[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <PH0PR04MB741645FA3CC7BE2C24711F269B549@PH0PR04MB7416.namprd04.prod.outlook.com>
Date: Wed, 28 Sep 2022 14:28:57 +0000
From: Johannes Thumshirn <Johannes.Thumshirn@....com>
To: John Garry <john.garry@...wei.com>,
"jejb@...ux.ibm.com" <jejb@...ux.ibm.com>,
"martin.petersen@...cle.com" <martin.petersen@...cle.com>,
"jinpu.wang@...ud.ionos.com" <jinpu.wang@...ud.ionos.com>,
Damien Le Moal <Damien.LeMoal@....com>
CC: "hare@...e.de" <hare@...e.de>,
"linux-scsi@...r.kernel.org" <linux-scsi@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linuxarm@...wei.com" <linuxarm@...wei.com>,
"ipylypiv@...gle.com" <ipylypiv@...gle.com>,
"changyuanl@...gle.com" <changyuanl@...gle.com>,
"hch@....de" <hch@....de>
Subject: Re: [PATCH 1/6] scsi: libsas: Add sas_task_find_rq()
On 28.09.22 15:56, John Garry wrote:
> On 28/09/2022 14:17, Johannes Thumshirn wrote:
>>> +static inline struct request *sas_task_find_rq(struct sas_task *task)
>>> +{
>>> + struct scsi_cmnd *scmd;
>>> +
>>> + if (!task || !task->uldd_task)
>>> + return NULL;
>> Is anyone actually calling sas_task_find_rq with a NULL task?
>
> Yeah, unfortunately. An example - the only one I think - is in the
> pm8001 driver: pm8001_ccb_alloc() function which takes a task pointer
> which may be NULL, and this function calls sas_task_find_rq()
>
>> That doesn't make a lot of sense from an API POV for me, having
>> the only argument allowed to be NULL (and not being a *free()
>> kind of function).
>
> I suppose that I could change to only call sas_task_find_rq() for
> non-NULL sas_task pointers (and remove the task check).
If this is done in only a few drivers I'd go that route TBH.
AFAICS hishi sas doesn't call sas_task_find_rq with a NULL task,
so it's only pm8001
Powered by blists - more mailing lists