linux-kernel - Re: [PATCH] drm/nouveau: check ioctl command codes better

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: <DB9ET7AG2N2D.YCVQRF9398V2@kernel.org>
Date: Fri, 11 Jul 2025 19:41:56 +0200
From: "Danilo Krummrich" <dakr@...nel.org>
To: "Arnd Bergmann" <arnd@...nel.org>
Cc: "Lyude Paul" <lyude@...hat.com>, "David Airlie" <airlied@...il.com>,
 "Simona Vetter" <simona@...ll.ch>, "Ben Skeggs" <bskeggs@...hat.com>, "Arnd
 Bergmann" <arnd@...db.de>, "Ben Skeggs" <bskeggs@...dia.com>, "Timur Tabi"
 <ttabi@...dia.com>, "Dave Airlie" <airlied@...hat.com>, "Thomas Zimmermann"
 <tzimmermann@...e.de>, <dri-devel@...ts.freedesktop.org>,
 <nouveau@...ts.freedesktop.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] drm/nouveau: check ioctl command codes better

On Fri Jul 11, 2025 at 9:24 AM CEST, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@...db.de>
>
> nouveau_drm_ioctl() only checks the _IOC_NR() bits in the DRM_NOUVEAU_NVIF
> command, but ignores the type and direction bits, so any command with
> '7' in the low eight bits gets passed into nouveau_abi16_ioctl() instead
> of drm_ioctl().
>
> Check for all the bits except the size that is handled inside of the handler.
>
> Fixes: 27111a23d01c ("drm/nouveau: expose the full object/event interfaces to userspace")
> Signed-off-by: Arnd Bergmann <arnd@...db.de>
> ---
>  drivers/gpu/drm/nouveau/nouveau_drm.c | 11 +++++------
>  1 file changed, 5 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
> index 1527b801f013..506eeb44f0d4 100644
> --- a/drivers/gpu/drm/nouveau/nouveau_drm.c
> +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
> @@ -1284,6 +1284,9 @@ nouveau_ioctls[] = {
>  	DRM_IOCTL_DEF_DRV(NOUVEAU_EXEC, nouveau_exec_ioctl_exec, DRM_RENDER_ALLOW),
>  };
>  
> +#define DRM_IOCTL_NOUVEAU_NVIV _IOC(_IOC_READ|_IOC_WRITE, DRM_IOCTL_BASE, \
> +				    DRM_COMMAND_BASE + DRM_NOUVEAU_NVIF, 0)

Is there any intention behind NVIV vs NVIF? Indicator that size is not
considered?

> +
>  long
>  nouveau_drm_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
>  {
> @@ -1297,14 +1300,10 @@ nouveau_drm_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
>  		return ret;
>  	}
>  
> -	switch (_IOC_NR(cmd) - DRM_COMMAND_BASE) {
> -	case DRM_NOUVEAU_NVIF:
> +	if ((cmd & ~IOCSIZE_MASK) == DRM_IOCTL_NOUVEAU_NVIV)
>  		ret = nouveau_abi16_ioctl(filp, (void __user *)arg, _IOC_SIZE(cmd));
> -		break;
> -	default:
> +	else
>  		ret = drm_ioctl(file, cmd, arg);
> -		break;
> -	}
>  
>  	pm_runtime_mark_last_busy(dev->dev);
>  	pm_runtime_put_autosuspend(dev->dev);