(I took this discussion off linux-arm-kernel because I think the majority on that list doesn't really care about this stuff.)
On Thu, Feb 9, 2012 at 1:19 PM, Rickard Andersson rickard.andersson@stericsson.com wrote:
it not clear how db8500-prcmu.c will be handled in the future, db8500-prcmu.c should most likely be split into several parts where one part only handles the communication with the PRCMU firmware.
Yes, I can answer so much: I am to author a letter to Greg (the device driver architecture maintainer) about "system controllers" and where to really put them.
There are maybe three pieces to the PRCMU: (1) PRCMU registers (MFD) (2) PRCMU firmware (MFD?) (3) System control (new subsystem? back to arch/arm/*?)
And then there are all subdrivers using it, like cpuidle, cpufreq, watchdog, hwmon and maybe a few more, which would use .h files for any/some of the above.
So it really needs a good structure and AFAIK Mattias is working on this.
As for merging this one patch I have one question mainly, and that is whether it was done on top of the patches I have pending to bring the mainline PRCMU code upto (one) date and that have already been submitted? (8 patches, cover letter says "DB8500 PRCMU update")
That set brings the current code up to where "internal development" i.e. http://www.igloocommunity.org/gitweb/?p=kernel/igloo-kernel.git%3Ba=shortlog... was at one point. (Now it's already progressed a bit I think...)
If we agree on the function signatures it doesn't matter much where they end up right now if you absolutely need it to go in, it will cause some churn but we can always move them about later, but the semantics for them are really crucial. (In the end it's not me but Samuel Ortiz who merges MFD patches BTW.)
Yours, Linus Walleij