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.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/1db285ec2ea4207ee0455b3f8e191a4fc58b9ade.1574452833.git.daniel@iogearbox.net

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index c3b29061284e542cd2a7c0451387b299dd030899..312983bf7faaa184ec1fc8ff0675e8a0adc4e9ad 100644 (file)
--- 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 ad80e9c6111c9f6cb4c152e45e985fca56a15faf..796b60d8cc6c1c1b78adc5fc427f96f8665ce9e4 100644 (file)
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -952,6 +952,9 @@ 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);
@@ -1055,6 +1058,13 @@ 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 07af9c1d9cf177ce132993ff5456ea8ce34861eb..608b7085e0c9296836c271f3f77dd9c0601c4740 100644 (file)
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -256,6 +256,7 @@ 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);
@@ -756,6 +757,39 @@ 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,