On 2022/06/07 5:38, Sergey Shtylyov wrote:
> Hello!
>
> On 6/6/22 5:42 AM, Damien Le Moal wrote:
>
>>> The {dma|pio|xfer}_mode sysfs files are incorrectly handled by the
>>> ata_bitfield_name_match() macro which leads to reading such kind of
>>> nonsense from them:
>>>
>>> $ cat /sys/class/ata_device/dev3.0/pio_mode
>>> XFER_UDMA_7, XFER_UDMA_6, XFER_UDMA_5, XFER_UDMA_4, XFER_MW_DMA_4,
>>> XFER_PIO_6, XFER_PIO_5, XFER_PIO_4, XFER_PIO_3, XFER_PIO_2, XFER_PIO_1,
>>> XFER_PIO_0
>>>
>>> Using the correct ata_bitfield_name_search() macro fixes that:
>>>
>>> $ cat /sys/class/ata_device/dev3.0/pio_mode
>>> XFER_PIO_4
>>
>> Looks good, but Documentation/ABI/testing/sysfs-ata says:
>
> Completely forgot that the sysfs files are documented as ABIs... :-(
> Hm, shouldn't that file be added to the libata's entry in MAINTAINERS?
>
>> pio_mode: (RO) Transfer modes supported by the device when
>> in PIO mode. Mostly used by PATA device.
>>
>> xfer_mode: (RO) Current transfer mode
>>
>> dma_mode: (RO) Transfer modes supported by the device when
>> in DMA mode. Mostly used by PATA device.
>>
>> which seems incorrect/badly worded for pio_mode and dma_mode. Since these
>> 2 sysfs attributes do not actually device the pio mask (list of supported
>
> Device?
advertise :)
>
>> pio modes) but the pio mode that will be used for that device, we should
>> reword, no ?
>
> Yes, of course. :-)
>
>> What about:
>>
>> pio_mode: (RO) Transfer mode used by the device when
>> in PIO mode. Mostly used by PATA device.
>>
>> xfer_mode: (RO) Current transfer mode
>>
>> dma_mode: (RO) Transfer mode used by the device when
>> in DMA mode. Mostly used by PATA device.
>
> Sounds quite tautological... :-)
> What about:
>
> {dma|pio}_mode: (RO) {DMA|PIO} transfer mode used by the device.
> Mostly used by PATA devices.
>
> I think this should be done in the same patch. Or would you prefer 2 patches?
Let's do 2 patches. Not sure if you can find a fixes tag for the doc update
though. But we should not aggregate the 2 attributes as you did. These doc files
have a defined format and may not be happy with that merged syntax.
>
> [...]
>
> MBR, Sergey
--
Damien Le Moal
Western Digital Research
