On Thu, Feb 01, 2018 at 01:25:20PM +0100, Andrew Jones wrote: > On Thu, Feb 01, 2018 at 11:48:31AM +0100, Christoffer Dall wrote: > > On Thu, Feb 01, 2018 at 11:42:22AM +0100, Andrew Jones wrote: > > > On Thu, Feb 01, 2018 at 10:59:54AM +0100, Christoffer Dall wrote: > > > > On Thu, Feb 1, 2018 at 10:33 AM, Ard Biesheuvel > > > > > If you could use dirty page tracking to only perform the cache > > > > > invalidation when the framebuffer memory has been updated, you can at > > > > > least limit the impact to cases where the framebuffer is actually > > > > > used, rather than sitting idle with a nice wallpaper image. > > > > > > Yes, this is the exact approach I took back when I experimented with > > > this. I must have screwed up my PoC in some way (like using the gcc > > > builtin), because it wasn't working for me... > > > > > Does that mean you have some code you feel like reviving and use to send > > out an RFC based on? ;) > > > > I don't usually delete things, but I do do some sort of copy on use > from old homedirs to new ones from time to time, letting stuff that > gets really old completely drop at some point. I might be able to > find the old work, or just hack a quick and ugly version from my > memory to share - hopefully whatever I did to cause it to fail last > time has been forgotten :-) >
If you have bandwidth to look at this that's great. Otherwise let us know, and I'll see if I can devote some time to it. Thanks, -Christoffer