All of lore.kernel.org
 help / color / mirror / Atom feed
* glib-2.0: do_compile, angstrom: gcc-cross ver.
@ 2010-07-11 13:58 lukas gorris
  2010-07-11 15:55 ` Koen Kooi
  0 siblings, 1 reply; 3+ messages in thread
From: lukas gorris @ 2010-07-11 13:58 UTC (permalink / raw)
  To: openembedded-devel

Hello,

I have just ran a build in org.openembedded.dev commit
fd049a8b9904348456b26ffa7ff25d834c477c6b and noticed a bug.

I am using armv7a based htcleo machine configuration and
angstrom-2008.1 distro. because angstrom still selects obsolete
gcc-cross versions for armv7a (see
http://bugs.openembedded.org/show_bug.cgi?id=5435#c21 ) I bump it
manually like this

PREFERRED_VERSION_gcc-cross = "4.4.4"
PREFERRED_VERSION_gcc-cross-initial = "4.4.4"
PREFERRED_VERSION_gcc-cross-intermediate = "4.4.4"
PREFERRED_VERSION_binutils-cross = "2.20.1"

compile error from glib-2.0

| gconvert.c:55:2: error: #error GNU libiconv not in use but included
iconv.h is from libiconv

full log:

http://tinderbox.openembedded.net/public/logs/task/6483371.txt

Khem said it's because my host lacks libiconv. Unfortunately I can't
test it now. Is this related to following commit?

211fb05d665e95192cf4d4b81dfa89fde04359cf

Maybe it's good to add libiconv in bitbake host sanity checker or
require it in some other way like installing a libiconv-native ?

Can somebody look into updating gcc-cross in angstrom for armv7 ?
(Hope this the right list for angstrom topic)

Thanks

Lukas




^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: glib-2.0: do_compile, angstrom: gcc-cross ver.
  2010-07-11 13:58 glib-2.0: do_compile, angstrom: gcc-cross ver lukas gorris
@ 2010-07-11 15:55 ` Koen Kooi
  2010-07-11 22:11   ` Khem Raj
  0 siblings, 1 reply; 3+ messages in thread
From: Koen Kooi @ 2010-07-11 15:55 UTC (permalink / raw)
  To: openembedded-devel

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 11-07-10 15:58, lukas gorris wrote:
> Hello,
> 
> I have just ran a build in org.openembedded.dev commit
> fd049a8b9904348456b26ffa7ff25d834c477c6b and noticed a bug.
> 
> I am using armv7a based htcleo machine configuration and
> angstrom-2008.1 distro. because angstrom still selects obsolete
> gcc-cross versions for armv7a

Excuse me!?!?!? Gcc 4.3.x still generates the fastest and best code for
armv7a, gcc 4.4.x generates slow and buggy code and well, let's not talk
about gcc 4.5.

I'm using binaries built by this 'obsolete' toolchain daily on armv7a
hardware without problems, using neon, vfp, etc, so I'm wondering what
this is all about.

This all reeks of broken kernel syndrome, are you using linux.inc? If
not, please do so, it fixes up a whole lot of bugs.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFMOemGMkyGM64RGpERAtpVAKCNitZFtUlGpbLtF/glkqhvs5GZQQCgrLx7
drZkWVNgaleAovXa7efOsVc=
=yPyi
-----END PGP SIGNATURE-----




^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: glib-2.0: do_compile, angstrom: gcc-cross ver.
  2010-07-11 15:55 ` Koen Kooi
@ 2010-07-11 22:11   ` Khem Raj
  0 siblings, 0 replies; 3+ messages in thread
From: Khem Raj @ 2010-07-11 22:11 UTC (permalink / raw)
  To: openembedded-devel

On Sun, Jul 11, 2010 at 8:55 AM, Koen Kooi <k.kooi@student.utwente.nl> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 11-07-10 15:58, lukas gorris wrote:
>> Hello,
>>
>> I have just ran a build in org.openembedded.dev commit
>> fd049a8b9904348456b26ffa7ff25d834c477c6b and noticed a bug.
>>
>> I am using armv7a based htcleo machine configuration and
>> angstrom-2008.1 distro. because angstrom still selects obsolete
>> gcc-cross versions for armv7a
>
> Excuse me!?!?!? Gcc 4.3.x still generates the fastest and best code for
> armv7a,

agreed.

gcc 4.4.x generates slow and buggy code and well,

may be, I don't have any data to support this claim but if you do have
something it would be nice if you shared that here.

 let's not talk
> about gcc 4.5.

gcc 4.5 has many fixes including the direct TLS register access through CP15
on armv7+ which should be some improvement for TLS access. May be its
worth to start trying it on armv7a and weed out any issues early in
its adoption.
and may be start using thumb2 if not done thus far for armv7+

>
> I'm using binaries built by this 'obsolete' toolchain daily on armv7a
> hardware without problems, using neon, vfp, etc, so I'm wondering what
> this is all about.
>
> This all reeks of broken kernel syndrome, are you using linux.inc? If
> not, please do so, it fixes up a whole lot of bugs.


> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.5 (Darwin)
>
> iD8DBQFMOemGMkyGM64RGpERAtpVAKCNitZFtUlGpbLtF/glkqhvs5GZQQCgrLx7
> drZkWVNgaleAovXa7efOsVc=
> =yPyi
> -----END PGP SIGNATURE-----
>
>
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2010-07-11 22:16 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-07-11 13:58 glib-2.0: do_compile, angstrom: gcc-cross ver lukas gorris
2010-07-11 15:55 ` Koen Kooi
2010-07-11 22:11   ` Khem Raj

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.