Re: [PATCH v2 19/20] scsi: ufs: Implement polling support

From: Bart Van Assche <bvanassche@acm.org>
To: Bean Huo <huobean@gmail.com>,
	"Martin K . Petersen" <martin.petersen@oracle.com>
Cc: Jaegeuk Kim <jaegeuk@kernel.org>,
	Adrian Hunter <adrian.hunter@intel.com>,
	linux-scsi@vger.kernel.org,
	"James E.J. Bottomley" <jejb@linux.ibm.com>,
	Bean Huo <beanhuo@micron.com>, Can Guo <cang@codeaurora.org>,
	Avri Altman <avri.altman@wdc.com>,
	Stanley Chu <stanley.chu@mediatek.com>,
	Asutosh Das <asutoshd@codeaurora.org>
Subject: Re: [PATCH v2 19/20] scsi: ufs: Implement polling support
Date: Tue, 30 Nov 2021 09:37:46 -0800	[thread overview]
Message-ID: <ca60e62b-d310-1f97-d61e-d0ba8bdd420b@acm.org> (raw)
In-Reply-To: <e0dc15c742c2f626a7149c3c44d53493fe1a9a44.camel@gmail.com>

On 11/30/21 12:43 AM, Bean Huo wrote:
> On Fri, 2021-11-19 at 11:57 -0800, Bart Van Assche wrote:
>> +	return IRQ_HANDLED;
>>   }
>>   
> 
> ufshcd_transfer_req_compl() will never return IRQ_NONE,
> but ufshcd_intr() needs this return to check the fake interrupt.

I don't think that it is possible to implement polling and detect
spurious interrupts without affecting performance negatively. Hence
the choice to never return IRQ_NONE. Is spurious interrupt detection
that important or is this only required for debugging host controllers?
I consider helping with hardware debugging as out of scope for the UFS
driver.

>> @@ -9437,6 +9481,7 @@ int ufshcd_alloc_host(struct device *dev,
>> struct ufs_hba **hba_handle)
>>   		err = -ENOMEM;
>>   		goto out_error;
>>   	}
>> +	host->nr_maps = 3;
> 
> I don't understand why not directly uses HCTX_MAX_TYPES, 3 is already
> maximum.

If new map types would be introduced in the future, we still need 3 maps.
Hence the choice for '3' instead of HCTX_MAX_TYPES.

Thanks,

Bart.