[meta-intel] [PATCH 3/3] lttng-modules: add bbappend to address 3.19.5 update

[Darren Hart]

On 5/12/15, 1:35 PM, "Saul Wold" <sgw at linux.intel.com> wrote:

>On 05/12/2015 01:27 PM, Darren Hart wrote:
>> On 5/11/15, 5:23 PM, "Saul Wold" <sgw at linux.intel.com> wrote:
>>
>>> The 3.19.5 kernel update included an incompatible change that caused
>>> lttng-modules
>>> to fail to build, the patch needed to be rebased due to changes in the
>>> associated
>>> Makefile.
>>>
>>> Signed-off-by: Saul Wold <sgw at linux.intel.com>
>>> ---
>>> ...modules-replace-KERNELDIR-with-KERNEL_SRC.patch | 71
>>> ++++++++++++++++++++++
>>> .../lttng/lttng-modules_2.6.%.bbappend             |  9 +++
>>> 2 files changed, 80 insertions(+)
>>> create mode 100644
>>> 
>>>common/recipes-kernel/lttng/lttng-modules/lttng-modules-replace-KERNELDI
>>>R-
>>> with-KERNEL_SRC.patch
>>> create mode 100644
>>> common/recipes-kernel/lttng/lttng-modules_2.6.%.bbappend
>>>
>>> diff --git
>>> 
>>>a/common/recipes-kernel/lttng/lttng-modules/lttng-modules-replace-KERNEL
>>>DI
>>> R-with-KERNEL_SRC.patch
>>> 
>>>b/common/recipes-kernel/lttng/lttng-modules/lttng-modules-replace-KERNEL
>>>DI
>>> R-with-KERNEL_SRC.patch
>>> new file mode 100644
>>> index 0000000..ccc6c54
>>> --- /dev/null
>>> +++
>>> 
>>>b/common/recipes-kernel/lttng/lttng-modules/lttng-modules-replace-KERNEL
>>>DI
>>> R-with-KERNEL_SRC.patch
>>> @@ -0,0 +1,71 @@
>>> +Upstream-Status: Inappropriate [embedded specific]
>>> +
>>> +lttng-modules: replace KERNELDIR with KERNEL_SRC
>>
>> Interesting, I thought we had something in place that would allow for
>>the
>> use of either or similar variables in the build. Apparently not.
>>
>> This works - but is fragile with version bumps. One thought would be to
>> define the KERNELDIR variable via the recipe based on KERNEL_SRC (via
>> EXTRA_OEMAKE), rather than patch the project Makefile. Was this approach
>> considered?
>>
>This was an existing patch that I rebased, I started to look into that
>with setting LTTNG_KERNELDIR in the recipe, but I got a failure so
>defaulted to the existing solution, I was going to circle back and look
>at this once I got meta-intel out.

Ah, I missed that.

Reviewed-by: Darren Hart <dvhart at linux.intel.com>

(same for the rest of this series)

Thanks Sau!

-- 
Darren Hart
Intel Open Source Technology Center