Commit a00196a2 authored by Nicholas Piggin's avatar Nicholas Piggin Committed by Michael Ellerman

powerpc/64s: __find_linux_pte() synchronization vs pmdp_invalidate()

The change to pmdp_invalidate() to mark the pmd with _PAGE_INVALID
broke the synchronisation against lock free lookups,
__find_linux_pte()'s pmd_none() check no longer returns true for such
cases.

Fix this by adding a check for this condition as well.

Fixes: da7ad366 ("powerpc/mm/book3s: Update pmd_present to look at _PAGE_PRESENT bit")
Cc: stable@vger.kernel.org # v4.20+
Suggested-by: default avatarAneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: default avatarNicholas Piggin <npiggin@gmail.com>
Reviewed-by: default avatarAneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: default avatarMichael Ellerman <mpe@ellerman.id.au>
parent 33258a1d
...@@ -372,13 +372,25 @@ pte_t *__find_linux_pte(pgd_t *pgdir, unsigned long ea, ...@@ -372,13 +372,25 @@ pte_t *__find_linux_pte(pgd_t *pgdir, unsigned long ea,
pdshift = PMD_SHIFT; pdshift = PMD_SHIFT;
pmdp = pmd_offset(&pud, ea); pmdp = pmd_offset(&pud, ea);
pmd = READ_ONCE(*pmdp); pmd = READ_ONCE(*pmdp);
/* /*
* A hugepage collapse is captured by pmd_none, because * A hugepage collapse is captured by this condition, see
* it mark the pmd none and do a hpte invalidate. * pmdp_collapse_flush.
*/ */
if (pmd_none(pmd)) if (pmd_none(pmd))
return NULL; return NULL;
#ifdef CONFIG_PPC_BOOK3S_64
/*
* A hugepage split is captured by this condition, see
* pmdp_invalidate.
*
* Huge page modification can be caught here too.
*/
if (pmd_is_serializing(pmd))
return NULL;
#endif
if (pmd_trans_huge(pmd) || pmd_devmap(pmd)) { if (pmd_trans_huge(pmd) || pmd_devmap(pmd)) {
if (is_thp) if (is_thp)
*is_thp = true; *is_thp = true;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment