Received-SPF: pass (google.com: domain of linux-kernel+bounces-207772-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Date: Mon, 10 Jun 2024 07:26:57 +0000
From: "Pankaj Raghav (Samsung)" <kernel@pankajraghav.com>
To: Zi Yan <ziy@nvidia.com>
Cc: willy@infradead.org, david@fromorbit.com, djwong@kernel.org,
	chandan.babu@oracle.com, brauner@kernel.org,
	akpm@linux-foundation.org, mcgrof@kernel.org, linux-mm@kvack.org,
	hare@suse.de, linux-kernel@vger.kernel.org,
	yang@os.amperecomputing.com, linux-xfs@vger.kernel.org,
	p.raghav@samsung.com, linux-fsdevel@vger.kernel.org, hch@lst.de,
	gost.dev@samsung.com, cl@os.amperecomputing.com,
	john.g.garry@oracle.com
Subject: Re: [PATCH v7 05/11] mm: split a folio in minimum folio order chunks
Message-ID: <20240610072657.erdzkedvbzj3gohu@quentin>
References: <20240607145902.1137853-1-kernel@pankajraghav.com>
 <20240607145902.1137853-6-kernel@pankajraghav.com>
 <75CCE180-EC90-4BDC-B5D8-0ED1B710BE49@nvidia.com>
 <20240607203026.zj3akxdjeykchnnf@quentin>
 <45567EBA-5856-4BBC-8C02-EAE03A676B94@nvidia.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <45567EBA-5856-4BBC-8C02-EAE03A676B94@nvidia.com>

On Fri, Jun 07, 2024 at 04:51:04PM -0400, Zi Yan wrote:
> On 7 Jun 2024, at 16:30, Pankaj Raghav (Samsung) wrote:
> >>> +		if (!folio->mapping) {
> >>> +			count_vm_event(THP_SPLIT_PAGE_FAILED);
> >>
> >> You should only increase this counter when the input folio is a THP, namely
> >> folio_test_pmd_mappable(folio) is true. For other large folios, we will
> >> need a separate counter. Something like MTHP_STAT_FILE_SPLIT_FAILED.
> >> See enum mthp_stat_item in include/linux/huge_mm.h.
> >>
> > Hmm, but we don't have mTHP support for non-anonymous memory right? In
> > that case it won't be applicable for file backed memory?
> 
> Large folio support in page cache precedes mTHP (large anonymous folio),
> thanks to willy's work. mTHP is more like a subset of large folio.
> There is no specific counters for page cache large folio. If you think
> it is worth tracking folios with orders between 0 and 9 (exclusive),
> you can add counters. Matthew, what is your take on this?

Got it. I think this is out of scope for this series but something we
could consider as a future enhancement?

In any case, we need to decide whether we need to count truncation as a 
VM event or not.