Linux virtualization list
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Maxime Ripard" <mripard@kernel.org>
Cc: dri-devel@lists.freedesktop.org, virtualization@lists.linux.dev,
	"Michael S. Tsirkin" <mst@redhat.com>,
	"Eugenio Perez" <eperezma@redhat.com>
Subject: Re: [PATCH RESEND 34/83] drm/qxl: Switch to drm_mode_config_create_initial_state
Date: Thu, 05 Nov 2026 19:32:32 +0000	[thread overview]
Message-ID: <sashiko-outbox-199921@kernel.org> (raw)
In-Reply-To: <20261105-drm-no-more-config-reset-v1-34-1f58d1e33380@kernel.org>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] drm/qxl: missing monitors object cleanup and UAF in qxl_modeset_init()
--

commit c9a17dafeda120436479a0975b3829acefce4188
Author: Maxime Ripard <mripard@kernel.org>

drm/qxl: Switch to drm_mode_config_create_initial_state

This commit replaces drm_mode_config_reset() with
drm_mode_config_create_initial_state() in qxl_modeset_init() to handle initial
KMS state creation at probe time.

> diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c
> index 204c47d613732..7f719e200d317 100644
> --- a/drivers/gpu/drm/qxl/qxl_display.c
> +++ b/drivers/gpu/drm/qxl/qxl_display.c
> @@ -1312,8 +1312,7 @@ int qxl_modeset_init(struct qxl_device *qdev)
>  	if (ret)
>  		return ret;
>  
> -	drm_mode_config_reset(&qdev->ddev);
> -	return 0;
> +	return drm_mode_config_create_initial_state(&qdev->ddev);
>  }

[Severity: High]
If drm_mode_config_create_initial_state() fails and returns an error here, can
this cause a use-after-free of qdev->monitors_config_bo and an unmapped MMIO
write to qdev->ram_header->monitors_config during probe teardown?

Earlier in qxl_modeset_init(), qxl_create_monitors_object() allocates, pins,
and vmaps qdev->monitors_config_bo, sets qdev->ram_header->monitors_config,
and allocates qdev->dumb_heads.

When qxl_modeset_init() returns an error to qxl_pci_probe(), the error path
jumps past qxl_modeset_fini() directly to the unload label:

drivers/gpu/drm/qxl/qxl_drv.c:qxl_pci_probe() {
	...
	ret = qxl_modeset_init(qdev);
	if (ret)
		goto unload;
	...
poll_fini:
	drm_kms_helper_poll_fini(&qdev->ddev);
	qxl_modeset_fini(qdev);
unload:
	qxl_device_fini(qdev);
}

In qxl_device_fini(), qxl_gem_fini() -> qxl_bo_force_delete() force-frees
qdev->monitors_config_bo while it is still pinned and vmapped (triggering
WARN_ON_ONCE(bo->map_count > 0) in qxl_ttm_bo_destroy() and leaving the
qdev->monitors_config_bo pointer non-NULL), and iounmap() unmaps
qdev->ram_header.

When qxl_pci_probe() then returns the error code, devres cleanup for
devm_drm_dev_alloc() runs qxl_drm_release() -> qxl_modeset_fini() ->
qxl_destroy_monitors_object():

drivers/gpu/drm/qxl/qxl_display.c:qxl_destroy_monitors_object() {
	if (!qdev->monitors_config_bo)
		return 0;
	...
	qdev->ram_header->monitors_config = 0;

	ret = qxl_bo_vunmap_and_unpin(qdev->monitors_config_bo);
	if (ret)
		return ret;

	qxl_bo_unref(&qdev->monitors_config_bo);
	return 0;
}

Because qdev->monitors_config_bo is still non-NULL, should qxl_modeset_init()
call qxl_destroy_monitors_object(qdev) on failure before returning?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261105-drm-no-more-config-reset-v1-0-1f58d1e33380@kernel.org?part=34

  reply	other threads:[~2026-11-05 19:32 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-05 15:29 [PATCH RESEND 00/83] drm: Switch drivers to drm_mode_config_create_initial_state Maxime Ripard
2026-11-05 15:29 ` [PATCH RESEND 34/83] drm/qxl: Switch " Maxime Ripard
2026-11-05 19:32   ` sashiko-bot [this message]
2026-11-05 15:30 ` [PATCH RESEND 61/83] drm/bochs: " Maxime Ripard
2026-11-05 21:25   ` sashiko-bot
2026-11-05 15:30 ` [PATCH RESEND 62/83] drm/cirrus-qemu: " Maxime Ripard
2026-11-05 21:26   ` sashiko-bot
2026-11-05 15:30 ` [PATCH RESEND 79/83] drm/virtio: " Maxime Ripard
2026-11-05 22:14   ` sashiko-bot

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=sashiko-outbox-199921@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=eperezma@redhat.com \
    --cc=mripard@kernel.org \
    --cc=mst@redhat.com \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=virtualization@lists.linux.dev \
    /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