From patchwork Thu May 21 05:11:40 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Wengang Wang X-Patchwork-Id: 6452231 Return-Path: X-Original-To: patchwork-linux-rdma@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork2.web.kernel.org (Postfix) with ESMTP id 057C5C0432 for ; Thu, 21 May 2015 05:11:44 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 42BB320426 for ; Thu, 21 May 2015 05:11:42 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id E81A92041A for ; Thu, 21 May 2015 05:11:40 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751140AbbEUFLj (ORCPT ); Thu, 21 May 2015 01:11:39 -0400 Received: from userp1040.oracle.com ([156.151.31.81]:27349 "EHLO userp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751193AbbEUFLi (ORCPT ); Thu, 21 May 2015 01:11:38 -0400 Received: from userv0021.oracle.com (userv0021.oracle.com [156.151.31.71]) by userp1040.oracle.com (Sentrion-MTA-4.3.2/Sentrion-MTA-4.3.2) with ESMTP id t4L5BcZp016159 (version=TLSv1 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Thu, 21 May 2015 05:11:38 GMT Received: from aserv0121.oracle.com (aserv0121.oracle.com [141.146.126.235]) by userv0021.oracle.com (8.13.8/8.13.8) with ESMTP id t4L5Bbc4032474 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=FAIL) for ; Thu, 21 May 2015 05:11:38 GMT Received: from abhmp0018.oracle.com (abhmp0018.oracle.com [141.146.116.24]) by aserv0121.oracle.com (8.13.8/8.13.8) with ESMTP id t4L5BbLK014576 for ; Thu, 21 May 2015 05:11:37 GMT Received: from oracle.com (/10.182.64.160) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Wed, 20 May 2015 22:11:37 -0700 From: Wengang Wang To: linux-rdma@vger.kernel.org Cc: wen.gang.wang@oracle.com Subject: [PATCH] rds: re-entry of rds_ib_xmit/rds_iw_xmit Date: Thu, 21 May 2015 13:11:40 +0800 Message-Id: <1432185100-5613-1-git-send-email-wen.gang.wang@oracle.com> X-Mailer: git-send-email 2.1.0 X-Source-IP: userv0021.oracle.com [156.151.31.71] Sender: linux-rdma-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, T_RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP The BUG_ON at line 452/453 is triggered in function rds_send_xmit. 441 while (ret) { 442 tmp = min_t(int, ret, sg->length - 443 conn->c_xmit_data_off); 444 conn->c_xmit_data_off += tmp; 445 ret -= tmp; 446 if (conn->c_xmit_data_off == sg->length) { 447 conn->c_xmit_data_off = 0; 448 sg++; 449 conn->c_xmit_sg++; 450 if (ret != 0 && conn->c_xmit_sg == rm->data.op_nents) 451 printk(KERN_ERR "conn %p rm %p sg %p ret %d\n", conn, rm, sg, ret); 452 BUG_ON(ret != 0 && 453 conn->c_xmit_sg == rm->data.op_nents); 454 } 455 } it is complaining the total sent length is bigger that we want to send. rds_ib_xmit() is wrong for the second entry for the same rds_message returning wrong value. the sg and off passed by rds_send_xmit to rds_ib_xmit is based on scatterlist.offset/length, but the rds_ib_xmit action is based on scatterlist.dma_address/dma_length. in case dma_length is larger than length there is problem. for the 2nd and later entries of rds_ib_xmit for same rds_message, at least one of the following two is wrong: 1) the scatterlist to start with, the choosen one can far beyond the correct one. 2) the offset to start with within the scatterlist. fix: add op_dmasg and op_dmaoff to rm_data_op structure indicating the scatterlist and offset within the it to start with for rds_ib_xmit respectively. op_dmasg and op_dmaoff are initialized to zero when doing dma mapping for the first see of the message and are changed when filling send slots. the same applies to rds_iw_xmit too. Signed-off-by: Wengang Wang --- net/rds/ib_send.c | 17 +++++++++++------ net/rds/iw_send.c | 18 +++++++++++------- net/rds/rds.h | 2 ++ 3 files changed, 24 insertions(+), 13 deletions(-) diff --git a/net/rds/ib_send.c b/net/rds/ib_send.c index bd3825d..1df6c84 100644 --- a/net/rds/ib_send.c +++ b/net/rds/ib_send.c @@ -605,6 +605,8 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm, } rds_message_addref(rm); + rm->data.op_dmasg = 0; + rm->data.op_dmaoff = 0; ic->i_data_op = &rm->data; /* Finalize the header */ @@ -658,7 +660,7 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm, send = &ic->i_sends[pos]; first = send; prev = NULL; - scat = &ic->i_data_op->op_sg[sg]; + scat = &ic->i_data_op->op_sg[rm->data.op_dmasg]; i = 0; do { unsigned int len = 0; @@ -680,17 +682,20 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm, /* Set up the data, if present */ if (i < work_alloc && scat != &rm->data.op_sg[rm->data.op_count]) { - len = min(RDS_FRAG_SIZE, ib_sg_dma_len(dev, scat) - off); + len = min(RDS_FRAG_SIZE, + ib_sg_dma_len(dev, scat) - rm->data.op_dmaoff); send->s_wr.num_sge = 2; - send->s_sge[1].addr = ib_sg_dma_address(dev, scat) + off; + send->s_sge[1].addr = ib_sg_dma_address(dev, scat); + send->s_sge[1].addr += rm->data.op_dmaoff; send->s_sge[1].length = len; bytes_sent += len; - off += len; - if (off == ib_sg_dma_len(dev, scat)) { + rm->data.op_dmaoff += len; + if (rm->data.op_dmaoff == ib_sg_dma_len(dev, scat)) { scat++; - off = 0; + rm->data.op_dmasg++; + rm->data.op_dmaoff = 0; } } diff --git a/net/rds/iw_send.c b/net/rds/iw_send.c index 1383478..334fe98 100644 --- a/net/rds/iw_send.c +++ b/net/rds/iw_send.c @@ -581,6 +581,8 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm, ic->i_unsignaled_wrs = rds_iw_sysctl_max_unsig_wrs; ic->i_unsignaled_bytes = rds_iw_sysctl_max_unsig_bytes; rds_message_addref(rm); + rm->data.op_dmasg = 0; + rm->data.op_dmaoff = 0; ic->i_rm = rm; /* Finalize the header */ @@ -622,7 +624,7 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm, send = &ic->i_sends[pos]; first = send; prev = NULL; - scat = &rm->data.op_sg[sg]; + scat = &rm->data.op_sg[rm->data.op_dmasg]; sent = 0; i = 0; @@ -656,10 +658,11 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm, send = &ic->i_sends[pos]; - len = min(RDS_FRAG_SIZE, ib_sg_dma_len(dev, scat) - off); + len = min(RDS_FRAG_SIZE, + ib_sg_dma_len(dev, scat) - rm->data.op_dmaoff); rds_iw_xmit_populate_wr(ic, send, pos, - ib_sg_dma_address(dev, scat) + off, len, - send_flags); + ib_sg_dma_address(dev, scat) + rm->data.op_dmaoff, len, + send_flags); /* * We want to delay signaling completions just enough to get @@ -687,10 +690,11 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm, &send->s_wr, send->s_wr.num_sge, send->s_wr.next); sent += len; - off += len; - if (off == ib_sg_dma_len(dev, scat)) { + rm->data.op_dmaoff += len; + if (rm->data.op_dmaoff == ib_sg_dma_len(dev, scat)) { scat++; - off = 0; + rm->data.op_dmaoff = 0; + rm->data.op_dmasg++; } add_header: diff --git a/net/rds/rds.h b/net/rds/rds.h index 0d41155..d2c6009 100644 --- a/net/rds/rds.h +++ b/net/rds/rds.h @@ -363,6 +363,8 @@ struct rds_message { unsigned int op_active:1; unsigned int op_nents; unsigned int op_count; + unsigned int op_dmasg; + unsigned int op_dmaoff; struct scatterlist *op_sg; } data; };