dax: Modify bdev_dax_pgoff() to handle NULL bdev
authorVivek Goyal <vgoyal@redhat.com>
Wed, 19 Aug 2020 22:19:39 +0000 (18:19 -0400)
committerMiklos Szeredi <mszeredi@redhat.com>
Thu, 10 Sep 2020 09:39:22 +0000 (11:39 +0200)
commit1a9d5d405962d134acb8efb4ccb4bc17805134c7
treea68ee728ad2e5f65ba5bd049360bdc395100d3a7
parent27bd6129a4c5b1269540c925bfda1c280cbb9f1d
dax: Modify bdev_dax_pgoff() to handle NULL bdev

virtiofs does not have a block device but it has dax device.
Modify bdev_dax_pgoff() to be able to handle that.

If there is no bdev, that means dax offset is 0. (It can't be a partition
block device starting at an offset in dax device).

This is little hackish. There have been discussions about getting rid
of dax not supporting partitions.

https://lore.kernel.org/linux-fsdevel/20200107125159.GA15745@infradead.org/

IMHO, this path can easily break exisitng users. For example
ioctl(BLKPG_ADD_PARTITION) will start breaking on block devices
supporting DAX. Also, I personally find it very useful to be able to
partition dax devices and still be able to use DAX.

Alternatively, I tried to store offset into dax device information in iomap
interface, but that got NACKed.

https://lore.kernel.org/linux-fsdevel/20200217133117.GB20444@infradead.org/

I can't think of a good path to solve this issue properly. So to make
progress, it seems this patch is least bad option for now and I hope
we can take it.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Vishal L Verma <vishal.l.verma@intel.com>
Cc: "Weiny, Ira" <ira.weiny@intel.com>
Cc: linux-nvdimm@lists.01.org
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
drivers/dax/super.c