The patch titled
Subject: writeback: avoid division by 0 in wb_update_dirty_ratelimit()
has been removed from the -mm tree. Its filename was
writeback-fix-bandwidth-estimate-for-spiky-workload-fix.patch
This patch was dropped because it was folded into writeback-fix-bandwidth-estimate-for-spiky-workload.patch
------------------------------------------------------
From: Jan Kara <jack@xxxxxxx>
Subject: writeback: avoid division by 0 in wb_update_dirty_ratelimit()
Fixup patch "writeback: Fix bandwidth estimate for spiky workload" which
introduced possibility of __wb_update_bandwidth() getting called at a
moment when 'elapsed' evaluates to 0.
Cc: Jan Kara <jack@xxxxxxx>
Cc: Michael Stapelberg <stapelberg+linux@xxxxxxxxxx>
Cc: Wu Fengguang <fengguang.wu@xxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/page-writeback.c | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
--- a/mm/page-writeback.c~writeback-fix-bandwidth-estimate-for-spiky-workload-fix
+++ a/mm/page-writeback.c
@@ -1336,11 +1336,19 @@ static void __wb_update_bandwidth(struct
{
struct bdi_writeback *wb = gdtc->wb;
unsigned long now = jiffies;
- unsigned long elapsed = now - wb->bw_time_stamp;
+ unsigned long elapsed;
unsigned long dirtied;
unsigned long written;
spin_lock(&wb->list_lock);
+
+ /*
+ * Lockless checks for elapsed time are racy and delayed update after
+ * IO completion doesn't do it at all (to make sure written pages are
+ * accounted reasonably quickly). Make sure elapsed >= 1 to avoid
+ * division errors.
+ */
+ elapsed = max(now - wb->bw_time_stamp, 1UL);
dirtied = percpu_counter_read(&wb->stat[WB_DIRTIED]);
written = percpu_counter_read(&wb->stat[WB_WRITTEN]);
_
Patches currently in -mm which might be from jack@xxxxxxx are
writeback-track-number-of-inodes-under-writeback.patch
writeback-reliably-update-bandwidth-estimation.patch
writeback-fix-bandwidth-estimate-for-spiky-workload.patch
writeback-rename-domain_update_bandwidth.patch
writeback-use-read_once-for-unlocked-reads-of-writeback-stats.patch
