Ben Hutchings <b...@decadent.org.uk> writes:
> On Mon, 2018-12-03 at 23:08 +0100, TS wrote:
>
>> Just out of curiosity since linux-image-amd64 4.18+100 has been uploaded to
>> unstable. This issue here seems not to be wide spread, or reproducible?
>
> Hard to tell. There are a few people that reported the same *symptom*,
> but that doesn't mean they found the same bug.
True of course. But in this case there seems to be a very strong
correlation with Gen4 Intel GPUs. I believe every one of the reports
have had at least a lspci dump showing one of those. And a couple of
the reports had stack traces pointing to gen4_render_ring_flush as well.
Given this, there is an extremely suspicious commit added in v4.18.20:
06e562e7f515 ("drm/i915/ringbuffer: Delay after EMIT_INVALIDATE for gen4/gen5")
I do have an old laptop with an affected chipset generation, and
verified that it had the same symptoms. But never got the time to
actually test any further. Still, I do think that there is good reason
to simply try a revert of that commit.
Bjørn
