Patchwork tcmode-default: disable ARMv7 Optimization for qt4-x11-free

login
register
mail settings
Submitter Saul Wold
Date June 1, 2011, 5:26 a.m.
Message ID <4DE5CDA3.1020000@linux.intel.com>
Download mbox | patch
Permalink /patch/5135/
State New, archived
Headers show

Comments

Saul Wold - June 1, 2011, 5:26 a.m.
[YOCTO #1030]

This diables the ARMv7 optimizations that cause the compiler to fail 
with the following error

| ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h: In function 
'void qt_removeObject(QObject*)':
| ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h:361:35: 
error: output number 1 not directly addressable

Signed-off-by: Saul Wold <sgw@linux.intel.com>
---
  meta/conf/distro/include/tcmode-default.inc |    1 +
  1 files changed, 1 insertions(+), 0 deletions(-)


  PREFERRED_VERSION_gcc ?= "${GCCVERSION}"
  PREFERRED_VERSION_gcc-cross ?= "${GCCVERSION}"
Khem Raj - June 1, 2011, 5:57 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 05/31/2011 10:26 PM, Saul Wold wrote:
> [YOCTO #1030]
> 
> This diables the ARMv7 optimizations that cause the compiler to fail
> with the following error
> 
> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h: In function
> 'void qt_removeObject(QObject*)':
> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h:361:35:
> error: output number 1 not directly addressable

this is a coding error you can fix the source code I guess instead of
pessimising the whole package

> 
> Signed-off-by: Saul Wold <sgw@linux.intel.com>
> ---
>  meta/conf/distro/include/tcmode-default.inc |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
> 
> diff --git a/meta/conf/distro/include/tcmode-default.inc
> b/meta/conf/distro/include/tcmode-default.inc
> index 6eb34ef..6393ffd 100644
> --- a/meta/conf/distro/include/tcmode-default.inc
> +++ b/meta/conf/distro/include/tcmode-default.inc
> @@ -29,6 +29,7 @@ PREFERRED_VERSION_u-boot-mkimage-native_powerpc ?=
> "2009.08"
>  # Temporary workaround for gcc 4.6.0 ICE with beagleboard
>  # gcc bug: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47719
>  TARGET_CC_ARCH_arm_pn-mesa-xlib :=
> "${@'${TARGET_CC_ARCH}'.replace('armv7-a','armv5')}"
> +TARGET_CC_ARCH_arm_pn-qt4-x11-free :=
> "${@'${TARGET_CC_ARCH}'.replace('armv7-a','armv5')}"
> 
>  PREFERRED_VERSION_gcc ?= "${GCCVERSION}"
>  PREFERRED_VERSION_gcc-cross ?= "${GCCVERSION}"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk3l1KwACgkQuwUzVZGdMxTBZACfTxYiNgPVbFuBDVzzLy8p2Vtr
NuQAn3MQg54GfX/s0+lyb7qXMn8BAcay
=gOaO
-----END PGP SIGNATURE-----
Martin Jansa - June 1, 2011, 6:10 a.m.
On Tue, May 31, 2011 at 10:57:00PM -0700, Khem Raj wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 05/31/2011 10:26 PM, Saul Wold wrote:
> > [YOCTO #1030]
> > 
> > This diables the ARMv7 optimizations that cause the compiler to fail
> > with the following error
> > 
> > | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h: In function
> > 'void qt_removeObject(QObject*)':
> > | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h:361:35:
> > error: output number 1 not directly addressable
> 
> this is a coding error you can fix the source code I guess instead of
> pessimising the whole package

FWIW: I haven't seen it in my armv7-a builds (nokia900), but I'm using
gcc-4.6 from meta-oe.

Regards,
Khem Raj - June 1, 2011, 6:14 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 05/31/2011 11:10 PM, Martin Jansa wrote:
> On Tue, May 31, 2011 at 10:57:00PM -0700, Khem Raj wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 05/31/2011 10:26 PM, Saul Wold wrote:
>>> [YOCTO #1030]
>>>
>>> This diables the ARMv7 optimizations that cause the compiler to fail
>>> with the following error
>>>
>>> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h: In function
>>> 'void qt_removeObject(QObject*)':
>>> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h:361:35:
>>> error: output number 1 not directly addressable
>>
>> this is a coding error you can fix the source code I guess instead of
>> pessimising the whole package
> 
> FWIW: I haven't seen it in my armv7-a builds (nokia900), but I'm using
> gcc-4.6 from meta-oe.
> 
> Regards,
> 

this may give some hints on how to fix it in qt-x11
https://bugs.launchpad.net/ubuntu/+source/qt4-x11/+bug/673085
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk3l2NQACgkQuwUzVZGdMxQ27gCeKEorKdMnEwd3rRqLAqQ6LX+s
90kAn3yx+/omTFUbM214frXzdDNQKEFR
=dO6F
-----END PGP SIGNATURE-----
Khem Raj - June 1, 2011, 6:21 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 05/31/2011 11:10 PM, Martin Jansa wrote:
> On Tue, May 31, 2011 at 10:57:00PM -0700, Khem Raj wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 05/31/2011 10:26 PM, Saul Wold wrote:
>>> [YOCTO #1030]
>>>
>>> This diables the ARMv7 optimizations that cause the compiler to fail
>>> with the following error
>>>
>>> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h: In function
>>> 'void qt_removeObject(QObject*)':
>>> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h:361:35:
>>> error: output number 1 not directly addressable
>>
>> this is a coding error you can fix the source code I guess instead of
>> pessimising the whole package
> 
> FWIW: I haven't seen it in my armv7-a builds (nokia900), but I'm using
> gcc-4.6 from meta-oe.
> 

this is because gcc 4.6 in meta-oe has fix for it its this patch here
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01477.html

a patch developed by linaro folks but now is committed to FSF as well.
this patch has not been backported to FSF 4.6 branch

> Regards,
> 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk3l2oUACgkQuwUzVZGdMxQa1ACfaHP5ur/Dlh78nAPI9ViqPaK1
G60An3/34c1z9ZoM3y8o62fvke32xI9J
=hDdS
-----END PGP SIGNATURE-----
Phil Blundell - June 1, 2011, 10:46 a.m.
On Tue, 2011-05-31 at 23:21 -0700, Khem Raj wrote:
> On 05/31/2011 11:10 PM, Martin Jansa wrote:
> > On Tue, May 31, 2011 at 10:57:00PM -0700, Khem Raj wrote:
> >> On 05/31/2011 10:26 PM, Saul Wold wrote:
> >>> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h:361:35:
> >>> error: output number 1 not directly addressable
> >>
> >> this is a coding error you can fix the source code I guess instead of
> >> pessimising the whole package
> > 
> > FWIW: I haven't seen it in my armv7-a builds (nokia900), but I'm using
> > gcc-4.6 from meta-oe.
>
> this is because gcc 4.6 in meta-oe has fix for it its this patch here
> http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01477.html

I'm a bit confused by this.  In the first quote above you said that this
was a coding error and ought to be fixed in qt, but now you seem to be
saying that it's a compiler bug and should be fixed in gcc.  If it's the
latter, can we get that patch applied to the oe-core version of gcc?

Either way, I agree, working around this by just frobbing the
architecture options seems like a bad idea.

p.
Saul Wold - June 1, 2011, 5:50 p.m.
On 06/01/2011 03:46 AM, Phil Blundell wrote:
> On Tue, 2011-05-31 at 23:21 -0700, Khem Raj wrote:
>> On 05/31/2011 11:10 PM, Martin Jansa wrote:
>>> On Tue, May 31, 2011 at 10:57:00PM -0700, Khem Raj wrote:
>>>> On 05/31/2011 10:26 PM, Saul Wold wrote:
>>>>> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h:361:35:
>>>>> error: output number 1 not directly addressable
>>>>
>>>> this is a coding error you can fix the source code I guess instead of
>>>> pessimising the whole package
>>>
>>> FWIW: I haven't seen it in my armv7-a builds (nokia900), but I'm using
>>> gcc-4.6 from meta-oe.
>>
>> this is because gcc 4.6 in meta-oe has fix for it its this patch here
>> http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01477.html
>
> I'm a bit confused by this.  In the first quote above you said that this
> was a coding error and ought to be fixed in qt, but now you seem to be
> saying that it's a compiler bug and should be fixed in gcc.  If it's the
> latter, can we get that patch applied to the oe-core version of gcc?
>
> Either way, I agree, working around this by just frobbing the
> architecture options seems like a bad idea.
>
> p.
>
Khem,

Is there any chance that you would be willing to throw me a patch that 
address this issue based on the gcc patch?  If not. let me know and I 
will try to muddle my way through it ;-)!

Sau!

>
>
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
>
Khem Raj - June 1, 2011, 6:31 p.m.
On Wed, Jun 1, 2011 at 10:50 AM, Saul Wold <sgw@linux.intel.com> wrote:
> On 06/01/2011 03:46 AM, Phil Blundell wrote:
>>
>> On Tue, 2011-05-31 at 23:21 -0700, Khem Raj wrote:
>>>
>>> On 05/31/2011 11:10 PM, Martin Jansa wrote:
>>>>
>>>> On Tue, May 31, 2011 at 10:57:00PM -0700, Khem Raj wrote:
>>>>>
>>>>> On 05/31/2011 10:26 PM, Saul Wold wrote:
>>>>>>
>>>>>> | ../../include/QtCore/../../src/corelib/arch/qatomic_arm.h:361:35:
>>>>>> error: output number 1 not directly addressable
>>>>>
>>>>> this is a coding error you can fix the source code I guess instead of
>>>>> pessimising the whole package
>>>>
>>>> FWIW: I haven't seen it in my armv7-a builds (nokia900), but I'm using
>>>> gcc-4.6 from meta-oe.
>>>
>>> this is because gcc 4.6 in meta-oe has fix for it its this patch here
>>> http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01477.html
>>
>> I'm a bit confused by this.  In the first quote above you said that this
>> was a coding error and ought to be fixed in qt, but now you seem to be
>> saying that it's a compiler bug and should be fixed in gcc.  If it's the
>> latter, can we get that patch applied to the oe-core version of gcc?
>>
>> Either way, I agree, working around this by just frobbing the
>> architecture options seems like a bad idea.
>>
>> p.
>>
> Khem,
>
> Is there any chance that you would be willing to throw me a patch that
> address this issue based on the gcc patch?  If not. let me know and I will
> try to muddle my way through it ;-)!

I dont have the environment to reproduce the error. If you can try to apply gcc
patch I mentioned to gcc then it should help.

>
> Sau!
>
>>
>>
>> _______________________________________________
>> Openembedded-core mailing list
>> Openembedded-core@lists.openembedded.org
>> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
>>
>

Patch

diff --git a/meta/conf/distro/include/tcmode-default.inc 
b/meta/conf/distro/include/tcmode-default.inc
index 6eb34ef..6393ffd 100644
--- a/meta/conf/distro/include/tcmode-default.inc
+++ b/meta/conf/distro/include/tcmode-default.inc
@@ -29,6 +29,7 @@  PREFERRED_VERSION_u-boot-mkimage-native_powerpc ?= 
"2009.08"
  # Temporary workaround for gcc 4.6.0 ICE with beagleboard
  # gcc bug: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47719
  TARGET_CC_ARCH_arm_pn-mesa-xlib := 
"${@'${TARGET_CC_ARCH}'.replace('armv7-a','armv5')}"
+TARGET_CC_ARCH_arm_pn-qt4-x11-free := 
"${@'${TARGET_CC_ARCH}'.replace('armv7-a','armv5')}"