linux-kernel - Re: [PATCH] block: Fix BLKRRPART regression

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <tnqg4la2bhbhfbty3aa74uorkfhz76v5sntd3md44lfctjhjb7@7qbx5z2o7hzm>
Date: Wed, 10 Apr 2024 05:04:15 +0000
From: Shinichiro Kawasaki <shinichiro.kawasaki@....com>
To: Saranya Muruganandam <saranyamohan@...gle.com>
CC: "hch@....de" <hch@....de>, "axboe@...nel.dk" <axboe@...nel.dk>,
	"linux-block@...r.kernel.org" <linux-block@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"stable@...r.kernel.org" <stable@...r.kernel.org>, "tj@...nel.org"
	<tj@...nel.org>, "yukuai1@...weicloud.com" <yukuai1@...weicloud.com>
Subject: Re: [PATCH] block: Fix BLKRRPART regression

Saranya, while I evaluate the corresponding patch for blktests, I found a
weird behavior, and had a chance to look into this patch. Please find my
comments in line.

On Apr 06, 2024 / 04:46, Saranya Muruganandam wrote:
> The BLKRRPART ioctl used to report errors such as EIO before we changed
> the blkdev_reread_part() logic.
> 
> Lets add a flag and capture the errors returned by bdev_disk_changed()
> when the flag is set. Setting this flag for the BLKRRPART path when we
> want the errors to be reported when rereading partitions on the disk.

Documentation/process/submitting-patches.rst guides to describe changes in
imperative mood, then the sentences above can be rephrased like,

   Add a flag...
   Set this flag...

> 
> Link: https://lore.kernel.org/all/20240320015134.GA14267@lst.de/
> Suggested-by: Christoph Hellwig <hch@....de>
> Tested: Tested by simulating failure to the block device and will
> propose a new test to blktests.
> Fixes: 4601b4b130de ("block: reopen the device in blkdev_reread_part")
> Reported-by: Saranya Muruganandam <saranyamohan@...gle.com>
> Signed-off-by: Saranya Muruganandam <saranyamohan@...gle.com>
> 
> Change-Id: Idf3d97390ed78061556f8468d10d6cab24ae20b1
> ---
>  block/bdev.c           | 29 +++++++++++++++++++----------
>  block/ioctl.c          |  3 ++-
>  include/linux/blkdev.h |  2 ++
>  3 files changed, 23 insertions(+), 11 deletions(-)
> 
> diff --git a/block/bdev.c b/block/bdev.c
> index 7a5f611c3d2e3..42940bced33bb 100644
> --- a/block/bdev.c
> +++ b/block/bdev.c
> @@ -652,6 +652,14 @@ static void blkdev_flush_mapping(struct block_device *bdev)
>  	bdev_write_inode(bdev);
>  }
>  
> +static void blkdev_put_whole(struct block_device *bdev)
> +{
> +	if (atomic_dec_and_test(&bdev->bd_openers))
> +		blkdev_flush_mapping(bdev);
> +	if (bdev->bd_disk->fops->release)
> +		bdev->bd_disk->fops->release(bdev->bd_disk);
> +}
> +
>  static int blkdev_get_whole(struct block_device *bdev, blk_mode_t mode)
>  {
>  	struct gendisk *disk = bdev->bd_disk;
> @@ -670,20 +678,21 @@ static int blkdev_get_whole(struct block_device *bdev, blk_mode_t mode)
>  
>  	if (!atomic_read(&bdev->bd_openers))
>  		set_init_blocksize(bdev);
> -	if (test_bit(GD_NEED_PART_SCAN, &disk->state))
> -		bdev_disk_changed(disk, false);
>  	atomic_inc(&bdev->bd_openers);
> +	if (test_bit(GD_NEED_PART_SCAN, &disk->state)) {
> +		/*
> +		 * Only return scanning errors if we are called from conexts

s/conexts/contexts/

> +		 * that explicitly want them, e.g. the BLKRRPART ioctl.
> +		 */
> +		ret = bdev_disk_changed(disk, false);
> +		if (ret && (mode & BLK_OPEN_STRICT_SCAN)) {
> +			blkdev_put_whole(bdev);
> +			return ret;
> +		}
> +	}
>  	return 0;
>  }
>  
> -static void blkdev_put_whole(struct block_device *bdev)
> -{
> -	if (atomic_dec_and_test(&bdev->bd_openers))
> -		blkdev_flush_mapping(bdev);
> -	if (bdev->bd_disk->fops->release)
> -		bdev->bd_disk->fops->release(bdev->bd_disk);
> -}
> -
>  static int blkdev_get_part(struct block_device *part, blk_mode_t mode)
>  {
>  	struct gendisk *disk = part->bd_disk;
> diff --git a/block/ioctl.c b/block/ioctl.c
> index 0c76137adcaaa..128f503828cee 100644
> --- a/block/ioctl.c
> +++ b/block/ioctl.c
> @@ -562,7 +562,8 @@ static int blkdev_common_ioctl(struct block_device *bdev, blk_mode_t mode,
>  			return -EACCES;
>  		if (bdev_is_partition(bdev))
>  			return -EINVAL;
> -		return disk_scan_partitions(bdev->bd_disk, mode);
> +		return disk_scan_partitions(bdev->bd_disk,
> +				mode | BLK_OPEN_STRICT_SCAN);
>  	case BLKTRACESTART:
>  	case BLKTRACESTOP:
>  	case BLKTRACETEARDOWN:
> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
> index c3e8f7cf96be9..3ed5e03109c29 100644
> --- a/include/linux/blkdev.h
> +++ b/include/linux/blkdev.h
> @@ -128,6 +128,8 @@ typedef unsigned int __bitwise blk_mode_t;
>  #define BLK_OPEN_WRITE_IOCTL	((__force blk_mode_t)(1 << 4))
>  /* open is exclusive wrt all other BLK_OPEN_WRITE opens to the device */
>  #define BLK_OPEN_RESTRICT_WRITES	((__force blk_mode_t)(1 << 5))
> +/* return partition scanning errors */
> +#define BLK_OPEN_STRICT_SCAN	((__force blk_mode_t)(1 << 5))

The line above assigns the same number for BLK_OPEN_STRICT_SCAN and
BLK_OPEN_RESTRICT_WRITES, then blockdev --rereadpt fails with EINVAL,
not EIO. I modified it to "1 << 6", then it looks working good.

>  
>  struct gendisk {
>  	/*
> -- 
> 2.44.0.478.gd926399ef9-goog
>