[PATCH 1/4] ASoC: sgtl5000: give it a ramping time before writting

Shawn Guo shawn.guo at linaro.org
Mon Jul 1 09:55:24 EDT 2013


On Mon, Jul 01, 2013 at 12:54:18PM +0200, Marek Vasut wrote:
> Dear Mark Brown,
> 
> > On Mon, Jul 01, 2013 at 04:16:08PM +0800, Shawn Guo wrote:
> > > +	/*
> > > +	 * It seems that sgtl5000 takes a longer time to get the registers
> > > +	 * ready for write than bread.  Let's give it a ramping time before
> > > +	 * the first write goes.
> > > +	 */
> > > +	msleep(50);
> > > +
> > > 
> > >  	/* Ensure sgtl5000 will start with sane register values */
> > >  	ret = sgtl5000_fill_defaults(sgtl5000);
> > >  	if (ret)
> > 
> > This seems like a really odd place to add the sleep - I'd have expected
> > this to be a part of or just after the reset operation.  It's a *really*
> > long sleep too, though if that's what you need that's what you need.
> 
> The delay might also be caused by the I2C driver, I have this in my ToDo to 
> check, but I didn't get to it yet. On the other hand, if subsequent IO does 
> work, then it's unlikely to be the case.

Right, since read is already working there.

Shawn




More information about the linux-arm-kernel mailing list