Messages in this thread | | | Date | Thu, 07 Nov 2013 12:10:35 -0500 | From | Rik van Riel <> | Subject | Re: [PATCH] staging: zsmalloc: Ensure handle is never 0 on success |
| |
On 11/07/2013 02:04 AM, Minchan Kim wrote:
> I'm guilty and I have been busy by other stuff. Sorry for that. > Fortunately, I discussed this issue with Hugh in this Linuxcon for a > long time(Thanks Hugh!) he felt zram's block device abstraction is > better design rather than frontswap backend stuff although it's a question > where we put zsmalloc. I will CC Hugh because many of things is related > to swap subsystem and his opinion is really important. > And I discussed it with Rik and he feel positive about zram.
To clarify that, I agree with Minchan that there are certain workloads where zram is probably more appropriate than zswap.
For most of the workloads that I am interested in, zswap will be more interesting, but zram seems to have its own niche, and I certainly do not want to hold back the embedded folks...
| |