I'm now testing my one-line patch from comment #5 on top of 3.2.0-67.101
(amd64, generic kernel flavour). So far it doesn't seem to make things
worse, but since I don't have a sure-fire way of triggering the bug it
may take a while to get experimental confirmation that it cures the issue.
(I'm reasonably confident about it based on my reading of the source code,
however. The various set_acl methods in 3.14 seem to be doing the
same thing as that patch.)
I'm now testing my one-line patch from comment #5 on top of 3.2.0-67.101
(amd64, generic kernel flavour). So far it doesn't seem to make things
worse, but since I don't have a sure-fire way of triggering the bug it
may take a while to get experimental confirmation that it cures the issue.
(I'm reasonably confident about it based on my reading of the source code,
however. The various set_acl methods in 3.14 seem to be doing the
same thing as that patch.)