Received-SPF: pass (google.com: domain of linux-kernel+bounces-211769-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Message-ID: <68b4dbc5-18ee-4dc7-85da-dd420df9bf16@gmail.com>
Date: Wed, 12 Jun 2024 23:12:56 +0800
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v5 2/3] scsi: scsi_error: Fix wrong statistic when print
 error info
To: Hannes Reinecke <hare@suse.de>,
 "James E . J . Bottomley" <James.Bottomley@HansenPartnership.com>,
 "Martin K . Petersen" <martin.petersen@oracle.com>,
 linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org
References: <20240605091731.3111195-1-haowenchao22@gmail.com>
 <20240605091731.3111195-3-haowenchao22@gmail.com>
 <15f0fb9b-3b30-413d-9f30-81c246b6bae1@suse.de>
Content-Language: en-US
From: Wenchao Hao <haowenchao22@gmail.com>
In-Reply-To: <15f0fb9b-3b30-413d-9f30-81c246b6bae1@suse.de>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

On 6/12/24 4:34 PM, Hannes Reinecke wrote:
> On 6/5/24 11:17, Wenchao Hao wrote:
>> shost_for_each_device() would skip devices which is in progress of
>> removing, so commands of these devices would be ignored in
>> scsi_eh_prt_fail_stats().
>>
>> Fix this issue by using shost_for_each_device_include_deleted()
>> to iterate devices in scsi_eh_prt_fail_stats().
>>
>> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
>> Signed-off-by: Wenchao Hao <haowenchao22@gmail.com>
>> ---
>>   drivers/scsi/scsi_error.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c
>> index 612489afe8d2..a61fd8af3b1f 100644
>> --- a/drivers/scsi/scsi_error.c
>> +++ b/drivers/scsi/scsi_error.c
>> @@ -409,7 +409,7 @@ static inline void scsi_eh_prt_fail_stats(struct Scsi_Host *shost,
>>       int cmd_cancel = 0;
>>       int devices_failed = 0;
>>   -    shost_for_each_device(sdev, shost) {
>> +    shost_for_each_device_include_deleted(sdev, shost) {
>>           list_for_each_entry(scmd, work_q, eh_entry) {
>>               if (scmd->device == sdev) {
>>                   ++total_failures;
> 
> That is wrong. We should rather add a failure counter to the SCSI host, and have the scsi device increase it every time a failure occurs.
> Then we can avoid this loop completely.
> 

This function would print the total failure commands and the number of device like following:

scsi host4: Total of 3 commands on 2 devices require eh work

Just add a failure counter to the SCSI host can not record the number of devices.

> Cheers,
> 
> Hannes