CVE-2026-89755

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> mm/migrate_device: clear stale mapping after freeing swapcache<br /> <br /> __migrate_device_pages() reads the folio mapping before calling<br /> folio_free_swap(). When folio_free_swap() succeeds, the folio is removed<br /> from the swap cache, but the saved mapping still points to swap_space.<br /> <br /> Passing the stale mapping to folio_migrate_mapping() makes it use the<br /> mapped-folio path for a folio that is no longer in swapcache. It can then<br /> operate on swap_space.i_pages with invalid reference accounting,<br /> eventually triggering a folio reference count BUG.<br /> <br /> After a successful split, nr still contains the number of pages in the<br /> original large folio, although each resulting page is now a separate<br /> order-0 folio. Reset nr to 1 so each split folio is processed separately,<br /> including its own swapcache removal and mapping lookup.<br /> <br /> Refresh the saved mapping after folio_free_swap() so the current folio<br /> state is used during migration.