diff mbox series

[v5,01/11] driver: platform: Add helper for safer setting of driver_override

Message ID 20220316150533.421349-2-krzysztof.kozlowski@canonical.com (mailing list archive)
State New, archived
Headers show
Series Fix broken usage of driver_override (and kfree of static memory) | expand

Commit Message

Krzysztof Kozlowski March 16, 2022, 3:05 p.m. UTC
Several core drivers and buses expect that driver_override is a
dynamically allocated memory thus later they can kfree() it.

However such assumption is not documented, there were in the past and
there are already users setting it to a string literal. This leads to
kfree() of static memory during device release (e.g. in error paths or
during unbind):

    kernel BUG at ../mm/slub.c:3960!
    Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
    ...
    (kfree) from [<c058da50>] (platform_device_release+0x88/0xb4)
    (platform_device_release) from [<c0585be0>] (device_release+0x2c/0x90)
    (device_release) from [<c0a69050>] (kobject_put+0xec/0x20c)
    (kobject_put) from [<c0f2f120>] (exynos5_clk_probe+0x154/0x18c)
    (exynos5_clk_probe) from [<c058de70>] (platform_drv_probe+0x6c/0xa4)
    (platform_drv_probe) from [<c058b7ac>] (really_probe+0x280/0x414)
    (really_probe) from [<c058baf4>] (driver_probe_device+0x78/0x1c4)
    (driver_probe_device) from [<c0589854>] (bus_for_each_drv+0x74/0xb8)
    (bus_for_each_drv) from [<c058b48c>] (__device_attach+0xd4/0x16c)
    (__device_attach) from [<c058a638>] (bus_probe_device+0x88/0x90)
    (bus_probe_device) from [<c05871fc>] (device_add+0x3dc/0x62c)
    (device_add) from [<c075ff10>] (of_platform_device_create_pdata+0x94/0xbc)
    (of_platform_device_create_pdata) from [<c07600ec>] (of_platform_bus_create+0x1a8/0x4fc)
    (of_platform_bus_create) from [<c0760150>] (of_platform_bus_create+0x20c/0x4fc)
    (of_platform_bus_create) from [<c07605f0>] (of_platform_populate+0x84/0x118)
    (of_platform_populate) from [<c0f3c964>] (of_platform_default_populate_init+0xa0/0xb8)
    (of_platform_default_populate_init) from [<c01031f8>] (do_one_initcall+0x8c/0x404)

Provide a helper which clearly documents the usage of driver_override.
This will allow later to reuse the helper and reduce the amount of
duplicated code.

Convert the platform driver to use a new helper and make the
driver_override field const char (it is not modified by the core).

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
---
 drivers/base/driver.c           | 56 +++++++++++++++++++++++++++++++++
 drivers/base/platform.c         | 28 +++--------------
 include/linux/device/driver.h   |  2 ++
 include/linux/platform_device.h |  6 +++-
 4 files changed, 67 insertions(+), 25 deletions(-)

Comments

Andy Shevchenko March 16, 2022, 3:54 p.m. UTC | #1
On Wed, Mar 16, 2022 at 5:06 PM Krzysztof Kozlowski
<krzysztof.kozlowski@canonical.com> wrote:

...

> +int driver_set_override(struct device *dev, const char **override,
> +                       const char *s, size_t len)
> +{
> +       const char *new, *old;
> +       char *cp;

> +       if (!dev || !override || !s)
> +               return -EINVAL;

Sorry, I didn't pay much attention on this. First of all, I would drop
dev checks and simply require that dev should be valid. Do you expect
this can be called when dev is invalid? I would like to hear if it's
anything but theoretical. Second one, is the !s requirement. Do I
understand correctly that the string must be always present? But then
how we NULify the override? Is it possible? Third one is absence of
len check. See below.

> +       /*
> +        * The stored value will be used in sysfs show callback (sysfs_emit()),
> +        * which has a length limit of PAGE_SIZE and adds a trailing newline.
> +        * Thus we can store one character less to avoid truncation during sysfs
> +        * show.
> +        */
> +       if (len >= (PAGE_SIZE - 1))
> +               return -EINVAL;

I would relax this to make sure we can use it if \n is within this limit.

> +       cp = strnchr(s, len, '\n');
> +       if (cp)
> +               len = cp - s;
> +
> +       new = kstrndup(s, len, GFP_KERNEL);

Here is a word about the len check.

> +       if (!new)

If len == 0, this won't trigger and you have something very
interesting as a result.

One way is to use ZERO_PTR_OR_NULL() another is explicitly check for 0
and issue a (different?) error code.

> +               return -ENOMEM;
> +
> +       device_lock(dev);
> +       old = *override;
> +       if (cp != s) {
> +               *override = new;
> +       } else {
> +               kfree(new);
> +               *override = NULL;
> +       }
> +       device_unlock(dev);
> +
> +       kfree(old);
> +
> +       return 0;
> +}
Krzysztof Kozlowski April 3, 2022, 6:26 p.m. UTC | #2
On 16/03/2022 16:54, Andy Shevchenko wrote:
> On Wed, Mar 16, 2022 at 5:06 PM Krzysztof Kozlowski
> <krzysztof.kozlowski@canonical.com> wrote:
> 
> ...
> 
>> +int driver_set_override(struct device *dev, const char **override,
>> +                       const char *s, size_t len)
>> +{
>> +       const char *new, *old;
>> +       char *cp;

I focused on some other topics, so I did not respond to your questions
for some time.

> 
>> +       if (!dev || !override || !s)
>> +               return -EINVAL;
> 
> Sorry, I didn't pay much attention on this. First of all, I would drop
> dev checks and simply require that dev should be valid. Do you expect
> this can be called when dev is invalid?

No, I can skip the check.

> I would like to hear if it's
> anything but theoretical. Second one, is the !s requirement. Do I
> understand correctly that the string must be always present? But then
> how we NULify the override? Is it possible?

I did not change the convention of this sysfs hook, so removing of
override is passing empty string "". Have in mind that his interface is
mainly for sysfs, not for other drivers.

> Third one is absence of
> len check. See below.
> 
>> +       /*
>> +        * The stored value will be used in sysfs show callback (sysfs_emit()),
>> +        * which has a length limit of PAGE_SIZE and adds a trailing newline.
>> +        * Thus we can store one character less to avoid truncation during sysfs
>> +        * show.
>> +        */
>> +       if (len >= (PAGE_SIZE - 1))
>> +               return -EINVAL;
> 
> I would relax this to make sure we can use it if \n is within this limit.

Relax in what way? Store more than PAGE_SIZE? This is a user-visible
string and the name of a driver.

> 
>> +       cp = strnchr(s, len, '\n');
>> +       if (cp)
>> +               len = cp - s;
>> +
>> +       new = kstrndup(s, len, GFP_KERNEL);
> 
> Here is a word about the len check.
> 
>> +       if (!new)
> 
> If len == 0, this won't trigger and you have something very
> interesting as a result.

True, empty string would be set as override. The API says that empty
string clears the override, so len==0 should be valid (just like "\n" is
ok).

> 
> One way is to use ZERO_PTR_OR_NULL() another is explicitly check for 0
> and issue a (different?) error code.



Best regards,
Krzysztof
diff mbox series

Patch

diff --git a/drivers/base/driver.c b/drivers/base/driver.c
index 8c0d33e182fd..d51df6d80829 100644
--- a/drivers/base/driver.c
+++ b/drivers/base/driver.c
@@ -30,6 +30,62 @@  static struct device *next_device(struct klist_iter *i)
 	return dev;
 }
 
+/**
+ * driver_set_override() - Helper to set or clear driver override.
+ * @dev: Device to change
+ * @override: Address of string to change (e.g. &device->driver_override);
+ *            The contents will be freed and hold newly allocated override.
+ * @s: NUL-terminated string, new driver name to force a match, pass empty
+ *     string to clear it
+ * @len: length of @s
+ *
+ * Helper to set or clear driver override in a device, intended for the cases
+ * when the driver_override field is allocated by driver/bus code.
+ *
+ * Returns: 0 on success or a negative error code on failure.
+ */
+int driver_set_override(struct device *dev, const char **override,
+			const char *s, size_t len)
+{
+	const char *new, *old;
+	char *cp;
+
+	if (!dev || !override || !s)
+		return -EINVAL;
+
+	/*
+	 * The stored value will be used in sysfs show callback (sysfs_emit()),
+	 * which has a length limit of PAGE_SIZE and adds a trailing newline.
+	 * Thus we can store one character less to avoid truncation during sysfs
+	 * show.
+	 */
+	if (len >= (PAGE_SIZE - 1))
+		return -EINVAL;
+
+	cp = strnchr(s, len, '\n');
+	if (cp)
+		len = cp - s;
+
+	new = kstrndup(s, len, GFP_KERNEL);
+	if (!new)
+		return -ENOMEM;
+
+	device_lock(dev);
+	old = *override;
+	if (cp != s) {
+		*override = new;
+	} else {
+		kfree(new);
+		*override = NULL;
+	}
+	device_unlock(dev);
+
+	kfree(old);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(driver_set_override);
+
 /**
  * driver_for_each_device - Iterator for devices bound to a driver.
  * @drv: Driver we're iterating.
diff --git a/drivers/base/platform.c b/drivers/base/platform.c
index 8cc272fd5c99..b684157b7f2f 100644
--- a/drivers/base/platform.c
+++ b/drivers/base/platform.c
@@ -1275,31 +1275,11 @@  static ssize_t driver_override_store(struct device *dev,
 				     const char *buf, size_t count)
 {
 	struct platform_device *pdev = to_platform_device(dev);
-	char *driver_override, *old, *cp;
-
-	/* We need to keep extra room for a newline */
-	if (count >= (PAGE_SIZE - 1))
-		return -EINVAL;
-
-	driver_override = kstrndup(buf, count, GFP_KERNEL);
-	if (!driver_override)
-		return -ENOMEM;
-
-	cp = strchr(driver_override, '\n');
-	if (cp)
-		*cp = '\0';
-
-	device_lock(dev);
-	old = pdev->driver_override;
-	if (strlen(driver_override)) {
-		pdev->driver_override = driver_override;
-	} else {
-		kfree(driver_override);
-		pdev->driver_override = NULL;
-	}
-	device_unlock(dev);
+	int ret;
 
-	kfree(old);
+	ret = driver_set_override(dev, &pdev->driver_override, buf, count);
+	if (ret)
+		return ret;
 
 	return count;
 }
diff --git a/include/linux/device/driver.h b/include/linux/device/driver.h
index 15e7c5e15d62..700453017e1c 100644
--- a/include/linux/device/driver.h
+++ b/include/linux/device/driver.h
@@ -151,6 +151,8 @@  extern int __must_check driver_create_file(struct device_driver *driver,
 extern void driver_remove_file(struct device_driver *driver,
 			       const struct driver_attribute *attr);
 
+int driver_set_override(struct device *dev, const char **override,
+			const char *s, size_t len);
 extern int __must_check driver_for_each_device(struct device_driver *drv,
 					       struct device *start,
 					       void *data,
diff --git a/include/linux/platform_device.h b/include/linux/platform_device.h
index 7c96f169d274..582d83ed9a91 100644
--- a/include/linux/platform_device.h
+++ b/include/linux/platform_device.h
@@ -31,7 +31,11 @@  struct platform_device {
 	struct resource	*resource;
 
 	const struct platform_device_id	*id_entry;
-	char *driver_override; /* Driver name to force a match */
+	/*
+	 * Driver name to force a match.  Do not set directly, because core
+	 * frees it.  Use driver_set_override() to set or clear it.
+	 */
+	const char *driver_override;
 
 	/* MFD cell pointer */
 	struct mfd_cell *mfd_cell;