On 22.03.24 11:19, Vlastimil Babka wrote:
On 3/21/24 15:24, Matthew Wilcox (Oracle) wrote:
The current folio_test_hugetlb() can be fooled by a concurrent folio split
into returning true for a folio which has never belonged to hugetlbfs.
This can't happen if the caller holds a refcount on it, but we have a
few places (memory-failure, compaction, procfs) which do not and should
not take a speculative reference.
Should we add metadata wrt closing the bug report from Luis?
https://lore.kernel.org/all/8fa1c95c-4749-33dd-42ba-243e492ab109@xxxxxxx/
I assume this wouldn't be fun wrt stable...
Since hugetlb pages do not use individual page mapcounts (they are always
fully mapped and use the entire_mapcount field to record the number
Wasn't there some discussions to allow partial mappings of hugetlb? What
would be the implications?
If we ever go that path, we really should avoid messing with any
subpages right from the start. We should make it work using a single
total mapcount per folio.
Anyhow, that's stuff for the future.
--
Cheers,
David / dhildenb
