kernel/linux.git/fs/gfs2, branch v6.2-rc2

Merge tag 'gfs2-v6.1-rc7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2

2022-12-17T14:18:04+00:00

Pull gfs2 updtaes from Andreas Gruenbacher: - Revert a change to delete_work_func() that has gone wrong in commit c412a97cf6c5 ("gfs2: Use TRY lock in gfs2_inode_lookup for UNLINKED inodes"). - Avoid dequeuing GL_ASYNC glock holders twice by first checking if the holder is still queued. - gfs2: Always check the inode size of inline inodes when reading in inodes to prevent corrupt filesystem images from causing weid errors. - Properly handle a race between gfs2_create_inode() and gfs2_inode_lookup() that causes insert_inode_locked4() to return -EBUSY. - Fix and clean up the interaction between gfs2_create_inode() and gfs2_evict_inode() by completely handling the inode deallocation and destruction in gfs2_evict_inode(). - Remove support for glock holder auto-demotion as we have no current plans of using this feature again. - And a few more minor cleanups and clarifications. * tag 'gfs2-v6.1-rc7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2: gfs2: Remove support for glock holder auto-demotion (2) gfs2: Remove support for glock holder auto-demotion gfs2: Minor gfs2_try_evict cleanup gfs2: Partially revert gfs2_inode_lookup change gfs2: Add gfs2_inode_lookup comment gfs2: Uninline and improve glock_{set,clear}_object gfs2: Simply dequeue iopen glock in gfs2_evict_inode gfs2: Clean up after gfs2_create_inode rework gfs2: Avoid dequeuing GL_ASYNC glock holders twice gfs2: Make gfs2_glock_hold return its glock argument gfs2: Always check inode size of inline inodes gfs2: Cosmetic gfs2_dinode_{in,out} cleanup gfs2: Handle -EBUSY result of insert_inode_locked4 gfs2: Fix and clean up create / evict interaction gfs2: Clean up initialization of "ip" in gfs2_create_inode gfs2: Get rid of ghs[] in gfs2_create_inode gfs2: Add extra error check in alloc_dinode

gfs2: Remove support for glock holder auto-demotion (2)

2022-12-15T11:41:22+00:00

As a follow-up to the previous commit, move the recovery related code in __gfs2_glock_dq() to gfs2_glock_dq() where it better fits. No functional change. Signed-off-by: Andreas Gruenbacher

gfs2: Remove support for glock holder auto-demotion

2022-12-15T11:41:22+00:00

Remove the support for glock holder auto-demotion (commit dc732906c245 and folow-ups) as we are not planning to use this feature, and the additional code therefore only adds unnecessary complexity. Signed-off-by: Andreas Gruenbacher

gfs2: Minor gfs2_try_evict cleanup

2022-12-10T12:06:04+00:00

In gfs2_try_evict(), when an inode can't be evicted, we are grabbing a temporary reference on the inode glock to poke that glock. That should be safe, but it's easier to just grab an inode reference as we already do earlier in this function. Signed-off-by: Andreas Gruenbacher

gfs2: Partially revert gfs2_inode_lookup change

2022-12-06T15:08:12+00:00

Commit c412a97cf6c5 changed delete_work_func() to always perform an inode lookup when gfs2_try_evict() fails. This doesn't make sense as a gfs2_try_evict() failure indicates that the inode is likely still in use. Revert that change. Fixes: c412a97cf6c5 ("gfs2: Use TRY lock in gfs2_inode_lookup for UNLINKED inodes") Signed-off-by: Andreas Gruenbacher

gfs2: Add gfs2_inode_lookup comment

2022-12-06T15:06:32+00:00

Add comment on when and why gfs2_cancel_delete_work() needs to be skipped in gfs2_inode_lookup(). Signed-off-by: Andreas Gruenbacher

gfs2: Uninline and improve glock_{set,clear}_object

2022-12-06T15:06:32+00:00

Those functions have reached a size at which having them inline isn't useful anymore, so uninline them. In addition, report the glock name on assertion failures. Signed-off-by: Andreas Gruenbacher

gfs2: Simply dequeue iopen glock in gfs2_evict_inode

2022-12-06T15:06:32+00:00

With the previous change, to simplify things, we can always just dequeue and uninitialize the iopen glock in gfs2_evict_inode() even if it isn't queued anymore. Signed-off-by: Andreas Gruenbacher

gfs2: Clean up after gfs2_create_inode rework

2022-12-06T15:06:31+00:00

Since commit 3d36e57ff768 ("gfs2: gfs2_create_inode rework"), gfs2_evict_inode() and gfs2_create_inode() / gfs2_inode_lookup() will synchronize via the inode hash table and we can be certain that once a new inode is inserted into the inode hash table(), gfs2_evict_inode() has completely destroyed any previous versions. We no longer need to worry about overlapping inode object lifespans. Update the code and comments accordingly. Signed-off-by: Andreas Gruenbacher

gfs2: Avoid dequeuing GL_ASYNC glock holders twice

2022-12-06T15:06:31+00:00

When a locking request fails, the associated glock holder is automatically dequeued from the list of active and waiting holders. For GL_ASYNC locking requests, this will obviously happen asynchronously and it can race with attempts to cancel that locking request via gfs2_glock_dq(). Therefore, don't forget to check if a locking request has already been dequeued in gfs2_glock_dq(). Signed-off-by: Andreas Gruenbacher