[PATCH Vivid SRU] UBUNTU: [Config] CONFIG_RTC_DRV_XGENE=y for only arm64

Dann Frazier dann.frazier at canonical.com
Wed Sep 30 15:10:29 UTC 2015


On Wed, Sep 30, 2015 at 8:02 AM, Stefan Bader
<stefan.bader at canonical.com> wrote:
> On 30.09.2015 15:50, Tim Gardner wrote:
>> On 09/30/2015 12:23 AM, Stefan Bader wrote:
>>> On 29.09.2015 23:27, tim.gardner at canonical.com wrote:
>>>> From: Tim Gardner <tim.gardner at canonical.com>
>>>>
>>>> BugLink: http://bugs.launchpad.net/bugs/1499869
>>>>
>>>> This driver only works on arm64 platforms.
>>>>
>>>> Signed-off-by: Tim Gardner <tim.gardner at canonical.com> ---
>>>> debian.master/abi/3.19.0-30.33/arm64/generic.modules          | 1
>>>> - debian.master/abi/3.19.0-30.33/armhf/generic-lpae.modules     |
>>>> 1 - debian.master/abi/3.19.0-30.33/armhf/generic.modules
>>>> | 1 - debian.master/abi/3.19.0-30.33/i386/generic.modules
>>>> | 1 - debian.master/abi/3.19.0-30.33/i386/lowlatency.modules
>>>> | 1 -
>>>> debian.master/abi/3.19.0-30.33/powerpc/powerpc-e500mc.modules | 1
>>>> - debian.master/abi/3.19.0-30.33/powerpc/powerpc-smp.modules    |
>>>> 1 - debian.master/abi/3.19.0-30.33/powerpc/powerpc64-emb.modules
>>>> | 1 -
>>>> debian.master/abi/3.19.0-30.33/powerpc/powerpc64-smp.modules  | 1
>>>> - debian.master/abi/3.19.0-30.33/ppc64el/generic.modules        |
>>>> 1 -
>>>
>>> Suppose this is cheating it out of the previous ABI definition.
>>> Wondering whether adding ignore module files would not be
>>> cleaner...
>>>
>>
>> I cannot think of why either method would have a tactical or strategic
>> advantage. Therefore I went with the surgical approach so that the
>> changes made were explicit. In my opinion, an ignore file obfuscates
>> the solution.
>>
> I was somehow thinking along the lines of the contents of the modules list in
> that abi directory represent the state of that upload. Which still had the
> module(s). And the goal is to drop them with the next upload. That is why I
> thought using ignore files to document the expected "will be missing on the next
> build" might be better...

Would it be generally useful to also check modules.builtin to avoid
bailing in the module->static case, or is that something we want to
require an explicit metadata update?




More information about the kernel-team mailing list