Re: sendbug retry - also kernel: protection fault trap

"not into names" <[email protected]>
Newsgroups gmane.os.openbsd.bugs
Message-ID <[email protected]>
On Sat Aug 15, 2026 at 4:55 AM CEST, Jonathan Gray wrote:
> On Fri, Aug 14, 2026 at 02:57:11PM +0200, not into names wrote:
>> hi. 
>> 
>> i block everything everywhere all the time all at once. sendbug could
>> have a retry so when it can't connect out i can allow the outgoing ips
>> and try again, so i don't have to type it a second time, which i won't.
>> so here short version:
>> 
>>     login: kernel: protection fault trap, code=0
>>     Stopped at      dma_fence_chain_walk+0xba:          movq        0x18(%rax),%r11
>>     dbb{0}>
>> 
>> had to press reset. i guess "login:" was still there from boot. i
>> was watching some webm i downloaded from the youtubes. nothing special,
>> do that all the time. was over sshfs though. also doing that all the
>> time.
>
> It is unclear which fence this is without a full trace.
>
> If you can reproduce it, does it still occur with this?
>
> Index: sys/dev/pci/drm/drm_linux.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/pci/drm/drm_linux.c,v
> diff -u -p -r1.144 drm_linux.c
> --- sys/dev/pci/drm/drm_linux.c	27 Jul 2026 02:24:44 -0000	1.144
> +++ sys/dev/pci/drm/drm_linux.c	15 Aug 2026 02:49:18 -0000
> @@ -1889,6 +1889,9 @@ dma_fence_is_signaled(struct dma_fence *
>  	if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags))
>  		return true;
>  
> +	if (fence->ops == NULL)
> +		return false;
> +
>  	if (fence->ops->signaled && fence->ops->signaled(fence)) {
>  		dma_fence_signal(fence);
>  		return true;

my apologies. i just learned about the trace stuff after i sent the bug
report. also i noticed some smtp thing was still sending after sendbug
exited, so the retry feature is already there it's just in the
background.

i can't reproduce it. as i said i do the same thing all the time. works
just fine. 

i've been using troonix for a while, but that's literally all i do. i
just use it. so i know what you gave me is some thing i use with patch,
right? but that's all i know. should it happen again i'll do the trace
stuff. baby steps.

thanks for your time.


oh, should i send the wsconsctl screen.display_off crash i mentioned?
that one is easyly reproducible. i'd do the trace and the panic. would be
a good exercise. though you'd get a picture then. i have very low
conscientiousness, if i type it all it wouldn't be accurate.
signature.asc (application/pgp-signature, 265 B)
-----BEGIN PGP SIGNATURE-----

iJEEABYKADkWIQTiKe7AVvrO9YysHhY/3YtseOrfiQUCaoAGyBsUgAAAAAAEAA5t
YW51MiwyLjUrMS4xMiwyLDMACgkQP92LbHjq34me8gD+JhqzyBBRKlg5i7/E/4bU
Q9ZhtBPTC+qDdt84lzjrW64A+wdPtPtym5O5+kgvGMc985+lhMayZrgjP0Bnjl7u
N08A
=vX0g
-----END PGP SIGNATURE-----
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.