From a58e0be6f6b3eb2079b0b8fedc9df6fa86869f1e Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 22 Mar 2013 12:52:59 -0400 Subject: [PATCH 1/6] SUNRPC: Remove extra xprt_put() While testing error cases where rpc_new_client() fails, I saw some oopses. If rpc_new_client() fails, it already invokes xprt_put(). Thus __rpc_clone_client() does not need to invoke it again. Introduced by commit 1b63a751 "SUNRPC: Refactor rpc_clone_client()" Fri Sep 14, 2012. Signed-off-by: Chuck Lever Cc: stable@vger.kernel.org [>=3.7] Signed-off-by: Trond Myklebust --- net/sunrpc/clnt.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index dcc446e7fbf6..9df26b785dc7 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -512,7 +512,7 @@ static struct rpc_clnt *__rpc_clone_client(struct rpc_create_args *args, new = rpc_new_client(args, xprt); if (IS_ERR(new)) { err = PTR_ERR(new); - goto out_put; + goto out_err; } atomic_inc(&clnt->cl_count); @@ -525,8 +525,6 @@ static struct rpc_clnt *__rpc_clone_client(struct rpc_create_args *args, new->cl_chatty = clnt->cl_chatty; return new; -out_put: - xprt_put(xprt); out_err: dprintk("RPC: %s: returned error %d\n", __func__, err); return ERR_PTR(err); From b193d59a4863ea670872d76dc99231ddeb598625 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 4 Apr 2013 15:55:00 -0400 Subject: [PATCH 2/6] NFSv4: Fix a memory leak in nfs4_discover_server_trunking When we assign a new rpc_client to clp->cl_rpcclient, we need to destroy the old one. Signed-off-by: Trond Myklebust Cc: Chuck Lever Cc: stable@vger.kernel.org [>=3.7] --- fs/nfs/nfs4state.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 6ace365c6334..d41a3518509f 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1886,7 +1886,13 @@ again: status = PTR_ERR(clnt); break; } - clp->cl_rpcclient = clnt; + /* Note: this is safe because we haven't yet marked the + * client as ready, so we are the only user of + * clp->cl_rpcclient + */ + clnt = xchg(&clp->cl_rpcclient, clnt); + rpc_shutdown_client(clnt); + clnt = clp->cl_rpcclient; goto again; case -NFS4ERR_MINOR_VERS_MISMATCH: From 7b1f1fd1842e6ede25183c267ae733a7f67f00bc Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 5 Apr 2013 16:11:11 -0400 Subject: [PATCH 3/6] NFSv4/4.1: Fix bugs in nfs4[01]_walk_client_list It is unsafe to use list_for_each_entry_safe() here, because when we drop the nn->nfs_client_lock, we pin the _current_ list entry and ensure that it stays in the list, but we don't do the same for the _next_ list entry. Use of list_for_each_entry() is therefore the correct thing to do. Also fix the refcounting in nfs41_walk_client_list(). Finally, ensure that the nfs_client has finished being initialised and, in the case of NFSv4.1, that the session is set up. Signed-off-by: Trond Myklebust Cc: Chuck Lever Cc: Bryan Schumaker Cc: stable@vger.kernel.org [>= 3.7] --- fs/nfs/nfs4client.c | 44 ++++++++++++++++++++++++++++---------------- 1 file changed, 28 insertions(+), 16 deletions(-) diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c index ac4fc9a8fdbc..c7b346f8cc44 100644 --- a/fs/nfs/nfs4client.c +++ b/fs/nfs/nfs4client.c @@ -300,7 +300,7 @@ int nfs40_walk_client_list(struct nfs_client *new, struct rpc_cred *cred) { struct nfs_net *nn = net_generic(new->cl_net, nfs_net_id); - struct nfs_client *pos, *n, *prev = NULL; + struct nfs_client *pos, *prev = NULL; struct nfs4_setclientid_res clid = { .clientid = new->cl_clientid, .confirm = new->cl_confirm, @@ -308,10 +308,23 @@ int nfs40_walk_client_list(struct nfs_client *new, int status = -NFS4ERR_STALE_CLIENTID; spin_lock(&nn->nfs_client_lock); - list_for_each_entry_safe(pos, n, &nn->nfs_client_list, cl_share_link) { + list_for_each_entry(pos, &nn->nfs_client_list, cl_share_link) { /* If "pos" isn't marked ready, we can't trust the * remaining fields in "pos" */ - if (pos->cl_cons_state < NFS_CS_READY) + if (pos->cl_cons_state > NFS_CS_READY) { + atomic_inc(&pos->cl_count); + spin_unlock(&nn->nfs_client_lock); + + if (prev) + nfs_put_client(prev); + prev = pos; + + status = nfs_wait_client_init_complete(pos); + spin_lock(&nn->nfs_client_lock); + if (status < 0) + continue; + } + if (pos->cl_cons_state != NFS_CS_READY) continue; if (pos->rpc_ops != new->rpc_ops) @@ -423,16 +436,16 @@ int nfs41_walk_client_list(struct nfs_client *new, struct rpc_cred *cred) { struct nfs_net *nn = net_generic(new->cl_net, nfs_net_id); - struct nfs_client *pos, *n, *prev = NULL; + struct nfs_client *pos, *prev = NULL; int status = -NFS4ERR_STALE_CLIENTID; spin_lock(&nn->nfs_client_lock); - list_for_each_entry_safe(pos, n, &nn->nfs_client_list, cl_share_link) { + list_for_each_entry(pos, &nn->nfs_client_list, cl_share_link) { /* If "pos" isn't marked ready, we can't trust the * remaining fields in "pos", especially the client * ID and serverowner fields. Wait for CREATE_SESSION * to finish. */ - if (pos->cl_cons_state < NFS_CS_READY) { + if (pos->cl_cons_state > NFS_CS_READY) { atomic_inc(&pos->cl_count); spin_unlock(&nn->nfs_client_lock); @@ -440,18 +453,17 @@ int nfs41_walk_client_list(struct nfs_client *new, nfs_put_client(prev); prev = pos; - nfs4_schedule_lease_recovery(pos); status = nfs_wait_client_init_complete(pos); - if (status < 0) { - nfs_put_client(pos); - spin_lock(&nn->nfs_client_lock); - continue; + if (status == 0) { + nfs4_schedule_lease_recovery(pos); + status = nfs4_wait_clnt_recover(pos); } - status = pos->cl_cons_state; spin_lock(&nn->nfs_client_lock); if (status < 0) continue; } + if (pos->cl_cons_state != NFS_CS_READY) + continue; if (pos->rpc_ops != new->rpc_ops) continue; @@ -469,17 +481,17 @@ int nfs41_walk_client_list(struct nfs_client *new, continue; atomic_inc(&pos->cl_count); - spin_unlock(&nn->nfs_client_lock); + *result = pos; dprintk("NFS: <-- %s using nfs_client = %p ({%d})\n", __func__, pos, atomic_read(&pos->cl_count)); - - *result = pos; - return 0; + break; } /* No matching nfs_client found. */ spin_unlock(&nn->nfs_client_lock); dprintk("NFS: <-- %s status = %d\n", __func__, status); + if (prev) + nfs_put_client(prev); return status; } #endif /* CONFIG_NFS_V4_1 */ From f05c124a70a4953a66acbd6d6c601ea1eb5d0fa7 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 5 Apr 2013 14:13:21 -0400 Subject: [PATCH 4/6] SUNRPC: Fix a potential memory leak in rpc_new_client If the call to rpciod_up() fails, we currently leak a reference to the struct rpc_xprt. As part of the fix, we also remove the redundant check for xprt!=NULL. This is already taken care of by the callers. Signed-off-by: Trond Myklebust --- net/sunrpc/clnt.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 9df26b785dc7..d5f35f15af98 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -304,10 +304,8 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args, stru err = rpciod_up(); if (err) goto out_no_rpciod; - err = -EINVAL; - if (!xprt) - goto out_no_xprt; + err = -EINVAL; if (args->version >= program->nrvers) goto out_err; version = program->version[args->version]; @@ -382,10 +380,9 @@ out_no_principal: out_no_stats: kfree(clnt); out_err: - xprt_put(xprt); -out_no_xprt: rpciod_down(); out_no_rpciod: + xprt_put(xprt); return ERR_PTR(err); } From fa332941c0c7c00e3420078268b7558d0ef792b5 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 9 Apr 2013 12:56:52 -0400 Subject: [PATCH 5/6] NFSv4: Fix another potential state manager deadlock Don't hold the NFSv4 sequence id while we check for open permission. The call to ACCESS may block due to reboot recovery. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 26431cf62ddb..0ad025eb523b 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -1046,6 +1046,7 @@ static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata) /* Save the delegation */ nfs4_stateid_copy(&stateid, &delegation->stateid); rcu_read_unlock(); + nfs_release_seqid(opendata->o_arg.seqid); ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode); if (ret != 0) goto out; From eb04e0ac198cec3bab407ad220438dfa65c19c67 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 10 Apr 2013 12:44:18 -0400 Subject: [PATCH 6/6] NFSv4: Doh! Typo in the fix to nfs41_walk_client_list Make sure that we set the status to 0 on success. Missed in testing because it never appears when doing multiple mounts to _different_ servers. Signed-off-by: Trond Myklebust Cc: # 3.7.x: 7b1f1fd: NFSv4/4.1: Fix bugs in nfs4[01]_walk_client_list --- fs/nfs/nfs4client.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c index c7b346f8cc44..66b6664dcd4c 100644 --- a/fs/nfs/nfs4client.c +++ b/fs/nfs/nfs4client.c @@ -482,6 +482,7 @@ int nfs41_walk_client_list(struct nfs_client *new, atomic_inc(&pos->cl_count); *result = pos; + status = 0; dprintk("NFS: <-- %s using nfs_client = %p ({%d})\n", __func__, pos, atomic_read(&pos->cl_count)); break;