Skip to contentSkip to conversation

ext4: fix log printing of ext4_mb_regular_allocator()

PATCH Applied brookxu 4 participants 4 messages Aug 15, 2020 – Aug 19, 2020 ext4 file system

lore
Share & tools
lore

brookxu

author
#0 lore raw

PATCH v3 1/1 ext4: fix log printing of ext4_mb_regular_allocator()

Fix log printing of ext4_mb_regular_allocator(),it may be an unintentional behavior.

V3:
It may be better to add a comma between start and len, which is convenient for script processing.

V2:
Add more valuable information, such as group, start, len, lost.

---
 fs/ext4/mballoc.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
fs/ext4/mballoc.c +6 -3
index c0a331e..70b110f 100644
@@ -2218,6 +2218,7 @@ static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac,
 	struct ext4_sb_info *sbi;
 	struct super_block *sb;
 	struct ext4_buddy e4b;
+	unsigned int lost;
 
 	sb = ac->ac_sb;
 	sbi = EXT4_SB(sb);
@@ -2341,22 +2342,24 @@ static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac,
 		 * We've been searching too long. Let's try to allocate
 		 * the best chunk we've found so far
 		 */
-
 		ext4_mb_try_best_found(ac, &e4b);
 		if (ac->ac_status != AC_STATUS_FOUND) {
 			/*
 			 * Someone more lucky has already allocated it.
 			 * The only thing we can do is just take first
 			 * found block(s)
-			printk(KERN_DEBUG "EXT4-fs: someone won our chunk\n");
 			 */
+			lost = (unsigned int)atomic_inc_return(&sbi->s_mb_lost_chunks);
+			mb_debug(sb, "lost chunk, group: %u, start: %d, len: %d, lost: %u\n",
+				 ac->ac_b_ex.fe_group, ac->ac_b_ex.fe_start,
+				 ac->ac_b_ex.fe_len, lost);
+
 			ac->ac_b_ex.fe_group = 0;
 			ac->ac_b_ex.fe_start = 0;
 			ac->ac_b_ex.fe_len = 0;
 			ac->ac_status = AC_STATUS_CONTINUE;
 			ac->ac_flags |= EXT4_MB_HINT_FIRST;
 			cr = 3;
-			atomic_inc(&sbi->s_mb_lost_chunks);
 			goto repeat;
 		}
 	}
Signature · 1 line
1.8.3.1
Reviewed-by on v3 1/1
#1 lore raw
Quoted 4 lines · brookxu

Fix log printing of ext4_mb_regular_allocator(),it may be an unintentional behavior.

Quoted 8 lines

V3:
It may be better to add a comma between start and len, which is convenient for script processing.

V2:
Add more valuable information, such as group, start, len, lost.

A very minor cleanup possible, but not really worthwhile to resubmit if Ted wants to apply it. Either way, it looks good.

Quoted 20 lines

---
fs/ext4/mballoc.c | 9 ++++++---

1 file changed, 6 insertions(+), 3 deletions(-)
fs/ext4/mballoc.c +1 -0
index c0a331e..70b110f 100644
@@ -2218,6 +2218,7 @@ static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac,
 	struct ext4_sb_info *sbi;
 	struct super_block *sb;
 	struct ext4_buddy e4b;
+	unsigned int lost;
 
 	sb = ac->ac_sb;
 	sbi = EXT4_SB(sb);
@@ -2341,22 +2342,24 @@ static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac,
 		 * We've been searching too long. Let's try to allocate
 		 * the best chunk we've found so far
 		 */

fs/ext4/mballoc.c static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac, · line 2352

Quoted 10 lines
fs/ext4/mballoc.c +1 -2
-
 		ext4_mb_try_best_found(ac, &e4b);
 		if (ac->ac_status != AC_STATUS_FOUND) {
 			/*
 			 * Someone more lucky has already allocated it.
 			 * The only thing we can do is just take first
 			 * found block(s)
-			printk(KERN_DEBUG "EXT4-fs: someone won our chunk\n");
 			 */
+			lost = (unsigned int)atomic_inc_return(&sbi->s_mb_lost_chunks);

(minor) no need for a typecast here? The return type is already "int", so it would be better to declare "int lost" and that would make the line a bit shorter.

Quoted 15 lines
fs/ext4/mballoc.c +4 -2
+			mb_debug(sb, "lost chunk, group: %u, start: %d, len: %d, lost: %u\n",
+				 ac->ac_b_ex.fe_group, ac->ac_b_ex.fe_start,
+				 ac->ac_b_ex.fe_len, lost);
+
 			ac->ac_b_ex.fe_group = 0;
 			ac->ac_b_ex.fe_start = 0;
 			ac->ac_b_ex.fe_len = 0;
 			ac->ac_status = AC_STATUS_CONTINUE;
 			ac->ac_flags |= EXT4_MB_HINT_FIRST;
 			cr = 3;
-			atomic_inc(&sbi->s_mb_lost_chunks);
 			goto repeat;
 		}
 	}
--
Quoted 3 lines

1.8.3.1

Cheers, Andreas

Reviewed-by on v3 1/1
#2 lore raw
In reply to brookxu
Quoted 11 lines · brookxu

Fix log printing of ext4_mb_regular_allocator(),it may be an unintentional behavior.

V3:
It may be better to add a comma between start and len, which is convenient for script processing.

V2:
Add more valuable information, such as group, start, len, lost.

LGTM, please feel free to add:

Quoted 45 lines
---
  fs/ext4/mballoc.c | 9 ++++++---
  1 file changed, 6 insertions(+), 3 deletions(-)
fs/ext4/mballoc.c +6 -3
index c0a331e..70b110f 100644
@@ -2218,6 +2218,7 @@ static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac,
  	struct ext4_sb_info *sbi;
  	struct super_block *sb;
  	struct ext4_buddy e4b;
+	unsigned int lost;
 
  	sb = ac->ac_sb;
  	sbi = EXT4_SB(sb);
@@ -2341,22 +2342,24 @@ static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac,
  		 * We've been searching too long. Let's try to allocate
  		 * the best chunk we've found so far
  		 */
-
  		ext4_mb_try_best_found(ac, &e4b);
  		if (ac->ac_status != AC_STATUS_FOUND) {
  			/*
  			 * Someone more lucky has already allocated it.
  			 * The only thing we can do is just take first
  			 * found block(s)
-			printk(KERN_DEBUG "EXT4-fs: someone won our chunk\n");
  			 */
+			lost = (unsigned int)atomic_inc_return(&sbi->s_mb_lost_chunks);
+			mb_debug(sb, "lost chunk, group: %u, start: %d, len: %d, lost: %u\n",
+				 ac->ac_b_ex.fe_group, ac->ac_b_ex.fe_start,
+				 ac->ac_b_ex.fe_len, lost);
+
  			ac->ac_b_ex.fe_group = 0;
  			ac->ac_b_ex.fe_start = 0;
  			ac->ac_b_ex.fe_len = 0;
  			ac->ac_status = AC_STATUS_CONTINUE;
  			ac->ac_flags |= EXT4_MB_HINT_FIRST;
  			cr = 3;
-			atomic_inc(&sbi->s_mb_lost_chunks);
  			goto repeat;
  		}
  	}
maintainer Applied
#3 lore raw
In reply to brookxu
Quoted 2 lines · brookxu

Fix log printing of ext4_mb_regular_allocator(),it may be an unintentional behavior.

Thanks, applied with Andreas's suggested improvement and with a reworded commit description.

- Ted