On 3/26/24 9:47 PM, Yonghong Song wrote:
On 3/26/24 6:07 PM, Alexei Starovoitov wrote:
On Tue, Mar 26, 2024 at 5:54 PM Martin KaFai Lau <martin.lau@xxxxxxxxx> wrote:
On 3/25/24 2:36 AM, syzbot wrote:
Hello,
syzbot found the following issue on:
HEAD commit: 5e74df2f8f15 Merge tag 'x86-urgent-2024-03-24' of git://gi..
git tree: upstream
console+strace: https://syzkaller.appspot.com/x/log.txt?x=148872a5180000
kernel config: https://syzkaller.appspot.com/x/.config?x=e6bd769cb793b98a
dashboard link: https://syzkaller.appspot.com/bug?extid=1a3cf6f08d68868f9db3
compiler: Debian clang version 15.0.6, GNU ld (GNU Binutils for
Debian) 2.40
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=15921a6e180000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=12e081f1180000
Downloadable assets:
disk image:
https://storage.googleapis.com/syzbot-assets/1a82880723a7/disk-5e74df2f.raw.xz
vmlinux:
https://storage.googleapis.com/syzbot-assets/fd3046ac43b9/vmlinux-5e74df2f.xz
kernel image:
https://storage.googleapis.com/syzbot-assets/2097be59cbc1/bzImage-5e74df2f.xz
IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+1a3cf6f08d68868f9db3@xxxxxxxxxxxxxxxxxxxxxxxxx
=====================================================
BUG: KMSAN: uninit-value in __dev_map_lookup_elem kernel/bpf/devmap.c:441
[inline]
BUG: KMSAN: uninit-value in dev_map_lookup_elem+0xf3/0x170
kernel/bpf/devmap.c:796
__dev_map_lookup_elem kernel/bpf/devmap.c:441 [inline]
dev_map_lookup_elem+0xf3/0x170 kernel/bpf/devmap.c:796
____bpf_map_lookup_elem kernel/bpf/helpers.c:42 [inline]
bpf_map_lookup_elem+0x5c/0x80 kernel/bpf/helpers.c:38
___bpf_prog_run+0x13fe/0xe0f0 kernel/bpf/core.c:1997
__bpf_prog_run256+0xb5/0xe0 kernel/bpf/core.c:2237
It should be in the interpreter mode.
The C reproducer is trying to run the following bpf prog:
0: (18) r0 = 0x0
2: (18) r1 = map[id:49]
4: (b7) r8 = 16777216
5: (7b) *(u64 *)(r10 -8) = r8
6: (bf) r2 = r10
7: (07) r2 += -229
^^^^^^^^^^
8: (b7) r3 = 8
9: (b7) r4 = 0
10: (85) call dev_map_lookup_elem#1543472
11: (95) exit
I think this KMSAN report (and a few others related to lookup/delete_elem)
should only happen in the interpreter mode.
Does it worth to suppress it by always initializing the stack in the interpreter
mode considering the interpreter is not very speed sensitive ?
Maybe we can mark it as initialized from kmsan pov ?
There are kasan_poison/unpoison helpers that may fit ?
Maybe use kmsan_unpoison_memory()?
Good idea. It should do. My qemu cannot boot with CONFIG_KMSAN somehow.
I will ask syzbot to test.
#syz test: git://git.kernel.org/pub/scm/linux/kernel/git/martin.lau/bpf-next.git interpreter.kmsan
