https://bugzilla.kernel.org/show_bug.cgi?id=196223 --- Comment #18 from Tomi Sarvela (tomi.p.sarvela@xxxxxxxxx) --- I managed to reproduce the bug on next-20170626 and earlier by running igt@gem_exec_suspend@basic-s3 test back-to-back, takes ~5 runs. Later next-kernels reproduce this issue on 1-2 runs. ./scripts/run-tests.sh -vt igt@gem_exec_suspend@basic-s3 -x devices Issue is seen when igt test returns 'pass', but doesn't drop back to command line. Repeated Ctrl-C gets to the command prompt. Cached commands work, but anything that needs disk access doesn't. linus-v4.12 handled ~10 runs without a problem, then same issue. Dmesg shows in usual case something like: [ 137.351829] nvme nvme0: I/O 197 QID 0 timeout, disable controller [ 137.464115] nvme nvme0: failed to set APST feature (-4) [ 137.861403] EXT4-fs warning (device nvme0n1p2): ext4_end_bio:322: I/O error 10 writing to inode 1453376 (offset 0 size 4096 starting block 5949696) [ 137.861406] Buffer I/O error on device nvme0n1p2, logical block 5818368 I'm concerned about the 'disable controller' part. Setup on this testing: KBL-7700K on MSI Z170 mb. Ubuntu base, Intel 600p M.2, one DP connected to monitor. -- You are receiving this mail because: You are watching the assignee of the bug.
