Re: [PATCH 1/2] zram: introduce compressed data writeback

[Barry Song]

On Sat, Nov 29, 2025 at 1:06 AM Sergey Senozhatsky
<senozhatsky@xxxxxxxxxxxx> wrote:
>
> From: Richard Chang <richardycc@xxxxxxxxxx>
>

Hi Richard, Sergey,

Thanks a lot for developing this. For years, people have been looking for
compressed data writeback to reduce I/O, such as compacting multiple compressed
blocks into a single page on block devices. I guess this patchset hasn’t reached
that point yet, right?

> zram stores all written back slots raw, which implies that
> during writeback zram first has to decompress slots (except
> for ZRAM_HUGE slots, which are raw already).  The problem
> with this approach is that not every written back page gets
> read back (either via read() or via page-fault), which means
> that zram basically wastes CPU cycles and battery decompressing
> such slots.  This changes with introduction of decompression

If a page is swapped out and never read again, does that actually indicate
a memory leak in userspace?

So the main benefit of this patch so far is actually avoiding decompression
for "leaked" anon pages, which might still have a pointer but are
never accessed again?

> on demand, in other words decompression on read()/page-fault.

Thanks
Barry