Bug 147474 - Mesa DRI drivers disable exec-shield
Mesa DRI drivers disable exec-shield
Status: CLOSED DUPLICATE of bug 147475
Product: Fedora
Classification: Fedora
Component: xorg-x11 (Show other bugs)
3
All Linux
medium Severity medium
: ---
: ---
Assigned To: X/OpenGL Maintenance List
David Lawrence
:
Depends On:
Blocks:
  Show dependency treegraph
 
Reported: 2005-02-08 07:06 EST by Mike A. Harris
Modified: 2007-11-30 17:10 EST (History)
0 users

See Also:
Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Environment:
Last Closed: 2006-02-21 14:08:09 EST
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---


Attachments (Terms of Use)

  None (edit)
Description Mike A. Harris 2005-02-08 07:06:31 EST
In XFree86 4.3.0, we had to patch the Mesa DRI drivers to call
mprotect and mark memory allocated for runtime code generation
with PROT_EXEC.  John Dennis created our patch that implemented
this, and I believe it was submitted upstream, however a recent
report by Arjan shows that Mesa still disables exec-shield.

xorg-x11-6.8.1-23.EL.i386     
./usr/X11R6/lib/modules/dri/gamma_dri.so: stack=RWE
xorg-x11-6.8.1-23.EL.i386     
./usr/X11R6/lib/modules/dri/i810_dri.so: stack=RWE
xorg-x11-6.8.1-23.EL.i386     
./usr/X11R6/lib/modules/dri/i915_dri.so: stack=RWE
xorg-x11-6.8.1-23.EL.i386      ./usr/X11R6/lib/modules/dri/mga_dri.so:
 stack=RWE
xorg-x11-6.8.1-23.EL.i386     
./usr/X11R6/lib/modules/dri/r128_dri.so: stack=RWE
xorg-x11-6.8.1-23.EL.i386     
./usr/X11R6/lib/modules/dri/r200_dri.so: stack=RWE
xorg-x11-6.8.1-23.EL.i386     
./usr/X11R6/lib/modules/dri/radeon_dri.so: stack=RWE
xorg-x11-6.8.1-23.EL.i386     
./usr/X11R6/lib/modules/dri/tdfx_dri.so: stack=RWE
xorg-x11-libs-6.8.1-23.EL.i386 ./usr/X11R6/lib/libOSMesa.so.4.0:     
  stack=RWE


It appears the patch never got accepted and integrated into Mesa CVS.
The old patch is still present in our current xorg rpm disabled.

Here are my own thoughts on the proper approach for someone to
tackle fixing this:

1) Open a discussion on the topic on Mesa-dev mailing list to find
   out if current Mesa CVS head or any existing stable Mesa releases
   have already fixed this (or look at the CVS code directly).

2) Assuming it is still not fixed in Mesa CVS, propose a solution
   on mesa-dev and discuss it until Brian et al. agree the solution
   will be acceptable into the Mesa CVS tree.

3) If someone goes ahead and does the work, great.  If not, then we
   should implement the solution discussed and agreed upon on
   mesa-dev, and submit it to Mesa.

4) Once the code is checked into Mesa CVS (and no sooner IMHO), we
   should then backport it to the Mesa in Xorg CVS head and get it
   checked into Xorg CVS, and into Mesa CVS stable branch of same
   Mesa release.

5) Once it is in Xorg CVS head, we should backport it to 6.8.x,
   and get it checked into both Mesa and Xorg CVS stable branches
   assuming it's accepted.


The final step is applying the patches to our rpms, which I would
place at step 6, to ensure that this is fully handled upstream
before we ship patched rpms again, so that we know we wont have
to port it again and again and again forever.
Comment 1 Mike A. Harris 2005-02-08 07:27:14 EST

*** This bug has been marked as a duplicate of 147475 ***
Comment 2 Red Hat Bugzilla 2006-02-21 14:08:09 EST
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.

Note You need to log in before you can comment on or make changes to this bug.