Skip to contentSkip to conversation

badblocks: Fix a nonsense WARN_ON() which checks whether a u64 variable < 0

PATCH Applied [email protected] 4 participants 8 messages Mar 9, 2025 – Mar 10, 2025 Block layer

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 badblocks: Fix a nonsense WARN_ON() which checks whether a u64 variable < 0

From: Coly Li <[email protected]>

In _badblocks_check(), there are lines of code like this,
1246 sectors -= len;
[snipped]
1251 WARN_ON(sectors < 0);

The WARN_ON() at line 1257 doesn't make sense because sectors is unsigned long long type and never to be <0.

Fix it by checking directly checking whether sectors is less than len.

---
 block/badblocks.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
block/badblocks.c +3 -2
index 673ef068423a..ece64e76fe8f 100644
@@ -1242,14 +1242,15 @@ static int _badblocks_check(struct badblocks *bb, sector_t s, sector_t sectors,
 	len = sectors;
 
 update_sectors:
+	/* This situation should never happen */
+	WARN_ON(sectors < len);
+
 	s += len;
 	sectors -= len;
 
 	if (sectors > 0)
 		goto re_check;
 
-	WARN_ON(sectors < 0);
-
 	if (unacked_badblocks > 0)
 		rv = -1;
 	else if (acked_badblocks > 0)
Signature · 1 line
2.47.2
#1 lore raw

PATCH 1/1 badblocks: Fix a nonsense WARN_ON() which checks whether a u64 variable < 0

Hi Jens,

Could you please take a look at it and pick this patch into the for-6.15/block branch? The patch is generated based on the for-6.15/block branch.

Thanks in advance.

Coly Li

Quoted 43 lines

From: Coly Li <[email protected]>

In _badblocks_check(), there are lines of code like this,
1246 sectors -= len;
[snipped]
1251 WARN_ON(sectors < 0);

The WARN_ON() at line 1257 doesn't make sense because sectors is unsigned long long type and never to be <0.

Fix it by checking directly checking whether sectors is less than len.

---
block/badblocks.c | 5 +++--

1 file changed, 3 insertions(+), 2 deletions(-)
block/badblocks.c +0 -0
index 673ef068423a..ece64e76fe8f 100644
@@ -1242,14 +1242,15 @@ static int _badblocks_check(struct badblocks *bb, sector_t s, sector_t sectors,
 	len = sectors;
 

update_sectors:

block/badblocks.c +3 -0
+	/* This situation should never happen */
+	WARN_ON(sectors < len);
+
	s += len;
	sectors -= len;

	if (sectors > 0)
		goto re_check;
block/badblocks.c +0 -2
-	WARN_ON(sectors < 0);
-

if (unacked_badblocks > 0)
rv = -1;
else if (acked_badblocks > 0)

Signature · 1 line
2.47.2
maintainer
#2 lore raw
Quoted 3 lines · Coly Li

Hi Jens,

Could you please take a look at it and pick this patch into the for-6.15/block branch? The patch is generated based on the for-6.15/block branch.

Just a heads-up - you don't need to send these emails outside of just sending the patch, I do get the patches. If I didn't, then that'd be a problem. If you feel patches need extra context, then just do a cover letter for them.

Signature · 1 line
Jens Axboe
#3 lore raw
Quoted 5 lines

2025年3月10日 21:42,Jens Axboe <[email protected]> 写道:

On 3/9/25 10:12 AM, Coly Li wrote:

Quoted 8 lines
Quoted 3 lines

Hi Jens,

Could you please take a look at it and pick this patch into the for-6.15/block branch? The patch is generated based on the for-6.15/block branch.

Just a heads-up - you don't need to send these emails outside of just sending the patch, I do get the patches. If I didn't, then that'd be a problem. If you feel patches need extra context, then just do a cover letter for them.

So if you are the receiver of the patch email, then I don’t need to worry that you will treat it as a normal patch for review. Can I take this as a rule?

Thanks.

Coly Li

maintainer
#4 lore raw
Quoted 9 lines · Coly Li
Quoted 9 lines

2025?3?10? 21:42?Jens Axboe <[email protected]> ???

Quoted 3 lines · Coly Li

Hi Jens,

Could you please take a look at it and pick this patch into the for-6.15/block branch? The patch is generated based on the for-6.15/block branch.

Quoted 8 lines
Quoted 4 lines

Just a heads-up - you don't need to send these emails outside of just sending the patch, I do get the patches. If I didn't, then that'd be a problem. If you feel patches need extra context, then just do a cover letter for them.

So if you are the receiver of the patch email, then I don?t need to worry that you will treat it as a normal patch for review. Can I take this as a rule?

Yes of course - I don't think I've ever seen anyone else send out a patch with a followup a few minutes later to apply it. Just send out the patch, and it should get applied. If it doesn't after a week or whatever, then feel free to send a reminder. But a reminder to apply it a few min after the original patch is a bit unusual and odd.

Signature · 1 line
Jens Axboe
#5 lore raw
Quoted 23 lines

2025年3月10日 21:49,Jens Axboe <[email protected]> 写道:

Quoted 9 lines · Coly Li
Quoted 9 lines

2025?3?10? 21:42?Jens Axboe <[email protected]> ???

Quoted 3 lines · Coly Li

Hi Jens,

Could you please take a look at it and pick this patch into the for-6.15/block branch? The patch is generated based on the for-6.15/block branch.

Quoted 8 lines
Quoted 4 lines

Just a heads-up - you don't need to send these emails outside of just sending the patch, I do get the patches. If I didn't, then that'd be a problem. If you feel patches need extra context, then just do a cover letter for them.

So if you are the receiver of the patch email, then I don?t need to worry that you will treat it as a normal patch for review. Can I take this as a rule?

Quoted 5 lines

Yes of course - I don't think I've ever seen anyone else send out a patch with a followup a few minutes later to apply it. Just send out the patch, and it should get applied. If it doesn't after a week or whatever, then feel free to send a reminder. But a reminder to apply it a few min after the original patch is a bit unusual and odd.

Copied. So normally I will not send a following email to explain the extra information. If it is necessary, I will compose a cover letter.

Thanks for the notice.

Coly Li

Reviewed-by on v1 1/1
#6 lore raw
In reply to [email protected]
Quoted 10 lines · 在 2025/03/10 0:05, [email protected] 写道:

From: Coly Li <[email protected]>

In _badblocks_check(), there are lines of code like this,
1246 sectors -= len;
[snipped]
1251 WARN_ON(sectors < 0);

The WARN_ON() at line 1257 doesn't make sense because sectors is unsigned long long type and never to be <0.

Quoted 8 lines

Fix it by checking directly checking whether sectors is less than len.

---
  block/badblocks.c | 5 +++--
  1 file changed, 3 insertions(+), 2 deletions(-)

LGTM

Quoted 23 lines
block/badblocks.c +3 -2
index 673ef068423a..ece64e76fe8f 100644
@@ -1242,14 +1242,15 @@ static int _badblocks_check(struct badblocks *bb, sector_t s, sector_t sectors,
  	len = sectors;
 
  update_sectors:
+	/* This situation should never happen */
+	WARN_ON(sectors < len);
+
  	s += len;
  	sectors -= len;
 
  	if (sectors > 0)
  		goto re_check;
 
-	WARN_ON(sectors < 0);
-
  	if (unacked_badblocks > 0)
  		rv = -1;
  	else if (acked_badblocks > 0)
maintainer Applied
#7 lore raw
In reply to [email protected]
Quoted 9 lines · On Sun, 09 Mar 2025 12:05:56 -0400, [email protected] wrote:

In _badblocks_check(), there are lines of code like this,
1246 sectors -= len;
[snipped]
1251 WARN_ON(sectors < 0);

The WARN_ON() at line 1257 doesn't make sense because sectors is unsigned long long type and never to be <0.

[...]

Applied, thanks!

[1/1] badblocks: Fix a nonsense WARN_ON() which checks whether a u64 variable < 0

      commit: 7e76336e14de9a2b67af96012ddd46c5676cf340

Best regards,

Signature · 1 line
Jens Axboe