<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv="Content-Type" content="text/html; charset=us-ascii"> <meta name="Generator" content="Microsoft Word 15 (filtered medium)"> <style></style> </head> <body lang="FI" link="#0563C1" vlink="#954F72" style="word-wrap:break-word"> <div class="WordSection1"> More information (excuse my top-posting):<o:p></o:p> <o:p> </o:p> - Issue happens in igt@gem_tiled_swapping@non-threaded Mlocking<o:p></o:p> phase, before “starting subtest” appears.<o:p></o:p> <o:p> </o:p> - Filesystem trashed is the one containing swapfile<o:p></o:p> <o:p> </o:p> - If swap is partition, it seems that the swap signature is correct even<o:p></o:p> after running the test, so for now I’m assuming that the issue has to do<o:p></o:p> with swapfile<o:p></o:p> <o:p> </o:p> - Bisection between 20210129 and 20210215 proved to be challenging,<o:p></o:p> because the kernels have pre-init hang, don’t leave dmesg and I don’t<o:p></o:p> have console on testing host. Petri’s suggestion to bisect between<o:p></o:p> CI_DRM_9817 and 9818 might work better<o:p></o:p> <o:p> </o:p> Regards,<o:p></o:p> <o:p> </o:p> Tomi Sarvela<o:p></o:p> <o:p> </o:p> <div style="border:none;border-left:solid blue 1.5pt;padding:0cm 0cm 0cm 4.0pt"> <div> <div style="border:none;border-top:solid #E1E1E1 1.0pt;padding:3.0pt 0cm 0cm 0cm"> From: Sarvela, Tomi P Sent: Tuesday, March 2, 2021 1:38 PM To: intel-gfx@lists.freedesktop.org Cc: Szwichtenberg, Radoslaw <radoslaw.szwichtenberg@intel.com> Subject: Public i915 CI shardruns are disabled<o:p></o:p> </div> </div> <o:p> </o:p> Hello,<o:p></o:p> <o:p> </o:p> The linux i915 CI shardruns have been disabled. This is due to the unfortunate<o:p></o:p> filesystem-corrupting bug first seen in linux-next 20210215, which now has<o:p></o:p> been merged to linus 5.12-rc1 and further on to DRM-Tip, first instance seen<o:p></o:p> in CI_DRM_9818. Last changes coming in were:<o:p></o:p> <o:p> </o:p> fb3b93df7979 drm-tip: 2021y-03m-01d-09h-36m-57s UTC integration manifest<o:p></o:p> 3b3c4086295b drm-tip: 2021y-03m-01d-08h-49m-06s UTC integration manifest<o:p></o:p> fe07bfda2fb9 Linux 5.12-rc1<o:p></o:p> <o:p> </o:p> More information can be seen at:<o:p></o:p> <a href="https://phoronix.com/scan.php?page=news_item&px=Linux-5.12-Early-Buggy-Issue">https://phoronix.com/scan.php?page=news_item&px=Linux-5.12-Early-Buggy-Issue</a><o:p></o:p> <o:p> </o:p> I’ve seen this bug happen regularly with (but not limited to) IGT test:<o:p></o:p> igt@gem_tiled_swapping@non-threaded<o:p></o:p> <o:p> </o:p> The range for bisection is linux-next 20210215 to 20210129 because the kernels <o:p></o:p> in-between taint the kernel and our i915 testing was not done. Hitting the bug <o:p></o:p> corrupts the underlying filesystem very thoroughly, wiping out large amount of <o:p></o:p> data from the beginning of the partition which leaves fsck sad with thousands of<o:p></o:p> items lost. Bisection of the IGT testlist was done with two root filesystems, where<o:p></o:p> testable kernel booted from 2. partition, and copy of the 2. partition was stored<o:p></o:p> on 1. partition and could be restored at will.<o:p></o:p> <o:p> </o:p> I’ll continue bisecting this bug on the linux-next tree again. If someone has more<o:p></o:p> information where this issue originates from, help would be appreciated.<o:p></o:p> <o:p> </o:p> Regards,<o:p></o:p> <o:p> </o:p> Tomi Sarvela<o:p></o:p> <o:p> </o:p> --<o:p></o:p> Intel Finland Oy - BIC 0357606-4 - Westendinkatu 7, 02160 Espoo<o:p></o:p> <o:p> </o:p> </div> </div> </body> </html>