Skip to contentSkip to conversation

f2fs: do not issue small discard commands during checkpoint

PATCH In review Jaegeuk Kim 3 participants 21 messages Jun 13, 2023 – Aug 8, 2023 F2FS file system

lore
Share & tools
lore

Jaegeuk Kim

author maintainer
#0 lore raw

PATCH 1/1 f2fs: do not issue small discard commands during checkpoint

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---
 fs/f2fs/segment.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
fs/f2fs/segment.c +2 -3
index 0c0c033c4bdd..ef46bb085385 100644
@@ -2178,7 +2178,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
 	}
 	mutex_unlock(&dirty_i->seglist_lock);
 
-	if (!f2fs_block_unit_discard(sbi))
+	if (!f2fs_block_unit_discard(sbi) || !force)
 		goto wakeup;
 
 	/* send small discards */
@@ -2192,8 +2192,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
 					sbi->blocks_per_seg, cur_pos);
 			len = next_pos - cur_pos;
 
-			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+			if (f2fs_sb_has_blkzoned(sbi) || len < cpc->trim_minlen)
 				goto skip;
 
 			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
Signature · 8 lines
2.41.0.162.gfafddb0af9-goog



_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
#1 lore raw

Hi Jaegeuk,

Quoted 13 lines

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---
 fs/f2fs/segment.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
fs/f2fs/segment.c +0 -0
index 0c0c033c4bdd..ef46bb085385 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2181

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2178,7 +2178,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
          }
          mutex_unlock(&dirty_i->seglist_lock);
 
-        if (!f2fs_block_unit_discard(sbi))
+        if (!f2fs_block_unit_discard(sbi) || !force)

If we don't handle the discard entries here, dcc->entry_list will still have them, so stale discard entries may be handled by trim, causing incorrect discards to be issued. Therefore, I think this patch should also prevent the creation of discard entries in add_discard_addrs(), unless it is a checkpoint caused by trim.
This would further reduce the latency caused by the creation of a discard entry.

Thanks,
Daejun

author maintainer
#2 lore raw

Hi Daejun,

Quoted 18 lines · Daejun Park

Hi Jaegeuk,

Quoted 16 lines

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---
 fs/f2fs/segment.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
fs/f2fs/segment.c +0 -0
index 0c0c033c4bdd..ef46bb085385 100644
@@ -2178,7 +2178,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
          }
          mutex_unlock(&dirty_i->seglist_lock);
 
Quoted 8 lines

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2181

Quoted 2 lines
fs/f2fs/segment.c +1 -1
-        if (!f2fs_block_unit_discard(sbi))
+        if (!f2fs_block_unit_discard(sbi) || !force)

If we don't handle the discard entries here, dcc->entry_list will still have them, so stale discard entries may be handled by trim, causing incorrect discards to be issued. Therefore, I think this patch should also prevent the creation of discard entries in add_discard_addrs(), unless it is a checkpoint caused by trim.
This would further reduce the latency caused by the creation of a discard entry.

I found this causes some objects were not reclaimed when removing the module. Hence I'm testing v2.

Quoted 4 lines

Thanks,
Daejun

author maintainer
#3 lore raw
In reply to Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

 fs/f2fs/segment.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +1 -1
index 8c7af8b4fc47..0457d620011f 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
 			len = next_pos - cur_pos;
 
 			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
 				goto skip;
 
 			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
Signature · 1 line
2.41.0.162.gfafddb0af9-goog
maintainer
#4 lore raw
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

  fs/f2fs/segment.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2197

Quoted 6 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
  			len = next_pos - cur_pos;
 
  			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
  				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Or, if we prefer to disable small_discards by default, what about below change:

From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
 fs/f2fs/segment.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
 			len = next_pos - cur_pos;
 
 			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
 				goto skip;
 
 			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
 	atomic_set(&dcc->queued_discard, 0);
 	atomic_set(&dcc->discard_cmd_cnt, 0);
 	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
 	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
 	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
 	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2198

Quoted 2 lines
fs/f2fs/segment.c +0 -0
  			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

author maintainer
#5 lore raw
Quoted 34 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

  fs/f2fs/segment.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
  			len = next_pos - cur_pos;
  			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
  				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small
Quoted 5 lines

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

Quoted 48 lines

From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
 fs/f2fs/segment.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
 			len = next_pos - cur_pos;
 
 			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
 				goto skip;
 
 			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
 	atomic_set(&dcc->queued_discard, 0);
 	atomic_set(&dcc->discard_cmd_cnt, 0);
 	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
 	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
 	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
 	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
  			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
maintainer
#6 lore raw
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

   fs/f2fs/segment.c | 2 +-
   1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
   			len = next_pos - cur_pos;
   			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
   				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Thanks,

Quoted 49 lines
Quoted 49 lines

From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
  fs/f2fs/segment.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
  			len = next_pos - cur_pos;
 
  			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
  				goto skip;
 
  			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
  	atomic_set(&dcc->queued_discard, 0);
  	atomic_set(&dcc->discard_cmd_cnt, 0);
  	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
  	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
  	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
  	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
   			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
author maintainer
#7 lore raw
Quoted 40 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

   fs/f2fs/segment.c | 2 +-
   1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
   			len = next_pos - cur_pos;
   			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
   				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 6 lines
Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

You got the point.

Quoted 52 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
  fs/f2fs/segment.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
  			len = next_pos - cur_pos;
 
  			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
  				goto skip;
 
  			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
  	atomic_set(&dcc->queued_discard, 0);
  	atomic_set(&dcc->discard_cmd_cnt, 0);
  	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
  	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
  	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
  	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
   			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

maintainer
#8 lore raw
Quoted 41 lines · Jaegeuk Kim
Quoted 40 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
    			len = next_pos - cur_pos;
    			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
    				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
   fs/f2fs/segment.c | 4 ++--
   1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
   			len = next_pos - cur_pos;
 
   			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
   				goto skip;
 
   			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
   	atomic_set(&dcc->queued_discard, 0);
   	atomic_set(&dcc->discard_cmd_cnt, 0);
   	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
   	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
   	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
   	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
    			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
maintainer
#9 lore raw
Quoted 51 lines · Chao Yu
Quoted 41 lines · Jaegeuk Kim
Quoted 40 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 18 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

    Change log from v1:
     - move the skip logic to avoid dangling objects

    fs/f2fs/segment.c | 2 +-

    1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
                 len = next_pos - cur_pos;
Quoted 4 lines

                if (f2fs_sb_has_blkzoned(sbi) || -                (force && len < cpc->trim_minlen)) +                    !force || len < cpc->trim_minlen) goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

You got the point.

Quoted 8 lines

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

Jaegeuk, any comments?

Thanks,

Quoted 92 lines

f2fs_flush_sit_entries()
{
...
            if (!(cpc->reason & CP_DISCARD)) { cpc->trim_start = segno; add_discard_addrs(sbi, cpc, false); }
...
}

add_discard_addrs()
{
...
    while (force || SM_I(sbi)->dcc_info->nr_discards <= SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...
        if (!de) {
            de = f2fs_kmem_cache_alloc(discard_entry_slab, GFP_F2FS_ZERO, true, NULL); de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start); list_add_tail(&de->list, head);
        }
...
    }
...

Thanks,

Quoted 52 lines
Quoted 52 lines

Thanks,

Quoted 48 lines
Quoted 48 lines

  From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")

---
   fs/f2fs/segment.c | 4 ++--

   1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +0 -0
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
                len = next_pos - cur_pos;
 

               if (f2fs_sb_has_blkzoned(sbi) || -                    !force || len < cpc->trim_minlen) +                (force && len < cpc->trim_minlen)) goto skip;

               f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

fs/f2fs/segment.c +1 -1
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
        atomic_set(&dcc->queued_discard, 0);
        atomic_set(&dcc->discard_cmd_cnt, 0);
        dcc->nr_discards = 0;
-    dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+    dcc->max_discards = 0;
        dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
        dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
        dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

                f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

author maintainer
#10 lore raw
In reply to Chao Yu
Quoted 46 lines · Chao Yu
Quoted 45 lines · Jaegeuk Kim
Quoted 44 lines · Chao Yu
Quoted 42 lines · Jaegeuk Kim
Quoted 34 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
    			len = next_pos - cur_pos;
    			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
    				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small
Quoted 5 lines

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Quoted 34 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}
Quoted 57 lines

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
   fs/f2fs/segment.c | 4 ++--
   1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
   			len = next_pos - cur_pos;
 
   			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
   				goto skip;
 
   			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
   	atomic_set(&dcc->queued_discard, 0);
   	atomic_set(&dcc->discard_cmd_cnt, 0);
   	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
   	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
   	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
   	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
    			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

maintainer
#11 lore raw
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
     			len = next_pos - cur_pos;
     			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
     				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

   From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001

From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
    fs/f2fs/segment.c | 4 ++--
    1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
    			len = next_pos - cur_pos;
 
    			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
    				goto skip;
 
    			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
    	atomic_set(&dcc->queued_discard, 0);
    	atomic_set(&dcc->discard_cmd_cnt, 0);
    	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
    	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
    	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
    	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
     			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
author maintainer
#12 lore raw
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
     			len = next_pos - cur_pos;
     			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
     				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Quoted 95 lines

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

   From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001

From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
    fs/f2fs/segment.c | 4 ++--
    1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
    			len = next_pos - cur_pos;
 
    			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
    				goto skip;
 
    			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
    	atomic_set(&dcc->queued_discard, 0);
    	atomic_set(&dcc->discard_cmd_cnt, 0);
    	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
    	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
    	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
    	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
     			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
maintainer
#13 lore raw
Quoted 88 lines · Jaegeuk Kim
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
      			len = next_pos - cur_pos;
      			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
      				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Quoted 5 lines
Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Let me explain more:

Previously, we have two mechanisms to cache & submit small discards:

a) set max small discard number in /sys/fs/f2fs/vdb/max_small_discards, and checkpoint will cache small discard candidates w/ configured maximum number.

b) call FITRIM ioctl, also, checkpoint in f2fs_trim_fs() will cache small discard candidates w/ configured discard granularity, but w/o limitation of number. FSTRIM interface is asynchronized, so it won't submit discard directly.

Finally, discard thread will submit them in background periodically.

So what I mean is the mechanism a) is broken, since no matter how we configure the sysfs entry /sys/fs/f2fs/vdb/max_small_discards, checkpoint will not cache small discard candidates any more.

So, it needs to fix max_small_discards sysfs functionality? or just drop the functionality?

Quoted 3 lines

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Sorry, I'm not sure I get your point, do you mean max_small_discards functionality is obsoleted, so it recommended to use fstrim to cache & submit small discards?

Let me know, if I'm missing something or misunderstanding the point.

Thanks,

Quoted 96 lines
Quoted 96 lines

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

    From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001

From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
     fs/f2fs/segment.c | 4 ++--
     1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
     			len = next_pos - cur_pos;
 
     			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
     				goto skip;
 
     			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
     	atomic_set(&dcc->queued_discard, 0);
     	atomic_set(&dcc->discard_cmd_cnt, 0);
     	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
     	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
     	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
     	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
      			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
maintainer
#14 lore raw

Comments?

Quoted 228 lines · Chao Yu
Quoted 88 lines · Jaegeuk Kim
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 18 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

      Change log from v1:
       - move the skip logic to avoid dangling objects

      fs/f2fs/segment.c | 2 +-

      1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
                   len = next_pos - cur_pos;
Quoted 4 lines

                  if (f2fs_sb_has_blkzoned(sbi) || -                (force && len < cpc->trim_minlen)) +                    !force || len < cpc->trim_minlen) goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
   0         .       .       .       .       .       .       .       .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
   0         3       1       .       .       .       .       .       .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
   0         .       .       .       .       .       .       .       .

Quoted 5 lines
Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
   0         .       .       .       .       .       .       .       .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Let me explain more:

Previously, we have two mechanisms to cache & submit small discards:

a) set max small discard number in /sys/fs/f2fs/vdb/max_small_discards, and checkpoint will cache small discard candidates w/ configured maximum number.

b) call FITRIM ioctl, also, checkpoint in f2fs_trim_fs() will cache small discard candidates w/ configured discard granularity, but w/o limitation of number. FSTRIM interface is asynchronized, so it won't submit discard directly.

Finally, discard thread will submit them in background periodically.

So what I mean is the mechanism a) is broken, since no matter how we configure the sysfs entry /sys/fs/f2fs/vdb/max_small_discards, checkpoint will not cache small discard candidates any more.

So, it needs to fix max_small_discards sysfs functionality? or just drop the functionality?

Quoted 3 lines

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Sorry, I'm not sure I get your point, do you mean max_small_discards functionality is obsoleted, so it recommended to use fstrim to cache & submit small discards?

Let me know, if I'm missing something or misunderstanding the point.

Thanks,

Quoted 95 lines
Quoted 95 lines

Thanks,

Quoted 91 lines
Quoted 91 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...
            if (!(cpc->reason & CP_DISCARD)) { cpc->trim_start = segno; add_discard_addrs(sbi, cpc, false); }
...
}

add_discard_addrs()
{
...
    while (force || SM_I(sbi)->dcc_info->nr_discards <= SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...
        if (!de) {
            de = f2fs_kmem_cache_alloc(discard_entry_slab, GFP_F2FS_ZERO, true, NULL); de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start); list_add_tail(&de->list, head);
        }
...
    }
...

Thanks,

Quoted 52 lines
Quoted 52 lines

Thanks,

Quoted 48 lines
Quoted 48 lines

    From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")

---
     fs/f2fs/segment.c | 4 ++--

     1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +0 -0
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
                  len = next_pos - cur_pos;
 

                 if (f2fs_sb_has_blkzoned(sbi) || -                    !force || len < cpc->trim_minlen) +                (force && len < cpc->trim_minlen)) goto skip;

                 f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

fs/f2fs/segment.c +1 -1
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
          atomic_set(&dcc->queued_discard, 0);
          atomic_set(&dcc->discard_cmd_cnt, 0);
          dcc->nr_discards = 0;
-    dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+    dcc->max_discards = 0;
          dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
          dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
          dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

                  f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

author maintainer
#15 lore raw
In reply to Chao Yu
Quoted 106 lines · Chao Yu
Quoted 88 lines · Jaegeuk Kim
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
      			len = next_pos - cur_pos;
      			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
      				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Quoted 5 lines
Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Let me explain more:

Previously, we have two mechanisms to cache & submit small discards:

a) set max small discard number in /sys/fs/f2fs/vdb/max_small_discards, and checkpoint will cache small discard candidates w/ configured maximum number.

b) call FITRIM ioctl, also, checkpoint in f2fs_trim_fs() will cache small discard candidates w/ configured discard granularity, but w/o limitation of number. FSTRIM interface is asynchronized, so it won't submit discard directly.

Quoted 5 lines

Finally, discard thread will submit them in background periodically.

So what I mean is the mechanism a) is broken, since no matter how we configure the sysfs entry /sys/fs/f2fs/vdb/max_small_discards, checkpoint will not cache small discard candidates any more.

Ok, it seems what I encountered before was adding this small discard even after issuing it by checkpoint. Thoughts?

node=0], cp [data=0, node=0, meta=0], app [read=0 (direct=0, buffered=0), mapped=0], compr(buffered=0, mapped=0)], fs [data=0, (gc_data=0, cdata=0), node=0, meta=0]

 f2fs_discard-25-752     [003] .....  9744.173085: f2fs_iostat_latency: dev = (254,51), iotype [peak lat.(ms)/avg lat.(ms)/count], rd_data [0/0/0], rd_node [0/0/0], rd_meta [0/0/0], wr_sync_data [0/0/0], wr_sync_node [0/0/0], wr_sync_meta [0/0/0], wr_async_data [0/0/0], wr_async_node [0/0/0], wr_async_meta [0/0/0]
 f2fs_discard-25-752     [003] .....  9744.173089: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0c6, blklen = 0x1
 f2fs_discard-25-752     [003] .....  9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
 f2fs_discard-25-752     [003] .....  9744.173126: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c116, blklen = 0x1
 f2fs_discard-25-752     [003] .....  9744.173140: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c121, blklen = 0x1
 f2fs_discard-25-752     [003] .....  9744.173154: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c15b, blklen = 0x1
 f2fs_discard-25-752     [003] .....  9744.173169: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c16e, blklen = 0x1
 f2fs_discard-25-752     [003] .....  9744.173183: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c181, blklen = 0x1
 f2fs_discard-25-752     [004] .....  9744.175272: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0a9, blklen = 0x1
 f2fs_discard-25-752     [004] .....  9744.175345: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0c6, blklen = 0x1
 f2fs_discard-25-752     [004] .....  9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
 f2fs_discard-25-752     [004] .....  9744.175352: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c116, blklen = 0x1
 f2fs_discard-25-752     [004] .....  9744.175354: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c121, blklen = 0x1
 f2fs_discard-25-752     [004] .....  9744.175360: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c15b, blklen = 0x1
 f2fs_discard-25-752     [004] .....  9744.175362: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c16e, blklen = 0x1
 f2fs_discard-25-752     [004] .....  9744.175367: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c181, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.228775: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c197, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.228950: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1ad, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.228965: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1b1, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.228994: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1b3, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.229006: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c27a, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.229017: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c2a3, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.229028: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c2ab, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.229041: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c304, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.230811: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c197, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.230926: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1ad, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.230929: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1b1, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.230932: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1b3, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.230940: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c27a, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.230945: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c2a3, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.230947: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c2ab, blklen = 0x1
 f2fs_discard-25-752     [002] .....  9744.230952: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c304, blklen = 0x1
Quoted 8 lines

So, it needs to fix max_small_discards sysfs functionality? or just drop the functionality?

Quoted 3 lines

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Quoted 103 lines

Sorry, I'm not sure I get your point, do you mean max_small_discards functionality is obsoleted, so it recommended to use fstrim to cache & submit small discards?

Let me know, if I'm missing something or misunderstanding the point.

Thanks,

Quoted 96 lines
Quoted 96 lines

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

    From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001

From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
     fs/f2fs/segment.c | 4 ++--
     1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
     			len = next_pos - cur_pos;
 
     			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
     				goto skip;
 
     			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
     	atomic_set(&dcc->queued_discard, 0);
     	atomic_set(&dcc->discard_cmd_cnt, 0);
     	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
     	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
     	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
     	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
      			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

maintainer
#16 lore raw
Quoted 107 lines · Jaegeuk Kim
Quoted 106 lines · Chao Yu
Quoted 88 lines · Jaegeuk Kim
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
       			len = next_pos - cur_pos;
       			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
       				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Quoted 5 lines
Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Let me explain more:

Previously, we have two mechanisms to cache & submit small discards:

a) set max small discard number in /sys/fs/f2fs/vdb/max_small_discards, and checkpoint will cache small discard candidates w/ configured maximum number.

b) call FITRIM ioctl, also, checkpoint in f2fs_trim_fs() will cache small discard candidates w/ configured discard granularity, but w/o limitation of number. FSTRIM interface is asynchronized, so it won't submit discard directly.

Quoted 8 lines
Quoted 5 lines

Finally, discard thread will submit them in background periodically.

So what I mean is the mechanism a) is broken, since no matter how we configure the sysfs entry /sys/fs/f2fs/vdb/max_small_discards, checkpoint will not cache small discard candidates any more.

Ok, it seems what I encountered before was adding this small discard even after issuing it by checkpoint. Thoughts?

Do you mean: in f2fs_clear_prefree_segments(), small discard may overlap segment granularity discard?

e.g.
- f2fs_clear_prefree_segments
- f2fs_issue_discard(0, 512) --- segment granularity discard
- f2fs_issue_discard(0, 1) --- small discard
- f2fs_issue_discard(5, 1) --- small discard

Thanks,

Quoted 146 lines

node=0], cp [data=0, node=0, meta=0], app [read=0 (direct=0, buffered=0), mapped=0], compr(buffered=0, mapped=0)], fs [data=0, (gc_data=0, cdata=0), node=0, meta=0]

  f2fs_discard-25-752     [003] .....  9744.173085: f2fs_iostat_latency: dev = (254,51), iotype [peak lat.(ms)/avg lat.(ms)/count], rd_data [0/0/0], rd_node [0/0/0], rd_meta [0/0/0], wr_sync_data [0/0/0], wr_sync_node [0/0/0], wr_sync_meta [0/0/0], wr_async_data [0/0/0], wr_async_node [0/0/0], wr_async_meta [0/0/0]
  f2fs_discard-25-752     [003] .....  9744.173089: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0c6, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173126: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c116, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173140: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c121, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173154: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c15b, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173169: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c16e, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173183: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c181, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175272: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0a9, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175345: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0c6, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175352: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c116, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175354: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c121, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175360: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c15b, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175362: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c16e, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175367: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c181, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.228775: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c197, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.228950: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1ad, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.228965: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1b1, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.228994: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1b3, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.229006: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c27a, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.229017: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c2a3, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.229028: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c2ab, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.229041: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c304, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230811: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c197, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230926: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1ad, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230929: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1b1, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230932: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1b3, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230940: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c27a, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230945: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c2a3, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230947: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c2ab, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230952: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c304, blklen = 0x1
Quoted 111 lines

So, it needs to fix max_small_discards sysfs functionality? or just drop the functionality?

Quoted 3 lines

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Sorry, I'm not sure I get your point, do you mean max_small_discards functionality is obsoleted, so it recommended to use fstrim to cache & submit small discards?

Let me know, if I'm missing something or misunderstanding the point.

Thanks,

Quoted 96 lines
Quoted 96 lines

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

     From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001

From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
      fs/f2fs/segment.c | 4 ++--
      1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
      			len = next_pos - cur_pos;
 
      			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
      				goto skip;
 
      			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
      	atomic_set(&dcc->queued_discard, 0);
      	atomic_set(&dcc->discard_cmd_cnt, 0);
      	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
      	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
      	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
      	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
       			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
author maintainer
#17 lore raw
Quoted 114 lines · Chao Yu
Quoted 113 lines · Jaegeuk Kim
Quoted 111 lines · Chao Yu
Quoted 88 lines · Jaegeuk Kim
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 17 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency insanely. Let's issue small discards only by trim.

---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

fs/f2fs/segment.c +0 -0
index 8c7af8b4fc47..0457d620011f 100644

fs/f2fs/segment.c void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, · line 2196

Quoted 5 lines
fs/f2fs/segment.c +1 -1
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
       			len = next_pos - cur_pos;
       			if (f2fs_sb_has_blkzoned(sbi) ||
-			    (force && len < cpc->trim_minlen))
+					!force || len < cpc->trim_minlen)
       				goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Quoted 5 lines
Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Let me explain more:

Previously, we have two mechanisms to cache & submit small discards:

a) set max small discard number in /sys/fs/f2fs/vdb/max_small_discards, and checkpoint will cache small discard candidates w/ configured maximum number.

b) call FITRIM ioctl, also, checkpoint in f2fs_trim_fs() will cache small discard candidates w/ configured discard granularity, but w/o limitation of number. FSTRIM interface is asynchronized, so it won't submit discard directly.

Finally, discard thread will submit them in background periodically.

So what I mean is the mechanism a) is broken, since no matter how we configure the sysfs entry /sys/fs/f2fs/vdb/max_small_discards, checkpoint will not cache small discard candidates any more.

Quoted 5 lines
Quoted 2 lines

Ok, it seems what I encountered before was adding this small discard even after issuing it by checkpoint. Thoughts?

Do you mean: in f2fs_clear_prefree_segments(), small discard may overlap segment granularity discard?

I didn't dig enough tho, don't think so. Somehow I got a loop as below which said the same LBA was issued and added back repeatedly, not seen this short log unfortunately.

Quoted 155 lines

e.g.
- f2fs_clear_prefree_segments
- f2fs_issue_discard(0, 512) --- segment granularity discard
- f2fs_issue_discard(0, 1) --- small discard
- f2fs_issue_discard(5, 1) --- small discard

Thanks,

Quoted 146 lines

node=0], cp [data=0, node=0, meta=0], app [read=0 (direct=0, buffered=0), mapped=0], compr(buffered=0, mapped=0)], fs [data=0, (gc_data=0, cdata=0), node=0, meta=0]

  f2fs_discard-25-752     [003] .....  9744.173085: f2fs_iostat_latency: dev = (254,51), iotype [peak lat.(ms)/avg lat.(ms)/count], rd_data [0/0/0], rd_node [0/0/0], rd_meta [0/0/0], wr_sync_data [0/0/0], wr_sync_node [0/0/0], wr_sync_meta [0/0/0], wr_async_data [0/0/0], wr_async_node [0/0/0], wr_async_meta [0/0/0]
  f2fs_discard-25-752     [003] .....  9744.173089: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0c6, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173126: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c116, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173140: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c121, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173154: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c15b, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173169: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c16e, blklen = 0x1
  f2fs_discard-25-752     [003] .....  9744.173183: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c181, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175272: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0a9, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175345: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0c6, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175352: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c116, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175354: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c121, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175360: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c15b, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175362: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c16e, blklen = 0x1
  f2fs_discard-25-752     [004] .....  9744.175367: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c181, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.228775: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c197, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.228950: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1ad, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.228965: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1b1, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.228994: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c1b3, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.229006: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c27a, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.229017: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c2a3, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.229028: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c2ab, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.229041: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c304, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230811: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c197, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230926: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1ad, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230929: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1b1, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230932: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c1b3, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230940: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c27a, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230945: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c2a3, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230947: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c2ab, blklen = 0x1
  f2fs_discard-25-752     [002] .....  9744.230952: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c304, blklen = 0x1
Quoted 111 lines

So, it needs to fix max_small_discards sysfs functionality? or just drop the functionality?

Quoted 3 lines

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Sorry, I'm not sure I get your point, do you mean max_small_discards functionality is obsoleted, so it recommended to use fstrim to cache & submit small discards?

Let me know, if I'm missing something or misunderstanding the point.

Thanks,

Quoted 96 lines
Quoted 96 lines

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

     From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001

From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
      fs/f2fs/segment.c | 4 ++--
      1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
      			len = next_pos - cur_pos;
 
      			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
      				goto skip;
 
      			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
      	atomic_set(&dcc->queued_discard, 0);
      	atomic_set(&dcc->discard_cmd_cnt, 0);
      	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
      	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
      	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
      	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line
fs/f2fs/segment.c +0 -0
       			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
maintainer
#18 lore raw
Quoted 133 lines · Jaegeuk Kim
Quoted 114 lines · Chao Yu
Quoted 113 lines · Jaegeuk Kim
Quoted 111 lines · Chao Yu
Quoted 88 lines · Jaegeuk Kim
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu
Quoted 22 lines · Jaegeuk Kim

If there're huge # of small discards, this will increase checkpoint latency
insanely. Let's issue small discards only by trim.

Signed-off-by: Jaegeuk Kim <[email protected]>
---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 8c7af8b4fc47..0457d620011f 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
len = next_pos - cur_pos;
if (f2fs_sb_has_blkzoned(sbi) ||
- (force && len < cpc->trim_minlen))
+ !force || len < cpc->trim_minlen)
goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting max_small_discards to zero? otherwise, w/ above change, max_small_discards logic may be broken?

What:           /sys/fs/f2fs/<disk>/max_small_discards
Date:           November 2013
Contact:        "Jaegeuk Kim" <[email protected]>
Description:    Controls the issue rate of discard commands that consist of small

blocks less than 2MB. The candidates to be discarded are cached until checkpoint is triggered, and issued during the checkpoint. By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Quoted 5 lines
Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Let me explain more:

Previously, we have two mechanisms to cache & submit small discards:

a) set max small discard number in /sys/fs/f2fs/vdb/max_small_discards, and checkpoint will cache small discard candidates w/ configured maximum number.

b) call FITRIM ioctl, also, checkpoint in f2fs_trim_fs() will cache small discard candidates w/ configured discard granularity, but w/o limitation of number. FSTRIM interface is asynchronized, so it won't submit discard directly.

Finally, discard thread will submit them in background periodically.

So what I mean is the mechanism a) is broken, since no matter how we configure the sysfs entry /sys/fs/f2fs/vdb/max_small_discards, checkpoint will not cache small discard candidates any more.

Quoted 5 lines
Quoted 2 lines

Ok, it seems what I encountered before was adding this small discard even after issuing it by checkpoint. Thoughts?

Do you mean: in f2fs_clear_prefree_segments(), small discard may overlap segment granularity discard?

I didn't dig enough tho, don't think so. Somehow I got a loop as below which said the same LBA was issued and added back repeatedly, not seen this short log unfortunately.

Quoted 8 lines

e.g.
- f2fs_clear_prefree_segments
- f2fs_issue_discard(0, 512) --- segment granularity discard
- f2fs_issue_discard(0, 1) --- small discard
- f2fs_issue_discard(5, 1) --- small discard

Thanks,

[snip]

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
   f2fs_discard-25-752     [003] .....  9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
   f2fs_discard-25-752     [004] .....  9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1

I don't see any loop via above log, am I missing something?

The traces were printed in below call paths, and it printed the same LBAs as expected?

- issue_discard_thread
- __issue_discard_cmd
- __submit_discard_cmd
- trace_f2fs_issue_discard
9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
- __wait_all_discard_cmd
- __wait_discard_cmd_range
- __remove_discard_cmd
- trace_f2fs_remove_discard
9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1

Thanks,

Quoted 112 lines
Quoted 112 lines
Quoted 112 lines
Quoted 112 lines

So, it needs to fix max_small_discards sysfs functionality? or just drop the functionality?

Quoted 3 lines

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Sorry, I'm not sure I get your point, do you mean max_small_discards functionality is obsoleted, so it recommended to use fstrim to cache & submit small discards?

Let me know, if I'm missing something or misunderstanding the point.

Thanks,

Quoted 96 lines
Quoted 96 lines

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

      From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001

From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order to disable small discard by default, so that if there're huge number of small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint"), due to it breaks small discard feature which may be configured via sysfs entry max_small_discards.

Fixes 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
---
       fs/f2fs/segment.c | 4 ++--
       1 file changed, 2 insertions(+), 2 deletions(-)
fs/f2fs/segment.c +2 -2
index 14c822e5c9c9..0a313368f18b 100644
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
       			len = next_pos - cur_pos;
 
       			if (f2fs_sb_has_blkzoned(sbi) ||
-					!force || len < cpc->trim_minlen)
+			    (force && len < cpc->trim_minlen))
       				goto skip;
 
       			f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
       	atomic_set(&dcc->queued_discard, 0);
       	atomic_set(&dcc->discard_cmd_cnt, 0);
       	dcc->nr_discards = 0;
-	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+	dcc->max_discards = 0;
       	dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
       	dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
       	dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
Signature · 1 line
2.40.1
Quoted 1 line

f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

author maintainer
#19 lore raw
Quoted 137 lines · Chao Yu
Quoted 133 lines · Jaegeuk Kim
Quoted 114 lines · Chao Yu
Quoted 113 lines · Jaegeuk Kim
Quoted 111 lines · Chao Yu
Quoted 88 lines · Jaegeuk Kim
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 39 lines · Jaegeuk Kim
Quoted 38 lines · Chao Yu

On 2023/6/15 0:10, Jaegeuk Kim wrote:
If there're huge # of small discards, this will increase checkpoint latency
insanely. Let's issue small discards only by trim.

Signed-off-by: Jaegeuk Kim <[email protected]>
---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 8c7af8b4fc47..0457d620011f 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
len = next_pos - cur_pos;
if (f2fs_sb_has_blkzoned(sbi) ||
- (force && len < cpc->trim_minlen))
+ !force || len < cpc->trim_minlen)
goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting
max_small_discards to zero? otherwise, w/ above change, max_small_discards
logic may be broken?

What: /sys/fs/f2fs/<disk>/max_small_discards
Date: November 2013
Contact: "Jaegeuk Kim" <[email protected]>
Description: Controls the issue rate of discard commands that consist of small
blocks less than 2MB. The candidates to be discarded are cached until
checkpoint is triggered, and issued during the checkpoint.
By default, it is disabled with 0.

Quoted 3 lines
Quoted 1 line

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Quoted 5 lines
Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Let me explain more:

Previously, we have two mechanisms to cache & submit small discards:

a) set max small discard number in /sys/fs/f2fs/vdb/max_small_discards, and checkpoint will cache small discard candidates w/ configured maximum number.

b) call FITRIM ioctl, also, checkpoint in f2fs_trim_fs() will cache small discard candidates w/ configured discard granularity, but w/o limitation of number. FSTRIM interface is asynchronized, so it won't submit discard directly.

Finally, discard thread will submit them in background periodically.

So what I mean is the mechanism a) is broken, since no matter how we configure the sysfs entry /sys/fs/f2fs/vdb/max_small_discards, checkpoint will not cache small discard candidates any more.

Quoted 5 lines
Quoted 2 lines

Ok, it seems what I encountered before was adding this small discard even after issuing it by checkpoint. Thoughts?

Do you mean: in f2fs_clear_prefree_segments(), small discard may overlap segment granularity discard?

I didn't dig enough tho, don't think so. Somehow I got a loop as below which said the same LBA was issued and added back repeatedly, not seen this short log unfortunately.

Quoted 8 lines

e.g.
- f2fs_clear_prefree_segments
- f2fs_issue_discard(0, 512) --- segment granularity discard
- f2fs_issue_discard(0, 1) --- small discard
- f2fs_issue_discard(5, 1) --- small discard

Thanks,

[snip]

Quoted 7 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
   f2fs_discard-25-752     [003] .....  9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
   f2fs_discard-25-752     [004] .....  9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1

I don't see any loop via above log, am I missing something?

The traces were printed in below call paths, and it printed the same LBAs as expected?

I said the log was cut to show the loop enoughly. I'll try to reproduce this later locally.

Quoted 12 lines

- issue_discard_thread
- __issue_discard_cmd
- __submit_discard_cmd
- trace_f2fs_issue_discard
9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
- __wait_all_discard_cmd
- __wait_discard_cmd_range
- __remove_discard_cmd
- trace_f2fs_remove_discard
9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1

Quoted 114 lines

Thanks,

Quoted 112 lines
Quoted 112 lines
Quoted 112 lines
Quoted 112 lines

So, it needs to fix max_small_discards sysfs functionality? or just drop the functionality?

Quoted 3 lines

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Sorry, I'm not sure I get your point, do you mean max_small_discards functionality is obsoleted, so it recommended to use fstrim to cache & submit small discards?

Let me know, if I'm missing something or misunderstanding the point.

Thanks,

Quoted 96 lines
Quoted 96 lines

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines
Quoted 49 lines

From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order
to disable small discard by default, so that if there're huge number of
small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard
commands during checkpoint"), due to it breaks small discard feature which
may be configured via sysfs entry max_small_discards.

Fixes: 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
Signed-off-by: Chao Yu <[email protected]>
---
fs/f2fs/segment.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 14c822e5c9c9..0a313368f18b 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
len = next_pos - cur_pos;

if (f2fs_sb_has_blkzoned(sbi) ||
- !force || len < cpc->trim_minlen)
+ (force && len < cpc->trim_minlen))
goto skip;

f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
atomic_set(&dcc->queued_discard, 0);
atomic_set(&dcc->discard_cmd_cnt, 0);
dcc->nr_discards = 0;
- dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+ dcc->max_discards = 0;
dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
--
2.40.1

f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

maintainer
#20 lore raw
Quoted 141 lines · Jaegeuk Kim
Quoted 140 lines · Chao Yu
Quoted 133 lines · Jaegeuk Kim
Quoted 114 lines · Chao Yu
Quoted 113 lines · Jaegeuk Kim
Quoted 111 lines · Chao Yu
Quoted 88 lines · Jaegeuk Kim
Quoted 87 lines · Chao Yu
Quoted 50 lines · Jaegeuk Kim
Quoted 49 lines · Chao Yu
Quoted 48 lines · Jaegeuk Kim
Quoted 46 lines · Chao Yu
Quoted 42 lines · Jaegeuk Kim

On 07/03, Chao Yu wrote:
On 2023/6/15 0:10, Jaegeuk Kim wrote:
If there're huge # of small discards, this will increase checkpoint latency
insanely. Let's issue small discards only by trim.

Signed-off-by: Jaegeuk Kim <[email protected]>
---

Change log from v1:
- move the skip logic to avoid dangling objects

fs/f2fs/segment.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 8c7af8b4fc47..0457d620011f 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
len = next_pos - cur_pos;
if (f2fs_sb_has_blkzoned(sbi) ||
- (force && len < cpc->trim_minlen))
+ !force || len < cpc->trim_minlen)
goto skip;

Sorry for late reply.

We have a configuration for such case, what do you think of setting
max_small_discards to zero? otherwise, w/ above change, max_small_discards
logic may be broken?

What: /sys/fs/f2fs/<disk>/max_small_discards
Date: November 2013
Contact: "Jaegeuk Kim" <[email protected]>
Description: Controls the issue rate of discard commands that consist of small
blocks less than 2MB. The candidates to be discarded are cached until
checkpoint is triggered, and issued during the checkpoint.
By default, it is disabled with 0.

Or, if we prefer to disable small_discards by default, what about below change:

I think small_discards is fine, but need to avoid long checkpoint latency only.

I didn't get you, do you mean we can still issue small discard by fstrim, so small_discards functionality is fine?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

You got the point.

Well, actually, what I mean is max_small_discards sysfs entry's functionality is broken. Now, the entry can not be used to control number of small discards committed by checkpoint.

Could you descrbie this problem first?

Oh, alright, actually, I've described this problem literally, but maybe it's not clear, let me give some examples as below:

echo 0 > /sys/fs/f2fs/vdb/max_small_discards xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

echo 100 > /sys/fs/f2fs/vdb/max_small_discards rm /mnt/f2fs/file
xfs_io -f /mnt/f2fs/file -c "pwrite 0 2m" -c "fsync" xfs_io /mnt/f2fs/file -c "fpunch 0 4k"
sync
cat /proc/fs/f2fs/vdb/discard_plist_info |head -2

Before the patch:

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 3 1 . . . . . .

After the patch:
Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

Quoted 5 lines
Quoted 5 lines

Discard pend list(Show diacrd_cmd count on each entry, .:not exist):
0 . . . . . . . .

So, now max_small_discards can not be used to control small discard number cached by checkpoint.

Let me explain more:

Previously, we have two mechanisms to cache & submit small discards:

a) set max small discard number in /sys/fs/f2fs/vdb/max_small_discards, and checkpoint will cache small discard candidates w/ configured maximum number.

b) call FITRIM ioctl, also, checkpoint in f2fs_trim_fs() will cache small discard candidates w/ configured discard granularity, but w/o limitation of number. FSTRIM interface is asynchronized, so it won't submit discard directly.

Finally, discard thread will submit them in background periodically.

So what I mean is the mechanism a) is broken, since no matter how we configure the sysfs entry /sys/fs/f2fs/vdb/max_small_discards, checkpoint will not cache small discard candidates any more.

Quoted 5 lines
Quoted 2 lines

Ok, it seems what I encountered before was adding this small discard even after issuing it by checkpoint. Thoughts?

Do you mean: in f2fs_clear_prefree_segments(), small discard may overlap segment granularity discard?

I didn't dig enough tho, don't think so. Somehow I got a loop as below which said the same LBA was issued and added back repeatedly, not seen this short log unfortunately.

Quoted 8 lines

e.g.
- f2fs_clear_prefree_segments
- f2fs_issue_discard(0, 512) --- segment granularity discard
- f2fs_issue_discard(0, 1) --- small discard
- f2fs_issue_discard(5, 1) --- small discard

Thanks,

[snip]

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
    f2fs_discard-25-752     [003] .....  9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
    f2fs_discard-25-752     [004] .....  9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
Quoted 6 lines
Quoted 4 lines

I don't see any loop via above log, am I missing something?

The traces were printed in below call paths, and it printed the same LBAs as expected?

I said the log was cut to show the loop enoughly. I'll try to reproduce this

Oh, I misunderstood it...

Quoted 1 line

later locally.

Alright.

Thanks,

Quoted 127 lines
Quoted 127 lines

- issue_discard_thread
- __issue_discard_cmd
- __submit_discard_cmd
- trace_f2fs_issue_discard
9744.173111: f2fs_issue_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1
- __wait_all_discard_cmd
- __wait_discard_cmd_range
- __remove_discard_cmd
- trace_f2fs_remove_discard
9744.175348: f2fs_remove_discard: dev = (254,51), blkstart = 0x18c0ca, blklen = 0x1

Thanks,

Quoted 112 lines
Quoted 112 lines
Quoted 112 lines
Quoted 112 lines

So, it needs to fix max_small_discards sysfs functionality? or just drop the functionality?

Quoted 3 lines

Since we do not submit small discards anymore during checkpoint. Why not relying on the discard thread to issue them?

Sorry, I'm not sure I get your point, do you mean max_small_discards functionality is obsoleted, so it recommended to use fstrim to cache & submit small discards?

Let me know, if I'm missing something or misunderstanding the point.

Thanks,

Quoted 96 lines
Quoted 96 lines

Thanks,

Quoted 92 lines
Quoted 92 lines

I think there is another way to achieve "avoid long checkpoint latency caused by committing huge # of small discards", the way is we can set max_small_discards to small value or zero, w/ such configuration, it will take checkpoint much less time or no time to committing small discard due to below control logic:

f2fs_flush_sit_entries()
{
...

			if (!(cpc->reason & CP_DISCARD)) {
				cpc->trim_start = segno;
				add_discard_addrs(sbi, cpc, false);
			}

...
}

add_discard_addrs()
{
...

	while (force || SM_I(sbi)->dcc_info->nr_discards <=
				SM_I(sbi)->dcc_info->max_discards) {

It will break the loop once nr_discards is larger than max_discards, if max_discards is set to zero, checkpoint won't take time to handle small discards.

...

		if (!de) {
			de = f2fs_kmem_cache_alloc(discard_entry_slab,
						GFP_F2FS_ZERO, true, NULL);
			de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
			list_add_tail(&de->list, head);
		}
...
	}

...

Thanks,

Quoted 53 lines
Quoted 53 lines

Thanks,

Quoted 49 lines

From eb89d9b56e817e3046d7fa17165b12416f09d456 Mon Sep 17 00:00:00 2001
From: Chao Yu <[email protected]>
Date: Mon, 3 Jul 2023 09:06:53 +0800
Subject: [PATCH] Revert "f2fs: enable small discard by default"

This reverts commit d618ebaf0aa83d175658aea5291e0c459d471d39 in order
to disable small discard by default, so that if there're huge number of
small discards, it will decrease checkpoint's latency obviously.

Also, this patch reverts 9ac00e7cef10 ("f2fs: do not issue small discard
commands during checkpoint"), due to it breaks small discard feature which
may be configured via sysfs entry max_small_discards.

Fixes: 9ac00e7cef10 ("f2fs: do not issue small discard commands during checkpoint")
Signed-off-by: Chao Yu <[email protected]>
---
fs/f2fs/segment.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 14c822e5c9c9..0a313368f18b 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -2193,7 +2193,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
len = next_pos - cur_pos;

if (f2fs_sb_has_blkzoned(sbi) ||
- !force || len < cpc->trim_minlen)
+ (force && len < cpc->trim_minlen))
goto skip;

f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
@@ -2269,7 +2269,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
atomic_set(&dcc->queued_discard, 0);
atomic_set(&dcc->discard_cmd_cnt, 0);
dcc->nr_discards = 0;
- dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
+ dcc->max_discards = 0;
dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
--
2.40.1

f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,