[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAGsJ_4zLv=HpPL3g085vUaMo8tZZnPZBGT_SfLVCV-10zn+D3Q@mail.gmail.com>
Date: Thu, 7 Nov 2024 10:53:11 +1300
From: Barry Song <21cnbao@...il.com>
To: Andrew Morton <akpm@...ux-foundation.org>,
Kanchana P Sridhar <kanchana.p.sridhar@...el.com>
Cc: Johannes Weiner <hannes@...xchg.org>, linux-mm@...ck.org, linux-kernel@...r.kernel.org,
Barry Song <v-songbaohua@...o.com>, Nhat Pham <nphamcs@...il.com>,
Usama Arif <usamaarif642@...il.com>, Chengming Zhou <chengming.zhou@...ux.dev>,
Yosry Ahmed <yosryahmed@...gle.com>, Hailong Liu <hailong.liu@...o.com>,
David Hildenbrand <david@...hat.com>, Hugh Dickins <hughd@...gle.com>,
Matthew Wilcox <willy@...radead.org>, Shakeel Butt <shakeel.butt@...ux.dev>,
Andi Kleen <ak@...ux.intel.com>, Baolin Wang <baolin.wang@...ux.alibaba.com>,
Chris Li <chrisl@...nel.org>, "Huang, Ying" <ying.huang@...el.com>,
Kairui Song <kasong@...cent.com>, Ryan Roberts <ryan.roberts@....com>
Subject: Re: [PATCH v3] mm: count zeromap read and set for swapout and swapin
On Thu, Nov 7, 2024 at 10:44 AM Andrew Morton <akpm@...ux-foundation.org> wrote:
>
> On Thu, 7 Nov 2024 10:00:47 +1300 Barry Song <21cnbao@...il.com> wrote:
>
> > On Thu, Nov 7, 2024 at 9:42 AM Andrew Morton <akpm@...ux-foundation.org> wrote:
> > >
> > > On Thu, 7 Nov 2024 09:01:14 +1300 Barry Song <21cnbao@...il.com> wrote:
> > >
> > > > Oops, it seems that it depends on Kanchana's 'mm: change count_objcg_event() to
> > > > count_objcg_events() for batch event updates,' which also isn't present in 6.12.
> > > >
> > > > Otherwise, it won't build, as reported here:
> > > > https://lore.kernel.org/linux-mm/CAGsJ_4whD31+Lk0m2uq-o=ygvkRsw1uXcPeqxBONV-RUXkeEzg@mail.gmail.com/
> > >
> > > argh.
> > >
> >
> > Apologies for the inconvenience.
> >
> > > > Hi Andrew,
> > > > What’s the best approach here? Should we include Kanchana's patch that extends
> > > > the nr argument for count_objcg_events() in 6.12-rc as well?
> > >
> > > Let's do the right thing here. I'll drop this patch from mm-hotfixes.
> > > Please send a v4 against Linus mainline fairly soon then I'll redo
> > > Kanchana's series around that.
> >
> > Alright. The question is whether we should integrate Kanchana's 'mm:
> > change count_objcg_event() to count_objcg_events() for batch event
> > updates' into 'mm: count zeromap read and set for swapout and swapin,'
> > or keep it as a separate patch as patch 1/2?
> >
> > I guess integration would be better, as hotfixes may not be ideal for a patch
> > series?
>
> I don't fully understand what you're asking here.
>
> I'm suggesting that you prepare a minimal patch that fixes the bug in
> Linus's kernel. Then we figure out what to do with Kanchana's 6.13-rc1
> material after the bugfix is sorted out.
Kanchana's commit 'mm: change count_objcg_event() to count_objcg_events()'
changes count_objcg_event() to count_objcg_events() and supports
nr_pages more than 1. This is what we need for the minimal patch of
fixing zeromap
as zeromap could be nr_pages > 1 for large folios.
So my question is that, do I combine Kanchana's change into my patch
and send a single patch, or do I send a patch series with 2 patches:
1: Kanchana's mm: change count_objcg_event() to count_objcg_events()
2: mm: count zeromap read and set for swapout and swapin
If we combine them into a single patch, I'll need to incorporate the changes
from 1 into 2. I'm also unsure how to acknowledge Kanchana's contribution
—perhaps mark it as co-developed?
Cc'd Kanchana as well.
Thanks
barry
Powered by blists - more mailing lists