[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Oleksandr Zhadan]

Back to the BR2_GCC_VERSION_4_9_X, because the default/latest GCC fails to build the Linux kernel.

Signed-off-by: Oleksandr Zhadan <oleks@arcturusnetworks.com>
---
 configs/arcturus_ucp1020_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/configs/arcturus_ucp1020_defconfig b/configs/arcturus_ucp1020_defconfig
index cc1297ce1..29e0f766b 100644
--- a/configs/arcturus_ucp1020_defconfig
+++ b/configs/arcturus_ucp1020_defconfig
@@ -10,6 +10,7 @@ BR2_TARGET_GENERIC_HOSTNAME="UCP1020"
 BR2_TARGET_GENERIC_ISSUE="Welcome to Arcturus uCP1020 System on Module"
 BR2_TARGET_GENERIC_GETTY_PORT="ttyS0"
 BR2_SYSTEM_DHCP="eth0"
+BR2_GCC_VERSION_4_9_X=y
 
 # Kernel
 BR2_LINUX_KERNEL=y
-- 
2.11.0

[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Thomas Petazzoni]

Hello,

On Wed, 12 Jul 2017 14:31:03 -0400, Oleksandr Zhadan wrote:

> Back to the BR2_GCC_VERSION_4_9_X, because the default/latest GCC fails to build the Linux kernel.
> 
> Signed-off-by: Oleksandr Zhadan <oleks@arcturusnetworks.com>

What failure do you have when building the Linux kernel ?

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

[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Oleksandr G Zhadan]

Hello Thomas,
On 07/13/2017 03:42 AM, Thomas Petazzoni wrote:
> Hello,
>
> On Wed, 12 Jul 2017 14:31:03 -0400, Oleksandr Zhadan wrote:
>
>> Back to the BR2_GCC_VERSION_4_9_X, because the default/latest GCC fails to build the Linux kernel.
>>
>> Signed-off-by: Oleksandr Zhadan <oleks@arcturusnetworks.com>
> What failure do you have when building the Linux kernel ?

In file included from include/linux/linkage.h:4:0,
                  from include/linux/kernel.h:6,
                  from arch/powerpc/kernel/ptrace.c:18:
arch/powerpc/kernel/ptrace.c: In function ?fpr_get?:
arch/powerpc/kernel/ptrace.c:379:24: error: index 32 denotes an offset 
greater than size of ?u64[32][1] {aka long long unsigned int[32][1]}? 
[-Werror=array-bounds]
         offsetof(struct thread_fp_state, fpr[32][0]));
....

arch/powerpc/kernel/ptrace.c: In function ?fpr_set?:
arch/powerpc/kernel/ptrace.c:407:24: error: index 32 denotes an offset 
greater than size of ?u64[32][1] {aka long long unsigned int[32][1]}? 
[-Werror=array-bounds]
         offsetof(struct thread_fp_state, fpr[32][0]));

>
> Thomas
P.S.
Did you have a chance to look into our "[Buildroot][PATCH v4 1/1] 
configs/arcturus_ucls1012a: new defconfig" patch from 07/06/2017 05:13 PM.

Thank you,

Oleks

[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Thomas Petazzoni]

Hello,

On Thu, 13 Jul 2017 09:31:25 -0400, Oleksandr G Zhadan wrote:

> >> Signed-off-by: Oleksandr Zhadan <oleks@arcturusnetworks.com>  
> > What failure do you have when building the Linux kernel ?  
> 
> In file included from include/linux/linkage.h:4:0,
>                   from include/linux/kernel.h:6,
>                   from arch/powerpc/kernel/ptrace.c:18:
> arch/powerpc/kernel/ptrace.c: In function ?fpr_get?:
> arch/powerpc/kernel/ptrace.c:379:24: error: index 32 denotes an offset 
> greater than size of ?u64[32][1] {aka long long unsigned int[32][1]}? 
> [-Werror=array-bounds]
>          offsetof(struct thread_fp_state, fpr[32][0]));
> ....
> 
> arch/powerpc/kernel/ptrace.c: In function ?fpr_set?:
> arch/powerpc/kernel/ptrace.c:407:24: error: index 32 denotes an offset 
> greater than size of ?u64[32][1] {aka long long unsigned int[32][1]}? 
> [-Werror=array-bounds]
>          offsetof(struct thread_fp_state, fpr[32][0]));

Then could you please backport Linux kernel commit
1e407ee3b21f981140491d5b8a36422979ca246f instead? It should fix this
build issue.

> > Thomas  
> P.S.
> Did you have a chance to look into our "[Buildroot][PATCH v4 1/1] 
> configs/arcturus_ucls1012a: new defconfig" patch from 07/06/2017 05:13 PM.

I had a brief look, it looked good. I'll review more carefully, and
hopefully apply it as time permits.

Best regards,

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

[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Oleksandr G Zhadan]

On 07/13/2017 09:38 AM, Thomas Petazzoni wrote:
> Hello,
>
> On Thu, 13 Jul 2017 09:31:25 -0400, Oleksandr G Zhadan wrote:
>
>>>> Signed-off-by: Oleksandr Zhadan <oleks@arcturusnetworks.com>
>>> What failure do you have when building the Linux kernel ?
>> In file included from include/linux/linkage.h:4:0,
>>                    from include/linux/kernel.h:6,
>>                    from arch/powerpc/kernel/ptrace.c:18:
>> arch/powerpc/kernel/ptrace.c: In function ?fpr_get?:
>> arch/powerpc/kernel/ptrace.c:379:24: error: index 32 denotes an offset
>> greater than size of ?u64[32][1] {aka long long unsigned int[32][1]}?
>> [-Werror=array-bounds]
>>           offsetof(struct thread_fp_state, fpr[32][0]));
>> ....
>>
>> arch/powerpc/kernel/ptrace.c: In function ?fpr_set?:
>> arch/powerpc/kernel/ptrace.c:407:24: error: index 32 denotes an offset
>> greater than size of ?u64[32][1] {aka long long unsigned int[32][1]}?
>> [-Werror=array-bounds]
>>           offsetof(struct thread_fp_state, fpr[32][0]));
> Then could you please backport Linux kernel commit
> 1e407ee3b21f981140491d5b8a36422979ca246f instead? It should fix this
> build issue.

Yes, I can, but we are using public kernel.
Should I add it as an extra patches to my 
board/arcturus/ppc-ucp1020/patches/linux/ directory ?

>>> Thomas
>> P.S.
>> Did you have a chance to look into our "[Buildroot][PATCH v4 1/1]
>> configs/arcturus_ucls1012a: new defconfig" patch from 07/06/2017 05:13 PM.
> I had a brief look, it looked good. I'll review more carefully, and
> hopefully apply it as time permits.
>
> Best regards,
>
> Thomas

Kind Regards,

Oleks

[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Thomas Petazzoni]

Hello,

On Thu, 13 Jul 2017 10:14:41 -0400, Oleksandr G Zhadan wrote:

> > Then could you please backport Linux kernel commit
> > 1e407ee3b21f981140491d5b8a36422979ca246f instead? It should fix this
> > build issue.  
> 
> Yes, I can, but we are using public kernel.

Well, the kernel is downloaded from:

+BR2_LINUX_KERNEL_CUSTOM_TARBALL_LOCATION="https://www.arcturusnetworks.com/OSS/kernel-ucls1012a-170706-13.tar.gz"

and you work for Arcturus Networks, so I guess you could adjust your
tarball to include the relevant patch, no?

> Should I add it as an extra patches to my 
> board/arcturus/ppc-ucp1020/patches/linux/ directory ?

If changing the tarball is not possible, then adding the patch in
Buildroot is indeed the other option.

Thanks!

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

[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Oleksandr G Zhadan]

Hi Thomas,

Looks like we mixed two architectures:

We have arcturus_ucp1020_defconfig (board/arcturus/ppc-ucp1020) already 
in buildroot tree for a while and we been using public kernel with our 
patches.
Recently I tested the build of it and found that with latest GCC-6 we 
have an issue to build image for this existing board (uCP1020), but it 
was building OK when the latest was GCC-4.9.  I agree with you that it 
is better to port GCC-6 issue fix and use GCC-6, I already tested it 
here and will added that linux-3.18 patch  into my board patch directory.
The updated patch for uCP1020 I will send soon.

The arcturus_ucls1012a_defconfig (board/arcturus/aarch64-ucls1012a) is 
another board support package we are trying to add to buildroot now.
And in patches for this board we are using 
"BR2_LINUX_KERNEL_CUSTOM_TARBALL_LOCATION="https://www.arcturusnetworks.com/OSS/kernel-ucls1012a-170706-13.tar.gz".
The v4 of the patch for uCLS1012a is waiting for your approval.

Thank you,

Oleks

On 07/13/2017 11:33 AM, Thomas Petazzoni wrote:
> Hello,
>
> On Thu, 13 Jul 2017 10:14:41 -0400, Oleksandr G Zhadan wrote:
>
>>> Then could you please backport Linux kernel commit
>>> 1e407ee3b21f981140491d5b8a36422979ca246f instead? It should fix this
>>> build issue.
>> Yes, I can, but we are using public kernel.
> Well, the kernel is downloaded from:
>
> +BR2_LINUX_KERNEL_CUSTOM_TARBALL_LOCATION="https://www.arcturusnetworks.com/OSS/kernel-ucls1012a-170706-13.tar.gz"
>
> and you work for Arcturus Networks, so I guess you could adjust your
> tarball to include the relevant patch, no?
>
>> Should I add it as an extra patches to my
>> board/arcturus/ppc-ucp1020/patches/linux/ directory ?
> If changing the tarball is not possible, then adding the patch in
> Buildroot is indeed the other option.
>
> Thanks!
>
> Thomas

[Buildroot] [PATCH 1/1] configs/arcturus_ucp1020_defconfig

[Thomas Petazzoni]

Hello,

On Thu, 13 Jul 2017 12:20:54 -0400, Oleksandr G Zhadan wrote:

> Looks like we mixed two architectures:
> 
> We have arcturus_ucp1020_defconfig (board/arcturus/ppc-ucp1020) already 
> in buildroot tree for a while and we been using public kernel with our 
> patches.
> Recently I tested the build of it and found that with latest GCC-6 we 
> have an issue to build image for this existing board (uCP1020), but it 
> was building OK when the latest was GCC-4.9.  I agree with you that it 
> is better to port GCC-6 issue fix and use GCC-6, I already tested it 
> here and will added that linux-3.18 patch  into my board patch directory.
> The updated patch for uCP1020 I will send soon.
> 
> The arcturus_ucls1012a_defconfig (board/arcturus/aarch64-ucls1012a) is 
> another board support package we are trying to add to buildroot now.
> And in patches for this board we are using 
> "BR2_LINUX_KERNEL_CUSTOM_TARBALL_LOCATION="https://www.arcturusnetworks.com/OSS/kernel-ucls1012a-170706-13.tar.gz".
> The v4 of the patch for uCLS1012a is waiting for your approval.

Indeed, I got confused by the two defconfigs. Thanks for clarifying the
situation!

Best regards,

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