On Fri, Aug 10 2018 at 11:23am -0400, Mikulas Patocka <[email protected]> wrote:
> dm-crypt should only increase device limits, it should not decrease them. > > This fixes a bug where the user could create a crypt device with 1024 > sector size on the top of scsi device that had 4096 logical sectors. The > limit 4096 would be lost and the user could incorrectly send 1024-I/Os to > the crypt device. > > Signed-off-by: Mikulas Patocka <[email protected]> > > --- > drivers/md/dm-crypt.c | 8 +++----- > 1 file changed, 3 insertions(+), 5 deletions(-) > > Index: linux-2.6/drivers/md/dm-crypt.c > =================================================================== > --- linux-2.6.orig/drivers/md/dm-crypt.c 2018-08-10 15:02:26.700000000 > +0200 > +++ linux-2.6/drivers/md/dm-crypt.c 2018-08-10 15:28:02.170000000 +0200 > @@ -3072,11 +3072,9 @@ static void crypt_io_hints(struct dm_tar > */ > limits->max_segment_size = PAGE_SIZE; > > - if (cc->sector_size != (1 << SECTOR_SHIFT)) { > - limits->logical_block_size = cc->sector_size; > - limits->physical_block_size = cc->sector_size; > - blk_limits_io_min(limits, cc->sector_size); > - } > + limits->logical_block_size = max(limits->logical_block_size, > cc->sector_size); > + limits->physical_block_size = max(limits->physical_block_size, > (unsigned)cc->sector_size); > + blk_limits_io_min(limits, cc->sector_size); > } > > static struct target_type crypt_target = { tweaked and staged for 4.19 here: https://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=dm-4.19&id=08575fd060d7f88f674509461b32d86b272093d9 I switched to using max_t() and stacked io_min rather than impose that it is cc->sector_size. -- dm-devel mailing list [email protected] https://www.redhat.com/mailman/listinfo/dm-devel
