On Wed, Oct 13, 2010 at 05:55:48AM -0700, Joe Perches wrote:
On Wed, 2010-10-13 at 13:33 +0100, Mark Brown wrote:
Are any of the contexts actually ambiguous? I have to say I don't find this useful. If the register I/O code knows anything about volumes I'd say we've probably messed up somewhere.
Looking at the 2 register access routines where volatile_register/vol is used, I'd say those routines aren't useful at all and the checks should be inline instead.
If you check the code again you will notice that these functons are all used in ops structures which presents obvious issues when trying to replace with open coded checks. If we remove these functions then the core won't be able to tell if it should read registers from the hardware or from the cache.
This sort of wrapper routine doesn't serve much of a purpose to me:
static int wm8904_volatile_register(unsigned int reg)
I hope their usefulness is clear given the above.
It's used as the field name for volatility in all the drivers I can remember that use a table to look volatility up in register properties.
I did a grep for vol in sound, I found the uses where it was for volatile and patched them.
Are there others? Perhaps, but I don't think so. Probably not in sound. Was there some other subsystem you meant? Maybe you have examples I missed?
You're missing the "for this" here, I believe. To reiterate what I said originally as far as I can remember every ASoC driver which uses a table to look up volatility uses this naming.