xorg-server 1.11.4 compile time error

[fnoord]

Updating xorg-server from 1.11.3 to 1.11.4 fails for me with some strange compiler-error stating that there is an out-of-bound array access in some part of the xorg server code.

Previous versions of xorg-server are still building without any problems; the active use flags are:
ipv6
nptl
udev
xorg

Build log contains a shitload of warnings (>13k lines) trailed by a compile time error:

In file included from /var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86Crtc.h:30:0,
from /var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:44:
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86Modes.h:102:1: warning: redundant redeclaration of ‘xf86ValidateModesFlags’
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86Modes.h:72:1: note: previous declaration of ‘xf86ValidateModesFlags’ was here
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:240:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:244:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:248:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:252:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:256:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:260:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:264:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:268:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:272:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:276:5: warning: initialization discards qualifiers from pointer target type
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c: In function ‘DDCModesFromCEAExtension’:
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:988:52: warning: declaration of ‘MonPtr’ shadows a global declaration
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/common/xf86str.h:229:12: warning: shadowed declaration is here
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c: In function ‘handle_detailed_modes’:
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:750:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:749:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:748:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:747:33: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:750:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:749:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:748:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:747:33: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:750:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:749:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:748:19: error: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:747:33: error: array subscript is above array bounds
make[4]: *** [xf86EdidModes.lo] Error 1
make[4]: Leaving directory `/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4_build/hw/xfree86/modes'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory `/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4_build/hw/xfree86'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4_build/hw/xfree86'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4_build/hw'
make: *** [all-recursive] Error 1

Emerge also states that:
* Running elibtoolize in: xorg-server-1.11.4/
* We've already been run in this tree; you should
* avoid this if possible (perhaps by filing a bug)

Are there any obvious reasons or known bugs with that, might that just be a configuration issue and/or should I file a bug about this?

[chithanh]

Do you have anything installed to /usr/local that could interfere?

[Khrisbie]

Same thing here.

I looked into the source code :

/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:750:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:749:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:748:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:747:33: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:750:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:749:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:748:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:747:33: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:750:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:749:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:748:19: erreur: array subscript is above array bounds
/var/tmp/portage/x11-base/xorg-server-1.11.4/work/xorg-server-1.11.4/hw/xfree86/modes/xf86EdidModes.c:747:33: erreur: array subscript is above array bounds

Looks like there is a patch for that (named "stay in array bounds in DDCModesFromEstIII()")...

http://patchwork.freedesktop.org/patch/7970/

[OPelerin]

I'm having the same pb too.

Is there a bug already open for that?
_________________
Olivier PELERIN

[fnoord]

@chithanh: I can not imagine anything in my /usr/local could interfere with xorg. Only things in my /usr/local are vim-supertab, crossdev for mingw and avr and truecrypt...

[mrpdaemon]

[Goeland86]

So I'm having this same issue, but I tried masking 11.4, then 11.3 when that one failed with the SAME problem, and now that I'm back down to 10.6 I *STILL* have the exact same problem.

I'm no expert, but I'd say this is not the xorg-server code itself, but a dependent lib that's fubarding the whole thing!

[OPelerin]

11-4-r1 is out and It went without any pb
_________________
Olivier PELERIN

[Zarhan]

This happens to me, even with the -r1. What's wrong? Haven't tried with -O2 yet, though.

[miket]

Something really screwy has gone on. I already have a working installation of xorg-server-1.11.2-r2 on a different machine, and it installed with no problem (months ago). I've been trying to get the same package to compile on a new machine, and no matter what I try, from using simpler USE flags and rebuilding system and world, the ebuild keeps failing. This is on a new install.

I decided to take it one step further. On this same new machine, I made a new-installation chroot, emerged gentoo-sources so that the xorg packages could find kernel sources, and tried one more time for xorg-server in this very vanilla environment. Still no dice.

I get the idea that something happened in a eclass or one of the packages on which xorg-server depends.

I guess I could try unkeywording ~1.11.4-r1 and seeing if that would work.

Ah, what a fun day this has been.

[gentoorockerfr]

problem here

[miket]

Well, I bit the bullet and followed the workaround in the bug report: compile with the -O2 switch, not the -O3 that I normally use. The bug reports that someone has a patch to make xorg-server build with the -O3 switch, but that's not in any of the versions I tried.

https://bugs.gentoo.org/show_bug.cgi?id=402285

If I were really ambitious, I would have made my own private ebuild for xorg-server 1.11.2-r2. As it was, I compiled that package--and that package only--with -O2. Heh. Give it a little oxygen. I did, and now I can breathe again.

[czernitko]

If anyone doesn't know about this portage feature, it might help someone.
If you are optimising with -O3 by default (in make.conf) and want to always optimise xorg-server with -O2, try this:

[miket]