From bfdd083bef8e47093190f4cbe5630ecffd7af79f Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Fri, 22 Nov 2019 21:07:57 +0100 Subject: [PATCH] UPSTREAM: bpf: Add initial poke descriptor table for jit images Add initial poke table data structures and management to the BPF prog that can later be used by JITs. Also add an instance of poke specific data for tail call maps; plan for later work is to extend this also for BPF static keys. Change-Id: Ia0c9e1e029cac3bd9ee028bccd2b50b4815947f5 Signed-off-by: Daniel Borkmann Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/1db285ec2ea4207ee0455b3f8e191a4fc58b9ade.1574452833.git.daniel@iogearbox.net --- include/linux/bpf.h | 20 ++++++++++++++++++++ include/linux/filter.h | 10 ++++++++++ kernel/bpf/core.c | 34 ++++++++++++++++++++++++++++++++++ 3 files changed, 64 insertions(+) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 788486bd8645..7e7c64753367 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -488,6 +488,24 @@ struct bpf_func_info_aux { bool unreliable; }; +enum bpf_jit_poke_reason { + BPF_POKE_REASON_TAIL_CALL, +}; + +/* Descriptor of pokes pointing /into/ the JITed image. */ +struct bpf_jit_poke_descriptor { + void *ip; + union { + struct { + struct bpf_map *map; + u32 key; + } tail_call; + }; + bool ip_stable; + u8 adj_off; + u16 reason; +}; + struct bpf_prog_aux { atomic64_t refcnt; u32 used_map_cnt; @@ -513,6 +531,8 @@ struct bpf_prog_aux { const char *attach_func_name; struct bpf_prog **func; void *jit_data; /* JIT specific data. arch dependent */ + struct bpf_jit_poke_descriptor *poke_tab; + u32 size_poke_tab; struct latch_tree_node ksym_tnode; struct list_head ksym_lnode; const struct bpf_prog_ops *ops; diff --git a/include/linux/filter.h b/include/linux/filter.h index 4413655c0f33..537d5a14e99d 100644 --- a/include/linux/filter.h +++ b/include/linux/filter.h @@ -1035,6 +1035,9 @@ void *bpf_jit_alloc_exec(unsigned long size); void bpf_jit_free_exec(void *addr); void bpf_jit_free(struct bpf_prog *fp); +int bpf_jit_add_poke_descriptor(struct bpf_prog *prog, + struct bpf_jit_poke_descriptor *poke); + int bpf_jit_get_func_addr(const struct bpf_prog *prog, const struct bpf_insn *insn, bool extra_pass, u64 *func_addr, bool *func_addr_fixed); @@ -1138,6 +1141,13 @@ static inline bool bpf_prog_ebpf_jited(const struct bpf_prog *fp) return false; } +static inline int +bpf_jit_add_poke_descriptor(struct bpf_prog *prog, + struct bpf_jit_poke_descriptor *poke) +{ + return -ENOTSUPP; +} + static inline void bpf_jit_free(struct bpf_prog *fp) { bpf_prog_unlock_free(fp); diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c index 33da13f94316..90e03944d18c 100644 --- a/kernel/bpf/core.c +++ b/kernel/bpf/core.c @@ -261,6 +261,7 @@ void __bpf_prog_free(struct bpf_prog *fp) { if (fp->aux) { free_percpu(fp->aux->stats); + kfree(fp->aux->poke_tab); kfree(fp->aux); } vfree(fp); @@ -766,6 +767,39 @@ int bpf_get_kallsym(unsigned int symnum, unsigned long *value, char *type, return ret; } +int bpf_jit_add_poke_descriptor(struct bpf_prog *prog, + struct bpf_jit_poke_descriptor *poke) +{ + struct bpf_jit_poke_descriptor *tab = prog->aux->poke_tab; + static const u32 poke_tab_max = 1024; + u32 slot = prog->aux->size_poke_tab; + u32 size = slot + 1; + + if (size > poke_tab_max) + return -ENOSPC; + if (poke->ip || poke->ip_stable || poke->adj_off) + return -EINVAL; + + switch (poke->reason) { + case BPF_POKE_REASON_TAIL_CALL: + if (!poke->tail_call.map) + return -EINVAL; + break; + default: + return -EINVAL; + } + + tab = krealloc(tab, size * sizeof(*poke), GFP_KERNEL); + if (!tab) + return -ENOMEM; + + memcpy(&tab[slot], poke, sizeof(*poke)); + prog->aux->size_poke_tab = size; + prog->aux->poke_tab = tab; + + return slot; +} + static atomic_long_t bpf_jit_current; /* Can be overridden by an arch's JIT compiler if it has a custom,