diff options
author | Stephen Rothwell <sfr@canb.auug.org.au> | 2015-07-24 11:02:32 +1000 |
---|---|---|
committer | Stephen Rothwell <sfr@canb.auug.org.au> | 2015-07-24 11:02:32 +1000 |
commit | a28428be52ac5e486191b63c50dd06464d4abb1b (patch) | |
tree | 1f1a0bfd5a3875cb271cb8dec30947adb297b3e6 /net | |
parent | 4f77f64eec81958dc48578e101e094c16879b785 (diff) | |
parent | a75d029a958875a05e53b7a53be60c898cb7c036 (diff) |
Merge remote-tracking branch 'nfsd/nfsd-next'
Diffstat (limited to 'net')
-rw-r--r-- | net/sunrpc/svc.c | 113 | ||||
-rw-r--r-- | net/sunrpc/svc_xprt.c | 10 | ||||
-rw-r--r-- | net/sunrpc/xprtrdma/svc_rdma_sendto.c | 83 | ||||
-rw-r--r-- | net/sunrpc/xprtrdma/svc_rdma_transport.c | 35 |
4 files changed, 143 insertions, 98 deletions
diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c index 5a16d8d8c831..a8f579df14d8 100644 --- a/net/sunrpc/svc.c +++ b/net/sunrpc/svc.c @@ -34,36 +34,19 @@ static void svc_unregister(const struct svc_serv *serv, struct net *net); -#define svc_serv_is_pooled(serv) ((serv)->sv_function) +#define svc_serv_is_pooled(serv) ((serv)->sv_ops->svo_function) -/* - * Mode for mapping cpus to pools. - */ -enum { - SVC_POOL_AUTO = -1, /* choose one of the others */ - SVC_POOL_GLOBAL, /* no mapping, just a single global pool - * (legacy & UP mode) */ - SVC_POOL_PERCPU, /* one pool per cpu */ - SVC_POOL_PERNODE /* one pool per numa node */ -}; #define SVC_POOL_DEFAULT SVC_POOL_GLOBAL /* * Structure for mapping cpus to pools and vice versa. * Setup once during sunrpc initialisation. */ -static struct svc_pool_map { - int count; /* How many svc_servs use us */ - int mode; /* Note: int not enum to avoid - * warnings about "enumeration value - * not handled in switch" */ - unsigned int npools; - unsigned int *pool_to; /* maps pool id to cpu or node */ - unsigned int *to_pool; /* maps cpu or node to pool id */ -} svc_pool_map = { - .count = 0, +struct svc_pool_map svc_pool_map = { .mode = SVC_POOL_DEFAULT }; +EXPORT_SYMBOL_GPL(svc_pool_map); + static DEFINE_MUTEX(svc_pool_map_mutex);/* protects svc_pool_map.count only */ static int @@ -236,7 +219,7 @@ svc_pool_map_init_pernode(struct svc_pool_map *m) * vice versa). Initialise the map if we're the first user. * Returns the number of pools. */ -static unsigned int +unsigned int svc_pool_map_get(void) { struct svc_pool_map *m = &svc_pool_map; @@ -271,7 +254,7 @@ svc_pool_map_get(void) mutex_unlock(&svc_pool_map_mutex); return m->npools; } - +EXPORT_SYMBOL_GPL(svc_pool_map_get); /* * Drop a reference to the global map of cpus to pools. @@ -280,7 +263,7 @@ svc_pool_map_get(void) * mode using the pool_mode module option without * rebooting or re-loading sunrpc.ko. */ -static void +void svc_pool_map_put(void) { struct svc_pool_map *m = &svc_pool_map; @@ -297,7 +280,7 @@ svc_pool_map_put(void) mutex_unlock(&svc_pool_map_mutex); } - +EXPORT_SYMBOL_GPL(svc_pool_map_put); static int svc_pool_map_get_node(unsigned int pidx) { @@ -423,7 +406,7 @@ EXPORT_SYMBOL_GPL(svc_bind); */ static struct svc_serv * __svc_create(struct svc_program *prog, unsigned int bufsize, int npools, - void (*shutdown)(struct svc_serv *serv, struct net *net)) + struct svc_serv_ops *ops) { struct svc_serv *serv; unsigned int vers; @@ -440,7 +423,7 @@ __svc_create(struct svc_program *prog, unsigned int bufsize, int npools, bufsize = RPCSVC_MAXPAYLOAD; serv->sv_max_payload = bufsize? bufsize : 4096; serv->sv_max_mesg = roundup(serv->sv_max_payload + PAGE_SIZE, PAGE_SIZE); - serv->sv_shutdown = shutdown; + serv->sv_ops = ops; xdrsize = 0; while (prog) { prog->pg_lovers = prog->pg_nvers-1; @@ -486,26 +469,22 @@ __svc_create(struct svc_program *prog, unsigned int bufsize, int npools, struct svc_serv * svc_create(struct svc_program *prog, unsigned int bufsize, - void (*shutdown)(struct svc_serv *serv, struct net *net)) + struct svc_serv_ops *ops) { - return __svc_create(prog, bufsize, /*npools*/1, shutdown); + return __svc_create(prog, bufsize, /*npools*/1, ops); } EXPORT_SYMBOL_GPL(svc_create); struct svc_serv * svc_create_pooled(struct svc_program *prog, unsigned int bufsize, - void (*shutdown)(struct svc_serv *serv, struct net *net), - svc_thread_fn func, struct module *mod) + struct svc_serv_ops *ops) { struct svc_serv *serv; unsigned int npools = svc_pool_map_get(); - serv = __svc_create(prog, bufsize, npools, shutdown); + serv = __svc_create(prog, bufsize, npools, ops); if (!serv) goto out_err; - - serv->sv_function = func; - serv->sv_module = mod; return serv; out_err: svc_pool_map_put(); @@ -517,8 +496,8 @@ void svc_shutdown_net(struct svc_serv *serv, struct net *net) { svc_close_net(serv, net); - if (serv->sv_shutdown) - serv->sv_shutdown(serv, net); + if (serv->sv_ops->svo_shutdown) + serv->sv_ops->svo_shutdown(serv, net); } EXPORT_SYMBOL_GPL(svc_shutdown_net); @@ -604,40 +583,52 @@ svc_release_buffer(struct svc_rqst *rqstp) } struct svc_rqst * -svc_prepare_thread(struct svc_serv *serv, struct svc_pool *pool, int node) +svc_rqst_alloc(struct svc_serv *serv, struct svc_pool *pool, int node) { struct svc_rqst *rqstp; rqstp = kzalloc_node(sizeof(*rqstp), GFP_KERNEL, node); if (!rqstp) - goto out_enomem; + return rqstp; - serv->sv_nrthreads++; __set_bit(RQ_BUSY, &rqstp->rq_flags); spin_lock_init(&rqstp->rq_lock); rqstp->rq_server = serv; rqstp->rq_pool = pool; - spin_lock_bh(&pool->sp_lock); - pool->sp_nrthreads++; - list_add_rcu(&rqstp->rq_all, &pool->sp_all_threads); - spin_unlock_bh(&pool->sp_lock); rqstp->rq_argp = kmalloc_node(serv->sv_xdrsize, GFP_KERNEL, node); if (!rqstp->rq_argp) - goto out_thread; + goto out_enomem; rqstp->rq_resp = kmalloc_node(serv->sv_xdrsize, GFP_KERNEL, node); if (!rqstp->rq_resp) - goto out_thread; + goto out_enomem; if (!svc_init_buffer(rqstp, serv->sv_max_mesg, node)) - goto out_thread; + goto out_enomem; return rqstp; -out_thread: - svc_exit_thread(rqstp); out_enomem: - return ERR_PTR(-ENOMEM); + svc_rqst_free(rqstp); + return NULL; +} +EXPORT_SYMBOL_GPL(svc_rqst_alloc); + +struct svc_rqst * +svc_prepare_thread(struct svc_serv *serv, struct svc_pool *pool, int node) +{ + struct svc_rqst *rqstp; + + rqstp = svc_rqst_alloc(serv, pool, node); + if (!rqstp) + return ERR_PTR(-ENOMEM); + + serv->sv_nrthreads++; + spin_lock_bh(&pool->sp_lock); + pool->sp_nrthreads++; + list_add_rcu(&rqstp->rq_all, &pool->sp_all_threads); + spin_unlock_bh(&pool->sp_lock); + return rqstp; } EXPORT_SYMBOL_GPL(svc_prepare_thread); @@ -739,12 +730,12 @@ svc_set_num_threads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) break; } - __module_get(serv->sv_module); - task = kthread_create_on_node(serv->sv_function, rqstp, + __module_get(serv->sv_ops->svo_module); + task = kthread_create_on_node(serv->sv_ops->svo_function, rqstp, node, "%s", serv->sv_name); if (IS_ERR(task)) { error = PTR_ERR(task); - module_put(serv->sv_module); + module_put(serv->sv_ops->svo_module); svc_exit_thread(rqstp); break; } @@ -772,15 +763,21 @@ EXPORT_SYMBOL_GPL(svc_set_num_threads); * mutex" for the service. */ void -svc_exit_thread(struct svc_rqst *rqstp) +svc_rqst_free(struct svc_rqst *rqstp) { - struct svc_serv *serv = rqstp->rq_server; - struct svc_pool *pool = rqstp->rq_pool; - svc_release_buffer(rqstp); kfree(rqstp->rq_resp); kfree(rqstp->rq_argp); kfree(rqstp->rq_auth_data); + kfree_rcu(rqstp, rq_rcu_head); +} +EXPORT_SYMBOL_GPL(svc_rqst_free); + +void +svc_exit_thread(struct svc_rqst *rqstp) +{ + struct svc_serv *serv = rqstp->rq_server; + struct svc_pool *pool = rqstp->rq_pool; spin_lock_bh(&pool->sp_lock); pool->sp_nrthreads--; @@ -788,7 +785,7 @@ svc_exit_thread(struct svc_rqst *rqstp) list_del_rcu(&rqstp->rq_all); spin_unlock_bh(&pool->sp_lock); - kfree_rcu(rqstp, rq_rcu_head); + svc_rqst_free(rqstp); /* Release the server */ if (serv) diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index 163ac45c3639..a6cbb2104667 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -24,7 +24,6 @@ static int svc_deferred_recv(struct svc_rqst *rqstp); static struct cache_deferred_req *svc_defer(struct cache_req *req); static void svc_age_temp_xprts(unsigned long closure); static void svc_delete_xprt(struct svc_xprt *xprt); -static void svc_xprt_do_enqueue(struct svc_xprt *xprt); /* apparently the "standard" is that clients close * idle connections after 5 minutes, servers after @@ -225,12 +224,12 @@ static void svc_xprt_received(struct svc_xprt *xprt) } /* As soon as we clear busy, the xprt could be closed and - * 'put', so we need a reference to call svc_xprt_do_enqueue with: + * 'put', so we need a reference to call svc_enqueue_xprt with: */ svc_xprt_get(xprt); smp_mb__before_atomic(); clear_bit(XPT_BUSY, &xprt->xpt_flags); - svc_xprt_do_enqueue(xprt); + xprt->xpt_server->sv_ops->svo_enqueue_xprt(xprt); svc_xprt_put(xprt); } @@ -320,7 +319,7 @@ static bool svc_xprt_has_something_to_do(struct svc_xprt *xprt) return false; } -static void svc_xprt_do_enqueue(struct svc_xprt *xprt) +void svc_xprt_do_enqueue(struct svc_xprt *xprt) { struct svc_pool *pool; struct svc_rqst *rqstp = NULL; @@ -402,6 +401,7 @@ redo_search: out: trace_svc_xprt_do_enqueue(xprt, rqstp); } +EXPORT_SYMBOL_GPL(svc_xprt_do_enqueue); /* * Queue up a transport with data pending. If there are idle nfsd @@ -412,7 +412,7 @@ void svc_xprt_enqueue(struct svc_xprt *xprt) { if (test_bit(XPT_BUSY, &xprt->xpt_flags)) return; - svc_xprt_do_enqueue(xprt); + xprt->xpt_server->sv_ops->svo_enqueue_xprt(xprt); } EXPORT_SYMBOL_GPL(svc_xprt_enqueue); diff --git a/net/sunrpc/xprtrdma/svc_rdma_sendto.c b/net/sunrpc/xprtrdma/svc_rdma_sendto.c index d25cd430f9ff..1dfae8317065 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_sendto.c +++ b/net/sunrpc/xprtrdma/svc_rdma_sendto.c @@ -136,6 +136,79 @@ static dma_addr_t dma_map_xdr(struct svcxprt_rdma *xprt, return dma_addr; } +/* Returns the address of the first read chunk or <nul> if no read chunk + * is present + */ +struct rpcrdma_read_chunk * +svc_rdma_get_read_chunk(struct rpcrdma_msg *rmsgp) +{ + struct rpcrdma_read_chunk *ch = + (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0]; + + if (ch->rc_discrim == xdr_zero) + return NULL; + return ch; +} + +/* Returns the address of the first read write array element or <nul> + * if no write array list is present + */ +static struct rpcrdma_write_array * +svc_rdma_get_write_array(struct rpcrdma_msg *rmsgp) +{ + if (rmsgp->rm_body.rm_chunks[0] != xdr_zero || + rmsgp->rm_body.rm_chunks[1] == xdr_zero) + return NULL; + return (struct rpcrdma_write_array *)&rmsgp->rm_body.rm_chunks[1]; +} + +/* Returns the address of the first reply array element or <nul> if no + * reply array is present + */ +static struct rpcrdma_write_array * +svc_rdma_get_reply_array(struct rpcrdma_msg *rmsgp) +{ + struct rpcrdma_read_chunk *rch; + struct rpcrdma_write_array *wr_ary; + struct rpcrdma_write_array *rp_ary; + + /* XXX: Need to fix when reply chunk may occur with read list + * and/or write list. + */ + if (rmsgp->rm_body.rm_chunks[0] != xdr_zero || + rmsgp->rm_body.rm_chunks[1] != xdr_zero) + return NULL; + + rch = svc_rdma_get_read_chunk(rmsgp); + if (rch) { + while (rch->rc_discrim != xdr_zero) + rch++; + + /* The reply chunk follows an empty write array located + * at 'rc_position' here. The reply array is at rc_target. + */ + rp_ary = (struct rpcrdma_write_array *)&rch->rc_target; + goto found_it; + } + + wr_ary = svc_rdma_get_write_array(rmsgp); + if (wr_ary) { + int chunk = be32_to_cpu(wr_ary->wc_nchunks); + + rp_ary = (struct rpcrdma_write_array *) + &wr_ary->wc_array[chunk].wc_target.rs_length; + goto found_it; + } + + /* No read list, no write list */ + rp_ary = (struct rpcrdma_write_array *)&rmsgp->rm_body.rm_chunks[2]; + + found_it: + if (rp_ary->wc_discrim == xdr_zero) + return NULL; + return rp_ary; +} + /* Assumptions: * - The specified write_len can be represented in sc_max_sge * PAGE_SIZE */ @@ -384,6 +457,7 @@ static int send_reply(struct svcxprt_rdma *rdma, int byte_count) { struct ib_send_wr send_wr; + u32 xdr_off; int sge_no; int sge_bytes; int page_no; @@ -418,8 +492,8 @@ static int send_reply(struct svcxprt_rdma *rdma, ctxt->direction = DMA_TO_DEVICE; /* Map the payload indicated by 'byte_count' */ + xdr_off = 0; for (sge_no = 1; byte_count && sge_no < vec->count; sge_no++) { - int xdr_off = 0; sge_bytes = min_t(size_t, vec->sge[sge_no].iov_len, byte_count); byte_count -= sge_bytes; ctxt->sge[sge_no].addr = @@ -457,6 +531,13 @@ static int send_reply(struct svcxprt_rdma *rdma, } rqstp->rq_next_page = rqstp->rq_respages + 1; + /* The loop above bumps sc_dma_used for each sge. The + * xdr_buf.tail gets a separate sge, but resides in the + * same page as xdr_buf.head. Don't count it twice. + */ + if (sge_no > ctxt->count) + atomic_dec(&rdma->sc_dma_used); + if (sge_no > rdma->sc_max_sge) { pr_err("svcrdma: Too many sges (%d)\n", sge_no); goto err; diff --git a/net/sunrpc/xprtrdma/svc_rdma_transport.c b/net/sunrpc/xprtrdma/svc_rdma_transport.c index 6b36279e4288..4054a9de6a91 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_transport.c +++ b/net/sunrpc/xprtrdma/svc_rdma_transport.c @@ -659,6 +659,7 @@ static int rdma_cma_handler(struct rdma_cm_id *cma_id, if (xprt) { set_bit(XPT_CLOSE, &xprt->xpt_flags); svc_xprt_enqueue(xprt); + svc_xprt_put(xprt); } break; default: @@ -1201,40 +1202,6 @@ static int svc_rdma_secure_port(struct svc_rqst *rqstp) return 1; } -/* - * Attempt to register the kvec representing the RPC memory with the - * device. - * - * Returns: - * NULL : The device does not support fastreg or there were no more - * fastreg mr. - * frmr : The kvec register request was successfully posted. - * <0 : An error was encountered attempting to register the kvec. - */ -int svc_rdma_fastreg(struct svcxprt_rdma *xprt, - struct svc_rdma_fastreg_mr *frmr) -{ - struct ib_send_wr fastreg_wr; - u8 key; - - /* Bump the key */ - key = (u8)(frmr->mr->lkey & 0x000000FF); - ib_update_fast_reg_key(frmr->mr, ++key); - - /* 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; - return svc_rdma_send(xprt, &fastreg_wr); -} - int svc_rdma_send(struct svcxprt_rdma *xprt, struct ib_send_wr *wr) { struct ib_send_wr *bad_wr, *n_wr; |