2.6.35.10-74 compilation (and build) problems

[JD]

On 01/07/2011 09:21 PM, JD wrote:
> On 01/07/2011 02:45 PM, JD wrote:
>> On 01/07/2011 11:15 AM, Don Zickus wrote:
>>> On Fri, Jan 07, 2011 at 01:36:59PM -0500, Jarod Wilson wrote:
>>>> I just implemented exactly what I described here locally, and it 
>>>> worked
>>>> perfectly. The config-local file needs to be a Source: file so that it
>>>> actually gets included in the source rpm, and I'm not sure why you're
>>>> doing anything touching debuginfo bits...
>>>>
>>>> My test case was to simply add CONFIG_VIDEO_VIA_CAMERA=y to my
>>>> config-local, as I know its disabled in the stock config, because it
>>>> doesn't build (due to a dependency on updated viafb bits in 2.6.37). I
>>>> created a source rpm with that modified config-local, then did an
>>>> rpmbuild, and it failed on drivers/media/video/via-camera.c exactly 
>>>> like
>>>> I'd expected it to. So I suspect you've either done something wrong 
>>>> with
>>>> your implementation or you execution. :)
>>> Jarod,
>>>
>>> Patch please, just don't make sure everyone is on the same page.
>>>
>>> JD,
>>>
>>> Can you provide your .config that you want to use to make sure we 
>>> aren't
>>> missing something obvious?
>>>
>>> Cheers,
>>> Don
>>
> Hi Don,
> Using the config file I already sent, I ran
> make all
>
> The make dies with error:
> .
> .
> .
> OBJCOPY arch/x86/vdso/vdso32-sysenter.so
> AS arch/x86/vdso/vdso32.o
> CC arch/x86/vdso/vdso32-setup.o
> VDSOSYM arch/x86/vdso/vdso32-int80-syms.lds
> VDSOSYM arch/x86/vdso/vdso32-sysenter-syms.lds
> VDSOSYM arch/x86/vdso/vdso32-syms.lds
> LD arch/x86/vdso/built-in.o
> LD arch/x86/built-in.o
> CC kernel/sched.o
> kernel/sched.c: In function ‘init_idle’:
> kernel/sched.c:5397: error: implicit declaration of function 
> ‘double_rq_lock’
> kernel/sched.c:5422: error: implicit declaration of function 
> ‘double_rq_unlock’
> make[1]: *** [kernel/sched.o] Error 1
> make: *** [kernel] Error 2
>
I re-ran make xconfig and re-enabled SMP
and restarted make all.
No compilation problems.
So, it seems that disabling SMP breaks the kernel compilation.
This is a rather loud evidence that the kernel devs have inserted
code (perhaps unwittingly) that is SMP dependent without proper
ifdefs for SMP and non-SMP.