On Wed, Apr 04, 2012 at 11:11:08PM +0200, Martin Jansa wrote:
> On Wed, Mar 28, 2012 at 04:33:47PM +0200, Martin Jansa wrote:
> > On Wed, Mar 28, 2012 at 07:15:32AM -0700, Khem Raj wrote:
> > > On Fri, Mar 23, 2012 at 7:26 PM, Martin Jansa <martin.jansa@gmail.com> wrote:
> > > > On Fri, Mar 23, 2012 at 02:17:02PM +0100, Martin Jansa wrote:
> > > >> On Fri, Mar 23, 2012 at 12:05:25PM +0000, Richard Purdie wrote:
> > > >> > On Thu, 2012-03-22 at 19:56 -0700, Khem Raj wrote:
> > > >> > > Get new patches and remove the one that got merged upstream
> > > >> > >
> > > >> > > Signed-off-by: Khem Raj <raj.khem@gmail.com>
> > > >> > > ---
> > > >> > >  .../eglibc/eglibc-2.15/GLRO_dl_debug_mask.patch    |  108 ++++++++++++++++++++
> > > >> > >  .../eglibc-2.15/armv4-eabi-compile-fix.patch       |   25 -----
> > > >> > >  .../eglibc/eglibc-2.15/initgroups_keys.patch       |   20 ++++
> > > >> > >  meta/recipes-core/eglibc/eglibc_2.15.bb            |    5 +-
> > > >> > >  4 files changed, 131 insertions(+), 27 deletions(-)
> > > >> > >  create mode 100644 meta/recipes-core/eglibc/eglibc-2.15/GLRO_dl_debug_mask.patch
> > > >> > >  delete mode 100644 meta/recipes-core/eglibc/eglibc-2.15/armv4-eabi-compile-fix.patch
> > > >> > >  create mode 100644 meta/recipes-core/eglibc/eglibc-2.15/initgroups_keys.patch
> > > >> >
> > > >> > Since 2.15 isn't the default I'm tempted to merge this despite the point
> > > >> > we're at in the release cycle. I'd like to give anyone using this the
> > > >> > opportunity to comment first though.
> > > >>
> > > >> I'm testing this on all archs I'm using (armv4t, armv5te, armv7a,
> > > >> x86-64) and will report tomorrow.
> > > >
> > > > Doesn't look related to this change, but I got interesing error on other
> > > > buildhost (on mine everything seems fine sofar including tests on
> > > > target), I don't have log from armv4t build on mine, because it's on tmpfs
> > > > and there was unexpected reboot when I wasn't home :/.
> > > 
> > > Hi Martin
> > > 
> > > Does it look good for your testing ?
> > 
> > Seems fine in runtime any hint about that build issue?
> > 
> > Today I got another one :/
> > 
> > "build-x86_64-oe-linux/shlib.lds:127: syntax error
> > while building eglibc
> > 
> > I had reports about this before, but wasn't able to reproduce it on my
> > machine, today I got it for first time (maybe because of building with
> > -j8 instead of -j4), so I've created simple testcase to compare interim 
> > outputs from shlib.lds creation:
> > http://build.shr-project.org/tests/jama/shlib-issue/
> 
> To follow this monolog, the problem is in file Makerules itself, gcc
> output is the same as well as output from sed, but Makerules file is
> partially rewritten during build and if shlib.lds gets generated in
> wrong time then it results in this
> 
> touch /OE/shr-core/tmp-eglibc/work/armv4t-oe-linux-gnueabi/eglibc/eglibc-2.15-r3+svnr17386/eglibc-2_15/libc/Makerules
> 
> after eglibc build fail is enough to recreate correct shlibs.lds

I was running core-image-minimal builds on different block devices to
provide some build time data for 
[yocto] Build time data
thread and it seems that this bug is more likely to happen on fast host
with slow IO

I've got it 4 times today (from 4 tries) when running with tmpdir on RAID5,
and only once (from 3 tries) when running with with tmpdir on tmpfs.

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