diff mbox series

[02/10] block: move discard checks into the ioctl handler

Message ID 20240307151157.466013-3-hch@lst.de (mailing list archive)
State New, archived
Headers show
Series [01/10] block: remove the discard_granularity check in __blkdev_issue_discard | expand

Commit Message

Christoph Hellwig March 7, 2024, 3:11 p.m. UTC
Most bio operations get basic sanity checking in submit_bio and anything
more complicated than that is done in the callers.  Discards are a bit
different from that in that a lot of checking is done in
__blkdev_issue_discard, and the specific errnos for that are returned
to userspace.  Move the checks that require specific errnos to the ioctl
handler instead, and just leave the basic sanity checking in submit_bio
for the other handlers.  This introduces two changes in behavior:

 1) the logical block size alignment check of the start and len is lost
    for non-ioctl callers.
    This matches what is done for other operations including reads and
    writes.  We should probably verify this for all bios, but for now
    make discards match the normal flow.
 2) for non-ioctl callers all errors are reported on I/O completion now
    instead of synchronously.  Callers in general mostly ignore or log
    errors so this will actually simplify the code once cleaned up

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-lib.c | 13 -------------
 block/ioctl.c   | 13 +++++++++----
 2 files changed, 9 insertions(+), 17 deletions(-)

Comments

Dave Chinner March 7, 2024, 9:33 p.m. UTC | #1
On Thu, Mar 07, 2024 at 08:11:49AM -0700, Christoph Hellwig wrote:
> Most bio operations get basic sanity checking in submit_bio and anything
> more complicated than that is done in the callers.  Discards are a bit
> different from that in that a lot of checking is done in
> __blkdev_issue_discard, and the specific errnos for that are returned
> to userspace.  Move the checks that require specific errnos to the ioctl
> handler instead, and just leave the basic sanity checking in submit_bio
> for the other handlers.  This introduces two changes in behavior:
> 
>  1) the logical block size alignment check of the start and len is lost
>     for non-ioctl callers.
>     This matches what is done for other operations including reads and
>     writes.  We should probably verify this for all bios, but for now
>     make discards match the normal flow.
>  2) for non-ioctl callers all errors are reported on I/O completion now
>     instead of synchronously.  Callers in general mostly ignore or log
>     errors so this will actually simplify the code once cleaned up
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

OK.

> ---
>  block/blk-lib.c | 13 -------------
>  block/ioctl.c   | 13 +++++++++----
>  2 files changed, 9 insertions(+), 17 deletions(-)
> 
> diff --git a/block/blk-lib.c b/block/blk-lib.c
> index f873eb9a886f63..50923508a32466 100644
> --- a/block/blk-lib.c
> +++ b/block/blk-lib.c
> @@ -59,19 +59,6 @@ int __blkdev_issue_discard(struct block_device *bdev, sector_t sector,
>  		sector_t nr_sects, gfp_t gfp_mask, struct bio **biop)
>  {
>  	struct bio *bio = *biop;
> -	sector_t bs_mask;
> -
> -	if (bdev_read_only(bdev))
> -		return -EPERM;
> -	if (!bdev_max_discard_sectors(bdev))
> -		return -EOPNOTSUPP;
> -
> -	bs_mask = (bdev_logical_block_size(bdev) >> 9) - 1;
> -	if ((sector | nr_sects) & bs_mask)
> -		return -EINVAL;
> -
> -	if (!nr_sects)
> -		return -EINVAL;
>  
>  	while (nr_sects) {
>  		sector_t req_sects =
> diff --git a/block/ioctl.c b/block/ioctl.c
> index de0cc0d215c633..1d5de0a890c5e8 100644
> --- a/block/ioctl.c
> +++ b/block/ioctl.c
> @@ -95,6 +95,8 @@ static int compat_blkpg_ioctl(struct block_device *bdev,
>  static int blk_ioctl_discard(struct block_device *bdev, blk_mode_t mode,
>  		unsigned long arg)
>  {
> +	sector_t bs_mask = (bdev_logical_block_size(bdev) >> SECTOR_SHIFT) - 1;
> +	sector_t sector, nr_sects;

This changes the alignment checks from a hard coded 512 byte sector
to the logical block size of the device. I don't see a problem with
this (it fixes a bug) but it should at least be mentioned in the
commit message.

-Dave.
Christoph Hellwig March 8, 2024, 3:22 p.m. UTC | #2
On Fri, Mar 08, 2024 at 08:33:08AM +1100, Dave Chinner wrote:
> >  static int blk_ioctl_discard(struct block_device *bdev, blk_mode_t mode,
> >  		unsigned long arg)
> >  {
> > +	sector_t bs_mask = (bdev_logical_block_size(bdev) >> SECTOR_SHIFT) - 1;
> > +	sector_t sector, nr_sects;
> 
> This changes the alignment checks from a hard coded 512 byte sector
> to the logical block size of the device. I don't see a problem with
> this (it fixes a bug) but it should at least be mentioned in the
> commit message.

Before the exact block size alignment check as done down in
__blkdev_issue_discard, it just moves up here now.  I guess I need to
make that more clear in the commit message.
Dave Chinner March 8, 2024, 9:16 p.m. UTC | #3
On Fri, Mar 08, 2024 at 04:22:44PM +0100, Christoph Hellwig wrote:
> On Fri, Mar 08, 2024 at 08:33:08AM +1100, Dave Chinner wrote:
> > >  static int blk_ioctl_discard(struct block_device *bdev, blk_mode_t mode,
> > >  		unsigned long arg)
> > >  {
> > > +	sector_t bs_mask = (bdev_logical_block_size(bdev) >> SECTOR_SHIFT) - 1;
> > > +	sector_t sector, nr_sects;
> > 
> > This changes the alignment checks from a hard coded 512 byte sector
> > to the logical block size of the device. I don't see a problem with
> > this (it fixes a bug) but it should at least be mentioned in the
> > commit message.
> 
> Before the exact block size alignment check as done down in
> __blkdev_issue_discard, it just moves up here now.  I guess I need to
> make that more clear in the commit message.

Ah, eyeball pattern matching fail on my part - you changed it from a
hard coded '9' to SECTOR_SHIFT as it moved (which is fine!), I just
missed that. All good!

-Dave.
diff mbox series

Patch

diff --git a/block/blk-lib.c b/block/blk-lib.c
index f873eb9a886f63..50923508a32466 100644
--- a/block/blk-lib.c
+++ b/block/blk-lib.c
@@ -59,19 +59,6 @@  int __blkdev_issue_discard(struct block_device *bdev, sector_t sector,
 		sector_t nr_sects, gfp_t gfp_mask, struct bio **biop)
 {
 	struct bio *bio = *biop;
-	sector_t bs_mask;
-
-	if (bdev_read_only(bdev))
-		return -EPERM;
-	if (!bdev_max_discard_sectors(bdev))
-		return -EOPNOTSUPP;
-
-	bs_mask = (bdev_logical_block_size(bdev) >> 9) - 1;
-	if ((sector | nr_sects) & bs_mask)
-		return -EINVAL;
-
-	if (!nr_sects)
-		return -EINVAL;
 
 	while (nr_sects) {
 		sector_t req_sects =
diff --git a/block/ioctl.c b/block/ioctl.c
index de0cc0d215c633..1d5de0a890c5e8 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -95,6 +95,8 @@  static int compat_blkpg_ioctl(struct block_device *bdev,
 static int blk_ioctl_discard(struct block_device *bdev, blk_mode_t mode,
 		unsigned long arg)
 {
+	sector_t bs_mask = (bdev_logical_block_size(bdev) >> SECTOR_SHIFT) - 1;
+	sector_t sector, nr_sects;
 	uint64_t range[2];
 	uint64_t start, len;
 	struct inode *inode = bdev->bd_inode;
@@ -105,18 +107,21 @@  static int blk_ioctl_discard(struct block_device *bdev, blk_mode_t mode,
 
 	if (!bdev_max_discard_sectors(bdev))
 		return -EOPNOTSUPP;
+	if (bdev_read_only(bdev))
+		return -EPERM;
 
 	if (copy_from_user(range, (void __user *)arg, sizeof(range)))
 		return -EFAULT;
 
 	start = range[0];
 	len = range[1];
+	sector = start >> SECTOR_SHIFT;
+	nr_sects = len >> SECTOR_SHIFT;
 
-	if (start & 511)
+	if (!nr_sects)
 		return -EINVAL;
-	if (len & 511)
+	if ((sector | nr_sects) & bs_mask)
 		return -EINVAL;
-
 	if (start + len > bdev_nr_bytes(bdev))
 		return -EINVAL;
 
@@ -124,7 +129,7 @@  static int blk_ioctl_discard(struct block_device *bdev, blk_mode_t mode,
 	err = truncate_bdev_range(bdev, mode, start, start + len - 1);
 	if (err)
 		goto fail;
-	err = blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_KERNEL);
+	err = blkdev_issue_discard(bdev, sector, nr_sects, GFP_KERNEL);
 fail:
 	filemap_invalidate_unlock(inode->i_mapping);
 	return err;