Summary: | [CI] igt@drv_selftest@mock_scatterlist - dmesg-warn - sg_alloc_table timed out (leading to an incomplete) | ||
---|---|---|---|
Product: | DRI | Reporter: | Martin Peres <martin.peres> |
Component: | DRM/Intel | Assignee: | Intel GFX Bugs mailing list <intel-gfx-bugs> |
Status: | CLOSED FIXED | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> |
Severity: | normal | ||
Priority: | medium | CC: | intel-gfx-bugs, tomi.p.sarvela |
Version: | XOrg git | ||
Hardware: | Other | ||
OS: | All | ||
Whiteboard: | ReadyForDev | ||
i915 platform: | GLK | i915 features: | GEM/Other |
Description
Martin Peres
2018-07-02 23:06:50 UTC
This is only seen on shard-glk9. Adding Tomi so as we can know what is different about this machine. shard-glk9 is production hardware and has same BIOS version and same amount of RAM as any other shard-glks. Any ideas what this could be about? What's unusual here? mock_scatterlist is CI's false positive as it erroneously thinks a pr_note is a warning, and the live_gtt is just CI declaring the system hung before the test is able to complete (too low a threshold on the hung task watchdog -- don't use the hung task watchdog, hard lockup definitely, soft lockup is debatable so needs a relaxed timeout, hung task has too many false positives). I do think this was conflating two known bugs. As we've taken measures to reduce the amount of memory used for live_gtt, let's sweep this under the carpet. It indeed got 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.