On Tue, Aug 10, 2021 at 08:54:13AM +0200, Greg Kroah-Hartman wrote: > > Greg, do you *really* want to backport it to 4.4 ? I mean, if nobody > > faces this issue in 4.4 I can see more risks with the fix that without > > for systems with low memory (or manually tuned memory usage). > > I always prefer to merge "known fixes" to the stable kernel trees as > somehow once one person hits a bug, everyone hits it, no matter how long > it was in hiding :) > > The additional memory usage here seems low to me, and we backported the > needed accounting changes there a long time ago, and we know that 4.4.y > is being used for build servers as well as other huge hosting providers. > The "tiny" systems that might be on 4.4 are not really updating > themselves to newer kernels from what I can tell. I agree that someone sticking to 4.4 surely doesn't want to update. > So I would prefer to take this patch, but am always willing to revert it > if someone reports problems with it, In this case I agree, given that such a problem, if any, would be quick to notice. Thanks, Willy