[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20221201101417.f6qm4v3m7ibh3l72@CAB-WSD-L081021>
Date: Thu, 1 Dec 2022 13:14:17 +0300
From: Dmitry Rokosov <ddrokosov@...rdevices.ru>
To: Sergey Senozhatsky <senozhatsky@...omium.org>
CC: Aleksey Romanov <AVRomanov@...rdevices.ru>,
Johannes Weiner <hannes@...xchg.org>,
"minchan@...nel.org" <minchan@...nel.org>,
"ngupta@...are.org" <ngupta@...are.org>,
"akpm@...ux-foundation.org" <akpm@...ux-foundation.org>,
"linux-mm@...ck.org" <linux-mm@...ck.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
kernel <kernel@...rdevices.ru>
Subject: Re: [RFC PATCH v1 0/4] Introduce merge identical pages mechanism
Hello Sergey,
Hope you are doing well. Really sorry for the ping.
Did you get a chance to see the patch series, my questions, and
thoughts?
On Wed, Nov 23, 2022 at 11:53:06AM +0300, Dmitry Rokosov wrote:
> Hello Sergey,
>
> Thank you for your quick and detailed support! Here is my two cents
> below.
>
> On Wed, Nov 23, 2022 at 01:13:55PM +0900, Sergey Senozhatsky wrote:
> > On (22/11/22 12:14), Aleksey Romanov wrote:
> > > > IIRC that was patent in question:
> > > >
> > > > https://patentimages.storage.googleapis.com/e2/66/9e/0ddbfae5c182ac/US9977598.pdf
> > >
> > > I think the patent is talking about "mapping the virtual address" (like
> > > in KSM). But zram works with the "handle" abstraction, which is a boxed
> > > pointer to the required object. I think my implementation and the patent
> > > is slightly different.
> > >
> > > Also, the patent speaks of "compressing" pages. In this case, we can add
> > > zs_merge() function (like zs_compact()), that is, remove the merge logic
> > > at the allocator level. zsmalloc doesn't say anything about what objects
> > > it can work with. Implementation at the zsmalloc level is possible,
> > > though more complicated that at the zram level.
> > >
> > > I believe that we can implement at least one of the options I proposed.
> > >
> > > What do you think?
> >
> > Oh, yeah, I'm not saying that we cannot have something like that
> > in zram/zsmalloc, just wanted to give some historical retrospective
> > on this and point at some implementation details that should be
> > considered.
>
> It's a very curious situation, I would say. I'm not so familiar with US
> patent law, but I suppose it should be based on some keywords and
> algorithms.
>
> If we speak in terms of algorithm Alexey patch is different a little bit
> from suggested in the patent paper. If we care about keywords, I think by
> moving Alexey same page merging algorithm to zsmalloc we lose
> "compressing" keyword, because zsmalloc operates with "objects" only,
> doesn't matter if they are compressed or not.
>
> Anyway, could you please suggest who can help to understand if it's safe
> to use such same page merging algorithm in the upstream or not?
> Maybe we can ask Linux Foundation lawyers to help us, just a guess.
> I'm sure we shouldn't decline helpful features and optimization without
> complete certainty about all restrictions.
--
Thank you,
Dmitry
Powered by blists - more mailing lists