[Buildroot] gcc 4.3.3 anyone?

[Buildroot] gcc 4.3.3 anyone?

[Sven Neumann]

Hi,

I've ran into what appears to be a bug in gcc 4.3.2 and reported it
here: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39076

Now I consider to try to update buildroot to use gcc 4.3.3. But as there
are a number of patches for each gcc version this could become quite
some work. That's why I'm asking if anyone has already done that and
wants to share his/her experience/patches?


Sven

[Buildroot] gcc 4.3.3 anyone?

[Peter Korsgaard]

>>>>> "Sven" == Sven Neumann <s.neumann@phase-zero.de> writes:

 Sven> Hi,
 Sven> I've ran into what appears to be a bug in gcc 4.3.2 and reported it
 Sven> here: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39076

 Sven> Now I consider to try to update buildroot to use gcc 4.3.3. But as there
 Sven> are a number of patches for each gcc version this could become quite
 Sven> some work. That's why I'm asking if anyone has already done that and
 Sven> wants to share his/her experience/patches?

I was going to bump the gcc version right after the release, but if
someone else want to do it, that's fine as well.

I wouldn't do it for the release though.

-- 
Bye, Peter Korsgaard

[Buildroot] gcc 4.3.3 anyone?

[Sven Neumann]

Hi,

On Mon, 2009-02-02 at 21:06 +0100, Peter Korsgaard wrote:

>  Sven> I've ran into what appears to be a bug in gcc 4.3.2 and reported it
>  Sven> here: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39076

The gcc people have fixed this problem in the gcc-4.3 branch. I verified
that the patch fixes the compiler segfault when compiling flac. Attached
is a patch for buildroot that incorporates this fix for gcc 4.3.2.

I also noticed that two of the patches in toolchain/gcc/4.3.2 don't end
in .patch and thus are not applied. Is this intentional?


Sven


-------------- next part --------------
A non-text attachment was scrubbed...
Name: gcc-bugfix.diff
Type: text/x-patch
Size: 908 bytes
Desc: not available
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20090205/bbc89c33/attachment.bin>

[Buildroot] gcc 4.3.3 anyone?

[Peter Korsgaard]

>>>>> "Sven" == Sven Neumann <s.neumann@phase-zero.de> writes:

 Sven> Hi,
 Sven> On Mon, 2009-02-02 at 21:06 +0100, Peter Korsgaard wrote:

 Sven> I've ran into what appears to be a bug in gcc 4.3.2 and reported it
 Sven> here: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39076

 Sven> The gcc people have fixed this problem in the gcc-4.3 branch. I verified
 Sven> that the patch fixes the compiler segfault when compiling flac. Attached
 Sven> is a patch for buildroot that incorporates this fix for gcc 4.3.2.

Thanks, will apply.

 Sven> I also noticed that two of the patches in toolchain/gcc/4.3.2 don't end
 Sven> in .patch and thus are not applied. Is this intentional?

No, not afaik. The 995-short-enums.diff comes from Bernhard and is a
fix for PR34205 which got tagged WONTFIX upstream
(http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34205), and afaik
shouldn't be needed now we don't use abi=iwmmxt anymore - Bernhard,
correct me if I'm wrong - Otherwise I'll remove it.

The 306-arm-funroll-loops.patch.arm is from Ulf (r24993), and not just
the name is wrong, it doesn't apply (wrong patch level). I've fixed
that up and am now doing a test build to check if it fixes the mplayer
ICE. If so, then I'll commit, otherwise I'll remove it.

Thanks for noticing!

-- 
Bye, Peter Korsgaard

[Buildroot] gcc 4.3.3 anyone?

[Peter Korsgaard]

>>>>> "Sven" == Sven Neumann <s.neumann@phase-zero.de> writes:

 Sven> Hi,
 Sven> On Mon, 2009-02-02 at 21:06 +0100, Peter Korsgaard wrote:

 Sven> I've ran into what appears to be a bug in gcc 4.3.2 and reported it
 Sven> here: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39076

 Sven> The gcc people have fixed this problem in the gcc-4.3 branch. I verified
 Sven> that the patch fixes the compiler segfault when compiling flac. Attached
 Sven> is a patch for buildroot that incorporates this fix for gcc 4.3.2.

 Sven> I also noticed that two of the patches in toolchain/gcc/4.3.2 don't end
 Sven> in .patch and thus are not applied. Is this intentional?

I checked, and your patch seems to be the same a 2nd hunk of Ulf's
306-arm-funroll-loops.patch, so it isn't needed anymore.

-- 
Bye, Peter Korsgaard