[alsa-devel] REGRESSION: Webcam mic for 145f:015b not longer working with 1.0.24 and above (OK with 1.0.23)

[Daniel Mack]

On 29.08.2012 20:59, Kostas wrote:
> On 29/08/2012 07:21 μμ, Daniel Mack wrote:
>> On 29.08.2012 18:03, Kostas wrote:
>>> On 29/08/2012 03:12 μμ, Daniel Mack wrote:
>>>> On 29.08.2012 14:04, Kostas wrote:
>>>>> On 29/08/2012 09:23 πμ, Daniel Mack wrote:
>>>>>> On 29.08.2012 02:13, Kostas wrote:
>>>> [...]
>>>>
>>>>>>> Command git bisect good 1.0.23 gives me the output
>>>>>>> fatal: Needed a single revision
>>>>>>> Bad rev input: 1.0.23
>>>>>>>
>>>>>>> How can i figure out the revision number?
>>>>>>>
>>>>>> When working with the kernel git, you have to take the kernel version
>>>>>> numbers.
>>>>>>
>>>>> Sorry i really can't understand what to do.
>>>>> The steps i did were
>>>>> git clone git://git.alsa-project.org/alsa-kernel.git alsa-kernel
>>>>> git bisect start alsa-kernel/
>>>>> git bisect bad
>>>>> git bisect good (Here i don't know what number to give)
>>>>>
>>>>> I will keep reading because the whole thing is interesting to me...but
>>>>> for now i can't understand anything.
>>>> No problem. According to your dmesg logs, kernel 2.6.38 was working for
>>>> you, and 3.2.0 isn't.
>>>>
>>>> First, clone the mainline Linux git:
>>>>
>>>>    git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux
>>>>    cd linux
>>>>
>>>> Test again that 2.6.38 is working:
>>>>
>>>>    git checkout v2.6.38
>>>>    compile + test
>>>>
>>>> test again that 3.2.0 is not working:
>>>>
>>>>    git checkout v3.2
>>>>    compile and test
>>>>
>>>> Then do the bisect:
>>>>
>>>>    git bisect start
>>>>    git bisect good v2.6.38
>>>>    git bisect bad v3.2
>>>>
>>>> Compile and test each step the bisect drops you to, and then call either
>>>> "git bisect good" or "git bisect bad". If a step cannot be compiled or
>>>> tested, call "git bisect skip".
>>>>
>>>> If all goes well, you will be told which commit is the first one that
>>>> does not work anymore.
>>>>
>>>>
>>>> Daniel
>>>>
>>>>
>>>>
>>>>
>>> I can't compile the kernel from git source.
>>> After i enter the command|CONCURRENCY_LEVEL=`getconf _NPROCESSORS_ONLN`
>>> fakeroot make-kpkg --initrd --append-to-version=-custom kernel_image
>>> kernel_headers
>>> |i get after a while this error
>>>
>>> /tmp/ccCeI2GY.s: Assembler messages:
>>> /tmp/ccCeI2GY.s: Error: .size expression for do_hypervisor_callback does
>>> not evaluate to a constant
>>> make[3]: *** [arch/x86/kernel/entry_64.o] Error 1
>>> make[2]: *** [arch/x86/kernel] Error 2
>>> make[1]: *** [arch/x86] Error 2
>>> make[1]: *** Waiting for unfinished jobs....
>>>    LD      init/mounts.o
>>>    LD      init/built-in.o
>>> make[1]: Leaving directory `/home/kostas/linux'
>>> make: *** [debian/stamp/build/kernel] Error 2
>>>
>>>
>>>
>> Set CONFIG_XEN=n in your .config
>>
>>
>>
> So many hours in front of the computer and still can't compile the 
> kernel...now it gives me the error
> 
> ld: drivers/watchdog/wm8350_wdt.o: bad reloc symbol index (0x20000033 >= 
> 0x50) for offset 0x7a28 in section `.debug_info'
> drivers/watchdog/wm8350_wdt.o: could not read symbols: Bad value
> make[2]: *** [drivers/watchdog/wm8350_wdt.ko] Error 1
> make[2]: *** Waiting for unfinished jobs....
> make[1]: *** [modules] Error 2
> make[1]: Leaving directory `/home/kostas/linux'
> make: *** [debian/stamp/build/kernel] Error 2

Yes, it can be hard at times, as you have to fight with all sorts of
build breakage.

> Advice me if you can what to do...i started to have a headache.
> Will continue tomorrow...

Run "make localmodconfig". That will generate a .config based on the
modules currently loaded in your kernel. That will also speed up the
bisect as way less drivers have to be built.


HTH,
Daniel