[PATCH] coda: change locking order in coda_file_write_iter()

[PATCH] coda: change locking order in coda_file_write_iter()

[Amir Goldstein]

The coda host file is a backing file for the coda inode on a different
filesystem than the coda inode.

Change the locking order to take the coda inode lock before taking
the backing host file freeze protection, same as in ovl_write_iter()
and in network filesystems that use cachefiles.

Link: https://lore.kernel.org/r/CAOQ4uxjcnwuF1gMxe64WLODGA_MyAy8x-DtqkCUxqVQKk3Xbng@mail.gmail.com/
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
---

Hi Jan,

Can you please ack this patch so that I can add it to my series
and send to Christian?

Thanks,
Amir.

 fs/coda/file.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/coda/file.c b/fs/coda/file.c
index 16acc58311ea..e62315c37386 100644
--- a/fs/coda/file.c
+++ b/fs/coda/file.c
@@ -79,14 +79,14 @@ coda_file_write_iter(struct kiocb *iocb, struct iov_iter *to)
 	if (ret)
 		goto finish_write;
 
-	file_start_write(host_file);
 	inode_lock(coda_inode);
+	file_start_write(host_file);
 	ret = vfs_iter_write(cfi->cfi_container, to, &iocb->ki_pos, 0);
 	coda_inode->i_size = file_inode(host_file)->i_size;
 	coda_inode->i_blocks = (coda_inode->i_size + 511) >> 9;
 	inode_set_mtime_to_ts(coda_inode, inode_set_ctime_current(coda_inode));
-	inode_unlock(coda_inode);
 	file_end_write(host_file);
+	inode_unlock(coda_inode);
 
 finish_write:
 	venus_access_intent(coda_inode->i_sb, coda_i2f(coda_inode),
-- 
2.34.1

Re: [PATCH] coda: change locking order in coda_file_write_iter()

[Jan Harkes]

On Mon, Nov 20, 2023 at 11:51:10AM +0200, Amir Goldstein wrote:
> The coda host file is a backing file for the coda inode on a different
> filesystem than the coda inode.
> 
> Change the locking order to take the coda inode lock before taking
> the backing host file freeze protection, same as in ovl_write_iter()
> and in network filesystems that use cachefiles.
> 
> Link: https://lore.kernel.org/r/CAOQ4uxjcnwuF1gMxe64WLODGA_MyAy8x-DtqkCUxqVQKk3Xbng@mail.gmail.com/
> Signed-off-by: Amir Goldstein <amir73il@gmail.com>

Acked-by: Jan Harkes <jaharkes@cs.cmu.edu

Looks good, it makes sense to use the same lock ordering as the other
file systems. Thanks for pushing this along.

Jan