On Wed, Nov 09, 2011 at 06:00:18PM +0200, Alexander Shishkin wrote: > On Wed, 9 Nov 2011 12:51:13 +0200, Felipe Balbi <balbi@xxxxxx> wrote: > > On Wed, Nov 09, 2011 at 12:48:33PM +0200, Heikki Krogerus wrote: > > > On Wed, Nov 09, 2011 at 12:25:23PM +0200, Felipe Balbi wrote: > > > > good to know that there's more coming. Since we're at the very begining > > > > of the 3.2 cycle, you could take some time and prototype a little more > > > > of what's coming and how things should look like in the long run. I'll > > > > leave this series cook for a little longer before applying. I'll also > > > > wait for updates on the comments I had. > > > > > > OK, let's wait with these. I'm not going to introduce more then very > > > basic support for multiple transceivers if you remember. I'm more > > > interested in introducing the generic OTG state machine. Is this > > > acceptable for you? > > > > > > We are working with the new ChipIdea driver that is meant to replace > > > all the existing drivers in Linux kernel for this same controller. I > > > want to introduce the state machine with that driver. Would it make more > > > sense to send all of these with the OTG sm support for the new driver? > > > > could be... but hey, instead of replacing the in-tree drivers, just > > re-factor that old code. I really don't want to apply 'git rm' patches. > > Clean what's already merged, please. > > I'm not sure if it's really possible to make one driver out of four > without removing stuff. We can try, of course. There will still be > platform code for each of the existing ones in any case. of course you will have to remove code, c'mon, that's not what I said ;-) What I don't want to see is a patch removing all four drivers and another patch adding a magically correct new driver. I want to see incremental patches moving towards consolidating the four different implementations we have in tree and slowly removing the duplicated code. -- balbi
Attachment:
signature.asc
Description: Digital signature
