[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <b2f04325-f116-8fb4-b8c8-313e19c18ea4@hisilicon.com>
Date: Fri, 28 Jan 2022 16:06:22 +0800
From: "chenxiang (M)" <chenxiang66@...ilicon.com>
To: John Garry <john.garry@...wei.com>,
"jejb@...ux.ibm.com" <jejb@...ux.ibm.com>,
"martin.petersen@...cle.com" <martin.petersen@...cle.com>,
"artur.paszkiewicz@...el.com" <artur.paszkiewicz@...el.com>,
"jinpu.wang@...ud.ionos.com" <jinpu.wang@...ud.ionos.com>,
"Ajish.Koshy@...rochip.com" <Ajish.Koshy@...rochip.com>
CC: yanaijie <yanaijie@...wei.com>,
"linux-doc@...r.kernel.org" <linux-doc@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-scsi@...r.kernel.org" <linux-scsi@...r.kernel.org>,
Linuxarm <linuxarm@...wei.com>,
"liuqi (BA)" <liuqi115@...wei.com>,
"Viswas.G@...rochip.com" <Viswas.G@...rochip.com>,
"damien.lemoal@...nsource.wdc.com" <damien.lemoal@...nsource.wdc.com>
Subject: Re: [PATCH 06/16] scsi: libsas: Add sas_task.tmf
在 2022/1/28 0:01, John Garry 写道:
> On 27/01/2022 12:55, chenxiang (M) wrote:
>>> - slot->tmf = tmf;
>>> - slot->is_internal = tmf;
>> In kernel 5.17-rc1(with above two lines on), it seems there is a issue
>> for ata disk reset command as tmf = null while it is internal command.
>
> ok, thanks for the notice. Would this fix it:
Yes, it should fix it, and i just notice that it seems solve the issue
in the last patch of this patchset (including this fix in the last patch).
>
> diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c
> b/drivers/scsi/hisi_sas/hisi_sas_main.c
> index a05ec7aece5a..0e12c9329ee5 100644
> --- a/drivers/scsi/hisi_sas/hisi_sas_main.c
> +++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
> @@ -427,8 +427,6 @@ void hisi_sas_task_deliver(struct hisi_hba *hisi_hba,
> cmd_hdr_base = hisi_hba->cmd_hdr[dlvry_queue];
> slot->cmd_hdr = &cmd_hdr_base[dlvry_queue_slot];
>
> - slot->tmf = tmf;
> - slot->is_internal = tmf;
> task->lldd_task = slot;
>
> memset(slot->cmd_hdr, 0, sizeof(struct hisi_sas_cmd_hdr));
> @@ -1380,12 +1378,13 @@ static int hisi_sas_softreset_ata_disk(struct
> domain_device *device)
> struct hisi_hba *hisi_hba = dev_to_hisi_hba(device);
> struct device *dev = hisi_hba->dev;
> int s = sizeof(struct host_to_dev_fis);
> + struct hisi_sas_tmf_task tmf = {};
>
> ata_for_each_link(link, ap, EDGE) {
> int pmp = sata_srst_pmp(link);
>
> hisi_sas_fill_ata_reset_cmd(link->device, 1, pmp, fis);
> - rc = hisi_sas_exec_internal_tmf_task(device, fis, s, NULL);
> + rc = hisi_sas_exec_internal_tmf_task(device, fis, s, &tmf);
> if (rc != TMF_RESP_FUNC_COMPLETE)
> break;
> }
> @@ -1396,7 +1395,7 @@ static int hisi_sas_softreset_ata_disk(struct
> domain_device *device)
>
> hisi_sas_fill_ata_reset_cmd(link->device, 0, pmp, fis);
> rc = hisi_sas_exec_internal_tmf_task(device, fis,
> - s, NULL);
> + s, &tmf);
> if (rc != TMF_RESP_FUNC_COMPLETE)
> dev_err(dev, "ata disk %016llx de-reset failed\n",
> SAS_ADDR(device->sas_addr));
>
>>
>>> task->lldd_task = slot;
>
>
>
> .
>
Powered by blists - more mailing lists