Commit f8b5ff2c authored May 17, 2013 by Christian Borntraeger Committed by Gleb Natapov May 21, 2013

s390: fix gmap_ipte_notifier vs. software dirty pages

On heavy paging load some guest cpus started to loop in gmap_ipte_notify.
This was visible as stalled cpus inside the guest. The gmap_ipte_notifier
tries to map a user page and then made sure that the pte is valid and
writable. Turns out that with the software change bit tracking the pte
can become read-only (and only software writable) if the page is clean.
Since we loop in this code, the page would stay clean and, therefore,
be never writable again.
Let us just use fixup_user_fault, that guarantees to call handle_mm_fault.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>

parent 7c470539

arch/s390/mm/pgtable.c

+1 −2

Original line number	Original line	Diff line number	Diff line
	@@ -677,8 +677,7 @@ int gmap_ipte_notify(struct gmap *gmap, unsigned long start, unsigned long len)
	break;		break;
	}		}
	/* Get the page mapped */		/* Get the page mapped */
	if (get_user_pages(current, gmap->mm, addr, 1, 1, 0,		if (fixup_user_fault(current, gmap->mm, addr, FAULT_FLAG_WRITE)) {
	NULL, NULL) != 1) {
	rc = -EFAULT;		rc = -EFAULT;
	break;		break;
	}		}

Admin message