fuse: Don't do file_update_time for write_operation.
PATCH 1/1 fuse: Don't do file_update_time for write_operation.
For direct-write and normal-write operations, after write they will call fuse_invalidate_attr. So it's make no sense to call file_update_time before writing.
In function file_update_time, there is other thing which update version of inode.For this i am not sure.
---fs/fuse/file.c | 4 ----
1 file changed, 4 deletions(-)
fs/fuse/file.c +0 -0
@@ -988,10 +988,6 @@ static ssize_t fuse_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
if (err)
goto out;
- err = file_update_time(file);
- if (err)
- goto out;
-
if (file->f_flags & O_DIRECT) {
written = generic_file_direct_write(iocb, iov, &nr_segs, pos, &iocb->ki_pos,
Signature · 1 line
1.8.2.rc2.4.g7799588
Miklos Szeredi
maintainerQuoted 5 lines · majianpeng
For direct-write and normal-write operations, after write they will call
fuse_invalidate_attr. So it's make no sense to callfile_update_timebefore writing.In function
file_update_time, there is other thing which update version of inode.For this i am not sure.
file_update_time will skip the actual file update anyway due to S_NOCMTIME. And when we will do cached writes we actually need the time updates. So I'm not applying this one.
Thanks,
Miklos
Quoted 24 lines
Signed-off-by Jianpeng Ma <[email protected]>---
fs/fuse/file.c| 4 ----1 file changed, 4 deletions(-)fs/fuse/file.c +0
-5@@ -988,10 +988,6 @@ static ssize_t fuse_file_aio_write(struct kiocb *iocb, const struct iovec *iov,if (err) goto out; - err =file_update_time (file); -if (err) - goto out; -if (file->f_flags &O_DIRECT ) { written = generic_file_direct_write(iocb, iov, &nr_segs, pos, &iocb->ki_pos, --1.8.2.rc2.4.g7799588