nfsd: Return the correct number of bytes written to the file
authorTrond Myklebust <trondmy@gmail.com>
Tue, 17 Dec 2019 17:33:33 +0000 (12:33 -0500)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 11 Feb 2020 12:34:13 +0000 (04:34 -0800)
commit 09a80f2aef06b7c86143f5c14efd3485e0d2c139 upstream.

We must allow for the fact that iov_iter_write() could have returned
a short write (e.g. if there was an ENOSPC issue).

Fixes: d890be159a71 "nfsd: Add I/O trace points in the NFSv4 write path"
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/nfsd/vfs.c

index 4fe8db3149506c59e098812de14ccebab7acdf61..80cededcd10d68899e654bc4d2af5f770d32a628 100644 (file)
@@ -1016,6 +1016,7 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file,
        host_err = vfs_iter_write(file, &iter, &pos, flags);
        if (host_err < 0)
                goto out_nfserr;
+       *cnt = host_err;
        nfsdstats.io_write += *cnt;
        fsnotify_modify(file);