Thank you Chris, those are excellent comments. I am replacing "memory pressure" with "memory usage". I had used "pressure" because "usage" is vague and here we specifically refer to non-reclaimable pages. But it's true that "pressure" has a different, specific meaning for the mm, so "usage" is better. The example should clarify the problem. I am also removing the workaround. I thought it may be useful, but as you said, it's quite debatable that it belongs here. On Fri, Dec 27, 2019 at 4:25 PM Chris Down <chris@xxxxxxxxxxxxxx> wrote: > > Oh, and: > > Luigi Semenzato writes: > >+system to be present for the underlying CPU architecture. Additionally, the > >+current implementation can enter the hibernation state only when memory > >+pressure is low (see "Limitations" below). > > This doesn't seem to make much sense to me. You claim it only works when > pressure is low, but then suggest people raise pressure to fix it. > > I suppose you mean "only when memory *usage* is low"? I mean, if memory > pressure is low, then it's likely to result in these cold pages sitting around > in the first place.
