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] [thread-next>] [day] [month] [year] [list]
Date:   Wed, 18 Jan 2023 18:07:19 +0100
From:   Eric Dumazet <edumazet@...gle.com>
To:     Frederick Lawler <fred@...udflare.com>
Cc:     netdev@...r.kernel.org, kernel-team@...udflare.com,
        davem@...emloft.net, yoshfuji@...ux-ipv6.org, dsahern@...nel.org,
        kuba@...nel.org, pabeni@...hat.com
Subject: Re: BUG: using __this_cpu_add() in preemptible in tcp_make_synack()

On Wed, Jan 18, 2023 at 5:56 PM Frederick Lawler <fred@...udflare.com> wrote:
>
> Hello,
>
> We've been testing Linux 6.1.4, and came across an intermittent "BUG:
> using __this_cpu_add() in preemptible" [1] in our services leveraging
> TCP_FASTOPEN and our kernel configured with:
>
> CONFIG_PREEMPT_BUILD=y
> CONFIG_PREEMPT_VOLUNTARY=y
> CONFIG_PREEMPT_COUNT=y
> CONFIG_PREEMPTION=y
> CONFIG_PREEMPT_DYNAMIC=y
> CONFIG_PREEMPT_RCU=y
> CONFIG_HAVE_PREEMPT_DYNAMIC=y
> CONFIG_HAVE_PREEMPT_DYNAMIC_CALL=y
> CONFIG_PREEMPT_NOTIFIERS=y
> CONFIG_DEBUG_PREEMPT=y
>
> I'm not sure how related this is to commit 0a375c822497 ("tcp:
> tcp_rtx_synack() can be called from process context
> "), as I haven't found a reliable reproducer yet.
>
> The stack trace below has audit_*(), but we have other traces with
> do_tcp_setsockopt() mixed in the backtrace. Once we get to
> tcp_rx_synack() in the __release_sock() path, we get the same problem.
>
> [1]:
> BUG: using __this_cpu_add() in preemptible [00000000] code: nginx-ssl/209282
> caller is tcp_make_synack+0x38d/0x5a0
> CPU: 3 PID: 209282 Comm: nginx-ssl Kdump: loaded Tainted: G           O
>       6.1.4-cloudflare-2023.1.2 #1
> Hardware name: Quanta Cloud Technology Inc. QuantaPlex T42S-2U/T42S-2U
> MB (Lewisburg-1G) CLX, BIOS 3B16.Q102 02/19/2020
> Call Trace:
> <TASK>
>   dump_stack_lvl+0x34/0x48
>   check_preemption_disabled+0xdd/0xe0
>   tcp_make_synack+0x38d/0x5a0
>   tcp_v4_send_synack+0x50/0x1f0
>   tcp_rtx_synack+0x55/0x140
>   ? load_balance+0xa91/0xd40
>   ? _copy_to_iter+0x1d6/0x560
>   inet_rtx_syn_ack+0x16/0x30
>   tcp_check_req+0x39f/0x660
>   tcp_rcv_state_process+0xa3/0x1020
>   ? tcp_mstamp_refresh+0xe/0x40
>   ? tcp_update_recv_tstamps+0x61/0x90
>   ? tcp_recvmsg_locked+0x1eb/0x960
>   tcp_v4_do_rcv+0xc6/0x280
>   __release_sock+0xb4/0xc0
>   release_sock+0x2b/0x90
>   tcp_recvmsg+0x7c/0x200
>   inet_recvmsg+0x52/0x130
>   __sys_recvfrom+0xa8/0x120
>   ? audit_filter_inodes.part.0+0x2e/0x110
>   ? auditd_test_task+0x3c/0x50
>   ? __audit_syscall_entry+0xd5/0x120
>   __x64_sys_recvfrom+0x20/0x30
>   do_syscall_64+0x38/0x90
>   entry_SYSCALL_64_after_hwframe+0x4b/0xb5
> RIP: 0033:0x7fafc19adc74
> Code: 89 4c 24 1c e8 2d 41 f8 ff 44 8b 54 24 1c 8b 3c 24 45 31 c9 89 c5
> 48 8b 54 24 10 48 8b 74 24 08 45 31 c0 b8 2d 00 00 00 0f 05 <48> 3d 00
> f0 ff ff 77 34 89 ef 48 89 04 24 e8 59 41 f8 ff 48 8b 04
> RSP: 002b:00007ffc42e18a20 EFLAGS: 00000246 ORIG_RAX: 000000000000002d
> RAX: ffffffffffffffda RBX: 00007faf26457870 RCX: 00007fafc19adc74
> RDX: 0000000000000001 RSI: 00007ffc42e18a60 RDI: 000000000000076a
> RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000002 R11: 0000000000000246 R12: 00007faf2261fd40
> R13: 00007faf26457870 R14: 00007faf1fe962c0 R15: 00007ffc42e18a60
> </TASK>
>
> Thanks,
> Fred

Thanks for the report

I guess this part has been missed in commit 0a375c822497ed6a

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 71d01cf3c13eb4bd3d314ef140568d2ffd6a499e..ba839e441450f195012a8d77cb9e5ed956962d2f
100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -3605,7 +3605,7 @@ struct sk_buff *tcp_make_synack(const struct
sock *sk, struct dst_entry *dst,
        th->window = htons(min(req->rsk_rcv_wnd, 65535U));
        tcp_options_write(th, NULL, &opts);
        th->doff = (tcp_header_size >> 2);
-       __TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTSEGS);
+       TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTSEGS);

 #ifdef CONFIG_TCP_MD5SIG
        /* Okay, we have all we need - do the md5 hash if needed */

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ