https://bugzilla.kernel.org/show_bug.cgi?id=114591
--- Comment #31 from Eric Sandeen <sandeen@xxxxxxxxxx> ---
Michael, thanks for the update. Knowing that it's the 3rd call definitely
narrows down the issue. In this case we have searched for a cached buffer,
failed to find one, allocated a new one, looked up the same range again, this
time we found a cached buffer which appeared in the interim, freed the one we
just allocated, and hit your BUG_ON when we are freeing its associated pages.
This seems a bit odd, because on this path nothing should have happened with
the pages in question; they were simply allocated and almost immediately freed,
as far as I can tell. I'll think on this some more.
(It's not my intention to be hostile to Nick, only to save you the wasted time
that many others have suffered as a result of Nick's unwillingness to even
compile-test patches before he proposes them for subsystems which he does not
understand. If there is more than a whiff of annoyance, it is only because the
behavior has gone on for a very long time, despite reasonable politeness and
attempts at mentoring in the beginning. I won't spend more time on this, but a
bit of googling will turn up the back story).
--
You are receiving this mail because:
You are the assignee for the bug.
|