Received: by 2002:a05:6358:53a8:b0:117:f937:c515 with SMTP id z40csp3665902rwe; Mon, 17 Apr 2023 01:23:58 -0700 (PDT) X-Google-Smtp-Source: AKy350bRIq2g6sV3BTM2Uuu5R+nMAUuVjV7ypXWf+I6GLFdNTIsWq52gjOZrRfFnlSMgmUwKdUNL X-Received: by 2002:a17:902:dac5:b0:1a2:3108:5cc9 with SMTP id q5-20020a170902dac500b001a231085cc9mr14617885plx.40.1681719837841; Mon, 17 Apr 2023 01:23:57 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1681719837; cv=none; d=google.com; s=arc-20160816; b=BwIKF7nW10oz+prNzI2xQVk/8gcvppjGj6l0dBC60T/4CHZvwhz08jPBkltpGaAp9w sMHlA5sFTIv8D23SzcANm6acdQuisYhLKd/7H8zbBtgKcqYXe3LUamo3VHC89s4sXuAK 1FEEUUiQ7lrHD5bq1/hkTh+L3PTnmYAE+9AJTHeY27BIwibdVkka9bsI7wJdMC3cb09D 7vLgq1edBjVZ5WnQZaaVjt7uO7bkHhfma6lth0q5dVZZ3okHBruFhBoJqaKtBsQ30tRy SRWFSpd/m1wi95o78hu3wOqnirWrPxHAGyPsn6j/OzTOYgJcnsqWUUXohE4KodxtLrne 3hzQ== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:content-transfer-encoding:in-reply-to:from :references:cc:to:content-language:subject:user-agent:mime-version :date:message-id; bh=wqWuwKEkDNpxmkvvO6WUqriNjU9N5KwmDt5hgUD+L0w=; b=KdWwsmRutB0lZ5FwnSqP9VWfbSx7m1qN/CCzVdqHeR41+9HTO8sQ1sJnvONoxbbFp0 MaHhd9RA46AFNfcDJPzm24m7LBYttdCmk/ndnQBKXT1uZE1wuJvLUuinqpU/rjXsKx29 7KPic3GZBckyXn7utdzavdgher80vBxtAJF6VJAs67E6QspPvrrAWafm7u13RKoECW9x IIojYb6Ez3z5bnZ0cMpFCX6iYgT55TS9Kwzu6s/YTqMv9p/cmpIQJEmGYH3+Shf9NY3x q6NdNIClcTIRLBrDkFPOcMpRUMoq7yBvObRVHXWlYMKanf4xP+BTsKl0JGB4CSTL/CLb cnQA== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=fail (p=QUARANTINE sp=QUARANTINE dis=NONE) header.from=huawei.com Return-Path: Received: from out1.vger.email (out1.vger.email. [2620:137:e000::1:20]) by mx.google.com with ESMTP id o1-20020a170902778100b0019a7a67d312si11091444pll.454.2023.04.17.01.23.47; Mon, 17 Apr 2023 01:23:57 -0700 (PDT) Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20; Authentication-Results: mx.google.com; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=fail (p=QUARANTINE sp=QUARANTINE dis=NONE) header.from=huawei.com Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230023AbjDQIVU (ORCPT + 99 others); Mon, 17 Apr 2023 04:21:20 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:37642 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229652AbjDQIVR (ORCPT ); Mon, 17 Apr 2023 04:21:17 -0400 Received: from szxga08-in.huawei.com (szxga08-in.huawei.com [45.249.212.255]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9031626B2; Mon, 17 Apr 2023 01:21:14 -0700 (PDT) Received: from dggpemm500012.china.huawei.com (unknown [172.30.72.56]) by szxga08-in.huawei.com (SkyGuard) with ESMTP id 4Q0Kfh0JT7z17T1m; Mon, 17 Apr 2023 16:17:32 +0800 (CST) Received: from [10.67.101.126] (10.67.101.126) by dggpemm500012.china.huawei.com (7.185.36.89) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2507.23; Mon, 17 Apr 2023 16:18:33 +0800 Message-ID: Date: Mon, 17 Apr 2023 16:18:33 +0800 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101 Thunderbird/91.3.1 Subject: Re: [PATCH] scsi: libsas: set tf to normal in sas_ata_device_link_abort() Content-Language: en-CA To: Jason Yan , , , CC: , , , , References: <20230407035618.25123-1-yangxingui@huawei.com> From: yangxingui In-Reply-To: Content-Type: text/plain; charset="UTF-8"; format=flowed Content-Transfer-Encoding: 8bit X-Originating-IP: [10.67.101.126] X-ClientProxiedBy: dggpemm500014.china.huawei.com (7.185.36.153) To dggpemm500012.china.huawei.com (7.185.36.89) X-CFilter-Loop: Reflected X-Spam-Status: No, score=-5.2 required=5.0 tests=BAYES_00,NICE_REPLY_A, RCVD_IN_DNSWL_MED,RCVD_IN_MSPIKE_H2,SPF_HELO_NONE,SPF_PASS, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on lindbergh.monkeyblade.net Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Jason On 2023/4/14 15:36, Jason Yan wrote: > On 2023/4/7 11:56, Xingui Yang wrote: >> If the disk returns UNC for more than five times within a short >> period, the >> number of retry times for other I/Os may reach scmd->allowed, and the >> default error "Illegal Request" is returned for other I/Os, as follows: >> >> [  273.801770] hisi_sas_v3_hw 0000:b4:02.0: erroneous completion disk >> err dev id=2 sas_addr=0x5000000000000605 CQ hdr: 0x400903 0x20103 0x0 >> 0x80470000 >> [  273.875286] sas: Enter sas_scsi_recover_host busy: 30 failed: 30 >> [  273.879895] sas: trying to find task 0x00000000d9cfc893 >> [  273.879896] sas: sas_scsi_find_task: aborting task 0x00000000d9cfc893 >> [  273.880054] sas: sas_scsi_find_task: task 0x00000000d9cfc893 is done >> [  273.880055] sas: sas_eh_handle_sas_errors: task 0x00000000d9cfc893 >> is done >> [  273.880236] ata6.00: failed command: READ FPDMA QUEUED >> [  273.880238] ata6.00: cmd 60/08:00:59:27:00/00:00:00:00:00/40 tag 22 >> ncq dma 4096 in >>                          res 41/04:00:20:00:00/00:00:00:00:00/00 Emask >> 0x1 (device error) >> [  273.880239] ata6.00: status: { DRDY ERR } >> [  273.880240] ata6.00: error: { ABRT } >> [  273.880241] ata6.00: failed command: READ FPDMA QUEUED >> [  273.880243] ata6.00: cmd 60/90:00:d1:26:00/00:00:00:00:00/40 tag 23 >> ncq dma 73728 in >>                          res 41/40:90:10:27:00/00:00:00:00:00/00 Emask >> 0x409 (media error) >> [  273.880245] ata6.00: status: { DRDY ERR } >> [  273.880246] ata6.00: error: { UNC } >> [  273.880247] ata6.00: failed command: READ FPDMA QUEUED >> [  273.880249] ata6.00: cmd 60/08:00:19:27:00/00:00:00:00:00/40 tag 24 >> ncq dma 4096 in >>                          res 41/04:00:20:00:00/00:00:00:00:00/00 Emask >> 0x1 (device error) >> [  273.880250] ata6.00: status: { DRDY ERR } >> [  273.880251] ata6.00: error: { ABRT } >> [  274.199477] scmd->retries: 3, scmd->allowed: 5 >> [  274.199478] scmd->retries: 3, scmd->allowed: 5 >> [  274.199479] scmd->retries: 3, scmd->allowed: 5 >> [  274.199481] scmd->retries: 3, scmd->allowed: 5 >> [  274.199482] scmd->retries: 3, scmd->allowed: 5 >> [  274.199483] scmd->retries: 2, scmd->allowed: 5 >> [  274.199484] scmd->retries: 3, scmd->allowed: 5 >> [  274.199485] scmd->retries: 3, scmd->allowed: 5 >> [  274.199486] scmd->retries: 5, scmd->allowed: 5 >> [  274.199487] scmd->retries: 2, scmd->allowed: 5 >> [  274.199488] scmd->retries: 2, scmd->allowed: 5 >> [  274.199524] sd 6:0:1:0: [sdb] tag#258 FAILED Result: >> hostbyte=DID_OK driverbyte=DRIVER_SENSE >> [  274.199527] sd 6:0:1:0: [sdb] tag#258 Sense Key : Illegal Request >> [current] >> [  274.199530] sd 6:0:1:0: [sdb] tag#258 Add. Sense: Unaligned write >> command >> [  274.199532] sd 6:0:1:0: [sdb] tag#258 CDB: Read(10) 28 00 00 00 27 >> 59 00 00 08 00 >> [  274.199535] print_req_error: I/O error, dev sdb, sector 10073 >> [  274.199573] sd 6:0:1:0: [sdb] tag#259 FAILED Result: >> hostbyte=DID_OK driverbyte=DRIVER_SENSE >> [  274.199574] sd 6:0:1:0: [sdb] tag#259 Sense Key : Medium Error >> [current] >> [  274.199576] sd 6:0:1:0: [sdb] tag#259 Add. Sense: Unrecovered read >> error - auto reallocate failed >> [  274.199578] sd 6:0:1:0: [sdb] tag#259 CDB: Read(10) 28 00 00 00 26 >> d1 00 00 90 00 >> [  274.199579] print_req_error: I/O error, dev sdb, sector 10000 >> [  274.199608] ata6: EH complete >> [  274.199615] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 30 >> tries: 1 >> >> As mentioned in ata_eh_qc_retry(), if qc->err_mask is zero then increment >> scmd->allowed. So set tf to normal may be better. > > Hi Xingui, > > If we increase scmd->allowed every time, and the device returns UNC for > too many times, will the other IO pending for too long and cause > hungtask? And also the runtime check in scsi_cmd_runtime_exceeced() will > not trigger since cmd->allowed is extended. > Thank you for your reply. In scenarios similar to UNC error, where a disk returns an error through D2H or SDB, no error is reported for other I/Os in the disk. In this case, AHCI will increase the number of retry times for other I/Os, and AHCI may face similar problems you say, but default failures may not be very good for users. In addition, for commands with pass through type, other I/Os are immediately returned with default errors and are not retried, but AHCI only report one single error I/O. Thanks, Xingui > Thanks, > Jason > >> >> Signed-off-by: Xingui Yang >> --- >>   drivers/scsi/libsas/sas_ata.c | 4 ++-- >>   1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/scsi/libsas/sas_ata.c >> b/drivers/scsi/libsas/sas_ata.c >> index 77714a495cbb..f5047e8dcb59 100644 >> --- a/drivers/scsi/libsas/sas_ata.c >> +++ b/drivers/scsi/libsas/sas_ata.c >> @@ -949,8 +949,8 @@ void sas_ata_device_link_abort(struct >> domain_device *device, bool force_reset) >>       unsigned long flags; >>       spin_lock_irqsave(ap->lock, flags); >> -    device->sata_dev.fis[2] = ATA_ERR | ATA_DRDY; /* tf status */ >> -    device->sata_dev.fis[3] = ATA_ABORTED; /* tf error */ >> +    device->sata_dev.fis[2] = ATA_DRDY; /* tf status */ >> +    device->sata_dev.fis[3] = 0;        /* tf error */ >>       link->eh_info.err_mask |= AC_ERR_DEV; >>       if (force_reset) >> > .