Summary: | [CI] igt@kms_cursor_legacy@2x-(long-)nonblocking-modeset-vs-cursor-atomic - incomplete | ||
---|---|---|---|
Product: | DRI | Reporter: | Martin Peres <martin.peres> |
Component: | DRM/Intel | Assignee: | Stanislav Lisovskiy <stanislav.lisovskiy> |
Status: | CLOSED WORKSFORME | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> |
Severity: | normal | ||
Priority: | high | CC: | intel-gfx-bugs |
Version: | XOrg git | ||
Hardware: | Other | ||
OS: | All | ||
Whiteboard: | ReadyForDev | ||
i915 platform: | SKL | i915 features: | display/Other |
Description
Martin Peres
2018-04-17 10:41:06 UTC
This has been reproduced on all subsequent runs. I will take a look at this, if nobody objects. Still not able to reproduce the oops issue, despite using exactly same test sequence(extracted from shards x0027, x0030), skl-6770hw hardware and even same kernel version. Those tests seem to fail more frequently with vblank_matches_assertion failure though. especially if drm.debug all bits are set(probably affects timings). However no similar crash happens. So still need info on that one. eu-addr2line on the backtrace points to encoder being NULL in check_digital_port_conflicts. However that should only be called when a modeset is requested. Still lets figure out what happens.. https://patchwork.freedesktop.org/series/43132/ (In reply to Maarten Lankhorst from comment #4) > eu-addr2line on the backtrace points to encoder being NULL in > check_digital_port_conflicts. > > However that should only be called when a modeset is requested. Still lets > figure out what happens.. > > https://patchwork.freedesktop.org/series/43132/ There are ioctls sent to change cursor position, simultaneously with disabling/enabling pipe2 in parallel. To me it looks like some concurrency issue, got same result with objdump. However the question is then why it is not reproducible. Not seen on CI lately. (In reply to Jani Saarinen from comment #6) > Not seen on CI lately. Agreed, it was very consistent, but not anymore (past ~12 runs). I guess we can call this fixed! |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.