From: Hannes Reinecke Date: Wed, 28 May 2025 06:45:34 +0000 (+0200) Subject: nvme-tcp: fix I/O stalls on congested sockets X-Git-Tag: v6.16-rc1~26^2^2~1 X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=f42d4796ee100fade86086d1cf98537fb4d326c8;p=thirdparty%2Fkernel%2Flinux.git nvme-tcp: fix I/O stalls on congested sockets When the socket is busy processing nvme_tcp_try_recv() might return -EAGAIN, but this doesn't automatically imply that the sending side is blocked, too. So check if there are pending requests once nvme_tcp_try_recv() returns -EAGAIN and continue with the sending loop to avoid I/O stalls. Signed-off-by: Hannes Reinecke Acked-by: Chris Leech Reviewed-by: Sagi Grimberg Signed-off-by: Christoph Hellwig --- diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c index b37028320edd..abf78dc27e84 100644 --- a/drivers/nvme/host/tcp.c +++ b/drivers/nvme/host/tcp.c @@ -1361,7 +1361,7 @@ static int nvme_tcp_try_recv(struct nvme_tcp_queue *queue) queue->nr_cqe = 0; consumed = sock->ops->read_sock(sk, &rd_desc, nvme_tcp_recv_skb); release_sock(sk); - return consumed; + return consumed == -EAGAIN ? 0 : consumed; } static void nvme_tcp_io_work(struct work_struct *w) @@ -1389,6 +1389,11 @@ static void nvme_tcp_io_work(struct work_struct *w) else if (unlikely(result < 0)) return; + /* did we get some space after spending time in recv? */ + if (nvme_tcp_queue_has_pending(queue) && + sk_stream_is_writeable(queue->sock->sk)) + pending = true; + if (!pending || !queue->rd_enabled) return;