[PATCH] md/raid1:Fix bug about fixing read errors.

[PATCH] md/raid1:Fix bug about fixing read errors.

[majianpeng]

From b5b091c186efcd32d191206181590be76a393cda Mon Sep 17 00:00:00 2001
From: majianpeng <majianpeng@gmail.com>
Date: Wed, 11 Apr 2012 19:57:01 +0800
Subject: [PATCH] md/raid1:Fix bug about fixing read errors.
 Add spare disk  which working when fix read errors.And if it can't read
 corrected data,it should make all disks badsectors.And no
 need to read,because this no disk can read.


Signed-off-by: majianpeng <majianpeng@gmail.com>
---
 drivers/md/raid1.c |   38 +++++++++++++++++++++++++++++---------
 1 files changed, 29 insertions(+), 9 deletions(-)

diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index d35e4c9..67ec686 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -1833,7 +1833,7 @@ static void sync_request_write(struct mddev *mddev, struct r1bio *r1_bio)
  *	3.	Performs writes following reads for array synchronising.
  */
 
-static void fix_read_error(struct r1conf *conf, int read_disk,
+static int  fix_read_error(struct r1conf *conf, int read_disk,
 			   sector_t sect, int sectors)
 {
 	struct mddev *mddev = conf->mddev;
@@ -1858,7 +1858,9 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
 
 			rdev = conf->mirrors[d].rdev;
 			if (rdev &&
-			    test_bit(In_sync, &rdev->flags) &&
+			    (test_bit(In_sync, &rdev->flags) ||
+			     (!test_bit(Faulty, &rdev->flags) &&
+			      (rdev->recovery_offset >= sect + s))) &&
 			    is_badblock(rdev, sect, s,
 					&first_bad, &bad_sectors) == 0 &&
 			    sync_page_io(rdev, sect, s<<9,
@@ -1873,10 +1875,15 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
 
 		if (!success) {
 			/* Cannot read from anywhere - mark it bad */
-			struct md_rdev *rdev = conf->mirrors[read_disk].rdev;
-			if (!rdev_set_badblocks(rdev, sect, s, 0))
-				md_error(mddev, rdev);
-			break;
+			struct md_rdev *rdev;
+			for (d = 0; d < conf->raid_disks * 2; d++) {
+				rdev = conf->mirrors[d].rdev;
+				if (!rdev)
+					continue;
+				else if (!rdev_set_badblocks(rdev, sect, s, 0))
+					md_error(mddev, rdev);
+			}
+			return 1;
 		}
 		/* write it back and re-read */
 		start = d;
@@ -1915,6 +1922,7 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
 		sectors -= s;
 		sect += s;
 	}
+	return 0;
 }
 
 static void bi_complete(struct bio *bio, int error)
@@ -2072,8 +2080,13 @@ static void handle_read_error(struct r1conf *conf, struct r1bio *r1_bio)
 	struct bio *bio;
 	char b[BDEVNAME_SIZE];
 	struct md_rdev *rdev;
+	int ret;
 
 	clear_bit(R1BIO_ReadError, &r1_bio->state);
+
+	bio = r1_bio->bios[r1_bio->read_disk];
+	bdevname(bio->bi_bdev, b);
+
 	/* we got a read error. Maybe the drive is bad.  Maybe just
 	 * the block and we can fix it.
 	 * We freeze all other IO, and try reading the block from
@@ -2084,14 +2097,21 @@ static void handle_read_error(struct r1conf *conf, struct r1bio *r1_bio)
 	 */
 	if (mddev->ro == 0) {
 		freeze_array(conf);
-		fix_read_error(conf, r1_bio->read_disk,
+		ret = fix_read_error(conf, r1_bio->read_disk,
 			       r1_bio->sector, r1_bio->sectors);
 		unfreeze_array(conf);
+		/*no need read more */
+		if (ret) {
+			printk(KERN_ALERT "md/raid1:%s: %s: unrecoverable I/O"
+			      " read error for block %llu\n", mdname(mddev),
+			      b, (unsigned long long)r1_bio->sector);
+			raid_end_bio_io(r1_bio);
+			return;
+
+		}
 	} else
 		md_error(mddev, conf->mirrors[r1_bio->read_disk].rdev);
 
-	bio = r1_bio->bios[r1_bio->read_disk];
-	bdevname(bio->bi_bdev, b);
 read_more:
 	disk = read_balance(conf, r1_bio, &max_sectors);
 	if (disk == -1) {
-- 
1.7.5.4

Re: [PATCH] md/raid1:Fix bug about fixing read errors.

[NeilBrown]

[-- Attachment #1: Type: text/plain, Size: 4817 bytes --]

On Wed, 11 Apr 2012 20:08:31 +0800 "majianpeng" <majianpeng@gmail.com> wrote:

> >From b5b091c186efcd32d191206181590be76a393cda Mon Sep 17 00:00:00 2001
> From: majianpeng <majianpeng@gmail.com>
> Date: Wed, 11 Apr 2012 19:57:01 +0800
> Subject: [PATCH] md/raid1:Fix bug about fixing read errors.
>  Add spare disk  which working when fix read errors.And if it can't read
>  corrected data,it should make all disks badsectors.And no
>  need to read,because this no disk can read.

Thanks for the patch.

However it seems to address multiple issues, and so should be multiple
patches.

Firstly, it allows fix_read_error to read from a disk that is being recovered.
It seems unlikely that this will ever be necessary, but it is theoretically
possible so I am happy with the patch.
I have applied a patch making just this change.


Secondly it records a bad block on every device if it cannot read from
anywhere.
I don't think this is necessary.  fix_read_error should only be addressing
the one read error.  It either fixes it or marks it as bad.
If there are other bad blocks on other disks they will be found and handled
eventually and adding extra code in here  just makes it more complex with
little gain.

So unless you can convince me that it is actually behaving wrongly, I won't
be applying the rest of the patch.

Thanks,

NeilBrown


> 
> 
> Signed-off-by: majianpeng <majianpeng@gmail.com>
> ---
>  drivers/md/raid1.c |   38 +++++++++++++++++++++++++++++---------
>  1 files changed, 29 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
> index d35e4c9..67ec686 100644
> --- a/drivers/md/raid1.c
> +++ b/drivers/md/raid1.c
> @@ -1833,7 +1833,7 @@ static void sync_request_write(struct mddev *mddev, struct r1bio *r1_bio)
>   *	3.	Performs writes following reads for array synchronising.
>   */
>  
> -static void fix_read_error(struct r1conf *conf, int read_disk,
> +static int  fix_read_error(struct r1conf *conf, int read_disk,
>  			   sector_t sect, int sectors)
>  {
>  	struct mddev *mddev = conf->mddev;
> @@ -1858,7 +1858,9 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
>  
>  			rdev = conf->mirrors[d].rdev;
>  			if (rdev &&
> -			    test_bit(In_sync, &rdev->flags) &&
> +			    (test_bit(In_sync, &rdev->flags) ||
> +			     (!test_bit(Faulty, &rdev->flags) &&
> +			      (rdev->recovery_offset >= sect + s))) &&
>  			    is_badblock(rdev, sect, s,
>  					&first_bad, &bad_sectors) == 0 &&
>  			    sync_page_io(rdev, sect, s<<9,
> @@ -1873,10 +1875,15 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
>  
>  		if (!success) {
>  			/* Cannot read from anywhere - mark it bad */
> -			struct md_rdev *rdev = conf->mirrors[read_disk].rdev;
> -			if (!rdev_set_badblocks(rdev, sect, s, 0))
> -				md_error(mddev, rdev);
> -			break;
> +			struct md_rdev *rdev;
> +			for (d = 0; d < conf->raid_disks * 2; d++) {
> +				rdev = conf->mirrors[d].rdev;
> +				if (!rdev)
> +					continue;
> +				else if (!rdev_set_badblocks(rdev, sect, s, 0))
> +					md_error(mddev, rdev);
> +			}
> +			return 1;
>  		}
>  		/* write it back and re-read */
>  		start = d;
> @@ -1915,6 +1922,7 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
>  		sectors -= s;
>  		sect += s;
>  	}
> +	return 0;
>  }
>  
>  static void bi_complete(struct bio *bio, int error)
> @@ -2072,8 +2080,13 @@ static void handle_read_error(struct r1conf *conf, struct r1bio *r1_bio)
>  	struct bio *bio;
>  	char b[BDEVNAME_SIZE];
>  	struct md_rdev *rdev;
> +	int ret;
>  
>  	clear_bit(R1BIO_ReadError, &r1_bio->state);
> +
> +	bio = r1_bio->bios[r1_bio->read_disk];
> +	bdevname(bio->bi_bdev, b);
> +
>  	/* we got a read error. Maybe the drive is bad.  Maybe just
>  	 * the block and we can fix it.
>  	 * We freeze all other IO, and try reading the block from
> @@ -2084,14 +2097,21 @@ static void handle_read_error(struct r1conf *conf, struct r1bio *r1_bio)
>  	 */
>  	if (mddev->ro == 0) {
>  		freeze_array(conf);
> -		fix_read_error(conf, r1_bio->read_disk,
> +		ret = fix_read_error(conf, r1_bio->read_disk,
>  			       r1_bio->sector, r1_bio->sectors);
>  		unfreeze_array(conf);
> +		/*no need read more */
> +		if (ret) {
> +			printk(KERN_ALERT "md/raid1:%s: %s: unrecoverable I/O"
> +			      " read error for block %llu\n", mdname(mddev),
> +			      b, (unsigned long long)r1_bio->sector);
> +			raid_end_bio_io(r1_bio);
> +			return;
> +
> +		}
>  	} else
>  		md_error(mddev, conf->mirrors[r1_bio->read_disk].rdev);
>  
> -	bio = r1_bio->bios[r1_bio->read_disk];
> -	bdevname(bio->bi_bdev, b);
>  read_more:
>  	disk = read_balance(conf, r1_bio, &max_sectors);
>  	if (disk == -1) {


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

Re: Re: [PATCH] md/raid1:Fix bug about fixing read errors.

[kedacomkernel]

Hi:

>>  However it seems to address multiple issues, and so should be multiple
>>patches.

>>Firstly, it allows fix_read_error to read from a disk that is being recovered.
>>It seems unlikely that this will ever be necessary, but it is theoretically
>>possible so I am happy with the patch.
>>I have applied a patch making just this change.
 This idea from the read_balance. 

>>Secondly it records a bad block on every device if it cannot read from
>>anywhere.
>>I don't think this is necessary.  fix_read_error should only be addressing
>>the one read error.  It either fixes it or marks it as bad.
>>If there are other bad blocks on other disks they will be found and handled
>>eventually and adding extra code in here  just makes it more complex with
>>little gain.
In this situation
1:no channce to recover
2: performance,why judge more read?
3:like the action  in function fix_sync_read_error().

>>So unless you can convince me that it is actually behaving wrongly, I won't
>>be applying the rest of the patch.