Re: e179d973 insta-panics in nl_send_one()

From: Jakob Alvermark <jakob_at_alvermark.net>
Date: Mon, 08 Jan 2024 09:40:52 UTC
On 1/6/24 23:34, Poul-Henning Kamp wrote:
> Addendum:
>
>     I have only installed the new kernel, userland is still from dec18.
>
> (Even if that is the cause, we should not panic on bad syscall args.)
>
> Poul-Henning Kamp writes:
>> With fresh current:
>>
>> 	commit e179d9739b1438ae9acb958f80a983eff7e3dce9
>> 	Author: Michael Tuexen <tuexen@FreeBSD.org>
>> 	Date:   Sat Jan 6 21:31:46 2024 +0100
>>
>> 	    tcpsso: support TIME_WAIT state
>>      
>> I get an insta-panic as soon as any network interface comes up:
>>
>> 	--- trap 0xc, rip = 0xffff...f80d97b78, rsp = 0x...
>> 	nl_send_one() at nl_send_one+0x18/frame 0xf....
>> 	nl_send_group() at nl_send_group+0x1bc/frame 0xf...
>> 	_nlmsg_flush() at _nlmsg_flush+0x37/frame 0xf...
>> 	rtnl_handle_ifevent() + 0xa1
>> 	if_attach_internal + 0x3df
>>
>> I have a picture of the full panic if desired...
>>
>> -- 
>> Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
>> phk@FreeBSD.ORG         | TCP/IP since RFC 956
>> FreeBSD committer       | BSD since 4.3-tahoe
>> Never attribute to malice what can adequately be explained by incompetence.


I get the same panic, with kernel and userland both installed.


Jakob