On Thu, Apr 02, 2009 at 02:06:08AM +0200, Rask Ingemann Lambertsen wrote: > On Wed, Apr 01, 2009 at 06:51:36PM +0100, Mark Brown wrote:
> > Ideally these drivers would just get merged; I strongly suspect that it > > will be possible to share much of the code here over many ARM variants. > > As far as I can tell the only difference here is the new tables for the > > S3C24xx and your workaround (for which see below)? > Yes, although I wonder if we want to make the frequency tables depend on > the machine? E.g. memory clocks above 100 MHz are not expected to work on > the GTA02, and we'd preferably not go much lower for performance reasons, so > core frequencies that are fine for a 133 MHz memory bus won't necessarily be > good on a GTA02. Yes, that might be required. I guess ideally we could get constraints through the clock framework or cpufreq would be a real platform device and take platform data. Some of it should fall out of restricting the voltages that can be selected. > > If this is the case would you be OK if I picked this up and merged it > > into my driver/series (which I'll be continuing to push to mainline)? > Yes, that would be fine with me. OK, I'll do so next time I respin - probably after -rc1 (which I expect will be next week some time).
