linux-kernel - Re: [RFC 0/1] Convert is_migrate_isolate_page() to is_migrate_isolate

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aBpho6pICFz-0Mbw@lg>
Date: Tue, 6 May 2025 12:23:15 -0700
From: Fan Ni <nifan.cxl@...il.com>
To: Matthew Wilcox <willy@...radead.org>
Cc: nifan.cxl@...il.com, muchun.song@...ux.dev, osalvador@...e.de,
	mcgrof@...nel.org, a.manzanares@...sung.com, dave@...olabs.net,
	akpm@...ux-foundation.org, david@...hat.com, linux-mm@...ck.org,
	linux-kernel@...r.kernel.org
Subject: Re: [RFC 0/1] Convert is_migrate_isolate_page() to
 is_migrate_isolate_folio()

On Tue, May 06, 2025 at 08:08:55PM +0100, Matthew Wilcox wrote:
> On Tue, May 06, 2025 at 11:38:28AM -0700, nifan.cxl@...il.com wrote:
> > From: Fan Ni <fan.ni@...sung.com>
> > 
> > Sending out this patch per Matthew Wilcox's suggestion 
> > that we need to convert is_migrate_isolate_page() to use folio
> > https://lore.kernel.org/linux-mm/Z_XmUrbxKtYmzmJ6@casper.infradead.org/
> 
> That's not what I said!
> 
> This is what I said:
> > >  
> > > -		if (is_migrate_isolate_page(&folio->page))
> > > +		if (is_migrate_isolate_page(folio_page(folio, 0)))
> > >  			continue;
> >
> > I think we need an is_migrate_isolate_folio() instead of this.
> 
> > However, when looking into the code, I have noticed that among the uers
> > of is_migrate_isolate_page(), in most cases the page passed in is from a 
> > a pageblock. 
> > I am not sure how we should proceed with these cases.
> > Should we deal with pageblock or just leave it as it is and only do the page
> > to folio conversion for the pages within?
> 
> Neither.  Add a folio_test_migrate_isolate() in addition to
> is_migrate_isolate_page().  Don't force a conversion as it's a
> legitimate question to ask of pages as well as of folios.
> And some of the pages you want to ask it of may well not be part of
> folios (they may be part of a slab or some other memdesc).

Oh. I misunderstood "we need ... instead of .." :-(.
Thanks for the the clarification.

Another separate question.

We have a free_frozen_pages(page, order), which have two types of users
1) head page and order directly from a struct folio; or
2) page pointer that does not neccesarily be the head page of a
folio and order that may not be directly related to a folio;

Does it make sense to introduce a dedicate function like
free_frozen_folio(struct folio *folio) to handle case 1)?

Fan



-- 
Fan Ni