On 8/15/21 12:07 AM, Christoph Hellwig wrote:
On Sat, Aug 14, 2021 at 02:13:09PM -0700, Guenter Roeck wrote:
On Wed, Aug 04, 2021 at 11:41:43AM +0200, Christoph Hellwig wrote:
device mapper needs to register holders before it is ready to do I/O.
Currently it does so by registering the disk early, which can leave
the disk and queue in a weird half state where the queue is registered
with the disk, except for sysfs and the elevator. And this state has
been a bit promlematic before, and will get more so when sorting out
the responsibilities between the queue and the disk.
Support registering holders on an initialized but not registered disk
instead by delaying the sysfs registration until the disk is registered.
Signed-off-by: Christoph Hellwig <hch@xxxxxx>
Reviewed-by: Mike Snitzer <snitzer@xxxxxxxxxx>
This patch results in lockdep splats when booting from flash.
Reverting it fixes the proboem.
Should be fixed by:
https://git.kernel.dk/cgit/linux-block/commit/?h=for-5.15/drivers&id=6e4df4c6488165637b95b9701cc862a42a3836ba
No, it doesn't. I could not apply this patch alone, so I applied the entire series
on top of next-20210813 and gave it another try.
f53c2d11ac98 (HEAD -> master) nbd: reduce the nbd_index_mutex scope
f2f5254b356f nbd: refactor device search and allocation in nbd_genl_connect
d5b03177e069 nbd: return the allocated nbd_device from nbd_dev_add
350b3f6a6e6b nbd: remove nbd_del_disk
49efbeb9de86 nbd: refactor device removal
cdd920eb7cf2 nbd: do del_gendisk() asynchronously for NBD_DESTROY_ON_DISCONNECT
4b358aabb93a (tag: next-20210813, origin/master, origin/HEAD) Add linux-next specific files for 20210813
Still:
...
[ 14.467748][ T1] Possible unsafe locking scenario:
[ 14.467748][ T1]
[ 14.467928][ T1] CPU0 CPU1
[ 14.468058][ T1] ---- ----
[ 14.468187][ T1] lock(&disk->open_mutex);
[ 14.468317][ T1] lock(mtd_table_mutex);
[ 14.468493][ T1] lock(&disk->open_mutex);
[ 14.468671][ T1] lock(mtd_table_mutex);
Guenter
