dri-devel Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Thierry Reding <thierry.reding@gmail.com>
To: Dave Airlie <airlied@redhat.com>
Cc: Benjamin Gaignard <benjamin.gaignard@linaro.org>,
	Daniel Vetter <daniel.vetter@ffwll.ch>,
	dri-devel@lists.freedesktop.org,
	Tomi Valkeinen <tomi.valkeinen@ti.com>,
	Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
	Russell King <rmk+kernel@arm.linux.org.uk>
Subject: [PATCH v2 7/8] drm: Sanitize DRM_IOCTL_MODE_CREATE_DUMB input
Date: Thu,  6 Nov 2014 16:49:20 +0100	[thread overview]
Message-ID: <1415288961-14464-8-git-send-email-thierry.reding@gmail.com> (raw)
In-Reply-To: <1415288961-14464-1-git-send-email-thierry.reding@gmail.com>

From: Thierry Reding <treding@nvidia.com>

Some drivers treat the pitch and size fields as inputs and will use them
as minima provided by userspace so that they are only overwritten if the
minimal requirements of the driver exceed them.

This can cause strange behaviour when applications don't zero out these
fields, causing whatever was on the stack to be passed to the IOCTL. In
a typical case this would become visible as a failed allocation if the
pitch or size were unusually high. But this could also cause more subtle
bugs like overallocating dumb framebuffers.

To prevent drivers from misusing these values, make the DRM core zero
out the pitch and size fields before passing the structure to the driver
implementation.

While at it, also set the output handle field to zero for good measure,
even though it's less likely to be abused.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>
---
Changes in v2:
- add a comment about why we can't simply reject non-zero output fields
---
 drivers/gpu/drm/drm_crtc.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 0f3c24c0981b..4d7a7699772d 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -4755,6 +4755,16 @@ int drm_mode_create_dumb_ioctl(struct drm_device *dev,
 	if (PAGE_ALIGN(size) == 0)
 		return -EINVAL;
 
+	/*
+	 * handle, pitch and size are output parameters. Zero them out to
+	 * prevent drivers from accidentally using uninitialized data. Since
+	 * not all existing userspace is clearing these fields properly we
+	 * cannot reject IOCTL with garbage in them.
+	 */
+	args->handle = 0;
+	args->pitch = 0;
+	args->size = 0;
+
 	return dev->driver->dumb_create(file_priv, dev, args);
 }
 
-- 
2.1.3

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

  parent reply	other threads:[~2014-11-06 15:50 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-11-05 13:25 [PATCH 0/7] drm: Sanitize DRM_IOCTL_MODE_CREATE_DUMB input Thierry Reding
2014-11-05 13:25 ` [PATCH 1/7] drm/gem: Fix a few kerneldoc typos Thierry Reding
2014-11-05 14:18   ` Daniel Vetter
2014-11-05 13:25 ` [PATCH 2/7] drm/doc: mm: Fix indentation Thierry Reding
2014-11-05 14:19   ` Daniel Vetter
2014-11-05 13:25 ` [PATCH 3/7] drm/doc: Add GEM/CMA helpers to kerneldoc Thierry Reding
2014-11-05 14:34   ` Daniel Vetter
2014-11-05 15:01     ` Thierry Reding
2014-11-05 15:04       ` Daniel Vetter
2014-11-05 15:16         ` Thierry Reding
2014-11-05 13:25 ` [PATCH 4/7] drm/cma: Introduce drm_gem_cma_dumb_create_internal() Thierry Reding
2014-11-05 14:36   ` Daniel Vetter
2014-11-05 13:25 ` [PATCH 5/7] drm/omap: gem: dumb: pitch is an output Thierry Reding
2014-11-05 14:38   ` Daniel Vetter
2014-11-05 13:25 ` [PATCH 6/7] drm/rcar: " Thierry Reding
2014-11-05 14:39   ` Daniel Vetter
2014-11-05 18:47   ` Laurent Pinchart
2014-11-06  0:54     ` Russell King - ARM Linux
2014-11-06 18:17       ` Laurent Pinchart
2014-11-05 13:25 ` [PATCH 7/7] drm: Sanitize DRM_IOCTL_MODE_CREATE_DUMB input Thierry Reding
2014-11-05 14:42   ` Daniel Vetter
2014-11-05 14:24 ` [PATCH 0/7] " Russell King - ARM Linux
2014-11-05 14:45   ` Thierry Reding
2014-11-05 15:01   ` Daniel Vetter
2014-11-06 15:49 ` [PATCH v2 0/8] " Thierry Reding
2014-11-06 15:49   ` [PATCH v2 1/8] drm/gem: Fix a few kerneldoc typos Thierry Reding
2014-11-06 15:49   ` [PATCH v2 2/8] drm/doc: mm: Fix indentation Thierry Reding
2014-11-06 15:49   ` [PATCH v2 3/8] drm/doc: Add GEM/CMA helpers to kerneldoc Thierry Reding
2014-11-06 20:05     ` Daniel Vetter
2014-11-06 15:49   ` [PATCH v2 4/8] drm/cma: Introduce drm_gem_cma_dumb_create_internal() Thierry Reding
2014-11-06 15:49   ` [PATCH v2 5/8] drm/omap: gem: dumb: pitch is an output Thierry Reding
2014-11-06 22:23     ` Rob Clark
2014-11-07  8:02     ` Tomi Valkeinen
2014-11-06 15:49   ` [PATCH v2 6/8] drm/rcar: " Thierry Reding
2014-11-06 15:49   ` Thierry Reding [this message]
2014-11-06 15:49   ` [PATCH v2 8/8] drm/cma: Remove call to drm_gem_free_mmap_offset() Thierry Reding
2014-11-06 20:06     ` Daniel Vetter

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1415288961-14464-8-git-send-email-thierry.reding@gmail.com \
    --to=thierry.reding@gmail.com \
    --cc=airlied@redhat.com \
    --cc=benjamin.gaignard@linaro.org \
    --cc=daniel.vetter@ffwll.ch \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=rmk+kernel@arm.linux.org.uk \
    --cc=tomi.valkeinen@ti.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox