-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 11/06/2013 11:44 AM, Vaibhav Bedia wrote:
On Wed, Nov 6, 2013 at 8:45 AM, Lokesh Vutla lokeshvutla@ti.com wrote:
On Wednesday 06 November 2013 06:27 PM, Vaibhav Bedia wrote:
[snip]
Are the steps really different enough to warrant a new file? Can't the changes be handled properly in the code? How has this been handled in OMAPx where DDR3 and LPDDR both are supported?
Initially Tom also suggested not to use a new file. I tried with not to add a new file, but I ended up with many #ifdefs. EMIF is new IP(reused from OMAP5) very different from AM33xx EMIF IP. So to make things more cleaner I had to use a new file..
It really looks a step backward. The new IP should be an update to the old version and not just a complete overhaul of the programming model that folks are familiar with.
It sounds like we need to re-think the EMIF code here since it's not all THAT different between the OMAP parts, the am33xx parts, the am43xx parts and the ti81xx parts (and even the am35xx parts, but I'm OK setting that aside).
And no, I'm not 100% happy with the OMAP code either, but that's mainly the bits where we say OMAP rev FOO means memory chip BAR.
- -- Tom