On Fri, Jul 17, 2020 at 03:12:26PM -0400, Josef Bacik wrote: > We've had a longstanding lockdep splat in open_fs_devices that we've been > ignoring for some time, since it's not actually a problem. However this has > been masking all the other lockdep issues that we have, since lockdep takes its > ball and goes home as soon as you hit one problem. That lockdep does not report anything past the first warning is making it hard to test, I got an unrelated splat with btrfs/078 so can't actually verify that the fstests run is clean with your patchset.
