[PATCHv3 0/4] loop: enable different physical blocksizes

[PATCHv3 0/4] loop: enable different physical blocksizes

[Hannes Reinecke]

Currently the loop driver just simulates 512-byte blocks. When
creating bootable images for virtual machines it might be required
to use a different physical blocksize (eg 4k for S/390 DASD), as
the some bootloaders (like lilo or zipl for S/390) need to know
the physical block addresses of the kernel and initrd.

This patchset extends the current LOOP_SET_STATUS64 ioctl to
set the logical and physical blocksize by re-using the existing
'init' fields, which are currently unused.

As usual, comments and reviews are welcome.

Changes to v1:
- Move LO_FLAGS_BLOCKSIZE definition
- Reshuffle patches
Changes to v2:
- Include reviews from Ming Lei

Hannes Reinecke (4):
  loop: Remove unused 'bdev' argument from loop_set_capacity
  loop: Enable correct physical blocksize
  loop: Add 'lo_logical_blocksize'
  loop: Pass logical blocksize in 'lo_init[0]' ioctl field

 drivers/block/loop.c      | 40 +++++++++++++++++++++++++++++++++-------
 drivers/block/loop.h      |  1 +
 include/uapi/linux/loop.h |  1 +
 3 files changed, 35 insertions(+), 7 deletions(-)

-- 
2.6.6

[PATCH 1/4] loop: Remove unused 'bdev' argument from loop_set_capacity

[Hannes Reinecke]

Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/loop.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index fa1b7a9..3008dec 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -1298,7 +1298,7 @@ loop_get_status64(struct loop_device *lo, struct loop_info64 __user *arg) {
 	return err;
 }
 
-static int loop_set_capacity(struct loop_device *lo, struct block_device *bdev)
+static int loop_set_capacity(struct loop_device *lo)
 {
 	if (unlikely(lo->lo_state != Lo_bound))
 		return -ENXIO;
@@ -1361,7 +1361,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode,
 	case LOOP_SET_CAPACITY:
 		err = -EPERM;
 		if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
-			err = loop_set_capacity(lo, bdev);
+			err = loop_set_capacity(lo);
 		break;
 	case LOOP_SET_DIRECT_IO:
 		err = -EPERM;
-- 
2.6.6

[PATCH 2/4] loop: Enable correct physical blocksize

[Hannes Reinecke]

When running on files the physical blocksize is actually 4k,
so we should be announcing it as such. This is enabled with
a new LO_FLAGS_BLOCKSIZE flag value to the existing
loop_set_status ioctl.

Signed-off-by: Hannes Reinecke <hare@suse.de>
---
 drivers/block/loop.c      | 9 ++++++++-
 include/uapi/linux/loop.h | 1 +
 2 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 3008dec..ff5c874 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -233,6 +233,8 @@ figure_loop_size(struct loop_device *lo, loff_t offset, loff_t sizelimit)
 		lo->lo_offset = offset;
 	if (lo->lo_sizelimit != sizelimit)
 		lo->lo_sizelimit = sizelimit;
+	if (lo->lo_flags & LO_FLAGS_BLOCKSIZE)
+		blk_queue_physical_block_size(lo->lo_queue, lo->lo_blocksize);
 	set_capacity(lo->lo_disk, x);
 	bd_set_size(bdev, (loff_t)get_capacity(bdev->bd_disk) << 9);
 	/* let user-space know about the new size */
@@ -1087,6 +1089,7 @@ loop_set_status(struct loop_device *lo, const struct loop_info64 *info)
 	int err;
 	struct loop_func_table *xfer;
 	kuid_t uid = current_uid();
+	int lo_flags = lo->lo_flags;
 
 	if (lo->lo_encrypt_key_size &&
 	    !uid_eq(lo->lo_key_owner, uid) &&
@@ -1116,8 +1119,12 @@ loop_set_status(struct loop_device *lo, const struct loop_info64 *info)
 	if (err)
 		return err;
 
+	if (info->lo_flags & LO_FLAGS_BLOCKSIZE)
+		lo->lo_flags |= LO_FLAGS_BLOCKSIZE;
+
 	if (lo->lo_offset != info->lo_offset ||
-	    lo->lo_sizelimit != info->lo_sizelimit)
+	    lo->lo_sizelimit != info->lo_sizelimit ||
+	    lo->lo_flags != lo_flags)
 		if (figure_loop_size(lo, info->lo_offset, info->lo_sizelimit))
 			return -EFBIG;
 
diff --git a/include/uapi/linux/loop.h b/include/uapi/linux/loop.h
index c8125ec..2691c1c 100644
--- a/include/uapi/linux/loop.h
+++ b/include/uapi/linux/loop.h
@@ -22,6 +22,7 @@ enum {
 	LO_FLAGS_AUTOCLEAR	= 4,
 	LO_FLAGS_PARTSCAN	= 8,
 	LO_FLAGS_DIRECT_IO	= 16,
+	LO_FLAGS_BLOCKSIZE	= 32,
 };
 
 #include <asm/posix_types.h>	/* for __kernel_old_dev_t */
-- 
2.6.6

[PATCH 3/4] loop: Add 'lo_logical_blocksize'

[Hannes Reinecke]

Add a new field 'lo_logical_blocksize' to hold the logical
blocksize of the loop device.

Signed-off-by: Hannes Reinecke <hare@suse.de>
---
 drivers/block/loop.c | 9 +++++++--
 drivers/block/loop.h | 1 +
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index ff5c874..a863667 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -233,8 +233,11 @@ figure_loop_size(struct loop_device *lo, loff_t offset, loff_t sizelimit)
 		lo->lo_offset = offset;
 	if (lo->lo_sizelimit != sizelimit)
 		lo->lo_sizelimit = sizelimit;
-	if (lo->lo_flags & LO_FLAGS_BLOCKSIZE)
+	if (lo->lo_flags & LO_FLAGS_BLOCKSIZE) {
 		blk_queue_physical_block_size(lo->lo_queue, lo->lo_blocksize);
+		blk_queue_logical_block_size(lo->lo_queue,
+					     lo->lo_logical_blocksize);
+	}
 	set_capacity(lo->lo_disk, x);
 	bd_set_size(bdev, (loff_t)get_capacity(bdev->bd_disk) << 9);
 	/* let user-space know about the new size */
@@ -816,6 +819,7 @@ static void loop_config_discard(struct loop_device *lo)
 	struct file *file = lo->lo_backing_file;
 	struct inode *inode = file->f_mapping->host;
 	struct request_queue *q = lo->lo_queue;
+	int lo_bits = blksize_bits(lo->lo_logical_blocksize);
 
 	/*
 	 * We use punch hole to reclaim the free space used by the
@@ -835,7 +839,7 @@ static void loop_config_discard(struct loop_device *lo)
 
 	q->limits.discard_granularity = inode->i_sb->s_blocksize;
 	q->limits.discard_alignment = 0;
-	blk_queue_max_discard_sectors(q, UINT_MAX >> 9);
+	blk_queue_max_discard_sectors(q, UINT_MAX >> lo_bits);
 	q->limits.discard_zeroes_data = 1;
 	queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
 }
@@ -924,6 +928,7 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
 
 	lo->use_dio = false;
 	lo->lo_blocksize = lo_blocksize;
+	lo->lo_logical_blocksize = 512;
 	lo->lo_device = bdev;
 	lo->lo_flags = lo_flags;
 	lo->lo_backing_file = file;
diff --git a/drivers/block/loop.h b/drivers/block/loop.h
index fb2237c..579f2f7 100644
--- a/drivers/block/loop.h
+++ b/drivers/block/loop.h
@@ -49,6 +49,7 @@ struct loop_device {
 	struct file *	lo_backing_file;
 	struct block_device *lo_device;
 	unsigned	lo_blocksize;
+	unsigned	lo_logical_blocksize;
 	void		*key_data; 
 
 	gfp_t		old_gfp_mask;
-- 
2.6.6

[PATCH 4/4] loop: Pass logical blocksize in 'lo_init[0]' ioctl field

[Hannes Reinecke]

The current LOOP_SET_STATUS64 ioctl has two unused fields
'init[2]', which can be used in conjunction with the
LO_FLAGS_BLOCKSIZE flag to pass in the new logical blocksize.

Signed-off-by: Hannes Reinecke <hare@suse.de>
---
 drivers/block/loop.c | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index a863667..9ad56bf 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -221,7 +221,8 @@ static void __loop_update_dio(struct loop_device *lo, bool dio)
 }
 
 static int
-figure_loop_size(struct loop_device *lo, loff_t offset, loff_t sizelimit)
+figure_loop_size(struct loop_device *lo, loff_t offset, loff_t sizelimit,
+		 loff_t logical_blocksize)
 {
 	loff_t size = get_size(offset, sizelimit, lo->lo_backing_file);
 	sector_t x = (sector_t)size;
@@ -234,6 +235,7 @@ figure_loop_size(struct loop_device *lo, loff_t offset, loff_t sizelimit)
 	if (lo->lo_sizelimit != sizelimit)
 		lo->lo_sizelimit = sizelimit;
 	if (lo->lo_flags & LO_FLAGS_BLOCKSIZE) {
+		lo->lo_logical_blocksize = logical_blocksize;
 		blk_queue_physical_block_size(lo->lo_queue, lo->lo_blocksize);
 		blk_queue_logical_block_size(lo->lo_queue,
 					     lo->lo_logical_blocksize);
@@ -1124,13 +1126,24 @@ loop_set_status(struct loop_device *lo, const struct loop_info64 *info)
 	if (err)
 		return err;
 
-	if (info->lo_flags & LO_FLAGS_BLOCKSIZE)
+	if (info->lo_flags & LO_FLAGS_BLOCKSIZE) {
 		lo->lo_flags |= LO_FLAGS_BLOCKSIZE;
+		if ((info->lo_init[0] != 512) &&
+		    (info->lo_init[0] != 1024) &&
+		    (info->lo_init[0] != 2048) &&
+		    (info->lo_init[0] != 4096))
+			return -EINVAL;
+		if (info->lo_init[0] > lo->lo_blocksize)
+			return -EINVAL;
+	}
 
 	if (lo->lo_offset != info->lo_offset ||
 	    lo->lo_sizelimit != info->lo_sizelimit ||
-	    lo->lo_flags != lo_flags)
-		if (figure_loop_size(lo, info->lo_offset, info->lo_sizelimit))
+	    lo->lo_flags != lo_flags ||
+	    ((lo->lo_flags & LO_FLAGS_BLOCKSIZE) &&
+	     (lo->lo_logical_blocksize != info->lo_init[0])))
+		if (figure_loop_size(lo, info->lo_offset, info->lo_sizelimit,
+				     info->lo_init[0]))
 			return -EFBIG;
 
 	loop_config_discard(lo);
@@ -1315,7 +1328,8 @@ static int loop_set_capacity(struct loop_device *lo)
 	if (unlikely(lo->lo_state != Lo_bound))
 		return -ENXIO;
 
-	return figure_loop_size(lo, lo->lo_offset, lo->lo_sizelimit);
+	return figure_loop_size(lo, lo->lo_offset, lo->lo_sizelimit,
+				lo->lo_logical_blocksize);
 }
 
 static int loop_set_dio(struct loop_device *lo, unsigned long arg)
-- 
2.6.6

Re: [PATCH 1/4] loop: Remove unused 'bdev' argument from loop_set_capacity

[Ming Lei]

On Tue, Nov 1, 2016 at 3:37 AM, Hannes Reinecke <hare@suse.de> wrote:
> Signed-off-by: Hannes Reinecke <hare@suse.de>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> ---
>  drivers/block/loop.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/block/loop.c b/drivers/block/loop.c
> index fa1b7a9..3008dec 100644
> --- a/drivers/block/loop.c
> +++ b/drivers/block/loop.c
> @@ -1298,7 +1298,7 @@ loop_get_status64(struct loop_device *lo, struct loop_info64 __user *arg) {
>         return err;
>  }
>
> -static int loop_set_capacity(struct loop_device *lo, struct block_device *bdev)
> +static int loop_set_capacity(struct loop_device *lo)
>  {
>         if (unlikely(lo->lo_state != Lo_bound))
>                 return -ENXIO;
> @@ -1361,7 +1361,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode,
>         case LOOP_SET_CAPACITY:
>                 err = -EPERM;
>                 if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
> -                       err = loop_set_capacity(lo, bdev);
> +                       err = loop_set_capacity(lo);
>                 break;
>         case LOOP_SET_DIRECT_IO:
>                 err = -EPERM;

Reviewed-by: Ming Lei <tom.leiming@gmail.com>


thanks,
Ming Lei

Re: [PATCH 2/4] loop: Enable correct physical blocksize

[Christoph Hellwig]

On Mon, Oct 31, 2016 at 08:37:17PM +0100, Hannes Reinecke wrote:
> When running on files the physical blocksize is actually 4k,

Is it?  Independent of that sentence the patch looks fine:

Reviewed-by: Christoph Hellwig <hch@lst.de>

Re: [PATCH 3/4] loop: Add 'lo_logical_blocksize'

[Christoph Hellwig]

On Mon, Oct 31, 2016 at 08:37:18PM +0100, Hannes Reinecke wrote:
> Add a new field 'lo_logical_blocksize' to hold the logical
> blocksize of the loop device.

Why do we use the same flag for both block sizes?  If there is a good
reason for that it should be documented and both should be in the same
patch.

Re: [PATCH 4/4] loop: Pass logical blocksize in 'lo_init[0]' ioctl field

[Christoph Hellwig]

On Mon, Oct 31, 2016 at 08:37:19PM +0100, Hannes Reinecke wrote:
> The current LOOP_SET_STATUS64 ioctl has two unused fields
> 'init[2]', which can be used in conjunction with the
> LO_FLAGS_BLOCKSIZE flag to pass in the new logical blocksize.

Can we give them sane field names while at it?  Also I see now reason
not to merge this with the previous two patches.

Re: [PATCH 3/4] loop: Add 'lo_logical_blocksize'

[Hannes Reinecke]

On 11/01/2016 03:01 PM, Christoph Hellwig wrote:
> On Mon, Oct 31, 2016 at 08:37:18PM +0100, Hannes Reinecke wrote:
>> Add a new field 'lo_logical_blocksize' to hold the logical
>> blocksize of the loop device.
>
> Why do we use the same flag for both block sizes?  If there is a good
> reason for that it should be documented and both should be in the same
> patch.
>
_Actually_ the flag enables you to set the logical blocksize.
However, if you want to set the logical blocksize you also have to set a 
matching physical blocksize.
And the physical blocksize itself isn't settable, rather taken from the 
underlying device/filesystem/whatever.
So the flag itself does not have a dual purpose, it's just out of 
necessity that you have to set the physical blocksize if you want to 
change the logical blocksize.

And I got pushback when I proposed to set the physical blocksize always,
so I made it dependent on the LO_BLOCKSIZE flag, too.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH 4/4] loop: Pass logical blocksize in 'lo_init[0]' ioctl field

[Hannes Reinecke]

On 11/01/2016 03:02 PM, Christoph Hellwig wrote:
> On Mon, Oct 31, 2016 at 08:37:19PM +0100, Hannes Reinecke wrote:
>> The current LOOP_SET_STATUS64 ioctl has two unused fields
>> 'init[2]', which can be used in conjunction with the
>> LO_FLAGS_BLOCKSIZE flag to pass in the new logical blocksize.
>
> Can we give them sane field names while at it?  Also I see now reason
> not to merge this with the previous two patches.
>
Sure I can rename it.
As for merging, I don't really care.

So for the next round I'll be merging them all together.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH 4/4] loop: Pass logical blocksize in 'lo_init[0]' ioctl field

[Hannes Reinecke]

On 11/01/2016 03:02 PM, Christoph Hellwig wrote:
> On Mon, Oct 31, 2016 at 08:37:19PM +0100, Hannes Reinecke wrote:
>> The current LOOP_SET_STATUS64 ioctl has two unused fields
>> 'init[2]', which can be used in conjunction with the
>> LO_FLAGS_BLOCKSIZE flag to pass in the new logical blocksize.
>
> Can we give them sane field names while at it?  Also I see now reason
> not to merge this with the previous two patches.
>
I'd rather not change the name; it's part of the exported ioctl 
interface, and changing would mean to change the ioctl interface.
Which would make quite some user-space programs barf.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

[PATCH 1/4] loop: Remove unused 'bdev' argument from loop_set_capacity

[Hannes Reinecke]

Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/loop.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index f7a4c9d..b9e12a0 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -1130,7 +1130,7 @@ loop_get_status64(struct loop_device *lo, struct loop_info64 __user *arg) {
 	return err;
 }
 
-static int loop_set_capacity(struct loop_device *lo, struct block_device *bdev)
+static int loop_set_capacity(struct loop_device *lo)
 {
 	if (unlikely(lo->lo_state != Lo_bound))
 		return -ENXIO;
@@ -1179,7 +1179,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode,
 	case LOOP_SET_CAPACITY:
 		err = -EPERM;
 		if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
-			err = loop_set_capacity(lo, bdev);
+			err = loop_set_capacity(lo);
 		break;
 	default:
 		err = lo->ioctl ? lo->ioctl(lo, cmd, arg) : -EINVAL;
-- 
1.8.4.5