ext4: Move to shared i_rwsem even without dioread_nolock mount opt
authorRitesh Harjani <riteshh@linux.ibm.com>
Thu, 12 Dec 2019 05:55:57 +0000 (11:25 +0530)
committerTheodore Ts'o <tytso@mit.edu>
Mon, 23 Dec 2019 04:57:27 +0000 (23:57 -0500)
commitbc6385dab125d20870f0eb9ca9e589f43abb3f56
treecd540e75d89db53d020513cfda559ef5fb5ed294
parentaa9714d0e39788d0688474c9d5f6a9a36159599f
ext4: Move to shared i_rwsem even without dioread_nolock mount opt

We were using shared locking only in case of dioread_nolock mount option in case
of DIO overwrites. This mount condition is not needed anymore with current code,
since:-

1. No race between buffered writes & DIO overwrites. Since buffIO writes takes
exclusive lock & DIO overwrites will take shared locking. Also DIO path will
make sure to flush and wait for any dirty page cache data.

2. No race between buffered reads & DIO overwrites, since there is no block
allocation that is possible with DIO overwrites. So no stale data exposure
should happen. Same is the case between DIO reads & DIO overwrites.

3. Also other paths like truncate is protected, since we wait there for any DIO
in flight to be over.

Reviewed-by: Jan Kara <jack@suse.cz>
Tested-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/20191212055557.11151-4-riteshh@linux.ibm.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
fs/ext4/file.c