linux-kernel - Re: [PATCH v6 1/7] drm/tilcdc: ensure nonatomic iowrite64 is not used

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: <2b41b02c-07e3-b927-f276-a29480f94285@ti.com>
Date:   Fri, 4 Aug 2017 16:03:12 +0300
From:   Tomi Valkeinen <tomi.valkeinen@...com>
To:     Logan Gunthorpe <logang@...tatee.com>,
        <linux-kernel@...r.kernel.org>, <linux-arch@...r.kernel.org>,
        <linux-ntb@...glegroups.com>, <linux-crypto@...r.kernel.org>,
        Jyri Sarha <jsarha@...com>
CC:     Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Andy Shevchenko <andy.shevchenko@...il.com>,
        Horia Geantă <horia.geanta@....com>,
        Stephen Bates <sbates@...thlin.com>,
        David Airlie <airlied@...ux.ie>
Subject: Re: [PATCH v6 1/7] drm/tilcdc: ensure nonatomic iowrite64 is not used

On 03/08/17 21:30, Logan Gunthorpe wrote:
> Add a check to ensure iowrite64 is only used if it is atomic.
> 
> It was decided in [1] that the tilcdc driver should not be using an
> atomic operation (so it was left out of this patchset). However, it turns
> out that through the drm code, a nonatomic header is actually included:
> 
> include/linux/io-64-nonatomic-lo-hi.h
> is included from include/drm/drm_os_linux.h:9:0,
>             from include/drm/drmP.h:74,
>             from include/drm/drm_modeset_helper.h:26,
>             from include/drm/drm_atomic_helper.h:33,
>             from drivers/gpu/drm/tilcdc/tilcdc_crtc.c:19:
> 
> And thus, without this change, this patchset would inadvertantly
> change the behaviour of the tilcdc driver.

I haven't really followed the discussion on this, but if the tilcdc's
use of iowrite64 causes (real) problems/complications elsewhere, I think
we could drop it.

The problem is that the HW has a race issue, and the two registers in
question should be written as close to each other as possible. We
thought a single 64bit write, writing to both registers in one go, would
improve that slightly, compared to two 32 bit writes.

Jyri, correct me if I'm wrong, but we have no proof that it actually
helps, and it might be that even if it helps, the difference is
theoretical. Probably if we ensure the irqs are off when we do two 32
bit writes, we're already close enough to the optimal case.

 Tomi

Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)