NVIDIA GPU driver infrastructure
 help / color / mirror / Atom feed
From: Danilo Krummrich <dakr@kernel.org>
To: dakr@kernel.org, aliceryhl@google.com,
	daniel.almeida@collabora.com, acourbot@nvidia.com,
	ecourtney@nvidia.com, ojeda@kernel.org, boqun@kernel.org,
	gary@garyguo.net, bjorn3_gh@protonmail.com, lossin@kernel.org,
	a.hindborg@kernel.org, tmgross@umich.edu,
	deborah.brouwer@collabora.com, boris.brezillon@collabora.com,
	lyude@redhat.com
Cc: driver-core@lists.linux.dev, linux-kernel@vger.kernel.org,
	nova-gpu@lists.linux.dev, dri-devel@lists.freedesktop.org,
	rust-for-linux@vger.kernel.org, stable@vger.kernel.org,
	sashiko-bot@kernel.org
Subject: [PATCH v5 16/19] drm: fix race between partial drm_dev_register() failure and ioctl
Date: Sun, 28 Jun 2026 16:53:36 +0200	[thread overview]
Message-ID: <20260628145406.2107056-17-dakr@kernel.org> (raw)
In-Reply-To: <20260628145406.2107056-1-dakr@kernel.org>

If drm_dev_register() fails after registering a minor (e.g. render minor
registered, primary minor fails), userspace could have opened the first
minor and entered a drm_dev_enter() critical section. Since the
unplugged flag was never set, the ioctl proceeds while the error path
tears down device resources.

Fix this by introducing drm_dev_synchronize_unplug(), which sets the
unplugged flag and waits for the SRCU barrier, ensuring all in-flight
drm_dev_enter() critical sections complete before cleanup proceeds; call
it on the error path of drm_dev_register().

Fixes: bee330f3d672 ("drm: Use srcu to protect drm_device.unplugged")
Cc: stable@vger.kernel.org
Reported-by: sashiko-bot@kernel.org
Closes: https://lore.kernel.org/all/20260620190648.2E9F61F000E9@smtp.kernel.org/
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
---
 drivers/gpu/drm/drm_drv.c | 34 +++++++++++++++++++++++++---------
 1 file changed, 25 insertions(+), 9 deletions(-)

diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 675675480da4..e890052061f3 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -473,6 +473,22 @@ void drm_dev_exit(int idx)
 }
 EXPORT_SYMBOL(drm_dev_exit);
 
+/*
+ * Mark the device as unplugged and wait for any in-flight drm_dev_enter()
+ * critical sections to complete.
+ */
+static void drm_dev_synchronize_unplug(struct drm_device *dev)
+{
+	/*
+	 * After synchronizing any critical read section is guaranteed to see
+	 * the new value of ->unplugged, and any critical section which might
+	 * still have seen the old value of ->unplugged is guaranteed to have
+	 * finished.
+	 */
+	dev->unplugged = true;
+	synchronize_srcu(&drm_unplug_srcu);
+}
+
 /**
  * drm_dev_unplug - unplug a DRM device
  * @dev: DRM device
@@ -485,15 +501,7 @@ EXPORT_SYMBOL(drm_dev_exit);
  */
 void drm_dev_unplug(struct drm_device *dev)
 {
-	/*
-	 * After synchronizing any critical read section is guaranteed to see
-	 * the new value of ->unplugged, and any critical section which might
-	 * still have seen the old value of ->unplugged is guaranteed to have
-	 * finished.
-	 */
-	dev->unplugged = true;
-	synchronize_srcu(&drm_unplug_srcu);
-
+	drm_dev_synchronize_unplug(dev);
 	drm_dev_unregister(dev);
 
 	/* Clear all CPU mappings pointing to this device */
@@ -1091,6 +1099,7 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags)
 		goto err_minors;
 
 	dev->registered = true;
+	dev->unplugged = false;
 
 	if (driver->load) {
 		ret = driver->load(dev, flags);
@@ -1118,6 +1127,13 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags)
 	if (dev->driver->unload)
 		dev->driver->unload(dev);
 err_minors:
+	/*
+	 * If a minor was registered before the failure, userspace could have
+	 * opened it and entered a drm_dev_enter() critical section. Ensure all
+	 * such sections complete before we clean up.
+	 */
+	drm_dev_synchronize_unplug(dev);
+
 	remove_compat_control_link(dev);
 	drm_minor_unregister(dev, DRM_MINOR_ACCEL);
 	drm_minor_unregister(dev, DRM_MINOR_PRIMARY);
-- 
2.54.0


  parent reply	other threads:[~2026-06-28 14:55 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-28 14:53 [PATCH v5 00/19] rust: drm: Higher-Ranked Lifetime private data Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 01/19] rust: drm: ioctl: fix unbounded lifetimes in ioctl handler arguments Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 02/19] rust: drm: rename Uninit DeviceContext to Normal Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 03/19] rust: faux: add Device type with AsBusDevice support Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 04/19] rust: drm: Add Driver::ParentDevice associated type Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 05/19] rust: drm: change default DeviceContext to Normal Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 06/19] rust: drm: restrict AlwaysRefCounted to Normal Device context Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 07/19] rust: drm: restrict AlwaysRefCounted to Normal GEM Object context Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 08/19] rust: drm/gem: remove DeviceContext from shmem::Object Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 09/19] rust: drm: split Deref for Device context typestates Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 10/19] rust: drm: pin ioctl Device reference to Normal context Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 11/19] rust: drm: add Ioctl device context typestate Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 12/19] rust: drm: Add RegistrationGuard for drm_dev_enter/exit critical sections Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 13/19] rust: drm: Wrap ioctl dispatch in RegistrationGuard Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 14/19] rust: drm: return ParentDevice from Device AsRef Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 15/19] rust: drm: add AsRef<ParentDevice<Bound>> for Device<Registered> Danilo Krummrich
2026-06-28 14:53 ` Danilo Krummrich [this message]
2026-06-28 14:53 ` [PATCH v5 17/19] rust: drm: Add RegistrationData to drm::Driver Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 18/19] rust: drm: Pass registration data to ioctl handlers Danilo Krummrich
2026-06-28 14:53 ` [PATCH v5 19/19] drm: nova: Use drm::Device<Registered> to access the parent bus device Danilo Krummrich

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=20260628145406.2107056-17-dakr@kernel.org \
    --to=dakr@kernel.org \
    --cc=a.hindborg@kernel.org \
    --cc=acourbot@nvidia.com \
    --cc=aliceryhl@google.com \
    --cc=bjorn3_gh@protonmail.com \
    --cc=boqun@kernel.org \
    --cc=boris.brezillon@collabora.com \
    --cc=daniel.almeida@collabora.com \
    --cc=deborah.brouwer@collabora.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=driver-core@lists.linux.dev \
    --cc=ecourtney@nvidia.com \
    --cc=gary@garyguo.net \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lossin@kernel.org \
    --cc=lyude@redhat.com \
    --cc=nova-gpu@lists.linux.dev \
    --cc=ojeda@kernel.org \
    --cc=rust-for-linux@vger.kernel.org \
    --cc=sashiko-bot@kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=tmgross@umich.edu \
    /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