Received-SPF: pass (google.com: domain of linux-kernel+bounces-190759-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Mon, 27 May 2024 15:54:36 +0200
From: Oscar Salvador <osalvador@suse.de>
To: David Hildenbrand <dhildenb@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>, linux-kernel@vger.kernel.org,
	linux-mm@kvack.org, Muchun Song <muchun.song@linux.dev>,
	Vishal Moola <vishal.moola@gmail.com>
Subject: Re: [PATCH] mm/hugetlb: Move vmf_anon_prepare upfront in hugetlb_wp
Message-ID: <ZlSQnLvqIILiayiW@localhost.localdomain>
References: <20240521073446.23185-1-osalvador@suse.de>
 <ZlRKI_tJ2CYhmekw@localhost.localdomain>
 <9f3371fa-0d46-4e93-a00b-0f5cb30b58bb@redhat.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <9f3371fa-0d46-4e93-a00b-0f5cb30b58bb@redhat.com>

On Mon, May 27, 2024 at 03:17:01PM +0200, David Hildenbrand wrote:
> gup_must_unshare() without a VMA is only used for GUP-fast. Before
> triggering a page fault we always fallback to GUP-slow first, where we have
> a VMA.

I see, thanks for explaining!

> IMHO we better not make assumptions that hugetlb_wp() will always already
> have an anon VMA.

Fine by me, I just wanted to have the picture more clear, because I could not
see how we could end up in such situation, but better be safe and than sorry.

@Andrew: This fix is to avoid an issue like [1], but for hugetlb_wp().

[1] https://lore.kernel.org/linux-mm/000000000000daf1e10615e64dcb@google.com/T/
 

-- 
Oscar Salvador
SUSE Labs