[ 030/150] zram: Fix deadlock bug in partial read/write

From: Greg Kroah-Hartman
Date: Tue Feb 26 2013 - 20:15:19 EST

3.8-stable review patch. If anyone has any objections, please let me know.


From: Minchan Kim <minchan@xxxxxxxxxx>

commit 7e5a5104c6af709a8d97d5f4711e7c917761d464 upstream.

Now zram allocates new page with GFP_KERNEL in zram I/O path
if IO is partial. Unfortunately, It may cause deadlock with
reclaim path like below.

write_page from fs
write_page from fs
fs_lock <-- deadlock

This patch fixes it by using GFP_NOIO. In read path, we
reorganize code flow so that kmap_atomic is called after the
GFP_NOIO allocation.

Acked-by: Jerome Marchand <jmarchand@xxxxxxxxxx>
Acked-by: Nitin Gupta <ngupta@xxxxxxxxxx>
[ penberg@xxxxxxxxxx: don't use GFP_ATOMIC ]
Signed-off-by: Pekka Enberg <penberg@xxxxxxxxxx>
Signed-off-by: Minchan Kim <minchan@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

drivers/staging/zram/zram_drv.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)

--- a/drivers/staging/zram/zram_drv.c
+++ b/drivers/staging/zram/zram_drv.c
@@ -228,11 +228,12 @@ static int zram_bvec_read(struct zram *z
return 0;

- user_mem = kmap_atomic(page);
if (is_partial_io(bvec))
/* Use a temporary buffer to decompress the page */
- uncmem = kmalloc(PAGE_SIZE, GFP_KERNEL);
- else
+ uncmem = kmalloc(PAGE_SIZE, GFP_NOIO);
+ user_mem = kmap_atomic(page);
+ if (!is_partial_io(bvec))
uncmem = user_mem;

if (!uncmem) {
@@ -279,7 +280,7 @@ static int zram_bvec_write(struct zram *
* This is a partial IO. We need to read the full page
* before to write the changes.
- uncmem = kmalloc(PAGE_SIZE, GFP_KERNEL);
+ uncmem = kmalloc(PAGE_SIZE, GFP_NOIO);
if (!uncmem) {
pr_info("Error allocating temp memory!\n");
ret = -ENOMEM;

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/