All of lore.kernel.org
 help / color / mirror / Atom feed
From: Chao Yu <chao@kernel.org>
To: Jaegeuk Kim <jaegeuk@kernel.org>
Cc: linux-f2fs-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] f2fs: fix to trigger a checkpoint in the end of foreground garbage collection
Date: Tue, 4 Apr 2023 18:46:59 +0800	[thread overview]
Message-ID: <08259184-d3c1-cb46-9751-e32e1bc6eb87@kernel.org> (raw)
In-Reply-To: <ZCsXRin7symPxIrn@google.com>

On 2023/4/4 2:13, Jaegeuk Kim wrote:
> On 03/24, Chao Yu wrote:
>> In order to reclaim free blocks in prefree sections before latter use.
> 
> We were supposed to do checkpoint as is?

It seems commit 6f8d4455060d ("f2fs: avoid fi->i_gc_rwsem[WRITE] lock in f2fs_gc")
changed that logic? It caused that if has_not_enough_free_secs() returns false,
it missed to call f2fs_write_checkpoint() before exit f2fs_gc()?

@@ -1110,15 +1116,23 @@ gc_more:
  	if (gc_type == FG_GC)
  		sbi->cur_victim_sec = NULL_SEGNO;

-	if (!sync) {
-		if (has_not_enough_free_secs(sbi, sec_freed, 0)) {
-			if (skipped_round > MAX_SKIP_ATOMIC_COUNT &&
-				skipped_round * 2 >= round)
-				f2fs_drop_inmem_pages_all(sbi, true);
+	if (sync)
+		goto stop;
+
+	if (has_not_enough_free_secs(sbi, sec_freed, 0)) {
+		if (skipped_round <= MAX_SKIP_GC_COUNT ||
+					skipped_round * 2 < round) {
  			segno = NULL_SEGNO;
  			goto gc_more;
  		}

+		if (first_skipped < last_skipped &&
+				(last_skipped - first_skipped) >
+						sbi->skipped_gc_rwsem) {
+			f2fs_drop_inmem_pages_all(sbi, true);
+			segno = NULL_SEGNO;
+			goto gc_more;
+		}
  		if (gc_type == FG_GC)
  			ret = f2fs_write_checkpoint(sbi, &cpc);
  	}

> 
>>
>> Fixes: 6f8d4455060d ("f2fs: avoid fi->i_gc_rwsem[WRITE] lock in f2fs_gc")
>> Signed-off-by: Chao Yu <chao@kernel.org>
>> ---
>>   fs/f2fs/f2fs.h    | 1 +
>>   fs/f2fs/gc.c      | 8 ++++++++
>>   fs/f2fs/segment.c | 1 +
>>   3 files changed, 10 insertions(+)
>>
>> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
>> index 53a005b420cf..b1515375cb4c 100644
>> --- a/fs/f2fs/f2fs.h
>> +++ b/fs/f2fs/f2fs.h
>> @@ -1269,6 +1269,7 @@ struct f2fs_gc_control {
>>   	unsigned int victim_segno;	/* target victim segment number */
>>   	int init_gc_type;		/* FG_GC or BG_GC */
>>   	bool no_bg_gc;			/* check the space and stop bg_gc */
>> +	bool reclaim_space;		/* trigger checkpoint to reclaim space */
>>   	bool should_migrate_blocks;	/* should migrate blocks */
>>   	bool err_gc_skipped;		/* return EAGAIN if GC skipped */
>>   	unsigned int nr_free_secs;	/* # of free sections to do GC */
>> diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
>> index 2996d38aa89c..5a451d3d512d 100644
>> --- a/fs/f2fs/gc.c
>> +++ b/fs/f2fs/gc.c
>> @@ -132,6 +132,7 @@ static int gc_thread_func(void *data)
>>   
>>   		gc_control.init_gc_type = sync_mode ? FG_GC : BG_GC;
>>   		gc_control.no_bg_gc = foreground;
>> +		gc_control.reclaim_space = foreground;
>>   		gc_control.nr_free_secs = foreground ? 1 : 0;
>>   
>>   		/* if return value is not zero, no victim was selected */
>> @@ -1880,6 +1881,13 @@ int f2fs_gc(struct f2fs_sb_info *sbi, struct f2fs_gc_control *gc_control)
>>   				(gc_type == FG_GC) ? sec_freed : 0, 0)) {
>>   		if (gc_type == FG_GC && sec_freed < gc_control->nr_free_secs)
>>   			goto go_gc_more;
>> +
>> +		/*
>> +		 * trigger a checkpoint in the end of foreground garbage
>> +		 * collection.
>> +		 */
>> +		if (gc_control->reclaim_space)
>> +			ret = f2fs_write_checkpoint(sbi, &cpc);
>>   		goto stop;
>>   	}
>>   
>> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
>> index 6c11789da884..b62af2ae1685 100644
>> --- a/fs/f2fs/segment.c
>> +++ b/fs/f2fs/segment.c
>> @@ -421,6 +421,7 @@ void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need)
>>   				.victim_segno = NULL_SEGNO,
>>   				.init_gc_type = BG_GC,
>>   				.no_bg_gc = true,
>> +				.reclaim_space = true,
>>   				.should_migrate_blocks = false,
>>   				.err_gc_skipped = false,
>>   				.nr_free_secs = 1 };
>> -- 
>> 2.25.1

WARNING: multiple messages have this Message-ID (diff)
From: Chao Yu <chao@kernel.org>
To: Jaegeuk Kim <jaegeuk@kernel.org>
Cc: linux-kernel@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net
Subject: Re: [f2fs-dev] [PATCH] f2fs: fix to trigger a checkpoint in the end of foreground garbage collection
Date: Tue, 4 Apr 2023 18:46:59 +0800	[thread overview]
Message-ID: <08259184-d3c1-cb46-9751-e32e1bc6eb87@kernel.org> (raw)
In-Reply-To: <ZCsXRin7symPxIrn@google.com>

On 2023/4/4 2:13, Jaegeuk Kim wrote:
> On 03/24, Chao Yu wrote:
>> In order to reclaim free blocks in prefree sections before latter use.
> 
> We were supposed to do checkpoint as is?

It seems commit 6f8d4455060d ("f2fs: avoid fi->i_gc_rwsem[WRITE] lock in f2fs_gc")
changed that logic? It caused that if has_not_enough_free_secs() returns false,
it missed to call f2fs_write_checkpoint() before exit f2fs_gc()?

@@ -1110,15 +1116,23 @@ gc_more:
  	if (gc_type == FG_GC)
  		sbi->cur_victim_sec = NULL_SEGNO;

-	if (!sync) {
-		if (has_not_enough_free_secs(sbi, sec_freed, 0)) {
-			if (skipped_round > MAX_SKIP_ATOMIC_COUNT &&
-				skipped_round * 2 >= round)
-				f2fs_drop_inmem_pages_all(sbi, true);
+	if (sync)
+		goto stop;
+
+	if (has_not_enough_free_secs(sbi, sec_freed, 0)) {
+		if (skipped_round <= MAX_SKIP_GC_COUNT ||
+					skipped_round * 2 < round) {
  			segno = NULL_SEGNO;
  			goto gc_more;
  		}

+		if (first_skipped < last_skipped &&
+				(last_skipped - first_skipped) >
+						sbi->skipped_gc_rwsem) {
+			f2fs_drop_inmem_pages_all(sbi, true);
+			segno = NULL_SEGNO;
+			goto gc_more;
+		}
  		if (gc_type == FG_GC)
  			ret = f2fs_write_checkpoint(sbi, &cpc);
  	}

> 
>>
>> Fixes: 6f8d4455060d ("f2fs: avoid fi->i_gc_rwsem[WRITE] lock in f2fs_gc")
>> Signed-off-by: Chao Yu <chao@kernel.org>
>> ---
>>   fs/f2fs/f2fs.h    | 1 +
>>   fs/f2fs/gc.c      | 8 ++++++++
>>   fs/f2fs/segment.c | 1 +
>>   3 files changed, 10 insertions(+)
>>
>> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
>> index 53a005b420cf..b1515375cb4c 100644
>> --- a/fs/f2fs/f2fs.h
>> +++ b/fs/f2fs/f2fs.h
>> @@ -1269,6 +1269,7 @@ struct f2fs_gc_control {
>>   	unsigned int victim_segno;	/* target victim segment number */
>>   	int init_gc_type;		/* FG_GC or BG_GC */
>>   	bool no_bg_gc;			/* check the space and stop bg_gc */
>> +	bool reclaim_space;		/* trigger checkpoint to reclaim space */
>>   	bool should_migrate_blocks;	/* should migrate blocks */
>>   	bool err_gc_skipped;		/* return EAGAIN if GC skipped */
>>   	unsigned int nr_free_secs;	/* # of free sections to do GC */
>> diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
>> index 2996d38aa89c..5a451d3d512d 100644
>> --- a/fs/f2fs/gc.c
>> +++ b/fs/f2fs/gc.c
>> @@ -132,6 +132,7 @@ static int gc_thread_func(void *data)
>>   
>>   		gc_control.init_gc_type = sync_mode ? FG_GC : BG_GC;
>>   		gc_control.no_bg_gc = foreground;
>> +		gc_control.reclaim_space = foreground;
>>   		gc_control.nr_free_secs = foreground ? 1 : 0;
>>   
>>   		/* if return value is not zero, no victim was selected */
>> @@ -1880,6 +1881,13 @@ int f2fs_gc(struct f2fs_sb_info *sbi, struct f2fs_gc_control *gc_control)
>>   				(gc_type == FG_GC) ? sec_freed : 0, 0)) {
>>   		if (gc_type == FG_GC && sec_freed < gc_control->nr_free_secs)
>>   			goto go_gc_more;
>> +
>> +		/*
>> +		 * trigger a checkpoint in the end of foreground garbage
>> +		 * collection.
>> +		 */
>> +		if (gc_control->reclaim_space)
>> +			ret = f2fs_write_checkpoint(sbi, &cpc);
>>   		goto stop;
>>   	}
>>   
>> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
>> index 6c11789da884..b62af2ae1685 100644
>> --- a/fs/f2fs/segment.c
>> +++ b/fs/f2fs/segment.c
>> @@ -421,6 +421,7 @@ void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need)
>>   				.victim_segno = NULL_SEGNO,
>>   				.init_gc_type = BG_GC,
>>   				.no_bg_gc = true,
>> +				.reclaim_space = true,
>>   				.should_migrate_blocks = false,
>>   				.err_gc_skipped = false,
>>   				.nr_free_secs = 1 };
>> -- 
>> 2.25.1


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

  reply	other threads:[~2023-04-04 10:47 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-24  7:10 [PATCH] f2fs: fix to trigger a checkpoint in the end of foreground garbage collection Chao Yu
2023-03-24  7:10 ` [f2fs-dev] " Chao Yu
2023-04-03 18:13 ` Jaegeuk Kim
2023-04-03 18:13   ` [f2fs-dev] " Jaegeuk Kim
2023-04-04 10:46   ` Chao Yu [this message]
2023-04-04 10:46     ` Chao Yu
2023-04-04 21:39 ` Jaegeuk Kim
2023-04-04 21:39   ` [f2fs-dev] " Jaegeuk Kim
2023-04-05  2:02   ` Chao Yu
2023-04-05  2:02     ` [f2fs-dev] " Chao Yu
2023-04-05 15:55     ` Jaegeuk Kim
2023-04-05 15:55       ` [f2fs-dev] " Jaegeuk Kim
2023-04-10 13:52       ` Chao Yu
2023-04-10 13:52         ` [f2fs-dev] " Chao Yu
2023-04-10 23:21         ` Jaegeuk Kim
2023-04-10 23:21           ` [f2fs-dev] " Jaegeuk Kim
2023-04-13  9:15           ` Chao Yu
2023-04-13  9:15             ` [f2fs-dev] " Chao Yu
2023-04-13 15:56             ` Jaegeuk Kim
2023-04-13 15:56               ` [f2fs-dev] " Jaegeuk Kim
2023-04-13 15:58               ` Jaegeuk Kim
2023-04-13 15:58                 ` Jaegeuk Kim
2023-04-13 19:25                 ` Jaegeuk Kim
2023-04-13 19:25                   ` Jaegeuk Kim
2023-04-18 15:51                   ` Chao Yu
2023-04-18 15:51                     ` Chao Yu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=08259184-d3c1-cb46-9751-e32e1bc6eb87@kernel.org \
    --to=chao@kernel.org \
    --cc=jaegeuk@kernel.org \
    --cc=linux-f2fs-devel@lists.sourceforge.net \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.