Re: network crash in nhop_free

From: Alexander V. Chernikov <melifaro_at_ipfw.ru>
Date: Mon, 30 Aug 2021 09:22:03 +0100
> On 30 Aug 2021, at 08:32, Andriy Gapon <avg_at_FreeBSD.org> wrote:
> 
> On 30/08/2021 10:28, Andriy Gapon wrote:
>> On 01/08/2021 16:36, Alexander V. Chernikov wrote:
>>> 
>>> 
>>>> On 10 Jul 2021, at 10:07, Andriy Gapon <avg_at_FreeBSD.org> wrote:
>>>> 
>>>> On 09/07/2021 00:02, Alexander V. Chernikov wrote:
>>>>> Hi Andriy,
>>>>> Could you by any chance provide a bit more info on the system networking configuration and the steps leading to panic?
>>>>> No chance for a coredump?
>>>>> destroy_nhgrp() suggests that there was a multipath route (default?) that was deleted.
>>>>> nhops are created with UMA_ALIGN_PTR, so I suspect there is a garbage inside nhgrp pointer..
>>>> 
>>>> I've just reproduced the problem and got a crash dump.
>>>> The new panic is a little bit different, but I think that it confirms your analysis.
>>>> Also, you are right about the multipath route, although its creation was not intentional.
>>> 
>>> Should be fixed by https://cgit.freebsd.org/src/commit/?id=054948bd81bb9e4e32449cf351b62e501b8831ff .
>> I have to report that, unfortunately, as of main bb958dcf3d8af3a033dacbf8133681c9b0c73b2f I can still reproduce the same panic using the same steps.
Thanks for reporting!
>> To be clear, as I reported two similar but still distinct panics, it's the first panic, "Misaligned access from kernel space!".
>> I should also add that the commit message does not really match my scenario.
>> In my case routes do not change quite fast.  I have generous pauses between starting and stopping ppp.
>> I have a feeling that there must something more deterministic that leads to the crash.
I tried to reproduced it with delays, then removed the delays and got a similar panic - that’s how I got to the previous bug.
I’ll set up an arm64 machine hopefully this week and try to reproduce it.
> 
> Some more details from the today's crash:
Is there any chance you can share kernel/core?
> 
> panic: Misaligned access from kernel space!
> cpuid = 0
> time = 1630308311
> KDB: stack backtrace:
> db_trace_self() at db_trace_self
> db_trace_self_wrapper() at db_trace_self_wrapper+0x30
> vpanic() at vpanic+0x184
> panic() at panic+0x44
> align_abort() at align_abort+0xb8
> handle_el1h_sync() at handle_el1h_sync+0x78
> --- exception, esr 0x96000021
> nhop_free() at nhop_free+0x100
> destroy_nhgrp() at destroy_nhgrp+0x38
> epoch_call_task() at epoch_call_task+0x158
> gtaskqueue_run_locked() at gtaskqueue_run_locked+0x178
> gtaskqueue_thread_loop() at gtaskqueue_thread_loop+0xc8
> fork_exit() at fork_exit+0x74
> fork_trampoline() at fork_trampoline+0x14
> Uptime: 11m17s
> Dumping 150 out of 998 MB:..3%..11%..22%..32%..43%..51%..62%..72%..83%
> 
> get_curthread () at /usr/devel/git/rock/sys/arm64/include/pcpu.h:68
> 68      /usr/devel/git/rock/sys/arm64/include/pcpu.h: No such file or directory.
> (kgdb) bt
> #0  get_curthread () at /usr/devel/git/rock/sys/arm64/include/pcpu.h:68
> #1  doadump (textdump=textdump_at_entry=1) at /usr/devel/git/rock/sys/kern/kern_shutdown.c:417
> #2  0xffff0000003bebf0 in kern_reboot (howto=260) at /usr/devel/git/rock/sys/kern/kern_shutdown.c:504
> #3  0xffff0000003bf10c in vpanic (fmt=<optimized out>, ap=...) at /usr/devel/git/rock/sys/kern/kern_shutdown.c:947
> #4  0xffff0000003bee3c in panic (fmt=0x0) at /usr/devel/git/rock/sys/kern/kern_shutdown.c:871
> #5  0xffff0000006c2054 in align_abort (td=<optimized out>, frame=<optimized out>, esr=2516582433, far=16045693110842147062, lower=<optimized out>) at /usr/devel/git/rock/sys/arm64/arm64/trap.c:212
> #6  <signal handler called>
> #7  atomic_fetchadd_32_llsc (p=0xdeadc0dedeadc0f6, val=4294967295) at /usr/devel/git/rock/sys/arm64/include/atomic.h:316
> #8  atomic_fetchadd_32 (p=<optimized out>, val=4294967295) at /usr/devel/git/rock/sys/arm64/include/atomic.h:316
> #9  refcount_releasen (count=0xdeadc0dedeadc0f6, n=1) at /usr/devel/git/rock/sys/sys/refcount.h:152
> #10 refcount_release (count=0xdeadc0dedeadc0f6) at /usr/devel/git/rock/sys/sys/refcount.h:174
> #11 nhop_free (nh=<optimized out>) at /usr/devel/git/rock/sys/net/route/nhop_ctl.c:669
> #12 0xffff000000506268 in free_nhgrp_nhops (nhg_priv=0xffffa00000d31b98) at /usr/devel/git/rock/sys/net/route/nhgrp_ctl.c:423
> #13 destroy_nhgrp (nhg_priv=0xffffa00000d31b98) at /usr/devel/git/rock/sys/net/route/nhgrp_ctl.c:380
> #14 0xffff000000405434 in epoch_call_task (arg=<optimized out>) at /usr/devel/git/rock/sys/kern/subr_epoch.c:819
> #15 0xffff000000408ee4 in gtaskqueue_run_locked (queue=queue_at_entry=0xffffa00000c03c00) at /usr/devel/git/rock/sys/kern/subr_gtaskqueue.c:371
> #16 0xffff000000408c38 in gtaskqueue_thread_loop (arg=arg_at_entry=0xffff000089b69008) at /usr/devel/git/rock/sys/kern/subr_gtaskqueue.c:547
> #17 0xffff00000037701c in fork_exit (callout=0xffff000000408b6c <gtaskqueue_thread_loop>, arg=0xffff000089b69008, frame=0xffff000087934990) at /usr/devel/git/rock/sys/kern/kern_fork.c:1087
> 
> -- 
> Andriy Gapon
Received on Mon Aug 30 2021 - 08:22:03 UTC

Original text of this message