<html> <head> <base href="https://bugs.freedesktop.org/"> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - [regression] 32bit steam games segfault when launched with DRI_PRIME=1" href="https://bugs.freedesktop.org/show_bug.cgi?id=101881#c5">Comment # 5</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - [regression] 32bit steam games segfault when launched with DRI_PRIME=1" href="https://bugs.freedesktop.org/show_bug.cgi?id=101881">bug 101881</a> from <a class="email" href="mailto:rafael.ristovski@gmail.com" title="Rafael Ristovski <rafael.ristovski@gmail.com>"> Rafael Ristovski</a> <pre>I can confirm a very similar issue on my SI Cape Verde (HD8850M GCN1.0) using the amdgpu driver with mesa. It happens even if I compile mesa's `xdemos` with -m32 in which case most of the binaries produce a segfault pointing to drmGetVersion(). Can you confirm the same thing on your end? Steps to reproduce: 1) Compile xdemos with -m32 2) Load `glthreads` in gdb 3) Switch DRI on inside gdb via 'set environment DRI_prime=1' 4) After you run the program ('r'), and it segfaults, try and see what 'bt' produces. For me its the following: #0 0xf7a9c001 in ?? () from /lib32/libc.so.6 #1 0xf7a9bceb in strdup () from /lib32/libc.so.6 #2 0xf7735c55 in drmGetVersion () from /usr/lib32/libdrm.so.2 #3 0xf71a1acc in amdgpu_winsys_create () from /usr/lib32/dri/radeonsi_dri.so #4 0xf69152d8 in ?? () from /usr/lib32/dri/radeonsi_dri.so #5 0xf6de8238 in ?? () from /usr/lib32/dri/radeonsi_dri.so #6 0xf6de2eb3 in ?? () from /usr/lib32/dri/radeonsi_dri.so #7 0xf7e0f0c5 in ?? () from /usr/lib32/libGL.so.1 #8 0xf7de040c in ?? () from /usr/lib32/libGL.so.1 #9 0xf7dde789 in glXChooseVisual () from /usr/lib32/libGL.so.1 #10 0x0804a483 in create_window () #11 0x0804abe9 in main () Note 64bit binaries work flawlessly. Using gcc 8 alpha here.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the assignee for the bug.</li> </ul> </body> </html>