From cbd0d51a3318583fabf03bccc7a987e158482361 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Tue, 7 Feb 2006 12:58:32 -0800 Subject: [PATCH] knfsd: fix nfs4_open lock leak I just noticed that my patch "don't create on open that fails due to ERR_GRACE" (recently commited as fb553c0f17444e090db951b96df4d2d71b4f4b6b) had an obvious problem that causes a deadlock on reboot recovery. Sending in this now since it seems like a clear 2.6.16 candidate.--b. We're returning with a lock held in some error cases. Signed-off-by: J. Bruce Fields Signed-off-by: Neil Brown Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nfsd/nfs4proc.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs/nfsd') diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index a00fe8686293..6d63f1d9e5f5 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -195,10 +195,12 @@ nfsd4_open(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_open /* Openowner is now set, so sequence id will get bumped. Now we need * these checks before we do any creates: */ + status = nfserr_grace; if (nfs4_in_grace() && open->op_claim_type != NFS4_OPEN_CLAIM_PREVIOUS) - return nfserr_grace; + goto out; + status = nfserr_no_grace; if (!nfs4_in_grace() && open->op_claim_type == NFS4_OPEN_CLAIM_PREVIOUS) - return nfserr_no_grace; + goto out; switch (open->op_claim_type) { case NFS4_OPEN_CLAIM_DELEGATE_CUR: -- cgit v1.2.3