On Sun, Jan 14, 2024 at 2:42 PM Nhat Pham <nphamcs@xxxxxxxxx> wrote: > > On Sun, Jan 14, 2024 at 10:49 AM Yosry Ahmed <yosryahmed@xxxxxxxxxx> wrote: > > > > On Fri, Jan 12, 2024 at 4:38 PM Nhat Pham <nphamcs@xxxxxxxxx> wrote: > > > > > > On Fri, Jan 12, 2024 at 3:37 PM Yosry Ahmed <yosryahmed@xxxxxxxxxx> wrote: > > > > > > > > On Fri, Jan 12, 2024 at 11:42 AM Nhat Pham <nphamcs@xxxxxxxxx> wrote: > > > > > > > > > > On Fri, Jan 12, 2024 at 11:31 AM Yosry Ahmed <yosryahmed@xxxxxxxxxx> wrote: > > > > > > > > > > > > The z3fold compressed pages allocator is not widely used, most users use > > > > > > zsmalloc. The only disadvantage of zsmalloc in comparison is the > > > > > > dependency on MMU, and zbud is a more common option for !MMU as it was > > > > > > the default zswap allocator for a long time. > > > > > > > > > > Johannes and I were chatting about this the other day. We might be > > > > > able to disable certain zsmalloc behavior in the case of !MMU, making > > > > > it available there too. Once that's happened, we can outright remove > > > > > z3fold and zbud, and have one allocator to rule them all? :) > > > > > > > > (Adding Sergey and Minchan for visibility) > > > > > > > > I didn't want to bring up the zsmalloc MMU dependency in this thread > > > > to reduce noise, but that's also what I had in mind. Sergey and I were > > > > also chatting about this the other day :) > > > > > > > > I thought deprecating z3fold is the low hanging fruit. Then, once we > > > > can sort out the MMU dependency in zsmalloc, we can go after zbud as > > > > well. > > > > > > Makes sense to me. Should we do the same thing to zbud? We probably > > > have even less of a case for it, no? > > > > Do you mean declare it as deprecated now? I initially thought that > > would only be appropriate to do after zsmalloc has no dependency on > > MMU, so that we can confidently say zbud has no practical use case. > > Ah, I misread your email. My bad. Personally, I'd like to declare both > (zbud and z3fold) as deprecated. That said, no strong feelings here - > as long as (eventually) we move towards retiring both of them. So my > original ACK still holds. > > Not entirely sure which should we remove first between zbud and z3fold > though. I was under the assumption that z3fold is slightly better, but > that could be my bias for shiny new things showing :) My rationale is that zbud was the default zswap allocator for a long time, so it's the one we should keep for now as it is more likely to have users. That said, I don't know of any users of either zbud or z3fold, and I am fine starting with deprecating either.
