On Fri, Jun 20, 2008 at 8:09 AM, Dan McGee wrote:

This is going to get us when we rebuild a kernel with gcc 4.3.1, so we should probably be aware:

http://kerneltrap.org/Linux/Compiler_Oops https://bugzilla.redhat.com/show_bug.cgi?id=451068

Looks like 2.6.25.6 is built with gcc 4.3.0 so we are safe now. Should we patch gcc with the upstream fix?

I'd say so. The kerneltrap article points fingers at GCC, and indicates it's only an i686 issue. I'm for patching gcc to fix this before we start recompiling things.

Am Sat, 21 Jun 2008 19:03:38 +0200 schrieb Andreas Radke :

Am Fri, 20 Jun 2008 08:09:35 -0500 schrieb "Dan McGee" :

...

This is going to get us when we rebuild a kernel with gcc 4.3.1, so we should probably be aware:

http://kerneltrap.org/Linux/Compiler_Oops https://bugzilla.redhat.com/show_bug.cgi?id=451068

Looks like 2.6.25.6 is built with gcc 4.3.0 so we are safe now. Should we patch gcc with the upstream fix?

-Dan

I'll have a look at it at the beginning of next week. Patching gcc or or changing to their latest stable tree snapshot should also be an option.

-Andy

gcc-{,libs-}4.3.1-2 based on the latest snapshot has hit testing and with all the latest upstream commits included it _should_ fix the regression. so kernel packagers give it a try before we start gcc signoff. -Andy

On Mon, Jul 7, 2008 at 1:07 PM, Andreas Radke wrote:

Am Sun, 29 Jun 2008 21:40:37 +0200 schrieb Andreas Radke :

...

gcc-{,libs-}4.3.1-2 based on the latest snapshot has hit testing and with all the latest upstream commits included it _should_ fix the regression.

so kernel packagers give it a try before we start gcc signoff.

-Andy

Lastest kernel has been built with it and has been moved to core already without user complains. Also Jan and me were using it for db/heimdal rebuilds. He gave me his signoff via IM.

Going to move gcc now.

Yay! Thanks a lot guys, you guys do a great job with the toolchain stuff. 8)