https://bugzilla.kernel.org/show_bug.cgi?id=196223 --- Comment #14 from Theodore Tso (tytso@xxxxxxx) --- Also for the record, there is also no difference in any ext4/jbd2 code between next-20170626, next-20170627, next-20170628, and next-20170630. So if next-20170726 is working for you, and next-201706028 is not, it very well be the case that the fact that we are blocked waiting for I/O in ext4 code is a symptom, not the cause of the problem, and if ext4 is the cause of the problem, it must be indirectly. So I'm not sure there's anything more I can do, especially since I don't have a reliable repro that I can run at this point. Something else that might be worth trying. What's in linux-next is slowing getting picked up by Linus. You might see if the tip of Linus's tree is passing your tests or not. If not, it will be much easier to do a true git bisection, unlike with linux-next since many patches get rebased. -- You are receiving this mail because: You are watching the assignee of the bug.
