Subject: From: Peter Zijlstra Date: Thu Jan 17 13:19:25 CET 2019 Signed-off-by: Peter Zijlstra (Intel) --- include/linux/bpf.h | 7 +- include/linux/filter.h | 42 ------------ kernel/bpf/core.c | 164 ++++--------------------------------------------- kernel/extable.c | 4 - kernel/kallsyms.c | 19 ----- 5 files changed, 22 insertions(+), 214 deletions(-) --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -13,7 +13,7 @@ #include #include #include -#include +#include #include #include @@ -307,8 +307,9 @@ struct bpf_prog_aux { bool offload_requested; struct bpf_prog **func; void *jit_data; /* JIT specific data. arch dependent */ - struct latch_tree_node ksym_tnode; - struct list_head ksym_lnode; + + struct kallsym_node ktn; + const struct bpf_prog_ops *ops; struct bpf_map **used_maps; struct bpf_prog *prog; --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -932,23 +932,6 @@ static inline bool bpf_jit_kallsyms_enab return false; } -const char *__bpf_address_lookup(unsigned long addr, unsigned long *size, - unsigned long *off, char *sym); -bool is_bpf_text_address(unsigned long addr); -int bpf_get_kallsym(unsigned int symnum, unsigned long *value, char *type, - char *sym); - -static inline const char * -bpf_address_lookup(unsigned long addr, unsigned long *size, - unsigned long *off, char **modname, char *sym) -{ - const char *ret = __bpf_address_lookup(addr, size, off, sym); - - if (ret && modname) - *modname = NULL; - return ret; -} - void bpf_prog_kallsyms_add(struct bpf_prog *fp); void bpf_prog_kallsyms_del(struct bpf_prog *fp); @@ -974,31 +957,6 @@ static inline bool bpf_jit_kallsyms_enab return false; } -static inline const char * -__bpf_address_lookup(unsigned long addr, unsigned long *size, - unsigned long *off, char *sym) -{ - return NULL; -} - -static inline bool is_bpf_text_address(unsigned long addr) -{ - return false; -} - -static inline int bpf_get_kallsym(unsigned int symnum, unsigned long *value, - char *type, char *sym) -{ - return -ERANGE; -} - -static inline const char * -bpf_address_lookup(unsigned long addr, unsigned long *size, - unsigned long *off, char **modname, char *sym) -{ - return NULL; -} - static inline void bpf_prog_kallsyms_add(struct bpf_prog *fp) { } --- a/kernel/bpf/core.c +++ b/kernel/bpf/core.c @@ -30,7 +30,6 @@ #include #include #include -#include #include #include #include @@ -100,8 +99,6 @@ struct bpf_prog *bpf_prog_alloc(unsigned fp->aux->prog = fp; fp->jit_requested = ebpf_jit_enabled(); - INIT_LIST_HEAD_RCU(&fp->aux->ksym_lnode); - return fp; } EXPORT_SYMBOL_GPL(bpf_prog_alloc); @@ -530,86 +527,35 @@ static void bpf_get_prog_name(const stru *sym = 0; } -static __always_inline unsigned long -bpf_get_prog_addr_start(struct latch_tree_node *n) -{ - unsigned long symbol_start, symbol_end; - const struct bpf_prog_aux *aux; - - aux = container_of(n, struct bpf_prog_aux, ksym_tnode); - bpf_get_prog_addr_region(aux->prog, &symbol_start, &symbol_end); - - return symbol_start; -} - -static __always_inline bool bpf_tree_less(struct latch_tree_node *a, - struct latch_tree_node *b) -{ - return bpf_get_prog_addr_start(a) < bpf_get_prog_addr_start(b); -} - -static __always_inline int bpf_tree_comp(void *key, struct latch_tree_node *n) -{ - unsigned long val = (unsigned long)key; - unsigned long symbol_start, symbol_end; - const struct bpf_prog_aux *aux; - - aux = container_of(n, struct bpf_prog_aux, ksym_tnode); - bpf_get_prog_addr_region(aux->prog, &symbol_start, &symbol_end); - - if (val < symbol_start) - return -1; - if (val >= symbol_end) - return 1; - - return 0; -} - -static const struct latch_tree_ops bpf_tree_ops = { - .less = bpf_tree_less, - .comp = bpf_tree_comp, -}; - -static DEFINE_SPINLOCK(bpf_lock); -static LIST_HEAD(bpf_kallsyms); -static struct latch_tree_root bpf_tree __cacheline_aligned; - -static void bpf_prog_ksym_node_add(struct bpf_prog_aux *aux) -{ - WARN_ON_ONCE(!list_empty(&aux->ksym_lnode)); - list_add_tail_rcu(&aux->ksym_lnode, &bpf_kallsyms); - latch_tree_insert(&aux->ksym_tnode, &bpf_tree, &bpf_tree_ops); -} - -static void bpf_prog_ksym_node_del(struct bpf_prog_aux *aux) -{ - if (list_empty(&aux->ksym_lnode)) - return; - - latch_tree_erase(&aux->ksym_tnode, &bpf_tree, &bpf_tree_ops); - list_del_rcu(&aux->ksym_lnode); -} static bool bpf_prog_kallsyms_candidate(const struct bpf_prog *fp) { return fp->jited && !bpf_prog_was_classic(fp); } -static bool bpf_prog_kallsyms_verify_off(const struct bpf_prog *fp) +static void bpf_kn_names(struct kallsym_node *kn, char *sym, char **modname) { - return list_empty(&fp->aux->ksym_lnode) || - fp->aux->ksym_lnode.prev == LIST_POISON2; + struct bpf_prog_aux *aux = container_of(kn, struct bpf_prog_aux, ktn); + + *modname = "eBPF-jit"; + bpf_get_prog_name(aux->prog, sym); } void bpf_prog_kallsyms_add(struct bpf_prog *fp) { + unsigned long sym_start, sym_end; + if (!bpf_prog_kallsyms_candidate(fp) || !capable(CAP_SYS_ADMIN)) return; - spin_lock_bh(&bpf_lock); - bpf_prog_ksym_node_add(fp->aux); - spin_unlock_bh(&bpf_lock); + bpf_get_prog_addr_region(fp, &sym_start, &sym_end); + + fp->aux->ktn.kn_addr = sym_start; + fp->aux->ktn.kn_len = sym_end - sym_start; + fp->aux->ktn.kn_names = bpf_kn_names; + + kallsym_tree_add(&fp->aux->ktn); } void bpf_prog_kallsyms_del(struct bpf_prog *fp) @@ -617,85 +563,7 @@ void bpf_prog_kallsyms_del(struct bpf_pr if (!bpf_prog_kallsyms_candidate(fp)) return; - spin_lock_bh(&bpf_lock); - bpf_prog_ksym_node_del(fp->aux); - spin_unlock_bh(&bpf_lock); -} - -static struct bpf_prog *bpf_prog_kallsyms_find(unsigned long addr) -{ - struct latch_tree_node *n; - - if (!bpf_jit_kallsyms_enabled()) - return NULL; - - n = latch_tree_find((void *)addr, &bpf_tree, &bpf_tree_ops); - return n ? - container_of(n, struct bpf_prog_aux, ksym_tnode)->prog : - NULL; -} - -const char *__bpf_address_lookup(unsigned long addr, unsigned long *size, - unsigned long *off, char *sym) -{ - unsigned long symbol_start, symbol_end; - struct bpf_prog *prog; - char *ret = NULL; - - rcu_read_lock(); - prog = bpf_prog_kallsyms_find(addr); - if (prog) { - bpf_get_prog_addr_region(prog, &symbol_start, &symbol_end); - bpf_get_prog_name(prog, sym); - - ret = sym; - if (size) - *size = symbol_end - symbol_start; - if (off) - *off = addr - symbol_start; - } - rcu_read_unlock(); - - return ret; -} - -bool is_bpf_text_address(unsigned long addr) -{ - bool ret; - - rcu_read_lock(); - ret = bpf_prog_kallsyms_find(addr) != NULL; - rcu_read_unlock(); - - return ret; -} - -int bpf_get_kallsym(unsigned int symnum, unsigned long *value, char *type, - char *sym) -{ - struct bpf_prog_aux *aux; - unsigned int it = 0; - int ret = -ERANGE; - - if (!bpf_jit_kallsyms_enabled()) - return ret; - - rcu_read_lock(); - list_for_each_entry_rcu(aux, &bpf_kallsyms, ksym_lnode) { - if (it++ != symnum) - continue; - - bpf_get_prog_name(aux->prog, sym); - - *value = (unsigned long)aux->prog->bpf_func; - *type = BPF_SYM_ELF_TYPE; - - ret = 0; - break; - } - rcu_read_unlock(); - - return ret; + kallsym_tree_del(&fp->aux->ktn); } static atomic_long_t bpf_jit_current; @@ -806,8 +674,6 @@ void __weak bpf_jit_free(struct bpf_prog bpf_jit_binary_unlock_ro(hdr); bpf_jit_binary_free(hdr); - - WARN_ON_ONCE(!bpf_prog_kallsyms_verify_off(fp)); } bpf_prog_unlock_free(fp); --- a/kernel/extable.c +++ b/kernel/extable.c @@ -135,7 +135,7 @@ int kernel_text_address(unsigned long ad * coming back from idle, or cpu on or offlining. * * is_module_text_address() as well as the kprobe slots - * and is_bpf_text_address() require RCU to be watching. + * and is_kallsym_tree_text_address() require RCU to be watching. */ no_rcu = !rcu_is_watching(); @@ -151,8 +151,6 @@ int kernel_text_address(unsigned long ad goto out; if (is_kprobe_optinsn_slot(addr) || is_kprobe_insn_slot(addr)) goto out; - if (is_bpf_text_address(addr)) - goto out; ret = 0; out: if (no_rcu) --- a/kernel/kallsyms.c +++ b/kernel/kallsyms.c @@ -425,8 +425,7 @@ int kallsyms_lookup_size_offset(unsigned if (is_ksym_addr(addr)) return !!get_symbol_pos(addr, symbolsize, offset); return !!module_address_lookup(addr, symbolsize, offset, NULL, namebuf) || - !!kallsym_tree_address_lookup(addr, symbolsize, offset, NULL, namebuf) || - !!__bpf_address_lookup(addr, symbolsize, offset, namebuf); + !!kallsym_tree_address_lookup(addr, symbolsize, offset, NULL, namebuf); } /* @@ -464,11 +463,6 @@ const char *kallsyms_lookup(unsigned lon if (!ret) ret = kallsym_tree_address_lookup(addr, symbolsize, offset, modname, namebuf); - - if (!ret) - ret = bpf_address_lookup(addr, symbolsize, - offset, modname, namebuf); - if (!ret) ret = ftrace_mod_address_lookup(addr, symbolsize, offset, modname, namebuf); @@ -673,15 +667,6 @@ static int get_ksymbol_ftrace_mod(struct return 1; } -static int get_ksymbol_bpf(struct kallsym_iter *iter) -{ - iter->module_name[0] = '\0'; - iter->exported = 0; - return bpf_get_kallsym(iter->pos - iter->pos_ftrace_mod_end, - &iter->value, &iter->type, - iter->name) < 0 ? 0 : 1; -} - /* Returns space to next name. */ static unsigned long get_ksymbol_core(struct kallsym_iter *iter) { @@ -734,7 +719,7 @@ static int update_iter_mod(struct kallsy get_ksymbol_ftrace_mod(iter)) return 1; - return get_ksymbol_bpf(iter); + return 0; } /* Returns false if pos at or past end of file. */