Skip to contentSkip to conversation

sane-toolchain.inc: Make gcc 4.5 & eglibc 2.12 as default sane versions

PATCH In review Khem Raj 4 participants 8 messages Sep 25, 2010 – Sep 28, 2010

lore
Share & tools
lore

Khem Raj

author
#0 lore raw

PATCH 1/1 sane-toolchain.inc: Make gcc 4.5 & eglibc 2.12 as default sane versions

GCC 4.5 has been tested quite a bit and so is eglibc 2.12 in SHR and developers.

---
 conf/distro/include/sane-toolchain.inc |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)
conf/distro/include/sane-toolchain.inc +3 -3
index de73922..6d460a2 100644
@@ -5,14 +5,14 @@
 #############################################################################
 
 # Base line versions, good for most things
-PREFERRED_GCC_VERSION                ?= "4.4.4"
+PREFERRED_GCC_VERSION                ?= "4.5"
 PREFERRED_BINUTILS                   ?= "2.20.1"
 
 # Prefer glibc 2.6 and uclibc 0.9.30, these have had the most testing.
 PREFERRED_VERSION_glibc             ?= "2.10.1"
 PREFERRED_VERSION_glibc-initial     ?= "2.10.1"
-PREFERRED_VERSION_eglibc             ?= "2.11"
-PREFERRED_VERSION_eglibc-initial     ?= "2.11"
+PREFERRED_VERSION_eglibc             ?= "2.12"
+PREFERRED_VERSION_eglibc-initial     ?= "2.12"
 PREFERRED_UCLIBC_VERSION            ?= "0.9.31"
 
 # Some systems need a special gcc version
Signature · 1 line
1.7.0.4
#3 lore raw
Quoted 8 lines · Enrico Scholz
Quoted 1 line · Khem Raj

GCC 4.5 has been tested quite a bit

Probably not on ARMv5 ;) There 5/3 == 3 holds and ld.so crashes due to related errors.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45817

That bug seems to be specific to --enable-target-optspace, which (afaik) isn't selected for armv5 by the current recipes. Are you actually seeing that bug on a toolchain built with the settings in OE?

p.

#4 lore raw
Quoted 5 lines · Phil Blundell
Quoted 1 line

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45817

That bug seems to be specific to --enable-target-optspace, which (afaik) isn't selected for armv5 by the current recipes. Are you actually seeing that bug on a toolchain built with the settings in OE?

I have local customizations but I can not see in

$ git grep -i optspace

on vanilla OE where optspace is disabled for armv5.

Enrico

#5 lore raw
Quoted 8 lines · Enrico Scholz
Quoted 5 lines · Phil Blundell
Quoted 1 line

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45817

That bug seems to be specific to --enable-target-optspace, which (afaik) isn't selected for armv5 by the current recipes. Are you actually seeing that bug on a toolchain built with the settings in OE?

Quoted 5 lines

I have local customizations but I can not see in

$ git grep -i optspace

on vanilla OE where optspace is disabled for armv5.

Oh yes, right, it does seem to be enabled globally in gcc-configure-common. I had somehow read that conditional backwards when I looked at it before.

If that's the case then yes, obviously this needs fixing before gcc 4.5 is made the default. The slightly odd thing is that the corresponding conditional in lib1funcs.asm seems to have been there basically forever (since r77171) and it isn't obvious to me why it has suddenly broken now.

p.

author
#6 lore raw
In reply to Enrico Scholz
Quoted 8 lines · Enrico Scholz
Quoted 1 line · Khem Raj

GCC 4.5 has been tested quite a bit

Probably not on ARMv5 ;) There 5/3 == 3 holds and ld.so crashes due to related errors.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45817

yes there are few other bugs open against gcc 4.5 Interestingly I did not hit this bug while building images for qemuarm which is armv5. Is it common enough occurring problem that we should not use gcc 4.5 ?

Quoted 9 lines

Enrico

_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel