From patchwork Thu Jun 8 10:24:57 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Selvin Xavier X-Patchwork-Id: 13271924 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id D3C63C7EE23 for ; Thu, 8 Jun 2023 10:37:20 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S236037AbjFHKhT (ORCPT ); Thu, 8 Jun 2023 06:37:19 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:49570 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235963AbjFHKhS (ORCPT ); Thu, 8 Jun 2023 06:37:18 -0400 Received: from mail-pl1-x62f.google.com (mail-pl1-x62f.google.com [IPv6:2607:f8b0:4864:20::62f]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id B45582D53 for ; Thu, 8 Jun 2023 03:37:13 -0700 (PDT) Received: by mail-pl1-x62f.google.com with SMTP id d9443c01a7336-1b025d26f4fso2752935ad.1 for ; Thu, 08 Jun 2023 03:37:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=broadcom.com; s=google; t=1686220633; x=1688812633; h=references:in-reply-to:message-id:date:subject:cc:to:from:from:to :cc:subject:date:message-id:reply-to; bh=2VsZ55l/p5N1O/trRd6wo49opUNJKzs0Yf4+iMapfe4=; b=U7WB141eME8/iGe7ltZKwuEpp4ibQQStXosGEfw1PP1zCDSIbCICMHZI7+9bYmNaau Hbkprj0DzBHCy9vmvgECAYoucj47Za831osv++ZFBiqOQy1vpe+k6Lyn0pC4t2/ikCqS 4eOy9EHY4fNS9goM+0kPr6W5t9cwuoBVtwPAM= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686220633; x=1688812633; h=references:in-reply-to:message-id:date:subject:cc:to:from :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=2VsZ55l/p5N1O/trRd6wo49opUNJKzs0Yf4+iMapfe4=; b=dSmxKyqnGPRjR/O3j9MAsjbtphkuPFI0NBWk5MpsscEARzZLKGPxDUThL5c+lkmQcO 2d0jJWJQYiZHkn2szpgMW9mh7jolb9yg+NSfaZYnMM1lQUZ7dKo1QCDMXGHMo3n3mOcV HWgb0JLOp6KkirxE7MljP0xEGwoAwGD+gmYHnGIiCmqU7ph2wqKmnY/npJSrIefW9BM2 EjcVdPYNWrJYoi5rc7gHRhdFSBtwnMPabS9+iNeXlkeieC2d6TJkNxHk2ZBgOvv26X3Q FhlaxNZp232eSg9p4Vs+nbGnkLexM0av3L4iRoZ4irlupljShf72mrApLcv95/CupwEo yZFA== X-Gm-Message-State: AC+VfDwaiiZrweWxTy/eZqA0hvq4GP9DeuhgH3egRtRc+94FfMyAfc1G Sl33ExOVnX3imt+oBqkqlQawFQ== X-Google-Smtp-Source: ACHHUZ50r/7QJvLvCBCRbpU5EzcNIWy8CFG7QlycDdmhAz/yn+zjQSwVqwJwN21FkuZVEUoxTBMSlA== X-Received: by 2002:a17:903:248:b0:1af:a349:3f0d with SMTP id j8-20020a170903024800b001afa3493f0dmr9554696plh.23.1686220632934; Thu, 08 Jun 2023 03:37:12 -0700 (PDT) Received: from dhcp-10-192-206-197.iig.avagotech.net.net ([192.19.234.250]) by smtp.gmail.com with ESMTPSA id jj11-20020a170903048b00b001a980a23802sm1128510plb.111.2023.06.08.03.37.10 (version=TLS1_2 cipher=ECDHE-ECDSA-AES128-GCM-SHA256 bits=128/128); Thu, 08 Jun 2023 03:37:12 -0700 (PDT) From: Selvin Xavier To: jgg@ziepe.ca, leon@kernel.org Cc: linux-rdma@vger.kernel.org, andrew.gospodarek@broadcom.com, Kashyap Desai , Selvin Xavier Subject: [PATCH for-next 06/17] RDMA/bnxt_re: Avoid the command wait if firmware is inactive Date: Thu, 8 Jun 2023 03:24:57 -0700 Message-Id: <1686219908-11181-7-git-send-email-selvin.xavier@broadcom.com> X-Mailer: git-send-email 2.5.5 In-Reply-To: <1686219908-11181-1-git-send-email-selvin.xavier@broadcom.com> References: <1686219908-11181-1-git-send-email-selvin.xavier@broadcom.com> Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org From: Kashyap Desai Add a check to avoid waiting if driver already detects a FW timeout. Return success for resource destroy in case the device is detached. Add helper function to map timeout error code to success. Signed-off-by: Kashyap Desai Signed-off-by: Selvin Xavier --- drivers/infiniband/hw/bnxt_re/qplib_rcfw.c | 52 +++++++++++++++++++++++++++--- 1 file changed, 48 insertions(+), 4 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c index 3b242cc..3c4f72a 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c @@ -54,9 +54,46 @@ static void bnxt_qplib_service_creq(struct tasklet_struct *t); /** + * bnxt_qplib_map_rc - map return type based on opcode + * @opcode - roce slow path opcode + * + * In some cases like firmware halt is detected, the driver is supposed to + * remap the error code of the timed out command. + * + * It is not safe to assume hardware is really inactive so certain opcodes + * like destroy qp etc are not safe to be returned success, but this function + * will be called when FW already reports a timeout. This would be possible + * only when FW crashes and resets. This will clear all the HW resources. + * + * Returns: + * 0 to communicate success to caller. + * Non zero error code to communicate failure to caller. + */ +static int bnxt_qplib_map_rc(u8 opcode) +{ + switch (opcode) { + case CMDQ_BASE_OPCODE_DESTROY_QP: + case CMDQ_BASE_OPCODE_DESTROY_SRQ: + case CMDQ_BASE_OPCODE_DESTROY_CQ: + case CMDQ_BASE_OPCODE_DEALLOCATE_KEY: + case CMDQ_BASE_OPCODE_DEREGISTER_MR: + case CMDQ_BASE_OPCODE_DELETE_GID: + case CMDQ_BASE_OPCODE_DESTROY_QP1: + case CMDQ_BASE_OPCODE_DESTROY_AH: + case CMDQ_BASE_OPCODE_DEINITIALIZE_FW: + case CMDQ_BASE_OPCODE_MODIFY_ROCE_CC: + case CMDQ_BASE_OPCODE_SET_LINK_AGGR_MODE: + return 0; + default: + return -ETIMEDOUT; + } +} + +/** * __wait_for_resp - Don't hold the cpu context and wait for response * @rcfw - rcfw channel instance of rdev * @cookie - cookie to track the command + * @opcode - rcfw submitted for given opcode * * Wait for command completion in sleepable context. * @@ -64,7 +101,7 @@ static void bnxt_qplib_service_creq(struct tasklet_struct *t); * 0 if command is completed by firmware. * Non zero error code for rest of the case. */ -static int __wait_for_resp(struct bnxt_qplib_rcfw *rcfw, u16 cookie) +static int __wait_for_resp(struct bnxt_qplib_rcfw *rcfw, u16 cookie, u8 opcode) { struct bnxt_qplib_cmdq_ctx *cmdq; u16 cbit; @@ -74,6 +111,9 @@ static int __wait_for_resp(struct bnxt_qplib_rcfw *rcfw, u16 cookie) cbit = cookie % rcfw->cmdq_depth; do { + if (test_bit(ERR_DEVICE_DETACHED, &cmdq->flags)) + return bnxt_qplib_map_rc(opcode); + /* Non zero means command completed */ ret = wait_event_timeout(cmdq->waitq, !test_bit(cbit, cmdq->cmdq_bitmap), @@ -94,6 +134,7 @@ static int __wait_for_resp(struct bnxt_qplib_rcfw *rcfw, u16 cookie) * __block_for_resp - hold the cpu context and wait for response * @rcfw - rcfw channel instance of rdev * @cookie - cookie to track the command + * @opcode - rcfw submitted for given opcode * * This function will hold the cpu (non-sleepable context) and * wait for command completion. Maximum holding interval is 8 second. @@ -102,7 +143,7 @@ static int __wait_for_resp(struct bnxt_qplib_rcfw *rcfw, u16 cookie) * -ETIMEOUT if command is not completed in specific time interval. * 0 if command is completed by firmware. */ -static int __block_for_resp(struct bnxt_qplib_rcfw *rcfw, u16 cookie) +static int __block_for_resp(struct bnxt_qplib_rcfw *rcfw, u16 cookie, u8 opcode) { struct bnxt_qplib_cmdq_ctx *cmdq = &rcfw->cmdq; unsigned long issue_time = 0; @@ -112,6 +153,9 @@ static int __block_for_resp(struct bnxt_qplib_rcfw *rcfw, u16 cookie) issue_time = jiffies; do { + if (test_bit(ERR_DEVICE_DETACHED, &cmdq->flags)) + return bnxt_qplib_map_rc(opcode); + udelay(1); bnxt_qplib_service_creq(&rcfw->creq.creq_tasklet); @@ -267,9 +311,9 @@ int bnxt_qplib_rcfw_send_message(struct bnxt_qplib_rcfw *rcfw, } while (retry_cnt--); if (msg->block) - rc = __block_for_resp(rcfw, cookie); + rc = __block_for_resp(rcfw, cookie, opcode); else - rc = __wait_for_resp(rcfw, cookie); + rc = __wait_for_resp(rcfw, cookie, opcode); if (rc) { /* timed out */ dev_err(&rcfw->pdev->dev, "cmdq[%#x]=%#x timedout (%d)msec\n",