linux-kernel - Re: [PATCH bpf-next 1/2] bpf: Fix tnum_overlap to check for zero mask first

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <769c6167e9e650348f92b90c538b93c565a9ae11.camel@linux.dev>
Date: Mon, 27 Oct 2025 23:35:05 +0800
From: KaFai Wan <kafai.wan@...ux.dev>
To: ast@...nel.org, daniel@...earbox.net, john.fastabend@...il.com, 
 andrii@...nel.org, martin.lau@...ux.dev, eddyz87@...il.com,
 song@...nel.org,  yonghong.song@...ux.dev, kpsingh@...nel.org,
 sdf@...ichev.me, haoluo@...gle.com,  jolsa@...nel.org, shuah@...nel.org,
 paul.chaignon@...il.com, m.shachnai@...il.com,  memxor@...il.com,
 harishankar.vishwanathan@...il.com, colin.i.king@...il.com, 
 luis.gerhorst@....de, shung-hsi.yu@...e.com, bpf@...r.kernel.org, 
 linux-kernel@...r.kernel.org, linux-kselftest@...r.kernel.org
Cc: syzbot+c950cc277150935cc0b5@...kaller.appspotmail.com
Subject: Re: [PATCH bpf-next 1/2] bpf: Fix tnum_overlap to check for zero
 mask first

Sorry, this patch is wrong, please ignore.

On Mon, 2025-10-27 at 00:38 +0800, KaFai Wan wrote:
> Syzbot reported a kernel warning due to a range invariant violation in
> the BPF verifier. The issue occurs when tnum_overlap() fails to detect
> that two tnums don't have any overlapping bits.
> 
> The problematic BPF program:
>    0: call bpf_get_prandom_u32
>    1: r6 = r0
>    2: r6 &= 0xFFFFFFFFFFFFFFF0
>    3: r7 = r0
>    4: r7 &= 0x07
>    5: r7 -= 0xFF
>    6: if r6 == r7 goto <exit>
> 
> After instruction 5, R7 has the range:
>    R7: u64=[0xffffffffffffff01, 0xffffffffffffff08] var_off=(0xffffffffffffff00; 0xf)
> 
> R6 and R7 don't overlap since they have no agreeing bits. However,
> is_branch_taken() fails to recognize this, causing the verifier to
> refine register bounds and end up with inconsistent bounds:
> 
>    6: if r6 == r7 goto <exit>
>    R6: u64=[0xffffffffffffff01, 0xffffffffffffff00] var_off=(0xffffffffffffff00, 0x0)
>    R7: u64=[0xffffffffffffff01, 0xffffffffffffff00] var_off=(0xffffffffffffff00, 0x0)
> 
> The root cause is that tnum_overlap() doesn't properly handle the case
> where the masks have no overlapping bits.
> 
> Fix this by adding an early check for zero mask intersection in tnum_overlap().
> 
> Reported-by: syzbot+c950cc277150935cc0b5@...kaller.appspotmail.com
> Fixes: f41345f47fb2 ("bpf: Use tnums for JEQ/JNE is_branch_taken logic")
> Signed-off-by: KaFai Wan <kafai.wan@...ux.dev>
> ---
>  kernel/bpf/tnum.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/kernel/bpf/tnum.c b/kernel/bpf/tnum.c
> index f8e70e9c3998..af2f38b4f840 100644
> --- a/kernel/bpf/tnum.c
> +++ b/kernel/bpf/tnum.c
> @@ -163,6 +163,8 @@ bool tnum_overlap(struct tnum a, struct tnum b)
>  {
>  	u64 mu;
>  
> +	if ((a.mask & b.mask) == 0)
> +		return false;
>  	mu = ~a.mask & ~b.mask;
>  	return (a.value & mu) == (b.value & mu);
>  }

-- 
Thanks,
KaFai