From 7a6fdeb2b1e93548854063c46c9724458564c76b Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Mon, 22 Dec 2014 19:14:26 +0300 Subject: libceph: nuke pool op infrastructure On Mon, Dec 22, 2014 at 5:35 PM, Sage Weil wrote: > On Mon, 22 Dec 2014, Ilya Dryomov wrote: >> Actually, pool op stuff has been unused for over two years - looks like >> it was added for rbd create_snap and that got ripped out in 2012. It's >> unlikely we'd ever need to manage pools or snaps from the kernel client >> so I think it makes sense to nuke it. Sage? > > Yep! Signed-off-by: Ilya Dryomov --- net/ceph/ceph_strings.c | 14 ----- net/ceph/debugfs.c | 2 - net/ceph/mon_client.c | 135 ++---------------------------------------------- 3 files changed, 3 insertions(+), 148 deletions(-) (limited to 'net') diff --git a/net/ceph/ceph_strings.c b/net/ceph/ceph_strings.c index 30560202f57b..139a9cb19b0c 100644 --- a/net/ceph/ceph_strings.c +++ b/net/ceph/ceph_strings.c @@ -42,17 +42,3 @@ const char *ceph_osd_state_name(int s) return "???"; } } - -const char *ceph_pool_op_name(int op) -{ - switch (op) { - case POOL_OP_CREATE: return "create"; - case POOL_OP_DELETE: return "delete"; - case POOL_OP_AUID_CHANGE: return "auid change"; - case POOL_OP_CREATE_SNAP: return "create snap"; - case POOL_OP_DELETE_SNAP: return "delete snap"; - case POOL_OP_CREATE_UNMANAGED_SNAP: return "create unmanaged snap"; - case POOL_OP_DELETE_UNMANAGED_SNAP: return "delete unmanaged snap"; - } - return "???"; -} diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c index d2d525529f87..14d9995097cc 100644 --- a/net/ceph/debugfs.c +++ b/net/ceph/debugfs.c @@ -127,8 +127,6 @@ static int monc_show(struct seq_file *s, void *p) op = le16_to_cpu(req->request->hdr.type); if (op == CEPH_MSG_STATFS) seq_printf(s, "%llu statfs\n", req->tid); - else if (op == CEPH_MSG_POOLOP) - seq_printf(s, "%llu poolop\n", req->tid); else if (op == CEPH_MSG_MON_GET_VERSION) seq_printf(s, "%llu mon_get_version", req->tid); else diff --git a/net/ceph/mon_client.c b/net/ceph/mon_client.c index f2148e22b148..02e105c6865f 100644 --- a/net/ceph/mon_client.c +++ b/net/ceph/mon_client.c @@ -410,7 +410,7 @@ out_unlocked: } /* - * generic requests (e.g., statfs, poolop) + * generic requests (currently statfs, mon_get_version) */ static struct ceph_mon_generic_request *__lookup_generic_req( struct ceph_mon_client *monc, u64 tid) @@ -569,7 +569,7 @@ static void handle_statfs_reply(struct ceph_mon_client *monc, return; bad: - pr_err("corrupt generic reply, tid %llu\n", tid); + pr_err("corrupt statfs reply, tid %llu\n", tid); ceph_msg_dump(msg); } @@ -588,7 +588,6 @@ int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf) kref_init(&req->kref); req->buf = buf; - req->buf_len = sizeof(*buf); init_completion(&req->completion); err = -ENOMEM; @@ -647,7 +646,7 @@ static void handle_get_version_reply(struct ceph_mon_client *monc, return; bad: - pr_err("corrupt mon_get_version reply\n"); + pr_err("corrupt mon_get_version reply, tid %llu\n", tid); ceph_msg_dump(msg); } @@ -670,7 +669,6 @@ int ceph_monc_do_get_version(struct ceph_mon_client *monc, const char *what, kref_init(&req->kref); req->buf = newest; - req->buf_len = sizeof(*newest); init_completion(&req->completion); req->request = ceph_msg_new(CEPH_MSG_MON_GET_VERSION, @@ -706,128 +704,6 @@ out: } EXPORT_SYMBOL(ceph_monc_do_get_version); -/* - * pool ops - */ -static int get_poolop_reply_buf(const char *src, size_t src_len, - char *dst, size_t dst_len) -{ - u32 buf_len; - - if (src_len != sizeof(u32) + dst_len) - return -EINVAL; - - buf_len = le32_to_cpu(*(__le32 *)src); - if (buf_len != dst_len) - return -EINVAL; - - memcpy(dst, src + sizeof(u32), dst_len); - return 0; -} - -static void handle_poolop_reply(struct ceph_mon_client *monc, - struct ceph_msg *msg) -{ - struct ceph_mon_generic_request *req; - struct ceph_mon_poolop_reply *reply = msg->front.iov_base; - u64 tid = le64_to_cpu(msg->hdr.tid); - - if (msg->front.iov_len < sizeof(*reply)) - goto bad; - dout("handle_poolop_reply %p tid %llu\n", msg, tid); - - mutex_lock(&monc->mutex); - req = __lookup_generic_req(monc, tid); - if (req) { - if (req->buf_len && - get_poolop_reply_buf(msg->front.iov_base + sizeof(*reply), - msg->front.iov_len - sizeof(*reply), - req->buf, req->buf_len) < 0) { - mutex_unlock(&monc->mutex); - goto bad; - } - req->result = le32_to_cpu(reply->reply_code); - get_generic_request(req); - } - mutex_unlock(&monc->mutex); - if (req) { - complete(&req->completion); - put_generic_request(req); - } - return; - -bad: - pr_err("corrupt generic reply, tid %llu\n", tid); - ceph_msg_dump(msg); -} - -/* - * Do a synchronous pool op. - */ -static int do_poolop(struct ceph_mon_client *monc, u32 op, - u32 pool, u64 snapid, - char *buf, int len) -{ - struct ceph_mon_generic_request *req; - struct ceph_mon_poolop *h; - int err; - - req = kzalloc(sizeof(*req), GFP_NOFS); - if (!req) - return -ENOMEM; - - kref_init(&req->kref); - req->buf = buf; - req->buf_len = len; - init_completion(&req->completion); - - err = -ENOMEM; - req->request = ceph_msg_new(CEPH_MSG_POOLOP, sizeof(*h), GFP_NOFS, - true); - if (!req->request) - goto out; - req->reply = ceph_msg_new(CEPH_MSG_POOLOP_REPLY, 1024, GFP_NOFS, - true); - if (!req->reply) - goto out; - - /* fill out request */ - req->request->hdr.version = cpu_to_le16(2); - h = req->request->front.iov_base; - h->monhdr.have_version = 0; - h->monhdr.session_mon = cpu_to_le16(-1); - h->monhdr.session_mon_tid = 0; - h->fsid = monc->monmap->fsid; - h->pool = cpu_to_le32(pool); - h->op = cpu_to_le32(op); - h->auid = 0; - h->snapid = cpu_to_le64(snapid); - h->name_len = 0; - - err = do_generic_request(monc, req); - -out: - kref_put(&req->kref, release_generic_request); - return err; -} - -int ceph_monc_create_snapid(struct ceph_mon_client *monc, - u32 pool, u64 *snapid) -{ - return do_poolop(monc, POOL_OP_CREATE_UNMANAGED_SNAP, - pool, 0, (char *)snapid, sizeof(*snapid)); - -} -EXPORT_SYMBOL(ceph_monc_create_snapid); - -int ceph_monc_delete_snapid(struct ceph_mon_client *monc, - u32 pool, u64 snapid) -{ - return do_poolop(monc, POOL_OP_CREATE_UNMANAGED_SNAP, - pool, snapid, NULL, 0); - -} - /* * Resend pending generic requests. */ @@ -1112,10 +988,6 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) handle_get_version_reply(monc, msg); break; - case CEPH_MSG_POOLOP_REPLY: - handle_poolop_reply(monc, msg); - break; - case CEPH_MSG_MON_MAP: ceph_monc_handle_map(monc, msg); break; @@ -1154,7 +1026,6 @@ static struct ceph_msg *mon_alloc_msg(struct ceph_connection *con, case CEPH_MSG_MON_SUBSCRIBE_ACK: m = ceph_msg_get(monc->m_subscribe_ack); break; - case CEPH_MSG_POOLOP_REPLY: case CEPH_MSG_STATFS_REPLY: return get_generic_reply(con, hdr, skip); case CEPH_MSG_AUTH_REPLY: -- cgit v1.2.3 From f646912d1044ace6c6c5b5785b2579177781873b Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Mon, 22 Dec 2014 19:35:17 +0300 Subject: libceph: use mon_client.c/put_generic_request() more Signed-off-by: Ilya Dryomov --- net/ceph/mon_client.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/ceph/mon_client.c b/net/ceph/mon_client.c index 02e105c6865f..2b3cf05e87b0 100644 --- a/net/ceph/mon_client.c +++ b/net/ceph/mon_client.c @@ -610,7 +610,7 @@ int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf) err = do_generic_request(monc, req); out: - kref_put(&req->kref, release_generic_request); + put_generic_request(req); return err; } EXPORT_SYMBOL(ceph_monc_do_statfs); @@ -699,7 +699,7 @@ int ceph_monc_do_get_version(struct ceph_mon_client *monc, const char *what, mutex_unlock(&monc->mutex); out: - kref_put(&req->kref, release_generic_request); + put_generic_request(req); return err; } EXPORT_SYMBOL(ceph_monc_do_get_version); -- cgit v1.2.3 From ba988f87f532cd2b8c4740aa8ec49056521ae833 Mon Sep 17 00:00:00 2001 From: Chaitanya Huilgol Date: Fri, 23 Jan 2015 16:41:25 +0530 Subject: libceph: tcp_nodelay support MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit TCP_NODELAY socket option set on connection sockets, disables Nagle’s algorithm and improves latency characteristics. tcp_nodelay(default)/notcp_nodelay option flags provided to enable/disable setting the socket option. Signed-off-by: Chaitanya Huilgol [idryomov@redhat.com: NO_TCP_NODELAY -> TCP_NODELAY, minor adjustments] Signed-off-by: Ilya Dryomov --- include/linux/ceph/libceph.h | 3 ++- include/linux/ceph/messenger.h | 4 +++- net/ceph/ceph_common.c | 16 +++++++++++++++- net/ceph/messenger.c | 14 +++++++++++++- 4 files changed, 33 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h index 8b11a79ca1cb..16fff9608848 100644 --- a/include/linux/ceph/libceph.h +++ b/include/linux/ceph/libceph.h @@ -30,8 +30,9 @@ #define CEPH_OPT_MYIP (1<<2) /* specified my ip */ #define CEPH_OPT_NOCRC (1<<3) /* no data crc on writes */ #define CEPH_OPT_NOMSGAUTH (1<<4) /* not require cephx message signature */ +#define CEPH_OPT_TCP_NODELAY (1<<5) /* TCP_NODELAY on TCP sockets */ -#define CEPH_OPT_DEFAULT (0) +#define CEPH_OPT_DEFAULT (CEPH_OPT_TCP_NODELAY) #define ceph_set_opt(client, opt) \ (client)->options->flags |= CEPH_OPT_##opt; diff --git a/include/linux/ceph/messenger.h b/include/linux/ceph/messenger.h index d9d396c16503..e15499422fdc 100644 --- a/include/linux/ceph/messenger.h +++ b/include/linux/ceph/messenger.h @@ -57,6 +57,7 @@ struct ceph_messenger { atomic_t stopping; bool nocrc; + bool tcp_nodelay; /* * the global_seq counts connections i (attempt to) initiate @@ -264,7 +265,8 @@ extern void ceph_messenger_init(struct ceph_messenger *msgr, struct ceph_entity_addr *myaddr, u64 supported_features, u64 required_features, - bool nocrc); + bool nocrc, + bool tcp_nodelay); extern void ceph_con_init(struct ceph_connection *con, void *private, const struct ceph_connection_operations *ops, diff --git a/net/ceph/ceph_common.c b/net/ceph/ceph_common.c index 5d5ab67f516d..ec565508e904 100644 --- a/net/ceph/ceph_common.c +++ b/net/ceph/ceph_common.c @@ -239,6 +239,8 @@ enum { Opt_nocrc, Opt_cephx_require_signatures, Opt_nocephx_require_signatures, + Opt_tcp_nodelay, + Opt_notcp_nodelay, }; static match_table_t opt_tokens = { @@ -259,6 +261,8 @@ static match_table_t opt_tokens = { {Opt_nocrc, "nocrc"}, {Opt_cephx_require_signatures, "cephx_require_signatures"}, {Opt_nocephx_require_signatures, "nocephx_require_signatures"}, + {Opt_tcp_nodelay, "tcp_nodelay"}, + {Opt_notcp_nodelay, "notcp_nodelay"}, {-1, NULL} }; @@ -457,6 +461,7 @@ ceph_parse_options(char *options, const char *dev_name, case Opt_nocrc: opt->flags |= CEPH_OPT_NOCRC; break; + case Opt_cephx_require_signatures: opt->flags &= ~CEPH_OPT_NOMSGAUTH; break; @@ -464,6 +469,13 @@ ceph_parse_options(char *options, const char *dev_name, opt->flags |= CEPH_OPT_NOMSGAUTH; break; + case Opt_tcp_nodelay: + opt->flags |= CEPH_OPT_TCP_NODELAY; + break; + case Opt_notcp_nodelay: + opt->flags &= ~CEPH_OPT_TCP_NODELAY; + break; + default: BUG_ON(token); } @@ -518,10 +530,12 @@ struct ceph_client *ceph_create_client(struct ceph_options *opt, void *private, /* msgr */ if (ceph_test_opt(client, MYIP)) myaddr = &client->options->my_addr; + ceph_messenger_init(&client->msgr, myaddr, client->supported_features, client->required_features, - ceph_test_opt(client, NOCRC)); + ceph_test_opt(client, NOCRC), + ceph_test_opt(client, TCP_NODELAY)); /* subsystems */ err = ceph_monc_init(&client->monc, client); diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c index 33a2f201e460..6b3f54ed65ba 100644 --- a/net/ceph/messenger.c +++ b/net/ceph/messenger.c @@ -510,6 +510,16 @@ static int ceph_tcp_connect(struct ceph_connection *con) return ret; } + if (con->msgr->tcp_nodelay) { + int optval = 1; + + ret = kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY, + (char *)&optval, sizeof(optval)); + if (ret) + pr_err("kernel_setsockopt(TCP_NODELAY) failed: %d", + ret); + } + sk_set_memalloc(sock->sk); con->sock = sock; @@ -2922,7 +2932,8 @@ void ceph_messenger_init(struct ceph_messenger *msgr, struct ceph_entity_addr *myaddr, u64 supported_features, u64 required_features, - bool nocrc) + bool nocrc, + bool tcp_nodelay) { msgr->supported_features = supported_features; msgr->required_features = required_features; @@ -2937,6 +2948,7 @@ void ceph_messenger_init(struct ceph_messenger *msgr, get_random_bytes(&msgr->inst.addr.nonce, sizeof(msgr->inst.addr.nonce)); encode_my_addr(msgr); msgr->nocrc = nocrc; + msgr->tcp_nodelay = tcp_nodelay; atomic_set(&msgr->stopping, 0); -- cgit v1.2.3 From 7eb71e0351fbb1b242ae70abb7bb17107fe2f792 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Tue, 17 Feb 2015 19:37:15 +0300 Subject: libceph: fix double __remove_osd() problem It turns out it's possible to get __remove_osd() called twice on the same OSD. That doesn't sit well with rb_erase() - depending on the shape of the tree we can get a NULL dereference, a soft lockup or a random crash at some point in the future as we end up touching freed memory. One scenario that I was able to reproduce is as follows: con_fault_finish() osd_reset() ceph_osdc_handle_map() kick_requests() reset_changed_osds() __reset_osd() __remove_osd() __kick_osd_requests() __reset_osd() __remove_osd() <-- !!! A case can be made that osd refcounting is imperfect and reworking it would be a proper resolution, but for now Sage and I decided to fix this by adding a safe guard around __remove_osd(). Fixes: http://tracker.ceph.com/issues/8087 Cc: Sage Weil Cc: stable@vger.kernel.org # 3.9+: 7c6e6fc53e73: libceph: assert both regular and lingering lists in __remove_osd() Cc: stable@vger.kernel.org # 3.9+: cc9f1f518cec: libceph: change from BUG to WARN for __remove_osd() asserts Cc: stable@vger.kernel.org # 3.9+ Signed-off-by: Ilya Dryomov Reviewed-by: Sage Weil Reviewed-by: Alex Elder --- net/ceph/osd_client.c | 26 ++++++++++++++++++-------- 1 file changed, 18 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 53299c7b0ca4..f693a2f8ac86 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -1048,14 +1048,24 @@ static void put_osd(struct ceph_osd *osd) */ static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) { - dout("__remove_osd %p\n", osd); + dout("%s %p osd%d\n", __func__, osd, osd->o_osd); WARN_ON(!list_empty(&osd->o_requests)); WARN_ON(!list_empty(&osd->o_linger_requests)); - rb_erase(&osd->o_node, &osdc->osds); list_del_init(&osd->o_osd_lru); - ceph_con_close(&osd->o_con); - put_osd(osd); + rb_erase(&osd->o_node, &osdc->osds); + RB_CLEAR_NODE(&osd->o_node); +} + +static void remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) +{ + dout("%s %p osd%d\n", __func__, osd, osd->o_osd); + + if (!RB_EMPTY_NODE(&osd->o_node)) { + ceph_con_close(&osd->o_con); + __remove_osd(osdc, osd); + put_osd(osd); + } } static void remove_all_osds(struct ceph_osd_client *osdc) @@ -1065,7 +1075,7 @@ static void remove_all_osds(struct ceph_osd_client *osdc) while (!RB_EMPTY_ROOT(&osdc->osds)) { struct ceph_osd *osd = rb_entry(rb_first(&osdc->osds), struct ceph_osd, o_node); - __remove_osd(osdc, osd); + remove_osd(osdc, osd); } mutex_unlock(&osdc->request_mutex); } @@ -1106,7 +1116,7 @@ static void remove_old_osds(struct ceph_osd_client *osdc) list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) { if (time_before(jiffies, osd->lru_ttl)) break; - __remove_osd(osdc, osd); + remove_osd(osdc, osd); } mutex_unlock(&osdc->request_mutex); } @@ -1121,8 +1131,7 @@ static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) dout("__reset_osd %p osd%d\n", osd, osd->o_osd); if (list_empty(&osd->o_requests) && list_empty(&osd->o_linger_requests)) { - __remove_osd(osdc, osd); - + remove_osd(osdc, osd); return -ENODEV; } @@ -1926,6 +1935,7 @@ static void reset_changed_osds(struct ceph_osd_client *osdc) { struct rb_node *p, *n; + dout("%s %p\n", __func__, osdc); for (p = rb_first(&osdc->osds); p; p = n) { struct ceph_osd *osd = rb_entry(p, struct ceph_osd, o_node); -- cgit v1.2.3 From b28ec2f37e6a2bbd0bdf74b39cb89c74e4ad17f3 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Mon, 16 Feb 2015 11:49:42 +0300 Subject: libceph: kfree() in put_osd() shouldn't depend on authorizer a255651d4cad ("ceph: ensure auth ops are defined before use") made kfree() in put_osd() conditional on the authorizer. A mechanical mistake most likely - fix it. Cc: Alex Elder Signed-off-by: Ilya Dryomov Reviewed-by: Sage Weil Reviewed-by: Alex Elder --- net/ceph/osd_client.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index f693a2f8ac86..41a4abc7e98e 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -1035,10 +1035,11 @@ static void put_osd(struct ceph_osd *osd) { dout("put_osd %p %d -> %d\n", osd, atomic_read(&osd->o_ref), atomic_read(&osd->o_ref) - 1); - if (atomic_dec_and_test(&osd->o_ref) && osd->o_auth.authorizer) { + if (atomic_dec_and_test(&osd->o_ref)) { struct ceph_auth_client *ac = osd->o_osdc->client->monc.auth; - ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer); + if (osd->o_auth.authorizer) + ceph_auth_destroy_authorizer(ac, osd->o_auth.authorizer); kfree(osd); } } -- cgit v1.2.3