Skip to contentSkip to conversation

f2fs: show more info if fail to issue discard

PATCH New Chao Yu 1 participant 3 messages May 19, 2017

Subsystems (2)
Stable branch F2FS file system
lore
Share & tools
lore

Chao Yu

author maintainer
#0 lore raw

PATCH 1/3 f2fs: show more info if fail to issue discard

From: Chao Yu <[email protected]>

Signed-off-by Chao Yu <[email protected]>
---
 fs/f2fs/segment.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/f2fs/segment.c +2 -1
index a29919330e27..d310b82caef1 100644
@@ -740,7 +740,8 @@ static void __remove_discard_cmd(struct f2fs_sb_info *sbi,
 
 	if (dc->error)
 		f2fs_msg(sbi->sb, KERN_INFO,
-				"Issue discard failed, ret: %d", dc->error);
+			"Issue discard(%u, %u, %u) failed, ret: %d",
+			dc->lstart, dc->start, dc->len, dc->error);
 	__detach_discard_cmd(dcc, dc);
 }
Signature · 6 lines
2.12.2.575.gb14f27f


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
author maintainer
#1 lore raw

PATCH 2/3 f2fs: wake up all waiters in f2fs_submit_discard_endio

From: Chao Yu <[email protected]>

There could be more than one waiter waiting discard IO completion, so we need use complete_all() instead of complete() in f2fs_submit_discard_endio to avoid hungtask.

Fixes ec9895add2c5 ("f2fs: don't hold cmd_lock during waiting discard

command")

Cc <[email protected]>
Signed-off-by Chao Yu <[email protected]>
---
 fs/f2fs/segment.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +1 -1
index d310b82caef1..2b11119d9071 100644
@@ -751,7 +751,7 @@ static void f2fs_submit_discard_endio(struct bio *bio)
 
 	dc->error = bio->bi_error;
 	dc->state = D_DONE;
-	complete(&dc->wait);
+	complete_all(&dc->wait);
 	bio_put(bio);
 }
Signature · 1 line
2.12.2.575.gb14f27f
author maintainer
#2 lore raw

PATCH 3/3 f2fs: wait discard IO completion without cmd_lock held

In reply to Chao Yu

From: Chao Yu <[email protected]>

Wait discard IO completion outside cmd_lock to avoid long latency of holding cmd_lock in IO busy scenario.

Signed-off-by Chao Yu <[email protected]>
---
 fs/f2fs/segment.c | 23 ++++++++++++++++++++---
 1 file changed, 20 insertions(+), 3 deletions(-)
fs/f2fs/segment.c +20 -3
index 2b11119d9071..2b926db0588b 100644
@@ -1005,17 +1005,34 @@ static void __wait_discard_cmd(struct f2fs_sb_info *sbi, bool wait_cond)
 	struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
 	struct list_head *wait_list = &(dcc->wait_list);
 	struct discard_cmd *dc, *tmp;
+	bool need_wait;
+
+next:
+	need_wait = false;
 
 	mutex_lock(&dcc->cmd_lock);
 	list_for_each_entry_safe(dc, tmp, wait_list, list) {
-		if (!wait_cond || dc->state == D_DONE) {
-			if (dc->ref)
-				continue;
+		if (!wait_cond || (dc->state == D_DONE && !dc->ref)) {
 			wait_for_completion_io(&dc->wait);
 			__remove_discard_cmd(sbi, dc);
+		} else {
+			dc->ref++;
+			need_wait = true;
+			break;
 		}
 	}
 	mutex_unlock(&dcc->cmd_lock);
+
+	if (need_wait) {
+		wait_for_completion_io(&dc->wait);
+		mutex_lock(&dcc->cmd_lock);
+		f2fs_bug_on(sbi, dc->state != D_DONE);
+		dc->ref--;
+		if (!dc->ref)
+			__remove_discard_cmd(sbi, dc);
+		mutex_unlock(&dcc->cmd_lock);
+		goto next;
+	}
 }
 
 /* This should be covered by global mutex, &sit_i->sentry_lock */
Signature · 6 lines
2.12.2.575.gb14f27f


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot