https://bugzilla.kernel.org/show_bug.cgi?id=217965 --- Comment #60 from Matthew Stapleton (matthew4196@xxxxxxxxx) --- It looks like echo "0" > /sys/fs/ext4/<dev>/mb_best_avail_max_trim_order stops flush from getting stuck, but the flush kworker still uses more cpu during the write operation. Hopefully I can get the perf results over the next few days. -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
