From f6310491c4cdb88af73aa551ec9df1f10a90c709 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Mon, 6 Jul 2026 09:29:21 -0400 Subject: [PATCH] sunrpc: route to a populated pool in svc_pool_for_cpu() svc_set_num_threads() spreads the requested threads evenly across the service's pools (base = nrservs / sv_nrpools). When a service runs fewer threads than it has pools -- e.g. an nfsd configured with fewer threads than the host has NUMA nodes while running in "pernode" or "percpu" mode -- the trailing pools are left with no threads at all. svc_xprt_enqueue() selects a pool from the CPU servicing the transport, queues the transport on that pool's sp_xprts, and only wakes a thread from the same pool. Each thread services exclusively its own pool, so a transport that lands on a threadless pool is enqueued on sp_xprts and never picked up: the connection hangs indefinitely. Have svc_pool_for_cpu() skip pools that currently have no threads, falling back to the next populated pool. This trades NUMA locality for a guarantee that the work is actually serviced. sp_nrthreads is only updated under the service mutex; the lockless read here is a best-effort routing hint, so annotate it with data_race(). Fixes: bfd241600a3b ("[PATCH] knfsd: make rpc threads pools numa aware") Cc: stable@vger.kernel.org Signed-off-by: Jeff Layton Link: https://patch.msgid.link/20260706-sunrpc-pool-mode-v5-1-6c4ee7cd89aa@kernel.org Signed-off-by: Chuck Lever --- net/sunrpc/svc.c | 29 ++++++++++++++++++++++++++++- 1 file changed, 28 insertions(+), 1 deletion(-) diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c index 88b5d6d67e6e..31641d95aa4d 100644 --- a/net/sunrpc/svc.c +++ b/net/sunrpc/svc.c @@ -402,6 +402,7 @@ struct svc_pool *svc_pool_for_cpu(struct svc_serv *serv) struct svc_pool_map *m = &svc_pool_map; int cpu = raw_smp_processor_id(); unsigned int pidx = 0; + unsigned int i; if (serv->sv_nrpools <= 1) return serv->sv_pools; @@ -414,8 +415,34 @@ struct svc_pool *svc_pool_for_cpu(struct svc_serv *serv) pidx = m->to_pool[cpu_to_node(cpu)]; break; } + pidx %= serv->sv_nrpools; - return &serv->sv_pools[pidx % serv->sv_nrpools]; + /* + * It's possible to have a pool with no threads. Userland can just set + * things up this way directly. Also, when threads are autodistributed + * they are spread evenly across the pools, but when there are fewer + * threads than pools some pools can end up with none. + * + * A transport enqueued on a threadless pool would never be picked up, + * since each thread only services its own pool. Fall back to the next + * populated pool, trading NUMA locality for a guarantee that the + * transport is serviced. + */ + for (i = 0; i < serv->sv_nrpools; i++) { + struct svc_pool *pool = &serv->sv_pools[pidx]; + + /* This is set under the service mutex and rarely ever + * changes. A data race here is harmless. + */ + if (data_race(pool->sp_nrthreads)) + return pool; + + if (++pidx >= serv->sv_nrpools) + pidx = 0; + } + + /* No pool has any threads; nothing can service the transport. */ + return &serv->sv_pools[pidx]; } static int svc_rpcb_setup(struct svc_serv *serv, struct net *net)