Re: [PATCH v6 1/8] drm: bridge: dw_mipi_dsi: add initial regmap infrastructure

[Adrian Ratiu <adrian.ratiu-ZGY8ohtN/8qB+jHODAdFcQ@public.gmane.org>]

On Thu, 16 Apr 2020, Enric Balletbo Serra <eballetbo-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> 
wrote:
> Hi Adrian, 
> 
> [snip] 
> 
>> >> 
>> >> +static void dw_mipi_dsi_get_hw_version(struct dw_mipi_dsi 
>> >> *dsi) +{ +       regmap_read(dsi->regs, DSI_VERSION, 
>> >> &dsi->hw_version); +       dsi->hw_version &= VERSION; + if 
>> >> (!dsi->hw_version) +               dev_err(dsi->dev, "Failed 
>> >> to read DSI hw version register\n"); 
>> > 
>> > Is this an error that should be ignored? If you can't get the 
>> > HW version, probably, there is something wrong with your 
>> > hardware so, don't you need to return an error? 
>> > 
>> 
>> After thinking a bit more about it, that error should be a 
>> warning. 
>> 
>> I added it because in some cases (for eg. if the peripheral 
>> clock is disabled) the reads can return 0 which is obviously an 
>> invalid version and the bridge will error in the next step when 
>> not finding a layout. 
>> 
> 
> If you'll error anyway, why wait? IIUC at this point the clock 
> *must* be enabled, and if not, something is wrong with the 
> driver, I don't see any advantage on delay the error. do you 
> have a use case where this is called and peripheral clock 
> disabled? 

There should be no real use-case (maybe malfunctioning HW), and we 
could error out here to catch driver bugs ASAP, so I'll go this 
route then :)

Thank you, much appreciated!

>
>> So I'll make this a warning in v7 and explicitely mention that
>> reads version == 0 can be caused by a disabled pclk.
>>
>
> -- Enric
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel