Re: [PATCH] nfsd: fix courtesy client with deny mode handling in nfs4_upgrade_open

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2/15/23 3:05 PM, dai.ngo@xxxxxxxxxx wrote:
Hi Jeff, Chuck,

On 2/3/23 10:18 AM, Jeff Layton wrote:
The nested if statements here make no sense, as you can never reach
"else" branch in the nested statement. Fix the error handling for
when there is a courtesy client that holds a conflicting deny mode.

Fixes: 3d69427151806 (NFSD: add support for share reservation conflict to courteous server)
Reported-by: 張智諺 <cc85nod@xxxxxxxxx>
Cc: Dai Ngo <dai.ngo@xxxxxxxxxx>
Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
---
  fs/nfsd/nfs4state.c | 21 +++++++++++----------
  1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index c39e43742dd6..af22dfdc6fcc 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -5282,16 +5282,17 @@ nfs4_upgrade_open(struct svc_rqst *rqstp, struct nfs4_file *fp,
      /* test and set deny mode */
      spin_lock(&fp->fi_lock);
      status = nfs4_file_check_deny(fp, open->op_share_deny);
-    if (status == nfs_ok) {
-        if (status != nfserr_share_denied) {
-            set_deny(open->op_share_deny, stp);
-            fp->fi_share_deny |=
-                (open->op_share_deny & NFS4_SHARE_DENY_BOTH);
-        } else {
-            if (nfs4_resolve_deny_conflicts_locked(fp, false,
-                    stp, open->op_share_deny, false))
-                status = nfserr_jukebox;
-        }
+    switch (status) {
+    case nfs_ok:
+        set_deny(open->op_share_deny, stp);
+        fp->fi_share_deny |=
+            (open->op_share_deny & NFS4_SHARE_DENY_BOTH);
+        break;
+    case nfserr_share_denied:
+        if (nfs4_resolve_deny_conflicts_locked(fp, false,
+                stp, open->op_share_deny, false))

While trying to write a pynfs test case to exercise this code path,
I realize that we don't need to call nfs4_resolve_deny_conflicts_locked
here since this is an open upgrade so it must comes from the same client
hence there is no conflict to resolve. Same behavior as OPEN_DOWNGRADE.

never mind, I found the scenario where this code path is executed:

Client1 opens fileX with WRITE, DENY_NONE
Client2 opens fileX with READ, DENY_NONE
Client2 opens fileX with READ, DENY_WRITE

-Dai


-Dai

+            status = nfserr_jukebox;
+        break;
      }
      spin_unlock(&fp->fi_lock);



[Index of Archives]     [Linux Filesystem Development]     [Linux USB Development]     [Linux Media Development]     [Video for Linux]     [Linux NILFS]     [Linux Audio Users]     [Yosemite Info]     [Linux SCSI]

  Powered by Linux