On Fri, 8 Jul 2022 00:21:45 +0200 Vlastimil Babka <vbabka@xxxxxxx> wrote: > >> > add missing local_unlock_irqrestore() on contention path > >> > >> Doh, that's true and something to fix, although patch 7 did remove the bug > >> later in the same series so that wouldn't explain the lkp report for patch > >> 7. The reason lkp test robot complained was AFAICS that it was testing v4, > >> as I just replied there. > > > > Sorry I didn't bother to reply until now: it did test v5, at this > > commit, not the whole series. > > I meant this report that appears to be for v4 (full series including patch 7): > https://lore.kernel.org/all/YsFk%2FqU+QtWun04h@xsang-OptiPlex-9020/ > That reported a bug due to missing unpin that was previously reported for v4 > and fixed in v5. > > I'm not aware of a lkp report for v5 (found only Dan's) but yeah, hitting > the (similar but not identical) bug fixed by this -fix would indeed be > possible in v5 if patch 7 was not applied. Patch 7 ("mm/page_alloc: replace local_lock with normal spinlock") has been dropped for now due to ongoing issues, most recently "BUG:sleeping_function_called_from_invalid_context_at_mm/gup.c". This droppage exposed the issue which this patch (mm-page_alloc-protect-pcp-lists-with-a-spinlock-fix.patch) addresses.
