[PATCH 0/4] ARM: amba: driver_override improvements

[PATCH 0/4] ARM: amba: driver_override improvements

[Geert Uytterhoeven]

	Hi Russell,

This patch series contains improvements to driver_override handling in
the AMBA bus driver, including two bugfixes that are based on similar
fixes for the PCI and platform buses.

Thanks!

Geert Uytterhoeven (4):
  ARM: amba: Make driver_override output consistent with other buses
  ARM: amba: Fix race condition with driver_override
  ARM: amba: Don't read past the end of sysfs "driver_override" buffer
  ARM: amba: Fix wrong indentation in driver_override_store()

 drivers/amba/bus.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)

-- 
2.7.4

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

[PATCH 1/4] ARM: amba: Make driver_override output consistent with other buses

[Geert Uytterhoeven]

For AMBA devices with unconfigured driver override, the
"driver_override" sysfs virtual file is empty, while it contains
"(null)" for platform and PCI devices.

Make AMBA consistent with other buses by dropping the test for a NULL
pointer.

Note that contrary to popular belief, sprintf() handles NULL pointers
fine; they are printed as "(null)".

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 drivers/amba/bus.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/drivers/amba/bus.c b/drivers/amba/bus.c
index 594c228d2f021123..6ffd778352e6d953 100644
--- a/drivers/amba/bus.c
+++ b/drivers/amba/bus.c
@@ -70,9 +70,6 @@ static ssize_t driver_override_show(struct device *_dev,
 {
 	struct amba_device *dev = to_amba_device(_dev);
 
-	if (!dev->driver_override)
-		return 0;
-
 	return sprintf(buf, "%s\n", dev->driver_override);
 }
 
-- 
2.7.4

[PATCH 2/4] ARM: amba: Fix race condition with driver_override

[Geert Uytterhoeven]

The driver_override implementation is susceptible to a race condition
when different threads are reading vs storing a different driver
override.  Add locking to avoid this race condition.

Cfr. commits 6265539776a0810b ("driver core: platform: fix race
condition with driver_override") and 9561475db680f714 ("PCI: Fix race
condition with driver_override").

Fixes: 3cf385713460eb2b ("ARM: 8256/1: driver coamba: add device binding path 'driver_override'")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 drivers/amba/bus.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/drivers/amba/bus.c b/drivers/amba/bus.c
index 6ffd778352e6d953..36c5653ced5742b7 100644
--- a/drivers/amba/bus.c
+++ b/drivers/amba/bus.c
@@ -69,8 +69,12 @@ static ssize_t driver_override_show(struct device *_dev,
 				    struct device_attribute *attr, char *buf)
 {
 	struct amba_device *dev = to_amba_device(_dev);
+	ssize_t len;
 
-	return sprintf(buf, "%s\n", dev->driver_override);
+	device_lock(_dev);
+	len = sprintf(buf, "%s\n", dev->driver_override);
+	device_unlock(_dev);
+	return len;
 }
 
 static ssize_t driver_override_store(struct device *_dev,
@@ -78,7 +82,7 @@ static ssize_t driver_override_store(struct device *_dev,
 				     const char *buf, size_t count)
 {
 	struct amba_device *dev = to_amba_device(_dev);
-	char *driver_override, *old = dev->driver_override, *cp;
+	char *driver_override, *old, *cp;
 
 	if (count > PATH_MAX)
 		return -EINVAL;
@@ -91,12 +95,15 @@ static ssize_t driver_override_store(struct device *_dev,
 	if (cp)
 		*cp = '\0';
 
+	device_lock(_dev);
+	old = dev->driver_override;
 	if (strlen(driver_override)) {
 		dev->driver_override = driver_override;
 	} else {
 	       kfree(driver_override);
 	       dev->driver_override = NULL;
 	}
+	device_unlock(_dev);
 
 	kfree(old);
 
-- 
2.7.4

Re: [PATCH 2/4] ARM: amba: Fix race condition with driver_override

[Todd Kjos]

+stable

what is the status of this patch? We'd like to get it into the android
common branches to fix possible double free.

On Fri, Jan 19, 2018 at 7:24 AM, Geert Uytterhoeven
<geert+renesas@glider.be> wrote:
> The driver_override implementation is susceptible to a race condition
> when different threads are reading vs storing a different driver
> override.  Add locking to avoid this race condition.
>
> Cfr. commits 6265539776a0810b ("driver core: platform: fix race
> condition with driver_override") and 9561475db680f714 ("PCI: Fix race
> condition with driver_override").
>
> Fixes: 3cf385713460eb2b ("ARM: 8256/1: driver coamba: add device binding path 'driver_override'")
> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
> ---
>  drivers/amba/bus.c | 11 +++++++++--
>  1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/amba/bus.c b/drivers/amba/bus.c
> index 6ffd778352e6d953..36c5653ced5742b7 100644
> --- a/drivers/amba/bus.c
> +++ b/drivers/amba/bus.c
> @@ -69,8 +69,12 @@ static ssize_t driver_override_show(struct device *_dev,
>                                     struct device_attribute *attr, char *buf)
>  {
>         struct amba_device *dev = to_amba_device(_dev);
> +       ssize_t len;
>
> -       return sprintf(buf, "%s\n", dev->driver_override);
> +       device_lock(_dev);
> +       len = sprintf(buf, "%s\n", dev->driver_override);
> +       device_unlock(_dev);
> +       return len;
>  }
>
>  static ssize_t driver_override_store(struct device *_dev,
> @@ -78,7 +82,7 @@ static ssize_t driver_override_store(struct device *_dev,
>                                      const char *buf, size_t count)
>  {
>         struct amba_device *dev = to_amba_device(_dev);
> -       char *driver_override, *old = dev->driver_override, *cp;
> +       char *driver_override, *old, *cp;
>
>         if (count > PATH_MAX)
>                 return -EINVAL;
> @@ -91,12 +95,15 @@ static ssize_t driver_override_store(struct device *_dev,
>         if (cp)
>                 *cp = '\0';
>
> +       device_lock(_dev);
> +       old = dev->driver_override;
>         if (strlen(driver_override)) {
>                 dev->driver_override = driver_override;
>         } else {
>                kfree(driver_override);
>                dev->driver_override = NULL;
>         }
> +       device_unlock(_dev);
>
>         kfree(old);
>
> --
> 2.7.4
>

Re: [PATCH 2/4] ARM: amba: Fix race condition with driver_override

[Geert Uytterhoeven]

Hi Todd,

On Fri, Mar 2, 2018 at 7:23 PM, Todd Kjos <tkjos@android.com> wrote:
> +stable
>
> what is the status of this patch? We'd like to get it into the android
> common branches to fix possible double free.

Thanks for your interest!

So far this patch (and the 3 others in the series) haven't received any
comments. If someone sends a Acked-by or Reviewed-by, I can submit them
to RMK's patch system.

Thanks!

> On Fri, Jan 19, 2018 at 7:24 AM, Geert Uytterhoeven
> <geert+renesas@glider.be> wrote:
>> The driver_override implementation is susceptible to a race condition
>> when different threads are reading vs storing a different driver
>> override.  Add locking to avoid this race condition.
>>
>> Cfr. commits 6265539776a0810b ("driver core: platform: fix race
>> condition with driver_override") and 9561475db680f714 ("PCI: Fix race
>> condition with driver_override").
>>
>> Fixes: 3cf385713460eb2b ("ARM: 8256/1: driver coamba: add device binding path 'driver_override'")
>> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

[PATCH 3/4] ARM: amba: Don't read past the end of sysfs "driver_override" buffer

[Geert Uytterhoeven]

When printing the driver_override parameter when it is 4095 and 4094
bytes long, the printing code would access invalid memory because we
need count + 1 bytes for printing.

Cfr. commits 4efe874aace57dba ("PCI: Don't read past the end of sysfs
"driver_override" buffer") and bf563b01c2895a4b ("driver core: platform:
Don't read past the end of "driver_override" buffer").

Fixes: 3cf385713460eb2b ("ARM: 8256/1: driver coamba: add device binding path 'driver_override'")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 drivers/amba/bus.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/amba/bus.c b/drivers/amba/bus.c
index 36c5653ced5742b7..4a3ac31c07d0ee49 100644
--- a/drivers/amba/bus.c
+++ b/drivers/amba/bus.c
@@ -84,7 +84,8 @@ static ssize_t driver_override_store(struct device *_dev,
 	struct amba_device *dev = to_amba_device(_dev);
 	char *driver_override, *old, *cp;
 
-	if (count > PATH_MAX)
+	/* We need to keep extra room for a newline */
+	if (count >= (PAGE_SIZE - 1))
 		return -EINVAL;
 
 	driver_override = kstrndup(buf, count, GFP_KERNEL);
-- 
2.7.4

[PATCH 4/4] ARM: amba: Fix wrong indentation in driver_override_store()

[Geert Uytterhoeven]

Indentation is one TAB and 7 spaces instead of 2 TABs.

Fixes: 3cf385713460eb2b ("ARM: 8256/1: driver coamba: add device binding path 'driver_override'")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 drivers/amba/bus.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/amba/bus.c b/drivers/amba/bus.c
index 4a3ac31c07d0ee49..842314a439fdd4eb 100644
--- a/drivers/amba/bus.c
+++ b/drivers/amba/bus.c
@@ -101,8 +101,8 @@ static ssize_t driver_override_store(struct device *_dev,
 	if (strlen(driver_override)) {
 		dev->driver_override = driver_override;
 	} else {
-	       kfree(driver_override);
-	       dev->driver_override = NULL;
+		kfree(driver_override);
+		dev->driver_override = NULL;
 	}
 	device_unlock(_dev);
 
-- 
2.7.4