At Tue, 22 Apr 2008 19:38:42 +0100, Mark Brown wrote:
On Tue, Apr 22, 2008 at 04:56:21PM +0200, Takashi Iwai wrote:
The problem is that it wasn't passed through mm tree at all.
Oh, I see. I had thought that you already maintained a separate -mm tree. Looking again it seems your linus tree is actually just for backported fixes rather than a general submission to Linus tree.
Hm? No, right now it's identical with the master branch, too.
I must've looked at it at the wrong point in the cycle when figuring these things out.
Actually I have a separate branch on my git tree. The master is for mm tree and for-linus is for Linus. The appropriate patches are copied (and rebased) to for-linus branch once after the merge window is opened. So, it's no no problem at all to have patches that are only for mm.
If the stuff has to cooperate with the changes in other trees, pass it seprately to Andrew (or give me a hint) so that it's put to mm tree rightly at the next time.
I could also ask Andrew to pull the ASoC upstream branch (which is based off the ALSA -mm branch) and keep any dodgy commits there without e-mailing you.
Well, Andrew prefers patches merged into the subsystem tree, so it's better to just mail me as before. Then we have one more review automatically.
My previous comment was about the patch that affects several subsystem / arch changes. In this case, the change isn't only in the subsystem tree. So, we'd need to cope with others anyway. Keeping such a patch in mm tree separately is one option.
thanks,
Takashi