Linux Trace Kernel
 help / color / mirror / Atom feed
* [PATCH] driver core: Disable driver overriding by default
@ 2026-11-16 11:08 Uwe Kleine-König
  2026-11-16 11:36 ` sashiko-bot
  0 siblings, 1 reply; 2+ messages in thread
From: Uwe Kleine-König @ 2026-11-16 11:08 UTC (permalink / raw)
  To: Greg Kroah-Hartman
  Cc: Jonathan Corbet, Shuah Khan, Randy Dunlap, Rafael J. Wysocki,
	Danilo Krummrich, Johan Hovold, Steven Rostedt, Masami Hiramatsu,
	Mathieu Desnoyers, David Lechner, Armin Wolf, Thierry Reding,
	Mike Rapoport (Microsoft), Arnd Bergmann, Karl Mehltretter,
	Paul E. McKenney, Borislav Petkov (AMD), Dapeng Mi, Marco Elver,
	Ethan Nelson-Moore, Jeff Layton, Bradley Morgan, Andrew Morton,
	Petr Mladek, Pnina Feder, Mete Durlu, Aaron Tomlin, Rio,
	Feng Tang, linux-doc, linux-kernel, driver-core,
	linux-trace-kernel, Mark Brown, Bjorn Helgaas, Alex Williamson

Driver overriding is useful only in a very limited set of situations
and then only with very few drivers that are designed to support that.

Disallow matching via driver_override unless the driver explicitly
allows it or the safe guard is disabled using the
"allow_driver_override=1" kernel parameter.

There are many drivers that don't expect to be force-bound and often
doing it anyhow results in a NULL pointer exception. To make debugging
such cases easier, add a new taint if such a forced match happens.

Implementation detail: device_match_driver_override() isn't a static
inline any more. It grew a certain complexity (e.g. a pr_info()) and so
was made a regular exported function.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com>
---
Hello,

this is a rebase to current next of the series
https://lore.kernel.org/all/cover.1790612298.git.u.kleine-koenig@baylibre.com/
. It combines the functionality of former patches #2 and #3 (#1 was a
fix that was already applied). I also addressed sashiko's comments. I think
this matches what most people want in the previous discussion.

As the rebase wasn't trivial and I merged the two patches, I didn't
apply the review tags I already received for v3.

If this patch is accepted, a number of drivers that are supposed to be
force-bound need to get their .support_driver_override property set. I'm
aware of the following:

 - spidev_spi_driver in drivers/spi/spidev.c
 - stub_driver in drivers/pci/pci-stub.c
 - uio_dmem_genirq in drivers/uio/uio_dmem_genirq.c
 - uio_pci_driver in drivers/uio/uio_pci_generic.c
 - uio_pdrv_genirq in drivers/uio/uio_pdrv_genirq.c
 - vfio_amba_driver in drivers/vfio/platform/vfio_amba.c
 - vfio_ccw_sch_driver in drivers/s390/cio/vfio_ccw_drv.c
 - vfio_pci_driver in drivers/vfio/pci/vfio_pci.c
 - vfio_platform_driver in drivers/vfio/platform/vfio_platform.c

This is not needed for the userspace i2c driver.

As the flag to set is only introduced in this patch and assuming this
patch is accepted, I suggest to create an immutable branch for it to
allow setting the needed flags in driver trees. Alternatively let all
these go together via a single tree or branch.

Best regards
Uwe

 .../admin-guide/kernel-parameters.txt         |  6 +++
 Documentation/admin-guide/tainted-kernels.rst | 54 ++++++++++---------
 drivers/base/bus.c                            | 50 +++++++++++++++++
 include/linux/device.h                        | 19 +------
 include/linux/device/driver.h                 |  2 +
 include/linux/panic.h                         |  3 +-
 include/trace/events/module.h                 |  3 +-
 kernel/panic.c                                |  3 +-
 tools/debugging/kernel-chktaint               |  9 ++++
 9 files changed, 104 insertions(+), 45 deletions(-)

diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index 85804cfe7ef6..aabc97feef9b 100644
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -424,6 +424,12 @@ Kernel parameters
 			do not want to use tracing_snapshot_alloc() as it needs
 			to be done where GFP_KERNEL allocations are allowed.
 
+	allow_driver_override=
+			When set to 1 allow matching drivers without
+			.support_driver_override = true to devices using
+			driver_override. This is a dangerous operation in
+			general allowing root to oops the kernel.
+
 	allow_mismatched_32bit_el0 [ARM64,EARLY]
 			Allow execve() of 32-bit applications and setting of the
 			PER_LINUX32 personality on systems where only a strict
diff --git a/Documentation/admin-guide/tainted-kernels.rst b/Documentation/admin-guide/tainted-kernels.rst
index fcb8b9734548..e6bab575ccef 100644
--- a/Documentation/admin-guide/tainted-kernels.rst
+++ b/Documentation/admin-guide/tainted-kernels.rst
@@ -74,35 +74,36 @@ a particular type of taint. It's best to leave that to the aforementioned
 script, but if you need something quick you can use this shell command to check
 which bits are set::
 
-	$ for i in $(seq 0 19); do echo $i $(($(cat /proc/sys/kernel/tainted)>>$i&1));done
+	$ for i in $(seq 0 20); do echo $i $(($(cat /proc/sys/kernel/tainted)>>$i&1));done
 
 Table for decoding tainted state
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-===  ===  ======  ========================================================
-Bit  Log  Number  Reason that got the kernel tainted
-===  ===  ======  ========================================================
-  0  G/P       1  proprietary module was loaded
-  1  _/F       2  module was force loaded
-  2  _/S       4  kernel running on an out of specification system
-  3  _/R       8  module was force unloaded
-  4  _/M      16  processor reported a Machine Check Exception (MCE)
-  5  _/B      32  bad page referenced or some unexpected page flags
-  6  _/U      64  taint requested by userspace application
-  7  _/D     128  kernel died recently, i.e. there was an OOPS or BUG
-  8  _/A     256  ACPI table overridden by user
-  9  _/W     512  kernel issued warning
- 10  _/C    1024  staging driver was loaded
- 11  _/I    2048  workaround for bug in platform firmware applied
- 12  _/O    4096  externally-built ("out-of-tree") module was loaded
- 13  _/E    8192  unsigned module was loaded
- 14  _/L   16384  soft lockup occurred
- 15  _/K   32768  kernel has been live patched
- 16  _/X   65536  auxiliary taint, defined for and used by distros
- 17  _/T  131072  kernel was built with the struct randomization plugin
- 18  _/N  262144  an in-kernel test has been run
- 19  _/J  524288  userspace used a mutating debug operation in fwctl
-===  ===  ======  ========================================================
+===  ===  =======  ========================================================
+Bit  Log   Number  Reason that got the kernel tainted
+===  ===  =======  ========================================================
+  0  G/P        1  proprietary module was loaded
+  1  _/F        2  module was force loaded
+  2  _/S        4  kernel running on an out of specification system
+  3  _/R        8  module was force unloaded
+  4  _/M       16  processor reported a Machine Check Exception (MCE)
+  5  _/B       32  bad page referenced or some unexpected page flags
+  6  _/U       64  taint requested by userspace application
+  7  _/D      128  kernel died recently, i.e. there was an OOPS or BUG
+  8  _/A      256  ACPI table overridden by user
+  9  _/W      512  kernel issued warning
+ 10  _/C     1024  staging driver was loaded
+ 11  _/I     2048  workaround for bug in platform firmware applied
+ 12  _/O     4096  externally-built ("out-of-tree") module was loaded
+ 13  _/E     8192  unsigned module was loaded
+ 14  _/L    16384  soft lockup occurred
+ 15  _/K    32768  kernel has been live patched
+ 16  _/X    65536  auxiliary taint, defined for and used by distros
+ 17  _/T   131072  kernel was built with the struct randomization plugin
+ 18  _/N   262144  an in-kernel test has been run
+ 19  _/J   524288  userspace used a mutating debug operation in fwctl
+ 20  _/Z  1048576  a driver was forced on a non-matching device
+===  ===  =======  ========================================================
 
 Note: The character ``_`` is representing a blank in this table to make reading
 easier.
@@ -189,3 +190,6 @@ More detailed explanation for tainting
  19) ``J`` if userspace opened /dev/fwctl/* and performed a FWTCL_RPC_DEBUG_WRITE
      to use the devices debugging features. Device debugging features could
      cause the device to malfunction in undefined ways.
+
+ 20) ``Z`` if a device was matched using the `driver_override` sysfs property
+     to a driver that didn't explicitly allow that.
diff --git a/drivers/base/bus.c b/drivers/base/bus.c
index 396097fdb316..5344b20488fd 100644
--- a/drivers/base/bus.c
+++ b/drivers/base/bus.c
@@ -603,6 +603,56 @@ int bus_add_device(struct device *dev)
 	return error;
 }
 
+static bool __read_mostly allow_driver_override;
+
+static int __init allow_driver_override_setup(char *str)
+{
+	if (!str || kstrtobool(str, &allow_driver_override) < 0)
+		allow_driver_override = true;
+
+	return 1;
+}
+__setup("allow_driver_override=", allow_driver_override_setup);
+
+/**
+ * device_match_driver_override() - Match a driver against the device's driver_override.
+ * @dev: device to check
+ * @drv: driver to match against
+ *
+ * Returns > 0 if a driver override is set and matches the given driver, 0 if a
+ * driver override is set but does not match, or < 0 if a driver override is not
+ * set at all.
+ */
+int device_match_driver_override(struct device *dev,
+				 const struct device_driver *drv)
+{
+	guard(spinlock)(&dev->driver_override.lock);
+	if (dev->driver_override.name) {
+		int ret = !strcmp(dev->driver_override.name, drv->name);
+
+		if (ret > 0 && !drv->support_driver_override) {
+			if (allow_driver_override) {
+				add_taint_module(drv->owner,
+						 TAINT_DRIVER_OVERRIDE, LOCKDEP_STILL_OK);
+			} else {
+				/*
+				 * If you hit this warning and you need to
+				 * quickly workaround this check, boot with
+				 * allow_driver_override=1 on the kernel
+				 * commandline. The right fix is to set the
+				 * driver's .support_driver_override property.
+				 */
+				dev_warn(dev, "Suppressed driver override binding to %s.\n", drv->name);
+				return -1;
+			}
+		}
+
+		return ret;
+	}
+	return -1;
+}
+EXPORT_SYMBOL_GPL(device_match_driver_override);
+
 static void __bus_probe_device(struct device *dev)
 {
 	struct subsys_private *sp = bus_to_subsys(dev->bus);
diff --git a/include/linux/device.h b/include/linux/device.h
index 6b082c638d45..05413c1a9454 100644
--- a/include/linux/device.h
+++ b/include/linux/device.h
@@ -907,23 +907,8 @@ static inline bool device_has_driver_override(struct device *dev)
 	return !!dev->driver_override.name;
 }
 
-/**
- * device_match_driver_override() - Match a driver against the device's driver_override.
- * @dev: device to check
- * @drv: driver to match against
- *
- * Returns > 0 if a driver override is set and matches the given driver, 0 if a
- * driver override is set but does not match, or < 0 if a driver override is not
- * set at all.
- */
-static inline int device_match_driver_override(struct device *dev,
-					       const struct device_driver *drv)
-{
-	guard(spinlock)(&dev->driver_override.lock);
-	if (dev->driver_override.name)
-		return !strcmp(dev->driver_override.name, drv->name);
-	return -1;
-}
+int device_match_driver_override(struct device *dev,
+				 const struct device_driver *drv);
 
 /**
  * device_iommu_mapped - Returns true when the device DMA is translated
diff --git a/include/linux/device/driver.h b/include/linux/device/driver.h
index 29fbc01ef06f..0153cfcbe1b9 100644
--- a/include/linux/device/driver.h
+++ b/include/linux/device/driver.h
@@ -56,6 +56,7 @@ enum probe_type {
  * @bus:	The bus which the device of this driver belongs to.
  * @owner:	The module owner.
  * @mod_name:	Used for built-in modules.
+ * @support_driver_override: driver_override only works if this is true.
  * @suppress_bind_attrs: Disables bind/unbind via sysfs.
  * @probe_type:	Type of the probe (synchronous or asynchronous) to use.
  * @of_match_table: The open firmware table.
@@ -104,6 +105,7 @@ struct device_driver {
 	struct module		*owner;
 	const char		*mod_name;	/* used for built-in modules */
 
+	bool support_driver_override;
 	bool suppress_bind_attrs;	/* disables bind/unbind via sysfs */
 	enum probe_type probe_type;
 
diff --git a/include/linux/panic.h b/include/linux/panic.h
index 17e61b61c45f..009232511cc7 100644
--- a/include/linux/panic.h
+++ b/include/linux/panic.h
@@ -89,7 +89,8 @@ static inline void set_arch_panic_timeout(int timeout, int arch_default_timeout)
 #define TAINT_RANDSTRUCT		17
 #define TAINT_TEST			18
 #define TAINT_FWCTL			19
-#define TAINT_FLAGS_COUNT		20
+#define TAINT_DRIVER_OVERRIDE		20
+#define TAINT_FLAGS_COUNT		21
 #define TAINT_FLAGS_MAX			((1UL << TAINT_FLAGS_COUNT) - 1)
 
 struct taint_flag {
diff --git a/include/trace/events/module.h b/include/trace/events/module.h
index e5a006be9dc6..d54ce3c7572b 100644
--- a/include/trace/events/module.h
+++ b/include/trace/events/module.h
@@ -26,7 +26,8 @@ struct module;
 	{ (1UL << TAINT_OOT_MODULE),		"O" },		\
 	{ (1UL << TAINT_FORCED_MODULE),		"F" },		\
 	{ (1UL << TAINT_CRAP),			"C" },		\
-	{ (1UL << TAINT_UNSIGNED_MODULE),	"E" })
+	{ (1UL << TAINT_UNSIGNED_MODULE),	"E" },		\
+	{ (1UL << TAINT_DRIVER_OVERRIDE),	"Z" })
 
 TRACE_EVENT(module_load,
 
diff --git a/kernel/panic.c b/kernel/panic.c
index 91e48d33b7db..81ea9ef9afb0 100644
--- a/kernel/panic.c
+++ b/kernel/panic.c
@@ -825,6 +825,7 @@ const struct taint_flag taint_flags[TAINT_FLAGS_COUNT] = {
 	TAINT_FLAG(RANDSTRUCT,			'T', ' '),
 	TAINT_FLAG(TEST,			'N', ' '),
 	TAINT_FLAG(FWCTL,			'J', ' '),
+	TAINT_FLAG(DRIVER_OVERRIDE,		'Z', ' '),
 };
 
 #undef TAINT_FLAG
@@ -861,7 +862,7 @@ static void print_tainted_seq(struct seq_buf *s, bool verbose)
  * exact size is allocated dynamically; the initial buffer remains
  * as a fallback if allocation fails.
  *
- * The verbose taint string currently requires up to 327 characters.
+ * The verbose taint string currently requires up to 348 characters.
  */
 #define INIT_TAINT_BUF_MAX 350
 
diff --git a/tools/debugging/kernel-chktaint b/tools/debugging/kernel-chktaint
index e1571c04afb5..56215e0f5112 100755
--- a/tools/debugging/kernel-chktaint
+++ b/tools/debugging/kernel-chktaint
@@ -211,6 +211,15 @@ else
 	addout "J"
 	echo " * fwctl's mutating debug interface was used (#19)"
 fi
+
+T=`expr $T / 2`
+if [ `expr $T % 2` -eq 0 ]; then
+	addout " "
+else
+	addout "Z"
+	echo " * A device was matched using driver_override to a driver not supporting that (#20)"
+fi
+
 echo "Raw taint value as int/string: $taint/'$out'"
 
 # report on any tainted loadable modules

base-commit: 84abb955b1015477aa11a96e2f3ac1b938b77f48
-- 
2.56.0.rc1


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH] driver core: Disable driver overriding by default
  2026-11-16 11:08 [PATCH] driver core: Disable driver overriding by default Uwe Kleine-König
@ 2026-11-16 11:36 ` sashiko-bot
  0 siblings, 0 replies; 2+ messages in thread
From: sashiko-bot @ 2026-11-16 11:36 UTC (permalink / raw)
  To: Uwe Kleine-König; +Cc: linux-trace-kernel

Thank you for your contribution! Sashiko AI review found 3 potential
issues to consider:
- [High] driver core: device_match_driver_override() returns -1 instead
  of 0
- [High] driver core: disabling driver_override by default breaks
  in-tree drivers
- [Low] driver core: allow_driver_override_setup() enables override on
  parse error
--

commit 06f24686a63b0776fbdf7d5d7bb34b7905ceba41
Author: Uwe Kleine-König <u.kleine-koenig@baylibre.com>

driver core: Disable driver overriding by default

This commit disables driver_override matching unless a driver sets
support_driver_override or the allow_driver_override=1 boot parameter is
used, and adds TAINT_DRIVER_OVERRIDE when an unsupported override match is
allowed.

> diff --git a/drivers/base/bus.c b/drivers/base/bus.c
> index 396097fdb3162..5344b20488fde 100644
> --- a/drivers/base/bus.c
> +++ b/drivers/base/bus.c
> @@ -603,6 +603,56 @@ int bus_add_device(struct device *dev)
>  	return error;
>  }
>  
> +static bool __read_mostly allow_driver_override;
> +
> +static int __init allow_driver_override_setup(char *str)
> +{
> +	if (!str || kstrtobool(str, &allow_driver_override) < 0)
> +		allow_driver_override = true;
> +
> +	return 1;
> +}
> +__setup("allow_driver_override=", allow_driver_override_setup);

[Severity: Low]
Does allow_driver_override_setup() enable allow_driver_override when
kstrtobool() fails?  Passing an unrecognized argument such as
allow_driver_override=disabled or allow_driver_override= causes kstrtobool()
to return -EINVAL, setting allow_driver_override to true instead of keeping
it disabled.

> +
> +/**
> + * device_match_driver_override() - Match a driver against the device's driver_override.
> + * @dev: device to check
> + * @drv: driver to match against
> + *
> + * Returns > 0 if a driver override is set and matches the given driver, 0 if a
> + * driver override is set but does not match, or < 0 if a driver override is not
> + * set at all.
> + */
> +int device_match_driver_override(struct device *dev,
> +				 const struct device_driver *drv)
> +{
> +	guard(spinlock)(&dev->driver_override.lock);
> +	if (dev->driver_override.name) {
> +		int ret = !strcmp(dev->driver_override.name, drv->name);
> +
> +		if (ret > 0 && !drv->support_driver_override) {

[Severity: High]
Does rejecting driver_override by default before any in-tree driver sets
support_driver_override break in-kernel callers of
device_set_driver_override() and git bisectability?

Several subsystems call device_set_driver_override() internally to bind
kernel-created devices whose names or ID tables do not match via normal bus
matching.  For example, i2s_create_secondary_device() creates a "%s-sec"
platform device, sets its override to "samsung-i2s", and calls
device_attach():

sound/soc/samsung/i2s.c:i2s_create_secondary_device() {
	...
	ret = device_set_driver_override(&pdev_sec->dev, "samsung-i2s");
	...
	ret = device_attach(&pdev_sec->dev);
	if (ret <= 0) {
		platform_device_unregister(priv->pdev_sec);
		dev_info(&pdev_sec->dev, "device_attach() failed\n");
		return ret;
	}
	...
}

Because samsung_i2s_driver does not set support_driver_override,
platform_match() fails, device_attach() returns 0 while priv->pdev_sec is
still NULL (leaking pdev_sec), and samsung_i2s_probe() then dereferences
priv->pdev_sec.  Other internal users such as imx_clk_scu_alloc_dev() and
rpmsg_ns_register_device() also fail to bind their drivers at boot, and all
VFIO, UIO, and stub driver_override workflows are broken at this commit.

In addition, PCI and CDX drivers that use PCI_DRIVER_OVERRIDE_DEVICE_VFIO()
or CDX_DEVICE_DRIVER_OVERRIDE() rely on device_match_driver_override()
returning > 0 so do_pci_match_id() and cdx_bus_match() match override_only
ID table entries, whereas setting support_driver_override on those drivers
would also allow pci_match_device() to bind arbitrary non-matching devices
via pci_device_id_any without tainting.

> +			if (allow_driver_override) {
> +				add_taint_module(drv->owner,
> +						 TAINT_DRIVER_OVERRIDE, LOCKDEP_STILL_OK);
> +			} else {
> +				/*
> +				 * If you hit this warning and you need to
> +				 * quickly workaround this check, boot with
> +				 * allow_driver_override=1 on the kernel
> +				 * commandline. The right fix is to set the
> +				 * driver's .support_driver_override property.
> +				 */
> +				dev_warn(dev, "Suppressed driver override binding to %s.\n", drv->name);
> +				return -1;

[Severity: High]
Should device_match_driver_override() return 0 rather than -1 when
suppressing a match?

Per the function comment, < 0 means no driver_override is set at all, while
0 means an override is set but does not match drv.  Returning -1 for the
matching driver while returning 0 for all other drivers causes callers to
fall through to normal bus matching for the overridden driver:

drivers/vdpa/vdpa.c:vdpa_dev_match() {
	int ret;

	/* Check override first, and if set, only use the named driver */
	ret = device_match_driver_override(dev, drv);
	if (ret >= 0)
		return ret;

	/* Currently devices must be supported by all vDPA bus drivers */
	return 1;
}

On vDPA (and similarly in css_bus_match(), ap_device_probe(), and wildcard
PCI drivers), returning -1 logs that the override was suppressed, skips
TAINT_DRIVER_OVERRIDE, and still binds the device to the overridden driver.
And because device_match_driver_override() runs before normal bus matching,
setting driver_override to a driver that also matches the device natively
(such as in of_qcom_slim_ngd_register() or when rebinding a device to its
native driver) either logs a false suppression warning before binding anyway
or taints the kernel with TAINT_DRIVER_OVERRIDE.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261116110853.3418077-2-u.kleine-koenig@baylibre.com?part=1

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2026-11-16 11:36 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-16 11:08 [PATCH] driver core: Disable driver overriding by default Uwe Kleine-König
2026-11-16 11:36 ` sashiko-bot

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox