]> git.ipfire.org Git - thirdparty/kernel/stable.git/commit
bpf: handle jset (if a & b ...) as a jump in CFG computation
authorEduard Zingerman <eddyz87@gmail.com>
Fri, 13 Jun 2025 17:53:30 +0000 (10:53 -0700)
committerAlexei Starovoitov <ast@kernel.org>
Fri, 13 Jun 2025 18:51:19 +0000 (11:51 -0700)
commit3157f7e2999616ac91f4d559a8566214f74000a5
tree8e49fb984396c53c58ef8ba50bed44b512836687
parente4c8f96adeb29a98f7848ee793085765cdede64c
bpf: handle jset (if a & b ...) as a jump in CFG computation

BPF_JSET is a conditional jump and currently verifier.c:can_jump()
does not know about that. This can lead to incorrect live registers
and SCC computation.

E.g. in the following example:

   1: r0 = 1;
   2: r2 = 2;
   3: if r1 & 0x7 goto +1;
   4: exit;
   5: r0 = r2;
   6: exit;

W/o this fix insn_successors(3) will return only (4), a jump to (5)
would be missed and r2 won't be marked as alive at (3).

Fixes: 14c8552db644 ("bpf: simple DFA-based live registers analysis")
Reported-by: syzbot+a36aac327960ff474804@syzkaller.appspotmail.com
Suggested-by: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20250613175331.3238739-1-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
kernel/bpf/verifier.c