Re: [PATCH 3/5] bpf: Add strict alignment flag for BPF_PROG_LOAD.
From: Daniel Borkmann <daniel@iogearbox.net>
Date: 2017-05-11 12:54:03
On 05/10/2017 09:09 PM, David Miller wrote:
Add a new field, "prog_flags", and an initial flag value BPF_F_STRCIT_ALIGNMENT. When set, the verifier will enforce strict pointer alignment regardless of the setting of CONFIG_EFFICIENT_UNALIGNED_ACCESS. The verifier, in this mode, will also use a fixed value of "2" in place of NET_IP_ALIGN. This facilitates test cases that will exercise and validate this part of the verifier even when run on architectures where alignment doesn't matter. Signed-off-by: David S. Miller <davem@davemloft.net>
[...]
quoted hunk ↗ jump to hunk
@@ -833,10 +838,12 @@ static int check_val_ptr_alignment(const struct bpf_reg_state *reg, return 0; } +static bool strict_alignment; + static int check_ptr_alignment(const struct bpf_reg_state *reg, int off, int size) { - bool strict = false; + bool strict = strict_alignment; if (!IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)) strict = true;@@ -3574,6 +3581,10 @@ int bpf_check(struct bpf_prog **prog, union bpf_attr *attr) } else { log_level = 0; } + if (attr->prog_flags & BPF_F_STRICT_ALIGNMENT) + strict_alignment = true; + else + strict_alignment = false;
Just minor nit: Can we move this into struct bpf_verifier_env here instead of global var? The only change it would need is in check_ptr_alignment() to pass the env from check_mem_access(). check_ptr_alignment() can then infer this from env.
quoted hunk ↗ jump to hunk
ret = replace_map_fd_with_map_ptr(env); if (ret < 0)@@ -3679,6 +3690,7 @@ int bpf_analyzer(struct bpf_prog *prog, const struct bpf_ext_analyzer_ops *ops, mutex_lock(&bpf_verifier_lock); log_level = 0; + strict_alignment = false; env->explored_states = kcalloc(env->prog->len, sizeof(struct bpf_verifier_state_list *),
Rest looks good: Acked-by: Daniel Borkmann <daniel@iogearbox.net>