PT-2026-152608
CVE-2026-99728
·
Published
2026-11-20
·
Updated
2026-11-20
Description by
Score
N/A
Not availableNo severity ratings or metrics are available. When they are, we'll update the corresponding info on the page.
In the Linux kernel, the following vulnerability has been resolved:
mm/huge memory: fix list corruption when deferred split scan() races with deferred split folio()
Since the deferred split queue became a list lru, the isolate callback
moves the folios it took a reference on to an on-stack list, linked
through folio-> deferred list, and deferred split scan() takes them off
that list with list del init() without holding the list lru lock.
deferred split folio() does not know about the on-stack list. It only
holds the list lru lock when it looks at the same list head, and it can
run for such a folio: a THP is queued when it is mapped, as a candidate
for the underused check, and is queued "again" when a part of it gets
unmapped:
CPU0 deferred split scan() CPU1 deferred split folio()
list del init(&folio-> deferred list)
next->prev = prev
prev->next = next
entry->next = entry
list lru lock irqsave()
folio set partially mapped()
list lru add()
list empty(item) -> true
list add tail(item, &l->list)
item->next = &l->list
item->prev = tail
entry->prev = entry
The folio now sits on the list lru with ->prev pointing to itself. The
next removal leaves the neighbours pointing to it, and they still do so
after the folio has been freed. With CONFIG DEBUG LIST:
list del corruption. prev->next should be ffd4000000270090, but was
ffd4000000278090. (prev=ffd4000000270090)
Fix it by grabbing the lock around list del init() so
deferred split folio() cannot race with it.
Found an issue in the description? Have something to add? Feel free to write us 👾