Re: gcc 4.7 + mips64 architecture support

[Martin Jansa <martin.jansa@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 5534 bytes --]

On Sun, Apr 29, 2012 at 08:36:05AM -0700, Khem Raj wrote:
> On Sun, Apr 29, 2012 at 2:12 AM, Martin Jansa <martin.jansa@gmail.com> wrote:
> > On Wed, Apr 25, 2012 at 04:31:53PM -0700, Khem Raj wrote:
> >> Hello
> >>
> >> Now that release is almost out of door. I would like some eyes on the
> >> gcc-4.7 work
> >> and mips64 architecture support that I have done meanwhile
> >>
> >> I have made sure that gcc-4.6 still keeps working and I have patches
> >> for gcc-4.5 too
> >> for meta-oe too. I have repeatedly built core-image-sato and
> >> core-image-minimal for uclibc/eglibc
> >> but more coverage would be nicer.
> >>
> >> The patches are posted here
> >>
> >> http://git.openembedded.org/openembedded-core-contrib/log/?h=kraj/misc
> >>
> >> I would appreciate any feedback or testing on this
> >
> > I got some errors while building gcc* recipes (overwriting header
> > files in work-shared or something like that) but that's not new, gcc-4.6
> > was failing the same - restarting build few times and it will succeed
> > once.
> 
> hmm more details on this one would be better. May be its a bug in gcc
> build system

I'll include log of our discussion on #yocto from yesterday:

20:56 < JaMa> khem: those gcc build issues are always about different constant or invalid file (expected } or something) so as I said before I believe it's some sort of race condition with parallel build in gcc or with parallel build of all those *gcc* recipes, but hard to reproduce because with second run it's usually gone
20:57 < khem> hmmm
20:57 < khem> JaMa: so the build fails saying what ?
20:57 < khem> first time
20:58 < JaMa> thanks to RP's fix I can at least rebuilt just the one which was failing instead of all *gcc* again
20:58 < JaMa> undeclared constant or expected } or something, I think each time the error is a bit different
20:58 < khem> JaMa: I have changed gcc-4.7 to avoid some overlappings which should help too
20:59 < khem> oh I see
20:59 < khem> that means a .h file is not generated yet and is being referenced or included
20:59 < khem> and you said its looking in srcdir ?
21:00 < JaMa> last time we talked about it, I've blamed do_headerfix because it was running in parallel with do_compile in some other *gcc* recipe
...
21:32 < JaMa> khem: I've found couple of old build logs with gcc failing http://build.shr-project.org/tests/jama/gcc-upgrade-issue/

Cheers,

> 
> >
> > My minimal shr-lite-image compiles now fine and works on om-gta02
> > (armv4t) sofar without segfaults.. so looks good.
> 
> nice
> 
> >
> > For normal image and feeds I've few more recipes which are failing with gcc-4.7:
> >
> > meta-openembedded/meta-efl/recipes-efl/webkit/webkit-efl_svn.bb
> >  | Unknown parameter a interfaceName for tags/attrs
> >  | make[2]: *** [DerivedSources/WebCore/HTMLNames.cpp] Error 255
> >  http://norman-schleicher.de/jenkins/job/shr-core-branches-om-gta02-shr-image/ws/shr-core/tmp-eglibc/work/armv4t-oe-linux-gnueabi/webkit-efl-1.7.3+svnr105069-r1/
> >  I'm upgrading this one instead of fixing it, because we had some
> >  issues in runtime with current version, but new version needs newer
> >  libsoup-2.4->glib-2.0->gobject-introspection->etc and fails with
> >  different compilation issues.. so maybe if you know how to easily fix
> >  it in current, we can, otherwise I'll continue to work on upgrade
> 
> I can surely provide you a fix once I look at sources. But seems  you
> have more than one problems that may not be fixed. So upgrading has a
> slight chance of fixing both of your issues and introducing many more
> :) but lets try upgrade route and if then
> we need 4.7 patch ping me.
> 
> >
> >  BTW: similar fix as you did for webkit-gtk is in upstream r109509
> >
> 
> Nice as long as the things are fixed I am happy.
> 
> 
> > meta-oe/recipes-extended/mml-widget/gtkmathview_0.8.0.bb
> >  ../../../src/frontend/common/TemplateBuilder.hh:54:42: error: redeclaration of 'SmartPtr<typename ElementBuilder::type> elem'
> >  ../../../src/frontend/common/TemplateBuilder.hh:50:49: error: 'SmartPtr<typename ElementBuilder::type> elem' previously declared here
> >  http://norman-schleicher.de/jenkins/job/shr-core-branches-om-gta02-shr-image/ws/shr-core/tmp-eglibc/work/armv4t-oe-linux-gnueabi/gtkmathview-0.8.0-r1/
> >
> > meta-smartphone/meta-shr/recipes-shr/3rdparty/numptyphysics_svn.bb
> >  fixed in shr branch
> 
> good
> 
> 
> >
> > meta-mozilla/recipes-mozilla/firefox/firefox_10.0.2.bb
> >  | /var/lib/jenkins/jobs/shr-core-branches/workspace/shr-core/tmp-eglibc/work/armv7a-vfp-neon-oe-linux-gnueabi/firefox-10.0.2-r1/mozilla-release/ipc/chromium/src/base/file_util.cc:228:35: error: 'ftruncate' was not declared in this scope
> >  probably easy to fix, adding Otavio to CC (maybe he has upgrade or fix
> >  ready)
> >
> 
> just #include <unistd.h> in ipc/chromium/src/base/file_util.cc should fix it
> 
> > Cheers,
> >
> > --
> > Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com
> >
> > _______________________________________________
> > Openembedded-core mailing list
> > Openembedded-core@lists.openembedded.org
> > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
> >
> 
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 205 bytes --]