fuse: O_DIRECT support for files

PATCH Applied Anand Avati 3 participants 3 messages Feb 17, 2012 – Mar 5, 2012

Subsystems (2)

VFS FUSE filesystem core

lore

Share & tools

lore

Anand Avati

author

2012-02-17 17:46 UTC #0 lore raw

PATCH v5 1/1 fuse: O_DIRECT support for files

Implement ->direct_IO() method in aops. The ->direct_IO() method combines the existing fuse_direct_read/fuse_direct_write methods to implement O_DIRECT functionality.

Reaching ->direct_IO() in the read path via generic_file_aio_read ensures proper synchronization with page cache with its existing framework.

Reaching ->direct_IO() in the write path via fuse_file_aio_write is made to come via generic_file_direct_write() which makes it play nice with the page cache w.r.t other mmap pages etc.

On files marked 'direct_io' by the filesystem server, IO always follows the fuse_direct_read/write path. There is no effect of fcntl(O_DIRECT) and it always succeeds.

On files not marked with 'direct_io' by the filesystem server, the IO path depends on O_DIRECT flag by the application. This can be passed at the time of open() as well as via fcntl().

Note that asynchronous O_DIRECT iocb jobs are completed synchronously always (this has been the case with FUSE even before this patch)

Signed-off-by Anand Avati <[email protected]>

---
 fs/fuse/dir.c  |    3 -
 fs/fuse/file.c |  129 ++++++++++++++++++++++++++++++++++++++++++++++++-------
 2 files changed, 112 insertions(+), 20 deletions(-)

fs/fuse/dir.c +0 -3

index 2066328..7e5dbd0 100644
@@ -387,9 +387,6 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry,
 	if (fc->no_create)
 		return -ENOSYS;
 
-	if (flags & O_DIRECT)
-		return -EINVAL;
-
 	forget = fuse_alloc_forget();
 	if (!forget)
 		return -ENOMEM;

fs/fuse/file.c +112 ~~-17~~

192 lines not on this page · raw message

Signature · 1 line

1.7.4.4

Jeff Moyer

Reviewed-by on v5 1/1

2012-03-01 20:24 UTC #1 lore raw

Quoted 15 lines · Anand Avati

Implement ->direct_IO() method in aops. The ->direct_IO() method combines the existing fuse_direct_read/fuse_direct_write methods to implement O_DIRECT functionality.
Reaching ->direct_IO() in the read path via generic_file_aio_read ensures proper synchronization with page cache with its existing framework.
Reaching ->direct_IO() in the write path via fuse_file_aio_write is made to come via generic_file_direct_write() which makes it play nice with the page cache w.r.t other mmap pages etc.
On files marked 'direct_io' by the filesystem server, IO always follows the fuse_direct_read/write path. There is no effect of fcntl(O_DIRECT) and it always succeeds.

Quoted 8 lines

On files not marked with 'direct_io' by the filesystem server, the IO path depends on O_DIRECT flag by the application. This can be passed at the time of open() as well as via fcntl().
Note that asynchronous O_DIRECT iocb jobs are completed synchronously always (this has been the case with FUSE even before this patch)
Signed-off-by Anand Avati <[email protected]>

Looks good to me.

Reviewed-by Jeff Moyer <[email protected]>

Miklos Szeredi

maintainer Applied

2012-03-05 16:04 UTC #2 lore raw

Quoted 21 lines · Jeff Moyer

Quoted 19 lines · Anand Avati
Implement ->direct_IO() method in aops. The ->direct_IO() method combines the existing fuse_direct_read/fuse_direct_write methods to implement O_DIRECT functionality.
Reaching ->direct_IO() in the read path via generic_file_aio_read ensures proper synchronization with page cache with its existing framework.
Reaching ->direct_IO() in the write path via fuse_file_aio_write is made to come via generic_file_direct_write() which makes it play nice with the page cache w.r.t other mmap pages etc.
On files marked 'direct_io' by the filesystem server, IO always follows the fuse_direct_read/write path. There is no effect of fcntl(O_DIRECT) and it always succeeds.
On files not marked with 'direct_io' by the filesystem server, the IO path depends on O_DIRECT flag by the application. This can be passed at the time of open() as well as via fcntl().

Quoted 8 lines

Quoted 4 lines
Note that asynchronous O_DIRECT iocb jobs are completed synchronously always (this has been the case with FUSE even before this patch)
Signed-off-by Anand Avati <[email protected]>
Looks good to me.
Reviewed-by Jeff Moyer <[email protected]>

Pushed to the for-next branch of the fuse tree, together with and updated version of the i_nlink patch.

Thanks,
Miklos