Bug 106910 - Primus Segfaults after updating Mesa to 18.1.1
Summary: Primus Segfaults after updating Mesa to 18.1.1
Status: RESOLVED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Intel 3D Bugs Mailing List
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-06-13 10:26 UTC by Alexander
Modified: 2018-08-20 19:05 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:


Attachments
Journalctl traces (5.49 KB, application/octet-stream)
2018-06-13 10:26 UTC, Alexander
Details

Description Alexander 2018-06-13 10:26:50 UTC
Created attachment 140148 [details]
Journalctl traces

After upgrading MESA to version 18.1.1 Primus (bridge for Bumblebee, the NVIDIA Optimus implementation) segfaults while trying to use applications that require GLX.Running same applications with Intel and VirtualGL (another bridge for Bumblebee that runs slower than Primus) works fine.After downgrading to Mesa 18.0.4 everything starts working again.Applications' logs don't say much.
The issue can be reproduced by:
1)Installing Bumblebee 3.2.1, MESA 18.1.1 and primus 20151110
2)Running any GLX application with primusrun
Comment 1 Alexander 2018-06-13 10:31:56 UTC
The same issue remains in git version as well.
Comment 2 Daniel Serpell 2018-06-20 16:26:33 UTC
There is a backtrace in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=901701:

----------------------------------------------------------------------------
Thread 2.3 "glxgears" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fffee092700 (LWP 10545)]
0x00007ffff02914a7 in do_blit_drawpixels (pixels=0x0, unpack=0x7fffe0150ac8, type=5121, format=32993, height=1080, width=1920, y=0, x=0, ctx=0x7fffe0147540) at ../../../../../../src/mesa/drivers/dri/i965/intel_pixel_draw.c:80
80	../../../../../../src/mesa/drivers/dri/i965/intel_pixel_draw.c: Toks failas ar aplankas neegzistuoja.
#0  0x00007ffff02914a7 in do_blit_drawpixels (pixels=0x0, unpack=0x7fffe0150ac8, type=5121, format=32993, height=1080, width=1920, y=0, x=0, ctx=0x7fffe0147540) at ../../../../../../src/mesa/drivers/dri/i965/intel_pixel_draw.c:80
        src_offset = <optimized out>
        src_buffer = <optimized out>
        dst_format = 3761243984
        src_stride = <optimized out>
        src_flip = <optimized out>
        pbo_mt = 0x0
        rb = <optimized out>
        src_format = MESA_FORMAT_B8G8R8A8_UNORM
#1  intelDrawPixels (ctx=0x7fffe0147540, x=0, y=0, width=1920, height=1080, format=32993, type=5121, unpack=0x7fffe0150ac8, pixels=0x0) at ../../../../../../src/mesa/drivers/dri/i965/intel_pixel_draw.c:167
        __func__ = "intelDrawPixels"
#2  0x00007fffefedaf33 in _mesa_DrawPixels (width=1920, height=1080, format=32993, type=5121, pixels=0x0) at ../../../src/mesa/main/drawpix.c:162
        err = <optimized out>
        ctx = 0x7fffe0147540
#3  0x00007ffff76b8345 in test_drawpixels_fast (dconfig=<optimized out>, ctx=0x7fffe00239c0, dpy=0x7fffe0000b20) at libglfork.cpp:362
        pbo = 1
        end = 8568.4529074310012
        iters = 0
        is_fast = <optimized out>
        width = 1920
        height = 1080
        pbattrs = {32833, 1920, 32832, 1080, 32795, 1, 0}
        pbuffer = 146800642
        pixeldata = <optimized out>
#4  display_work (vd=<optimized out>) at libglfork.cpp:402
        drawable = 144703490
        width = 300
        height = 300
        quad_vertex_coords = {-1, -1, -1, 1, 1, 1, 1, -1}
        quad_texture_coords = {0, 0, 0, 1, 1, 1, 1, 0}
        textures = {0, 0}
        pbos = {0, 0}
        ctex = 0
        state_names = {0x7ffff76c1244 "wait", 0x7ffff76c1249 "upload", 0x7ffff76c1250 "draw+swap", 0x0}
        profiler = {name = 0x7ffff76c1203 "display", state_names = 0x7ffff78d12e0 <display_work(void*)::state_names>, state_time = {0, 0, 0, 0, 0, 0}, prev_timestamp = 8568.2315354050006, print_timestamp = 8568.2315354050006, state = 0, nframes = 0, width = 0, height = 0}
        ddpy = 0x7fffe0000b20
        __PRETTY_FUNCTION__ = "void* display_work(void*)"
        dconfigs = 0x7fffe000fff0
        context = 0x7fffe00239c0
        use_textures = <optimized out>
#5  0x00007ffff62455aa in start_thread (arg=0x7fffee092700) at pthread_create.c:463
        pd = 0x7fffee092700
        now = <optimized out>
        unwind_buf = {cancel_jmp_buf = {{jmp_buf = {140737186965248, 6747949740819525267, 140737488347470, 140737488347471, 140737488347472, 0, -6747988065853329773, -6747970307990442349}, mask_was_saved = 0}}, priv = {pad = {0x0, 0x0, 0x0, 0x0}, data = {prev = 0x0, cleanup = 0x0, canceltype = 0}}}
        not_first_call = <optimized out>
#6  0x00007ffff6aebcbf in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95
----------------------------------------------------------------------------
Comment 3 Kenneth Graunke 2018-06-21 18:45:34 UTC
FWIW, I can't reproduce this...using primus 20151110 and bumblebee 3.2.1, X server 1.19.6+13+gd0d1a694f-2 from Archlinux, and either Mesa master or 18.1.1, and 'primusrun glxgears' is working fine for me.
Comment 4 Gavin Troy 2018-06-26 23:19:28 UTC
Xorg server 1.20 is also required (introduced by xorg-server 8d0d89715984e321315631dd6667e05813d26e03).

In mesa, this is introduced by "069fdd5f9facbd72fb6a289696c7b74e3237e70f egl/x11: Support DRI3 v1.1".

The xcb_dri3_get_supported_modifiers_reply() in that change is returning an error when using primusrun.

Building without HAVE_DRI3_MODIFIERS, I do not see the issue.
Comment 5 MiksKalnins 2018-07-16 02:32:19 UTC
Problem still exists in 18.1.4
Comment 6 CPU 2018-07-19 11:47:56 UTC
I can confirm that there is not issue if building without HAVE_DRI3_MODIFIERS.
Comment 7 Tomas Janousek 2018-08-07 06:59:49 UTC
So if this is DRI3-related, export LIBGL_DRI3_DISABLE=1 must surely be a workaround. It certainly does help in my configuration. :-)
Comment 8 Alexander 2018-08-10 13:01:08 UTC
Building MESA with LIBGL_DRI3_DISABLE and without HAVE_DRI3_MODIFIERS work fine on my config.Thanks a lot!
Comment 9 Gavin Troy 2018-08-20 19:05:38 UTC
This is fixed. The underlying issue was shared with https://bugs.freedesktop.org/show_bug.cgi?id=107117


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.