| Message ID | 20240611051929.513387-6-hch@lst.de (mailing list archive) |
|---|---|
| State | Not Applicable |
| Headers | show |
| Series | [01/26] sd: fix sd_is_zoned | expand |
On 6/11/24 2:19 PM, Christoph Hellwig wrote: > The LOOP_CONFIGURE path automatically upgrades the block size to that > of the underlying file for O_DIRECT file descriptors, but the > LOOP_SET_BLOCK_SIZE path does not. Fix this by lifting the code to > pick the block size into common code. s/lock/block in the commit title. > > Signed-off-by: Christoph Hellwig <hch@lst.de> > --- > drivers/block/loop.c | 25 +++++++++++++++---------- > 1 file changed, 15 insertions(+), 10 deletions(-) > > diff --git a/drivers/block/loop.c b/drivers/block/loop.c > index c658282454af1b..4f6d8514d19bd6 100644 > --- a/drivers/block/loop.c > +++ b/drivers/block/loop.c > @@ -975,10 +975,24 @@ loop_set_status_from_info(struct loop_device *lo, > return 0; > } > > +static unsigned short loop_default_blocksize(struct loop_device *lo, > + struct block_device *backing_bdev) > +{ > + /* In case of direct I/O, match underlying block size */ > + if ((lo->lo_backing_file->f_flags & O_DIRECT) && backing_bdev) > + return bdev_logical_block_size(backing_bdev); > + return 512; Nit: SECTOR_SIZE ? > +} > + > static int loop_reconfigure_limits(struct loop_device *lo, unsigned short bsize) > { > + struct file *file = lo->lo_backing_file; > + struct inode *inode = file->f_mapping->host; > struct queue_limits lim; > > + if (!bsize) > + bsize = loop_default_blocksize(lo, inode->i_sb->s_bdev); If bsize is specified and there is a backing dev used with direct IO, should it be checked that bsize is a multiple of bdev_logical_block_size(backing_bdev) ? > + > lim = queue_limits_start_update(lo->lo_queue); > lim.logical_block_size = bsize; > lim.physical_block_size = bsize; > @@ -997,7 +1011,6 @@ static int loop_configure(struct loop_device *lo, blk_mode_t mode, > int error; > loff_t size; > bool partscan; > - unsigned short bsize; > bool is_loop; > > if (!file) > @@ -1076,15 +1089,7 @@ static int loop_configure(struct loop_device *lo, blk_mode_t mode, > if (!(lo->lo_flags & LO_FLAGS_READ_ONLY) && file->f_op->fsync) > blk_queue_write_cache(lo->lo_queue, true, false); > > - if (config->block_size) > - bsize = config->block_size; > - else if ((lo->lo_backing_file->f_flags & O_DIRECT) && inode->i_sb->s_bdev) > - /* In case of direct I/O, match underlying block size */ > - bsize = bdev_logical_block_size(inode->i_sb->s_bdev); > - else > - bsize = 512; > - > - error = loop_reconfigure_limits(lo, bsize); > + error = loop_reconfigure_limits(lo, config->block_size); > if (WARN_ON_ONCE(error)) > goto out_unlock; >
On Tue, Jun 11, 2024 at 02:56:59PM +0900, Damien Le Moal wrote: > > + if (!bsize) > > + bsize = loop_default_blocksize(lo, inode->i_sb->s_bdev); > > If bsize is specified and there is a backing dev used with direct IO, should it > be checked that bsize is a multiple of bdev_logical_block_size(backing_bdev) ? For direct I/O that check would be useful. For buffered I/O we can do read-modify-write cycles. However this series is already huge and not primarily about improving the loop driver parameter validation, so I'll defer this for now.
On 6/11/24 07:19, Christoph Hellwig wrote: > The LOOP_CONFIGURE path automatically upgrades the block size to that > of the underlying file for O_DIRECT file descriptors, but the > LOOP_SET_BLOCK_SIZE path does not. Fix this by lifting the code to > pick the block size into common code. > > Signed-off-by: Christoph Hellwig <hch@lst.de> > --- > drivers/block/loop.c | 25 +++++++++++++++---------- > 1 file changed, 15 insertions(+), 10 deletions(-) > Reviewed-by: Hannes Reinecke <hare@suse.de> Cheers, Hannes
On 6/10/24 10:19 PM, Christoph Hellwig wrote: > The LOOP_CONFIGURE path automatically upgrades the block size to that > of the underlying file for O_DIRECT file descriptors, but the > LOOP_SET_BLOCK_SIZE path does not. Fix this by lifting the code to > pick the block size into common code. Reviewed-by: Bart Van Assche <bvanassche@acm.org>
diff --git a/drivers/block/loop.c b/drivers/block/loop.c index c658282454af1b..4f6d8514d19bd6 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -975,10 +975,24 @@ loop_set_status_from_info(struct loop_device *lo, return 0; } +static unsigned short loop_default_blocksize(struct loop_device *lo, + struct block_device *backing_bdev) +{ + /* In case of direct I/O, match underlying block size */ + if ((lo->lo_backing_file->f_flags & O_DIRECT) && backing_bdev) + return bdev_logical_block_size(backing_bdev); + return 512; +} + static int loop_reconfigure_limits(struct loop_device *lo, unsigned short bsize) { + struct file *file = lo->lo_backing_file; + struct inode *inode = file->f_mapping->host; struct queue_limits lim; + if (!bsize) + bsize = loop_default_blocksize(lo, inode->i_sb->s_bdev); + lim = queue_limits_start_update(lo->lo_queue); lim.logical_block_size = bsize; lim.physical_block_size = bsize; @@ -997,7 +1011,6 @@ static int loop_configure(struct loop_device *lo, blk_mode_t mode, int error; loff_t size; bool partscan; - unsigned short bsize; bool is_loop; if (!file) @@ -1076,15 +1089,7 @@ static int loop_configure(struct loop_device *lo, blk_mode_t mode, if (!(lo->lo_flags & LO_FLAGS_READ_ONLY) && file->f_op->fsync) blk_queue_write_cache(lo->lo_queue, true, false); - if (config->block_size) - bsize = config->block_size; - else if ((lo->lo_backing_file->f_flags & O_DIRECT) && inode->i_sb->s_bdev) - /* In case of direct I/O, match underlying block size */ - bsize = bdev_logical_block_size(inode->i_sb->s_bdev); - else - bsize = 512; - - error = loop_reconfigure_limits(lo, bsize); + error = loop_reconfigure_limits(lo, config->block_size); if (WARN_ON_ONCE(error)) goto out_unlock;
The LOOP_CONFIGURE path automatically upgrades the block size to that of the underlying file for O_DIRECT file descriptors, but the LOOP_SET_BLOCK_SIZE path does not. Fix this by lifting the code to pick the block size into common code. Signed-off-by: Christoph Hellwig <hch@lst.de> --- drivers/block/loop.c | 25 +++++++++++++++---------- 1 file changed, 15 insertions(+), 10 deletions(-)