[PATCH 1/2] backports: add definition for USB_SPEED_SUPER_PLUS

[PATCH 1/2] backports: add definition for USB_SPEED_SUPER_PLUS

[Arend van Spriel]

Speed definition for usb3.1 has been added in 4.6 kernel. Add
definiton for older kernels.

Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
---
 backport/backport-include/linux/usb/ch9.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/backport/backport-include/linux/usb/ch9.h b/backport/backport-include/linux/usb/ch9.h
index 135e24c..c42fc30 100644
--- a/backport/backport-include/linux/usb/ch9.h
+++ b/backport/backport-include/linux/usb/ch9.h
@@ -21,4 +21,8 @@ static inline int usb_endpoint_maxp(const struct usb_endpoint_descriptor *epd)
 }
 #endif /* < 3.2 */
 
+#if LINUX_VERSION_CODE < KERNEL_VERSION(4,6,0)
+#define USB_SPEED_SUPER_PLUS	(USB_SPEED_SUPER + 1)
+#endif /* < 4.6 */
+
 #endif /* __BACKPORT__LINUX_USB_CH9_H */
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe backports" in

[PATCH 2/2] backports: patches: update devcoredump patch

[Arend van Spriel]

Due to changes in include/linux/devcoredump.h the patch needs to
be updated.

Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
---
This patch is another shot at devcoredump patch. It just fixes
applying the patch. However, the resulting file now uses the
scatterlist function sg_pcopy_to_buffer() so that will need a
backport for kernels before v3.11.

Regards,
Arend
---
 patches/backport-adjustments/devcoredump.patch | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/patches/backport-adjustments/devcoredump.patch b/patches/backport-adjustments/devcoredump.patch
index b438f32..397b317 100644
--- a/patches/backport-adjustments/devcoredump.patch
+++ b/patches/backport-adjustments/devcoredump.patch
@@ -97,18 +97,18 @@
 -__exitcall(devcoredump_exit);
 --- a/include/linux/backport-devcoredump.h
 +++ b/include/linux/backport-devcoredump.h
-@@ -5,7 +5,7 @@
- #include <linux/module.h>
- #include <linux/vmalloc.h>
+@@ -66,7 +66,7 @@ static inline void _devcd_free_sgtable(struct scatterlist *tab
+ }
+ 
  
 -#ifdef CONFIG_DEV_COREDUMP
 +#ifdef CPTCFG_BPAUTO_WANT_DEV_COREDUMP
- void dev_coredumpv(struct device *dev, const void *data, size_t datalen,
- 		   gfp_t gfp);
+ void dev_coredumpv(struct device *dev, void *data, size_t datalen,
+                   gfp_t gfp);
  
-@@ -30,6 +30,6 @@ dev_coredumpm(struct device *dev, struct
+@@ -100,6 +100,6 @@ static inline void dev_coredumpsg(struct device *dev, struct
  {
- 	free(data);
+        _devcd_free_sgtable(table);
  }
 -#endif /* CONFIG_DEV_COREDUMP */
 +#endif /* CPTCFG_BPAUTO_WANT_DEV_COREDUMP */
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe backports" in

Re: [PATCH 2/2] backports: patches: update devcoredump patch

[Johannes Berg]

On Sun, 2016-06-05 at 11:30 +0200, Arend van Spriel wrote:
> Due to changes in include/linux/devcoredump.h the patch needs to
> be updated.
> 
> Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
> ---
> This patch is another shot at devcoredump patch. It just fixes
> applying the patch. However, the resulting file now uses the
> scatterlist function sg_pcopy_to_buffer() so that will need a
> backport for kernels before v3.11.
> 
I think, in order to make the new dev_coredumpsg() [or whatever it's
called] function generally available, we have to also adjust the kernel
version where the backported one gets used.

johannes
--
To unsubscribe from this list: send the line "unsubscribe backports" in

Re: [PATCH 2/2] backports: patches: update devcoredump patch

[Arend Van Spriel]

On 5-6-2016 17:00, Johannes Berg wrote:
> On Sun, 2016-06-05 at 11:30 +0200, Arend van Spriel wrote:
>> Due to changes in include/linux/devcoredump.h the patch needs to
>> be updated.
>>
>> Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
>> ---
>> This patch is another shot at devcoredump patch. It just fixes
>> applying the patch. However, the resulting file now uses the
>> scatterlist function sg_pcopy_to_buffer() so that will need a
>> backport for kernels before v3.11.
>>
> I think, in order to make the new dev_coredumpsg() [or whatever it's
> called] function generally available, we have to also adjust the kernel
> version where the backported one gets used.

True. As there are no users right now we could opt to remove the new
function using devcoredump.patch. Or can we expect a driver using it soon.

Regards,
Arend
--
To unsubscribe from this list: send the line "unsubscribe backports" in

Re: [PATCH 2/2] backports: patches: update devcoredump patch

[Johannes Berg]

> > I think, in order to make the new dev_coredumpsg() [or whatever
> > it's
> > called] function generally available, we have to also adjust the
> > kernel
> > version where the backported one gets used.
> True. As there are no users right now we could opt to remove the new
> function using devcoredump.patch. Or can we expect a driver using it
> soon.
> 

I have a patch for our driver to use it, which is basically only
waiting for this to get into backports :-)

johannes
--
To unsubscribe from this list: send the line "unsubscribe backports" in

Re: [PATCH 2/2] backports: patches: update devcoredump patch

[Arend Van Spriel]

On 6-6-2016 11:00, Johannes Berg wrote:
> 
>>> I think, in order to make the new dev_coredumpsg() [or whatever
>>> it's
>>> called] function generally available, we have to also adjust the
>>> kernel
>>> version where the backported one gets used.
>> True. As there are no users right now we could opt to remove the new
>> function using devcoredump.patch. Or can we expect a driver using it
>> soon.
>>
> 
> I have a patch for our driver to use it, which is basically only
> waiting for this to get into backports :-)

So how about incorporating lib/scatterlist.c in backports?

Regards,
Arend
--
To unsubscribe from this list: send the line "unsubscribe backports" in

Re: [PATCH 2/2] backports: patches: update devcoredump patch

[Johannes Berg]

> > I have a patch for our driver to use it, which is basically only
> > waiting for this to get into backports :-)
> So how about incorporating lib/scatterlist.c in backports?
> 

We need the whole thing? I haven't really had time to look into this,
sorry.

johannes
--
To unsubscribe from this list: send the line "unsubscribe backports" in

Re: [PATCH 2/2] backports: patches: update devcoredump patch

[Arend Van Spriel]

/On 6-6-2016 11:12, Johannes Berg wrote:
> 
>>> I have a patch for our driver to use it, which is basically only
>>> waiting for this to get into backports :-)
>> So how about incorporating lib/scatterlist.c in backports?
>>
> 
> We need the whole thing? I haven't really had time to look into this,
> sorry.

Here call chain:

dev_coredumpsg() ->
	sg_pcopy_to_buffer() ->
		sg_copy_buffer() ->
			sg_miter_{start,skip,next,stop}() ->
				sg_miter_get_next_page()
				sg_page_iter_page()
				__sg_page_iter_next()
				__sg_page_iter_start()

So not everything is needed. About 350 lines from the 750 lines.

Regards,
Arend
--
To unsubscribe from this list: send the line "unsubscribe backports" in

Re: [PATCH 1/2] backports: add definition for USB_SPEED_SUPER_PLUS

[Hauke Mehrtens]

On 06/05/2016 11:30 AM, Arend van Spriel wrote:
> Speed definition for usb3.1 has been added in 4.6 kernel. Add
> definiton for older kernels.
> 
> Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
> ---
>  backport/backport-include/linux/usb/ch9.h | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/backport/backport-include/linux/usb/ch9.h b/backport/backport-include/linux/usb/ch9.h
> index 135e24c..c42fc30 100644
> --- a/backport/backport-include/linux/usb/ch9.h
> +++ b/backport/backport-include/linux/usb/ch9.h
> @@ -21,4 +21,8 @@ static inline int usb_endpoint_maxp(const struct usb_endpoint_descriptor *epd)
>  }
>  #endif /* < 3.2 */
>  
> +#if LINUX_VERSION_CODE < KERNEL_VERSION(4,6,0)
> +#define USB_SPEED_SUPER_PLUS	(USB_SPEED_SUPER + 1)
> +#endif /* < 4.6 */
> +
>  #endif /* __BACKPORT__LINUX_USB_CH9_H */
> 

This is in a emum in the mainline kernel. I think this will cause some
warnings when backporting it this way.

Hauke
--
To unsubscribe from this list: send the line "unsubscribe backports" in

Re: [PATCH 1/2] backports: add definition for USB_SPEED_SUPER_PLUS

[Arend Van Spriel]

On 10-7-2016 23:16, Hauke Mehrtens wrote:
> On 06/05/2016 11:30 AM, Arend van Spriel wrote:
>> Speed definition for usb3.1 has been added in 4.6 kernel. Add
>> definiton for older kernels.
>>
>> Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
>> ---
>>  backport/backport-include/linux/usb/ch9.h | 4 ++++
>>  1 file changed, 4 insertions(+)
>>
>> diff --git a/backport/backport-include/linux/usb/ch9.h b/backport/backport-include/linux/usb/ch9.h
>> index 135e24c..c42fc30 100644
>> --- a/backport/backport-include/linux/usb/ch9.h
>> +++ b/backport/backport-include/linux/usb/ch9.h
>> @@ -21,4 +21,8 @@ static inline int usb_endpoint_maxp(const struct usb_endpoint_descriptor *epd)
>>  }
>>  #endif /* < 3.2 */
>>  
>> +#if LINUX_VERSION_CODE < KERNEL_VERSION(4,6,0)
>> +#define USB_SPEED_SUPER_PLUS	(USB_SPEED_SUPER + 1)
>> +#endif /* < 4.6 */
>> +
>>  #endif /* __BACKPORT__LINUX_USB_CH9_H */
>>
> 
> This is in a emum in the mainline kernel. I think this will cause some
> warnings when backporting it this way.

The way it is used in brcmfmac I do not get a warning, but I agree it
would not work if this value is used calling a function that expects
enum usb_device_speed. So how would you backport this?

Regards,
Arend
--
To unsubscribe from this list: send the line "unsubscribe backports" in