On Thu, Oct 26, 2017 at 09:39:39PM +0200, Takashi Iwai wrote:
IMO, my PCM unbind workaround can be postponed to 4.15, as the unbind is still broken even with this. It can be combined with other unbind fixes.
Also, topology fixes are only potential issues, and no urgent fixes for the bugs we're hitting now. So they can be postponed, too.
For me I'd prioritize getting things into the core over other things (and both fixes are small) but these are the ones you want to drop. I'd say the topology bugs are just straight up bugs, someone with the ability to replace the topology file can readily trigger them (as could disk corruption) and the null pointer dereference is definitely severe enough that we should just fix it.
I guess we could try to drop all the resource leak fixes but several of them are in with other patches in branches.
One Skylake jack name fix is actually wrong. It sets "HDMI/DP, pcm=%d Jack", but the right name is without space, i.e. "HDMI/DP,pcm=%d Jack".
So how did they test that fix then... bah.
All others seem device-specific small fixes, so we can take them, I suppose, as long as they are a kind of stable-targeted fixes. (Basically the fixes with Fixes tag are candidates for stable.)
I'm not insisting people add fixes tags for everything, that way lies madness.