Set priority to high. It has an impact on prime category.

Also reproduced on KBL-U. 5 subcases are fail:
prime_busy@after-render
prime_busy@after-vebox
prime_busy@before-blt
prime_busy@before-default
prime_busy@before-render

./prime_busy --r after-render
IGT-Version: 1.15-g0404de4 (x86_64) (Linux: 4.7.0-rc5-patch-enable-guc-on-kbl+ x86_64)
(prime_busy:4240) CRITICAL: Test assertion failure function one, file prime_busy.c:136:
(prime_busy:4240) CRITICAL: Failed assertion: prime_busy(&pfd[SCRATCH], false)
Stack trace:
  #0 [__igt_fail_assert+0xf1]
  #1 [one+0x554]
  #2 [__real_main164+0x253]
  #3 [main+0x29]
  #4 [__libc_start_main+0xf0]
  #5 [_start+0x29]
  #6 [<unknown>+0x29]
Subtest after-render failed.
**** DEBUG ****
(prime_busy:4240) ioctl-wrappers-DEBUG: Test requirement passed: gem_has_ring(fd, ring)
(prime_busy:4240) DEBUG: Test requirement passed: !(gen == 6 && e->exec_id == I915_EXEC_BSD)
(prime_busy:4240) DEBUG: Test requirement passed: __gem_execbuf(fd, &execbuf) == 0
(prime_busy:4240) CRITICAL: Test assertion failure function one, file prime_busy.c:136:
(prime_busy:4240) CRITICAL: Failed assertion: prime_busy(&pfd[SCRATCH], false)
****  END  ****
Subtest after-render: FAIL (0.002s)

I will attach dmesg.

(In reply to cprigent from comment #2)
> Also reproduced on KBL-U. 5 subcases are fail:
> prime_busy@after-render
> prime_busy@after-vebox
> prime_busy@before-blt
> prime_busy@before-default
> prime_busy@before-render

Hmm, they should all fail.

issue is still present with the following config:

 Software information
============================================
Kernel version                  : 4.7.0-rc5-nightly+
Linux distribution              : Ubuntu 16.04 LTS
Architecture                    : 64-bit
xf86-video-intel version        : 2.99.917
Xorg-Xserver version            : 1.18.99.1
DRM version                     : 2.4.68
VAAPI version                   : Intel i965 driver for Intel(R) Broxton - 1.7.2.pre1 (1.7.0-53-gbcde10d)
Cairo version                   : 1.15.2
Intel GPU Tools version         : 
Kernel driver in use            : i915
Bios revision                   : 144.10
KSC revision                    : 1.6


 Hardware information
============================================
Platform                        : BXT-P
Motherboard model               : Broxton P
Motherboard manufacturer        : Intel Corp.
GPU Card                        : Intel Corporation Device 5a84 (rev 0a) (prog-if 00 [VGA controller])

 Firmwares information
============================================
DMC version                     : 1.7
GUC version found               : 8.7

Should be fixed by https://patchwork.freedesktop.org/series/9603/

(In reply to Chris Wilson from comment #5)
> Should be fixed by https://patchwork.freedesktop.org/series/9603/

Hi Chris :

with the series all prime_busy family pass now

kernel
======================
commit bdad748a827cef1d0605569008a6df80aed1eca3
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Thu Jul 7 09:42:10 2016 +0100

    Revert "drm/i915: Clean up associated VMAs on context destruction"

    This reverts commit e9f24d5fb7cf3628b195b18ff3ac4e37937ceeae.

    The patch was only a stop-gap measure that fixed half the problem - the
    leak of the fbcon when restarting X. A complete solution required
    releasing the VMA when the object itself was closed rather than rely on
    file/process exit. The previous patches add the VMA tracking necessary
    to do close them along with the object, context or file, and so the time
    has come to remove the partial fix.

    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

sorry for my last comment, no all tests are pass, the following tests are skopepd :

before-bsd1
after-bsd1
before-bsd2
after-bsd2

(In reply to Chris Wilson from comment #5)
> Should be fixed by https://patchwork.freedesktop.org/series/9603/

Chris could you tell us when the path is on nightly branch ?

Thanks

*** This bug has been marked as a duplicate of bug 97076 ***