From f4f58ed19b9ee303985f47352f795631a979294c Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 23 Jan 2016 14:12:52 -0500 Subject: NFSv4.x: Remove hard coded slotids in callback channel Instead, use the values encoded in the slot table itself. Signed-off-by: Trond Myklebust --- fs/nfs/callback_proc.c | 6 +++--- fs/nfs/nfs4proc.c | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index f0939d097406..83a66a8f40f2 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -361,7 +361,7 @@ validate_seqid(struct nfs4_slot_table *tbl, struct cb_sequenceargs * args) dprintk("%s enter. slotid %u seqid %u\n", __func__, args->csa_slotid, args->csa_sequenceid); - if (args->csa_slotid >= NFS41_BC_MAX_CALLBACKS) + if (args->csa_slotid > tbl->server_highest_slotid) return htonl(NFS4ERR_BADSLOT); slot = tbl->slots + args->csa_slotid; @@ -489,8 +489,8 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, sizeof(res->csr_sessionid)); res->csr_sequenceid = args->csa_sequenceid; res->csr_slotid = args->csa_slotid; - res->csr_highestslotid = NFS41_BC_MAX_CALLBACKS - 1; - res->csr_target_highestslotid = NFS41_BC_MAX_CALLBACKS - 1; + res->csr_highestslotid = tbl->server_highest_slotid; + res->csr_target_highestslotid = tbl->target_highest_slotid; status = validate_seqid(tbl, args); if (status) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 4bfc33ad0563..be685e236bd5 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -7319,7 +7319,7 @@ static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args) args->bc_attrs.max_resp_sz = PAGE_SIZE; args->bc_attrs.max_resp_sz_cached = 0; args->bc_attrs.max_ops = NFS4_MAX_BACK_CHANNEL_OPS; - args->bc_attrs.max_reqs = 1; + args->bc_attrs.max_reqs = NFS41_BC_MAX_CALLBACKS; dprintk("%s: Back Channel : max_rqst_sz=%u max_resp_sz=%u " "max_resp_sz_cached=%u max_ops=%u max_reqs=%u\n", -- cgit v1.2.3 From f74a834a0e1b35e39e917a1a88a62817610ee20b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 23 Jan 2016 17:15:37 -0500 Subject: NFSv4.x: CB_SEQUENCE should return NFS4ERR_DELAY if still executing See RFC5661 Section 2.10.6.2: if retrying a request, and the old one is still in progress, we must return NFS4ERR_DELAY as the reply to sequence. Signed-off-by: Trond Myklebust --- fs/nfs/callback_proc.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 83a66a8f40f2..e0844fdbd9ac 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -375,6 +375,8 @@ validate_seqid(struct nfs4_slot_table *tbl, struct cb_sequenceargs * args) if (args->csa_sequenceid == slot->seq_nr) { dprintk("%s seqid %u is a replay\n", __func__, args->csa_sequenceid); + if (tbl->highest_used_slotid != NFS4_NO_SLOT) + return htonl(NFS4ERR_DELAY); /* Signal process_op to set this error on next op */ if (args->csa_cachethis == 0) return htonl(NFS4ERR_RETRY_UNCACHED_REP); -- cgit v1.2.3 From 80f9642724af5dfab7d330481fa22e07fde084da Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 23 Jan 2016 15:21:08 -0500 Subject: NFSv4.x: Enforce the ca_maxresponsesize_cached on the back channel We have no duplicate reply cache, so we always set the back channel ca_maxresponsesize_cached to zero when negotiating the session. That means we should always error out as soon as we see the server set args->csa_cachethis. Signed-off-by: Trond Myklebust --- fs/nfs/callback_proc.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index e0844fdbd9ac..345df6309017 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -381,9 +381,8 @@ validate_seqid(struct nfs4_slot_table *tbl, struct cb_sequenceargs * args) if (args->csa_cachethis == 0) return htonl(NFS4ERR_RETRY_UNCACHED_REP); - /* The ca_maxresponsesize_cached is 0 with no DRC */ - else if (args->csa_cachethis == 1) - return htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE); + /* Liar! We never allowed you to set csa_cachethis != 0 */ + return htonl(NFS4ERR_SEQ_FALSE_RETRY); } /* Wraparound */ @@ -500,6 +499,10 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, cps->slotid = args->csa_slotid; + /* The ca_maxresponsesize_cached is 0 with no DRC */ + if (args->csa_cachethis != 0) + return htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE); + /* * Check for pending referring calls. If a match is found, a * related callback was received before the response to the original -- cgit v1.2.3 From 5f83d86cf531d737ba2ca9c3cc500ff331fbd43e Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 23 Jan 2016 16:57:58 -0500 Subject: NFSv4.x: Fix wraparound issues when validing the callback sequence id We need to make sure that we don't allow args->csa_sequenceid == 0. Signed-off-by: Trond Myklebust --- fs/nfs/callback_proc.c | 43 +++++++++++++++++-------------------------- 1 file changed, 17 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 345df6309017..79c93b3bbfec 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -354,23 +354,15 @@ out: * a single outstanding callback request at a time. */ static __be32 -validate_seqid(struct nfs4_slot_table *tbl, struct cb_sequenceargs * args) +validate_seqid(const struct nfs4_slot_table *tbl, const struct nfs4_slot *slot, + const struct cb_sequenceargs * args) { - struct nfs4_slot *slot; - - dprintk("%s enter. slotid %u seqid %u\n", - __func__, args->csa_slotid, args->csa_sequenceid); + dprintk("%s enter. slotid %u seqid %u, slot table seqid: %u\n", + __func__, args->csa_slotid, args->csa_sequenceid, slot->seq_nr); if (args->csa_slotid > tbl->server_highest_slotid) return htonl(NFS4ERR_BADSLOT); - slot = tbl->slots + args->csa_slotid; - dprintk("%s slot table seqid: %u\n", __func__, slot->seq_nr); - - /* Normal */ - if (likely(args->csa_sequenceid == slot->seq_nr + 1)) - goto out_ok; - /* Replay */ if (args->csa_sequenceid == slot->seq_nr) { dprintk("%s seqid %u is a replay\n", @@ -386,16 +378,14 @@ validate_seqid(struct nfs4_slot_table *tbl, struct cb_sequenceargs * args) } /* Wraparound */ - if (args->csa_sequenceid == 1 && (slot->seq_nr + 1) == 0) { - slot->seq_nr = 1; - goto out_ok; - } + if (unlikely(slot->seq_nr == 0xFFFFFFFFU)) { + if (args->csa_sequenceid == 1) + return htonl(NFS4_OK); + } else if (likely(args->csa_sequenceid == slot->seq_nr + 1)) + return htonl(NFS4_OK); /* Misordered request */ return htonl(NFS4ERR_SEQ_MISORDERED); -out_ok: - tbl->highest_used_slotid = args->csa_slotid; - return htonl(NFS4_OK); } /* @@ -486,6 +476,13 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, goto out_unlock; } + status = validate_seqid(tbl, slot, args); + if (status) + goto out_unlock; + + cps->slotid = args->csa_slotid; + tbl->highest_used_slotid = args->csa_slotid; + memcpy(&res->csr_sessionid, &args->csa_sessionid, sizeof(res->csr_sessionid)); res->csr_sequenceid = args->csa_sequenceid; @@ -493,12 +490,6 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, res->csr_highestslotid = tbl->server_highest_slotid; res->csr_target_highestslotid = tbl->target_highest_slotid; - status = validate_seqid(tbl, args); - if (status) - goto out_unlock; - - cps->slotid = args->csa_slotid; - /* The ca_maxresponsesize_cached is 0 with no DRC */ if (args->csa_cachethis != 0) return htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE); @@ -518,7 +509,7 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, * If CB_SEQUENCE returns an error, then the state of the slot * (sequence ID, cached reply) MUST NOT change. */ - slot->seq_nr++; + slot->seq_nr = args->csa_sequenceid; out_unlock: spin_unlock(&tbl->slot_tbl_lock); -- cgit v1.2.3 From 810d82e6830100615e7481813a862d26ffcff6bd Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 23 Jan 2016 15:18:18 -0500 Subject: NFSv4.x: Allow multiple callbacks in flight Hook the callback channel into the same session management machinery as we use for the forward channel. Signed-off-by: Trond Myklebust --- fs/nfs/callback.h | 3 ++- fs/nfs/callback_proc.c | 14 +++++++++---- fs/nfs/callback_xdr.c | 12 ++++++----- fs/nfs/nfs4session.c | 54 +++++++++++++++++++++++++++++++++++++++----------- fs/nfs/nfs4session.h | 8 ++++++++ 5 files changed, 69 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback.h b/fs/nfs/callback.h index ff8195bd75ea..5fe1cecbf9f0 100644 --- a/fs/nfs/callback.h +++ b/fs/nfs/callback.h @@ -37,10 +37,11 @@ enum nfs4_callback_opnum { OP_CB_ILLEGAL = 10044, }; +struct nfs4_slot; struct cb_process_state { __be32 drc_status; struct nfs_client *clp; - u32 slotid; + struct nfs4_slot *slot; u32 minorversion; struct net *net; }; diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 79c93b3bbfec..efd079d28946 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -367,7 +367,7 @@ validate_seqid(const struct nfs4_slot_table *tbl, const struct nfs4_slot *slot, if (args->csa_sequenceid == slot->seq_nr) { dprintk("%s seqid %u is a replay\n", __func__, args->csa_sequenceid); - if (tbl->highest_used_slotid != NFS4_NO_SLOT) + if (nfs4_test_locked_slot(tbl, slot->slot_nr)) return htonl(NFS4ERR_DELAY); /* Signal process_op to set this error on next op */ if (args->csa_cachethis == 0) @@ -476,12 +476,18 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, goto out_unlock; } + status = htonl(NFS4ERR_BADSLOT); + slot = nfs4_lookup_slot(tbl, args->csa_slotid); + if (IS_ERR(slot)) + goto out_unlock; status = validate_seqid(tbl, slot, args); if (status) goto out_unlock; - - cps->slotid = args->csa_slotid; - tbl->highest_used_slotid = args->csa_slotid; + if (!nfs4_try_to_lock_slot(tbl, slot)) { + status = htonl(NFS4ERR_DELAY); + goto out_unlock; + } + cps->slot = slot; memcpy(&res->csr_sessionid, &args->csa_sessionid, sizeof(res->csr_sessionid)); diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c index 646cdac73488..976c90608e56 100644 --- a/fs/nfs/callback_xdr.c +++ b/fs/nfs/callback_xdr.c @@ -752,7 +752,8 @@ preprocess_nfs41_op(int nop, unsigned int op_nr, struct callback_op **op) return htonl(NFS_OK); } -static void nfs4_callback_free_slot(struct nfs4_session *session) +static void nfs4_callback_free_slot(struct nfs4_session *session, + struct nfs4_slot *slot) { struct nfs4_slot_table *tbl = &session->bc_slot_table; @@ -761,15 +762,17 @@ static void nfs4_callback_free_slot(struct nfs4_session *session) * Let the state manager know callback processing done. * A single slot, so highest used slotid is either 0 or -1 */ - tbl->highest_used_slotid = NFS4_NO_SLOT; + nfs4_free_slot(tbl, slot); nfs4_slot_tbl_drain_complete(tbl); spin_unlock(&tbl->slot_tbl_lock); } static void nfs4_cb_free_slot(struct cb_process_state *cps) { - if (cps->slotid != NFS4_NO_SLOT) - nfs4_callback_free_slot(cps->clp->cl_session); + if (cps->slot) { + nfs4_callback_free_slot(cps->clp->cl_session, cps->slot); + cps->slot = NULL; + } } #else /* CONFIG_NFS_V4_1 */ @@ -893,7 +896,6 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r struct cb_process_state cps = { .drc_status = 0, .clp = NULL, - .slotid = NFS4_NO_SLOT, .net = SVC_NET(rqstp), }; unsigned int nops = 0; diff --git a/fs/nfs/nfs4session.c b/fs/nfs/nfs4session.c index e23366effcfb..332d06e64fa9 100644 --- a/fs/nfs/nfs4session.c +++ b/fs/nfs/nfs4session.c @@ -135,6 +135,43 @@ static struct nfs4_slot *nfs4_find_or_create_slot(struct nfs4_slot_table *tbl, return ERR_PTR(-ENOMEM); } +static void nfs4_lock_slot(struct nfs4_slot_table *tbl, + struct nfs4_slot *slot) +{ + u32 slotid = slot->slot_nr; + + __set_bit(slotid, tbl->used_slots); + if (slotid > tbl->highest_used_slotid || + tbl->highest_used_slotid == NFS4_NO_SLOT) + tbl->highest_used_slotid = slotid; + slot->generation = tbl->generation; +} + +/* + * nfs4_try_to_lock_slot - Given a slot try to allocate it + * + * Note: must be called with the slot_tbl_lock held. + */ +bool nfs4_try_to_lock_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot) +{ + if (nfs4_test_locked_slot(tbl, slot->slot_nr)) + return false; + nfs4_lock_slot(tbl, slot); + return true; +} + +/* + * nfs4_lookup_slot - Find a slot but don't allocate it + * + * Note: must be called with the slot_tbl_lock held. + */ +struct nfs4_slot *nfs4_lookup_slot(struct nfs4_slot_table *tbl, u32 slotid) +{ + if (slotid <= tbl->max_slotid) + return nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT); + return ERR_PTR(-E2BIG); +} + /* * nfs4_alloc_slot - efficiently look for a free slot * @@ -153,18 +190,11 @@ struct nfs4_slot *nfs4_alloc_slot(struct nfs4_slot_table *tbl) __func__, tbl->used_slots[0], tbl->highest_used_slotid, tbl->max_slotid + 1); slotid = find_first_zero_bit(tbl->used_slots, tbl->max_slotid + 1); - if (slotid > tbl->max_slotid) - goto out; - ret = nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT); - if (IS_ERR(ret)) - goto out; - __set_bit(slotid, tbl->used_slots); - if (slotid > tbl->highest_used_slotid || - tbl->highest_used_slotid == NFS4_NO_SLOT) - tbl->highest_used_slotid = slotid; - ret->generation = tbl->generation; - -out: + if (slotid <= tbl->max_slotid) { + ret = nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT); + if (!IS_ERR(ret)) + nfs4_lock_slot(tbl, ret); + } dprintk("<-- %s used_slots=%04lx highest_used=%u slotid=%u\n", __func__, tbl->used_slots[0], tbl->highest_used_slotid, !IS_ERR(ret) ? ret->slot_nr : NFS4_NO_SLOT); diff --git a/fs/nfs/nfs4session.h b/fs/nfs/nfs4session.h index e3ea2c5324d6..5b51298d1d03 100644 --- a/fs/nfs/nfs4session.h +++ b/fs/nfs/nfs4session.h @@ -77,6 +77,8 @@ extern int nfs4_setup_slot_table(struct nfs4_slot_table *tbl, unsigned int max_reqs, const char *queue); extern void nfs4_shutdown_slot_table(struct nfs4_slot_table *tbl); extern struct nfs4_slot *nfs4_alloc_slot(struct nfs4_slot_table *tbl); +extern struct nfs4_slot *nfs4_lookup_slot(struct nfs4_slot_table *tbl, u32 slotid); +extern bool nfs4_try_to_lock_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot); extern void nfs4_free_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot); extern void nfs4_slot_tbl_drain_complete(struct nfs4_slot_table *tbl); bool nfs41_wake_and_assign_slot(struct nfs4_slot_table *tbl, @@ -88,6 +90,12 @@ static inline bool nfs4_slot_tbl_draining(struct nfs4_slot_table *tbl) return !!test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state); } +static inline bool nfs4_test_locked_slot(const struct nfs4_slot_table *tbl, + u32 slotid) +{ + return !!test_bit(slotid, tbl->used_slots); +} + #if defined(CONFIG_NFS_V4_1) extern void nfs41_set_target_slotid(struct nfs4_slot_table *tbl, u32 target_highest_slotid); -- cgit v1.2.3 From e5003b2f6a7554713a84401d1f6ac546ab5a110d Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 1 Feb 2016 12:01:51 -0500 Subject: NFSv4.x: Fix NFS4ERR_RETRY_UNCACHED_REP in nfs4_callback_sequence We need to initialize cb_sequenceres information when reporting a NFS4ERR_RETRY_UNCACHED_REP error, since that will apply to the next operation, not to the CB_SEQUENCE itself. Reported-by: Kinglong Mee Signed-off-by: Trond Myklebust --- fs/nfs/callback_proc.c | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index efd079d28946..618ced381a14 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -464,6 +464,12 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, tbl = &clp->cl_session->bc_slot_table; slot = tbl->slots + args->csa_slotid; + /* Set up res before grabbing the spinlock */ + memcpy(&res->csr_sessionid, &args->csa_sessionid, + sizeof(res->csr_sessionid)); + res->csr_sequenceid = args->csa_sequenceid; + res->csr_slotid = args->csa_slotid; + spin_lock(&tbl->slot_tbl_lock); /* state manager is resetting the session */ if (test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state)) { @@ -480,6 +486,10 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, slot = nfs4_lookup_slot(tbl, args->csa_slotid); if (IS_ERR(slot)) goto out_unlock; + + res->csr_highestslotid = tbl->server_highest_slotid; + res->csr_target_highestslotid = tbl->target_highest_slotid; + status = validate_seqid(tbl, slot, args); if (status) goto out_unlock; @@ -489,13 +499,6 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, } cps->slot = slot; - memcpy(&res->csr_sessionid, &args->csa_sessionid, - sizeof(res->csr_sessionid)); - res->csr_sequenceid = args->csa_sequenceid; - res->csr_slotid = args->csa_slotid; - res->csr_highestslotid = tbl->server_highest_slotid; - res->csr_target_highestslotid = tbl->target_highest_slotid; - /* The ca_maxresponsesize_cached is 0 with no DRC */ if (args->csa_cachethis != 0) return htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE); -- cgit v1.2.3 From d9ddbf5d9225a417076e89ef650dd276165dfd07 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 30 Jan 2016 22:58:24 -0500 Subject: NFSv4.1: nfs4_proc_bind_conn_to_session must iterate over all connections Use the new helper to ensure that nfs4_proc_bind_conn_to_session() is called for all connections. However ensure that we only set the backchannel flag for the connection pointed to by rpc_clnt->cl_xprt. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 51 ++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 48 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 4bfc33ad0563..36471e013a73 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -6783,12 +6783,16 @@ nfs41_same_server_scope(struct nfs41_server_scope *a, } /* - * nfs4_proc_bind_conn_to_session() + * nfs4_proc_bind_one_conn_to_session() * * The 4.1 client currently uses the same TCP connection for the * fore and backchannel. */ -int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred) +static +int nfs4_proc_bind_one_conn_to_session(struct rpc_clnt *clnt, + struct rpc_xprt *xprt, + struct nfs_client *clp, + struct rpc_cred *cred) { int status; struct nfs41_bind_conn_to_session_args args = { @@ -6803,6 +6807,13 @@ int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred .rpc_resp = &res, .rpc_cred = cred, }; + struct rpc_task_setup task_setup_data = { + .rpc_client = clnt, + .rpc_xprt = xprt, + .rpc_message = &msg, + .flags = RPC_TASK_TIMEOUT, + }; + struct rpc_task *task; dprintk("--> %s\n", __func__); @@ -6810,7 +6821,16 @@ int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred if (!(clp->cl_session->flags & SESSION4_BACK_CHAN)) args.dir = NFS4_CDFC4_FORE; - status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT); + /* Do not set the backchannel flag unless this is clnt->cl_xprt */ + if (xprt != rcu_access_pointer(clnt->cl_xprt)) + args.dir = NFS4_CDFC4_FORE; + + task = rpc_run_task(&task_setup_data); + if (!IS_ERR(task)) { + status = task->tk_status; + rpc_put_task(task); + } else + status = PTR_ERR(task); trace_nfs4_bind_conn_to_session(clp, status); if (status == 0) { if (memcmp(res.sessionid.data, @@ -6837,6 +6857,31 @@ out: return status; } +struct rpc_bind_conn_calldata { + struct nfs_client *clp; + struct rpc_cred *cred; +}; + +static int +nfs4_proc_bind_conn_to_session_callback(struct rpc_clnt *clnt, + struct rpc_xprt *xprt, + void *calldata) +{ + struct rpc_bind_conn_calldata *p = calldata; + + return nfs4_proc_bind_one_conn_to_session(clnt, xprt, p->clp, p->cred); +} + +int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred) +{ + struct rpc_bind_conn_calldata data = { + .clp = clp, + .cred = cred, + }; + return rpc_clnt_iterate_for_each_xprt(clp->cl_rpcclient, + nfs4_proc_bind_conn_to_session_callback, &data); +} + /* * Minimum set of SP4_MACH_CRED operations from RFC 5661 in the enforce map * and operations we'd like to see to enable certain features in the allow map -- cgit v1.2.3 From fc821d59209d96e80f30f670a3acac9582dfdefd Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 31 Jan 2016 14:53:08 -0500 Subject: pnfs/NFSv4.1: Add multipath capabilities to pNFS flexfiles servers over NFSv3 This adds multipathing to pNFS over NFSv3 as described in the flexfiles draft spec. Ideally, we'd like to do the same for pNFS files, but the NFSv4.1 protocol requires a call to EXCHANGE_ID in order to test that the connection can do session trunking. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs_nfs.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index 81ac6480f9e7..4aaed890048f 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -606,12 +606,22 @@ static int _nfs4_pnfs_v3_ds_connect(struct nfs_server *mds_srv, dprintk("%s: DS %s: trying address %s\n", __func__, ds->ds_remotestr, da->da_remotestr); - clp = get_v3_ds_connect(mds_srv->nfs_client, + if (!IS_ERR(clp)) { + struct xprt_create xprt_args = { + .ident = XPRT_TRANSPORT_TCP, + .net = clp->cl_net, + .dstaddr = (struct sockaddr *)&da->da_addr, + .addrlen = da->da_addrlen, + .servername = clp->cl_hostname, + }; + /* Add this address as an alias */ + rpc_clnt_add_xprt(clp->cl_rpcclient, &xprt_args, + rpc_clnt_test_and_add_xprt, NULL); + } else + clp = get_v3_ds_connect(mds_srv->nfs_client, (struct sockaddr *)&da->da_addr, da->da_addrlen, IPPROTO_TCP, timeo, retrans, au_flavor); - if (!IS_ERR(clp)) - break; } if (IS_ERR(clp)) { -- cgit v1.2.3 From 02a95dee8cf03a73982f9cabf549c04ae354bed7 Mon Sep 17 00:00:00 2001 From: Andy Adamson Date: Fri, 5 Feb 2016 16:08:37 -0500 Subject: NFS add callback_ops to nfs4_proc_bind_conn_to_session_callback Fix oops when NULL callback_ops pointer accessed in rpc_init_task Signed-off-by: Andy Adamson Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 36471e013a73..dbc3dc2f65cd 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -6782,6 +6782,15 @@ nfs41_same_server_scope(struct nfs41_server_scope *a, return false; } +static void +nfs4_bind_one_conn_to_session_done(struct rpc_task *task, void *calldata) +{ +} + +static const struct rpc_call_ops nfs4_bind_one_conn_to_session_ops = { + .rpc_call_done = &nfs4_bind_one_conn_to_session_done, +}; + /* * nfs4_proc_bind_one_conn_to_session() * @@ -6810,6 +6819,7 @@ int nfs4_proc_bind_one_conn_to_session(struct rpc_clnt *clnt, struct rpc_task_setup task_setup_data = { .rpc_client = clnt, .rpc_xprt = xprt, + .callback_ops = &nfs4_bind_one_conn_to_session_ops, .rpc_message = &msg, .flags = RPC_TASK_TIMEOUT, }; -- cgit v1.2.3 From 4ff79bc7098fab71e5957d48d31d2036c234e506 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 2 Mar 2016 17:35:54 +0100 Subject: nfs: remove nfs4_file_fsync The only difference to nfs_file_fsync is the call to pnfs_sync_inode. But pnfs_sync_inode is just an inline that calls a pNFS layout driver method if CONFIG_PNFS is designed, and thus can be called just fine from the core NFS module. Signed-off-by: Christoph Hellwig Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 8 +++++--- fs/nfs/internal.h | 2 +- fs/nfs/nfs4file.c | 33 +-------------------------------- 3 files changed, 7 insertions(+), 36 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 748bb813b8ec..cdf862188618 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -233,7 +233,7 @@ EXPORT_SYMBOL_GPL(nfs_file_mmap); * nfs_file_write() that a write error occurred, and hence cause it to * fall back to doing a synchronous write. */ -int +static int nfs_file_fsync_commit(struct file *file, loff_t start, loff_t end, int datasync) { struct nfs_open_context *ctx = nfs_file_open_context(file); @@ -263,9 +263,8 @@ nfs_file_fsync_commit(struct file *file, loff_t start, loff_t end, int datasync) out: return ret; } -EXPORT_SYMBOL_GPL(nfs_file_fsync_commit); -static int +int nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) { int ret; @@ -280,6 +279,8 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) break; inode_lock(inode); ret = nfs_file_fsync_commit(file, start, end, datasync); + if (!ret) + ret = pnfs_sync_inode(inode, !!datasync); inode_unlock(inode); /* * If nfs_file_fsync_commit detected a server reboot, then @@ -293,6 +294,7 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) trace_nfs_fsync_exit(inode, ret); return ret; } +EXPORT_SYMBOL_GPL(nfs_file_fsync); /* * Decide whether a read/modify/write cycle may be more efficient diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 9a547aa3ec8e..c41c5f5baf03 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -358,7 +358,7 @@ int nfs_mknod(struct inode *, struct dentry *, umode_t, dev_t); int nfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *); /* file.c */ -int nfs_file_fsync_commit(struct file *, loff_t, loff_t, int); +int nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync); loff_t nfs_file_llseek(struct file *, loff_t, int); ssize_t nfs_file_read(struct kiocb *, struct iov_iter *); ssize_t nfs_file_splice_read(struct file *, loff_t *, struct pipe_inode_info *, diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c index 57ca1c8039c1..22c35abbee9d 100644 --- a/fs/nfs/nfs4file.c +++ b/fs/nfs/nfs4file.c @@ -128,37 +128,6 @@ nfs4_file_flush(struct file *file, fl_owner_t id) return vfs_fsync(file, 0); } -static int -nfs4_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) -{ - int ret; - struct inode *inode = file_inode(file); - - trace_nfs_fsync_enter(inode); - - nfs_inode_dio_wait(inode); - do { - ret = filemap_write_and_wait_range(inode->i_mapping, start, end); - if (ret != 0) - break; - inode_lock(inode); - ret = nfs_file_fsync_commit(file, start, end, datasync); - if (!ret) - ret = pnfs_sync_inode(inode, !!datasync); - inode_unlock(inode); - /* - * If nfs_file_fsync_commit detected a server reboot, then - * resend all dirty pages that might have been covered by - * the NFS_CONTEXT_RESEND_WRITES flag - */ - start = 0; - end = LLONG_MAX; - } while (ret == -EAGAIN); - - trace_nfs_fsync_exit(inode, ret); - return ret; -} - #ifdef CONFIG_NFS_V4_2 static loff_t nfs4_file_llseek(struct file *filep, loff_t offset, int whence) { @@ -266,7 +235,7 @@ const struct file_operations nfs4_file_operations = { .open = nfs4_file_open, .flush = nfs4_file_flush, .release = nfs_file_release, - .fsync = nfs4_file_fsync, + .fsync = nfs_file_fsync, .lock = nfs_lock, .flock = nfs_flock, .splice_read = nfs_file_splice_read, -- cgit v1.2.3 From 95d9f6c3edd60aab1bb163f081262645f8bb8dc0 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 2 Mar 2016 17:35:55 +0100 Subject: nfs: remove nfs_inode_dio_wait Just call inode_dio_wait directly instead of through a pointless wrapper. Signed-off-by: Christoph Hellwig Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 4 ++-- fs/nfs/inode.c | 2 +- fs/nfs/internal.h | 4 ---- 3 files changed, 3 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index cdf862188618..89bf093d342a 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -272,7 +272,7 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) trace_nfs_fsync_enter(inode); - nfs_inode_dio_wait(inode); + inode_dio_wait(inode); do { ret = filemap_write_and_wait_range(inode->i_mapping, start, end); if (ret != 0) @@ -370,7 +370,7 @@ start: /* * Wait for O_DIRECT to complete */ - nfs_inode_dio_wait(mapping->host); + inode_dio_wait(mapping->host); page = grab_cache_page_write_begin(mapping, index, flags); if (!page) diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 86faecf8f328..33d18c411905 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -141,7 +141,7 @@ void nfs_evict_inode(struct inode *inode) int nfs_sync_inode(struct inode *inode) { - nfs_inode_dio_wait(inode); + inode_dio_wait(inode); return nfs_wb_all(inode); } EXPORT_SYMBOL_GPL(nfs_sync_inode); diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index c41c5f5baf03..565f8135ae1f 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -515,10 +515,6 @@ extern int nfs_sillyrename(struct inode *dir, struct dentry *dentry); /* direct.c */ void nfs_init_cinfo_from_dreq(struct nfs_commit_info *cinfo, struct nfs_direct_req *dreq); -static inline void nfs_inode_dio_wait(struct inode *inode) -{ - inode_dio_wait(inode); -} extern ssize_t nfs_dreq_bytes_left(struct nfs_direct_req *dreq); /* nfs4proc.c */ -- cgit v1.2.3 From 849dc3244c916545790bfb9055625a3719061c92 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Wed, 24 Feb 2016 15:28:29 -0500 Subject: nfs4: nfs4_ff_layout_prepare_ds should return NULL if connection failed I hit the following oops out of the blue while testing with flexfiles: BUG: unable to handle kernel NULL pointer dereference at 00000000000000e8 IP: [] nfs4_ff_find_or_create_ds_client+0x48/0x50 [nfs_layout_flexfiles] PGD 44031067 PUD 5062d067 PMD 0 Oops: 0000 [#1] SMP Modules linked in: nfsv3 nfs_layout_flexfiles tun rpcsec_gss_krb5 nfsv4 dns_resolver nfs fscache dcdbas nfsd auth_rpcgss nfs_acl lockd grace sunrpc ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 xt_conntrack ebtable_nat ebtable_broute bridge stp llc ebtable_filter ebtables ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 ip6table_mangle ip6table_security ip6table_raw ip6table_filter ip6_tables iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle iptable_security iptable_raw bonding ipmi_devintf ipmi_msghandler snd_hda_codec_generic virtio_balloon ppdev snd_hda_intel snd_hda_controller snd_hda_codec iosf_mbi crct10dif_pclmul crc32_pclmul ghash_clmulni_intel snd_hda_core parport_pc snd_hwdep parport snd_seq snd_seq_device snd_pcm snd_timer acpi_cpufreq snd soundcore i2c_piix4 xfs libcrc32c joydev virtio_net virtio_console qxl drm_kms_helper ttm crc32c_intel drm virtio_pci serio_raw ata_generic virtio_ring virtio pata_acpi CPU: 0 PID: 19138 Comm: test5 Not tainted 4.1.9-100.pd.90.el7.x86_64 #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.8.2-20150714_191134- 04/01/2014 task: ffff88007b70cf00 ti: ffff88004cc44000 task.ti: ffff88004cc44000 RIP: 0010:[] [] nfs4_ff_find_or_create_ds_client+0x48/0x50 [nfs_layout_flexfiles] RSP: 0018:ffff88004cc47890 EFLAGS: 00010246 RAX: 0000000000000003 RBX: ffff880050932300 RCX: ffff88006978f488 RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff88003e0e8540 RBP: ffff88004cc47908 R08: 0000000000000000 R09: 0000000000000000 R10: ffff88007ff8c758 R11: 0000000000000005 R12: ffff88003e0e8540 R13: 0000000000000000 R14: ffff88006978f488 R15: ffff88004431cc80 FS: 00007fea40c7c740(0000) GS:ffff88007fc00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00000000000000e8 CR3: 0000000044318000 CR4: 00000000000406f0 Stack: ffffffffa048c934 ffff880050932310 0000000100000001 ffff88006978f510 ffff88006978f3c8 ffff88003e56cd90 ffff88004cc479d0 00000020a052aff0 000000000004b000 ffff88004cc47908 ffff880050932300 ffff88004cc479d0 Call Trace: [] ? ff_layout_write_pagelist+0x64/0x220 [nfs_layout_flexfiles] [] pnfs_generic_pg_writepages+0xaf/0x1b0 [nfsv4] [] nfs_pageio_doio+0x27/0x60 [nfs] [] nfs_pageio_complete_mirror+0x54/0xa0 [nfs] [] nfs_pageio_complete+0x2d/0x90 [nfs] [] nfs_writepage_locked+0x8d/0xe0 [nfs] [] ? page_referenced_one+0x1a0/0x1a0 [] nfs_wb_single_page+0xf7/0x190 [nfs] [] nfs_launder_page+0x41/0x90 [nfs] [] invalidate_inode_pages2_range+0x340/0x3a0 [] invalidate_inode_pages2+0x17/0x20 [] nfs_release+0x9e/0xb0 [nfs] [] nfs_file_release+0x3d/0x60 [nfs] [] __fput+0xdc/0x1e0 [] ____fput+0xe/0x10 [] task_work_run+0xa7/0xe0 [] get_signal+0x565/0x600 [] ? __filemap_fdatawrite_range+0x65/0x90 [] do_signal+0x37/0x730 [] ? nfs4_file_fsync+0x81/0x150 [nfsv4] [] ? vfs_fsync_range+0x3b/0xb0 [] ? __audit_syscall_exit+0x1e6/0x280 [] do_notify_resume+0x5f/0xa0 [] int_signal+0x12/0x17 Code: 48 8b 40 70 8b 00 83 f8 03 74 20 83 f8 04 75 13 55 48 89 ce 48 89 d7 48 89 e5 e8 14 0f 0e 00 5d c3 66 90 0f 0b 66 0f 1f 44 00 00 <48> 8b 82 e8 00 00 00 c3 66 66 66 66 90 55 48 89 e5 41 57 41 56 RIP [] nfs4_ff_find_or_create_ds_client+0x48/0x50 [nfs_layout_flexfiles] RSP CR2: 00000000000000e8 When the DS connection attempt fails, nfs4_ff_layout_prepare_ds marks it for the error but then just returns the ds as if it were usable. The comments though say: /* Upon return, either ds is connected, or ds is NULL */ Ensure that we set the return pointer to NULL in the event that the connection attempt fails. Signed-off-by: Jeff Layton Signed-off-by: Trond Myklebust --- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index eb370460ce20..add0e5a70bd6 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -418,6 +418,8 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, pnfs_error_mark_layout_for_return(ino, lseg); } else pnfs_error_mark_layout_for_return(ino, lseg); + ds = NULL; + goto out; } out_update_creds: if (ff_layout_update_mirror_cred(mirror, ds)) -- cgit v1.2.3 From f35592a97460f57d1089fd674176e9f34ba311f2 Mon Sep 17 00:00:00 2001 From: Kinglong Mee Date: Sat, 13 Feb 2016 21:51:31 +0800 Subject: nfs/blocklayout: make sure making a aligned read request Only treat write goes up to the inode size as aligned request, because it always write PAGE_CACHE_SIZE, but read a dynamic size. Signed-off-by: Kinglong Mee Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/blocklayout.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c index ddd0138f410c..8bc870e4c467 100644 --- a/fs/nfs/blocklayout/blocklayout.c +++ b/fs/nfs/blocklayout/blocklayout.c @@ -743,7 +743,7 @@ bl_set_layoutdriver(struct nfs_server *server, const struct nfs_fh *fh) static bool is_aligned_req(struct nfs_pageio_descriptor *pgio, - struct nfs_page *req, unsigned int alignment) + struct nfs_page *req, unsigned int alignment, bool is_write) { /* * Always accept buffered writes, higher layers take care of the @@ -758,7 +758,8 @@ is_aligned_req(struct nfs_pageio_descriptor *pgio, if (IS_ALIGNED(req->wb_bytes, alignment)) return true; - if (req_offset(req) + req->wb_bytes == i_size_read(pgio->pg_inode)) { + if (is_write && + (req_offset(req) + req->wb_bytes == i_size_read(pgio->pg_inode))) { /* * If the write goes up to the inode size, just write * the full page. Data past the inode size is @@ -775,7 +776,7 @@ is_aligned_req(struct nfs_pageio_descriptor *pgio, static void bl_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req) { - if (!is_aligned_req(pgio, req, SECTOR_SIZE)) { + if (!is_aligned_req(pgio, req, SECTOR_SIZE, false)) { nfs_pageio_reset_read_mds(pgio); return; } @@ -791,7 +792,7 @@ static size_t bl_pg_test_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req) { - if (!is_aligned_req(pgio, req, SECTOR_SIZE)) + if (!is_aligned_req(pgio, req, SECTOR_SIZE, false)) return 0; return pnfs_generic_pg_test(pgio, prev, req); } @@ -824,7 +825,7 @@ bl_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req) { u64 wb_size; - if (!is_aligned_req(pgio, req, PAGE_SIZE)) { + if (!is_aligned_req(pgio, req, PAGE_SIZE, true)) { nfs_pageio_reset_write_mds(pgio); return; } @@ -846,7 +847,7 @@ static size_t bl_pg_test_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req) { - if (!is_aligned_req(pgio, req, PAGE_SIZE)) + if (!is_aligned_req(pgio, req, PAGE_SIZE, true)) return 0; return pnfs_generic_pg_test(pgio, prev, req); } -- cgit v1.2.3