Sorry for the message before I can not retrieve my password to reissue it in English, so I created another account ...
back after the tests (sorry I forgot to warn that I was french and I use a google translator so tell me if you do not understand)
First problem solved, IUMMO is enabled in my bios and after disabling, no more AMD-VI error message regarding SATA PCI card
I found my problem but I do not see how to solve it. In fact, my problem comes from the update I made via omv-update. After updating a problem firmware with the GPU of (I have an athlon 200GE)
Here is the dmesg report:
[15,108840] EDAC amd64: 0 Node: RDC ECC Memory Buffer Disabled.
[15.108842] EDAC amd64: ECC disabled in the BIOS or no ECC capability, the module will not be loaded.
Enable ECC checking or force the module by setting 'ecc_enable_override'.
(Note that you may notice the substitution of the training agent for the Unknown Effects.)
[15,171789] amdgpu 0000: 07: 00.0: firmware: failure to load amdgpu / raven_gpu_info.bin (-2)
[15,171979] firmware_class: See wiki .debian.org / Firmware for more information on missing firmware.
[15.172184] amdgpu 0000: 07: 00.0: Direct firmware upload for amdgpu / raven_gpu_info.bin failed with -2 error
[15.172188] amdgpu 0000: 07: 00.0: Could not load firmware gpu_info "amdgpu / raven_gpu_info" "
[15.172372] amdgpu 0000: 07: 00.0: Fatal error during the
[15.172548] [drm] amdgpu: finishing device.
[15.172949] amps: probes of 0000: 07: 00.0 failed with error -2
the first restart error (one-to-time) but there is even one time
the second is finished black and plants omv
at startup, I have the choice between the two versions of debian 4.1.21-1, it is version 4.19.0.0,
I tried again with version 4.1.21-1, it seems to work well.
I do not understand in the error message it seems that we talk about raven gpu. Me with my amd 200Ge I have a:
Graphics processor model
Radeon ™ Vega 3 Graphics
maybe the problem comes from the? no ?
if you have an idea, I'm interested