From mboxrd@z Thu Jan 1 00:00:00 1970 From: Simon Derr Subject: [PATCH 00/10] Improve 9P/RDMA Date: Tue, 2 Jul 2013 15:11:14 +0200 Message-ID: <1372770684-25573-1-git-send-email-simon.derr@bull.net> Cc: simon.derr@bull.net, ericvh@gmail.com To: netdev@vger.kernel.org Return-path: Received: from ecfrec.frec.bull.fr ([129.183.4.8]:40990 "EHLO ecfrec.frec.bull.fr" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752550Ab3GBNl5 (ORCPT ); Tue, 2 Jul 2013 09:41:57 -0400 Sender: netdev-owner@vger.kernel.org List-ID: Hi, This is a series of patches which aim to make 9P/RDMA better. Eric Van Hensbergen asked that I send them to netdev to get a wider review. They deal mainly with fcall allocation/deallocation, RQ locking, error handling, and above all buffer management. This has been tested intensively with the nfs-ganesha server. Regards, Simon Simon Derr (10): 9P: Fix fcall allocation for rdma 9P/RDMA: rdma_request() needs not allocate req->rc 9pnet: refactor struct p9_fcall alloc code 9P/RDMA: increase P9_RDMA_MAXSIZE to 1MB 9P/RDMA: Protect against duplicate replies 9P/RDMA: Use a semaphore to protect the RQ 9P/RDMA: Do not free req->rc in error handling in rdma_request() 9P/RDMA: Improve error handling in rdma_request 9P/RDMA: count posted buffers without a pending request 9P: Add cancelled() to the transport functions. include/net/9p/transport.h | 3 + net/9p/client.c | 65 +++++++++++++--------- net/9p/trans_rdma.c | 132 ++++++++++++++++++++++++++++---------------- 3 files changed, 126 insertions(+), 74 deletions(-) -- 1.7.2.2