diff options
author | Paul Mundt <lethal@linux-sh.org> | 2011-03-24 15:17:25 +0900 |
---|---|---|
committer | Paul Mundt <lethal@linux-sh.org> | 2011-03-24 15:17:25 +0900 |
commit | a3d3362287fbe96fe90abdb5c6d1a35471129a8c (patch) | |
tree | ad3c85ed1feef470c66599eb514e30f43c2db5dd /net | |
parent | fb7f045ace0624f1e59a7db8497e460bd54b1cbc (diff) | |
parent | 4bbba111d94781d34081c37856bbc5eb33f6c72a (diff) |
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6 into sh-latest
Diffstat (limited to 'net')
33 files changed, 780 insertions, 321 deletions
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index ae610f046de..e34ea9e5e28 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -720,6 +720,7 @@ static int vlan_dev_init(struct net_device *dev) dev->fcoe_ddp_xid = real_dev->fcoe_ddp_xid; #endif + dev->needed_headroom = real_dev->needed_headroom; if (real_dev->features & NETIF_F_HW_VLAN_TX) { dev->header_ops = real_dev->header_ops; dev->hard_header_len = real_dev->hard_header_len; diff --git a/net/9p/client.c b/net/9p/client.c index 347ec0cd271..2ccbf04d37d 100644 --- a/net/9p/client.c +++ b/net/9p/client.c @@ -223,7 +223,7 @@ static struct p9_req_t *p9_tag_alloc(struct p9_client *c, u16 tag) req = &c->reqs[row][col]; if (!req->tc) { - req->wq = kmalloc(sizeof(wait_queue_head_t), GFP_KERNEL); + req->wq = kmalloc(sizeof(wait_queue_head_t), GFP_NOFS); if (!req->wq) { printk(KERN_ERR "Couldn't grow tag array\n"); return ERR_PTR(-ENOMEM); @@ -233,17 +233,17 @@ static struct p9_req_t *p9_tag_alloc(struct p9_client *c, u16 tag) P9_TRANS_PREF_PAYLOAD_SEP) { int alloc_msize = min(c->msize, 4096); req->tc = kmalloc(sizeof(struct p9_fcall)+alloc_msize, - GFP_KERNEL); + GFP_NOFS); req->tc->capacity = alloc_msize; req->rc = kmalloc(sizeof(struct p9_fcall)+alloc_msize, - GFP_KERNEL); + GFP_NOFS); req->rc->capacity = alloc_msize; } else { req->tc = kmalloc(sizeof(struct p9_fcall)+c->msize, - GFP_KERNEL); + GFP_NOFS); req->tc->capacity = c->msize; req->rc = kmalloc(sizeof(struct p9_fcall)+c->msize, - GFP_KERNEL); + GFP_NOFS); req->rc->capacity = c->msize; } if ((!req->tc) || (!req->rc)) { diff --git a/net/9p/protocol.c b/net/9p/protocol.c index 2ce515b859b..8a4084fa8b5 100644 --- a/net/9p/protocol.c +++ b/net/9p/protocol.c @@ -205,7 +205,7 @@ p9pdu_vreadf(struct p9_fcall *pdu, int proto_version, const char *fmt, if (errcode) break; - *sptr = kmalloc(len + 1, GFP_KERNEL); + *sptr = kmalloc(len + 1, GFP_NOFS); if (*sptr == NULL) { errcode = -EFAULT; break; @@ -273,7 +273,7 @@ p9pdu_vreadf(struct p9_fcall *pdu, int proto_version, const char *fmt, if (!errcode) { *wnames = kmalloc(sizeof(char *) * *nwname, - GFP_KERNEL); + GFP_NOFS); if (!*wnames) errcode = -ENOMEM; } @@ -317,7 +317,7 @@ p9pdu_vreadf(struct p9_fcall *pdu, int proto_version, const char *fmt, *wqids = kmalloc(*nwqid * sizeof(struct p9_qid), - GFP_KERNEL); + GFP_NOFS); if (*wqids == NULL) errcode = -ENOMEM; } diff --git a/net/9p/trans_common.c b/net/9p/trans_common.c index d62b9aa58df..9172ab78fcb 100644 --- a/net/9p/trans_common.c +++ b/net/9p/trans_common.c @@ -41,9 +41,9 @@ EXPORT_SYMBOL(p9_release_req_pages); int p9_nr_pages(struct p9_req_t *req) { - int start_page, end_page; - start_page = (unsigned long long)req->tc->pubuf >> PAGE_SHIFT; - end_page = ((unsigned long long)req->tc->pubuf + req->tc->pbuf_size + + unsigned long start_page, end_page; + start_page = (unsigned long)req->tc->pubuf >> PAGE_SHIFT; + end_page = ((unsigned long)req->tc->pubuf + req->tc->pbuf_size + PAGE_SIZE - 1) >> PAGE_SHIFT; return end_page - start_page; } @@ -69,8 +69,8 @@ p9_payload_gup(struct p9_req_t *req, size_t *pdata_off, int *pdata_len, *pdata_off = (size_t)req->tc->pubuf & (PAGE_SIZE-1); if (*pdata_off) - first_page_bytes = min((PAGE_SIZE - *pdata_off), - req->tc->pbuf_size); + first_page_bytes = min(((size_t)PAGE_SIZE - *pdata_off), + req->tc->pbuf_size); rpinfo = req->tc->private; pdata_mapped_pages = get_user_pages_fast((unsigned long)req->tc->pubuf, diff --git a/net/9p/trans_fd.c b/net/9p/trans_fd.c index a30471e5174..aa5672b15ea 100644 --- a/net/9p/trans_fd.c +++ b/net/9p/trans_fd.c @@ -350,7 +350,7 @@ static void p9_read_work(struct work_struct *work) if (m->req->rc == NULL) { m->req->rc = kmalloc(sizeof(struct p9_fcall) + - m->client->msize, GFP_KERNEL); + m->client->msize, GFP_NOFS); if (!m->req->rc) { m->req = NULL; err = -ENOMEM; diff --git a/net/9p/trans_rdma.c b/net/9p/trans_rdma.c index 29a54ccd213..150e0c4bbf4 100644 --- a/net/9p/trans_rdma.c +++ b/net/9p/trans_rdma.c @@ -424,7 +424,7 @@ static int rdma_request(struct p9_client *client, struct p9_req_t *req) struct p9_rdma_context *rpl_context = NULL; /* Allocate an fcall for the reply */ - rpl_context = kmalloc(sizeof *rpl_context, GFP_KERNEL); + rpl_context = kmalloc(sizeof *rpl_context, GFP_NOFS); if (!rpl_context) { err = -ENOMEM; goto err_close; @@ -437,7 +437,7 @@ static int rdma_request(struct p9_client *client, struct p9_req_t *req) */ if (!req->rc) { req->rc = kmalloc(sizeof(struct p9_fcall)+client->msize, - GFP_KERNEL); + GFP_NOFS); if (req->rc) { req->rc->sdata = (char *) req->rc + sizeof(struct p9_fcall); @@ -468,7 +468,7 @@ static int rdma_request(struct p9_client *client, struct p9_req_t *req) req->rc = NULL; /* Post the request */ - c = kmalloc(sizeof *c, GFP_KERNEL); + c = kmalloc(sizeof *c, GFP_NOFS); if (!c) { err = -ENOMEM; goto err_free1; diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c index 9b550ed9c71..e8f046b0718 100644 --- a/net/9p/trans_virtio.c +++ b/net/9p/trans_virtio.c @@ -43,6 +43,7 @@ #include <net/9p/client.h> #include <net/9p/transport.h> #include <linux/scatterlist.h> +#include <linux/swap.h> #include <linux/virtio.h> #include <linux/virtio_9p.h> #include "trans_common.h" @@ -51,6 +52,8 @@ /* a single mutex to manage channel initialization and attachment */ static DEFINE_MUTEX(virtio_9p_lock); +static DECLARE_WAIT_QUEUE_HEAD(vp_wq); +static atomic_t vp_pinned = ATOMIC_INIT(0); /** * struct virtio_chan - per-instance transport information @@ -78,7 +81,10 @@ struct virtio_chan { struct virtqueue *vq; int ring_bufs_avail; wait_queue_head_t *vc_wq; - + /* This is global limit. Since we don't have a global structure, + * will be placing it in each channel. + */ + int p9_max_pages; /* Scatterlist: can be too big for stack. */ struct scatterlist sg[VIRTQUEUE_NUM]; @@ -141,34 +147,36 @@ static void req_done(struct virtqueue *vq) P9_DPRINTK(P9_DEBUG_TRANS, ": request done\n"); - do { + while (1) { spin_lock_irqsave(&chan->lock, flags); rc = virtqueue_get_buf(chan->vq, &len); - if (rc != NULL) { - if (!chan->ring_bufs_avail) { - chan->ring_bufs_avail = 1; - wake_up(chan->vc_wq); - } - spin_unlock_irqrestore(&chan->lock, flags); - P9_DPRINTK(P9_DEBUG_TRANS, ": rc %p\n", rc); - P9_DPRINTK(P9_DEBUG_TRANS, ": lookup tag %d\n", - rc->tag); - req = p9_tag_lookup(chan->client, rc->tag); - req->status = REQ_STATUS_RCVD; - if (req->tc->private) { - struct trans_rpage_info *rp = req->tc->private; - /*Release pages */ - p9_release_req_pages(rp); - if (rp->rp_alloc) - kfree(rp); - req->tc->private = NULL; - } - p9_client_cb(chan->client, req); - } else { + if (rc == NULL) { spin_unlock_irqrestore(&chan->lock, flags); + break; + } + + chan->ring_bufs_avail = 1; + spin_unlock_irqrestore(&chan->lock, flags); + /* Wakeup if anyone waiting for VirtIO ring space. */ + wake_up(chan->vc_wq); + P9_DPRINTK(P9_DEBUG_TRANS, ": rc %p\n", rc); + P9_DPRINTK(P9_DEBUG_TRANS, ": lookup tag %d\n", rc->tag); + req = p9_tag_lookup(chan->client, rc->tag); + if (req->tc->private) { + struct trans_rpage_info *rp = req->tc->private; + int p = rp->rp_nr_pages; + /*Release pages */ + p9_release_req_pages(rp); + atomic_sub(p, &vp_pinned); + wake_up(&vp_wq); + if (rp->rp_alloc) + kfree(rp); + req->tc->private = NULL; } - } while (rc != NULL); + req->status = REQ_STATUS_RCVD; + p9_client_cb(chan->client, req); + } } /** @@ -263,7 +271,6 @@ p9_virtio_request(struct p9_client *client, struct p9_req_t *req) P9_DPRINTK(P9_DEBUG_TRANS, "9p debug: virtio request\n"); -req_retry: req->status = REQ_STATUS_SENT; if (req->tc->pbuf_size && (req->tc->pubuf && P9_IS_USER_CONTEXT)) { @@ -271,6 +278,14 @@ req_retry: int rpinfo_size = sizeof(struct trans_rpage_info) + sizeof(struct page *) * nr_pages; + if (atomic_read(&vp_pinned) >= chan->p9_max_pages) { + err = wait_event_interruptible(vp_wq, + atomic_read(&vp_pinned) < chan->p9_max_pages); + if (err == -ERESTARTSYS) + return err; + P9_DPRINTK(P9_DEBUG_TRANS, "9p: May gup pages now.\n"); + } + if (rpinfo_size <= (req->tc->capacity - req->tc->size)) { /* We can use sdata */ req->tc->private = req->tc->sdata + req->tc->size; @@ -293,9 +308,12 @@ req_retry: if (rpinfo->rp_alloc) kfree(rpinfo); return err; + } else { + atomic_add(rpinfo->rp_nr_pages, &vp_pinned); } } +req_retry_pinned: spin_lock_irqsave(&chan->lock, flags); /* Handle out VirtIO ring buffers */ @@ -356,7 +374,7 @@ req_retry: return err; P9_DPRINTK(P9_DEBUG_TRANS, "9p:Retry virtio request\n"); - goto req_retry; + goto req_retry_pinned; } else { spin_unlock_irqrestore(&chan->lock, flags); P9_DPRINTK(P9_DEBUG_TRANS, @@ -453,6 +471,8 @@ static int p9_virtio_probe(struct virtio_device *vdev) } init_waitqueue_head(chan->vc_wq); chan->ring_bufs_avail = 1; + /* Ceiling limit to avoid denial of service attacks */ + chan->p9_max_pages = nr_free_buffer_pages()/4; mutex_lock(&virtio_9p_lock); list_add_tail(&chan->chan_list, &virtio_chan_list); diff --git a/net/9p/util.c b/net/9p/util.c index e048701a72d..b84619b5ba2 100644 --- a/net/9p/util.c +++ b/net/9p/util.c @@ -92,7 +92,7 @@ int p9_idpool_get(struct p9_idpool *p) unsigned long flags; retry: - if (idr_pre_get(&p->pool, GFP_KERNEL) == 0) + if (idr_pre_get(&p->pool, GFP_NOFS) == 0) return 0; spin_lock_irqsave(&p->lock, flags); diff --git a/net/appletalk/ddp.c b/net/appletalk/ddp.c index 3d4f4b04340..206e771e82d 100644 --- a/net/appletalk/ddp.c +++ b/net/appletalk/ddp.c @@ -1051,6 +1051,7 @@ static int atalk_release(struct socket *sock) { struct sock *sk = sock->sk; + sock_hold(sk); lock_sock(sk); if (sk) { sock_orphan(sk); @@ -1058,6 +1059,8 @@ static int atalk_release(struct socket *sock) atalk_destroy_socket(sk); } release_sock(sk); + sock_put(sk); + return 0; } diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c index f97af5590ba..008ff6c4eec 100644 --- a/net/bridge/br_netfilter.c +++ b/net/bridge/br_netfilter.c @@ -739,6 +739,9 @@ static unsigned int br_nf_forward_ip(unsigned int hook, struct sk_buff *skb, nf_bridge->mask |= BRNF_PKT_TYPE; } + if (br_parse_ip_options(skb)) + return NF_DROP; + /* The physdev module checks on this */ nf_bridge->mask |= BRNF_BRIDGED; nf_bridge->physoutdev = skb->dev; diff --git a/net/ceph/armor.c b/net/ceph/armor.c index eb2a666b0be..1fc1ee11dfa 100644 --- a/net/ceph/armor.c +++ b/net/ceph/armor.c @@ -78,8 +78,10 @@ int ceph_unarmor(char *dst, const char *src, const char *end) while (src < end) { int a, b, c, d; - if (src < end && src[0] == '\n') + if (src[0] == '\n') { src++; + continue; + } if (src + 4 > end) return -EINVAL; a = decode_bits(src[0]); diff --git a/net/ceph/ceph_common.c b/net/ceph/ceph_common.c index f3e4a13fea0..95f96ab94bb 100644 --- a/net/ceph/ceph_common.c +++ b/net/ceph/ceph_common.c @@ -62,6 +62,7 @@ const char *ceph_msg_type_name(int type) case CEPH_MSG_OSD_MAP: return "osd_map"; case CEPH_MSG_OSD_OP: return "osd_op"; case CEPH_MSG_OSD_OPREPLY: return "osd_opreply"; + case CEPH_MSG_WATCH_NOTIFY: return "watch_notify"; default: return "unknown"; } } diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 3e20a122ffa..02212ed5085 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -22,10 +22,15 @@ #define OSD_OPREPLY_FRONT_LEN 512 static const struct ceph_connection_operations osd_con_ops; -static int __kick_requests(struct ceph_osd_client *osdc, - struct ceph_osd *kickosd); -static void kick_requests(struct ceph_osd_client *osdc, struct ceph_osd *osd); +static void send_queued(struct ceph_osd_client *osdc); +static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd); +static void __register_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req); +static void __unregister_linger_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req); +static int __send_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req); static int op_needs_trail(int op) { @@ -34,6 +39,7 @@ static int op_needs_trail(int op) case CEPH_OSD_OP_SETXATTR: case CEPH_OSD_OP_CMPXATTR: case CEPH_OSD_OP_CALL: + case CEPH_OSD_OP_NOTIFY: return 1; default: return 0; @@ -209,6 +215,8 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, init_completion(&req->r_completion); init_completion(&req->r_safe_completion); INIT_LIST_HEAD(&req->r_unsafe_item); + INIT_LIST_HEAD(&req->r_linger_item); + INIT_LIST_HEAD(&req->r_linger_osd); req->r_flags = flags; WARN_ON((flags & (CEPH_OSD_FLAG_READ|CEPH_OSD_FLAG_WRITE)) == 0); @@ -315,6 +323,24 @@ static void osd_req_encode_op(struct ceph_osd_request *req, break; case CEPH_OSD_OP_STARTSYNC: break; + case CEPH_OSD_OP_NOTIFY: + { + __le32 prot_ver = cpu_to_le32(src->watch.prot_ver); + __le32 timeout = cpu_to_le32(src->watch.timeout); + + BUG_ON(!req->r_trail); + + ceph_pagelist_append(req->r_trail, + &prot_ver, sizeof(prot_ver)); + ceph_pagelist_append(req->r_trail, + &timeout, sizeof(timeout)); + } + case CEPH_OSD_OP_NOTIFY_ACK: + case CEPH_OSD_OP_WATCH: + dst->watch.cookie = cpu_to_le64(src->watch.cookie); + dst->watch.ver = cpu_to_le64(src->watch.ver); + dst->watch.flag = src->watch.flag; + break; default: pr_err("unrecognized osd opcode %d\n", dst->op); WARN_ON(1); @@ -529,6 +555,45 @@ __lookup_request_ge(struct ceph_osd_client *osdc, return NULL; } +/* + * Resubmit requests pending on the given osd. + */ +static void __kick_osd_requests(struct ceph_osd_client *osdc, + struct ceph_osd *osd) +{ + struct ceph_osd_request *req, *nreq; + int err; + + dout("__kick_osd_requests osd%d\n", osd->o_osd); + err = __reset_osd(osdc, osd); + if (err == -EAGAIN) + return; + + list_for_each_entry(req, &osd->o_requests, r_osd_item) { + list_move(&req->r_req_lru_item, &osdc->req_unsent); + dout("requeued %p tid %llu osd%d\n", req, req->r_tid, + osd->o_osd); + if (!req->r_linger) + req->r_flags |= CEPH_OSD_FLAG_RETRY; + } + + list_for_each_entry_safe(req, nreq, &osd->o_linger_requests, + r_linger_osd) { + __unregister_linger_request(osdc, req); + __register_request(osdc, req); + list_move(&req->r_req_lru_item, &osdc->req_unsent); + dout("requeued lingering %p tid %llu osd%d\n", req, req->r_tid, + osd->o_osd); + } +} + +static void kick_osd_requests(struct ceph_osd_client *osdc, + struct ceph_osd *kickosd) +{ + mutex_lock(&osdc->request_mutex); + __kick_osd_requests(osdc, kickosd); + mutex_unlock(&osdc->request_mutex); +} /* * If the osd connection drops, we need to resubmit all requests. @@ -543,7 +608,8 @@ static void osd_reset(struct ceph_connection *con) dout("osd_reset osd%d\n", osd->o_osd); osdc = osd->o_osdc; down_read(&osdc->map_sem); - kick_requests(osdc, osd); + kick_osd_requests(osdc, osd); + send_queued(osdc); up_read(&osdc->map_sem); } @@ -561,6 +627,7 @@ static struct ceph_osd *create_osd(struct ceph_osd_client *osdc) atomic_set(&osd->o_ref, 1); osd->o_osdc = osdc; INIT_LIST_HEAD(&osd->o_requests); + INIT_LIST_HEAD(&osd->o_linger_requests); INIT_LIST_HEAD(&osd->o_osd_lru); osd->o_incarnation = 1; @@ -650,7 +717,8 @@ static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) int ret = 0; dout("__reset_osd %p osd%d\n", osd, osd->o_osd); - if (list_empty(&osd->o_requests)) { + if (list_empty(&osd->o_requests) && + list_empty(&osd->o_linger_requests)) { __remove_osd(osdc, osd); } else if (memcmp(&osdc->osdmap->osd_addr[osd->o_osd], &osd->o_con.peer_addr, @@ -723,10 +791,9 @@ static void __cancel_osd_timeout(struct ceph_osd_client *osdc) * Register request, assign tid. If this is the first request, set up * the timeout event. */ -static void register_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) +static void __register_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) { - mutex_lock(&osdc->request_mutex); req->r_tid = ++osdc->last_tid; req->r_request->hdr.tid = cpu_to_le64(req->r_tid); INIT_LIST_HEAD(&req->r_req_lru_item); @@ -740,6 +807,13 @@ static void register_request(struct ceph_osd_client *osdc, dout(" first request, scheduling timeout\n"); __schedule_osd_timeout(osdc); } +} + +static void register_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + mutex_lock(&osdc->request_mutex); + __register_request(osdc, req); mutex_unlock(&osdc->request_mutex); } @@ -758,9 +832,14 @@ static void __unregister_request(struct ceph_osd_client *osdc, ceph_con_revoke(&req->r_osd->o_con, req->r_request); list_del_init(&req->r_osd_item); - if (list_empty(&req->r_osd->o_requests)) + if (list_empty(&req->r_osd->o_requests) && + list_empty(&req->r_osd->o_linger_requests)) { + dout("moving osd to %p lru\n", req->r_osd); __move_osd_to_lru(osdc, req->r_osd); - req->r_osd = NULL; + } + if (list_empty(&req->r_osd_item) && + list_empty(&req->r_linger_item)) + req->r_osd = NULL; } ceph_osdc_put_request(req); @@ -781,20 +860,72 @@ static void __cancel_request(struct ceph_osd_request *req) ceph_con_revoke(&req->r_osd->o_con, req->r_request); req->r_sent = 0; } - list_del_init(&req->r_req_lru_item); } +static void __register_linger_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + dout("__register_linger_request %p\n", req); + list_add_tail(&req->r_linger_item, &osdc->req_linger); + list_add_tail(&req->r_linger_osd, &req->r_osd->o_linger_requests); +} + +static void __unregister_linger_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + dout("__unregister_linger_request %p\n", req); + if (req->r_osd) { + list_del_init(&req->r_linger_item); + list_del_init(&req->r_linger_osd); + + if (list_empty(&req->r_osd->o_requests) && + list_empty(&req->r_osd->o_linger_requests)) { + dout("moving osd to %p lru\n", req->r_osd); + __move_osd_to_lru(osdc, req->r_osd); + } + req->r_osd = NULL; + } +} + +void ceph_osdc_unregister_linger_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + mutex_lock(&osdc->request_mutex); + if (req->r_linger) { + __unregister_linger_request(osdc, req); + ceph_osdc_put_request(req); + } + mutex_unlock(&osdc->request_mutex); +} +EXPORT_SYMBOL(ceph_osdc_unregister_linger_request); + +void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) +{ + if (!req->r_linger) { + dout("set_request_linger %p\n", req); + req->r_linger = 1; + /* + * caller is now responsible for calling + * unregister_linger_request + */ + ceph_osdc_get_request(req); + } +} +EXPORT_SYMBOL(ceph_osdc_set_request_linger); + /* * Pick an osd (the first 'up' osd in the pg), allocate the osd struct * (as needed), and set the request r_osd appropriately. If there is - * no up osd, set r_osd to NULL. + * no up osd, set r_osd to NULL. Move the request to the appropiate list + * (unsent, homeless) or leave on in-flight lru. * * Return 0 if unchanged, 1 if changed, or negative on error. * * Caller should hold map_sem for read and request_mutex. */ -static int __map_osds(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) +static int __map_request(struct ceph_osd_client *osdc, + struct ceph_osd_request *req) { struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base; struct ceph_pg pgid; @@ -802,11 +933,13 @@ static int __map_osds(struct ceph_osd_client *osdc, int o = -1, num = 0; int err; - dout("map_osds %p tid %lld\n", req, req->r_tid); + dout("map_request %p tid %lld\n", req, req->r_tid); err = ceph_calc_object_layout(&reqhead->layout, req->r_oid, &req->r_file_layout, osdc->osdmap); - if (err) + if (err) { + list_move(&req->r_req_lru_item, &osdc->req_notarget); return err; + } pgid = reqhead->layout.ol_pgid; req->r_pgid = pgid; @@ -823,7 +956,7 @@ static int __map_osds(struct ceph_osd_client *osdc, (req->r_osd == NULL && o == -1)) return 0; /* no change */ - dout("map_osds tid %llu pgid %d.%x osd%d (was osd%d)\n", + dout("map_request tid %llu pgid %d.%x osd%d (was osd%d)\n", req->r_tid, le32_to_cpu(pgid.pool), le16_to_cpu(pgid.ps), o, req->r_osd ? req->r_osd->o_osd : -1); @@ -841,10 +974,12 @@ static int __map_osds(struct ceph_osd_client *osdc, if (!req->r_osd && o >= 0) { err = -ENOMEM; req->r_osd = create_osd(osdc); - if (!req->r_osd) + if (!req->r_osd) { + list_move(&req->r_req_lru_item, &osdc->req_notarget); goto out; + } - dout("map_osds osd %p is osd%d\n", req->r_osd, o); + dout("map_request osd %p is osd%d\n", req->r_osd, o); req->r_osd->o_osd = o; req->r_osd->o_con.peer_name.num = cpu_to_le64(o); __insert_osd(osdc, req->r_osd); @@ -855,6 +990,9 @@ static int __map_osds(struct ceph_osd_client *osdc, if (req->r_osd) { __remove_osd_from_lru(req->r_osd); list_add(&req->r_osd_item, &req->r_osd->o_requests); + list_move(&req->r_req_lru_item, &osdc->req_unsent); + } else { + list_move(&req->r_req_lru_item, &osdc->req_notarget); } err = 1; /* osd or pg changed */ @@ -869,16 +1007,6 @@ static int __send_request(struct ceph_osd_client *osdc, struct ceph_osd_request *req) { struct ceph_osd_request_head *reqhead; - int err; - - err = __map_osds(osdc, req); - if (err < 0) - return err; - if (req->r_osd == NULL) { - dout("send_request %p no up osds in pg\n", req); - ceph_monc_request_next_osdmap(&osdc->client->monc); - return 0; - } dout("send_request %p tid %llu to osd%d flags %d\n", req, req->r_tid, req->r_osd->o_osd, req->r_flags); @@ -898,6 +1026,21 @@ static int __send_request(struct ceph_osd_client *osdc, } /* + * Send any requests in the queue (req_unsent). + */ +static void send_queued(struct ceph_osd_client *osdc) +{ + struct ceph_osd_request *req, *tmp; + + dout("send_queued\n"); + mutex_lock(&osdc->request_mutex); + list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item) { + __send_request(osdc, req); + } + mutex_unlock(&osdc->request_mutex); +} + +/* * Timeout callback, called every N seconds when 1 or more osd * requests has been active for more than N seconds. When this * happens, we ping all OSDs with requests who have timed out to @@ -916,30 +1059,13 @@ static void handle_timeout(struct work_struct *work) unsigned long keepalive = osdc->client->options->osd_keepalive_timeout * HZ; unsigned long last_stamp = 0; - struct rb_node *p; struct list_head slow_osds; - dout("timeout\n"); down_read(&osdc->map_sem); ceph_monc_request_next_osdmap(&osdc->client->monc); mutex_lock(&osdc->request_mutex); - for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { - req = rb_entry(p, struct ceph_osd_request, r_node); - - if (req->r_resend) { - int err; - - dout("osdc resending prev failed %lld\n", req->r_tid); - err = __send_request(osdc, req); - if (err) - dout("osdc failed again on %lld\n", req->r_tid); - else - req->r_resend = false; - continue; - } - } /* * reset osds that appear to be _really_ unresponsive. this @@ -963,7 +1089,7 @@ static void handle_timeout(struct work_struct *work) BUG_ON(!osd); pr_warning(" tid %llu timed out on osd%d, will reset osd\n", req->r_tid, osd->o_osd); - __kick_requests(osdc, osd); + __kick_osd_requests(osdc, osd); } /* @@ -991,7 +1117,7 @@ static void handle_timeout(struct work_struct *work) __schedule_osd_timeout(osdc); mutex_unlock(&osdc->request_mutex); - + send_queued(osdc); up_read(&osdc->map_sem); } @@ -1035,7 +1161,6 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg, numops * sizeof(struct ceph_osd_op)) goto bad; dout("handle_reply %p tid %llu result %d\n", msg, tid, (int)result); - /* lookup */ mutex_lock(&osdc->request_mutex); req = __lookup_request(osdc, tid); @@ -1079,6 +1204,9 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg, dout("handle_reply tid %llu flags %d\n", tid, flags); + if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK)) + __re |