Date:   Sun, 30 Oct 2022 09:27:46 -0700
From:   Eric Dumazet <edumazet@...gle.com>
To:     Wei Chen <harperchen1110@...il.com>
Cc:     davem@...emloft.net, yoshfuji@...ux-ipv6.org, dsahern@...nel.org,
        kuba@...nel.org, ast@...nel.org, daniel@...earbox.net,
        andrii@...nel.org, kafai@...com, songliubraving@...com, yhs@...com,
        john.fastabend@...il.com, kpsingh@...nel.org,
        netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        bpf@...r.kernel.org
Subject: Re: BUG: unable to handle kernel paging request in tcp_retransmit_timer

On Sun, Oct 30, 2022 at 2:28 AM Wei Chen <harperchen1110@...il.com> wrote:
>
> Dear Linux Developer,
>
> Recently when using our tool to fuzz kernel, the following crash was triggered:
>
> HEAD commit: 64570fbc14f8 Linux 5.15-rc5

This is a quite old kernel. Please do not send reports on old rc kernels.

> git tree: upstream
> compiler: gcc 8.0.1
> console output:
> https://drive.google.com/file/d/1wVTAdDoOo8KqTaGm1v8SaKuv1V8Pt9qs/view?usp=share_link
> kernel config: https://drive.google.com/file/d/1uDOeEYgJDcLiSOrx9W8v2bqZ6uOA_55t/view?usp=share_link
>
> Unfortunately, I don't have any reproducer for this crash yet.

We already have syzbot reports like this one.

The important missing part is a reproducer, really.

See recent work that has been done recently in order to find the root
cause for these issue(s) in net-next.

0cafd77dcd03 net: add a refcount tracker for kernel sockets
d1e96cc4fbe0 mptcp: fix tracking issue in mptcp_subflow_create_socket()

Make sure to use a recent tree, if you really want your fuzzer to
participate in the effort.
Also enable:

CONFIG_NET_DEV_REFCNT_TRACKER=y


Thanks.

>
> IMPORTANT: if you fix the bug, please add the following tag to the commit:
> Reported-by: Wei Chen <harperchen1110@...il.com>
>
> BUG: unable to handle page fault for address: ffffe8ff3fa5f268
> #PF: supervisor write access in kernel mode
> #PF: error_code(0x0002) - not-present page
> PGD 983f067 P4D 983f067 PUD afce067 PMD 4e244067 PTE 0
> Oops: 0002 [#1] PREEMPT SMP
> CPU: 0 PID: 6544 Comm: syz-fuzzer Not tainted 5.15.0-rc5 #1
> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
> rel-1.13.0-48-gd9c812dda519-prebuilt.qemu.org 04/01/2014
> RIP: 0010:tcp_retransmit_timer+0x4c5/0x1540
> Code: 31 e7 ff ff e9 65 fd ff ff e8 b7 75 3c fd 48 c7 c7 26 1c ee 85
> e8 8b fa bc 00 48 8b 43 30 bf 1f 00 00 00 48 8b 80 58 02 00 00 <65> 48
> ff 80 40 01 00 00 44 0f b6 73 12 48 8b 43 30 44 89 f6 48 89
> RSP: 0000:ffffc90000807cc0 EFLAGS: 00010202
> RAX: 0000607ec1e5f128 RBX: ffff8880156c0000 RCX: ffff888011480000
> RDX: 0000000000000000 RSI: 0000000000000101 RDI: 000000000000001f
> RBP: ffff8880156c0120 R08: ffffffff8400fda9 R09: 0000000000000000
> R10: 0000000000000005 R11: 0000000080000001 R12: 0000000080000001
> R13: ffff88810cd1b280 R14: ffff888029b5f400 R15: ffff8880156c0278
> FS:  000000c000030c90(0000) GS:ffff88807dc00000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: ffffe8ff3fa5f268 CR3: 0000000015c0b000 CR4: 00000000003506f0
> Call Trace:
>  tcp_write_timer_handler+0x132/0x420
>  tcp_write_timer+0x179/0x230
>  call_timer_fn+0xe8/0x510
>  run_timer_softirq+0x423/0xa40
>  __do_softirq+0xe2/0x56b
>  irq_exit_rcu+0xb6/0xf0
>  sysvec_apic_timer_interrupt+0x52/0xc0
>  asm_sysvec_apic_timer_interrupt+0x12/0x20
> RIP: 0033:0x415543
> Code: 48 8b 1d a0 e8 76 01 84 03 48 8b 14 d3 48 85 d2 74 1d 48 89 c3
> 48 c1 e8 0d 48 25 ff 1f 00 00 48 8b 8c c2 00 00 20 00 48 89 d8 <e9> 6c
> fe ff ff 31 c9 e9 65 fe ff ff cc cc cc cc cc cc cc cc cc cc
> RSP: 002b:000000c00003de70 EFLAGS: 00000202
> RAX: 000000c004cc8600 RBX: 000000c004cc8600 RCX: 00007f27b2e23400
> RDX: 00007f27b2e3b000 RSI: 0000000000000001 RDI: 00000000000dcf40
> RBP: 000000c00003de98 R08: 00007f27b303afff R09: 000000c004beb6c0
> R10: 000000c000021e98 R11: 0000000000000008 R12: 000000c004cc8600
> R13: 000000c000001200 R14: 0000000000c4de75 R15: 0000000000000000
> Modules linked in:
> CR2: ffffe8ff3fa5f268
> ---[ end trace 8795388675688c1b ]---
> RIP: 0010:tcp_retransmit_timer+0x4c5/0x1540
> Code: 31 e7 ff ff e9 65 fd ff ff e8 b7 75 3c fd 48 c7 c7 26 1c ee 85
> e8 8b fa bc 00 48 8b 43 30 bf 1f 00 00 00 48 8b 80 58 02 00 00 <65> 48
> ff 80 40 01 00 00 44 0f b6 73 12 48 8b 43 30 44 89 f6 48 89
> RSP: 0000:ffffc90000807cc0 EFLAGS: 00010202
> RAX: 0000607ec1e5f128 RBX: ffff8880156c0000 RCX: ffff888011480000
> RDX: 0000000000000000 RSI: 0000000000000101 RDI: 000000000000001f
> RBP: ffff8880156c0120 R08: ffffffff8400fda9 R09: 0000000000000000
> R10: 0000000000000005 R11: 0000000080000001 R12: 0000000080000001
> R13: ffff88810cd1b280 R14: ffff888029b5f400 R15: ffff8880156c0278
> FS:  000000c000030c90(0000) GS:ffff88807dc00000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: ffffe8ff3fa5f268 CR3: 0000000015c0b000 CR4: 00000000003506f0
> ----------------
> Code disassembly (best guess), 4 bytes skipped:
>    0: e9 65 fd ff ff        jmpq   0xfffffd6a
>    5: e8 b7 75 3c fd        callq  0xfd3c75c1
>    a: 48 c7 c7 26 1c ee 85 mov    $0xffffffff85ee1c26,%rdi
>   11: e8 8b fa bc 00        callq  0xbcfaa1
>   16: 48 8b 43 30          mov    0x30(%rbx),%rax
>   1a: bf 1f 00 00 00        mov    $0x1f,%edi
>   1f: 48 8b 80 58 02 00 00 mov    0x258(%rax),%rax
> * 26: 65 48 ff 80 40 01 00 incq   %gs:0x140(%rax) <-- trapping instruction
>   2d: 00
>   2e: 44 0f b6 73 12        movzbl 0x12(%rbx),%r14d
>   33: 48 8b 43 30          mov    0x30(%rbx),%rax
>   37: 44 89 f6              mov    %r14d,%esi
>   3a: 48                    rex.W
>   3b: 89                    .byte 0x89
>
> Best,
> Wei

Powered by blists - more mailing lists