aboutsummaryrefslogtreecommitdiffstats
path: root/mm
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2015-03-25 15:40:21 -0700
committerLinus Torvalds <torvalds@linux-foundation.org>2015-03-25 15:40:21 -0700
commitb8517e98305e3c76fa293133826afe39a690edcd (patch)
tree55ce7797d169cd4f015b02a19c8762a72152951e /mm
parentMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux (diff)
parentwriteback: fix possible underflow in write bandwidth calculation (diff)
downloadlinux-dev-b8517e98305e3c76fa293133826afe39a690edcd.tar.xz
linux-dev-b8517e98305e3c76fa293133826afe39a690edcd.zip
Merge branch 'for-linus' of git://git.kernel.dk/linux-block
Pull block layer fixes from Jens Axboe: "A small collection of fixes that has been gathered over the last few weeks. This contains: - A one-liner fix for NVMe, fixing a missing list_head init that could makes us oops on hitting recovery at load time. - Two small blk-mq fixes: - Fixup a bad goto jump on error handling. - Fix for oopsing if running out of reserved tags. - A memory leak fix for NBD. - Two small writeback fixes from Tejun, fixing a missing init to INITIAL_JIFFIES, and a possible underflow introduced recently. - A core merge fixup in sg gap detection, where rq->biotail was indexed with the count of rq->bio" * 'for-linus' of git://git.kernel.dk/linux-block: writeback: fix possible underflow in write bandwidth calculation NVMe: Initialize device list head before starting Fix bug in blk_rq_merge_ok blkmq: Fix NULL pointer deref when all reserved tags in blk-mq: fix use of incorrect goto label in blk_mq_init_queue error path nbd: fix possible memory leak writeback: add missing INITIAL_JIFFIES init in global_update_bandwidth()
Diffstat (limited to 'mm')
-rw-r--r--mm/page-writeback.c7
1 files changed, 5 insertions, 2 deletions
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index 45e187b2d971..644bcb665773 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -857,8 +857,11 @@ static void bdi_update_write_bandwidth(struct backing_dev_info *bdi,
* bw * elapsed + write_bandwidth * (period - elapsed)
* write_bandwidth = ---------------------------------------------------
* period
+ *
+ * @written may have decreased due to account_page_redirty().
+ * Avoid underflowing @bw calculation.
*/
- bw = written - bdi->written_stamp;
+ bw = written - min(written, bdi->written_stamp);
bw *= HZ;
if (unlikely(elapsed > period)) {
do_div(bw, elapsed);
@@ -922,7 +925,7 @@ static void global_update_bandwidth(unsigned long thresh,
unsigned long now)
{
static DEFINE_SPINLOCK(dirty_lock);
- static unsigned long update_time;
+ static unsigned long update_time = INITIAL_JIFFIES;
/*
* check locklessly first to optimize away locking for the most time