Bug 105172

Summary: [bisected] Unable to start X on Vega (EE) AMDGPU(0): amdgpu_setup_kernel_mem failed
Product: DRI Reporter: James Harvey <lothmordor>
Component: DRM/AMDgpuAssignee: Default DRI bug account <dri-devel>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: medium    
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
xorg log
none
dmesg with amdgpu.dc_log=1 drm.debug=6 none

Description James Harvey 2018-02-20 08:43:59 UTC
Created attachment 137457 [details]
xorg log

# first bad commit: [5b565e0e5a9872f8c5a459ce53f8d6a4b19a1a66] drm/amdgpu: expose the VA above the hole to userspace

commit 5b565e0e5a9872f8c5a459ce53f8d6a4b19a1a66 (refs/bisect/bad)
Author: Christian König <christian.koenig@amd.com>
Date:   Tue Nov 7 12:03:31 2017 +0100

    drm/amdgpu: expose the VA above the hole to userspace
    
    Let userspace know how much area we have above the 48bit VA hole on
    Vega10.
    
    Signed-off-by: Christian König <christian.koenig@amd.com>
    Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
    Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

My system:

Software versions:
    Linux 4.15.0-rc2+ x86_64
    OpenGL core profile version string: 4.5 (Core Profile) Mesa 18.1.0-devel (git-7be5f30bb1)

GPU hardware:
    OpenGL renderer string: Radeon RX Vega (VEGA10 / DRM 3.23.0 / 4.15.0-rc2+, LLVM 7.0.0)
    03:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Vega 10 XT [Radeon RX Vega 64] [1002:687f] (rev c3)

CPU hardware:
    AMD Phenom(tm) II X4 955 Processor

Weston also fails to start:

weston-launch: spawned weston with pid: 2501
Date: 2018-02-19 PST
[21:52:45.451] weston 3.0.0
               http://wayland.freedesktop.org
               Bug reports to: https://bugs.freedesktop.org/enter_bug.cgi?product=Wayland&component=weston&version=3.0.0
               Build: unknown (not built from git or tarball)
[21:52:45.451] Command line: /usr/bin/weston
[21:52:45.451] OS: Linux, 4.16.0-rc2, #2 SMP PREEMPT Mon Feb 19 19:32:23 PST 2018, x86_64
[21:52:45.452] Using config file '/home/james/.config/weston.ini'
[21:52:45.452] Output repaint window is 7 ms maximum.
[21:52:45.452] Loading module '/usr/lib64/libweston-3/drm-backend.so'
[21:52:45.453] initializing drm backend
weston-launch: opened /dev/dri/card0: ret: 0, fd: 4
[21:52:45.454] using /dev/dri/card0
[21:52:45.455] Loading module '/usr/lib64/libweston-3/gl-renderer.so'
radeonsi: Failed to create a context.
[21:52:45.531] EGL client extensions: EGL_EXT_client_extensions
               EGL_EXT_platform_base EGL_KHR_client_get_all_proc_addresses
               EGL_KHR_debug EGL_EXT_platform_wayland EGL_EXT_platform_x11
               EGL_MESA_platform_gbm EGL_MESA_platform_surfaceless
[21:52:45.531] warning: neither EGL_EXT_swap_buffers_with_damage or EGL_KHR_swap_buffers_with_damage is supported. Performance could be aff$
[21:52:45.531] EGL_KHR_surfaceless_context available
radeonsi: Failed to create a context.
[21:52:45.531] failed to create context
[21:52:45.531] EGL error state: EGL_BAD_ALLOC (0x3003)
[21:52:45.531] EGL error state: EGL_SUCCESS (0x3000)
Comment 1 James Harvey 2018-02-20 08:45:33 UTC
Created attachment 137458 [details]
dmesg with amdgpu.dc_log=1 drm.debug=6
Comment 2 Christian König 2018-02-20 08:55:59 UTC
Have you updated libdrm recently as well?
Comment 3 James Harvey 2018-02-20 09:03:06 UTC
I last updated libdrm from git on the 14th, I'll update it and see if that helps.
Comment 5 Christian König 2018-02-20 09:08:44 UTC
Yeah that commit came to my mind as well.
Comment 6 James Harvey 2018-02-20 10:18:45 UTC
Thanks for your patience guys.

Using kernel 4.16-rc2 (but with the bisected commit reverted) X and weston work.

Updating libdrm to master-33a2851, X and weston work again (yay!).

Updating also mesa (master-af5f232), now I'm getting hard lockups.  Downgrading the kernel to 4.15-rc9 or using the 4.16-rc2 kernel with 5b565e0 reverted and things work again.  Maybe a new related bug, but in mesa?

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.