From 863bc087d69814dc4113ca9ec91ce1895b53480e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 23 Oct 2020 12:11:59 +0200 Subject: [PATCH 002/323] stop_machine: Add function and caller debug info Origin: https://www.kernel.org/pub/linux/kernel/projects/rt/5.10/older/patches-5.10.204-rt100.tar.xz Crashes in stop-machine are hard to connect to the calling code, add a little something to help with that. Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Sebastian Andrzej Siewior --- include/linux/stop_machine.h | 5 +++++ kernel/stop_machine.c | 23 ++++++++++++++++++++--- lib/dump_stack.c | 2 ++ 3 files changed, 27 insertions(+), 3 deletions(-) --- a/include/linux/stop_machine.h +++ b/include/linux/stop_machine.h @@ -24,6 +24,7 @@ struct cpu_stop_work { struct list_head list; /* cpu_stopper->works */ cpu_stop_fn_t fn; + unsigned long caller; void *arg; struct cpu_stop_done *done; }; @@ -36,6 +37,8 @@ void stop_machine_unpark(int cpu); void stop_machine_yield(const struct cpumask *cpumask); +extern void print_stop_info(const char *log_lvl, struct task_struct *task); + #else /* CONFIG_SMP */ #include @@ -80,6 +83,8 @@ return false; } +static inline void print_stop_info(const char *log_lvl, struct task_struct *task) { } + #endif /* CONFIG_SMP */ /* --- a/kernel/stop_machine.c +++ b/kernel/stop_machine.c @@ -42,11 +42,23 @@ struct list_head works; /* list of pending works */ struct cpu_stop_work stop_work; /* for stop_cpus */ + unsigned long caller; + cpu_stop_fn_t fn; }; static DEFINE_PER_CPU(struct cpu_stopper, cpu_stopper); static bool stop_machine_initialized = false; +void print_stop_info(const char *log_lvl, struct task_struct *task) +{ + struct cpu_stopper *stopper = this_cpu_ptr(&cpu_stopper); + + if (task != stopper->thread) + return; + + printk("%sStopper: %pS <- %pS\n", log_lvl, stopper->fn, (void *)stopper->caller); +} + /* static data for stop_cpus */ static DEFINE_MUTEX(stop_cpus_mutex); static bool stop_cpus_in_progress; @@ -123,7 +135,7 @@ int stop_one_cpu(unsigned int cpu, cpu_stop_fn_t fn, void *arg) { struct cpu_stop_done done; - struct cpu_stop_work work = { .fn = fn, .arg = arg, .done = &done }; + struct cpu_stop_work work = { .fn = fn, .arg = arg, .done = &done, .caller = _RET_IP_ }; cpu_stop_init_done(&done, 1); if (!cpu_stop_queue_work(cpu, &work)) @@ -331,7 +343,8 @@ work1 = work2 = (struct cpu_stop_work){ .fn = multi_cpu_stop, .arg = &msdata, - .done = &done + .done = &done, + .caller = _RET_IP_, }; cpu_stop_init_done(&done, 2); @@ -367,7 +380,7 @@ bool stop_one_cpu_nowait(unsigned int cpu, cpu_stop_fn_t fn, void *arg, struct cpu_stop_work *work_buf) { - *work_buf = (struct cpu_stop_work){ .fn = fn, .arg = arg, }; + *work_buf = (struct cpu_stop_work){ .fn = fn, .arg = arg, .caller = _RET_IP_, }; return cpu_stop_queue_work(cpu, work_buf); } @@ -487,6 +500,8 @@ int ret; /* cpu stop callbacks must not sleep, make in_atomic() == T */ + stopper->caller = work->caller; + stopper->fn = fn; preempt_count_inc(); ret = fn(arg); if (done) { @@ -495,6 +510,8 @@ cpu_stop_signal_done(done); } preempt_count_dec(); + stopper->fn = NULL; + stopper->caller = 0; WARN_ONCE(preempt_count(), "cpu_stop: %ps(%p) leaked preempt count\n", fn, arg); goto repeat; --- a/lib/dump_stack.c +++ b/lib/dump_stack.c @@ -12,6 +12,7 @@ #include #include #include +#include #include static char dump_stack_arch_desc_str[128]; @@ -59,6 +60,7 @@ log_lvl, dump_stack_arch_desc_str); print_worker_info(log_lvl, current); + print_stop_info(log_lvl, current); } /**