[Buildroot] NPTL compile error without MMU (ARM)

[Buildroot] NPTL compile error without MMU (ARM)

[Sangjung]

Dear developers.

Hi. I found out that nptl make a compile error when MMU (i.e. 'Toolchain 
- Enable MMU support')  is disabled.

Detailed error message is as below.

-------------------------------------------------------------------------------- 

In file included from 
libpthread/nptl/sysdeps/unix/sysv/linux/arm/fork.c:31:0:
libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:61:17: error: 
'fork' undeclared here (not in a function)
libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:63:1: error: 
'__libc_fork' redeclared as different kind of symbol
libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:61:23: note: 
previous declaration of '__libc_fork' was here
libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:228:1: error: 
'fork' redeclared as different kind of symbol
libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:227:1: note: 
previous declaration of 'fork' was here
libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:229:1: error: 
'__EI_fork' aliased to undefined symbol '__GI_fork'
make[1]: *** [libpthread/nptl/sysdeps/unix/sysv/linux/arm/fork.os] Error 1
-------------------------------------------------------------------------------- 



This problem is already reported by Axel Lin.
(Link: http://lists.uclibc.org/pipermail/uclibc-cvs/2013-May/031094.html)

However, it looks like that this issue is not solved for a long time.
If someone let me know the alternative solution, I'd be really happy.

Thank you in advances.

BRs,
Sangjung

[Buildroot] NPTL compile error without MMU (ARM)

[Thomas Petazzoni]

Dear Sangjung,

On Thu, 27 Mar 2014 15:33:41 +0900, Sangjung wrote:
> Dear developers.
> 
> Hi. I found out that nptl make a compile error when MMU (i.e. 'Toolchain 
> - Enable MMU support')  is disabled.
> 
> Detailed error message is as below.
> 
> -------------------------------------------------------------------------------- 
> 
> In file included from 
> libpthread/nptl/sysdeps/unix/sysv/linux/arm/fork.c:31:0:
> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:61:17: error: 
> 'fork' undeclared here (not in a function)
> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:63:1: error: 
> '__libc_fork' redeclared as different kind of symbol
> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:61:23: note: 
> previous declaration of '__libc_fork' was here
> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:228:1: error: 
> 'fork' redeclared as different kind of symbol
> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:227:1: note: 
> previous declaration of 'fork' was here
> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:229:1: error: 
> '__EI_fork' aliased to undefined symbol '__GI_fork'
> make[1]: *** [libpthread/nptl/sysdeps/unix/sysv/linux/arm/fork.os] Error 1
> -------------------------------------------------------------------------------- 
> 
> This problem is already reported by Axel Lin.
> (Link: http://lists.uclibc.org/pipermail/uclibc-cvs/2013-May/031094.html)
> 
> However, it looks like that this issue is not solved for a long time.
> If someone let me know the alternative solution, I'd be really happy.

Thanks for your report. The problem really looks like a uClibc problem,
so I would suggest you to take the uClibc configuration file in
output/build/uclibc-<version>/.config, and sent the report to the
uClibc developers. I am also on the uClibc mailing list, so I will be
able to add additional Buildroot-specific details if requested by the
uClibc developers.

Outside of this, it is interesting to see that you are looking at the
ARM non-MMU support. Do you have an actual hardware platform? In fact,
I have been interested by looking into ARM non-MMU support in Buildroot
for a long time, but I have never found a hardware platform, or
emulated platform, to test things on. If this is something you can help
with, I'm pretty sure the ARM non-MMU support would make more
significant progress.

Thanks!

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

[Buildroot] NPTL compile error without MMU (ARM)

[Sangjung]

Dear Thomas.


On 03/27/2014 03:40 PM, Thomas Petazzoni wrote:
> Thanks for your report. The problem really looks like a uClibc problem,
> so I would suggest you to take the uClibc configuration file in
> output/build/uclibc-<version>/.config, and sent the report to the
> uClibc developers. I am also on the uClibc mailing list, so I will be
> able to add additional Buildroot-specific details if requested by the
> uClibc developers.

Thank you for your quick reply.
I will also report this issue to uClibc mailing list ASAP.

>
> Outside of this, it is interesting to see that you are looking at the
> ARM non-MMU support. Do you have an actual hardware platform? In fact,
> I have been interested by looking into ARM non-MMU support in Buildroot
> for a long time, but I have never found a hardware platform, or
> emulated platform, to test things on. If this is something you can help
> with, I'm pretty sure the ARM non-MMU support would make more
> significant progress.

Our target device is for very small wearable thing.
Because of various issue such as power consumption and cost,
ARM Cortex M (that is MMU-less) series are commonly used.


BRs
Sangjung

[Buildroot] NPTL compile error without MMU (ARM)

[Thomas Petazzoni]

Dear Sangjung,

On Thu, 27 Mar 2014 16:11:29 +0900, Sangjung wrote:

> On 03/27/2014 03:40 PM, Thomas Petazzoni wrote:
> > Thanks for your report. The problem really looks like a uClibc problem,
> > so I would suggest you to take the uClibc configuration file in
> > output/build/uclibc-<version>/.config, and sent the report to the
> > uClibc developers. I am also on the uClibc mailing list, so I will be
> > able to add additional Buildroot-specific details if requested by the
> > uClibc developers.
> 
> Thank you for your quick reply.
> I will also report this issue to uClibc mailing list ASAP.

Excellent, thanks!

> > Outside of this, it is interesting to see that you are looking at the
> > ARM non-MMU support. Do you have an actual hardware platform? In fact,
> > I have been interested by looking into ARM non-MMU support in Buildroot
> > for a long time, but I have never found a hardware platform, or
> > emulated platform, to test things on. If this is something you can help
> > with, I'm pretty sure the ARM non-MMU support would make more
> > significant progress.
> 
> Our target device is for very small wearable thing.
> Because of various issue such as power consumption and cost,
> ARM Cortex M (that is MMU-less) series are commonly used.

So I suppose these platforms have external RAM and Flash, right?
Because most of the Cortex-M platforms I've seen have only the internal
RAM and Flash of the Cortex-M processor, and the amount available
internally is usually not sufficient to run Linux.

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

[Buildroot] NPTL compile error without MMU (ARM)

[Sangjung Woo]

Dear Thomas


On 03/27/2014 03:40 PM, Thomas Petazzoni wrote:
> Dear Sangjung,
>
> On Thu, 27 Mar 2014 15:33:41 +0900, Sangjung wrote:
>> Dear developers.
>>
>> Hi. I found out that nptl make a compile error when MMU (i.e. 'Toolchain
>> - Enable MMU support')  is disabled.
>>
>> Detailed error message is as below.
>>
>> --------------------------------------------------------------------------------
>>
>> In file included from
>> libpthread/nptl/sysdeps/unix/sysv/linux/arm/fork.c:31:0:
>> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:61:17: error:
>> 'fork' undeclared here (not in a function)
>> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:63:1: error:
>> '__libc_fork' redeclared as different kind of symbol
>> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:61:23: note:
>> previous declaration of '__libc_fork' was here
>> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:228:1: error:
>> 'fork' redeclared as different kind of symbol
>> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:227:1: note:
>> previous declaration of 'fork' was here
>> libpthread/nptl/sysdeps/unix/sysv/linux/arm/../fork.c:229:1: error:
>> '__EI_fork' aliased to undefined symbol '__GI_fork'
>> make[1]: *** [libpthread/nptl/sysdeps/unix/sysv/linux/arm/fork.os] Error 1
>> --------------------------------------------------------------------------------
>>
>> This problem is already reported by Axel Lin.
>> (Link: http://lists.uclibc.org/pipermail/uclibc-cvs/2013-May/031094.html)
>>
>> However, it looks like that this issue is not solved for a long time.
>> If someone let me know the alternative solution, I'd be really happy.
> Thanks for your report. The problem really looks like a uClibc problem,
> so I would suggest you to take the uClibc configuration file in
> output/build/uclibc-<version>/.config, and sent the report to the
> uClibc developers. I am also on the uClibc mailing list, so I will be
> able to add additional Buildroot-specific details if requested by the
> uClibc developers.


After looking into this issue for a while,
in my opinion, buildroot generated the wrong configuration for uclibc.

I also found out that the _SAME_ problem already reported in _uclibc_ 
Bugzilla.
(URL: https://bugs.uclibc.org/show_bug.cgi?id=6224)


What if you want to reproduce this error,
'qemu_arm_versatile_defconfig' could be used as below.
Then, you will face the same errors that I reported.

---------------------------------------------
$ make qemu_arm_versatile_defconfig
Disable MMU (i.e. 'BR2_USE_MMU is not set' in .config)
$ make
---------------------------------------------


If I misunderstood something, let me know it.

BRs
Sangjung.