diff options
Diffstat (limited to 'tools/lib/bpf/btf_dump.c')
| -rw-r--r-- | tools/lib/bpf/btf_dump.c | 1531 |
1 files changed, 1531 insertions, 0 deletions
diff --git a/tools/lib/bpf/btf_dump.c b/tools/lib/bpf/btf_dump.c new file mode 100644 index 000000000..61aa2c47f --- /dev/null +++ b/tools/lib/bpf/btf_dump.c @@ -0,0 +1,1531 @@ +// SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) + +/* + * BTF-to-C type converter. + * + * Copyright (c) 2019 Facebook + */ + +#include <stdbool.h> +#include <stddef.h> +#include <stdlib.h> +#include <string.h> +#include <errno.h> +#include <linux/err.h> +#include <linux/btf.h> +#include <linux/kernel.h> +#include "btf.h" +#include "hashmap.h" +#include "libbpf.h" +#include "libbpf_internal.h" + +static const char PREFIXES[] = "\t\t\t\t\t\t\t\t\t\t\t\t\t"; +static const size_t PREFIX_CNT = sizeof(PREFIXES) - 1; + +static const char *pfx(int lvl) +{ + return lvl >= PREFIX_CNT ? PREFIXES : &PREFIXES[PREFIX_CNT - lvl]; +} + +enum btf_dump_type_order_state { + NOT_ORDERED, + ORDERING, + ORDERED, +}; + +enum btf_dump_type_emit_state { + NOT_EMITTED, + EMITTING, + EMITTED, +}; + +/* per-type auxiliary state */ +struct btf_dump_type_aux_state { + /* topological sorting state */ + enum btf_dump_type_order_state order_state: 2; + /* emitting state used to determine the need for forward declaration */ + enum btf_dump_type_emit_state emit_state: 2; + /* whether forward declaration was already emitted */ + __u8 fwd_emitted: 1; + /* whether unique non-duplicate name was already assigned */ + __u8 name_resolved: 1; + /* whether type is referenced from any other type */ + __u8 referenced: 1; +}; + +struct btf_dump { + const struct btf *btf; + const struct btf_ext *btf_ext; + btf_dump_printf_fn_t printf_fn; + struct btf_dump_opts opts; + int ptr_sz; + bool strip_mods; + int last_id; + + /* per-type auxiliary state */ + struct btf_dump_type_aux_state *type_states; + size_t type_states_cap; + /* per-type optional cached unique name, must be freed, if present */ + const char **cached_names; + size_t cached_names_cap; + + /* topo-sorted list of dependent type definitions */ + __u32 *emit_queue; + int emit_queue_cap; + int emit_queue_cnt; + + /* + * stack of type declarations (e.g., chain of modifiers, arrays, + * funcs, etc) + */ + __u32 *decl_stack; + int decl_stack_cap; + int decl_stack_cnt; + + /* maps struct/union/enum name to a number of name occurrences */ + struct hashmap *type_names; + /* + * maps typedef identifiers and enum value names to a number of such + * name occurrences + */ + struct hashmap *ident_names; +}; + +static size_t str_hash_fn(const void *key, void *ctx) +{ + return str_hash(key); +} + +static bool str_equal_fn(const void *a, const void *b, void *ctx) +{ + return strcmp(a, b) == 0; +} + +static const char *btf_name_of(const struct btf_dump *d, __u32 name_off) +{ + return btf__name_by_offset(d->btf, name_off); +} + +static void btf_dump_printf(const struct btf_dump *d, const char *fmt, ...) +{ + va_list args; + + va_start(args, fmt); + d->printf_fn(d->opts.ctx, fmt, args); + va_end(args); +} + +static int btf_dump_mark_referenced(struct btf_dump *d); +static int btf_dump_resize(struct btf_dump *d); + +struct btf_dump *btf_dump__new(const struct btf *btf, + const struct btf_ext *btf_ext, + const struct btf_dump_opts *opts, + btf_dump_printf_fn_t printf_fn) +{ + struct btf_dump *d; + int err; + + d = calloc(1, sizeof(struct btf_dump)); + if (!d) + return ERR_PTR(-ENOMEM); + + d->btf = btf; + d->btf_ext = btf_ext; + d->printf_fn = printf_fn; + d->opts.ctx = opts ? opts->ctx : NULL; + d->ptr_sz = btf__pointer_size(btf) ? : sizeof(void *); + + d->type_names = hashmap__new(str_hash_fn, str_equal_fn, NULL); + if (IS_ERR(d->type_names)) { + err = PTR_ERR(d->type_names); + d->type_names = NULL; + goto err; + } + d->ident_names = hashmap__new(str_hash_fn, str_equal_fn, NULL); + if (IS_ERR(d->ident_names)) { + err = PTR_ERR(d->ident_names); + d->ident_names = NULL; + goto err; + } + + err = btf_dump_resize(d); + if (err) + goto err; + + return d; +err: + btf_dump__free(d); + return ERR_PTR(err); +} + +static int btf_dump_resize(struct btf_dump *d) +{ + int err, last_id = btf__get_nr_types(d->btf); + + if (last_id <= d->last_id) + return 0; + + if (btf_ensure_mem((void **)&d->type_states, &d->type_states_cap, + sizeof(*d->type_states), last_id + 1)) + return -ENOMEM; + if (btf_ensure_mem((void **)&d->cached_names, &d->cached_names_cap, + sizeof(*d->cached_names), last_id + 1)) + return -ENOMEM; + + if (d->last_id == 0) { + /* VOID is special */ + d->type_states[0].order_state = ORDERED; + d->type_states[0].emit_state = EMITTED; + } + + /* eagerly determine referenced types for anon enums */ + err = btf_dump_mark_referenced(d); + if (err) + return err; + + d->last_id = last_id; + return 0; +} + +static void btf_dump_free_names(struct hashmap *map) +{ + size_t bkt; + struct hashmap_entry *cur; + + hashmap__for_each_entry(map, cur, bkt) + free((void *)cur->key); + + hashmap__free(map); +} + +void btf_dump__free(struct btf_dump *d) +{ + int i; + + if (IS_ERR_OR_NULL(d)) + return; + + free(d->type_states); + if (d->cached_names) { + /* any set cached name is owned by us and should be freed */ + for (i = 0; i <= d->last_id; i++) { + if (d->cached_names[i]) + free((void *)d->cached_names[i]); + } + } + free(d->cached_names); + free(d->emit_queue); + free(d->decl_stack); + btf_dump_free_names(d->type_names); + btf_dump_free_names(d->ident_names); + + free(d); +} + +static int btf_dump_order_type(struct btf_dump *d, __u32 id, bool through_ptr); +static void btf_dump_emit_type(struct btf_dump *d, __u32 id, __u32 cont_id); + +/* + * Dump BTF type in a compilable C syntax, including all the necessary + * dependent types, necessary for compilation. If some of the dependent types + * were already emitted as part of previous btf_dump__dump_type() invocation + * for another type, they won't be emitted again. This API allows callers to + * filter out BTF types according to user-defined criterias and emitted only + * minimal subset of types, necessary to compile everything. Full struct/union + * definitions will still be emitted, even if the only usage is through + * pointer and could be satisfied with just a forward declaration. + * + * Dumping is done in two high-level passes: + * 1. Topologically sort type definitions to satisfy C rules of compilation. + * 2. Emit type definitions in C syntax. + * + * Returns 0 on success; <0, otherwise. + */ +int btf_dump__dump_type(struct btf_dump *d, __u32 id) +{ + int err, i; + + if (id > btf__get_nr_types(d->btf)) + return -EINVAL; + + err = btf_dump_resize(d); + if (err) + return err; + + d->emit_queue_cnt = 0; + err = btf_dump_order_type(d, id, false); + if (err < 0) + return err; + + for (i = 0; i < d->emit_queue_cnt; i++) + btf_dump_emit_type(d, d->emit_queue[i], 0 /*top-level*/); + + return 0; +} + +/* + * Mark all types that are referenced from any other type. This is used to + * determine top-level anonymous enums that need to be emitted as an + * independent type declarations. + * Anonymous enums come in two flavors: either embedded in a struct's field + * definition, in which case they have to be declared inline as part of field + * type declaration; or as a top-level anonymous enum, typically used for + * declaring global constants. It's impossible to distinguish between two + * without knowning whether given enum type was referenced from other type: + * top-level anonymous enum won't be referenced by anything, while embedded + * one will. + */ +static int btf_dump_mark_referenced(struct btf_dump *d) +{ + int i, j, n = btf__get_nr_types(d->btf); + const struct btf_type *t; + __u16 vlen; + + for (i = d->last_id + 1; i <= n; i++) { + t = btf__type_by_id(d->btf, i); + vlen = btf_vlen(t); + + switch (btf_kind(t)) { + case BTF_KIND_INT: + case BTF_KIND_ENUM: + case BTF_KIND_FWD: + break; + + case BTF_KIND_VOLATILE: + case BTF_KIND_CONST: + case BTF_KIND_RESTRICT: + case BTF_KIND_PTR: + case BTF_KIND_TYPEDEF: + case BTF_KIND_FUNC: + case BTF_KIND_VAR: + d->type_states[t->type].referenced = 1; + break; + + case BTF_KIND_ARRAY: { + const struct btf_array *a = btf_array(t); + + d->type_states[a->index_type].referenced = 1; + d->type_states[a->type].referenced = 1; + break; + } + case BTF_KIND_STRUCT: + case BTF_KIND_UNION: { + const struct btf_member *m = btf_members(t); + + for (j = 0; j < vlen; j++, m++) + d->type_states[m->type].referenced = 1; + break; + } + case BTF_KIND_FUNC_PROTO: { + const struct btf_param *p = btf_params(t); + + for (j = 0; j < vlen; j++, p++) + d->type_states[p->type].referenced = 1; + break; + } + case BTF_KIND_DATASEC: { + const struct btf_var_secinfo *v = btf_var_secinfos(t); + + for (j = 0; j < vlen; j++, v++) + d->type_states[v->type].referenced = 1; + break; + } + default: + return -EINVAL; + } + } + return 0; +} + +static int btf_dump_add_emit_queue_id(struct btf_dump *d, __u32 id) +{ + __u32 *new_queue; + size_t new_cap; + + if (d->emit_queue_cnt >= d->emit_queue_cap) { + new_cap = max(16, d->emit_queue_cap * 3 / 2); + new_queue = libbpf_reallocarray(d->emit_queue, new_cap, sizeof(new_queue[0])); + if (!new_queue) + return -ENOMEM; + d->emit_queue = new_queue; + d->emit_queue_cap = new_cap; + } + + d->emit_queue[d->emit_queue_cnt++] = id; + return 0; +} + +/* + * Determine order of emitting dependent types and specified type to satisfy + * C compilation rules. This is done through topological sorting with an + * additional complication which comes from C rules. The main idea for C is + * that if some type is "embedded" into a struct/union, it's size needs to be + * known at the time of definition of containing type. E.g., for: + * + * struct A {}; + * struct B { struct A x; } + * + * struct A *HAS* to be defined before struct B, because it's "embedded", + * i.e., it is part of struct B layout. But in the following case: + * + * struct A; + * struct B { struct A *x; } + * struct A {}; + * + * it's enough to just have a forward declaration of struct A at the time of + * struct B definition, as struct B has a pointer to struct A, so the size of + * field x is known without knowing struct A size: it's sizeof(void *). + * + * Unfortunately, there are some trickier cases we need to handle, e.g.: + * + * struct A {}; // if this was forward-declaration: compilation error + * struct B { + * struct { // anonymous struct + * struct A y; + * } *x; + * }; + * + * In this case, struct B's field x is a pointer, so it's size is known + * regardless of the size of (anonymous) struct it points to. But because this + * struct is anonymous and thus defined inline inside struct B, *and* it + * embeds struct A, compiler requires full definition of struct A to be known + * before struct B can be defined. This creates a transitive dependency + * between struct A and struct B. If struct A was forward-declared before + * struct B definition and fully defined after struct B definition, that would + * trigger compilation error. + * + * All this means that while we are doing topological sorting on BTF type + * graph, we need to determine relationships between different types (graph + * nodes): + * - weak link (relationship) between X and Y, if Y *CAN* be + * forward-declared at the point of X definition; + * - strong link, if Y *HAS* to be fully-defined before X can be defined. + * + * The rule is as follows. Given a chain of BTF types from X to Y, if there is + * BTF_KIND_PTR type in the chain and at least one non-anonymous type + * Z (excluding X, including Y), then link is weak. Otherwise, it's strong. + * Weak/strong relationship is determined recursively during DFS traversal and + * is returned as a result from btf_dump_order_type(). + * + * btf_dump_order_type() is trying to avoid unnecessary forward declarations, + * but it is not guaranteeing that no extraneous forward declarations will be + * emitted. + * + * To avoid extra work, algorithm marks some of BTF types as ORDERED, when + * it's done with them, but not for all (e.g., VOLATILE, CONST, RESTRICT, + * ARRAY, FUNC_PROTO), as weak/strong semantics for those depends on the + * entire graph path, so depending where from one came to that BTF type, it + * might cause weak or strong ordering. For types like STRUCT/UNION/INT/ENUM, + * once they are processed, there is no need to do it again, so they are + * marked as ORDERED. We can mark PTR as ORDERED as well, as it semi-forces + * weak link, unless subsequent referenced STRUCT/UNION/ENUM is anonymous. But + * in any case, once those are processed, no need to do it again, as the + * result won't change. + * + * Returns: + * - 1, if type is part of strong link (so there is strong topological + * ordering requirements); + * - 0, if type is part of weak link (so can be satisfied through forward + * declaration); + * - <0, on error (e.g., unsatisfiable type loop detected). + */ +static int btf_dump_order_type(struct btf_dump *d, __u32 id, bool through_ptr) +{ + /* + * Order state is used to detect strong link cycles, but only for BTF + * kinds that are or could be an independent definition (i.e., + * stand-alone fwd decl, enum, typedef, struct, union). Ptrs, arrays, + * func_protos, modifiers are just means to get to these definitions. + * Int/void don't need definitions, they are assumed to be always + * properly defined. We also ignore datasec, var, and funcs for now. + * So for all non-defining kinds, we never even set ordering state, + * for defining kinds we set ORDERING and subsequently ORDERED if it + * forms a strong link. + */ + struct btf_dump_type_aux_state *tstate = &d->type_states[id]; + const struct btf_type *t; + __u16 vlen; + int err, i; + + /* return true, letting typedefs know that it's ok to be emitted */ + if (tstate->order_state == ORDERED) + return 1; + + t = btf__type_by_id(d->btf, id); + + if (tstate->order_state == ORDERING) { + /* type loop, but resolvable through fwd declaration */ + if (btf_is_composite(t) && through_ptr && t->name_off != 0) + return 0; + pr_warn("unsatisfiable type cycle, id:[%u]\n", id); + return -ELOOP; + } + + switch (btf_kind(t)) { + case BTF_KIND_INT: + tstate->order_state = ORDERED; + return 0; + + case BTF_KIND_PTR: + err = btf_dump_order_type(d, t->type, true); + tstate->order_state = ORDERED; + return err; + + case BTF_KIND_ARRAY: + return btf_dump_order_type(d, btf_array(t)->type, false); + + case BTF_KIND_STRUCT: + case BTF_KIND_UNION: { + const struct btf_member *m = btf_members(t); + /* + * struct/union is part of strong link, only if it's embedded + * (so no ptr in a path) or it's anonymous (so has to be + * defined inline, even if declared through ptr) + */ + if (through_ptr && t->name_off != 0) + return 0; + + tstate->order_state = ORDERING; + + vlen = btf_vlen(t); + for (i = 0; i < vlen; i++, m++) { + err = btf_dump_order_type(d, m->type, false); + if (err < 0) + return err; + } + + if (t->name_off != 0) { + err = btf_dump_add_emit_queue_id(d, id); + if (err < 0) + return err; + } + + tstate->order_state = ORDERED; + return 1; + } + case BTF_KIND_ENUM: + case BTF_KIND_FWD: + /* + * non-anonymous or non-referenced enums are top-level + * declarations and should be emitted. Same logic can be + * applied to FWDs, it won't hurt anyways. + */ + if (t->name_off != 0 || !tstate->referenced) { + err = btf_dump_add_emit_queue_id(d, id); + if (err) + return err; + } + tstate->order_state = ORDERED; + return 1; + + case BTF_KIND_TYPEDEF: { + int is_strong; + + is_strong = btf_dump_order_type(d, t->type, through_ptr); + if (is_strong < 0) + return is_strong; + + /* typedef is similar to struct/union w.r.t. fwd-decls */ + if (through_ptr && !is_strong) + return 0; + + /* typedef is always a named definition */ + err = btf_dump_add_emit_queue_id(d, id); + if (err) + return err; + + d->type_states[id].order_state = ORDERED; + return 1; + } + case BTF_KIND_VOLATILE: + case BTF_KIND_CONST: + case BTF_KIND_RESTRICT: + return btf_dump_order_type(d, t->type, through_ptr); + + case BTF_KIND_FUNC_PROTO: { + const struct btf_param *p = btf_params(t); + bool is_strong; + + err = btf_dump_order_type(d, t->type, through_ptr); + if (err < 0) + return err; + is_strong = err > 0; + + vlen = btf_vlen(t); + for (i = 0; i < vlen; i++, p++) { + err = btf_dump_order_type(d, p->type, through_ptr); + if (err < 0) + return err; + if (err > 0) + is_strong = true; + } + return is_strong; + } + case BTF_KIND_FUNC: + case BTF_KIND_VAR: + case BTF_KIND_DATASEC: + d->type_states[id].order_state = ORDERED; + return 0; + + default: + return -EINVAL; + } +} + +static void btf_dump_emit_missing_aliases(struct btf_dump *d, __u32 id, + const struct btf_type *t); + +static void btf_dump_emit_struct_fwd(struct btf_dump *d, __u32 id, + const struct btf_type *t); +static void btf_dump_emit_struct_def(struct btf_dump *d, __u32 id, + const struct btf_type *t, int lvl); + +static void btf_dump_emit_enum_fwd(struct btf_dump *d, __u32 id, + const struct btf_type *t); +static void btf_dump_emit_enum_def(struct btf_dump *d, __u32 id, + const struct btf_type *t, int lvl); + +static void btf_dump_emit_fwd_def(struct btf_dump *d, __u32 id, + const struct btf_type *t); + +static void btf_dump_emit_typedef_def(struct btf_dump *d, __u32 id, + const struct btf_type *t, int lvl); + +/* a local view into a shared stack */ +struct id_stack { + const __u32 *ids; + int cnt; +}; + +static void btf_dump_emit_type_decl(struct btf_dump *d, __u32 id, + const char *fname, int lvl); +static void btf_dump_emit_type_chain(struct btf_dump *d, + struct id_stack *decl_stack, + const char *fname, int lvl); + +static const char *btf_dump_type_name(struct btf_dump *d, __u32 id); +static const char *btf_dump_ident_name(struct btf_dump *d, __u32 id); +static size_t btf_dump_name_dups(struct btf_dump *d, struct hashmap *name_map, + const char *orig_name); + +static bool btf_dump_is_blacklisted(struct btf_dump *d, __u32 id) +{ + const struct btf_type *t = btf__type_by_id(d->btf, id); + + /* __builtin_va_list is a compiler built-in, which causes compilation + * errors, when compiling w/ different compiler, then used to compile + * original code (e.g., GCC to compile kernel, Clang to use generated + * C header from BTF). As it is built-in, it should be already defined + * properly internally in compiler. + */ + if (t->name_off == 0) + return false; + return strcmp(btf_name_of(d, t->name_off), "__builtin_va_list") == 0; +} + +/* + * Emit C-syntax definitions of types from chains of BTF types. + * + * High-level handling of determining necessary forward declarations are handled + * by btf_dump_emit_type() itself, but all nitty-gritty details of emitting type + * declarations/definitions in C syntax are handled by a combo of + * btf_dump_emit_type_decl()/btf_dump_emit_type_chain() w/ delegation to + * corresponding btf_dump_emit_*_{def,fwd}() functions. + * + * We also keep track of "containing struct/union type ID" to determine when + * we reference it from inside and thus can avoid emitting unnecessary forward + * declaration. + * + * This algorithm is designed in such a way, that even if some error occurs + * (either technical, e.g., out of memory, or logical, i.e., malformed BTF + * that doesn't comply to C rules completely), algorithm will try to proceed + * and produce as much meaningful output as possible. + */ +static void btf_dump_emit_type(struct btf_dump *d, __u32 id, __u32 cont_id) +{ + struct btf_dump_type_aux_state *tstate = &d->type_states[id]; + bool top_level_def = cont_id == 0; + const struct btf_type *t; + __u16 kind; + + if (tstate->emit_state == EMITTED) + return; + + t = btf__type_by_id(d->btf, id); + kind = btf_kind(t); + + if (tstate->emit_state == EMITTING) { + if (tstate->fwd_emitted) + return; + + switch (kind) { + case BTF_KIND_STRUCT: + case BTF_KIND_UNION: + /* + * if we are referencing a struct/union that we are + * part of - then no need for fwd declaration + */ + if (id == cont_id) + return; + if (t->name_off == 0) { + pr_warn("anonymous struct/union loop, id:[%u]\n", + id); + return; + } + btf_dump_emit_struct_fwd(d, id, t); + btf_dump_printf(d, ";\n\n"); + tstate->fwd_emitted = 1; + break; + case BTF_KIND_TYPEDEF: + /* + * for typedef fwd_emitted means typedef definition + * was emitted, but it can be used only for "weak" + * references through pointer only, not for embedding + */ + if (!btf_dump_is_blacklisted(d, id)) { + btf_dump_emit_typedef_def(d, id, t, 0); + btf_dump_printf(d, ";\n\n"); + } + tstate->fwd_emitted = 1; + break; + default: + break; + } + + return; + } + + switch (kind) { + case BTF_KIND_INT: + /* Emit type alias definitions if necessary */ + btf_dump_emit_missing_aliases(d, id, t); + + tstate->emit_state = EMITTED; + break; + case BTF_KIND_ENUM: + if (top_level_def) { + btf_dump_emit_enum_def(d, id, t, 0); + btf_dump_printf(d, ";\n\n"); + } + tstate->emit_state = EMITTED; + break; + case BTF_KIND_PTR: + case BTF_KIND_VOLATILE: + case BTF_KIND_CONST: + case BTF_KIND_RESTRICT: + btf_dump_emit_type(d, t->type, cont_id); + break; + case BTF_KIND_ARRAY: + btf_dump_emit_type(d, btf_array(t)->type, cont_id); + break; + case BTF_KIND_FWD: + btf_dump_emit_fwd_def(d, id, t); + btf_dump_printf(d, ";\n\n"); + tstate->emit_state = EMITTED; + break; + case BTF_KIND_TYPEDEF: + tstate->emit_state = EMITTING; + btf_dump_emit_type(d, t->type, id); + /* + * typedef can server as both definition and forward + * declaration; at this stage someone depends on + * typedef as a forward declaration (refers to it + * through pointer), so unless we already did it, + * emit typedef as a forward declaration + */ + if (!tstate->fwd_emitted && !btf_dump_is_blacklisted(d, id)) { + btf_dump_emit_typedef_def(d, id, t, 0); + btf_dump_printf(d, ";\n\n"); + } + tstate->emit_state = EMITTED; + break; + case BTF_KIND_STRUCT: + case BTF_KIND_UNION: + tstate->emit_state = EMITTING; + /* if it's a top-level struct/union definition or struct/union + * is anonymous, then in C we'll be emitting all fields and + * their types (as opposed to just `struct X`), so we need to + * make sure that all types, referenced from struct/union + * members have necessary forward-declarations, where + * applicable + */ + if (top_level_def || t->name_off == 0) { + const struct btf_member *m = btf_members(t); + __u16 vlen = btf_vlen(t); + int i, new_cont_id; + + new_cont_id = t->name_off == 0 ? cont_id : id; + for (i = 0; i < vlen; i++, m++) + btf_dump_emit_type(d, m->type, new_cont_id); + } else if (!tstate->fwd_emitted && id != cont_id) { + btf_dump_emit_struct_fwd(d, id, t); + btf_dump_printf(d, ";\n\n"); + tstate->fwd_emitted = 1; + } + + if (top_level_def) { + btf_dump_emit_struct_def(d, id, t, 0); + btf_dump_printf(d, ";\n\n"); + tstate->emit_state = EMITTED; + } else { + tstate->emit_state = NOT_EMITTED; + } + break; + case BTF_KIND_FUNC_PROTO: { + const struct btf_param *p = btf_params(t); + __u16 vlen = btf_vlen(t); + int i; + + btf_dump_emit_type(d, t->type, cont_id); + for (i = 0; i < vlen; i++, p++) + btf_dump_emit_type(d, p->type, cont_id); + + break; + } + default: + break; + } +} + +static bool btf_is_struct_packed(const struct btf *btf, __u32 id, + const struct btf_type *t) +{ + const struct btf_member *m; + int max_align = 1, align, i, bit_sz; + __u16 vlen; + + m = btf_members(t); + vlen = btf_vlen(t); + /* all non-bitfield fields have to be naturally aligned */ + for (i = 0; i < vlen; i++, m++) { + align = btf__align_of(btf, m->type); + bit_sz = btf_member_bitfield_size(t, i); + if (align && bit_sz == 0 && m->offset % (8 * align) != 0) + return true; + max_align = max(align, max_align); + } + /* size of a non-packed struct has to be a multiple of its alignment */ + if (t->size % max_align != 0) + return true; + /* + * if original struct was marked as packed, but its layout is + * naturally aligned, we'll detect that it's not packed + */ + return false; +} + +static void btf_dump_emit_bit_padding(const struct btf_dump *d, + int cur_off, int next_off, int next_align, + bool in_bitfield, int lvl) +{ + const struct { + const char *name; + int bits; + } pads[] = { + {"long", d->ptr_sz * 8}, {"int", 32}, {"short", 16}, {"char", 8} + }; + int new_off, pad_bits, bits, i; + const char *pad_type; + + if (cur_off >= next_off) + return; /* no gap */ + + /* For filling out padding we want to take advantage of + * natural alignment rules to minimize unnecessary explicit + * padding. First, we find the largest type (among long, int, + * short, or char) that can be used to force naturally aligned + * boundary. Once determined, we'll use such type to fill in + * the remaining padding gap. In some cases we can rely on + * compiler filling some gaps, but sometimes we need to force + * alignment to close natural alignment with markers like + * `long: 0` (this is always the case for bitfields). Note + * that even if struct itself has, let's say 4-byte alignment + * (i.e., it only uses up to int-aligned types), using `long: + * X;` explicit padding doesn't actually change struct's + * overall alignment requirements, but compiler does take into + * account that type's (long, in this example) natural + * alignment requirements when adding implicit padding. We use + * this fact heavily and don't worry about ruining correct + * struct alignment requirement. + */ + for (i = 0; i < ARRAY_SIZE(pads); i++) { + pad_bits = pads[i].bits; + pad_type = pads[i].name; + + new_off = roundup(cur_off, pad_bits); + if (new_off <= next_off) + break; + } + + if (new_off > cur_off && new_off <= next_off) { + /* We need explicit `<type>: 0` aligning mark if next + * field is right on alignment offset and its + * alignment requirement is less strict than <type>'s + * alignment (so compiler won't naturally align to the + * offset we expect), or if subsequent `<type>: X`, + * will actually completely fit in the remaining hole, + * making compiler basically ignore `<type>: X` + * completely. + */ + if (in_bitfield || + (new_off == next_off && roundup(cur_off, next_align * 8) != new_off) || + (new_off != next_off && next_off - new_off <= new_off - cur_off)) + /* but for bitfields we'll emit explicit bit count */ + btf_dump_printf(d, "\n%s%s: %d;", pfx(lvl), pad_type, + in_bitfield ? new_off - cur_off : 0); + cur_off = new_off; + } + + /* Now we know we start at naturally aligned offset for a chosen + * padding type (long, int, short, or char), and so the rest is just + * a straightforward filling of remaining padding gap with full + * `<type>: sizeof(<type>);` markers, except for the last one, which + * might need smaller than sizeof(<type>) padding. + */ + while (cur_off != next_off) { + bits = min(next_off - cur_off, pad_bits); + if (bits == pad_bits) { + btf_dump_printf(d, "\n%s%s: %d;", pfx(lvl), pad_type, pad_bits); + cur_off += bits; + continue; + } + /* For the remainder padding that doesn't cover entire + * pad_type bit length, we pick the smallest necessary type. + * This is pure aesthetics, we could have just used `long`, + * but having smallest necessary one communicates better the + * scale of the padding gap. + */ + for (i = ARRAY_SIZE(pads) - 1; i >= 0; i--) { + pad_type = pads[i].name; + pad_bits = pads[i].bits; + if (pad_bits < bits) + continue; + + btf_dump_printf(d, "\n%s%s: %d;", pfx(lvl), pad_type, bits); + cur_off += bits; + break; + } + } +} + +static void btf_dump_emit_struct_fwd(struct btf_dump *d, __u32 id, + const struct btf_type *t) +{ + btf_dump_printf(d, "%s %s", + btf_is_struct(t) ? "struct" : "union", + btf_dump_type_name(d, id)); +} + +static void btf_dump_emit_struct_def(struct btf_dump *d, + __u32 id, + const struct btf_type *t, + int lvl) +{ + const struct btf_member *m = btf_members(t); + bool is_struct = btf_is_struct(t); + bool packed, prev_bitfield = false; + int align, i, off = 0; + __u16 vlen = btf_vlen(t); + + align = btf__align_of(d->btf, id); + packed = is_struct ? btf_is_struct_packed(d->btf, id, t) : 0; + + btf_dump_printf(d, "%s%s%s {", + is_struct ? "struct" : "union", + t->name_off ? " " : "", + btf_dump_type_name(d, id)); + + for (i = 0; i < vlen; i++, m++) { + const char *fname; + int m_off, m_sz, m_align; + bool in_bitfield; + + fname = btf_name_of(d, m->name_off); + m_sz = btf_member_bitfield_size(t, i); + m_off = btf_member_bit_offset(t, i); + m_align = packed ? 1 : btf__align_of(d->btf, m->type); + + in_bitfield = prev_bitfield && m_sz != 0; + + btf_dump_emit_bit_padding(d, off, m_off, m_align, in_bitfield, lvl + 1); + btf_dump_printf(d, "\n%s", pfx(lvl + 1)); + btf_dump_emit_type_decl(d, m->type, fname, lvl + 1); + + if (m_sz) { + btf_dump_printf(d, ": %d", m_sz); + off = m_off + m_sz; + prev_bitfield = true; + } else { + m_sz = max((__s64)0, btf__resolve_size(d->btf, m->type)); + off = m_off + m_sz * 8; + prev_bitfield = false; + } + + btf_dump_printf(d, ";"); + } + + /* pad at the end, if necessary */ + if (is_struct) + btf_dump_emit_bit_padding(d, off, t->size * 8, align, false, lvl + 1); + + /* + * Keep `struct empty {}` on a single line, + * only print newline when there are regular or padding fields. + */ + if (vlen || t->size) { + btf_dump_printf(d, "\n"); + btf_dump_printf(d, "%s}", pfx(lvl)); + } else { + btf_dump_printf(d, "}"); + } + if (packed) + btf_dump_printf(d, " __attribute__((packed))"); +} + +static const char *missing_base_types[][2] = { + /* + * GCC emits typedefs to its internal __PolyX_t types when compiling Arm + * SIMD intrinsics. Alias them to standard base types. + */ + { "__Poly8_t", "unsigned char" }, + { "__Poly16_t", "unsigned short" }, + { "__Poly64_t", "unsigned long long" }, + { "__Poly128_t", "unsigned __int128" }, +}; + +static void btf_dump_emit_missing_aliases(struct btf_dump *d, __u32 id, + const struct btf_type *t) +{ + const char *name = btf_dump_type_name(d, id); + int i; + + for (i = 0; i < ARRAY_SIZE(missing_base_types); i++) { + if (strcmp(name, missing_base_types[i][0]) == 0) { + btf_dump_printf(d, "typedef %s %s;\n\n", + missing_base_types[i][1], name); + break; + } + } +} + +static void btf_dump_emit_enum_fwd(struct btf_dump *d, __u32 id, + const struct btf_type *t) +{ + btf_dump_printf(d, "enum %s", btf_dump_type_name(d, id)); +} + +static void btf_dump_emit_enum_def(struct btf_dump *d, __u32 id, + const struct btf_type *t, + int lvl) +{ + const struct btf_enum *v = btf_enum(t); + __u16 vlen = btf_vlen(t); + const char *name; + size_t dup_cnt; + int i; + + btf_dump_printf(d, "enum%s%s", + t->name_off ? " " : "", + btf_dump_type_name(d, id)); + + if (vlen) { + btf_dump_printf(d, " {"); + for (i = 0; i < vlen; i++, v++) { + name = btf_name_of(d, v->name_off); + /* enumerators share namespace with typedef idents */ + dup_cnt = btf_dump_name_dups(d, d->ident_names, name); + if (dup_cnt > 1) { + btf_dump_printf(d, "\n%s%s___%zu = %u,", + pfx(lvl + 1), name, dup_cnt, + (__u32)v->val); + } else { + btf_dump_printf(d, "\n%s%s = %u,", + pfx(lvl + 1), name, + (__u32)v->val); + } + } + btf_dump_printf(d, "\n%s}", pfx(lvl)); + } +} + +static void btf_dump_emit_fwd_def(struct btf_dump *d, __u32 id, + const struct btf_type *t) +{ + const char *name = btf_dump_type_name(d, id); + + if (btf_kflag(t)) + btf_dump_printf(d, "union %s", name); + else + btf_dump_printf(d, "struct %s", name); +} + +static void btf_dump_emit_typedef_def(struct btf_dump *d, __u32 id, + const struct btf_type *t, int lvl) +{ + const char *name = btf_dump_ident_name(d, id); + + /* + * Old GCC versions are emitting invalid typedef for __gnuc_va_list + * pointing to VOID. This generates warnings from btf_dump() and + * results in uncompilable header file, so we are fixing it up here + * with valid typedef into __builtin_va_list. + */ + if (t->type == 0 && strcmp(name, "__gnuc_va_list") == 0) { + btf_dump_printf(d, "typedef __builtin_va_list __gnuc_va_list"); + return; + } + + btf_dump_printf(d, "typedef "); + btf_dump_emit_type_decl(d, t->type, name, lvl); +} + +static int btf_dump_push_decl_stack_id(struct btf_dump *d, __u32 id) +{ + __u32 *new_stack; + size_t new_cap; + + if (d->decl_stack_cnt >= d->decl_stack_cap) { + new_cap = max(16, d->decl_stack_cap * 3 / 2); + new_stack = libbpf_reallocarray(d->decl_stack, new_cap, sizeof(new_stack[0])); + if (!new_stack) + return -ENOMEM; + d->decl_stack = new_stack; + d->decl_stack_cap = new_cap; + } + + d->decl_stack[d->decl_stack_cnt++] = id; + + return 0; +} + +/* + * Emit type declaration (e.g., field type declaration in a struct or argument + * declaration in function prototype) in correct C syntax. + * + * For most types it's trivial, but there are few quirky type declaration + * cases worth mentioning: + * - function prototypes (especially nesting of function prototypes); + * - arrays; + * - const/volatile/restrict for pointers vs other types. + * + * For a good discussion of *PARSING* C syntax (as a human), see + * Peter van der Linden's "Expert C Programming: Deep C Secrets", + * Ch.3 "Unscrambling Declarations in C". + * + * It won't help with BTF to C conversion much, though, as it's an opposite + * problem. So we came up with this algorithm in reverse to van der Linden's + * parsing algorithm. It goes from structured BTF representation of type + * declaration to a valid compilable C syntax. + * + * For instance, consider this C typedef: + * typedef const int * const * arr[10] arr_t; + * It will be represented in BTF with this chain of BTF types: + * [typedef] -> [array] -> [ptr] -> [const] -> [ptr] -> [const] -> [int] + * + * Notice how [const] modifier always goes before type it modifies in BTF type + * graph, but in C syntax, const/volatile/restrict modifiers are written to + * the right of pointers, but to the left of other types. There are also other + * quirks, like function pointers, arrays of them, functions returning other + * functions, etc. + * + * We handle that by pushing all the types to a stack, until we hit "terminal" + * type (int/enum/struct/union/fwd). Then depending on the kind of a type on + * top of a stack, modifiers are handled differently. Array/function pointers + * have also wildly different syntax and how nesting of them are done. See + * code for authoritative definition. + * + * To avoid allocating new stack for each independent chain of BTF types, we + * share one bigger stack, with each chain working only on its own local view + * of a stack frame. Some care is required to "pop" stack frames after + * processing type declaration chain. + */ +int btf_dump__emit_type_decl(struct btf_dump *d, __u32 id, + const struct btf_dump_emit_type_decl_opts *opts) +{ + const char *fname; + int lvl, err; + + if (!OPTS_VALID(opts, btf_dump_emit_type_decl_opts)) + return -EINVAL; + + err = btf_dump_resize(d); + if (err) + return -EINVAL; + + fname = OPTS_GET(opts, field_name, ""); + lvl = OPTS_GET(opts, indent_level, 0); + d->strip_mods = OPTS_GET(opts, strip_mods, false); + btf_dump_emit_type_decl(d, id, fname, lvl); + d->strip_mods = false; + return 0; +} + +static void btf_dump_emit_type_decl(struct btf_dump *d, __u32 id, + const char *fname, int lvl) +{ + struct id_stack decl_stack; + const struct btf_type *t; + int err, stack_start; + + stack_start = d->decl_stack_cnt; + for (;;) { + t = btf__type_by_id(d->btf, id); + if (d->strip_mods && btf_is_mod(t)) + goto skip_mod; + + err = btf_dump_push_decl_stack_id(d, id); + if (err < 0) { + /* + * if we don't have enough memory for entire type decl + * chain, restore stack, emit warning, and try to + * proceed nevertheless + */ + pr_warn("not enough memory for decl stack:%d", err); + d->decl_stack_cnt = stack_start; + return; + } +skip_mod: + /* VOID */ + if (id == 0) + break; + + switch (btf_kind(t)) { + case BTF_KIND_PTR: + case BTF_KIND_VOLATILE: + case BTF_KIND_CONST: + case BTF_KIND_RESTRICT: + case BTF_KIND_FUNC_PROTO: + id = t->type; + break; + case BTF_KIND_ARRAY: + id = btf_array(t)->type; + break; + case BTF_KIND_INT: + case BTF_KIND_ENUM: + case BTF_KIND_FWD: + case BTF_KIND_STRUCT: + case BTF_KIND_UNION: + case BTF_KIND_TYPEDEF: + goto done; + default: + pr_warn("unexpected type in decl chain, kind:%u, id:[%u]\n", + btf_kind(t), id); + goto done; + } + } +done: + /* + * We might be inside a chain of declarations (e.g., array of function + * pointers returning anonymous (so inlined) structs, having another + * array field). Each of those needs its own "stack frame" to handle + * emitting of declarations. Those stack frames are non-overlapping + * portions of shared btf_dump->decl_stack. To make it a bit nicer to + * handle this set of nested stacks, we create a view corresponding to + * our own "stack frame" and work with it as an independent stack. + * We'll need to clean up after emit_type_chain() returns, though. + */ + decl_stack.ids = d->decl_stack + stack_start; + decl_stack.cnt = d->decl_stack_cnt - stack_start; + btf_dump_emit_type_chain(d, &decl_stack, fname, lvl); + /* + * emit_type_chain() guarantees that it will pop its entire decl_stack + * frame before returning. But it works with a read-only view into + * decl_stack, so it doesn't actually pop anything from the + * perspective of shared btf_dump->decl_stack, per se. We need to + * reset decl_stack state to how it was before us to avoid it growing + * all the time. + */ + d->decl_stack_cnt = stack_start; +} + +static void btf_dump_emit_mods(struct btf_dump *d, struct id_stack *decl_stack) +{ + const struct btf_type *t; + __u32 id; + + while (decl_stack->cnt) { + id = decl_stack->ids[decl_stack->cnt - 1]; + t = btf__type_by_id(d->btf, id); + + switch (btf_kind(t)) { + case BTF_KIND_VOLATILE: + btf_dump_printf(d, "volatile "); + break; + case BTF_KIND_CONST: + btf_dump_printf(d, "const "); + break; + case BTF_KIND_RESTRICT: + btf_dump_printf(d, "restrict "); + break; + default: + return; + } + decl_stack->cnt--; + } +} + +static void btf_dump_drop_mods(struct btf_dump *d, struct id_stack *decl_stack) +{ + const struct btf_type *t; + __u32 id; + + while (decl_stack->cnt) { + id = decl_stack->ids[decl_stack->cnt - 1]; + t = btf__type_by_id(d->btf, id); + if (!btf_is_mod(t)) + return; + decl_stack->cnt--; + } +} + +static void btf_dump_emit_name(const struct btf_dump *d, + const char *name, bool last_was_ptr) +{ + bool separate = name[0] && !last_was_ptr; + + btf_dump_printf(d, "%s%s", separate ? " " : "", name); +} + +static void btf_dump_emit_type_chain(struct btf_dump *d, + struct id_stack *decls, + const char *fname, int lvl) +{ + /* + * last_was_ptr is used to determine if we need to separate pointer + * asterisk (*) from previous part of type signature with space, so + * that we get `int ***`, instead of `int * * *`. We default to true + * for cases where we have single pointer in a chain. E.g., in ptr -> + * func_proto case. func_proto will start a new emit_type_chain call + * with just ptr, which should be emitted as (*) or (*<fname>), so we + * don't want to prepend space for that last pointer. + */ + bool last_was_ptr = true; + const struct btf_type *t; + const char *name; + __u16 kind; + __u32 id; + + while (decls->cnt) { + id = decls->ids[--decls->cnt]; + if (id == 0) { + /* VOID is a special snowflake */ + btf_dump_emit_mods(d, decls); + btf_dump_printf(d, "void"); + last_was_ptr = false; + continue; + } + + t = btf__type_by_id(d->btf, id); + kind = btf_kind(t); + + switch (kind) { + case BTF_KIND_INT: + btf_dump_emit_mods(d, decls); + name = btf_name_of(d, t->name_off); + btf_dump_printf(d, "%s", name); + break; + case BTF_KIND_STRUCT: + case BTF_KIND_UNION: + btf_dump_emit_mods(d, decls); + /* inline anonymous struct/union */ + if (t->name_off == 0) + btf_dump_emit_struct_def(d, id, t, lvl); + else + btf_dump_emit_struct_fwd(d, id, t); + break; + case BTF_KIND_ENUM: + btf_dump_emit_mods(d, decls); + /* inline anonymous enum */ + if (t->name_off == 0) + btf_dump_emit_enum_def(d, id, t, lvl); + else + btf_dump_emit_enum_fwd(d, id, t); + break; + case BTF_KIND_FWD: + btf_dump_emit_mods(d, decls); + btf_dump_emit_fwd_def(d, id, t); + break; + case BTF_KIND_TYPEDEF: + btf_dump_emit_mods(d, decls); + btf_dump_printf(d, "%s", btf_dump_ident_name(d, id)); + break; + case BTF_KIND_PTR: + btf_dump_printf(d, "%s", last_was_ptr ? "*" : " *"); + break; + case BTF_KIND_VOLATILE: + btf_dump_printf(d, " volatile"); + break; + case BTF_KIND_CONST: + btf_dump_printf(d, " const"); + break; + case BTF_KIND_RESTRICT: + btf_dump_printf(d, " restrict"); + break; + case BTF_KIND_ARRAY: { + const struct btf_array *a = btf_array(t); + const struct btf_type *next_t; + __u32 next_id; + bool multidim; + /* + * GCC has a bug + * (https://gcc.gnu.org/bugzilla/show_bug.cgi?id=8354) + * which causes it to emit extra const/volatile + * modifiers for an array, if array's element type has + * const/volatile modifiers. Clang doesn't do that. + * In general, it doesn't seem very meaningful to have + * a const/volatile modifier for array, so we are + * going to silently skip them here. + */ + btf_dump_drop_mods(d, decls); + + if (decls->cnt == 0) { + btf_dump_emit_name(d, fname, last_was_ptr); + btf_dump_printf(d, "[%u]", a->nelems); + return; + } + + next_id = decls->ids[decls->cnt - 1]; + next_t = btf__type_by_id(d->btf, next_id); + multidim = btf_is_array(next_t); + /* we need space if we have named non-pointer */ + if (fname[0] && !last_was_ptr) + btf_dump_printf(d, " "); + /* no parentheses for multi-dimensional array */ + if (!multidim) + btf_dump_printf(d, "("); + btf_dump_emit_type_chain(d, decls, fname, lvl); + if (!multidim) + btf_dump_printf(d, ")"); + btf_dump_printf(d, "[%u]", a->nelems); + return; + } + case BTF_KIND_FUNC_PROTO: { + const struct btf_param *p = btf_params(t); + __u16 vlen = btf_vlen(t); + int i; + + /* + * GCC emits extra volatile qualifier for + * __attribute__((noreturn)) function pointers. Clang + * doesn't do it. It's a GCC quirk for backwards + * compatibility with code written for GCC <2.5. So, + * similarly to extra qualifiers for array, just drop + * them, instead of handling them. + */ + btf_dump_drop_mods(d, decls); + if (decls->cnt) { + btf_dump_printf(d, " ("); + btf_dump_emit_type_chain(d, decls, fname, lvl); + btf_dump_printf(d, ")"); + } else { + btf_dump_emit_name(d, fname, last_was_ptr); + } + btf_dump_printf(d, "("); + /* + * Clang for BPF target generates func_proto with no + * args as a func_proto with a single void arg (e.g., + * `int (*f)(void)` vs just `int (*f)()`). We are + * going to pretend there are no args for such case. + */ + if (vlen == 1 && p->type == 0) { + btf_dump_printf(d, ")"); + return; + } + + for (i = 0; i < vlen; i++, p++) { + if (i > 0) + btf_dump_printf(d, ", "); + + /* last arg of type void is vararg */ + if (i == vlen - 1 && p->type == 0) { + btf_dump_printf(d, "..."); + break; + } + + name = btf_name_of(d, p->name_off); + btf_dump_emit_type_decl(d, p->type, name, lvl); + } + + btf_dump_printf(d, ")"); + return; + } + default: + pr_warn("unexpected type in decl chain, kind:%u, id:[%u]\n", + kind, id); + return; + } + + last_was_ptr = kind == BTF_KIND_PTR; + } + + btf_dump_emit_name(d, fname, last_was_ptr); +} + +/* return number of duplicates (occurrences) of a given name */ +static size_t btf_dump_name_dups(struct btf_dump *d, struct hashmap *name_map, + const char *orig_name) +{ + char *old_name, *new_name; + size_t dup_cnt = 0; + int err; + + new_name = strdup(orig_name); + if (!new_name) + return 1; + + hashmap__find(name_map, orig_name, (void **)&dup_cnt); + dup_cnt++; + + err = hashmap__set(name_map, new_name, (void *)dup_cnt, + (const void **)&old_name, NULL); + if (err) + free(new_name); + + free(old_name); + + return dup_cnt; +} + +static const char *btf_dump_resolve_name(struct btf_dump *d, __u32 id, + struct hashmap *name_map) +{ + struct btf_dump_type_aux_state *s = &d->type_states[id]; + const struct btf_type *t = btf__type_by_id(d->btf, id); + const char *orig_name = btf_name_of(d, t->name_off); + const char **cached_name = &d->cached_names[id]; + size_t dup_cnt; + + if (t->name_off == 0) + return ""; + + if (s->name_resolved) + return *cached_name ? *cached_name : orig_name; + + if (btf_is_fwd(t) || (btf_is_enum(t) && btf_vlen(t) == 0)) { + s->name_resolved = 1; + return orig_name; + } + + dup_cnt = btf_dump_name_dups(d, name_map, orig_name); + if (dup_cnt > 1) { + const size_t max_len = 256; + char new_name[max_len]; + + snprintf(new_name, max_len, "%s___%zu", orig_name, dup_cnt); + *cached_name = strdup(new_name); + } + + s->name_resolved = 1; + return *cached_name ? *cached_name : orig_name; +} + +static const char *btf_dump_type_name(struct btf_dump *d, __u32 id) +{ + return btf_dump_resolve_name(d, id, d->type_names); +} + +static const char *btf_dump_ident_name(struct btf_dump *d, __u32 id) +{ + return btf_dump_resolve_name(d, id, d->ident_names); +} |