Bug 212369 - AMDGPU: GPU hangs with '*ERROR* Couldn't update BO_VA (-12)' on MIPS64
Summary: AMDGPU: GPU hangs with '*ERROR* Couldn't update BO_VA (-12)' on MIPS64
Status: RESOLVED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - non Intel) (show other bugs)
Hardware: Mips64 Linux
: P1 normal
Assignee: drivers_video-dri
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2021-03-20 18:14 UTC by Xi Ruoyao
Modified: 2021-04-18 17:40 UTC (History)
0 users

See Also:
Kernel Version: 5.12.0-rc3+
Subsystem:
Regression: Yes
Bisected commit-id:

Attachments
kernel config (122.30 KB, text/plain)
2021-03-20 18:21 UTC, Xi Ruoyao 		Details
Add an attachment (proposed patch, testcase, etc.)

Description Xi Ruoyao 2021-03-20 18:14:12 UTC 
With Linus' kernel (git 1c273e10), when I try to run some program using GL (for example, gtk3-demo's GL view demo) from Xorg, GPU hangs and the system becomes inoperable.  Using SysRq key to reboot the system to gather the stack backtrace:

Mar 20 18:51:59 xry111-A1901 kernel: [drm:amdgpu_gem_va_ioctl] *ERROR* Couldn't update BO_VA (-12)
Mar 20 18:51:59 xry111-A1901 kernel: ------------[ cut here ]------------
Mar 20 18:51:59 xry111-A1901 kernel: WARNING: CPU: 1 PID: 2649 at drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:1457 amdgpu_bo_gpu_offset+0x140/0x180
Mar 20 18:51:59 xry111-A1901 kernel: Modules linked in: vfat fat mousedev hid_generic usbhid snd_hda_codec_hdmi snd_hda_intel snd_intel_dspcfg snd_hda_codec snd_hwdep snd_hda_core fuse ip_t>
Mar 20 18:51:59 xry111-A1901 kernel: CPU: 1 PID: 2649 Comm: X:cs0 Not tainted 5.12.0-rc3+ #10
Mar 20 18:51:59 xry111-A1901 kernel: Hardware name: SANCOG SANCOG-QingLong-L402-V1.0-DeskTop/SANCOG-QingLong-L402-V1.0-DeskTop, BIOS Kunlun-L402-V4.1.4 05/25/2020
Mar 20 18:51:59 xry111-A1901 kernel: Stack : 0000000000000000 0000000000000000 0000000000000008 98000001347cb8c8
Mar 20 18:51:59 xry111-A1901 kernel:         d59008b3440befaa 0000000000000000 0000000000000000 ffffffff84e034c0
Mar 20 18:51:59 xry111-A1901 kernel:         98000001347cb700 0000000000000001 98000001347cb788 0000000000000000
Mar 20 18:51:59 xry111-A1901 kernel:         20534f4942202c70 0000000000000010 ffffffff83bf89a0 98000001347cb6f8
Mar 20 18:51:59 xry111-A1901 kernel:         2d302e31562d3230 ffffffff84f40000 ffff000000000000 0000000000000001
Mar 20 18:51:59 xry111-A1901 kernel:         ffffffff84e00000 0000000000000000 00000000e39b8000 00000000000f7ded
Mar 20 18:51:59 xry111-A1901 kernel:         fffffffffffffff4 0000000000000000 98000001b47cb787 0000000000006000
Mar 20 18:51:59 xry111-A1901 kernel:         9800000121134000 98000001347c8000 98000001347cb8c0 0000000000000003
Mar 20 18:51:59 xry111-A1901 kernel:         ffffffff843f20ec 0000000000000000 0000000000000003 ffffffff84f492c3
Mar 20 18:51:59 xry111-A1901 kernel:         ffffffff8530c070 ffffffff84e034c0 ffffffff8379b950 d59008b3440befaa
Mar 20 18:51:59 xry111-A1901 kernel:         ...
Mar 20 18:51:59 xry111-A1901 kernel: Call Trace:
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff8379b950>] show_stack+0x40/0x130
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff843f20ec>] dump_stack+0xb4/0xe0
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff843edd60>] __warn+0xc0/0x114
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff843ede28>] warn_slowpath_fmt+0x74/0xc4
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83d2b7c0>] amdgpu_bo_gpu_offset+0x140/0x180
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83d3514c>] amdgpu_cs_ioctl+0x1154/0x1bb8
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83ccfc64>] drm_ioctl_kernel+0xcc/0x120
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83cd0158>] drm_ioctl+0x220/0x408
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83d0fb58>] amdgpu_drm_ioctl+0x58/0x98
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff839bd224>] sys_ioctl+0xcc/0xe8
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff837a7210>] syscall_common+0x34/0x58
Mar 20 18:51:59 xry111-A1901 kernel: 
Mar 20 18:51:59 xry111-A1901 kernel: ---[ end trace 678210b5f2b1f5e4 ]---
Mar 20 18:51:59 xry111-A1901 kernel: [drm:amdgpu_cs_ioctl] *ERROR* Not enough memory for command submission!
Mar 20 18:51:59 xry111-A1901 kernel: CPU 1 Unable to handle kernel paging request at virtual address 0000000000000008, epc == ffffffff83d414a0, ra == ffffffff83d41434
Mar 20 18:51:59 xry111-A1901 kernel: Oops[#1]:
Mar 20 18:51:59 xry111-A1901 kernel: CPU: 1 PID: 2649 Comm: X:cs0 Tainted: G        W         5.12.0-rc3+ #10
Mar 20 18:51:59 xry111-A1901 kernel: Hardware name: SANCOG SANCOG-QingLong-L402-V1.0-DeskTop/SANCOG-QingLong-L402-V1.0-DeskTop, BIOS Kunlun-L402-V4.1.4 05/25/2020
Mar 20 18:51:59 xry111-A1901 kernel: $ 0   : 0000000000000000 0000000000000001 98000001345dc3f0 98000001483f1000
Mar 20 18:51:59 xry111-A1901 kernel: $ 4   : 980000027aa36c00 98000001345dc3b8 98000001483f1068 0000000000000001
Mar 20 18:51:59 xry111-A1901 kernel: $ 8   : 000000000006a600 9800000148b3f880 0000000000000000 0000000000288a35
Mar 20 18:51:59 xry111-A1901 kernel: $12   : 0000000000000000 ffffffff83bf89a8 00000000000000eb 00000000000000eb
Mar 20 18:51:59 xry111-A1901 kernel: $16   : 98000001345dc3d8 9800000120780000 6db6db6db6db0000 98000001483f1048
Mar 20 18:51:59 xry111-A1901 kernel: $20   : ffffffff84f40000 98000001347cbc10 0000000000000000 98000001483f1000
Mar 20 18:51:59 xry111-A1901 kernel: $24   : ffffffff84411960 0000000000000000                                  
Mar 20 18:51:59 xry111-A1901 kernel: $28   : 98000001347c8000 98000001347cbb40 ffffffff84f40000 ffffffff83d41434
Mar 20 18:51:59 xry111-A1901 kernel: Hi    : 0000000000000002
Mar 20 18:51:59 xry111-A1901 kernel: Lo    : 0000000000103c06
Mar 20 18:51:59 xry111-A1901 kernel: epc   : ffffffff83d414a0 amdgpu_vm_update_pdes+0xf0/0x268
Mar 20 18:51:59 xry111-A1901 kernel: ra    : ffffffff83d41434 amdgpu_vm_update_pdes+0x84/0x268
Mar 20 18:51:59 xry111-A1901 kernel: Status: 5400cce3        KX SX UX KERNEL EXL IE 
Mar 20 18:51:59 xry111-A1901 kernel: Cause : 10000008 (ExcCode 02)
Mar 20 18:51:59 xry111-A1901 kernel: BadVA : 0000000000000008
Mar 20 18:51:59 xry111-A1901 kernel: PrId  : 0014c004 (ICT Loongson-3)
Mar 20 18:51:59 xry111-A1901 kernel: Modules linked in: vfat fat mousedev hid_generic usbhid snd_hda_codec_hdmi snd_hda_intel snd_intel_dspcfg snd_hda_codec snd_hwdep snd_hda_core fuse ip_t>
Mar 20 18:51:59 xry111-A1901 kernel: Process X:cs0 (pid: 2649, threadinfo=00000000a6d48d91, task=0000000007ecd99e, tls=000000ffe7b0a6c0)
Mar 20 18:51:59 xry111-A1901 kernel: Stack : 980000012cb70158 0400000000000076 980000013421b780 ffffffff83d41b50
Mar 20 18:51:59 xry111-A1901 kernel:         0000000000000000 9800000120780000 98000001483f1000 0000000000000000
Mar 20 18:51:59 xry111-A1901 kernel:         0000000000000000 980000027aa37000 0000000000000100 d59008b3440befaa
Mar 20 18:51:59 xry111-A1901 kernel:         0000000000000002 0000000000000000 ffffffff84f40000 98000001483f1000
Mar 20 18:51:59 xry111-A1901 kernel:         9800000120780000 98000001347cbc10 980000013442a180 980000012d666858
Mar 20 18:51:59 xry111-A1901 kernel:         ffffffff84f40000 ffffffff83d31dc8 98000001347cbc48 98000001347cbc28
Mar 20 18:51:59 xry111-A1901 kernel:         98000001347cbc00 98000001347cbc00 9800000120988f00 0000000000038623
Mar 20 18:51:59 xry111-A1901 kernel:         0001000000000002 98000001347cbbf0 98000001347cbc48 980000012d666858
Mar 20 18:51:59 xry111-A1901 kernel:         9800000100000000 98000001347cbc28 98000001347cbbf0 980000012cb70058
Mar 20 18:51:59 xry111-A1901 kernel:         0000000000000004 9800000134928032 0000000000000000 0000000000000000
Mar 20 18:51:59 xry111-A1901 kernel:         ...
Mar 20 18:51:59 xry111-A1901 kernel: Call Trace:
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83d414a0>] amdgpu_vm_update_pdes+0xf0/0x268
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83d31dc8>] amdgpu_gem_va_ioctl+0x3d8/0x430
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83ccfc64>] drm_ioctl_kernel+0xcc/0x120
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83cd0158>] drm_ioctl+0x220/0x408
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff83d0fb58>] amdgpu_drm_ioctl+0x58/0x98
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff839bd224>] sys_ioctl+0xcc/0xe8
Mar 20 18:51:59 xry111-A1901 kernel: [<ffffffff837a7210>] syscall_common+0x34/0x58
Mar 20 18:51:59 xry111-A1901 kernel: 
Mar 20 18:51:59 xry111-A1901 kernel: Code: 12c00002  00000000  ded602a8 <dede0008> dfc202b0  10400004  00001825  dc4202b0  1440fffe 
Mar 20 18:51:59 xry111-A1901 kernel: 
Mar 20 18:51:59 xry111-A1901 kernel: ---[ end trace 678210b5f2b1f5e5 ]---

The system is MIPS64 (Loongson 3A4000) and the GPU is RX550 (640SP).

Linux-5.10.17 works properly, but 5.11.0 also has this issue.  I tried bisection but it gone into some commits before 5.10.0 (due to merging commits), which can't boot my system.

My guess is some change between 5.10.0 and 5.11.0 has broke non-4K-page systems.
Comment 1 Xi Ruoyao 2021-03-20 18:21:04 UTC 
Created attachment 295963 [details]
kernel config
Comment 2 Xi Ruoyao 2021-03-23 16:54:33 UTC 
With some workarounds, bisection gives:

first bad commit: [5b8c596976d4338942dd889b66cd06dc766424e1] Merge tag 'amd-drm-next-5.11-2020-11-05' of git://people.freedesktop.org/~agd5f/linux into drm-next
Comment 3 Xi Ruoyao 2021-04-18 17:40:51 UTC 
Fixed at 566c6e25f957ebdb0b6e8073ee291049118f47fb.
Note You need to log in before you can comment on or make changes to this bug.