From: Thomas Zimmermann <tzimmermann@suse.de>
To: Sasha Levin <sashal@kernel.org>, stable@vger.kernel.org
Cc: Javier Martinez Canillas <javierm@redhat.com>,
Alex Deucher <alexander.deucher@amd.com>,
dri-devel@lists.freedesktop.org, nouveau@lists.freedesktop.org,
amd-gfx@lists.freedesktop.org, linux-fbdev@vger.kernel.org
Subject: Re: [PATCH 6.6.y] drm, fbcon, vga_switcheroo: Avoid race condition in fbcon setup
Date: Tue, 2 Dec 2025 08:42:52 +0100 [thread overview]
Message-ID: <250b44b3-e75b-4f8d-af44-d3a985ea554c@suse.de> (raw)
In-Reply-To: <20251201225123.1298682-1-sashal@kernel.org>
Hi,
thanks for backporting.
You also have to remove the calls to vga_switcheroo_client_fb_set() from
these files
https://elixir.bootlin.com/linux/v6.6/source/drivers/gpu/drm/i915/display/intel_fbdev.c
https://elixir.bootlin.com/linux/v6.6/source/drivers/gpu/drm/radeon/radeon_fbdev.c
Best regards
Thomas
Am 01.12.25 um 23:51 schrieb Sasha Levin:
> From: Thomas Zimmermann <tzimmermann@suse.de>
>
> [ Upstream commit eb76d0f5553575599561010f24c277cc5b31d003 ]
>
> Protect vga_switcheroo_client_fb_set() with console lock. Avoids OOB
> access in fbcon_remap_all(). Without holding the console lock the call
> races with switching outputs.
>
> VGA switcheroo calls fbcon_remap_all() when switching clients. The fbcon
> function uses struct fb_info.node, which is set by register_framebuffer().
> As the fb-helper code currently sets up VGA switcheroo before registering
> the framebuffer, the value of node is -1 and therefore not a legal value.
> For example, fbcon uses the value within set_con2fb_map() [1] as an index
> into an array.
>
> Moving vga_switcheroo_client_fb_set() after register_framebuffer() can
> result in VGA switching that does not switch fbcon correctly.
>
> Therefore move vga_switcheroo_client_fb_set() under fbcon_fb_registered(),
> which already holds the console lock. Fbdev calls fbcon_fb_registered()
> from within register_framebuffer(). Serializes the helper with VGA
> switcheroo's call to fbcon_remap_all().
>
> Although vga_switcheroo_client_fb_set() takes an instance of struct fb_info
> as parameter, it really only needs the contained fbcon state. Moving the
> call to fbcon initialization is therefore cleaner than before. Only amdgpu,
> i915, nouveau and radeon support vga_switcheroo. For all other drivers,
> this change does nothing.
>
> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
> Link: https://elixir.bootlin.com/linux/v6.17/source/drivers/video/fbdev/core/fbcon.c#L2942 # [1]
> Fixes: 6a9ee8af344e ("vga_switcheroo: initial implementation (v15)")
> Acked-by: Javier Martinez Canillas <javierm@redhat.com>
> Acked-by: Alex Deucher <alexander.deucher@amd.com>
> Cc: dri-devel@lists.freedesktop.org
> Cc: nouveau@lists.freedesktop.org
> Cc: amd-gfx@lists.freedesktop.org
> Cc: linux-fbdev@vger.kernel.org
> Cc: <stable@vger.kernel.org> # v2.6.34+
> Link: https://patch.msgid.link/20251105161549.98836-1-tzimmermann@suse.de
> [ drm_fb_helper_unregister_info() lacks vga_switcheroo code ]
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
> drivers/gpu/drm/drm_fb_helper.c | 7 -------
> drivers/video/fbdev/core/fbcon.c | 9 +++++++++
> 2 files changed, 9 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
> index b507c1c008a3e..3891837a78414 100644
> --- a/drivers/gpu/drm/drm_fb_helper.c
> +++ b/drivers/gpu/drm/drm_fb_helper.c
> @@ -30,9 +30,7 @@
> #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
>
> #include <linux/console.h>
> -#include <linux/pci.h>
> #include <linux/sysrq.h>
> -#include <linux/vga_switcheroo.h>
>
> #include <drm/drm_atomic.h>
> #include <drm/drm_drv.h>
> @@ -1668,7 +1666,6 @@ static int drm_fb_helper_find_sizes(struct drm_fb_helper *fb_helper,
> static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper)
> {
> struct drm_client_dev *client = &fb_helper->client;
> - struct drm_device *dev = fb_helper->dev;
> struct drm_fb_helper_surface_size sizes;
> int ret;
>
> @@ -1687,10 +1684,6 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper)
>
> strcpy(fb_helper->fb->comm, "[fbcon]");
>
> - /* Set the fb info for vgaswitcheroo clients. Does nothing otherwise. */
> - if (dev_is_pci(dev->dev))
> - vga_switcheroo_client_fb_set(to_pci_dev(dev->dev), fb_helper->info);
> -
> return 0;
> }
>
> diff --git a/drivers/video/fbdev/core/fbcon.c b/drivers/video/fbdev/core/fbcon.c
> index 78a5b22c8d150..8b2c3065c0c26 100644
> --- a/drivers/video/fbdev/core/fbcon.c
> +++ b/drivers/video/fbdev/core/fbcon.c
> @@ -64,6 +64,7 @@
> #include <linux/console.h>
> #include <linux/string.h>
> #include <linux/kd.h>
> +#include <linux/pci.h>
> #include <linux/slab.h>
> #include <linux/fb.h>
> #include <linux/fbcon.h>
> @@ -75,6 +76,7 @@
> #include <linux/interrupt.h>
> #include <linux/crc32.h> /* For counting font checksums */
> #include <linux/uaccess.h>
> +#include <linux/vga_switcheroo.h>
> #include <asm/irq.h>
>
> #include "fbcon.h"
> @@ -2913,6 +2915,9 @@ void fbcon_fb_unregistered(struct fb_info *info)
>
> console_lock();
>
> + if (info->device && dev_is_pci(info->device))
> + vga_switcheroo_client_fb_set(to_pci_dev(info->device), NULL);
> +
> fbcon_registered_fb[info->node] = NULL;
> fbcon_num_registered_fb--;
>
> @@ -3046,6 +3051,10 @@ static int do_fb_registered(struct fb_info *info)
> }
> }
>
> + /* Set the fb info for vga_switcheroo clients. Does nothing otherwise. */
> + if (info->device && dev_is_pci(info->device))
> + vga_switcheroo_client_fb_set(to_pci_dev(info->device), info);
> +
> return ret;
> }
>
--
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Jochen Jaser, Andrew McDonald, Werner Knoblich, (HRB 36809, AG Nürnberg)
next prev parent reply other threads:[~2025-12-02 7:42 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-01 16:01 FAILED: patch "[PATCH] drm, fbcon, vga_switcheroo: Avoid race condition in fbcon" failed to apply to 6.6-stable tree gregkh
2025-12-01 22:51 ` [PATCH 6.6.y] drm, fbcon, vga_switcheroo: Avoid race condition in fbcon setup Sasha Levin
2025-12-02 7:42 ` Thomas Zimmermann [this message]
2025-12-02 15:53 ` Greg KH
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=250b44b3-e75b-4f8d-af44-d3a985ea554c@suse.de \
--to=tzimmermann@suse.de \
--cc=alexander.deucher@amd.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=javierm@redhat.com \
--cc=linux-fbdev@vger.kernel.org \
--cc=nouveau@lists.freedesktop.org \
--cc=sashal@kernel.org \
--cc=stable@vger.kernel.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.