fs/buffer.c: Add generic_buffers_fsync*() implementation
authorRitesh Harjani (IBM) <ritesh.list@gmail.com>
Fri, 21 Apr 2023 09:46:12 +0000 (15:16 +0530)
committerJan Kara <jack@suse.cz>
Tue, 16 May 2023 09:32:42 +0000 (11:32 +0200)
commit31b2ebc0929e964f4edfbfa7129d43f7e3c17165
tree7dffe7cab84d3b01029d2e0a5c1ebd38b290e411
parentfcced95b6ba2a507a83b8b3e0358a8ac16b13e35
fs/buffer.c: Add generic_buffers_fsync*() implementation

Some of the higher layers like iomap takes inode_lock() when calling
generic_write_sync().
Also writeback already happens from other paths without inode lock,
so it's difficult to say that we really need sync_mapping_buffers() to
take any inode locking here. Having said that, let's add
generic_buffers_fsync/_noflush() implementation in buffer.c with no
inode_lock/unlock() for now so that filesystems like ext2 and
ext4's nojournal mode can use it.

Ext4 when got converted to iomap for direct-io already copied it's own
variant of __generic_file_fsync() without lock.

This patch adds generic_buffers_fsync()
& generic_buffers_fsync_noflush() implementations for use in filesystems
like ext2 & ext4 respectively.

Later we can review other filesystems as well to see if we can make
generic_buffers_fsync/_noflush() which does not take any inode_lock() as
the default path.

Tested-by: Disha Goel <disgoel@linux.ibm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Message-Id: <d573408ac8408627d23a3d2d166e748c172c4c9e.1682069716.git.ritesh.list@gmail.com>
fs/buffer.c
include/linux/buffer_head.h