https://bugzilla.kernel.org/show_bug.cgi?id=217965 Matthew Stapleton (matthew4196@xxxxxxxxx) changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |matthew4196@xxxxxxxxx --- Comment #52 from Matthew Stapleton (matthew4196@xxxxxxxxx) --- I'm a bit late to reporting this, but it might be related to this bug report. Since noticing this 100% cpu flush problem on my computer with a non-raid Samsung SSD 990 PRO NMVe and kernel 6.6.3 (and now 6.6.8) I decided to do a git bisect and seemed to track down the problem to commit: 7e170922f06bf46effa7c57f6035fc463d6edc7e if that helps, although I haven't tested if disabling CR1_5 also resolves the problem. Running "tar xf linux-6.6.8.tar.xz" is enough to trigger 100% cpu for a few seconds after tar is finished on my system, although it doesn't noticeably trigger the problem on my server with raid 6 and hard drives. It also looks like my ext4 filesystem is mounting with stripe=32733 according to /proc/mounts, although dumpe2fs -h doesn't show any stripe settings on the filesystem. I'll try the patch and see if it fixes the problem. -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
