GCC4 and volatile effects on xorg-x11

[Lonni J Friedman]

On 7/17/05, Jim Cornette <jim-cornette at insight.rr.com> wrote:
> Hi,
> I am new to the list and could not find discussions from the last two
> months on GCC4 optimizations which cause misbehaving xorg-x11 servers on
> various cards. The problems seem to effect Intel, MGA, cyberblade and
> varous other drivers. Replacing the libvgahw.a file with a version
> compiled for a previous distribution of Fedora Core 3 will get X up and
> running again. Recent tests show that using lesser optimization will
> allow X to work on a larger amount of hardware, but not all hardware is
> made to work again with gcc4 and reduced optimization.
> 
> A thread on the gcc list that refers to this problem is listed below.
> 
> http://gcc.gnu.org/ml/gcc/2005-07/msg00699.html
> 
> I use the compiled binaries from distributions mostly. I do not know a
> lot about coding. I think people are working on resolving the changes
> which caused the many failures with X on certain hardware.
> 
> Ny troubles are with an Intel 815 displaying a blue border with wrapped
> text when changing to a virtual terminal. On an Intel 865G, the
> terminals are blank. Changing the libvgahw.a fixes both issues. Using
> lessor compilation errors fixes the Intel 815 but not the Intel 865G.
> 
> X seems to work great in the GUI. The problem is with the virtual terminals.
> 

FWIW, I'm also seeing identical behavior here with those two intel
chipsets.  These two bugs seem to be relevant:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=161242
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=160420


-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
L. Friedman                                    netllama at gmail.com
LlamaLand                       http://netllama.linux-sxs.org