Kexec results in fatal error during GPU init
On a Dell OptiPlex 5055 with Oland [Radeon HD 8570 / R7 240/340 OEM] [1002:6611], running Linux 5.13-rc1, and starting Linux 5.13-rc2 with kexec shows the errors below:
$ dmesg --level=err
[ 25.831254] amdgpu 0000:05:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0x5062ff4000 flags=0x0000]
[ 25.844072] amdgpu 0000:05:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0x5062ff4100 flags=0x0000]
[ 25.856618] amdgpu 0000:05:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0x5062ff4200 flags=0x0000]
[ 25.868989] amdgpu 0000:05:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0x5062ff4300 flags=0x0000]
[ 25.881880] amdgpu 0000:05:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0x5062ff4400 flags=0x0000]
[ 25.894663] amdgpu 0000:05:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0x5062ff4500 flags=0x0000]
[ 25.907409] amdgpu 0000:05:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0x5062ff4600 flags=0x0000]
[ 25.920265] amdgpu 0000:05:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0x5062ff4700 flags=0x0000]
[ 26.824980] amdgpu 0000:05:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring gfx test failed (-110)
[ 26.837443] [drm:amdgpu_device_init [amdgpu]] *ERROR* hw_init of IP block <gfx_v6_0> failed -110
[ 26.848696] amdgpu 0000:05:00.0: amdgpu: amdgpu_device_ip_init failed
[ 26.856892] amdgpu 0000:05:00.0: amdgpu: Fatal error during GPU initDoing a full reset, Linux logs no errors.
Running Linux 5.13-rc2, and starting Linux 5.13-rc2 with kexec shows the errors below:
$ dmesg --level=err
[ 23.976881] [drm:uvd_v3_1_hw_init [amdgpu]] *ERROR* amdgpu: UVD Firmware validate fail (-22).
[ 23.986878] [drm:amdgpu_device_init [amdgpu]] *ERROR* hw_init of IP block <uvd_v3_1> failed -22
[ 23.996991] amdgpu 0000:05:00.0: amdgpu: amdgpu_device_ip_init failed
[ 24.004417] amdgpu 0000:05:00.0: amdgpu: Fatal error during GPU init