linux-kernel - Re: [Intel-gfx] [PATCH] drm/i915: Selectively enable self-reclaim

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] [day] [month] [year] [list]

Message-ID: <AANLkTin8sbiCQCOUatWAadfLBKuOXUQHKAH1Bs0aTztp@mail.gmail.com>
Date:	Sun, 18 Jul 2010 09:59:28 -0700
From:	Linus Torvalds <torvalds@...ux-foundation.org>
To:	"M. Vefa Bicakci" <bicave@...eronline.com>
Cc:	Dave Airlie <airlied@...il.com>,
	Chris Wilson <chris@...is-wilson.co.uk>, earny@...4u.de,
	Roman Jarosz <kedgedev@...il.com>,
	intel-gfx@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
	jcnengel@...glemail.com,
	"A. Boulan" <arnaud.boulan@...ertysurf.fr>,
	Hugh Dickins <hugh.dickins@...cali.co.uk>,
	Pekka Enberg <penberg@...helsinki.fi>,
	A Rojas <nqn1976list@...il.com>,
	KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>,
	rientjes@...gle.com, michael@...nelt.co.at, stable@...nel.org
Subject: Re: [Intel-gfx] [PATCH] drm/i915: Selectively enable self-reclaim

On Sun, Jul 18, 2010 at 7:27 AM, M. Vefa Bicakci <bicave@...eronline.com> wrote:
>
> After hours of testing I came up with the following result: We need
> to have the __GFP_RECLAIMABLE flag in addition to GFP_HIGHUSER.

Thanks for the extensive testing, and I'm committing the one-liner to
add it, and cc'ing it to stable. I'm pretty certain that there is
something overly fragile in the i915 driver that this flag makes so
much of a difference, but at the same time I'm actually happy that
it's that reclaimable flag, because at least that one was always the
"conceptually makes sense" one.

So I suspected it would be some low-memory issue and the flag that
woudl turn out to matter would be the NOMEMALLOC one, but I'm happy to
have been wrong. Adding __GFP_RECLAIMABLE is sane, although I really
would like to understand why the i915 driver apparently cares so
deeply about the allocation/freeing patterns.

But whatever. Thanks again for being such a thorough tester,

                        Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/