[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <830b10df-f697-409d-9ec4-37bc0196cec7@redhat.com>
Date: Wed, 4 Jun 2025 15:18:36 +0200
From: David Hildenbrand <david@...hat.com>
To: Jason Gunthorpe <jgg@...pe.ca>
Cc: Hyesoo Yu <hyesoo.yu@...sung.com>, janghyuck.kim@...sung.com,
zhaoyang.huang@...soc.com, jaewon31.kim@...il.com,
Andrew Morton <akpm@...ux-foundation.org>, John Hubbard
<jhubbard@...dia.com>, Peter Xu <peterx@...hat.com>, linux-mm@...ck.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm: gup: fail migration when no migratable page to
prevent CMA pinning
On 04.06.25 15:11, Jason Gunthorpe wrote:
> On Wed, Jun 04, 2025 at 12:07:21PM +0200, David Hildenbrand wrote:
>
>>> Instead of retrying, this patch explicitly fails the migration attempt
>>> (-EBUSY) if no movable pages are found and unpinnable pages remain.
>>> This avoids infinite loops and gives user a clear signal to retry,
>>> rather then spinning inside kernel.
>>
>> Hmmm, that means we will return EBUSY to the caller. Are all users actually
>> prepared to deal with that?
>
> I don't think anyone is really prepared to deal with GUP temporarily
> failing..
>
> Kernel is expected to sort it out. We tolerated the existing temporary
> failure due to its rarity and lack of a solution only.
>
> Either it can be gup'd or not. There should be no retry.
Right, so -EAGAIN should be used for now such that GUP itself will retry
and can be canceled on signal if required.
--
Cheers,
David / dhildenb
Powered by blists - more mailing lists