Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1515842 > unrolled thread
| Started by | Hugh Dickins <hughd@google.com> |
|---|---|
| First post | 2016-11-07 05:10 +0100 |
| Last post | 2016-11-07 12:40 +0100 |
| Articles | 2 — 2 participants |
Back to article view | Back to linux.kernel
[PATCH] shmem: fix pageflags after swapping DMA32 object Hugh Dickins <hughd@google.com> - 2016-11-07 05:10 +0100
Re: [PATCH] shmem: fix pageflags after swapping DMA32 object "Kirill A. Shutemov" <kirill@shutemov.name> - 2016-11-07 12:40 +0100
| From | Hugh Dickins <hughd@google.com> |
|---|---|
| Date | 2016-11-07 05:10 +0100 |
| Subject | [PATCH] shmem: fix pageflags after swapping DMA32 object |
| Message-ID | <sAJpT-6m8-7@gated-at.bofh.it> |
If shmem_alloc_page() does not set PageLocked and PageSwapBacked, then
shmem_replace_page() needs to do so for itself. Without this, it puts
newpage on the wrong lru, re-unlocks the unlocked newpage, and system
descends into "Bad page" reports and freeze; or if CONFIG_DEBUG_VM=y,
it hits an earlier VM_BUG_ON_PAGE(!PageLocked), depending on config.
But shmem_replace_page() is not a common path: it's only called when
swapin (or swapoff) finds the page was already read into an unsuitable
zone: usually all zones are suitable, but gem objects for a few drm
devices (gma500, omapdrm, crestline, broadwater) require zone DMA32
if there's more than 4GB of ram.
Fixes: 800d8c63b2e9 ("shmem: add huge pages support")
Cc: stable@vger.kernel.org # v4.8
Signed-off-by: Hugh Dickins <hughd@google.com>
---
mm/shmem.c | 2 ++
1 file changed, 2 insertions(+)
--- 4.9-rc4/mm/shmem.c 2016-10-15 12:52:13.157533478 -0700
+++ linux/mm/shmem.c 2016-11-06 12:45:49.626193769 -0800
@@ -1483,6 +1483,8 @@ static int shmem_replace_page(struct pag
copy_highpage(newpage, oldpage);
flush_dcache_page(newpage);
+ __SetPageLocked(newpage);
+ __SetPageSwapBacked(newpage);
SetPageUptodate(newpage);
set_page_private(newpage, swap_index);
SetPageSwapCache(newpage);
[toc] | [next] | [standalone]
| From | "Kirill A. Shutemov" <kirill@shutemov.name> |
|---|---|
| Date | 2016-11-07 12:40 +0100 |
| Message-ID | <sAQro-2rT-7@gated-at.bofh.it> |
| In reply to | #1515842 |
On Sun, Nov 06, 2016 at 08:08:29PM -0800, Hugh Dickins wrote:
> If shmem_alloc_page() does not set PageLocked and PageSwapBacked, then
> shmem_replace_page() needs to do so for itself. Without this, it puts
> newpage on the wrong lru, re-unlocks the unlocked newpage, and system
> descends into "Bad page" reports and freeze; or if CONFIG_DEBUG_VM=y,
> it hits an earlier VM_BUG_ON_PAGE(!PageLocked), depending on config.
>
> But shmem_replace_page() is not a common path: it's only called when
> swapin (or swapoff) finds the page was already read into an unsuitable
> zone: usually all zones are suitable, but gem objects for a few drm
> devices (gma500, omapdrm, crestline, broadwater) require zone DMA32
> if there's more than 4GB of ram.
>
> Fixes: 800d8c63b2e9 ("shmem: add huge pages support")
> Cc: stable@vger.kernel.org # v4.8
> Signed-off-by: Hugh Dickins <hughd@google.com>
Sorry for that.
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
--
Kirill A. Shutemov
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web