From patchwork Tue Apr 21 21:27:50 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ondrej Zary X-Patchwork-Id: 6253841 Return-Path: X-Original-To: patchwork-linux-scsi@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 01AC2BF4A6 for ; Tue, 21 Apr 2015 21:28:16 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 26DAC20304 for ; Tue, 21 Apr 2015 21:28:15 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id E76A7202D1 for ; Tue, 21 Apr 2015 21:28:13 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933060AbbDUV2N (ORCPT ); Tue, 21 Apr 2015 17:28:13 -0400 Received: from ns.gsystem.sk ([62.176.172.50]:55301 "EHLO gsystem.sk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932871AbbDUV2M (ORCPT ); Tue, 21 Apr 2015 17:28:12 -0400 Received: from stip-static-68.213-81-217.telecom.sk ([213.81.217.68] helo=gsql.ggedos.sk) by gsystem.sk with esmtpsa (TLS1.1:DHE_RSA_AES_256_CBC_SHA1:256) (Exim 4.80) (envelope-from ) id 1Ykfi5-0000sK-IO; Tue, 21 Apr 2015 23:28:01 +0200 From: Ondrej Zary To: Dan Carpenter Cc: James Bottomley , linux-scsi@vger.kernel.org Subject: [PATCH] aha1542: Allocate memory before taking a lock Date: Tue, 21 Apr 2015 23:27:50 +0200 Message-Id: <1429651670-5466-1-git-send-email-linux@rainbow-software.org> X-Mailer: git-send-email 1.7.10.4 Sender: linux-scsi-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-scsi@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=ham 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 driver currently calls kmalloc with GFP_KERNEL while holding a lock. Fix it by doing the allocation earlier, before taking the lock. Tested on AHA-1542B. Signed-off-by: Ondrej Zary --- drivers/scsi/aha1542.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) diff --git a/drivers/scsi/aha1542.c b/drivers/scsi/aha1542.c index ec43276..b95d277 100644 --- a/drivers/scsi/aha1542.c +++ b/drivers/scsi/aha1542.c @@ -375,9 +375,10 @@ static int aha1542_queuecommand(struct Scsi_Host *sh, struct scsi_cmnd *cmd) u8 lun = cmd->device->lun; unsigned long flags; int bufflen = scsi_bufflen(cmd); - int mbo; + int mbo, sg_count; struct mailbox *mb = aha1542->mb; struct ccb *ccb = aha1542->ccb; + struct chain *cptr; if (*cmd->cmnd == REQUEST_SENSE) { /* Don't do the command - we have the sense data already */ @@ -397,6 +398,13 @@ static int aha1542_queuecommand(struct Scsi_Host *sh, struct scsi_cmnd *cmd) print_hex_dump_bytes("command: ", DUMP_PREFIX_NONE, cmd->cmnd, cmd->cmd_len); } #endif + if (bufflen) { /* allocate memory before taking host_lock */ + sg_count = scsi_sg_count(cmd); + cptr = kmalloc(sizeof(*cptr) * sg_count, GFP_KERNEL | GFP_DMA); + if (!cptr) + return SCSI_MLQUEUE_HOST_BUSY; + } + /* Use the outgoing mailboxes in a round-robin fashion, because this is how the host adapter will scan for them */ @@ -441,19 +449,10 @@ static int aha1542_queuecommand(struct Scsi_Host *sh, struct scsi_cmnd *cmd) if (bufflen) { struct scatterlist *sg; - struct chain *cptr; - int i, sg_count = scsi_sg_count(cmd); + int i; ccb[mbo].op = 2; /* SCSI Initiator Command w/scatter-gather */ - cmd->host_scribble = kmalloc(sizeof(*cptr)*sg_count, - GFP_KERNEL | GFP_DMA); - cptr = (struct chain *) cmd->host_scribble; - if (cptr == NULL) { - /* free the claimed mailbox slot */ - aha1542->int_cmds[mbo] = NULL; - spin_unlock_irqrestore(sh->host_lock, flags); - return SCSI_MLQUEUE_HOST_BUSY; - } + cmd->host_scribble = (void *)cptr; scsi_for_each_sg(cmd, sg, sg_count, i) { any2scsi(cptr[i].dataptr, isa_page_to_bus(sg_page(sg)) + sg->offset);