Natty ti-omap4

[Bryan Wu]

On Tue, Dec 7, 2010 at 11:24 PM, Tim Gardner <tim.gardner at canonical.com> wrote:
> On 12/07/2010 07:09 AM, Tim Gardner wrote:
>>
>> On 12/06/2010 09:58 PM, Bryan Wu wrote:
>>>
>>> On Tue, Dec 7, 2010 at 7:59 AM, Tim Gardner<tcanonical at tpi.com>   wrote:
>>>>
>>>> On 12/06/2010 04:29 PM, Dechesne, Nicolas wrote:
>>>>>
>>>>> Tim,
>>>>>
>>>>> On Mon, Dec 6, 2010 at 10:59 PM, Tim Gardner<tim.gardner at canonical.com>
>>>>>   wrote:
>>>>>>
>>>>>> Bryan - I pushed a natty ti-omap4 branch which was initially based on
>>>>>> Maverick ti-omap4 2.6.35-903.19. The new package version is
>>>>>> 2.6.35-1100.1, but I'm still working on some build problems. I should
>>>>>
>>>>> is there any difference between -903.19 and -1100.1? Can you explain
>>>>> what you've done?
>>>>>
>>>>
>>>> No difference between Maverick and Natty other then packaging. The ABI
>>>> has
>>>> to be different so that there is no conflict with binary package names.
>>>>
>>>
>>> Great, that's a good start point for us. I've found the reason of the
>>> building failure.
>>> There is no 1100.0 change item in debian.ti-omap4/changelog.
>>>
>>>>>> have it fixed tomorrow.
>>>
>>> I think you will fix it, since it needs to revert the 1100.1 commit.
>>>
>>>>>>
>>>>>> Any status updates on rolling ti-omap4 forward to 2.6.37?
>>>>>
>>>>> we don't have, and won't have a .37 BSP for OMAP4. TI is focusing on
>>>>> .35 for now because we are adding support for power management. the
>>>>> current plan is for us to deliver a new .35 kernel with power
>>>>> management for inclusion in natty. Sebastien will be working with
>>>>> Bryan to make this. We will have everything ported over to .38 in
>>>>> February, and we will submit pull request for natty at that time.
>>>>>
>>>>
>>>> OK, I guess we'll just carry the .35 kernel versions until y'all are
>>>> ready
>>>> with .38
>>>>
>>>
>>> Thanks Nico and Tim, I will try to handle that with help from Sebastien.
>>>
>>
>> OK, I banged on HEAD a little, so you'll have to reset --hard. It now
>> builds correctly.
>>
>> rtg
>
> Alas, a tools issue that does not show up with the cross compiler:
>
> http://launchpadlibrarian.net/60273139/buildlog_ubuntu-natty-armel.linux-ti-omap4_2.6.35-1100.1_FAILEDTOBUILD.txt.gz
>

For this issue:
---
  AS      arch/arm/mach-omap2/omap-headsmp.o
/build/buildd/linux-ti-omap4-2.6.35/arch/arm/mach-omap2/omap-headsmp.S:
Assembler messages:
/build/buildd/linux-ti-omap4-2.6.35/arch/arm/mach-omap2/omap-headsmp.S:36:
Error: selected processor does not support ARM mode `smc #0'
make[3]: *** [arch/arm/mach-omap2/omap-headsmp.o] Error 1
make[2]: *** [arch/arm/mach-omap2] Error 2
make[1]: *** [sub-make] Error 2
make[1]: Leaving directory `/build/buildd/linux-ti-omap4-2.6.35'
make: *** [/build/buildd/linux-ti-omap4-2.6.35/debian/stamps/stamp-build-omap4]
Error 2
dpkg-buildpackage: error: debian/rules build gave error exit status 2
---
I think it's a issue of compiler, which is gcc 4.5.1 in Natty. When we
use gcc 4.5.1 to build 2.6.35 ARM kernel, it pops up. Gcc 4.5.1 can
build the 2.6.37 based Linaro kernel without this error.

I can build it in my Maverick machine with Linaro GCC 4.4.5 successfully.

I've pinged Marcin Juszkiewicz in the IRC about this and copied him
into this email.

Marcin, do you know any fix for this issue? Thanks.

Best Regards,
-- 
Bryan Wu <bryan.wu at canonical.com>
Kernel Developer    +86.138-1617-6545 Mobile
Ubuntu Kernel Team
Canonical Ltd.      www.canonical.com
Ubuntu - Linux for human beings | www.ubuntu.com