There were multiple places in dm's __bind() function where it could fail and not completely roll back, leaving the device using the the old table, but with device limits and resources from the new table. Additionally, unused mempools for request-based devices were not always freed immediately. Finally, there were a number of issues with switching zoned tables that emulate zone append (in other words, dm-crypt on top of zoned devices). dm_blk_report_zones() could be called while the device was suspended and modifying zoned resources or could possibly fail to end a srcu read section. More importantly, blk_revalidate_disk_zones() would never get called when updating a zoned table. This could cause the dm device to see the wrong zone write offsets, not have a large enough zwplugs reserved in its mempool, or read invalid memory when checking the conventional zones bitmap. This patchset fixes these issues. It does not make it so that device-mapper is able to load any zoned table from any other zoned table. Zoned dm-crypt devices can be safely grown and shrunk, but reloading a zoned dm-crypt device to, for instance, point at a completely different underlying device won't work correctly. IO might fail since the zone write offsets of the dm-crypt device will not be updated for all the existing zones with plugs. If the new device's zone offsets don't match the old device's offsets, IO to the zone will fail. If the ability to switch tables from a zoned dm-crypt device to an abritry other zoned dm-crypt device is important to people, it could be done as long as there are no plugged zones when dm suspends. This patchset also doesn't touch the code for switching from a zoned to a non-zoned device. Switching from a zoned dm-crypt device to a non-zoned device is problematic if there are plugged zones, since the underlying device will not be prepared to handle these plugged writes. Switching to a target that does not pass down IOs, like the dm-error target, is fine. So is switching when there are no plugged zones, except that we do not free the zoned resources in this case, even though we safely could. If people are interested in removing some of these limitations, I can send patches for them, but I'm not sure how much extra code we want, just to support niche zoned dm-crypt reloads. Benjamin Marzinski (7): dm: don't change md if dm_table_set_restrictions() fails dm: free table mempools if not used in __bind dm: handle failures in dm_table_set_restrictions dm: fix dm_blk_report_zones blk-zoned: clean up zone settings for devices without zwplugs blk-zoned: modify blk_revalidate_disk_zones for bio-based drivers dm: allow devices to revalidate existing zones block/blk-zoned.c | 78 ++++++++++++++++++++++-------------------- block/blk.h | 4 --- drivers/md/dm-core.h | 1 + drivers/md/dm-table.c | 24 ++++++++----- drivers/md/dm-zone.c | 75 ++++++++++++++++++++++++++-------------- drivers/md/dm.c | 30 ++++++++-------- drivers/md/dm.h | 1 + include/linux/blkdev.h | 4 +++ 8 files changed, 128 insertions(+), 89 deletions(-) -- 2.48.1
