[PATCH v2 0/2] mfd: viperboard: cleanups and fix for cache line sharing

[PATCH v2 0/2] mfd: viperboard: cleanups and fix for cache line sharing

[Octavian Purdila]

Changes since v1:

 * split cache sharing fix and cleanups into separate patches

Octavian Purdila (2):
  mfd: viperboard: allocate I/O buffer separately
  mfd: viperboard: remove redundant OOM message and NULL pointer check

 drivers/mfd/viperboard.c       | 16 ++++++++++------
 include/linux/mfd/viperboard.h |  2 +-
 2 files changed, 11 insertions(+), 7 deletions(-)

-- 
1.9.1

[PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Octavian Purdila]

Currently the I/O buffer is allocated part of the device status
structure, potentially sharing the same cache line with other members
in this structure.

Allocate the buffer separately, to avoid the I/O operations corrupting
the device status structure due to cache line sharing.

Compiled tested only as I don't have access to hardware.

Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
---
 drivers/mfd/viperboard.c       | 8 ++++++++
 include/linux/mfd/viperboard.h | 2 +-
 2 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
index e00f534..5f62f4e 100644
--- a/drivers/mfd/viperboard.c
+++ b/drivers/mfd/viperboard.c
@@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
 		return -ENOMEM;
 	}
 
+	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
+	if (vb->buf == NULL) {
+		ret = -ENOMEM;
+		goto error;
+	}
+
 	mutex_init(&vb->lock);
 
 	vb->usb_dev = usb_get_dev(interface_to_usbdev(interface));
@@ -105,6 +111,7 @@ static int vprbrd_probe(struct usb_interface *interface,
 error:
 	if (vb) {
 		usb_put_dev(vb->usb_dev);
+		kfree(vb->buf);
 		kfree(vb);
 	}
 
@@ -118,6 +125,7 @@ static void vprbrd_disconnect(struct usb_interface *interface)
 	mfd_remove_devices(&interface->dev);
 	usb_set_intfdata(interface, NULL);
 	usb_put_dev(vb->usb_dev);
+	kfree(vb->buf);
 	kfree(vb);
 
 	dev_dbg(&interface->dev, "disconnected\n");
diff --git a/include/linux/mfd/viperboard.h b/include/linux/mfd/viperboard.h
index 1934528..af928d0 100644
--- a/include/linux/mfd/viperboard.h
+++ b/include/linux/mfd/viperboard.h
@@ -103,7 +103,7 @@ struct vprbrd_i2c_addr_msg {
 struct vprbrd {
 	struct usb_device *usb_dev; /* the usb device for this device */
 	struct mutex lock;
-	u8 buf[sizeof(struct vprbrd_i2c_write_msg)];
+	u8 *buf;
 	struct platform_device pdev;
 };
 
-- 
1.9.1

[PATCH v2 2/2] mfd: viperboard: remove redundant OOM message and NULL pointer check

[Octavian Purdila]

Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
---
 drivers/mfd/viperboard.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)

diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
index 5f62f4e..d27c131 100644
--- a/drivers/mfd/viperboard.c
+++ b/drivers/mfd/viperboard.c
@@ -59,10 +59,8 @@ static int vprbrd_probe(struct usb_interface *interface,
 
 	/* allocate memory for our device state and initialize it */
 	vb = kzalloc(sizeof(*vb), GFP_KERNEL);
-	if (vb == NULL) {
-		dev_err(&interface->dev, "Out of memory\n");
+	if (vb == NULL)
 		return -ENOMEM;
-	}
 
 	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
 	if (vb->buf == NULL) {
@@ -109,11 +107,9 @@ static int vprbrd_probe(struct usb_interface *interface,
 	return 0;
 
 error:
-	if (vb) {
-		usb_put_dev(vb->usb_dev);
-		kfree(vb->buf);
-		kfree(vb);
-	}
+	usb_put_dev(vb->usb_dev);
+	kfree(vb->buf);
+	kfree(vb);
 
 	return ret;
 }
-- 
1.9.1

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Johan Hovold]

On Mon, Sep 22, 2014 at 10:39:18PM +0300, Octavian Purdila wrote:
> Currently the I/O buffer is allocated part of the device status
> structure, potentially sharing the same cache line with other members
> in this structure.
> 
> Allocate the buffer separately, to avoid the I/O operations corrupting
> the device status structure due to cache line sharing.
> 
> Compiled tested only as I don't have access to hardware.
> 
> Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
> ---
>  drivers/mfd/viperboard.c       | 8 ++++++++
>  include/linux/mfd/viperboard.h | 2 +-
>  2 files changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
> index e00f534..5f62f4e 100644
> --- a/drivers/mfd/viperboard.c
> +++ b/drivers/mfd/viperboard.c
> @@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
>  		return -ENOMEM;
>  	}
>  
> +	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
> +	if (vb->buf == NULL) {
> +		ret = -ENOMEM;
> +		goto error;

This will cause a kref imbalance as you have a usb_put_dev in error,
but haven't done the get yet.

> +	}
> +
>  	mutex_init(&vb->lock);
>  
>  	vb->usb_dev = usb_get_dev(interface_to_usbdev(interface));

Here's the get.

Johan

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Johan Hovold]

On Tue, Sep 23, 2014 at 09:35:41AM +0200, Johan Hovold wrote:
> On Mon, Sep 22, 2014 at 10:39:18PM +0300, Octavian Purdila wrote:
> > Currently the I/O buffer is allocated part of the device status
> > structure, potentially sharing the same cache line with other members
> > in this structure.
> > 
> > Allocate the buffer separately, to avoid the I/O operations corrupting
> > the device status structure due to cache line sharing.
> > 
> > Compiled tested only as I don't have access to hardware.
> > 
> > Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
> > ---
> >  drivers/mfd/viperboard.c       | 8 ++++++++
> >  include/linux/mfd/viperboard.h | 2 +-
> >  2 files changed, 9 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
> > index e00f534..5f62f4e 100644
> > --- a/drivers/mfd/viperboard.c
> > +++ b/drivers/mfd/viperboard.c
> > @@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
> >  		return -ENOMEM;
> >  	}
> >  
> > +	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
> > +	if (vb->buf == NULL) {
> > +		ret = -ENOMEM;
> > +		goto error;
> 
> This will cause a kref imbalance as you have a usb_put_dev in error,
> but haven't done the get yet.

Nevermind. This isn't problem as the usb device is null.

Haven't had my morning coffee yet. ;)

Johan

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Johan Hovold]

On Mon, Sep 22, 2014 at 10:39:18PM +0300, Octavian Purdila wrote:
> Currently the I/O buffer is allocated part of the device status
> structure, potentially sharing the same cache line with other members
> in this structure.
> 
> Allocate the buffer separately, to avoid the I/O operations corrupting
> the device status structure due to cache line sharing.
> 
> Compiled tested only as I don't have access to hardware.
> 
> Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>

Reviewed-by: Johan Hovold <johan@kernel.org>

> ---
>  drivers/mfd/viperboard.c       | 8 ++++++++
>  include/linux/mfd/viperboard.h | 2 +-
>  2 files changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
> index e00f534..5f62f4e 100644
> --- a/drivers/mfd/viperboard.c
> +++ b/drivers/mfd/viperboard.c
> @@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
>  		return -ENOMEM;
>  	}
>  
> +	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
> +	if (vb->buf == NULL) {
> +		ret = -ENOMEM;
> +		goto error;
> +	}
> +
>  	mutex_init(&vb->lock);
>  
>  	vb->usb_dev = usb_get_dev(interface_to_usbdev(interface));
> @@ -105,6 +111,7 @@ static int vprbrd_probe(struct usb_interface *interface,
>  error:
>  	if (vb) {
>  		usb_put_dev(vb->usb_dev);
> +		kfree(vb->buf);
>  		kfree(vb);
>  	}
>  
> @@ -118,6 +125,7 @@ static void vprbrd_disconnect(struct usb_interface *interface)
>  	mfd_remove_devices(&interface->dev);
>  	usb_set_intfdata(interface, NULL);
>  	usb_put_dev(vb->usb_dev);
> +	kfree(vb->buf);
>  	kfree(vb);
>  
>  	dev_dbg(&interface->dev, "disconnected\n");
> diff --git a/include/linux/mfd/viperboard.h b/include/linux/mfd/viperboard.h
> index 1934528..af928d0 100644
> --- a/include/linux/mfd/viperboard.h
> +++ b/include/linux/mfd/viperboard.h
> @@ -103,7 +103,7 @@ struct vprbrd_i2c_addr_msg {
>  struct vprbrd {
>  	struct usb_device *usb_dev; /* the usb device for this device */
>  	struct mutex lock;
> -	u8 buf[sizeof(struct vprbrd_i2c_write_msg)];
> +	u8 *buf;
>  	struct platform_device pdev;
>  };

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Lee Jones]

On Mon, 22 Sep 2014, Octavian Purdila wrote:

> Currently the I/O buffer is allocated part of the device status
> structure, potentially sharing the same cache line with other members
> in this structure.
> 
> Allocate the buffer separately, to avoid the I/O operations corrupting
> the device status structure due to cache line sharing.
> 
> Compiled tested only as I don't have access to hardware.
> 
> Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
> ---
>  drivers/mfd/viperboard.c       | 8 ++++++++
>  include/linux/mfd/viperboard.h | 2 +-
>  2 files changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
> index e00f534..5f62f4e 100644
> --- a/drivers/mfd/viperboard.c
> +++ b/drivers/mfd/viperboard.c
> @@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
>  		return -ENOMEM;
>  	}
>  
> +	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);

Can you obtain the 'struct device' first then use managed resources
(devm_*)?

> +	if (vb->buf == NULL) {
> +		ret = -ENOMEM;
> +		goto error;
> +	}
> +
>  	mutex_init(&vb->lock);
>  
>  	vb->usb_dev = usb_get_dev(interface_to_usbdev(interface));
> @@ -105,6 +111,7 @@ static int vprbrd_probe(struct usb_interface *interface,
>  error:
>  	if (vb) {
>  		usb_put_dev(vb->usb_dev);
> +		kfree(vb->buf);

Then you don't need this.

>  		kfree(vb);
>  	}
>  
> @@ -118,6 +125,7 @@ static void vprbrd_disconnect(struct usb_interface *interface)
>  	mfd_remove_devices(&interface->dev);
>  	usb_set_intfdata(interface, NULL);
>  	usb_put_dev(vb->usb_dev);
> +	kfree(vb->buf);

Or this.

>  	kfree(vb);
>  
>  	dev_dbg(&interface->dev, "disconnected\n");
> diff --git a/include/linux/mfd/viperboard.h b/include/linux/mfd/viperboard.h
> index 1934528..af928d0 100644
> --- a/include/linux/mfd/viperboard.h
> +++ b/include/linux/mfd/viperboard.h
> @@ -103,7 +103,7 @@ struct vprbrd_i2c_addr_msg {
>  struct vprbrd {
>  	struct usb_device *usb_dev; /* the usb device for this device */
>  	struct mutex lock;
> -	u8 buf[sizeof(struct vprbrd_i2c_write_msg)];
> +	u8 *buf;
>  	struct platform_device pdev;
>  };
>  

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog

Re: [PATCH v2 2/2] mfd: viperboard: remove redundant OOM message and NULL pointer check

[Lee Jones]

On Mon, 22 Sep 2014, Octavian Purdila wrote:

> Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
> ---
>  drivers/mfd/viperboard.c | 12 ++++--------
>  1 file changed, 4 insertions(+), 8 deletions(-)

Looks fine, but won't apply without patch 1.

> diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
> index 5f62f4e..d27c131 100644
> --- a/drivers/mfd/viperboard.c
> +++ b/drivers/mfd/viperboard.c
> @@ -59,10 +59,8 @@ static int vprbrd_probe(struct usb_interface *interface,
>  
>  	/* allocate memory for our device state and initialize it */
>  	vb = kzalloc(sizeof(*vb), GFP_KERNEL);
> -	if (vb == NULL) {
> -		dev_err(&interface->dev, "Out of memory\n");
> +	if (vb == NULL)
>  		return -ENOMEM;
> -	}
>  
>  	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
>  	if (vb->buf == NULL) {
> @@ -109,11 +107,9 @@ static int vprbrd_probe(struct usb_interface *interface,
>  	return 0;
>  
>  error:
> -	if (vb) {
> -		usb_put_dev(vb->usb_dev);
> -		kfree(vb->buf);
> -		kfree(vb);
> -	}
> +	usb_put_dev(vb->usb_dev);
> +	kfree(vb->buf);
> +	kfree(vb);
>  
>  	return ret;
>  }

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Johan Hovold]

On Wed, Sep 24, 2014 at 11:12:06AM +0100, Lee Jones wrote:
> On Mon, 22 Sep 2014, Octavian Purdila wrote:
> 
> > Currently the I/O buffer is allocated part of the device status
> > structure, potentially sharing the same cache line with other members
> > in this structure.
> > 
> > Allocate the buffer separately, to avoid the I/O operations corrupting
> > the device status structure due to cache line sharing.
> > 
> > Compiled tested only as I don't have access to hardware.
> > 
> > Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
> > ---
> >  drivers/mfd/viperboard.c       | 8 ++++++++
> >  include/linux/mfd/viperboard.h | 2 +-
> >  2 files changed, 9 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
> > index e00f534..5f62f4e 100644
> > --- a/drivers/mfd/viperboard.c
> > +++ b/drivers/mfd/viperboard.c
> > @@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
> >  		return -ENOMEM;
> >  	}
> >  
> > +	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
> 
> Can you obtain the 'struct device' first then use managed resources
> (devm_*)?

I think any devres conversion should be done in a follow-up patch and
not be included in the fix (e.g. in order to facilitate backporting). We
also don't want to mix allocation schemes.

Johan

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Lee Jones]

On Wed, 24 Sep 2014, Johan Hovold wrote:

> On Wed, Sep 24, 2014 at 11:12:06AM +0100, Lee Jones wrote:
> > On Mon, 22 Sep 2014, Octavian Purdila wrote:
> > 
> > > Currently the I/O buffer is allocated part of the device status
> > > structure, potentially sharing the same cache line with other members
> > > in this structure.
> > > 
> > > Allocate the buffer separately, to avoid the I/O operations corrupting
> > > the device status structure due to cache line sharing.
> > > 
> > > Compiled tested only as I don't have access to hardware.
> > > 
> > > Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
> > > ---
> > >  drivers/mfd/viperboard.c       | 8 ++++++++
> > >  include/linux/mfd/viperboard.h | 2 +-
> > >  2 files changed, 9 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
> > > index e00f534..5f62f4e 100644
> > > --- a/drivers/mfd/viperboard.c
> > > +++ b/drivers/mfd/viperboard.c
> > > @@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
> > >  		return -ENOMEM;
> > >  	}
> > >  
> > > +	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
> > 
> > Can you obtain the 'struct device' first then use managed resources
> > (devm_*)?
> 
> I think any devres conversion should be done in a follow-up patch and
> not be included in the fix (e.g. in order to facilitate backporting). We
> also don't want to mix allocation schemes.

I agree, but equally I'm not keen on accepting this patch as I believe
it could be done better.

Please submit two patches, one converting to shared resources and this
being the subsequent one, fixed up to do the right thing.

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Johan Hovold]

On Wed, Sep 24, 2014 at 01:00:02PM +0100, Lee Jones wrote:
> On Wed, 24 Sep 2014, Johan Hovold wrote:
> 
> > On Wed, Sep 24, 2014 at 11:12:06AM +0100, Lee Jones wrote:
> > > On Mon, 22 Sep 2014, Octavian Purdila wrote:
> > > 
> > > > Currently the I/O buffer is allocated part of the device status
> > > > structure, potentially sharing the same cache line with other members
> > > > in this structure.
> > > > 
> > > > Allocate the buffer separately, to avoid the I/O operations corrupting
> > > > the device status structure due to cache line sharing.
> > > > 
> > > > Compiled tested only as I don't have access to hardware.
> > > > 
> > > > Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
> > > > ---
> > > >  drivers/mfd/viperboard.c       | 8 ++++++++
> > > >  include/linux/mfd/viperboard.h | 2 +-
> > > >  2 files changed, 9 insertions(+), 1 deletion(-)
> > > > 
> > > > diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
> > > > index e00f534..5f62f4e 100644
> > > > --- a/drivers/mfd/viperboard.c
> > > > +++ b/drivers/mfd/viperboard.c
> > > > @@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
> > > >  		return -ENOMEM;
> > > >  	}
> > > >  
> > > > +	vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
> > > 
> > > Can you obtain the 'struct device' first then use managed resources
> > > (devm_*)?
> > 
> > I think any devres conversion should be done in a follow-up patch and
> > not be included in the fix (e.g. in order to facilitate backporting). We
> > also don't want to mix allocation schemes.
> 
> I agree, but equally I'm not keen on accepting this patch as I believe
> it could be done better.
>
> Please submit two patches, one converting to shared resources and this
> being the subsequent one, fixed up to do the right thing.

A buffer-corruption fix is a candidate for stable, whereas a devres
conversion (clean up) is not. Hence the former should not depend on the
latter.

Johan

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Octavian Purdila]

On Wed, Sep 24, 2014 at 3:26 PM, Johan Hovold <johan@kernel.org> wrote:
> On Wed, Sep 24, 2014 at 01:00:02PM +0100, Lee Jones wrote:
>> On Wed, 24 Sep 2014, Johan Hovold wrote:
>>
>> > On Wed, Sep 24, 2014 at 11:12:06AM +0100, Lee Jones wrote:
>> > > On Mon, 22 Sep 2014, Octavian Purdila wrote:
>> > >
>> > > > Currently the I/O buffer is allocated part of the device status
>> > > > structure, potentially sharing the same cache line with other members
>> > > > in this structure.
>> > > >
>> > > > Allocate the buffer separately, to avoid the I/O operations corrupting
>> > > > the device status structure due to cache line sharing.
>> > > >
>> > > > Compiled tested only as I don't have access to hardware.
>> > > >
>> > > > Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
>> > > > ---
>> > > >  drivers/mfd/viperboard.c       | 8 ++++++++
>> > > >  include/linux/mfd/viperboard.h | 2 +-
>> > > >  2 files changed, 9 insertions(+), 1 deletion(-)
>> > > >
>> > > > diff --git a/drivers/mfd/viperboard.c b/drivers/mfd/viperboard.c
>> > > > index e00f534..5f62f4e 100644
>> > > > --- a/drivers/mfd/viperboard.c
>> > > > +++ b/drivers/mfd/viperboard.c
>> > > > @@ -64,6 +64,12 @@ static int vprbrd_probe(struct usb_interface *interface,
>> > > >                 return -ENOMEM;
>> > > >         }
>> > > >
>> > > > +       vb->buf = kzalloc(sizeof(struct vprbrd_i2c_write_msg), GFP_KERNEL);
>> > >
>> > > Can you obtain the 'struct device' first then use managed resources
>> > > (devm_*)?
>> >
>> > I think any devres conversion should be done in a follow-up patch and
>> > not be included in the fix (e.g. in order to facilitate backporting). We
>> > also don't want to mix allocation schemes.
>>
>> I agree, but equally I'm not keen on accepting this patch as I believe
>> it could be done better.
>>
>> Please submit two patches, one converting to shared resources and this
>> being the subsequent one, fixed up to do the right thing.
>
> A buffer-corruption fix is a candidate for stable, whereas a devres
> conversion (clean up) is not. Hence the former should not depend on the
> latter.
>

I can follow-up with a v3 3 patch series: first for the fix, second
for the OOM & error path cleanup, third for devm conversion.

I think this will address both issues (backport and better final result).

Re: [PATCH v2 1/2] mfd: viperboard: allocate I/O buffer separately

[Johan Hovold]

On Wed, Sep 24, 2014 at 03:34:08PM +0300, Octavian Purdila wrote:

> I can follow-up with a v3 3 patch series: first for the fix, second
> for the OOM & error path cleanup, third for devm conversion.

I'd include the error-path clean up bit in the devres conversion as that
is what it's really all about.

Johan