Messages in this thread | | | Date | Wed, 30 Jan 2013 10:57:34 +0800 | From | Mark Brown <> | Subject | Re: [GIT PULL][FOR REVIEW] Regulators: Bring the AB8500 into the 20th century |
| |
On Tue, Jan 29, 2013 at 04:41:31PM +0000, Lee Jones wrote:
> If you'd like me to sent the patches to the MLs for a better review > platform I can do, but I'd like to get permission before I go and > start upsetting people.
> This branch contains 111 patches and has already gone though multiple > cycles of fixing-up, squashing, splitting and revert removal.
So, some highlights from a quick scan through:
- Most of the OF stuff needs to go, OF isn't just a dumping ground for typing in all the platform data with no thought about abstraction. Things like -input-microvolt for example, and the mode switching needs to be completely reworked.
- Send the arch/arm stuff separately, supplies with NULL for the device are generally not a good sign though.
- The -debug code looks like it's all NIH stuff, large chunks of it are reproducing stuff that's already there, it looks to have copies of all the data defining the regulators that must be duplicated in the regulator drivers themselves (which isn't a triumph) and the overwhelming bulk of it looks totally generic so it really shouldn't be in a driver.
- I can't help but think this would get a lot less verbose with regmap, there's a lot of functions that . The no-bus support just merged would make that a lot easier.
- dbx500-prcmu looks worrying, it appears to be replicating bits of the core regulator API for some unclear reason.
So probably no point in posting the series as-is. I'd recommend starting off by pulling out the core feature additions to the regulator driver and the -ext driver, that should make things a lot smaller. [unhandled content-type:application/pgp-signature] | |