diff options
Diffstat (limited to 'net')
38 files changed, 1036 insertions, 1311 deletions
diff --git a/net/atm/pppoatm.c b/net/atm/pppoatm.c index 8c93267ce969..c4e09846d1de 100644 --- a/net/atm/pppoatm.c +++ b/net/atm/pppoatm.c @@ -252,7 +252,7 @@ static int pppoatm_may_send(struct pppoatm_vcc *pvcc, int size) * we need to ensure there's a memory barrier after it. The bit * *must* be set before we do the atomic_inc() on pvcc->inflight. * There's no smp_mb__after_set_bit(), so it's this or abuse - * smp_mb__after_clear_bit(). + * smp_mb__after_atomic(). */ test_and_set_bit(BLOCKED, &pvcc->blocked); diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 15010a230b6d..682f33a38366 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -45,7 +45,7 @@ static void hci_cc_inquiry_cancel(struct hci_dev *hdev, struct sk_buff *skb) return; clear_bit(HCI_INQUIRY, &hdev->flags); - smp_mb__after_clear_bit(); /* wake_up_bit advises about this barrier */ + smp_mb__after_atomic(); /* wake_up_bit advises about this barrier */ wake_up_bit(&hdev->flags, HCI_INQUIRY); hci_conn_check_pending(hdev); @@ -1768,7 +1768,7 @@ static void hci_inquiry_complete_evt(struct hci_dev *hdev, struct sk_buff *skb) if (!test_and_clear_bit(HCI_INQUIRY, &hdev->flags)) return; - smp_mb__after_clear_bit(); /* wake_up_bit advises about this barrier */ + smp_mb__after_atomic(); /* wake_up_bit advises about this barrier */ wake_up_bit(&hdev->flags, HCI_INQUIRY); if (!test_bit(HCI_MGMT, &hdev->dev_flags)) diff --git a/net/ceph/pagevec.c b/net/ceph/pagevec.c index 815a2249cfa9..555013034f7a 100644 --- a/net/ceph/pagevec.c +++ b/net/ceph/pagevec.c @@ -53,7 +53,10 @@ void ceph_put_page_vector(struct page **pages, int num_pages, bool dirty) set_page_dirty_lock(pages[i]); put_page(pages[i]); } - kfree(pages); + if (is_vmalloc_addr(pages)) + vfree(pages); + else + kfree(pages); } EXPORT_SYMBOL(ceph_put_page_vector); @@ -165,36 +168,6 @@ void ceph_copy_from_page_vector(struct page **pages, EXPORT_SYMBOL(ceph_copy_from_page_vector); /* - * copy user data from a page vector into a user pointer - */ -int ceph_copy_page_vector_to_user(struct page **pages, - void __user *data, - loff_t off, size_t len) -{ - int i = 0; - int po = off & ~PAGE_CACHE_MASK; - int left = len; - int l, bad; - - while (left > 0) { - l = min_t(int, left, PAGE_CACHE_SIZE-po); - bad = copy_to_user(data, page_address(pages[i]) + po, l); - if (bad == l) - return -EFAULT; - data += l - bad; - left -= l - bad; - if (po) { - po += l - bad; - if (po == PAGE_CACHE_SIZE) - po = 0; - } - i++; - } - return len; -} -EXPORT_SYMBOL(ceph_copy_page_vector_to_user); - -/* * Zero an extent within a page vector. Offset is relative to the * start of the first page. */ diff --git a/net/core/dev.c b/net/core/dev.c index fb8b0546485b..8908a68db449 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1326,7 +1326,7 @@ static int __dev_close_many(struct list_head *head) * dev->stop() will invoke napi_disable() on all of it's * napi_struct instances on this device. */ - smp_mb__after_clear_bit(); /* Commit netif_running(). */ + smp_mb__after_atomic(); /* Commit netif_running(). */ } dev_deactivate_many(head); @@ -3356,7 +3356,7 @@ static void net_tx_action(struct softirq_action *h) root_lock = qdisc_lock(q); if (spin_trylock(root_lock)) { - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(__QDISC_STATE_SCHED, &q->state); qdisc_run(q); @@ -3366,7 +3366,7 @@ static void net_tx_action(struct softirq_action *h) &q->state)) { __netif_reschedule(q); } else { - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(__QDISC_STATE_SCHED, &q->state); } @@ -4258,7 +4258,7 @@ void __napi_complete(struct napi_struct *n) BUG_ON(n->gro_list); list_del(&n->poll_list); - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(NAPI_STATE_SCHED, &n->state); } EXPORT_SYMBOL(__napi_complete); diff --git a/net/core/link_watch.c b/net/core/link_watch.c index 9c3a839322ba..bd0767e6b2b3 100644 --- a/net/core/link_watch.c +++ b/net/core/link_watch.c @@ -147,7 +147,7 @@ static void linkwatch_do_dev(struct net_device *dev) * Make sure the above read is complete since it can be * rewritten as soon as we clear the bit below. */ - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); /* We are about to handle this device, * so new events can be accepted diff --git a/net/core/netclassid_cgroup.c b/net/core/netclassid_cgroup.c index 22931e1b99b4..30d903b19c62 100644 --- a/net/core/netclassid_cgroup.c +++ b/net/core/netclassid_cgroup.c @@ -42,7 +42,7 @@ cgrp_css_alloc(struct cgroup_subsys_state *parent_css) static int cgrp_css_online(struct cgroup_subsys_state *css) { struct cgroup_cls_state *cs = css_cls_state(css); - struct cgroup_cls_state *parent = css_cls_state(css_parent(css)); + struct cgroup_cls_state *parent = css_cls_state(css->parent); if (parent) cs->classid = parent->classid; diff --git a/net/core/netprio_cgroup.c b/net/core/netprio_cgroup.c index 3825f669147b..2f385b9bccc0 100644 --- a/net/core/netprio_cgroup.c +++ b/net/core/netprio_cgroup.c @@ -140,7 +140,7 @@ cgrp_css_alloc(struct cgroup_subsys_state *parent_css) static int cgrp_css_online(struct cgroup_subsys_state *css) { - struct cgroup_subsys_state *parent_css = css_parent(css); + struct cgroup_subsys_state *parent_css = css->parent; struct net_device *dev; int ret = 0; @@ -185,15 +185,15 @@ static int read_priomap(struct seq_file *sf, void *v) return 0; } -static int write_priomap(struct cgroup_subsys_state *css, struct cftype *cft, - char *buffer) +static ssize_t write_priomap(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) { char devname[IFNAMSIZ + 1]; struct net_device *dev; u32 prio; int ret; - if (sscanf(buffer, "%"__stringify(IFNAMSIZ)"s %u", devname, &prio) != 2) + if (sscanf(buf, "%"__stringify(IFNAMSIZ)"s %u", devname, &prio) != 2) return -EINVAL; dev = dev_get_by_name(&init_net, devname); @@ -202,11 +202,11 @@ static int write_priomap(struct cgroup_subsys_state *css, struct cftype *cft, rtnl_lock(); - ret = netprio_set_prio(css, dev, prio); + ret = netprio_set_prio(of_css(of), dev, prio); rtnl_unlock(); dev_put(dev); - return ret; + return ret ?: nbytes; } static int update_netprio(const void *v, struct file *file, unsigned n) @@ -239,7 +239,7 @@ static struct cftype ss_files[] = { { .name = "ifpriomap", .seq_show = read_priomap, - .write_string = write_priomap, + .write = write_priomap, }, { } /* terminate */ }; diff --git a/net/ipv4/inetpeer.c b/net/ipv4/inetpeer.c index 48f424465112..56cd458a1b8c 100644 --- a/net/ipv4/inetpeer.c +++ b/net/ipv4/inetpeer.c @@ -522,7 +522,7 @@ EXPORT_SYMBOL_GPL(inet_getpeer); void inet_putpeer(struct inet_peer *p) { p->dtime = (__u32)jiffies; - smp_mb__before_atomic_dec(); + smp_mb__before_atomic(); atomic_dec(&p->refcnt); } EXPORT_SYMBOL_GPL(inet_putpeer); diff --git a/net/ipv4/tcp_memcontrol.c b/net/ipv4/tcp_memcontrol.c index d4f015ad6c84..f7a2ec3ac584 100644 --- a/net/ipv4/tcp_memcontrol.c +++ b/net/ipv4/tcp_memcontrol.c @@ -102,17 +102,19 @@ static int tcp_update_limit(struct mem_cgroup *memcg, u64 val) return 0; } -static int tcp_cgroup_write(struct cgroup_subsys_state *css, struct cftype *cft, - char *buffer) +static ssize_t tcp_cgroup_write(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) { - struct mem_cgroup *memcg = mem_cgroup_from_css(css); + struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); unsigned long long val; int ret = 0; - switch (cft->private) { + buf = strstrip(buf); + + switch (of_cft(of)->private) { case RES_LIMIT: /* see memcontrol.c */ - ret = res_counter_memparse_write_strategy(buffer, &val); + ret = res_counter_memparse_write_strategy(buf, &val); if (ret) break; ret = tcp_update_limit(memcg, val); @@ -121,7 +123,7 @@ static int tcp_cgroup_write(struct cgroup_subsys_state *css, struct cftype *cft, ret = -EINVAL; break; } - return ret; + return ret ?: nbytes; } static u64 tcp_read_stat(struct mem_cgroup *memcg, int type, u64 default_val) @@ -168,17 +170,18 @@ static u64 tcp_cgroup_read(struct cgroup_subsys_state *css, struct cftype *cft) return val; } -static int tcp_cgroup_reset(struct cgroup_subsys_state *css, unsigned int event) +static ssize_t tcp_cgroup_reset(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) { struct mem_cgroup *memcg; struct cg_proto *cg_proto; - memcg = mem_cgroup_from_css(css); + memcg = mem_cgroup_from_css(of_css(of)); cg_proto = tcp_prot.proto_cgroup(memcg); if (!cg_proto) - return 0; + return nbytes; - switch (event) { + switch (of_cft(of)->private) { case RES_MAX_USAGE: res_counter_reset_max(&cg_proto->memory_allocated); break; @@ -187,13 +190,13 @@ static int tcp_cgroup_reset(struct cgroup_subsys_state *css, unsigned int event) break; } - return 0; + return nbytes; } static struct cftype tcp_files[] = { { .name = "kmem.tcp.limit_in_bytes", - .write_string = tcp_cgroup_write, + .write = tcp_cgroup_write, .read_u64 = tcp_cgroup_read, .private = RES_LIMIT, }, @@ -205,13 +208,13 @@ static struct cftype tcp_files[] = { { .name = "kmem.tcp.failcnt", .private = RES_FAILCNT, - .trigger = tcp_cgroup_reset, + .write = tcp_cgroup_reset, .read_u64 = tcp_cgroup_read, }, { .name = "kmem.tcp.max_usage_in_bytes", .private = RES_MAX_USAGE, - .trigger = tcp_cgroup_reset, + .write = tcp_cgroup_reset, .read_u64 = tcp_cgroup_read, }, { } /* terminate */ diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 12d6016bdd9a..2d340bd2cd3d 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1930,10 +1930,8 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle, /* It is possible TX completion already happened * before we set TSQ_THROTTLED, so we must * test again the condition. - * We abuse smp_mb__after_clear_bit() because - * there is no smp_mb__after_set_bit() yet */ - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); if (atomic_read(&sk->sk_wmem_alloc) > limit) break; } diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index 75421f2ba8be..1f4f954c4b47 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -914,7 +914,7 @@ void nf_conntrack_free(struct nf_conn *ct) nf_ct_ext_destroy(ct); nf_ct_ext_free(ct); kmem_cache_free(net->ct.nf_conntrack_cachep, ct); - smp_mb__before_atomic_dec(); + smp_mb__before_atomic(); atomic_dec(&net->ct.count); } EXPORT_SYMBOL_GPL(nf_conntrack_free); diff --git a/net/netfilter/xt_repldata.h b/net/netfilter/xt_repldata.h index 6efe4e5a81c6..8fd324116e6f 100644 --- a/net/netfilter/xt_repldata.h +++ b/net/netfilter/xt_repldata.h @@ -5,23 +5,35 @@ * they serve as the hanging-off data accessed through repl.data[]. */ +/* tbl has the following structure equivalent, but is C99 compliant: + * struct { + * struct type##_replace repl; + * struct type##_standard entries[nhooks]; + * struct type##_error term; + * } *tbl; + */ + #define xt_alloc_initial_table(type, typ2) ({ \ unsigned int hook_mask = info->valid_hooks; \ unsigned int nhooks = hweight32(hook_mask); \ unsigned int bytes = 0, hooknum = 0, i = 0; \ struct { \ struct type##_replace repl; \ - struct type##_standard entries[nhooks]; \ - struct type##_error term; \ - } *tbl = kzalloc(sizeof(*tbl), GFP_KERNEL); \ + struct type##_standard entries[]; \ + } *tbl; \ + struct type##_error *term; \ + size_t term_offset = (offsetof(typeof(*tbl), entries[nhooks]) + \ + __alignof__(*term) - 1) & ~(__alignof__(*term) - 1); \ + tbl = kzalloc(term_offset + sizeof(*term), GFP_KERNEL); \ if (tbl == NULL) \ return NULL; \ + term = (struct type##_error *)&(((char *)tbl)[term_offset]); \ strncpy(tbl->repl.name, info->name, sizeof(tbl->repl.name)); \ - tbl->term = (struct type##_error)typ2##_ERROR_INIT; \ + *term = (struct type##_error)typ2##_ERROR_INIT; \ tbl->repl.valid_hooks = hook_mask; \ tbl->repl.num_entries = nhooks + 1; \ tbl->repl.size = nhooks * sizeof(struct type##_standard) + \ - sizeof(struct type##_error); \ + sizeof(struct type##_error); \ for (; hook_mask != 0; hook_mask >>= 1, ++hooknum) { \ if (!(hook_mask & 1)) \ continue; \ diff --git a/net/rds/ib_recv.c b/net/rds/ib_recv.c index b7ebe23cdedf..d67de453c35a 100644 --- a/net/rds/ib_recv.c +++ b/net/rds/ib_recv.c @@ -598,7 +598,7 @@ static void rds_ib_set_ack(struct rds_ib_connection *ic, u64 seq, { atomic64_set(&ic->i_ack_next, seq); if (ack_required) { - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); set_bit(IB_ACK_REQUESTED, &ic->i_ack_flags); } } @@ -606,7 +606,7 @@ static void rds_ib_set_ack(struct rds_ib_connection *ic, u64 seq, static u64 rds_ib_get_ack(struct rds_ib_connection *ic) { clear_bit(IB_ACK_REQUESTED, &ic->i_ack_flags); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); return atomic64_read(&ic->i_ack_next); } diff --git a/net/rds/iw_recv.c b/net/rds/iw_recv.c index 45033358358e..aa8bf6786008 100644 --- a/net/rds/iw_recv.c +++ b/net/rds/iw_recv.c @@ -429,7 +429,7 @@ static void rds_iw_set_ack(struct rds_iw_connection *ic, u64 seq, { atomic64_set(&ic->i_ack_next, seq); if (ack_required) { - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); set_bit(IB_ACK_REQUESTED, &ic->i_ack_flags); } } @@ -437,7 +437,7 @@ static void rds_iw_set_ack(struct rds_iw_connection *ic, u64 seq, static u64 rds_iw_get_ack(struct rds_iw_connection *ic) { clear_bit(IB_ACK_REQUESTED, &ic->i_ack_flags); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); return atomic64_read(&ic->i_ack_next); } diff --git a/net/rds/send.c b/net/rds/send.c index a82fb660ec00..23718160d71e 100644 --- a/net/rds/send.c +++ b/net/rds/send.c @@ -107,7 +107,7 @@ static int acquire_in_xmit(struct rds_connection *conn) static void release_in_xmit(struct rds_connection *conn) { clear_bit(RDS_IN_XMIT, &conn->c_flags); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); /* * We don't use wait_on_bit()/wake_up_bit() because our waking is in a * hot path and finding waiters is very rare. We don't want to walk @@ -661,7 +661,7 @@ void rds_send_drop_acked(struct rds_connection *conn, u64 ack, /* order flag updates with spin locks */ if (!list_empty(&list)) - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); spin_unlock_irqrestore(&conn->c_lock, flags); @@ -691,7 +691,7 @@ void rds_send_drop_to(struct rds_sock *rs, struct sockaddr_in *dest) } /* order flag updates with the rs lock */ - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); spin_unlock_irqrestore(&rs->rs_lock, flags); diff --git a/net/rds/tcp_send.c b/net/rds/tcp_send.c index 81cf5a4c5e40..53b17ca0dff5 100644 --- a/net/rds/tcp_send.c +++ b/net/rds/tcp_send.c @@ -93,7 +93,7 @@ int rds_tcp_xmit(struct rds_connection *conn, struct rds_message *rm, rm->m_ack_seq = tc->t_last_sent_nxt + sizeof(struct rds_header) + be32_to_cpu(rm->m_inc.i_hdr.h_len) - 1; - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); set_bit(RDS_MSG_HAS_ACK_SEQ, &rm->m_flags); tc->t_last_expected_una = rm->m_ack_seq + 1; diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c index 5285ead196c0..247e973544bf 100644 --- a/net/sunrpc/auth.c +++ b/net/sunrpc/auth.c @@ -296,7 +296,7 @@ static void rpcauth_unhash_cred_locked(struct rpc_cred *cred) { hlist_del_rcu(&cred->cr_hash); - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(RPCAUTH_CRED_HASHED, &cred->cr_flags); } diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c index 36e431ee1c90..b6e440baccc3 100644 --- a/net/sunrpc/auth_gss/auth_gss.c +++ b/net/sunrpc/auth_gss/auth_gss.c @@ -143,7 +143,7 @@ gss_cred_set_ctx(struct rpc_cred *cred, struct gss_cl_ctx *ctx) gss_get_ctx(ctx); rcu_assign_pointer(gss_cred->gc_ctx, ctx); set_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags); - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(RPCAUTH_CRED_NEW, &cred->cr_flags); } diff --git a/net/sunrpc/auth_gss/gss_mech_switch.c b/net/sunrpc/auth_gss/gss_mech_switch.c index 27ce26240932..92d5ab99fbf3 100644 --- a/net/sunrpc/auth_gss/gss_mech_switch.c +++ b/net/sunrpc/auth_gss/gss_mech_switch.c @@ -218,10 +218,8 @@ static struct gss_api_mech *_gss_mech_get_by_pseudoflavor(u32 pseudoflavor) spin_lock(®istered_mechs_lock); list_for_each_entry(pos, ®istered_mechs, gm_list) { - if (!mech_supports_pseudoflavor(pos, pseudoflavor)) { - module_put(pos->gm_owner); + if (!mech_supports_pseudoflavor(pos, pseudoflavor)) continue; - } if (try_module_get(pos->gm_owner)) gm = pos; break; diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c index 0f73f4507746..4ce5eccec1f6 100644 --- a/net/sunrpc/auth_gss/svcauth_gss.c +++ b/net/sunrpc/auth_gss/svcauth_gss.c @@ -1503,6 +1503,7 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp) if (unwrap_integ_data(rqstp, &rqstp->rq_arg, gc->gc_seq, rsci->mechctx)) goto garbage_args; + rqstp->rq_auth_slack = RPC_MAX_AUTH_SIZE; break; case RPC_GSS_SVC_PRIVACY: /* placeholders for length and seq. number: */ @@ -1511,6 +1512,7 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp) if (unwrap_priv_data(rqstp, &rqstp->rq_arg, gc->gc_seq, rsci->mechctx)) goto garbage_args; + rqstp->rq_auth_slack = RPC_MAX_AUTH_SIZE * 2; break; default: goto auth_err; diff --git a/net/sunrpc/backchannel_rqst.c b/net/sunrpc/backchannel_rqst.c index 3513d559bc45..9761a0da964d 100644 --- a/net/sunrpc/backchannel_rqst.c +++ b/net/sunrpc/backchannel_rqst.c @@ -244,10 +244,10 @@ void xprt_free_bc_request(struct rpc_rqst *req) dprintk("RPC: free backchannel req=%p\n", req); req->rq_connect_cookie = xprt->connect_cookie - 1; - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); WARN_ON_ONCE(!test_bit(RPC_BC_PA_IN_USE, &req->rq_bc_pa_state)); clear_bit(RPC_BC_PA_IN_USE, &req->rq_bc_pa_state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); if (!xprt_need_to_requeue(xprt)) { /* diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index ae333c1845bb..066362141133 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -374,7 +374,7 @@ void sunrpc_destroy_cache_detail(struct cache_detail *cd) } return; out: - printk(KERN_ERR "nfsd: failed to unregister %s cache\n", cd->name); + printk(KERN_ERR "RPC: failed to unregister %s cache\n", cd->name); } EXPORT_SYMBOL_GPL(sunrpc_destroy_cache_detail); diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c index 25578afe1548..c0365c14b858 100644 --- a/net/sunrpc/sched.c +++ b/net/sunrpc/sched.c @@ -832,7 +832,8 @@ static void rpc_async_schedule(struct work_struct *work) * @size: requested byte size * * To prevent rpciod from hanging, this allocator never sleeps, - * returning NULL if the request cannot be serviced immediately. + * returning NULL and suppressing warning if the request cannot be serviced + * immediately. * The caller can arrange to sleep in a way that is safe for rpciod. * * Most requests are 'small' (under 2KiB) and can be serviced from a @@ -845,7 +846,7 @@ static void rpc_async_schedule(struct work_struct *work) void *rpc_malloc(struct rpc_task *task, size_t size) { struct rpc_buffer *buf; - gfp_t gfp = GFP_NOWAIT; + gfp_t gfp = GFP_NOWAIT | __GFP_NOWARN; if (RPC_IS_SWAPPER(task)) gfp |= __GFP_MEMALLOC; diff --git a/net/sunrpc/sunrpc.h b/net/sunrpc/sunrpc.h index 14c9f6d1c5ff..f2b7cb540e61 100644 --- a/net/sunrpc/sunrpc.h +++ b/net/sunrpc/sunrpc.h @@ -43,6 +43,19 @@ static inline int rpc_reply_expected(struct rpc_task *task) (task->tk_msg.rpc_proc->p_decode != NULL); } +static inline int sock_is_loopback(struct sock *sk) +{ + struct dst_entry *dst; + int loopback = 0; + rcu_read_lock(); + dst = rcu_dereference(sk->sk_dst_cache); + if (dst && dst->dev && + (dst->dev->features & NETIF_F_LOOPBACK)) + loopback = 1; + rcu_read_unlock(); + return loopback; +} + int svc_send_common(struct socket *sock, struct xdr_buf *xdr, struct page *headpage, unsigned long headoffset, struct page *tailpage, unsigned long tailoffset); diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index 06c6ff0cb911..b4737fbdec13 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -597,6 +597,7 @@ static int svc_alloc_arg(struct svc_rqst *rqstp) } rqstp->rq_pages[i] = p; } + rqstp->rq_page_end = &rqstp->rq_pages[i]; rqstp->rq_pages[i++] = NULL; /* this might be seen in nfs_read_actor */ /* Make arg->head point to first page and arg->pages point to rest */ @@ -730,6 +731,8 @@ static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt) newxpt = xprt->xpt_ops->xpo_accept(xprt); if (newxpt) svc_add_new_temp_xprt(serv, newxpt); + else + module_put(xprt->xpt_class->xcl_owner); } else if (xprt->xpt_ops->xpo_has_wspace(xprt)) { /* XPT_DATA|XPT_DEFERRED case: */ dprintk("svc: server %p, pool %u, transport %p, inuse=%d\n", @@ -793,7 +796,7 @@ int svc_recv(struct svc_rqst *rqstp, long timeout) clear_bit(XPT_OLD, &xprt->xpt_flags); - rqstp->rq_secure = svc_port_is_privileged(svc_addr(rqstp)); + rqstp->rq_secure = xprt->xpt_ops->xpo_secure_port(rqstp); rqstp->rq_chandle.defer = svc_defer; if (serv->sv_stats) diff --git a/net/sunrpc/svcauth.c b/net/sunrpc/svcauth.c index 2af7b0cba43a..79c0f3459b5c 100644 --- a/net/sunrpc/svcauth.c +++ b/net/sunrpc/svcauth.c @@ -54,6 +54,8 @@ svc_authenticate(struct svc_rqst *rqstp, __be32 *authp) } spin_unlock(&authtab_lock); + rqstp->rq_auth_slack = 0; + rqstp->rq_authop = aops; return aops->accept(rqstp, authp); } diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 43bcb4699d69..b507cd327d9b 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -400,6 +400,12 @@ static void svc_sock_setbufsize(struct socket *sock, unsigned int snd, release_sock(sock->sk); #endif } + +static int svc_sock_secure_port(struct svc_rqst *rqstp) +{ + return svc_port_is_privileged(svc_addr(rqstp)); +} + /* * INET callback when data has been received on the socket. */ @@ -678,6 +684,7 @@ static struct svc_xprt_ops svc_udp_ops = { .xpo_prep_reply_hdr = svc_udp_prep_reply_hdr, .xpo_has_wspace = svc_udp_has_wspace, .xpo_accept = svc_udp_accept, + .xpo_secure_port = svc_sock_secure_port, }; static struct svc_xprt_class svc_udp_class = { @@ -842,8 +849,7 @@ static struct svc_xprt *svc_tcp_accept(struct svc_xprt *xprt) * tell us anything. For now just warn about unpriv connections. */ if (!svc_port_is_privileged(sin)) { - dprintk(KERN_WARNING - "%s: connect from unprivileged port: %s\n", + dprintk("%s: connect from unprivileged port: %s\n", serv->sv_name, __svc_print_addr(sin, buf, sizeof(buf))); } @@ -867,6 +873,10 @@ static struct svc_xprt *svc_tcp_accept(struct svc_xprt *xprt) } svc_xprt_set_local(&newsvsk->sk_xprt, sin, slen); + if (sock_is_loopback(newsock->sk)) + set_bit(XPT_LOCAL, &newsvsk->sk_xprt.xpt_flags); + else + clear_bit(XPT_LOCAL, &newsvsk->sk_xprt.xpt_flags); if (serv->sv_stats) serv->sv_stats->nettcpconn++; @@ -1112,6 +1122,7 @@ static int svc_tcp_recvfrom(struct svc_rqst *rqstp) rqstp->rq_xprt_ctxt = NULL; rqstp->rq_prot = IPPROTO_TCP; + rqstp->rq_local = !!test_bit(XPT_LOCAL, &svsk->sk_xprt.xpt_flags); p = (__be32 *)rqstp->rq_arg.head[0].iov_base; calldir = p[1]; @@ -1234,6 +1245,7 @@ static struct svc_xprt_ops svc_tcp_bc_ops = { .xpo_detach = svc_bc_tcp_sock_detach, .xpo_free = svc_bc_sock_free, .xpo_prep_reply_hdr = svc_tcp_prep_reply_hdr, + .xpo_secure_port = svc_sock_secure_port, }; static struct svc_xprt_class svc_tcp_bc_class = { @@ -1272,6 +1284,7 @@ static struct svc_xprt_ops svc_tcp_ops = { .xpo_prep_reply_hdr = svc_tcp_prep_reply_hdr, .xpo_has_wspace = svc_tcp_has_wspace, .xpo_accept = svc_tcp_accept, + .xpo_secure_port = svc_sock_secure_port, }; static struct svc_xprt_class svc_tcp_class = { diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c index dd97ba3c4456..23fb4e75e245 100644 --- a/net/sunrpc/xdr.c +++ b/net/sunrpc/xdr.c @@ -462,6 +462,7 @@ void xdr_init_encode(struct xdr_stream *xdr, struct xdr_buf *buf, __be32 *p) struct kvec *iov = buf->head; int scratch_len = buf->buflen - buf->page_len - buf->tail[0].iov_len; + xdr_set_scratch_buffer(xdr, NULL, 0); BUG_ON(scratch_len < 0); xdr->buf = buf; xdr->iov = iov; @@ -482,6 +483,73 @@ void xdr_init_encode(struct xdr_stream *xdr, struct xdr_buf *buf, __be32 *p) EXPORT_SYMBOL_GPL(xdr_init_encode); /** + * xdr_commit_encode - Ensure all data is written to buffer + * @xdr: pointer to xdr_stream + * + * We handle encoding across page boundaries by giving the caller a + * temporary location to write to, then later copying the data into + * place; xdr_commit_encode does that copying. + * + * Normally the caller doesn't need to call this directly, as the + * following xdr_reserve_space will do it. But an explicit call may be + * required at the end of encoding, or any other time when the xdr_buf + * data might be read. + */ +void xdr_commit_encode(struct xdr_stream *xdr) +{ + int shift = xdr->scratch.iov_len; + void *page; + + if (shift == 0) + return; + page = page_address(*xdr->page_ptr); + memcpy(xdr->scratch.iov_base, page, shift); + memmove(page, page + shift, (void *)xdr->p - page); + xdr->scratch.iov_len = 0; +} +EXPORT_SYMBOL_GPL(xdr_commit_encode); + +__be32 *xdr_get_next_encode_buffer(struct xdr_stream *xdr, size_t nbytes) +{ + static __be32 *p; + int space_left; + int frag1bytes, frag2bytes; + + if (nbytes > PAGE_SIZE) + return NULL; /* Bigger buffers require special handling */ + if (xdr->buf->len + nbytes > xdr->buf->buflen) + return NULL; /* Sorry, we're totally out of space */ + frag1bytes = (xdr->end - xdr->p) << 2; + frag2bytes = nbytes - frag1bytes; + if (xdr->iov) + xdr->iov->iov_len += frag1bytes; + else + xdr->buf->page_len += frag1bytes; + xdr->page_ptr++; + xdr->iov = NULL; + /* + * If the last encode didn't end exactly on a page boundary, the + * next one will straddle boundaries. Encode into the next + * page, then copy it back later in xdr_commit_encode. We use + * the "scratch" iov to track any temporarily unused fragment of + * space at the end of the previous buffer: + */ + xdr->scratch.iov_base = xdr->p; + xdr->scratch.iov_len = frag1bytes; + p = page_address(*xdr->page_ptr); + /* + * Note this is where the next encode will start after we've + * shifted this one back: + */ + xdr->p = (void *)p + frag2bytes; + space_left = xdr->buf->buflen - xdr->buf->len; + xdr->end = (void *)p + min_t(int, space_left, PAGE_SIZE); + xdr->buf->page_len += frag2bytes; + xdr->buf->len += nbytes; + return p; +} + +/** * xdr_reserve_space - Reserve buffer space for sending * @xdr: pointer to xdr_stream * @nbytes: number of bytes to reserve @@ -495,20 +563,122 @@ __be32 * xdr_reserve_space(struct xdr_stream *xdr, size_t nbytes) __be32 *p = xdr->p; __be32 *q; + xdr_commit_encode(xdr); /* align nbytes on the next 32-bit boundary */ nbytes += 3; nbytes &= ~3; q = p + (nbytes >> 2); if (unlikely(q > xdr->end || q < p)) - return NULL; + return xdr_get_next_encode_buffer(xdr, nbytes); xdr->p = q; - xdr->iov->iov_len += nbytes; + if (xdr->iov) + xdr->iov->iov_len += nbytes; + else + xdr->buf->page_len += nbytes; xdr->buf->len += nbytes; return p; } EXPORT_SYMBOL_GPL(xdr_reserve_space); /** + * xdr_truncate_encode - truncate an encode buffer + * @xdr: pointer to xdr_stream + * @len: new length of buffer + * + * Truncates the xdr stream, so that xdr->buf->len == len, + * and xdr->p points at offset len from the start of the buffer, and + * head, tail, and page lengths are adjusted to correspond. + * + * If this means moving xdr->p to a different buffer, we assume that + * that the end pointer should be set to the end of the current page, + * except in the case of the head buffer when we assume the head + * buffer's current length represents the end of the available buffer. + * + * This is *not* safe to use on a buffer that already has inlined page + * cache pages (as in a zero-copy server read reply), except for the + * simple case of truncating from one position in the tail to another. + * + */ +void xdr_truncate_encode(struct xdr_stream *xdr, size_t len) +{ + struct xdr_buf *buf = xdr->buf; + struct kvec *head = buf->head; + struct kvec *tail = buf->tail; + int fraglen; + int new, old; + + if (len > buf->len) { + WARN_ON_ONCE(1); + return; + } + xdr_commit_encode(xdr); + + fraglen = min_t(int, buf->len - len, tail->iov_len); + tail->iov_len -= fraglen; + buf->len -= fraglen; + if (tail->iov_len && buf->len == len) { + xdr->p = tail->iov_base + tail->iov_len; + /* xdr->end, xdr->iov should be set already */ + return; + } + WARN_ON_ONCE(fraglen); + fraglen = min_t(int, buf->len - len, buf->page_len); + buf->page_len -= fraglen; + buf->len -= fraglen; + + new = buf->page_base + buf->page_len; + old = new + fraglen; + xdr->page_ptr -= (old >> PAGE_SHIFT) - (new >> PAGE_SHIFT); + + if (buf->page_len && buf->len == len) { + xdr->p = page_address(*xdr->page_ptr); + xdr->end = (void *)xdr->p + PAGE_SIZE; + xdr->p = (void *)xdr->p + (new % PAGE_SIZE); + /* xdr->iov should already be NULL */ + return; + } + if (fraglen) { + xdr->end = head->iov_base + head->iov_len; + xdr->page_ptr--; + } + /* (otherwise assume xdr->end is already set) */ + head->iov_len = len; + buf->len = len; + xdr->p = head->iov_base + head->iov_len; + xdr->iov = buf->head; +} +EXPORT_SYMBOL(xdr_truncate_encode); + +/** + * xdr_restrict_buflen - decrease available buffer space + * @xdr: pointer to xdr_stream + * @newbuflen: new maximum number of bytes available + * + * Adjust our idea of how much space is available in the buffer. + * If we've already used too much space in the buffer, returns -1. + * If the available space is already smaller than newbuflen, returns 0 + * and does nothing. Otherwise, adjusts xdr->buf->buflen to newbuflen + * and ensures xdr->end is set at most offset newbuflen from the start + * of the buffer. + */ +int xdr_restrict_buflen(struct xdr_stream *xdr, int newbuflen) +{ + struct xdr_buf *buf = xdr->buf; + int left_in_this_buf = (void *)xdr->end - (void *)xdr->p; + int end_offset = buf->len + left_in_this_buf; + + if (newbuflen < 0 || newbuflen < buf->len) + return -1; + if (newbuflen > buf->buflen) + return 0; + if (newbuflen < end_offset) + xdr->end = (void *)xdr->end + newbuflen - end_offset; + buf->buflen = newbuflen; + return 0; +} +EXPORT_SYMBOL(xdr_restrict_buflen); + +/** * xdr_write_pages - Insert a list of pages into an XDR buffer for sending * @xdr: pointer to xdr_stream * @pages: list of pages diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index d173f79947c6..c3b2b3369e52 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -71,24 +71,6 @@ static void xprt_destroy(struct rpc_xprt *xprt); static DEFINE_SPINLOCK(xprt_list_lock); static LIST_HEAD(xprt_list); -/* - * The transport code maintains an estimate on the maximum number of out- - * standing RPC requests, using a smoothed version of the congestion - * avoidance implemented in 44BSD. This is basically the Van Jacobson - * congestion algorithm: If a retransmit occurs, the congestion window is - * halved; otherwise, it is incremented by 1/cwnd when - * - * - a reply is received and - * - a full number of requests are outstanding and - * - the congestion window hasn't been updated recently. - */ -#define RPC_CWNDSHIFT (8U) -#define RPC_CWNDSCALE (1U << RPC_CWNDSHIFT) -#define RPC_INITCWND RPC_CWNDSCALE -#define RPC_MAXCWND(xprt) ((xprt)->max_reqs << RPC_CWNDSHIFT) - -#define RPCXPRT_CONGESTED(xprt) ((xprt)->cong >= (xprt)->cwnd) - /** * xprt_register_transport - register a transport implementation * @transport: transport to register @@ -230,9 +212,9 @@ static void xprt_clear_locked(struct rpc_xprt *xprt) { xprt->snd_task = NULL; if (!test_bit(XPRT_CLOSE_WAIT, &xprt->state)) { - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(XPRT_LOCKED, &xprt->state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); } else queue_work(rpciod_workqueue, &xprt->task_cleanup); } @@ -446,7 +428,15 @@ EXPORT_SYMBOL_GPL(xprt_release_rqst_cong); * @task: recently completed RPC request used to adjust window * @result: result code of completed RPC request * - * We use a time-smoothed congestion estimator to avoid heavy oscillation. + * The transport code maintains an estimate on the maximum number of out- + * standing RPC requests, using a smoothed version of the congestion + * avoidance implemented in 44BSD. This is basically the Van Jacobson + * congestion algorithm: If a retransmit occurs, the congestion window is + * halved; otherwise, it is incremented by 1/cwnd when + * + * - a reply is received and + * - a full number of requests are outstanding and + * - the congestion window hasn't been updated recently. */ void xprt_adjust_cwnd(struct rpc_xprt *xprt, struct rpc_task *task, int result) { diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c index 96ead526b125..693966d3f33b 100644 --- a/net/sunrpc/xprtrdma/rpc_rdma.c +++ b/net/sunrpc/xprtrdma/rpc_rdma.c @@ -78,8 +78,7 @@ static const char transfertypes[][12] = { * elements. Segments are then coalesced when registered, if possible * within the selected memreg mode. * - * Note, this routine is never called if the connection's memory - * registration strategy is 0 (bounce buffers). + * Returns positive number of segments converted, or a negative errno. */ static int @@ -102,10 +101,17 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos, page_base = xdrbuf->page_base & ~PAGE_MASK; p = 0; while (len && n < nsegs) { + if (!ppages[p]) { + /* alloc the pagelist for receiving buffer */ + ppages[p] = alloc_page(GFP_ATOMIC); + if (!ppages[p]) + return -ENOMEM; + } seg[n].mr_page = ppages[p]; seg[n].mr_offset = (void *)(unsigned long) page_base; seg[n].mr_len = min_t(u32, PAGE_SIZE - page_base, len); - BUG_ON(seg[n].mr_len > PAGE_SIZE); + if (seg[n].mr_len > PAGE_SIZE) + return -EIO; len -= seg[n].mr_len; ++n; ++p; @@ -114,7 +120,7 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos, /* Message overflows the seg array */ if (len && n == nsegs) - return 0; + return -EIO; if (xdrbuf->tail[0].iov_len) { /* the rpcrdma protocol allows us to omit any trailing @@ -123,7 +129,7 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos, return n; if (n == nsegs) /* Tail remains, but we're out of segments */ - return 0; + return -EIO; seg[n].mr_page = NULL; seg[n].mr_offset = xdrbuf->tail[0].iov_base; seg[n].mr_len = xdrbuf->tail[0].iov_len; @@ -164,15 +170,17 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos, * Reply chunk (a counted array): * N elements: * 1 - N - HLOO - HLOO - ... - HLOO + * + * Returns positive RPC/RDMA header size, or negative errno. */ -static unsigned int +static ssize_t rpcrdma_create_chunks(struct rpc_rqst *rqst, struct xdr_buf *target, struct rpcrdma_msg *headerp, enum rpcrdma_chunktype type) { struct rpcrdma_req *req = rpcr_to_rdmar(rqst); struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(rqst->rq_xprt); - int nsegs, nchunks = 0; + int n, nsegs, nchunks = 0; unsigned int pos; struct rpcrdma_mr_seg *seg = req->rl_segments; struct rpcrdma_read_chunk *cur_rchunk = NULL; @@ -198,12 +206,11 @@ rpcrdma_create_chunks(struct rpc_rqst *rqst, struct xdr_buf *target, pos = target->head[0].iov_len; nsegs = rpcrdma_convert_iovs(target, pos, type, seg, RPCRDMA_MAX_SEGS); - if (nsegs == 0) - return 0; + if (nsegs < 0) + return nsegs; do { - /* bind/register the memory, then build chunk from result. */ - int n = rpcrdma_register_external(seg, nsegs, + n = rpcrdma_register_external(seg, nsegs, cur_wchunk != NULL, r_xprt); if (n <= 0) goto out; @@ -248,10 +255,6 @@ rpcrdma_create_chunks(struct rpc_rqst *rqst, struct xdr_buf *target, /* success. all failures return above */ req->rl_nchunks = nchunks; - BUG_ON(nchunks == 0); - BUG_ON((r_xprt->rx_ia.ri_memreg_strategy == RPCRDMA_FRMR) - && (nchunks > 3)); - /* * finish off header. If write, marshal discrim and nchunks. */ @@ -278,8 +281,8 @@ rpcrdma_create_chunks(struct rpc_rqst *rqst, struct xdr_buf *target, out: for (pos = 0; nchunks--;) pos += rpcrdma_deregister_external( - &req->rl_segments[pos], r_xprt, NULL); - return 0; + &req->rl_segments[pos], r_xprt); + return n; } /* @@ -361,6 +364,8 @@ rpcrdma_inline_pullup(struct rpc_rqst *rqst, int pad) * [1] -- the RPC header/data, marshaled by RPC and the NFS protocol. * [2] -- optional padding. * [3] -- if padded, header only in [1] and data here. + * + * Returns zero on success, otherwise a negative errno. */ int @@ -370,7 +375,8 @@ rpcrdma_marshal_req(struct rpc_rqst *rqst) struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); struct rpcrdma_req *req = rpcr_to_rdmar(rqst); char *base; - size_t hdrlen, rpclen, padlen; + size_t rpclen, padlen; + ssize_t hdrlen; enum rpcrdma_chunktype rtype, wtype; struct rpcrdma_msg *headerp; @@ -441,14 +447,10 @@ rpcrdma_marshal_req(struct rpc_rqst *rqst) /* The following simplification is not true forever */ if (rtype != rpcrdma_noch && wtype == rpcrdma_replych) wtype = rpcrdma_noch; - BUG_ON(rtype != rpcrdma_noch && wtype != rpcrdma_noch); - - if (r_xprt->rx_ia.ri_memreg_strategy == RPCRDMA_BOUNCEBUFFERS && - (rtype != rpcrdma_noch || wtype != rpcrdma_noch)) { - /* forced to "pure inline"? */ - dprintk("RPC: %s: too much data (%d/%d) for inline\n", - __func__, rqst->rq_rcv_buf.len, rqst->rq_snd_buf.len); - return -1; + if (rtype != rpcrdma_noch && wtype != rpcrdma_noch) { + dprintk("RPC: %s: cannot marshal multiple chunk lists\n", + __func__); + return -EIO; } hdrlen = 28; /*sizeof *headerp;*/ @@ -474,8 +476,11 @@ rpcrdma_marshal_req(struct rpc_rqst *rqst) headerp->rm_body.rm_padded.rm_pempty[1] = xdr_zero; headerp->rm_body.rm_padded.rm_pempty[2] = xdr_zero; hdrlen += 2 * sizeof(u32); /* extra words in padhdr */ - BUG_ON(wtype != rpcrdma_noch); - + if (wtype != rpcrdma_noch) { + dprintk("RPC: %s: invalid chunk list\n", + __func__); + return -EIO; + } } else { headerp->rm_body.rm_nochunks.rm_empty[0] = xdr_zero; headerp->rm_body.rm_nochunks.rm_empty[1] = xdr_zero; @@ -492,8 +497,7 @@ rpcrdma_marshal_req(struct rpc_rqst *rqst) * on receive. Therefore, we request a reply chunk * for non-writes wherever feasible and efficient. */ - if (wtype == rpcrdma_noch && - r_xprt->rx_ia.ri_memreg_strategy > RPCRDMA_REGISTER) + if (wtype == rpcrdma_noch) wtype = rpcrdma_replych; } } @@ -511,9 +515,8 @@ rpcrdma_marshal_req(struct rpc_rqst *rqst) hdrlen = rpcrdma_create_chunks(rqst, &rqst->rq_rcv_buf, headerp, wtype); } - - if (hdrlen == 0) - return -1; + if (hdrlen < 0) + return hdrlen; dprintk("RPC: %s: %s: hdrlen %zd rpclen %zd padlen %zd" " headerp 0x%p base 0x%p lkey 0x%x\n", @@ -680,15 +683,11 @@ rpcrdma_inline_fixup(struct rpc_rqst *rqst, char *srcp, int copy_len, int pad) rqst->rq_private_buf = rqst->rq_rcv_buf; } -/* - * This function is called when an async event is posted to - * the connection which changes the connection state. All it - * does at this point is mark the connection up/down, the rpc - * timers do the rest. - */ void -rpcrdma_conn_func(struct rpcrdma_ep *ep) +rpcrdma_connect_worker(struct work_struct *work) { + struct rpcrdma_ep *ep = + container_of(work, struct rpcrdma_ep, rep_connect_worker.work); struct rpc_xprt *xprt = ep->rep_xprt; spin_lock_bh(&xprt->transport_lock); @@ -705,13 +704,15 @@ rpcrdma_conn_func(struct rpcrdma_ep *ep) } /* - * This function is called when memory window unbind which we are waiting - * for completes. Just use rr_func (zeroed by upcall) to signal completion. + * This function is called when an async event is posted to + * the connection which changes the connection state. All it + * does at this point is mark the connection up/down, the rpc + * timers do the rest. */ -static void -rpcrdma_unbind_func(struct rpcrdma_rep *rep) +void +rpcrdma_conn_func(struct rpcrdma_ep *ep) { - wake_up(&rep->rr_unbind); + schedule_delayed_work(&ep->rep_connect_worker, 0); } /* @@ -728,7 +729,8 @@ rpcrdma_reply_handler(struct rpcrdma_rep *rep) struct rpc_xprt *xprt = rep->rr_xprt; struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); __be32 *iptr; - int i, rdmalen, status; + int rdmalen, status; + unsigned long cwnd; /* Check status. If bad, signal disconnect and return rep to pool */ if (rep->rr_len == ~0U) { @@ -783,6 +785,7 @@ repost: /* from here on, the reply is no longer an orphan */ req->rl_reply = rep; + xprt->reestablish_timeout = 0; /* check for expected message types */ /* The order of some of these tests is important. */ @@ -857,26 +860,10 @@ badheader: break; } - /* If using mw bind, start the deregister process now. */ - /* (Note: if mr_free(), cannot perform it here, in tasklet context) */ - if (req->rl_nchunks) switch (r_xprt->rx_ia.ri_memreg_strategy) { - case RPCRDMA_MEMWINDOWS: - for (i = 0; req->rl_nchunks-- > 1;) - i += rpcrdma_deregister_external( - &req->rl_segments[i], r_xprt, NULL); - /* Optionally wait (not here) for unbinds to complete */ - rep->rr_func = rpcrdma_unbind_func; - (void) rpcrdma_deregister_external(&req->rl_segments[i], - r_xprt, rep); - break; - case RPCRDMA_MEMWINDOWS_ASYNC: - for (i = 0; req->rl_nchunks--;) - i += rpcrdma_deregister_external(&req->rl_segments[i], - r_xprt, NULL); - break; - default: - break; - } + cwnd = xprt->cwnd; + xprt->cwnd = atomic_read(&r_xprt->rx_buf.rb_credits) << RPC_CWNDSHIFT; + if (xprt->cwnd > cwnd) + xprt_release_rqst_cong(rqst->rq_task); dprintk("RPC: %s: xprt_complete_rqst(0x%p, 0x%p, %d)\n", __func__, xprt, rqst, status); diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c index 8d904e4eef15..8f92a61ee2df 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c +++ b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c @@ -1,4 +1,5 @@ /* + * Copyright (c) 2014 Open Grid Computing, Inc. All rights reserved. * Copyright (c) 2005-2006 Network Appliance, Inc. All rights reserved. * * This software is available to you under a choice of one of two @@ -69,7 +70,8 @@ static void rdma_build_arg_xdr(struct svc_rqst *rqstp, /* Set up the XDR head */ rqstp->rq_arg.head[0].iov_base = page_address(page); - rqstp->rq_arg.head[0].iov_len = min(byte_count, ctxt->sge[0].length); + rqstp->rq_arg.head[0].iov_len = + min_t(size_t, byte_count, ctxt->sge[0].length); rqstp->rq_arg.len = byte_count; rqstp->rq_arg.buflen = byte_count; @@ -85,7 +87,7 @@ static void rdma_build_arg_xdr(struct svc_rqst *rqstp, page = ctxt->pages[sge_no]; put_page(rqstp->rq_pages[sge_no]); rqstp->rq_pages[sge_no] = page; - bc -= min(bc, ctxt->sge[sge_no].length); + bc -= min_t(u32, bc, ctxt->sge[sge_no].length); rqstp->rq_arg.buflen += ctxt->sge[sge_no].length; sge_no++; } @@ -113,291 +115,265 @@ static void rdma_build_arg_xdr(struct svc_rqst *rqstp, rqstp->rq_arg.tail[0].iov_len = 0; } -/* Encode a read-chunk-list as an array of IB SGE - * - * Assumptions: - * - chunk[0]->position points to pages[0] at an offset of 0 - * - pages[] is not physically or virtually contiguous and consists of - * PAGE_SIZE elements. - * - * Output: - * - sge array pointing into pages[] array. - * - chunk_sge array specifying sge index and count for each - * chunk in the read list - * - */ -static int map_read_chunks(struct svcxprt_rdma *xprt, - struct svc_rqst *rqstp, - struct svc_rdma_op_ctxt *head, - struct rpcrdma_msg *rmsgp, - struct svc_rdma_req_map *rpl_map, - struct svc_rdma_req_map *chl_map, - int ch_count, - int byte_count) +static int rdma_read_max_sge(struct svcxprt_rdma *xprt, int sge_count) { - int sge_no; - int sge_bytes; - int page_off; - int page_no; - int ch_bytes; - int ch_no; - struct rpcrdma_read_chunk *ch; + if (rdma_node_get_transport(xprt->sc_cm_id->device->node_type) == + RDMA_TRANSPORT_IWARP) + return 1; + else + return min_t(int, sge_count, xprt->sc_max_sge); +} - sge_no = 0; - page_no = 0; - page_off = 0; - ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0]; - ch_no = 0; - ch_bytes = ntohl(ch->rc_target.rs_length); - head->arg.head[0] = rqstp->rq_arg.head[0]; - head->arg.tail[0] = rqstp->rq_arg.tail[0]; - head->arg.pages = &head->pages[head->count]; - head->hdr_count = head->count; /* save count of hdr pages */ - head->arg.page_base = 0; - head->arg.page_len = ch_bytes; - head->arg.len = rqstp->rq_arg.len + ch_bytes; - head->arg.buflen = rqstp->rq_arg.buflen + ch_bytes; - head->count++; - chl_map->ch[0].start = 0; - while (byte_count) { - rpl_map->sge[sge_no].iov_base = - page_address(rqstp->rq_arg.pages[page_no]) + page_off; - sge_bytes = min_t(int, PAGE_SIZE-page_off, ch_bytes); - rpl_map->sge[sge_no].iov_len = sge_bytes; - /* - * Don't bump head->count here because the same page - * may be used by multiple SGE. - */ - head->arg.pages[page_no] = rqstp->rq_arg.pages[page_no]; - rqstp->rq_respages = &rqstp->rq_arg.pages[page_no+1]; +typedef int (*rdma_reader_fn)(struct svcxprt_rdma *xprt, + struct svc_rqst *rqstp, + struct svc_rdma_op_ctxt *head, + int *page_no, + u32 *page_offset, + u32 rs_handle, + u32 rs_length, + u64 rs_offset, + int last); + +/* Issue an RDMA_READ using the local lkey to map the data sink */ +static int rdma_read_chunk_lcl(struct svcxprt_rdma *xprt, + struct svc_rqst *rqstp, + struct svc_rdma_op_ctxt *head, + int *page_no, + u32 *page_offset, + u32 rs_handle, + u32 rs_length, + u64 rs_offset, + int last) +{ + struct ib_send_wr read_wr; + int pages_needed = PAGE_ALIGN(*page_offset + rs_length) >> PAGE_SHIFT; + struct svc_rdma_op_ctxt *ctxt = svc_rdma_get_context(xprt); + int ret, read, pno; + u32 pg_off = *page_offset; + u32 pg_no = *page_no; + + ctxt->direction = DMA_FROM_DEVICE; + ctxt->read_hdr = head; + pages_needed = + min_t(int, pages_needed, rdma_read_max_sge(xprt, pages_needed)); + read = min_t(int, pages_needed << PAGE_SHIFT, rs_length); + + for (pno = 0; pno < pages_needed; pno++) { + int len = min_t(int, rs_length, PAGE_SIZE - pg_off); + + head->arg.pages[pg_no] = rqstp->rq_arg.pages[pg_no]; + head->arg.page_len += len; + head->arg.len += len; + if (!pg_off) + head->count++; + rqstp->rq_respages = &rqstp->rq_arg.pages[pg_no+1]; rqstp->rq_next_page = rqstp->rq_respages + 1; + ctxt->sge[pno].addr = + ib_dma_map_page(xprt->sc_cm_id->device, + head->arg.pages[pg_no], pg_off, + PAGE_SIZE - pg_off, + DMA_FROM_DEVICE); + ret = ib_dma_mapping_error(xprt->sc_cm_id->device, + ctxt->sge[pno].addr); + if (ret) + goto err; + atomic_inc(&xprt->sc_dma_used); - byte_count -= sge_bytes; - ch_bytes -= sge_bytes; - sge_no++; - /* - * If all bytes for this chunk have been mapped to an - * SGE, move to the next SGE - */ - if (ch_bytes == 0) { - chl_map->ch[ch_no].count = - sge_no - chl_map->ch[ch_no].start; - ch_no++; - ch++; - chl_map->ch[ch_no].start = sge_no; - ch_bytes = ntohl(ch->rc_target.rs_length); - /* If bytes remaining account for next chunk */ - if (byte_count) { - head->arg.page_len += ch_bytes; - head->arg.len += ch_bytes; - head->arg.buflen += ch_bytes; - } + /* The lkey here is either a local dma lkey or a dma_mr lkey */ + ctxt->sge[pno].lkey = xprt->sc_dma_lkey; + ctxt->sge[pno].length = len; + ctxt->count++; + + /* adjust offset and wrap to next page if needed */ + pg_off += len; + if (pg_off == PAGE_SIZE) { + pg_off = 0; + pg_no++; } - /* - * If this SGE consumed all of the page, move to the - * next page - */ - if ((sge_bytes + page_off) == PAGE_SIZE) { - page_no++; - page_off = 0; - /* - * If there are still bytes left to map, bump - * the page count - */ - if (byte_count) - head->count++; - } else - page_off += sge_bytes; + rs_length -= len; } - BUG_ON(byte_count != 0); - return sge_no; + + if (last && rs_length == 0) + set_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags); + else + clear_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags); + + memset(&read_wr, 0, sizeof(read_wr)); + read_wr.wr_id = (unsigned long)ctxt; + read_wr.opcode = IB_WR_RDMA_READ; + ctxt->wr_op = read_wr.opcode; + read_wr.send_flags = IB_SEND_SIGNALED; + read_wr.wr.rdma.rkey = rs_handle; + read_wr.wr.rdma.remote_addr = rs_offset; + read_wr.sg_list = ctxt->sge; + read_wr.num_sge = pages_needed; + + ret = svc_rdma_send(xprt, &read_wr); + if (ret) { + pr_err("svcrdma: Error %d posting RDMA_READ\n", ret); + set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags); + goto err; + } + + /* return current location in page array */ + *page_no = pg_no; + *page_offset = pg_off; + ret = read; + atomic_inc(&rdma_stat_read); + return ret; + err: + svc_rdma_unmap_dma(ctxt); + svc_rdma_put_context(ctxt, 0); + return ret; } -/* Map a read-chunk-list to an XDR and fast register the page-list. - * - * Assumptions: - * - chunk[0] position points to pages[0] at an offset of 0 - * - pages[] will be made physically contiguous by creating a one-off memory - * region using the fastreg verb. - * - byte_count is # of bytes in read-chunk-list - * - ch_count is # of chunks in read-chunk-list - * - * Output: - * - sge array pointing into pages[] array. - * - chunk_sge array specifying sge index and count for each - * chunk in the read list - */ -static int fast_reg_read_chunks(struct svcxprt_rdma *xprt, +/* Issue an RDMA_READ using an FRMR to map the data sink */ +static int rdma_read_chunk_frmr(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp, struct svc_rdma_op_ctxt *head, - struct rpcrdma_msg *rmsgp, - struct svc_rdma_req_map *rpl_map, - struct svc_rdma_req_map *chl_map, - int ch_count, - int byte_count) + int *page_no, + u32 *page_offset, + u32 rs_handle, + u32 rs_length, + u64 rs_offset, + int last) { - int page_no; - int ch_no; - u32 offset; - struct rpcrdma_read_chunk *ch; - struct svc_rdma_fastreg_mr *frmr; - int ret = 0; + struct ib_send_wr read_wr; + struct ib_send_wr inv_wr; + struct ib_send_wr fastreg_wr; + u8 key; + int pages_needed = PAGE_ALIGN(*page_offset + rs_length) >> PAGE_SHIFT; + struct svc_rdma_op_ctxt *ctxt = svc_rdma_get_context(xprt); + struct svc_rdma_fastreg_mr *frmr = svc_rdma_get_frmr(xprt); + int ret, read, pno; + u32 pg_off = *page_offset; + u32 pg_no = *page_no; - frmr = svc_rdma_get_frmr(xprt); if (IS_ERR(frmr)) return -ENOMEM; - head->frmr = frmr; - head->arg.head[0] = rqstp->rq_arg.head[0]; - head->arg.tail[0] = rqstp->rq_arg.tail[0]; - head->arg.pages = &head->pages[head->count]; - head->hdr_count = head->count; /* save count of hdr pages */ - head->arg.page_base = 0; - head->arg.page_len = byte_count; - head->arg.len = rqstp->rq_arg.len + byte_count; - head->arg.buflen = rqstp->rq_arg.buflen + byte_count; + ctxt->direction = DMA_FROM_DEVICE; + ctxt->frmr = frmr; + pages_needed = min_t(int, pages_needed, xprt->sc_frmr_pg_list_len); + read = min_t(int, pages_needed << PAGE_SHIFT, rs_length); - /* Fast register the page list */ - frmr->kva = page_address(rqstp->rq_arg.pages[0]); + frmr->kva = page_address(rqstp->rq_arg.pages[pg_no]); frmr->direction = DMA_FROM_DEVICE; frmr->access_flags = (IB_ACCESS_LOCAL_WRITE|IB_ACCESS_REMOTE_WRITE); - frmr->map_len = byte_count; - frmr->page_list_len = PAGE_ALIGN(byte_count) >> PAGE_SHIFT; - for (page_no = 0; page_no < frmr->page_list_len; page_no++) { - frmr->page_list->page_list[page_no] = + frmr->map_len = pages_needed << PAGE_SHIFT; + frmr->page_list_len = pages_needed; + + for (pno = 0; pno < pages_needed; pno++) { + int len = min_t(int, rs_length, PAGE_SIZE - pg_off); + + head->arg.pages[pg_no] = rqstp->rq_arg.pages[pg_no]; + head->arg.page_len += len; + head->arg.len += len; + if (!pg_off) + head->count++; + rqstp->rq_respages = &rqstp->rq_arg.pages[pg_no+1]; + rqstp->rq_next_page = rqstp->rq_respages + 1; + frmr->page_list->page_list[pno] = ib_dma_map_page(xprt->sc_cm_id->device, - rqstp->rq_arg.pages[page_no], 0, + head->arg.pages[pg_no], 0, PAGE_SIZE, DMA_FROM_DEVICE); - if (ib_dma_mapping_error(xprt->sc_cm_id->device, - frmr->page_list->page_list[page_no])) - goto fatal_err; + ret = ib_dma_mapping_error(xprt->sc_cm_id->device, + frmr->page_list->page_list[pno]); + if (ret) + goto err; atomic_inc(&xprt->sc_dma_used); - head->arg.pages[page_no] = rqstp->rq_arg.pages[page_no]; - } - head->count += page_no; - - /* rq_respages points one past arg pages */ - rqstp->rq_respages = &rqstp->rq_arg.pages[page_no]; - rqstp->rq_next_page = rqstp->rq_respages + 1; - /* Create the reply and chunk maps */ - offset = 0; - ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0]; - for (ch_no = 0; ch_no < ch_count; ch_no++) { - int len = ntohl(ch->rc_target.rs_length); - rpl_map->sge[ch_no].iov_base = frmr->kva + offset; - rpl_map->sge[ch_no].iov_len = len; - chl_map->ch[ch_no].count = 1; - chl_map->ch[ch_no].start = ch_no; - offset += len; - ch++; + /* adjust offset and wrap to next page if needed */ + pg_off += len; + if (pg_off == PAGE_SIZE) { + pg_off = 0; + pg_no++; + } + rs_length -= len; } - ret = svc_rdma_fastreg(xprt, frmr); - if (ret) - goto fatal_err; - - return ch_no; - - fatal_err: - printk("svcrdma: error fast registering xdr for xprt %p", xprt); - svc_rdma_put_frmr(xprt, frmr); - return -EIO; -} - -static int rdma_set_ctxt_sge(struct svcxprt_rdma *xprt, - struct svc_rdma_op_ctxt *ctxt, - struct svc_rdma_fastreg_mr *frmr, - struct kvec *vec, - u64 *sgl_offset, - int count) -{ - int i; - unsigned long off; + if (last && rs_length == 0) + set_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags); + else + clear_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags); - ctxt->count = count; - ctxt->direction = DMA_FROM_DEVICE; - for (i = 0; i < count; i++) { - ctxt->sge[i].length = 0; /* in case map fails */ - if (!frmr) { - BUG_ON(!virt_to_page(vec[i].iov_base)); - off = (unsigned long)vec[i].iov_base & ~PAGE_MASK; - ctxt->sge[i].addr = - ib_dma_map_page(xprt->sc_cm_id->device, - virt_to_page(vec[i].iov_base), - off, - vec[i].iov_len, - DMA_FROM_DEVICE); - if (ib_dma_mapping_error(xprt->sc_cm_id->device, - ctxt->sge[i].addr)) - return -EINVAL; - ctxt->sge[i].lkey = xprt->sc_dma_lkey; - atomic_inc(&xprt->sc_dma_used); - } else { - ctxt->sge[i].addr = (unsigned long)vec[i].iov_base; - ctxt->sge[i].lkey = frmr->mr->lkey; - } - ctxt->sge[i].length = vec[i].iov_len; - *sgl_offset = *sgl_offset + vec[i].iov_len; + /* Bump the key */ + key = (u8)(frmr->mr->lkey & 0x000000FF); + ib_update_fast_reg_key(frmr->mr, ++key); + + ctxt->sge[0].addr = (unsigned long)frmr->kva + *page_offset; + ctxt->sge[0].lkey = frmr->mr->lkey; + ctxt->sge[0].length = read; + ctxt->count = 1; + ctxt->read_hdr = head; + + /* Prepare FASTREG WR */ + memset(&fastreg_wr, 0, sizeof(fastreg_wr)); + fastreg_wr.opcode = IB_WR_FAST_REG_MR; + fastreg_wr.send_flags = IB_SEND_SIGNALED; + fastreg_wr.wr.fast_reg.iova_start = (unsigned long)frmr->kva; + fastreg_wr.wr.fast_reg.page_list = frmr->page_list; + fastreg_wr.wr.fast_reg.page_list_len = frmr->page_list_len; + fastreg_wr.wr.fast_reg.page_shift = PAGE_SHIFT; + fastreg_wr.wr.fast_reg.length = frmr->map_len; + fastreg_wr.wr.fast_reg.access_flags = frmr->access_flags; + fastreg_wr.wr.fast_reg.rkey = frmr->mr->lkey; + fastreg_wr.next = &read_wr; + + /* Prepare RDMA_READ */ + memset(&read_wr, 0, sizeof(read_wr)); + read_wr.send_flags = IB_SEND_SIGNALED; + read_wr.wr.rdma.rkey = rs_handle; + read_wr.wr.rdma.remote_addr = rs_offset; + read_wr.sg_list = ctxt->sge; + read_wr.num_sge = 1; + if (xprt->sc_dev_caps & SVCRDMA_DEVCAP_READ_W_INV) { + read_wr.opcode = IB_WR_RDMA_READ_WITH_INV; + read_wr.wr_id = (unsigned long)ctxt; + read_wr.ex.invalidate_rkey = ctxt->frmr->mr->lkey; + } else { + read_wr.opcode = IB_WR_RDMA_READ; + read_wr.next = &inv_wr; + /* Prepare invalidate */ + memset(&inv_wr, 0, sizeof(inv_wr)); + inv_wr.wr_id = (unsigned long)ctxt; + inv_wr.opcode = IB_WR_LOCAL_INV; + inv_wr.send_flags = IB_SEND_SIGNALED | IB_SEND_FENCE; + inv_wr.ex.invalidate_rkey = frmr->mr->lkey; + } + ctxt->wr_op = read_wr.opcode; + + /* Post the chain */ + ret = svc_rdma_send(xprt, &fastreg_wr); + if (ret) { + pr_err("svcrdma: Error %d posting RDMA_READ\n", ret); + set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags); + goto err; } - return 0; -} -static int rdma_read_max_sge(struct svcxprt_rdma *xprt, int sge_count) -{ - if ((rdma_node_get_transport(xprt->sc_cm_id->device->node_type) == - RDMA_TRANSPORT_IWARP) && - sge_count > 1) - return 1; - else - return min_t(int, sge_count, xprt->sc_max_sge); + /* return current location in page array */ + *page_no = pg_no; + *page_offset = pg_off; + ret = read; + atomic_inc(&rdma_stat_read); + return ret; + err: + svc_rdma_unmap_dma(ctxt); + svc_rdma_put_context(ctxt, 0); + svc_rdma_put_frmr(xprt, frmr); + return ret; } -/* - * Use RDMA_READ to read data from the advertised client buffer into the - * XDR stream starting at rq_arg.head[0].iov_base. - * Each chunk in the array - * contains the following fields: - * discrim - '1', This isn't used for data placement - * position - The xdr stream offset (the same for every chunk) - * handle - RMR for client memory region - * length - data transfer length - * offset - 64 bit tagged offset in remote memory region - * - * On our side, we need to read into a pagelist. The first page immediately - * follows the RPC header. - * - * This function returns: - * 0 - No error and no read-list found. - * - * 1 - Successful read-list processing. The data is not yet in - * the pagelist and therefore the RPC request must be deferred. The - * I/O completion will enqueue the transport again and - * svc_rdma_recvfrom will complete the request. - * - * <0 - Error processing/posting read-list. - * - * NOTE: The ctxt must not be touched after the last WR has been posted - * because the I/O completion processing may occur on another - * processor and free / modify the context. Ne touche pas! - */ -static int rdma_read_xdr(struct svcxprt_rdma *xprt, - struct rpcrdma_msg *rmsgp, - struct svc_rqst *rqstp, - struct svc_rdma_op_ctxt *hdr_ctxt) +static int rdma_read_chunks(struct svcxprt_rdma *xprt, + struct rpcrdma_msg *rmsgp, + struct svc_rqst *rqstp, + struct svc_rdma_op_ctxt *head) { - struct ib_send_wr read_wr; - struct ib_send_wr inv_wr; - int err = 0; - int ch_no; - int ch_count; - int byte_count; - int sge_count; - u64 sgl_offset; + int page_no, ch_count, ret; struct rpcrdma_read_chunk *ch; - struct svc_rdma_op_ctxt *ctxt = NULL; - struct svc_rdma_req_map *rpl_map; - struct svc_rdma_req_map *chl_map; + u32 page_offset, byte_count; + u64 rs_offset; + rdma_reader_fn reader; /* If no read list is present, return 0 */ ch = svc_rdma_get_read_chunk(rmsgp); @@ -408,122 +384,55 @@ static int rdma_read_xdr(struct svcxprt_rdma *xprt, if (ch_count > RPCSVC_MAXPAGES) return -EINVAL; - /* Allocate temporary reply and chunk maps */ - rpl_map = svc_rdma_get_req_map(); - chl_map = svc_rdma_get_req_map(); + /* The request is completed when the RDMA_READs complete. The + * head context keeps all the pages that comprise the + * request. + */ + head->arg.head[0] = rqstp->rq_arg.head[0]; + head->arg.tail[0] = rqstp->rq_arg.tail[0]; + head->arg.pages = &head->pages[head->count]; + head->hdr_count = head->count; + head->arg.page_base = 0; + head->arg.page_len = 0; + head->arg.len = rqstp->rq_arg.len; + head->arg.buflen = rqstp->rq_arg.buflen; - if (!xprt->sc_frmr_pg_list_len) - sge_count = map_read_chunks(xprt, rqstp, hdr_ctxt, rmsgp, - rpl_map, chl_map, ch_count, - byte_count); + /* Use FRMR if supported */ + if (xprt->sc_dev_caps & SVCRDMA_DEVCAP_FAST_REG) + reader = rdma_read_chunk_frmr; else - sge_count = fast_reg_read_chunks(xprt, rqstp, hdr_ctxt, rmsgp, - rpl_map, chl_map, ch_count, - byte_count); - if (sge_count < 0) { - err = -EIO; - goto out; - } - - sgl_offset = 0; - ch_no = 0; + reader = rdma_read_chunk_lcl; + page_no = 0; page_offset = 0; for (ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0]; - ch->rc_discrim != 0; ch++, ch_no++) { - u64 rs_offset; -next_sge: - ctxt = svc_rdma_get_context(xprt); - ctxt->direction = DMA_FROM_DEVICE; - ctxt->frmr = hdr_ctxt->frmr; - ctxt->read_hdr = NULL; - clear_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags); - clear_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags); + ch->rc_discrim != 0; ch++) { - /* Prepare READ WR */ - memset(&read_wr, 0, sizeof read_wr); - read_wr.wr_id = (unsigned long)ctxt; - read_wr.opcode = IB_WR_RDMA_READ; - ctxt->wr_op = read_wr.opcode; - read_wr.send_flags = IB_SEND_SIGNALED; - read_wr.wr.rdma.rkey = ntohl(ch->rc_target.rs_handle); xdr_decode_hyper((__be32 *)&ch->rc_target.rs_offset, &rs_offset); - read_wr.wr.rdma.remote_addr = rs_offset + sgl_offset; - read_wr.sg_list = ctxt->sge; - read_wr.num_sge = - rdma_read_max_sge(xprt, chl_map->ch[ch_no].count); - err = rdma_set_ctxt_sge(xprt, ctxt, hdr_ctxt->frmr, - &rpl_map->sge[chl_map->ch[ch_no].start], - &sgl_offset, - read_wr.num_sge); - if (err) { - svc_rdma_unmap_dma(ctxt); - svc_rdma_put_context(ctxt, 0); - goto out; - } - if (((ch+1)->rc_discrim == 0) && - (read_wr.num_sge == chl_map->ch[ch_no].count)) { - /* - * Mark the last RDMA_READ with a bit to - * indicate all RPC data has been fetched from - * the client and the RPC needs to be enqueued. - */ - set_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags); - if (hdr_ctxt->frmr) { - set_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags); - /* - * Invalidate the local MR used to map the data - * sink. - */ - if (xprt->sc_dev_caps & - SVCRDMA_DEVCAP_READ_W_INV) { - read_wr.opcode = - IB_WR_RDMA_READ_WITH_INV; - ctxt->wr_op = read_wr.opcode; - read_wr.ex.invalidate_rkey = - ctxt->frmr->mr->lkey; - } else { - /* Prepare INVALIDATE WR */ - memset(&inv_wr, 0, sizeof inv_wr); - inv_wr.opcode = IB_WR_LOCAL_INV; - inv_wr.send_flags = IB_SEND_SIGNALED; - inv_wr.ex.invalidate_rkey = - hdr_ctxt->frmr->mr->lkey; - read_wr.next = &inv_wr; - } - } - ctxt->read_hdr = hdr_ctxt; - } - /* Post the read */ - err = svc_rdma_send(xprt, &read_wr); - if (err) { - printk(KERN_ERR "svcrdma: Error %d posting RDMA_READ\n", - err); - set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags); - svc_rdma_unmap_dma(ctxt); - svc_rdma_put_context(ctxt, 0); - goto out; + byte_count = ntohl(ch->rc_target.rs_length); + + while (byte_count > 0) { + ret = reader(xprt, rqstp, head, + &page_no, &page_offset, + ntohl(ch->rc_target.rs_handle), + byte_count, rs_offset, + ((ch+1)->rc_discrim == 0) /* last */ + ); + if (ret < 0) + goto err; + byte_count -= ret; + rs_offset += ret; + head->arg.buflen += ret; } - atomic_inc(&rdma_stat_read); - - if (read_wr.num_sge < chl_map->ch[ch_no].count) { - chl_map->ch[ch_no].count -= read_wr.num_sge; - chl_map->ch[ch_no].start += read_wr.num_sge; - goto next_sge; - } - sgl_offset = 0; - err = 1; } - - out: - svc_rdma_put_req_map(rpl_map); - svc_rdma_put_req_map(chl_map); - + ret = 1; + err: /* Detach arg pages. svc_recv will replenish them */ - for (ch_no = 0; &rqstp->rq_pages[ch_no] < rqstp->rq_respages; ch_no++) - rqstp->rq_pages[ch_no] = NULL; + for (page_no = 0; + &rqstp->rq_pages[page_no] < rqstp->rq_respages; page_no++) + rqstp->rq_pages[page_no] = NULL; - return err; + return ret; } static int rdma_read_complete(struct svc_rqst *rqstp, @@ -595,13 +504,9 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp) struct svc_rdma_op_ctxt, dto_q); list_del_init(&ctxt->dto_q); - } - if (ctxt) { spin_unlock_bh(&rdma_xprt->sc_rq_dto_lock); return rdma_read_complete(rqstp, ctxt); - } - - if (!list_empty(&rdma_xprt->sc_rq_dto_q)) { + } else if (!list_empty(&rdma_xprt->sc_rq_dto_q)) { ctxt = list_entry(rdma_xprt->sc_rq_dto_q.next, struct svc_rdma_op_ctxt, dto_q); @@ -621,7 +526,6 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp) if (test_bit(XPT_CLOSE, &xprt->xpt_flags)) goto close_out; - BUG_ON(ret); goto out; } dprintk("svcrdma: processing ctxt=%p on xprt=%p, rqstp=%p, status=%d\n", @@ -644,12 +548,11 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp) } /* Read read-list data. */ - ret = rdma_read_xdr(rdma_xprt, rmsgp, rqstp, ctxt); + ret = rdma_read_chunks(rdma_xprt, rmsgp, rqstp, ctxt); if (ret > 0) { /* read-list posted, defer until data received from client. */ goto defer; - } - if (ret < 0) { + } else if (ret < 0) { /* Post of read-list failed, free context. */ svc_rdma_put_context(ctxt, 1); return 0; diff --git a/net/sunrpc/xprtrdma/svc_rdma_sendto.c b/net/sunrpc/xprtrdma/svc_rdma_sendto.c index 7e024a51617e..49fd21a5c215 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_sendto.c +++ b/net/sunrpc/xprtrdma/svc_rdma_sendto.c @@ -1,4 +1,5 @@ /* + * Copyright (c) 2014 Open Grid Computing, Inc. All rights reserved. * Copyright (c) 2005-2006 Network Appliance, Inc. All rights reserved. * * This software is available to you under a choice of one of two @@ -49,152 +50,6 @@ #define RPCDBG_FACILITY RPCDBG_SVCXPRT -/* Encode an XDR as an array of IB SGE - * - * Assumptions: - * - head[0] is physically contiguous. - * - tail[0] is physically contiguous. - * - pages[] is not physically or virtually contiguous and consists of - * PAGE_SIZE elements. - * - * Output: - * SGE[0] reserved for RCPRDMA header - * SGE[1] data from xdr->head[] - * SGE[2..sge_count-2] data from xdr->pages[] - * SGE[sge_count-1] data from xdr->tail. - * - * The max SGE we need is the length of the XDR / pagesize + one for - * head + one for tail + one for RPCRDMA header. Since RPCSVC_MAXPAGES - * reserves a page for both the request and the reply header, and this - * array is only concerned with the reply we are assured that we have - * on extra page for the RPCRMDA header. - */ -static int fast_reg_xdr(struct svcxprt_rdma *xprt, - struct xdr_buf *xdr, - struct svc_rdma_req_map *vec) -{ - int sge_no; - u32 sge_bytes; - u32 page_bytes; - u32 page_off; - int page_no = 0; - u8 *frva; - struct svc_rdma_fastreg_mr *frmr; - - frmr = svc_rdma_get_frmr(xprt); - if (IS_ERR(frmr)) - return -ENOMEM; - vec->frmr = frmr; - - /* Skip the RPCRDMA header */ - sge_no = 1; - - /* Map the head. */ - frva = (void *)((unsigned long)(xdr->head[0].iov_base) & PAGE_MASK); - vec->sge[sge_no].iov_base = xdr->head[0].iov_base; - vec->sge[sge_no].iov_len = xdr->head[0].iov_len; - vec->count = 2; - sge_no++; - - /* Map the XDR head */ - frmr->kva = frva; - frmr->direction = DMA_TO_DEVICE; - frmr->access_flags = 0; - frmr->map_len = PAGE_SIZE; - frmr->page_list_len = 1; - page_off = (unsigned long)xdr->head[0].iov_base & ~PAGE_MASK; - frmr->page_list->page_list[page_no] = - ib_dma_map_page(xprt->sc_cm_id->device, - virt_to_page(xdr->head[0].iov_base), - page_off, - PAGE_SIZE - page_off, - DMA_TO_DEVICE); - if (ib_dma_mapping_error(xprt->sc_cm_id->device, - frmr->page_list->page_list[page_no])) - goto fatal_err; - atomic_inc(&xprt->sc_dma_used); - - /* Map the XDR page list */ - page_off = xdr->page_base; - page_bytes = xdr->page_len + page_off; - if (!page_bytes) - goto encode_tail; - - /* Map the pages */ - vec->sge[sge_no].iov_base = frva + frmr->map_len + page_off; - vec->sge[sge_no].iov_len = page_bytes; - sge_no++; - while (page_bytes) { - struct page *page; - - page = xdr->pages[page_no++]; - sge_bytes = min_t(u32, page_bytes, (PAGE_SIZE - page_off)); - page_bytes -= sge_bytes; - - frmr->page_list->page_list[page_no] = - ib_dma_map_page(xprt->sc_cm_id->device, - page, page_off, - sge_bytes, DMA_TO_DEVICE); - if (ib_dma_mapping_error(xprt->sc_cm_id->device, - frmr->page_list->page_list[page_no])) - goto fatal_err; - - atomic_inc(&xprt->sc_dma_used); - page_off = 0; /* reset for next time through loop */ - frmr->map_len += PAGE_SIZE; - frmr->page_list_len++; - } - vec->count++; - - encode_tail: - /* Map tail */ - if (0 == xdr->tail[0].iov_len) - goto done; - - vec->count++; - vec->sge[sge_no].iov_len = xdr->tail[0].iov_len; - - if (((unsigned long)xdr->tail[0].iov_base & PAGE_MASK) == - ((unsigned long)xdr->head[0].iov_base & PAGE_MASK)) { - /* - * If head and tail use the same page, we don't need - * to map it again. - */ - vec->sge[sge_no].iov_base = xdr->tail[0].iov_base; - } else { - void *va; - - /* Map another page for the tail */ - page_off = (unsigned long)xdr->tail[0].iov_base & ~PAGE_MASK; - va = (void *)((unsigned long)xdr->tail[0].iov_base & PAGE_MASK); - vec->sge[sge_no].iov_base = frva + frmr->map_len + page_off; - - frmr->page_list->page_list[page_no] = - ib_dma_map_page(xprt->sc_cm_id->device, virt_to_page(va), - page_off, - PAGE_SIZE, - DMA_TO_DEVICE); - if (ib_dma_mapping_error(xprt->sc_cm_id->device, - frmr->page_list->page_list[page_no])) - goto fatal_err; - atomic_inc(&xprt->sc_dma_used); - frmr->map_len += PAGE_SIZE; - frmr->page_list_len++; - } - - done: - if (svc_rdma_fastreg(xprt, frmr)) - goto fatal_err; - - return 0; - - fatal_err: - printk("svcrdma: Error fast registering memory for xprt %p\n", xprt); - vec->frmr = NULL; - svc_rdma_put_frmr(xprt, frmr); - return -EIO; -} - static int map_xdr(struct svcxprt_rdma *xprt, struct xdr_buf *xdr, struct svc_rdma_req_map *vec) @@ -208,9 +63,6 @@ static int map_xdr(struct svcxprt_rdma *xprt, BUG_ON(xdr->len != (xdr->head[0].iov_len + xdr->page_len + xdr->tail[0].iov_len)); - if (xprt->sc_frmr_pg_list_len) - return fast_reg_xdr(xprt, xdr, vec); - /* Skip the first sge, this is for the RPCRDMA header */ sge_no = 1; @@ -282,8 +134,6 @@ static dma_addr_t dma_map_xdr(struct svcxprt_rdma *xprt, } /* Assumptions: - * - We are using FRMR - * - or - * - The specified write_len can be represented in sc_max_sge * PAGE_SIZE */ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp, @@ -327,23 +177,16 @@ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp, sge_bytes = min_t(size_t, bc, vec->sge[xdr_sge_no].iov_len-sge_off); sge[sge_no].length = sge_bytes; - if (!vec->frmr) { - sge[sge_no].addr = - dma_map_xdr(xprt, &rqstp->rq_res, xdr_off, - sge_bytes, DMA_TO_DEVICE); - xdr_off += sge_bytes; - if (ib_dma_mapping_error(xprt->sc_cm_id->device, - sge[sge_no].addr)) - goto err; - atomic_inc(&xprt->sc_dma_used); - sge[sge_no].lkey = xprt->sc_dma_lkey; - } else { - sge[sge_no].addr = (unsigned long) - vec->sge[xdr_sge_no].iov_base + sge_off; - sge[sge_no].lkey = vec->frmr->mr->lkey; - } + sge[sge_no].addr = + dma_map_xdr(xprt, &rqstp->rq_res, xdr_off, + sge_bytes, DMA_TO_DEVICE); + xdr_off += sge_bytes; + if (ib_dma_mapping_error(xprt->sc_cm_id->device, + sge[sge_no].addr)) + goto err; + atomic_inc(&xprt->sc_dma_used); + sge[sge_no].lkey = xprt->sc_dma_lkey; ctxt->count++; - ctxt->frmr = vec->frmr; sge_off = 0; sge_no++; xdr_sge_no++; @@ -369,7 +212,6 @@ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp, return 0; err: svc_rdma_unmap_dma(ctxt); - svc_rdma_put_frmr(xprt, vec->frmr); svc_rdma_put_context(ctxt, 0); /* Fatal error, close transport */ return -EIO; @@ -397,10 +239,7 @@ static int send_write_chunks(struct svcxprt_rdma *xprt, res_ary = (struct rpcrdma_write_array *) &rdma_resp->rm_body.rm_chunks[1]; - if (vec->frmr) - max_write = vec->frmr->map_len; - else - max_write = xprt->sc_max_sge * PAGE_SIZE; + max_write = xprt->sc_max_sge * PAGE_SIZE; /* Write chunks start at the pagelist */ for (xdr_off = rqstp->rq_res.head[0].iov_len, chunk_no = 0; @@ -472,10 +311,7 @@ static int send_reply_chunks(struct svcxprt_rdma *xprt, res_ary = (struct rpcrdma_write_array *) &rdma_resp->rm_body.rm_chunks[2]; - if (vec->frmr) - max_write = vec->frmr->map_len; - else - max_write = xprt->sc_max_sge * PAGE_SIZE; + max_write = xprt->sc_max_sge * PAGE_SIZE; /* xdr offset starts at RPC message */ nchunks = ntohl(arg_ary->wc_nchunks); @@ -545,7 +381,6 @@ static int send_reply(struct svcxprt_rdma *rdma, int byte_count) { struct ib_send_wr send_wr; - struct ib_send_wr inv_wr; int sge_no; int sge_bytes; int page_no; @@ -559,7 +394,6 @@ static int send_reply(struct svcxprt_rdma *rdma, "svcrdma: could not post a receive buffer, err=%d." "Closing transport %p.\n", ret, rdma); set_bit(XPT_CLOSE, &rdma->sc_xprt.xpt_flags); - svc_rdma_put_frmr(rdma, vec->frmr); svc_rdma_put_context(ctxt, 0); return -ENOTCONN; } @@ -567,11 +401,6 @@ static int send_reply(struct svcxprt_rdma *rdma, /* Prepare the context */ ctxt->pages[0] = page; ctxt->count = 1; - ctxt->frmr = vec->frmr; - if (vec->frmr) - set_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags); - else - clear_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags); /* Prepare the SGE for the RPCRDMA Header */ ctxt->sge[0].lkey = rdma->sc_dma_lkey; @@ -590,21 +419,15 @@ static int send_reply(struct svcxprt_rdma *rdma, int xdr_off = 0; sge_bytes = min_t(size_t, vec->sge[sge_no].iov_len, byte_count); byte_count -= sge_bytes; - if (!vec->frmr) { - ctxt->sge[sge_no].addr = - dma_map_xdr(rdma, &rqstp->rq_res, xdr_off, - sge_bytes, DMA_TO_DEVICE); - xdr_off += sge_bytes; - if (ib_dma_mapping_error(rdma->sc_cm_id->device, - ctxt->sge[sge_no].addr)) - goto err; - atomic_inc(&rdma->sc_dma_used); - ctxt->sge[sge_no].lkey = rdma->sc_dma_lkey; - } else { - ctxt->sge[sge_no].addr = (unsigned long) - vec->sge[sge_no].iov_base; - ctxt->sge[sge_no].lkey = vec->frmr->mr->lkey; - } + ctxt->sge[sge_no].addr = + dma_map_xdr(rdma, &rqstp->rq_res, xdr_off, + sge_bytes, DMA_TO_DEVICE); + xdr_off += sge_bytes; + if (ib_dma_mapping_error(rdma->sc_cm_id->device, + ctxt->sge[sge_no].addr)) + goto err; + atomic_inc(&rdma->sc_dma_used); + ctxt->sge[sge_no].lkey = rdma->sc_dma_lkey; ctxt->sge[sge_no].length = sge_bytes; } BUG_ON(byte_count != 0); @@ -627,6 +450,7 @@ static int send_reply(struct svcxprt_rdma *rdma, ctxt->sge[page_no+1].length = 0; } rqstp->rq_next_page = rqstp->rq_respages + 1; + BUG_ON(sge_no > rdma->sc_max_sge); memset(&send_wr, 0, sizeof send_wr); ctxt->wr_op = IB_WR_SEND; @@ -635,15 +459,6 @@ static int send_reply(struct svcxprt_rdma *rdma, send_wr.num_sge = sge_no; send_wr.opcode = IB_WR_SEND; send_wr.send_flags = IB_SEND_SIGNALED; - if (vec->frmr) { - /* Prepare INVALIDATE WR */ - memset(&inv_wr, 0, sizeof inv_wr); - inv_wr.opcode = IB_WR_LOCAL_INV; - inv_wr.send_flags = IB_SEND_SIGNALED; - inv_wr.ex.invalidate_rkey = - vec->frmr->mr->lkey; - send_wr.next = &inv_wr; - } ret = svc_rdma_send(rdma, &send_wr); if (ret) @@ -653,7 +468,6 @@ static int send_reply(struct svcxprt_rdma *rdma, err: svc_rdma_unmap_dma(ctxt); - svc_rdma_put_frmr(rdma, vec->frmr); svc_rdma_put_context(ctxt, 1); return -EIO; } diff --git a/net/sunrpc/xprtrdma/svc_rdma_transport.c b/net/sunrpc/xprtrdma/svc_rdma_transport.c index 25688fa2207f..e7323fbbd348 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_transport.c +++ b/net/sunrpc/xprtrdma/svc_rdma_transport.c @@ -1,4 +1,5 @@ /* + * Copyright (c) 2014 Open Grid Computing, Inc. All rights reserved. * Copyright (c) 2005-2007 Network Appliance, Inc. All rights reserved. * * This software is available to you under a choice of one of two @@ -65,6 +66,7 @@ static void dto_tasklet_func(unsigned long data); static void svc_rdma_detach(struct svc_xprt *xprt); static void svc_rdma_free(struct svc_xprt *xprt); static int svc_rdma_has_wspace(struct svc_xprt *xprt); +static int svc_rdma_secure_port(struct svc_rqst *); static void rq_cq_reap(struct svcxprt_rdma *xprt); static void sq_cq_reap(struct svcxprt_rdma *xprt); @@ -82,6 +84,7 @@ static struct svc_xprt_ops svc_rdma_ops = { .xpo_prep_reply_hdr = svc_rdma_prep_reply_hdr, .xpo_has_wspace = svc_rdma_has_wspace, .xpo_accept = svc_rdma_accept, + .xpo_secure_port = svc_rdma_secure_port, }; struct svc_xprt_class svc_rdma_class = { @@ -160,7 +163,6 @@ struct svc_rdma_req_map *svc_rdma_get_req_map(void) schedule_timeout_uninterruptible(msecs_to_jiffies(500)); } map->count = 0; - map->frmr = NULL; return map; } @@ -336,22 +338,21 @@ static void process_context(struct svcxprt_rdma *xprt, switch (ctxt->wr_op) { case IB_WR_SEND: - if (test_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags)) - svc_rdma_put_frmr(xprt, ctxt->frmr); + BUG_ON(ctxt->frmr); svc_rdma_put_context(ctxt, 1); break; case IB_WR_RDMA_WRITE: + BUG_ON(ctxt->frmr); svc_rdma_put_context(ctxt, 0); break; case IB_WR_RDMA_READ: case IB_WR_RDMA_READ_WITH_INV: + svc_rdma_put_frmr(xprt, ctxt->frmr); if (test_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags)) { struct svc_rdma_op_ctxt *read_hdr = ctxt->read_hdr; BUG_ON(!read_hdr); - if (test_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags)) - svc_rdma_put_frmr(xprt, ctxt->frmr); spin_lock_bh(&xprt->sc_rq_dto_lock); set_bit(XPT_DATA, &xprt->sc_xprt.xpt_flags); list_add_tail(&read_hdr->dto_q, @@ -363,6 +364,7 @@ static void process_context(struct svcxprt_rdma *xprt, break; default: + BUG_ON(1); printk(KERN_ERR "svcrdma: unexpected completion type, " "opcode=%d\n", ctxt->wr_op); @@ -378,29 +380,42 @@ static void process_context(struct svcxprt_rdma *xprt, static void sq_cq_reap(struct svcxprt_rdma *xprt) { struct svc_rdma_op_ctxt *ctxt = NULL; - struct ib_wc wc; + struct ib_wc wc_a[6]; + struct ib_wc *wc; struct ib_cq *cq = xprt->sc_sq_cq; int ret; + memset(wc_a, 0, sizeof(wc_a)); + if (!test_and_clear_bit(RDMAXPRT_SQ_PENDING, &xprt->sc_flags)) return; ib_req_notify_cq(xprt->sc_sq_cq, IB_CQ_NEXT_COMP); atomic_inc(&rdma_stat_sq_poll); - while ((ret = ib_poll_cq(cq, 1, &wc)) > 0) { - if (wc.status != IB_WC_SUCCESS) - /* Close the transport */ - set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags); + while ((ret = ib_poll_cq(cq, ARRAY_SIZE(wc_a), wc_a)) > 0) { + int i; - /* Decrement used SQ WR count */ - atomic_dec(&xprt->sc_sq_count); - wake_up(&xprt->sc_send_wait); + for (i = 0; i < ret; i++) { + wc = &wc_a[i]; + if (wc->status != IB_WC_SUCCESS) { + dprintk("svcrdma: sq wc err status %d\n", + wc->status); - ctxt = (struct svc_rdma_op_ctxt *)(unsigned long)wc.wr_id; - if (ctxt) - process_context(xprt, ctxt); + /* Close the transport */ + set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags); + } - svc_xprt_put(&xprt->sc_xprt); + /* Decrement used SQ WR count */ + atomic_dec(&xprt->sc_sq_count); + wake_up(&xprt->sc_send_wait); + + ctxt = (struct svc_rdma_op_ctxt *) + (unsigned long)wc->wr_id; + if (ctxt) + process_context(xprt, ctxt); + + svc_xprt_put(&xprt->sc_xprt); + } } if (ctxt) @@ -993,7 +1008,11 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt) need_dma_mr = 0; break; case RDMA_TRANSPORT_IB: - if (!(devattr.device_cap_flags & IB_DEVICE_LOCAL_DMA_LKEY)) { + if (!(newxprt->sc_dev_caps & SVCRDMA_DEVCAP_FAST_REG)) { + need_dma_mr = 1; + dma_mr_acc = IB_ACCESS_LOCAL_WRITE; + } else if (!(devattr.device_cap_flags & + IB_DEVICE_LOCAL_DMA_LKEY)) { need_dma_mr = 1; dma_mr_acc = IB_ACCESS_LOCAL_WRITE; } else @@ -1190,14 +1209,7 @@ static int svc_rdma_has_wspace(struct svc_xprt *xprt) container_of(xprt, struct svcxprt_rdma, sc_xprt); /* - * If there are fewer SQ WR available than required to send a - * simple response, return false. - */ - if ((rdma->sc_sq_depth - atomic_read(&rdma->sc_sq_count) < 3)) - return 0; - - /* - * ...or there are already waiters on the SQ, + * If there are already waiters on the SQ, * return false. */ if (waitqueue_active(&rdma->sc_send_wait)) @@ -1207,6 +1219,11 @@ static int svc_rdma_has_wspace(struct svc_xprt *xprt) return 1; } +static int svc_rdma_secure_port(struct svc_rqst *rqstp) +{ + return 1; +} + /* * Attempt to register the kvec representing the RPC memory with the * device. diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c index 1eb9c468d0c9..66f91f0d071a 100644 --- a/net/sunrpc/xprtrdma/transport.c +++ b/net/sunrpc/xprtrdma/transport.c @@ -149,6 +149,11 @@ static struct ctl_table sunrpc_table[] = { #endif +#define RPCRDMA_BIND_TO (60U * HZ) +#define RPCRDMA_INIT_REEST_TO (5U * HZ) +#define RPCRDMA_MAX_REEST_TO (30U * HZ) +#define RPCRDMA_IDLE_DISC_TO (5U * 60 * HZ) + static struct rpc_xprt_ops xprt_rdma_procs; /* forward reference */ static void @@ -229,7 +234,6 @@ static void xprt_rdma_destroy(struct rpc_xprt *xprt) { struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); - int rc; dprintk("RPC: %s: called\n", __func__); @@ -238,10 +242,7 @@ xprt_rdma_destroy(struct rpc_xprt *xprt) xprt_clear_connected(xprt); rpcrdma_buffer_destroy(&r_xprt->rx_buf); - rc = rpcrdma_ep_destroy(&r_xprt->rx_ep, &r_xprt->rx_ia); - if (rc) - dprintk("RPC: %s: rpcrdma_ep_destroy returned %i\n", - __func__, rc); + rpcrdma_ep_destroy(&r_xprt->rx_ep, &r_xprt->rx_ia); rpcrdma_ia_close(&r_xprt->rx_ia); xprt_rdma_free_addresses(xprt); @@ -289,9 +290,9 @@ xprt_setup_rdma(struct xprt_create *args) /* 60 second timeout, no retries */ xprt->timeout = &xprt_rdma_default_timeout; - xprt->bind_timeout = (60U * HZ); - xprt->reestablish_timeout = (5U * HZ); - xprt->idle_timeout = (5U * 60 * HZ); + xprt->bind_timeout = RPCRDMA_BIND_TO; + xprt->reestablish_timeout = RPCRDMA_INIT_REEST_TO; + xprt->idle_timeout = RPCRDMA_IDLE_DISC_TO; xprt->resvport = 0; /* privileged port not needed */ xprt->tsh_size = 0; /* RPC-RDMA handles framing */ @@ -391,7 +392,7 @@ out4: xprt_rdma_free_addresses(xprt); rc = -EINVAL; out3: - (void) rpcrdma_ep_destroy(new_ep, &new_xprt->rx_ia); + rpcrdma_ep_destroy(new_ep, &new_xprt->rx_ia); out2: rpcrdma_ia_close(&new_xprt->rx_ia); out1: @@ -436,10 +437,10 @@ xprt_rdma_connect(struct rpc_xprt *xprt, struct rpc_task *task) schedule_delayed_work(&r_xprt->rdma_connect, xprt->reestablish_timeout); xprt->reestablish_timeout <<= 1; - if (xprt->reestablish_timeout > (30 * HZ)) - xprt->reestablish_timeout = (30 * HZ); - else if (xprt->reestablish_timeout < (5 * HZ)) - xprt->reestablish_timeout = (5 * HZ); + if (xprt->reestablish_timeout > RPCRDMA_MAX_REEST_TO) + xprt->reestablish_timeout = RPCRDMA_MAX_REEST_TO; + else if (xprt->reestablish_timeout < RPCRDMA_INIT_REEST_TO) + xprt->reestablish_timeout = RPCRDMA_INIT_REEST_TO; } else { schedule_delayed_work(&r_xprt->rdma_connect, 0); if (!RPC_IS_ASYNC(task)) @@ -447,23 +448,6 @@ xprt_rdma_connect(struct rpc_xprt *xprt, struct rpc_task *task) } } -static int -xprt_rdma_reserve_xprt(struct rpc_xprt *xprt, struct rpc_task *task) -{ - struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); - int credits = atomic_read(&r_xprt->rx_buf.rb_credits); - - /* == RPC_CWNDSCALE @ init, but *after* setup */ - if (r_xprt->rx_buf.rb_cwndscale == 0UL) { - r_xprt->rx_buf.rb_cwndscale = xprt->cwnd; - dprintk("RPC: %s: cwndscale %lu\n", __func__, - r_xprt->rx_buf.rb_cwndscale); - BUG_ON(r_xprt->rx_buf.rb_cwndscale <= 0); - } - xprt->cwnd = credits * r_xprt->rx_buf.rb_cwndscale; - return xprt_reserve_xprt_cong(xprt, task); -} - /* * The RDMA allocate/free functions need the task structure as a place * to hide the struct rpcrdma_req, which is necessary for the actual send/recv @@ -479,7 +463,8 @@ xprt_rdma_allocate(struct rpc_task *task, size_t size) struct rpcrdma_req *req, *nreq; req = rpcrdma_buffer_get(&rpcx_to_rdmax(xprt)->rx_buf); - BUG_ON(NULL == req); + if (req == NULL) + return NULL; if (size > req->rl_size) { dprintk("RPC: %s: size %zd too large for buffer[%zd]: " @@ -503,18 +488,6 @@ xprt_rdma_allocate(struct rpc_task *task, size_t size) * If the allocation or registration fails, the RPC framework * will (doggedly) retry. */ - if (rpcx_to_rdmax(xprt)->rx_ia.ri_memreg_strategy == - RPCRDMA_BOUNCEBUFFERS) { - /* forced to "pure inline" */ - dprintk("RPC: %s: too much data (%zd) for inline " - "(r/w max %d/%d)\n", __func__, size, - rpcx_to_rdmad(xprt).inline_rsize, - rpcx_to_rdmad(xprt).inline_wsize); - size = req->rl_size; - rpc_exit(task, -EIO); /* fail the operation */ - rpcx_to_rdmax(xprt)->rx_stats.failed_marshal_count++; - goto out; - } if (task->tk_flags & RPC_TASK_SWAPPER) nreq = kmalloc(sizeof *req + size, GFP_ATOMIC); else @@ -543,7 +516,6 @@ xprt_rdma_allocate(struct rpc_task *task, size_t size) req = nreq; } dprintk("RPC: %s: size %zd, request 0x%p\n", __func__, size, req); -out: req->rl_connect_cookie = 0; /* our reserved value */ return req->rl_xdr_buf; @@ -579,9 +551,7 @@ xprt_rdma_free(void *buffer) __func__, rep, (rep && rep->rr_func) ? " (with waiter)" : ""); /* - * Finish the deregistration. When using mw bind, this was - * begun in rpcrdma_reply_handler(). In all other modes, we - * do it here, in thread context. The process is considered + * Finish the deregistration. The process is considered * complete when the rr_func vector becomes NULL - this * was put in place during rpcrdma_reply_handler() - the wait * call below will not block if the dereg is "done". If @@ -590,12 +560,7 @@ xprt_rdma_free(void *buffer) for (i = 0; req->rl_nchunks;) { --req->rl_nchunks; i += rpcrdma_deregister_external( - &req->rl_segments[i], r_xprt, NULL); - } - - if (rep && wait_event_interruptible(rep->rr_unbind, !rep->rr_func)) { - rep->rr_func = NULL; /* abandon the callback */ - req->rl_reply = NULL; + &req->rl_segments[i], r_xprt); } if (req->rl_iov.length == 0) { /* see allocate above */ @@ -630,13 +595,12 @@ xprt_rdma_send_request(struct rpc_task *task) struct rpc_xprt *xprt = rqst->rq_xprt; struct rpcrdma_req *req = rpcr_to_rdmar(rqst); struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); + int rc; - /* marshal the send itself */ - if (req->rl_niovs == 0 && rpcrdma_marshal_req(rqst) != 0) { - r_xprt->rx_stats.failed_marshal_count++; - dprintk("RPC: %s: rpcrdma_marshal_req failed\n", - __func__); - return -EIO; + if (req->rl_niovs == 0) { + rc = rpcrdma_marshal_req(rqst); + if (rc < 0) + goto failed_marshal; } if (req->rl_reply == NULL) /* e.g. reconnection */ @@ -660,6 +624,12 @@ xprt_rdma_send_request(struct rpc_task *task) rqst->rq_bytes_sent = 0; return 0; +failed_marshal: + r_xprt->rx_stats.failed_marshal_count++; + dprintk("RPC: %s: rpcrdma_marshal_req failed, status %i\n", + __func__, rc); + if (rc == -EIO) + return -EIO; drop_connection: xprt_disconnect_done(xprt); return -ENOTCONN; /* implies disconnect */ @@ -705,7 +675,7 @@ static void xprt_rdma_print_stats(struct rpc_xprt *xprt, struct seq_file *seq) */ static struct rpc_xprt_ops xprt_rdma_procs = { - .reserve_xprt = xprt_rdma_reserve_xprt, + .reserve_xprt = xprt_reserve_xprt_cong, .release_xprt = xprt_release_xprt_cong, /* sunrpc/xprt.c */ .alloc_slot = xprt_alloc_slot, .release_request = xprt_release_rqst_cong, /* ditto */ diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c index 93726560eaa8..13dbd1c389ff 100644 --- a/net/sunrpc/xprtrdma/verbs.c +++ b/net/sunrpc/xprtrdma/verbs.c @@ -48,8 +48,8 @@ */ #include <linux/interrupt.h> -#include <linux/pci.h> /* for Tavor hack below */ #include <linux/slab.h> +#include <asm/bitops.h> #include "xprt_rdma.h" @@ -142,98 +142,139 @@ rpcrdma_cq_async_error_upcall(struct ib_event *event, void *context) } } -static inline -void rpcrdma_event_process(struct ib_wc *wc) +static void +rpcrdma_sendcq_process_wc(struct ib_wc *wc) { - struct rpcrdma_mw *frmr; - struct rpcrdma_rep *rep = - (struct rpcrdma_rep *)(unsigned long) wc->wr_id; + struct rpcrdma_mw *frmr = (struct rpcrdma_mw *)(unsigned long)wc->wr_id; - dprintk("RPC: %s: event rep %p status %X opcode %X length %u\n", - __func__, rep, wc->status, wc->opcode, wc->byte_len); + dprintk("RPC: %s: frmr %p status %X opcode %d\n", + __func__, frmr, wc->status, wc->opcode); - if (!rep) /* send or bind completion that we don't care about */ + if (wc->wr_id == 0ULL) return; - - if (IB_WC_SUCCESS != wc->status) { - dprintk("RPC: %s: WC opcode %d status %X, connection lost\n", - __func__, wc->opcode, wc->status); - rep->rr_len = ~0U; - if (wc->opcode != IB_WC_FAST_REG_MR && wc->opcode != IB_WC_LOCAL_INV) - rpcrdma_schedule_tasklet(rep); + if (wc->status != IB_WC_SUCCESS) return; - } - switch (wc->opcode) { - case IB_WC_FAST_REG_MR: - frmr = (struct rpcrdma_mw *)(unsigned long)wc->wr_id; + if (wc->opcode == IB_WC_FAST_REG_MR) frmr->r.frmr.state = FRMR_IS_VALID; - break; - case IB_WC_LOCAL_INV: - frmr = (struct rpcrdma_mw *)(unsigned long)wc->wr_id; + else if (wc->opcode == IB_WC_LOCAL_INV) frmr->r.frmr.state = FRMR_IS_INVALID; - break; - case IB_WC_RECV: - rep->rr_len = wc->byte_len; - ib_dma_sync_single_for_cpu( - rdmab_to_ia(rep->rr_buffer)->ri_id->device, - rep->rr_iov.addr, rep->rr_len, DMA_FROM_DEVICE); - /* Keep (only) the most recent credits, after check validity */ - if (rep->rr_len >= 16) { - struct rpcrdma_msg *p = - (struct rpcrdma_msg *) rep->rr_base; - unsigned int credits = ntohl(p->rm_credit); - if (credits == 0) { - dprintk("RPC: %s: server" - " dropped credits to 0!\n", __func__); - /* don't deadlock */ - credits = 1; - } else if (credits > rep->rr_buffer->rb_max_requests) { - dprintk("RPC: %s: server" - " over-crediting: %d (%d)\n", - __func__, credits, - rep->rr_buffer->rb_max_requests); - credits = rep->rr_buffer->rb_max_requests; - } - atomic_set(&rep->rr_buffer->rb_credits, credits); - } - /* fall through */ - case IB_WC_BIND_MW: - rpcrdma_schedule_tasklet(rep); - break; - default: - dprintk("RPC: %s: unexpected WC event %X\n", - __func__, wc->opcode); - break; - } } -static inline int -rpcrdma_cq_poll(struct ib_cq *cq) +static int +rpcrdma_sendcq_poll(struct ib_cq *cq, struct rpcrdma_ep *ep) { - struct ib_wc wc; - int rc; + struct ib_wc *wcs; + int budget, count, rc; - for (;;) { - rc = ib_poll_cq(cq, 1, &wc); - if (rc < 0) { - dprintk("RPC: %s: ib_poll_cq failed %i\n", - __func__, rc); + budget = RPCRDMA_WC_BUDGET / RPCRDMA_POLLSIZE; + do { + wcs = ep->rep_send_wcs; + + rc = ib_poll_cq(cq, RPCRDMA_POLLSIZE, wcs); + if (rc <= 0) return rc; - } - if (rc == 0) - break; - rpcrdma_event_process(&wc); + count = rc; + while (count-- > 0) + rpcrdma_sendcq_process_wc(wcs++); + } while (rc == RPCRDMA_POLLSIZE && --budget); + return 0; +} + +/* + * Handle send, fast_reg_mr, and local_inv completions. + * + * Send events are typically suppressed and thus do not result + * in an upcall. Occasionally one is signaled, however. This + * prevents the provider's completion queue from wrapping and + * losing a completion. + */ +static void +rpcrdma_sendcq_upcall(struct ib_cq *cq, void *cq_context) +{ + struct rpcrdma_ep *ep = (struct rpcrdma_ep *)cq_context; + int rc; + + rc = rpcrdma_sendcq_poll(cq, ep); + if (rc) { + dprintk("RPC: %s: ib_poll_cq failed: %i\n", + __func__, rc); + return; } + rc = ib_req_notify_cq(cq, + IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS); + if (rc == 0) + return; + if (rc < 0) { + dprintk("RPC: %s: ib_req_notify_cq failed: %i\n", + __func__, rc); + return; + } + + rpcrdma_sendcq_poll(cq, ep); +} + +static void +rpcrdma_recvcq_process_wc(struct ib_wc *wc) +{ + struct rpcrdma_rep *rep = + (struct rpcrdma_rep *)(unsigned long)wc->wr_id; + + dprintk("RPC: %s: rep %p status %X opcode %X length %u\n", + __func__, rep, wc->status, wc->opcode, wc->byte_len); + + if (wc->status != IB_WC_SUCCESS) { + rep->rr_len = ~0U; + goto out_schedule; + } + if (wc->opcode != IB_WC_RECV) + return; + + rep->rr_len = wc->byte_len; + ib_dma_sync_single_for_cpu(rdmab_to_ia(rep->rr_buffer)->ri_id->device, + rep->rr_iov.addr, rep->rr_len, DMA_FROM_DEVICE); + + if (rep->rr_len >= 16) { + struct rpcrdma_msg *p = (struct rpcrdma_msg *)rep->rr_base; + unsigned int credits = ntohl(p->rm_credit); + + if (credits == 0) + credits = 1; /* don't deadlock */ + else if (credits > rep->rr_buffer->rb_max_requests) + credits = rep->rr_buffer->rb_max_requests; + atomic_set(&rep->rr_buffer->rb_credits, credits); + } + +out_schedule: + rpcrdma_schedule_tasklet(rep); +} + +static int +rpcrdma_recvcq_poll(struct ib_cq *cq, struct rpcrdma_ep *ep) +{ + struct ib_wc *wcs; + int budget, count, rc; + + budget = RPCRDMA_WC_BUDGET / RPCRDMA_POLLSIZE; + do { + wcs = ep->rep_recv_wcs; + + rc = ib_poll_cq(cq, RPCRDMA_POLLSIZE, wcs); + if (rc <= 0) + return rc; + + count = rc; + while (count-- > 0) + rpcrdma_recvcq_process_wc(wcs++); + } while (rc == RPCRDMA_POLLSIZE && --budget); return 0; } /* - * rpcrdma_cq_event_upcall + * Handle receive completions. * - * This upcall handles recv, send, bind and unbind events. * It is reentrant but processes single events in order to maintain * ordering of receives to keep server credits. * @@ -242,26 +283,31 @@ rpcrdma_cq_poll(struct ib_cq *cq) * connection shutdown. That is, the structures required for * the completion of the reply handler must remain intact until * all memory has been reclaimed. - * - * Note that send events are suppressed and do not result in an upcall. */ static void -rpcrdma_cq_event_upcall(struct ib_cq *cq, void *context) +rpcrdma_recvcq_upcall(struct ib_cq *cq, void *cq_context) { + struct rpcrdma_ep *ep = (struct rpcrdma_ep *)cq_context; int rc; - rc = rpcrdma_cq_poll(cq); - if (rc) + rc = rpcrdma_recvcq_poll(cq, ep); + if (rc) { + dprintk("RPC: %s: ib_poll_cq failed: %i\n", + __func__, rc); return; + } - rc = ib_req_notify_cq(cq, IB_CQ_NEXT_COMP); - if (rc) { - dprintk("RPC: %s: ib_req_notify_cq failed %i\n", + rc = ib_req_notify_cq(cq, + IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS); + if (rc == 0) + return; + if (rc < 0) { + dprintk("RPC: %s: ib_req_notify_cq failed: %i\n", __func__, rc); return; } - rpcrdma_cq_poll(cq); + rpcrdma_recvcq_poll(cq, ep); } #ifdef RPC_DEBUG @@ -493,54 +539,32 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg) ia->ri_dma_lkey = ia->ri_id->device->local_dma_lkey; } - switch (memreg) { - case RPCRDMA_MEMWINDOWS: - case RPCRDMA_MEMWINDOWS_ASYNC: - if (!(devattr.device_cap_flags & IB_DEVICE_MEM_WINDOW)) { - dprintk("RPC: %s: MEMWINDOWS registration " - "specified but not supported by adapter, " - "using slower RPCRDMA_REGISTER\n", - __func__); - memreg = RPCRDMA_REGISTER; - } - break; - case RPCRDMA_MTHCAFMR: - if (!ia->ri_id->device->alloc_fmr) { -#if RPCRDMA_PERSISTENT_REGISTRATION - dprintk("RPC: %s: MTHCAFMR registration " - "specified but not supported by adapter, " - "using riskier RPCRDMA_ALLPHYSICAL\n", - __func__); - memreg = RPCRDMA_ALLPHYSICAL; -#else - dprintk("RPC: %s: MTHCAFMR registration " - "specified but not supported by adapter, " - "using slower RPCRDMA_REGISTER\n", - __func__); - memreg = RPCRDMA_REGISTER; -#endif - } - break; - case RPCRDMA_FRMR: + if (memreg == RPCRDMA_FRMR) { /* Requires both frmr reg and local dma lkey */ if ((devattr.device_cap_flags & (IB_DEVICE_MEM_MGT_EXTENSIONS|IB_DEVICE_LOCAL_DMA_LKEY)) != (IB_DEVICE_MEM_MGT_EXTENSIONS|IB_DEVICE_LOCAL_DMA_LKEY)) { -#if RPCRDMA_PERSISTENT_REGISTRATION dprintk("RPC: %s: FRMR registration " - "specified but not supported by adapter, " - "using riskier RPCRDMA_ALLPHYSICAL\n", - __func__); + "not supported by HCA\n", __func__); + memreg = RPCRDMA_MTHCAFMR; + } else { + /* Mind the ia limit on FRMR page list depth */ + ia->ri_max_frmr_depth = min_t(unsigned int, + RPCRDMA_MAX_DATA_SEGS, + devattr.max_fast_reg_page_list_len); + } + } + if (memreg == RPCRDMA_MTHCAFMR) { + if (!ia->ri_id->device->alloc_fmr) { + dprintk("RPC: %s: MTHCAFMR registration " + "not supported by HCA\n", __func__); +#if RPCRDMA_PERSISTENT_REGISTRATION memreg = RPCRDMA_ALLPHYSICAL; #else - dprintk("RPC: %s: FRMR registration " - "specified but not supported by adapter, " - "using slower RPCRDMA_REGISTER\n", - __func__); - memreg = RPCRDMA_REGISTER; + rc = -ENOMEM; + goto out2; #endif } - break; } /* @@ -552,8 +576,6 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg) * adapter. */ switch (memreg) { - case RPCRDMA_BOUNCEBUFFERS: - case RPCRDMA_REGISTER: case RPCRDMA_FRMR: break; #if RPCRDMA_PERSISTENT_REGISTRATION @@ -563,30 +585,26 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg) IB_ACCESS_REMOTE_READ; goto register_setup; #endif - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: - mem_priv = IB_ACCESS_LOCAL_WRITE | - IB_ACCESS_MW_BIND; - goto register_setup; case RPCRDMA_MTHCAFMR: if (ia->ri_have_dma_lkey) break; mem_priv = IB_ACCESS_LOCAL_WRITE; +#if RPCRDMA_PERSISTENT_REGISTRATION register_setup: +#endif ia->ri_bind_mem = ib_get_dma_mr(ia->ri_pd, mem_priv); if (IS_ERR(ia->ri_bind_mem)) { printk(KERN_ALERT "%s: ib_get_dma_mr for " - "phys register failed with %lX\n\t" - "Will continue with degraded performance\n", + "phys register failed with %lX\n", __func__, PTR_ERR(ia->ri_bind_mem)); - memreg = RPCRDMA_REGISTER; - ia->ri_bind_mem = NULL; + rc = -ENOMEM; + goto out2; } break; default: - printk(KERN_ERR "%s: invalid memory registration mode %d\n", - __func__, memreg); - rc = -EINVAL; + printk(KERN_ERR "RPC: Unsupported memory " + "registration mode: %d\n", memreg); + rc = -ENOMEM; goto out2; } dprintk("RPC: %s: memory registration strategy is %d\n", @@ -640,6 +658,7 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, struct rpcrdma_create_data_internal *cdata) { struct ib_device_attr devattr; + struct ib_cq *sendcq, *recvcq; int rc, err; rc = ib_query_device(ia->ri_id->device, &devattr); @@ -659,32 +678,42 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, ep->rep_attr.srq = NULL; ep->rep_attr.cap.max_send_wr = cdata->max_requests; switch (ia->ri_memreg_strategy) { - case RPCRDMA_FRMR: + case RPCRDMA_FRMR: { + int depth = 7; + /* Add room for frmr register and invalidate WRs. * 1. FRMR reg WR for head * 2. FRMR invalidate WR for head - * 3. FRMR reg WR for pagelist - * 4. FRMR invalidate WR for pagelist + * 3. N FRMR reg WRs for pagelist + * 4. N FRMR invalidate WRs for pagelist * 5. FRMR reg WR for tail * 6. FRMR invalidate WR for tail * 7. The RDMA_SEND WR */ - ep->rep_attr.cap.max_send_wr *= 7; + + /* Calculate N if the device max FRMR depth is smaller than + * RPCRDMA_MAX_DATA_SEGS. + */ + if (ia->ri_max_frmr_depth < RPCRDMA_MAX_DATA_SEGS) { + int delta = RPCRDMA_MAX_DATA_SEGS - + ia->ri_max_frmr_depth; + + do { + depth += 2; /* FRMR reg + invalidate */ + delta -= ia->ri_max_frmr_depth; + } while (delta > 0); + + } + ep->rep_attr.cap.max_send_wr *= depth; if (ep->rep_attr.cap.max_send_wr > devattr.max_qp_wr) { - cdata->max_requests = devattr.max_qp_wr / 7; + cdata->max_requests = devattr.max_qp_wr / depth; if (!cdata->max_requests) return -EINVAL; - ep->rep_attr.cap.max_send_wr = cdata->max_requests * 7; + ep->rep_attr.cap.max_send_wr = cdata->max_requests * + depth; } break; - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: - /* Add room for mw_binds+unbinds - overkill! */ - ep->rep_attr.cap.max_send_wr++; - ep->rep_attr.cap.max_send_wr *= (2 * RPCRDMA_MAX_SEGS); - if (ep->rep_attr.cap.max_send_wr > devattr.max_qp_wr) - return -EINVAL; - break; + } default: break; } @@ -705,46 +734,51 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, ep->rep_attr.cap.max_recv_sge); /* set trigger for requesting send completion */ - ep->rep_cqinit = ep->rep_attr.cap.max_send_wr/2 /* - 1*/; - switch (ia->ri_memreg_strategy) { - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: - ep->rep_cqinit -= RPCRDMA_MAX_SEGS; - break; - default: - break; - } + ep->rep_cqinit = ep->rep_attr.cap.max_send_wr/2 - 1; if (ep->rep_cqinit <= 2) ep->rep_cqinit = 0; INIT_CQCOUNT(ep); ep->rep_ia = ia; init_waitqueue_head(&ep->rep_connect_wait); + INIT_DELAYED_WORK(&ep->rep_connect_worker, rpcrdma_connect_worker); - /* - * Create a single cq for receive dto and mw_bind (only ever - * care about unbind, really). Send completions are suppressed. - * Use single threaded tasklet upcalls to maintain ordering. - */ - ep->rep_cq = ib_create_cq(ia->ri_id->device, rpcrdma_cq_event_upcall, - rpcrdma_cq_async_error_upcall, NULL, - ep->rep_attr.cap.max_recv_wr + + sendcq = ib_create_cq(ia->ri_id->device, rpcrdma_sendcq_upcall, + rpcrdma_cq_async_error_upcall, ep, ep->rep_attr.cap.max_send_wr + 1, 0); - if (IS_ERR(ep->rep_cq)) { - rc = PTR_ERR(ep->rep_cq); - dprintk("RPC: %s: ib_create_cq failed: %i\n", + if (IS_ERR(sendcq)) { + rc = PTR_ERR(sendcq); + dprintk("RPC: %s: failed to create send CQ: %i\n", __func__, rc); goto out1; } - rc = ib_req_notify_cq(ep->rep_cq, IB_CQ_NEXT_COMP); + rc = ib_req_notify_cq(sendcq, IB_CQ_NEXT_COMP); + if (rc) { + dprintk("RPC: %s: ib_req_notify_cq failed: %i\n", + __func__, rc); + goto out2; + } + + recvcq = ib_create_cq(ia->ri_id->device, rpcrdma_recvcq_upcall, + rpcrdma_cq_async_error_upcall, ep, + ep->rep_attr.cap.max_recv_wr + 1, 0); + if (IS_ERR(recvcq)) { + rc = PTR_ERR(recvcq); + dprintk("RPC: %s: failed to create recv CQ: %i\n", + __func__, rc); + goto out2; + } + + rc = ib_req_notify_cq(recvcq, IB_CQ_NEXT_COMP); if (rc) { dprintk("RPC: %s: ib_req_notify_cq failed: %i\n", __func__, rc); + ib_destroy_cq(recvcq); goto out2; } - ep->rep_attr.send_cq = ep->rep_cq; - ep->rep_attr.recv_cq = ep->rep_cq; + ep->rep_attr.send_cq = sendcq; + ep->rep_attr.recv_cq = recvcq; /* Initialize cma parameters */ @@ -754,9 +788,7 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, /* Client offers RDMA Read but does not initiate */ ep->rep_remote_cma.initiator_depth = 0; - if (ia->ri_memreg_strategy == RPCRDMA_BOUNCEBUFFERS) - ep->rep_remote_cma.responder_resources = 0; - else if (devattr.max_qp_rd_atom > 32) /* arbitrary but <= 255 */ + if (devattr.max_qp_rd_atom > 32) /* arbitrary but <= 255 */ ep->rep_remote_cma.responder_resources = 32; else ep->rep_remote_cma.responder_resources = devattr.max_qp_rd_atom; @@ -768,7 +800,7 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, return 0; out2: - err = ib_destroy_cq(ep->rep_cq); + err = ib_destroy_cq(sendcq); if (err) dprintk("RPC: %s: ib_destroy_cq returned %i\n", __func__, err); @@ -782,11 +814,8 @@ out1: * Disconnect and destroy endpoint. After this, the only * valid operations on the ep are to free it (if dynamically * allocated) or re-create it. - * - * The caller's error handling must be sure to not leak the endpoint - * if this function fails. */ -int +void rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) { int rc; @@ -794,6 +823,8 @@ rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) dprintk("RPC: %s: entering, connected is %d\n", __func__, ep->rep_connected); + cancel_delayed_work_sync(&ep->rep_connect_worker); + if (ia->ri_id->qp) { rc = rpcrdma_ep_disconnect(ep, ia); if (rc) @@ -809,13 +840,17 @@ rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) ep->rep_pad_mr = NULL; } - rpcrdma_clean_cq(ep->rep_cq); - rc = ib_destroy_cq(ep->rep_cq); + rpcrdma_clean_cq(ep->rep_attr.recv_cq); + rc = ib_destroy_cq(ep->rep_attr.recv_cq); if (rc) dprintk("RPC: %s: ib_destroy_cq returned %i\n", __func__, rc); - return rc; + rpcrdma_clean_cq(ep->rep_attr.send_cq); + rc = ib_destroy_cq(ep->rep_attr.send_cq); + if (rc) + dprintk("RPC: %s: ib_destroy_cq returned %i\n", + __func__, rc); } /* @@ -831,17 +866,20 @@ rpcrdma_ep_connect(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) if (ep->rep_connected != 0) { struct rpcrdma_xprt *xprt; retry: + dprintk("RPC: %s: reconnecting...\n", __func__); rc = rpcrdma_ep_disconnect(ep, ia); if (rc && rc != -ENOTCONN) dprintk("RPC: %s: rpcrdma_ep_disconnect" " status %i\n", __func__, rc); - rpcrdma_clean_cq(ep->rep_cq); + + rpcrdma_clean_cq(ep->rep_attr.recv_cq); + rpcrdma_clean_cq(ep->rep_attr.send_cq); xprt = container_of(ia, struct rpcrdma_xprt, rx_ia); id = rpcrdma_create_id(xprt, ia, (struct sockaddr *)&xprt->rx_data.addr); if (IS_ERR(id)) { - rc = PTR_ERR(id); + rc = -EHOSTUNREACH; goto out; } /* TEMP TEMP TEMP - fail if new device: @@ -855,35 +893,32 @@ retry: printk("RPC: %s: can't reconnect on " "different device!\n", __func__); rdma_destroy_id(id); - rc = -ENETDOWN; + rc = -ENETUNREACH; goto out; } /* END TEMP */ + rc = rdma_create_qp(id, ia->ri_pd, &ep->rep_attr); + if (rc) { + dprintk("RPC: %s: rdma_create_qp failed %i\n", + __func__, rc); + rdma_destroy_id(id); + rc = -ENETUNREACH; + goto out; + } rdma_destroy_qp(ia->ri_id); rdma_destroy_id(ia->ri_id); ia->ri_id = id; + } else { + dprintk("RPC: %s: connecting...\n", __func__); + rc = rdma_create_qp(ia->ri_id, ia->ri_pd, &ep->rep_attr); + if (rc) { + dprintk("RPC: %s: rdma_create_qp failed %i\n", + __func__, rc); + /* do not update ep->rep_connected */ + return -ENETUNREACH; + } } - rc = rdma_create_qp(ia->ri_id, ia->ri_pd, &ep->rep_attr); - if (rc) { - dprintk("RPC: %s: rdma_create_qp failed %i\n", - __func__, rc); - goto out; - } - -/* XXX Tavor device performs badly with 2K MTU! */ -if (strnicmp(ia->ri_id->device->dma_device->bus->name, "pci", 3) == 0) { - struct pci_dev *pcid = to_pci_dev(ia->ri_id->device->dma_device); - if (pcid->device == PCI_DEVICE_ID_MELLANOX_TAVOR && - (pcid->vendor == PCI_VENDOR_ID_MELLANOX || - pcid->vendor == PCI_VENDOR_ID_TOPSPIN)) { - struct ib_qp_attr attr = { - .path_mtu = IB_MTU_1024 - }; - rc = ib_modify_qp(ia->ri_id->qp, &attr, IB_QP_PATH_MTU); - } -} - ep->rep_connected = 0; rc = rdma_connect(ia->ri_id, &ep->rep_remote_cma); @@ -944,7 +979,8 @@ rpcrdma_ep_disconnect(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) { int rc; - rpcrdma_clean_cq(ep->rep_cq); + rpcrdma_clean_cq(ep->rep_attr.recv_cq); + rpcrdma_clean_cq(ep->rep_attr.send_cq); rc = rdma_disconnect(ia->ri_id); if (!rc) { /* returns without wait if not connected */ @@ -967,7 +1003,7 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, struct rpcrdma_create_data_internal *cdata) { char *p; - size_t len; + size_t len, rlen, wlen; int i, rc; struct rpcrdma_mw *r; @@ -997,11 +1033,6 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep, len += (buf->rb_max_requests + 1) * RPCRDMA_MAX_SEGS * sizeof(struct rpcrdma_mw); break; - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: - len += (buf->rb_max_requests + 1) * RPCRDMA_MAX_SEGS * - sizeof(struct rpcrdma_mw); - break; default: break; } @@ -1032,32 +1063,29 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep, } p += cdata->padding; - /* - * Allocate the fmr's, or mw's for mw_bind chunk registration. - * We "cycle" the mw's in order to minimize rkey reuse, - * and also reduce unbind-to-bind collision. - */ INIT_LIST_HEAD(&buf->rb_mws); r = (struct rpcrdma_mw *)p; switch (ia->ri_memreg_strategy) { case RPCRDMA_FRMR: for (i = buf->rb_max_requests * RPCRDMA_MAX_SEGS; i; i--) { r->r.frmr.fr_mr = ib_alloc_fast_reg_mr(ia->ri_pd, - RPCRDMA_MAX_SEGS); + ia->ri_max_frmr_depth); if (IS_ERR(r->r.frmr.fr_mr)) { rc = PTR_ERR(r->r.frmr.fr_mr); dprintk("RPC: %s: ib_alloc_fast_reg_mr" " failed %i\n", __func__, rc); goto out; } - r->r.frmr.fr_pgl = - ib_alloc_fast_reg_page_list(ia->ri_id->device, - RPCRDMA_MAX_SEGS); + r->r.frmr.fr_pgl = ib_alloc_fast_reg_page_list( + ia->ri_id->device, + ia->ri_max_frmr_depth); if (IS_ERR(r->r.frmr.fr_pgl)) { rc = PTR_ERR(r->r.frmr.fr_pgl); dprintk("RPC: %s: " "ib_alloc_fast_reg_page_list " "failed %i\n", __func__, rc); + + ib_dereg_mr(r->r.frmr.fr_mr); goto out; } list_add(&r->mw_list, &buf->rb_mws); @@ -1082,21 +1110,6 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep, ++r; } break; - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: - /* Allocate one extra request's worth, for full cycling */ - for (i = (buf->rb_max_requests+1) * RPCRDMA_MAX_SEGS; i; i--) { - r->r.mw = ib_alloc_mw(ia->ri_pd, IB_MW_TYPE_1); - if (IS_ERR(r->r.mw)) { - rc = PTR_ERR(r->r.mw); - dprintk("RPC: %s: ib_alloc_mw" - " failed %i\n", __func__, rc); - goto out; - } - list_add(&r->mw_list, &buf->rb_mws); - ++r; - } - break; default: break; } @@ -1105,16 +1118,16 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep, * Allocate/init the request/reply buffers. Doing this * using kmalloc for now -- one for each buf. */ + wlen = 1 << fls(cdata->inline_wsize + sizeof(struct rpcrdma_req)); + rlen = 1 << fls(cdata->inline_rsize + sizeof(struct rpcrdma_rep)); + dprintk("RPC: %s: wlen = %zu, rlen = %zu\n", + __func__, wlen, rlen); + for (i = 0; i < buf->rb_max_requests; i++) { struct rpcrdma_req *req; struct rpcrdma_rep *rep; - len = cdata->inline_wsize + sizeof(struct rpcrdma_req); - /* RPC layer requests *double* size + 1K RPC_SLACK_SPACE! */ - /* Typical ~2400b, so rounding up saves work later */ - if (len < 4096) - len = 4096; - req = kmalloc(len, GFP_KERNEL); + req = kmalloc(wlen, GFP_KERNEL); if (req == NULL) { dprintk("RPC: %s: request buffer %d alloc" " failed\n", __func__, i); @@ -1126,16 +1139,16 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep, buf->rb_send_bufs[i]->rl_buffer = buf; rc = rpcrdma_register_internal(ia, req->rl_base, - len - offsetof(struct rpcrdma_req, rl_base), + wlen - offsetof(struct rpcrdma_req, rl_base), &buf->rb_send_bufs[i]->rl_handle, &buf->rb_send_bufs[i]->rl_iov); if (rc) goto out; - buf->rb_send_bufs[i]->rl_size = len-sizeof(struct rpcrdma_req); + buf->rb_send_bufs[i]->rl_size = wlen - + sizeof(struct rpcrdma_req); - len = cdata->inline_rsize + sizeof(struct rpcrdma_rep); - rep = kmalloc(len, GFP_KERNEL); + rep = kmalloc(rlen, GFP_KERNEL); if (rep == NULL) { dprintk("RPC: %s: reply buffer %d alloc failed\n", __func__, i); @@ -1145,10 +1158,9 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep, memset(rep, 0, sizeof(struct rpcrdma_rep)); buf->rb_recv_bufs[i] = rep; buf->rb_recv_bufs[i]->rr_buffer = buf; - init_waitqueue_head(&rep->rr_unbind); rc = rpcrdma_register_internal(ia, rep->rr_base, - len - offsetof(struct rpcrdma_rep, rr_base), + rlen - offsetof(struct rpcrdma_rep, rr_base), &buf->rb_recv_bufs[i]->rr_handle, &buf->rb_recv_bufs[i]->rr_iov); if (rc) @@ -1179,7 +1191,6 @@ rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf) /* clean up in reverse order from create * 1. recv mr memory (mr free, then kfree) - * 1a. bind mw memory * 2. send mr memory (mr free, then kfree) * 3. padding (if any) [moved to rpcrdma_ep_destroy] * 4. arrays @@ -1194,41 +1205,6 @@ rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf) kfree(buf->rb_recv_bufs[i]); } if (buf->rb_send_bufs && buf->rb_send_bufs[i]) { - while (!list_empty(&buf->rb_mws)) { - r = list_entry(buf->rb_mws.next, - struct rpcrdma_mw, mw_list); - list_del(&r->mw_list); - switch (ia->ri_memreg_strategy) { - case RPCRDMA_FRMR: - rc = ib_dereg_mr(r->r.frmr.fr_mr); - if (rc) - dprintk("RPC: %s:" - " ib_dereg_mr" - " failed %i\n", - __func__, rc); - ib_free_fast_reg_page_list(r->r.frmr.fr_pgl); - break; - case RPCRDMA_MTHCAFMR: - rc = ib_dealloc_fmr(r->r.fmr); - if (rc) - dprintk("RPC: %s:" - " ib_dealloc_fmr" - " failed %i\n", - __func__, rc); - break; - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: - rc = ib_dealloc_mw(r->r.mw); - if (rc) - dprintk("RPC: %s:" - " ib_dealloc_mw" - " failed %i\n", - __func__, rc); - break; - default: - break; - } - } rpcrdma_deregister_internal(ia, buf->rb_send_bufs[i]->rl_handle, &buf->rb_send_bufs[i]->rl_iov); @@ -1236,6 +1212,33 @@ rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf) } } + while (!list_empty(&buf->rb_mws)) { + r = list_entry(buf->rb_mws.next, + struct rpcrdma_mw, mw_list); + list_del(&r->mw_list); + switch (ia->ri_memreg_strategy) { + case RPCRDMA_FRMR: + rc = ib_dereg_mr(r->r.frmr.fr_mr); + if (rc) + dprintk("RPC: %s:" + " ib_dereg_mr" + " failed %i\n", + __func__, rc); + ib_free_fast_reg_page_list(r->r.frmr.fr_pgl); + break; + case RPCRDMA_MTHCAFMR: + rc = ib_dealloc_fmr(r->r.fmr); + if (rc) + dprintk("RPC: %s:" + " ib_dealloc_fmr" + " failed %i\n", + __func__, rc); + break; + default: + break; + } + } + kfree(buf->rb_pool); } @@ -1299,21 +1302,17 @@ rpcrdma_buffer_put(struct rpcrdma_req *req) int i; unsigned long flags; - BUG_ON(req->rl_nchunks != 0); spin_lock_irqsave(&buffers->rb_lock, flags); buffers->rb_send_bufs[--buffers->rb_send_index] = req; req->rl_niovs = 0; if (req->rl_reply) { buffers->rb_recv_bufs[--buffers->rb_recv_index] = req->rl_reply; - init_waitqueue_head(&req->rl_reply->rr_unbind); req->rl_reply->rr_func = NULL; req->rl_reply = NULL; } switch (ia->ri_memreg_strategy) { case RPCRDMA_FRMR: case RPCRDMA_MTHCAFMR: - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: /* * Cycle mw's back in reverse order, and "spin" them. * This delays and scrambles reuse as much as possible. @@ -1358,8 +1357,7 @@ rpcrdma_recv_buffer_get(struct rpcrdma_req *req) /* * Put reply buffers back into pool when not attached to - * request. This happens in error conditions, and when - * aborting unbinds. Pre-decrement counter/array index. + * request. This happens in error conditions. */ void rpcrdma_recv_buffer_put(struct rpcrdma_rep *rep) @@ -1498,8 +1496,8 @@ rpcrdma_register_frmr_external(struct rpcrdma_mr_seg *seg, seg1->mr_offset -= pageoff; /* start of page */ seg1->mr_len += pageoff; len = -pageoff; - if (*nsegs > RPCRDMA_MAX_DATA_SEGS) - *nsegs = RPCRDMA_MAX_DATA_SEGS; + if (*nsegs > ia->ri_max_frmr_depth) + *nsegs = ia->ri_max_frmr_depth; for (page_no = i = 0; i < *nsegs;) { rpcrdma_map_one(ia, seg, writing); pa = seg->mr_dma; @@ -1536,10 +1534,6 @@ rpcrdma_register_frmr_external(struct rpcrdma_mr_seg *seg, } else post_wr = &frmr_wr; - /* Bump the key */ - key = (u8)(seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey & 0x000000FF); - ib_update_fast_reg_key(seg1->mr_chunk.rl_mw->r.frmr.fr_mr, ++key); - /* Prepare FRMR WR */ memset(&frmr_wr, 0, sizeof frmr_wr); frmr_wr.wr_id = (unsigned long)(void *)seg1->mr_chunk.rl_mw; @@ -1550,7 +1544,16 @@ rpcrdma_register_frmr_external(struct rpcrdma_mr_seg *seg, frmr_wr.wr.fast_reg.page_list_len = page_no; frmr_wr.wr.fast_reg.page_shift = PAGE_SHIFT; frmr_wr.wr.fast_reg.length = page_no << PAGE_SHIFT; - BUG_ON(frmr_wr.wr.fast_reg.length < len); + if (frmr_wr.wr.fast_reg.length < len) { + while (seg1->mr_nsegs--) + rpcrdma_unmap_one(ia, seg++); + return -EIO; + } + + /* Bump the key */ + key = (u8)(seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey & 0x000000FF); + ib_update_fast_reg_key(seg1->mr_chunk.rl_mw->r.frmr.fr_mr, ++key); + frmr_wr.wr.fast_reg.access_flags = (writing ? IB_ACCESS_REMOTE_WRITE | IB_ACCESS_LOCAL_WRITE : IB_ACCESS_REMOTE_READ); @@ -1661,135 +1664,6 @@ rpcrdma_deregister_fmr_external(struct rpcrdma_mr_seg *seg, return rc; } -static int -rpcrdma_register_memwin_external(struct rpcrdma_mr_seg *seg, - int *nsegs, int writing, struct rpcrdma_ia *ia, - struct rpcrdma_xprt *r_xprt) -{ - int mem_priv = (writing ? IB_ACCESS_REMOTE_WRITE : - IB_ACCESS_REMOTE_READ); - struct ib_mw_bind param; - int rc; - - *nsegs = 1; - rpcrdma_map_one(ia, seg, writing); - param.bind_info.mr = ia->ri_bind_mem; - param.wr_id = 0ULL; /* no send cookie */ - param.bind_info.addr = seg->mr_dma; - param.bind_info.length = seg->mr_len; - param.send_flags = 0; - param.bind_info.mw_access_flags = mem_priv; - - DECR_CQCOUNT(&r_xprt->rx_ep); - rc = ib_bind_mw(ia->ri_id->qp, seg->mr_chunk.rl_mw->r.mw, ¶m); - if (rc) { - dprintk("RPC: %s: failed ib_bind_mw " - "%u@0x%llx status %i\n", - __func__, seg->mr_len, - (unsigned long long)seg->mr_dma, rc); - rpcrdma_unmap_one(ia, seg); - } else { - seg->mr_rkey = seg->mr_chunk.rl_mw->r.mw->rkey; - seg->mr_base = param.bind_info.addr; - seg->mr_nsegs = 1; - } - return rc; -} - -static int -rpcrdma_deregister_memwin_external(struct rpcrdma_mr_seg *seg, - struct rpcrdma_ia *ia, - struct rpcrdma_xprt *r_xprt, void **r) -{ - struct ib_mw_bind param; - LIST_HEAD(l); - int rc; - - BUG_ON(seg->mr_nsegs != 1); - param.bind_info.mr = ia->ri_bind_mem; - param.bind_info.addr = 0ULL; /* unbind */ - param.bind_info.length = 0; - param.bind_info.mw_access_flags = 0; - if (*r) { - param.wr_id = (u64) (unsigned long) *r; - param.send_flags = IB_SEND_SIGNALED; - INIT_CQCOUNT(&r_xprt->rx_ep); - } else { - param.wr_id = 0ULL; - param.send_flags = 0; - DECR_CQCOUNT(&r_xprt->rx_ep); - } - rc = ib_bind_mw(ia->ri_id->qp, seg->mr_chunk.rl_mw->r.mw, ¶m); - rpcrdma_unmap_one(ia, seg); - if (rc) - dprintk("RPC: %s: failed ib_(un)bind_mw," - " status %i\n", __func__, rc); - else - *r = NULL; /* will upcall on completion */ - return rc; -} - -static int -rpcrdma_register_default_external(struct rpcrdma_mr_seg *seg, - int *nsegs, int writing, struct rpcrdma_ia *ia) -{ - int mem_priv = (writing ? IB_ACCESS_REMOTE_WRITE : - IB_ACCESS_REMOTE_READ); - struct rpcrdma_mr_seg *seg1 = seg; - struct ib_phys_buf ipb[RPCRDMA_MAX_DATA_SEGS]; - int len, i, rc = 0; - - if (*nsegs > RPCRDMA_MAX_DATA_SEGS) - *nsegs = RPCRDMA_MAX_DATA_SEGS; - for (len = 0, i = 0; i < *nsegs;) { - rpcrdma_map_one(ia, seg, writing); - ipb[i].addr = seg->mr_dma; - ipb[i].size = seg->mr_len; - len += seg->mr_len; - ++seg; - ++i; - /* Check for holes */ - if ((i < *nsegs && offset_in_page(seg->mr_offset)) || - offset_in_page((seg-1)->mr_offset+(seg-1)->mr_len)) - break; - } - seg1->mr_base = seg1->mr_dma; - seg1->mr_chunk.rl_mr = ib_reg_phys_mr(ia->ri_pd, - ipb, i, mem_priv, &seg1->mr_base); - if (IS_ERR(seg1->mr_chunk.rl_mr)) { - rc = PTR_ERR(seg1->mr_chunk.rl_mr); - dprintk("RPC: %s: failed ib_reg_phys_mr " - "%u@0x%llx (%d)... status %i\n", - __func__, len, - (unsigned long long)seg1->mr_dma, i, rc); - while (i--) - rpcrdma_unmap_one(ia, --seg); - } else { - seg1->mr_rkey = seg1->mr_chunk.rl_mr->rkey; - seg1->mr_nsegs = i; - seg1->mr_len = len; - } - *nsegs = i; - return rc; -} - -static int -rpcrdma_deregister_default_external(struct rpcrdma_mr_seg *seg, - struct rpcrdma_ia *ia) -{ - struct rpcrdma_mr_seg *seg1 = seg; - int rc; - - rc = ib_dereg_mr(seg1->mr_chunk.rl_mr); - seg1->mr_chunk.rl_mr = NULL; - while (seg1->mr_nsegs--) - rpcrdma_unmap_one(ia, seg++); - if (rc) - dprintk("RPC: %s: failed ib_dereg_mr," - " status %i\n", __func__, rc); - return rc; -} - int rpcrdma_register_external(struct rpcrdma_mr_seg *seg, int nsegs, int writing, struct rpcrdma_xprt *r_xprt) @@ -1819,16 +1693,8 @@ rpcrdma_register_external(struct rpcrdma_mr_seg *seg, rc = rpcrdma_register_fmr_external(seg, &nsegs, writing, ia); break; - /* Registration using memory windows */ - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: - rc = rpcrdma_register_memwin_external(seg, &nsegs, writing, ia, r_xprt); - break; - - /* Default registration each time */ default: - rc = rpcrdma_register_default_external(seg, &nsegs, writing, ia); - break; + return -1; } if (rc) return -1; @@ -1838,7 +1704,7 @@ rpcrdma_register_external(struct rpcrdma_mr_seg *seg, int rpcrdma_deregister_external(struct rpcrdma_mr_seg *seg, - struct rpcrdma_xprt *r_xprt, void *r) + struct rpcrdma_xprt *r_xprt) { struct rpcrdma_ia *ia = &r_xprt->rx_ia; int nsegs = seg->mr_nsegs, rc; @@ -1847,9 +1713,7 @@ rpcrdma_deregister_external(struct rpcrdma_mr_seg *seg, #if RPCRDMA_PERSISTENT_REGISTRATION case RPCRDMA_ALLPHYSICAL: - BUG_ON(nsegs != 1); rpcrdma_unmap_one(ia, seg); - rc = 0; break; #endif @@ -1861,21 +1725,9 @@ rpcrdma_deregister_external(struct rpcrdma_mr_seg *seg, rc = rpcrdma_deregister_fmr_external(seg, ia); break; - case RPCRDMA_MEMWINDOWS_ASYNC: - case RPCRDMA_MEMWINDOWS: - rc = rpcrdma_deregister_memwin_external(seg, ia, r_xprt, &r); - break; - default: - rc = rpcrdma_deregister_default_external(seg, ia); break; } - if (r) { - struct rpcrdma_rep *rep = r; - void (*func)(struct rpcrdma_rep *) = rep->rr_func; - rep->rr_func = NULL; - func(rep); /* dereg done, callback now */ - } return nsegs; } @@ -1950,7 +1802,6 @@ rpcrdma_ep_post_recv(struct rpcrdma_ia *ia, ib_dma_sync_single_for_cpu(ia->ri_id->device, rep->rr_iov.addr, rep->rr_iov.length, DMA_BIDIRECTIONAL); - DECR_CQCOUNT(ep); rc = ib_post_recv(ia->ri_id->qp, &recv_wr, &recv_wr_fail); if (rc) diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h index cc1445dc1d1a..89e7cd479705 100644 --- a/net/sunrpc/xprtrdma/xprt_rdma.h +++ b/net/sunrpc/xprtrdma/xprt_rdma.h @@ -43,6 +43,7 @@ #include <linux/wait.h> /* wait_queue_head_t, etc */ #include <linux/spinlock.h> /* spinlock_t, etc */ #include <linux/atomic.h> /* atomic_t, etc */ +#include <linux/workqueue.h> /* struct work_struct */ #include <rdma/rdma_cm.h> /* RDMA connection api */ #include <rdma/ib_verbs.h> /* RDMA verbs api */ @@ -66,18 +67,21 @@ struct rpcrdma_ia { struct completion ri_done; int ri_async_rc; enum rpcrdma_memreg ri_memreg_strategy; + unsigned int ri_max_frmr_depth; }; /* * RDMA Endpoint -- one per transport instance */ +#define RPCRDMA_WC_BUDGET (128) +#define RPCRDMA_POLLSIZE (16) + struct rpcrdma_ep { atomic_t rep_cqcount; int rep_cqinit; int rep_connected; struct rpcrdma_ia *rep_ia; - struct ib_cq *rep_cq; struct ib_qp_init_attr rep_attr; wait_queue_head_t rep_connect_wait; struct ib_sge rep_pad; /* holds zeroed pad */ @@ -86,6 +90,9 @@ struct rpcrdma_ep { struct rpc_xprt *rep_xprt; /* for rep_func */ struct rdma_conn_param rep_remote_cma; struct sockaddr_storage rep_remote_addr; + struct delayed_work rep_connect_worker; + struct ib_wc rep_send_wcs[RPCRDMA_POLLSIZE]; + struct ib_wc rep_recv_wcs[RPCRDMA_POLLSIZE]; }; #define INIT_CQCOUNT(ep) atomic_set(&(ep)->rep_cqcount, (ep)->rep_cqinit) @@ -124,7 +131,6 @@ struct rpcrdma_rep { struct rpc_xprt *rr_xprt; /* needed for request/reply matching */ void (*rr_func)(struct rpcrdma_rep *);/* called by tasklet in softint */ struct list_head rr_list; /* tasklet list */ - wait_queue_head_t rr_unbind; /* optional unbind wait */ struct ib_sge rr_iov; /* for posting */ struct ib_mr *rr_handle; /* handle for mem in rr_iov */ char rr_base[MAX_RPCRDMAHDR]; /* minimal inline receive buffer */ @@ -159,7 +165,6 @@ struct rpcrdma_mr_seg { /* chunk descriptors */ struct ib_mr *rl_mr; /* if registered directly */ struct rpcrdma_mw { /* if registered from region */ union { - struct ib_mw *mw; struct ib_fmr *fmr; struct { struct ib_fast_reg_page_list *fr_pgl; @@ -207,7 +212,6 @@ struct rpcrdma_req { struct rpcrdma_buffer { spinlock_t rb_lock; /* protects indexes */ atomic_t rb_credits; /* most recent server credits */ - unsigned long rb_cwndscale; /* cached framework rpc_cwndscale */ int rb_max_requests;/* client max requests */ struct list_head rb_mws; /* optional memory windows/fmrs/frmrs */ int rb_send_index; @@ -300,7 +304,7 @@ void rpcrdma_ia_close(struct rpcrdma_ia *); */ int rpcrdma_ep_create(struct rpcrdma_ep *, struct rpcrdma_ia *, struct rpcrdma_create_data_internal *); -int rpcrdma_ep_destroy(struct rpcrdma_ep *, struct rpcrdma_ia *); +void rpcrdma_ep_destroy(struct rpcrdma_ep *, struct rpcrdma_ia *); int rpcrdma_ep_connect(struct rpcrdma_ep *, struct rpcrdma_ia *); int rpcrdma_ep_disconnect(struct rpcrdma_ep *, struct rpcrdma_ia *); @@ -330,11 +334,12 @@ int rpcrdma_deregister_internal(struct rpcrdma_ia *, int rpcrdma_register_external(struct rpcrdma_mr_seg *, int, int, struct rpcrdma_xprt *); int rpcrdma_deregister_external(struct rpcrdma_mr_seg *, - struct rpcrdma_xprt *, void *); + struct rpcrdma_xprt *); /* * RPC/RDMA connection management calls - xprtrdma/rpc_rdma.c */ +void rpcrdma_connect_worker(struct work_struct *); void rpcrdma_conn_func(struct rpcrdma_ep *); void rpcrdma_reply_handler(struct rpcrdma_rep *); diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 25a3dcf15cae..402a7e9a16b7 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -893,11 +893,11 @@ static void xs_close(struct rpc_xprt *xprt) xs_reset_transport(transport); xprt->reestablish_timeout = 0; - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(XPRT_CONNECTION_ABORT, &xprt->state); clear_bit(XPRT_CLOSE_WAIT, &xprt->state); clear_bit(XPRT_CLOSING, &xprt->state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); xprt_disconnect_done(xprt); } @@ -1497,12 +1497,12 @@ static void xs_tcp_cancel_linger_timeout(struct rpc_xprt *xprt) static void xs_sock_reset_connection_flags(struct rpc_xprt *xprt) { - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(XPRT_CONNECTION_ABORT, &xprt->state); clear_bit(XPRT_CONNECTION_CLOSE, &xprt->state); clear_bit(XPRT_CLOSE_WAIT, &xprt->state); clear_bit(XPRT_CLOSING, &xprt->state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); } static void xs_sock_mark_closed(struct rpc_xprt *xprt) @@ -1556,10 +1556,10 @@ static void xs_tcp_state_change(struct sock *sk) xprt->connect_cookie++; xprt->reestablish_timeout = 0; set_bit(XPRT_CLOSING, &xprt->state); - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(XPRT_CONNECTED, &xprt->state); clear_bit(XPRT_CLOSE_WAIT, &xprt->state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); xs_tcp_schedule_linger_timeout(xprt, xs_tcp_fin_timeout); break; case TCP_CLOSE_WAIT: @@ -1578,9 +1578,9 @@ static void xs_tcp_state_change(struct sock *sk) case TCP_LAST_ACK: set_bit(XPRT_CLOSING, &xprt->state); xs_tcp_schedule_linger_timeout(xprt, xs_tcp_fin_timeout); - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(XPRT_CONNECTED, &xprt->state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); break; case TCP_CLOSE: xs_tcp_cancel_linger_timeout(xprt); diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index bb7e8ba821f4..749f80c21e22 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -1207,7 +1207,7 @@ restart: sk->sk_state = TCP_ESTABLISHED; sock_hold(newsk); - smp_mb__after_atomic_inc(); /* sock_hold() does an atomic_inc() */ + smp_mb__after_atomic(); /* sock_hold() does an atomic_inc() */ unix_peer(sk) = newsk; unix_state_unlock(sk); |