[Mesa-dev] [Bug 107351] Android 8.1: radv segfault with 3Dmark vulkan benchmarks

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Mon Jul 23 19:45:45 UTC 2018


https://bugs.freedesktop.org/show_bug.cgi?id=107351

            Bug ID: 107351
           Summary: Android 8.1: radv segfault with 3Dmark vulkan
                    benchmarks
           Product: Mesa
           Version: git
          Hardware: x86-64 (AMD64)
                OS: other
            Status: NEW
          Severity: major
          Priority: medium
         Component: Drivers/Vulkan/radeon
          Assignee: mesa-dev at lists.freedesktop.org
          Reporter: issor.oruam at gmail.com
        QA Contact: mesa-dev at lists.freedesktop.org

Created attachment 140796
  --> https://bugs.freedesktop.org/attachment.cgi?id=140796&action=edit
logcat of api_overhead test

Hi,

while performing some test on vulkan radv HAL on Android 8.1 (oreo-x86)
build a segfault was noticed and traced with 3Dmark vulkan benchmark.

The build is experimental and based on hwcomposer.drm + gralloc.gbm
and tested on HD7790 gpu (Bonaire)

The same 3Dmark vulkan benchmarks work without segfault on Intel anv running on
Skylake GT2, so it is better to investigate why radv crashes.

Inspection of crash with addr2line -Cfe returns last line of code:
src/amd/vulkan/radv_image.c: 1289

Mauro Rossi
android-x86 team


--------- beginning of crash
07-22 21:31:15.938  4949  4995 F libc    : Fatal signal 11 (SIGSEGV), code 1,
fault addr 0x1c in tid 4995 (Thread-9), pid 4949 (cation:workload)
...
07-22 21:31:15.956  5007  5007 F DEBUG   : *** *** *** *** *** *** *** *** ***
*** *** *** *** *** *** ***
07-22 21:31:15.956  5007  5007 F DEBUG   : Build fingerprint:
'Android-x86/android_x86_64/x86_64:8.1.0/OPM6.171019.030.E1/maur07141051:userdebug/test-keys'
07-22 21:31:15.956  5007  5007 F DEBUG   : Revision: '0'
07-22 21:31:15.956  5007  5007 F DEBUG   : ABI: 'x86'
07-22 21:31:15.956  5007  5007 F DEBUG   : pid: 4949, tid: 4995, name: Thread-9
 >>> com.futuremark.dmandroid.application:workload <<<
07-22 21:31:15.956  5007  5007 F DEBUG   : signal 11 (SIGSEGV), code 1
(SEGV_MAPERR), fault addr 0x1c
07-22 21:31:15.956  5007  5007 F DEBUG   : Cause: null pointer dereference
07-22 21:31:15.956  5007  5007 F DEBUG   :     eax 00000000  ebx c65b690c  ecx
c2a6d0ac  edx 00000097
07-22 21:31:15.956  5007  5007 F DEBUG   :     esi c2a6d0ac  edi c1fbbb00
07-22 21:31:15.956  5007  5007 F DEBUG   :     xcs 00000023  xds 0000002b  xes
0000002b  xfs 0000006b  xss 0000002b
07-22 21:31:15.956  5007  5007 F DEBUG   :     eip c64e0d47  ebp c2b7eb38  esp
c2b7eaf0  flags 00010286
07-22 21:31:16.071  5007  5007 F DEBUG   : 
07-22 21:31:16.071  5007  5007 F DEBUG   : backtrace:
07-22 21:31:16.071  5007  5007 F DEBUG   :     #00 pc 000f1d47 
/system/vendor/lib/hw/vulkan.radv.so (radv_image_create+199)
07-22 21:31:16.071  5007  5007 F DEBUG   :     #01 pc 000f2f26 
/system/vendor/lib/hw/vulkan.radv.so (radv_CreateImage+166)
07-22 21:31:16.071  5007  5007 F DEBUG   :     #02 pc 0000b59a 
/android/system/lib/libvulkan.so

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20180723/9fb0f4e7/attachment.html>


More information about the mesa-dev mailing list