[alsa-devel] [PATCH] ASoC: add xtensa xtfpga I2S interface and platform

Max Filippov jcmvbkbc at gmail.com
Wed Oct 29 15:19:05 CET 2014


On Wed, Oct 29, 2014 at 12:34 AM, Mark Brown <broonie at kernel.org> wrote:
> On Tue, Oct 28, 2014 at 09:11:34PM +0300, Max Filippov wrote:
>> On Tue, Oct 28, 2014 at 8:38 PM, Mark Brown <broonie at kernel.org> wrote:
>
>> > You *really* need to explain how it's supposed to work - right now it's
>> > not at all obvious, like I say the fact that this is a rarely used idiom
>> > is not helping.  For example when we tear down the stream we just assign
>> > the pointer in _stop() but don't bother with a sync until the stream is
>> > closed - why?
>
>> Because we can't wait in stop and syncing is not time critical, we can
>> do it any time before the stream becomes invalid.
>
> To be clear: the important part is that someone reading the code can
> understand what's going on.

Ok, I'll change it.

>> >> hw_params callback can change MCLK rate, so it has to disable and
>> >> enable the clock anyway, and since enable can fail it does not guarantee
>> >> that the clock will be left in the same state. Or should I adjust MCLK rate
>> >> w/o disabling the clock?
>
>> > So yet again: why not just enable the clock only when the device is in
>> > use?  If it's being configured it stands to reason that the device isn't
>> > actively in use...
>
>> Mark, I don't get it, sorry ): My clock synthesizer is I2C controlled, so
>> I can't prepare/unprepare it in the trigger callback. When should I do it?
>
> Runtime PM is the normal way of doing it.

Ok, thanks.

>> >> The level field in the control register is 4 bit wide, so the allowed range of
>> >> level is 0..15. FIFO size is 8192 entries, level = 1 corresponds to
>> >> FIFO size / 2, level = 14 -- to FIFO size of 0. I guess this function
>> >> won't get period_size of 0?
>
>> > So if the IP gets changed and the code gets blown up this could well
>> > explode then...  which doesn't seem entirely unlikely considering this
>> > is a FPGA platform so presumably this is easy to update.  To repeat this
>> > is about clarity and the code looking like it's probably hiding bugs as
>> > much as if the code actually works if you really sit down and study it.
>
>> The calculation does not depend on the actual hardware, but on the
>> constant definitions in the same file. They need to be updated if the
>> hardware changes. I'll try to rewrite it in a cleaner way.
>
> Right, my point is that if someone changes the hardware they'll just
> update the constants and then things will break.

Ok, I've rewritten it in a safer manner.

-- 
Thanks.
-- Max


More information about the Alsa-devel mailing list