Filed from caserun (https://tcms.engineering.redhat.com/run/294277/#caserun_13783579)

Version-Release number of selected component (if applicable):
RHEL-7.3-20160914.1
xorg-x11-server-Xorg-1.17.2-22.el7.x86_64
xorg-x11-drv-ati-7.6.1-3.20160215gitd41fccc.el7.x86_64
kernel-3.10.0-506.el7.x86_64
linux-firmware-20160830-49.git7534e19.el7.noarch
package mesa is not installed
mesa-dri-drivers-11.2.2-2.20160614.el7.x86_64
package xorg-x11-glamor is not installed


Steps to Reproduce: 
 1. Do a fresh boot up
 2. Install ResultsDB client:
    # wget -O /etc/yum.repos.d/resultsdb-client.repo http://
    download.eng.brq.redhat.com/scratch/mdomonko/resultsdb/
    resultsdb-client.repo
    # yum install resultsdb-wrappers


 1. Run as a regular user:
    $ piglit-run_and_report



Actual results: 
ABRT says "glslparsetest killed by SIGSEGV" (piglit takes a while but does complete)


Expected results:
After the test completes, the script should ask you a couple of questions and
then report your results to ResultsDB for analysis and archiving.

Note: If the script terminates without asking any questions or Xorg crashes/
freezes during the test, this test case FAILED.

When answering the questions, please follow these hints:

  * "Do you want to have a script set prefill values for identifier, and TCMS
    run ID? [Y/n]"
    Answer "n"
  * Don't change the prefilled metric
  * Identifier is just a name you want to give to this report, similar to the
    Test Run Summary in TCMS (e.g. "Intel on ThinkPad T520")
  * We haven't decided yet on what determines the overall PASS/FAIL status so
    set this field to PASS for now :)
  * "Do you want to skip trying to prefill login and just enter it manually? [y
    /N]"
    Answer "y" and then enter your Kerberos login

After the script finishes, the result of the analysis is printed to the
terminal. If the outcome is

  * GOOD or WARN, set this test case to PASSED and you're done
  * REGRESS, set this test case to FAILED and file a bug report against the
    mesa component (include the provided Diff URL, if any)

Note from David Airlie: Its unlikely we'll ever see a perfect piglit run on any
GL hardware, its probably more of a regression suite, so if a test passes in
RHEL6.1 it should continue to pass in RHEL6.2