p; + cpumask_var_t mask; + int rc; + + if (!alloc_cpumask_var(&mask, GFP_KERNEL)) + return -ENOMEM; + + p.sets = req->ctrset; + cpumask_and(mask, &req->mask, cpu_online_mask); + on_each_cpu_mask(mask, cfset_cpu_read, &p, 1); + rc = cfset_all_copy(arg, mask); + free_cpumask_var(mask); + return rc; +} + +static long cfset_ioctl_read(unsigned long arg, struct cfset_request *req) +{ + int ret = -ENODATA; + + if (req && req->ctrset) + ret = cfset_all_read(arg, req); + return ret; +} + +static long cfset_ioctl_stop(struct file *file) +{ + struct cfset_request *req = file->private_data; + int ret = -ENXIO; + + if (req) { + cfset_all_stop(req); + cfset_session_del(req); + kfree(req); + file->private_data = NULL; + ret = 0; + } + return ret; +} + +static long cfset_ioctl_start(unsigned long arg, struct file *file) +{ + struct s390_ctrset_start __user *ustart; + struct s390_ctrset_start start; + struct cfset_request *preq; + void __user *umask; + unsigned int len; + int ret = 0; + size_t need; + + if (file->private_data) + return -EBUSY; + ustart = (struct s390_ctrset_start __user *)arg; + if (copy_from_user(&start, ustart, sizeof(start))) + return -EFAULT; + if (start.version != S390_HWCTR_START_VERSION) + return -EINVAL; + if (start.counter_sets & ~(cpumf_ctr_ctl[CPUMF_CTR_SET_BASIC] | + cpumf_ctr_ctl[CPUMF_CTR_SET_USER] | + cpumf_ctr_ctl[CPUMF_CTR_SET_CRYPTO] | + cpumf_ctr_ctl[CPUMF_CTR_SET_EXT] | + cpumf_ctr_ctl[CPUMF_CTR_SET_MT_DIAG])) + return -EINVAL; /* Invalid counter set */ + if (!start.counter_sets) + return -EINVAL; /* No counter set at all? */ + + preq = kzalloc(sizeof(*preq), GFP_KERNEL); + if (!preq) + return -ENOMEM; + cpumask_clear(&preq->mask); + len = min_t(u64, start.cpumask_len, cpumask_size()); + umask = (void __user *)start.cpumask; + if (copy_from_user(&preq->mask, umask, len)) { + kfree(preq); + return -EFAULT; + } + if (cpumask_empty(&preq->mask)) { + kfree(preq); + return -EINVAL; + } + need = cfset_needspace(start.counter_sets); + if (put_user(need, &ustart->data_bytes)) { + kfree(preq); + return -EFAULT; + } + preq->ctrset = start.counter_sets; + ret = cfset_all_start(preq); + if (!ret) { + cfset_session_add(preq); + file->private_data = preq; + } else { + kfree(preq); + } + return ret; +} + +/* Entry point to the /dev/hwctr device interface. + * The ioctl system call supports three subcommands: + * S390_HWCTR_START: Start the specified counter sets on a CPU list. The + * counter set keeps running until explicitly stopped. Returns the number + * of bytes needed to store the counter values. If another S390_HWCTR_START + * ioctl subcommand is called without a previous S390_HWCTR_STOP stop + * command on the same file descriptor, -EBUSY is returned. + * S390_HWCTR_READ: Read the counter set values from specified CPU list given + * with the S390_HWCTR_START command. + * S390_HWCTR_STOP: Stops the counter sets on the CPU list given with the + * previous S390_HWCTR_START subcommand. + */ +static long cfset_ioctl(struct file *file, unsigned int cmd, unsigned long arg) +{ + int ret; + + cpus_read_lock(); + mutex_lock(&cfset_ctrset_mutex); + switch (cmd) { + case S390_HWCTR_START: + ret = cfset_ioctl_start(arg, file); + break; + case S390_HWCTR_STOP: + ret = cfset_ioctl_stop(file); + break; + case S390_HWCTR_READ: + ret = cfset_ioctl_read(arg, file->private_data); + break; + default: + ret = -ENOTTY; + break; + } + mutex_unlock(&cfset_ctrset_mutex); + cpus_read_unlock(); + return ret; +} + +static const struct file_operations cfset_fops = { + .owner = THIS_MODULE, + .open = cfset_open, + .release = cfset_release, + .unlocked_ioctl = cfset_ioctl, + .compat_ioctl = cfset_ioctl, + .llseek = no_llseek +}; + +static struct miscdevice cfset_dev = { + .name = S390_HWCTR_DEVICE, + .minor = MISC_DYNAMIC_MINOR, + .fops = &cfset_fops, + .mode = 0666, +}; + +/* Hotplug add of a CPU. Scan through all active processes and add + * that CPU to the list of CPUs supplied with ioctl(..., START, ...). + */ +static int cfset_online_cpu(unsigned int cpu) +{ + struct cfset_call_on_cpu_parm p; + struct cfset_request *rp; + + if (!list_empty(&cfset_session.head)) { + list_for_each_entry(rp, &cfset_session.head, node) { + p.sets = rp->ctrset; + cfset_ioctl_on(&p); + cpumask_set_cpu(cpu, &rp->mask); + } + } + return 0; +} + +/* Hotplug remove of a CPU. Scan through all active processes and clear + * that CPU from the list of CPUs supplied with ioctl(..., START, ...). + * Adjust reference counts. + */ +static int cfset_offline_cpu(unsigned int cpu) +{ + struct cfset_call_on_cpu_parm p; + struct cfset_request *rp; + + if (!list_empty(&cfset_session.head)) { + list_for_each_entry(rp, &cfset_session.head, node) { + p.sets = rp->ctrset; + cfset_ioctl_off(&p); + cpumask_clear_cpu(cpu, &rp->mask); + } + } + return 0; +} + +static void cfdiag_read(struct perf_event *event) +{ +} + +static int get_authctrsets(void) +{ + unsigned long auth = 0; + enum cpumf_ctr_set i; + + for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) { + if (cpumf_ctr_info.auth_ctl & cpumf_ctr_ctl[i]) + auth |= cpumf_ctr_ctl[i]; + } + return auth; +} + +/* Setup the event. Test for authorized counter sets and only include counter + * sets which are authorized at the time of the setup. Including unauthorized + * counter sets result in specification exception (and panic). + */ +static int cfdiag_event_init2(struct perf_event *event) +{ + struct perf_event_attr *attr = &event->attr; + int err = 0; + + /* Set sample_period to indicate sampling */ + event->hw.config = attr->config; + event->hw.sample_period = attr->sample_period; + local64_set(&event->hw.period_left, event->hw.sample_period); + local64_set(&event->count, 0); + event->hw.last_period = event->hw.sample_period; + + /* Add all authorized counter sets to config_base. The + * the hardware init function is either called per-cpu or just once + * for all CPUS (event->cpu == -1). This depends on the whether + * counting is started for all CPUs or on a per workload base where + * the perf event moves from one CPU to another CPU. + * Checking the authorization on any CPU is fine as the hardware + * applies the same authorization settings to all CPUs. + */ + event->hw.config_base = get_authctrsets(); + + /* No authorized counter sets, nothing to count/sample */ + if (!event->hw.config_base) + err = -EINVAL; + + return err; +} + +static int cfdiag_event_init(struct perf_event *event) +{ + struct perf_event_attr *attr = &event->attr; + int err = -ENOENT; + + if (event->attr.config != PERF_EVENT_CPUM_CF_DIAG || + event->attr.type != event->pmu->type) + goto out; + + /* Raw events are used to access counters directly, + * hence do not permit excludes. + * This event is useless without PERF_SAMPLE_RAW to return counter set + * values as raw data. + */ + if (attr->exclude_kernel || attr->exclude_user || attr->exclude_hv || + !(attr->sample_type & (PERF_SAMPLE_CPU | PERF_SAMPLE_RAW))) { + err = -EOPNOTSUPP; + goto out; + } + + /* Initialize for using the CPU-measurement counter facility */ + if (cpum_cf_alloc(event->cpu)) + return -ENOMEM; + event->destroy = hw_perf_event_destroy; + + err = cfdiag_event_init2(event); + if (unlikely(err)) + event->destroy(event); +out: + return err; +} + +/* Create cf_diag/events/CF_DIAG event sysfs file. This counter is used + * to collect the complete counter sets for a scheduled process. Target + * are complete counter sets attached as raw data to the artificial event. + * This results in complete counter sets available when a process is + * scheduled. Contains the delta of every counter while the process was + * running. + */ +CPUMF_EVENT_ATTR(CF_DIAG, CF_DIAG, PERF_EVENT_CPUM_CF_DIAG); + +static struct attribute *cfdiag_events_attr[] = { + CPUMF_EVENT_PTR(CF_DIAG, CF_DIAG), + NULL, +}; + +PMU_FORMAT_ATTR(event, "config:0-63"); + +static struct attribute *cfdiag_format_attr[] = { + &format_attr_event.attr, + NULL, +}; + +static struct attribute_group cfdiag_events_group = { + .name = "events", + .attrs = cfdiag_events_attr, +}; +static struct attribute_group cfdiag_format_group = { + .name = "format", + .attrs = cfdiag_format_attr, +}; +static const struct attribute_group *cfdiag_attr_groups[] = { + &cfdiag_events_group, + &cfdiag_format_group, + NULL, +}; + +/* Performance monitoring unit for event CF_DIAG. Since this event + * is also started and stopped via the perf_event_open() system call, use + * the same event enable/disable call back functions. They do not + * have a pointer to the perf_event strcture as first parameter. + * + * The functions XXX_add, XXX_del, XXX_start and XXX_stop are also common. + * Reuse them and distinguish the event (always first parameter) via + * 'config' member. + */ +static struct pmu cf_diag = { + .task_ctx_nr = perf_sw_context, + .event_init = cfdiag_event_init, + .pmu_enable = cpumf_pmu_enable, + .pmu_disable = cpumf_pmu_disable, + .add = cpumf_pmu_add, + .del = cpumf_pmu_del, + .start = cpumf_pmu_start, + .stop = cpumf_pmu_stop, + .read = cfdiag_read, + + .attr_groups = cfdiag_attr_groups +}; + +/* Calculate memory needed to store all counter sets together with header and + * trailer data. This is independent of the counter set authorization which + * can vary depending on the configuration. + */ +static size_t cfdiag_maxsize(struct cpumf_ctr_info *info) +{ + size_t max_size = sizeof(struct cf_trailer_entry); + enum cpumf_ctr_set i; + + for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) { + size_t size = cpum_cf_read_setsize(i); + + if (size) + max_size += size * sizeof(u64) + + sizeof(struct cf_ctrset_entry); + } + return max_size; +} + +/* Get the CPU speed, try sampling facility first and CPU attributes second. */ +static void cfdiag_get_cpu_speed(void) +{ + unsigned long mhz; + + if (cpum_sf_avail()) { /* Sampling facility first */ + struct hws_qsi_info_block si; + + memset(&si, 0, sizeof(si)); + if (!qsi(&si)) { + cfdiag_cpu_speed = si.cpu_speed; + return; + } + } + + /* Fallback: CPU speed extract static part. Used in case + * CPU Measurement Sampling Facility is turned off. + */ + mhz = __ecag(ECAG_CPU_ATTRIBUTE, 0); + if (mhz != -1UL) + cfdiag_cpu_speed = mhz & 0xffffffff; +} + +static int cfset_init(void) +{ + size_t need; + int rc; + + cfdiag_get_cpu_speed(); + /* Make sure the counter set data fits into predefined buffer. */ + need = cfdiag_maxsize(&cpumf_ctr_info); + if (need > sizeof(((struct cpu_cf_events *)0)->start)) { + pr_err("Insufficient memory for PMU(cpum_cf_diag) need=%zu\n", + need); + return -ENOMEM; + } + + rc = misc_register(&cfset_dev); + if (rc) { + pr_err("Registration of /dev/%s failed rc=%i\n", + cfset_dev.name, rc); + goto out; + } + + rc = perf_pmu_register(&cf_diag, "cpum_cf_diag", -1); + if (rc) { + misc_deregister(&cfset_dev); + pr_err("Registration of PMU(cpum_cf_diag) failed with rc=%i\n", + rc); + } +out: + return rc; +} + +device_initcall(cpumf_pmu_init); diff --git a/arch/s390/kernel/perf_cpum_cf_events.c b/arch/s390/kernel/perf_cpum_cf_events.c new file mode 100644 index 0000000000..0d64aafd15 --- /dev/null +++ b/arch/s390/kernel/perf_cpum_cf_events.c @@ -0,0 +1,909 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Perf PMU sysfs events attributes for available CPU-measurement counters + * + */ + +#include <linux/slab.h> +#include <linux/perf_event.h> +#include <asm/cpu_mf.h> + + +/* BEGIN: CPUM_CF COUNTER DEFINITIONS =================================== */ + +CPUMF_EVENT_ATTR(cf_fvn1, CPU_CYCLES, 0x0000); +CPUMF_EVENT_ATTR(cf_fvn1, INSTRUCTIONS, 0x0001); +CPUMF_EVENT_ATTR(cf_fvn1, L1I_DIR_WRITES, 0x0002); +CPUMF_EVENT_ATTR(cf_fvn1, L1I_PENALTY_CYCLES, 0x0003); +CPUMF_EVENT_ATTR(cf_fvn1, PROBLEM_STATE_CPU_CYCLES, 0x0020); +CPUMF_EVENT_ATTR(cf_fvn1, PROBLEM_STATE_INSTRUCTIONS, 0x0021); +CPUMF_EVENT_ATTR(cf_fvn1, PROBLEM_STATE_L1I_DIR_WRITES, 0x0022); +CPUMF_EVENT_ATTR(cf_fvn1, PROBLEM_STATE_L1I_PENALTY_CYCLES, 0x0023); +CPUMF_EVENT_ATTR(cf_fvn1, PROBLEM_STATE_L1D_DIR_WRITES, 0x0024); +CPUMF_EVENT_ATTR(cf_fvn1, PROBLEM_STATE_L1D_PENALTY_CYCLES, 0x0025); +CPUMF_EVENT_ATTR(cf_fvn1, L1D_DIR_WRITES, 0x0004); +CPUMF_EVENT_ATTR(cf_fvn1, L1D_PENALTY_CYCLES, 0x0005); +CPUMF_EVENT_ATTR(cf_fvn3, CPU_CYCLES, 0x0000); +CPUMF_EVENT_ATTR(cf_fvn3, INSTRUCTIONS, 0x0001); +CPUMF_EVENT_ATTR(cf_fvn3, L1I_DIR_WRITES, 0x0002); +CPUMF_EVENT_ATTR(cf_fvn3, L1I_PENALTY_CYCLES, 0x0003); +CPUMF_EVENT_ATTR(cf_fvn3, PROBLEM_STATE_CPU_CYCLES, 0x0020); +CPUMF_EVENT_ATTR(cf_fvn3, PROBLEM_STATE_INSTRUCTIONS, 0x0021); +CPUMF_EVENT_ATTR(cf_fvn3, L1D_DIR_WRITES, 0x0004); +CPUMF_EVENT_ATTR(cf_fvn3, L1D_PENALTY_CYCLES, 0x0005); +CPUMF_EVENT_ATTR(cf_svn_12345, PRNG_FUNCTIONS, 0x0040); +CPUMF_EVENT_ATTR(cf_svn_12345, PRNG_CYCLES, 0x0041); +CPUMF_EVENT_ATTR(cf_svn_12345, PRNG_BLOCKED_FUNCTIONS, 0x0042); +CPUMF_EVENT_ATTR(cf_svn_12345, PRNG_BLOCKED_CYCLES, 0x0043); +CPUMF_EVENT_ATTR(cf_svn_12345, SHA_FUNCTIONS, 0x0044); +CPUMF_EVENT_ATTR(cf_svn_12345, SHA_CYCLES, 0x0045); +CPUMF_EVENT_ATTR(cf_svn_12345, SHA_BLOCKED_FUNCTIONS, 0x0046); +CPUMF_EVENT_ATTR(cf_svn_12345, SHA_BLOCKED_CYCLES, 0x0047); +CPUMF_EVENT_ATTR(cf_svn_12345, DEA_FUNCTIONS, 0x0048); +CPUMF_EVENT_ATTR(cf_svn_12345, DEA_CYCLES, 0x0049); +CPUMF_EVENT_ATTR(cf_svn_12345, DEA_BLOCKED_FUNCTIONS, 0x004a); +CPUMF_EVENT_ATTR(cf_svn_12345, DEA_BLOCKED_CYCLES, 0x004b); +CPUMF_EVENT_ATTR(cf_svn_12345, AES_FUNCTIONS, 0x004c); +CPUMF_EVENT_ATTR(cf_svn_12345, AES_CYCLES, 0x004d); +CPUMF_EVENT_ATTR(cf_svn_12345, AES_BLOCKED_FUNCTIONS, 0x004e); +CPUMF_EVENT_ATTR(cf_svn_12345, AES_BLOCKED_CYCLES, 0x004f); +CPUMF_EVENT_ATTR(cf_svn_6, ECC_FUNCTION_COUNT, 0x0050); +CPUMF_EVENT_ATTR(cf_svn_6, ECC_CYCLES_COUNT, 0x0051); +CPUMF_EVENT_ATTR(cf_svn_6, ECC_BLOCKED_FUNCTION_COUNT, 0x0052); +CPUMF_EVENT_ATTR(cf_svn_6, ECC_BLOCKED_CYCLES_COUNT, 0x0053); +CPUMF_EVENT_ATTR(cf_z10, L1I_L2_SOURCED_WRITES, 0x0080); +CPUMF_EVENT_ATTR(cf_z10, L1D_L2_SOURCED_WRITES, 0x0081); +CPUMF_EVENT_ATTR(cf_z10, L1I_L3_LOCAL_WRITES, 0x0082); +CPUMF_EVENT_ATTR(cf_z10, L1D_L3_LOCAL_WRITES, 0x0083); +CPUMF_EVENT_ATTR(cf_z10, L1I_L3_REMOTE_WRITES, 0x0084); +CPUMF_EVENT_ATTR(cf_z10, L1D_L3_REMOTE_WRITES, 0x0085); +CPUMF_EVENT_ATTR(cf_z10, L1D_LMEM_SOURCED_WRITES, 0x0086); +CPUMF_EVENT_ATTR(cf_z10, L1I_LMEM_SOURCED_WRITES, 0x0087); +CPUMF_EVENT_ATTR(cf_z10, L1D_RO_EXCL_WRITES, 0x0088); +CPUMF_EVENT_ATTR(cf_z10, L1I_CACHELINE_INVALIDATES, 0x0089); +CPUMF_EVENT_ATTR(cf_z10, ITLB1_WRITES, 0x008a); +CPUMF_EVENT_ATTR(cf_z10, DTLB1_WRITES, 0x008b); +CPUMF_EVENT_ATTR(cf_z10, TLB2_PTE_WRITES, 0x008c); +CPUMF_EVENT_ATTR(cf_z10, TLB2_CRSTE_WRITES, 0x008d); +CPUMF_EVENT_ATTR(cf_z10, TLB2_CRSTE_HPAGE_WRITES, 0x008e); +CPUMF_EVENT_ATTR(cf_z10, ITLB1_MISSES, 0x0091); +CPUMF_EVENT_ATTR(cf_z10, DTLB1_MISSES, 0x0092); +CPUMF_EVENT_ATTR(cf_z10, L2C_STORES_SENT, 0x0093); +CPUMF_EVENT_ATTR(cf_z196, L1D_L2_SOURCED_WRITES, 0x0080); +CPUMF_EVENT_ATTR(cf_z196, L1I_L2_SOURCED_WRITES, 0x0081); +CPUMF_EVENT_ATTR(cf_z196, DTLB1_MISSES, 0x0082); +CPUMF_EVENT_ATTR(cf_z196, ITLB1_MISSES, 0x0083); +CPUMF_EVENT_ATTR(cf_z196, L2C_STORES_SENT, 0x0085); +CPUMF_EVENT_ATTR(cf_z196, L1D_OFFBOOK_L3_SOURCED_WRITES, 0x0086); +CPUMF_EVENT_ATTR(cf_z196, L1D_ONBOOK_L4_SOURCED_WRITES, 0x0087); +CPUMF_EVENT_ATTR(cf_z196, L1I_ONBOOK_L4_SOURCED_WRITES, 0x0088); +CPUMF_EVENT_ATTR(cf_z196, L1D_RO_EXCL_WRITES, 0x0089); +CPUMF_EVENT_ATTR(cf_z196, L1D_OFFBOOK_L4_SOURCED_WRITES, 0x008a); +CPUMF_EVENT_ATTR(cf_z196, L1I_OFFBOOK_L4_SOURCED_WRITES, 0x008b); +CPUMF_EVENT_ATTR(cf_z196, DTLB1_HPAGE_WRITES, 0x008c); +CPUMF_EVENT_ATTR(cf_z196, L1D_LMEM_SOURCED_WRITES, 0x008d); +CPUMF_EVENT_ATTR(cf_z196, L1I_LMEM_SOURCED_WRITES, 0x008e); +CPUMF_EVENT_ATTR(cf_z196, L1I_OFFBOOK_L3_SOURCED_WRITES, 0x008f); +CPUMF_EVENT_ATTR(cf_z196, DTLB1_WRITES, 0x0090); +CPUMF_EVENT_ATTR(cf_z196, ITLB1_WRITES, 0x0091); +CPUMF_EVENT_ATTR(cf_z196, TLB2_PTE_WRITES, 0x0092); +CPUMF_EVENT_ATTR(cf_z196, TLB2_CRSTE_HPAGE_WRITES, 0x0093); +CPUMF_EVENT_ATTR(cf_z196, TLB2_CRSTE_WRITES, 0x0094); +CPUMF_EVENT_ATTR(cf_z196, L1D_ONCHIP_L3_SOURCED_WRITES, 0x0096); +CPUMF_EVENT_ATTR(cf_z196, L1D_OFFCHIP_L3_SOURCED_WRITES, 0x0098); +CPUMF_EVENT_ATTR(cf_z196, L1I_ONCHIP_L3_SOURCED_WRITES, 0x0099); +CPUMF_EVENT_ATTR(cf_z196, L1I_OFFCHIP_L3_SOURCED_WRITES, 0x009b); +CPUMF_EVENT_ATTR(cf_zec12, DTLB1_MISSES, 0x0080); +CPUMF_EVENT_ATTR(cf_zec12, ITLB1_MISSES, 0x0081); +CPUMF_EVENT_ATTR(cf_zec12, L1D_L2I_SOURCED_WRITES, 0x0082); +CPUMF_EVENT_ATTR(cf_zec12, L1I_L2I_SOURCED_WRITES, 0x0083); +CPUMF_EVENT_ATTR(cf_zec12, L1D_L2D_SOURCED_WRITES, 0x0084); +CPUMF_EVENT_ATTR(cf_zec12, DTLB1_WRITES, 0x0085); +CPUMF_EVENT_ATTR(cf_zec12, L1D_LMEM_SOURCED_WRITES, 0x0087); +CPUMF_EVENT_ATTR(cf_zec12, L1I_LMEM_SOURCED_WRITES, 0x0089); +CPUMF_EVENT_ATTR(cf_zec12, L1D_RO_EXCL_WRITES, 0x008a); +CPUMF_EVENT_ATTR(cf_zec12, DTLB1_HPAGE_WRITES, 0x008b); +CPUMF_EVENT_ATTR(cf_zec12, ITLB1_WRITES, 0x008c); +CPUMF_EVENT_ATTR(cf_zec12, TLB2_PTE_WRITES, 0x008d); +CPUMF_EVENT_ATTR(cf_zec12, TLB2_CRSTE_HPAGE_WRITES, 0x008e); +CPUMF_EVENT_ATTR(cf_zec12, TLB2_CRSTE_WRITES, 0x008f); +CPUMF_EVENT_ATTR(cf_zec12, L1D_ONCHIP_L3_SOURCED_WRITES, 0x0090); +CPUMF_EVENT_ATTR(cf_zec12, L1D_OFFCHIP_L3_SOURCED_WRITES, 0x0091); +CPUMF_EVENT_ATTR(cf_zec12, L1D_OFFBOOK_L3_SOURCED_WRITES, 0x0092); +CPUMF_EVENT_ATTR(cf_zec12, L1D_ONBOOK_L4_SOURCED_WRITES, 0x0093); +CPUMF_EVENT_ATTR(cf_zec12, L1D_OFFBOOK_L4_SOURCED_WRITES, 0x0094); +CPUMF_EVENT_ATTR(cf_zec12, TX_NC_TEND, 0x0095); +CPUMF_EVENT_ATTR(cf_zec12, L1D_ONCHIP_L3_SOURCED_WRITES_IV, 0x0096); +CPUMF_EVENT_ATTR(cf_zec12, L1D_OFFCHIP_L3_SOURCED_WRITES_IV, 0x0097); +CPUMF_EVENT_ATTR(cf_zec12, L1D_OFFBOOK_L3_SOURCED_WRITES_IV, 0x0098); +CPUMF_EVENT_ATTR(cf_zec12, L1I_ONCHIP_L3_SOURCED_WRITES, 0x0099); +CPUMF_EVENT_ATTR(cf_zec12, L1I_OFFCHIP_L3_SOURCED_WRITES, 0x009a); +CPUMF_EVENT_ATTR(cf_zec12, L1I_OFFBOOK_L3_SOURCED_WRITES, 0x009b); +CPUMF_EVENT_ATTR(cf_zec12, L1I_ONBOOK_L4_SOURCED_WRITES, 0x009c); +CPUMF_EVENT_ATTR(cf_zec12, L1I_OFFBOOK_L4_SOURCED_WRITES, 0x009d); +CPUMF_EVENT_ATTR(cf_zec12, TX_C_TEND, 0x009e); +CPUMF_EVENT_ATTR(cf_zec12, L1I_ONCHIP_L3_SOURCED_WRITES_IV, 0x009f); +CPUMF_EVENT_ATTR(cf_zec12, L1I_OFFCHIP_L3_SOURCED_WRITES_IV, 0x00a0); +CPUMF_EVENT_ATTR(cf_zec12, L1I_OFFBOOK_L3_SOURCED_WRITES_IV, 0x00a1); +CPUMF_EVENT_ATTR(cf_zec12, TX_NC_TABORT, 0x00b1); +CPUMF_EVENT_ATTR(cf_zec12, TX_C_TABORT_NO_SPECIAL, 0x00b2); +CPUMF_EVENT_ATTR(cf_zec12, TX_C_TABORT_SPECIAL, 0x00b3); +CPUMF_EVENT_ATTR(cf_z13, L1D_RO_EXCL_WRITES, 0x0080); +CPUMF_EVENT_ATTR(cf_z13, DTLB1_WRITES, 0x0081); +CPUMF_EVENT_ATTR(cf_z13, DTLB1_MISSES, 0x0082); +CPUMF_EVENT_ATTR(cf_z13, DTLB1_HPAGE_WRITES, 0x0083); +CPUMF_EVENT_ATTR(cf_z13, DTLB1_GPAGE_WRITES, 0x0084); +CPUMF_EVENT_ATTR(cf_z13, L1D_L2D_SOURCED_WRITES, 0x0085); +CPUMF_EVENT_ATTR(cf_z13, ITLB1_WRITES, 0x0086); +CPUMF_EVENT_ATTR(cf_z13, ITLB1_MISSES, 0x0087); +CPUMF_EVENT_ATTR(cf_z13, L1I_L2I_SOURCED_WRITES, 0x0088); +CPUMF_EVENT_ATTR(cf_z13, TLB2_PTE_WRITES, 0x0089); +CPUMF_EVENT_ATTR(cf_z13, TLB2_CRSTE_HPAGE_WRITES, 0x008a); +CPUMF_EVENT_ATTR(cf_z13, TLB2_CRSTE_WRITES, 0x008b); +CPUMF_EVENT_ATTR(cf_z13, TX_C_TEND, 0x008c); +CPUMF_EVENT_ATTR(cf_z13, TX_NC_TEND, 0x008d); +CPUMF_EVENT_ATTR(cf_z13, L1C_TLB1_MISSES, 0x008f); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONCHIP_L3_SOURCED_WRITES, 0x0090); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONCHIP_L3_SOURCED_WRITES_IV, 0x0091); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONNODE_L4_SOURCED_WRITES, 0x0092); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONNODE_L3_SOURCED_WRITES_IV, 0x0093); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONNODE_L3_SOURCED_WRITES, 0x0094); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONDRAWER_L4_SOURCED_WRITES, 0x0095); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONDRAWER_L3_SOURCED_WRITES_IV, 0x0096); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONDRAWER_L3_SOURCED_WRITES, 0x0097); +CPUMF_EVENT_ATTR(cf_z13, L1D_OFFDRAWER_SCOL_L4_SOURCED_WRITES, 0x0098); +CPUMF_EVENT_ATTR(cf_z13, L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV, 0x0099); +CPUMF_EVENT_ATTR(cf_z13, L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES, 0x009a); +CPUMF_EVENT_ATTR(cf_z13, L1D_OFFDRAWER_FCOL_L4_SOURCED_WRITES, 0x009b); +CPUMF_EVENT_ATTR(cf_z13, L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV, 0x009c); +CPUMF_EVENT_ATTR(cf_z13, L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES, 0x009d); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONNODE_MEM_SOURCED_WRITES, 0x009e); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONDRAWER_MEM_SOURCED_WRITES, 0x009f); +CPUMF_EVENT_ATTR(cf_z13, L1D_OFFDRAWER_MEM_SOURCED_WRITES, 0x00a0); +CPUMF_EVENT_ATTR(cf_z13, L1D_ONCHIP_MEM_SOURCED_WRITES, 0x00a1); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONCHIP_L3_SOURCED_WRITES, 0x00a2); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONCHIP_L3_SOURCED_WRITES_IV, 0x00a3); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONNODE_L4_SOURCED_WRITES, 0x00a4); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONNODE_L3_SOURCED_WRITES_IV, 0x00a5); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONNODE_L3_SOURCED_WRITES, 0x00a6); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONDRAWER_L4_SOURCED_WRITES, 0x00a7); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONDRAWER_L3_SOURCED_WRITES_IV, 0x00a8); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONDRAWER_L3_SOURCED_WRITES, 0x00a9); +CPUMF_EVENT_ATTR(cf_z13, L1I_OFFDRAWER_SCOL_L4_SOURCED_WRITES, 0x00aa); +CPUMF_EVENT_ATTR(cf_z13, L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV, 0x00ab); +CPUMF_EVENT_ATTR(cf_z13, L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES, 0x00ac); +CPUMF_EVENT_ATTR(cf_z13, L1I_OFFDRAWER_FCOL_L4_SOURCED_WRITES, 0x00ad); +CPUMF_EVENT_ATTR(cf_z13, L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV, 0x00ae); +CPUMF_EVENT_ATTR(cf_z13, L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES, 0x00af); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONNODE_MEM_SOURCED_WRITES, 0x00b0); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONDRAWER_MEM_SOURCED_WRITES, 0x00b1); +CPUMF_EVENT_ATTR(cf_z13, L1I_OFFDRAWER_MEM_SOURCED_WRITES, 0x00b2); +CPUMF_EVENT_ATTR(cf_z13, L1I_ONCHIP_MEM_SOURCED_WRITES, 0x00b3); +CPUMF_EVENT_ATTR(cf_z13, TX_NC_TABORT, 0x00da); +CPUMF_EVENT_ATTR(cf_z13, TX_C_TABORT_NO_SPECIAL, 0x00db); +CPUMF_EVENT_ATTR(cf_z13, TX_C_TABORT_SPECIAL, 0x00dc); +CPUMF_EVENT_ATTR(cf_z13, MT_DIAG_CYCLES_ONE_THR_ACTIVE, 0x01c0); +CPUMF_EVENT_ATTR(cf_z13, MT_DIAG_CYCLES_TWO_THR_ACTIVE, 0x01c1); +CPUMF_EVENT_ATTR(cf_z14, L1D_RO_EXCL_WRITES, 0x0080); +CPUMF_EVENT_ATTR(cf_z14, DTLB2_WRITES, 0x0081); +CPUMF_EVENT_ATTR(cf_z14, DTLB2_MISSES, 0x0082); +CPUMF_EVENT_ATTR(cf_z14, DTLB2_HPAGE_WRITES, 0x0083); +CPUMF_EVENT_ATTR(cf_z14, DTLB2_GPAGE_WRITES, 0x0084); +CPUMF_EVENT_ATTR(cf_z14, L1D_L2D_SOURCED_WRITES, 0x0085); +CPUMF_EVENT_ATTR(cf_z14, ITLB2_WRITES, 0x0086); +CPUMF_EVENT_ATTR(cf_z14, ITLB2_MISSES, 0x0087); +CPUMF_EVENT_ATTR(cf_z14, L1I_L2I_SOURCED_WRITES, 0x0088); +CPUMF_EVENT_ATTR(cf_z14, TLB2_PTE_WRITES, 0x0089); +CPUMF_EVENT_ATTR(cf_z14, TLB2_CRSTE_WRITES, 0x008a); +CPUMF_EVENT_ATTR(cf_z14, TLB2_ENGINES_BUSY, 0x008b); +CPUMF_EVENT_ATTR(cf_z14, TX_C_TEND, 0x008c); +CPUMF_EVENT_ATTR(cf_z14, TX_NC_TEND, 0x008d); +CPUMF_EVENT_ATTR(cf_z14, L1C_TLB2_MISSES, 0x008f); +CPUMF_EVENT_ATTR(cf_z14, L1D_ONCHIP_L3_SOURCED_WRITES, 0x0090); +CPUMF_EVENT_ATTR(cf_z14, L1D_ONCHIP_MEMORY_SOURCED_WRITES, 0x0091); +CPUMF_EVENT_ATTR(cf_z14, L1D_ONCHIP_L3_SOURCED_WRITES_IV, 0x0092); +CPUMF_EVENT_ATTR(cf_z14, L1D_ONCLUSTER_L3_SOURCED_WRITES, 0x0093); +CPUMF_EVENT_ATTR(cf_z14, L1D_ONCLUSTER_MEMORY_SOURCED_WRITES, 0x0094); +CPUMF_EVENT_ATTR(cf_z14, L1D_ONCLUSTER_L3_SOURCED_WRITES_IV, 0x0095); +CPUMF_EVENT_ATTR(cf_z14, L1D_OFFCLUSTER_L3_SOURCED_WRITES, 0x0096); +CPUMF_EVENT_ATTR(cf_z14, L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES, 0x0097); +CPUMF_EVENT_ATTR(cf_z14, L1D_OFFCLUSTER_L3_SOURCED_WRITES_IV, 0x0098); +CPUMF_EVENT_ATTR(cf_z14, L1D_OFFDRAWER_L3_SOURCED_WRITES, 0x0099); +CPUMF_EVENT_ATTR(cf_z14, L1D_OFFDRAWER_MEMORY_SOURCED_WRITES, 0x009a); +CPUMF_EVENT_ATTR(cf_z14, L1D_OFFDRAWER_L3_SOURCED_WRITES_IV, 0x009b); +CPUMF_EVENT_ATTR(cf_z14, L1D_ONDRAWER_L4_SOURCED_WRITES, 0x009c); +CPUMF_EVENT_ATTR(cf_z14, L1D_OFFDRAWER_L4_SOURCED_WRITES, 0x009d); +CPUMF_EVENT_ATTR(cf_z14, L1D_ONCHIP_L3_SOURCED_WRITES_RO, 0x009e); +CPUMF_EVENT_ATTR(cf_z14, L1I_ONCHIP_L3_SOURCED_WRITES, 0x00a2); +CPUMF_EVENT_ATTR(cf_z14, L1I_ONCHIP_MEMORY_SOURCED_WRITES, 0x00a3); +CPUMF_EVENT_ATTR(cf_z14, L1I_ONCHIP_L3_SOURCED_WRITES_IV, 0x00a4); +CPUMF_EVENT_ATTR(cf_z14, L1I_ONCLUSTER_L3_SOURCED_WRITES, 0x00a5); +CPUMF_EVENT_ATTR(cf_z14, L1I_ONCLUSTER_MEMORY_SOURCED_WRITES, 0x00a6); +CPUMF_EVENT_ATTR(cf_z14, L1I_ONCLUSTER_L3_SOURCED_WRITES_IV, 0x00a7); +CPUMF_EVENT_ATTR(cf_z14, L1I_OFFCLUSTER_L3_SOURCED_WRITES, 0x00a8); +CPUMF_EVENT_ATTR(cf_z14, L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES, 0x00a9); +CPUMF_EVENT_ATTR(cf_z14, L1I_OFFCLUSTER_L3_SOURCED_WRITES_IV, 0x00aa); +CPUMF_EVENT_ATTR(cf_z14, L1I_OFFDRAWER_L3_SOURCED_WRITES, 0x00ab); +CPUMF_EVENT_ATTR(cf_z14, L1I_OFFDRAWER_MEMORY_SOURCED_WRITES, 0x00ac); +CPUMF_EVENT_ATTR(cf_z14, L1I_OFFDRAWER_L3_SOURCED_WRITES_IV, 0x00ad); +CPUMF_EVENT_ATTR(cf_z14, L1I_ONDRAWER_L4_SOURCED_WRITES, 0x00ae); +CPUMF_EVENT_ATTR(cf_z14, L1I_OFFDRAWER_L4_SOURCED_WRITES, 0x00af); +CPUMF_EVENT_ATTR(cf_z14, BCD_DFP_EXECUTION_SLOTS, 0x00e0); +CPUMF_EVENT_ATTR(cf_z14, VX_BCD_EXECUTION_SLOTS, 0x00e1); +CPUMF_EVENT_ATTR(cf_z14, DECIMAL_INSTRUCTIONS, 0x00e2); +CPUMF_EVENT_ATTR(cf_z14, LAST_HOST_TRANSLATIONS, 0x00e8); +CPUMF_EVENT_ATTR(cf_z14, TX_NC_TABORT, 0x00f3); +CPUMF_EVENT_ATTR(cf_z14, TX_C_TABORT_NO_SPECIAL, 0x00f4); +CPUMF_EVENT_ATTR(cf_z14, TX_C_TABORT_SPECIAL, 0x00f5); +CPUMF_EVENT_ATTR(cf_z14, MT_DIAG_CYCLES_ONE_THR_ACTIVE, 0x01c0); +CPUMF_EVENT_ATTR(cf_z14, MT_DIAG_CYCLES_TWO_THR_ACTIVE, 0x01c1); + +CPUMF_EVENT_ATTR(cf_z15, L1D_RO_EXCL_WRITES, 0x0080); +CPUMF_EVENT_ATTR(cf_z15, DTLB2_WRITES, 0x0081); +CPUMF_EVENT_ATTR(cf_z15, DTLB2_MISSES, 0x0082); +CPUMF_EVENT_ATTR(cf_z15, DTLB2_HPAGE_WRITES, 0x0083); +CPUMF_EVENT_ATTR(cf_z15, DTLB2_GPAGE_WRITES, 0x0084); +CPUMF_EVENT_ATTR(cf_z15, L1D_L2D_SOURCED_WRITES, 0x0085); +CPUMF_EVENT_ATTR(cf_z15, ITLB2_WRITES, 0x0086); +CPUMF_EVENT_ATTR(cf_z15, ITLB2_MISSES, 0x0087); +CPUMF_EVENT_ATTR(cf_z15, L1I_L2I_SOURCED_WRITES, 0x0088); +CPUMF_EVENT_ATTR(cf_z15, TLB2_PTE_WRITES, 0x0089); +CPUMF_EVENT_ATTR(cf_z15, TLB2_CRSTE_WRITES, 0x008a); +CPUMF_EVENT_ATTR(cf_z15, TLB2_ENGINES_BUSY, 0x008b); +CPUMF_EVENT_ATTR(cf_z15, TX_C_TEND, 0x008c); +CPUMF_EVENT_ATTR(cf_z15, TX_NC_TEND, 0x008d); +CPUMF_EVENT_ATTR(cf_z15, L1C_TLB2_MISSES, 0x008f); +CPUMF_EVENT_ATTR(cf_z15, L1D_ONCHIP_L3_SOURCED_WRITES, 0x0090); +CPUMF_EVENT_ATTR(cf_z15, L1D_ONCHIP_MEMORY_SOURCED_WRITES, 0x0091); +CPUMF_EVENT_ATTR(cf_z15, L1D_ONCHIP_L3_SOURCED_WRITES_IV, 0x0092); +CPUMF_EVENT_ATTR(cf_z15, L1D_ONCLUSTER_L3_SOURCED_WRITES, 0x0093); +CPUMF_EVENT_ATTR(cf_z15, L1D_ONCLUSTER_MEMORY_SOURCED_WRITES, 0x0094); +CPUMF_EVENT_ATTR(cf_z15, L1D_ONCLUSTER_L3_SOURCED_WRITES_IV, 0x0095); +CPUMF_EVENT_ATTR(cf_z15, L1D_OFFCLUSTER_L3_SOURCED_WRITES, 0x0096); +CPUMF_EVENT_ATTR(cf_z15, L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES, 0x0097); +CPUMF_EVENT_ATTR(cf_z15, L1D_OFFCLUSTER_L3_SOURCED_WRITES_IV, 0x0098); +CPUMF_EVENT_ATTR(cf_z15, L1D_OFFDRAWER_L3_SOURCED_WRITES, 0x0099); +CPUMF_EVENT_ATTR(cf_z15, L1D_OFFDRAWER_MEMORY_SOURCED_WRITES, 0x009a); +CPUMF_EVENT_ATTR(cf_z15, L1D_OFFDRAWER_L3_SOURCED_WRITES_IV, 0x009b); +CPUMF_EVENT_ATTR(cf_z15, L1D_ONDRAWER_L4_SOURCED_WRITES, 0x009c); +CPUMF_EVENT_ATTR(cf_z15, L1D_OFFDRAWER_L4_SOURCED_WRITES, 0x009d); +CPUMF_EVENT_ATTR(cf_z15, L1D_ONCHIP_L3_SOURCED_WRITES_RO, 0x009e); +CPUMF_EVENT_ATTR(cf_z15, L1I_ONCHIP_L3_SOURCED_WRITES, 0x00a2); +CPUMF_EVENT_ATTR(cf_z15, L1I_ONCHIP_MEMORY_SOURCED_WRITES, 0x00a3); +CPUMF_EVENT_ATTR(cf_z15, L1I_ONCHIP_L3_SOURCED_WRITES_IV, 0x00a4); +CPUMF_EVENT_ATTR(cf_z15, L1I_ONCLUSTER_L3_SOURCED_WRITES, 0x00a5); +CPUMF_EVENT_ATTR(cf_z15, L1I_ONCLUSTER_MEMORY_SOURCED_WRITES, 0x00a6); +CPUMF_EVENT_ATTR(cf_z15, L1I_ONCLUSTER_L3_SOURCED_WRITES_IV, 0x00a7); +CPUMF_EVENT_ATTR(cf_z15, L1I_OFFCLUSTER_L3_SOURCED_WRITES, 0x00a8); +CPUMF_EVENT_ATTR(cf_z15, L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES, 0x00a9); +CPUMF_EVENT_ATTR(cf_z15, L1I_OFFCLUSTER_L3_SOURCED_WRITES_IV, 0x00aa); +CPUMF_EVENT_ATTR(cf_z15, L1I_OFFDRAWER_L3_SOURCED_WRITES, 0x00ab); +CPUMF_EVENT_ATTR(cf_z15, L1I_OFFDRAWER_MEMORY_SOURCED_WRITES, 0x00ac); +CPUMF_EVENT_ATTR(cf_z15, L1I_OFFDRAWER_L3_SOURCED_WRITES_IV, 0x00ad); +CPUMF_EVENT_ATTR(cf_z15, L1I_ONDRAWER_L4_SOURCED_WRITES, 0x00ae); +CPUMF_EVENT_ATTR(cf_z15, L1I_OFFDRAWER_L4_SOURCED_WRITES, 0x00af); +CPUMF_EVENT_ATTR(cf_z15, BCD_DFP_EXECUTION_SLOTS, 0x00e0); +CPUMF_EVENT_ATTR(cf_z15, VX_BCD_EXECUTION_SLOTS, 0x00e1); +CPUMF_EVENT_ATTR(cf_z15, DECIMAL_INSTRUCTIONS, 0x00e2); +CPUMF_EVENT_ATTR(cf_z15, LAST_HOST_TRANSLATIONS, 0x00e8); +CPUMF_EVENT_ATTR(cf_z15, TX_NC_TABORT, 0x00f3); +CPUMF_EVENT_ATTR(cf_z15, TX_C_TABORT_NO_SPECIAL, 0x00f4); +CPUMF_EVENT_ATTR(cf_z15, TX_C_TABORT_SPECIAL, 0x00f5); +CPUMF_EVENT_ATTR(cf_z15, DFLT_ACCESS, 0x00f7); +CPUMF_EVENT_ATTR(cf_z15, DFLT_CYCLES, 0x00fc); +CPUMF_EVENT_ATTR(cf_z15, DFLT_CC, 0x00108); +CPUMF_EVENT_ATTR(cf_z15, DFLT_CCFINISH, 0x00109); +CPUMF_EVENT_ATTR(cf_z15, MT_DIAG_CYCLES_ONE_THR_ACTIVE, 0x01c0); +CPUMF_EVENT_ATTR(cf_z15, MT_DIAG_CYCLES_TWO_THR_ACTIVE, 0x01c1); +CPUMF_EVENT_ATTR(cf_z16, L1D_RO_EXCL_WRITES, 0x0080); +CPUMF_EVENT_ATTR(cf_z16, DTLB2_WRITES, 0x0081); +CPUMF_EVENT_ATTR(cf_z16, DTLB2_MISSES, 0x0082); +CPUMF_EVENT_ATTR(cf_z16, CRSTE_1MB_WRITES, 0x0083); +CPUMF_EVENT_ATTR(cf_z16, DTLB2_GPAGE_WRITES, 0x0084); +CPUMF_EVENT_ATTR(cf_z16, ITLB2_WRITES, 0x0086); +CPUMF_EVENT_ATTR(cf_z16, ITLB2_MISSES, 0x0087); +CPUMF_EVENT_ATTR(cf_z16, TLB2_PTE_WRITES, 0x0089); +CPUMF_EVENT_ATTR(cf_z16, TLB2_CRSTE_WRITES, 0x008a); +CPUMF_EVENT_ATTR(cf_z16, TLB2_ENGINES_BUSY, 0x008b); +CPUMF_EVENT_ATTR(cf_z16, TX_C_TEND, 0x008c); +CPUMF_EVENT_ATTR(cf_z16, TX_NC_TEND, 0x008d); +CPUMF_EVENT_ATTR(cf_z16, L1C_TLB2_MISSES, 0x008f); +CPUMF_EVENT_ATTR(cf_z16, DCW_REQ, 0x0091); +CPUMF_EVENT_ATTR(cf_z16, DCW_REQ_IV, 0x0092); +CPUMF_EVENT_ATTR(cf_z16, DCW_REQ_CHIP_HIT, 0x0093); +CPUMF_EVENT_ATTR(cf_z16, DCW_REQ_DRAWER_HIT, 0x0094); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_CHIP, 0x0095); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_CHIP_IV, 0x0096); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_CHIP_CHIP_HIT, 0x0097); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_CHIP_DRAWER_HIT, 0x0098); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_MODULE, 0x0099); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_DRAWER, 0x009a); +CPUMF_EVENT_ATTR(cf_z16, DCW_OFF_DRAWER, 0x009b); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_CHIP_MEMORY, 0x009c); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_MODULE_MEMORY, 0x009d); +CPUMF_EVENT_ATTR(cf_z16, DCW_ON_DRAWER_MEMORY, 0x009e); +CPUMF_EVENT_ATTR(cf_z16, DCW_OFF_DRAWER_MEMORY, 0x009f); +CPUMF_EVENT_ATTR(cf_z16, IDCW_ON_MODULE_IV, 0x00a0); +CPUMF_EVENT_ATTR(cf_z16, IDCW_ON_MODULE_CHIP_HIT, 0x00a1); +CPUMF_EVENT_ATTR(cf_z16, IDCW_ON_MODULE_DRAWER_HIT, 0x00a2); +CPUMF_EVENT_ATTR(cf_z16, IDCW_ON_DRAWER_IV, 0x00a3); +CPUMF_EVENT_ATTR(cf_z16, IDCW_ON_DRAWER_CHIP_HIT, 0x00a4); +CPUMF_EVENT_ATTR(cf_z16, IDCW_ON_DRAWER_DRAWER_HIT, 0x00a5); +CPUMF_EVENT_ATTR(cf_z16, IDCW_OFF_DRAWER_IV, 0x00a6); +CPUMF_EVENT_ATTR(cf_z16, IDCW_OFF_DRAWER_CHIP_HIT, 0x00a7); +CPUMF_EVENT_ATTR(cf_z16, IDCW_OFF_DRAWER_DRAWER_HIT, 0x00a8); +CPUMF_EVENT_ATTR(cf_z16, ICW_REQ, 0x00a9); +CPUMF_EVENT_ATTR(cf_z16, ICW_REQ_IV, 0x00aa); +CPUMF_EVENT_ATTR(cf_z16, ICW_REQ_CHIP_HIT, 0x00ab); +CPUMF_EVENT_ATTR(cf_z16, ICW_REQ_DRAWER_HIT, 0x00ac); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_CHIP, 0x00ad); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_CHIP_IV, 0x00ae); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_CHIP_CHIP_HIT, 0x00af); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_CHIP_DRAWER_HIT, 0x00b0); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_MODULE, 0x00b1); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_DRAWER, 0x00b2); +CPUMF_EVENT_ATTR(cf_z16, ICW_OFF_DRAWER, 0x00b3); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_CHIP_MEMORY, 0x00b4); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_MODULE_MEMORY, 0x00b5); +CPUMF_EVENT_ATTR(cf_z16, ICW_ON_DRAWER_MEMORY, 0x00b6); +CPUMF_EVENT_ATTR(cf_z16, ICW_OFF_DRAWER_MEMORY, 0x00b7); +CPUMF_EVENT_ATTR(cf_z16, BCD_DFP_EXECUTION_SLOTS, 0x00e0); +CPUMF_EVENT_ATTR(cf_z16, VX_BCD_EXECUTION_SLOTS, 0x00e1); +CPUMF_EVENT_ATTR(cf_z16, DECIMAL_INSTRUCTIONS, 0x00e2); +CPUMF_EVENT_ATTR(cf_z16, LAST_HOST_TRANSLATIONS, 0x00e8); +CPUMF_EVENT_ATTR(cf_z16, TX_NC_TABORT, 0x00f4); +CPUMF_EVENT_ATTR(cf_z16, TX_C_TABORT_NO_SPECIAL, 0x00f5); +CPUMF_EVENT_ATTR(cf_z16, TX_C_TABORT_SPECIAL, 0x00f6); +CPUMF_EVENT_ATTR(cf_z16, DFLT_ACCESS, 0x00f8); +CPUMF_EVENT_ATTR(cf_z16, DFLT_CYCLES, 0x00fd); +CPUMF_EVENT_ATTR(cf_z16, SORTL, 0x0100); +CPUMF_EVENT_ATTR(cf_z16, DFLT_CC, 0x0109); +CPUMF_EVENT_ATTR(cf_z16, DFLT_CCFINISH, 0x010a); +CPUMF_EVENT_ATTR(cf_z16, NNPA_INVOCATIONS, 0x010b); +CPUMF_EVENT_ATTR(cf_z16, NNPA_COMPLETIONS, 0x010c); +CPUMF_EVENT_ATTR(cf_z16, NNPA_WAIT_LOCK, 0x010d); +CPUMF_EVENT_ATTR(cf_z16, NNPA_HOLD_LOCK, 0x010e); +CPUMF_EVENT_ATTR(cf_z16, MT_DIAG_CYCLES_ONE_THR_ACTIVE, 0x01c0); +CPUMF_EVENT_ATTR(cf_z16, MT_DIAG_CYCLES_TWO_THR_ACTIVE, 0x01c1); + +static struct attribute *cpumcf_fvn1_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_fvn1, CPU_CYCLES), + CPUMF_EVENT_PTR(cf_fvn1, INSTRUCTIONS), + CPUMF_EVENT_PTR(cf_fvn1, L1I_DIR_WRITES), + CPUMF_EVENT_PTR(cf_fvn1, L1I_PENALTY_CYCLES), + CPUMF_EVENT_PTR(cf_fvn1, PROBLEM_STATE_CPU_CYCLES), + CPUMF_EVENT_PTR(cf_fvn1, PROBLEM_STATE_INSTRUCTIONS), + CPUMF_EVENT_PTR(cf_fvn1, PROBLEM_STATE_L1I_DIR_WRITES), + CPUMF_EVENT_PTR(cf_fvn1, PROBLEM_STATE_L1I_PENALTY_CYCLES), + CPUMF_EVENT_PTR(cf_fvn1, PROBLEM_STATE_L1D_DIR_WRITES), + CPUMF_EVENT_PTR(cf_fvn1, PROBLEM_STATE_L1D_PENALTY_CYCLES), + CPUMF_EVENT_PTR(cf_fvn1, L1D_DIR_WRITES), + CPUMF_EVENT_PTR(cf_fvn1, L1D_PENALTY_CYCLES), + NULL, +}; + +static struct attribute *cpumcf_fvn3_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_fvn3, CPU_CYCLES), + CPUMF_EVENT_PTR(cf_fvn3, INSTRUCTIONS), + CPUMF_EVENT_PTR(cf_fvn3, L1I_DIR_WRITES), + CPUMF_EVENT_PTR(cf_fvn3, L1I_PENALTY_CYCLES), + CPUMF_EVENT_PTR(cf_fvn3, PROBLEM_STATE_CPU_CYCLES), + CPUMF_EVENT_PTR(cf_fvn3, PROBLEM_STATE_INSTRUCTIONS), + CPUMF_EVENT_PTR(cf_fvn3, L1D_DIR_WRITES), + CPUMF_EVENT_PTR(cf_fvn3, L1D_PENALTY_CYCLES), + NULL, +}; + +static struct attribute *cpumcf_svn_12345_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_svn_12345, PRNG_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, PRNG_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, PRNG_BLOCKED_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, PRNG_BLOCKED_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, SHA_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, SHA_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, SHA_BLOCKED_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, SHA_BLOCKED_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, DEA_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, DEA_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, DEA_BLOCKED_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, DEA_BLOCKED_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, AES_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, AES_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, AES_BLOCKED_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, AES_BLOCKED_CYCLES), + NULL, +}; + +static struct attribute *cpumcf_svn_67_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_svn_12345, PRNG_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, PRNG_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, PRNG_BLOCKED_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, PRNG_BLOCKED_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, SHA_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, SHA_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, SHA_BLOCKED_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, SHA_BLOCKED_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, DEA_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, DEA_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, DEA_BLOCKED_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, DEA_BLOCKED_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, AES_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, AES_CYCLES), + CPUMF_EVENT_PTR(cf_svn_12345, AES_BLOCKED_FUNCTIONS), + CPUMF_EVENT_PTR(cf_svn_12345, AES_BLOCKED_CYCLES), + CPUMF_EVENT_PTR(cf_svn_6, ECC_FUNCTION_COUNT), + CPUMF_EVENT_PTR(cf_svn_6, ECC_CYCLES_COUNT), + CPUMF_EVENT_PTR(cf_svn_6, ECC_BLOCKED_FUNCTION_COUNT), + CPUMF_EVENT_PTR(cf_svn_6, ECC_BLOCKED_CYCLES_COUNT), + NULL, +}; + +static struct attribute *cpumcf_z10_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_z10, L1I_L2_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1D_L2_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1I_L3_LOCAL_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1D_L3_LOCAL_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1I_L3_REMOTE_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1D_L3_REMOTE_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1D_LMEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1I_LMEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1D_RO_EXCL_WRITES), + CPUMF_EVENT_PTR(cf_z10, L1I_CACHELINE_INVALIDATES), + CPUMF_EVENT_PTR(cf_z10, ITLB1_WRITES), + CPUMF_EVENT_PTR(cf_z10, DTLB1_WRITES), + CPUMF_EVENT_PTR(cf_z10, TLB2_PTE_WRITES), + CPUMF_EVENT_PTR(cf_z10, TLB2_CRSTE_WRITES), + CPUMF_EVENT_PTR(cf_z10, TLB2_CRSTE_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z10, ITLB1_MISSES), + CPUMF_EVENT_PTR(cf_z10, DTLB1_MISSES), + CPUMF_EVENT_PTR(cf_z10, L2C_STORES_SENT), + NULL, +}; + +static struct attribute *cpumcf_z196_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_z196, L1D_L2_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1I_L2_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, DTLB1_MISSES), + CPUMF_EVENT_PTR(cf_z196, ITLB1_MISSES), + CPUMF_EVENT_PTR(cf_z196, L2C_STORES_SENT), + CPUMF_EVENT_PTR(cf_z196, L1D_OFFBOOK_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1D_ONBOOK_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1I_ONBOOK_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1D_RO_EXCL_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1D_OFFBOOK_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1I_OFFBOOK_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, DTLB1_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1D_LMEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1I_LMEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1I_OFFBOOK_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, DTLB1_WRITES), + CPUMF_EVENT_PTR(cf_z196, ITLB1_WRITES), + CPUMF_EVENT_PTR(cf_z196, TLB2_PTE_WRITES), + CPUMF_EVENT_PTR(cf_z196, TLB2_CRSTE_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z196, TLB2_CRSTE_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1D_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1D_OFFCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1I_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z196, L1I_OFFCHIP_L3_SOURCED_WRITES), + NULL, +}; + +static struct attribute *cpumcf_zec12_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_zec12, DTLB1_MISSES), + CPUMF_EVENT_PTR(cf_zec12, ITLB1_MISSES), + CPUMF_EVENT_PTR(cf_zec12, L1D_L2I_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1I_L2I_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1D_L2D_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, DTLB1_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1D_LMEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1I_LMEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1D_RO_EXCL_WRITES), + CPUMF_EVENT_PTR(cf_zec12, DTLB1_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_zec12, ITLB1_WRITES), + CPUMF_EVENT_PTR(cf_zec12, TLB2_PTE_WRITES), + CPUMF_EVENT_PTR(cf_zec12, TLB2_CRSTE_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_zec12, TLB2_CRSTE_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1D_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1D_OFFCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1D_OFFBOOK_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1D_ONBOOK_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1D_OFFBOOK_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, TX_NC_TEND), + CPUMF_EVENT_PTR(cf_zec12, L1D_ONCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_zec12, L1D_OFFCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_zec12, L1D_OFFBOOK_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_zec12, L1I_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1I_OFFCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1I_OFFBOOK_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1I_ONBOOK_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, L1I_OFFBOOK_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_zec12, TX_C_TEND), + CPUMF_EVENT_PTR(cf_zec12, L1I_ONCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_zec12, L1I_OFFCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_zec12, L1I_OFFBOOK_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_zec12, TX_NC_TABORT), + CPUMF_EVENT_PTR(cf_zec12, TX_C_TABORT_NO_SPECIAL), + CPUMF_EVENT_PTR(cf_zec12, TX_C_TABORT_SPECIAL), + NULL, +}; + +static struct attribute *cpumcf_z13_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_z13, L1D_RO_EXCL_WRITES), + CPUMF_EVENT_PTR(cf_z13, DTLB1_WRITES), + CPUMF_EVENT_PTR(cf_z13, DTLB1_MISSES), + CPUMF_EVENT_PTR(cf_z13, DTLB1_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z13, DTLB1_GPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_L2D_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, ITLB1_WRITES), + CPUMF_EVENT_PTR(cf_z13, ITLB1_MISSES), + CPUMF_EVENT_PTR(cf_z13, L1I_L2I_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, TLB2_PTE_WRITES), + CPUMF_EVENT_PTR(cf_z13, TLB2_CRSTE_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z13, TLB2_CRSTE_WRITES), + CPUMF_EVENT_PTR(cf_z13, TX_C_TEND), + CPUMF_EVENT_PTR(cf_z13, TX_NC_TEND), + CPUMF_EVENT_PTR(cf_z13, L1C_TLB1_MISSES), + CPUMF_EVENT_PTR(cf_z13, L1D_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_ONCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1D_ONNODE_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_ONNODE_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1D_ONNODE_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_ONDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_ONDRAWER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1D_ONDRAWER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_OFFDRAWER_SCOL_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_OFFDRAWER_FCOL_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_ONNODE_MEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_ONDRAWER_MEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_OFFDRAWER_MEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1D_ONCHIP_MEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_ONCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1I_ONNODE_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_ONNODE_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1I_ONNODE_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_ONDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_ONDRAWER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1I_ONDRAWER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_OFFDRAWER_SCOL_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_OFFDRAWER_FCOL_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z13, L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_ONNODE_MEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_ONDRAWER_MEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_OFFDRAWER_MEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, L1I_ONCHIP_MEM_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z13, TX_NC_TABORT), + CPUMF_EVENT_PTR(cf_z13, TX_C_TABORT_NO_SPECIAL), + CPUMF_EVENT_PTR(cf_z13, TX_C_TABORT_SPECIAL), + CPUMF_EVENT_PTR(cf_z13, MT_DIAG_CYCLES_ONE_THR_ACTIVE), + CPUMF_EVENT_PTR(cf_z13, MT_DIAG_CYCLES_TWO_THR_ACTIVE), + NULL, +}; + +static struct attribute *cpumcf_z14_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_z14, L1D_RO_EXCL_WRITES), + CPUMF_EVENT_PTR(cf_z14, DTLB2_WRITES), + CPUMF_EVENT_PTR(cf_z14, DTLB2_MISSES), + CPUMF_EVENT_PTR(cf_z14, DTLB2_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z14, DTLB2_GPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_L2D_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, ITLB2_WRITES), + CPUMF_EVENT_PTR(cf_z14, ITLB2_MISSES), + CPUMF_EVENT_PTR(cf_z14, L1I_L2I_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, TLB2_PTE_WRITES), + CPUMF_EVENT_PTR(cf_z14, TLB2_CRSTE_WRITES), + CPUMF_EVENT_PTR(cf_z14, TLB2_ENGINES_BUSY), + CPUMF_EVENT_PTR(cf_z14, TX_C_TEND), + CPUMF_EVENT_PTR(cf_z14, TX_NC_TEND), + CPUMF_EVENT_PTR(cf_z14, L1C_TLB2_MISSES), + CPUMF_EVENT_PTR(cf_z14, L1D_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_ONCHIP_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_ONCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z14, L1D_ONCLUSTER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_ONCLUSTER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_ONCLUSTER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z14, L1D_OFFCLUSTER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_OFFCLUSTER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z14, L1D_OFFDRAWER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_OFFDRAWER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_OFFDRAWER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z14, L1D_ONDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_OFFDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1D_ONCHIP_L3_SOURCED_WRITES_RO), + CPUMF_EVENT_PTR(cf_z14, L1I_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_ONCHIP_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_ONCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z14, L1I_ONCLUSTER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_ONCLUSTER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_ONCLUSTER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z14, L1I_OFFCLUSTER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_OFFCLUSTER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z14, L1I_OFFDRAWER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_OFFDRAWER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_OFFDRAWER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z14, L1I_ONDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, L1I_OFFDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z14, BCD_DFP_EXECUTION_SLOTS), + CPUMF_EVENT_PTR(cf_z14, VX_BCD_EXECUTION_SLOTS), + CPUMF_EVENT_PTR(cf_z14, DECIMAL_INSTRUCTIONS), + CPUMF_EVENT_PTR(cf_z14, LAST_HOST_TRANSLATIONS), + CPUMF_EVENT_PTR(cf_z14, TX_NC_TABORT), + CPUMF_EVENT_PTR(cf_z14, TX_C_TABORT_NO_SPECIAL), + CPUMF_EVENT_PTR(cf_z14, TX_C_TABORT_SPECIAL), + CPUMF_EVENT_PTR(cf_z14, MT_DIAG_CYCLES_ONE_THR_ACTIVE), + CPUMF_EVENT_PTR(cf_z14, MT_DIAG_CYCLES_TWO_THR_ACTIVE), + NULL, +}; + +static struct attribute *cpumcf_z15_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_z15, L1D_RO_EXCL_WRITES), + CPUMF_EVENT_PTR(cf_z15, DTLB2_WRITES), + CPUMF_EVENT_PTR(cf_z15, DTLB2_MISSES), + CPUMF_EVENT_PTR(cf_z15, DTLB2_HPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z15, DTLB2_GPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_L2D_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, ITLB2_WRITES), + CPUMF_EVENT_PTR(cf_z15, ITLB2_MISSES), + CPUMF_EVENT_PTR(cf_z15, L1I_L2I_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, TLB2_PTE_WRITES), + CPUMF_EVENT_PTR(cf_z15, TLB2_CRSTE_WRITES), + CPUMF_EVENT_PTR(cf_z15, TLB2_ENGINES_BUSY), + CPUMF_EVENT_PTR(cf_z15, TX_C_TEND), + CPUMF_EVENT_PTR(cf_z15, TX_NC_TEND), + CPUMF_EVENT_PTR(cf_z15, L1C_TLB2_MISSES), + CPUMF_EVENT_PTR(cf_z15, L1D_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_ONCHIP_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_ONCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z15, L1D_ONCLUSTER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_ONCLUSTER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_ONCLUSTER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z15, L1D_OFFCLUSTER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_OFFCLUSTER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z15, L1D_OFFDRAWER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_OFFDRAWER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_OFFDRAWER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z15, L1D_ONDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_OFFDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1D_ONCHIP_L3_SOURCED_WRITES_RO), + CPUMF_EVENT_PTR(cf_z15, L1I_ONCHIP_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_ONCHIP_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_ONCHIP_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z15, L1I_ONCLUSTER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_ONCLUSTER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_ONCLUSTER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z15, L1I_OFFCLUSTER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_OFFCLUSTER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z15, L1I_OFFDRAWER_L3_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_OFFDRAWER_MEMORY_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_OFFDRAWER_L3_SOURCED_WRITES_IV), + CPUMF_EVENT_PTR(cf_z15, L1I_ONDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, L1I_OFFDRAWER_L4_SOURCED_WRITES), + CPUMF_EVENT_PTR(cf_z15, BCD_DFP_EXECUTION_SLOTS), + CPUMF_EVENT_PTR(cf_z15, VX_BCD_EXECUTION_SLOTS), + CPUMF_EVENT_PTR(cf_z15, DECIMAL_INSTRUCTIONS), + CPUMF_EVENT_PTR(cf_z15, LAST_HOST_TRANSLATIONS), + CPUMF_EVENT_PTR(cf_z15, TX_NC_TABORT), + CPUMF_EVENT_PTR(cf_z15, TX_C_TABORT_NO_SPECIAL), + CPUMF_EVENT_PTR(cf_z15, TX_C_TABORT_SPECIAL), + CPUMF_EVENT_PTR(cf_z15, DFLT_ACCESS), + CPUMF_EVENT_PTR(cf_z15, DFLT_CYCLES), + CPUMF_EVENT_PTR(cf_z15, DFLT_CC), + CPUMF_EVENT_PTR(cf_z15, DFLT_CCFINISH), + CPUMF_EVENT_PTR(cf_z15, MT_DIAG_CYCLES_ONE_THR_ACTIVE), + CPUMF_EVENT_PTR(cf_z15, MT_DIAG_CYCLES_TWO_THR_ACTIVE), + NULL, +}; + +static struct attribute *cpumcf_z16_pmu_event_attr[] __initdata = { + CPUMF_EVENT_PTR(cf_z16, L1D_RO_EXCL_WRITES), + CPUMF_EVENT_PTR(cf_z16, DTLB2_WRITES), + CPUMF_EVENT_PTR(cf_z16, DTLB2_MISSES), + CPUMF_EVENT_PTR(cf_z16, CRSTE_1MB_WRITES), + CPUMF_EVENT_PTR(cf_z16, DTLB2_GPAGE_WRITES), + CPUMF_EVENT_PTR(cf_z16, ITLB2_WRITES), + CPUMF_EVENT_PTR(cf_z16, ITLB2_MISSES), + CPUMF_EVENT_PTR(cf_z16, TLB2_PTE_WRITES), + CPUMF_EVENT_PTR(cf_z16, TLB2_CRSTE_WRITES), + CPUMF_EVENT_PTR(cf_z16, TLB2_ENGINES_BUSY), + CPUMF_EVENT_PTR(cf_z16, TX_C_TEND), + CPUMF_EVENT_PTR(cf_z16, TX_NC_TEND), + CPUMF_EVENT_PTR(cf_z16, L1C_TLB2_MISSES), + CPUMF_EVENT_PTR(cf_z16, DCW_REQ), + CPUMF_EVENT_PTR(cf_z16, DCW_REQ_IV), + CPUMF_EVENT_PTR(cf_z16, DCW_REQ_CHIP_HIT), + CPUMF_EVENT_PTR(cf_z16, DCW_REQ_DRAWER_HIT), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_CHIP), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_CHIP_IV), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_CHIP_CHIP_HIT), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_CHIP_DRAWER_HIT), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_MODULE), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_DRAWER), + CPUMF_EVENT_PTR(cf_z16, DCW_OFF_DRAWER), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_CHIP_MEMORY), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_MODULE_MEMORY), + CPUMF_EVENT_PTR(cf_z16, DCW_ON_DRAWER_MEMORY), + CPUMF_EVENT_PTR(cf_z16, DCW_OFF_DRAWER_MEMORY), + CPUMF_EVENT_PTR(cf_z16, IDCW_ON_MODULE_IV), + CPUMF_EVENT_PTR(cf_z16, IDCW_ON_MODULE_CHIP_HIT), + CPUMF_EVENT_PTR(cf_z16, IDCW_ON_MODULE_DRAWER_HIT), + CPUMF_EVENT_PTR(cf_z16, IDCW_ON_DRAWER_IV), + CPUMF_EVENT_PTR(cf_z16, IDCW_ON_DRAWER_CHIP_HIT), + CPUMF_EVENT_PTR(cf_z16, IDCW_ON_DRAWER_DRAWER_HIT), + CPUMF_EVENT_PTR(cf_z16, IDCW_OFF_DRAWER_IV), + CPUMF_EVENT_PTR(cf_z16, IDCW_OFF_DRAWER_CHIP_HIT), + CPUMF_EVENT_PTR(cf_z16, IDCW_OFF_DRAWER_DRAWER_HIT), + CPUMF_EVENT_PTR(cf_z16, ICW_REQ), + CPUMF_EVENT_PTR(cf_z16, ICW_REQ_IV), + CPUMF_EVENT_PTR(cf_z16, ICW_REQ_CHIP_HIT), + CPUMF_EVENT_PTR(cf_z16, ICW_REQ_DRAWER_HIT), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_CHIP), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_CHIP_IV), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_CHIP_CHIP_HIT), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_CHIP_DRAWER_HIT), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_MODULE), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_DRAWER), + CPUMF_EVENT_PTR(cf_z16, ICW_OFF_DRAWER), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_CHIP_MEMORY), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_MODULE_MEMORY), + CPUMF_EVENT_PTR(cf_z16, ICW_ON_DRAWER_MEMORY), + CPUMF_EVENT_PTR(cf_z16, ICW_OFF_DRAWER_MEMORY), + CPUMF_EVENT_PTR(cf_z16, BCD_DFP_EXECUTION_SLOTS), + CPUMF_EVENT_PTR(cf_z16, VX_BCD_EXECUTION_SLOTS), + CPUMF_EVENT_PTR(cf_z16, DECIMAL_INSTRUCTIONS), + CPUMF_EVENT_PTR(cf_z16, LAST_HOST_TRANSLATIONS), + CPUMF_EVENT_PTR(cf_z16, TX_NC_TABORT), + CPUMF_EVENT_PTR(cf_z16, TX_C_TABORT_NO_SPECIAL), + CPUMF_EVENT_PTR(cf_z16, TX_C_TABORT_SPECIAL), + CPUMF_EVENT_PTR(cf_z16, DFLT_ACCESS), + CPUMF_EVENT_PTR(cf_z16, DFLT_CYCLES), + CPUMF_EVENT_PTR(cf_z16, SORTL), + CPUMF_EVENT_PTR(cf_z16, DFLT_CC), + CPUMF_EVENT_PTR(cf_z16, DFLT_CCFINISH), + CPUMF_EVENT_PTR(cf_z16, NNPA_INVOCATIONS), + CPUMF_EVENT_PTR(cf_z16, NNPA_COMPLETIONS), + CPUMF_EVENT_PTR(cf_z16, NNPA_WAIT_LOCK), + CPUMF_EVENT_PTR(cf_z16, NNPA_HOLD_LOCK), + CPUMF_EVENT_PTR(cf_z16, MT_DIAG_CYCLES_ONE_THR_ACTIVE), + CPUMF_EVENT_PTR(cf_z16, MT_DIAG_CYCLES_TWO_THR_ACTIVE), + NULL, +}; + +/* END: CPUM_CF COUNTER DEFINITIONS ===================================== */ + +static struct attribute_group cpumcf_pmu_events_group = { + .name = "events", +}; + +PMU_FORMAT_ATTR(event, "config:0-63"); + +static struct attribute *cpumcf_pmu_format_attr[] = { + &format_attr_event.attr, + NULL, +}; + +static struct attribute_group cpumcf_pmu_format_group = { + .name = "format", + .attrs = cpumcf_pmu_format_attr, +}; + +static const struct attribute_group *cpumcf_pmu_attr_groups[] = { + &cpumcf_pmu_events_group, + &cpumcf_pmu_format_group, + NULL, +}; + + +static __init struct attribute **merge_attr(struct attribute **a, + struct attribute **b, + struct attribute **c) +{ + struct attribute **new; + int j, i; + + for (j = 0; a[j]; j++) + ; + for (i = 0; b[i]; i++) + j++; + for (i = 0; c[i]; i++) + j++; + j++; + + new = kmalloc_array(j, sizeof(struct attribute *), GFP_KERNEL); + if (!new) + return NULL; + j = 0; + for (i = 0; a[i]; i++) + new[j++] = a[i]; + for (i = 0; b[i]; i++) + new[j++] = b[i]; + for (i = 0; c[i]; i++) + new[j++] = c[i]; + new[j] = NULL; + + return new; +} + +__init const struct attribute_group **cpumf_cf_event_group(void) +{ + struct attribute **combined, **model, **cfvn, **csvn; + struct attribute *none[] = { NULL }; + struct cpumf_ctr_info ci; + struct cpuid cpu_id; + + /* Determine generic counters set(s) */ + qctri(&ci); + switch (ci.cfvn) { + case 1: + cfvn = cpumcf_fvn1_pmu_event_attr; + break; + case 3: + cfvn = cpumcf_fvn3_pmu_event_attr; + break; + default: + cfvn = none; + } + + /* Determine version specific crypto set */ + switch (ci.csvn) { + case 1 ... 5: + csvn = cpumcf_svn_12345_pmu_event_attr; + break; + case 6 ... 7: + csvn = cpumcf_svn_67_pmu_event_attr; + break; + default: + csvn = none; + } + + /* Determine model-specific counter set(s) */ + get_cpu_id(&cpu_id); + switch (cpu_id.machine) { + case 0x2097: + case 0x2098: + model = cpumcf_z10_pmu_event_attr; + break; + case 0x2817: + case 0x2818: + model = cpumcf_z196_pmu_event_attr; + break; + case 0x2827: + case 0x2828: + model = cpumcf_zec12_pmu_event_attr; + break; + case 0x2964: + case 0x2965: + model = cpumcf_z13_pmu_event_attr; + break; + case 0x3906: + case 0x3907: + model = cpumcf_z14_pmu_event_attr; + break; + case 0x8561: + case 0x8562: + model = cpumcf_z15_pmu_event_attr; + break; + case 0x3931: + case 0x3932: + model = cpumcf_z16_pmu_event_attr; + break; + default: + model = none; + break; + } + + combined = merge_attr(cfvn, csvn, model); + if (combined) + cpumcf_pmu_events_group.attrs = combined; + return cpumcf_pmu_attr_groups; +} diff --git a/arch/s390/kernel/perf_cpum_sf.c b/arch/s390/kernel/perf_cpum_sf.c new file mode 100644 index 0000000000..06efad5b4f --- /dev/null +++ b/arch/s390/kernel/perf_cpum_sf.c @@ -0,0 +1,2280 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Performance event support for the System z CPU-measurement Sampling Facility + * + * Copyright IBM Corp. 2013, 2018 + * Author(s): Hendrik Brueckner <brueckner@linux.vnet.ibm.com> + */ +#define KMSG_COMPONENT "cpum_sf" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/kernel.h> +#include <linux/kernel_stat.h> +#include <linux/perf_event.h> +#include <linux/percpu.h> +#include <linux/pid.h> +#include <linux/notifier.h> +#include <linux/export.h> +#include <linux/slab.h> +#include <linux/mm.h> +#include <linux/moduleparam.h> +#include <asm/cpu_mf.h> +#include <asm/irq.h> +#include <asm/debug.h> +#include <asm/timex.h> +#include <linux/io.h> + +/* Minimum number of sample-data-block-tables: + * At least one table is required for the sampling buffer structure. + * A single table contains up to 511 pointers to sample-data-blocks. + */ +#define CPUM_SF_MIN_SDBT 1 + +/* Number of sample-data-blocks per sample-data-block-table (SDBT): + * A table contains SDB pointers (8 bytes) and one table-link entry + * that points to the origin of the next SDBT. + */ +#define CPUM_SF_SDB_PER_TABLE ((PAGE_SIZE - 8) / 8) + +/* Maximum page offset for an SDBT table-link entry: + * If this page offset is reached, a table-link entry to the next SDBT + * must be added. + */ +#define CPUM_SF_SDBT_TL_OFFSET (CPUM_SF_SDB_PER_TABLE * 8) +static inline int require_table_link(const void *sdbt) +{ + return ((unsigned long)sdbt & ~PAGE_MASK) == CPUM_SF_SDBT_TL_OFFSET; +} + +/* Minimum and maximum sampling buffer sizes: + * + * This number represents the maximum size of the sampling buffer taking + * the number of sample-data-block-tables into account. Note that these + * numbers apply to the basic-sampling function only. + * The maximum number of SDBs is increased by CPUM_SF_SDB_DIAG_FACTOR if + * the diagnostic-sampling function is active. + * + * Sampling buffer size Buffer characteristics + * --------------------------------------------------- + * 64KB == 16 pages (4KB per page) + * 1 page for SDB-tables + * 15 pages for SDBs + * + * 32MB == 8192 pages (4KB per page) + * 16 pages for SDB-tables + * 8176 pages for SDBs + */ +static unsigned long __read_mostly CPUM_SF_MIN_SDB = 15; +static unsigned long __read_mostly CPUM_SF_MAX_SDB = 8176; +static unsigned long __read_mostly CPUM_SF_SDB_DIAG_FACTOR = 1; + +struct sf_buffer { + unsigned long *sdbt; /* Sample-data-block-table origin */ + /* buffer characteristics (required for buffer increments) */ + unsigned long num_sdb; /* Number of sample-data-blocks */ + unsigned long num_sdbt; /* Number of sample-data-block-tables */ + unsigned long *tail; /* last sample-data-block-table */ +}; + +struct aux_buffer { + struct sf_buffer sfb; + unsigned long head; /* index of SDB of buffer head */ + unsigned long alert_mark; /* index of SDB of alert request position */ + unsigned long empty_mark; /* mark of SDB not marked full */ + unsigned long *sdb_index; /* SDB address for fast lookup */ + unsigned long *sdbt_index; /* SDBT address for fast lookup */ +}; + +struct cpu_hw_sf { + /* CPU-measurement sampling information block */ + struct hws_qsi_info_block qsi; + /* CPU-measurement sampling control block */ + struct hws_lsctl_request_block lsctl; + struct sf_buffer sfb; /* Sampling buffer */ + unsigned int flags; /* Status flags */ + struct perf_event *event; /* Scheduled perf event */ + struct perf_output_handle handle; /* AUX buffer output handle */ +}; +static DEFINE_PER_CPU(struct cpu_hw_sf, cpu_hw_sf); + +/* Debug feature */ +static debug_info_t *sfdbg; + +/* Sampling control helper functions */ +static inline unsigned long freq_to_sample_rate(struct hws_qsi_info_block *qsi, + unsigned long freq) +{ + return (USEC_PER_SEC / freq) * qsi->cpu_speed; +} + +static inline unsigned long sample_rate_to_freq(struct hws_qsi_info_block *qsi, + unsigned long rate) +{ + return USEC_PER_SEC * qsi->cpu_speed / rate; +} + +/* Return TOD timestamp contained in an trailer entry */ +static inline unsigned long long trailer_timestamp(struct hws_trailer_entry *te) +{ + /* TOD in STCKE format */ + if (te->header.t) + return *((unsigned long long *)&te->timestamp[1]); + + /* TOD in STCK format */ + return *((unsigned long long *)&te->timestamp[0]); +} + +/* Return pointer to trailer entry of an sample data block */ +static inline struct hws_trailer_entry *trailer_entry_ptr(unsigned long v) +{ + void *ret; + + ret = (void *)v; + ret += PAGE_SIZE; + ret -= sizeof(struct hws_trailer_entry); + + return ret; +} + +/* + * Return true if the entry in the sample data block table (sdbt) + * is a link to the next sdbt + */ +static inline int is_link_entry(unsigned long *s) +{ + return *s & 0x1UL ? 1 : 0; +} + +/* Return pointer to the linked sdbt */ +static inline unsigned long *get_next_sdbt(unsigned long *s) +{ + return phys_to_virt(*s & ~0x1UL); +} + +/* + * sf_disable() - Switch off sampling facility + */ +static int sf_disable(void) +{ + struct hws_lsctl_request_block sreq; + + memset(&sreq, 0, sizeof(sreq)); + return lsctl(&sreq); +} + +/* + * sf_buffer_available() - Check for an allocated sampling buffer + */ +static int sf_buffer_available(struct cpu_hw_sf *cpuhw) +{ + return !!cpuhw->sfb.sdbt; +} + +/* + * deallocate sampling facility buffer + */ +static void free_sampling_buffer(struct sf_buffer *sfb) +{ + unsigned long *sdbt, *curr; + + if (!sfb->sdbt) + return; + + sdbt = sfb->sdbt; + curr = sdbt; + + /* Free the SDBT after all SDBs are processed... */ + while (1) { + if (!*curr || !sdbt) + break; + + /* Process table-link entries */ + if (is_link_entry(curr)) { + curr = get_next_sdbt(curr); + if (sdbt) + free_page((unsigned long)sdbt); + + /* If the origin is reached, sampling buffer is freed */ + if (curr == sfb->sdbt) + break; + else + sdbt = curr; + } else { + /* Process SDB pointer */ + if (*curr) { + free_page((unsigned long)phys_to_virt(*curr)); + curr++; + } + } + } + + debug_sprintf_event(sfdbg, 5, "%s: freed sdbt %#lx\n", __func__, + (unsigned long)sfb->sdbt); + memset(sfb, 0, sizeof(*sfb)); +} + +static int alloc_sample_data_block(unsigned long *sdbt, gfp_t gfp_flags) +{ + struct hws_trailer_entry *te; + unsigned long sdb; + + /* Allocate and initialize sample-data-block */ + sdb = get_zeroed_page(gfp_flags); + if (!sdb) + return -ENOMEM; + te = trailer_entry_ptr(sdb); + te->header.a = 1; + + /* Link SDB into the sample-data-block-table */ + *sdbt = virt_to_phys((void *)sdb); + + return 0; +} + +/* + * realloc_sampling_buffer() - extend sampler memory + * + * Allocates new sample-data-blocks and adds them to the specified sampling + * buffer memory. + * + * Important: This modifies the sampling buffer and must be called when the + * sampling facility is disabled. + * + * Returns zero on success, non-zero otherwise. + */ +static int realloc_sampling_buffer(struct sf_buffer *sfb, + unsigned long num_sdb, gfp_t gfp_flags) +{ + int i, rc; + unsigned long *new, *tail, *tail_prev = NULL; + + if (!sfb->sdbt || !sfb->tail) + return -EINVAL; + + if (!is_link_entry(sfb->tail)) + return -EINVAL; + + /* Append to the existing sampling buffer, overwriting the table-link + * register. + * The tail variables always points to the "tail" (last and table-link) + * entry in an SDB-table. + */ + tail = sfb->tail; + + /* Do a sanity check whether the table-link entry points to + * the sampling buffer origin. + */ + if (sfb->sdbt != get_next_sdbt(tail)) { + debug_sprintf_event(sfdbg, 3, "%s: " + "sampling buffer is not linked: origin %#lx" + " tail %#lx\n", __func__, + (unsigned long)sfb->sdbt, + (unsigned long)tail); + return -EINVAL; + } + + /* Allocate remaining SDBs */ + rc = 0; + for (i = 0; i < num_sdb; i++) { + /* Allocate a new SDB-table if it is full. */ + if (require_table_link(tail)) { + new = (unsigned long *)get_zeroed_page(gfp_flags); + if (!new) { + rc = -ENOMEM; + break; + } + sfb->num_sdbt++; + /* Link current page to tail of chain */ + *tail = virt_to_phys((void *)new) + 1; + tail_prev = tail; + tail = new; + } + + /* Allocate a new sample-data-block. + * If there is not enough memory, stop the realloc process + * and simply use what was allocated. If this is a temporary + * issue, a new realloc call (if required) might succeed. + */ + rc = alloc_sample_data_block(tail, gfp_flags); + if (rc) { + /* Undo last SDBT. An SDBT with no SDB at its first + * entry but with an SDBT entry instead can not be + * handled by the interrupt handler code. + * Avoid this situation. + */ + if (tail_prev) { + sfb->num_sdbt--; + free_page((unsigned long)new); + tail = tail_prev; + } + break; + } + sfb->num_sdb++; + tail++; + tail_prev = new = NULL; /* Allocated at least one SBD */ + } + + /* Link sampling buffer to its origin */ + *tail = virt_to_phys(sfb->sdbt) + 1; + sfb->tail = tail; + + debug_sprintf_event(sfdbg, 4, "%s: new buffer" + " settings: sdbt %lu sdb %lu\n", __func__, + sfb->num_sdbt, sfb->num_sdb); + return rc; +} + +/* + * allocate_sampling_buffer() - allocate sampler memory + * + * Allocates and initializes a sampling buffer structure using the + * specified number of sample-data-blocks (SDB). For each allocation, + * a 4K page is used. The number of sample-data-block-tables (SDBT) + * are calculated from SDBs. + * Also set the ALERT_REQ mask in each SDBs trailer. + * + * Returns zero on success, non-zero otherwise. + */ +static int alloc_sampling_buffer(struct sf_buffer *sfb, unsigned long num_sdb) +{ + int rc; + + if (sfb->sdbt) + return -EINVAL; + + /* Allocate the sample-data-block-table origin */ + sfb->sdbt = (unsigned long *)get_zeroed_page(GFP_KERNEL); + if (!sfb->sdbt) + return -ENOMEM; + sfb->num_sdb = 0; + sfb->num_sdbt = 1; + + /* Link the table origin to point to itself to prepare for + * realloc_sampling_buffer() invocation. + */ + sfb->tail = sfb->sdbt; + *sfb->tail = virt_to_phys((void *)sfb->sdbt) + 1; + + /* Allocate requested number of sample-data-blocks */ + rc = realloc_sampling_buffer(sfb, num_sdb, GFP_KERNEL); + if (rc) { + free_sampling_buffer(sfb); + debug_sprintf_event(sfdbg, 4, "%s: " + "realloc_sampling_buffer failed with rc %i\n", + __func__, rc); + } else + debug_sprintf_event(sfdbg, 4, + "%s: tear %#lx dear %#lx\n", __func__, + (unsigned long)sfb->sdbt, (unsigned long)*sfb->sdbt); + return rc; +} + +static void sfb_set_limits(unsigned long min, unsigned long max) +{ + struct hws_qsi_info_block si; + + CPUM_SF_MIN_SDB = min; + CPUM_SF_MAX_SDB = max; + + memset(&si, 0, sizeof(si)); + if (!qsi(&si)) + CPUM_SF_SDB_DIAG_FACTOR = DIV_ROUND_UP(si.dsdes, si.bsdes); +} + +static unsigned long sfb_max_limit(struct hw_perf_event *hwc) +{ + return SAMPL_DIAG_MODE(hwc) ? CPUM_SF_MAX_SDB * CPUM_SF_SDB_DIAG_FACTOR + : CPUM_SF_MAX_SDB; +} + +static unsigned long sfb_pending_allocs(struct sf_buffer *sfb, + struct hw_perf_event *hwc) +{ + if (!sfb->sdbt) + return SFB_ALLOC_REG(hwc); + if (SFB_ALLOC_REG(hwc) > sfb->num_sdb) + return SFB_ALLOC_REG(hwc) - sfb->num_sdb; + return 0; +} + +static int sfb_has_pending_allocs(struct sf_buffer *sfb, + struct hw_perf_event *hwc) +{ + return sfb_pending_allocs(sfb, hwc) > 0; +} + +static void sfb_account_allocs(unsigned long num, struct hw_perf_event *hwc) +{ + /* Limit the number of SDBs to not exceed the maximum */ + num = min_t(unsigned long, num, sfb_max_limit(hwc) - SFB_ALLOC_REG(hwc)); + if (num) + SFB_ALLOC_REG(hwc) += num; +} + +static void sfb_init_allocs(unsigned long num, struct hw_perf_event *hwc) +{ + SFB_ALLOC_REG(hwc) = 0; + sfb_account_allocs(num, hwc); +} + +static void deallocate_buffers(struct cpu_hw_sf *cpuhw) +{ + if (cpuhw->sfb.sdbt) + free_sampling_buffer(&cpuhw->sfb); +} + +static int allocate_buffers(struct cpu_hw_sf *cpuhw, struct hw_perf_event *hwc) +{ + unsigned long n_sdb, freq; + size_t sample_size; + + /* Calculate sampling buffers using 4K pages + * + * 1. The sampling size is 32 bytes for basic sampling. This size + * is the same for all machine types. Diagnostic + * sampling uses auxlilary data buffer setup which provides the + * memory for SDBs using linux common code auxiliary trace + * setup. + * + * 2. Function alloc_sampling_buffer() sets the Alert Request + * Control indicator to trigger a measurement-alert to harvest + * sample-data-blocks (SDB). This is done per SDB. This + * measurement alert interrupt fires quick enough to handle + * one SDB, on very high frequency and work loads there might + * be 2 to 3 SBDs available for sample processing. + * Currently there is no need for setup alert request on every + * n-th page. This is counterproductive as one IRQ triggers + * a very high number of samples to be processed at one IRQ. + * + * 3. Use the sampling frequency as input. + * Compute the number of SDBs and ensure a minimum + * of CPUM_SF_MIN_SDB. Depending on frequency add some more + * SDBs to handle a higher sampling rate. + * Use a minimum of CPUM_SF_MIN_SDB and allow for 100 samples + * (one SDB) for every 10000 HZ frequency increment. + * + * 4. Compute the number of sample-data-block-tables (SDBT) and + * ensure a minimum of CPUM_SF_MIN_SDBT (one table can manage up + * to 511 SDBs). + */ + sample_size = sizeof(struct hws_basic_entry); + freq = sample_rate_to_freq(&cpuhw->qsi, SAMPL_RATE(hwc)); + n_sdb = CPUM_SF_MIN_SDB + DIV_ROUND_UP(freq, 10000); + + /* If there is already a sampling buffer allocated, it is very likely + * that the sampling facility is enabled too. If the event to be + * initialized requires a greater sampling buffer, the allocation must + * be postponed. Changing the sampling buffer requires the sampling + * facility to be in the disabled state. So, account the number of + * required SDBs and let cpumsf_pmu_enable() resize the buffer just + * before the event is started. + */ + sfb_init_allocs(n_sdb, hwc); + if (sf_buffer_available(cpuhw)) + return 0; + + debug_sprintf_event(sfdbg, 3, + "%s: rate %lu f %lu sdb %lu/%lu" + " sample_size %lu cpuhw %p\n", __func__, + SAMPL_RATE(hwc), freq, n_sdb, sfb_max_limit(hwc), + sample_size, cpuhw); + + return alloc_sampling_buffer(&cpuhw->sfb, + sfb_pending_allocs(&cpuhw->sfb, hwc)); +} + +static unsigned long min_percent(unsigned int percent, unsigned long base, + unsigned long min) +{ + return min_t(unsigned long, min, DIV_ROUND_UP(percent * base, 100)); +} + +static unsigned long compute_sfb_extent(unsigned long ratio, unsigned long base) +{ + /* Use a percentage-based approach to extend the sampling facility + * buffer. Accept up to 5% sample data loss. + * Vary the extents between 1% to 5% of the current number of + * sample-data-blocks. + */ + if (ratio <= 5) + return 0; + if (ratio <= 25) + return min_percent(1, base, 1); + if (ratio <= 50) + return min_percent(1, base, 1); + if (ratio <= 75) + return min_percent(2, base, 2); + if (ratio <= 100) + return min_percent(3, base, 3); + if (ratio <= 250) + return min_percent(4, base, 4); + + return min_percent(5, base, 8); +} + +static void sfb_account_overflows(struct cpu_hw_sf *cpuhw, + struct hw_perf_event *hwc) +{ + unsigned long ratio, num; + + if (!OVERFLOW_REG(hwc)) + return; + + /* The sample_overflow contains the average number of sample data + * that has been lost because sample-data-blocks were full. + * + * Calculate the total number of sample data entries that has been + * discarded. Then calculate the ratio of lost samples to total samples + * per second in percent. + */ + ratio = DIV_ROUND_UP(100 * OVERFLOW_REG(hwc) * cpuhw->sfb.num_sdb, + sample_rate_to_freq(&cpuhw->qsi, SAMPL_RATE(hwc))); + + /* Compute number of sample-data-blocks */ + num = compute_sfb_extent(ratio, cpuhw->sfb.num_sdb); + if (num) + sfb_account_allocs(num, hwc); + + debug_sprintf_event(sfdbg, 5, "%s: overflow %llu ratio %lu num %lu\n", + __func__, OVERFLOW_REG(hwc), ratio, num); + OVERFLOW_REG(hwc) = 0; +} + +/* extend_sampling_buffer() - Extend sampling buffer + * @sfb: Sampling buffer structure (for local CPU) + * @hwc: Perf event hardware structure + * + * Use this function to extend the sampling buffer based on the overflow counter + * and postponed allocation extents stored in the specified Perf event hardware. + * + * Important: This function disables the sampling facility in order to safely + * change the sampling buffer structure. Do not call this function + * when the PMU is active. + */ +static void extend_sampling_buffer(struct sf_buffer *sfb, + struct hw_perf_event *hwc) +{ + unsigned long num, num_old; + int rc; + + num = sfb_pending_allocs(sfb, hwc); + if (!num) + return; + num_old = sfb->num_sdb; + + /* Disable the sampling facility to reset any states and also + * clear pending measurement alerts. + */ + sf_disable(); + + /* Extend the sampling buffer. + * This memory allocation typically happens in an atomic context when + * called by perf. Because this is a reallocation, it is fine if the + * new SDB-request cannot be satisfied immediately. + */ + rc = realloc_sampling_buffer(sfb, num, GFP_ATOMIC); + if (rc) + debug_sprintf_event(sfdbg, 5, "%s: realloc failed with rc %i\n", + __func__, rc); + + if (sfb_has_pending_allocs(sfb, hwc)) + debug_sprintf_event(sfdbg, 5, "%s: " + "req %lu alloc %lu remaining %lu\n", + __func__, num, sfb->num_sdb - num_old, + sfb_pending_allocs(sfb, hwc)); +} + +/* Number of perf events counting hardware events */ +static atomic_t num_events; +/* Used to avoid races in calling reserve/release_cpumf_hardware */ +static DEFINE_MUTEX(pmc_reserve_mutex); + +#define PMC_INIT 0 +#define PMC_RELEASE 1 +#define PMC_FAILURE 2 +static void setup_pmc_cpu(void *flags) +{ + struct cpu_hw_sf *cpusf = this_cpu_ptr(&cpu_hw_sf); + int err = 0; + + switch (*((int *)flags)) { + case PMC_INIT: + memset(cpusf, 0, sizeof(*cpusf)); + err = qsi(&cpusf->qsi); + if (err) + break; + cpusf->flags |= PMU_F_RESERVED; + err = sf_disable(); + break; + case PMC_RELEASE: + cpusf->flags &= ~PMU_F_RESERVED; + err = sf_disable(); + if (!err) + deallocate_buffers(cpusf); + break; + } + if (err) { + *((int *)flags) |= PMC_FAILURE; + pr_err("Switching off the sampling facility failed with rc %i\n", err); + } +} + +static void release_pmc_hardware(void) +{ + int flags = PMC_RELEASE; + + irq_subclass_unregister(IRQ_SUBCLASS_MEASUREMENT_ALERT); + on_each_cpu(setup_pmc_cpu, &flags, 1); +} + +static int reserve_pmc_hardware(void) +{ + int flags = PMC_INIT; + + on_each_cpu(setup_pmc_cpu, &flags, 1); + if (flags & PMC_FAILURE) { + release_pmc_hardware(); + return -ENODEV; + } + irq_subclass_register(IRQ_SUBCLASS_MEASUREMENT_ALERT); + + return 0; +} + +static void hw_perf_event_destroy(struct perf_event *event) +{ + /* Release PMC if this is the last perf event */ + if (!atomic_add_unless(&num_events, -1, 1)) { + mutex_lock(&pmc_reserve_mutex); + if (atomic_dec_return(&num_events) == 0) + release_pmc_hardware(); + mutex_unlock(&pmc_reserve_mutex); + } +} + +static void hw_init_period(struct hw_perf_event *hwc, u64 period) +{ + hwc->sample_period = period; + hwc->last_period = hwc->sample_period; + local64_set(&hwc->period_left, hwc->sample_period); +} + +static unsigned long hw_limit_rate(const struct hws_qsi_info_block *si, + unsigned long rate) +{ + return clamp_t(unsigned long, rate, + si->min_sampl_rate, si->max_sampl_rate); +} + +static u32 cpumsf_pid_type(struct perf_event *event, + u32 pid, enum pid_type type) +{ + struct task_struct *tsk; + + /* Idle process */ + if (!pid) + goto out; + + tsk = find_task_by_pid_ns(pid, &init_pid_ns); + pid = -1; + if (tsk) { + /* + * Only top level events contain the pid namespace in which + * they are created. + */ + if (event->parent) + event = event->parent; + pid = __task_pid_nr_ns(tsk, type, event->ns); + /* + * See also 1d953111b648 + * "perf/core: Don't report zero PIDs for exiting tasks". + */ + if (!pid && !pid_alive(tsk)) + pid = -1; + } +out: + return pid; +} + +static void cpumsf_output_event_pid(struct perf_event *event, + struct perf_sample_data *data, + struct pt_regs *regs) +{ + u32 pid; + struct perf_event_header header; + struct perf_output_handle handle; + + /* + * Obtain the PID from the basic-sampling data entry and + * correct the data->tid_entry.pid value. + */ + pid = data->tid_entry.pid; + + /* Protect callchain buffers, tasks */ + rcu_read_lock(); + + perf_prepare_sample(data, event, regs); + perf_prepare_header(&header, data, event, regs); + if (perf_output_begin(&handle, data, event, header.size)) + goto out; + + /* Update the process ID (see also kernel/events/core.c) */ + data->tid_entry.pid = cpumsf_pid_type(event, pid, PIDTYPE_TGID); + data->tid_entry.tid = cpumsf_pid_type(event, pid, PIDTYPE_PID); + + perf_output_sample(&handle, &header, data, event); + perf_output_end(&handle); +out: + rcu_read_unlock(); +} + +static unsigned long getrate(bool freq, unsigned long sample, + struct hws_qsi_info_block *si) +{ + unsigned long rate; + + if (freq) { + rate = freq_to_sample_rate(si, sample); + rate = hw_limit_rate(si, rate); + } else { + /* The min/max sampling rates specifies the valid range + * of sample periods. If the specified sample period is + * out of range, limit the period to the range boundary. + */ + rate = hw_limit_rate(si, sample); + + /* The perf core maintains a maximum sample rate that is + * configurable through the sysctl interface. Ensure the + * sampling rate does not exceed this value. This also helps + * to avoid throttling when pushing samples with + * perf_event_overflow(). + */ + if (sample_rate_to_freq(si, rate) > + sysctl_perf_event_sample_rate) { + debug_sprintf_event(sfdbg, 1, "%s: " + "Sampling rate exceeds maximum " + "perf sample rate\n", __func__); + rate = 0; + } + } + return rate; +} + +/* The sampling information (si) contains information about the + * min/max sampling intervals and the CPU speed. So calculate the + * correct sampling interval and avoid the whole period adjust + * feedback loop. + * + * Since the CPU Measurement sampling facility can not handle frequency + * calculate the sampling interval when frequency is specified using + * this formula: + * interval := cpu_speed * 1000000 / sample_freq + * + * Returns errno on bad input and zero on success with parameter interval + * set to the correct sampling rate. + * + * Note: This function turns off freq bit to avoid calling function + * perf_adjust_period(). This causes frequency adjustment in the common + * code part which causes tremendous variations in the counter values. + */ +static int __hw_perf_event_init_rate(struct perf_event *event, + struct hws_qsi_info_block *si) +{ + struct perf_event_attr *attr = &event->attr; + struct hw_perf_event *hwc = &event->hw; + unsigned long rate; + + if (attr->freq) { + if (!attr->sample_freq) + return -EINVAL; + rate = getrate(attr->freq, attr->sample_freq, si); + attr->freq = 0; /* Don't call perf_adjust_period() */ + SAMPL_FLAGS(hwc) |= PERF_CPUM_SF_FREQ_MODE; + } else { + rate = getrate(attr->freq, attr->sample_period, si); + if (!rate) + return -EINVAL; + } + attr->sample_period = rate; + SAMPL_RATE(hwc) = rate; + hw_init_period(hwc, SAMPL_RATE(hwc)); + debug_sprintf_event(sfdbg, 4, "%s: cpu %d period %#llx freq %d,%#lx\n", + __func__, event->cpu, event->attr.sample_period, + event->attr.freq, SAMPLE_FREQ_MODE(hwc)); + return 0; +} + +static int __hw_perf_event_init(struct perf_event *event) +{ + struct cpu_hw_sf *cpuhw; + struct hws_qsi_info_block si; + struct perf_event_attr *attr = &event->attr; + struct hw_perf_event *hwc = &event->hw; + int cpu, err; + + /* Reserve CPU-measurement sampling facility */ + err = 0; + if (!atomic_inc_not_zero(&num_events)) { + mutex_lock(&pmc_reserve_mutex); + if (atomic_read(&num_events) == 0 && reserve_pmc_hardware()) + err = -EBUSY; + else + atomic_inc(&num_events); + mutex_unlock(&pmc_reserve_mutex); + } + event->destroy = hw_perf_event_destroy; + + if (err) + goto out; + + /* Access per-CPU sampling information (query sampling info) */ + /* + * The event->cpu value can be -1 to count on every CPU, for example, + * when attaching to a task. If this is specified, use the query + * sampling info from the current CPU, otherwise use event->cpu to + * retrieve the per-CPU information. + * Later, cpuhw indicates whether to allocate sampling buffers for a + * particular CPU (cpuhw!=NULL) or each online CPU (cpuw==NULL). + */ + memset(&si, 0, sizeof(si)); + cpuhw = NULL; + if (event->cpu == -1) + qsi(&si); + else { + /* Event is pinned to a particular CPU, retrieve the per-CPU + * sampling structure for accessing the CPU-specific QSI. + */ + cpuhw = &per_cpu(cpu_hw_sf, event->cpu); + si = cpuhw->qsi; + } + + /* Check sampling facility authorization and, if not authorized, + * fall back to other PMUs. It is safe to check any CPU because + * the authorization is identical for all configured CPUs. + */ + if (!si.as) { + err = -ENOENT; + goto out; + } + + if (si.ribm & CPU_MF_SF_RIBM_NOTAV) { + pr_warn("CPU Measurement Facility sampling is temporarily not available\n"); + err = -EBUSY; + goto out; + } + + /* Always enable basic sampling */ + SAMPL_FLAGS(hwc) = PERF_CPUM_SF_BASIC_MODE; + + /* Check if diagnostic sampling is requested. Deny if the required + * sampling authorization is missing. + */ + if (attr->config == PERF_EVENT_CPUM_SF_DIAG) { + if (!si.ad) { + err = -EPERM; + goto out; + } + SAMPL_FLAGS(hwc) |= PERF_CPUM_SF_DIAG_MODE; + } + + err = __hw_perf_event_init_rate(event, &si); + if (err) + goto out; + + /* Initialize sample data overflow accounting */ + hwc->extra_reg.reg = REG_OVERFLOW; + OVERFLOW_REG(hwc) = 0; + + /* Use AUX buffer. No need to allocate it by ourself */ + if (attr->config == PERF_EVENT_CPUM_SF_DIAG) + return 0; + + /* Allocate the per-CPU sampling buffer using the CPU information + * from the event. If the event is not pinned to a particular + * CPU (event->cpu == -1; or cpuhw == NULL), allocate sampling + * buffers for each online CPU. + */ + if (cpuhw) + /* Event is pinned to a particular CPU */ + err = allocate_buffers(cpuhw, hwc); + else { + /* Event is not pinned, allocate sampling buffer on + * each online CPU + */ + for_each_online_cpu(cpu) { + cpuhw = &per_cpu(cpu_hw_sf, cpu); + err = allocate_buffers(cpuhw, hwc); + if (err) + break; + } + } + + /* If PID/TID sampling is active, replace the default overflow + * handler to extract and resolve the PIDs from the basic-sampling + * data entries. + */ + if (event->attr.sample_type & PERF_SAMPLE_TID) + if (is_default_overflow_handler(event)) + event->overflow_handler = cpumsf_output_event_pid; +out: + return err; +} + +static bool is_callchain_event(struct perf_event *event) +{ + u64 sample_type = event->attr.sample_type; + + return sample_type & (PERF_SAMPLE_CALLCHAIN | PERF_SAMPLE_REGS_USER | + PERF_SAMPLE_STACK_USER); +} + +static int cpumsf_pmu_event_init(struct perf_event *event) +{ + int err; + + /* No support for taken branch sampling */ + /* No support for callchain, stacks and registers */ + if (has_branch_stack(event) || is_callchain_event(event)) + return -EOPNOTSUPP; + + switch (event->attr.type) { + case PERF_TYPE_RAW: + if ((event->attr.config != PERF_EVENT_CPUM_SF) && + (event->attr.config != PERF_EVENT_CPUM_SF_DIAG)) + return -ENOENT; + break; + case PERF_TYPE_HARDWARE: + /* Support sampling of CPU cycles in addition to the + * counter facility. However, the counter facility + * is more precise and, hence, restrict this PMU to + * sampling events only. + */ + if (event->attr.config != PERF_COUNT_HW_CPU_CYCLES) + return -ENOENT; + if (!is_sampling_event(event)) + return -ENOENT; + break; + default: + return -ENOENT; + } + + /* Force reset of idle/hv excludes regardless of what the + * user requested. + */ + if (event->attr.exclude_hv) + event->attr.exclude_hv = 0; + if (event->attr.exclude_idle) + event->attr.exclude_idle = 0; + + err = __hw_perf_event_init(event); + if (unlikely(err)) + if (event->destroy) + event->destroy(event); + return err; +} + +static void cpumsf_pmu_enable(struct pmu *pmu) +{ + struct cpu_hw_sf *cpuhw = this_cpu_ptr(&cpu_hw_sf); + struct hw_perf_event *hwc; + int err; + + if (cpuhw->flags & PMU_F_ENABLED) + return; + + if (cpuhw->flags & PMU_F_ERR_MASK) + return; + + /* Check whether to extent the sampling buffer. + * + * Two conditions trigger an increase of the sampling buffer for a + * perf event: + * 1. Postponed buffer allocations from the event initialization. + * 2. Sampling overflows that contribute to pending allocations. + * + * Note that the extend_sampling_buffer() function disables the sampling + * facility, but it can be fully re-enabled using sampling controls that + * have been saved in cpumsf_pmu_disable(). + */ + if (cpuhw->event) { + hwc = &cpuhw->event->hw; + if (!(SAMPL_DIAG_MODE(hwc))) { + /* + * Account number of overflow-designated + * buffer extents + */ + sfb_account_overflows(cpuhw, hwc); + extend_sampling_buffer(&cpuhw->sfb, hwc); + } + /* Rate may be adjusted with ioctl() */ + cpuhw->lsctl.interval = SAMPL_RATE(&cpuhw->event->hw); + } + + /* (Re)enable the PMU and sampling facility */ + cpuhw->flags |= PMU_F_ENABLED; + barrier(); + + err = lsctl(&cpuhw->lsctl); + if (err) { + cpuhw->flags &= ~PMU_F_ENABLED; + pr_err("Loading sampling controls failed: op 1 err %i\n", err); + return; + } + + /* Load current program parameter */ + lpp(&S390_lowcore.lpp); + + debug_sprintf_event(sfdbg, 6, "%s: es %i cs %i ed %i cd %i " + "interval %#lx tear %#lx dear %#lx\n", __func__, + cpuhw->lsctl.es, cpuhw->lsctl.cs, cpuhw->lsctl.ed, + cpuhw->lsctl.cd, cpuhw->lsctl.interval, + cpuhw->lsctl.tear, cpuhw->lsctl.dear); +} + +static void cpumsf_pmu_disable(struct pmu *pmu) +{ + struct cpu_hw_sf *cpuhw = this_cpu_ptr(&cpu_hw_sf); + struct hws_lsctl_request_block inactive; + struct hws_qsi_info_block si; + int err; + + if (!(cpuhw->flags & PMU_F_ENABLED)) + return; + + if (cpuhw->flags & PMU_F_ERR_MASK) + return; + + /* Switch off sampling activation control */ + inactive = cpuhw->lsctl; + inactive.cs = 0; + inactive.cd = 0; + + err = lsctl(&inactive); + if (err) { + pr_err("Loading sampling controls failed: op 2 err %i\n", err); + return; + } + + /* Save state of TEAR and DEAR register contents */ + err = qsi(&si); + if (!err) { + /* TEAR/DEAR values are valid only if the sampling facility is + * enabled. Note that cpumsf_pmu_disable() might be called even + * for a disabled sampling facility because cpumsf_pmu_enable() + * controls the enable/disable state. + */ + if (si.es) { + cpuhw->lsctl.tear = si.tear; + cpuhw->lsctl.dear = si.dear; + } + } else + debug_sprintf_event(sfdbg, 3, "%s: qsi() failed with err %i\n", + __func__, err); + + cpuhw->flags &= ~PMU_F_ENABLED; +} + +/* perf_exclude_event() - Filter event + * @event: The perf event + * @regs: pt_regs structure + * @sde_regs: Sample-data-entry (sde) regs structure + * + * Filter perf events according to their exclude specification. + * + * Return non-zero if the event shall be excluded. + */ +static int perf_exclude_event(struct perf_event *event, struct pt_regs *regs, + struct perf_sf_sde_regs *sde_regs) +{ + if (event->attr.exclude_user && user_mode(regs)) + return 1; + if (event->attr.exclude_kernel && !user_mode(regs)) + return 1; + if (event->attr.exclude_guest && sde_regs->in_guest) + return 1; + if (event->attr.exclude_host && !sde_regs->in_guest) + return 1; + return 0; +} + +/* perf_push_sample() - Push samples to perf + * @event: The perf event + * @sample: Hardware sample data + * + * Use the hardware sample data to create perf event sample. The sample + * is the pushed to the event subsystem and the function checks for + * possible event overflows. If an event overflow occurs, the PMU is + * stopped. + * + * Return non-zero if an event overflow occurred. + */ +static int perf_push_sample(struct perf_event *event, + struct hws_basic_entry *basic) +{ + int overflow; + struct pt_regs regs; + struct perf_sf_sde_regs *sde_regs; + struct perf_sample_data data; + + /* Setup perf sample */ + perf_sample_data_init(&data, 0, event->hw.last_period); + + /* Setup pt_regs to look like an CPU-measurement external interrupt + * using the Program Request Alert code. The regs.int_parm_long + * field which is unused contains additional sample-data-entry related + * indicators. + */ + memset(®s, 0, sizeof(regs)); + regs.int_code = 0x1407; + regs.int_parm = CPU_MF_INT_SF_PRA; + sde_regs = (struct perf_sf_sde_regs *) ®s.int_parm_long; + + psw_bits(regs.psw).ia = basic->ia; + psw_bits(regs.psw).dat = basic->T; + psw_bits(regs.psw).wait = basic->W; + psw_bits(regs.psw).pstate = basic->P; + psw_bits(regs.psw).as = basic->AS; + + /* + * Use the hardware provided configuration level to decide if the + * sample belongs to a guest or host. If that is not available, + * fall back to the following heuristics: + * A non-zero guest program parameter always indicates a guest + * sample. Some early samples or samples from guests without + * lpp usage would be misaccounted to the host. We use the asn + * value as an addon heuristic to detect most of these guest samples. + * If the value differs from 0xffff (the host value), we assume to + * be a KVM guest. + */ + switch (basic->CL) { + case 1: /* logical partition */ + sde_regs->in_guest = 0; + break; + case 2: /* virtual machine */ + sde_regs->in_guest = 1; + break; + default: /* old machine, use heuristics */ + if (basic->gpp || basic->prim_asn != 0xffff) + sde_regs->in_guest = 1; + break; + } + + /* + * Store the PID value from the sample-data-entry to be + * processed and resolved by cpumsf_output_event_pid(). + */ + data.tid_entry.pid = basic->hpp & LPP_PID_MASK; + + overflow = 0; + if (perf_exclude_event(event, ®s, sde_regs)) + goto out; + if (perf_event_overflow(event, &data, ®s)) { + overflow = 1; + event->pmu->stop(event, 0); + } + perf_event_update_userpage(event); +out: + return overflow; +} + +static void perf_event_count_update(struct perf_event *event, u64 count) +{ + local64_add(count, &event->count); +} + +/* hw_collect_samples() - Walk through a sample-data-block and collect samples + * @event: The perf event + * @sdbt: Sample-data-block table + * @overflow: Event overflow counter + * + * Walks through a sample-data-block and collects sampling data entries that are + * then pushed to the perf event subsystem. Depending on the sampling function, + * there can be either basic-sampling or combined-sampling data entries. A + * combined-sampling data entry consists of a basic- and a diagnostic-sampling + * data entry. The sampling function is determined by the flags in the perf + * event hardware structure. The function always works with a combined-sampling + * data entry but ignores the the diagnostic portion if it is not available. + * + * Note that the implementation focuses on basic-sampling data entries and, if + * such an entry is not valid, the entire combined-sampling data entry is + * ignored. + * + * The overflow variables counts the number of samples that has been discarded + * due to a perf event overflow. + */ +static void hw_collect_samples(struct perf_event *event, unsigned long *sdbt, + unsigned long long *overflow) +{ + struct hws_trailer_entry *te; + struct hws_basic_entry *sample; + + te = trailer_entry_ptr((unsigned long)sdbt); + sample = (struct hws_basic_entry *)sdbt; + while ((unsigned long *)sample < (unsigned long *)te) { + /* Check for an empty sample */ + if (!sample->def || sample->LS) + break; + + /* Update perf event period */ + perf_event_count_update(event, SAMPL_RATE(&event->hw)); + + /* Check whether sample is valid */ + if (sample->def == 0x0001) { + /* If an event overflow occurred, the PMU is stopped to + * throttle event delivery. Remaining sample data is + * discarded. + */ + if (!*overflow) { + /* Check whether sample is consistent */ + if (sample->I == 0 && sample->W == 0) { + /* Deliver sample data to perf */ + *overflow = perf_push_sample(event, + sample); + } + } else + /* Count discarded samples */ + *overflow += 1; + } else { + debug_sprintf_event(sfdbg, 4, + "%s: Found unknown" + " sampling data entry: te->f %i" + " basic.def %#4x (%p)\n", __func__, + te->header.f, sample->def, sample); + /* Sample slot is not yet written or other record. + * + * This condition can occur if the buffer was reused + * from a combined basic- and diagnostic-sampling. + * If only basic-sampling is then active, entries are + * written into the larger diagnostic entries. + * This is typically the case for sample-data-blocks + * that are not full. Stop processing if the first + * invalid format was detected. + */ + if (!te->header.f) + break; + } + + /* Reset sample slot and advance to next sample */ + sample->def = 0; + sample++; + } +} + +/* hw_perf_event_update() - Process sampling buffer + * @event: The perf event + * @flush_all: Flag to also flush partially filled sample-data-blocks + * + * Processes the sampling buffer and create perf event samples. + * The sampling buffer position are retrieved and saved in the TEAR_REG + * register of the specified perf event. + * + * Only full sample-data-blocks are processed. Specify the flush_all flag + * to also walk through partially filled sample-data-blocks. + */ +static void hw_perf_event_update(struct perf_event *event, int flush_all) +{ + unsigned long long event_overflow, sampl_overflow, num_sdb; + union hws_trailer_header old, prev, new; + struct hw_perf_event *hwc = &event->hw; + struct hws_trailer_entry *te; + unsigned long *sdbt, sdb; + int done; + + /* + * AUX buffer is used when in diagnostic sampling mode. + * No perf events/samples are created. + */ + if (SAMPL_DIAG_MODE(&event->hw)) + return; + + sdbt = (unsigned long *)TEAR_REG(hwc); + done = event_overflow = sampl_overflow = num_sdb = 0; + while (!done) { + /* Get the trailer entry of the sample-data-block */ + sdb = (unsigned long)phys_to_virt(*sdbt); + te = trailer_entry_ptr(sdb); + + /* Leave loop if no more work to do (block full indicator) */ + if (!te->header.f) { + done = 1; + if (!flush_all) + break; + } + + /* Check the sample overflow count */ + if (te->header.overflow) + /* Account sample overflows and, if a particular limit + * is reached, extend the sampling buffer. + * For details, see sfb_account_overflows(). + */ + sampl_overflow += te->header.overflow; + + /* Timestamps are valid for full sample-data-blocks only */ + debug_sprintf_event(sfdbg, 6, "%s: sdbt %#lx/%#lx " + "overflow %llu timestamp %#llx\n", + __func__, sdb, (unsigned long)sdbt, + te->header.overflow, + (te->header.f) ? trailer_timestamp(te) : 0ULL); + + /* Collect all samples from a single sample-data-block and + * flag if an (perf) event overflow happened. If so, the PMU + * is stopped and remaining samples will be discarded. + */ + hw_collect_samples(event, (unsigned long *)sdb, &event_overflow); + num_sdb++; + + /* Reset trailer (using compare-double-and-swap) */ + prev.val = READ_ONCE_ALIGNED_128(te->header.val); + do { + old.val = prev.val; + new.val = prev.val; + new.f = 0; + new.a = 1; + new.overflow = 0; + prev.val = cmpxchg128(&te->header.val, old.val, new.val); + } while (prev.val != old.val); + + /* Advance to next sample-data-block */ + sdbt++; + if (is_link_entry(sdbt)) + sdbt = get_next_sdbt(sdbt); + + /* Update event hardware registers */ + TEAR_REG(hwc) = (unsigned long) sdbt; + + /* Stop processing sample-data if all samples of the current + * sample-data-block were flushed even if it was not full. + */ + if (flush_all && done) + break; + } + + /* Account sample overflows in the event hardware structure */ + if (sampl_overflow) + OVERFLOW_REG(hwc) = DIV_ROUND_UP(OVERFLOW_REG(hwc) + + sampl_overflow, 1 + num_sdb); + + /* Perf_event_overflow() and perf_event_account_interrupt() limit + * the interrupt rate to an upper limit. Roughly 1000 samples per + * task tick. + * Hitting this limit results in a large number + * of throttled REF_REPORT_THROTTLE entries and the samples + * are dropped. + * Slightly increase the interval to avoid hitting this limit. + */ + if (event_overflow) { + SAMPL_RATE(hwc) += DIV_ROUND_UP(SAMPL_RATE(hwc), 10); + debug_sprintf_event(sfdbg, 1, "%s: rate adjustment %ld\n", + __func__, + DIV_ROUND_UP(SAMPL_RATE(hwc), 10)); + } + + if (sampl_overflow || event_overflow) + debug_sprintf_event(sfdbg, 4, "%s: " + "overflows: sample %llu event %llu" + " total %llu num_sdb %llu\n", + __func__, sampl_overflow, event_overflow, + OVERFLOW_REG(hwc), num_sdb); +} + +static inline unsigned long aux_sdb_index(struct aux_buffer *aux, + unsigned long i) +{ + return i % aux->sfb.num_sdb; +} + +static inline unsigned long aux_sdb_num(unsigned long start, unsigned long end) +{ + return end >= start ? end - start + 1 : 0; +} + +static inline unsigned long aux_sdb_num_alert(struct aux_buffer *aux) +{ + return aux_sdb_num(aux->head, aux->alert_mark); +} + +static inline unsigned long aux_sdb_num_empty(struct aux_buffer *aux) +{ + return aux_sdb_num(aux->head, aux->empty_mark); +} + +/* + * Get trailer entry by index of SDB. + */ +static struct hws_trailer_entry *aux_sdb_trailer(struct aux_buffer *aux, + unsigned long index) +{ + unsigned long sdb; + + index = aux_sdb_index(aux, index); + sdb = aux->sdb_index[index]; + return trailer_entry_ptr(sdb); +} + +/* + * Finish sampling on the cpu. Called by cpumsf_pmu_del() with pmu + * disabled. Collect the full SDBs in AUX buffer which have not reached + * the point of alert indicator. And ignore the SDBs which are not + * full. + * + * 1. Scan SDBs to see how much data is there and consume them. + * 2. Remove alert indicator in the buffer. + */ +static void aux_output_end(struct perf_output_handle *handle) +{ + unsigned long i, range_scan, idx; + struct aux_buffer *aux; + struct hws_trailer_entry *te; + + aux = perf_get_aux(handle); + if (!aux) + return; + + range_scan = aux_sdb_num_alert(aux); + for (i = 0, idx = aux->head; i < range_scan; i++, idx++) { + te = aux_sdb_trailer(aux, idx); + if (!te->header.f) + break; + } + /* i is num of SDBs which are full */ + perf_aux_output_end(handle, i << PAGE_SHIFT); + + /* Remove alert indicators in the buffer */ + te = aux_sdb_trailer(aux, aux->alert_mark); + te->header.a = 0; + + debug_sprintf_event(sfdbg, 6, "%s: SDBs %ld range %ld head %ld\n", + __func__, i, range_scan, aux->head); +} + +/* + * Start sampling on the CPU. Called by cpumsf_pmu_add() when an event + * is first added to the CPU or rescheduled again to the CPU. It is called + * with pmu disabled. + * + * 1. Reset the trailer of SDBs to get ready for new data. + * 2. Tell the hardware where to put the data by reset the SDBs buffer + * head(tear/dear). + */ +static int aux_output_begin(struct perf_output_handle *handle, + struct aux_buffer *aux, + struct cpu_hw_sf *cpuhw) +{ + unsigned long range, i, range_scan, idx, head, base, offset; + struct hws_trailer_entry *te; + + if (WARN_ON_ONCE(handle->head & ~PAGE_MASK)) + return -EINVAL; + + aux->head = handle->head >> PAGE_SHIFT; + range = (handle->size + 1) >> PAGE_SHIFT; + if (range <= 1) + return -ENOMEM; + + /* + * SDBs between aux->head and aux->empty_mark are already ready + * for new data. range_scan is num of SDBs not within them. + */ + debug_sprintf_event(sfdbg, 6, + "%s: range %ld head %ld alert %ld empty %ld\n", + __func__, range, aux->head, aux->alert_mark, + aux->empty_mark); + if (range > aux_sdb_num_empty(aux)) { + range_scan = range - aux_sdb_num_empty(aux); + idx = aux->empty_mark + 1; + for (i = 0; i < range_scan; i++, idx++) { + te = aux_sdb_trailer(aux, idx); + te->header.f = 0; + te->header.a = 0; + te->header.overflow = 0; + } + /* Save the position of empty SDBs */ + aux->empty_mark = aux->head + range - 1; + } + + /* Set alert indicator */ + aux->alert_mark = aux->head + range/2 - 1; + te = aux_sdb_trailer(aux, aux->alert_mark); + te->header.a = 1; + + /* Reset hardware buffer head */ + head = aux_sdb_index(aux, aux->head); + base = aux->sdbt_index[head / CPUM_SF_SDB_PER_TABLE]; + offset = head % CPUM_SF_SDB_PER_TABLE; + cpuhw->lsctl.tear = virt_to_phys((void *)base) + offset * sizeof(unsigned long); + cpuhw->lsctl.dear = virt_to_phys((void *)aux->sdb_index[head]); + + debug_sprintf_event(sfdbg, 6, "%s: head %ld alert %ld empty %ld " + "index %ld tear %#lx dear %#lx\n", __func__, + aux->head, aux->alert_mark, aux->empty_mark, + head / CPUM_SF_SDB_PER_TABLE, + cpuhw->lsctl.tear, cpuhw->lsctl.dear); + + return 0; +} + +/* + * Set alert indicator on SDB at index @alert_index while sampler is running. + * + * Return true if successfully. + * Return false if full indicator is already set by hardware sampler. + */ +static bool aux_set_alert(struct aux_buffer *aux, unsigned long alert_index, + unsigned long long *overflow) +{ + union hws_trailer_header old, prev, new; + struct hws_trailer_entry *te; + + te = aux_sdb_trailer(aux, alert_index); + prev.val = READ_ONCE_ALIGNED_128(te->header.val); + do { + old.val = prev.val; + new.val = prev.val; + *overflow = old.overflow; + if (old.f) { + /* + * SDB is already set by hardware. + * Abort and try to set somewhere + * behind. + */ + return false; + } + new.a = 1; + new.overflow = 0; + prev.val = cmpxchg128(&te->header.val, old.val, new.val); + } while (prev.val != old.val); + return true; +} + +/* + * aux_reset_buffer() - Scan and setup SDBs for new samples + * @aux: The AUX buffer to set + * @range: The range of SDBs to scan started from aux->head + * @overflow: Set to overflow count + * + * Set alert indicator on the SDB at index of aux->alert_mark. If this SDB is + * marked as empty, check if it is already set full by the hardware sampler. + * If yes, that means new data is already there before we can set an alert + * indicator. Caller should try to set alert indicator to some position behind. + * + * Scan the SDBs in AUX buffer from behind aux->empty_mark. They are used + * previously and have already been consumed by user space. Reset these SDBs + * (clear full indicator and alert indicator) for new data. + * If aux->alert_mark fall in this area, just set it. Overflow count is + * recorded while scanning. + * + * SDBs between aux->head and aux->empty_mark are already reset at last time. + * and ready for new samples. So scanning on this area could be skipped. + * + * Return true if alert indicator is set successfully and false if not. + */ +static bool aux_reset_buffer(struct aux_buffer *aux, unsigned long range, + unsigned long long *overflow) +{ + unsigned long i, range_scan, idx, idx_old; + union hws_trailer_header old, prev, new; + unsigned long long orig_overflow; + struct hws_trailer_entry *te; + + debug_sprintf_event(sfdbg, 6, "%s: range %ld head %ld alert %ld " + "empty %ld\n", __func__, range, aux->head, + aux->alert_mark, aux->empty_mark); + if (range <= aux_sdb_num_empty(aux)) + /* + * No need to scan. All SDBs in range are marked as empty. + * Just set alert indicator. Should check race with hardware + * sampler. + */ + return aux_set_alert(aux, aux->alert_mark, overflow); + + if (aux->alert_mark <= aux->empty_mark) + /* + * Set alert indicator on empty SDB. Should check race + * with hardware sampler. + */ + if (!aux_set_alert(aux, aux->alert_mark, overflow)) + return false; + + /* + * Scan the SDBs to clear full and alert indicator used previously. + * Start scanning from one SDB behind empty_mark. If the new alert + * indicator fall into this range, set it. + */ + range_scan = range - aux_sdb_num_empty(aux); + idx_old = idx = aux->empty_mark + 1; + for (i = 0; i < range_scan; i++, idx++) { + te = aux_sdb_trailer(aux, idx); + prev.val = READ_ONCE_ALIGNED_128(te->header.val); + do { + old.val = prev.val; + new.val = prev.val; + orig_overflow = old.overflow; + new.f = 0; + new.overflow = 0; + if (idx == aux->alert_mark) + new.a = 1; + else + new.a = 0; + prev.val = cmpxchg128(&te->header.val, old.val, new.val); + } while (prev.val != old.val); + *overflow += orig_overflow; + } + + /* Update empty_mark to new position */ + aux->empty_mark = aux->head + range - 1; + + debug_sprintf_event(sfdbg, 6, "%s: range_scan %ld idx %ld..%ld " + "empty %ld\n", __func__, range_scan, idx_old, + idx - 1, aux->empty_mark); + return true; +} + +/* + * Measurement alert handler for diagnostic mode sampling. + */ +static void hw_collect_aux(struct cpu_hw_sf *cpuhw) +{ + struct aux_buffer *aux; + int done = 0; + unsigned long range = 0, size; + unsigned long long overflow = 0; + struct perf_output_handle *handle = &cpuhw->handle; + unsigned long num_sdb; + + aux = perf_get_aux(handle); + if (WARN_ON_ONCE(!aux)) + return; + + /* Inform user space new data arrived */ + size = aux_sdb_num_alert(aux) << PAGE_SHIFT; + debug_sprintf_event(sfdbg, 6, "%s: #alert %ld\n", __func__, + size >> PAGE_SHIFT); + perf_aux_output_end(handle, size); + + num_sdb = aux->sfb.num_sdb; + while (!done) { + /* Get an output handle */ + aux = perf_aux_output_begin(handle, cpuhw->event); + if (handle->size == 0) { + pr_err("The AUX buffer with %lu pages for the " + "diagnostic-sampling mode is full\n", + num_sdb); + break; + } + if (WARN_ON_ONCE(!aux)) + return; + + /* Update head and alert_mark to new position */ + aux->head = handle->head >> PAGE_SHIFT; + range = (handle->size + 1) >> PAGE_SHIFT; + if (range == 1) + aux->alert_mark = aux->head; + else + aux->alert_mark = aux->head + range/2 - 1; + + if (aux_reset_buffer(aux, range, &overflow)) { + if (!overflow) { + done = 1; + break; + } + size = range << PAGE_SHIFT; + perf_aux_output_end(&cpuhw->handle, size); + pr_err("Sample data caused the AUX buffer with %lu " + "pages to overflow\n", aux->sfb.num_sdb); + debug_sprintf_event(sfdbg, 1, "%s: head %ld range %ld " + "overflow %lld\n", __func__, + aux->head, range, overflow); + } else { + size = aux_sdb_num_alert(aux) << PAGE_SHIFT; + perf_aux_output_end(&cpuhw->handle, size); + debug_sprintf_event(sfdbg, 6, "%s: head %ld alert %ld " + "already full, try another\n", + __func__, + aux->head, aux->alert_mark); + } + } + + if (done) + debug_sprintf_event(sfdbg, 6, "%s: head %ld alert %ld " + "empty %ld\n", __func__, aux->head, + aux->alert_mark, aux->empty_mark); +} + +/* + * Callback when freeing AUX buffers. + */ +static void aux_buffer_free(void *data) +{ + struct aux_buffer *aux = data; + unsigned long i, num_sdbt; + + if (!aux) + return; + + /* Free SDBT. SDB is freed by the caller */ + num_sdbt = aux->sfb.num_sdbt; + for (i = 0; i < num_sdbt; i++) + free_page(aux->sdbt_index[i]); + + kfree(aux->sdbt_index); + kfree(aux->sdb_index); + kfree(aux); + + debug_sprintf_event(sfdbg, 4, "%s: SDBTs %lu\n", __func__, num_sdbt); +} + +static void aux_sdb_init(unsigned long sdb) +{ + struct hws_trailer_entry *te; + + te = trailer_entry_ptr(sdb); + + /* Save clock base */ + te->clock_base = 1; + te->progusage2 = tod_clock_base.tod; +} + +/* + * aux_buffer_setup() - Setup AUX buffer for diagnostic mode sampling + * @event: Event the buffer is setup for, event->cpu == -1 means current + * @pages: Array of pointers to buffer pages passed from perf core + * @nr_pages: Total pages + * @snapshot: Flag for snapshot mode + * + * This is the callback when setup an event using AUX buffer. Perf tool can + * trigger this by an additional mmap() call on the event. Unlike the buffer + * for basic samples, AUX buffer belongs to the event. It is scheduled with + * the task among online cpus when it is a per-thread event. + * + * Return the private AUX buffer structure if success or NULL if fails. + */ +static void *aux_buffer_setup(struct perf_event *event, void **pages, + int nr_pages, bool snapshot) +{ + struct sf_buffer *sfb; + struct aux_buffer *aux; + unsigned long *new, *tail; + int i, n_sdbt; + + if (!nr_pages || !pages) + return NULL; + + if (nr_pages > CPUM_SF_MAX_SDB * CPUM_SF_SDB_DIAG_FACTOR) { + pr_err("AUX buffer size (%i pages) is larger than the " + "maximum sampling buffer limit\n", + nr_pages); + return NULL; + } else if (nr_pages < CPUM_SF_MIN_SDB * CPUM_SF_SDB_DIAG_FACTOR) { + pr_err("AUX buffer size (%i pages) is less than the " + "minimum sampling buffer limit\n", + nr_pages); + return NULL; + } + + /* Allocate aux_buffer struct for the event */ + aux = kzalloc(sizeof(struct aux_buffer), GFP_KERNEL); + if (!aux) + goto no_aux; + sfb = &aux->sfb; + + /* Allocate sdbt_index for fast reference */ + n_sdbt = DIV_ROUND_UP(nr_pages, CPUM_SF_SDB_PER_TABLE); + aux->sdbt_index = kmalloc_array(n_sdbt, sizeof(void *), GFP_KERNEL); + if (!aux->sdbt_index) + goto no_sdbt_index; + + /* Allocate sdb_index for fast reference */ + aux->sdb_index = kmalloc_array(nr_pages, sizeof(void *), GFP_KERNEL); + if (!aux->sdb_index) + goto no_sdb_index; + + /* Allocate the first SDBT */ + sfb->num_sdbt = 0; + sfb->sdbt = (unsigned long *)get_zeroed_page(GFP_KERNEL); + if (!sfb->sdbt) + goto no_sdbt; + aux->sdbt_index[sfb->num_sdbt++] = (unsigned long)sfb->sdbt; + tail = sfb->tail = sfb->sdbt; + + /* + * Link the provided pages of AUX buffer to SDBT. + * Allocate SDBT if needed. + */ + for (i = 0; i < nr_pages; i++, tail++) { + if (require_table_link(tail)) { + new = (unsigned long *)get_zeroed_page(GFP_KERNEL); + if (!new) + goto no_sdbt; + aux->sdbt_index[sfb->num_sdbt++] = (unsigned long)new; + /* Link current page to tail of chain */ + *tail = virt_to_phys(new) + 1; + tail = new; + } + /* Tail is the entry in a SDBT */ + *tail = virt_to_phys(pages[i]); + aux->sdb_index[i] = (unsigned long)pages[i]; + aux_sdb_init((unsigned long)pages[i]); + } + sfb->num_sdb = nr_pages; + + /* Link the last entry in the SDBT to the first SDBT */ + *tail = virt_to_phys(sfb->sdbt) + 1; + sfb->tail = tail; + + /* + * Initial all SDBs are zeroed. Mark it as empty. + * So there is no need to clear the full indicator + * when this event is first added. + */ + aux->empty_mark = sfb->num_sdb - 1; + + debug_sprintf_event(sfdbg, 4, "%s: SDBTs %lu SDBs %lu\n", __func__, + sfb->num_sdbt, sfb->num_sdb); + + return aux; + +no_sdbt: + /* SDBs (AUX buffer pages) are freed by caller */ + for (i = 0; i < sfb->num_sdbt; i++) + free_page(aux->sdbt_index[i]); + kfree(aux->sdb_index); +no_sdb_index: + kfree(aux->sdbt_index); +no_sdbt_index: + kfree(aux); +no_aux: + return NULL; +} + +static void cpumsf_pmu_read(struct perf_event *event) +{ + /* Nothing to do ... updates are interrupt-driven */ +} + +/* Check if the new sampling period/frequency is appropriate. + * + * Return non-zero on error and zero on passed checks. + */ +static int cpumsf_pmu_check_period(struct perf_event *event, u64 value) +{ + struct hws_qsi_info_block si; + unsigned long rate; + bool do_freq; + + memset(&si, 0, sizeof(si)); + if (event->cpu == -1) { + if (qsi(&si)) + return -ENODEV; + } else { + /* Event is pinned to a particular CPU, retrieve the per-CPU + * sampling structure for accessing the CPU-specific QSI. + */ + struct cpu_hw_sf *cpuhw = &per_cpu(cpu_hw_sf, event->cpu); + + si = cpuhw->qsi; + } + + do_freq = !!SAMPLE_FREQ_MODE(&event->hw); + rate = getrate(do_freq, value, &si); + if (!rate) + return -EINVAL; + + event->attr.sample_period = rate; + SAMPL_RATE(&event->hw) = rate; + hw_init_period(&event->hw, SAMPL_RATE(&event->hw)); + debug_sprintf_event(sfdbg, 4, "%s:" + " cpu %d value %#llx period %#llx freq %d\n", + __func__, event->cpu, value, + event->attr.sample_period, do_freq); + return 0; +} + +/* Activate sampling control. + * Next call of pmu_enable() starts sampling. + */ +static void cpumsf_pmu_start(struct perf_event *event, int flags) +{ + struct cpu_hw_sf *cpuhw = this_cpu_ptr(&cpu_hw_sf); + + if (WARN_ON_ONCE(!(event->hw.state & PERF_HES_STOPPED))) + return; + + if (flags & PERF_EF_RELOAD) + WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE)); + + perf_pmu_disable(event->pmu); + event->hw.state = 0; + cpuhw->lsctl.cs = 1; + if (SAMPL_DIAG_MODE(&event->hw)) + cpuhw->lsctl.cd = 1; + perf_pmu_enable(event->pmu); +} + +/* Deactivate sampling control. + * Next call of pmu_enable() stops sampling. + */ +static void cpumsf_pmu_stop(struct perf_event *event, int flags) +{ + struct cpu_hw_sf *cpuhw = this_cpu_ptr(&cpu_hw_sf); + + if (event->hw.state & PERF_HES_STOPPED) + return; + + perf_pmu_disable(event->pmu); + cpuhw->lsctl.cs = 0; + cpuhw->lsctl.cd = 0; + event->hw.state |= PERF_HES_STOPPED; + + if ((flags & PERF_EF_UPDATE) && !(event->hw.state & PERF_HES_UPTODATE)) { + hw_perf_event_update(event, 1); + event->hw.state |= PERF_HES_UPTODATE; + } + perf_pmu_enable(event->pmu); +} + +static int cpumsf_pmu_add(struct perf_event *event, int flags) +{ + struct cpu_hw_sf *cpuhw = this_cpu_ptr(&cpu_hw_sf); + struct aux_buffer *aux; + int err; + + if (cpuhw->flags & PMU_F_IN_USE) + return -EAGAIN; + + if (!SAMPL_DIAG_MODE(&event->hw) && !cpuhw->sfb.sdbt) + return -EINVAL; + + err = 0; + perf_pmu_disable(event->pmu); + + event->hw.state = PERF_HES_UPTODATE | PERF_HES_STOPPED; + + /* Set up sampling controls. Always program the sampling register + * using the SDB-table start. Reset TEAR_REG event hardware register + * that is used by hw_perf_event_update() to store the sampling buffer + * position after samples have been flushed. + */ + cpuhw->lsctl.s = 0; + cpuhw->lsctl.h = 1; + cpuhw->lsctl.interval = SAMPL_RATE(&event->hw); + if (!SAMPL_DIAG_MODE(&event->hw)) { + cpuhw->lsctl.tear = virt_to_phys(cpuhw->sfb.sdbt); + cpuhw->lsctl.dear = *(unsigned long *)cpuhw->sfb.sdbt; + TEAR_REG(&event->hw) = (unsigned long)cpuhw->sfb.sdbt; + } + + /* Ensure sampling functions are in the disabled state. If disabled, + * switch on sampling enable control. */ + if (WARN_ON_ONCE(cpuhw->lsctl.es == 1 || cpuhw->lsctl.ed == 1)) { + err = -EAGAIN; + goto out; + } + if (SAMPL_DIAG_MODE(&event->hw)) { + aux = perf_aux_output_begin(&cpuhw->handle, event); + if (!aux) { + err = -EINVAL; + goto out; + } + err = aux_output_begin(&cpuhw->handle, aux, cpuhw); + if (err) + goto out; + cpuhw->lsctl.ed = 1; + } + cpuhw->lsctl.es = 1; + + /* Set in_use flag and store event */ + cpuhw->event = event; + cpuhw->flags |= PMU_F_IN_USE; + + if (flags & PERF_EF_START) + cpumsf_pmu_start(event, PERF_EF_RELOAD); +out: + perf_event_update_userpage(event); + perf_pmu_enable(event->pmu); + return err; +} + +static void cpumsf_pmu_del(struct perf_event *event, int flags) +{ + struct cpu_hw_sf *cpuhw = this_cpu_ptr(&cpu_hw_sf); + + perf_pmu_disable(event->pmu); + cpumsf_pmu_stop(event, PERF_EF_UPDATE); + + cpuhw->lsctl.es = 0; + cpuhw->lsctl.ed = 0; + cpuhw->flags &= ~PMU_F_IN_USE; + cpuhw->event = NULL; + + if (SAMPL_DIAG_MODE(&event->hw)) + aux_output_end(&cpuhw->handle); + perf_event_update_userpage(event); + perf_pmu_enable(event->pmu); +} + +CPUMF_EVENT_ATTR(SF, SF_CYCLES_BASIC, PERF_EVENT_CPUM_SF); +CPUMF_EVENT_ATTR(SF, SF_CYCLES_BASIC_DIAG, PERF_EVENT_CPUM_SF_DIAG); + +/* Attribute list for CPU_SF. + * + * The availablitiy depends on the CPU_MF sampling facility authorization + * for basic + diagnositic samples. This is determined at initialization + * time by the sampling facility device driver. + * If the authorization for basic samples is turned off, it should be + * also turned off for diagnostic sampling. + * + * During initialization of the device driver, check the authorization + * level for diagnostic sampling and installs the attribute + * file for diagnostic sampling if necessary. + * + * For now install a placeholder to reference all possible attributes: + * SF_CYCLES_BASIC and SF_CYCLES_BASIC_DIAG. + * Add another entry for the final NULL pointer. + */ +enum { + SF_CYCLES_BASIC_ATTR_IDX = 0, + SF_CYCLES_BASIC_DIAG_ATTR_IDX, + SF_CYCLES_ATTR_MAX +}; + +static struct attribute *cpumsf_pmu_events_attr[SF_CYCLES_ATTR_MAX + 1] = { + [SF_CYCLES_BASIC_ATTR_IDX] = CPUMF_EVENT_PTR(SF, SF_CYCLES_BASIC) +}; + +PMU_FORMAT_ATTR(event, "config:0-63"); + +static struct attribute *cpumsf_pmu_format_attr[] = { + &format_attr_event.attr, + NULL, +}; + +static struct attribute_group cpumsf_pmu_events_group = { + .name = "events", + .attrs = cpumsf_pmu_events_attr, +}; + +static struct attribute_group cpumsf_pmu_format_group = { + .name = "format", + .attrs = cpumsf_pmu_format_attr, +}; + +static const struct attribute_group *cpumsf_pmu_attr_groups[] = { + &cpumsf_pmu_events_group, + &cpumsf_pmu_format_group, + NULL, +}; + +static struct pmu cpumf_sampling = { + .pmu_enable = cpumsf_pmu_enable, + .pmu_disable = cpumsf_pmu_disable, + + .event_init = cpumsf_pmu_event_init, + .add = cpumsf_pmu_add, + .del = cpumsf_pmu_del, + + .start = cpumsf_pmu_start, + .stop = cpumsf_pmu_stop, + .read = cpumsf_pmu_read, + + .attr_groups = cpumsf_pmu_attr_groups, + + .setup_aux = aux_buffer_setup, + .free_aux = aux_buffer_free, + + .check_period = cpumsf_pmu_check_period, +}; + +static void cpumf_measurement_alert(struct ext_code ext_code, + unsigned int alert, unsigned long unused) +{ + struct cpu_hw_sf *cpuhw; + + if (!(alert & CPU_MF_INT_SF_MASK)) + return; + inc_irq_stat(IRQEXT_CMS); + cpuhw = this_cpu_ptr(&cpu_hw_sf); + + /* Measurement alerts are shared and might happen when the PMU + * is not reserved. Ignore these alerts in this case. */ + if (!(cpuhw->flags & PMU_F_RESERVED)) + return; + + /* The processing below must take care of multiple alert events that + * might be indicated concurrently. */ + + /* Program alert request */ + if (alert & CPU_MF_INT_SF_PRA) { + if (cpuhw->flags & PMU_F_IN_USE) + if (SAMPL_DIAG_MODE(&cpuhw->event->hw)) + hw_collect_aux(cpuhw); + else + hw_perf_event_update(cpuhw->event, 0); + else + WARN_ON_ONCE(!(cpuhw->flags & PMU_F_IN_USE)); + } + + /* Report measurement alerts only for non-PRA codes */ + if (alert != CPU_MF_INT_SF_PRA) + debug_sprintf_event(sfdbg, 6, "%s: alert %#x\n", __func__, + alert); + + /* Sampling authorization change request */ + if (alert & CPU_MF_INT_SF_SACA) + qsi(&cpuhw->qsi); + + /* Loss of sample data due to high-priority machine activities */ + if (alert & CPU_MF_INT_SF_LSDA) { + pr_err("Sample data was lost\n"); + cpuhw->flags |= PMU_F_ERR_LSDA; + sf_disable(); + } + + /* Invalid sampling buffer entry */ + if (alert & (CPU_MF_INT_SF_IAE|CPU_MF_INT_SF_ISE)) { + pr_err("A sampling buffer entry is incorrect (alert=0x%x)\n", + alert); + cpuhw->flags |= PMU_F_ERR_IBE; + sf_disable(); + } +} + +static int cpusf_pmu_setup(unsigned int cpu, int flags) +{ + /* Ignore the notification if no events are scheduled on the PMU. + * This might be racy... + */ + if (!atomic_read(&num_events)) + return 0; + + local_irq_disable(); + setup_pmc_cpu(&flags); + local_irq_enable(); + return 0; +} + +static int s390_pmu_sf_online_cpu(unsigned int cpu) +{ + return cpusf_pmu_setup(cpu, PMC_INIT); +} + +static int s390_pmu_sf_offline_cpu(unsigned int cpu) +{ + return cpusf_pmu_setup(cpu, PMC_RELEASE); +} + +static int param_get_sfb_size(char *buffer, const struct kernel_param *kp) +{ + if (!cpum_sf_avail()) + return -ENODEV; + return sprintf(buffer, "%lu,%lu", CPUM_SF_MIN_SDB, CPUM_SF_MAX_SDB); +} + +static int param_set_sfb_size(const char *val, const struct kernel_param *kp) +{ + int rc; + unsigned long min, max; + + if (!cpum_sf_avail()) + return -ENODEV; + if (!val || !strlen(val)) + return -EINVAL; + + /* Valid parameter values: "min,max" or "max" */ + min = CPUM_SF_MIN_SDB; + max = CPUM_SF_MAX_SDB; + if (strchr(val, ',')) + rc = (sscanf(val, "%lu,%lu", &min, &max) == 2) ? 0 : -EINVAL; + else + rc = kstrtoul(val, 10, &max); + + if (min < 2 || min >= max || max > get_num_physpages()) + rc = -EINVAL; + if (rc) + return rc; + + sfb_set_limits(min, max); + pr_info("The sampling buffer limits have changed to: " + "min %lu max %lu (diag %lu)\n", + CPUM_SF_MIN_SDB, CPUM_SF_MAX_SDB, CPUM_SF_SDB_DIAG_FACTOR); + return 0; +} + +#define param_check_sfb_size(name, p) __param_check(name, p, void) +static const struct kernel_param_ops param_ops_sfb_size = { + .set = param_set_sfb_size, + .get = param_get_sfb_size, +}; + +#define RS_INIT_FAILURE_QSI 0x0001 +#define RS_INIT_FAILURE_BSDES 0x0002 +#define RS_INIT_FAILURE_ALRT 0x0003 +#define RS_INIT_FAILURE_PERF 0x0004 +static void __init pr_cpumsf_err(unsigned int reason) +{ + pr_err("Sampling facility support for perf is not available: " + "reason %#x\n", reason); +} + +static int __init init_cpum_sampling_pmu(void) +{ + struct hws_qsi_info_block si; + int err; + + if (!cpum_sf_avail()) + return -ENODEV; + + memset(&si, 0, sizeof(si)); + if (qsi(&si)) { + pr_cpumsf_err(RS_INIT_FAILURE_QSI); + return -ENODEV; + } + + if (!si.as && !si.ad) + return -ENODEV; + + if (si.bsdes != sizeof(struct hws_basic_entry)) { + pr_cpumsf_err(RS_INIT_FAILURE_BSDES); + return -EINVAL; + } + + if (si.ad) { + sfb_set_limits(CPUM_SF_MIN_SDB, CPUM_SF_MAX_SDB); + /* Sampling of diagnostic data authorized, + * install event into attribute list of PMU device. + */ + cpumsf_pmu_events_attr[SF_CYCLES_BASIC_DIAG_ATTR_IDX] = + CPUMF_EVENT_PTR(SF, SF_CYCLES_BASIC_DIAG); + } + + sfdbg = debug_register(KMSG_COMPONENT, 2, 1, 80); + if (!sfdbg) { + pr_err("Registering for s390dbf failed\n"); + return -ENOMEM; + } + debug_register_view(sfdbg, &debug_sprintf_view); + + err = register_external_irq(EXT_IRQ_MEASURE_ALERT, + cpumf_measurement_alert); + if (err) { + pr_cpumsf_err(RS_INIT_FAILURE_ALRT); + debug_unregister(sfdbg); + goto out; + } + + err = perf_pmu_register(&cpumf_sampling, "cpum_sf", PERF_TYPE_RAW); + if (err) { + pr_cpumsf_err(RS_INIT_FAILURE_PERF); + unregister_external_irq(EXT_IRQ_MEASURE_ALERT, + cpumf_measurement_alert); + debug_unregister(sfdbg); + goto out; + } + + cpuhp_setup_state(CPUHP_AP_PERF_S390_SF_ONLINE, "perf/s390/sf:online", + s390_pmu_sf_online_cpu, s390_pmu_sf_offline_cpu); +out: + return err; +} + +arch_initcall(init_cpum_sampling_pmu); +core_param(cpum_sfb_size, CPUM_SF_MAX_SDB, sfb_size, 0644); diff --git a/arch/s390/kernel/perf_event.c b/arch/s390/kernel/perf_event.c new file mode 100644 index 0000000000..c27321cb09 --- /dev/null +++ b/arch/s390/kernel/perf_event.c @@ -0,0 +1,223 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Performance event support for s390x + * + * Copyright IBM Corp. 2012, 2013 + * Author(s): Hendrik Brueckner <brueckner@linux.vnet.ibm.com> + */ +#define KMSG_COMPONENT "perf" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/kernel.h> +#include <linux/perf_event.h> +#include <linux/kvm_host.h> +#include <linux/percpu.h> +#include <linux/export.h> +#include <linux/seq_file.h> +#include <linux/spinlock.h> +#include <linux/sysfs.h> +#include <asm/irq.h> +#include <asm/cpu_mf.h> +#include <asm/lowcore.h> +#include <asm/processor.h> +#include <asm/sysinfo.h> +#include <asm/unwind.h> + +static struct kvm_s390_sie_block *sie_block(struct pt_regs *regs) +{ + struct stack_frame *stack = (struct stack_frame *) regs->gprs[15]; + + if (!stack) + return NULL; + + return (struct kvm_s390_sie_block *)stack->sie_control_block; +} + +static bool is_in_guest(struct pt_regs *regs) +{ + if (user_mode(regs)) + return false; +#if IS_ENABLED(CONFIG_KVM) + return instruction_pointer(regs) == (unsigned long) &sie_exit; +#else + return false; +#endif +} + +static unsigned long guest_is_user_mode(struct pt_regs *regs) +{ + return sie_block(regs)->gpsw.mask & PSW_MASK_PSTATE; +} + +static unsigned long instruction_pointer_guest(struct pt_regs *regs) +{ + return sie_block(regs)->gpsw.addr; +} + +unsigned long perf_instruction_pointer(struct pt_regs *regs) +{ + return is_in_guest(regs) ? instruction_pointer_guest(regs) + : instruction_pointer(regs); +} + +static unsigned long perf_misc_guest_flags(struct pt_regs *regs) +{ + return guest_is_user_mode(regs) ? PERF_RECORD_MISC_GUEST_USER + : PERF_RECORD_MISC_GUEST_KERNEL; +} + +static unsigned long perf_misc_flags_sf(struct pt_regs *regs) +{ + struct perf_sf_sde_regs *sde_regs; + unsigned long flags; + + sde_regs = (struct perf_sf_sde_regs *) ®s->int_parm_long; + if (sde_regs->in_guest) + flags = user_mode(regs) ? PERF_RECORD_MISC_GUEST_USER + : PERF_RECORD_MISC_GUEST_KERNEL; + else + flags = user_mode(regs) ? PERF_RECORD_MISC_USER + : PERF_RECORD_MISC_KERNEL; + return flags; +} + +unsigned long perf_misc_flags(struct pt_regs *regs) +{ + /* Check if the cpum_sf PMU has created the pt_regs structure. + * In this case, perf misc flags can be easily extracted. Otherwise, + * do regular checks on the pt_regs content. + */ + if (regs->int_code == 0x1407 && regs->int_parm == CPU_MF_INT_SF_PRA) + if (!regs->gprs[15]) + return perf_misc_flags_sf(regs); + + if (is_in_guest(regs)) + return perf_misc_guest_flags(regs); + + return user_mode(regs) ? PERF_RECORD_MISC_USER + : PERF_RECORD_MISC_KERNEL; +} + +static void print_debug_cf(void) +{ + struct cpumf_ctr_info cf_info; + int cpu = smp_processor_id(); + + memset(&cf_info, 0, sizeof(cf_info)); + if (!qctri(&cf_info)) + pr_info("CPU[%i] CPUM_CF: ver=%u.%u A=%04x E=%04x C=%04x\n", + cpu, cf_info.cfvn, cf_info.csvn, + cf_info.auth_ctl, cf_info.enable_ctl, cf_info.act_ctl); +} + +static void print_debug_sf(void) +{ + struct hws_qsi_info_block si; + int cpu = smp_processor_id(); + + memset(&si, 0, sizeof(si)); + if (qsi(&si)) + return; + + pr_info("CPU[%i] CPUM_SF: basic=%i diag=%i min=%lu max=%lu cpu_speed=%u\n", + cpu, si.as, si.ad, si.min_sampl_rate, si.max_sampl_rate, + si.cpu_speed); + + if (si.as) + pr_info("CPU[%i] CPUM_SF: Basic-sampling: a=%i e=%i c=%i" + " bsdes=%i tear=%016lx dear=%016lx\n", cpu, + si.as, si.es, si.cs, si.bsdes, si.tear, si.dear); + if (si.ad) + pr_info("CPU[%i] CPUM_SF: Diagnostic-sampling: a=%i e=%i c=%i" + " dsdes=%i tear=%016lx dear=%016lx\n", cpu, + si.ad, si.ed, si.cd, si.dsdes, si.tear, si.dear); +} + +void perf_event_print_debug(void) +{ + unsigned long flags; + + local_irq_save(flags); + if (cpum_cf_avail()) + print_debug_cf(); + if (cpum_sf_avail()) + print_debug_sf(); + local_irq_restore(flags); +} + +/* Service level infrastructure */ +static void sl_print_counter(struct seq_file *m) +{ + struct cpumf_ctr_info ci; + + memset(&ci, 0, sizeof(ci)); + if (qctri(&ci)) + return; + + seq_printf(m, "CPU-MF: Counter facility: version=%u.%u " + "authorization=%04x\n", ci.cfvn, ci.csvn, ci.auth_ctl); +} + +static void sl_print_sampling(struct seq_file *m) +{ + struct hws_qsi_info_block si; + + memset(&si, 0, sizeof(si)); + if (qsi(&si)) + return; + + if (!si.as && !si.ad) + return; + + seq_printf(m, "CPU-MF: Sampling facility: min_rate=%lu max_rate=%lu" + " cpu_speed=%u\n", si.min_sampl_rate, si.max_sampl_rate, + si.cpu_speed); + if (si.as) + seq_printf(m, "CPU-MF: Sampling facility: mode=basic" + " sample_size=%u\n", si.bsdes); + if (si.ad) + seq_printf(m, "CPU-MF: Sampling facility: mode=diagnostic" + " sample_size=%u\n", si.dsdes); +} + +static void service_level_perf_print(struct seq_file *m, + struct service_level *sl) +{ + if (cpum_cf_avail()) + sl_print_counter(m); + if (cpum_sf_avail()) + sl_print_sampling(m); +} + +static struct service_level service_level_perf = { + .seq_print = service_level_perf_print, +}; + +static int __init service_level_perf_register(void) +{ + return register_service_level(&service_level_perf); +} +arch_initcall(service_level_perf_register); + +void perf_callchain_kernel(struct perf_callchain_entry_ctx *entry, + struct pt_regs *regs) +{ + struct unwind_state state; + unsigned long addr; + + unwind_for_each_frame(&state, current, regs, 0) { + addr = unwind_get_return_address(&state); + if (!addr || perf_callchain_store(entry, addr)) + return; + } +} + +/* Perf definitions for PMU event attributes in sysfs */ +ssize_t cpumf_events_sysfs_show(struct device *dev, + struct device_attribute *attr, char *page) +{ + struct perf_pmu_events_attr *pmu_attr; + + pmu_attr = container_of(attr, struct perf_pmu_events_attr, attr); + return sprintf(page, "event=0x%04llx\n", pmu_attr->id); +} diff --git a/arch/s390/kernel/perf_pai_crypto.c b/arch/s390/kernel/perf_pai_crypto.c new file mode 100644 index 0000000000..fe7d1774de --- /dev/null +++ b/arch/s390/kernel/perf_pai_crypto.c @@ -0,0 +1,698 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Performance event support - Processor Activity Instrumentation Facility + * + * Copyright IBM Corp. 2022 + * Author(s): Thomas Richter <tmricht@linux.ibm.com> + */ +#define KMSG_COMPONENT "pai_crypto" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/kernel.h> +#include <linux/kernel_stat.h> +#include <linux/percpu.h> +#include <linux/notifier.h> +#include <linux/init.h> +#include <linux/export.h> +#include <linux/io.h> +#include <linux/perf_event.h> + +#include <asm/ctl_reg.h> +#include <asm/pai.h> +#include <asm/debug.h> + +static debug_info_t *cfm_dbg; +static unsigned int paicrypt_cnt; /* Size of the mapped counter sets */ + /* extracted with QPACI instruction */ + +DEFINE_STATIC_KEY_FALSE(pai_key); + +struct pai_userdata { + u16 num; + u64 value; +} __packed; + +struct paicrypt_map { + unsigned long *page; /* Page for CPU to store counters */ + struct pai_userdata *save; /* Page to store no-zero counters */ + unsigned int active_events; /* # of PAI crypto users */ + refcount_t refcnt; /* Reference count mapped buffers */ + enum paievt_mode mode; /* Type of event */ + struct perf_event *event; /* Perf event for sampling */ +}; + +static DEFINE_PER_CPU(struct paicrypt_map, paicrypt_map); + +/* Release the PMU if event is the last perf event */ +static DEFINE_MUTEX(pai_reserve_mutex); + +/* Adjust usage counters and remove allocated memory when all users are + * gone. + */ +static void paicrypt_event_destroy(struct perf_event *event) +{ + struct paicrypt_map *cpump = per_cpu_ptr(&paicrypt_map, event->cpu); + + cpump->event = NULL; + static_branch_dec(&pai_key); + mutex_lock(&pai_reserve_mutex); + debug_sprintf_event(cfm_dbg, 5, "%s event %#llx cpu %d users %d" + " mode %d refcnt %u\n", __func__, + event->attr.config, event->cpu, + cpump->active_events, cpump->mode, + refcount_read(&cpump->refcnt)); + if (refcount_dec_and_test(&cpump->refcnt)) { + debug_sprintf_event(cfm_dbg, 4, "%s page %#lx save %p\n", + __func__, (unsigned long)cpump->page, + cpump->save); + free_page((unsigned long)cpump->page); + cpump->page = NULL; + kvfree(cpump->save); + cpump->save = NULL; + cpump->mode = PAI_MODE_NONE; + } + mutex_unlock(&pai_reserve_mutex); +} + +static u64 paicrypt_getctr(struct paicrypt_map *cpump, int nr, bool kernel) +{ + if (kernel) + nr += PAI_CRYPTO_MAXCTR; + return cpump->page[nr]; +} + +/* Read the counter values. Return value from location in CMP. For event + * CRYPTO_ALL sum up all events. + */ +static u64 paicrypt_getdata(struct perf_event *event, bool kernel) +{ + struct paicrypt_map *cpump = this_cpu_ptr(&paicrypt_map); + u64 sum = 0; + int i; + + if (event->attr.config != PAI_CRYPTO_BASE) { + return paicrypt_getctr(cpump, + event->attr.config - PAI_CRYPTO_BASE, + kernel); + } + + for (i = 1; i <= paicrypt_cnt; i++) { + u64 val = paicrypt_getctr(cpump, i, kernel); + + if (!val) + continue; + sum += val; + } + return sum; +} + +static u64 paicrypt_getall(struct perf_event *event) +{ + u64 sum = 0; + + if (!event->attr.exclude_kernel) + sum += paicrypt_getdata(event, true); + if (!event->attr.exclude_user) + sum += paicrypt_getdata(event, false); + + return sum; +} + +/* Used to avoid races in checking concurrent access of counting and + * sampling for crypto events + * + * Only one instance of event pai_crypto/CRYPTO_ALL/ for sampling is + * allowed and when this event is running, no counting event is allowed. + * Several counting events are allowed in parallel, but no sampling event + * is allowed while one (or more) counting events are running. + * + * This function is called in process context and it is save to block. + * When the event initialization functions fails, no other call back will + * be invoked. + * + * Allocate the memory for the event. + */ +static int paicrypt_busy(struct perf_event_attr *a, struct paicrypt_map *cpump) +{ + int rc = 0; + + mutex_lock(&pai_reserve_mutex); + if (a->sample_period) { /* Sampling requested */ + if (cpump->mode != PAI_MODE_NONE) + rc = -EBUSY; /* ... sampling/counting active */ + } else { /* Counting requested */ + if (cpump->mode == PAI_MODE_SAMPLING) + rc = -EBUSY; /* ... and sampling active */ + } + if (rc) + goto unlock; + + /* Allocate memory for counter page and counter extraction. + * Only the first counting event has to allocate a page. + */ + if (cpump->page) { + refcount_inc(&cpump->refcnt); + goto unlock; + } + + rc = -ENOMEM; + cpump->page = (unsigned long *)get_zeroed_page(GFP_KERNEL); + if (!cpump->page) + goto unlock; + cpump->save = kvmalloc_array(paicrypt_cnt + 1, + sizeof(struct pai_userdata), GFP_KERNEL); + if (!cpump->save) { + free_page((unsigned long)cpump->page); + cpump->page = NULL; + goto unlock; + } + rc = 0; + refcount_set(&cpump->refcnt, 1); + +unlock: + /* If rc is non-zero, do not set mode and reference count */ + if (!rc) { + cpump->mode = a->sample_period ? PAI_MODE_SAMPLING + : PAI_MODE_COUNTING; + } + debug_sprintf_event(cfm_dbg, 5, "%s sample_period %#llx users %d" + " mode %d refcnt %u page %#lx save %p rc %d\n", + __func__, a->sample_period, cpump->active_events, + cpump->mode, refcount_read(&cpump->refcnt), + (unsigned long)cpump->page, cpump->save, rc); + mutex_unlock(&pai_reserve_mutex); + return rc; +} + +/* Might be called on different CPU than the one the event is intended for. */ +static int paicrypt_event_init(struct perf_event *event) +{ + struct perf_event_attr *a = &event->attr; + struct paicrypt_map *cpump; + int rc; + + /* PAI crypto PMU registered as PERF_TYPE_RAW, check event type */ + if (a->type != PERF_TYPE_RAW && event->pmu->type != a->type) + return -ENOENT; + /* PAI crypto event must be in valid range */ + if (a->config < PAI_CRYPTO_BASE || + a->config > PAI_CRYPTO_BASE + paicrypt_cnt) + return -EINVAL; + /* Allow only CPU wide operation, no process context for now. */ + if (event->hw.target || event->cpu == -1) + return -ENOENT; + /* Allow only CRYPTO_ALL for sampling. */ + if (a->sample_period && a->config != PAI_CRYPTO_BASE) + return -EINVAL; + + cpump = per_cpu_ptr(&paicrypt_map, event->cpu); + rc = paicrypt_busy(a, cpump); + if (rc) + return rc; + + /* Event initialization sets last_tag to 0. When later on the events + * are deleted and re-added, do not reset the event count value to zero. + * Events are added, deleted and re-added when 2 or more events + * are active at the same time. + */ + event->hw.last_tag = 0; + cpump->event = event; + event->destroy = paicrypt_event_destroy; + + if (a->sample_period) { + a->sample_period = 1; + a->freq = 0; + /* Register for paicrypt_sched_task() to be called */ + event->attach_state |= PERF_ATTACH_SCHED_CB; + /* Add raw data which contain the memory mapped counters */ + a->sample_type |= PERF_SAMPLE_RAW; + /* Turn off inheritance */ + a->inherit = 0; + } + + static_branch_inc(&pai_key); + return 0; +} + +static void paicrypt_read(struct perf_event *event) +{ + u64 prev, new, delta; + + prev = local64_read(&event->hw.prev_count); + new = paicrypt_getall(event); + local64_set(&event->hw.prev_count, new); + delta = (prev <= new) ? new - prev + : (-1ULL - prev) + new + 1; /* overflow */ + local64_add(delta, &event->count); +} + +static void paicrypt_start(struct perf_event *event, int flags) +{ + u64 sum; + + if (!event->hw.last_tag) { + event->hw.last_tag = 1; + sum = paicrypt_getall(event); /* Get current value */ + local64_set(&event->count, 0); + local64_set(&event->hw.prev_count, sum); + } +} + +static int paicrypt_add(struct perf_event *event, int flags) +{ + struct paicrypt_map *cpump = this_cpu_ptr(&paicrypt_map); + unsigned long ccd; + + if (++cpump->active_events == 1) { + ccd = virt_to_phys(cpump->page) | PAI_CRYPTO_KERNEL_OFFSET; + WRITE_ONCE(S390_lowcore.ccd, ccd); + __ctl_set_bit(0, 50); + } + cpump->event = event; + if (flags & PERF_EF_START && !event->attr.sample_period) { + /* Only counting needs initial counter value */ + paicrypt_start(event, PERF_EF_RELOAD); + } + event->hw.state = 0; + if (event->attr.sample_period) + perf_sched_cb_inc(event->pmu); + return 0; +} + +static void paicrypt_stop(struct perf_event *event, int flags) +{ + paicrypt_read(event); + event->hw.state = PERF_HES_STOPPED; +} + +static void paicrypt_del(struct perf_event *event, int flags) +{ + struct paicrypt_map *cpump = this_cpu_ptr(&paicrypt_map); + + if (event->attr.sample_period) + perf_sched_cb_dec(event->pmu); + if (!event->attr.sample_period) + /* Only counting needs to read counter */ + paicrypt_stop(event, PERF_EF_UPDATE); + if (--cpump->active_events == 0) { + __ctl_clear_bit(0, 50); + WRITE_ONCE(S390_lowcore.ccd, 0); + } +} + +/* Create raw data and save it in buffer. Returns number of bytes copied. + * Saves only positive counter entries of the form + * 2 bytes: Number of counter + * 8 bytes: Value of counter + */ +static size_t paicrypt_copy(struct pai_userdata *userdata, + struct paicrypt_map *cpump, + bool exclude_user, bool exclude_kernel) +{ + int i, outidx = 0; + + for (i = 1; i <= paicrypt_cnt; i++) { + u64 val = 0; + + if (!exclude_kernel) + val += paicrypt_getctr(cpump, i, true); + if (!exclude_user) + val += paicrypt_getctr(cpump, i, false); + if (val) { + userdata[outidx].num = i; + userdata[outidx].value = val; + outidx++; + } + } + return outidx * sizeof(struct pai_userdata); +} + +static int paicrypt_push_sample(void) +{ + struct paicrypt_map *cpump = this_cpu_ptr(&paicrypt_map); + struct perf_event *event = cpump->event; + struct perf_sample_data data; + struct perf_raw_record raw; + struct pt_regs regs; + size_t rawsize; + int overflow; + + if (!cpump->event) /* No event active */ + return 0; + rawsize = paicrypt_copy(cpump->save, cpump, + cpump->event->attr.exclude_user, + cpump->event->attr.exclude_kernel); + if (!rawsize) /* No incremented counters */ + return 0; + + /* Setup perf sample */ + memset(®s, 0, sizeof(regs)); + memset(&raw, 0, sizeof(raw)); + memset(&data, 0, sizeof(data)); + perf_sample_data_init(&data, 0, event->hw.last_period); + if (event->attr.sample_type & PERF_SAMPLE_TID) { + data.tid_entry.pid = task_tgid_nr(current); + data.tid_entry.tid = task_pid_nr(current); + } + if (event->attr.sample_type & PERF_SAMPLE_TIME) + data.time = event->clock(); + if (event->attr.sample_type & (PERF_SAMPLE_ID | PERF_SAMPLE_IDENTIFIER)) + data.id = event->id; + if (event->attr.sample_type & PERF_SAMPLE_CPU) { + data.cpu_entry.cpu = smp_processor_id(); + data.cpu_entry.reserved = 0; + } + if (event->attr.sample_type & PERF_SAMPLE_RAW) { + raw.frag.size = rawsize; + raw.frag.data = cpump->save; + perf_sample_save_raw_data(&data, &raw); + } + + overflow = perf_event_overflow(event, &data, ®s); + perf_event_update_userpage(event); + /* Clear lowcore page after read */ + memset(cpump->page, 0, PAGE_SIZE); + return overflow; +} + +/* Called on schedule-in and schedule-out. No access to event structure, + * but for sampling only event CRYPTO_ALL is allowed. + */ +static void paicrypt_sched_task(struct perf_event_pmu_context *pmu_ctx, bool sched_in) +{ + /* We started with a clean page on event installation. So read out + * results on schedule_out and if page was dirty, clear values. + */ + if (!sched_in) + paicrypt_push_sample(); +} + +/* Attribute definitions for paicrypt interface. As with other CPU + * Measurement Facilities, there is one attribute per mapped counter. + * The number of mapped counters may vary per machine generation. Use + * the QUERY PROCESSOR ACTIVITY COUNTER INFORMATION (QPACI) instruction + * to determine the number of mapped counters. The instructions returns + * a positive number, which is the highest number of supported counters. + * All counters less than this number are also supported, there are no + * holes. A returned number of zero means no support for mapped counters. + * + * The identification of the counter is a unique number. The chosen range + * is 0x1000 + offset in mapped kernel page. + * All CPU Measurement Facility counters identifiers must be unique and + * the numbers from 0 to 496 are already used for the CPU Measurement + * Counter facility. Numbers 0xb0000, 0xbc000 and 0xbd000 are already + * used for the CPU Measurement Sampling facility. + */ +PMU_FORMAT_ATTR(event, "config:0-63"); + +static struct attribute *paicrypt_format_attr[] = { + &format_attr_event.attr, + NULL, +}; + +static struct attribute_group paicrypt_events_group = { + .name = "events", + .attrs = NULL /* Filled in attr_event_init() */ +}; + +static struct attribute_group paicrypt_format_group = { + .name = "format", + .attrs = paicrypt_format_attr, +}; + +static const struct attribute_group *paicrypt_attr_groups[] = { + &paicrypt_events_group, + &paicrypt_format_group, + NULL, +}; + +/* Performance monitoring unit for mapped counters */ +static struct pmu paicrypt = { + .task_ctx_nr = perf_invalid_context, + .event_init = paicrypt_event_init, + .add = paicrypt_add, + .del = paicrypt_del, + .start = paicrypt_start, + .stop = paicrypt_stop, + .read = paicrypt_read, + .sched_task = paicrypt_sched_task, + .attr_groups = paicrypt_attr_groups +}; + +/* List of symbolic PAI counter names. */ +static const char * const paicrypt_ctrnames[] = { + [0] = "CRYPTO_ALL", + [1] = "KM_DEA", + [2] = "KM_TDEA_128", + [3] = "KM_TDEA_192", + [4] = "KM_ENCRYPTED_DEA", + [5] = "KM_ENCRYPTED_TDEA_128", + [6] = "KM_ENCRYPTED_TDEA_192", + [7] = "KM_AES_128", + [8] = "KM_AES_192", + [9] = "KM_AES_256", + [10] = "KM_ENCRYPTED_AES_128", + [11] = "KM_ENCRYPTED_AES_192", + [12] = "KM_ENCRYPTED_AES_256", + [13] = "KM_XTS_AES_128", + [14] = "KM_XTS_AES_256", + [15] = "KM_XTS_ENCRYPTED_AES_128", + [16] = "KM_XTS_ENCRYPTED_AES_256", + [17] = "KMC_DEA", + [18] = "KMC_TDEA_128", + [19] = "KMC_TDEA_192", + [20] = "KMC_ENCRYPTED_DEA", + [21] = "KMC_ENCRYPTED_TDEA_128", + [22] = "KMC_ENCRYPTED_TDEA_192", + [23] = "KMC_AES_128", + [24] = "KMC_AES_192", + [25] = "KMC_AES_256", + [26] = "KMC_ENCRYPTED_AES_128", + [27] = "KMC_ENCRYPTED_AES_192", + [28] = "KMC_ENCRYPTED_AES_256", + [29] = "KMC_PRNG", + [30] = "KMA_GCM_AES_128", + [31] = "KMA_GCM_AES_192", + [32] = "KMA_GCM_AES_256", + [33] = "KMA_GCM_ENCRYPTED_AES_128", + [34] = "KMA_GCM_ENCRYPTED_AES_192", + [35] = "KMA_GCM_ENCRYPTED_AES_256", + [36] = "KMF_DEA", + [37] = "KMF_TDEA_128", + [38] = "KMF_TDEA_192", + [39] = "KMF_ENCRYPTED_DEA", + [40] = "KMF_ENCRYPTED_TDEA_128", + [41] = "KMF_ENCRYPTED_TDEA_192", + [42] = "KMF_AES_128", + [43] = "KMF_AES_192", + [44] = "KMF_AES_256", + [45] = "KMF_ENCRYPTED_AES_128", + [46] = "KMF_ENCRYPTED_AES_192", + [47] = "KMF_ENCRYPTED_AES_256", + [48] = "KMCTR_DEA", + [49] = "KMCTR_TDEA_128", + [50] = "KMCTR_TDEA_192", + [51] = "KMCTR_ENCRYPTED_DEA", + [52] = "KMCTR_ENCRYPTED_TDEA_128", + [53] = "KMCTR_ENCRYPTED_TDEA_192", + [54] = "KMCTR_AES_128", + [55] = "KMCTR_AES_192", + [56] = "KMCTR_AES_256", + [57] = "KMCTR_ENCRYPTED_AES_128", + [58] = "KMCTR_ENCRYPTED_AES_192", + [59] = "KMCTR_ENCRYPTED_AES_256", + [60] = "KMO_DEA", + [61] = "KMO_TDEA_128", + [62] = "KMO_TDEA_192", + [63] = "KMO_ENCRYPTED_DEA", + [64] = "KMO_ENCRYPTED_TDEA_128", + [65] = "KMO_ENCRYPTED_TDEA_192", + [66] = "KMO_AES_128", + [67] = "KMO_AES_192", + [68] = "KMO_AES_256", + [69] = "KMO_ENCRYPTED_AES_128", + [70] = "KMO_ENCRYPTED_AES_192", + [71] = "KMO_ENCRYPTED_AES_256", + [72] = "KIMD_SHA_1", + [73] = "KIMD_SHA_256", + [74] = "KIMD_SHA_512", + [75] = "KIMD_SHA3_224", + [76] = "KIMD_SHA3_256", + [77] = "KIMD_SHA3_384", + [78] = "KIMD_SHA3_512", + [79] = "KIMD_SHAKE_128", + [80] = "KIMD_SHAKE_256", + [81] = "KIMD_GHASH", + [82] = "KLMD_SHA_1", + [83] = "KLMD_SHA_256", + [84] = "KLMD_SHA_512", + [85] = "KLMD_SHA3_224", + [86] = "KLMD_SHA3_256", + [87] = "KLMD_SHA3_384", + [88] = "KLMD_SHA3_512", + [89] = "KLMD_SHAKE_128", + [90] = "KLMD_SHAKE_256", + [91] = "KMAC_DEA", + [92] = "KMAC_TDEA_128", + [93] = "KMAC_TDEA_192", + [94] = "KMAC_ENCRYPTED_DEA", + [95] = "KMAC_ENCRYPTED_TDEA_128", + [96] = "KMAC_ENCRYPTED_TDEA_192", + [97] = "KMAC_AES_128", + [98] = "KMAC_AES_192", + [99] = "KMAC_AES_256", + [100] = "KMAC_ENCRYPTED_AES_128", + [101] = "KMAC_ENCRYPTED_AES_192", + [102] = "KMAC_ENCRYPTED_AES_256", + [103] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_DEA", + [104] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_TDEA_128", + [105] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_TDEA_192", + [106] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_ENCRYPTED_DEA", + [107] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_ENCRYPTED_TDEA_128", + [108] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_ENCRYPTED_TDEA_192", + [109] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_AES_128", + [110] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_AES_192", + [111] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_AES_256", + [112] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_ENCRYPTED_AES_128", + [113] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_ENCRYPTED_AES_192", + [114] = "PCC_COMPUTE_LAST_BLOCK_CMAC_USING_ENCRYPTED_AES_256A", + [115] = "PCC_COMPUTE_XTS_PARAMETER_USING_AES_128", + [116] = "PCC_COMPUTE_XTS_PARAMETER_USING_AES_256", + [117] = "PCC_COMPUTE_XTS_PARAMETER_USING_ENCRYPTED_AES_128", + [118] = "PCC_COMPUTE_XTS_PARAMETER_USING_ENCRYPTED_AES_256", + [119] = "PCC_SCALAR_MULTIPLY_P256", + [120] = "PCC_SCALAR_MULTIPLY_P384", + [121] = "PCC_SCALAR_MULTIPLY_P521", + [122] = "PCC_SCALAR_MULTIPLY_ED25519", + [123] = "PCC_SCALAR_MULTIPLY_ED448", + [124] = "PCC_SCALAR_MULTIPLY_X25519", + [125] = "PCC_SCALAR_MULTIPLY_X448", + [126] = "PRNO_SHA_512_DRNG", + [127] = "PRNO_TRNG_QUERY_RAW_TO_CONDITIONED_RATIO", + [128] = "PRNO_TRNG", + [129] = "KDSA_ECDSA_VERIFY_P256", + [130] = "KDSA_ECDSA_VERIFY_P384", + [131] = "KDSA_ECDSA_VERIFY_P521", + [132] = "KDSA_ECDSA_SIGN_P256", + [133] = "KDSA_ECDSA_SIGN_P384", + [134] = "KDSA_ECDSA_SIGN_P521", + [135] = "KDSA_ENCRYPTED_ECDSA_SIGN_P256", + [136] = "KDSA_ENCRYPTED_ECDSA_SIGN_P384", + [137] = "KDSA_ENCRYPTED_ECDSA_SIGN_P521", + [138] = "KDSA_EDDSA_VERIFY_ED25519", + [139] = "KDSA_EDDSA_VERIFY_ED448", + [140] = "KDSA_EDDSA_SIGN_ED25519", + [141] = "KDSA_EDDSA_SIGN_ED448", + [142] = "KDSA_ENCRYPTED_EDDSA_SIGN_ED25519", + [143] = "KDSA_ENCRYPTED_EDDSA_SIGN_ED448", + [144] = "PCKMO_ENCRYPT_DEA_KEY", + [145] = "PCKMO_ENCRYPT_TDEA_128_KEY", + [146] = "PCKMO_ENCRYPT_TDEA_192_KEY", + [147] = "PCKMO_ENCRYPT_AES_128_KEY", + [148] = "PCKMO_ENCRYPT_AES_192_KEY", + [149] = "PCKMO_ENCRYPT_AES_256_KEY", + [150] = "PCKMO_ENCRYPT_ECC_P256_KEY", + [151] = "PCKMO_ENCRYPT_ECC_P384_KEY", + [152] = "PCKMO_ENCRYPT_ECC_P521_KEY", + [153] = "PCKMO_ENCRYPT_ECC_ED25519_KEY", + [154] = "PCKMO_ENCRYPT_ECC_ED448_KEY", + [155] = "IBM_RESERVED_155", + [156] = "IBM_RESERVED_156", +}; + +static void __init attr_event_free(struct attribute **attrs, int num) +{ + struct perf_pmu_events_attr *pa; + int i; + + for (i = 0; i < num; i++) { + struct device_attribute *dap; + + dap = container_of(attrs[i], struct device_attribute, attr); + pa = container_of(dap, struct perf_pmu_events_attr, attr); + kfree(pa); + } + kfree(attrs); +} + +static int __init attr_event_init_one(struct attribute **attrs, int num) +{ + struct perf_pmu_events_attr *pa; + + pa = kzalloc(sizeof(*pa), GFP_KERNEL); + if (!pa) + return -ENOMEM; + + sysfs_attr_init(&pa->attr.attr); + pa->id = PAI_CRYPTO_BASE + num; + pa->attr.attr.name = paicrypt_ctrnames[num]; + pa->attr.attr.mode = 0444; + pa->attr.show = cpumf_events_sysfs_show; + pa->attr.store = NULL; + attrs[num] = &pa->attr.attr; + return 0; +} + +/* Create PMU sysfs event attributes on the fly. */ +static int __init attr_event_init(void) +{ + struct attribute **attrs; + int ret, i; + + attrs = kmalloc_array(ARRAY_SIZE(paicrypt_ctrnames) + 1, sizeof(*attrs), + GFP_KERNEL); + if (!attrs) + return -ENOMEM; + for (i = 0; i < ARRAY_SIZE(paicrypt_ctrnames); i++) { + ret = attr_event_init_one(attrs, i); + if (ret) { + attr_event_free(attrs, i - 1); + return ret; + } + } + attrs[i] = NULL; + paicrypt_events_group.attrs = attrs; + return 0; +} + +static int __init paicrypt_init(void) +{ + struct qpaci_info_block ib; + int rc; + + if (!test_facility(196)) + return 0; + + qpaci(&ib); + paicrypt_cnt = ib.num_cc; + if (paicrypt_cnt == 0) + return 0; + if (paicrypt_cnt >= PAI_CRYPTO_MAXCTR) + paicrypt_cnt = PAI_CRYPTO_MAXCTR - 1; + + rc = attr_event_init(); /* Export known PAI crypto events */ + if (rc) { + pr_err("Creation of PMU pai_crypto /sysfs failed\n"); + return rc; + } + + /* Setup s390dbf facility */ + cfm_dbg = debug_register(KMSG_COMPONENT, 2, 256, 128); + if (!cfm_dbg) { + pr_err("Registration of s390dbf pai_crypto failed\n"); + return -ENOMEM; + } + debug_register_view(cfm_dbg, &debug_sprintf_view); + + rc = perf_pmu_register(&paicrypt, "pai_crypto", -1); + if (rc) { + pr_err("Registering the pai_crypto PMU failed with rc=%i\n", + rc); + debug_unregister_view(cfm_dbg, &debug_sprintf_view); + debug_unregister(cfm_dbg); + return rc; + } + return 0; +} + +device_initcall(paicrypt_init); diff --git a/arch/s390/kernel/perf_pai_ext.c b/arch/s390/kernel/perf_pai_ext.c new file mode 100644 index 0000000000..c57c1a2032 --- /dev/null +++ b/arch/s390/kernel/perf_pai_ext.c @@ -0,0 +1,667 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Performance event support - Processor Activity Instrumentation Extension + * Facility + * + * Copyright IBM Corp. 2022 + * Author(s): Thomas Richter <tmricht@linux.ibm.com> + */ +#define KMSG_COMPONENT "pai_ext" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/kernel.h> +#include <linux/kernel_stat.h> +#include <linux/percpu.h> +#include <linux/notifier.h> +#include <linux/init.h> +#include <linux/export.h> +#include <linux/io.h> +#include <linux/perf_event.h> + +#include <asm/ctl_reg.h> +#include <asm/pai.h> +#include <asm/debug.h> + +#define PAIE1_CB_SZ 0x200 /* Size of PAIE1 control block */ +#define PAIE1_CTRBLOCK_SZ 0x400 /* Size of PAIE1 counter blocks */ + +static debug_info_t *paiext_dbg; +static unsigned int paiext_cnt; /* Extracted with QPACI instruction */ + +struct pai_userdata { + u16 num; + u64 value; +} __packed; + +/* Create the PAI extension 1 control block area. + * The PAI extension control block 1 is pointed to by lowcore + * address 0x1508 for each CPU. This control block is 512 bytes in size + * and requires a 512 byte boundary alignment. + */ +struct paiext_cb { /* PAI extension 1 control block */ + u64 header; /* Not used */ + u64 reserved1; + u64 acc; /* Addr to analytics counter control block */ + u8 reserved2[488]; +} __packed; + +struct paiext_map { + unsigned long *area; /* Area for CPU to store counters */ + struct pai_userdata *save; /* Area to store non-zero counters */ + enum paievt_mode mode; /* Type of event */ + unsigned int active_events; /* # of PAI Extension users */ + refcount_t refcnt; + struct perf_event *event; /* Perf event for sampling */ + struct paiext_cb *paiext_cb; /* PAI extension control block area */ +}; + +struct paiext_mapptr { + struct paiext_map *mapptr; +}; + +static struct paiext_root { /* Anchor to per CPU data */ + refcount_t refcnt; /* Overall active events */ + struct paiext_mapptr __percpu *mapptr; +} paiext_root; + +/* Free per CPU data when the last event is removed. */ +static void paiext_root_free(void) +{ + if (refcount_dec_and_test(&paiext_root.refcnt)) { + free_percpu(paiext_root.mapptr); + paiext_root.mapptr = NULL; + } +} + +/* On initialization of first event also allocate per CPU data dynamically. + * Start with an array of pointers, the array size is the maximum number of + * CPUs possible, which might be larger than the number of CPUs currently + * online. + */ +static int paiext_root_alloc(void) +{ + if (!refcount_inc_not_zero(&paiext_root.refcnt)) { + /* The memory is already zeroed. */ + paiext_root.mapptr = alloc_percpu(struct paiext_mapptr); + if (!paiext_root.mapptr) { + /* Returning without refcnt adjustment is ok. The + * error code is handled by paiext_alloc() which + * decrements refcnt when an event can not be + * created. + */ + return -ENOMEM; + } + refcount_set(&paiext_root.refcnt, 1); + } + return 0; +} + +/* Protects against concurrent increment of sampler and counter member + * increments at the same time and prohibits concurrent execution of + * counting and sampling events. + * Ensures that analytics counter block is deallocated only when the + * sampling and counting on that cpu is zero. + * For details see paiext_alloc(). + */ +static DEFINE_MUTEX(paiext_reserve_mutex); + +/* Free all memory allocated for event counting/sampling setup */ +static void paiext_free(struct paiext_mapptr *mp) +{ + kfree(mp->mapptr->area); + kfree(mp->mapptr->paiext_cb); + kvfree(mp->mapptr->save); + kfree(mp->mapptr); + mp->mapptr = NULL; +} + +/* Release the PMU if event is the last perf event */ +static void paiext_event_destroy(struct perf_event *event) +{ + struct paiext_mapptr *mp = per_cpu_ptr(paiext_root.mapptr, event->cpu); + struct paiext_map *cpump = mp->mapptr; + + mutex_lock(&paiext_reserve_mutex); + cpump->event = NULL; + if (refcount_dec_and_test(&cpump->refcnt)) /* Last reference gone */ + paiext_free(mp); + paiext_root_free(); + mutex_unlock(&paiext_reserve_mutex); + debug_sprintf_event(paiext_dbg, 4, "%s cpu %d mapptr %p\n", __func__, + event->cpu, mp->mapptr); + +} + +/* Used to avoid races in checking concurrent access of counting and + * sampling for pai_extension events. + * + * Only one instance of event pai_ext/NNPA_ALL/ for sampling is + * allowed and when this event is running, no counting event is allowed. + * Several counting events are allowed in parallel, but no sampling event + * is allowed while one (or more) counting events are running. + * + * This function is called in process context and it is safe to block. + * When the event initialization functions fails, no other call back will + * be invoked. + * + * Allocate the memory for the event. + */ +static int paiext_alloc(struct perf_event_attr *a, struct perf_event *event) +{ + struct paiext_mapptr *mp; + struct paiext_map *cpump; + int rc; + + mutex_lock(&paiext_reserve_mutex); + + rc = paiext_root_alloc(); + if (rc) + goto unlock; + + mp = per_cpu_ptr(paiext_root.mapptr, event->cpu); + cpump = mp->mapptr; + if (!cpump) { /* Paiext_map allocated? */ + rc = -ENOMEM; + cpump = kzalloc(sizeof(*cpump), GFP_KERNEL); + if (!cpump) + goto undo; + + /* Allocate memory for counter area and counter extraction. + * These are + * - a 512 byte block and requires 512 byte boundary alignment. + * - a 1KB byte block and requires 1KB boundary alignment. + * Only the first counting event has to allocate the area. + * + * Note: This works with commit 59bb47985c1d by default. + * Backporting this to kernels without this commit might + * need adjustment. + */ + mp->mapptr = cpump; + cpump->area = kzalloc(PAIE1_CTRBLOCK_SZ, GFP_KERNEL); + cpump->paiext_cb = kzalloc(PAIE1_CB_SZ, GFP_KERNEL); + cpump->save = kvmalloc_array(paiext_cnt + 1, + sizeof(struct pai_userdata), + GFP_KERNEL); + if (!cpump->save || !cpump->area || !cpump->paiext_cb) { + paiext_free(mp); + goto undo; + } + refcount_set(&cpump->refcnt, 1); + cpump->mode = a->sample_period ? PAI_MODE_SAMPLING + : PAI_MODE_COUNTING; + } else { + /* Multiple invocation, check what is active. + * Supported are multiple counter events or only one sampling + * event concurrently at any one time. + */ + if (cpump->mode == PAI_MODE_SAMPLING || + (cpump->mode == PAI_MODE_COUNTING && a->sample_period)) { + rc = -EBUSY; + goto undo; + } + refcount_inc(&cpump->refcnt); + } + + rc = 0; + cpump->event = event; + +undo: + if (rc) { + /* Error in allocation of event, decrement anchor. Since + * the event in not created, its destroy() function is never + * invoked. Adjust the reference counter for the anchor. + */ + paiext_root_free(); + } +unlock: + mutex_unlock(&paiext_reserve_mutex); + /* If rc is non-zero, no increment of counter/sampler was done. */ + return rc; +} + +/* The PAI extension 1 control block supports up to 128 entries. Return + * the index within PAIE1_CB given the event number. Also validate event + * number. + */ +static int paiext_event_valid(struct perf_event *event) +{ + u64 cfg = event->attr.config; + + if (cfg >= PAI_NNPA_BASE && cfg <= PAI_NNPA_BASE + paiext_cnt) { + /* Offset NNPA in paiext_cb */ + event->hw.config_base = offsetof(struct paiext_cb, acc); + return 0; + } + return -EINVAL; +} + +/* Might be called on different CPU than the one the event is intended for. */ +static int paiext_event_init(struct perf_event *event) +{ + struct perf_event_attr *a = &event->attr; + int rc; + + /* PMU pai_ext registered as PERF_TYPE_RAW, check event type */ + if (a->type != PERF_TYPE_RAW && event->pmu->type != a->type) + return -ENOENT; + /* PAI extension event must be valid and in supported range */ + rc = paiext_event_valid(event); + if (rc) + return rc; + /* Allow only CPU wide operation, no process context for now. */ + if (event->hw.target || event->cpu == -1) + return -ENOENT; + /* Allow only event NNPA_ALL for sampling. */ + if (a->sample_period && a->config != PAI_NNPA_BASE) + return -EINVAL; + /* Prohibit exclude_user event selection */ + if (a->exclude_user) + return -EINVAL; + + rc = paiext_alloc(a, event); + if (rc) + return rc; + event->hw.last_tag = 0; + event->destroy = paiext_event_destroy; + + if (a->sample_period) { + a->sample_period = 1; + a->freq = 0; + /* Register for paicrypt_sched_task() to be called */ + event->attach_state |= PERF_ATTACH_SCHED_CB; + /* Add raw data which are the memory mapped counters */ + a->sample_type |= PERF_SAMPLE_RAW; + /* Turn off inheritance */ + a->inherit = 0; + } + + return 0; +} + +static u64 paiext_getctr(struct paiext_map *cpump, int nr) +{ + return cpump->area[nr]; +} + +/* Read the counter values. Return value from location in buffer. For event + * NNPA_ALL sum up all events. + */ +static u64 paiext_getdata(struct perf_event *event) +{ + struct paiext_mapptr *mp = this_cpu_ptr(paiext_root.mapptr); + struct paiext_map *cpump = mp->mapptr; + u64 sum = 0; + int i; + + if (event->attr.config != PAI_NNPA_BASE) + return paiext_getctr(cpump, event->attr.config - PAI_NNPA_BASE); + + for (i = 1; i <= paiext_cnt; i++) + sum += paiext_getctr(cpump, i); + + return sum; +} + +static u64 paiext_getall(struct perf_event *event) +{ + return paiext_getdata(event); +} + +static void paiext_read(struct perf_event *event) +{ + u64 prev, new, delta; + + prev = local64_read(&event->hw.prev_count); + new = paiext_getall(event); + local64_set(&event->hw.prev_count, new); + delta = new - prev; + local64_add(delta, &event->count); +} + +static void paiext_start(struct perf_event *event, int flags) +{ + u64 sum; + + if (event->hw.last_tag) + return; + event->hw.last_tag = 1; + sum = paiext_getall(event); /* Get current value */ + local64_set(&event->hw.prev_count, sum); + local64_set(&event->count, 0); +} + +static int paiext_add(struct perf_event *event, int flags) +{ + struct paiext_mapptr *mp = this_cpu_ptr(paiext_root.mapptr); + struct paiext_map *cpump = mp->mapptr; + struct paiext_cb *pcb = cpump->paiext_cb; + + if (++cpump->active_events == 1) { + S390_lowcore.aicd = virt_to_phys(cpump->paiext_cb); + pcb->acc = virt_to_phys(cpump->area) | 0x1; + /* Enable CPU instruction lookup for PAIE1 control block */ + __ctl_set_bit(0, 49); + debug_sprintf_event(paiext_dbg, 4, "%s 1508 %llx acc %llx\n", + __func__, S390_lowcore.aicd, pcb->acc); + } + if (flags & PERF_EF_START && !event->attr.sample_period) { + /* Only counting needs initial counter value */ + paiext_start(event, PERF_EF_RELOAD); + } + event->hw.state = 0; + if (event->attr.sample_period) { + cpump->event = event; + perf_sched_cb_inc(event->pmu); + } + return 0; +} + +static void paiext_stop(struct perf_event *event, int flags) +{ + paiext_read(event); + event->hw.state = PERF_HES_STOPPED; +} + +static void paiext_del(struct perf_event *event, int flags) +{ + struct paiext_mapptr *mp = this_cpu_ptr(paiext_root.mapptr); + struct paiext_map *cpump = mp->mapptr; + struct paiext_cb *pcb = cpump->paiext_cb; + + if (event->attr.sample_period) + perf_sched_cb_dec(event->pmu); + if (!event->attr.sample_period) { + /* Only counting needs to read counter */ + paiext_stop(event, PERF_EF_UPDATE); + } + if (--cpump->active_events == 0) { + /* Disable CPU instruction lookup for PAIE1 control block */ + __ctl_clear_bit(0, 49); + pcb->acc = 0; + S390_lowcore.aicd = 0; + debug_sprintf_event(paiext_dbg, 4, "%s 1508 %llx acc %llx\n", + __func__, S390_lowcore.aicd, pcb->acc); + } +} + +/* Create raw data and save it in buffer. Returns number of bytes copied. + * Saves only positive counter entries of the form + * 2 bytes: Number of counter + * 8 bytes: Value of counter + */ +static size_t paiext_copy(struct paiext_map *cpump) +{ + struct pai_userdata *userdata = cpump->save; + int i, outidx = 0; + + for (i = 1; i <= paiext_cnt; i++) { + u64 val = paiext_getctr(cpump, i); + + if (val) { + userdata[outidx].num = i; + userdata[outidx].value = val; + outidx++; + } + } + return outidx * sizeof(*userdata); +} + +/* Write sample when one or more counters values are nonzero. + * + * Note: The function paiext_sched_task() and paiext_push_sample() are not + * invoked after function paiext_del() has been called because of function + * perf_sched_cb_dec(). + * The function paiext_sched_task() and paiext_push_sample() are only + * called when sampling is active. Function perf_sched_cb_inc() + * has been invoked to install function paiext_sched_task() as call back + * to run at context switch time (see paiext_add()). + * + * This causes function perf_event_context_sched_out() and + * perf_event_context_sched_in() to check whether the PMU has installed an + * sched_task() callback. That callback is not active after paiext_del() + * returns and has deleted the event on that CPU. + */ +static int paiext_push_sample(void) +{ + struct paiext_mapptr *mp = this_cpu_ptr(paiext_root.mapptr); + struct paiext_map *cpump = mp->mapptr; + struct perf_event *event = cpump->event; + struct perf_sample_data data; + struct perf_raw_record raw; + struct pt_regs regs; + size_t rawsize; + int overflow; + + rawsize = paiext_copy(cpump); + if (!rawsize) /* No incremented counters */ + return 0; + + /* Setup perf sample */ + memset(®s, 0, sizeof(regs)); + memset(&raw, 0, sizeof(raw)); + memset(&data, 0, sizeof(data)); + perf_sample_data_init(&data, 0, event->hw.last_period); + if (event->attr.sample_type & PERF_SAMPLE_TID) { + data.tid_entry.pid = task_tgid_nr(current); + data.tid_entry.tid = task_pid_nr(current); + } + if (event->attr.sample_type & PERF_SAMPLE_TIME) + data.time = event->clock(); + if (event->attr.sample_type & (PERF_SAMPLE_ID | PERF_SAMPLE_IDENTIFIER)) + data.id = event->id; + if (event->attr.sample_type & PERF_SAMPLE_CPU) + data.cpu_entry.cpu = smp_processor_id(); + if (event->attr.sample_type & PERF_SAMPLE_RAW) { + raw.frag.size = rawsize; + raw.frag.data = cpump->save; + perf_sample_save_raw_data(&data, &raw); + } + + overflow = perf_event_overflow(event, &data, ®s); + perf_event_update_userpage(event); + /* Clear lowcore area after read */ + memset(cpump->area, 0, PAIE1_CTRBLOCK_SZ); + return overflow; +} + +/* Called on schedule-in and schedule-out. No access to event structure, + * but for sampling only event NNPA_ALL is allowed. + */ +static void paiext_sched_task(struct perf_event_pmu_context *pmu_ctx, bool sched_in) +{ + /* We started with a clean page on event installation. So read out + * results on schedule_out and if page was dirty, clear values. + */ + if (!sched_in) + paiext_push_sample(); +} + +/* Attribute definitions for pai extension1 interface. As with other CPU + * Measurement Facilities, there is one attribute per mapped counter. + * The number of mapped counters may vary per machine generation. Use + * the QUERY PROCESSOR ACTIVITY COUNTER INFORMATION (QPACI) instruction + * to determine the number of mapped counters. The instructions returns + * a positive number, which is the highest number of supported counters. + * All counters less than this number are also supported, there are no + * holes. A returned number of zero means no support for mapped counters. + * + * The identification of the counter is a unique number. The chosen range + * is 0x1800 + offset in mapped kernel page. + * All CPU Measurement Facility counters identifiers must be unique and + * the numbers from 0 to 496 are already used for the CPU Measurement + * Counter facility. Number 0x1000 to 0x103e are used for PAI cryptography + * counters. + * Numbers 0xb0000, 0xbc000 and 0xbd000 are already + * used for the CPU Measurement Sampling facility. + */ +PMU_FORMAT_ATTR(event, "config:0-63"); + +static struct attribute *paiext_format_attr[] = { + &format_attr_event.attr, + NULL, +}; + +static struct attribute_group paiext_events_group = { + .name = "events", + .attrs = NULL, /* Filled in attr_event_init() */ +}; + +static struct attribute_group paiext_format_group = { + .name = "format", + .attrs = paiext_format_attr, +}; + +static const struct attribute_group *paiext_attr_groups[] = { + &paiext_events_group, + &paiext_format_group, + NULL, +}; + +/* Performance monitoring unit for mapped counters */ +static struct pmu paiext = { + .task_ctx_nr = perf_invalid_context, + .event_init = paiext_event_init, + .add = paiext_add, + .del = paiext_del, + .start = paiext_start, + .stop = paiext_stop, + .read = paiext_read, + .sched_task = paiext_sched_task, + .attr_groups = paiext_attr_groups, +}; + +/* List of symbolic PAI extension 1 NNPA counter names. */ +static const char * const paiext_ctrnames[] = { + [0] = "NNPA_ALL", + [1] = "NNPA_ADD", + [2] = "NNPA_SUB", + [3] = "NNPA_MUL", + [4] = "NNPA_DIV", + [5] = "NNPA_MIN", + [6] = "NNPA_MAX", + [7] = "NNPA_LOG", + [8] = "NNPA_EXP", + [9] = "NNPA_IBM_RESERVED_9", + [10] = "NNPA_RELU", + [11] = "NNPA_TANH", + [12] = "NNPA_SIGMOID", + [13] = "NNPA_SOFTMAX", + [14] = "NNPA_BATCHNORM", + [15] = "NNPA_MAXPOOL2D", + [16] = "NNPA_AVGPOOL2D", + [17] = "NNPA_LSTMACT", + [18] = "NNPA_GRUACT", + [19] = "NNPA_CONVOLUTION", + [20] = "NNPA_MATMUL_OP", + [21] = "NNPA_MATMUL_OP_BCAST23", + [22] = "NNPA_SMALLBATCH", + [23] = "NNPA_LARGEDIM", + [24] = "NNPA_SMALLTENSOR", + [25] = "NNPA_1MFRAME", + [26] = "NNPA_2GFRAME", + [27] = "NNPA_ACCESSEXCEPT", +}; + +static void __init attr_event_free(struct attribute **attrs, int num) +{ + struct perf_pmu_events_attr *pa; + struct device_attribute *dap; + int i; + + for (i = 0; i < num; i++) { + dap = container_of(attrs[i], struct device_attribute, attr); + pa = container_of(dap, struct perf_pmu_events_attr, attr); + kfree(pa); + } + kfree(attrs); +} + +static int __init attr_event_init_one(struct attribute **attrs, int num) +{ + struct perf_pmu_events_attr *pa; + + pa = kzalloc(sizeof(*pa), GFP_KERNEL); + if (!pa) + return -ENOMEM; + + sysfs_attr_init(&pa->attr.attr); + pa->id = PAI_NNPA_BASE + num; + pa->attr.attr.name = paiext_ctrnames[num]; + pa->attr.attr.mode = 0444; + pa->attr.show = cpumf_events_sysfs_show; + pa->attr.store = NULL; + attrs[num] = &pa->attr.attr; + return 0; +} + +/* Create PMU sysfs event attributes on the fly. */ +static int __init attr_event_init(void) +{ + struct attribute **attrs; + int ret, i; + + attrs = kmalloc_array(ARRAY_SIZE(paiext_ctrnames) + 1, sizeof(*attrs), + GFP_KERNEL); + if (!attrs) + return -ENOMEM; + for (i = 0; i < ARRAY_SIZE(paiext_ctrnames); i++) { + ret = attr_event_init_one(attrs, i); + if (ret) { + attr_event_free(attrs, i - 1); + return ret; + } + } + attrs[i] = NULL; + paiext_events_group.attrs = attrs; + return 0; +} + +static int __init paiext_init(void) +{ + struct qpaci_info_block ib; + int rc = -ENOMEM; + + if (!test_facility(197)) + return 0; + + qpaci(&ib); + paiext_cnt = ib.num_nnpa; + if (paiext_cnt >= PAI_NNPA_MAXCTR) + paiext_cnt = PAI_NNPA_MAXCTR; + if (!paiext_cnt) + return 0; + + rc = attr_event_init(); + if (rc) { + pr_err("Creation of PMU " KMSG_COMPONENT " /sysfs failed\n"); + return rc; + } + + /* Setup s390dbf facility */ + paiext_dbg = debug_register(KMSG_COMPONENT, 2, 256, 128); + if (!paiext_dbg) { + pr_err("Registration of s390dbf " KMSG_COMPONENT " failed\n"); + rc = -ENOMEM; + goto out_init; + } + debug_register_view(paiext_dbg, &debug_sprintf_view); + + rc = perf_pmu_register(&paiext, KMSG_COMPONENT, -1); + if (rc) { + pr_err("Registration of " KMSG_COMPONENT " PMU failed with " + "rc=%i\n", rc); + goto out_pmu; + } + + return 0; + +out_pmu: + debug_unregister_view(paiext_dbg, &debug_sprintf_view); + debug_unregister(paiext_dbg); +out_init: + attr_event_free(paiext_events_group.attrs, + ARRAY_SIZE(paiext_ctrnames) + 1); + return rc; +} + +device_initcall(paiext_init); diff --git a/arch/s390/kernel/perf_regs.c b/arch/s390/kernel/perf_regs.c new file mode 100644 index 0000000000..6e9e5d5e92 --- /dev/null +++ b/arch/s390/kernel/perf_regs.c @@ -0,0 +1,68 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/perf_event.h> +#include <linux/perf_regs.h> +#include <linux/kernel.h> +#include <linux/errno.h> +#include <linux/bug.h> +#include <asm/ptrace.h> +#include <asm/fpu/api.h> +#include <asm/fpu/types.h> + +u64 perf_reg_value(struct pt_regs *regs, int idx) +{ + freg_t fp; + + if (idx >= PERF_REG_S390_R0 && idx <= PERF_REG_S390_R15) + return regs->gprs[idx]; + + if (idx >= PERF_REG_S390_FP0 && idx <= PERF_REG_S390_FP15) { + if (!user_mode(regs)) + return 0; + + idx -= PERF_REG_S390_FP0; + fp = MACHINE_HAS_VX ? *(freg_t *)(current->thread.fpu.vxrs + idx) + : current->thread.fpu.fprs[idx]; + return fp.ui; + } + + if (idx == PERF_REG_S390_MASK) + return regs->psw.mask; + if (idx == PERF_REG_S390_PC) + return regs->psw.addr; + + WARN_ON_ONCE((u32)idx >= PERF_REG_S390_MAX); + return 0; +} + +#define REG_RESERVED (~((1UL << PERF_REG_S390_MAX) - 1)) + +int perf_reg_validate(u64 mask) +{ + if (!mask || mask & REG_RESERVED) + return -EINVAL; + + return 0; +} + +u64 perf_reg_abi(struct task_struct *task) +{ + if (test_tsk_thread_flag(task, TIF_31BIT)) + return PERF_SAMPLE_REGS_ABI_32; + + return PERF_SAMPLE_REGS_ABI_64; +} + +void perf_get_regs_user(struct perf_regs *regs_user, + struct pt_regs *regs) +{ + /* + * Use the regs from the first interruption and let + * perf_sample_regs_intr() handle interrupts (regs == get_irq_regs()). + * + * Also save FPU registers for user-space tasks only. + */ + regs_user->regs = task_pt_regs(current); + if (user_mode(regs_user->regs)) + save_fpu_regs(); + regs_user->abi = perf_reg_abi(current); +} diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c new file mode 100644 index 0000000000..2580004177 --- /dev/null +++ b/arch/s390/kernel/process.c @@ -0,0 +1,240 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * This file handles the architecture dependent parts of process handling. + * + * Copyright IBM Corp. 1999, 2009 + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>, + * Hartmut Penner <hp@de.ibm.com>, + * Denis Joseph Barrow, + */ + +#include <linux/elf-randomize.h> +#include <linux/compiler.h> +#include <linux/cpu.h> +#include <linux/sched.h> +#include <linux/sched/debug.h> +#include <linux/sched/task.h> +#include <linux/sched/task_stack.h> +#include <linux/kernel.h> +#include <linux/mm.h> +#include <linux/elfcore.h> +#include <linux/smp.h> +#include <linux/slab.h> +#include <linux/interrupt.h> +#include <linux/tick.h> +#include <linux/personality.h> +#include <linux/syscalls.h> +#include <linux/compat.h> +#include <linux/kprobes.h> +#include <linux/random.h> +#include <linux/export.h> +#include <linux/init_task.h> +#include <linux/entry-common.h> +#include <linux/io.h> +#include <asm/cpu_mf.h> +#include <asm/processor.h> +#include <asm/vtimer.h> +#include <asm/exec.h> +#include <asm/irq.h> +#include <asm/nmi.h> +#include <asm/smp.h> +#include <asm/stacktrace.h> +#include <asm/switch_to.h> +#include <asm/runtime_instr.h> +#include <asm/unwind.h> +#include "entry.h" + +void ret_from_fork(void) asm("ret_from_fork"); + +void __ret_from_fork(struct task_struct *prev, struct pt_regs *regs) +{ + void (*func)(void *arg); + + schedule_tail(prev); + + if (!user_mode(regs)) { + /* Kernel thread */ + func = (void *)regs->gprs[9]; + func((void *)regs->gprs[10]); + } + clear_pt_regs_flag(regs, PIF_SYSCALL); + syscall_exit_to_user_mode(regs); +} + +void flush_thread(void) +{ +} + +void arch_setup_new_exec(void) +{ + if (S390_lowcore.current_pid != current->pid) { + S390_lowcore.current_pid = current->pid; + if (test_facility(40)) + lpp(&S390_lowcore.lpp); + } +} + +void arch_release_task_struct(struct task_struct *tsk) +{ + runtime_instr_release(tsk); + guarded_storage_release(tsk); +} + +int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src) +{ + /* + * Save the floating-point or vector register state of the current + * task and set the CIF_FPU flag to lazy restore the FPU register + * state when returning to user space. + */ + save_fpu_regs(); + + memcpy(dst, src, arch_task_struct_size); + dst->thread.fpu.regs = dst->thread.fpu.fprs; + + /* + * Don't transfer over the runtime instrumentation or the guarded + * storage control block pointers. These fields are cleared here instead + * of in copy_thread() to avoid premature freeing of associated memory + * on fork() failure. Wait to clear the RI flag because ->stack still + * refers to the source thread. + */ + dst->thread.ri_cb = NULL; + dst->thread.gs_cb = NULL; + dst->thread.gs_bc_cb = NULL; + + return 0; +} + +int copy_thread(struct task_struct *p, const struct kernel_clone_args *args) +{ + unsigned long clone_flags = args->flags; + unsigned long new_stackp = args->stack; + unsigned long tls = args->tls; + struct fake_frame + { + struct stack_frame sf; + struct pt_regs childregs; + } *frame; + + frame = container_of(task_pt_regs(p), struct fake_frame, childregs); + p->thread.ksp = (unsigned long) frame; + /* Save access registers to new thread structure. */ + save_access_regs(&p->thread.acrs[0]); + /* start new process with ar4 pointing to the correct address space */ + /* Don't copy debug registers */ + memset(&p->thread.per_user, 0, sizeof(p->thread.per_user)); + memset(&p->thread.per_event, 0, sizeof(p->thread.per_event)); + clear_tsk_thread_flag(p, TIF_SINGLE_STEP); + p->thread.per_flags = 0; + /* Initialize per thread user and system timer values */ + p->thread.user_timer = 0; + p->thread.guest_timer = 0; + p->thread.system_timer = 0; + p->thread.hardirq_timer = 0; + p->thread.softirq_timer = 0; + p->thread.last_break = 1; + + frame->sf.back_chain = 0; + frame->sf.gprs[11 - 6] = (unsigned long)&frame->childregs; + frame->sf.gprs[12 - 6] = (unsigned long)p; + /* new return point is ret_from_fork */ + frame->sf.gprs[14 - 6] = (unsigned long)ret_from_fork; + /* fake return stack for resume(), don't go back to schedule */ + frame->sf.gprs[15 - 6] = (unsigned long)frame; + + /* Store access registers to kernel stack of new process. */ + if (unlikely(args->fn)) { + /* kernel thread */ + memset(&frame->childregs, 0, sizeof(struct pt_regs)); + frame->childregs.psw.mask = PSW_KERNEL_BITS | PSW_MASK_IO | + PSW_MASK_EXT | PSW_MASK_MCHECK; + frame->childregs.gprs[9] = (unsigned long)args->fn; + frame->childregs.gprs[10] = (unsigned long)args->fn_arg; + frame->childregs.orig_gpr2 = -1; + frame->childregs.last_break = 1; + return 0; + } + frame->childregs = *current_pt_regs(); + frame->childregs.gprs[2] = 0; /* child returns 0 on fork. */ + frame->childregs.flags = 0; + if (new_stackp) + frame->childregs.gprs[15] = new_stackp; + /* + * Clear the runtime instrumentation flag after the above childregs + * copy. The CB pointer was already cleared in arch_dup_task_struct(). + */ + frame->childregs.psw.mask &= ~PSW_MASK_RI; + + /* Set a new TLS ? */ + if (clone_flags & CLONE_SETTLS) { + if (is_compat_task()) { + p->thread.acrs[0] = (unsigned int)tls; + } else { + p->thread.acrs[0] = (unsigned int)(tls >> 32); + p->thread.acrs[1] = (unsigned int)tls; + } + } + /* + * s390 stores the svc return address in arch_data when calling + * sigreturn()/restart_syscall() via vdso. 1 means no valid address + * stored. + */ + p->restart_block.arch_data = 1; + return 0; +} + +void execve_tail(void) +{ + current->thread.fpu.fpc = 0; + asm volatile("sfpc %0" : : "d" (0)); +} + +unsigned long __get_wchan(struct task_struct *p) +{ + struct unwind_state state; + unsigned long ip = 0; + + if (!task_stack_page(p)) + return 0; + + if (!try_get_task_stack(p)) + return 0; + + unwind_for_each_frame(&state, p, NULL, 0) { + if (state.stack_info.type != STACK_TYPE_TASK) { + ip = 0; + break; + } + + ip = unwind_get_return_address(&state); + if (!ip) + break; + + if (!in_sched_functions(ip)) + break; + } + + put_task_stack(p); + return ip; +} + +unsigned long arch_align_stack(unsigned long sp) +{ + if (!(current->personality & ADDR_NO_RANDOMIZE) && randomize_va_space) + sp -= get_random_u32_below(PAGE_SIZE); + return sp & ~0xf; +} + +static inline unsigned long brk_rnd(void) +{ + return (get_random_u16() & BRK_RND_MASK) << PAGE_SHIFT; +} + +unsigned long arch_randomize_brk(struct mm_struct *mm) +{ + unsigned long ret; + + ret = PAGE_ALIGN(mm->brk + brk_rnd()); + return (ret > mm->brk) ? ret : mm->brk; +} diff --git a/arch/s390/kernel/processor.c b/arch/s390/kernel/processor.c new file mode 100644 index 0000000000..0a999c8226 --- /dev/null +++ b/arch/s390/kernel/processor.c @@ -0,0 +1,366 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2008 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com) + */ + +#define KMSG_COMPONENT "cpu" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/stop_machine.h> +#include <linux/bitops.h> +#include <linux/kernel.h> +#include <linux/random.h> +#include <linux/sched/mm.h> +#include <linux/init.h> +#include <linux/seq_file.h> +#include <linux/mm_types.h> +#include <linux/delay.h> +#include <linux/cpu.h> + +#include <asm/diag.h> +#include <asm/facility.h> +#include <asm/elf.h> +#include <asm/lowcore.h> +#include <asm/param.h> +#include <asm/sclp.h> +#include <asm/smp.h> + +unsigned long __read_mostly elf_hwcap; +char elf_platform[ELF_PLATFORM_SIZE]; + +struct cpu_info { + unsigned int cpu_mhz_dynamic; + unsigned int cpu_mhz_static; + struct cpuid cpu_id; +}; + +static DEFINE_PER_CPU(struct cpu_info, cpu_info); +static DEFINE_PER_CPU(int, cpu_relax_retry); + +static bool machine_has_cpu_mhz; + +void __init cpu_detect_mhz_feature(void) +{ + if (test_facility(34) && __ecag(ECAG_CPU_ATTRIBUTE, 0) != -1UL) + machine_has_cpu_mhz = true; +} + +static void update_cpu_mhz(void *arg) +{ + unsigned long mhz; + struct cpu_info *c; + + mhz = __ecag(ECAG_CPU_ATTRIBUTE, 0); + c = this_cpu_ptr(&cpu_info); + c->cpu_mhz_dynamic = mhz >> 32; + c->cpu_mhz_static = mhz & 0xffffffff; +} + +void s390_update_cpu_mhz(void) +{ + s390_adjust_jiffies(); + if (machine_has_cpu_mhz) + on_each_cpu(update_cpu_mhz, NULL, 0); +} + +void notrace stop_machine_yield(const struct cpumask *cpumask) +{ + int cpu, this_cpu; + + this_cpu = smp_processor_id(); + if (__this_cpu_inc_return(cpu_relax_retry) >= spin_retry) { + __this_cpu_write(cpu_relax_retry, 0); + cpu = cpumask_next_wrap(this_cpu, cpumask, this_cpu, false); + if (cpu >= nr_cpu_ids) + return; + if (arch_vcpu_is_preempted(cpu)) + smp_yield_cpu(cpu); + } +} + +/* + * cpu_init - initializes state that is per-CPU. + */ +void cpu_init(void) +{ + struct cpuid *id = this_cpu_ptr(&cpu_info.cpu_id); + + get_cpu_id(id); + if (machine_has_cpu_mhz) + update_cpu_mhz(NULL); + mmgrab(&init_mm); + current->active_mm = &init_mm; + BUG_ON(current->mm); + enter_lazy_tlb(&init_mm, current); +} + +static void show_facilities(struct seq_file *m) +{ + unsigned int bit; + + seq_puts(m, "facilities :"); + for_each_set_bit_inv(bit, (long *)&stfle_fac_list, MAX_FACILITY_BIT) + seq_printf(m, " %d", bit); + seq_putc(m, '\n'); +} + +static void show_cpu_summary(struct seq_file *m, void *v) +{ + static const char *hwcap_str[] = { + [HWCAP_NR_ESAN3] = "esan3", + [HWCAP_NR_ZARCH] = "zarch", + [HWCAP_NR_STFLE] = "stfle", + [HWCAP_NR_MSA] = "msa", + [HWCAP_NR_LDISP] = "ldisp", + [HWCAP_NR_EIMM] = "eimm", + [HWCAP_NR_DFP] = "dfp", + [HWCAP_NR_HPAGE] = "edat", + [HWCAP_NR_ETF3EH] = "etf3eh", + [HWCAP_NR_HIGH_GPRS] = "highgprs", + [HWCAP_NR_TE] = "te", + [HWCAP_NR_VXRS] = "vx", + [HWCAP_NR_VXRS_BCD] = "vxd", + [HWCAP_NR_VXRS_EXT] = "vxe", + [HWCAP_NR_GS] = "gs", + [HWCAP_NR_VXRS_EXT2] = "vxe2", + [HWCAP_NR_VXRS_PDE] = "vxp", + [HWCAP_NR_SORT] = "sort", + [HWCAP_NR_DFLT] = "dflt", + [HWCAP_NR_VXRS_PDE2] = "vxp2", + [HWCAP_NR_NNPA] = "nnpa", + [HWCAP_NR_PCI_MIO] = "pcimio", + [HWCAP_NR_SIE] = "sie", + }; + int i, cpu; + + BUILD_BUG_ON(ARRAY_SIZE(hwcap_str) != HWCAP_NR_MAX); + seq_printf(m, "vendor_id : IBM/S390\n" + "# processors : %i\n" + "bogomips per cpu: %lu.%02lu\n", + num_online_cpus(), loops_per_jiffy/(500000/HZ), + (loops_per_jiffy/(5000/HZ))%100); + seq_printf(m, "max thread id : %d\n", smp_cpu_mtid); + seq_puts(m, "features\t: "); + for (i = 0; i < ARRAY_SIZE(hwcap_str); i++) + if (hwcap_str[i] && (elf_hwcap & (1UL << i))) + seq_printf(m, "%s ", hwcap_str[i]); + seq_puts(m, "\n"); + show_facilities(m); + show_cacheinfo(m); + for_each_online_cpu(cpu) { + struct cpuid *id = &per_cpu(cpu_info.cpu_id, cpu); + + seq_printf(m, "processor %d: " + "version = %02X, " + "identification = %06X, " + "machine = %04X\n", + cpu, id->version, id->ident, id->machine); + } +} + +static int __init setup_hwcaps(void) +{ + /* instructions named N3, "backported" to esa-mode */ + elf_hwcap |= HWCAP_ESAN3; + + /* z/Architecture mode active */ + elf_hwcap |= HWCAP_ZARCH; + + /* store-facility-list-extended */ + if (test_facility(7)) + elf_hwcap |= HWCAP_STFLE; + + /* message-security assist */ + if (test_facility(17)) + elf_hwcap |= HWCAP_MSA; + + /* long-displacement */ + if (test_facility(19)) + elf_hwcap |= HWCAP_LDISP; + + /* extended-immediate */ + elf_hwcap |= HWCAP_EIMM; + + /* extended-translation facility 3 enhancement */ + if (test_facility(22) && test_facility(30)) + elf_hwcap |= HWCAP_ETF3EH; + + /* decimal floating point & perform floating point operation */ + if (test_facility(42) && test_facility(44)) + elf_hwcap |= HWCAP_DFP; + + /* huge page support */ + if (MACHINE_HAS_EDAT1) + elf_hwcap |= HWCAP_HPAGE; + + /* 64-bit register support for 31-bit processes */ + elf_hwcap |= HWCAP_HIGH_GPRS; + + /* transactional execution */ + if (MACHINE_HAS_TE) + elf_hwcap |= HWCAP_TE; + + /* + * Vector extension can be disabled with the "novx" parameter. + * Use MACHINE_HAS_VX instead of facility bit 129. + */ + if (MACHINE_HAS_VX) { + elf_hwcap |= HWCAP_VXRS; + if (test_facility(134)) + elf_hwcap |= HWCAP_VXRS_BCD; + if (test_facility(135)) + elf_hwcap |= HWCAP_VXRS_EXT; + if (test_facility(148)) + elf_hwcap |= HWCAP_VXRS_EXT2; + if (test_facility(152)) + elf_hwcap |= HWCAP_VXRS_PDE; + if (test_facility(192)) + elf_hwcap |= HWCAP_VXRS_PDE2; + } + + if (test_facility(150)) + elf_hwcap |= HWCAP_SORT; + + if (test_facility(151)) + elf_hwcap |= HWCAP_DFLT; + + if (test_facility(165)) + elf_hwcap |= HWCAP_NNPA; + + /* guarded storage */ + if (MACHINE_HAS_GS) + elf_hwcap |= HWCAP_GS; + + if (MACHINE_HAS_PCI_MIO) + elf_hwcap |= HWCAP_PCI_MIO; + + /* virtualization support */ + if (sclp.has_sief2) + elf_hwcap |= HWCAP_SIE; + + return 0; +} +arch_initcall(setup_hwcaps); + +static int __init setup_elf_platform(void) +{ + struct cpuid cpu_id; + + get_cpu_id(&cpu_id); + add_device_randomness(&cpu_id, sizeof(cpu_id)); + switch (cpu_id.machine) { + default: /* Use "z10" as default. */ + strcpy(elf_platform, "z10"); + break; + case 0x2817: + case 0x2818: + strcpy(elf_platform, "z196"); + break; + case 0x2827: + case 0x2828: + strcpy(elf_platform, "zEC12"); + break; + case 0x2964: + case 0x2965: + strcpy(elf_platform, "z13"); + break; + case 0x3906: + case 0x3907: + strcpy(elf_platform, "z14"); + break; + case 0x8561: + case 0x8562: + strcpy(elf_platform, "z15"); + break; + case 0x3931: + case 0x3932: + strcpy(elf_platform, "z16"); + break; + } + return 0; +} +arch_initcall(setup_elf_platform); + +static void show_cpu_topology(struct seq_file *m, unsigned long n) +{ +#ifdef CONFIG_SCHED_TOPOLOGY + seq_printf(m, "physical id : %d\n", topology_physical_package_id(n)); + seq_printf(m, "core id : %d\n", topology_core_id(n)); + seq_printf(m, "book id : %d\n", topology_book_id(n)); + seq_printf(m, "drawer id : %d\n", topology_drawer_id(n)); + seq_printf(m, "dedicated : %d\n", topology_cpu_dedicated(n)); + seq_printf(m, "address : %d\n", smp_cpu_get_cpu_address(n)); + seq_printf(m, "siblings : %d\n", cpumask_weight(topology_core_cpumask(n))); + seq_printf(m, "cpu cores : %d\n", topology_booted_cores(n)); +#endif /* CONFIG_SCHED_TOPOLOGY */ +} + +static void show_cpu_ids(struct seq_file *m, unsigned long n) +{ + struct cpuid *id = &per_cpu(cpu_info.cpu_id, n); + + seq_printf(m, "version : %02X\n", id->version); + seq_printf(m, "identification : %06X\n", id->ident); + seq_printf(m, "machine : %04X\n", id->machine); +} + +static void show_cpu_mhz(struct seq_file *m, unsigned long n) +{ + struct cpu_info *c = per_cpu_ptr(&cpu_info, n); + + if (!machine_has_cpu_mhz) + return; + seq_printf(m, "cpu MHz dynamic : %d\n", c->cpu_mhz_dynamic); + seq_printf(m, "cpu MHz static : %d\n", c->cpu_mhz_static); +} + +/* + * show_cpuinfo - Get information on one CPU for use by procfs. + */ +static int show_cpuinfo(struct seq_file *m, void *v) +{ + unsigned long n = (unsigned long) v - 1; + unsigned long first = cpumask_first(cpu_online_mask); + + if (n == first) + show_cpu_summary(m, v); + seq_printf(m, "\ncpu number : %ld\n", n); + show_cpu_topology(m, n); + show_cpu_ids(m, n); + show_cpu_mhz(m, n); + return 0; +} + +static inline void *c_update(loff_t *pos) +{ + if (*pos) + *pos = cpumask_next(*pos - 1, cpu_online_mask); + else + *pos = cpumask_first(cpu_online_mask); + return *pos < nr_cpu_ids ? (void *)*pos + 1 : NULL; +} + +static void *c_start(struct seq_file *m, loff_t *pos) +{ + cpus_read_lock(); + return c_update(pos); +} + +static void *c_next(struct seq_file *m, void *v, loff_t *pos) +{ + ++*pos; + return c_update(pos); +} + +static void c_stop(struct seq_file *m, void *v) +{ + cpus_read_unlock(); +} + +const struct seq_operations cpuinfo_op = { + .start = c_start, + .next = c_next, + .stop = c_stop, + .show = show_cpuinfo, +}; diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c new file mode 100644 index 0000000000..ea244a73ef --- /dev/null +++ b/arch/s390/kernel/ptrace.c @@ -0,0 +1,1608 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Ptrace user space interface. + * + * Copyright IBM Corp. 1999, 2010 + * Author(s): Denis Joseph Barrow + * Martin Schwidefsky (schwidefsky@de.ibm.com) + */ + +#include "asm/ptrace.h" +#include <linux/kernel.h> +#include <linux/sched.h> +#include <linux/sched/task_stack.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/errno.h> +#include <linux/ptrace.h> +#include <linux/user.h> +#include <linux/security.h> +#include <linux/audit.h> +#include <linux/signal.h> +#include <linux/elf.h> +#include <linux/regset.h> +#include <linux/seccomp.h> +#include <linux/compat.h> +#include <trace/syscall.h> +#include <asm/page.h> +#include <linux/uaccess.h> +#include <asm/unistd.h> +#include <asm/switch_to.h> +#include <asm/runtime_instr.h> +#include <asm/facility.h> + +#include "entry.h" + +#ifdef CONFIG_COMPAT +#include "compat_ptrace.h" +#endif + +void update_cr_regs(struct task_struct *task) +{ + struct pt_regs *regs = task_pt_regs(task); + struct thread_struct *thread = &task->thread; + struct per_regs old, new; + union ctlreg0 cr0_old, cr0_new; + union ctlreg2 cr2_old, cr2_new; + int cr0_changed, cr2_changed; + + __ctl_store(cr0_old.val, 0, 0); + __ctl_store(cr2_old.val, 2, 2); + cr0_new = cr0_old; + cr2_new = cr2_old; + /* Take care of the enable/disable of transactional execution. */ + if (MACHINE_HAS_TE) { + /* Set or clear transaction execution TXC bit 8. */ + cr0_new.tcx = 1; + if (task->thread.per_flags & PER_FLAG_NO_TE) + cr0_new.tcx = 0; + /* Set or clear transaction execution TDC bits 62 and 63. */ + cr2_new.tdc = 0; + if (task->thread.per_flags & PER_FLAG_TE_ABORT_RAND) { + if (task->thread.per_flags & PER_FLAG_TE_ABORT_RAND_TEND) + cr2_new.tdc = 1; + else + cr2_new.tdc = 2; + } + } + /* Take care of enable/disable of guarded storage. */ + if (MACHINE_HAS_GS) { + cr2_new.gse = 0; + if (task->thread.gs_cb) + cr2_new.gse = 1; + } + /* Load control register 0/2 iff changed */ + cr0_changed = cr0_new.val != cr0_old.val; + cr2_changed = cr2_new.val != cr2_old.val; + if (cr0_changed) + __ctl_load(cr0_new.val, 0, 0); + if (cr2_changed) + __ctl_load(cr2_new.val, 2, 2); + /* Copy user specified PER registers */ + new.control = thread->per_user.control; + new.start = thread->per_user.start; + new.end = thread->per_user.end; + + /* merge TIF_SINGLE_STEP into user specified PER registers. */ + if (test_tsk_thread_flag(task, TIF_SINGLE_STEP) || + test_tsk_thread_flag(task, TIF_UPROBE_SINGLESTEP)) { + if (test_tsk_thread_flag(task, TIF_BLOCK_STEP)) + new.control |= PER_EVENT_BRANCH; + else + new.control |= PER_EVENT_IFETCH; + new.control |= PER_CONTROL_SUSPENSION; + new.control |= PER_EVENT_TRANSACTION_END; + if (test_tsk_thread_flag(task, TIF_UPROBE_SINGLESTEP)) + new.control |= PER_EVENT_IFETCH; + new.start = 0; + new.end = -1UL; + } + + /* Take care of the PER enablement bit in the PSW. */ + if (!(new.control & PER_EVENT_MASK)) { + regs->psw.mask &= ~PSW_MASK_PER; + return; + } + regs->psw.mask |= PSW_MASK_PER; + __ctl_store(old, 9, 11); + if (memcmp(&new, &old, sizeof(struct per_regs)) != 0) + __ctl_load(new, 9, 11); +} + +void user_enable_single_step(struct task_struct *task) +{ + clear_tsk_thread_flag(task, TIF_BLOCK_STEP); + set_tsk_thread_flag(task, TIF_SINGLE_STEP); +} + +void user_disable_single_step(struct task_struct *task) +{ + clear_tsk_thread_flag(task, TIF_BLOCK_STEP); + clear_tsk_thread_flag(task, TIF_SINGLE_STEP); +} + +void user_enable_block_step(struct task_struct *task) +{ + set_tsk_thread_flag(task, TIF_SINGLE_STEP); + set_tsk_thread_flag(task, TIF_BLOCK_STEP); +} + +/* + * Called by kernel/ptrace.c when detaching.. + * + * Clear all debugging related fields. + */ +void ptrace_disable(struct task_struct *task) +{ + memset(&task->thread.per_user, 0, sizeof(task->thread.per_user)); + memset(&task->thread.per_event, 0, sizeof(task->thread.per_event)); + clear_tsk_thread_flag(task, TIF_SINGLE_STEP); + clear_tsk_thread_flag(task, TIF_PER_TRAP); + task->thread.per_flags = 0; +} + +#define __ADDR_MASK 7 + +static inline unsigned long __peek_user_per(struct task_struct *child, + addr_t addr) +{ + if (addr == offsetof(struct per_struct_kernel, cr9)) + /* Control bits of the active per set. */ + return test_thread_flag(TIF_SINGLE_STEP) ? + PER_EVENT_IFETCH : child->thread.per_user.control; + else if (addr == offsetof(struct per_struct_kernel, cr10)) + /* Start address of the active per set. */ + return test_thread_flag(TIF_SINGLE_STEP) ? + 0 : child->thread.per_user.start; + else if (addr == offsetof(struct per_struct_kernel, cr11)) + /* End address of the active per set. */ + return test_thread_flag(TIF_SINGLE_STEP) ? + -1UL : child->thread.per_user.end; + else if (addr == offsetof(struct per_struct_kernel, bits)) + /* Single-step bit. */ + return test_thread_flag(TIF_SINGLE_STEP) ? + (1UL << (BITS_PER_LONG - 1)) : 0; + else if (addr == offsetof(struct per_struct_kernel, starting_addr)) + /* Start address of the user specified per set. */ + return child->thread.per_user.start; + else if (addr == offsetof(struct per_struct_kernel, ending_addr)) + /* End address of the user specified per set. */ + return child->thread.per_user.end; + else if (addr == offsetof(struct per_struct_kernel, perc_atmid)) + /* PER code, ATMID and AI of the last PER trap */ + return (unsigned long) + child->thread.per_event.cause << (BITS_PER_LONG - 16); + else if (addr == offsetof(struct per_struct_kernel, address)) + /* Address of the last PER trap */ + return child->thread.per_event.address; + else if (addr == offsetof(struct per_struct_kernel, access_id)) + /* Access id of the last PER trap */ + return (unsigned long) + child->thread.per_event.paid << (BITS_PER_LONG - 8); + return 0; +} + +/* + * Read the word at offset addr from the user area of a process. The + * trouble here is that the information is littered over different + * locations. The process registers are found on the kernel stack, + * the floating point stuff and the trace settings are stored in + * the task structure. In addition the different structures in + * struct user contain pad bytes that should be read as zeroes. + * Lovely... + */ +static unsigned long __peek_user(struct task_struct *child, addr_t addr) +{ + addr_t offset, tmp; + + if (addr < offsetof(struct user, regs.acrs)) { + /* + * psw and gprs are stored on the stack + */ + tmp = *(addr_t *)((addr_t) &task_pt_regs(child)->psw + addr); + if (addr == offsetof(struct user, regs.psw.mask)) { + /* Return a clean psw mask. */ + tmp &= PSW_MASK_USER | PSW_MASK_RI; + tmp |= PSW_USER_BITS; + } + + } else if (addr < offsetof(struct user, regs.orig_gpr2)) { + /* + * access registers are stored in the thread structure + */ + offset = addr - offsetof(struct user, regs.acrs); + /* + * Very special case: old & broken 64 bit gdb reading + * from acrs[15]. Result is a 64 bit value. Read the + * 32 bit acrs[15] value and shift it by 32. Sick... + */ + if (addr == offsetof(struct user, regs.acrs[15])) + tmp = ((unsigned long) child->thread.acrs[15]) << 32; + else + tmp = *(addr_t *)((addr_t) &child->thread.acrs + offset); + + } else if (addr == offsetof(struct user, regs.orig_gpr2)) { + /* + * orig_gpr2 is stored on the kernel stack + */ + tmp = (addr_t) task_pt_regs(child)->orig_gpr2; + + } else if (addr < offsetof(struct user, regs.fp_regs)) { + /* + * prevent reads of padding hole between + * orig_gpr2 and fp_regs on s390. + */ + tmp = 0; + + } else if (addr == offsetof(struct user, regs.fp_regs.fpc)) { + /* + * floating point control reg. is in the thread structure + */ + tmp = child->thread.fpu.fpc; + tmp <<= BITS_PER_LONG - 32; + + } else if (addr < offsetof(struct user, regs.fp_regs) + sizeof(s390_fp_regs)) { + /* + * floating point regs. are either in child->thread.fpu + * or the child->thread.fpu.vxrs array + */ + offset = addr - offsetof(struct user, regs.fp_regs.fprs); + if (MACHINE_HAS_VX) + tmp = *(addr_t *) + ((addr_t) child->thread.fpu.vxrs + 2*offset); + else + tmp = *(addr_t *) + ((addr_t) child->thread.fpu.fprs + offset); + + } else if (addr < offsetof(struct user, regs.per_info) + sizeof(per_struct)) { + /* + * Handle access to the per_info structure. + */ + addr -= offsetof(struct user, regs.per_info); + tmp = __peek_user_per(child, addr); + + } else + tmp = 0; + + return tmp; +} + +static int +peek_user(struct task_struct *child, addr_t addr, addr_t data) +{ + addr_t tmp, mask; + + /* + * Stupid gdb peeks/pokes the access registers in 64 bit with + * an alignment of 4. Programmers from hell... + */ + mask = __ADDR_MASK; + if (addr >= offsetof(struct user, regs.acrs) && + addr < offsetof(struct user, regs.orig_gpr2)) + mask = 3; + if ((addr & mask) || addr > sizeof(struct user) - __ADDR_MASK) + return -EIO; + + tmp = __peek_user(child, addr); + return put_user(tmp, (addr_t __user *) data); +} + +static inline void __poke_user_per(struct task_struct *child, + addr_t addr, addr_t data) +{ + /* + * There are only three fields in the per_info struct that the + * debugger user can write to. + * 1) cr9: the debugger wants to set a new PER event mask + * 2) starting_addr: the debugger wants to set a new starting + * address to use with the PER event mask. + * 3) ending_addr: the debugger wants to set a new ending + * address to use with the PER event mask. + * The user specified PER event mask and the start and end + * addresses are used only if single stepping is not in effect. + * Writes to any other field in per_info are ignored. + */ + if (addr == offsetof(struct per_struct_kernel, cr9)) + /* PER event mask of the user specified per set. */ + child->thread.per_user.control = + data & (PER_EVENT_MASK | PER_CONTROL_MASK); + else if (addr == offsetof(struct per_struct_kernel, starting_addr)) + /* Starting address of the user specified per set. */ + child->thread.per_user.start = data; + else if (addr == offsetof(struct per_struct_kernel, ending_addr)) + /* Ending address of the user specified per set. */ + child->thread.per_user.end = data; +} + +/* + * Write a word to the user area of a process at location addr. This + * operation does have an additional problem compared to peek_user. + * Stores to the program status word and on the floating point + * control register needs to get checked for validity. + */ +static int __poke_user(struct task_struct *child, addr_t addr, addr_t data) +{ + addr_t offset; + + + if (addr < offsetof(struct user, regs.acrs)) { + struct pt_regs *regs = task_pt_regs(child); + /* + * psw and gprs are stored on the stack + */ + if (addr == offsetof(struct user, regs.psw.mask)) { + unsigned long mask = PSW_MASK_USER; + + mask |= is_ri_task(child) ? PSW_MASK_RI : 0; + if ((data ^ PSW_USER_BITS) & ~mask) + /* Invalid psw mask. */ + return -EINVAL; + if ((data & PSW_MASK_ASC) == PSW_ASC_HOME) + /* Invalid address-space-control bits */ + return -EINVAL; + if ((data & PSW_MASK_EA) && !(data & PSW_MASK_BA)) + /* Invalid addressing mode bits */ + return -EINVAL; + } + + if (test_pt_regs_flag(regs, PIF_SYSCALL) && + addr == offsetof(struct user, regs.gprs[2])) { + struct pt_regs *regs = task_pt_regs(child); + + regs->int_code = 0x20000 | (data & 0xffff); + } + *(addr_t *)((addr_t) ®s->psw + addr) = data; + } else if (addr < offsetof(struct user, regs.orig_gpr2)) { + /* + * access registers are stored in the thread structure + */ + offset = addr - offsetof(struct user, regs.acrs); + /* + * Very special case: old & broken 64 bit gdb writing + * to acrs[15] with a 64 bit value. Ignore the lower + * half of the value and write the upper 32 bit to + * acrs[15]. Sick... + */ + if (addr == offsetof(struct user, regs.acrs[15])) + child->thread.acrs[15] = (unsigned int) (data >> 32); + else + *(addr_t *)((addr_t) &child->thread.acrs + offset) = data; + + } else if (addr == offsetof(struct user, regs.orig_gpr2)) { + /* + * orig_gpr2 is stored on the kernel stack + */ + task_pt_regs(child)->orig_gpr2 = data; + + } else if (addr < offsetof(struct user, regs.fp_regs)) { + /* + * prevent writes of padding hole between + * orig_gpr2 and fp_regs on s390. + */ + return 0; + + } else if (addr == offsetof(struct user, regs.fp_regs.fpc)) { + /* + * floating point control reg. is in the thread structure + */ + if ((unsigned int) data != 0 || + test_fp_ctl(data >> (BITS_PER_LONG - 32))) + return -EINVAL; + child->thread.fpu.fpc = data >> (BITS_PER_LONG - 32); + + } else if (addr < offsetof(struct user, regs.fp_regs) + sizeof(s390_fp_regs)) { + /* + * floating point regs. are either in child->thread.fpu + * or the child->thread.fpu.vxrs array + */ + offset = addr - offsetof(struct user, regs.fp_regs.fprs); + if (MACHINE_HAS_VX) + *(addr_t *)((addr_t) + child->thread.fpu.vxrs + 2*offset) = data; + else + *(addr_t *)((addr_t) + child->thread.fpu.fprs + offset) = data; + + } else if (addr < offsetof(struct user, regs.per_info) + sizeof(per_struct)) { + /* + * Handle access to the per_info structure. + */ + addr -= offsetof(struct user, regs.per_info); + __poke_user_per(child, addr, data); + + } + + return 0; +} + +static int poke_user(struct task_struct *child, addr_t addr, addr_t data) +{ + addr_t mask; + + /* + * Stupid gdb peeks/pokes the access registers in 64 bit with + * an alignment of 4. Programmers from hell indeed... + */ + mask = __ADDR_MASK; + if (addr >= offsetof(struct user, regs.acrs) && + addr < offsetof(struct user, regs.orig_gpr2)) + mask = 3; + if ((addr & mask) || addr > sizeof(struct user) - __ADDR_MASK) + return -EIO; + + return __poke_user(child, addr, data); +} + +long arch_ptrace(struct task_struct *child, long request, + unsigned long addr, unsigned long data) +{ + ptrace_area parea; + int copied, ret; + + switch (request) { + case PTRACE_PEEKUSR: + /* read the word at location addr in the USER area. */ + return peek_user(child, addr, data); + + case PTRACE_POKEUSR: + /* write the word at location addr in the USER area */ + return poke_user(child, addr, data); + + case PTRACE_PEEKUSR_AREA: + case PTRACE_POKEUSR_AREA: + if (copy_from_user(&parea, (void __force __user *) addr, + sizeof(parea))) + return -EFAULT; + addr = parea.kernel_addr; + data = parea.process_addr; + copied = 0; + while (copied < parea.len) { + if (request == PTRACE_PEEKUSR_AREA) + ret = peek_user(child, addr, data); + else { + addr_t utmp; + if (get_user(utmp, + (addr_t __force __user *) data)) + return -EFAULT; + ret = poke_user(child, addr, utmp); + } + if (ret) + return ret; + addr += sizeof(unsigned long); + data += sizeof(unsigned long); + copied += sizeof(unsigned long); + } + return 0; + case PTRACE_GET_LAST_BREAK: + return put_user(child->thread.last_break, (unsigned long __user *)data); + case PTRACE_ENABLE_TE: + if (!MACHINE_HAS_TE) + return -EIO; + child->thread.per_flags &= ~PER_FLAG_NO_TE; + return 0; + case PTRACE_DISABLE_TE: + if (!MACHINE_HAS_TE) + return -EIO; + child->thread.per_flags |= PER_FLAG_NO_TE; + child->thread.per_flags &= ~PER_FLAG_TE_ABORT_RAND; + return 0; + case PTRACE_TE_ABORT_RAND: + if (!MACHINE_HAS_TE || (child->thread.per_flags & PER_FLAG_NO_TE)) + return -EIO; + switch (data) { + case 0UL: + child->thread.per_flags &= ~PER_FLAG_TE_ABORT_RAND; + break; + case 1UL: + child->thread.per_flags |= PER_FLAG_TE_ABORT_RAND; + child->thread.per_flags |= PER_FLAG_TE_ABORT_RAND_TEND; + break; + case 2UL: + child->thread.per_flags |= PER_FLAG_TE_ABORT_RAND; + child->thread.per_flags &= ~PER_FLAG_TE_ABORT_RAND_TEND; + break; + default: + return -EINVAL; + } + return 0; + default: + return ptrace_request(child, request, addr, data); + } +} + +#ifdef CONFIG_COMPAT +/* + * Now the fun part starts... a 31 bit program running in the + * 31 bit emulation tracing another program. PTRACE_PEEKTEXT, + * PTRACE_PEEKDATA, PTRACE_POKETEXT and PTRACE_POKEDATA are easy + * to handle, the difference to the 64 bit versions of the requests + * is that the access is done in multiples of 4 byte instead of + * 8 bytes (sizeof(unsigned long) on 31/64 bit). + * The ugly part are PTRACE_PEEKUSR, PTRACE_PEEKUSR_AREA, + * PTRACE_POKEUSR and PTRACE_POKEUSR_AREA. If the traced program + * is a 31 bit program too, the content of struct user can be + * emulated. A 31 bit program peeking into the struct user of + * a 64 bit program is a no-no. + */ + +/* + * Same as peek_user_per but for a 31 bit program. + */ +static inline __u32 __peek_user_per_compat(struct task_struct *child, + addr_t addr) +{ + if (addr == offsetof(struct compat_per_struct_kernel, cr9)) + /* Control bits of the active per set. */ + return (__u32) test_thread_flag(TIF_SINGLE_STEP) ? + PER_EVENT_IFETCH : child->thread.per_user.control; + else if (addr == offsetof(struct compat_per_struct_kernel, cr10)) + /* Start address of the active per set. */ + return (__u32) test_thread_flag(TIF_SINGLE_STEP) ? + 0 : child->thread.per_user.start; + else if (addr == offsetof(struct compat_per_struct_kernel, cr11)) + /* End address of the active per set. */ + return test_thread_flag(TIF_SINGLE_STEP) ? + PSW32_ADDR_INSN : child->thread.per_user.end; + else if (addr == offsetof(struct compat_per_struct_kernel, bits)) + /* Single-step bit. */ + return (__u32) test_thread_flag(TIF_SINGLE_STEP) ? + 0x80000000 : 0; + else if (addr == offsetof(struct compat_per_struct_kernel, starting_addr)) + /* Start address of the user specified per set. */ + return (__u32) child->thread.per_user.start; + else if (addr == offsetof(struct compat_per_struct_kernel, ending_addr)) + /* End address of the user specified per set. */ + return (__u32) child->thread.per_user.end; + else if (addr == offsetof(struct compat_per_struct_kernel, perc_atmid)) + /* PER code, ATMID and AI of the last PER trap */ + return (__u32) child->thread.per_event.cause << 16; + else if (addr == offsetof(struct compat_per_struct_kernel, address)) + /* Address of the last PER trap */ + return (__u32) child->thread.per_event.address; + else if (addr == offsetof(struct compat_per_struct_kernel, access_id)) + /* Access id of the last PER trap */ + return (__u32) child->thread.per_event.paid << 24; + return 0; +} + +/* + * Same as peek_user but for a 31 bit program. + */ +static u32 __peek_user_compat(struct task_struct *child, addr_t addr) +{ + addr_t offset; + __u32 tmp; + + if (addr < offsetof(struct compat_user, regs.acrs)) { + struct pt_regs *regs = task_pt_regs(child); + /* + * psw and gprs are stored on the stack + */ + if (addr == offsetof(struct compat_user, regs.psw.mask)) { + /* Fake a 31 bit psw mask. */ + tmp = (__u32)(regs->psw.mask >> 32); + tmp &= PSW32_MASK_USER | PSW32_MASK_RI; + tmp |= PSW32_USER_BITS; + } else if (addr == offsetof(struct compat_user, regs.psw.addr)) { + /* Fake a 31 bit psw address. */ + tmp = (__u32) regs->psw.addr | + (__u32)(regs->psw.mask & PSW_MASK_BA); + } else { + /* gpr 0-15 */ + tmp = *(__u32 *)((addr_t) ®s->psw + addr*2 + 4); + } + } else if (addr < offsetof(struct compat_user, regs.orig_gpr2)) { + /* + * access registers are stored in the thread structure + */ + offset = addr - offsetof(struct compat_user, regs.acrs); + tmp = *(__u32*)((addr_t) &child->thread.acrs + offset); + + } else if (addr == offsetof(struct compat_user, regs.orig_gpr2)) { + /* + * orig_gpr2 is stored on the kernel stack + */ + tmp = *(__u32*)((addr_t) &task_pt_regs(child)->orig_gpr2 + 4); + + } else if (addr < offsetof(struct compat_user, regs.fp_regs)) { + /* + * prevent reads of padding hole between + * orig_gpr2 and fp_regs on s390. + */ + tmp = 0; + + } else if (addr == offsetof(struct compat_user, regs.fp_regs.fpc)) { + /* + * floating point control reg. is in the thread structure + */ + tmp = child->thread.fpu.fpc; + + } else if (addr < offsetof(struct compat_user, regs.fp_regs) + sizeof(s390_fp_regs)) { + /* + * floating point regs. are either in child->thread.fpu + * or the child->thread.fpu.vxrs array + */ + offset = addr - offsetof(struct compat_user, regs.fp_regs.fprs); + if (MACHINE_HAS_VX) + tmp = *(__u32 *) + ((addr_t) child->thread.fpu.vxrs + 2*offset); + else + tmp = *(__u32 *) + ((addr_t) child->thread.fpu.fprs + offset); + + } else if (addr < offsetof(struct compat_user, regs.per_info) + sizeof(struct compat_per_struct_kernel)) { + /* + * Handle access to the per_info structure. + */ + addr -= offsetof(struct compat_user, regs.per_info); + tmp = __peek_user_per_compat(child, addr); + + } else + tmp = 0; + + return tmp; +} + +static int peek_user_compat(struct task_struct *child, + addr_t addr, addr_t data) +{ + __u32 tmp; + + if (!is_compat_task() || (addr & 3) || addr > sizeof(struct user) - 3) + return -EIO; + + tmp = __peek_user_compat(child, addr); + return put_user(tmp, (__u32 __user *) data); +} + +/* + * Same as poke_user_per but for a 31 bit program. + */ +static inline void __poke_user_per_compat(struct task_struct *child, + addr_t addr, __u32 data) +{ + if (addr == offsetof(struct compat_per_struct_kernel, cr9)) + /* PER event mask of the user specified per set. */ + child->thread.per_user.control = + data & (PER_EVENT_MASK | PER_CONTROL_MASK); + else if (addr == offsetof(struct compat_per_struct_kernel, starting_addr)) + /* Starting address of the user specified per set. */ + child->thread.per_user.start = data; + else if (addr == offsetof(struct compat_per_struct_kernel, ending_addr)) + /* Ending address of the user specified per set. */ + child->thread.per_user.end = data; +} + +/* + * Same as poke_user but for a 31 bit program. + */ +static int __poke_user_compat(struct task_struct *child, + addr_t addr, addr_t data) +{ + __u32 tmp = (__u32) data; + addr_t offset; + + if (addr < offsetof(struct compat_user, regs.acrs)) { + struct pt_regs *regs = task_pt_regs(child); + /* + * psw, gprs, acrs and orig_gpr2 are stored on the stack + */ + if (addr == offsetof(struct compat_user, regs.psw.mask)) { + __u32 mask = PSW32_MASK_USER; + + mask |= is_ri_task(child) ? PSW32_MASK_RI : 0; + /* Build a 64 bit psw mask from 31 bit mask. */ + if ((tmp ^ PSW32_USER_BITS) & ~mask) + /* Invalid psw mask. */ + return -EINVAL; + if ((data & PSW32_MASK_ASC) == PSW32_ASC_HOME) + /* Invalid address-space-control bits */ + return -EINVAL; + regs->psw.mask = (regs->psw.mask & ~PSW_MASK_USER) | + (regs->psw.mask & PSW_MASK_BA) | + (__u64)(tmp & mask) << 32; + } else if (addr == offsetof(struct compat_user, regs.psw.addr)) { + /* Build a 64 bit psw address from 31 bit address. */ + regs->psw.addr = (__u64) tmp & PSW32_ADDR_INSN; + /* Transfer 31 bit amode bit to psw mask. */ + regs->psw.mask = (regs->psw.mask & ~PSW_MASK_BA) | + (__u64)(tmp & PSW32_ADDR_AMODE); + } else { + if (test_pt_regs_flag(regs, PIF_SYSCALL) && + addr == offsetof(struct compat_user, regs.gprs[2])) { + struct pt_regs *regs = task_pt_regs(child); + + regs->int_code = 0x20000 | (data & 0xffff); + } + /* gpr 0-15 */ + *(__u32*)((addr_t) ®s->psw + addr*2 + 4) = tmp; + } + } else if (addr < offsetof(struct compat_user, regs.orig_gpr2)) { + /* + * access registers are stored in the thread structure + */ + offset = addr - offsetof(struct compat_user, regs.acrs); + *(__u32*)((addr_t) &child->thread.acrs + offset) = tmp; + + } else if (addr == offsetof(struct compat_user, regs.orig_gpr2)) { + /* + * orig_gpr2 is stored on the kernel stack + */ + *(__u32*)((addr_t) &task_pt_regs(child)->orig_gpr2 + 4) = tmp; + + } else if (addr < offsetof(struct compat_user, regs.fp_regs)) { + /* + * prevent writess of padding hole between + * orig_gpr2 and fp_regs on s390. + */ + return 0; + + } else if (addr == offsetof(struct compat_user, regs.fp_regs.fpc)) { + /* + * floating point control reg. is in the thread structure + */ + if (test_fp_ctl(tmp)) + return -EINVAL; + child->thread.fpu.fpc = data; + + } else if (addr < offsetof(struct compat_user, regs.fp_regs) + sizeof(s390_fp_regs)) { + /* + * floating point regs. are either in child->thread.fpu + * or the child->thread.fpu.vxrs array + */ + offset = addr - offsetof(struct compat_user, regs.fp_regs.fprs); + if (MACHINE_HAS_VX) + *(__u32 *)((addr_t) + child->thread.fpu.vxrs + 2*offset) = tmp; + else + *(__u32 *)((addr_t) + child->thread.fpu.fprs + offset) = tmp; + + } else if (addr < offsetof(struct compat_user, regs.per_info) + sizeof(struct compat_per_struct_kernel)) { + /* + * Handle access to the per_info structure. + */ + addr -= offsetof(struct compat_user, regs.per_info); + __poke_user_per_compat(child, addr, data); + } + + return 0; +} + +static int poke_user_compat(struct task_struct *child, + addr_t addr, addr_t data) +{ + if (!is_compat_task() || (addr & 3) || + addr > sizeof(struct compat_user) - 3) + return -EIO; + + return __poke_user_compat(child, addr, data); +} + +long compat_arch_ptrace(struct task_struct *child, compat_long_t request, + compat_ulong_t caddr, compat_ulong_t cdata) +{ + unsigned long addr = caddr; + unsigned long data = cdata; + compat_ptrace_area parea; + int copied, ret; + + switch (request) { + case PTRACE_PEEKUSR: + /* read the word at location addr in the USER area. */ + return peek_user_compat(child, addr, data); + + case PTRACE_POKEUSR: + /* write the word at location addr in the USER area */ + return poke_user_compat(child, addr, data); + + case PTRACE_PEEKUSR_AREA: + case PTRACE_POKEUSR_AREA: + if (copy_from_user(&parea, (void __force __user *) addr, + sizeof(parea))) + return -EFAULT; + addr = parea.kernel_addr; + data = parea.process_addr; + copied = 0; + while (copied < parea.len) { + if (request == PTRACE_PEEKUSR_AREA) + ret = peek_user_compat(child, addr, data); + else { + __u32 utmp; + if (get_user(utmp, + (__u32 __force __user *) data)) + return -EFAULT; + ret = poke_user_compat(child, addr, utmp); + } + if (ret) + return ret; + addr += sizeof(unsigned int); + data += sizeof(unsigned int); + copied += sizeof(unsigned int); + } + return 0; + case PTRACE_GET_LAST_BREAK: + return put_user(child->thread.last_break, (unsigned int __user *)data); + } + return compat_ptrace_request(child, request, addr, data); +} +#endif + +/* + * user_regset definitions. + */ + +static int s390_regs_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + unsigned pos; + if (target == current) + save_access_regs(target->thread.acrs); + + for (pos = 0; pos < sizeof(s390_regs); pos += sizeof(long)) + membuf_store(&to, __peek_user(target, pos)); + return 0; +} + +static int s390_regs_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + int rc = 0; + + if (target == current) + save_access_regs(target->thread.acrs); + + if (kbuf) { + const unsigned long *k = kbuf; + while (count > 0 && !rc) { + rc = __poke_user(target, pos, *k++); + count -= sizeof(*k); + pos += sizeof(*k); + } + } else { + const unsigned long __user *u = ubuf; + while (count > 0 && !rc) { + unsigned long word; + rc = __get_user(word, u++); + if (rc) + break; + rc = __poke_user(target, pos, word); + count -= sizeof(*u); + pos += sizeof(*u); + } + } + + if (rc == 0 && target == current) + restore_access_regs(target->thread.acrs); + + return rc; +} + +static int s390_fpregs_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + _s390_fp_regs fp_regs; + + if (target == current) + save_fpu_regs(); + + fp_regs.fpc = target->thread.fpu.fpc; + fpregs_store(&fp_regs, &target->thread.fpu); + + return membuf_write(&to, &fp_regs, sizeof(fp_regs)); +} + +static int s390_fpregs_set(struct task_struct *target, + const struct user_regset *regset, unsigned int pos, + unsigned int count, const void *kbuf, + const void __user *ubuf) +{ + int rc = 0; + freg_t fprs[__NUM_FPRS]; + + if (target == current) + save_fpu_regs(); + + if (MACHINE_HAS_VX) + convert_vx_to_fp(fprs, target->thread.fpu.vxrs); + else + memcpy(&fprs, target->thread.fpu.fprs, sizeof(fprs)); + + /* If setting FPC, must validate it first. */ + if (count > 0 && pos < offsetof(s390_fp_regs, fprs)) { + u32 ufpc[2] = { target->thread.fpu.fpc, 0 }; + rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, &ufpc, + 0, offsetof(s390_fp_regs, fprs)); + if (rc) + return rc; + if (ufpc[1] != 0 || test_fp_ctl(ufpc[0])) + return -EINVAL; + target->thread.fpu.fpc = ufpc[0]; + } + + if (rc == 0 && count > 0) + rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, + fprs, offsetof(s390_fp_regs, fprs), -1); + if (rc) + return rc; + + if (MACHINE_HAS_VX) + convert_fp_to_vx(target->thread.fpu.vxrs, fprs); + else + memcpy(target->thread.fpu.fprs, &fprs, sizeof(fprs)); + + return rc; +} + +static int s390_last_break_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + return membuf_store(&to, target->thread.last_break); +} + +static int s390_last_break_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + return 0; +} + +static int s390_tdb_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + struct pt_regs *regs = task_pt_regs(target); + size_t size; + + if (!(regs->int_code & 0x200)) + return -ENODATA; + size = sizeof(target->thread.trap_tdb.data); + return membuf_write(&to, target->thread.trap_tdb.data, size); +} + +static int s390_tdb_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + return 0; +} + +static int s390_vxrs_low_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + __u64 vxrs[__NUM_VXRS_LOW]; + int i; + + if (!MACHINE_HAS_VX) + return -ENODEV; + if (target == current) + save_fpu_regs(); + for (i = 0; i < __NUM_VXRS_LOW; i++) + vxrs[i] = target->thread.fpu.vxrs[i].low; + return membuf_write(&to, vxrs, sizeof(vxrs)); +} + +static int s390_vxrs_low_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + __u64 vxrs[__NUM_VXRS_LOW]; + int i, rc; + + if (!MACHINE_HAS_VX) + return -ENODEV; + if (target == current) + save_fpu_regs(); + + for (i = 0; i < __NUM_VXRS_LOW; i++) + vxrs[i] = target->thread.fpu.vxrs[i].low; + + rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, vxrs, 0, -1); + if (rc == 0) + for (i = 0; i < __NUM_VXRS_LOW; i++) + target->thread.fpu.vxrs[i].low = vxrs[i]; + + return rc; +} + +static int s390_vxrs_high_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + if (!MACHINE_HAS_VX) + return -ENODEV; + if (target == current) + save_fpu_regs(); + return membuf_write(&to, target->thread.fpu.vxrs + __NUM_VXRS_LOW, + __NUM_VXRS_HIGH * sizeof(__vector128)); +} + +static int s390_vxrs_high_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + int rc; + + if (!MACHINE_HAS_VX) + return -ENODEV; + if (target == current) + save_fpu_regs(); + + rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, + target->thread.fpu.vxrs + __NUM_VXRS_LOW, 0, -1); + return rc; +} + +static int s390_system_call_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + return membuf_store(&to, target->thread.system_call); +} + +static int s390_system_call_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + unsigned int *data = &target->thread.system_call; + return user_regset_copyin(&pos, &count, &kbuf, &ubuf, + data, 0, sizeof(unsigned int)); +} + +static int s390_gs_cb_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + struct gs_cb *data = target->thread.gs_cb; + + if (!MACHINE_HAS_GS) + return -ENODEV; + if (!data) + return -ENODATA; + if (target == current) + save_gs_cb(data); + return membuf_write(&to, data, sizeof(struct gs_cb)); +} + +static int s390_gs_cb_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + struct gs_cb gs_cb = { }, *data = NULL; + int rc; + + if (!MACHINE_HAS_GS) + return -ENODEV; + if (!target->thread.gs_cb) { + data = kzalloc(sizeof(*data), GFP_KERNEL); + if (!data) + return -ENOMEM; + } + if (!target->thread.gs_cb) + gs_cb.gsd = 25; + else if (target == current) + save_gs_cb(&gs_cb); + else + gs_cb = *target->thread.gs_cb; + rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, + &gs_cb, 0, sizeof(gs_cb)); + if (rc) { + kfree(data); + return -EFAULT; + } + preempt_disable(); + if (!target->thread.gs_cb) + target->thread.gs_cb = data; + *target->thread.gs_cb = gs_cb; + if (target == current) { + __ctl_set_bit(2, 4); + restore_gs_cb(target->thread.gs_cb); + } + preempt_enable(); + return rc; +} + +static int s390_gs_bc_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + struct gs_cb *data = target->thread.gs_bc_cb; + + if (!MACHINE_HAS_GS) + return -ENODEV; + if (!data) + return -ENODATA; + return membuf_write(&to, data, sizeof(struct gs_cb)); +} + +static int s390_gs_bc_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + struct gs_cb *data = target->thread.gs_bc_cb; + + if (!MACHINE_HAS_GS) + return -ENODEV; + if (!data) { + data = kzalloc(sizeof(*data), GFP_KERNEL); + if (!data) + return -ENOMEM; + target->thread.gs_bc_cb = data; + } + return user_regset_copyin(&pos, &count, &kbuf, &ubuf, + data, 0, sizeof(struct gs_cb)); +} + +static bool is_ri_cb_valid(struct runtime_instr_cb *cb) +{ + return (cb->rca & 0x1f) == 0 && + (cb->roa & 0xfff) == 0 && + (cb->rla & 0xfff) == 0xfff && + cb->s == 1 && + cb->k == 1 && + cb->h == 0 && + cb->reserved1 == 0 && + cb->ps == 1 && + cb->qs == 0 && + cb->pc == 1 && + cb->qc == 0 && + cb->reserved2 == 0 && + cb->reserved3 == 0 && + cb->reserved4 == 0 && + cb->reserved5 == 0 && + cb->reserved6 == 0 && + cb->reserved7 == 0 && + cb->reserved8 == 0 && + cb->rla >= cb->roa && + cb->rca >= cb->roa && + cb->rca <= cb->rla+1 && + cb->m < 3; +} + +static int s390_runtime_instr_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + struct runtime_instr_cb *data = target->thread.ri_cb; + + if (!test_facility(64)) + return -ENODEV; + if (!data) + return -ENODATA; + + return membuf_write(&to, data, sizeof(struct runtime_instr_cb)); +} + +static int s390_runtime_instr_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + struct runtime_instr_cb ri_cb = { }, *data = NULL; + int rc; + + if (!test_facility(64)) + return -ENODEV; + + if (!target->thread.ri_cb) { + data = kzalloc(sizeof(*data), GFP_KERNEL); + if (!data) + return -ENOMEM; + } + + if (target->thread.ri_cb) { + if (target == current) + store_runtime_instr_cb(&ri_cb); + else + ri_cb = *target->thread.ri_cb; + } + + rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, + &ri_cb, 0, sizeof(struct runtime_instr_cb)); + if (rc) { + kfree(data); + return -EFAULT; + } + + if (!is_ri_cb_valid(&ri_cb)) { + kfree(data); + return -EINVAL; + } + /* + * Override access key in any case, since user space should + * not be able to set it, nor should it care about it. + */ + ri_cb.key = PAGE_DEFAULT_KEY >> 4; + preempt_disable(); + if (!target->thread.ri_cb) + target->thread.ri_cb = data; + *target->thread.ri_cb = ri_cb; + if (target == current) + load_runtime_instr_cb(target->thread.ri_cb); + preempt_enable(); + + return 0; +} + +static const struct user_regset s390_regsets[] = { + { + .core_note_type = NT_PRSTATUS, + .n = sizeof(s390_regs) / sizeof(long), + .size = sizeof(long), + .align = sizeof(long), + .regset_get = s390_regs_get, + .set = s390_regs_set, + }, + { + .core_note_type = NT_PRFPREG, + .n = sizeof(s390_fp_regs) / sizeof(long), + .size = sizeof(long), + .align = sizeof(long), + .regset_get = s390_fpregs_get, + .set = s390_fpregs_set, + }, + { + .core_note_type = NT_S390_SYSTEM_CALL, + .n = 1, + .size = sizeof(unsigned int), + .align = sizeof(unsigned int), + .regset_get = s390_system_call_get, + .set = s390_system_call_set, + }, + { + .core_note_type = NT_S390_LAST_BREAK, + .n = 1, + .size = sizeof(long), + .align = sizeof(long), + .regset_get = s390_last_break_get, + .set = s390_last_break_set, + }, + { + .core_note_type = NT_S390_TDB, + .n = 1, + .size = 256, + .align = 1, + .regset_get = s390_tdb_get, + .set = s390_tdb_set, + }, + { + .core_note_type = NT_S390_VXRS_LOW, + .n = __NUM_VXRS_LOW, + .size = sizeof(__u64), + .align = sizeof(__u64), + .regset_get = s390_vxrs_low_get, + .set = s390_vxrs_low_set, + }, + { + .core_note_type = NT_S390_VXRS_HIGH, + .n = __NUM_VXRS_HIGH, + .size = sizeof(__vector128), + .align = sizeof(__vector128), + .regset_get = s390_vxrs_high_get, + .set = s390_vxrs_high_set, + }, + { + .core_note_type = NT_S390_GS_CB, + .n = sizeof(struct gs_cb) / sizeof(__u64), + .size = sizeof(__u64), + .align = sizeof(__u64), + .regset_get = s390_gs_cb_get, + .set = s390_gs_cb_set, + }, + { + .core_note_type = NT_S390_GS_BC, + .n = sizeof(struct gs_cb) / sizeof(__u64), + .size = sizeof(__u64), + .align = sizeof(__u64), + .regset_get = s390_gs_bc_get, + .set = s390_gs_bc_set, + }, + { + .core_note_type = NT_S390_RI_CB, + .n = sizeof(struct runtime_instr_cb) / sizeof(__u64), + .size = sizeof(__u64), + .align = sizeof(__u64), + .regset_get = s390_runtime_instr_get, + .set = s390_runtime_instr_set, + }, +}; + +static const struct user_regset_view user_s390_view = { + .name = "s390x", + .e_machine = EM_S390, + .regsets = s390_regsets, + .n = ARRAY_SIZE(s390_regsets) +}; + +#ifdef CONFIG_COMPAT +static int s390_compat_regs_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + unsigned n; + + if (target == current) + save_access_regs(target->thread.acrs); + + for (n = 0; n < sizeof(s390_compat_regs); n += sizeof(compat_ulong_t)) + membuf_store(&to, __peek_user_compat(target, n)); + return 0; +} + +static int s390_compat_regs_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + int rc = 0; + + if (target == current) + save_access_regs(target->thread.acrs); + + if (kbuf) { + const compat_ulong_t *k = kbuf; + while (count > 0 && !rc) { + rc = __poke_user_compat(target, pos, *k++); + count -= sizeof(*k); + pos += sizeof(*k); + } + } else { + const compat_ulong_t __user *u = ubuf; + while (count > 0 && !rc) { + compat_ulong_t word; + rc = __get_user(word, u++); + if (rc) + break; + rc = __poke_user_compat(target, pos, word); + count -= sizeof(*u); + pos += sizeof(*u); + } + } + + if (rc == 0 && target == current) + restore_access_regs(target->thread.acrs); + + return rc; +} + +static int s390_compat_regs_high_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + compat_ulong_t *gprs_high; + int i; + + gprs_high = (compat_ulong_t *)task_pt_regs(target)->gprs; + for (i = 0; i < NUM_GPRS; i++, gprs_high += 2) + membuf_store(&to, *gprs_high); + return 0; +} + +static int s390_compat_regs_high_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + compat_ulong_t *gprs_high; + int rc = 0; + + gprs_high = (compat_ulong_t *) + &task_pt_regs(target)->gprs[pos / sizeof(compat_ulong_t)]; + if (kbuf) { + const compat_ulong_t *k = kbuf; + while (count > 0) { + *gprs_high = *k++; + *gprs_high += 2; + count -= sizeof(*k); + } + } else { + const compat_ulong_t __user *u = ubuf; + while (count > 0 && !rc) { + unsigned long word; + rc = __get_user(word, u++); + if (rc) + break; + *gprs_high = word; + *gprs_high += 2; + count -= sizeof(*u); + } + } + + return rc; +} + +static int s390_compat_last_break_get(struct task_struct *target, + const struct user_regset *regset, + struct membuf to) +{ + compat_ulong_t last_break = target->thread.last_break; + + return membuf_store(&to, (unsigned long)last_break); +} + +static int s390_compat_last_break_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + return 0; +} + +static const struct user_regset s390_compat_regsets[] = { + { + .core_note_type = NT_PRSTATUS, + .n = sizeof(s390_compat_regs) / sizeof(compat_long_t), + .size = sizeof(compat_long_t), + .align = sizeof(compat_long_t), + .regset_get = s390_compat_regs_get, + .set = s390_compat_regs_set, + }, + { + .core_note_type = NT_PRFPREG, + .n = sizeof(s390_fp_regs) / sizeof(compat_long_t), + .size = sizeof(compat_long_t), + .align = sizeof(compat_long_t), + .regset_get = s390_fpregs_get, + .set = s390_fpregs_set, + }, + { + .core_note_type = NT_S390_SYSTEM_CALL, + .n = 1, + .size = sizeof(compat_uint_t), + .align = sizeof(compat_uint_t), + .regset_get = s390_system_call_get, + .set = s390_system_call_set, + }, + { + .core_note_type = NT_S390_LAST_BREAK, + .n = 1, + .size = sizeof(long), + .align = sizeof(long), + .regset_get = s390_compat_last_break_get, + .set = s390_compat_last_break_set, + }, + { + .core_note_type = NT_S390_TDB, + .n = 1, + .size = 256, + .align = 1, + .regset_get = s390_tdb_get, + .set = s390_tdb_set, + }, + { + .core_note_type = NT_S390_VXRS_LOW, + .n = __NUM_VXRS_LOW, + .size = sizeof(__u64), + .align = sizeof(__u64), + .regset_get = s390_vxrs_low_get, + .set = s390_vxrs_low_set, + }, + { + .core_note_type = NT_S390_VXRS_HIGH, + .n = __NUM_VXRS_HIGH, + .size = sizeof(__vector128), + .align = sizeof(__vector128), + .regset_get = s390_vxrs_high_get, + .set = s390_vxrs_high_set, + }, + { + .core_note_type = NT_S390_HIGH_GPRS, + .n = sizeof(s390_compat_regs_high) / sizeof(compat_long_t), + .size = sizeof(compat_long_t), + .align = sizeof(compat_long_t), + .regset_get = s390_compat_regs_high_get, + .set = s390_compat_regs_high_set, + }, + { + .core_note_type = NT_S390_GS_CB, + .n = sizeof(struct gs_cb) / sizeof(__u64), + .size = sizeof(__u64), + .align = sizeof(__u64), + .regset_get = s390_gs_cb_get, + .set = s390_gs_cb_set, + }, + { + .core_note_type = NT_S390_GS_BC, + .n = sizeof(struct gs_cb) / sizeof(__u64), + .size = sizeof(__u64), + .align = sizeof(__u64), + .regset_get = s390_gs_bc_get, + .set = s390_gs_bc_set, + }, + { + .core_note_type = NT_S390_RI_CB, + .n = sizeof(struct runtime_instr_cb) / sizeof(__u64), + .size = sizeof(__u64), + .align = sizeof(__u64), + .regset_get = s390_runtime_instr_get, + .set = s390_runtime_instr_set, + }, +}; + +static const struct user_regset_view user_s390_compat_view = { + .name = "s390", + .e_machine = EM_S390, + .regsets = s390_compat_regsets, + .n = ARRAY_SIZE(s390_compat_regsets) +}; +#endif + +const struct user_regset_view *task_user_regset_view(struct task_struct *task) +{ +#ifdef CONFIG_COMPAT + if (test_tsk_thread_flag(task, TIF_31BIT)) + return &user_s390_compat_view; +#endif + return &user_s390_view; +} + +static const char *gpr_names[NUM_GPRS] = { + "r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7", + "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15", +}; + +unsigned long regs_get_register(struct pt_regs *regs, unsigned int offset) +{ + if (offset >= NUM_GPRS) + return 0; + return regs->gprs[offset]; +} + +int regs_query_register_offset(const char *name) +{ + unsigned long offset; + + if (!name || *name != 'r') + return -EINVAL; + if (kstrtoul(name + 1, 10, &offset)) + return -EINVAL; + if (offset >= NUM_GPRS) + return -EINVAL; + return offset; +} + +const char *regs_query_register_name(unsigned int offset) +{ + if (offset >= NUM_GPRS) + return NULL; + return gpr_names[offset]; +} + +static int regs_within_kernel_stack(struct pt_regs *regs, unsigned long addr) +{ + unsigned long ksp = kernel_stack_pointer(regs); + + return (addr & ~(THREAD_SIZE - 1)) == (ksp & ~(THREAD_SIZE - 1)); +} + +/** + * regs_get_kernel_stack_nth() - get Nth entry of the stack + * @regs:pt_regs which contains kernel stack pointer. + * @n:stack entry number. + * + * regs_get_kernel_stack_nth() returns @n th entry of the kernel stack which + * is specifined by @regs. If the @n th entry is NOT in the kernel stack, + * this returns 0. + */ +unsigned long regs_get_kernel_stack_nth(struct pt_regs *regs, unsigned int n) +{ + unsigned long addr; + + addr = kernel_stack_pointer(regs) + n * sizeof(long); + if (!regs_within_kernel_stack(regs, addr)) + return 0; + return *(unsigned long *)addr; +} diff --git a/arch/s390/kernel/reipl.S b/arch/s390/kernel/reipl.S new file mode 100644 index 0000000000..88087a32eb --- /dev/null +++ b/arch/s390/kernel/reipl.S @@ -0,0 +1,81 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright IBM Corp 2000, 2011 + * Author(s): Holger Smolinski <Holger.Smolinski@de.ibm.com>, + * Denis Joseph Barrow, + */ + +#include <linux/linkage.h> +#include <asm/asm-offsets.h> +#include <asm/nospec-insn.h> +#include <asm/sigp.h> + + GEN_BR_THUNK %r9 + +# +# Issue "store status" for the current CPU to its prefix page +# and call passed function afterwards +# +# r2 = Function to be called after store status +# r3 = Parameter for function +# +SYM_CODE_START(store_status) + /* Save register one and load save area base */ + stg %r1,__LC_SAVE_AREA_RESTART + /* General purpose registers */ + lghi %r1,__LC_GPREGS_SAVE_AREA + stmg %r0,%r15,0(%r1) + mvc 8(8,%r1),__LC_SAVE_AREA_RESTART + /* Control registers */ + lghi %r1,__LC_CREGS_SAVE_AREA + stctg %c0,%c15,0(%r1) + /* Access registers */ + lghi %r1,__LC_AREGS_SAVE_AREA + stam %a0,%a15,0(%r1) + /* Floating point registers */ + lghi %r1,__LC_FPREGS_SAVE_AREA + std %f0, 0x00(%r1) + std %f1, 0x08(%r1) + std %f2, 0x10(%r1) + std %f3, 0x18(%r1) + std %f4, 0x20(%r1) + std %f5, 0x28(%r1) + std %f6, 0x30(%r1) + std %f7, 0x38(%r1) + std %f8, 0x40(%r1) + std %f9, 0x48(%r1) + std %f10,0x50(%r1) + std %f11,0x58(%r1) + std %f12,0x60(%r1) + std %f13,0x68(%r1) + std %f14,0x70(%r1) + std %f15,0x78(%r1) + /* Floating point control register */ + lghi %r1,__LC_FP_CREG_SAVE_AREA + stfpc 0(%r1) + /* CPU timer */ + lghi %r1,__LC_CPU_TIMER_SAVE_AREA + stpt 0(%r1) + /* Store prefix register */ + lghi %r1,__LC_PREFIX_SAVE_AREA + stpx 0(%r1) + /* Clock comparator - seven bytes */ + lghi %r1,__LC_CLOCK_COMP_SAVE_AREA + larl %r4,clkcmp + stckc 0(%r4) + mvc 1(7,%r1),1(%r4) + /* Program status word */ + lghi %r1,__LC_PSW_SAVE_AREA + epsw %r4,%r5 + st %r4,0(%r1) + st %r5,4(%r1) + stg %r2,8(%r1) + lgr %r9,%r2 + lgr %r2,%r3 + BR_EX %r9 +SYM_CODE_END(store_status) + + .section .bss + .balign 8 +SYM_DATA_LOCAL(clkcmp, .quad 0x0000000000000000) + .previous diff --git a/arch/s390/kernel/relocate_kernel.S b/arch/s390/kernel/relocate_kernel.S new file mode 100644 index 0000000000..0ae297c82a --- /dev/null +++ b/arch/s390/kernel/relocate_kernel.S @@ -0,0 +1,76 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright IBM Corp. 2005 + * + * Author(s): Rolf Adelsberger + * + */ + +#include <linux/linkage.h> +#include <asm/page.h> +#include <asm/sigp.h> + +/* + * moves the new kernel to its destination... + * %r2 = pointer to first kimage_entry_t + * %r3 = start address - where to jump to after the job is done... + * %r4 = subcode + * + * %r5 will be used as temp. storage + * %r6 holds the destination address + * %r7 = PAGE_SIZE + * %r8 holds the source address + * %r9 = PAGE_SIZE + * + * 0xf000 is a page_mask + */ + + .text +SYM_CODE_START(relocate_kernel) + basr %r13,0 # base address +.base: + lghi %r7,PAGE_SIZE # load PAGE_SIZE in r7 + lghi %r9,PAGE_SIZE # load PAGE_SIZE in r9 + lg %r5,0(%r2) # read another word for indirection page + aghi %r2,8 # increment pointer + tml %r5,0x1 # is it a destination page? + je .indir_check # NO, goto "indir_check" + lgr %r6,%r5 # r6 = r5 + nill %r6,0xf000 # mask it out and... + j .base # ...next iteration +.indir_check: + tml %r5,0x2 # is it a indirection page? + je .done_test # NO, goto "done_test" + nill %r5,0xf000 # YES, mask out, + lgr %r2,%r5 # move it into the right register, + j .base # and read next... +.done_test: + tml %r5,0x4 # is it the done indicator? + je .source_test # NO! Well, then it should be the source indicator... + j .done # ok, lets finish it here... +.source_test: + tml %r5,0x8 # it should be a source indicator... + je .base # NO, ignore it... + lgr %r8,%r5 # r8 = r5 + nill %r8,0xf000 # masking +0: mvcle %r6,%r8,0x0 # copy PAGE_SIZE bytes from r8 to r6 - pad with 0 + jo 0b + j .base +.done: + lgr %r0,%r4 # subcode + cghi %r3,0 + je .diag + la %r4,load_psw-.base(%r13) # load psw-address into the register + o %r3,4(%r4) # or load address into psw + st %r3,4(%r4) + mvc 0(8,%r0),0(%r4) # copy psw to absolute address 0 +.diag: + diag %r0,%r0,0x308 +SYM_CODE_END(relocate_kernel) + + .balign 8 +SYM_DATA_START_LOCAL(load_psw) + .long 0x00080000,0x80000000 +SYM_DATA_END_LABEL(load_psw, SYM_L_LOCAL, relocate_kernel_end) + .balign 8 +SYM_DATA(relocate_kernel_len, .quad relocate_kernel_end - relocate_kernel) diff --git a/arch/s390/kernel/rethook.c b/arch/s390/kernel/rethook.c new file mode 100644 index 0000000000..af10e6bdd3 --- /dev/null +++ b/arch/s390/kernel/rethook.c @@ -0,0 +1,34 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +#include <linux/rethook.h> +#include <linux/kprobes.h> +#include "rethook.h" + +void arch_rethook_prepare(struct rethook_node *rh, struct pt_regs *regs, bool mcount) +{ + rh->ret_addr = regs->gprs[14]; + rh->frame = regs->gprs[15]; + + /* Replace the return addr with trampoline addr */ + regs->gprs[14] = (unsigned long)&arch_rethook_trampoline; +} +NOKPROBE_SYMBOL(arch_rethook_prepare); + +void arch_rethook_fixup_return(struct pt_regs *regs, + unsigned long correct_ret_addr) +{ + /* Replace fake return address with real one. */ + regs->gprs[14] = correct_ret_addr; +} +NOKPROBE_SYMBOL(arch_rethook_fixup_return); + +/* + * Called from arch_rethook_trampoline + */ +unsigned long arch_rethook_trampoline_callback(struct pt_regs *regs) +{ + return rethook_trampoline_handler(regs, regs->gprs[15]); +} +NOKPROBE_SYMBOL(arch_rethook_trampoline_callback); + +/* assembler function that handles the rethook must not be probed itself */ +NOKPROBE_SYMBOL(arch_rethook_trampoline); diff --git a/arch/s390/kernel/rethook.h b/arch/s390/kernel/rethook.h new file mode 100644 index 0000000000..32f069eed3 --- /dev/null +++ b/arch/s390/kernel/rethook.h @@ -0,0 +1,7 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +#ifndef __S390_RETHOOK_H +#define __S390_RETHOOK_H + +unsigned long arch_rethook_trampoline_callback(struct pt_regs *regs); + +#endif diff --git a/arch/s390/kernel/runtime_instr.c b/arch/s390/kernel/runtime_instr.c new file mode 100644 index 0000000000..1788a5454b --- /dev/null +++ b/arch/s390/kernel/runtime_instr.c @@ -0,0 +1,102 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2012 + * Author(s): Jan Glauber <jang@linux.vnet.ibm.com> + */ + +#include <linux/kernel.h> +#include <linux/syscalls.h> +#include <linux/signal.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/init.h> +#include <linux/errno.h> +#include <linux/kernel_stat.h> +#include <linux/sched/task_stack.h> + +#include <asm/runtime_instr.h> +#include <asm/cpu_mf.h> +#include <asm/irq.h> + +#include "entry.h" + +/* empty control block to disable RI by loading it */ +struct runtime_instr_cb runtime_instr_empty_cb; + +void runtime_instr_release(struct task_struct *tsk) +{ + kfree(tsk->thread.ri_cb); +} + +static void disable_runtime_instr(void) +{ + struct task_struct *task = current; + struct pt_regs *regs; + + if (!task->thread.ri_cb) + return; + regs = task_pt_regs(task); + preempt_disable(); + load_runtime_instr_cb(&runtime_instr_empty_cb); + kfree(task->thread.ri_cb); + task->thread.ri_cb = NULL; + preempt_enable(); + + /* + * Make sure the RI bit is deleted from the PSW. If the user did not + * switch off RI before the system call the process will get a + * specification exception otherwise. + */ + regs->psw.mask &= ~PSW_MASK_RI; +} + +static void init_runtime_instr_cb(struct runtime_instr_cb *cb) +{ + cb->rla = 0xfff; + cb->s = 1; + cb->k = 1; + cb->ps = 1; + cb->pc = 1; + cb->key = PAGE_DEFAULT_KEY >> 4; + cb->v = 1; +} + +/* + * The signum argument is unused. In older kernels it was used to + * specify a real-time signal. For backwards compatibility user space + * should pass a valid real-time signal number (the signum argument + * was checked in older kernels). + */ +SYSCALL_DEFINE2(s390_runtime_instr, int, command, int, signum) +{ + struct runtime_instr_cb *cb; + + if (!test_facility(64)) + return -EOPNOTSUPP; + + if (command == S390_RUNTIME_INSTR_STOP) { + disable_runtime_instr(); + return 0; + } + + if (command != S390_RUNTIME_INSTR_START) + return -EINVAL; + + if (!current->thread.ri_cb) { + cb = kzalloc(sizeof(*cb), GFP_KERNEL); + if (!cb) + return -ENOMEM; + } else { + cb = current->thread.ri_cb; + memset(cb, 0, sizeof(*cb)); + } + + init_runtime_instr_cb(cb); + + /* now load the control block to make it available */ + preempt_disable(); + current->thread.ri_cb = cb; + load_runtime_instr_cb(cb); + preempt_enable(); + return 0; +} diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c new file mode 100644 index 0000000000..de6ad0fb23 --- /dev/null +++ b/arch/s390/kernel/setup.c @@ -0,0 +1,1012 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * S390 version + * Copyright IBM Corp. 1999, 2012 + * Author(s): Hartmut Penner (hp@de.ibm.com), + * Martin Schwidefsky (schwidefsky@de.ibm.com) + * + * Derived from "arch/i386/kernel/setup.c" + * Copyright (C) 1995, Linus Torvalds + */ + +/* + * This file handles the architecture-dependent parts of initialization + */ + +#define KMSG_COMPONENT "setup" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/errno.h> +#include <linux/export.h> +#include <linux/sched.h> +#include <linux/sched/task.h> +#include <linux/cpu.h> +#include <linux/kernel.h> +#include <linux/memblock.h> +#include <linux/mm.h> +#include <linux/stddef.h> +#include <linux/unistd.h> +#include <linux/ptrace.h> +#include <linux/random.h> +#include <linux/user.h> +#include <linux/tty.h> +#include <linux/ioport.h> +#include <linux/delay.h> +#include <linux/init.h> +#include <linux/initrd.h> +#include <linux/root_dev.h> +#include <linux/console.h> +#include <linux/kernel_stat.h> +#include <linux/dma-map-ops.h> +#include <linux/device.h> +#include <linux/notifier.h> +#include <linux/pfn.h> +#include <linux/ctype.h> +#include <linux/reboot.h> +#include <linux/topology.h> +#include <linux/kexec.h> +#include <linux/crash_dump.h> +#include <linux/memory.h> +#include <linux/compat.h> +#include <linux/start_kernel.h> +#include <linux/hugetlb.h> +#include <linux/kmemleak.h> + +#include <asm/archrandom.h> +#include <asm/boot_data.h> +#include <asm/ipl.h> +#include <asm/facility.h> +#include <asm/smp.h> +#include <asm/mmu_context.h> +#include <asm/cpcmd.h> +#include <asm/abs_lowcore.h> +#include <asm/nmi.h> +#include <asm/irq.h> +#include <asm/page.h> +#include <asm/ptrace.h> +#include <asm/sections.h> +#include <asm/ebcdic.h> +#include <asm/diag.h> +#include <asm/os_info.h> +#include <asm/sclp.h> +#include <asm/stacktrace.h> +#include <asm/sysinfo.h> +#include <asm/numa.h> +#include <asm/alternative.h> +#include <asm/nospec-branch.h> +#include <asm/physmem_info.h> +#include <asm/maccess.h> +#include <asm/uv.h> +#include <asm/asm-offsets.h> +#include "entry.h" + +/* + * Machine setup.. + */ +unsigned int console_mode = 0; +EXPORT_SYMBOL(console_mode); + +unsigned int console_devno = -1; +EXPORT_SYMBOL(console_devno); + +unsigned int console_irq = -1; +EXPORT_SYMBOL(console_irq); + +/* + * Some code and data needs to stay below 2 GB, even when the kernel would be + * relocated above 2 GB, because it has to use 31 bit addresses. + * Such code and data is part of the .amode31 section. + */ +char __amode31_ref *__samode31 = _samode31; +char __amode31_ref *__eamode31 = _eamode31; +char __amode31_ref *__stext_amode31 = _stext_amode31; +char __amode31_ref *__etext_amode31 = _etext_amode31; +struct exception_table_entry __amode31_ref *__start_amode31_ex_table = _start_amode31_ex_table; +struct exception_table_entry __amode31_ref *__stop_amode31_ex_table = _stop_amode31_ex_table; + +/* + * Control registers CR2, CR5 and CR15 are initialized with addresses + * of tables that must be placed below 2G which is handled by the AMODE31 + * sections. + * Because the AMODE31 sections are relocated below 2G at startup, + * the content of control registers CR2, CR5 and CR15 must be updated + * with new addresses after the relocation. The initial initialization of + * control registers occurs in head64.S and then gets updated again after AMODE31 + * relocation. We must access the relevant AMODE31 tables indirectly via + * pointers placed in the .amode31.refs linker section. Those pointers get + * updated automatically during AMODE31 relocation and always contain a valid + * address within AMODE31 sections. + */ + +static __amode31_data u32 __ctl_duct_amode31[16] __aligned(64); + +static __amode31_data u64 __ctl_aste_amode31[8] __aligned(64) = { + [1] = 0xffffffffffffffff +}; + +static __amode31_data u32 __ctl_duald_amode31[32] __aligned(128) = { + 0x80000000, 0, 0, 0, + 0x80000000, 0, 0, 0, + 0x80000000, 0, 0, 0, + 0x80000000, 0, 0, 0, + 0x80000000, 0, 0, 0, + 0x80000000, 0, 0, 0, + 0x80000000, 0, 0, 0, + 0x80000000, 0, 0, 0 +}; + +static __amode31_data u32 __ctl_linkage_stack_amode31[8] __aligned(64) = { + 0, 0, 0x89000000, 0, + 0, 0, 0x8a000000, 0 +}; + +static u64 __amode31_ref *__ctl_aste = __ctl_aste_amode31; +static u32 __amode31_ref *__ctl_duald = __ctl_duald_amode31; +static u32 __amode31_ref *__ctl_linkage_stack = __ctl_linkage_stack_amode31; +static u32 __amode31_ref *__ctl_duct = __ctl_duct_amode31; + +unsigned long __bootdata_preserved(max_mappable); +unsigned long __bootdata(ident_map_size); +struct physmem_info __bootdata(physmem_info); + +unsigned long __bootdata_preserved(__kaslr_offset); +int __bootdata_preserved(__kaslr_enabled); +unsigned int __bootdata_preserved(zlib_dfltcc_support); +EXPORT_SYMBOL(zlib_dfltcc_support); +u64 __bootdata_preserved(stfle_fac_list[16]); +EXPORT_SYMBOL(stfle_fac_list); +u64 __bootdata_preserved(alt_stfle_fac_list[16]); +struct oldmem_data __bootdata_preserved(oldmem_data); + +unsigned long VMALLOC_START; +EXPORT_SYMBOL(VMALLOC_START); + +unsigned long VMALLOC_END; +EXPORT_SYMBOL(VMALLOC_END); + +struct page *vmemmap; +EXPORT_SYMBOL(vmemmap); +unsigned long vmemmap_size; + +unsigned long MODULES_VADDR; +unsigned long MODULES_END; + +/* An array with a pointer to the lowcore of every CPU. */ +struct lowcore *lowcore_ptr[NR_CPUS]; +EXPORT_SYMBOL(lowcore_ptr); + +DEFINE_STATIC_KEY_FALSE(cpu_has_bear); + +/* + * The Write Back bit position in the physaddr is given by the SLPC PCI. + * Leaving the mask zero always uses write through which is safe + */ +unsigned long mio_wb_bit_mask __ro_after_init; + +/* + * This is set up by the setup-routine at boot-time + * for S390 need to find out, what we have to setup + * using address 0x10400 ... + */ + +#include <asm/setup.h> + +/* + * condev= and conmode= setup parameter. + */ + +static int __init condev_setup(char *str) +{ + int vdev; + + vdev = simple_strtoul(str, &str, 0); + if (vdev >= 0 && vdev < 65536) { + console_devno = vdev; + console_irq = -1; + } + return 1; +} + +__setup("condev=", condev_setup); + +static void __init set_preferred_console(void) +{ + if (CONSOLE_IS_3215 || CONSOLE_IS_SCLP) + add_preferred_console("ttyS", 0, NULL); + else if (CONSOLE_IS_3270) + add_preferred_console("tty3270", 0, NULL); + else if (CONSOLE_IS_VT220) + add_preferred_console("ttysclp", 0, NULL); + else if (CONSOLE_IS_HVC) + add_preferred_console("hvc", 0, NULL); +} + +static int __init conmode_setup(char *str) +{ +#if defined(CONFIG_SCLP_CONSOLE) || defined(CONFIG_SCLP_VT220_CONSOLE) + if (!strcmp(str, "hwc") || !strcmp(str, "sclp")) + SET_CONSOLE_SCLP; +#endif +#if defined(CONFIG_TN3215_CONSOLE) + if (!strcmp(str, "3215")) + SET_CONSOLE_3215; +#endif +#if defined(CONFIG_TN3270_CONSOLE) + if (!strcmp(str, "3270")) + SET_CONSOLE_3270; +#endif + set_preferred_console(); + return 1; +} + +__setup("conmode=", conmode_setup); + +static void __init conmode_default(void) +{ + char query_buffer[1024]; + char *ptr; + + if (MACHINE_IS_VM) { + cpcmd("QUERY CONSOLE", query_buffer, 1024, NULL); + console_devno = simple_strtoul(query_buffer + 5, NULL, 16); + ptr = strstr(query_buffer, "SUBCHANNEL ="); + console_irq = simple_strtoul(ptr + 13, NULL, 16); + cpcmd("QUERY TERM", query_buffer, 1024, NULL); + ptr = strstr(query_buffer, "CONMODE"); + /* + * Set the conmode to 3215 so that the device recognition + * will set the cu_type of the console to 3215. If the + * conmode is 3270 and we don't set it back then both + * 3215 and the 3270 driver will try to access the console + * device (3215 as console and 3270 as normal tty). + */ + cpcmd("TERM CONMODE 3215", NULL, 0, NULL); + if (ptr == NULL) { +#if defined(CONFIG_SCLP_CONSOLE) || defined(CONFIG_SCLP_VT220_CONSOLE) + SET_CONSOLE_SCLP; +#endif + return; + } + if (str_has_prefix(ptr + 8, "3270")) { +#if defined(CONFIG_TN3270_CONSOLE) + SET_CONSOLE_3270; +#elif defined(CONFIG_TN3215_CONSOLE) + SET_CONSOLE_3215; +#elif defined(CONFIG_SCLP_CONSOLE) || defined(CONFIG_SCLP_VT220_CONSOLE) + SET_CONSOLE_SCLP; +#endif + } else if (str_has_prefix(ptr + 8, "3215")) { +#if defined(CONFIG_TN3215_CONSOLE) + SET_CONSOLE_3215; +#elif defined(CONFIG_TN3270_CONSOLE) + SET_CONSOLE_3270; +#elif defined(CONFIG_SCLP_CONSOLE) || defined(CONFIG_SCLP_VT220_CONSOLE) + SET_CONSOLE_SCLP; +#endif + } + } else if (MACHINE_IS_KVM) { + if (sclp.has_vt220 && IS_ENABLED(CONFIG_SCLP_VT220_CONSOLE)) + SET_CONSOLE_VT220; + else if (sclp.has_linemode && IS_ENABLED(CONFIG_SCLP_CONSOLE)) + SET_CONSOLE_SCLP; + else + SET_CONSOLE_HVC; + } else { +#if defined(CONFIG_SCLP_CONSOLE) || defined(CONFIG_SCLP_VT220_CONSOLE) + SET_CONSOLE_SCLP; +#endif + } +} + +#ifdef CONFIG_CRASH_DUMP +static void __init setup_zfcpdump(void) +{ + if (!is_ipl_type_dump()) + return; + if (oldmem_data.start) + return; + strcat(boot_command_line, " cio_ignore=all,!ipldev,!condev"); + console_loglevel = 2; +} +#else +static inline void setup_zfcpdump(void) {} +#endif /* CONFIG_CRASH_DUMP */ + + /* + * Reboot, halt and power_off stubs. They just call _machine_restart, + * _machine_halt or _machine_power_off. + */ + +void machine_restart(char *command) +{ + if ((!in_interrupt() && !in_atomic()) || oops_in_progress) + /* + * Only unblank the console if we are called in enabled + * context or a bust_spinlocks cleared the way for us. + */ + console_unblank(); + _machine_restart(command); +} + +void machine_halt(void) +{ + if (!in_interrupt() || oops_in_progress) + /* + * Only unblank the console if we are called in enabled + * context or a bust_spinlocks cleared the way for us. + */ + console_unblank(); + _machine_halt(); +} + +void machine_power_off(void) +{ + if (!in_interrupt() || oops_in_progress) + /* + * Only unblank the console if we are called in enabled + * context or a bust_spinlocks cleared the way for us. + */ + console_unblank(); + _machine_power_off(); +} + +/* + * Dummy power off function. + */ +void (*pm_power_off)(void) = machine_power_off; +EXPORT_SYMBOL_GPL(pm_power_off); + +void *restart_stack; + +unsigned long stack_alloc(void) +{ +#ifdef CONFIG_VMAP_STACK + void *ret; + + ret = __vmalloc_node(THREAD_SIZE, THREAD_SIZE, THREADINFO_GFP, + NUMA_NO_NODE, __builtin_return_address(0)); + kmemleak_not_leak(ret); + return (unsigned long)ret; +#else + return __get_free_pages(GFP_KERNEL, THREAD_SIZE_ORDER); +#endif +} + +void stack_free(unsigned long stack) +{ +#ifdef CONFIG_VMAP_STACK + vfree((void *) stack); +#else + free_pages(stack, THREAD_SIZE_ORDER); +#endif +} + +void __init __noreturn arch_call_rest_init(void) +{ + smp_reinit_ipl_cpu(); + rest_init(); +} + +static unsigned long __init stack_alloc_early(void) +{ + unsigned long stack; + + stack = (unsigned long)memblock_alloc(THREAD_SIZE, THREAD_SIZE); + if (!stack) { + panic("%s: Failed to allocate %lu bytes align=0x%lx\n", + __func__, THREAD_SIZE, THREAD_SIZE); + } + return stack; +} + +static void __init setup_lowcore(void) +{ + struct lowcore *lc, *abs_lc; + + /* + * Setup lowcore for boot cpu + */ + BUILD_BUG_ON(sizeof(struct lowcore) != LC_PAGES * PAGE_SIZE); + lc = memblock_alloc_low(sizeof(*lc), sizeof(*lc)); + if (!lc) + panic("%s: Failed to allocate %zu bytes align=%zx\n", + __func__, sizeof(*lc), sizeof(*lc)); + + lc->restart_psw.mask = PSW_KERNEL_BITS & ~PSW_MASK_DAT; + lc->restart_psw.addr = __pa(restart_int_handler); + lc->external_new_psw.mask = PSW_KERNEL_BITS | PSW_MASK_MCHECK; + lc->external_new_psw.addr = (unsigned long) ext_int_handler; + lc->svc_new_psw.mask = PSW_KERNEL_BITS | PSW_MASK_MCHECK; + lc->svc_new_psw.addr = (unsigned long) system_call; + lc->program_new_psw.mask = PSW_KERNEL_BITS | PSW_MASK_MCHECK; + lc->program_new_psw.addr = (unsigned long) pgm_check_handler; + lc->mcck_new_psw.mask = PSW_KERNEL_BITS; + lc->mcck_new_psw.addr = (unsigned long) mcck_int_handler; + lc->io_new_psw.mask = PSW_KERNEL_BITS | PSW_MASK_MCHECK; + lc->io_new_psw.addr = (unsigned long) io_int_handler; + lc->clock_comparator = clock_comparator_max; + lc->current_task = (unsigned long)&init_task; + lc->lpp = LPP_MAGIC; + lc->machine_flags = S390_lowcore.machine_flags; + lc->preempt_count = S390_lowcore.preempt_count; + nmi_alloc_mcesa_early(&lc->mcesad); + lc->sys_enter_timer = S390_lowcore.sys_enter_timer; + lc->exit_timer = S390_lowcore.exit_timer; + lc->user_timer = S390_lowcore.user_timer; + lc->system_timer = S390_lowcore.system_timer; + lc->steal_timer = S390_lowcore.steal_timer; + lc->last_update_timer = S390_lowcore.last_update_timer; + lc->last_update_clock = S390_lowcore.last_update_clock; + /* + * Allocate the global restart stack which is the same for + * all CPUs in case *one* of them does a PSW restart. + */ + restart_stack = (void *)(stack_alloc_early() + STACK_INIT_OFFSET); + lc->mcck_stack = stack_alloc_early() + STACK_INIT_OFFSET; + lc->async_stack = stack_alloc_early() + STACK_INIT_OFFSET; + lc->nodat_stack = stack_alloc_early() + STACK_INIT_OFFSET; + lc->kernel_stack = S390_lowcore.kernel_stack; + /* + * Set up PSW restart to call ipl.c:do_restart(). Copy the relevant + * restart data to the absolute zero lowcore. This is necessary if + * PSW restart is done on an offline CPU that has lowcore zero. + */ + lc->restart_stack = (unsigned long) restart_stack; + lc->restart_fn = (unsigned long) do_restart; + lc->restart_data = 0; + lc->restart_source = -1U; + __ctl_store(lc->cregs_save_area, 0, 15); + lc->spinlock_lockval = arch_spin_lockval(0); + lc->spinlock_index = 0; + arch_spin_lock_setup(0); + lc->return_lpswe = gen_lpswe(__LC_RETURN_PSW); + lc->return_mcck_lpswe = gen_lpswe(__LC_RETURN_MCCK_PSW); + lc->preempt_count = PREEMPT_DISABLED; + lc->kernel_asce = S390_lowcore.kernel_asce; + lc->user_asce = S390_lowcore.user_asce; + + abs_lc = get_abs_lowcore(); + abs_lc->restart_stack = lc->restart_stack; + abs_lc->restart_fn = lc->restart_fn; + abs_lc->restart_data = lc->restart_data; + abs_lc->restart_source = lc->restart_source; + abs_lc->restart_psw = lc->restart_psw; + abs_lc->restart_flags = RESTART_FLAG_CTLREGS; + memcpy(abs_lc->cregs_save_area, lc->cregs_save_area, sizeof(abs_lc->cregs_save_area)); + abs_lc->program_new_psw = lc->program_new_psw; + abs_lc->mcesad = lc->mcesad; + put_abs_lowcore(abs_lc); + + set_prefix(__pa(lc)); + lowcore_ptr[0] = lc; + if (abs_lowcore_map(0, lowcore_ptr[0], false)) + panic("Couldn't setup absolute lowcore"); +} + +static struct resource code_resource = { + .name = "Kernel code", + .flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM, +}; + +static struct resource data_resource = { + .name = "Kernel data", + .flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM, +}; + +static struct resource bss_resource = { + .name = "Kernel bss", + .flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM, +}; + +static struct resource __initdata *standard_resources[] = { + &code_resource, + &data_resource, + &bss_resource, +}; + +static void __init setup_resources(void) +{ + struct resource *res, *std_res, *sub_res; + phys_addr_t start, end; + int j; + u64 i; + + code_resource.start = (unsigned long) _text; + code_resource.end = (unsigned long) _etext - 1; + data_resource.start = (unsigned long) _etext; + data_resource.end = (unsigned long) _edata - 1; + bss_resource.start = (unsigned long) __bss_start; + bss_resource.end = (unsigned long) __bss_stop - 1; + + for_each_mem_range(i, &start, &end) { + res = memblock_alloc(sizeof(*res), 8); + if (!res) + panic("%s: Failed to allocate %zu bytes align=0x%x\n", + __func__, sizeof(*res), 8); + res->flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM; + + res->name = "System RAM"; + res->start = start; + /* + * In memblock, end points to the first byte after the + * range while in resources, end points to the last byte in + * the range. + */ + res->end = end - 1; + request_resource(&iomem_resource, res); + + for (j = 0; j < ARRAY_SIZE(standard_resources); j++) { + std_res = standard_resources[j]; + if (std_res->start < res->start || + std_res->start > res->end) + continue; + if (std_res->end > res->end) { + sub_res = memblock_alloc(sizeof(*sub_res), 8); + if (!sub_res) + panic("%s: Failed to allocate %zu bytes align=0x%x\n", + __func__, sizeof(*sub_res), 8); + *sub_res = *std_res; + sub_res->end = res->end; + std_res->start = res->end + 1; + request_resource(res, sub_res); + } else { + request_resource(res, std_res); + } + } + } +#ifdef CONFIG_CRASH_DUMP + /* + * Re-add removed crash kernel memory as reserved memory. This makes + * sure it will be mapped with the identity mapping and struct pages + * will be created, so it can be resized later on. + * However add it later since the crash kernel resource should not be + * part of the System RAM resource. + */ + if (crashk_res.end) { + memblock_add_node(crashk_res.start, resource_size(&crashk_res), + 0, MEMBLOCK_NONE); + memblock_reserve(crashk_res.start, resource_size(&crashk_res)); + insert_resource(&iomem_resource, &crashk_res); + } +#endif +} + +static void __init setup_memory_end(void) +{ + max_pfn = max_low_pfn = PFN_DOWN(ident_map_size); + pr_notice("The maximum memory size is %luMB\n", ident_map_size >> 20); +} + +#ifdef CONFIG_CRASH_DUMP + +/* + * When kdump is enabled, we have to ensure that no memory from the area + * [0 - crashkernel memory size] is set offline - it will be exchanged with + * the crashkernel memory region when kdump is triggered. The crashkernel + * memory region can never get offlined (pages are unmovable). + */ +static int kdump_mem_notifier(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct memory_notify *arg = data; + + if (action != MEM_GOING_OFFLINE) + return NOTIFY_OK; + if (arg->start_pfn < PFN_DOWN(resource_size(&crashk_res))) + return NOTIFY_BAD; + return NOTIFY_OK; +} + +static struct notifier_block kdump_mem_nb = { + .notifier_call = kdump_mem_notifier, +}; + +#endif + +/* + * Reserve page tables created by decompressor + */ +static void __init reserve_pgtables(void) +{ + unsigned long start, end; + struct reserved_range *range; + + for_each_physmem_reserved_type_range(RR_VMEM, range, &start, &end) + memblock_reserve(start, end - start); +} + +/* + * Reserve memory for kdump kernel to be loaded with kexec + */ +static void __init reserve_crashkernel(void) +{ +#ifdef CONFIG_CRASH_DUMP + unsigned long long crash_base, crash_size; + phys_addr_t low, high; + int rc; + + rc = parse_crashkernel(boot_command_line, ident_map_size, &crash_size, + &crash_base); + + crash_base = ALIGN(crash_base, KEXEC_CRASH_MEM_ALIGN); + crash_size = ALIGN(crash_size, KEXEC_CRASH_MEM_ALIGN); + if (rc || crash_size == 0) + return; + + if (memblock.memory.regions[0].size < crash_size) { + pr_info("crashkernel reservation failed: %s\n", + "first memory chunk must be at least crashkernel size"); + return; + } + + low = crash_base ?: oldmem_data.start; + high = low + crash_size; + if (low >= oldmem_data.start && high <= oldmem_data.start + oldmem_data.size) { + /* The crashkernel fits into OLDMEM, reuse OLDMEM */ + crash_base = low; + } else { + /* Find suitable area in free memory */ + low = max_t(unsigned long, crash_size, sclp.hsa_size); + high = crash_base ? crash_base + crash_size : ULONG_MAX; + + if (crash_base && crash_base < low) { + pr_info("crashkernel reservation failed: %s\n", + "crash_base too low"); + return; + } + low = crash_base ?: low; + crash_base = memblock_phys_alloc_range(crash_size, + KEXEC_CRASH_MEM_ALIGN, + low, high); + } + + if (!crash_base) { + pr_info("crashkernel reservation failed: %s\n", + "no suitable area found"); + return; + } + + if (register_memory_notifier(&kdump_mem_nb)) { + memblock_phys_free(crash_base, crash_size); + return; + } + + if (!oldmem_data.start && MACHINE_IS_VM) + diag10_range(PFN_DOWN(crash_base), PFN_DOWN(crash_size)); + crashk_res.start = crash_base; + crashk_res.end = crash_base + crash_size - 1; + memblock_remove(crash_base, crash_size); + pr_info("Reserving %lluMB of memory at %lluMB " + "for crashkernel (System RAM: %luMB)\n", + crash_size >> 20, crash_base >> 20, + (unsigned long)memblock.memory.total_size >> 20); + os_info_crashkernel_add(crash_base, crash_size); +#endif +} + +/* + * Reserve the initrd from being used by memblock + */ +static void __init reserve_initrd(void) +{ + unsigned long addr, size; + + if (!IS_ENABLED(CONFIG_BLK_DEV_INITRD) || !get_physmem_reserved(RR_INITRD, &addr, &size)) + return; + initrd_start = (unsigned long)__va(addr); + initrd_end = initrd_start + size; + memblock_reserve(addr, size); +} + +/* + * Reserve the memory area used to pass the certificate lists + */ +static void __init reserve_certificate_list(void) +{ + if (ipl_cert_list_addr) + memblock_reserve(ipl_cert_list_addr, ipl_cert_list_size); +} + +static void __init reserve_physmem_info(void) +{ + unsigned long addr, size; + + if (get_physmem_reserved(RR_MEM_DETECT_EXTENDED, &addr, &size)) + memblock_reserve(addr, size); +} + +static void __init free_physmem_info(void) +{ + unsigned long addr, size; + + if (get_physmem_reserved(RR_MEM_DETECT_EXTENDED, &addr, &size)) + memblock_phys_free(addr, size); +} + +static void __init memblock_add_physmem_info(void) +{ + unsigned long start, end; + int i; + + pr_debug("physmem info source: %s (%hhd)\n", + get_physmem_info_source(), physmem_info.info_source); + /* keep memblock lists close to the kernel */ + memblock_set_bottom_up(true); + for_each_physmem_usable_range(i, &start, &end) + memblock_add(start, end - start); + for_each_physmem_online_range(i, &start, &end) + memblock_physmem_add(start, end - start); + memblock_set_bottom_up(false); + memblock_set_node(0, ULONG_MAX, &memblock.memory, 0); +} + +/* + * Reserve memory used for lowcore/command line/kernel image. + */ +static void __init reserve_kernel(void) +{ + memblock_reserve(0, STARTUP_NORMAL_OFFSET); + memblock_reserve(OLDMEM_BASE, sizeof(unsigned long)); + memblock_reserve(OLDMEM_SIZE, sizeof(unsigned long)); + memblock_reserve(physmem_info.reserved[RR_AMODE31].start, __eamode31 - __samode31); + memblock_reserve(__pa(sclp_early_sccb), EXT_SCCB_READ_SCP); + memblock_reserve(__pa(_stext), _end - _stext); +} + +static void __init setup_memory(void) +{ + phys_addr_t start, end; + u64 i; + + /* + * Init storage key for present memory + */ + for_each_mem_range(i, &start, &end) + storage_key_init_range(start, end); + + psw_set_key(PAGE_DEFAULT_KEY); +} + +static void __init relocate_amode31_section(void) +{ + unsigned long amode31_size = __eamode31 - __samode31; + long amode31_offset, *ptr; + + amode31_offset = physmem_info.reserved[RR_AMODE31].start - (unsigned long)__samode31; + pr_info("Relocating AMODE31 section of size 0x%08lx\n", amode31_size); + + /* Move original AMODE31 section to the new one */ + memmove((void *)physmem_info.reserved[RR_AMODE31].start, __samode31, amode31_size); + /* Zero out the old AMODE31 section to catch invalid accesses within it */ + memset(__samode31, 0, amode31_size); + + /* Update all AMODE31 region references */ + for (ptr = _start_amode31_refs; ptr != _end_amode31_refs; ptr++) + *ptr += amode31_offset; +} + +/* This must be called after AMODE31 relocation */ +static void __init setup_cr(void) +{ + union ctlreg2 cr2; + union ctlreg5 cr5; + union ctlreg15 cr15; + + __ctl_duct[1] = (unsigned long)__ctl_aste; + __ctl_duct[2] = (unsigned long)__ctl_aste; + __ctl_duct[4] = (unsigned long)__ctl_duald; + + /* Update control registers CR2, CR5 and CR15 */ + __ctl_store(cr2.val, 2, 2); + __ctl_store(cr5.val, 5, 5); + __ctl_store(cr15.val, 15, 15); + cr2.ducto = (unsigned long)__ctl_duct >> 6; + cr5.pasteo = (unsigned long)__ctl_duct >> 6; + cr15.lsea = (unsigned long)__ctl_linkage_stack >> 3; + __ctl_load(cr2.val, 2, 2); + __ctl_load(cr5.val, 5, 5); + __ctl_load(cr15.val, 15, 15); +} + +/* + * Add system information as device randomness + */ +static void __init setup_randomness(void) +{ + struct sysinfo_3_2_2 *vmms; + + vmms = memblock_alloc(PAGE_SIZE, PAGE_SIZE); + if (!vmms) + panic("Failed to allocate memory for sysinfo structure\n"); + if (stsi(vmms, 3, 2, 2) == 0 && vmms->count) + add_device_randomness(&vmms->vm, sizeof(vmms->vm[0]) * vmms->count); + memblock_free(vmms, PAGE_SIZE); + + if (cpacf_query_func(CPACF_PRNO, CPACF_PRNO_TRNG)) + static_branch_enable(&s390_arch_random_available); +} + +/* + * Find the correct size for the task_struct. This depends on + * the size of the struct fpu at the end of the thread_struct + * which is embedded in the task_struct. + */ +static void __init setup_task_size(void) +{ + int task_size = sizeof(struct task_struct); + + if (!MACHINE_HAS_VX) { + task_size -= sizeof(__vector128) * __NUM_VXRS; + task_size += sizeof(freg_t) * __NUM_FPRS; + } + arch_task_struct_size = task_size; +} + +/* + * Issue diagnose 318 to set the control program name and + * version codes. + */ +static void __init setup_control_program_code(void) +{ + union diag318_info diag318_info = { + .cpnc = CPNC_LINUX, + .cpvc = 0, + }; + + if (!sclp.has_diag318) + return; + + diag_stat_inc(DIAG_STAT_X318); + asm volatile("diag %0,0,0x318\n" : : "d" (diag318_info.val)); +} + +/* + * Print the component list from the IPL report + */ +static void __init log_component_list(void) +{ + struct ipl_rb_component_entry *ptr, *end; + char *str; + + if (!early_ipl_comp_list_addr) + return; + if (ipl_block.hdr.flags & IPL_PL_FLAG_SIPL) + pr_info("Linux is running with Secure-IPL enabled\n"); + else + pr_info("Linux is running with Secure-IPL disabled\n"); + ptr = __va(early_ipl_comp_list_addr); + end = (void *) ptr + early_ipl_comp_list_size; + pr_info("The IPL report contains the following components:\n"); + while (ptr < end) { + if (ptr->flags & IPL_RB_COMPONENT_FLAG_SIGNED) { + if (ptr->flags & IPL_RB_COMPONENT_FLAG_VERIFIED) + str = "signed, verified"; + else + str = "signed, verification failed"; + } else { + str = "not signed"; + } + pr_info("%016llx - %016llx (%s)\n", + ptr->addr, ptr->addr + ptr->len, str); + ptr++; + } +} + +/* + * Setup function called from init/main.c just after the banner + * was printed. + */ + +void __init setup_arch(char **cmdline_p) +{ + /* + * print what head.S has found out about the machine + */ + if (MACHINE_IS_VM) + pr_info("Linux is running as a z/VM " + "guest operating system in 64-bit mode\n"); + else if (MACHINE_IS_KVM) + pr_info("Linux is running under KVM in 64-bit mode\n"); + else if (MACHINE_IS_LPAR) + pr_info("Linux is running natively in 64-bit mode\n"); + else + pr_info("Linux is running as a guest in 64-bit mode\n"); + + log_component_list(); + + /* Have one command line that is parsed and saved in /proc/cmdline */ + /* boot_command_line has been already set up in early.c */ + *cmdline_p = boot_command_line; + + ROOT_DEV = Root_RAM0; + + setup_initial_init_mm(_text, _etext, _edata, _end); + + if (IS_ENABLED(CONFIG_EXPOLINE_AUTO)) + nospec_auto_detect(); + + jump_label_init(); + parse_early_param(); +#ifdef CONFIG_CRASH_DUMP + /* Deactivate elfcorehdr= kernel parameter */ + elfcorehdr_addr = ELFCORE_ADDR_MAX; +#endif + + os_info_init(); + setup_ipl(); + setup_task_size(); + setup_control_program_code(); + + /* Do some memory reservations *before* memory is added to memblock */ + reserve_pgtables(); + reserve_kernel(); + reserve_initrd(); + reserve_certificate_list(); + reserve_physmem_info(); + memblock_set_current_limit(ident_map_size); + memblock_allow_resize(); + + /* Get information about *all* installed memory */ + memblock_add_physmem_info(); + + free_physmem_info(); + setup_memory_end(); + memblock_dump_all(); + setup_memory(); + + relocate_amode31_section(); + setup_cr(); + setup_uv(); + dma_contiguous_reserve(ident_map_size); + vmcp_cma_reserve(); + if (MACHINE_HAS_EDAT2) + hugetlb_cma_reserve(PUD_SHIFT - PAGE_SHIFT); + + reserve_crashkernel(); +#ifdef CONFIG_CRASH_DUMP + /* + * Be aware that smp_save_dump_secondary_cpus() triggers a system reset. + * Therefore CPU and device initialization should be done afterwards. + */ + smp_save_dump_secondary_cpus(); +#endif + + setup_resources(); + setup_lowcore(); + smp_fill_possible_mask(); + cpu_detect_mhz_feature(); + cpu_init(); + numa_setup(); + smp_detect_cpus(); + topology_init_early(); + + if (test_facility(193)) + static_branch_enable(&cpu_has_bear); + + /* + * Create kernel page tables. + */ + paging_init(); + + /* + * After paging_init created the kernel page table, the new PSWs + * in lowcore can now run with DAT enabled. + */ +#ifdef CONFIG_CRASH_DUMP + smp_save_dump_ipl_cpu(); +#endif + + /* Setup default console */ + conmode_default(); + set_preferred_console(); + + apply_alternative_instructions(); + if (IS_ENABLED(CONFIG_EXPOLINE)) + nospec_init_branches(); + + /* Setup zfcp/nvme dump support */ + setup_zfcpdump(); + + /* Add system specific data to the random pool */ + setup_randomness(); +} diff --git a/arch/s390/kernel/signal.c b/arch/s390/kernel/signal.c new file mode 100644 index 0000000000..d63557d386 --- /dev/null +++ b/arch/s390/kernel/signal.c @@ -0,0 +1,534 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 1999, 2006 + * Author(s): Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com) + * + * Based on Intel version + * + * Copyright (C) 1991, 1992 Linus Torvalds + * + * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson + */ + +#include <linux/sched.h> +#include <linux/sched/task_stack.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/kernel.h> +#include <linux/signal.h> +#include <linux/entry-common.h> +#include <linux/errno.h> +#include <linux/wait.h> +#include <linux/ptrace.h> +#include <linux/unistd.h> +#include <linux/stddef.h> +#include <linux/tty.h> +#include <linux/personality.h> +#include <linux/binfmts.h> +#include <linux/syscalls.h> +#include <linux/compat.h> +#include <asm/ucontext.h> +#include <linux/uaccess.h> +#include <asm/lowcore.h> +#include <asm/switch_to.h> +#include <asm/vdso.h> +#include "entry.h" + +/* + * Layout of an old-style signal-frame: + * ----------------------------------------- + * | save area (_SIGNAL_FRAMESIZE) | + * ----------------------------------------- + * | struct sigcontext | + * | oldmask | + * | _sigregs * | + * ----------------------------------------- + * | _sigregs with | + * | _s390_regs_common | + * | _s390_fp_regs | + * ----------------------------------------- + * | int signo | + * ----------------------------------------- + * | _sigregs_ext with | + * | gprs_high 64 byte (opt) | + * | vxrs_low 128 byte (opt) | + * | vxrs_high 256 byte (opt) | + * | reserved 128 byte (opt) | + * ----------------------------------------- + * | __u16 svc_insn | + * ----------------------------------------- + * The svc_insn entry with the sigreturn system call opcode does not + * have a fixed position and moves if gprs_high or vxrs exist. + * Future extensions will be added to _sigregs_ext. + */ +struct sigframe +{ + __u8 callee_used_stack[__SIGNAL_FRAMESIZE]; + struct sigcontext sc; + _sigregs sregs; + int signo; + _sigregs_ext sregs_ext; + __u16 svc_insn; /* Offset of svc_insn is NOT fixed! */ +}; + +/* + * Layout of an rt signal-frame: + * ----------------------------------------- + * | save area (_SIGNAL_FRAMESIZE) | + * ----------------------------------------- + * | svc __NR_rt_sigreturn 2 byte | + * ----------------------------------------- + * | struct siginfo | + * ----------------------------------------- + * | struct ucontext_extended with | + * | unsigned long uc_flags | + * | struct ucontext *uc_link | + * | stack_t uc_stack | + * | _sigregs uc_mcontext with | + * | _s390_regs_common | + * | _s390_fp_regs | + * | sigset_t uc_sigmask | + * | _sigregs_ext uc_mcontext_ext | + * | gprs_high 64 byte (opt) | + * | vxrs_low 128 byte (opt) | + * | vxrs_high 256 byte (opt)| + * | reserved 128 byte (opt) | + * ----------------------------------------- + * Future extensions will be added to _sigregs_ext. + */ +struct rt_sigframe +{ + __u8 callee_used_stack[__SIGNAL_FRAMESIZE]; + __u16 svc_insn; + struct siginfo info; + struct ucontext_extended uc; +}; + +/* Store registers needed to create the signal frame */ +static void store_sigregs(void) +{ + save_access_regs(current->thread.acrs); + save_fpu_regs(); +} + +/* Load registers after signal return */ +static void load_sigregs(void) +{ + restore_access_regs(current->thread.acrs); +} + +/* Returns non-zero on fault. */ +static int save_sigregs(struct pt_regs *regs, _sigregs __user *sregs) +{ + _sigregs user_sregs; + + /* Copy a 'clean' PSW mask to the user to avoid leaking + information about whether PER is currently on. */ + user_sregs.regs.psw.mask = PSW_USER_BITS | + (regs->psw.mask & (PSW_MASK_USER | PSW_MASK_RI)); + user_sregs.regs.psw.addr = regs->psw.addr; + memcpy(&user_sregs.regs.gprs, ®s->gprs, sizeof(sregs->regs.gprs)); + memcpy(&user_sregs.regs.acrs, current->thread.acrs, + sizeof(user_sregs.regs.acrs)); + fpregs_store(&user_sregs.fpregs, ¤t->thread.fpu); + if (__copy_to_user(sregs, &user_sregs, sizeof(_sigregs))) + return -EFAULT; + return 0; +} + +static int restore_sigregs(struct pt_regs *regs, _sigregs __user *sregs) +{ + _sigregs user_sregs; + + /* Always make any pending restarted system call return -EINTR */ + current->restart_block.fn = do_no_restart_syscall; + + if (__copy_from_user(&user_sregs, sregs, sizeof(user_sregs))) + return -EFAULT; + + if (!is_ri_task(current) && (user_sregs.regs.psw.mask & PSW_MASK_RI)) + return -EINVAL; + + /* Test the floating-point-control word. */ + if (test_fp_ctl(user_sregs.fpregs.fpc)) + return -EINVAL; + + /* Use regs->psw.mask instead of PSW_USER_BITS to preserve PER bit. */ + regs->psw.mask = (regs->psw.mask & ~(PSW_MASK_USER | PSW_MASK_RI)) | + (user_sregs.regs.psw.mask & (PSW_MASK_USER | PSW_MASK_RI)); + /* Check for invalid user address space control. */ + if ((regs->psw.mask & PSW_MASK_ASC) == PSW_ASC_HOME) + regs->psw.mask = PSW_ASC_PRIMARY | + (regs->psw.mask & ~PSW_MASK_ASC); + /* Check for invalid amode */ + if (regs->psw.mask & PSW_MASK_EA) + regs->psw.mask |= PSW_MASK_BA; + regs->psw.addr = user_sregs.regs.psw.addr; + memcpy(®s->gprs, &user_sregs.regs.gprs, sizeof(sregs->regs.gprs)); + memcpy(¤t->thread.acrs, &user_sregs.regs.acrs, + sizeof(current->thread.acrs)); + + fpregs_load(&user_sregs.fpregs, ¤t->thread.fpu); + + clear_pt_regs_flag(regs, PIF_SYSCALL); /* No longer in a system call */ + return 0; +} + +/* Returns non-zero on fault. */ +static int save_sigregs_ext(struct pt_regs *regs, + _sigregs_ext __user *sregs_ext) +{ + __u64 vxrs[__NUM_VXRS_LOW]; + int i; + + /* Save vector registers to signal stack */ + if (MACHINE_HAS_VX) { + for (i = 0; i < __NUM_VXRS_LOW; i++) + vxrs[i] = current->thread.fpu.vxrs[i].low; + if (__copy_to_user(&sregs_ext->vxrs_low, vxrs, + sizeof(sregs_ext->vxrs_low)) || + __copy_to_user(&sregs_ext->vxrs_high, + current->thread.fpu.vxrs + __NUM_VXRS_LOW, + sizeof(sregs_ext->vxrs_high))) + return -EFAULT; + } + return 0; +} + +static int restore_sigregs_ext(struct pt_regs *regs, + _sigregs_ext __user *sregs_ext) +{ + __u64 vxrs[__NUM_VXRS_LOW]; + int i; + + /* Restore vector registers from signal stack */ + if (MACHINE_HAS_VX) { + if (__copy_from_user(vxrs, &sregs_ext->vxrs_low, + sizeof(sregs_ext->vxrs_low)) || + __copy_from_user(current->thread.fpu.vxrs + __NUM_VXRS_LOW, + &sregs_ext->vxrs_high, + sizeof(sregs_ext->vxrs_high))) + return -EFAULT; + for (i = 0; i < __NUM_VXRS_LOW; i++) + current->thread.fpu.vxrs[i].low = vxrs[i]; + } + return 0; +} + +SYSCALL_DEFINE0(sigreturn) +{ + struct pt_regs *regs = task_pt_regs(current); + struct sigframe __user *frame = + (struct sigframe __user *) regs->gprs[15]; + sigset_t set; + + if (__copy_from_user(&set.sig, &frame->sc.oldmask, _SIGMASK_COPY_SIZE)) + goto badframe; + set_current_blocked(&set); + save_fpu_regs(); + if (restore_sigregs(regs, &frame->sregs)) + goto badframe; + if (restore_sigregs_ext(regs, &frame->sregs_ext)) + goto badframe; + load_sigregs(); + return regs->gprs[2]; +badframe: + force_sig(SIGSEGV); + return 0; +} + +SYSCALL_DEFINE0(rt_sigreturn) +{ + struct pt_regs *regs = task_pt_regs(current); + struct rt_sigframe __user *frame = + (struct rt_sigframe __user *)regs->gprs[15]; + sigset_t set; + + if (__copy_from_user(&set.sig, &frame->uc.uc_sigmask, sizeof(set))) + goto badframe; + set_current_blocked(&set); + if (restore_altstack(&frame->uc.uc_stack)) + goto badframe; + save_fpu_regs(); + if (restore_sigregs(regs, &frame->uc.uc_mcontext)) + goto badframe; + if (restore_sigregs_ext(regs, &frame->uc.uc_mcontext_ext)) + goto badframe; + load_sigregs(); + return regs->gprs[2]; +badframe: + force_sig(SIGSEGV); + return 0; +} + +/* + * Determine which stack to use.. + */ +static inline void __user * +get_sigframe(struct k_sigaction *ka, struct pt_regs * regs, size_t frame_size) +{ + unsigned long sp; + + /* Default to using normal stack */ + sp = regs->gprs[15]; + + /* Overflow on alternate signal stack gives SIGSEGV. */ + if (on_sig_stack(sp) && !on_sig_stack((sp - frame_size) & -8UL)) + return (void __user *) -1UL; + + /* This is the X/Open sanctioned signal stack switching. */ + if (ka->sa.sa_flags & SA_ONSTACK) { + if (! sas_ss_flags(sp)) + sp = current->sas_ss_sp + current->sas_ss_size; + } + + return (void __user *)((sp - frame_size) & -8ul); +} + +static int setup_frame(int sig, struct k_sigaction *ka, + sigset_t *set, struct pt_regs * regs) +{ + struct sigframe __user *frame; + struct sigcontext sc; + unsigned long restorer; + size_t frame_size; + + /* + * gprs_high are only present for a 31-bit task running on + * a 64-bit kernel (see compat_signal.c) but the space for + * gprs_high need to be allocated if vector registers are + * included in the signal frame on a 31-bit system. + */ + frame_size = sizeof(*frame) - sizeof(frame->sregs_ext); + if (MACHINE_HAS_VX) + frame_size += sizeof(frame->sregs_ext); + frame = get_sigframe(ka, regs, frame_size); + if (frame == (void __user *) -1UL) + return -EFAULT; + + /* Set up backchain. */ + if (__put_user(regs->gprs[15], (addr_t __user *) frame)) + return -EFAULT; + + /* Create struct sigcontext on the signal stack */ + memcpy(&sc.oldmask, &set->sig, _SIGMASK_COPY_SIZE); + sc.sregs = (_sigregs __user __force *) &frame->sregs; + if (__copy_to_user(&frame->sc, &sc, sizeof(frame->sc))) + return -EFAULT; + + /* Store registers needed to create the signal frame */ + store_sigregs(); + + /* Create _sigregs on the signal stack */ + if (save_sigregs(regs, &frame->sregs)) + return -EFAULT; + + /* Place signal number on stack to allow backtrace from handler. */ + if (__put_user(regs->gprs[2], (int __user *) &frame->signo)) + return -EFAULT; + + /* Create _sigregs_ext on the signal stack */ + if (save_sigregs_ext(regs, &frame->sregs_ext)) + return -EFAULT; + + /* Set up to return from userspace. If provided, use a stub + already in userspace. */ + if (ka->sa.sa_flags & SA_RESTORER) + restorer = (unsigned long) ka->sa.sa_restorer; + else + restorer = VDSO64_SYMBOL(current, sigreturn); + + /* Set up registers for signal handler */ + regs->gprs[14] = restorer; + regs->gprs[15] = (unsigned long) frame; + /* Force default amode and default user address space control. */ + regs->psw.mask = PSW_MASK_EA | PSW_MASK_BA | + (PSW_USER_BITS & PSW_MASK_ASC) | + (regs->psw.mask & ~PSW_MASK_ASC); + regs->psw.addr = (unsigned long) ka->sa.sa_handler; + + regs->gprs[2] = sig; + regs->gprs[3] = (unsigned long) &frame->sc; + + /* We forgot to include these in the sigcontext. + To avoid breaking binary compatibility, they are passed as args. */ + if (sig == SIGSEGV || sig == SIGBUS || sig == SIGILL || + sig == SIGTRAP || sig == SIGFPE) { + /* set extra registers only for synchronous signals */ + regs->gprs[4] = regs->int_code & 127; + regs->gprs[5] = regs->int_parm_long; + regs->gprs[6] = current->thread.last_break; + } + return 0; +} + +static int setup_rt_frame(struct ksignal *ksig, sigset_t *set, + struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + unsigned long uc_flags, restorer; + size_t frame_size; + + frame_size = sizeof(struct rt_sigframe) - sizeof(_sigregs_ext); + /* + * gprs_high are only present for a 31-bit task running on + * a 64-bit kernel (see compat_signal.c) but the space for + * gprs_high need to be allocated if vector registers are + * included in the signal frame on a 31-bit system. + */ + uc_flags = 0; + if (MACHINE_HAS_VX) { + frame_size += sizeof(_sigregs_ext); + uc_flags |= UC_VXRS; + } + frame = get_sigframe(&ksig->ka, regs, frame_size); + if (frame == (void __user *) -1UL) + return -EFAULT; + + /* Set up backchain. */ + if (__put_user(regs->gprs[15], (addr_t __user *) frame)) + return -EFAULT; + + /* Set up to return from userspace. If provided, use a stub + already in userspace. */ + if (ksig->ka.sa.sa_flags & SA_RESTORER) + restorer = (unsigned long) ksig->ka.sa.sa_restorer; + else + restorer = VDSO64_SYMBOL(current, rt_sigreturn); + + /* Create siginfo on the signal stack */ + if (copy_siginfo_to_user(&frame->info, &ksig->info)) + return -EFAULT; + + /* Store registers needed to create the signal frame */ + store_sigregs(); + + /* Create ucontext on the signal stack. */ + if (__put_user(uc_flags, &frame->uc.uc_flags) || + __put_user(NULL, &frame->uc.uc_link) || + __save_altstack(&frame->uc.uc_stack, regs->gprs[15]) || + save_sigregs(regs, &frame->uc.uc_mcontext) || + __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)) || + save_sigregs_ext(regs, &frame->uc.uc_mcontext_ext)) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->gprs[14] = restorer; + regs->gprs[15] = (unsigned long) frame; + /* Force default amode and default user address space control. */ + regs->psw.mask = PSW_MASK_EA | PSW_MASK_BA | + (PSW_USER_BITS & PSW_MASK_ASC) | + (regs->psw.mask & ~PSW_MASK_ASC); + regs->psw.addr = (unsigned long) ksig->ka.sa.sa_handler; + + regs->gprs[2] = ksig->sig; + regs->gprs[3] = (unsigned long) &frame->info; + regs->gprs[4] = (unsigned long) &frame->uc; + regs->gprs[5] = current->thread.last_break; + return 0; +} + +static void handle_signal(struct ksignal *ksig, sigset_t *oldset, + struct pt_regs *regs) +{ + int ret; + + /* Set up the stack frame */ + if (ksig->ka.sa.sa_flags & SA_SIGINFO) + ret = setup_rt_frame(ksig, oldset, regs); + else + ret = setup_frame(ksig->sig, &ksig->ka, oldset, regs); + + signal_setup_done(ret, ksig, test_thread_flag(TIF_SINGLE_STEP)); +} + +/* + * Note that 'init' is a special process: it doesn't get signals it doesn't + * want to handle. Thus you cannot kill init even with a SIGKILL even by + * mistake. + * + * Note that we go through the signals twice: once to check the signals that + * the kernel can handle, and then we build all the user-level signal handling + * stack-frames in one go after that. + */ + +void arch_do_signal_or_restart(struct pt_regs *regs) +{ + struct ksignal ksig; + sigset_t *oldset = sigmask_to_save(); + + /* + * Get signal to deliver. When running under ptrace, at this point + * the debugger may change all our registers, including the system + * call information. + */ + current->thread.system_call = + test_pt_regs_flag(regs, PIF_SYSCALL) ? regs->int_code : 0; + + if (get_signal(&ksig)) { + /* Whee! Actually deliver the signal. */ + if (current->thread.system_call) { + regs->int_code = current->thread.system_call; + /* Check for system call restarting. */ + switch (regs->gprs[2]) { + case -ERESTART_RESTARTBLOCK: + case -ERESTARTNOHAND: + regs->gprs[2] = -EINTR; + break; + case -ERESTARTSYS: + if (!(ksig.ka.sa.sa_flags & SA_RESTART)) { + regs->gprs[2] = -EINTR; + break; + } + fallthrough; + case -ERESTARTNOINTR: + regs->gprs[2] = regs->orig_gpr2; + regs->psw.addr = + __rewind_psw(regs->psw, + regs->int_code >> 16); + break; + } + } + /* No longer in a system call */ + clear_pt_regs_flag(regs, PIF_SYSCALL); + + rseq_signal_deliver(&ksig, regs); + if (is_compat_task()) + handle_signal32(&ksig, oldset, regs); + else + handle_signal(&ksig, oldset, regs); + return; + } + + /* No handlers present - check for system call restart */ + clear_pt_regs_flag(regs, PIF_SYSCALL); + if (current->thread.system_call) { + regs->int_code = current->thread.system_call; + switch (regs->gprs[2]) { + case -ERESTART_RESTARTBLOCK: + /* Restart with sys_restart_syscall */ + regs->gprs[2] = regs->orig_gpr2; + current->restart_block.arch_data = regs->psw.addr; + if (is_compat_task()) + regs->psw.addr = VDSO32_SYMBOL(current, restart_syscall); + else + regs->psw.addr = VDSO64_SYMBOL(current, restart_syscall); + if (test_thread_flag(TIF_SINGLE_STEP)) + clear_thread_flag(TIF_PER_TRAP); + break; + case -ERESTARTNOHAND: + case -ERESTARTSYS: + case -ERESTARTNOINTR: + regs->gprs[2] = regs->orig_gpr2; + regs->psw.addr = __rewind_psw(regs->psw, regs->int_code >> 16); + if (test_thread_flag(TIF_SINGLE_STEP)) + clear_thread_flag(TIF_PER_TRAP); + break; + } + } + + /* + * If there's no signal to deliver, we just put the saved sigmask back. + */ + restore_saved_sigmask(); +} diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c new file mode 100644 index 0000000000..a4edb7ea66 --- /dev/null +++ b/arch/s390/kernel/smp.c @@ -0,0 +1,1317 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * SMP related functions + * + * Copyright IBM Corp. 1999, 2012 + * Author(s): Denis Joseph Barrow, + * Martin Schwidefsky <schwidefsky@de.ibm.com>, + * + * based on other smp stuff by + * (c) 1995 Alan Cox, CymruNET Ltd <alan@cymru.net> + * (c) 1998 Ingo Molnar + * + * The code outside of smp.c uses logical cpu numbers, only smp.c does + * the translation of logical to physical cpu ids. All new code that + * operates on physical cpu numbers needs to go into smp.c. + */ + +#define KMSG_COMPONENT "cpu" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/workqueue.h> +#include <linux/memblock.h> +#include <linux/export.h> +#include <linux/init.h> +#include <linux/mm.h> +#include <linux/err.h> +#include <linux/spinlock.h> +#include <linux/kernel_stat.h> +#include <linux/delay.h> +#include <linux/interrupt.h> +#include <linux/irqflags.h> +#include <linux/irq_work.h> +#include <linux/cpu.h> +#include <linux/slab.h> +#include <linux/sched/hotplug.h> +#include <linux/sched/task_stack.h> +#include <linux/crash_dump.h> +#include <linux/kprobes.h> +#include <asm/asm-offsets.h> +#include <asm/pfault.h> +#include <asm/diag.h> +#include <asm/switch_to.h> +#include <asm/facility.h> +#include <asm/ipl.h> +#include <asm/setup.h> +#include <asm/irq.h> +#include <asm/tlbflush.h> +#include <asm/vtimer.h> +#include <asm/abs_lowcore.h> +#include <asm/sclp.h> +#include <asm/debug.h> +#include <asm/os_info.h> +#include <asm/sigp.h> +#include <asm/idle.h> +#include <asm/nmi.h> +#include <asm/stacktrace.h> +#include <asm/topology.h> +#include <asm/vdso.h> +#include <asm/maccess.h> +#include "entry.h" + +enum { + ec_schedule = 0, + ec_call_function_single, + ec_stop_cpu, + ec_mcck_pending, + ec_irq_work, +}; + +enum { + CPU_STATE_STANDBY, + CPU_STATE_CONFIGURED, +}; + +static DEFINE_PER_CPU(struct cpu *, cpu_device); + +struct pcpu { + unsigned long ec_mask; /* bit mask for ec_xxx functions */ + unsigned long ec_clk; /* sigp timestamp for ec_xxx */ + signed char state; /* physical cpu state */ + signed char polarization; /* physical polarization */ + u16 address; /* physical cpu address */ +}; + +static u8 boot_core_type; +static struct pcpu pcpu_devices[NR_CPUS]; + +unsigned int smp_cpu_mt_shift; +EXPORT_SYMBOL(smp_cpu_mt_shift); + +unsigned int smp_cpu_mtid; +EXPORT_SYMBOL(smp_cpu_mtid); + +#ifdef CONFIG_CRASH_DUMP +__vector128 __initdata boot_cpu_vector_save_area[__NUM_VXRS]; +#endif + +static unsigned int smp_max_threads __initdata = -1U; +cpumask_t cpu_setup_mask; + +static int __init early_nosmt(char *s) +{ + smp_max_threads = 1; + return 0; +} +early_param("nosmt", early_nosmt); + +static int __init early_smt(char *s) +{ + get_option(&s, &smp_max_threads); + return 0; +} +early_param("smt", early_smt); + +/* + * The smp_cpu_state_mutex must be held when changing the state or polarization + * member of a pcpu data structure within the pcpu_devices array. + */ +DEFINE_MUTEX(smp_cpu_state_mutex); + +/* + * Signal processor helper functions. + */ +static inline int __pcpu_sigp_relax(u16 addr, u8 order, unsigned long parm) +{ + int cc; + + while (1) { + cc = __pcpu_sigp(addr, order, parm, NULL); + if (cc != SIGP_CC_BUSY) + return cc; + cpu_relax(); + } +} + +static int pcpu_sigp_retry(struct pcpu *pcpu, u8 order, u32 parm) +{ + int cc, retry; + + for (retry = 0; ; retry++) { + cc = __pcpu_sigp(pcpu->address, order, parm, NULL); + if (cc != SIGP_CC_BUSY) + break; + if (retry >= 3) + udelay(10); + } + return cc; +} + +static inline int pcpu_stopped(struct pcpu *pcpu) +{ + u32 status; + + if (__pcpu_sigp(pcpu->address, SIGP_SENSE, + 0, &status) != SIGP_CC_STATUS_STORED) + return 0; + return !!(status & (SIGP_STATUS_CHECK_STOP|SIGP_STATUS_STOPPED)); +} + +static inline int pcpu_running(struct pcpu *pcpu) +{ + if (__pcpu_sigp(pcpu->address, SIGP_SENSE_RUNNING, + 0, NULL) != SIGP_CC_STATUS_STORED) + return 1; + /* Status stored condition code is equivalent to cpu not running. */ + return 0; +} + +/* + * Find struct pcpu by cpu address. + */ +static struct pcpu *pcpu_find_address(const struct cpumask *mask, u16 address) +{ + int cpu; + + for_each_cpu(cpu, mask) + if (pcpu_devices[cpu].address == address) + return pcpu_devices + cpu; + return NULL; +} + +static void pcpu_ec_call(struct pcpu *pcpu, int ec_bit) +{ + int order; + + if (test_and_set_bit(ec_bit, &pcpu->ec_mask)) + return; + order = pcpu_running(pcpu) ? SIGP_EXTERNAL_CALL : SIGP_EMERGENCY_SIGNAL; + pcpu->ec_clk = get_tod_clock_fast(); + pcpu_sigp_retry(pcpu, order, 0); +} + +static int pcpu_alloc_lowcore(struct pcpu *pcpu, int cpu) +{ + unsigned long async_stack, nodat_stack, mcck_stack; + struct lowcore *lc; + + lc = (struct lowcore *) __get_free_pages(GFP_KERNEL | GFP_DMA, LC_ORDER); + nodat_stack = __get_free_pages(GFP_KERNEL, THREAD_SIZE_ORDER); + async_stack = stack_alloc(); + mcck_stack = stack_alloc(); + if (!lc || !nodat_stack || !async_stack || !mcck_stack) + goto out; + memcpy(lc, &S390_lowcore, 512); + memset((char *) lc + 512, 0, sizeof(*lc) - 512); + lc->async_stack = async_stack + STACK_INIT_OFFSET; + lc->nodat_stack = nodat_stack + STACK_INIT_OFFSET; + lc->mcck_stack = mcck_stack + STACK_INIT_OFFSET; + lc->cpu_nr = cpu; + lc->spinlock_lockval = arch_spin_lockval(cpu); + lc->spinlock_index = 0; + lc->return_lpswe = gen_lpswe(__LC_RETURN_PSW); + lc->return_mcck_lpswe = gen_lpswe(__LC_RETURN_MCCK_PSW); + lc->preempt_count = PREEMPT_DISABLED; + if (nmi_alloc_mcesa(&lc->mcesad)) + goto out; + if (abs_lowcore_map(cpu, lc, true)) + goto out_mcesa; + lowcore_ptr[cpu] = lc; + pcpu_sigp_retry(pcpu, SIGP_SET_PREFIX, __pa(lc)); + return 0; + +out_mcesa: + nmi_free_mcesa(&lc->mcesad); +out: + stack_free(mcck_stack); + stack_free(async_stack); + free_pages(nodat_stack, THREAD_SIZE_ORDER); + free_pages((unsigned long) lc, LC_ORDER); + return -ENOMEM; +} + +static void pcpu_free_lowcore(struct pcpu *pcpu) +{ + unsigned long async_stack, nodat_stack, mcck_stack; + struct lowcore *lc; + int cpu; + + cpu = pcpu - pcpu_devices; + lc = lowcore_ptr[cpu]; + nodat_stack = lc->nodat_stack - STACK_INIT_OFFSET; + async_stack = lc->async_stack - STACK_INIT_OFFSET; + mcck_stack = lc->mcck_stack - STACK_INIT_OFFSET; + pcpu_sigp_retry(pcpu, SIGP_SET_PREFIX, 0); + lowcore_ptr[cpu] = NULL; + abs_lowcore_unmap(cpu); + nmi_free_mcesa(&lc->mcesad); + stack_free(async_stack); + stack_free(mcck_stack); + free_pages(nodat_stack, THREAD_SIZE_ORDER); + free_pages((unsigned long) lc, LC_ORDER); +} + +static void pcpu_prepare_secondary(struct pcpu *pcpu, int cpu) +{ + struct lowcore *lc, *abs_lc; + + lc = lowcore_ptr[cpu]; + cpumask_set_cpu(cpu, &init_mm.context.cpu_attach_mask); + cpumask_set_cpu(cpu, mm_cpumask(&init_mm)); + lc->cpu_nr = cpu; + lc->restart_flags = RESTART_FLAG_CTLREGS; + lc->spinlock_lockval = arch_spin_lockval(cpu); + lc->spinlock_index = 0; + lc->percpu_offset = __per_cpu_offset[cpu]; + lc->kernel_asce = S390_lowcore.kernel_asce; + lc->user_asce = s390_invalid_asce; + lc->machine_flags = S390_lowcore.machine_flags; + lc->user_timer = lc->system_timer = + lc->steal_timer = lc->avg_steal_timer = 0; + abs_lc = get_abs_lowcore(); + memcpy(lc->cregs_save_area, abs_lc->cregs_save_area, sizeof(lc->cregs_save_area)); + put_abs_lowcore(abs_lc); + lc->cregs_save_area[1] = lc->kernel_asce; + lc->cregs_save_area[7] = lc->user_asce; + save_access_regs((unsigned int *) lc->access_regs_save_area); + arch_spin_lock_setup(cpu); +} + +static void pcpu_attach_task(struct pcpu *pcpu, struct task_struct *tsk) +{ + struct lowcore *lc; + int cpu; + + cpu = pcpu - pcpu_devices; + lc = lowcore_ptr[cpu]; + lc->kernel_stack = (unsigned long)task_stack_page(tsk) + STACK_INIT_OFFSET; + lc->current_task = (unsigned long)tsk; + lc->lpp = LPP_MAGIC; + lc->current_pid = tsk->pid; + lc->user_timer = tsk->thread.user_timer; + lc->guest_timer = tsk->thread.guest_timer; + lc->system_timer = tsk->thread.system_timer; + lc->hardirq_timer = tsk->thread.hardirq_timer; + lc->softirq_timer = tsk->thread.softirq_timer; + lc->steal_timer = 0; +} + +static void pcpu_start_fn(struct pcpu *pcpu, void (*func)(void *), void *data) +{ + struct lowcore *lc; + int cpu; + + cpu = pcpu - pcpu_devices; + lc = lowcore_ptr[cpu]; + lc->restart_stack = lc->kernel_stack; + lc->restart_fn = (unsigned long) func; + lc->restart_data = (unsigned long) data; + lc->restart_source = -1U; + pcpu_sigp_retry(pcpu, SIGP_RESTART, 0); +} + +typedef void (pcpu_delegate_fn)(void *); + +/* + * Call function via PSW restart on pcpu and stop the current cpu. + */ +static void __pcpu_delegate(pcpu_delegate_fn *func, void *data) +{ + func(data); /* should not return */ +} + +static void pcpu_delegate(struct pcpu *pcpu, + pcpu_delegate_fn *func, + void *data, unsigned long stack) +{ + struct lowcore *lc, *abs_lc; + unsigned int source_cpu; + + lc = lowcore_ptr[pcpu - pcpu_devices]; + source_cpu = stap(); + + if (pcpu->address == source_cpu) { + call_on_stack(2, stack, void, __pcpu_delegate, + pcpu_delegate_fn *, func, void *, data); + } + /* Stop target cpu (if func returns this stops the current cpu). */ + pcpu_sigp_retry(pcpu, SIGP_STOP, 0); + pcpu_sigp_retry(pcpu, SIGP_CPU_RESET, 0); + /* Restart func on the target cpu and stop the current cpu. */ + if (lc) { + lc->restart_stack = stack; + lc->restart_fn = (unsigned long)func; + lc->restart_data = (unsigned long)data; + lc->restart_source = source_cpu; + } else { + abs_lc = get_abs_lowcore(); + abs_lc->restart_stack = stack; + abs_lc->restart_fn = (unsigned long)func; + abs_lc->restart_data = (unsigned long)data; + abs_lc->restart_source = source_cpu; + put_abs_lowcore(abs_lc); + } + asm volatile( + "0: sigp 0,%0,%2 # sigp restart to target cpu\n" + " brc 2,0b # busy, try again\n" + "1: sigp 0,%1,%3 # sigp stop to current cpu\n" + " brc 2,1b # busy, try again\n" + : : "d" (pcpu->address), "d" (source_cpu), + "K" (SIGP_RESTART), "K" (SIGP_STOP) + : "0", "1", "cc"); + for (;;) ; +} + +/* + * Enable additional logical cpus for multi-threading. + */ +static int pcpu_set_smt(unsigned int mtid) +{ + int cc; + + if (smp_cpu_mtid == mtid) + return 0; + cc = __pcpu_sigp(0, SIGP_SET_MULTI_THREADING, mtid, NULL); + if (cc == 0) { + smp_cpu_mtid = mtid; + smp_cpu_mt_shift = 0; + while (smp_cpu_mtid >= (1U << smp_cpu_mt_shift)) + smp_cpu_mt_shift++; + pcpu_devices[0].address = stap(); + } + return cc; +} + +/* + * Call function on an online CPU. + */ +void smp_call_online_cpu(void (*func)(void *), void *data) +{ + struct pcpu *pcpu; + + /* Use the current cpu if it is online. */ + pcpu = pcpu_find_address(cpu_online_mask, stap()); + if (!pcpu) + /* Use the first online cpu. */ + pcpu = pcpu_devices + cpumask_first(cpu_online_mask); + pcpu_delegate(pcpu, func, data, (unsigned long) restart_stack); +} + +/* + * Call function on the ipl CPU. + */ +void smp_call_ipl_cpu(void (*func)(void *), void *data) +{ + struct lowcore *lc = lowcore_ptr[0]; + + if (pcpu_devices[0].address == stap()) + lc = &S390_lowcore; + + pcpu_delegate(&pcpu_devices[0], func, data, + lc->nodat_stack); +} + +int smp_find_processor_id(u16 address) +{ + int cpu; + + for_each_present_cpu(cpu) + if (pcpu_devices[cpu].address == address) + return cpu; + return -1; +} + +void schedule_mcck_handler(void) +{ + pcpu_ec_call(pcpu_devices + smp_processor_id(), ec_mcck_pending); +} + +bool notrace arch_vcpu_is_preempted(int cpu) +{ + if (test_cpu_flag_of(CIF_ENABLED_WAIT, cpu)) + return false; + if (pcpu_running(pcpu_devices + cpu)) + return false; + return true; +} +EXPORT_SYMBOL(arch_vcpu_is_preempted); + +void notrace smp_yield_cpu(int cpu) +{ + if (!MACHINE_HAS_DIAG9C) + return; + diag_stat_inc_norecursion(DIAG_STAT_X09C); + asm volatile("diag %0,0,0x9c" + : : "d" (pcpu_devices[cpu].address)); +} +EXPORT_SYMBOL_GPL(smp_yield_cpu); + +/* + * Send cpus emergency shutdown signal. This gives the cpus the + * opportunity to complete outstanding interrupts. + */ +void notrace smp_emergency_stop(void) +{ + static arch_spinlock_t lock = __ARCH_SPIN_LOCK_UNLOCKED; + static cpumask_t cpumask; + u64 end; + int cpu; + + arch_spin_lock(&lock); + cpumask_copy(&cpumask, cpu_online_mask); + cpumask_clear_cpu(smp_processor_id(), &cpumask); + + end = get_tod_clock() + (1000000UL << 12); + for_each_cpu(cpu, &cpumask) { + struct pcpu *pcpu = pcpu_devices + cpu; + set_bit(ec_stop_cpu, &pcpu->ec_mask); + while (__pcpu_sigp(pcpu->address, SIGP_EMERGENCY_SIGNAL, + 0, NULL) == SIGP_CC_BUSY && + get_tod_clock() < end) + cpu_relax(); + } + while (get_tod_clock() < end) { + for_each_cpu(cpu, &cpumask) + if (pcpu_stopped(pcpu_devices + cpu)) + cpumask_clear_cpu(cpu, &cpumask); + if (cpumask_empty(&cpumask)) + break; + cpu_relax(); + } + arch_spin_unlock(&lock); +} +NOKPROBE_SYMBOL(smp_emergency_stop); + +/* + * Stop all cpus but the current one. + */ +void smp_send_stop(void) +{ + int cpu; + + /* Disable all interrupts/machine checks */ + __load_psw_mask(PSW_KERNEL_BITS); + trace_hardirqs_off(); + + debug_set_critical(); + + if (oops_in_progress) + smp_emergency_stop(); + + /* stop all processors */ + for_each_online_cpu(cpu) { + if (cpu == smp_processor_id()) + continue; + pcpu_sigp_retry(pcpu_devices + cpu, SIGP_STOP, 0); + while (!pcpu_stopped(pcpu_devices + cpu)) + cpu_relax(); + } +} + +/* + * This is the main routine where commands issued by other + * cpus are handled. + */ +static void smp_handle_ext_call(void) +{ + unsigned long bits; + + /* handle bit signal external calls */ + bits = xchg(&pcpu_devices[smp_processor_id()].ec_mask, 0); + if (test_bit(ec_stop_cpu, &bits)) + smp_stop_cpu(); + if (test_bit(ec_schedule, &bits)) + scheduler_ipi(); + if (test_bit(ec_call_function_single, &bits)) + generic_smp_call_function_single_interrupt(); + if (test_bit(ec_mcck_pending, &bits)) + s390_handle_mcck(); + if (test_bit(ec_irq_work, &bits)) + irq_work_run(); +} + +static void do_ext_call_interrupt(struct ext_code ext_code, + unsigned int param32, unsigned long param64) +{ + inc_irq_stat(ext_code.code == 0x1202 ? IRQEXT_EXC : IRQEXT_EMS); + smp_handle_ext_call(); +} + +void arch_send_call_function_ipi_mask(const struct cpumask *mask) +{ + int cpu; + + for_each_cpu(cpu, mask) + pcpu_ec_call(pcpu_devices + cpu, ec_call_function_single); +} + +void arch_send_call_function_single_ipi(int cpu) +{ + pcpu_ec_call(pcpu_devices + cpu, ec_call_function_single); +} + +/* + * this function sends a 'reschedule' IPI to another CPU. + * it goes straight through and wastes no time serializing + * anything. Worst case is that we lose a reschedule ... + */ +void arch_smp_send_reschedule(int cpu) +{ + pcpu_ec_call(pcpu_devices + cpu, ec_schedule); +} + +#ifdef CONFIG_IRQ_WORK +void arch_irq_work_raise(void) +{ + pcpu_ec_call(pcpu_devices + smp_processor_id(), ec_irq_work); +} +#endif + +/* + * parameter area for the set/clear control bit callbacks + */ +struct ec_creg_mask_parms { + unsigned long orval; + unsigned long andval; + int cr; +}; + +/* + * callback for setting/clearing control bits + */ +static void smp_ctl_bit_callback(void *info) +{ + struct ec_creg_mask_parms *pp = info; + unsigned long cregs[16]; + + __ctl_store(cregs, 0, 15); + cregs[pp->cr] = (cregs[pp->cr] & pp->andval) | pp->orval; + __ctl_load(cregs, 0, 15); +} + +static DEFINE_SPINLOCK(ctl_lock); + +void smp_ctl_set_clear_bit(int cr, int bit, bool set) +{ + struct ec_creg_mask_parms parms = { .cr = cr, }; + struct lowcore *abs_lc; + u64 ctlreg; + + if (set) { + parms.orval = 1UL << bit; + parms.andval = -1UL; + } else { + parms.orval = 0; + parms.andval = ~(1UL << bit); + } + spin_lock(&ctl_lock); + abs_lc = get_abs_lowcore(); + ctlreg = abs_lc->cregs_save_area[cr]; + ctlreg = (ctlreg & parms.andval) | parms.orval; + abs_lc->cregs_save_area[cr] = ctlreg; + put_abs_lowcore(abs_lc); + on_each_cpu(smp_ctl_bit_callback, &parms, 1); + spin_unlock(&ctl_lock); +} +EXPORT_SYMBOL(smp_ctl_set_clear_bit); + +#ifdef CONFIG_CRASH_DUMP + +int smp_store_status(int cpu) +{ + struct lowcore *lc; + struct pcpu *pcpu; + unsigned long pa; + + pcpu = pcpu_devices + cpu; + lc = lowcore_ptr[cpu]; + pa = __pa(&lc->floating_pt_save_area); + if (__pcpu_sigp_relax(pcpu->address, SIGP_STORE_STATUS_AT_ADDRESS, + pa) != SIGP_CC_ORDER_CODE_ACCEPTED) + return -EIO; + if (!MACHINE_HAS_VX && !MACHINE_HAS_GS) + return 0; + pa = lc->mcesad & MCESA_ORIGIN_MASK; + if (MACHINE_HAS_GS) + pa |= lc->mcesad & MCESA_LC_MASK; + if (__pcpu_sigp_relax(pcpu->address, SIGP_STORE_ADDITIONAL_STATUS, + pa) != SIGP_CC_ORDER_CODE_ACCEPTED) + return -EIO; + return 0; +} + +/* + * Collect CPU state of the previous, crashed system. + * There are four cases: + * 1) standard zfcp/nvme dump + * condition: OLDMEM_BASE == NULL && is_ipl_type_dump() == true + * The state for all CPUs except the boot CPU needs to be collected + * with sigp stop-and-store-status. The boot CPU state is located in + * the absolute lowcore of the memory stored in the HSA. The zcore code + * will copy the boot CPU state from the HSA. + * 2) stand-alone kdump for SCSI/NVMe (zfcp/nvme dump with swapped memory) + * condition: OLDMEM_BASE != NULL && is_ipl_type_dump() == true + * The state for all CPUs except the boot CPU needs to be collected + * with sigp stop-and-store-status. The firmware or the boot-loader + * stored the registers of the boot CPU in the absolute lowcore in the + * memory of the old system. + * 3) kdump and the old kernel did not store the CPU state, + * or stand-alone kdump for DASD + * condition: OLDMEM_BASE != NULL && !is_kdump_kernel() + * The state for all CPUs except the boot CPU needs to be collected + * with sigp stop-and-store-status. The kexec code or the boot-loader + * stored the registers of the boot CPU in the memory of the old system. + * 4) kdump and the old kernel stored the CPU state + * condition: OLDMEM_BASE != NULL && is_kdump_kernel() + * This case does not exist for s390 anymore, setup_arch explicitly + * deactivates the elfcorehdr= kernel parameter + */ +static bool dump_available(void) +{ + return oldmem_data.start || is_ipl_type_dump(); +} + +void __init smp_save_dump_ipl_cpu(void) +{ + struct save_area *sa; + void *regs; + + if (!dump_available()) + return; + sa = save_area_alloc(true); + regs = memblock_alloc(512, 8); + if (!sa || !regs) + panic("could not allocate memory for boot CPU save area\n"); + copy_oldmem_kernel(regs, __LC_FPREGS_SAVE_AREA, 512); + save_area_add_regs(sa, regs); + memblock_free(regs, 512); + if (MACHINE_HAS_VX) + save_area_add_vxrs(sa, boot_cpu_vector_save_area); +} + +void __init smp_save_dump_secondary_cpus(void) +{ + int addr, boot_cpu_addr, max_cpu_addr; + struct save_area *sa; + void *page; + + if (!dump_available()) + return; + /* Allocate a page as dumping area for the store status sigps */ + page = memblock_alloc_low(PAGE_SIZE, PAGE_SIZE); + if (!page) + panic("ERROR: Failed to allocate %lx bytes below %lx\n", + PAGE_SIZE, 1UL << 31); + + /* Set multi-threading state to the previous system. */ + pcpu_set_smt(sclp.mtid_prev); + boot_cpu_addr = stap(); + max_cpu_addr = SCLP_MAX_CORES << sclp.mtid_prev; + for (addr = 0; addr <= max_cpu_addr; addr++) { + if (addr == boot_cpu_addr) + continue; + if (__pcpu_sigp_relax(addr, SIGP_SENSE, 0) == + SIGP_CC_NOT_OPERATIONAL) + continue; + sa = save_area_alloc(false); + if (!sa) + panic("could not allocate memory for save area\n"); + __pcpu_sigp_relax(addr, SIGP_STORE_STATUS_AT_ADDRESS, __pa(page)); + save_area_add_regs(sa, page); + if (MACHINE_HAS_VX) { + __pcpu_sigp_relax(addr, SIGP_STORE_ADDITIONAL_STATUS, __pa(page)); + save_area_add_vxrs(sa, page); + } + } + memblock_free(page, PAGE_SIZE); + diag_amode31_ops.diag308_reset(); + pcpu_set_smt(0); +} +#endif /* CONFIG_CRASH_DUMP */ + +void smp_cpu_set_polarization(int cpu, int val) +{ + pcpu_devices[cpu].polarization = val; +} + +int smp_cpu_get_polarization(int cpu) +{ + return pcpu_devices[cpu].polarization; +} + +int smp_cpu_get_cpu_address(int cpu) +{ + return pcpu_devices[cpu].address; +} + +static void __ref smp_get_core_info(struct sclp_core_info *info, int early) +{ + static int use_sigp_detection; + int address; + + if (use_sigp_detection || sclp_get_core_info(info, early)) { + use_sigp_detection = 1; + for (address = 0; + address < (SCLP_MAX_CORES << smp_cpu_mt_shift); + address += (1U << smp_cpu_mt_shift)) { + if (__pcpu_sigp_relax(address, SIGP_SENSE, 0) == + SIGP_CC_NOT_OPERATIONAL) + continue; + info->core[info->configured].core_id = + address >> smp_cpu_mt_shift; + info->configured++; + } + info->combined = info->configured; + } +} + +static int smp_add_present_cpu(int cpu); + +static int smp_add_core(struct sclp_core_entry *core, cpumask_t *avail, + bool configured, bool early) +{ + struct pcpu *pcpu; + int cpu, nr, i; + u16 address; + + nr = 0; + if (sclp.has_core_type && core->type != boot_core_type) + return nr; + cpu = cpumask_first(avail); + address = core->core_id << smp_cpu_mt_shift; + for (i = 0; (i <= smp_cpu_mtid) && (cpu < nr_cpu_ids); i++) { + if (pcpu_find_address(cpu_present_mask, address + i)) + continue; + pcpu = pcpu_devices + cpu; + pcpu->address = address + i; + if (configured) + pcpu->state = CPU_STATE_CONFIGURED; + else + pcpu->state = CPU_STATE_STANDBY; + smp_cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); + set_cpu_present(cpu, true); + if (!early && smp_add_present_cpu(cpu) != 0) + set_cpu_present(cpu, false); + else + nr++; + cpumask_clear_cpu(cpu, avail); + cpu = cpumask_next(cpu, avail); + } + return nr; +} + +static int __smp_rescan_cpus(struct sclp_core_info *info, bool early) +{ + struct sclp_core_entry *core; + static cpumask_t avail; + bool configured; + u16 core_id; + int nr, i; + + cpus_read_lock(); + mutex_lock(&smp_cpu_state_mutex); + nr = 0; + cpumask_xor(&avail, cpu_possible_mask, cpu_present_mask); + /* + * Add IPL core first (which got logical CPU number 0) to make sure + * that all SMT threads get subsequent logical CPU numbers. + */ + if (early) { + core_id = pcpu_devices[0].address >> smp_cpu_mt_shift; + for (i = 0; i < info->configured; i++) { + core = &info->core[i]; + if (core->core_id == core_id) { + nr += smp_add_core(core, &avail, true, early); + break; + } + } + } + for (i = 0; i < info->combined; i++) { + configured = i < info->configured; + nr += smp_add_core(&info->core[i], &avail, configured, early); + } + mutex_unlock(&smp_cpu_state_mutex); + cpus_read_unlock(); + return nr; +} + +void __init smp_detect_cpus(void) +{ + unsigned int cpu, mtid, c_cpus, s_cpus; + struct sclp_core_info *info; + u16 address; + + /* Get CPU information */ + info = memblock_alloc(sizeof(*info), 8); + if (!info) + panic("%s: Failed to allocate %zu bytes align=0x%x\n", + __func__, sizeof(*info), 8); + smp_get_core_info(info, 1); + /* Find boot CPU type */ + if (sclp.has_core_type) { + address = stap(); + for (cpu = 0; cpu < info->combined; cpu++) + if (info->core[cpu].core_id == address) { + /* The boot cpu dictates the cpu type. */ + boot_core_type = info->core[cpu].type; + break; + } + if (cpu >= info->combined) + panic("Could not find boot CPU type"); + } + + /* Set multi-threading state for the current system */ + mtid = boot_core_type ? sclp.mtid : sclp.mtid_cp; + mtid = (mtid < smp_max_threads) ? mtid : smp_max_threads - 1; + pcpu_set_smt(mtid); + + /* Print number of CPUs */ + c_cpus = s_cpus = 0; + for (cpu = 0; cpu < info->combined; cpu++) { + if (sclp.has_core_type && + info->core[cpu].type != boot_core_type) + continue; + if (cpu < info->configured) + c_cpus += smp_cpu_mtid + 1; + else + s_cpus += smp_cpu_mtid + 1; + } + pr_info("%d configured CPUs, %d standby CPUs\n", c_cpus, s_cpus); + + /* Add CPUs present at boot */ + __smp_rescan_cpus(info, true); + memblock_free(info, sizeof(*info)); +} + +/* + * Activate a secondary processor. + */ +static void smp_start_secondary(void *cpuvoid) +{ + int cpu = raw_smp_processor_id(); + + S390_lowcore.last_update_clock = get_tod_clock(); + S390_lowcore.restart_stack = (unsigned long)restart_stack; + S390_lowcore.restart_fn = (unsigned long)do_restart; + S390_lowcore.restart_data = 0; + S390_lowcore.restart_source = -1U; + S390_lowcore.restart_flags = 0; + restore_access_regs(S390_lowcore.access_regs_save_area); + cpu_init(); + rcu_cpu_starting(cpu); + init_cpu_timer(); + vtime_init(); + vdso_getcpu_init(); + pfault_init(); + cpumask_set_cpu(cpu, &cpu_setup_mask); + update_cpu_masks(); + notify_cpu_starting(cpu); + if (topology_cpu_dedicated(cpu)) + set_cpu_flag(CIF_DEDICATED_CPU); + else + clear_cpu_flag(CIF_DEDICATED_CPU); + set_cpu_online(cpu, true); + inc_irq_stat(CPU_RST); + local_irq_enable(); + cpu_startup_entry(CPUHP_AP_ONLINE_IDLE); +} + +/* Upping and downing of CPUs */ +int __cpu_up(unsigned int cpu, struct task_struct *tidle) +{ + struct pcpu *pcpu = pcpu_devices + cpu; + int rc; + + if (pcpu->state != CPU_STATE_CONFIGURED) + return -EIO; + if (pcpu_sigp_retry(pcpu, SIGP_INITIAL_CPU_RESET, 0) != + SIGP_CC_ORDER_CODE_ACCEPTED) + return -EIO; + + rc = pcpu_alloc_lowcore(pcpu, cpu); + if (rc) + return rc; + /* + * Make sure global control register contents do not change + * until new CPU has initialized control registers. + */ + spin_lock(&ctl_lock); + pcpu_prepare_secondary(pcpu, cpu); + pcpu_attach_task(pcpu, tidle); + pcpu_start_fn(pcpu, smp_start_secondary, NULL); + /* Wait until cpu puts itself in the online & active maps */ + while (!cpu_online(cpu)) + cpu_relax(); + spin_unlock(&ctl_lock); + return 0; +} + +static unsigned int setup_possible_cpus __initdata; + +static int __init _setup_possible_cpus(char *s) +{ + get_option(&s, &setup_possible_cpus); + return 0; +} +early_param("possible_cpus", _setup_possible_cpus); + +int __cpu_disable(void) +{ + unsigned long cregs[16]; + int cpu; + + /* Handle possible pending IPIs */ + smp_handle_ext_call(); + cpu = smp_processor_id(); + set_cpu_online(cpu, false); + cpumask_clear_cpu(cpu, &cpu_setup_mask); + update_cpu_masks(); + /* Disable pseudo page faults on this cpu. */ + pfault_fini(); + /* Disable interrupt sources via control register. */ + __ctl_store(cregs, 0, 15); + cregs[0] &= ~0x0000ee70UL; /* disable all external interrupts */ + cregs[6] &= ~0xff000000UL; /* disable all I/O interrupts */ + cregs[14] &= ~0x1f000000UL; /* disable most machine checks */ + __ctl_load(cregs, 0, 15); + clear_cpu_flag(CIF_NOHZ_DELAY); + return 0; +} + +void __cpu_die(unsigned int cpu) +{ + struct pcpu *pcpu; + + /* Wait until target cpu is down */ + pcpu = pcpu_devices + cpu; + while (!pcpu_stopped(pcpu)) + cpu_relax(); + pcpu_free_lowcore(pcpu); + cpumask_clear_cpu(cpu, mm_cpumask(&init_mm)); + cpumask_clear_cpu(cpu, &init_mm.context.cpu_attach_mask); +} + +void __noreturn cpu_die(void) +{ + idle_task_exit(); + pcpu_sigp_retry(pcpu_devices + smp_processor_id(), SIGP_STOP, 0); + for (;;) ; +} + +void __init smp_fill_possible_mask(void) +{ + unsigned int possible, sclp_max, cpu; + + sclp_max = max(sclp.mtid, sclp.mtid_cp) + 1; + sclp_max = min(smp_max_threads, sclp_max); + sclp_max = (sclp.max_cores * sclp_max) ?: nr_cpu_ids; + possible = setup_possible_cpus ?: nr_cpu_ids; + possible = min(possible, sclp_max); + for (cpu = 0; cpu < possible && cpu < nr_cpu_ids; cpu++) + set_cpu_possible(cpu, true); +} + +void __init smp_prepare_cpus(unsigned int max_cpus) +{ + /* request the 0x1201 emergency signal external interrupt */ + if (register_external_irq(EXT_IRQ_EMERGENCY_SIG, do_ext_call_interrupt)) + panic("Couldn't request external interrupt 0x1201"); + /* request the 0x1202 external call external interrupt */ + if (register_external_irq(EXT_IRQ_EXTERNAL_CALL, do_ext_call_interrupt)) + panic("Couldn't request external interrupt 0x1202"); +} + +void __init smp_prepare_boot_cpu(void) +{ + struct pcpu *pcpu = pcpu_devices; + + WARN_ON(!cpu_present(0) || !cpu_online(0)); + pcpu->state = CPU_STATE_CONFIGURED; + S390_lowcore.percpu_offset = __per_cpu_offset[0]; + smp_cpu_set_polarization(0, POLARIZATION_UNKNOWN); +} + +void __init smp_setup_processor_id(void) +{ + pcpu_devices[0].address = stap(); + S390_lowcore.cpu_nr = 0; + S390_lowcore.spinlock_lockval = arch_spin_lockval(0); + S390_lowcore.spinlock_index = 0; +} + +/* + * the frequency of the profiling timer can be changed + * by writing a multiplier value into /proc/profile. + * + * usually you want to run this on all CPUs ;) + */ +int setup_profiling_timer(unsigned int multiplier) +{ + return 0; +} + +static ssize_t cpu_configure_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + ssize_t count; + + mutex_lock(&smp_cpu_state_mutex); + count = sprintf(buf, "%d\n", pcpu_devices[dev->id].state); + mutex_unlock(&smp_cpu_state_mutex); + return count; +} + +static ssize_t cpu_configure_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct pcpu *pcpu; + int cpu, val, rc, i; + char delim; + + if (sscanf(buf, "%d %c", &val, &delim) != 1) + return -EINVAL; + if (val != 0 && val != 1) + return -EINVAL; + cpus_read_lock(); + mutex_lock(&smp_cpu_state_mutex); + rc = -EBUSY; + /* disallow configuration changes of online cpus and cpu 0 */ + cpu = dev->id; + cpu = smp_get_base_cpu(cpu); + if (cpu == 0) + goto out; + for (i = 0; i <= smp_cpu_mtid; i++) + if (cpu_online(cpu + i)) + goto out; + pcpu = pcpu_devices + cpu; + rc = 0; + switch (val) { + case 0: + if (pcpu->state != CPU_STATE_CONFIGURED) + break; + rc = sclp_core_deconfigure(pcpu->address >> smp_cpu_mt_shift); + if (rc) + break; + for (i = 0; i <= smp_cpu_mtid; i++) { + if (cpu + i >= nr_cpu_ids || !cpu_present(cpu + i)) + continue; + pcpu[i].state = CPU_STATE_STANDBY; + smp_cpu_set_polarization(cpu + i, + POLARIZATION_UNKNOWN); + } + topology_expect_change(); + break; + case 1: + if (pcpu->state != CPU_STATE_STANDBY) + break; + rc = sclp_core_configure(pcpu->address >> smp_cpu_mt_shift); + if (rc) + break; + for (i = 0; i <= smp_cpu_mtid; i++) { + if (cpu + i >= nr_cpu_ids || !cpu_present(cpu + i)) + continue; + pcpu[i].state = CPU_STATE_CONFIGURED; + smp_cpu_set_polarization(cpu + i, + POLARIZATION_UNKNOWN); + } + topology_expect_change(); + break; + default: + break; + } +out: + mutex_unlock(&smp_cpu_state_mutex); + cpus_read_unlock(); + return rc ? rc : count; +} +static DEVICE_ATTR(configure, 0644, cpu_configure_show, cpu_configure_store); + +static ssize_t show_cpu_address(struct device *dev, + struct device_attribute *attr, char *buf) +{ + return sprintf(buf, "%d\n", pcpu_devices[dev->id].address); +} +static DEVICE_ATTR(address, 0444, show_cpu_address, NULL); + +static struct attribute *cpu_common_attrs[] = { + &dev_attr_configure.attr, + &dev_attr_address.attr, + NULL, +}; + +static struct attribute_group cpu_common_attr_group = { + .attrs = cpu_common_attrs, +}; + +static struct attribute *cpu_online_attrs[] = { + &dev_attr_idle_count.attr, + &dev_attr_idle_time_us.attr, + NULL, +}; + +static struct attribute_group cpu_online_attr_group = { + .attrs = cpu_online_attrs, +}; + +static int smp_cpu_online(unsigned int cpu) +{ + struct device *s = &per_cpu(cpu_device, cpu)->dev; + + return sysfs_create_group(&s->kobj, &cpu_online_attr_group); +} + +static int smp_cpu_pre_down(unsigned int cpu) +{ + struct device *s = &per_cpu(cpu_device, cpu)->dev; + + sysfs_remove_group(&s->kobj, &cpu_online_attr_group); + return 0; +} + +static int smp_add_present_cpu(int cpu) +{ + struct device *s; + struct cpu *c; + int rc; + + c = kzalloc(sizeof(*c), GFP_KERNEL); + if (!c) + return -ENOMEM; + per_cpu(cpu_device, cpu) = c; + s = &c->dev; + c->hotpluggable = 1; + rc = register_cpu(c, cpu); + if (rc) + goto out; + rc = sysfs_create_group(&s->kobj, &cpu_common_attr_group); + if (rc) + goto out_cpu; + rc = topology_cpu_init(c); + if (rc) + goto out_topology; + return 0; + +out_topology: + sysfs_remove_group(&s->kobj, &cpu_common_attr_group); +out_cpu: + unregister_cpu(c); +out: + return rc; +} + +int __ref smp_rescan_cpus(void) +{ + struct sclp_core_info *info; + int nr; + + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) + return -ENOMEM; + smp_get_core_info(info, 0); + nr = __smp_rescan_cpus(info, false); + kfree(info); + if (nr) + topology_schedule_update(); + return 0; +} + +static ssize_t __ref rescan_store(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + int rc; + + rc = lock_device_hotplug_sysfs(); + if (rc) + return rc; + rc = smp_rescan_cpus(); + unlock_device_hotplug(); + return rc ? rc : count; +} +static DEVICE_ATTR_WO(rescan); + +static int __init s390_smp_init(void) +{ + struct device *dev_root; + int cpu, rc = 0; + + dev_root = bus_get_dev_root(&cpu_subsys); + if (dev_root) { + rc = device_create_file(dev_root, &dev_attr_rescan); + put_device(dev_root); + if (rc) + return rc; + } + + for_each_present_cpu(cpu) { + rc = smp_add_present_cpu(cpu); + if (rc) + goto out; + } + + rc = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "s390/smp:online", + smp_cpu_online, smp_cpu_pre_down); + rc = rc <= 0 ? rc : 0; +out: + return rc; +} +subsys_initcall(s390_smp_init); + +static __always_inline void set_new_lowcore(struct lowcore *lc) +{ + union register_pair dst, src; + u32 pfx; + + src.even = (unsigned long) &S390_lowcore; + src.odd = sizeof(S390_lowcore); + dst.even = (unsigned long) lc; + dst.odd = sizeof(*lc); + pfx = __pa(lc); + + asm volatile( + " mvcl %[dst],%[src]\n" + " spx %[pfx]\n" + : [dst] "+&d" (dst.pair), [src] "+&d" (src.pair) + : [pfx] "Q" (pfx) + : "memory", "cc"); +} + +int __init smp_reinit_ipl_cpu(void) +{ + unsigned long async_stack, nodat_stack, mcck_stack; + struct lowcore *lc, *lc_ipl; + unsigned long flags, cr0; + u64 mcesad; + + lc_ipl = lowcore_ptr[0]; + lc = (struct lowcore *) __get_free_pages(GFP_KERNEL | GFP_DMA, LC_ORDER); + nodat_stack = __get_free_pages(GFP_KERNEL, THREAD_SIZE_ORDER); + async_stack = stack_alloc(); + mcck_stack = stack_alloc(); + if (!lc || !nodat_stack || !async_stack || !mcck_stack || nmi_alloc_mcesa(&mcesad)) + panic("Couldn't allocate memory"); + + local_irq_save(flags); + local_mcck_disable(); + set_new_lowcore(lc); + S390_lowcore.nodat_stack = nodat_stack + STACK_INIT_OFFSET; + S390_lowcore.async_stack = async_stack + STACK_INIT_OFFSET; + S390_lowcore.mcck_stack = mcck_stack + STACK_INIT_OFFSET; + __ctl_store(cr0, 0, 0); + __ctl_clear_bit(0, 28); /* disable lowcore protection */ + S390_lowcore.mcesad = mcesad; + __ctl_load(cr0, 0, 0); + if (abs_lowcore_map(0, lc, false)) + panic("Couldn't remap absolute lowcore"); + lowcore_ptr[0] = lc; + local_mcck_enable(); + local_irq_restore(flags); + + memblock_free_late(__pa(lc_ipl->mcck_stack - STACK_INIT_OFFSET), THREAD_SIZE); + memblock_free_late(__pa(lc_ipl->async_stack - STACK_INIT_OFFSET), THREAD_SIZE); + memblock_free_late(__pa(lc_ipl->nodat_stack - STACK_INIT_OFFSET), THREAD_SIZE); + memblock_free_late(__pa(lc_ipl), sizeof(*lc_ipl)); + return 0; +} diff --git a/arch/s390/kernel/stacktrace.c b/arch/s390/kernel/stacktrace.c new file mode 100644 index 0000000000..0787010139 --- /dev/null +++ b/arch/s390/kernel/stacktrace.c @@ -0,0 +1,60 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Stack trace management functions + * + * Copyright IBM Corp. 2006 + */ + +#include <linux/stacktrace.h> +#include <asm/stacktrace.h> +#include <asm/unwind.h> +#include <asm/kprobes.h> + +void arch_stack_walk(stack_trace_consume_fn consume_entry, void *cookie, + struct task_struct *task, struct pt_regs *regs) +{ + struct unwind_state state; + unsigned long addr; + + unwind_for_each_frame(&state, task, regs, 0) { + addr = unwind_get_return_address(&state); + if (!addr || !consume_entry(cookie, addr)) + break; + } +} + +int arch_stack_walk_reliable(stack_trace_consume_fn consume_entry, + void *cookie, struct task_struct *task) +{ + struct unwind_state state; + unsigned long addr; + + unwind_for_each_frame(&state, task, NULL, 0) { + if (state.stack_info.type != STACK_TYPE_TASK) + return -EINVAL; + + if (state.regs) + return -EINVAL; + + addr = unwind_get_return_address(&state); + if (!addr) + return -EINVAL; + +#ifdef CONFIG_RETHOOK + /* + * Mark stacktraces with krethook functions on them + * as unreliable. + */ + if (state.ip == (unsigned long)arch_rethook_trampoline) + return -EINVAL; +#endif + + if (!consume_entry(cookie, addr)) + return -EINVAL; + } + + /* Check for stack corruption */ + if (unwind_error(&state)) + return -EINVAL; + return 0; +} diff --git a/arch/s390/kernel/sthyi.c b/arch/s390/kernel/sthyi.c new file mode 100644 index 0000000000..30bb20461d --- /dev/null +++ b/arch/s390/kernel/sthyi.c @@ -0,0 +1,517 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * store hypervisor information instruction emulation functions. + * + * Copyright IBM Corp. 2016 + * Author(s): Janosch Frank <frankja@linux.vnet.ibm.com> + */ +#include <linux/errno.h> +#include <linux/pagemap.h> +#include <linux/vmalloc.h> +#include <linux/syscalls.h> +#include <linux/mutex.h> +#include <asm/asm-offsets.h> +#include <asm/sclp.h> +#include <asm/diag.h> +#include <asm/sysinfo.h> +#include <asm/ebcdic.h> +#include <asm/facility.h> +#include <asm/sthyi.h> +#include "entry.h" + +#define DED_WEIGHT 0xffff +/* + * CP and IFL as EBCDIC strings, SP/0x40 determines the end of string + * as they are justified with spaces. + */ +#define CP 0xc3d7404040404040UL +#define IFL 0xc9c6d34040404040UL + +enum hdr_flags { + HDR_NOT_LPAR = 0x10, + HDR_STACK_INCM = 0x20, + HDR_STSI_UNAV = 0x40, + HDR_PERF_UNAV = 0x80, +}; + +enum mac_validity { + MAC_NAME_VLD = 0x20, + MAC_ID_VLD = 0x40, + MAC_CNT_VLD = 0x80, +}; + +enum par_flag { + PAR_MT_EN = 0x80, +}; + +enum par_validity { + PAR_GRP_VLD = 0x08, + PAR_ID_VLD = 0x10, + PAR_ABS_VLD = 0x20, + PAR_WGHT_VLD = 0x40, + PAR_PCNT_VLD = 0x80, +}; + +struct hdr_sctn { + u8 infhflg1; + u8 infhflg2; /* reserved */ + u8 infhval1; /* reserved */ + u8 infhval2; /* reserved */ + u8 reserved[3]; + u8 infhygct; + u16 infhtotl; + u16 infhdln; + u16 infmoff; + u16 infmlen; + u16 infpoff; + u16 infplen; + u16 infhoff1; + u16 infhlen1; + u16 infgoff1; + u16 infglen1; + u16 infhoff2; + u16 infhlen2; + u16 infgoff2; + u16 infglen2; + u16 infhoff3; + u16 infhlen3; + u16 infgoff3; + u16 infglen3; + u8 reserved2[4]; +} __packed; + +struct mac_sctn { + u8 infmflg1; /* reserved */ + u8 infmflg2; /* reserved */ + u8 infmval1; + u8 infmval2; /* reserved */ + u16 infmscps; + u16 infmdcps; + u16 infmsifl; + u16 infmdifl; + char infmname[8]; + char infmtype[4]; + char infmmanu[16]; + char infmseq[16]; + char infmpman[4]; + u8 reserved[4]; +} __packed; + +struct par_sctn { + u8 infpflg1; + u8 infpflg2; /* reserved */ + u8 infpval1; + u8 infpval2; /* reserved */ + u16 infppnum; + u16 infpscps; + u16 infpdcps; + u16 infpsifl; + u16 infpdifl; + u16 reserved; + char infppnam[8]; + u32 infpwbcp; + u32 infpabcp; + u32 infpwbif; + u32 infpabif; + char infplgnm[8]; + u32 infplgcp; + u32 infplgif; +} __packed; + +struct sthyi_sctns { + struct hdr_sctn hdr; + struct mac_sctn mac; + struct par_sctn par; +} __packed; + +struct cpu_inf { + u64 lpar_cap; + u64 lpar_grp_cap; + u64 lpar_weight; + u64 all_weight; + int cpu_num_ded; + int cpu_num_shd; +}; + +struct lpar_cpu_inf { + struct cpu_inf cp; + struct cpu_inf ifl; +}; + +/* + * STHYI requires extensive locking in the higher hypervisors + * and is very computational/memory expensive. Therefore we + * cache the retrieved data whose valid period is 1s. + */ +#define CACHE_VALID_JIFFIES HZ + +struct sthyi_info { + void *info; + unsigned long end; +}; + +static DEFINE_MUTEX(sthyi_mutex); +static struct sthyi_info sthyi_cache; + +static inline u64 cpu_id(u8 ctidx, void *diag224_buf) +{ + return *((u64 *)(diag224_buf + (ctidx + 1) * DIAG204_CPU_NAME_LEN)); +} + +/* + * Scales the cpu capping from the lpar range to the one expected in + * sthyi data. + * + * diag204 reports a cap in hundredths of processor units. + * z/VM's range for one core is 0 - 0x10000. + */ +static u32 scale_cap(u32 in) +{ + return (0x10000 * in) / 100; +} + +static void fill_hdr(struct sthyi_sctns *sctns) +{ + sctns->hdr.infhdln = sizeof(sctns->hdr); + sctns->hdr.infmoff = sizeof(sctns->hdr); + sctns->hdr.infmlen = sizeof(sctns->mac); + sctns->hdr.infplen = sizeof(sctns->par); + sctns->hdr.infpoff = sctns->hdr.infhdln + sctns->hdr.infmlen; + sctns->hdr.infhtotl = sctns->hdr.infpoff + sctns->hdr.infplen; +} + +static void fill_stsi_mac(struct sthyi_sctns *sctns, + struct sysinfo_1_1_1 *sysinfo) +{ + sclp_ocf_cpc_name_copy(sctns->mac.infmname); + if (*(u64 *)sctns->mac.infmname != 0) + sctns->mac.infmval1 |= MAC_NAME_VLD; + + if (stsi(sysinfo, 1, 1, 1)) + return; + + memcpy(sctns->mac.infmtype, sysinfo->type, sizeof(sctns->mac.infmtype)); + memcpy(sctns->mac.infmmanu, sysinfo->manufacturer, sizeof(sctns->mac.infmmanu)); + memcpy(sctns->mac.infmpman, sysinfo->plant, sizeof(sctns->mac.infmpman)); + memcpy(sctns->mac.infmseq, sysinfo->sequence, sizeof(sctns->mac.infmseq)); + + sctns->mac.infmval1 |= MAC_ID_VLD; +} + +static void fill_stsi_par(struct sthyi_sctns *sctns, + struct sysinfo_2_2_2 *sysinfo) +{ + if (stsi(sysinfo, 2, 2, 2)) + return; + + sctns->par.infppnum = sysinfo->lpar_number; + memcpy(sctns->par.infppnam, sysinfo->name, sizeof(sctns->par.infppnam)); + + sctns->par.infpval1 |= PAR_ID_VLD; +} + +static void fill_stsi(struct sthyi_sctns *sctns) +{ + void *sysinfo; + + /* Errors are handled through the validity bits in the response. */ + sysinfo = (void *)__get_free_page(GFP_KERNEL); + if (!sysinfo) + return; + + fill_stsi_mac(sctns, sysinfo); + fill_stsi_par(sctns, sysinfo); + + free_pages((unsigned long)sysinfo, 0); +} + +static void fill_diag_mac(struct sthyi_sctns *sctns, + struct diag204_x_phys_block *block, + void *diag224_buf) +{ + int i; + + for (i = 0; i < block->hdr.cpus; i++) { + switch (cpu_id(block->cpus[i].ctidx, diag224_buf)) { + case CP: + if (block->cpus[i].weight == DED_WEIGHT) + sctns->mac.infmdcps++; + else + sctns->mac.infmscps++; + break; + case IFL: + if (block->cpus[i].weight == DED_WEIGHT) + sctns->mac.infmdifl++; + else + sctns->mac.infmsifl++; + break; + } + } + sctns->mac.infmval1 |= MAC_CNT_VLD; +} + +/* Returns a pointer to the the next partition block. */ +static struct diag204_x_part_block *lpar_cpu_inf(struct lpar_cpu_inf *part_inf, + bool this_lpar, + void *diag224_buf, + struct diag204_x_part_block *block) +{ + int i, capped = 0, weight_cp = 0, weight_ifl = 0; + struct cpu_inf *cpu_inf; + + for (i = 0; i < block->hdr.rcpus; i++) { + if (!(block->cpus[i].cflag & DIAG204_CPU_ONLINE)) + continue; + + switch (cpu_id(block->cpus[i].ctidx, diag224_buf)) { + case CP: + cpu_inf = &part_inf->cp; + if (block->cpus[i].cur_weight < DED_WEIGHT) + weight_cp |= block->cpus[i].cur_weight; + break; + case IFL: + cpu_inf = &part_inf->ifl; + if (block->cpus[i].cur_weight < DED_WEIGHT) + weight_ifl |= block->cpus[i].cur_weight; + break; + default: + continue; + } + + if (!this_lpar) + continue; + + capped |= block->cpus[i].cflag & DIAG204_CPU_CAPPED; + cpu_inf->lpar_cap |= block->cpus[i].cpu_type_cap; + cpu_inf->lpar_grp_cap |= block->cpus[i].group_cpu_type_cap; + + if (block->cpus[i].weight == DED_WEIGHT) + cpu_inf->cpu_num_ded += 1; + else + cpu_inf->cpu_num_shd += 1; + } + + if (this_lpar && capped) { + part_inf->cp.lpar_weight = weight_cp; + part_inf->ifl.lpar_weight = weight_ifl; + } + part_inf->cp.all_weight += weight_cp; + part_inf->ifl.all_weight += weight_ifl; + return (struct diag204_x_part_block *)&block->cpus[i]; +} + +static void fill_diag(struct sthyi_sctns *sctns) +{ + int i, r, pages; + bool this_lpar; + void *diag204_buf; + void *diag224_buf = NULL; + struct diag204_x_info_blk_hdr *ti_hdr; + struct diag204_x_part_block *part_block; + struct diag204_x_phys_block *phys_block; + struct lpar_cpu_inf lpar_inf = {}; + + /* Errors are handled through the validity bits in the response. */ + pages = diag204((unsigned long)DIAG204_SUBC_RSI | + (unsigned long)DIAG204_INFO_EXT, 0, NULL); + if (pages <= 0) + return; + + diag204_buf = __vmalloc_node(array_size(pages, PAGE_SIZE), + PAGE_SIZE, GFP_KERNEL, NUMA_NO_NODE, + __builtin_return_address(0)); + if (!diag204_buf) + return; + + r = diag204((unsigned long)DIAG204_SUBC_STIB7 | + (unsigned long)DIAG204_INFO_EXT, pages, diag204_buf); + if (r < 0) + goto out; + + diag224_buf = (void *)__get_free_page(GFP_KERNEL | GFP_DMA); + if (!diag224_buf || diag224(diag224_buf)) + goto out; + + ti_hdr = diag204_buf; + part_block = diag204_buf + sizeof(*ti_hdr); + + for (i = 0; i < ti_hdr->npar; i++) { + /* + * For the calling lpar we also need to get the cpu + * caps and weights. The time information block header + * specifies the offset to the partition block of the + * caller lpar, so we know when we process its data. + */ + this_lpar = (void *)part_block - diag204_buf == ti_hdr->this_part; + part_block = lpar_cpu_inf(&lpar_inf, this_lpar, diag224_buf, + part_block); + } + + phys_block = (struct diag204_x_phys_block *)part_block; + part_block = diag204_buf + ti_hdr->this_part; + if (part_block->hdr.mtid) + sctns->par.infpflg1 = PAR_MT_EN; + + sctns->par.infpval1 |= PAR_GRP_VLD; + sctns->par.infplgcp = scale_cap(lpar_inf.cp.lpar_grp_cap); + sctns->par.infplgif = scale_cap(lpar_inf.ifl.lpar_grp_cap); + memcpy(sctns->par.infplgnm, part_block->hdr.hardware_group_name, + sizeof(sctns->par.infplgnm)); + + sctns->par.infpscps = lpar_inf.cp.cpu_num_shd; + sctns->par.infpdcps = lpar_inf.cp.cpu_num_ded; + sctns->par.infpsifl = lpar_inf.ifl.cpu_num_shd; + sctns->par.infpdifl = lpar_inf.ifl.cpu_num_ded; + sctns->par.infpval1 |= PAR_PCNT_VLD; + + sctns->par.infpabcp = scale_cap(lpar_inf.cp.lpar_cap); + sctns->par.infpabif = scale_cap(lpar_inf.ifl.lpar_cap); + sctns->par.infpval1 |= PAR_ABS_VLD; + + /* + * Everything below needs global performance data to be + * meaningful. + */ + if (!(ti_hdr->flags & DIAG204_LPAR_PHYS_FLG)) { + sctns->hdr.infhflg1 |= HDR_PERF_UNAV; + goto out; + } + + fill_diag_mac(sctns, phys_block, diag224_buf); + + if (lpar_inf.cp.lpar_weight) { + sctns->par.infpwbcp = sctns->mac.infmscps * 0x10000 * + lpar_inf.cp.lpar_weight / lpar_inf.cp.all_weight; + } + + if (lpar_inf.ifl.lpar_weight) { + sctns->par.infpwbif = sctns->mac.infmsifl * 0x10000 * + lpar_inf.ifl.lpar_weight / lpar_inf.ifl.all_weight; + } + sctns->par.infpval1 |= PAR_WGHT_VLD; + +out: + free_page((unsigned long)diag224_buf); + vfree(diag204_buf); +} + +static int sthyi(u64 vaddr, u64 *rc) +{ + union register_pair r1 = { .even = 0, }; /* subcode */ + union register_pair r2 = { .even = vaddr, }; + int cc; + + asm volatile( + ".insn rre,0xB2560000,%[r1],%[r2]\n" + "ipm %[cc]\n" + "srl %[cc],28\n" + : [cc] "=&d" (cc), [r2] "+&d" (r2.pair) + : [r1] "d" (r1.pair) + : "memory", "cc"); + *rc = r2.odd; + return cc; +} + +static int fill_dst(void *dst, u64 *rc) +{ + struct sthyi_sctns *sctns = (struct sthyi_sctns *)dst; + + /* + * If the facility is on, we don't want to emulate the instruction. + * We ask the hypervisor to provide the data. + */ + if (test_facility(74)) + return sthyi((u64)dst, rc); + + fill_hdr(sctns); + fill_stsi(sctns); + fill_diag(sctns); + *rc = 0; + return 0; +} + +static int sthyi_init_cache(void) +{ + if (sthyi_cache.info) + return 0; + sthyi_cache.info = (void *)get_zeroed_page(GFP_KERNEL); + if (!sthyi_cache.info) + return -ENOMEM; + sthyi_cache.end = jiffies - 1; /* expired */ + return 0; +} + +static int sthyi_update_cache(u64 *rc) +{ + int r; + + memset(sthyi_cache.info, 0, PAGE_SIZE); + r = fill_dst(sthyi_cache.info, rc); + if (r) + return r; + sthyi_cache.end = jiffies + CACHE_VALID_JIFFIES; + return r; +} + +/* + * sthyi_fill - Fill page with data returned by the STHYI instruction + * + * @dst: Pointer to zeroed page + * @rc: Pointer for storing the return code of the instruction + * + * Fills the destination with system information returned by the STHYI + * instruction. The data is generated by emulation or execution of STHYI, + * if available. The return value is either a negative error value or + * the condition code that would be returned, the rc parameter is the + * return code which is passed in register R2 + 1. + */ +int sthyi_fill(void *dst, u64 *rc) +{ + int r; + + mutex_lock(&sthyi_mutex); + r = sthyi_init_cache(); + if (r) + goto out; + + if (time_is_before_jiffies(sthyi_cache.end)) { + /* cache expired */ + r = sthyi_update_cache(rc); + if (r) + goto out; + } + *rc = 0; + memcpy(dst, sthyi_cache.info, PAGE_SIZE); +out: + mutex_unlock(&sthyi_mutex); + return r; +} +EXPORT_SYMBOL_GPL(sthyi_fill); + +SYSCALL_DEFINE4(s390_sthyi, unsigned long, function_code, void __user *, buffer, + u64 __user *, return_code, unsigned long, flags) +{ + u64 sthyi_rc; + void *info; + int r; + + if (flags) + return -EINVAL; + if (function_code != STHYI_FC_CP_IFL_CAP) + return -EOPNOTSUPP; + info = (void *)get_zeroed_page(GFP_KERNEL); + if (!info) + return -ENOMEM; + r = sthyi_fill(info, &sthyi_rc); + if (r < 0) + goto out; + if (return_code && put_user(sthyi_rc, return_code)) { + r = -EFAULT; + goto out; + } + if (copy_to_user(buffer, info, PAGE_SIZE)) + r = -EFAULT; +out: + free_page((unsigned long)info); + return r; +} diff --git a/arch/s390/kernel/syscall.c b/arch/s390/kernel/syscall.c new file mode 100644 index 0000000000..dc2355c623 --- /dev/null +++ b/arch/s390/kernel/syscall.c @@ -0,0 +1,170 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * S390 version + * Copyright IBM Corp. 1999, 2000 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), + * Thomas Spatzier (tspat@de.ibm.com) + * + * Derived from "arch/i386/kernel/sys_i386.c" + * + * This file contains various random system calls that + * have a non-standard calling sequence on the Linux/s390 + * platform. + */ + +#include <linux/errno.h> +#include <linux/sched.h> +#include <linux/mm.h> +#include <linux/fs.h> +#include <linux/smp.h> +#include <linux/sem.h> +#include <linux/msg.h> +#include <linux/shm.h> +#include <linux/stat.h> +#include <linux/syscalls.h> +#include <linux/mman.h> +#include <linux/file.h> +#include <linux/utsname.h> +#include <linux/personality.h> +#include <linux/unistd.h> +#include <linux/ipc.h> +#include <linux/uaccess.h> +#include <linux/string.h> +#include <linux/thread_info.h> +#include <linux/entry-common.h> + +#include <asm/ptrace.h> +#include <asm/vtime.h> + +#include "entry.h" + +/* + * Perform the mmap() system call. Linux for S/390 isn't able to handle more + * than 5 system call parameters, so this system call uses a memory block + * for parameter passing. + */ + +struct s390_mmap_arg_struct { + unsigned long addr; + unsigned long len; + unsigned long prot; + unsigned long flags; + unsigned long fd; + unsigned long offset; +}; + +SYSCALL_DEFINE1(mmap2, struct s390_mmap_arg_struct __user *, arg) +{ + struct s390_mmap_arg_struct a; + int error = -EFAULT; + + if (copy_from_user(&a, arg, sizeof(a))) + goto out; + error = ksys_mmap_pgoff(a.addr, a.len, a.prot, a.flags, a.fd, a.offset); +out: + return error; +} + +#ifdef CONFIG_SYSVIPC +/* + * sys_ipc() is the de-multiplexer for the SysV IPC calls. + */ +SYSCALL_DEFINE5(s390_ipc, uint, call, int, first, unsigned long, second, + unsigned long, third, void __user *, ptr) +{ + if (call >> 16) + return -EINVAL; + /* The s390 sys_ipc variant has only five parameters instead of six + * like the generic variant. The only difference is the handling of + * the SEMTIMEDOP subcall where on s390 the third parameter is used + * as a pointer to a struct timespec where the generic variant uses + * the fifth parameter. + * Therefore we can call the generic variant by simply passing the + * third parameter also as fifth parameter. + */ + return ksys_ipc(call, first, second, third, ptr, third); +} +#endif /* CONFIG_SYSVIPC */ + +SYSCALL_DEFINE1(s390_personality, unsigned int, personality) +{ + unsigned int ret = current->personality; + + if (personality(current->personality) == PER_LINUX32 && + personality(personality) == PER_LINUX) + personality |= PER_LINUX32; + + if (personality != 0xffffffff) + set_personality(personality); + + if (personality(ret) == PER_LINUX32) + ret &= ~PER_LINUX32; + + return ret; +} + +SYSCALL_DEFINE0(ni_syscall) +{ + return -ENOSYS; +} + +static void do_syscall(struct pt_regs *regs) +{ + unsigned long nr; + + nr = regs->int_code & 0xffff; + if (!nr) { + nr = regs->gprs[1] & 0xffff; + regs->int_code &= ~0xffffUL; + regs->int_code |= nr; + } + + regs->gprs[2] = nr; + + if (nr == __NR_restart_syscall && !(current->restart_block.arch_data & 1)) { + regs->psw.addr = current->restart_block.arch_data; + current->restart_block.arch_data = 1; + } + nr = syscall_enter_from_user_mode_work(regs, nr); + + /* + * In the s390 ptrace ABI, both the syscall number and the return value + * use gpr2. However, userspace puts the syscall number either in the + * svc instruction itself, or uses gpr1. To make at least skipping syscalls + * work, the ptrace code sets PIF_SYSCALL_RET_SET, which is checked here + * and if set, the syscall will be skipped. + */ + + if (unlikely(test_and_clear_pt_regs_flag(regs, PIF_SYSCALL_RET_SET))) + goto out; + regs->gprs[2] = -ENOSYS; + if (likely(nr >= NR_syscalls)) + goto out; + do { + regs->gprs[2] = current->thread.sys_call_table[nr](regs); + } while (test_and_clear_pt_regs_flag(regs, PIF_EXECVE_PGSTE_RESTART)); +out: + syscall_exit_to_user_mode_work(regs); +} + +void noinstr __do_syscall(struct pt_regs *regs, int per_trap) +{ + add_random_kstack_offset(); + enter_from_user_mode(regs); + regs->psw = S390_lowcore.svc_old_psw; + regs->int_code = S390_lowcore.svc_int_code; + update_timer_sys(); + if (static_branch_likely(&cpu_has_bear)) + current->thread.last_break = regs->last_break; + + local_irq_enable(); + regs->orig_gpr2 = regs->gprs[2]; + + if (per_trap) + set_thread_flag(TIF_PER_TRAP); + + regs->flags = 0; + set_pt_regs_flag(regs, PIF_SYSCALL); + do_syscall(regs); + exit_to_user_mode(); +} diff --git a/arch/s390/kernel/syscalls/Makefile b/arch/s390/kernel/syscalls/Makefile new file mode 100644 index 0000000000..fb85e79794 --- /dev/null +++ b/arch/s390/kernel/syscalls/Makefile @@ -0,0 +1,45 @@ +# SPDX-License-Identifier: GPL-2.0 + +gen := arch/$(ARCH)/include/generated +kapi := $(gen)/asm +uapi := $(gen)/uapi/asm + +syscall := $(srctree)/$(src)/syscall.tbl +systbl := $(srctree)/$(src)/syscalltbl + +gen-y := $(kapi)/syscall_table.h +kapi-hdrs-y := $(kapi)/unistd_nr.h +uapi-hdrs-y := $(uapi)/unistd_32.h +uapi-hdrs-y += $(uapi)/unistd_64.h + +targets += $(addprefix ../../../,$(gen-y) $(kapi-hdrs-y) $(uapi-hdrs-y)) + +PHONY += kapi uapi + +kapi: $(gen-y) $(kapi-hdrs-y) +uapi: $(uapi-hdrs-y) + + +# Create output directory if not already present +$(shell mkdir -p $(uapi) $(kapi)) + +filechk_syshdr = $(CONFIG_SHELL) '$(systbl)' -H -a $(syshdr_abi_$(basetarget)) -f "$2" < $< + +filechk_sysnr = $(CONFIG_SHELL) '$(systbl)' -N -a $(sysnr_abi_$(basetarget)) < $< + +filechk_syscalls = $(CONFIG_SHELL) '$(systbl)' -S < $< + +syshdr_abi_unistd_32 := common,32 +$(uapi)/unistd_32.h: $(syscall) FORCE + $(call filechk,syshdr,$@) + +syshdr_abi_unistd_64 := common,64 +$(uapi)/unistd_64.h: $(syscall) FORCE + $(call filechk,syshdr,$@) + +$(kapi)/syscall_table.h: $(syscall) FORCE + $(call filechk,syscalls) + +sysnr_abi_unistd_nr := common,32,64 +$(kapi)/unistd_nr.h: $(syscall) FORCE + $(call filechk,sysnr) diff --git a/arch/s390/kernel/syscalls/syscall.tbl b/arch/s390/kernel/syscalls/syscall.tbl new file mode 100644 index 0000000000..0122cc1569 --- /dev/null +++ b/arch/s390/kernel/syscalls/syscall.tbl @@ -0,0 +1,457 @@ +# SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note +# +# System call table for s390 +# +# Format: +# +# <nr> <abi> <syscall> <entry-64bit> <compat-entry> +# +# where <abi> can be common, 64, or 32 + +1 common exit sys_exit sys_exit +2 common fork sys_fork sys_fork +3 common read sys_read compat_sys_s390_read +4 common write sys_write compat_sys_s390_write +5 common open sys_open compat_sys_open +6 common close sys_close sys_close +7 common restart_syscall sys_restart_syscall sys_restart_syscall +8 common creat sys_creat sys_creat +9 common link sys_link sys_link +10 common unlink sys_unlink sys_unlink +11 common execve sys_execve compat_sys_execve +12 common chdir sys_chdir sys_chdir +13 32 time - sys_time32 +14 common mknod sys_mknod sys_mknod +15 common chmod sys_chmod sys_chmod +16 32 lchown - sys_lchown16 +19 common lseek sys_lseek compat_sys_lseek +20 common getpid sys_getpid sys_getpid +21 common mount sys_mount sys_mount +22 common umount sys_oldumount sys_oldumount +23 32 setuid - sys_setuid16 +24 32 getuid - sys_getuid16 +25 32 stime - sys_stime32 +26 common ptrace sys_ptrace compat_sys_ptrace +27 common alarm sys_alarm sys_alarm +29 common pause sys_pause sys_pause +30 common utime sys_utime sys_utime32 +33 common access sys_access sys_access +34 common nice sys_nice sys_nice +36 common sync sys_sync sys_sync +37 common kill sys_kill sys_kill +38 common rename sys_rename sys_rename +39 common mkdir sys_mkdir sys_mkdir +40 common rmdir sys_rmdir sys_rmdir +41 common dup sys_dup sys_dup +42 common pipe sys_pipe sys_pipe +43 common times sys_times compat_sys_times +45 common brk sys_brk sys_brk +46 32 setgid - sys_setgid16 +47 32 getgid - sys_getgid16 +48 common signal sys_signal sys_signal +49 32 geteuid - sys_geteuid16 +50 32 getegid - sys_getegid16 +51 common acct sys_acct sys_acct +52 common umount2 sys_umount sys_umount +54 common ioctl sys_ioctl compat_sys_ioctl +55 common fcntl sys_fcntl compat_sys_fcntl +57 common setpgid sys_setpgid sys_setpgid +60 common umask sys_umask sys_umask +61 common chroot sys_chroot sys_chroot +62 common ustat sys_ustat compat_sys_ustat +63 common dup2 sys_dup2 sys_dup2 +64 common getppid sys_getppid sys_getppid +65 common getpgrp sys_getpgrp sys_getpgrp +66 common setsid sys_setsid sys_setsid +67 common sigaction sys_sigaction compat_sys_sigaction +70 32 setreuid - sys_setreuid16 +71 32 setregid - sys_setregid16 +72 common sigsuspend sys_sigsuspend sys_sigsuspend +73 common sigpending sys_sigpending compat_sys_sigpending +74 common sethostname sys_sethostname sys_sethostname +75 common setrlimit sys_setrlimit compat_sys_setrlimit +76 32 getrlimit - compat_sys_old_getrlimit +77 common getrusage sys_getrusage compat_sys_getrusage +78 common gettimeofday sys_gettimeofday compat_sys_gettimeofday +79 common settimeofday sys_settimeofday compat_sys_settimeofday +80 32 getgroups - sys_getgroups16 +81 32 setgroups - sys_setgroups16 +83 common symlink sys_symlink sys_symlink +85 common readlink sys_readlink sys_readlink +86 common uselib sys_uselib sys_uselib +87 common swapon sys_swapon sys_swapon +88 common reboot sys_reboot sys_reboot +89 common readdir - compat_sys_old_readdir +90 common mmap sys_old_mmap compat_sys_s390_old_mmap +91 common munmap sys_munmap sys_munmap +92 common truncate sys_truncate compat_sys_truncate +93 common ftruncate sys_ftruncate compat_sys_ftruncate +94 common fchmod sys_fchmod sys_fchmod +95 32 fchown - sys_fchown16 +96 common getpriority sys_getpriority sys_getpriority +97 common setpriority sys_setpriority sys_setpriority +99 common statfs sys_statfs compat_sys_statfs +100 common fstatfs sys_fstatfs compat_sys_fstatfs +101 32 ioperm - - +102 common socketcall sys_socketcall compat_sys_socketcall +103 common syslog sys_syslog sys_syslog +104 common setitimer sys_setitimer compat_sys_setitimer +105 common getitimer sys_getitimer compat_sys_getitimer +106 common stat sys_newstat compat_sys_newstat +107 common lstat sys_newlstat compat_sys_newlstat +108 common fstat sys_newfstat compat_sys_newfstat +110 common lookup_dcookie sys_lookup_dcookie compat_sys_lookup_dcookie +111 common vhangup sys_vhangup sys_vhangup +112 common idle - - +114 common wait4 sys_wait4 compat_sys_wait4 +115 common swapoff sys_swapoff sys_swapoff +116 common sysinfo sys_sysinfo compat_sys_sysinfo +117 common ipc sys_s390_ipc compat_sys_s390_ipc +118 common fsync sys_fsync sys_fsync +119 common sigreturn sys_sigreturn compat_sys_sigreturn +120 common clone sys_clone sys_clone +121 common setdomainname sys_setdomainname sys_setdomainname +122 common uname sys_newuname sys_newuname +124 common adjtimex sys_adjtimex sys_adjtimex_time32 +125 common mprotect sys_mprotect sys_mprotect +126 common sigprocmask sys_sigprocmask compat_sys_sigprocmask +127 common create_module - - +128 common init_module sys_init_module sys_init_module +129 common delete_module sys_delete_module sys_delete_module +130 common get_kernel_syms - - +131 common quotactl sys_quotactl sys_quotactl +132 common getpgid sys_getpgid sys_getpgid +133 common fchdir sys_fchdir sys_fchdir +134 common bdflush sys_ni_syscall sys_ni_syscall +135 common sysfs sys_sysfs sys_sysfs +136 common personality sys_s390_personality sys_s390_personality +137 common afs_syscall - - +138 32 setfsuid - sys_setfsuid16 +139 32 setfsgid - sys_setfsgid16 +140 32 _llseek - sys_llseek +141 common getdents sys_getdents compat_sys_getdents +142 32 _newselect - compat_sys_select +142 64 select sys_select - +143 common flock sys_flock sys_flock +144 common msync sys_msync sys_msync +145 common readv sys_readv sys_readv +146 common writev sys_writev sys_writev +147 common getsid sys_getsid sys_getsid +148 common fdatasync sys_fdatasync sys_fdatasync +149 common _sysctl - - +150 common mlock sys_mlock sys_mlock +151 common munlock sys_munlock sys_munlock +152 common mlockall sys_mlockall sys_mlockall +153 common munlockall sys_munlockall sys_munlockall +154 common sched_setparam sys_sched_setparam sys_sched_setparam +155 common sched_getparam sys_sched_getparam sys_sched_getparam +156 common sched_setscheduler sys_sched_setscheduler sys_sched_setscheduler +157 common sched_getscheduler sys_sched_getscheduler sys_sched_getscheduler +158 common sched_yield sys_sched_yield sys_sched_yield +159 common sched_get_priority_max sys_sched_get_priority_max sys_sched_get_priority_max +160 common sched_get_priority_min sys_sched_get_priority_min sys_sched_get_priority_min +161 common sched_rr_get_interval sys_sched_rr_get_interval sys_sched_rr_get_interval_time32 +162 common nanosleep sys_nanosleep sys_nanosleep_time32 +163 common mremap sys_mremap sys_mremap +164 32 setresuid - sys_setresuid16 +165 32 getresuid - sys_getresuid16 +167 common query_module - - +168 common poll sys_poll sys_poll +169 common nfsservctl - - +170 32 setresgid - sys_setresgid16 +171 32 getresgid - sys_getresgid16 +172 common prctl sys_prctl sys_prctl +173 common rt_sigreturn sys_rt_sigreturn compat_sys_rt_sigreturn +174 common rt_sigaction sys_rt_sigaction compat_sys_rt_sigaction +175 common rt_sigprocmask sys_rt_sigprocmask compat_sys_rt_sigprocmask +176 common rt_sigpending sys_rt_sigpending compat_sys_rt_sigpending +177 common rt_sigtimedwait sys_rt_sigtimedwait compat_sys_rt_sigtimedwait_time32 +178 common rt_sigqueueinfo sys_rt_sigqueueinfo compat_sys_rt_sigqueueinfo +179 common rt_sigsuspend sys_rt_sigsuspend compat_sys_rt_sigsuspend +180 common pread64 sys_pread64 compat_sys_s390_pread64 +181 common pwrite64 sys_pwrite64 compat_sys_s390_pwrite64 +182 32 chown - sys_chown16 +183 common getcwd sys_getcwd sys_getcwd +184 common capget sys_capget sys_capget +185 common capset sys_capset sys_capset +186 common sigaltstack sys_sigaltstack compat_sys_sigaltstack +187 common sendfile sys_sendfile64 compat_sys_sendfile +188 common getpmsg - - +189 common putpmsg - - +190 common vfork sys_vfork sys_vfork +191 32 ugetrlimit - compat_sys_getrlimit +191 64 getrlimit sys_getrlimit - +192 32 mmap2 - compat_sys_s390_mmap2 +193 32 truncate64 - compat_sys_s390_truncate64 +194 32 ftruncate64 - compat_sys_s390_ftruncate64 +195 32 stat64 - compat_sys_s390_stat64 +196 32 lstat64 - compat_sys_s390_lstat64 +197 32 fstat64 - compat_sys_s390_fstat64 +198 32 lchown32 - sys_lchown +198 64 lchown sys_lchown - +199 32 getuid32 - sys_getuid +199 64 getuid sys_getuid - +200 32 getgid32 - sys_getgid +200 64 getgid sys_getgid - +201 32 geteuid32 - sys_geteuid +201 64 geteuid sys_geteuid - +202 32 getegid32 - sys_getegid +202 64 getegid sys_getegid - +203 32 setreuid32 - sys_setreuid +203 64 setreuid sys_setreuid - +204 32 setregid32 - sys_setregid +204 64 setregid sys_setregid - +205 32 getgroups32 - sys_getgroups +205 64 getgroups sys_getgroups - +206 32 setgroups32 - sys_setgroups +206 64 setgroups sys_setgroups - +207 32 fchown32 - sys_fchown +207 64 fchown sys_fchown - +208 32 setresuid32 - sys_setresuid +208 64 setresuid sys_setresuid - +209 32 getresuid32 - sys_getresuid +209 64 getresuid sys_getresuid - +210 32 setresgid32 - sys_setresgid +210 64 setresgid sys_setresgid - +211 32 getresgid32 - sys_getresgid +211 64 getresgid sys_getresgid - +212 32 chown32 - sys_chown +212 64 chown sys_chown - +213 32 setuid32 - sys_setuid +213 64 setuid sys_setuid - +214 32 setgid32 - sys_setgid +214 64 setgid sys_setgid - +215 32 setfsuid32 - sys_setfsuid +215 64 setfsuid sys_setfsuid - +216 32 setfsgid32 - sys_setfsgid +216 64 setfsgid sys_setfsgid - +217 common pivot_root sys_pivot_root sys_pivot_root +218 common mincore sys_mincore sys_mincore +219 common madvise sys_madvise sys_madvise +220 common getdents64 sys_getdents64 sys_getdents64 +221 32 fcntl64 - compat_sys_fcntl64 +222 common readahead sys_readahead compat_sys_s390_readahead +223 32 sendfile64 - compat_sys_sendfile64 +224 common setxattr sys_setxattr sys_setxattr +225 common lsetxattr sys_lsetxattr sys_lsetxattr +226 common fsetxattr sys_fsetxattr sys_fsetxattr +227 common getxattr sys_getxattr sys_getxattr +228 common lgetxattr sys_lgetxattr sys_lgetxattr +229 common fgetxattr sys_fgetxattr sys_fgetxattr +230 common listxattr sys_listxattr sys_listxattr +231 common llistxattr sys_llistxattr sys_llistxattr +232 common flistxattr sys_flistxattr sys_flistxattr +233 common removexattr sys_removexattr sys_removexattr +234 common lremovexattr sys_lremovexattr sys_lremovexattr +235 common fremovexattr sys_fremovexattr sys_fremovexattr +236 common gettid sys_gettid sys_gettid +237 common tkill sys_tkill sys_tkill +238 common futex sys_futex sys_futex_time32 +239 common sched_setaffinity sys_sched_setaffinity compat_sys_sched_setaffinity +240 common sched_getaffinity sys_sched_getaffinity compat_sys_sched_getaffinity +241 common tgkill sys_tgkill sys_tgkill +243 common io_setup sys_io_setup compat_sys_io_setup +244 common io_destroy sys_io_destroy sys_io_destroy +245 common io_getevents sys_io_getevents sys_io_getevents_time32 +246 common io_submit sys_io_submit compat_sys_io_submit +247 common io_cancel sys_io_cancel sys_io_cancel +248 common exit_group sys_exit_group sys_exit_group +249 common epoll_create sys_epoll_create sys_epoll_create +250 common epoll_ctl sys_epoll_ctl sys_epoll_ctl +251 common epoll_wait sys_epoll_wait sys_epoll_wait +252 common set_tid_address sys_set_tid_address sys_set_tid_address +253 common fadvise64 sys_fadvise64_64 compat_sys_s390_fadvise64 +254 common timer_create sys_timer_create compat_sys_timer_create +255 common timer_settime sys_timer_settime sys_timer_settime32 +256 common timer_gettime sys_timer_gettime sys_timer_gettime32 +257 common timer_getoverrun sys_timer_getoverrun sys_timer_getoverrun +258 common timer_delete sys_timer_delete sys_timer_delete +259 common clock_settime sys_clock_settime sys_clock_settime32 +260 common clock_gettime sys_clock_gettime sys_clock_gettime32 +261 common clock_getres sys_clock_getres sys_clock_getres_time32 +262 common clock_nanosleep sys_clock_nanosleep sys_clock_nanosleep_time32 +264 32 fadvise64_64 - compat_sys_s390_fadvise64_64 +265 common statfs64 sys_statfs64 compat_sys_statfs64 +266 common fstatfs64 sys_fstatfs64 compat_sys_fstatfs64 +267 common remap_file_pages sys_remap_file_pages sys_remap_file_pages +268 common mbind sys_mbind sys_mbind +269 common get_mempolicy sys_get_mempolicy sys_get_mempolicy +270 common set_mempolicy sys_set_mempolicy sys_set_mempolicy +271 common mq_open sys_mq_open compat_sys_mq_open +272 common mq_unlink sys_mq_unlink sys_mq_unlink +273 common mq_timedsend sys_mq_timedsend sys_mq_timedsend_time32 +274 common mq_timedreceive sys_mq_timedreceive sys_mq_timedreceive_time32 +275 common mq_notify sys_mq_notify compat_sys_mq_notify +276 common mq_getsetattr sys_mq_getsetattr compat_sys_mq_getsetattr +277 common kexec_load sys_kexec_load compat_sys_kexec_load +278 common add_key sys_add_key sys_add_key +279 common request_key sys_request_key sys_request_key +280 common keyctl sys_keyctl compat_sys_keyctl +281 common waitid sys_waitid compat_sys_waitid +282 common ioprio_set sys_ioprio_set sys_ioprio_set +283 common ioprio_get sys_ioprio_get sys_ioprio_get +284 common inotify_init sys_inotify_init sys_inotify_init +285 common inotify_add_watch sys_inotify_add_watch sys_inotify_add_watch +286 common inotify_rm_watch sys_inotify_rm_watch sys_inotify_rm_watch +287 common migrate_pages sys_migrate_pages sys_migrate_pages +288 common openat sys_openat compat_sys_openat +289 common mkdirat sys_mkdirat sys_mkdirat +290 common mknodat sys_mknodat sys_mknodat +291 common fchownat sys_fchownat sys_fchownat +292 common futimesat sys_futimesat sys_futimesat_time32 +293 32 fstatat64 - compat_sys_s390_fstatat64 +293 64 newfstatat sys_newfstatat - +294 common unlinkat sys_unlinkat sys_unlinkat +295 common renameat sys_renameat sys_renameat +296 common linkat sys_linkat sys_linkat +297 common symlinkat sys_symlinkat sys_symlinkat +298 common readlinkat sys_readlinkat sys_readlinkat +299 common fchmodat sys_fchmodat sys_fchmodat +300 common faccessat sys_faccessat sys_faccessat +301 common pselect6 sys_pselect6 compat_sys_pselect6_time32 +302 common ppoll sys_ppoll compat_sys_ppoll_time32 +303 common unshare sys_unshare sys_unshare +304 common set_robust_list sys_set_robust_list compat_sys_set_robust_list +305 common get_robust_list sys_get_robust_list compat_sys_get_robust_list +306 common splice sys_splice sys_splice +307 common sync_file_range sys_sync_file_range compat_sys_s390_sync_file_range +308 common tee sys_tee sys_tee +309 common vmsplice sys_vmsplice sys_vmsplice +310 common move_pages sys_move_pages sys_move_pages +311 common getcpu sys_getcpu sys_getcpu +312 common epoll_pwait sys_epoll_pwait compat_sys_epoll_pwait +313 common utimes sys_utimes sys_utimes_time32 +314 common fallocate sys_fallocate compat_sys_s390_fallocate +315 common utimensat sys_utimensat sys_utimensat_time32 +316 common signalfd sys_signalfd compat_sys_signalfd +317 common timerfd - - +318 common eventfd sys_eventfd sys_eventfd +319 common timerfd_create sys_timerfd_create sys_timerfd_create +320 common timerfd_settime sys_timerfd_settime sys_timerfd_settime32 +321 common timerfd_gettime sys_timerfd_gettime sys_timerfd_gettime32 +322 common signalfd4 sys_signalfd4 compat_sys_signalfd4 +323 common eventfd2 sys_eventfd2 sys_eventfd2 +324 common inotify_init1 sys_inotify_init1 sys_inotify_init1 +325 common pipe2 sys_pipe2 sys_pipe2 +326 common dup3 sys_dup3 sys_dup3 +327 common epoll_create1 sys_epoll_create1 sys_epoll_create1 +328 common preadv sys_preadv compat_sys_preadv +329 common pwritev sys_pwritev compat_sys_pwritev +330 common rt_tgsigqueueinfo sys_rt_tgsigqueueinfo compat_sys_rt_tgsigqueueinfo +331 common perf_event_open sys_perf_event_open sys_perf_event_open +332 common fanotify_init sys_fanotify_init sys_fanotify_init +333 common fanotify_mark sys_fanotify_mark compat_sys_fanotify_mark +334 common prlimit64 sys_prlimit64 sys_prlimit64 +335 common name_to_handle_at sys_name_to_handle_at sys_name_to_handle_at +336 common open_by_handle_at sys_open_by_handle_at compat_sys_open_by_handle_at +337 common clock_adjtime sys_clock_adjtime sys_clock_adjtime32 +338 common syncfs sys_syncfs sys_syncfs +339 common setns sys_setns sys_setns +340 common process_vm_readv sys_process_vm_readv sys_process_vm_readv +341 common process_vm_writev sys_process_vm_writev sys_process_vm_writev +342 common s390_runtime_instr sys_s390_runtime_instr sys_s390_runtime_instr +343 common kcmp sys_kcmp sys_kcmp +344 common finit_module sys_finit_module sys_finit_module +345 common sched_setattr sys_sched_setattr sys_sched_setattr +346 common sched_getattr sys_sched_getattr sys_sched_getattr +347 common renameat2 sys_renameat2 sys_renameat2 +348 common seccomp sys_seccomp sys_seccomp +349 common getrandom sys_getrandom sys_getrandom +350 common memfd_create sys_memfd_create sys_memfd_create +351 common bpf sys_bpf sys_bpf +352 common s390_pci_mmio_write sys_s390_pci_mmio_write sys_s390_pci_mmio_write +353 common s390_pci_mmio_read sys_s390_pci_mmio_read sys_s390_pci_mmio_read +354 common execveat sys_execveat compat_sys_execveat +355 common userfaultfd sys_userfaultfd sys_userfaultfd +356 common membarrier sys_membarrier sys_membarrier +357 common recvmmsg sys_recvmmsg compat_sys_recvmmsg_time32 +358 common sendmmsg sys_sendmmsg compat_sys_sendmmsg +359 common socket sys_socket sys_socket +360 common socketpair sys_socketpair sys_socketpair +361 common bind sys_bind sys_bind +362 common connect sys_connect sys_connect +363 common listen sys_listen sys_listen +364 common accept4 sys_accept4 sys_accept4 +365 common getsockopt sys_getsockopt sys_getsockopt +366 common setsockopt sys_setsockopt sys_setsockopt +367 common getsockname sys_getsockname sys_getsockname +368 common getpeername sys_getpeername sys_getpeername +369 common sendto sys_sendto sys_sendto +370 common sendmsg sys_sendmsg compat_sys_sendmsg +371 common recvfrom sys_recvfrom compat_sys_recvfrom +372 common recvmsg sys_recvmsg compat_sys_recvmsg +373 common shutdown sys_shutdown sys_shutdown +374 common mlock2 sys_mlock2 sys_mlock2 +375 common copy_file_range sys_copy_file_range sys_copy_file_range +376 common preadv2 sys_preadv2 compat_sys_preadv2 +377 common pwritev2 sys_pwritev2 compat_sys_pwritev2 +378 common s390_guarded_storage sys_s390_guarded_storage sys_s390_guarded_storage +379 common statx sys_statx sys_statx +380 common s390_sthyi sys_s390_sthyi sys_s390_sthyi +381 common kexec_file_load sys_kexec_file_load sys_kexec_file_load +382 common io_pgetevents sys_io_pgetevents compat_sys_io_pgetevents +383 common rseq sys_rseq sys_rseq +384 common pkey_mprotect sys_pkey_mprotect sys_pkey_mprotect +385 common pkey_alloc sys_pkey_alloc sys_pkey_alloc +386 common pkey_free sys_pkey_free sys_pkey_free +# room for arch specific syscalls +392 64 semtimedop sys_semtimedop - +393 common semget sys_semget sys_semget +394 common semctl sys_semctl compat_sys_semctl +395 common shmget sys_shmget sys_shmget +396 common shmctl sys_shmctl compat_sys_shmctl +397 common shmat sys_shmat compat_sys_shmat +398 common shmdt sys_shmdt sys_shmdt +399 common msgget sys_msgget sys_msgget +400 common msgsnd sys_msgsnd compat_sys_msgsnd +401 common msgrcv sys_msgrcv compat_sys_msgrcv +402 common msgctl sys_msgctl compat_sys_msgctl +403 32 clock_gettime64 - sys_clock_gettime +404 32 clock_settime64 - sys_clock_settime +405 32 clock_adjtime64 - sys_clock_adjtime +406 32 clock_getres_time64 - sys_clock_getres +407 32 clock_nanosleep_time64 - sys_clock_nanosleep +408 32 timer_gettime64 - sys_timer_gettime +409 32 timer_settime64 - sys_timer_settime +410 32 timerfd_gettime64 - sys_timerfd_gettime +411 32 timerfd_settime64 - sys_timerfd_settime +412 32 utimensat_time64 - sys_utimensat +413 32 pselect6_time64 - compat_sys_pselect6_time64 +414 32 ppoll_time64 - compat_sys_ppoll_time64 +416 32 io_pgetevents_time64 - sys_io_pgetevents +417 32 recvmmsg_time64 - compat_sys_recvmmsg_time64 +418 32 mq_timedsend_time64 - sys_mq_timedsend +419 32 mq_timedreceive_time64 - sys_mq_timedreceive +420 32 semtimedop_time64 - sys_semtimedop +421 32 rt_sigtimedwait_time64 - compat_sys_rt_sigtimedwait_time64 +422 32 futex_time64 - sys_futex +423 32 sched_rr_get_interval_time64 - sys_sched_rr_get_interval +424 common pidfd_send_signal sys_pidfd_send_signal sys_pidfd_send_signal +425 common io_uring_setup sys_io_uring_setup sys_io_uring_setup +426 common io_uring_enter sys_io_uring_enter sys_io_uring_enter +427 common io_uring_register sys_io_uring_register sys_io_uring_register +428 common open_tree sys_open_tree sys_open_tree +429 common move_mount sys_move_mount sys_move_mount +430 common fsopen sys_fsopen sys_fsopen +431 common fsconfig sys_fsconfig sys_fsconfig +432 common fsmount sys_fsmount sys_fsmount +433 common fspick sys_fspick sys_fspick +434 common pidfd_open sys_pidfd_open sys_pidfd_open +435 common clone3 sys_clone3 sys_clone3 +436 common close_range sys_close_range sys_close_range +437 common openat2 sys_openat2 sys_openat2 +438 common pidfd_getfd sys_pidfd_getfd sys_pidfd_getfd +439 common faccessat2 sys_faccessat2 sys_faccessat2 +440 common process_madvise sys_process_madvise sys_process_madvise +441 common epoll_pwait2 sys_epoll_pwait2 compat_sys_epoll_pwait2 +442 common mount_setattr sys_mount_setattr sys_mount_setattr +443 common quotactl_fd sys_quotactl_fd sys_quotactl_fd +444 common landlock_create_ruleset sys_landlock_create_ruleset sys_landlock_create_ruleset +445 common landlock_add_rule sys_landlock_add_rule sys_landlock_add_rule +446 common landlock_restrict_self sys_landlock_restrict_self sys_landlock_restrict_self +447 common memfd_secret sys_memfd_secret sys_memfd_secret +448 common process_mrelease sys_process_mrelease sys_process_mrelease +449 common futex_waitv sys_futex_waitv sys_futex_waitv +450 common set_mempolicy_home_node sys_set_mempolicy_home_node sys_set_mempolicy_home_node +451 common cachestat sys_cachestat sys_cachestat +452 common fchmodat2 sys_fchmodat2 sys_fchmodat2 diff --git a/arch/s390/kernel/syscalls/syscalltbl b/arch/s390/kernel/syscalls/syscalltbl new file mode 100755 index 0000000000..fbac1732f8 --- /dev/null +++ b/arch/s390/kernel/syscalls/syscalltbl @@ -0,0 +1,232 @@ +#!/bin/sh +# SPDX-License-Identifier: GPL-2.0 +# +# Generate system call table and header files +# +# Copyright IBM Corp. 2018 +# Author(s): Hendrik Brueckner <brueckner@linux.vnet.ibm.com> + +# +# File path to the system call table definition. +# You can set the path with the -i option. If omitted, +# system call table definitions are read from standard input. +# +SYSCALL_TBL="" + + +create_syscall_table_entries() +{ + local nr abi name entry64 entry32 _ignore + local temp=$(mktemp ${TMPDIR:-/tmp}/syscalltbl-common.XXXXXXXXX) + + ( + # + # Initialize with 0 to create an NI_SYSCALL for 0 + # + local prev_nr=0 prev_32=sys_ni_syscall prev_64=sys_ni_syscall + while read nr abi name entry64 entry32 _ignore; do + test x$entry32 = x- && entry32=sys_ni_syscall + test x$entry64 = x- && entry64=sys_ni_syscall + + if test $prev_nr -eq $nr; then + # + # Same syscall but different ABI, just update + # the respective entry point + # + case $abi in + 32) + prev_32=$entry32 + ;; + 64) + prev_64=$entry64 + ;; + esac + continue; + else + printf "%d\t%s\t%s\n" $prev_nr $prev_64 $prev_32 + fi + + prev_nr=$nr + prev_64=$entry64 + prev_32=$entry32 + done + printf "%d\t%s\t%s\n" $prev_nr $prev_64 $prev_32 + ) >> $temp + + # + # Check for duplicate syscall numbers + # + if ! cat $temp |cut -f1 |uniq -d 2>&1; then + echo "Error: generated system call table contains duplicate entries: $temp" >&2 + exit 1 + fi + + # + # Generate syscall table + # + prev_nr=0 + while read nr entry64 entry32; do + while test $prev_nr -lt $((nr - 1)); do + printf "NI_SYSCALL\n" + prev_nr=$((prev_nr + 1)) + done + if test x$entry64 = xsys_ni_syscall && + test x$entry32 = xsys_ni_syscall; then + printf "NI_SYSCALL\n" + else + printf "SYSCALL(%s,%s)\n" $entry64 $entry32 + fi + prev_nr=$nr + done < $temp + rm $temp +} + +generate_syscall_table() +{ + cat <<-EoHEADER + /* SPDX-License-Identifier: GPL-2.0 */ + /* + * Definitions for sys_call_table, each line represents an + * entry in the table in the form + * SYSCALL(64 bit syscall, 31 bit emulated syscall) + * + * This file is meant to be included from entry.S. + */ + + #define NI_SYSCALL SYSCALL(sys_ni_syscall,sys_ni_syscall) + +EoHEADER + grep -Ev '^(#|[[:blank:]]*$)' $SYSCALL_TBL \ + |sort -k1 -n \ + |create_syscall_table_entries +} + +create_header_defines() +{ + local nr abi name _ignore + + while read nr abi name _ignore; do + printf "#define __NR_%s %d\n" $name $nr + done +} + +normalize_fileguard() +{ + local fileguard="$1" + + echo "$1" |tr '[[:lower:]]' '[[:upper:]]' \ + |sed -e 's/[^A-Z0-9_]/_/g' -e 's/__/_/g' +} + +generate_syscall_header() +{ + local abis=$(echo "($1)" | tr ',' '|') + local filename="$2" + local fileguard suffix + + if test "$filename"; then + fileguard=$(normalize_fileguard "__UAPI_ASM_S390_$2") + else + case "$abis" in + *64*) suffix=64 ;; + *32*) suffix=32 ;; + esac + fileguard=$(normalize_fileguard "__UAPI_ASM_S390_SYSCALLS_$suffix") + fi + + cat <<-EoHEADER + /* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ + #ifndef ${fileguard} + #define ${fileguard} + +EoHEADER + + grep -E "^[[:digit:]]+[[:space:]]+${abis}" $SYSCALL_TBL \ + |sort -k1 -n \ + |create_header_defines + + cat <<-EoFOOTER + + #endif /* ${fileguard} */ +EoFOOTER +} + +__max_syscall_nr() +{ + local abis=$(echo "($1)" | tr ',' '|') + + grep -E "^[[:digit:]]+[[:space:]]+${abis}" $SYSCALL_TBL \ + |sed -ne 's/^\([[:digit:]]*\)[[:space:]].*/\1/p' \ + |sort -n \ + |tail -1 +} + + +generate_syscall_nr() +{ + local abis="$1" + local max_syscall_nr num_syscalls + + max_syscall_nr=$(__max_syscall_nr "$abis") + num_syscalls=$((max_syscall_nr + 1)) + + cat <<-EoHEADER + /* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ + #ifndef __ASM_S390_SYSCALLS_NR + #define __ASM_S390_SYSCALLS_NR + + #define NR_syscalls ${num_syscalls} + + #endif /* __ASM_S390_SYSCALLS_NR */ +EoHEADER +} + + +# +# Parse command line arguments +# +do_syscall_header="" +do_syscall_table="" +do_syscall_nr="" +output_file="" +abi_list="common,64" +filename="" +while getopts ":HNSXi:a:f:" arg; do + case $arg in + a) + abi_list="$OPTARG" + ;; + i) + SYSCALL_TBL="$OPTARG" + ;; + f) + filename=${OPTARG##*/} + ;; + H) + do_syscall_header=1 + ;; + N) + do_syscall_nr=1 + ;; + S) + do_syscall_table=1 + ;; + X) + set -x + ;; + :) + echo "Missing argument for -$OPTARG" >&2 + exit 1 + ;; + \?) + echo "Invalid option specified" >&2 + exit 1 + ;; + esac +done + +test "$do_syscall_header" && generate_syscall_header "$abi_list" "$filename" +test "$do_syscall_table" && generate_syscall_table +test "$do_syscall_nr" && generate_syscall_nr "$abi_list" + +exit 0 diff --git a/arch/s390/kernel/sysinfo.c b/arch/s390/kernel/sysinfo.c new file mode 100644 index 0000000000..b5e364358c --- /dev/null +++ b/arch/s390/kernel/sysinfo.c @@ -0,0 +1,570 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2001, 2009 + * Author(s): Ulrich Weigand <Ulrich.Weigand@de.ibm.com>, + * Martin Schwidefsky <schwidefsky@de.ibm.com>, + */ + +#include <linux/debugfs.h> +#include <linux/kernel.h> +#include <linux/mm.h> +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <linux/init.h> +#include <linux/delay.h> +#include <linux/export.h> +#include <linux/slab.h> +#include <asm/asm-extable.h> +#include <asm/ebcdic.h> +#include <asm/debug.h> +#include <asm/sysinfo.h> +#include <asm/cpcmd.h> +#include <asm/topology.h> +#include <asm/fpu/api.h> + +int topology_max_mnest; + +static inline int __stsi(void *sysinfo, int fc, int sel1, int sel2, int *lvl) +{ + int r0 = (fc << 28) | sel1; + int rc = 0; + + asm volatile( + " lr 0,%[r0]\n" + " lr 1,%[r1]\n" + " stsi 0(%[sysinfo])\n" + "0: jz 2f\n" + "1: lhi %[rc],%[retval]\n" + "2: lr %[r0],0\n" + EX_TABLE(0b, 1b) + : [r0] "+d" (r0), [rc] "+d" (rc) + : [r1] "d" (sel2), + [sysinfo] "a" (sysinfo), + [retval] "K" (-EOPNOTSUPP) + : "cc", "0", "1", "memory"); + *lvl = ((unsigned int) r0) >> 28; + return rc; +} + +/* + * stsi - store system information + * + * Returns the current configuration level if function code 0 was specified. + * Otherwise returns 0 on success or a negative value on error. + */ +int stsi(void *sysinfo, int fc, int sel1, int sel2) +{ + int lvl, rc; + + rc = __stsi(sysinfo, fc, sel1, sel2, &lvl); + if (rc) + return rc; + return fc ? 0 : lvl; +} +EXPORT_SYMBOL(stsi); + +#ifdef CONFIG_PROC_FS + +static bool convert_ext_name(unsigned char encoding, char *name, size_t len) +{ + switch (encoding) { + case 1: /* EBCDIC */ + EBCASC(name, len); + break; + case 2: /* UTF-8 */ + break; + default: + return false; + } + return true; +} + +static void stsi_1_1_1(struct seq_file *m, struct sysinfo_1_1_1 *info) +{ + int i; + + if (stsi(info, 1, 1, 1)) + return; + EBCASC(info->manufacturer, sizeof(info->manufacturer)); + EBCASC(info->type, sizeof(info->type)); + EBCASC(info->model, sizeof(info->model)); + EBCASC(info->sequence, sizeof(info->sequence)); + EBCASC(info->plant, sizeof(info->plant)); + EBCASC(info->model_capacity, sizeof(info->model_capacity)); + EBCASC(info->model_perm_cap, sizeof(info->model_perm_cap)); + EBCASC(info->model_temp_cap, sizeof(info->model_temp_cap)); + seq_printf(m, "Manufacturer: %-16.16s\n", info->manufacturer); + seq_printf(m, "Type: %-4.4s\n", info->type); + if (info->lic) + seq_printf(m, "LIC Identifier: %016lx\n", info->lic); + /* + * Sigh: the model field has been renamed with System z9 + * to model_capacity and a new model field has been added + * after the plant field. To avoid confusing older programs + * the "Model:" prints "model_capacity model" or just + * "model_capacity" if the model string is empty . + */ + seq_printf(m, "Model: %-16.16s", info->model_capacity); + if (info->model[0] != '\0') + seq_printf(m, " %-16.16s", info->model); + seq_putc(m, '\n'); + seq_printf(m, "Sequence Code: %-16.16s\n", info->sequence); + seq_printf(m, "Plant: %-4.4s\n", info->plant); + seq_printf(m, "Model Capacity: %-16.16s %08u\n", + info->model_capacity, info->model_cap_rating); + if (info->model_perm_cap_rating) + seq_printf(m, "Model Perm. Capacity: %-16.16s %08u\n", + info->model_perm_cap, + info->model_perm_cap_rating); + if (info->model_temp_cap_rating) + seq_printf(m, "Model Temp. Capacity: %-16.16s %08u\n", + info->model_temp_cap, + info->model_temp_cap_rating); + if (info->ncr) + seq_printf(m, "Nominal Cap. Rating: %08u\n", info->ncr); + if (info->npr) + seq_printf(m, "Nominal Perm. Rating: %08u\n", info->npr); + if (info->ntr) + seq_printf(m, "Nominal Temp. Rating: %08u\n", info->ntr); + if (info->cai) { + seq_printf(m, "Capacity Adj. Ind.: %d\n", info->cai); + seq_printf(m, "Capacity Ch. Reason: %d\n", info->ccr); + seq_printf(m, "Capacity Transient: %d\n", info->t); + } + if (info->p) { + for (i = 1; i <= ARRAY_SIZE(info->typepct); i++) { + seq_printf(m, "Type %d Percentage: %d\n", + i, info->typepct[i - 1]); + } + } +} + +static void stsi_15_1_x(struct seq_file *m, struct sysinfo_15_1_x *info) +{ + int i; + + seq_putc(m, '\n'); + if (!MACHINE_HAS_TOPOLOGY) + return; + if (stsi(info, 15, 1, topology_max_mnest)) + return; + seq_printf(m, "CPU Topology HW: "); + for (i = 0; i < TOPOLOGY_NR_MAG; i++) + seq_printf(m, " %d", info->mag[i]); + seq_putc(m, '\n'); +#ifdef CONFIG_SCHED_TOPOLOGY + store_topology(info); + seq_printf(m, "CPU Topology SW: "); + for (i = 0; i < TOPOLOGY_NR_MAG; i++) + seq_printf(m, " %d", info->mag[i]); + seq_putc(m, '\n'); +#endif +} + +static void stsi_1_2_2(struct seq_file *m, struct sysinfo_1_2_2 *info) +{ + struct sysinfo_1_2_2_extension *ext; + int i; + + if (stsi(info, 1, 2, 2)) + return; + ext = (struct sysinfo_1_2_2_extension *) + ((unsigned long) info + info->acc_offset); + seq_printf(m, "CPUs Total: %d\n", info->cpus_total); + seq_printf(m, "CPUs Configured: %d\n", info->cpus_configured); + seq_printf(m, "CPUs Standby: %d\n", info->cpus_standby); + seq_printf(m, "CPUs Reserved: %d\n", info->cpus_reserved); + if (info->mt_installed) { + seq_printf(m, "CPUs G-MTID: %d\n", info->mt_gtid); + seq_printf(m, "CPUs S-MTID: %d\n", info->mt_stid); + } + /* + * Sigh 2. According to the specification the alternate + * capability field is a 32 bit floating point number + * if the higher order 8 bits are not zero. Printing + * a floating point number in the kernel is a no-no, + * always print the number as 32 bit unsigned integer. + * The user-space needs to know about the strange + * encoding of the alternate cpu capability. + */ + seq_printf(m, "Capability: %u", info->capability); + if (info->format == 1) + seq_printf(m, " %u", ext->alt_capability); + seq_putc(m, '\n'); + if (info->nominal_cap) + seq_printf(m, "Nominal Capability: %d\n", info->nominal_cap); + if (info->secondary_cap) + seq_printf(m, "Secondary Capability: %d\n", info->secondary_cap); + for (i = 2; i <= info->cpus_total; i++) { + seq_printf(m, "Adjustment %02d-way: %u", + i, info->adjustment[i-2]); + if (info->format == 1) + seq_printf(m, " %u", ext->alt_adjustment[i-2]); + seq_putc(m, '\n'); + } +} + +static void stsi_2_2_2(struct seq_file *m, struct sysinfo_2_2_2 *info) +{ + if (stsi(info, 2, 2, 2)) + return; + EBCASC(info->name, sizeof(info->name)); + seq_putc(m, '\n'); + seq_printf(m, "LPAR Number: %d\n", info->lpar_number); + seq_printf(m, "LPAR Characteristics: "); + if (info->characteristics & LPAR_CHAR_DEDICATED) + seq_printf(m, "Dedicated "); + if (info->characteristics & LPAR_CHAR_SHARED) + seq_printf(m, "Shared "); + if (info->characteristics & LPAR_CHAR_LIMITED) + seq_printf(m, "Limited "); + seq_putc(m, '\n'); + seq_printf(m, "LPAR Name: %-8.8s\n", info->name); + seq_printf(m, "LPAR Adjustment: %d\n", info->caf); + seq_printf(m, "LPAR CPUs Total: %d\n", info->cpus_total); + seq_printf(m, "LPAR CPUs Configured: %d\n", info->cpus_configured); + seq_printf(m, "LPAR CPUs Standby: %d\n", info->cpus_standby); + seq_printf(m, "LPAR CPUs Reserved: %d\n", info->cpus_reserved); + seq_printf(m, "LPAR CPUs Dedicated: %d\n", info->cpus_dedicated); + seq_printf(m, "LPAR CPUs Shared: %d\n", info->cpus_shared); + if (info->mt_installed) { + seq_printf(m, "LPAR CPUs G-MTID: %d\n", info->mt_gtid); + seq_printf(m, "LPAR CPUs S-MTID: %d\n", info->mt_stid); + seq_printf(m, "LPAR CPUs PS-MTID: %d\n", info->mt_psmtid); + } + if (convert_ext_name(info->vsne, info->ext_name, sizeof(info->ext_name))) { + seq_printf(m, "LPAR Extended Name: %-.256s\n", info->ext_name); + seq_printf(m, "LPAR UUID: %pUb\n", &info->uuid); + } +} + +static void print_ext_name(struct seq_file *m, int lvl, + struct sysinfo_3_2_2 *info) +{ + size_t len = sizeof(info->ext_names[lvl]); + + if (!convert_ext_name(info->vm[lvl].evmne, info->ext_names[lvl], len)) + return; + seq_printf(m, "VM%02d Extended Name: %-.256s\n", lvl, + info->ext_names[lvl]); +} + +static void print_uuid(struct seq_file *m, int i, struct sysinfo_3_2_2 *info) +{ + if (uuid_is_null(&info->vm[i].uuid)) + return; + seq_printf(m, "VM%02d UUID: %pUb\n", i, &info->vm[i].uuid); +} + +static void stsi_3_2_2(struct seq_file *m, struct sysinfo_3_2_2 *info) +{ + int i; + + if (stsi(info, 3, 2, 2)) + return; + for (i = 0; i < info->count; i++) { + EBCASC(info->vm[i].name, sizeof(info->vm[i].name)); + EBCASC(info->vm[i].cpi, sizeof(info->vm[i].cpi)); + seq_putc(m, '\n'); + seq_printf(m, "VM%02d Name: %-8.8s\n", i, info->vm[i].name); + seq_printf(m, "VM%02d Control Program: %-16.16s\n", i, info->vm[i].cpi); + seq_printf(m, "VM%02d Adjustment: %d\n", i, info->vm[i].caf); + seq_printf(m, "VM%02d CPUs Total: %d\n", i, info->vm[i].cpus_total); + seq_printf(m, "VM%02d CPUs Configured: %d\n", i, info->vm[i].cpus_configured); + seq_printf(m, "VM%02d CPUs Standby: %d\n", i, info->vm[i].cpus_standby); + seq_printf(m, "VM%02d CPUs Reserved: %d\n", i, info->vm[i].cpus_reserved); + print_ext_name(m, i, info); + print_uuid(m, i, info); + } +} + +static int sysinfo_show(struct seq_file *m, void *v) +{ + void *info = (void *)get_zeroed_page(GFP_KERNEL); + int level; + + if (!info) + return 0; + level = stsi(NULL, 0, 0, 0); + if (level >= 1) + stsi_1_1_1(m, info); + if (level >= 1) + stsi_15_1_x(m, info); + if (level >= 1) + stsi_1_2_2(m, info); + if (level >= 2) + stsi_2_2_2(m, info); + if (level >= 3) + stsi_3_2_2(m, info); + free_page((unsigned long)info); + return 0; +} + +static int __init sysinfo_create_proc(void) +{ + proc_create_single("sysinfo", 0444, NULL, sysinfo_show); + return 0; +} +device_initcall(sysinfo_create_proc); + +#endif /* CONFIG_PROC_FS */ + +/* + * Service levels interface. + */ + +static DECLARE_RWSEM(service_level_sem); +static LIST_HEAD(service_level_list); + +int register_service_level(struct service_level *slr) +{ + struct service_level *ptr; + + down_write(&service_level_sem); + list_for_each_entry(ptr, &service_level_list, list) + if (ptr == slr) { + up_write(&service_level_sem); + return -EEXIST; + } + list_add_tail(&slr->list, &service_level_list); + up_write(&service_level_sem); + return 0; +} +EXPORT_SYMBOL(register_service_level); + +int unregister_service_level(struct service_level *slr) +{ + struct service_level *ptr, *next; + int rc = -ENOENT; + + down_write(&service_level_sem); + list_for_each_entry_safe(ptr, next, &service_level_list, list) { + if (ptr != slr) + continue; + list_del(&ptr->list); + rc = 0; + break; + } + up_write(&service_level_sem); + return rc; +} +EXPORT_SYMBOL(unregister_service_level); + +static void *service_level_start(struct seq_file *m, loff_t *pos) +{ + down_read(&service_level_sem); + return seq_list_start(&service_level_list, *pos); +} + +static void *service_level_next(struct seq_file *m, void *p, loff_t *pos) +{ + return seq_list_next(p, &service_level_list, pos); +} + +static void service_level_stop(struct seq_file *m, void *p) +{ + up_read(&service_level_sem); +} + +static int service_level_show(struct seq_file *m, void *p) +{ + struct service_level *slr; + + slr = list_entry(p, struct service_level, list); + slr->seq_print(m, slr); + return 0; +} + +static const struct seq_operations service_level_seq_ops = { + .start = service_level_start, + .next = service_level_next, + .stop = service_level_stop, + .show = service_level_show +}; + +static void service_level_vm_print(struct seq_file *m, + struct service_level *slr) +{ + char *query_buffer, *str; + + query_buffer = kmalloc(1024, GFP_KERNEL | GFP_DMA); + if (!query_buffer) + return; + cpcmd("QUERY CPLEVEL", query_buffer, 1024, NULL); + str = strchr(query_buffer, '\n'); + if (str) + *str = 0; + seq_printf(m, "VM: %s\n", query_buffer); + kfree(query_buffer); +} + +static struct service_level service_level_vm = { + .seq_print = service_level_vm_print +}; + +static __init int create_proc_service_level(void) +{ + proc_create_seq("service_levels", 0, NULL, &service_level_seq_ops); + if (MACHINE_IS_VM) + register_service_level(&service_level_vm); + return 0; +} +subsys_initcall(create_proc_service_level); + +/* + * CPU capability might have changed. Therefore recalculate loops_per_jiffy. + */ +void s390_adjust_jiffies(void) +{ + struct sysinfo_1_2_2 *info; + unsigned long capability; + struct kernel_fpu fpu; + + info = (void *) get_zeroed_page(GFP_KERNEL); + if (!info) + return; + + if (stsi(info, 1, 2, 2) == 0) { + /* + * Major sigh. The cpu capability encoding is "special". + * If the first 9 bits of info->capability are 0 then it + * is a 32 bit unsigned integer in the range 0 .. 2^23. + * If the first 9 bits are != 0 then it is a 32 bit float. + * In addition a lower value indicates a proportionally + * higher cpu capacity. Bogomips are the other way round. + * To get to a halfway suitable number we divide 1e7 + * by the cpu capability number. Yes, that means a floating + * point division .. + */ + kernel_fpu_begin(&fpu, KERNEL_FPR); + asm volatile( + " sfpc %3\n" + " l %0,%1\n" + " tmlh %0,0xff80\n" + " jnz 0f\n" + " cefbr %%f2,%0\n" + " j 1f\n" + "0: le %%f2,%1\n" + "1: cefbr %%f0,%2\n" + " debr %%f0,%%f2\n" + " cgebr %0,5,%%f0\n" + : "=&d" (capability) + : "Q" (info->capability), "d" (10000000), "d" (0) + : "cc" + ); + kernel_fpu_end(&fpu, KERNEL_FPR); + } else + /* + * Really old machine without stsi block for basic + * cpu information. Report 42.0 bogomips. + */ + capability = 42; + loops_per_jiffy = capability * (500000/HZ); + free_page((unsigned long) info); +} + +/* + * calibrate the delay loop + */ +void calibrate_delay(void) +{ + s390_adjust_jiffies(); + /* Print the good old Bogomips line .. */ + printk(KERN_DEBUG "Calibrating delay loop (skipped)... " + "%lu.%02lu BogoMIPS preset\n", loops_per_jiffy/(500000/HZ), + (loops_per_jiffy/(5000/HZ)) % 100); +} + +#ifdef CONFIG_DEBUG_FS + +#define STSI_FILE(fc, s1, s2) \ +static int stsi_open_##fc##_##s1##_##s2(struct inode *inode, struct file *file)\ +{ \ + file->private_data = (void *) get_zeroed_page(GFP_KERNEL); \ + if (!file->private_data) \ + return -ENOMEM; \ + if (stsi(file->private_data, fc, s1, s2)) { \ + free_page((unsigned long)file->private_data); \ + file->private_data = NULL; \ + return -EACCES; \ + } \ + return nonseekable_open(inode, file); \ +} \ + \ +static const struct file_operations stsi_##fc##_##s1##_##s2##_fs_ops = { \ + .open = stsi_open_##fc##_##s1##_##s2, \ + .release = stsi_release, \ + .read = stsi_read, \ + .llseek = no_llseek, \ +}; + +static int stsi_release(struct inode *inode, struct file *file) +{ + free_page((unsigned long)file->private_data); + return 0; +} + +static ssize_t stsi_read(struct file *file, char __user *buf, size_t size, loff_t *ppos) +{ + return simple_read_from_buffer(buf, size, ppos, file->private_data, PAGE_SIZE); +} + +STSI_FILE( 1, 1, 1); +STSI_FILE( 1, 2, 1); +STSI_FILE( 1, 2, 2); +STSI_FILE( 2, 2, 1); +STSI_FILE( 2, 2, 2); +STSI_FILE( 3, 2, 2); +STSI_FILE(15, 1, 2); +STSI_FILE(15, 1, 3); +STSI_FILE(15, 1, 4); +STSI_FILE(15, 1, 5); +STSI_FILE(15, 1, 6); + +struct stsi_file { + const struct file_operations *fops; + char *name; +}; + +static struct stsi_file stsi_file[] __initdata = { + {.fops = &stsi_1_1_1_fs_ops, .name = "1_1_1"}, + {.fops = &stsi_1_2_1_fs_ops, .name = "1_2_1"}, + {.fops = &stsi_1_2_2_fs_ops, .name = "1_2_2"}, + {.fops = &stsi_2_2_1_fs_ops, .name = "2_2_1"}, + {.fops = &stsi_2_2_2_fs_ops, .name = "2_2_2"}, + {.fops = &stsi_3_2_2_fs_ops, .name = "3_2_2"}, + {.fops = &stsi_15_1_2_fs_ops, .name = "15_1_2"}, + {.fops = &stsi_15_1_3_fs_ops, .name = "15_1_3"}, + {.fops = &stsi_15_1_4_fs_ops, .name = "15_1_4"}, + {.fops = &stsi_15_1_5_fs_ops, .name = "15_1_5"}, + {.fops = &stsi_15_1_6_fs_ops, .name = "15_1_6"}, +}; + +static u8 stsi_0_0_0; + +static __init int stsi_init_debugfs(void) +{ + struct dentry *stsi_root; + struct stsi_file *sf; + int lvl, i; + + stsi_root = debugfs_create_dir("stsi", arch_debugfs_dir); + lvl = stsi(NULL, 0, 0, 0); + if (lvl > 0) + stsi_0_0_0 = lvl; + debugfs_create_u8("0_0_0", 0400, stsi_root, &stsi_0_0_0); + for (i = 0; i < ARRAY_SIZE(stsi_file); i++) { + sf = &stsi_file[i]; + debugfs_create_file(sf->name, 0400, stsi_root, NULL, sf->fops); + } + if (IS_ENABLED(CONFIG_SCHED_TOPOLOGY) && MACHINE_HAS_TOPOLOGY) { + char link_to[10]; + + sprintf(link_to, "15_1_%d", topology_mnest_limit()); + debugfs_create_symlink("topology", stsi_root, link_to); + } + return 0; +} +device_initcall(stsi_init_debugfs); + +#endif /* CONFIG_DEBUG_FS */ diff --git a/arch/s390/kernel/text_amode31.S b/arch/s390/kernel/text_amode31.S new file mode 100644 index 0000000000..14c6d25c03 --- /dev/null +++ b/arch/s390/kernel/text_amode31.S @@ -0,0 +1,159 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Code that needs to run below 2 GB. + * + * Copyright IBM Corp. 2019 + */ + +#include <linux/linkage.h> +#include <asm/asm-extable.h> +#include <asm/errno.h> +#include <asm/sigp.h> + + .section .amode31.text,"ax" +/* + * Simplified version of expoline thunk. The normal thunks can not be used here, + * because they might be more than 2 GB away, and not reachable by the relative + * branch. No comdat, exrl, etc. optimizations used here, because it only + * affects a few functions that are not performance-relevant. + */ + .macro BR_EX_AMODE31_r14 + larl %r1,0f + ex 0,0(%r1) + j . +0: br %r14 + .endm + +/* + * int _diag14_amode31(unsigned long rx, unsigned long ry1, unsigned long subcode) + */ +SYM_FUNC_START(_diag14_amode31) + lgr %r1,%r2 + lgr %r2,%r3 + lgr %r3,%r4 + lhi %r5,-EIO + sam31 + diag %r1,%r2,0x14 +.Ldiag14_ex: + ipm %r5 + srl %r5,28 +.Ldiag14_fault: + sam64 + lgfr %r2,%r5 + BR_EX_AMODE31_r14 + EX_TABLE_AMODE31(.Ldiag14_ex, .Ldiag14_fault) +SYM_FUNC_END(_diag14_amode31) + +/* + * int _diag210_amode31(struct diag210 *addr) + */ +SYM_FUNC_START(_diag210_amode31) + lgr %r1,%r2 + lhi %r2,-1 + sam31 + diag %r1,%r0,0x210 +.Ldiag210_ex: + ipm %r2 + srl %r2,28 +.Ldiag210_fault: + sam64 + lgfr %r2,%r2 + BR_EX_AMODE31_r14 + EX_TABLE_AMODE31(.Ldiag210_ex, .Ldiag210_fault) +SYM_FUNC_END(_diag210_amode31) + +/* + * int diag8c(struct diag8c *addr, struct ccw_dev_id *devno, size_t len) +*/ +SYM_FUNC_START(_diag8c_amode31) + llgf %r3,0(%r3) + sam31 + diag %r2,%r4,0x8c +.Ldiag8c_ex: + sam64 + lgfr %r2,%r3 + BR_EX_AMODE31_r14 + EX_TABLE_AMODE31(.Ldiag8c_ex, .Ldiag8c_ex) +SYM_FUNC_END(_diag8c_amode31) +/* + * int _diag26c_amode31(void *req, void *resp, enum diag26c_sc subcode) + */ +SYM_FUNC_START(_diag26c_amode31) + lghi %r5,-EOPNOTSUPP + sam31 + diag %r2,%r4,0x26c +.Ldiag26c_ex: + sam64 + lgfr %r2,%r5 + BR_EX_AMODE31_r14 + EX_TABLE_AMODE31(.Ldiag26c_ex, .Ldiag26c_ex) +SYM_FUNC_END(_diag26c_amode31) + +/* + * void _diag0c_amode31(struct hypfs_diag0c_entry *entry) + */ +SYM_FUNC_START(_diag0c_amode31) + sam31 + diag %r2,%r2,0x0c + sam64 + BR_EX_AMODE31_r14 +SYM_FUNC_END(_diag0c_amode31) + +/* + * void _diag308_reset_amode31(void) + * + * Calls diag 308 subcode 1 and continues execution + */ +SYM_FUNC_START(_diag308_reset_amode31) + larl %r4,ctlregs # Save control registers + stctg %c0,%c15,0(%r4) + lg %r2,0(%r4) # Disable lowcore protection + nilh %r2,0xefff + larl %r4,ctlreg0 + stg %r2,0(%r4) + lctlg %c0,%c0,0(%r4) + larl %r4,fpctl # Floating point control register + stfpc 0(%r4) + larl %r4,prefix # Save prefix register + stpx 0(%r4) + larl %r4,prefix_zero # Set prefix register to 0 + spx 0(%r4) + larl %r4,continue_psw # Save PSW flags + epsw %r2,%r3 + stm %r2,%r3,0(%r4) + larl %r4,.Lrestart_part2 # Setup restart PSW at absolute 0 + larl %r3,restart_diag308_psw + og %r4,0(%r3) # Save PSW + lghi %r3,0 + sturg %r4,%r3 # Use sturg, because of large pages + lghi %r1,1 + lghi %r0,0 + diag %r0,%r1,0x308 +.Lrestart_part2: + lhi %r0,0 # Load r0 with zero + lhi %r1,2 # Use mode 2 = ESAME (dump) + sigp %r1,%r0,SIGP_SET_ARCHITECTURE # Switch to ESAME mode + sam64 # Switch to 64 bit addressing mode + larl %r4,ctlregs # Restore control registers + lctlg %c0,%c15,0(%r4) + larl %r4,fpctl # Restore floating point ctl register + lfpc 0(%r4) + larl %r4,prefix # Restore prefix register + spx 0(%r4) + larl %r4,continue_psw # Restore PSW flags + larl %r2,.Lcontinue + stg %r2,8(%r4) + lpswe 0(%r4) +.Lcontinue: + BR_EX_AMODE31_r14 +SYM_FUNC_END(_diag308_reset_amode31) + + .section .amode31.data,"aw",@progbits + .balign 8 +SYM_DATA_LOCAL(restart_diag308_psw, .long 0x00080000,0x80000000) +SYM_DATA_LOCAL(continue_psw, .quad 0,0) +SYM_DATA_LOCAL(ctlreg0, .quad 0) +SYM_DATA_LOCAL(ctlregs, .fill 16,8,0) +SYM_DATA_LOCAL(fpctl, .long 0) +SYM_DATA_LOCAL(prefix, .long 0) +SYM_DATA_LOCAL(prefix_zero, .long 0) diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c new file mode 100644 index 0000000000..d34d3548c0 --- /dev/null +++ b/arch/s390/kernel/time.c @@ -0,0 +1,944 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Time of day based timer functions. + * + * S390 version + * Copyright IBM Corp. 1999, 2008 + * Author(s): Hartmut Penner (hp@de.ibm.com), + * Martin Schwidefsky (schwidefsky@de.ibm.com), + * Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com) + * + * Derived from "arch/i386/kernel/time.c" + * Copyright (C) 1991, 1992, 1995 Linus Torvalds + */ + +#define KMSG_COMPONENT "time" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/kernel_stat.h> +#include <linux/errno.h> +#include <linux/export.h> +#include <linux/sched.h> +#include <linux/sched/clock.h> +#include <linux/kernel.h> +#include <linux/param.h> +#include <linux/string.h> +#include <linux/mm.h> +#include <linux/interrupt.h> +#include <linux/cpu.h> +#include <linux/stop_machine.h> +#include <linux/time.h> +#include <linux/device.h> +#include <linux/delay.h> +#include <linux/init.h> +#include <linux/smp.h> +#include <linux/types.h> +#include <linux/profile.h> +#include <linux/timex.h> +#include <linux/notifier.h> +#include <linux/timekeeper_internal.h> +#include <linux/clockchips.h> +#include <linux/gfp.h> +#include <linux/kprobes.h> +#include <linux/uaccess.h> +#include <vdso/vsyscall.h> +#include <vdso/clocksource.h> +#include <vdso/helpers.h> +#include <asm/facility.h> +#include <asm/delay.h> +#include <asm/div64.h> +#include <asm/vdso.h> +#include <asm/irq.h> +#include <asm/irq_regs.h> +#include <asm/vtimer.h> +#include <asm/stp.h> +#include <asm/cio.h> +#include "entry.h" + +union tod_clock tod_clock_base __section(".data"); +EXPORT_SYMBOL_GPL(tod_clock_base); + +u64 clock_comparator_max = -1ULL; +EXPORT_SYMBOL_GPL(clock_comparator_max); + +static DEFINE_PER_CPU(struct clock_event_device, comparators); + +ATOMIC_NOTIFIER_HEAD(s390_epoch_delta_notifier); +EXPORT_SYMBOL(s390_epoch_delta_notifier); + +unsigned char ptff_function_mask[16]; + +static unsigned long lpar_offset; +static unsigned long initial_leap_seconds; +static unsigned long tod_steering_end; +static long tod_steering_delta; + +/* + * Get time offsets with PTFF + */ +void __init time_early_init(void) +{ + struct ptff_qto qto; + struct ptff_qui qui; + int cs; + + /* Initialize TOD steering parameters */ + tod_steering_end = tod_clock_base.tod; + for (cs = 0; cs < CS_BASES; cs++) + vdso_data[cs].arch_data.tod_steering_end = tod_steering_end; + + if (!test_facility(28)) + return; + + ptff(&ptff_function_mask, sizeof(ptff_function_mask), PTFF_QAF); + + /* get LPAR offset */ + if (ptff_query(PTFF_QTO) && ptff(&qto, sizeof(qto), PTFF_QTO) == 0) + lpar_offset = qto.tod_epoch_difference; + + /* get initial leap seconds */ + if (ptff_query(PTFF_QUI) && ptff(&qui, sizeof(qui), PTFF_QUI) == 0) + initial_leap_seconds = (unsigned long) + ((long) qui.old_leap * 4096000000L); +} + +unsigned long long noinstr sched_clock_noinstr(void) +{ + return tod_to_ns(__get_tod_clock_monotonic()); +} + +/* + * Scheduler clock - returns current time in nanosec units. + */ +unsigned long long notrace sched_clock(void) +{ + return tod_to_ns(get_tod_clock_monotonic()); +} +NOKPROBE_SYMBOL(sched_clock); + +static void ext_to_timespec64(union tod_clock *clk, struct timespec64 *xt) +{ + unsigned long rem, sec, nsec; + + sec = clk->us; + rem = do_div(sec, 1000000); + nsec = ((clk->sus + (rem << 12)) * 125) >> 9; + xt->tv_sec = sec; + xt->tv_nsec = nsec; +} + +void clock_comparator_work(void) +{ + struct clock_event_device *cd; + + S390_lowcore.clock_comparator = clock_comparator_max; + cd = this_cpu_ptr(&comparators); + cd->event_handler(cd); +} + +static int s390_next_event(unsigned long delta, + struct clock_event_device *evt) +{ + S390_lowcore.clock_comparator = get_tod_clock() + delta; + set_clock_comparator(S390_lowcore.clock_comparator); + return 0; +} + +/* + * Set up lowcore and control register of the current cpu to + * enable TOD clock and clock comparator interrupts. + */ +void init_cpu_timer(void) +{ + struct clock_event_device *cd; + int cpu; + + S390_lowcore.clock_comparator = clock_comparator_max; + set_clock_comparator(S390_lowcore.clock_comparator); + + cpu = smp_processor_id(); + cd = &per_cpu(comparators, cpu); + cd->name = "comparator"; + cd->features = CLOCK_EVT_FEAT_ONESHOT; + cd->mult = 16777; + cd->shift = 12; + cd->min_delta_ns = 1; + cd->min_delta_ticks = 1; + cd->max_delta_ns = LONG_MAX; + cd->max_delta_ticks = ULONG_MAX; + cd->rating = 400; + cd->cpumask = cpumask_of(cpu); + cd->set_next_event = s390_next_event; + + clockevents_register_device(cd); + + /* Enable clock comparator timer interrupt. */ + __ctl_set_bit(0,11); + + /* Always allow the timing alert external interrupt. */ + __ctl_set_bit(0, 4); +} + +static void clock_comparator_interrupt(struct ext_code ext_code, + unsigned int param32, + unsigned long param64) +{ + inc_irq_stat(IRQEXT_CLK); + if (S390_lowcore.clock_comparator == clock_comparator_max) + set_clock_comparator(S390_lowcore.clock_comparator); +} + +static void stp_timing_alert(struct stp_irq_parm *); + +static void timing_alert_interrupt(struct ext_code ext_code, + unsigned int param32, unsigned long param64) +{ + inc_irq_stat(IRQEXT_TLA); + if (param32 & 0x00038000) + stp_timing_alert((struct stp_irq_parm *) ¶m32); +} + +static void stp_reset(void); + +void read_persistent_clock64(struct timespec64 *ts) +{ + union tod_clock clk; + u64 delta; + + delta = initial_leap_seconds + TOD_UNIX_EPOCH; + store_tod_clock_ext(&clk); + clk.eitod -= delta; + ext_to_timespec64(&clk, ts); +} + +void __init read_persistent_wall_and_boot_offset(struct timespec64 *wall_time, + struct timespec64 *boot_offset) +{ + struct timespec64 boot_time; + union tod_clock clk; + u64 delta; + + delta = initial_leap_seconds + TOD_UNIX_EPOCH; + clk = tod_clock_base; + clk.eitod -= delta; + ext_to_timespec64(&clk, &boot_time); + + read_persistent_clock64(wall_time); + *boot_offset = timespec64_sub(*wall_time, boot_time); +} + +static u64 read_tod_clock(struct clocksource *cs) +{ + unsigned long now, adj; + + preempt_disable(); /* protect from changes to steering parameters */ + now = get_tod_clock(); + adj = tod_steering_end - now; + if (unlikely((s64) adj > 0)) + /* + * manually steer by 1 cycle every 2^16 cycles. This + * corresponds to shifting the tod delta by 15. 1s is + * therefore steered in ~9h. The adjust will decrease + * over time, until it finally reaches 0. + */ + now += (tod_steering_delta < 0) ? (adj >> 15) : -(adj >> 15); + preempt_enable(); + return now; +} + +static struct clocksource clocksource_tod = { + .name = "tod", + .rating = 400, + .read = read_tod_clock, + .mask = CLOCKSOURCE_MASK(64), + .mult = 1000, + .shift = 12, + .flags = CLOCK_SOURCE_IS_CONTINUOUS, + .vdso_clock_mode = VDSO_CLOCKMODE_TOD, +}; + +struct clocksource * __init clocksource_default_clock(void) +{ + return &clocksource_tod; +} + +/* + * Initialize the TOD clock and the CPU timer of + * the boot cpu. + */ +void __init time_init(void) +{ + /* Reset time synchronization interfaces. */ + stp_reset(); + + /* request the clock comparator external interrupt */ + if (register_external_irq(EXT_IRQ_CLK_COMP, clock_comparator_interrupt)) + panic("Couldn't request external interrupt 0x1004"); + + /* request the timing alert external interrupt */ + if (register_external_irq(EXT_IRQ_TIMING_ALERT, timing_alert_interrupt)) + panic("Couldn't request external interrupt 0x1406"); + + if (__clocksource_register(&clocksource_tod) != 0) + panic("Could not register TOD clock source"); + + /* Enable TOD clock interrupts on the boot cpu. */ + init_cpu_timer(); + + /* Enable cpu timer interrupts on the boot cpu. */ + vtime_init(); +} + +static DEFINE_PER_CPU(atomic_t, clock_sync_word); +static DEFINE_MUTEX(stp_mutex); +static unsigned long clock_sync_flags; + +#define CLOCK_SYNC_HAS_STP 0 +#define CLOCK_SYNC_STP 1 +#define CLOCK_SYNC_STPINFO_VALID 2 + +/* + * The get_clock function for the physical clock. It will get the current + * TOD clock, subtract the LPAR offset and write the result to *clock. + * The function returns 0 if the clock is in sync with the external time + * source. If the clock mode is local it will return -EOPNOTSUPP and + * -EAGAIN if the clock is not in sync with the external reference. + */ +int get_phys_clock(unsigned long *clock) +{ + atomic_t *sw_ptr; + unsigned int sw0, sw1; + + sw_ptr = &get_cpu_var(clock_sync_word); + sw0 = atomic_read(sw_ptr); + *clock = get_tod_clock() - lpar_offset; + sw1 = atomic_read(sw_ptr); + put_cpu_var(clock_sync_word); + if (sw0 == sw1 && (sw0 & 0x80000000U)) + /* Success: time is in sync. */ + return 0; + if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags)) + return -EOPNOTSUPP; + if (!test_bit(CLOCK_SYNC_STP, &clock_sync_flags)) + return -EACCES; + return -EAGAIN; +} +EXPORT_SYMBOL(get_phys_clock); + +/* + * Make get_phys_clock() return -EAGAIN. + */ +static void disable_sync_clock(void *dummy) +{ + atomic_t *sw_ptr = this_cpu_ptr(&clock_sync_word); + /* + * Clear the in-sync bit 2^31. All get_phys_clock calls will + * fail until the sync bit is turned back on. In addition + * increase the "sequence" counter to avoid the race of an + * stp event and the complete recovery against get_phys_clock. + */ + atomic_andnot(0x80000000, sw_ptr); + atomic_inc(sw_ptr); +} + +/* + * Make get_phys_clock() return 0 again. + * Needs to be called from a context disabled for preemption. + */ +static void enable_sync_clock(void) +{ + atomic_t *sw_ptr = this_cpu_ptr(&clock_sync_word); + atomic_or(0x80000000, sw_ptr); +} + +/* + * Function to check if the clock is in sync. + */ +static inline int check_sync_clock(void) +{ + atomic_t *sw_ptr; + int rc; + + sw_ptr = &get_cpu_var(clock_sync_word); + rc = (atomic_read(sw_ptr) & 0x80000000U) != 0; + put_cpu_var(clock_sync_word); + return rc; +} + +/* + * Apply clock delta to the global data structures. + * This is called once on the CPU that performed the clock sync. + */ +static void clock_sync_global(long delta) +{ + unsigned long now, adj; + struct ptff_qto qto; + int cs; + + /* Fixup the monotonic sched clock. */ + tod_clock_base.eitod += delta; + /* Adjust TOD steering parameters. */ + now = get_tod_clock(); + adj = tod_steering_end - now; + if (unlikely((s64) adj >= 0)) + /* Calculate how much of the old adjustment is left. */ + tod_steering_delta = (tod_steering_delta < 0) ? + -(adj >> 15) : (adj >> 15); + tod_steering_delta += delta; + if ((abs(tod_steering_delta) >> 48) != 0) + panic("TOD clock sync offset %li is too large to drift\n", + tod_steering_delta); + tod_steering_end = now + (abs(tod_steering_delta) << 15); + for (cs = 0; cs < CS_BASES; cs++) { + vdso_data[cs].arch_data.tod_steering_end = tod_steering_end; + vdso_data[cs].arch_data.tod_steering_delta = tod_steering_delta; + } + + /* Update LPAR offset. */ + if (ptff_query(PTFF_QTO) && ptff(&qto, sizeof(qto), PTFF_QTO) == 0) + lpar_offset = qto.tod_epoch_difference; + /* Call the TOD clock change notifier. */ + atomic_notifier_call_chain(&s390_epoch_delta_notifier, 0, &delta); +} + +/* + * Apply clock delta to the per-CPU data structures of this CPU. + * This is called for each online CPU after the call to clock_sync_global. + */ +static void clock_sync_local(long delta) +{ + /* Add the delta to the clock comparator. */ + if (S390_lowcore.clock_comparator != clock_comparator_max) { + S390_lowcore.clock_comparator += delta; + set_clock_comparator(S390_lowcore.clock_comparator); + } + /* Adjust the last_update_clock time-stamp. */ + S390_lowcore.last_update_clock += delta; +} + +/* Single threaded workqueue used for stp sync events */ +static struct workqueue_struct *time_sync_wq; + +static void __init time_init_wq(void) +{ + if (time_sync_wq) + return; + time_sync_wq = create_singlethread_workqueue("timesync"); +} + +struct clock_sync_data { + atomic_t cpus; + int in_sync; + long clock_delta; +}; + +/* + * Server Time Protocol (STP) code. + */ +static bool stp_online; +static struct stp_sstpi stp_info; +static void *stp_page; + +static void stp_work_fn(struct work_struct *work); +static DECLARE_WORK(stp_work, stp_work_fn); +static struct timer_list stp_timer; + +static int __init early_parse_stp(char *p) +{ + return kstrtobool(p, &stp_online); +} +early_param("stp", early_parse_stp); + +/* + * Reset STP attachment. + */ +static void __init stp_reset(void) +{ + int rc; + + stp_page = (void *) get_zeroed_page(GFP_ATOMIC); + rc = chsc_sstpc(stp_page, STP_OP_CTRL, 0x0000, NULL); + if (rc == 0) + set_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags); + else if (stp_online) { + pr_warn("The real or virtual hardware system does not provide an STP interface\n"); + free_page((unsigned long) stp_page); + stp_page = NULL; + stp_online = false; + } +} + +static void stp_timeout(struct timer_list *unused) +{ + queue_work(time_sync_wq, &stp_work); +} + +static int __init stp_init(void) +{ + if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags)) + return 0; + timer_setup(&stp_timer, stp_timeout, 0); + time_init_wq(); + if (!stp_online) + return 0; + queue_work(time_sync_wq, &stp_work); + return 0; +} + +arch_initcall(stp_init); + +/* + * STP timing alert. There are three causes: + * 1) timing status change + * 2) link availability change + * 3) time control parameter change + * In all three cases we are only interested in the clock source state. + * If a STP clock source is now available use it. + */ +static void stp_timing_alert(struct stp_irq_parm *intparm) +{ + if (intparm->tsc || intparm->lac || intparm->tcpc) + queue_work(time_sync_wq, &stp_work); +} + +/* + * STP sync check machine check. This is called when the timing state + * changes from the synchronized state to the unsynchronized state. + * After a STP sync check the clock is not in sync. The machine check + * is broadcasted to all cpus at the same time. + */ +int stp_sync_check(void) +{ + disable_sync_clock(NULL); + return 1; +} + +/* + * STP island condition machine check. This is called when an attached + * server attempts to communicate over an STP link and the servers + * have matching CTN ids and have a valid stratum-1 configuration + * but the configurations do not match. + */ +int stp_island_check(void) +{ + disable_sync_clock(NULL); + return 1; +} + +void stp_queue_work(void) +{ + queue_work(time_sync_wq, &stp_work); +} + +static int __store_stpinfo(void) +{ + int rc = chsc_sstpi(stp_page, &stp_info, sizeof(struct stp_sstpi)); + + if (rc) + clear_bit(CLOCK_SYNC_STPINFO_VALID, &clock_sync_flags); + else + set_bit(CLOCK_SYNC_STPINFO_VALID, &clock_sync_flags); + return rc; +} + +static int stpinfo_valid(void) +{ + return stp_online && test_bit(CLOCK_SYNC_STPINFO_VALID, &clock_sync_flags); +} + +static int stp_sync_clock(void *data) +{ + struct clock_sync_data *sync = data; + long clock_delta, flags; + static int first; + int rc; + + enable_sync_clock(); + if (xchg(&first, 1) == 0) { + /* Wait until all other cpus entered the sync function. */ + while (atomic_read(&sync->cpus) != 0) + cpu_relax(); + rc = 0; + if (stp_info.todoff || stp_info.tmd != 2) { + flags = vdso_update_begin(); + rc = chsc_sstpc(stp_page, STP_OP_SYNC, 0, + &clock_delta); + if (rc == 0) { + sync->clock_delta = clock_delta; + clock_sync_global(clock_delta); + rc = __store_stpinfo(); + if (rc == 0 && stp_info.tmd != 2) + rc = -EAGAIN; + } + vdso_update_end(flags); + } + sync->in_sync = rc ? -EAGAIN : 1; + xchg(&first, 0); + } else { + /* Slave */ + atomic_dec(&sync->cpus); + /* Wait for in_sync to be set. */ + while (READ_ONCE(sync->in_sync) == 0) + __udelay(1); + } + if (sync->in_sync != 1) + /* Didn't work. Clear per-cpu in sync bit again. */ + disable_sync_clock(NULL); + /* Apply clock delta to per-CPU fields of this CPU. */ + clock_sync_local(sync->clock_delta); + + return 0; +} + +static int stp_clear_leap(void) +{ + struct __kernel_timex txc; + int ret; + + memset(&txc, 0, sizeof(txc)); + + ret = do_adjtimex(&txc); + if (ret < 0) + return ret; + + txc.modes = ADJ_STATUS; + txc.status &= ~(STA_INS|STA_DEL); + return do_adjtimex(&txc); +} + +static void stp_check_leap(void) +{ + struct stp_stzi stzi; + struct stp_lsoib *lsoib = &stzi.lsoib; + struct __kernel_timex txc; + int64_t timediff; + int leapdiff, ret; + + if (!stp_info.lu || !check_sync_clock()) { + /* + * Either a scheduled leap second was removed by the operator, + * or STP is out of sync. In both cases, clear the leap second + * kernel flags. + */ + if (stp_clear_leap() < 0) + pr_err("failed to clear leap second flags\n"); + return; + } + + if (chsc_stzi(stp_page, &stzi, sizeof(stzi))) { + pr_err("stzi failed\n"); + return; + } + + timediff = tod_to_ns(lsoib->nlsout - get_tod_clock()) / NSEC_PER_SEC; + leapdiff = lsoib->nlso - lsoib->also; + + if (leapdiff != 1 && leapdiff != -1) { + pr_err("Cannot schedule %d leap seconds\n", leapdiff); + return; + } + + if (timediff < 0) { + if (stp_clear_leap() < 0) + pr_err("failed to clear leap second flags\n"); + } else if (timediff < 7200) { + memset(&txc, 0, sizeof(txc)); + ret = do_adjtimex(&txc); + if (ret < 0) + return; + + txc.modes = ADJ_STATUS; + if (leapdiff > 0) + txc.status |= STA_INS; + else + txc.status |= STA_DEL; + ret = do_adjtimex(&txc); + if (ret < 0) + pr_err("failed to set leap second flags\n"); + /* arm Timer to clear leap second flags */ + mod_timer(&stp_timer, jiffies + msecs_to_jiffies(14400 * MSEC_PER_SEC)); + } else { + /* The day the leap second is scheduled for hasn't been reached. Retry + * in one hour. + */ + mod_timer(&stp_timer, jiffies + msecs_to_jiffies(3600 * MSEC_PER_SEC)); + } +} + +/* + * STP work. Check for the STP state and take over the clock + * synchronization if the STP clock source is usable. + */ +static void stp_work_fn(struct work_struct *work) +{ + struct clock_sync_data stp_sync; + int rc; + + /* prevent multiple execution. */ + mutex_lock(&stp_mutex); + + if (!stp_online) { + chsc_sstpc(stp_page, STP_OP_CTRL, 0x0000, NULL); + del_timer_sync(&stp_timer); + goto out_unlock; + } + + rc = chsc_sstpc(stp_page, STP_OP_CTRL, 0xf0e0, NULL); + if (rc) + goto out_unlock; + + rc = __store_stpinfo(); + if (rc || stp_info.c == 0) + goto out_unlock; + + /* Skip synchronization if the clock is already in sync. */ + if (!check_sync_clock()) { + memset(&stp_sync, 0, sizeof(stp_sync)); + cpus_read_lock(); + atomic_set(&stp_sync.cpus, num_online_cpus() - 1); + stop_machine_cpuslocked(stp_sync_clock, &stp_sync, cpu_online_mask); + cpus_read_unlock(); + } + + if (!check_sync_clock()) + /* + * There is a usable clock but the synchronization failed. + * Retry after a second. + */ + mod_timer(&stp_timer, jiffies + msecs_to_jiffies(MSEC_PER_SEC)); + else if (stp_info.lu) + stp_check_leap(); + +out_unlock: + mutex_unlock(&stp_mutex); +} + +/* + * STP subsys sysfs interface functions + */ +static struct bus_type stp_subsys = { + .name = "stp", + .dev_name = "stp", +}; + +static ssize_t ctn_id_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid()) + ret = sprintf(buf, "%016lx\n", + *(unsigned long *) stp_info.ctnid); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(ctn_id); + +static ssize_t ctn_type_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid()) + ret = sprintf(buf, "%i\n", stp_info.ctn); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(ctn_type); + +static ssize_t dst_offset_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid() && (stp_info.vbits & 0x2000)) + ret = sprintf(buf, "%i\n", (int)(s16) stp_info.dsto); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(dst_offset); + +static ssize_t leap_seconds_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid() && (stp_info.vbits & 0x8000)) + ret = sprintf(buf, "%i\n", (int)(s16) stp_info.leaps); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(leap_seconds); + +static ssize_t leap_seconds_scheduled_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct stp_stzi stzi; + ssize_t ret; + + mutex_lock(&stp_mutex); + if (!stpinfo_valid() || !(stp_info.vbits & 0x8000) || !stp_info.lu) { + mutex_unlock(&stp_mutex); + return -ENODATA; + } + + ret = chsc_stzi(stp_page, &stzi, sizeof(stzi)); + mutex_unlock(&stp_mutex); + if (ret < 0) + return ret; + + if (!stzi.lsoib.p) + return sprintf(buf, "0,0\n"); + + return sprintf(buf, "%lu,%d\n", + tod_to_ns(stzi.lsoib.nlsout - TOD_UNIX_EPOCH) / NSEC_PER_SEC, + stzi.lsoib.nlso - stzi.lsoib.also); +} + +static DEVICE_ATTR_RO(leap_seconds_scheduled); + +static ssize_t stratum_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid()) + ret = sprintf(buf, "%i\n", (int)(s16) stp_info.stratum); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(stratum); + +static ssize_t time_offset_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid() && (stp_info.vbits & 0x0800)) + ret = sprintf(buf, "%i\n", (int) stp_info.tto); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(time_offset); + +static ssize_t time_zone_offset_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid() && (stp_info.vbits & 0x4000)) + ret = sprintf(buf, "%i\n", (int)(s16) stp_info.tzo); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(time_zone_offset); + +static ssize_t timing_mode_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid()) + ret = sprintf(buf, "%i\n", stp_info.tmd); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(timing_mode); + +static ssize_t timing_state_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t ret = -ENODATA; + + mutex_lock(&stp_mutex); + if (stpinfo_valid()) + ret = sprintf(buf, "%i\n", stp_info.tst); + mutex_unlock(&stp_mutex); + return ret; +} + +static DEVICE_ATTR_RO(timing_state); + +static ssize_t online_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + return sprintf(buf, "%i\n", stp_online); +} + +static ssize_t online_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + unsigned int value; + + value = simple_strtoul(buf, NULL, 0); + if (value != 0 && value != 1) + return -EINVAL; + if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags)) + return -EOPNOTSUPP; + mutex_lock(&stp_mutex); + stp_online = value; + if (stp_online) + set_bit(CLOCK_SYNC_STP, &clock_sync_flags); + else + clear_bit(CLOCK_SYNC_STP, &clock_sync_flags); + queue_work(time_sync_wq, &stp_work); + mutex_unlock(&stp_mutex); + return count; +} + +/* + * Can't use DEVICE_ATTR because the attribute should be named + * stp/online but dev_attr_online already exists in this file .. + */ +static DEVICE_ATTR_RW(online); + +static struct attribute *stp_dev_attrs[] = { + &dev_attr_ctn_id.attr, + &dev_attr_ctn_type.attr, + &dev_attr_dst_offset.attr, + &dev_attr_leap_seconds.attr, + &dev_attr_online.attr, + &dev_attr_leap_seconds_scheduled.attr, + &dev_attr_stratum.attr, + &dev_attr_time_offset.attr, + &dev_attr_time_zone_offset.attr, + &dev_attr_timing_mode.attr, + &dev_attr_timing_state.attr, + NULL +}; +ATTRIBUTE_GROUPS(stp_dev); + +static int __init stp_init_sysfs(void) +{ + return subsys_system_register(&stp_subsys, stp_dev_groups); +} + +device_initcall(stp_init_sysfs); diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c new file mode 100644 index 0000000000..68adf1de88 --- /dev/null +++ b/arch/s390/kernel/topology.c @@ -0,0 +1,661 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright IBM Corp. 2007, 2011 + */ + +#define KMSG_COMPONENT "cpu" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/workqueue.h> +#include <linux/memblock.h> +#include <linux/uaccess.h> +#include <linux/sysctl.h> +#include <linux/cpuset.h> +#include <linux/device.h> +#include <linux/export.h> +#include <linux/kernel.h> +#include <linux/sched.h> +#include <linux/sched/topology.h> +#include <linux/delay.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/cpu.h> +#include <linux/smp.h> +#include <linux/mm.h> +#include <linux/nodemask.h> +#include <linux/node.h> +#include <asm/sysinfo.h> + +#define PTF_HORIZONTAL (0UL) +#define PTF_VERTICAL (1UL) +#define PTF_CHECK (2UL) + +enum { + TOPOLOGY_MODE_HW, + TOPOLOGY_MODE_SINGLE, + TOPOLOGY_MODE_PACKAGE, + TOPOLOGY_MODE_UNINITIALIZED +}; + +struct mask_info { + struct mask_info *next; + unsigned char id; + cpumask_t mask; +}; + +static int topology_mode = TOPOLOGY_MODE_UNINITIALIZED; +static void set_topology_timer(void); +static void topology_work_fn(struct work_struct *work); +static struct sysinfo_15_1_x *tl_info; + +static DECLARE_WORK(topology_work, topology_work_fn); + +/* + * Socket/Book linked lists and cpu_topology updates are + * protected by "sched_domains_mutex". + */ +static struct mask_info socket_info; +static struct mask_info book_info; +static struct mask_info drawer_info; + +struct cpu_topology_s390 cpu_topology[NR_CPUS]; +EXPORT_SYMBOL_GPL(cpu_topology); + +static void cpu_group_map(cpumask_t *dst, struct mask_info *info, unsigned int cpu) +{ + static cpumask_t mask; + + cpumask_clear(&mask); + if (!cpumask_test_cpu(cpu, &cpu_setup_mask)) + goto out; + cpumask_set_cpu(cpu, &mask); + switch (topology_mode) { + case TOPOLOGY_MODE_HW: + while (info) { + if (cpumask_test_cpu(cpu, &info->mask)) { + cpumask_copy(&mask, &info->mask); + break; + } + info = info->next; + } + break; + case TOPOLOGY_MODE_PACKAGE: + cpumask_copy(&mask, cpu_present_mask); + break; + default: + fallthrough; + case TOPOLOGY_MODE_SINGLE: + break; + } + cpumask_and(&mask, &mask, &cpu_setup_mask); +out: + cpumask_copy(dst, &mask); +} + +static void cpu_thread_map(cpumask_t *dst, unsigned int cpu) +{ + static cpumask_t mask; + unsigned int max_cpu; + + cpumask_clear(&mask); + if (!cpumask_test_cpu(cpu, &cpu_setup_mask)) + goto out; + cpumask_set_cpu(cpu, &mask); + if (topology_mode != TOPOLOGY_MODE_HW) + goto out; + cpu -= cpu % (smp_cpu_mtid + 1); + max_cpu = min(cpu + smp_cpu_mtid, nr_cpu_ids - 1); + for (; cpu <= max_cpu; cpu++) { + if (cpumask_test_cpu(cpu, &cpu_setup_mask)) + cpumask_set_cpu(cpu, &mask); + } +out: + cpumask_copy(dst, &mask); +} + +#define TOPOLOGY_CORE_BITS 64 + +static void add_cpus_to_mask(struct topology_core *tl_core, + struct mask_info *drawer, + struct mask_info *book, + struct mask_info *socket) +{ + struct cpu_topology_s390 *topo; + unsigned int core; + + for_each_set_bit(core, &tl_core->mask, TOPOLOGY_CORE_BITS) { + unsigned int max_cpu, rcore; + int cpu; + + rcore = TOPOLOGY_CORE_BITS - 1 - core + tl_core->origin; + cpu = smp_find_processor_id(rcore << smp_cpu_mt_shift); + if (cpu < 0) + continue; + max_cpu = min(cpu + smp_cpu_mtid, nr_cpu_ids - 1); + for (; cpu <= max_cpu; cpu++) { + topo = &cpu_topology[cpu]; + topo->drawer_id = drawer->id; + topo->book_id = book->id; + topo->socket_id = socket->id; + topo->core_id = rcore; + topo->thread_id = cpu; + topo->dedicated = tl_core->d; + cpumask_set_cpu(cpu, &drawer->mask); + cpumask_set_cpu(cpu, &book->mask); + cpumask_set_cpu(cpu, &socket->mask); + smp_cpu_set_polarization(cpu, tl_core->pp); + } + } +} + +static void clear_masks(void) +{ + struct mask_info *info; + + info = &socket_info; + while (info) { + cpumask_clear(&info->mask); + info = info->next; + } + info = &book_info; + while (info) { + cpumask_clear(&info->mask); + info = info->next; + } + info = &drawer_info; + while (info) { + cpumask_clear(&info->mask); + info = info->next; + } +} + +static union topology_entry *next_tle(union topology_entry *tle) +{ + if (!tle->nl) + return (union topology_entry *)((struct topology_core *)tle + 1); + return (union topology_entry *)((struct topology_container *)tle + 1); +} + +static void tl_to_masks(struct sysinfo_15_1_x *info) +{ + struct mask_info *socket = &socket_info; + struct mask_info *book = &book_info; + struct mask_info *drawer = &drawer_info; + union topology_entry *tle, *end; + + clear_masks(); + tle = info->tle; + end = (union topology_entry *)((unsigned long)info + info->length); + while (tle < end) { + switch (tle->nl) { + case 3: + drawer = drawer->next; + drawer->id = tle->container.id; + break; + case 2: + book = book->next; + book->id = tle->container.id; + break; + case 1: + socket = socket->next; + socket->id = tle->container.id; + break; + case 0: + add_cpus_to_mask(&tle->cpu, drawer, book, socket); + break; + default: + clear_masks(); + return; + } + tle = next_tle(tle); + } +} + +static void topology_update_polarization_simple(void) +{ + int cpu; + + for_each_possible_cpu(cpu) + smp_cpu_set_polarization(cpu, POLARIZATION_HRZ); +} + +static int ptf(unsigned long fc) +{ + int rc; + + asm volatile( + " .insn rre,0xb9a20000,%1,%1\n" + " ipm %0\n" + " srl %0,28\n" + : "=d" (rc) + : "d" (fc) : "cc"); + return rc; +} + +int topology_set_cpu_management(int fc) +{ + int cpu, rc; + + if (!MACHINE_HAS_TOPOLOGY) + return -EOPNOTSUPP; + if (fc) + rc = ptf(PTF_VERTICAL); + else + rc = ptf(PTF_HORIZONTAL); + if (rc) + return -EBUSY; + for_each_possible_cpu(cpu) + smp_cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); + return rc; +} + +void update_cpu_masks(void) +{ + struct cpu_topology_s390 *topo, *topo_package, *topo_sibling; + int cpu, sibling, pkg_first, smt_first, id; + + for_each_possible_cpu(cpu) { + topo = &cpu_topology[cpu]; + cpu_thread_map(&topo->thread_mask, cpu); + cpu_group_map(&topo->core_mask, &socket_info, cpu); + cpu_group_map(&topo->book_mask, &book_info, cpu); + cpu_group_map(&topo->drawer_mask, &drawer_info, cpu); + topo->booted_cores = 0; + if (topology_mode != TOPOLOGY_MODE_HW) { + id = topology_mode == TOPOLOGY_MODE_PACKAGE ? 0 : cpu; + topo->thread_id = cpu; + topo->core_id = cpu; + topo->socket_id = id; + topo->book_id = id; + topo->drawer_id = id; + } + } + for_each_online_cpu(cpu) { + topo = &cpu_topology[cpu]; + pkg_first = cpumask_first(&topo->core_mask); + topo_package = &cpu_topology[pkg_first]; + if (cpu == pkg_first) { + for_each_cpu(sibling, &topo->core_mask) { + topo_sibling = &cpu_topology[sibling]; + smt_first = cpumask_first(&topo_sibling->thread_mask); + if (sibling == smt_first) + topo_package->booted_cores++; + } + } else { + topo->booted_cores = topo_package->booted_cores; + } + } +} + +void store_topology(struct sysinfo_15_1_x *info) +{ + stsi(info, 15, 1, topology_mnest_limit()); +} + +static void __arch_update_dedicated_flag(void *arg) +{ + if (topology_cpu_dedicated(smp_processor_id())) + set_cpu_flag(CIF_DEDICATED_CPU); + else + clear_cpu_flag(CIF_DEDICATED_CPU); +} + +static int __arch_update_cpu_topology(void) +{ + struct sysinfo_15_1_x *info = tl_info; + int rc = 0; + + mutex_lock(&smp_cpu_state_mutex); + if (MACHINE_HAS_TOPOLOGY) { + rc = 1; + store_topology(info); + tl_to_masks(info); + } + update_cpu_masks(); + if (!MACHINE_HAS_TOPOLOGY) + topology_update_polarization_simple(); + mutex_unlock(&smp_cpu_state_mutex); + return rc; +} + +int arch_update_cpu_topology(void) +{ + struct device *dev; + int cpu, rc; + + rc = __arch_update_cpu_topology(); + on_each_cpu(__arch_update_dedicated_flag, NULL, 0); + for_each_online_cpu(cpu) { + dev = get_cpu_device(cpu); + if (dev) + kobject_uevent(&dev->kobj, KOBJ_CHANGE); + } + return rc; +} + +static void topology_work_fn(struct work_struct *work) +{ + rebuild_sched_domains(); +} + +void topology_schedule_update(void) +{ + schedule_work(&topology_work); +} + +static void topology_flush_work(void) +{ + flush_work(&topology_work); +} + +static void topology_timer_fn(struct timer_list *unused) +{ + if (ptf(PTF_CHECK)) + topology_schedule_update(); + set_topology_timer(); +} + +static struct timer_list topology_timer; + +static atomic_t topology_poll = ATOMIC_INIT(0); + +static void set_topology_timer(void) +{ + if (atomic_add_unless(&topology_poll, -1, 0)) + mod_timer(&topology_timer, jiffies + msecs_to_jiffies(100)); + else + mod_timer(&topology_timer, jiffies + msecs_to_jiffies(60 * MSEC_PER_SEC)); +} + +void topology_expect_change(void) +{ + if (!MACHINE_HAS_TOPOLOGY) + return; + /* This is racy, but it doesn't matter since it is just a heuristic. + * Worst case is that we poll in a higher frequency for a bit longer. + */ + if (atomic_read(&topology_poll) > 60) + return; + atomic_add(60, &topology_poll); + set_topology_timer(); +} + +static int cpu_management; + +static ssize_t dispatching_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + ssize_t count; + + mutex_lock(&smp_cpu_state_mutex); + count = sprintf(buf, "%d\n", cpu_management); + mutex_unlock(&smp_cpu_state_mutex); + return count; +} + +static ssize_t dispatching_store(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + int val, rc; + char delim; + + if (sscanf(buf, "%d %c", &val, &delim) != 1) + return -EINVAL; + if (val != 0 && val != 1) + return -EINVAL; + rc = 0; + cpus_read_lock(); + mutex_lock(&smp_cpu_state_mutex); + if (cpu_management == val) + goto out; + rc = topology_set_cpu_management(val); + if (rc) + goto out; + cpu_management = val; + topology_expect_change(); +out: + mutex_unlock(&smp_cpu_state_mutex); + cpus_read_unlock(); + return rc ? rc : count; +} +static DEVICE_ATTR_RW(dispatching); + +static ssize_t cpu_polarization_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + int cpu = dev->id; + ssize_t count; + + mutex_lock(&smp_cpu_state_mutex); + switch (smp_cpu_get_polarization(cpu)) { + case POLARIZATION_HRZ: + count = sprintf(buf, "horizontal\n"); + break; + case POLARIZATION_VL: + count = sprintf(buf, "vertical:low\n"); + break; + case POLARIZATION_VM: + count = sprintf(buf, "vertical:medium\n"); + break; + case POLARIZATION_VH: + count = sprintf(buf, "vertical:high\n"); + break; + default: + count = sprintf(buf, "unknown\n"); + break; + } + mutex_unlock(&smp_cpu_state_mutex); + return count; +} +static DEVICE_ATTR(polarization, 0444, cpu_polarization_show, NULL); + +static struct attribute *topology_cpu_attrs[] = { + &dev_attr_polarization.attr, + NULL, +}; + +static struct attribute_group topology_cpu_attr_group = { + .attrs = topology_cpu_attrs, +}; + +static ssize_t cpu_dedicated_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + int cpu = dev->id; + ssize_t count; + + mutex_lock(&smp_cpu_state_mutex); + count = sprintf(buf, "%d\n", topology_cpu_dedicated(cpu)); + mutex_unlock(&smp_cpu_state_mutex); + return count; +} +static DEVICE_ATTR(dedicated, 0444, cpu_dedicated_show, NULL); + +static struct attribute *topology_extra_cpu_attrs[] = { + &dev_attr_dedicated.attr, + NULL, +}; + +static struct attribute_group topology_extra_cpu_attr_group = { + .attrs = topology_extra_cpu_attrs, +}; + +int topology_cpu_init(struct cpu *cpu) +{ + int rc; + + rc = sysfs_create_group(&cpu->dev.kobj, &topology_cpu_attr_group); + if (rc || !MACHINE_HAS_TOPOLOGY) + return rc; + rc = sysfs_create_group(&cpu->dev.kobj, &topology_extra_cpu_attr_group); + if (rc) + sysfs_remove_group(&cpu->dev.kobj, &topology_cpu_attr_group); + return rc; +} + +static const struct cpumask *cpu_thread_mask(int cpu) +{ + return &cpu_topology[cpu].thread_mask; +} + + +const struct cpumask *cpu_coregroup_mask(int cpu) +{ + return &cpu_topology[cpu].core_mask; +} + +static const struct cpumask *cpu_book_mask(int cpu) +{ + return &cpu_topology[cpu].book_mask; +} + +static const struct cpumask *cpu_drawer_mask(int cpu) +{ + return &cpu_topology[cpu].drawer_mask; +} + +static struct sched_domain_topology_level s390_topology[] = { + { cpu_thread_mask, cpu_smt_flags, SD_INIT_NAME(SMT) }, + { cpu_coregroup_mask, cpu_core_flags, SD_INIT_NAME(MC) }, + { cpu_book_mask, SD_INIT_NAME(BOOK) }, + { cpu_drawer_mask, SD_INIT_NAME(DRAWER) }, + { cpu_cpu_mask, SD_INIT_NAME(DIE) }, + { NULL, }, +}; + +static void __init alloc_masks(struct sysinfo_15_1_x *info, + struct mask_info *mask, int offset) +{ + int i, nr_masks; + + nr_masks = info->mag[TOPOLOGY_NR_MAG - offset]; + for (i = 0; i < info->mnest - offset; i++) + nr_masks *= info->mag[TOPOLOGY_NR_MAG - offset - 1 - i]; + nr_masks = max(nr_masks, 1); + for (i = 0; i < nr_masks; i++) { + mask->next = memblock_alloc(sizeof(*mask->next), 8); + if (!mask->next) + panic("%s: Failed to allocate %zu bytes align=0x%x\n", + __func__, sizeof(*mask->next), 8); + mask = mask->next; + } +} + +void __init topology_init_early(void) +{ + struct sysinfo_15_1_x *info; + + set_sched_topology(s390_topology); + if (topology_mode == TOPOLOGY_MODE_UNINITIALIZED) { + if (MACHINE_HAS_TOPOLOGY) + topology_mode = TOPOLOGY_MODE_HW; + else + topology_mode = TOPOLOGY_MODE_SINGLE; + } + if (!MACHINE_HAS_TOPOLOGY) + goto out; + tl_info = memblock_alloc(PAGE_SIZE, PAGE_SIZE); + if (!tl_info) + panic("%s: Failed to allocate %lu bytes align=0x%lx\n", + __func__, PAGE_SIZE, PAGE_SIZE); + info = tl_info; + store_topology(info); + pr_info("The CPU configuration topology of the machine is: %d %d %d %d %d %d / %d\n", + info->mag[0], info->mag[1], info->mag[2], info->mag[3], + info->mag[4], info->mag[5], info->mnest); + alloc_masks(info, &socket_info, 1); + alloc_masks(info, &book_info, 2); + alloc_masks(info, &drawer_info, 3); +out: + cpumask_set_cpu(0, &cpu_setup_mask); + __arch_update_cpu_topology(); + __arch_update_dedicated_flag(NULL); +} + +static inline int topology_get_mode(int enabled) +{ + if (!enabled) + return TOPOLOGY_MODE_SINGLE; + return MACHINE_HAS_TOPOLOGY ? TOPOLOGY_MODE_HW : TOPOLOGY_MODE_PACKAGE; +} + +static inline int topology_is_enabled(void) +{ + return topology_mode != TOPOLOGY_MODE_SINGLE; +} + +static int __init topology_setup(char *str) +{ + bool enabled; + int rc; + + rc = kstrtobool(str, &enabled); + if (rc) + return rc; + topology_mode = topology_get_mode(enabled); + return 0; +} +early_param("topology", topology_setup); + +static int topology_ctl_handler(struct ctl_table *ctl, int write, + void *buffer, size_t *lenp, loff_t *ppos) +{ + int enabled = topology_is_enabled(); + int new_mode; + int rc; + struct ctl_table ctl_entry = { + .procname = ctl->procname, + .data = &enabled, + .maxlen = sizeof(int), + .extra1 = SYSCTL_ZERO, + .extra2 = SYSCTL_ONE, + }; + + rc = proc_douintvec_minmax(&ctl_entry, write, buffer, lenp, ppos); + if (rc < 0 || !write) + return rc; + + mutex_lock(&smp_cpu_state_mutex); + new_mode = topology_get_mode(enabled); + if (topology_mode != new_mode) { + topology_mode = new_mode; + topology_schedule_update(); + } + mutex_unlock(&smp_cpu_state_mutex); + topology_flush_work(); + + return rc; +} + +static struct ctl_table topology_ctl_table[] = { + { + .procname = "topology", + .mode = 0644, + .proc_handler = topology_ctl_handler, + }, + { }, +}; + +static int __init topology_init(void) +{ + struct device *dev_root; + int rc = 0; + + timer_setup(&topology_timer, topology_timer_fn, TIMER_DEFERRABLE); + if (MACHINE_HAS_TOPOLOGY) + set_topology_timer(); + else + topology_update_polarization_simple(); + register_sysctl("s390", topology_ctl_table); + + dev_root = bus_get_dev_root(&cpu_subsys); + if (dev_root) { + rc = device_create_file(dev_root, &dev_attr_dispatching); + put_device(dev_root); + } + return rc; +} +device_initcall(topology_init); diff --git a/arch/s390/kernel/trace.c b/arch/s390/kernel/trace.c new file mode 100644 index 0000000000..11a669f3cc --- /dev/null +++ b/arch/s390/kernel/trace.c @@ -0,0 +1,33 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Tracepoint definitions for s390 + * + * Copyright IBM Corp. 2015 + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> + */ + +#include <linux/percpu.h> +#define CREATE_TRACE_POINTS +#include <asm/trace/diag.h> + +EXPORT_TRACEPOINT_SYMBOL(s390_diagnose); + +static DEFINE_PER_CPU(unsigned int, diagnose_trace_depth); + +void notrace trace_s390_diagnose_norecursion(int diag_nr) +{ + unsigned long flags; + unsigned int *depth; + + /* Avoid lockdep recursion. */ + if (IS_ENABLED(CONFIG_LOCKDEP)) + return; + local_irq_save(flags); + depth = this_cpu_ptr(&diagnose_trace_depth); + if (*depth == 0) { + (*depth)++; + trace_s390_diagnose(diag_nr); + (*depth)--; + } + local_irq_restore(flags); +} diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c new file mode 100644 index 0000000000..1d2aa448d1 --- /dev/null +++ b/arch/s390/kernel/traps.c @@ -0,0 +1,403 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * S390 version + * Copyright IBM Corp. 1999, 2000 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), + * Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com), + * + * Derived from "arch/i386/kernel/traps.c" + * Copyright (C) 1991, 1992 Linus Torvalds + */ + +/* + * 'Traps.c' handles hardware traps and faults after we have saved some + * state in 'asm.s'. + */ +#include "asm/irqflags.h" +#include "asm/ptrace.h" +#include <linux/kprobes.h> +#include <linux/kdebug.h> +#include <linux/randomize_kstack.h> +#include <linux/extable.h> +#include <linux/ptrace.h> +#include <linux/sched.h> +#include <linux/sched/debug.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/uaccess.h> +#include <linux/cpu.h> +#include <linux/entry-common.h> +#include <asm/asm-extable.h> +#include <asm/fpu/api.h> +#include <asm/vtime.h> +#include "entry.h" + +static inline void __user *get_trap_ip(struct pt_regs *regs) +{ + unsigned long address; + + if (regs->int_code & 0x200) + address = current->thread.trap_tdb.data[3]; + else + address = regs->psw.addr; + return (void __user *) (address - (regs->int_code >> 16)); +} + +int is_valid_bugaddr(unsigned long addr) +{ + return 1; +} + +void do_report_trap(struct pt_regs *regs, int si_signo, int si_code, char *str) +{ + if (user_mode(regs)) { + force_sig_fault(si_signo, si_code, get_trap_ip(regs)); + report_user_fault(regs, si_signo, 0); + } else { + if (!fixup_exception(regs)) + die(regs, str); + } +} + +static void do_trap(struct pt_regs *regs, int si_signo, int si_code, char *str) +{ + if (notify_die(DIE_TRAP, str, regs, 0, + regs->int_code, si_signo) == NOTIFY_STOP) + return; + do_report_trap(regs, si_signo, si_code, str); +} +NOKPROBE_SYMBOL(do_trap); + +void do_per_trap(struct pt_regs *regs) +{ + if (notify_die(DIE_SSTEP, "sstep", regs, 0, 0, SIGTRAP) == NOTIFY_STOP) + return; + if (!current->ptrace) + return; + force_sig_fault(SIGTRAP, TRAP_HWBKPT, + (void __force __user *) current->thread.per_event.address); +} +NOKPROBE_SYMBOL(do_per_trap); + +static void default_trap_handler(struct pt_regs *regs) +{ + if (user_mode(regs)) { + report_user_fault(regs, SIGSEGV, 0); + force_exit_sig(SIGSEGV); + } else + die(regs, "Unknown program exception"); +} + +#define DO_ERROR_INFO(name, signr, sicode, str) \ +static void name(struct pt_regs *regs) \ +{ \ + do_trap(regs, signr, sicode, str); \ +} + +DO_ERROR_INFO(addressing_exception, SIGILL, ILL_ILLADR, + "addressing exception") +DO_ERROR_INFO(execute_exception, SIGILL, ILL_ILLOPN, + "execute exception") +DO_ERROR_INFO(divide_exception, SIGFPE, FPE_INTDIV, + "fixpoint divide exception") +DO_ERROR_INFO(overflow_exception, SIGFPE, FPE_INTOVF, + "fixpoint overflow exception") +DO_ERROR_INFO(hfp_overflow_exception, SIGFPE, FPE_FLTOVF, + "HFP overflow exception") +DO_ERROR_INFO(hfp_underflow_exception, SIGFPE, FPE_FLTUND, + "HFP underflow exception") +DO_ERROR_INFO(hfp_significance_exception, SIGFPE, FPE_FLTRES, + "HFP significance exception") +DO_ERROR_INFO(hfp_divide_exception, SIGFPE, FPE_FLTDIV, + "HFP divide exception") +DO_ERROR_INFO(hfp_sqrt_exception, SIGFPE, FPE_FLTINV, + "HFP square root exception") +DO_ERROR_INFO(operand_exception, SIGILL, ILL_ILLOPN, + "operand exception") +DO_ERROR_INFO(privileged_op, SIGILL, ILL_PRVOPC, + "privileged operation") +DO_ERROR_INFO(special_op_exception, SIGILL, ILL_ILLOPN, + "special operation exception") +DO_ERROR_INFO(transaction_exception, SIGILL, ILL_ILLOPN, + "transaction constraint exception") + +static inline void do_fp_trap(struct pt_regs *regs, __u32 fpc) +{ + int si_code = 0; + /* FPC[2] is Data Exception Code */ + if ((fpc & 0x00000300) == 0) { + /* bits 6 and 7 of DXC are 0 iff IEEE exception */ + if (fpc & 0x8000) /* invalid fp operation */ + si_code = FPE_FLTINV; + else if (fpc & 0x4000) /* div by 0 */ + si_code = FPE_FLTDIV; + else if (fpc & 0x2000) /* overflow */ + si_code = FPE_FLTOVF; + else if (fpc & 0x1000) /* underflow */ + si_code = FPE_FLTUND; + else if (fpc & 0x0800) /* inexact */ + si_code = FPE_FLTRES; + } + do_trap(regs, SIGFPE, si_code, "floating point exception"); +} + +static void translation_specification_exception(struct pt_regs *regs) +{ + /* May never happen. */ + panic("Translation-Specification Exception"); +} + +static void illegal_op(struct pt_regs *regs) +{ + __u8 opcode[6]; + __u16 __user *location; + int is_uprobe_insn = 0; + int signal = 0; + + location = get_trap_ip(regs); + + if (user_mode(regs)) { + if (get_user(*((__u16 *) opcode), (__u16 __user *) location)) + return; + if (*((__u16 *) opcode) == S390_BREAKPOINT_U16) { + if (current->ptrace) + force_sig_fault(SIGTRAP, TRAP_BRKPT, location); + else + signal = SIGILL; +#ifdef CONFIG_UPROBES + } else if (*((__u16 *) opcode) == UPROBE_SWBP_INSN) { + is_uprobe_insn = 1; +#endif + } else + signal = SIGILL; + } + /* + * We got either an illegal op in kernel mode, or user space trapped + * on a uprobes illegal instruction. See if kprobes or uprobes picks + * it up. If not, SIGILL. + */ + if (is_uprobe_insn || !user_mode(regs)) { + if (notify_die(DIE_BPT, "bpt", regs, 0, + 3, SIGTRAP) != NOTIFY_STOP) + signal = SIGILL; + } + if (signal) + do_trap(regs, signal, ILL_ILLOPC, "illegal operation"); +} +NOKPROBE_SYMBOL(illegal_op); + +DO_ERROR_INFO(specification_exception, SIGILL, ILL_ILLOPN, + "specification exception"); + +static void vector_exception(struct pt_regs *regs) +{ + int si_code, vic; + + if (!MACHINE_HAS_VX) { + do_trap(regs, SIGILL, ILL_ILLOPN, "illegal operation"); + return; + } + + /* get vector interrupt code from fpc */ + save_fpu_regs(); + vic = (current->thread.fpu.fpc & 0xf00) >> 8; + switch (vic) { + case 1: /* invalid vector operation */ + si_code = FPE_FLTINV; + break; + case 2: /* division by zero */ + si_code = FPE_FLTDIV; + break; + case 3: /* overflow */ + si_code = FPE_FLTOVF; + break; + case 4: /* underflow */ + si_code = FPE_FLTUND; + break; + case 5: /* inexact */ + si_code = FPE_FLTRES; + break; + default: /* unknown cause */ + si_code = 0; + } + do_trap(regs, SIGFPE, si_code, "vector exception"); +} + +static void data_exception(struct pt_regs *regs) +{ + save_fpu_regs(); + if (current->thread.fpu.fpc & FPC_DXC_MASK) + do_fp_trap(regs, current->thread.fpu.fpc); + else + do_trap(regs, SIGILL, ILL_ILLOPN, "data exception"); +} + +static void space_switch_exception(struct pt_regs *regs) +{ + /* Set user psw back to home space mode. */ + if (user_mode(regs)) + regs->psw.mask |= PSW_ASC_HOME; + /* Send SIGILL. */ + do_trap(regs, SIGILL, ILL_PRVOPC, "space switch event"); +} + +static void monitor_event_exception(struct pt_regs *regs) +{ + if (user_mode(regs)) + return; + + switch (report_bug(regs->psw.addr - (regs->int_code >> 16), regs)) { + case BUG_TRAP_TYPE_NONE: + fixup_exception(regs); + break; + case BUG_TRAP_TYPE_WARN: + break; + case BUG_TRAP_TYPE_BUG: + die(regs, "monitor event"); + break; + } +} + +void kernel_stack_overflow(struct pt_regs *regs) +{ + bust_spinlocks(1); + printk("Kernel stack overflow.\n"); + show_regs(regs); + bust_spinlocks(0); + panic("Corrupt kernel stack, can't continue."); +} +NOKPROBE_SYMBOL(kernel_stack_overflow); + +static void __init test_monitor_call(void) +{ + int val = 1; + + if (!IS_ENABLED(CONFIG_BUG)) + return; + asm volatile( + " mc 0,0\n" + "0: xgr %0,%0\n" + "1:\n" + EX_TABLE(0b,1b) + : "+d" (val)); + if (!val) + panic("Monitor call doesn't work!\n"); +} + +void __init trap_init(void) +{ + local_mcck_enable(); + test_monitor_call(); +} + +static void (*pgm_check_table[128])(struct pt_regs *regs); + +void noinstr __do_pgm_check(struct pt_regs *regs) +{ + unsigned int trapnr; + irqentry_state_t state; + + regs->int_code = S390_lowcore.pgm_int_code; + regs->int_parm_long = S390_lowcore.trans_exc_code; + + state = irqentry_enter(regs); + + if (user_mode(regs)) { + update_timer_sys(); + if (!static_branch_likely(&cpu_has_bear)) { + if (regs->last_break < 4096) + regs->last_break = 1; + } + current->thread.last_break = regs->last_break; + } + + if (S390_lowcore.pgm_code & 0x0200) { + /* transaction abort */ + current->thread.trap_tdb = S390_lowcore.pgm_tdb; + } + + if (S390_lowcore.pgm_code & PGM_INT_CODE_PER) { + if (user_mode(regs)) { + struct per_event *ev = ¤t->thread.per_event; + + set_thread_flag(TIF_PER_TRAP); + ev->address = S390_lowcore.per_address; + ev->cause = S390_lowcore.per_code_combined; + ev->paid = S390_lowcore.per_access_id; + } else { + /* PER event in kernel is kprobes */ + __arch_local_irq_ssm(regs->psw.mask & ~PSW_MASK_PER); + do_per_trap(regs); + goto out; + } + } + + if (!irqs_disabled_flags(regs->psw.mask)) + trace_hardirqs_on(); + __arch_local_irq_ssm(regs->psw.mask & ~PSW_MASK_PER); + + trapnr = regs->int_code & PGM_INT_CODE_MASK; + if (trapnr) + pgm_check_table[trapnr](regs); +out: + local_irq_disable(); + irqentry_exit(regs, state); +} + +/* + * The program check table contains exactly 128 (0x00-0x7f) entries. Each + * line defines the function to be called corresponding to the program check + * interruption code. + */ +static void (*pgm_check_table[128])(struct pt_regs *regs) = { + [0x00] = default_trap_handler, + [0x01] = illegal_op, + [0x02] = privileged_op, + [0x03] = execute_exception, + [0x04] = do_protection_exception, + [0x05] = addressing_exception, + [0x06] = specification_exception, + [0x07] = data_exception, + [0x08] = overflow_exception, + [0x09] = divide_exception, + [0x0a] = overflow_exception, + [0x0b] = divide_exception, + [0x0c] = hfp_overflow_exception, + [0x0d] = hfp_underflow_exception, + [0x0e] = hfp_significance_exception, + [0x0f] = hfp_divide_exception, + [0x10] = do_dat_exception, + [0x11] = do_dat_exception, + [0x12] = translation_specification_exception, + [0x13] = special_op_exception, + [0x14] = default_trap_handler, + [0x15] = operand_exception, + [0x16] = default_trap_handler, + [0x17] = default_trap_handler, + [0x18] = transaction_exception, + [0x19] = default_trap_handler, + [0x1a] = default_trap_handler, + [0x1b] = vector_exception, + [0x1c] = space_switch_exception, + [0x1d] = hfp_sqrt_exception, + [0x1e ... 0x37] = default_trap_handler, + [0x38] = do_dat_exception, + [0x39] = do_dat_exception, + [0x3a] = do_dat_exception, + [0x3b] = do_dat_exception, + [0x3c] = default_trap_handler, + [0x3d] = do_secure_storage_access, + [0x3e] = do_non_secure_storage_access, + [0x3f] = do_secure_storage_violation, + [0x40] = monitor_event_exception, + [0x41 ... 0x7f] = default_trap_handler, +}; + +#define COND_TRAP(x) asm( \ + ".weak " __stringify(x) "\n\t" \ + ".set " __stringify(x) "," \ + __stringify(default_trap_handler)) + +COND_TRAP(do_secure_storage_access); +COND_TRAP(do_non_secure_storage_access); +COND_TRAP(do_secure_storage_violation); diff --git a/arch/s390/kernel/unwind_bc.c b/arch/s390/kernel/unwind_bc.c new file mode 100644 index 0000000000..0ece156fdd --- /dev/null +++ b/arch/s390/kernel/unwind_bc.c @@ -0,0 +1,178 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#include <linux/sched.h> +#include <linux/sched/task.h> +#include <linux/sched/task_stack.h> +#include <linux/interrupt.h> +#include <asm/sections.h> +#include <asm/ptrace.h> +#include <asm/bitops.h> +#include <asm/stacktrace.h> +#include <asm/unwind.h> + +unsigned long unwind_get_return_address(struct unwind_state *state) +{ + if (unwind_done(state)) + return 0; + return __kernel_text_address(state->ip) ? state->ip : 0; +} +EXPORT_SYMBOL_GPL(unwind_get_return_address); + +static bool outside_of_stack(struct unwind_state *state, unsigned long sp) +{ + return (sp <= state->sp) || + (sp > state->stack_info.end - sizeof(struct stack_frame)); +} + +static bool update_stack_info(struct unwind_state *state, unsigned long sp) +{ + struct stack_info *info = &state->stack_info; + unsigned long *mask = &state->stack_mask; + + /* New stack pointer leaves the current stack */ + if (get_stack_info(sp, state->task, info, mask) != 0 || + !on_stack(info, sp, sizeof(struct stack_frame))) + /* 'sp' does not point to a valid stack */ + return false; + return true; +} + +static inline bool is_final_pt_regs(struct unwind_state *state, + struct pt_regs *regs) +{ + /* user mode or kernel thread pt_regs at the bottom of task stack */ + if (task_pt_regs(state->task) == regs) + return true; + + /* user mode pt_regs at the bottom of irq stack */ + return state->stack_info.type == STACK_TYPE_IRQ && + state->stack_info.end - sizeof(struct pt_regs) == (unsigned long)regs && + READ_ONCE_NOCHECK(regs->psw.mask) & PSW_MASK_PSTATE; +} + +bool unwind_next_frame(struct unwind_state *state) +{ + struct stack_info *info = &state->stack_info; + struct stack_frame *sf; + struct pt_regs *regs; + unsigned long sp, ip; + bool reliable; + + regs = state->regs; + if (unlikely(regs)) { + sp = state->sp; + sf = (struct stack_frame *) sp; + ip = READ_ONCE_NOCHECK(sf->gprs[8]); + reliable = false; + regs = NULL; + /* skip bogus %r14 or if is the same as regs->psw.addr */ + if (!__kernel_text_address(ip) || state->ip == unwind_recover_ret_addr(state, ip)) { + state->regs = NULL; + return unwind_next_frame(state); + } + } else { + sf = (struct stack_frame *) state->sp; + sp = READ_ONCE_NOCHECK(sf->back_chain); + if (likely(sp)) { + /* Non-zero back-chain points to the previous frame */ + if (unlikely(outside_of_stack(state, sp))) { + if (!update_stack_info(state, sp)) + goto out_err; + } + sf = (struct stack_frame *) sp; + ip = READ_ONCE_NOCHECK(sf->gprs[8]); + reliable = true; + } else { + /* No back-chain, look for a pt_regs structure */ + sp = state->sp + STACK_FRAME_OVERHEAD; + if (!on_stack(info, sp, sizeof(struct pt_regs))) + goto out_err; + regs = (struct pt_regs *) sp; + if (is_final_pt_regs(state, regs)) + goto out_stop; + ip = READ_ONCE_NOCHECK(regs->psw.addr); + sp = READ_ONCE_NOCHECK(regs->gprs[15]); + if (unlikely(outside_of_stack(state, sp))) { + if (!update_stack_info(state, sp)) + goto out_err; + } + reliable = true; + } + } + + /* Sanity check: ABI requires SP to be aligned 8 bytes. */ + if (sp & 0x7) + goto out_err; + + /* Update unwind state */ + state->sp = sp; + state->regs = regs; + state->reliable = reliable; + state->ip = unwind_recover_ret_addr(state, ip); + return true; + +out_err: + state->error = true; +out_stop: + state->stack_info.type = STACK_TYPE_UNKNOWN; + return false; +} +EXPORT_SYMBOL_GPL(unwind_next_frame); + +void __unwind_start(struct unwind_state *state, struct task_struct *task, + struct pt_regs *regs, unsigned long first_frame) +{ + struct stack_info *info = &state->stack_info; + struct stack_frame *sf; + unsigned long ip, sp; + + memset(state, 0, sizeof(*state)); + state->task = task; + state->regs = regs; + + /* Don't even attempt to start from user mode regs: */ + if (regs && user_mode(regs)) { + info->type = STACK_TYPE_UNKNOWN; + return; + } + + /* Get the instruction pointer from pt_regs or the stack frame */ + if (regs) { + ip = regs->psw.addr; + sp = regs->gprs[15]; + } else if (task == current) { + sp = current_frame_address(); + } else { + sp = task->thread.ksp; + } + + /* Get current stack pointer and initialize stack info */ + if (!update_stack_info(state, sp)) { + /* Something is wrong with the stack pointer */ + info->type = STACK_TYPE_UNKNOWN; + state->error = true; + return; + } + + if (!regs) { + /* Stack frame is within valid stack */ + sf = (struct stack_frame *)sp; + ip = READ_ONCE_NOCHECK(sf->gprs[8]); + } + + /* Update unwind state */ + state->sp = sp; + state->reliable = true; + state->ip = unwind_recover_ret_addr(state, ip); + + if (!first_frame) + return; + /* Skip through the call chain to the specified starting frame */ + while (!unwind_done(state)) { + if (on_stack(&state->stack_info, first_frame, sizeof(struct stack_frame))) { + if (state->sp >= first_frame) + break; + } + unwind_next_frame(state); + } +} +EXPORT_SYMBOL_GPL(__unwind_start); diff --git a/arch/s390/kernel/uprobes.c b/arch/s390/kernel/uprobes.c new file mode 100644 index 0000000000..b88345ef8b --- /dev/null +++ b/arch/s390/kernel/uprobes.c @@ -0,0 +1,388 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * User-space Probes (UProbes) for s390 + * + * Copyright IBM Corp. 2014 + * Author(s): Jan Willeke, + */ + +#include <linux/uaccess.h> +#include <linux/uprobes.h> +#include <linux/compat.h> +#include <linux/kdebug.h> +#include <linux/sched/task_stack.h> + +#include <asm/switch_to.h> +#include <asm/facility.h> +#include <asm/kprobes.h> +#include <asm/dis.h> +#include "entry.h" + +#define UPROBE_TRAP_NR UINT_MAX + +int arch_uprobe_analyze_insn(struct arch_uprobe *auprobe, struct mm_struct *mm, + unsigned long addr) +{ + return probe_is_prohibited_opcode(auprobe->insn); +} + +int arch_uprobe_pre_xol(struct arch_uprobe *auprobe, struct pt_regs *regs) +{ + if (psw_bits(regs->psw).eaba == PSW_BITS_AMODE_24BIT) + return -EINVAL; + if (!is_compat_task() && psw_bits(regs->psw).eaba == PSW_BITS_AMODE_31BIT) + return -EINVAL; + clear_thread_flag(TIF_PER_TRAP); + auprobe->saved_per = psw_bits(regs->psw).per; + auprobe->saved_int_code = regs->int_code; + regs->int_code = UPROBE_TRAP_NR; + regs->psw.addr = current->utask->xol_vaddr; + set_tsk_thread_flag(current, TIF_UPROBE_SINGLESTEP); + update_cr_regs(current); + return 0; +} + +bool arch_uprobe_xol_was_trapped(struct task_struct *tsk) +{ + struct pt_regs *regs = task_pt_regs(tsk); + + if (regs->int_code != UPROBE_TRAP_NR) + return true; + return false; +} + +static int check_per_event(unsigned short cause, unsigned long control, + struct pt_regs *regs) +{ + if (!(regs->psw.mask & PSW_MASK_PER)) + return 0; + /* user space single step */ + if (control == 0) + return 1; + /* over indication for storage alteration */ + if ((control & 0x20200000) && (cause & 0x2000)) + return 1; + if (cause & 0x8000) { + /* all branches */ + if ((control & 0x80800000) == 0x80000000) + return 1; + /* branch into selected range */ + if (((control & 0x80800000) == 0x80800000) && + regs->psw.addr >= current->thread.per_user.start && + regs->psw.addr <= current->thread.per_user.end) + return 1; + } + return 0; +} + +int arch_uprobe_post_xol(struct arch_uprobe *auprobe, struct pt_regs *regs) +{ + int fixup = probe_get_fixup_type(auprobe->insn); + struct uprobe_task *utask = current->utask; + + clear_tsk_thread_flag(current, TIF_UPROBE_SINGLESTEP); + update_cr_regs(current); + psw_bits(regs->psw).per = auprobe->saved_per; + regs->int_code = auprobe->saved_int_code; + + if (fixup & FIXUP_PSW_NORMAL) + regs->psw.addr += utask->vaddr - utask->xol_vaddr; + if (fixup & FIXUP_RETURN_REGISTER) { + int reg = (auprobe->insn[0] & 0xf0) >> 4; + + regs->gprs[reg] += utask->vaddr - utask->xol_vaddr; + } + if (fixup & FIXUP_BRANCH_NOT_TAKEN) { + int ilen = insn_length(auprobe->insn[0] >> 8); + + if (regs->psw.addr - utask->xol_vaddr == ilen) + regs->psw.addr = utask->vaddr + ilen; + } + if (check_per_event(current->thread.per_event.cause, + current->thread.per_user.control, regs)) { + /* fix per address */ + current->thread.per_event.address = utask->vaddr; + /* trigger per event */ + set_thread_flag(TIF_PER_TRAP); + } + return 0; +} + +int arch_uprobe_exception_notify(struct notifier_block *self, unsigned long val, + void *data) +{ + struct die_args *args = data; + struct pt_regs *regs = args->regs; + + if (!user_mode(regs)) + return NOTIFY_DONE; + if (regs->int_code & 0x200) /* Trap during transaction */ + return NOTIFY_DONE; + switch (val) { + case DIE_BPT: + if (uprobe_pre_sstep_notifier(regs)) + return NOTIFY_STOP; + break; + case DIE_SSTEP: + if (uprobe_post_sstep_notifier(regs)) + return NOTIFY_STOP; + break; + default: + break; + } + return NOTIFY_DONE; +} + +void arch_uprobe_abort_xol(struct arch_uprobe *auprobe, struct pt_regs *regs) +{ + clear_thread_flag(TIF_UPROBE_SINGLESTEP); + regs->int_code = auprobe->saved_int_code; + regs->psw.addr = current->utask->vaddr; + current->thread.per_event.address = current->utask->vaddr; +} + +unsigned long arch_uretprobe_hijack_return_addr(unsigned long trampoline, + struct pt_regs *regs) +{ + unsigned long orig; + + orig = regs->gprs[14]; + regs->gprs[14] = trampoline; + return orig; +} + +bool arch_uretprobe_is_alive(struct return_instance *ret, enum rp_check ctx, + struct pt_regs *regs) +{ + if (ctx == RP_CHECK_CHAIN_CALL) + return user_stack_pointer(regs) <= ret->stack; + else + return user_stack_pointer(regs) < ret->stack; +} + +/* Instruction Emulation */ + +static void adjust_psw_addr(psw_t *psw, unsigned long len) +{ + psw->addr = __rewind_psw(*psw, -len); +} + +#define EMU_ILLEGAL_OP 1 +#define EMU_SPECIFICATION 2 +#define EMU_ADDRESSING 3 + +#define emu_load_ril(ptr, output) \ +({ \ + unsigned int mask = sizeof(*(ptr)) - 1; \ + __typeof__(*(ptr)) input; \ + int __rc = 0; \ + \ + if ((u64 __force)ptr & mask) \ + __rc = EMU_SPECIFICATION; \ + else if (get_user(input, ptr)) \ + __rc = EMU_ADDRESSING; \ + else \ + *(output) = input; \ + __rc; \ +}) + +#define emu_store_ril(regs, ptr, input) \ +({ \ + unsigned int mask = sizeof(*(ptr)) - 1; \ + __typeof__(ptr) __ptr = (ptr); \ + int __rc = 0; \ + \ + if ((u64 __force)__ptr & mask) \ + __rc = EMU_SPECIFICATION; \ + else if (put_user(*(input), __ptr)) \ + __rc = EMU_ADDRESSING; \ + if (__rc == 0) \ + sim_stor_event(regs, \ + (void __force *)__ptr, \ + mask + 1); \ + __rc; \ +}) + +#define emu_cmp_ril(regs, ptr, cmp) \ +({ \ + unsigned int mask = sizeof(*(ptr)) - 1; \ + __typeof__(*(ptr)) input; \ + int __rc = 0; \ + \ + if ((u64 __force)ptr & mask) \ + __rc = EMU_SPECIFICATION; \ + else if (get_user(input, ptr)) \ + __rc = EMU_ADDRESSING; \ + else if (input > *(cmp)) \ + psw_bits((regs)->psw).cc = 1; \ + else if (input < *(cmp)) \ + psw_bits((regs)->psw).cc = 2; \ + else \ + psw_bits((regs)->psw).cc = 0; \ + __rc; \ +}) + +struct insn_ril { + u8 opc0; + u8 reg : 4; + u8 opc1 : 4; + s32 disp; +} __packed; + +union split_register { + u64 u64; + u32 u32[2]; + u16 u16[4]; + s64 s64; + s32 s32[2]; + s16 s16[4]; +}; + +/* + * If user per registers are setup to trace storage alterations and an + * emulated store took place on a fitting address a user trap is generated. + */ +static void sim_stor_event(struct pt_regs *regs, void *addr, int len) +{ + if (!(regs->psw.mask & PSW_MASK_PER)) + return; + if (!(current->thread.per_user.control & PER_EVENT_STORE)) + return; + if ((void *)current->thread.per_user.start > (addr + len)) + return; + if ((void *)current->thread.per_user.end < addr) + return; + current->thread.per_event.address = regs->psw.addr; + current->thread.per_event.cause = PER_EVENT_STORE >> 16; + set_thread_flag(TIF_PER_TRAP); +} + +/* + * pc relative instructions are emulated, since parameters may not be + * accessible from the xol area due to range limitations. + */ +static void handle_insn_ril(struct arch_uprobe *auprobe, struct pt_regs *regs) +{ + union split_register *rx; + struct insn_ril *insn; + unsigned int ilen; + void *uptr; + int rc = 0; + + insn = (struct insn_ril *) &auprobe->insn; + rx = (union split_register *) ®s->gprs[insn->reg]; + uptr = (void *)(regs->psw.addr + (insn->disp * 2)); + ilen = insn_length(insn->opc0); + + switch (insn->opc0) { + case 0xc0: + switch (insn->opc1) { + case 0x00: /* larl */ + rx->u64 = (unsigned long)uptr; + break; + } + break; + case 0xc4: + switch (insn->opc1) { + case 0x02: /* llhrl */ + rc = emu_load_ril((u16 __user *)uptr, &rx->u32[1]); + break; + case 0x04: /* lghrl */ + rc = emu_load_ril((s16 __user *)uptr, &rx->u64); + break; + case 0x05: /* lhrl */ + rc = emu_load_ril((s16 __user *)uptr, &rx->u32[1]); + break; + case 0x06: /* llghrl */ + rc = emu_load_ril((u16 __user *)uptr, &rx->u64); + break; + case 0x08: /* lgrl */ + rc = emu_load_ril((u64 __user *)uptr, &rx->u64); + break; + case 0x0c: /* lgfrl */ + rc = emu_load_ril((s32 __user *)uptr, &rx->u64); + break; + case 0x0d: /* lrl */ + rc = emu_load_ril((u32 __user *)uptr, &rx->u32[1]); + break; + case 0x0e: /* llgfrl */ + rc = emu_load_ril((u32 __user *)uptr, &rx->u64); + break; + case 0x07: /* sthrl */ + rc = emu_store_ril(regs, (u16 __user *)uptr, &rx->u16[3]); + break; + case 0x0b: /* stgrl */ + rc = emu_store_ril(regs, (u64 __user *)uptr, &rx->u64); + break; + case 0x0f: /* strl */ + rc = emu_store_ril(regs, (u32 __user *)uptr, &rx->u32[1]); + break; + } + break; + case 0xc6: + switch (insn->opc1) { + case 0x04: /* cghrl */ + rc = emu_cmp_ril(regs, (s16 __user *)uptr, &rx->s64); + break; + case 0x05: /* chrl */ + rc = emu_cmp_ril(regs, (s16 __user *)uptr, &rx->s32[1]); + break; + case 0x06: /* clghrl */ + rc = emu_cmp_ril(regs, (u16 __user *)uptr, &rx->u64); + break; + case 0x07: /* clhrl */ + rc = emu_cmp_ril(regs, (u16 __user *)uptr, &rx->u32[1]); + break; + case 0x08: /* cgrl */ + rc = emu_cmp_ril(regs, (s64 __user *)uptr, &rx->s64); + break; + case 0x0a: /* clgrl */ + rc = emu_cmp_ril(regs, (u64 __user *)uptr, &rx->u64); + break; + case 0x0c: /* cgfrl */ + rc = emu_cmp_ril(regs, (s32 __user *)uptr, &rx->s64); + break; + case 0x0d: /* crl */ + rc = emu_cmp_ril(regs, (s32 __user *)uptr, &rx->s32[1]); + break; + case 0x0e: /* clgfrl */ + rc = emu_cmp_ril(regs, (u32 __user *)uptr, &rx->u64); + break; + case 0x0f: /* clrl */ + rc = emu_cmp_ril(regs, (u32 __user *)uptr, &rx->u32[1]); + break; + } + break; + } + adjust_psw_addr(®s->psw, ilen); + switch (rc) { + case EMU_ILLEGAL_OP: + regs->int_code = ilen << 16 | 0x0001; + do_report_trap(regs, SIGILL, ILL_ILLOPC, NULL); + break; + case EMU_SPECIFICATION: + regs->int_code = ilen << 16 | 0x0006; + do_report_trap(regs, SIGILL, ILL_ILLOPC , NULL); + break; + case EMU_ADDRESSING: + regs->int_code = ilen << 16 | 0x0005; + do_report_trap(regs, SIGSEGV, SEGV_MAPERR, NULL); + break; + } +} + +bool arch_uprobe_skip_sstep(struct arch_uprobe *auprobe, struct pt_regs *regs) +{ + if ((psw_bits(regs->psw).eaba == PSW_BITS_AMODE_24BIT) || + ((psw_bits(regs->psw).eaba == PSW_BITS_AMODE_31BIT) && + !is_compat_task())) { + regs->psw.addr = __rewind_psw(regs->psw, UPROBE_SWBP_INSN_SIZE); + do_report_trap(regs, SIGILL, ILL_ILLADR, NULL); + return true; + } + if (probe_is_insn_relative_long(auprobe->insn)) { + handle_insn_ril(auprobe, regs); + return true; + } + return false; +} diff --git a/arch/s390/kernel/uv.c b/arch/s390/kernel/uv.c new file mode 100644 index 0000000000..fc07bc39e6 --- /dev/null +++ b/arch/s390/kernel/uv.c @@ -0,0 +1,717 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Common Ultravisor functions and initialization + * + * Copyright IBM Corp. 2019, 2020 + */ +#define KMSG_COMPONENT "prot_virt" +#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt + +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/sizes.h> +#include <linux/bitmap.h> +#include <linux/memblock.h> +#include <linux/pagemap.h> +#include <linux/swap.h> +#include <asm/facility.h> +#include <asm/sections.h> +#include <asm/uv.h> + +/* the bootdata_preserved fields come from ones in arch/s390/boot/uv.c */ +#ifdef CONFIG_PROTECTED_VIRTUALIZATION_GUEST +int __bootdata_preserved(prot_virt_guest); +#endif + +/* + * uv_info contains both host and guest information but it's currently only + * expected to be used within modules if it's the KVM module or for + * any PV guest module. + * + * The kernel itself will write these values once in uv_query_info() + * and then make some of them readable via a sysfs interface. + */ +struct uv_info __bootdata_preserved(uv_info); +EXPORT_SYMBOL(uv_info); + +#if IS_ENABLED(CONFIG_KVM) +int __bootdata_preserved(prot_virt_host); +EXPORT_SYMBOL(prot_virt_host); + +static int __init uv_init(phys_addr_t stor_base, unsigned long stor_len) +{ + struct uv_cb_init uvcb = { + .header.cmd = UVC_CMD_INIT_UV, + .header.len = sizeof(uvcb), + .stor_origin = stor_base, + .stor_len = stor_len, + }; + + if (uv_call(0, (uint64_t)&uvcb)) { + pr_err("Ultravisor init failed with rc: 0x%x rrc: 0%x\n", + uvcb.header.rc, uvcb.header.rrc); + return -1; + } + return 0; +} + +void __init setup_uv(void) +{ + void *uv_stor_base; + + if (!is_prot_virt_host()) + return; + + uv_stor_base = memblock_alloc_try_nid( + uv_info.uv_base_stor_len, SZ_1M, SZ_2G, + MEMBLOCK_ALLOC_ACCESSIBLE, NUMA_NO_NODE); + if (!uv_stor_base) { + pr_warn("Failed to reserve %lu bytes for ultravisor base storage\n", + uv_info.uv_base_stor_len); + goto fail; + } + + if (uv_init(__pa(uv_stor_base), uv_info.uv_base_stor_len)) { + memblock_free(uv_stor_base, uv_info.uv_base_stor_len); + goto fail; + } + + pr_info("Reserving %luMB as ultravisor base storage\n", + uv_info.uv_base_stor_len >> 20); + return; +fail: + pr_info("Disabling support for protected virtualization"); + prot_virt_host = 0; +} + +/* + * Requests the Ultravisor to pin the page in the shared state. This will + * cause an intercept when the guest attempts to unshare the pinned page. + */ +int uv_pin_shared(unsigned long paddr) +{ + struct uv_cb_cfs uvcb = { + .header.cmd = UVC_CMD_PIN_PAGE_SHARED, + .header.len = sizeof(uvcb), + .paddr = paddr, + }; + + if (uv_call(0, (u64)&uvcb)) + return -EINVAL; + return 0; +} +EXPORT_SYMBOL_GPL(uv_pin_shared); + +/* + * Requests the Ultravisor to destroy a guest page and make it + * accessible to the host. The destroy clears the page instead of + * exporting. + * + * @paddr: Absolute host address of page to be destroyed + */ +static int uv_destroy_page(unsigned long paddr) +{ + struct uv_cb_cfs uvcb = { + .header.cmd = UVC_CMD_DESTR_SEC_STOR, + .header.len = sizeof(uvcb), + .paddr = paddr + }; + + if (uv_call(0, (u64)&uvcb)) { + /* + * Older firmware uses 107/d as an indication of a non secure + * page. Let us emulate the newer variant (no-op). + */ + if (uvcb.header.rc == 0x107 && uvcb.header.rrc == 0xd) + return 0; + return -EINVAL; + } + return 0; +} + +/* + * The caller must already hold a reference to the page + */ +int uv_destroy_owned_page(unsigned long paddr) +{ + struct page *page = phys_to_page(paddr); + int rc; + + get_page(page); + rc = uv_destroy_page(paddr); + if (!rc) + clear_bit(PG_arch_1, &page->flags); + put_page(page); + return rc; +} + +/* + * Requests the Ultravisor to encrypt a guest page and make it + * accessible to the host for paging (export). + * + * @paddr: Absolute host address of page to be exported + */ +int uv_convert_from_secure(unsigned long paddr) +{ + struct uv_cb_cfs uvcb = { + .header.cmd = UVC_CMD_CONV_FROM_SEC_STOR, + .header.len = sizeof(uvcb), + .paddr = paddr + }; + + if (uv_call(0, (u64)&uvcb)) + return -EINVAL; + return 0; +} + +/* + * The caller must already hold a reference to the page + */ +int uv_convert_owned_from_secure(unsigned long paddr) +{ + struct page *page = phys_to_page(paddr); + int rc; + + get_page(page); + rc = uv_convert_from_secure(paddr); + if (!rc) + clear_bit(PG_arch_1, &page->flags); + put_page(page); + return rc; +} + +/* + * Calculate the expected ref_count for a page that would otherwise have no + * further pins. This was cribbed from similar functions in other places in + * the kernel, but with some slight modifications. We know that a secure + * page can not be a huge page for example. + */ +static int expected_page_refs(struct page *page) +{ + int res; + + res = page_mapcount(page); + if (PageSwapCache(page)) { + res++; + } else if (page_mapping(page)) { + res++; + if (page_has_private(page)) + res++; + } + return res; +} + +static int make_page_secure(struct page *page, struct uv_cb_header *uvcb) +{ + int expected, cc = 0; + + if (PageWriteback(page)) + return -EAGAIN; + expected = expected_page_refs(page); + if (!page_ref_freeze(page, expected)) + return -EBUSY; + set_bit(PG_arch_1, &page->flags); + /* + * If the UVC does not succeed or fail immediately, we don't want to + * loop for long, or we might get stall notifications. + * On the other hand, this is a complex scenario and we are holding a lot of + * locks, so we can't easily sleep and reschedule. We try only once, + * and if the UVC returned busy or partial completion, we return + * -EAGAIN and we let the callers deal with it. + */ + cc = __uv_call(0, (u64)uvcb); + page_ref_unfreeze(page, expected); + /* + * Return -ENXIO if the page was not mapped, -EINVAL for other errors. + * If busy or partially completed, return -EAGAIN. + */ + if (cc == UVC_CC_OK) + return 0; + else if (cc == UVC_CC_BUSY || cc == UVC_CC_PARTIAL) + return -EAGAIN; + return uvcb->rc == 0x10a ? -ENXIO : -EINVAL; +} + +/** + * should_export_before_import - Determine whether an export is needed + * before an import-like operation + * @uvcb: the Ultravisor control block of the UVC to be performed + * @mm: the mm of the process + * + * Returns whether an export is needed before every import-like operation. + * This is needed for shared pages, which don't trigger a secure storage + * exception when accessed from a different guest. + * + * Although considered as one, the Unpin Page UVC is not an actual import, + * so it is not affected. + * + * No export is needed also when there is only one protected VM, because the + * page cannot belong to the wrong VM in that case (there is no "other VM" + * it can belong to). + * + * Return: true if an export is needed before every import, otherwise false. + */ +static bool should_export_before_import(struct uv_cb_header *uvcb, struct mm_struct *mm) +{ + /* + * The misc feature indicates, among other things, that importing a + * shared page from a different protected VM will automatically also + * transfer its ownership. + */ + if (uv_has_feature(BIT_UV_FEAT_MISC)) + return false; + if (uvcb->cmd == UVC_CMD_UNPIN_PAGE_SHARED) + return false; + return atomic_read(&mm->context.protected_count) > 1; +} + +/* + * Requests the Ultravisor to make a page accessible to a guest. + * If it's brought in the first time, it will be cleared. If + * it has been exported before, it will be decrypted and integrity + * checked. + */ +int gmap_make_secure(struct gmap *gmap, unsigned long gaddr, void *uvcb) +{ + struct vm_area_struct *vma; + bool local_drain = false; + spinlock_t *ptelock; + unsigned long uaddr; + struct page *page; + pte_t *ptep; + int rc; + +again: + rc = -EFAULT; + mmap_read_lock(gmap->mm); + + uaddr = __gmap_translate(gmap, gaddr); + if (IS_ERR_VALUE(uaddr)) + goto out; + vma = vma_lookup(gmap->mm, uaddr); + if (!vma) + goto out; + /* + * Secure pages cannot be huge and userspace should not combine both. + * In case userspace does it anyway this will result in an -EFAULT for + * the unpack. The guest is thus never reaching secure mode. If + * userspace is playing dirty tricky with mapping huge pages later + * on this will result in a segmentation fault. + */ + if (is_vm_hugetlb_page(vma)) + goto out; + + rc = -ENXIO; + ptep = get_locked_pte(gmap->mm, uaddr, &ptelock); + if (!ptep) + goto out; + if (pte_present(*ptep) && !(pte_val(*ptep) & _PAGE_INVALID) && pte_write(*ptep)) { + page = pte_page(*ptep); + rc = -EAGAIN; + if (trylock_page(page)) { + if (should_export_before_import(uvcb, gmap->mm)) + uv_convert_from_secure(page_to_phys(page)); + rc = make_page_secure(page, uvcb); + unlock_page(page); + } + } + pte_unmap_unlock(ptep, ptelock); +out: + mmap_read_unlock(gmap->mm); + + if (rc == -EAGAIN) { + /* + * If we are here because the UVC returned busy or partial + * completion, this is just a useless check, but it is safe. + */ + wait_on_page_writeback(page); + } else if (rc == -EBUSY) { + /* + * If we have tried a local drain and the page refcount + * still does not match our expected safe value, try with a + * system wide drain. This is needed if the pagevecs holding + * the page are on a different CPU. + */ + if (local_drain) { + lru_add_drain_all(); + /* We give up here, and let the caller try again */ + return -EAGAIN; + } + /* + * We are here if the page refcount does not match the + * expected safe value. The main culprits are usually + * pagevecs. With lru_add_drain() we drain the pagevecs + * on the local CPU so that hopefully the refcount will + * reach the expected safe value. + */ + lru_add_drain(); + local_drain = true; + /* And now we try again immediately after draining */ + goto again; + } else if (rc == -ENXIO) { + if (gmap_fault(gmap, gaddr, FAULT_FLAG_WRITE)) + return -EFAULT; + return -EAGAIN; + } + return rc; +} +EXPORT_SYMBOL_GPL(gmap_make_secure); + +int gmap_convert_to_secure(struct gmap *gmap, unsigned long gaddr) +{ + struct uv_cb_cts uvcb = { + .header.cmd = UVC_CMD_CONV_TO_SEC_STOR, + .header.len = sizeof(uvcb), + .guest_handle = gmap->guest_handle, + .gaddr = gaddr, + }; + + return gmap_make_secure(gmap, gaddr, &uvcb); +} +EXPORT_SYMBOL_GPL(gmap_convert_to_secure); + +/** + * gmap_destroy_page - Destroy a guest page. + * @gmap: the gmap of the guest + * @gaddr: the guest address to destroy + * + * An attempt will be made to destroy the given guest page. If the attempt + * fails, an attempt is made to export the page. If both attempts fail, an + * appropriate error is returned. + */ +int gmap_destroy_page(struct gmap *gmap, unsigned long gaddr) +{ + struct vm_area_struct *vma; + unsigned long uaddr; + struct page *page; + int rc; + + rc = -EFAULT; + mmap_read_lock(gmap->mm); + + uaddr = __gmap_translate(gmap, gaddr); + if (IS_ERR_VALUE(uaddr)) + goto out; + vma = vma_lookup(gmap->mm, uaddr); + if (!vma) + goto out; + /* + * Huge pages should not be able to become secure + */ + if (is_vm_hugetlb_page(vma)) + goto out; + + rc = 0; + /* we take an extra reference here */ + page = follow_page(vma, uaddr, FOLL_WRITE | FOLL_GET); + if (IS_ERR_OR_NULL(page)) + goto out; + rc = uv_destroy_owned_page(page_to_phys(page)); + /* + * Fault handlers can race; it is possible that two CPUs will fault + * on the same secure page. One CPU can destroy the page, reboot, + * re-enter secure mode and import it, while the second CPU was + * stuck at the beginning of the handler. At some point the second + * CPU will be able to progress, and it will not be able to destroy + * the page. In that case we do not want to terminate the process, + * we instead try to export the page. + */ + if (rc) + rc = uv_convert_owned_from_secure(page_to_phys(page)); + put_page(page); +out: + mmap_read_unlock(gmap->mm); + return rc; +} +EXPORT_SYMBOL_GPL(gmap_destroy_page); + +/* + * To be called with the page locked or with an extra reference! This will + * prevent gmap_make_secure from touching the page concurrently. Having 2 + * parallel make_page_accessible is fine, as the UV calls will become a + * no-op if the page is already exported. + */ +int arch_make_page_accessible(struct page *page) +{ + int rc = 0; + + /* Hugepage cannot be protected, so nothing to do */ + if (PageHuge(page)) + return 0; + + /* + * PG_arch_1 is used in 3 places: + * 1. for kernel page tables during early boot + * 2. for storage keys of huge pages and KVM + * 3. As an indication that this page might be secure. This can + * overindicate, e.g. we set the bit before calling + * convert_to_secure. + * As secure pages are never huge, all 3 variants can co-exists. + */ + if (!test_bit(PG_arch_1, &page->flags)) + return 0; + + rc = uv_pin_shared(page_to_phys(page)); + if (!rc) { + clear_bit(PG_arch_1, &page->flags); + return 0; + } + + rc = uv_convert_from_secure(page_to_phys(page)); + if (!rc) { + clear_bit(PG_arch_1, &page->flags); + return 0; + } + + return rc; +} +EXPORT_SYMBOL_GPL(arch_make_page_accessible); + +#endif + +#if defined(CONFIG_PROTECTED_VIRTUALIZATION_GUEST) || IS_ENABLED(CONFIG_KVM) +static ssize_t uv_query_facilities(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n%lx\n%lx\n%lx\n", + uv_info.inst_calls_list[0], + uv_info.inst_calls_list[1], + uv_info.inst_calls_list[2], + uv_info.inst_calls_list[3]); +} + +static struct kobj_attribute uv_query_facilities_attr = + __ATTR(facilities, 0444, uv_query_facilities, NULL); + +static ssize_t uv_query_supp_se_hdr_ver(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.supp_se_hdr_ver); +} + +static struct kobj_attribute uv_query_supp_se_hdr_ver_attr = + __ATTR(supp_se_hdr_ver, 0444, uv_query_supp_se_hdr_ver, NULL); + +static ssize_t uv_query_supp_se_hdr_pcf(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.supp_se_hdr_pcf); +} + +static struct kobj_attribute uv_query_supp_se_hdr_pcf_attr = + __ATTR(supp_se_hdr_pcf, 0444, uv_query_supp_se_hdr_pcf, NULL); + +static ssize_t uv_query_dump_cpu_len(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.guest_cpu_stor_len); +} + +static struct kobj_attribute uv_query_dump_cpu_len_attr = + __ATTR(uv_query_dump_cpu_len, 0444, uv_query_dump_cpu_len, NULL); + +static ssize_t uv_query_dump_storage_state_len(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.conf_dump_storage_state_len); +} + +static struct kobj_attribute uv_query_dump_storage_state_len_attr = + __ATTR(dump_storage_state_len, 0444, uv_query_dump_storage_state_len, NULL); + +static ssize_t uv_query_dump_finalize_len(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.conf_dump_finalize_len); +} + +static struct kobj_attribute uv_query_dump_finalize_len_attr = + __ATTR(dump_finalize_len, 0444, uv_query_dump_finalize_len, NULL); + +static ssize_t uv_query_feature_indications(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.uv_feature_indications); +} + +static struct kobj_attribute uv_query_feature_indications_attr = + __ATTR(feature_indications, 0444, uv_query_feature_indications, NULL); + +static ssize_t uv_query_max_guest_cpus(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%d\n", uv_info.max_guest_cpu_id + 1); +} + +static struct kobj_attribute uv_query_max_guest_cpus_attr = + __ATTR(max_cpus, 0444, uv_query_max_guest_cpus, NULL); + +static ssize_t uv_query_max_guest_vms(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%d\n", uv_info.max_num_sec_conf); +} + +static struct kobj_attribute uv_query_max_guest_vms_attr = + __ATTR(max_guests, 0444, uv_query_max_guest_vms, NULL); + +static ssize_t uv_query_max_guest_addr(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.max_sec_stor_addr); +} + +static struct kobj_attribute uv_query_max_guest_addr_attr = + __ATTR(max_address, 0444, uv_query_max_guest_addr, NULL); + +static ssize_t uv_query_supp_att_req_hdr_ver(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.supp_att_req_hdr_ver); +} + +static struct kobj_attribute uv_query_supp_att_req_hdr_ver_attr = + __ATTR(supp_att_req_hdr_ver, 0444, uv_query_supp_att_req_hdr_ver, NULL); + +static ssize_t uv_query_supp_att_pflags(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.supp_att_pflags); +} + +static struct kobj_attribute uv_query_supp_att_pflags_attr = + __ATTR(supp_att_pflags, 0444, uv_query_supp_att_pflags, NULL); + +static ssize_t uv_query_supp_add_secret_req_ver(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.supp_add_secret_req_ver); +} + +static struct kobj_attribute uv_query_supp_add_secret_req_ver_attr = + __ATTR(supp_add_secret_req_ver, 0444, uv_query_supp_add_secret_req_ver, NULL); + +static ssize_t uv_query_supp_add_secret_pcf(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.supp_add_secret_pcf); +} + +static struct kobj_attribute uv_query_supp_add_secret_pcf_attr = + __ATTR(supp_add_secret_pcf, 0444, uv_query_supp_add_secret_pcf, NULL); + +static ssize_t uv_query_supp_secret_types(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%lx\n", uv_info.supp_secret_types); +} + +static struct kobj_attribute uv_query_supp_secret_types_attr = + __ATTR(supp_secret_types, 0444, uv_query_supp_secret_types, NULL); + +static ssize_t uv_query_max_secrets(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%d\n", uv_info.max_secrets); +} + +static struct kobj_attribute uv_query_max_secrets_attr = + __ATTR(max_secrets, 0444, uv_query_max_secrets, NULL); + +static struct attribute *uv_query_attrs[] = { + &uv_query_facilities_attr.attr, + &uv_query_feature_indications_attr.attr, + &uv_query_max_guest_cpus_attr.attr, + &uv_query_max_guest_vms_attr.attr, + &uv_query_max_guest_addr_attr.attr, + &uv_query_supp_se_hdr_ver_attr.attr, + &uv_query_supp_se_hdr_pcf_attr.attr, + &uv_query_dump_storage_state_len_attr.attr, + &uv_query_dump_finalize_len_attr.attr, + &uv_query_dump_cpu_len_attr.attr, + &uv_query_supp_att_req_hdr_ver_attr.attr, + &uv_query_supp_att_pflags_attr.attr, + &uv_query_supp_add_secret_req_ver_attr.attr, + &uv_query_supp_add_secret_pcf_attr.attr, + &uv_query_supp_secret_types_attr.attr, + &uv_query_max_secrets_attr.attr, + NULL, +}; + +static struct attribute_group uv_query_attr_group = { + .attrs = uv_query_attrs, +}; + +static ssize_t uv_is_prot_virt_guest(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + int val = 0; + +#ifdef CONFIG_PROTECTED_VIRTUALIZATION_GUEST + val = prot_virt_guest; +#endif + return sysfs_emit(buf, "%d\n", val); +} + +static ssize_t uv_is_prot_virt_host(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + int val = 0; + +#if IS_ENABLED(CONFIG_KVM) + val = prot_virt_host; +#endif + + return sysfs_emit(buf, "%d\n", val); +} + +static struct kobj_attribute uv_prot_virt_guest = + __ATTR(prot_virt_guest, 0444, uv_is_prot_virt_guest, NULL); + +static struct kobj_attribute uv_prot_virt_host = + __ATTR(prot_virt_host, 0444, uv_is_prot_virt_host, NULL); + +static const struct attribute *uv_prot_virt_attrs[] = { + &uv_prot_virt_guest.attr, + &uv_prot_virt_host.attr, + NULL, +}; + +static struct kset *uv_query_kset; +static struct kobject *uv_kobj; + +static int __init uv_info_init(void) +{ + int rc = -ENOMEM; + + if (!test_facility(158)) + return 0; + + uv_kobj = kobject_create_and_add("uv", firmware_kobj); + if (!uv_kobj) + return -ENOMEM; + + rc = sysfs_create_files(uv_kobj, uv_prot_virt_attrs); + if (rc) + goto out_kobj; + + uv_query_kset = kset_create_and_add("query", NULL, uv_kobj); + if (!uv_query_kset) { + rc = -ENOMEM; + goto out_ind_files; + } + + rc = sysfs_create_group(&uv_query_kset->kobj, &uv_query_attr_group); + if (!rc) + return 0; + + kset_unregister(uv_query_kset); +out_ind_files: + sysfs_remove_files(uv_kobj, uv_prot_virt_attrs); +out_kobj: + kobject_del(uv_kobj); + kobject_put(uv_kobj); + return rc; +} +device_initcall(uv_info_init); +#endif diff --git a/arch/s390/kernel/vdso.c b/arch/s390/kernel/vdso.c new file mode 100644 index 0000000000..bbaefd84f1 --- /dev/null +++ b/arch/s390/kernel/vdso.c @@ -0,0 +1,258 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * vdso setup for s390 + * + * Copyright IBM Corp. 2008 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com) + */ + +#include <linux/binfmts.h> +#include <linux/compat.h> +#include <linux/elf.h> +#include <linux/errno.h> +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/smp.h> +#include <linux/time_namespace.h> +#include <linux/random.h> +#include <vdso/datapage.h> +#include <asm/vdso.h> + +extern char vdso64_start[], vdso64_end[]; +extern char vdso32_start[], vdso32_end[]; + +static struct vm_special_mapping vvar_mapping; + +static union { + struct vdso_data data[CS_BASES]; + u8 page[PAGE_SIZE]; +} vdso_data_store __page_aligned_data; + +struct vdso_data *vdso_data = vdso_data_store.data; + +enum vvar_pages { + VVAR_DATA_PAGE_OFFSET, + VVAR_TIMENS_PAGE_OFFSET, + VVAR_NR_PAGES, +}; + +#ifdef CONFIG_TIME_NS +struct vdso_data *arch_get_vdso_data(void *vvar_page) +{ + return (struct vdso_data *)(vvar_page); +} + +/* + * The VVAR page layout depends on whether a task belongs to the root or + * non-root time namespace. Whenever a task changes its namespace, the VVAR + * page tables are cleared and then they will be re-faulted with a + * corresponding layout. + * See also the comment near timens_setup_vdso_data() for details. + */ +int vdso_join_timens(struct task_struct *task, struct time_namespace *ns) +{ + struct mm_struct *mm = task->mm; + VMA_ITERATOR(vmi, mm, 0); + struct vm_area_struct *vma; + + mmap_read_lock(mm); + for_each_vma(vmi, vma) { + if (!vma_is_special_mapping(vma, &vvar_mapping)) + continue; + zap_vma_pages(vma); + break; + } + mmap_read_unlock(mm); + return 0; +} +#endif + +static vm_fault_t vvar_fault(const struct vm_special_mapping *sm, + struct vm_area_struct *vma, struct vm_fault *vmf) +{ + struct page *timens_page = find_timens_vvar_page(vma); + unsigned long addr, pfn; + vm_fault_t err; + + switch (vmf->pgoff) { + case VVAR_DATA_PAGE_OFFSET: + pfn = virt_to_pfn(vdso_data); + if (timens_page) { + /* + * Fault in VVAR page too, since it will be accessed + * to get clock data anyway. + */ + addr = vmf->address + VVAR_TIMENS_PAGE_OFFSET * PAGE_SIZE; + err = vmf_insert_pfn(vma, addr, pfn); + if (unlikely(err & VM_FAULT_ERROR)) + return err; + pfn = page_to_pfn(timens_page); + } + break; +#ifdef CONFIG_TIME_NS + case VVAR_TIMENS_PAGE_OFFSET: + /* + * If a task belongs to a time namespace then a namespace + * specific VVAR is mapped with the VVAR_DATA_PAGE_OFFSET and + * the real VVAR page is mapped with the VVAR_TIMENS_PAGE_OFFSET + * offset. + * See also the comment near timens_setup_vdso_data(). + */ + if (!timens_page) + return VM_FAULT_SIGBUS; + pfn = virt_to_pfn(vdso_data); + break; +#endif /* CONFIG_TIME_NS */ + default: + return VM_FAULT_SIGBUS; + } + return vmf_insert_pfn(vma, vmf->address, pfn); +} + +static int vdso_mremap(const struct vm_special_mapping *sm, + struct vm_area_struct *vma) +{ + current->mm->context.vdso_base = vma->vm_start; + return 0; +} + +static struct vm_special_mapping vvar_mapping = { + .name = "[vvar]", + .fault = vvar_fault, +}; + +static struct vm_special_mapping vdso64_mapping = { + .name = "[vdso]", + .mremap = vdso_mremap, +}; + +static struct vm_special_mapping vdso32_mapping = { + .name = "[vdso]", + .mremap = vdso_mremap, +}; + +int vdso_getcpu_init(void) +{ + set_tod_programmable_field(smp_processor_id()); + return 0; +} +early_initcall(vdso_getcpu_init); /* Must be called before SMP init */ + +static int map_vdso(unsigned long addr, unsigned long vdso_mapping_len) +{ + unsigned long vvar_start, vdso_text_start, vdso_text_len; + struct vm_special_mapping *vdso_mapping; + struct mm_struct *mm = current->mm; + struct vm_area_struct *vma; + int rc; + + BUILD_BUG_ON(VVAR_NR_PAGES != __VVAR_PAGES); + if (mmap_write_lock_killable(mm)) + return -EINTR; + + if (is_compat_task()) { + vdso_text_len = vdso32_end - vdso32_start; + vdso_mapping = &vdso32_mapping; + } else { + vdso_text_len = vdso64_end - vdso64_start; + vdso_mapping = &vdso64_mapping; + } + vvar_start = get_unmapped_area(NULL, addr, vdso_mapping_len, 0, 0); + rc = vvar_start; + if (IS_ERR_VALUE(vvar_start)) + goto out; + vma = _install_special_mapping(mm, vvar_start, VVAR_NR_PAGES*PAGE_SIZE, + VM_READ|VM_MAYREAD|VM_IO|VM_DONTDUMP| + VM_PFNMAP, + &vvar_mapping); + rc = PTR_ERR(vma); + if (IS_ERR(vma)) + goto out; + vdso_text_start = vvar_start + VVAR_NR_PAGES * PAGE_SIZE; + /* VM_MAYWRITE for COW so gdb can set breakpoints */ + vma = _install_special_mapping(mm, vdso_text_start, vdso_text_len, + VM_READ|VM_EXEC| + VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC, + vdso_mapping); + if (IS_ERR(vma)) { + do_munmap(mm, vvar_start, PAGE_SIZE, NULL); + rc = PTR_ERR(vma); + } else { + current->mm->context.vdso_base = vdso_text_start; + rc = 0; + } +out: + mmap_write_unlock(mm); + return rc; +} + +static unsigned long vdso_addr(unsigned long start, unsigned long len) +{ + unsigned long addr, end, offset; + + /* + * Round up the start address. It can start out unaligned as a result + * of stack start randomization. + */ + start = PAGE_ALIGN(start); + + /* Round the lowest possible end address up to a PMD boundary. */ + end = (start + len + PMD_SIZE - 1) & PMD_MASK; + if (end >= VDSO_BASE) + end = VDSO_BASE; + end -= len; + + if (end > start) { + offset = get_random_u32_below(((end - start) >> PAGE_SHIFT) + 1); + addr = start + (offset << PAGE_SHIFT); + } else { + addr = start; + } + return addr; +} + +unsigned long vdso_size(void) +{ + unsigned long size = VVAR_NR_PAGES * PAGE_SIZE; + + if (is_compat_task()) + size += vdso32_end - vdso32_start; + else + size += vdso64_end - vdso64_start; + return PAGE_ALIGN(size); +} + +int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp) +{ + unsigned long addr = VDSO_BASE; + unsigned long size = vdso_size(); + + if (current->flags & PF_RANDOMIZE) + addr = vdso_addr(current->mm->start_stack + PAGE_SIZE, size); + return map_vdso(addr, size); +} + +static struct page ** __init vdso_setup_pages(void *start, void *end) +{ + int pages = (end - start) >> PAGE_SHIFT; + struct page **pagelist; + int i; + + pagelist = kcalloc(pages + 1, sizeof(struct page *), GFP_KERNEL); + if (!pagelist) + panic("%s: Cannot allocate page list for VDSO", __func__); + for (i = 0; i < pages; i++) + pagelist[i] = virt_to_page(start + i * PAGE_SIZE); + return pagelist; +} + +static int __init vdso_init(void) +{ + vdso64_mapping.pages = vdso_setup_pages(vdso64_start, vdso64_end); + if (IS_ENABLED(CONFIG_COMPAT)) + vdso32_mapping.pages = vdso_setup_pages(vdso32_start, vdso32_end); + return 0; +} +arch_initcall(vdso_init); diff --git a/arch/s390/kernel/vdso32/.gitignore b/arch/s390/kernel/vdso32/.gitignore new file mode 100644 index 0000000000..5167384843 --- /dev/null +++ b/arch/s390/kernel/vdso32/.gitignore @@ -0,0 +1,2 @@ +# SPDX-License-Identifier: GPL-2.0-only +vdso32.lds diff --git a/arch/s390/kernel/vdso32/Makefile b/arch/s390/kernel/vdso32/Makefile new file mode 100644 index 0000000000..23e868b79a --- /dev/null +++ b/arch/s390/kernel/vdso32/Makefile @@ -0,0 +1,80 @@ +# SPDX-License-Identifier: GPL-2.0 +# List of files in the vdso + +KCOV_INSTRUMENT := n + +# Include the generic Makefile to check the built vdso. +include $(srctree)/lib/vdso/Makefile +obj-vdso32 = vdso_user_wrapper-32.o note-32.o + +# Build rules + +targets := $(obj-vdso32) vdso32.so vdso32.so.dbg +obj-vdso32 := $(addprefix $(obj)/, $(obj-vdso32)) + +KBUILD_AFLAGS += -DBUILD_VDSO +KBUILD_CFLAGS += -DBUILD_VDSO -DDISABLE_BRANCH_PROFILING + +KBUILD_AFLAGS_32 := $(filter-out -m64,$(KBUILD_AFLAGS)) +KBUILD_AFLAGS_32 += -m31 -s + +KBUILD_CFLAGS_32 := $(filter-out -m64,$(KBUILD_CFLAGS)) +KBUILD_CFLAGS_32 := $(filter-out -mno-pic-data-is-text-relative,$(KBUILD_CFLAGS_32)) +KBUILD_CFLAGS_32 += -m31 -fPIC -shared -fno-common -fno-builtin + +LDFLAGS_vdso32.so.dbg += -fPIC -shared -soname=linux-vdso32.so.1 \ + --hash-style=both --build-id=sha1 -melf_s390 -T + +$(targets:%=$(obj)/%.dbg): KBUILD_CFLAGS = $(KBUILD_CFLAGS_32) +$(targets:%=$(obj)/%.dbg): KBUILD_AFLAGS = $(KBUILD_AFLAGS_32) + +obj-y += vdso32_wrapper.o +targets += vdso32.lds +CPPFLAGS_vdso32.lds += -P -C -U$(ARCH) + +# Disable gcov profiling, ubsan and kasan for VDSO code +GCOV_PROFILE := n +UBSAN_SANITIZE := n +KASAN_SANITIZE := n +KCSAN_SANITIZE := n + +# Force dependency (incbin is bad) +$(obj)/vdso32_wrapper.o : $(obj)/vdso32.so + +quiet_cmd_vdso_and_check = VDSO $@ + cmd_vdso_and_check = $(cmd_ld); $(cmd_vdso_check) + +$(obj)/vdso32.so.dbg: $(src)/vdso32.lds $(obj-vdso32) FORCE + $(call if_changed,vdso_and_check) + +# strip rule for the .so file +$(obj)/%.so: OBJCOPYFLAGS := -S +$(obj)/%.so: $(obj)/%.so.dbg FORCE + $(call if_changed,objcopy) + +$(obj-vdso32): %-32.o: %.S FORCE + $(call if_changed_dep,vdso32as) + +# actual build commands +quiet_cmd_vdso32as = VDSO32A $@ + cmd_vdso32as = $(CC) $(a_flags) -c -o $@ $< +quiet_cmd_vdso32cc = VDSO32C $@ + cmd_vdso32cc = $(CC) $(c_flags) -c -o $@ $< + +# install commands for the unstripped file +quiet_cmd_vdso_install = INSTALL $@ + cmd_vdso_install = cp $(obj)/$@.dbg $(MODLIB)/vdso/$@ + +vdso32.so: $(obj)/vdso32.so.dbg + @mkdir -p $(MODLIB)/vdso + $(call cmd,vdso_install) + +vdso_install: vdso32.so + +# Generate VDSO offsets using helper script +gen-vdsosym := $(srctree)/$(src)/gen_vdso_offsets.sh +quiet_cmd_vdsosym = VDSOSYM $@ + cmd_vdsosym = $(NM) $< | $(gen-vdsosym) | LC_ALL=C sort > $@ + +include/generated/vdso32-offsets.h: $(obj)/vdso32.so.dbg FORCE + $(call if_changed,vdsosym) diff --git a/arch/s390/kernel/vdso32/gen_vdso_offsets.sh b/arch/s390/kernel/vdso32/gen_vdso_offsets.sh new file mode 100755 index 0000000000..9c4f951e22 --- /dev/null +++ b/arch/s390/kernel/vdso32/gen_vdso_offsets.sh @@ -0,0 +1,15 @@ +#!/bin/sh +# SPDX-License-Identifier: GPL-2.0 + +# +# Match symbols in the DSO that look like VDSO_*; produce a header file +# of constant offsets into the shared object. +# +# Doing this inside the Makefile will break the $(filter-out) function, +# causing Kbuild to rebuild the vdso-offsets header file every time. +# +# Inspired by arm64 version. +# + +LC_ALL=C +sed -n 's/\([0-9a-f]*\) . __kernel_compat_\(.*\)/\#define vdso32_offset_\2\t0x\1/p' diff --git a/arch/s390/kernel/vdso32/note.S b/arch/s390/kernel/vdso32/note.S new file mode 100644 index 0000000000..db19d0680a --- /dev/null +++ b/arch/s390/kernel/vdso32/note.S @@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * This supplies .note.* sections to go into the PT_NOTE inside the vDSO text. + * Here we can supply some information useful to userland. + */ + +#include <linux/uts.h> +#include <linux/version.h> +#include <linux/elfnote.h> + +ELFNOTE_START(Linux, 0, "a") + .long LINUX_VERSION_CODE +ELFNOTE_END diff --git a/arch/s390/kernel/vdso32/vdso32.lds.S b/arch/s390/kernel/vdso32/vdso32.lds.S new file mode 100644 index 0000000000..edf5ff1deb --- /dev/null +++ b/arch/s390/kernel/vdso32/vdso32.lds.S @@ -0,0 +1,142 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * This is the infamous ld script for the 64 bits vdso + * library + */ + +#include <asm/page.h> +#include <asm/vdso.h> + +OUTPUT_FORMAT("elf32-s390", "elf32-s390", "elf32-s390") +OUTPUT_ARCH(s390:31-bit) +ENTRY(_start) + +SECTIONS +{ + PROVIDE(_vdso_data = . - __VVAR_PAGES * PAGE_SIZE); +#ifdef CONFIG_TIME_NS + PROVIDE(_timens_data = _vdso_data + PAGE_SIZE); +#endif + . = VDSO_LBASE + SIZEOF_HEADERS; + + .hash : { *(.hash) } :text + .gnu.hash : { *(.gnu.hash) } + .dynsym : { *(.dynsym) } + .dynstr : { *(.dynstr) } + .gnu.version : { *(.gnu.version) } + .gnu.version_d : { *(.gnu.version_d) } + .gnu.version_r : { *(.gnu.version_r) } + + .note : { *(.note.*) } :text :note + + . = ALIGN(16); + .text : { + *(.text .stub .text.* .gnu.linkonce.t.*) + } :text + PROVIDE(__etext = .); + PROVIDE(_etext = .); + PROVIDE(etext = .); + + /* + * Other stuff is appended to the text segment: + */ + .rodata : { *(.rodata .rodata.* .gnu.linkonce.r.*) } + .rodata1 : { *(.rodata1) } + + .dynamic : { *(.dynamic) } :text :dynamic + + .eh_frame_hdr : { *(.eh_frame_hdr) } :text :eh_frame_hdr + .eh_frame : { KEEP (*(.eh_frame)) } :text + .gcc_except_table : { *(.gcc_except_table .gcc_except_table.*) } + + .rela.dyn ALIGN(8) : { *(.rela.dyn) } + .got ALIGN(8) : { *(.got .toc) } + .got.plt ALIGN(8) : { *(.got.plt) } + + _end = .; + PROVIDE(end = .); + + /* + * Stabs debugging sections are here too. + */ + .stab 0 : { *(.stab) } + .stabstr 0 : { *(.stabstr) } + .stab.excl 0 : { *(.stab.excl) } + .stab.exclstr 0 : { *(.stab.exclstr) } + .stab.index 0 : { *(.stab.index) } + .stab.indexstr 0 : { *(.stab.indexstr) } + .comment 0 : { *(.comment) } + + /* + * DWARF debug sections. + * Symbols in the DWARF debugging sections are relative to the + * beginning of the section so we begin them at 0. + */ + /* DWARF 1 */ + .debug 0 : { *(.debug) } + .line 0 : { *(.line) } + /* GNU DWARF 1 extensions */ + .debug_srcinfo 0 : { *(.debug_srcinfo) } + .debug_sfnames 0 : { *(.debug_sfnames) } + /* DWARF 1.1 and DWARF 2 */ + .debug_aranges 0 : { *(.debug_aranges) } + .debug_pubnames 0 : { *(.debug_pubnames) } + /* DWARF 2 */ + .debug_info 0 : { *(.debug_info .gnu.linkonce.wi.*) } + .debug_abbrev 0 : { *(.debug_abbrev) } + .debug_line 0 : { *(.debug_line) } + .debug_frame 0 : { *(.debug_frame) } + .debug_str 0 : { *(.debug_str) } + .debug_loc 0 : { *(.debug_loc) } + .debug_macinfo 0 : { *(.debug_macinfo) } + /* SGI/MIPS DWARF 2 extensions */ + .debug_weaknames 0 : { *(.debug_weaknames) } + .debug_funcnames 0 : { *(.debug_funcnames) } + .debug_typenames 0 : { *(.debug_typenames) } + .debug_varnames 0 : { *(.debug_varnames) } + /* DWARF 3 */ + .debug_pubtypes 0 : { *(.debug_pubtypes) } + .debug_ranges 0 : { *(.debug_ranges) } + .gnu.attributes 0 : { KEEP (*(.gnu.attributes)) } + + /DISCARD/ : { + *(.note.GNU-stack) + *(.branch_lt) + *(.data .data.* .gnu.linkonce.d.* .sdata*) + *(.bss .sbss .dynbss .dynsbss) + } +} + +/* + * Very old versions of ld do not recognize this name token; use the constant. + */ +#define PT_GNU_EH_FRAME 0x6474e550 + +/* + * We must supply the ELF program headers explicitly to get just one + * PT_LOAD segment, and set the flags explicitly to make segments read-only. + */ +PHDRS +{ + text PT_LOAD FILEHDR PHDRS FLAGS(5); /* PF_R|PF_X */ + dynamic PT_DYNAMIC FLAGS(4); /* PF_R */ + note PT_NOTE FLAGS(4); /* PF_R */ + eh_frame_hdr PT_GNU_EH_FRAME; +} + +/* + * This controls what symbols we export from the DSO. + */ +VERSION +{ + VDSO_VERSION_STRING { + global: + /* + * Has to be there for the kernel to find + */ + __kernel_compat_restart_syscall; + __kernel_compat_rt_sigreturn; + __kernel_compat_sigreturn; + local: *; + }; +} diff --git a/arch/s390/kernel/vdso32/vdso32_wrapper.S b/arch/s390/kernel/vdso32/vdso32_wrapper.S new file mode 100644 index 0000000000..de2fb93047 --- /dev/null +++ b/arch/s390/kernel/vdso32/vdso32_wrapper.S @@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#include <linux/init.h> +#include <linux/linkage.h> +#include <asm/page.h> + + __PAGE_ALIGNED_DATA + + .globl vdso32_start, vdso32_end + .balign PAGE_SIZE +vdso32_start: + .incbin "arch/s390/kernel/vdso32/vdso32.so" + .balign PAGE_SIZE +vdso32_end: + + .previous diff --git a/arch/s390/kernel/vdso32/vdso_user_wrapper.S b/arch/s390/kernel/vdso32/vdso_user_wrapper.S new file mode 100644 index 0000000000..2e645003fd --- /dev/null +++ b/arch/s390/kernel/vdso32/vdso_user_wrapper.S @@ -0,0 +1,22 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#include <linux/linkage.h> +#include <asm/unistd.h> +#include <asm/dwarf.h> + +.macro vdso_syscall func,syscall + .globl __kernel_compat_\func + .type __kernel_compat_\func,@function + __ALIGN +__kernel_compat_\func: + CFI_STARTPROC + svc \syscall + /* Make sure we notice when a syscall returns, which shouldn't happen */ + .word 0 + CFI_ENDPROC + .size __kernel_compat_\func,.-__kernel_compat_\func +.endm + +vdso_syscall restart_syscall,__NR_restart_syscall +vdso_syscall sigreturn,__NR_sigreturn +vdso_syscall rt_sigreturn,__NR_rt_sigreturn diff --git a/arch/s390/kernel/vdso64/.gitignore b/arch/s390/kernel/vdso64/.gitignore new file mode 100644 index 0000000000..4ec80685fe --- /dev/null +++ b/arch/s390/kernel/vdso64/.gitignore @@ -0,0 +1,2 @@ +# SPDX-License-Identifier: GPL-2.0-only +vdso64.lds diff --git a/arch/s390/kernel/vdso64/Makefile b/arch/s390/kernel/vdso64/Makefile new file mode 100644 index 0000000000..fc1c6ff817 --- /dev/null +++ b/arch/s390/kernel/vdso64/Makefile @@ -0,0 +1,89 @@ +# SPDX-License-Identifier: GPL-2.0 +# List of files in the vdso + +KCOV_INSTRUMENT := n + +# Include the generic Makefile to check the built vdso. +include $(srctree)/lib/vdso/Makefile +obj-vdso64 = vdso_user_wrapper.o note.o +obj-cvdso64 = vdso64_generic.o getcpu.o +VDSO_CFLAGS_REMOVE := -pg $(CC_FLAGS_FTRACE) $(CC_FLAGS_EXPOLINE) $(CC_FLAGS_CHECK_STACK) +CFLAGS_REMOVE_getcpu.o = $(VDSO_CFLAGS_REMOVE) +CFLAGS_REMOVE_vdso64_generic.o = $(VDSO_CFLAGS_REMOVE) + +# Build rules + +targets := $(obj-vdso64) $(obj-cvdso64) vdso64.so vdso64.so.dbg +obj-vdso64 := $(addprefix $(obj)/, $(obj-vdso64)) +obj-cvdso64 := $(addprefix $(obj)/, $(obj-cvdso64)) + +KBUILD_AFLAGS += -DBUILD_VDSO +KBUILD_CFLAGS += -DBUILD_VDSO -DDISABLE_BRANCH_PROFILING + +KBUILD_AFLAGS_64 := $(filter-out -m64,$(KBUILD_AFLAGS)) +KBUILD_AFLAGS_64 += -m64 + +KBUILD_CFLAGS_64 := $(filter-out -m64,$(KBUILD_CFLAGS)) +KBUILD_CFLAGS_64 := $(filter-out -mno-pic-data-is-text-relative,$(KBUILD_CFLAGS_64)) +KBUILD_CFLAGS_64 += -m64 -fPIC -fno-common -fno-builtin +ldflags-y := -fPIC -shared -soname=linux-vdso64.so.1 \ + --hash-style=both --build-id=sha1 -T + +$(targets:%=$(obj)/%.dbg): KBUILD_CFLAGS = $(KBUILD_CFLAGS_64) +$(targets:%=$(obj)/%.dbg): KBUILD_AFLAGS = $(KBUILD_AFLAGS_64) + +obj-y += vdso64_wrapper.o +targets += vdso64.lds +CPPFLAGS_vdso64.lds += -P -C -U$(ARCH) + +# Disable gcov profiling, ubsan and kasan for VDSO code +GCOV_PROFILE := n +UBSAN_SANITIZE := n +KASAN_SANITIZE := n +KCSAN_SANITIZE := n + +# Force dependency (incbin is bad) +$(obj)/vdso64_wrapper.o : $(obj)/vdso64.so + +quiet_cmd_vdso_and_check = VDSO $@ + cmd_vdso_and_check = $(cmd_ld); $(cmd_vdso_check) + +# link rule for the .so file, .lds has to be first +$(obj)/vdso64.so.dbg: $(src)/vdso64.lds $(obj-vdso64) $(obj-cvdso64) FORCE + $(call if_changed,vdso_and_check) + +# strip rule for the .so file +$(obj)/%.so: OBJCOPYFLAGS := -S +$(obj)/%.so: $(obj)/%.so.dbg FORCE + $(call if_changed,objcopy) + +# assembly rules for the .S files +$(obj-vdso64): %.o: %.S FORCE + $(call if_changed_dep,vdso64as) + +$(obj-cvdso64): %.o: %.c FORCE + $(call if_changed_dep,vdso64cc) + +# actual build commands +quiet_cmd_vdso64as = VDSO64A $@ + cmd_vdso64as = $(CC) $(a_flags) -c -o $@ $< +quiet_cmd_vdso64cc = VDSO64C $@ + cmd_vdso64cc = $(CC) $(c_flags) -c -o $@ $< + +# install commands for the unstripped file +quiet_cmd_vdso_install = INSTALL $@ + cmd_vdso_install = cp $(obj)/$@.dbg $(MODLIB)/vdso/$@ + +vdso64.so: $(obj)/vdso64.so.dbg + @mkdir -p $(MODLIB)/vdso + $(call cmd,vdso_install) + +vdso_install: vdso64.so + +# Generate VDSO offsets using helper script +gen-vdsosym := $(srctree)/$(src)/gen_vdso_offsets.sh +quiet_cmd_vdsosym = VDSOSYM $@ + cmd_vdsosym = $(NM) $< | $(gen-vdsosym) | LC_ALL=C sort > $@ + +include/generated/vdso64-offsets.h: $(obj)/vdso64.so.dbg FORCE + $(call if_changed,vdsosym) diff --git a/arch/s390/kernel/vdso64/gen_vdso_offsets.sh b/arch/s390/kernel/vdso64/gen_vdso_offsets.sh new file mode 100755 index 0000000000..37f05cb38d --- /dev/null +++ b/arch/s390/kernel/vdso64/gen_vdso_offsets.sh @@ -0,0 +1,15 @@ +#!/bin/sh +# SPDX-License-Identifier: GPL-2.0 + +# +# Match symbols in the DSO that look like VDSO_*; produce a header file +# of constant offsets into the shared object. +# +# Doing this inside the Makefile will break the $(filter-out) function, +# causing Kbuild to rebuild the vdso-offsets header file every time. +# +# Inspired by arm64 version. +# + +LC_ALL=C +sed -n 's/\([0-9a-f]*\) . __kernel_\(.*\)/\#define vdso64_offset_\2\t0x\1/p' diff --git a/arch/s390/kernel/vdso64/getcpu.c b/arch/s390/kernel/vdso64/getcpu.c new file mode 100644 index 0000000000..5c5d4a848b --- /dev/null +++ b/arch/s390/kernel/vdso64/getcpu.c @@ -0,0 +1,21 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright IBM Corp. 2020 */ + +#include <linux/compiler.h> +#include <linux/getcpu.h> +#include <asm/timex.h> +#include "vdso.h" + +int __s390_vdso_getcpu(unsigned *cpu, unsigned *node, struct getcpu_cache *unused) +{ + union tod_clock clk; + + /* CPU number is stored in the programmable field of the TOD clock */ + store_tod_clock_ext(&clk); + if (cpu) + *cpu = clk.pf; + /* NUMA node is always zero */ + if (node) + *node = 0; + return 0; +} diff --git a/arch/s390/kernel/vdso64/note.S b/arch/s390/kernel/vdso64/note.S new file mode 100644 index 0000000000..db19d0680a --- /dev/null +++ b/arch/s390/kernel/vdso64/note.S @@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * This supplies .note.* sections to go into the PT_NOTE inside the vDSO text. + * Here we can supply some information useful to userland. + */ + +#include <linux/uts.h> +#include <linux/version.h> +#include <linux/elfnote.h> + +ELFNOTE_START(Linux, 0, "a") + .long LINUX_VERSION_CODE +ELFNOTE_END diff --git a/arch/s390/kernel/vdso64/vdso.h b/arch/s390/kernel/vdso64/vdso.h new file mode 100644 index 0000000000..34c7a2312f --- /dev/null +++ b/arch/s390/kernel/vdso64/vdso.h @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __ARCH_S390_KERNEL_VDSO64_VDSO_H +#define __ARCH_S390_KERNEL_VDSO64_VDSO_H + +#include <vdso/datapage.h> + +struct getcpu_cache; + +int __s390_vdso_getcpu(unsigned *cpu, unsigned *node, struct getcpu_cache *unused); +int __s390_vdso_gettimeofday(struct __kernel_old_timeval *tv, struct timezone *tz); +int __s390_vdso_clock_gettime(clockid_t clock, struct __kernel_timespec *ts); +int __s390_vdso_clock_getres(clockid_t clock, struct __kernel_timespec *ts); + +#endif /* __ARCH_S390_KERNEL_VDSO64_VDSO_H */ diff --git a/arch/s390/kernel/vdso64/vdso64.lds.S b/arch/s390/kernel/vdso64/vdso64.lds.S new file mode 100644 index 0000000000..4461ea151e --- /dev/null +++ b/arch/s390/kernel/vdso64/vdso64.lds.S @@ -0,0 +1,146 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * This is the infamous ld script for the 64 bits vdso + * library + */ + +#include <asm/page.h> +#include <asm/vdso.h> + +OUTPUT_FORMAT("elf64-s390", "elf64-s390", "elf64-s390") +OUTPUT_ARCH(s390:64-bit) +ENTRY(_start) + +SECTIONS +{ + PROVIDE(_vdso_data = . - __VVAR_PAGES * PAGE_SIZE); +#ifdef CONFIG_TIME_NS + PROVIDE(_timens_data = _vdso_data + PAGE_SIZE); +#endif + . = VDSO_LBASE + SIZEOF_HEADERS; + + .hash : { *(.hash) } :text + .gnu.hash : { *(.gnu.hash) } + .dynsym : { *(.dynsym) } + .dynstr : { *(.dynstr) } + .gnu.version : { *(.gnu.version) } + .gnu.version_d : { *(.gnu.version_d) } + .gnu.version_r : { *(.gnu.version_r) } + + .note : { *(.note.*) } :text :note + + . = ALIGN(16); + .text : { + *(.text .stub .text.* .gnu.linkonce.t.*) + } :text + PROVIDE(__etext = .); + PROVIDE(_etext = .); + PROVIDE(etext = .); + + /* + * Other stuff is appended to the text segment: + */ + .rodata : { *(.rodata .rodata.* .gnu.linkonce.r.*) } + .rodata1 : { *(.rodata1) } + + .dynamic : { *(.dynamic) } :text :dynamic + + .eh_frame_hdr : { *(.eh_frame_hdr) } :text :eh_frame_hdr + .eh_frame : { KEEP (*(.eh_frame)) } :text + .gcc_except_table : { *(.gcc_except_table .gcc_except_table.*) } + + .rela.dyn ALIGN(8) : { *(.rela.dyn) } + .got ALIGN(8) : { *(.got .toc) } + .got.plt ALIGN(8) : { *(.got.plt) } + + _end = .; + PROVIDE(end = .); + + /* + * Stabs debugging sections are here too. + */ + .stab 0 : { *(.stab) } + .stabstr 0 : { *(.stabstr) } + .stab.excl 0 : { *(.stab.excl) } + .stab.exclstr 0 : { *(.stab.exclstr) } + .stab.index 0 : { *(.stab.index) } + .stab.indexstr 0 : { *(.stab.indexstr) } + .comment 0 : { *(.comment) } + + /* + * DWARF debug sections. + * Symbols in the DWARF debugging sections are relative to the + * beginning of the section so we begin them at 0. + */ + /* DWARF 1 */ + .debug 0 : { *(.debug) } + .line 0 : { *(.line) } + /* GNU DWARF 1 extensions */ + .debug_srcinfo 0 : { *(.debug_srcinfo) } + .debug_sfnames 0 : { *(.debug_sfnames) } + /* DWARF 1.1 and DWARF 2 */ + .debug_aranges 0 : { *(.debug_aranges) } + .debug_pubnames 0 : { *(.debug_pubnames) } + /* DWARF 2 */ + .debug_info 0 : { *(.debug_info .gnu.linkonce.wi.*) } + .debug_abbrev 0 : { *(.debug_abbrev) } + .debug_line 0 : { *(.debug_line) } + .debug_frame 0 : { *(.debug_frame) } + .debug_str 0 : { *(.debug_str) } + .debug_loc 0 : { *(.debug_loc) } + .debug_macinfo 0 : { *(.debug_macinfo) } + /* SGI/MIPS DWARF 2 extensions */ + .debug_weaknames 0 : { *(.debug_weaknames) } + .debug_funcnames 0 : { *(.debug_funcnames) } + .debug_typenames 0 : { *(.debug_typenames) } + .debug_varnames 0 : { *(.debug_varnames) } + /* DWARF 3 */ + .debug_pubtypes 0 : { *(.debug_pubtypes) } + .debug_ranges 0 : { *(.debug_ranges) } + .gnu.attributes 0 : { KEEP (*(.gnu.attributes)) } + + /DISCARD/ : { + *(.note.GNU-stack) + *(.branch_lt) + *(.data .data.* .gnu.linkonce.d.* .sdata*) + *(.bss .sbss .dynbss .dynsbss) + } +} + +/* + * Very old versions of ld do not recognize this name token; use the constant. + */ +#define PT_GNU_EH_FRAME 0x6474e550 + +/* + * We must supply the ELF program headers explicitly to get just one + * PT_LOAD segment, and set the flags explicitly to make segments read-only. + */ +PHDRS +{ + text PT_LOAD FILEHDR PHDRS FLAGS(5); /* PF_R|PF_X */ + dynamic PT_DYNAMIC FLAGS(4); /* PF_R */ + note PT_NOTE FLAGS(4); /* PF_R */ + eh_frame_hdr PT_GNU_EH_FRAME; +} + +/* + * This controls what symbols we export from the DSO. + */ +VERSION +{ + VDSO_VERSION_STRING { + global: + /* + * Has to be there for the kernel to find + */ + __kernel_gettimeofday; + __kernel_clock_gettime; + __kernel_clock_getres; + __kernel_getcpu; + __kernel_restart_syscall; + __kernel_rt_sigreturn; + __kernel_sigreturn; + local: *; + }; +} diff --git a/arch/s390/kernel/vdso64/vdso64_generic.c b/arch/s390/kernel/vdso64/vdso64_generic.c new file mode 100644 index 0000000000..a9aa75643c --- /dev/null +++ b/arch/s390/kernel/vdso64/vdso64_generic.c @@ -0,0 +1,19 @@ +// SPDX-License-Identifier: GPL-2.0 +#include "../../../../lib/vdso/gettimeofday.c" +#include "vdso.h" + +int __s390_vdso_gettimeofday(struct __kernel_old_timeval *tv, + struct timezone *tz) +{ + return __cvdso_gettimeofday(tv, tz); +} + +int __s390_vdso_clock_gettime(clockid_t clock, struct __kernel_timespec *ts) +{ + return __cvdso_clock_gettime(clock, ts); +} + +int __s390_vdso_clock_getres(clockid_t clock, struct __kernel_timespec *ts) +{ + return __cvdso_clock_getres(clock, ts); +} diff --git a/arch/s390/kernel/vdso64/vdso64_wrapper.S b/arch/s390/kernel/vdso64/vdso64_wrapper.S new file mode 100644 index 0000000000..6721849986 --- /dev/null +++ b/arch/s390/kernel/vdso64/vdso64_wrapper.S @@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#include <linux/init.h> +#include <linux/linkage.h> +#include <asm/page.h> + + __PAGE_ALIGNED_DATA + + .globl vdso64_start, vdso64_end + .balign PAGE_SIZE +vdso64_start: + .incbin "arch/s390/kernel/vdso64/vdso64.so" + .balign PAGE_SIZE +vdso64_end: + + .previous diff --git a/arch/s390/kernel/vdso64/vdso_user_wrapper.S b/arch/s390/kernel/vdso64/vdso_user_wrapper.S new file mode 100644 index 0000000000..57f62596e5 --- /dev/null +++ b/arch/s390/kernel/vdso64/vdso_user_wrapper.S @@ -0,0 +1,57 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#include <linux/linkage.h> +#include <asm/vdso.h> +#include <asm/unistd.h> +#include <asm/asm-offsets.h> +#include <asm/dwarf.h> +#include <asm/ptrace.h> + +#define WRAPPER_FRAME_SIZE (STACK_FRAME_OVERHEAD+8) + +/* + * Older glibc version called vdso without allocating a stackframe. This wrapper + * is just used to allocate a stackframe. See + * https://sourceware.org/git/?p=glibc.git;a=commit;h=478593e6374f3818da39332260dc453cb19cfa1e + * for details. + */ +.macro vdso_func func + .globl __kernel_\func + .type __kernel_\func,@function + __ALIGN +__kernel_\func: + CFI_STARTPROC + aghi %r15,-WRAPPER_FRAME_SIZE + CFI_DEF_CFA_OFFSET (STACK_FRAME_OVERHEAD + WRAPPER_FRAME_SIZE) + CFI_VAL_OFFSET 15, -STACK_FRAME_OVERHEAD + stg %r14,STACK_FRAME_OVERHEAD(%r15) + brasl %r14,__s390_vdso_\func + lg %r14,STACK_FRAME_OVERHEAD(%r15) + aghi %r15,WRAPPER_FRAME_SIZE + CFI_DEF_CFA_OFFSET STACK_FRAME_OVERHEAD + CFI_RESTORE 15 + br %r14 + CFI_ENDPROC + .size __kernel_\func,.-__kernel_\func +.endm + +vdso_func gettimeofday +vdso_func clock_getres +vdso_func clock_gettime +vdso_func getcpu + +.macro vdso_syscall func,syscall + .globl __kernel_\func + .type __kernel_\func,@function + __ALIGN +__kernel_\func: + CFI_STARTPROC + svc \syscall + /* Make sure we notice when a syscall returns, which shouldn't happen */ + .word 0 + CFI_ENDPROC + .size __kernel_\func,.-__kernel_\func +.endm + +vdso_syscall restart_syscall,__NR_restart_syscall +vdso_syscall sigreturn,__NR_sigreturn +vdso_syscall rt_sigreturn,__NR_rt_sigreturn diff --git a/arch/s390/kernel/vmlinux.lds.S b/arch/s390/kernel/vmlinux.lds.S new file mode 100644 index 0000000000..2ae201ebf9 --- /dev/null +++ b/arch/s390/kernel/vmlinux.lds.S @@ -0,0 +1,243 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* ld script to make s390 Linux kernel + * Written by Martin Schwidefsky (schwidefsky@de.ibm.com) + */ + +#include <asm/thread_info.h> +#include <asm/page.h> +#include <asm/ftrace.lds.h> + +/* + * Put .bss..swapper_pg_dir as the first thing in .bss. This will + * make sure it has 16k alignment. + */ +#define BSS_FIRST_SECTIONS *(.bss..swapper_pg_dir) \ + *(.bss..invalid_pg_dir) + +#define RO_EXCEPTION_TABLE_ALIGN 16 + +/* Handle ro_after_init data on our own. */ +#define RO_AFTER_INIT_DATA + +#define RUNTIME_DISCARD_EXIT + +#define EMITS_PT_NOTE + +#include <asm-generic/vmlinux.lds.h> +#include <asm/vmlinux.lds.h> + +OUTPUT_FORMAT("elf64-s390", "elf64-s390", "elf64-s390") +OUTPUT_ARCH(s390:64-bit) +ENTRY(startup_continue) +jiffies = jiffies_64; + +PHDRS { + text PT_LOAD FLAGS(5); /* R_E */ + data PT_LOAD FLAGS(7); /* RWE */ + note PT_NOTE FLAGS(0); /* ___ */ +} + +SECTIONS +{ + . = 0x100000; + .text : { + _stext = .; /* Start of text section */ + _text = .; /* Text and read-only data */ + HEAD_TEXT + TEXT_TEXT + SCHED_TEXT + LOCK_TEXT + KPROBES_TEXT + IRQENTRY_TEXT + SOFTIRQENTRY_TEXT + FTRACE_HOTPATCH_TRAMPOLINES_TEXT + *(.text.*_indirect_*) + *(.gnu.warning) + . = ALIGN(PAGE_SIZE); + _etext = .; /* End of text section */ + } :text = 0x0700 + + RO_DATA(PAGE_SIZE) + + . = ALIGN(PAGE_SIZE); + _sdata = .; /* Start of data section */ + + . = ALIGN(PAGE_SIZE); + __start_ro_after_init = .; + .data..ro_after_init : { + *(.data..ro_after_init) + JUMP_TABLE_DATA + } :data + . = ALIGN(PAGE_SIZE); + __end_ro_after_init = .; + + RW_DATA(0x100, PAGE_SIZE, THREAD_SIZE) + BOOT_DATA_PRESERVED + + . = ALIGN(8); + .amode31.refs : { + _start_amode31_refs = .; + *(.amode31.refs) + _end_amode31_refs = .; + } + + . = ALIGN(PAGE_SIZE); + _edata = .; /* End of data section */ + + /* will be freed after init */ + . = ALIGN(PAGE_SIZE); /* Init code and data */ + __init_begin = .; + + . = ALIGN(PAGE_SIZE); + .init.text : AT(ADDR(.init.text) - LOAD_OFFSET) { + _sinittext = .; + INIT_TEXT + . = ALIGN(PAGE_SIZE); + _einittext = .; + } + + /* + * .exit.text is discarded at runtime, not link time, + * to deal with references from __bug_table + */ + .exit.text : { + EXIT_TEXT + } + + .exit.data : { + EXIT_DATA + } + + /* + * struct alt_inst entries. From the header (alternative.h): + * "Alternative instructions for different CPU types or capabilities" + * Think locking instructions on spinlocks. + * Note, that it is a part of __init region. + */ + . = ALIGN(8); + .altinstructions : { + __alt_instructions = .; + *(.altinstructions) + __alt_instructions_end = .; + } + + /* + * And here are the replacement instructions. The linker sticks + * them as binary blobs. The .altinstructions has enough data to + * get the address and the length of them to patch the kernel safely. + * Note, that it is a part of __init region. + */ + .altinstr_replacement : { + *(.altinstr_replacement) + } + + /* + * Table with the patch locations to undo expolines + */ + . = ALIGN(4); + .nospec_call_table : { + __nospec_call_start = . ; + *(.s390_indirect*) + __nospec_call_end = . ; + } + .nospec_return_table : { + __nospec_return_start = . ; + *(.s390_return*) + __nospec_return_end = . ; + } + + BOOT_DATA + + /* + * .amode31 section for code, data, ex_table that need to stay + * below 2 GB, even when the kernel is relocated above 2 GB. + */ + . = ALIGN(PAGE_SIZE); + _samode31 = .; + .amode31.text : { + _stext_amode31 = .; + *(.amode31.text) + *(.amode31.text.*_indirect_*) + . = ALIGN(PAGE_SIZE); + _etext_amode31 = .; + } + . = ALIGN(16); + .amode31.ex_table : { + _start_amode31_ex_table = .; + KEEP(*(.amode31.ex_table)) + _stop_amode31_ex_table = .; + } + . = ALIGN(PAGE_SIZE); + .amode31.data : { + *(.amode31.data) + } + . = ALIGN(PAGE_SIZE); + _eamode31 = .; + + /* early.c uses stsi, which requires page aligned data. */ + . = ALIGN(PAGE_SIZE); + INIT_DATA_SECTION(0x100) + + PERCPU_SECTION(0x100) + + .dynsym ALIGN(8) : { + __dynsym_start = .; + *(.dynsym) + __dynsym_end = .; + } + .rela.dyn ALIGN(8) : { + __rela_dyn_start = .; + *(.rela*) + __rela_dyn_end = .; + } + + . = ALIGN(PAGE_SIZE); + __init_end = .; /* freed after init ends here */ + + BSS_SECTION(PAGE_SIZE, 4 * PAGE_SIZE, PAGE_SIZE) + + . = ALIGN(PAGE_SIZE); + _end = . ; + + /* + * uncompressed image info used by the decompressor + * it should match struct vmlinux_info + */ + .vmlinux.info 0 (INFO) : { + QUAD(_stext) /* default_lma */ + QUAD(startup_continue) /* entry */ + QUAD(__bss_start - _stext) /* image_size */ + QUAD(__bss_stop - __bss_start) /* bss_size */ + QUAD(__boot_data_start) /* bootdata_off */ + QUAD(__boot_data_end - __boot_data_start) /* bootdata_size */ + QUAD(__boot_data_preserved_start) /* bootdata_preserved_off */ + QUAD(__boot_data_preserved_end - + __boot_data_preserved_start) /* bootdata_preserved_size */ + QUAD(__dynsym_start) /* dynsym_start */ + QUAD(__rela_dyn_start) /* rela_dyn_start */ + QUAD(__rela_dyn_end) /* rela_dyn_end */ + QUAD(_eamode31 - _samode31) /* amode31_size */ + QUAD(init_mm) + QUAD(swapper_pg_dir) + QUAD(invalid_pg_dir) +#ifdef CONFIG_KASAN + QUAD(kasan_early_shadow_page) + QUAD(kasan_early_shadow_pte) + QUAD(kasan_early_shadow_pmd) + QUAD(kasan_early_shadow_pud) + QUAD(kasan_early_shadow_p4d) +#endif + } :NONE + + /* Debugging sections. */ + STABS_DEBUG + DWARF_DEBUG + ELF_DETAILS + + /* Sections to be discarded */ + DISCARDS + /DISCARD/ : { + *(.eh_frame) + *(.interp) + } +} diff --git a/arch/s390/kernel/vtime.c b/arch/s390/kernel/vtime.c new file mode 100644 index 0000000000..e0a88dcaf5 --- /dev/null +++ b/arch/s390/kernel/vtime.c @@ -0,0 +1,454 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Virtual cpu timer based timer functions. + * + * Copyright IBM Corp. 2004, 2012 + * Author(s): Jan Glauber <jan.glauber@de.ibm.com> + */ + +#include <linux/kernel_stat.h> +#include <linux/export.h> +#include <linux/kernel.h> +#include <linux/timex.h> +#include <linux/types.h> +#include <linux/time.h> +#include <asm/alternative.h> +#include <asm/cputime.h> +#include <asm/vtimer.h> +#include <asm/vtime.h> +#include <asm/cpu_mf.h> +#include <asm/smp.h> + +#include "entry.h" + +static void virt_timer_expire(void); + +static LIST_HEAD(virt_timer_list); +static DEFINE_SPINLOCK(virt_timer_lock); +static atomic64_t virt_timer_current; +static atomic64_t virt_timer_elapsed; + +DEFINE_PER_CPU(u64, mt_cycles[8]); +static DEFINE_PER_CPU(u64, mt_scaling_mult) = { 1 }; +static DEFINE_PER_CPU(u64, mt_scaling_div) = { 1 }; +static DEFINE_PER_CPU(u64, mt_scaling_jiffies); + +static inline u64 get_vtimer(void) +{ + u64 timer; + + asm volatile("stpt %0" : "=Q" (timer)); + return timer; +} + +static inline void set_vtimer(u64 expires) +{ + u64 timer; + + asm volatile( + " stpt %0\n" /* Store current cpu timer value */ + " spt %1" /* Set new value imm. afterwards */ + : "=Q" (timer) : "Q" (expires)); + S390_lowcore.system_timer += S390_lowcore.last_update_timer - timer; + S390_lowcore.last_update_timer = expires; +} + +static inline int virt_timer_forward(u64 elapsed) +{ + BUG_ON(!irqs_disabled()); + + if (list_empty(&virt_timer_list)) + return 0; + elapsed = atomic64_add_return(elapsed, &virt_timer_elapsed); + return elapsed >= atomic64_read(&virt_timer_current); +} + +static void update_mt_scaling(void) +{ + u64 cycles_new[8], *cycles_old; + u64 delta, fac, mult, div; + int i; + + stcctm(MT_DIAG, smp_cpu_mtid + 1, cycles_new); + cycles_old = this_cpu_ptr(mt_cycles); + fac = 1; + mult = div = 0; + for (i = 0; i <= smp_cpu_mtid; i++) { + delta = cycles_new[i] - cycles_old[i]; + div += delta; + mult *= i + 1; + mult += delta * fac; + fac *= i + 1; + } + div *= fac; + if (div > 0) { + /* Update scaling factor */ + __this_cpu_write(mt_scaling_mult, mult); + __this_cpu_write(mt_scaling_div, div); + memcpy(cycles_old, cycles_new, + sizeof(u64) * (smp_cpu_mtid + 1)); + } + __this_cpu_write(mt_scaling_jiffies, jiffies_64); +} + +static inline u64 update_tsk_timer(unsigned long *tsk_vtime, u64 new) +{ + u64 delta; + + delta = new - *tsk_vtime; + *tsk_vtime = new; + return delta; +} + + +static inline u64 scale_vtime(u64 vtime) +{ + u64 mult = __this_cpu_read(mt_scaling_mult); + u64 div = __this_cpu_read(mt_scaling_div); + + if (smp_cpu_mtid) + return vtime * mult / div; + return vtime; +} + +static void account_system_index_scaled(struct task_struct *p, u64 cputime, + enum cpu_usage_stat index) +{ + p->stimescaled += cputime_to_nsecs(scale_vtime(cputime)); + account_system_index_time(p, cputime_to_nsecs(cputime), index); +} + +/* + * Update process times based on virtual cpu times stored by entry.S + * to the lowcore fields user_timer, system_timer & steal_clock. + */ +static int do_account_vtime(struct task_struct *tsk) +{ + u64 timer, clock, user, guest, system, hardirq, softirq; + + timer = S390_lowcore.last_update_timer; + clock = S390_lowcore.last_update_clock; + asm volatile( + " stpt %0\n" /* Store current cpu timer value */ + " stckf %1" /* Store current tod clock value */ + : "=Q" (S390_lowcore.last_update_timer), + "=Q" (S390_lowcore.last_update_clock) + : : "cc"); + clock = S390_lowcore.last_update_clock - clock; + timer -= S390_lowcore.last_update_timer; + + if (hardirq_count()) + S390_lowcore.hardirq_timer += timer; + else + S390_lowcore.system_timer += timer; + + /* Update MT utilization calculation */ + if (smp_cpu_mtid && + time_after64(jiffies_64, this_cpu_read(mt_scaling_jiffies))) + update_mt_scaling(); + + /* Calculate cputime delta */ + user = update_tsk_timer(&tsk->thread.user_timer, + READ_ONCE(S390_lowcore.user_timer)); + guest = update_tsk_timer(&tsk->thread.guest_timer, + READ_ONCE(S390_lowcore.guest_timer)); + system = update_tsk_timer(&tsk->thread.system_timer, + READ_ONCE(S390_lowcore.system_timer)); + hardirq = update_tsk_timer(&tsk->thread.hardirq_timer, + READ_ONCE(S390_lowcore.hardirq_timer)); + softirq = update_tsk_timer(&tsk->thread.softirq_timer, + READ_ONCE(S390_lowcore.softirq_timer)); + S390_lowcore.steal_timer += + clock - user - guest - system - hardirq - softirq; + + /* Push account value */ + if (user) { + account_user_time(tsk, cputime_to_nsecs(user)); + tsk->utimescaled += cputime_to_nsecs(scale_vtime(user)); + } + + if (guest) { + account_guest_time(tsk, cputime_to_nsecs(guest)); + tsk->utimescaled += cputime_to_nsecs(scale_vtime(guest)); + } + + if (system) + account_system_index_scaled(tsk, system, CPUTIME_SYSTEM); + if (hardirq) + account_system_index_scaled(tsk, hardirq, CPUTIME_IRQ); + if (softirq) + account_system_index_scaled(tsk, softirq, CPUTIME_SOFTIRQ); + + return virt_timer_forward(user + guest + system + hardirq + softirq); +} + +void vtime_task_switch(struct task_struct *prev) +{ + do_account_vtime(prev); + prev->thread.user_timer = S390_lowcore.user_timer; + prev->thread.guest_timer = S390_lowcore.guest_timer; + prev->thread.system_timer = S390_lowcore.system_timer; + prev->thread.hardirq_timer = S390_lowcore.hardirq_timer; + prev->thread.softirq_timer = S390_lowcore.softirq_timer; + S390_lowcore.user_timer = current->thread.user_timer; + S390_lowcore.guest_timer = current->thread.guest_timer; + S390_lowcore.system_timer = current->thread.system_timer; + S390_lowcore.hardirq_timer = current->thread.hardirq_timer; + S390_lowcore.softirq_timer = current->thread.softirq_timer; +} + +/* + * In s390, accounting pending user time also implies + * accounting system time in order to correctly compute + * the stolen time accounting. + */ +void vtime_flush(struct task_struct *tsk) +{ + u64 steal, avg_steal; + + if (do_account_vtime(tsk)) + virt_timer_expire(); + + steal = S390_lowcore.steal_timer; + avg_steal = S390_lowcore.avg_steal_timer / 2; + if ((s64) steal > 0) { + S390_lowcore.steal_timer = 0; + account_steal_time(cputime_to_nsecs(steal)); + avg_steal += steal; + } + S390_lowcore.avg_steal_timer = avg_steal; +} + +static u64 vtime_delta(void) +{ + u64 timer = S390_lowcore.last_update_timer; + + S390_lowcore.last_update_timer = get_vtimer(); + + return timer - S390_lowcore.last_update_timer; +} + +/* + * Update process times based on virtual cpu times stored by entry.S + * to the lowcore fields user_timer, system_timer & steal_clock. + */ +void vtime_account_kernel(struct task_struct *tsk) +{ + u64 delta = vtime_delta(); + + if (tsk->flags & PF_VCPU) + S390_lowcore.guest_timer += delta; + else + S390_lowcore.system_timer += delta; + + virt_timer_forward(delta); +} +EXPORT_SYMBOL_GPL(vtime_account_kernel); + +void vtime_account_softirq(struct task_struct *tsk) +{ + u64 delta = vtime_delta(); + + S390_lowcore.softirq_timer += delta; + + virt_timer_forward(delta); +} + +void vtime_account_hardirq(struct task_struct *tsk) +{ + u64 delta = vtime_delta(); + + S390_lowcore.hardirq_timer += delta; + + virt_timer_forward(delta); +} + +/* + * Sorted add to a list. List is linear searched until first bigger + * element is found. + */ +static void list_add_sorted(struct vtimer_list *timer, struct list_head *head) +{ + struct vtimer_list *tmp; + + list_for_each_entry(tmp, head, entry) { + if (tmp->expires > timer->expires) { + list_add_tail(&timer->entry, &tmp->entry); + return; + } + } + list_add_tail(&timer->entry, head); +} + +/* + * Handler for expired virtual CPU timer. + */ +static void virt_timer_expire(void) +{ + struct vtimer_list *timer, *tmp; + unsigned long elapsed; + LIST_HEAD(cb_list); + + /* walk timer list, fire all expired timers */ + spin_lock(&virt_timer_lock); + elapsed = atomic64_read(&virt_timer_elapsed); + list_for_each_entry_safe(timer, tmp, &virt_timer_list, entry) { + if (timer->expires < elapsed) + /* move expired timer to the callback queue */ + list_move_tail(&timer->entry, &cb_list); + else + timer->expires -= elapsed; + } + if (!list_empty(&virt_timer_list)) { + timer = list_first_entry(&virt_timer_list, + struct vtimer_list, entry); + atomic64_set(&virt_timer_current, timer->expires); + } + atomic64_sub(elapsed, &virt_timer_elapsed); + spin_unlock(&virt_timer_lock); + + /* Do callbacks and recharge periodic timers */ + list_for_each_entry_safe(timer, tmp, &cb_list, entry) { + list_del_init(&timer->entry); + timer->function(timer->data); + if (timer->interval) { + /* Recharge interval timer */ + timer->expires = timer->interval + + atomic64_read(&virt_timer_elapsed); + spin_lock(&virt_timer_lock); + list_add_sorted(timer, &virt_timer_list); + spin_unlock(&virt_timer_lock); + } + } +} + +void init_virt_timer(struct vtimer_list *timer) +{ + timer->function = NULL; + INIT_LIST_HEAD(&timer->entry); +} +EXPORT_SYMBOL(init_virt_timer); + +static inline int vtimer_pending(struct vtimer_list *timer) +{ + return !list_empty(&timer->entry); +} + +static void internal_add_vtimer(struct vtimer_list *timer) +{ + if (list_empty(&virt_timer_list)) { + /* First timer, just program it. */ + atomic64_set(&virt_timer_current, timer->expires); + atomic64_set(&virt_timer_elapsed, 0); + list_add(&timer->entry, &virt_timer_list); + } else { + /* Update timer against current base. */ + timer->expires += atomic64_read(&virt_timer_elapsed); + if (likely((s64) timer->expires < + (s64) atomic64_read(&virt_timer_current))) + /* The new timer expires before the current timer. */ + atomic64_set(&virt_timer_current, timer->expires); + /* Insert new timer into the list. */ + list_add_sorted(timer, &virt_timer_list); + } +} + +static void __add_vtimer(struct vtimer_list *timer, int periodic) +{ + unsigned long flags; + + timer->interval = periodic ? timer->expires : 0; + spin_lock_irqsave(&virt_timer_lock, flags); + internal_add_vtimer(timer); + spin_unlock_irqrestore(&virt_timer_lock, flags); +} + +/* + * add_virt_timer - add a oneshot virtual CPU timer + */ +void add_virt_timer(struct vtimer_list *timer) +{ + __add_vtimer(timer, 0); +} +EXPORT_SYMBOL(add_virt_timer); + +/* + * add_virt_timer_int - add an interval virtual CPU timer + */ +void add_virt_timer_periodic(struct vtimer_list *timer) +{ + __add_vtimer(timer, 1); +} +EXPORT_SYMBOL(add_virt_timer_periodic); + +static int __mod_vtimer(struct vtimer_list *timer, u64 expires, int periodic) +{ + unsigned long flags; + int rc; + + BUG_ON(!timer->function); + + if (timer->expires == expires && vtimer_pending(timer)) + return 1; + spin_lock_irqsave(&virt_timer_lock, flags); + rc = vtimer_pending(timer); + if (rc) + list_del_init(&timer->entry); + timer->interval = periodic ? expires : 0; + timer->expires = expires; + internal_add_vtimer(timer); + spin_unlock_irqrestore(&virt_timer_lock, flags); + return rc; +} + +/* + * returns whether it has modified a pending timer (1) or not (0) + */ +int mod_virt_timer(struct vtimer_list *timer, u64 expires) +{ + return __mod_vtimer(timer, expires, 0); +} +EXPORT_SYMBOL(mod_virt_timer); + +/* + * returns whether it has modified a pending timer (1) or not (0) + */ +int mod_virt_timer_periodic(struct vtimer_list *timer, u64 expires) +{ + return __mod_vtimer(timer, expires, 1); +} +EXPORT_SYMBOL(mod_virt_timer_periodic); + +/* + * Delete a virtual timer. + * + * returns whether the deleted timer was pending (1) or not (0) + */ +int del_virt_timer(struct vtimer_list *timer) +{ + unsigned long flags; + + if (!vtimer_pending(timer)) + return 0; + spin_lock_irqsave(&virt_timer_lock, flags); + list_del_init(&timer->entry); + spin_unlock_irqrestore(&virt_timer_lock, flags); + return 1; +} +EXPORT_SYMBOL(del_virt_timer); + +/* + * Start the virtual CPU timer on the current CPU. + */ +void vtime_init(void) +{ + /* set initial cpu timer */ + set_vtimer(VTIMER_MAX_SLICE); + /* Setup initial MT scaling values */ + if (smp_cpu_mtid) { + __this_cpu_write(mt_scaling_jiffies, jiffies); + __this_cpu_write(mt_scaling_mult, 1); + __this_cpu_write(mt_scaling_div, 1); + stcctm(MT_DIAG, smp_cpu_mtid + 1, this_cpu_ptr(mt_cycles)); + } +} |