[Support] ich777 - AMD Vendor Reset, CoralTPU, hpsahba,...


Recommended Posts

1 hour ago, alturismo said:

i think someone around @ich777 also has the same issue with rocket lake cpu

11 hours ago, naxos said:

I've got an i5-11400

Tone mapping seems to be a little buggy with 11th Gen and Plex currently.

 

11 hours ago, naxos said:

is forced to transcode at 2 Mbps which takes it down to SD

Please make sure that your port forwarding is set right:

  • Internal IP
  • Port
  • Protocol
  • Like 1
Link to comment
On 12/8/2021 at 1:23 AM, ich777 said:

I will try to reflash my card (Mellanox Connect X3) and will see if the same happens to me and report back.

 

Where you able to flash this exact card in the past or did you flashed other cards?

 

Do you use this command: 'mstflint -d 01:00.0 -i *.bin burn' with the asterisk or do you specify the full filename?

 

EDIT: Flash went fine without any error:

mstflint -d 01:00.0 -i fw-ConnectX3-rel-2_42_5000-MCX311A-XCA_Ax-FlexBoot-3.4.752.bin burn

    Current FW version on flash:  2.42.5000
    New FW version:               2.42.5000

    Note: The new FW version is the same as the current FW version on flash.

 Do you want to continue ? (y/n) [n] : y

Burn process will not be failsafe. No checks will be performed.
ALL flash, including the Invariant Sector will be overwritten.
If this process fails, computer may remain in an inoperable state.

 Do you want to continue ? (y/n) [n] : y
Burning FS2 FW image without signatures - OK  
Restoring signature                     - OK

 

Also removed the boot ROM after that:

mstflint -d 01:00.0 --allow_rom_change drom

-I- Preparing to remove ROM ...
Removing ROM image    - OK  
Restoring signature  - OK

 

Same errors with bin name

 

-W- DMA burning is not supported due to BME is unset (Bus Master Enable).
FSMST_INITIALIZE -   OK
Writing Boot image component -   OK
-E- Burning FS3 image failed: Unexpected state

 

 

Link to comment

Any idea what firewire chipset devices would be supported by this? I believe that the two main chipsets are Texas Instruments and Via.

 

I believe that Texas Instruments cards are generally preferred, but I have an older unRAID server running 6.4.1 with the old LibreElec DVB kernel and in this server I have a card that reports as a VT6315 Firewire Contoller, which would mean it has a Via chipset, and it works fine.

Link to comment
42 minutes ago, wayner said:

I believe that Texas Instruments cards are generally preferred, but I have an older unRAID server running 6.4.1 with the old LibreElec DVB kernel and in this server I have a card that reports as a VT6315 Firewire Contoller, which would mean it has a Via chipset, and it works fine.

This should work fine too as long as it reports as OHCI device.

 

One idea would be that you take a empty USB key install unRAID on it and boot from that one, (don't start the Array,) register trail, install the CA App, download the Firewire Driver plugin and pull the Diagnostics and post it here or issue 'lspci -v' from the terminal and see if the driver is loaded successfully.

Link to comment

Hi sir, not sure if anyone already reported this:

This plugin seems not working with my 11700T CPU, I can't the /dev/dri after restart and saw some error and warning in log like:

Dec 21 21:31:12 OMNI kernel: ACPI: Early table checksum verification disabled
Dec 21 21:31:12 OMNI kernel: ACPI BIOS Error (bug): Could not resolve symbol [\TBTS], AE_NOT_FOUND (20200925/psargs-330)
Dec 21 21:31:12 OMNI kernel: ACPI: Ignoring error and continuing table load
Dec 21 21:31:12 OMNI kernel: ACPI Error: Skipping While/If block (20200925/psloop-427)
Dec 21 21:31:12 OMNI kernel: ACPI BIOS Error (bug): Could not resolve symbol [\TBTS], AE_NOT_FOUND (20200925/psargs-330)
Dec 21 21:31:12 OMNI kernel: ACPI: Ignoring error and continuing table load
Dec 21 21:31:12 OMNI kernel: ACPI Error: Skipping While/If block (20200925/psloop-427)
Dec 21 21:31:12 OMNI kernel: floppy0: no floppy controllers found
Dec 21 21:31:12 OMNI kernel: random: 7 urandom warning(s) missed due to ratelimiting
Dec 21 21:31:13 OMNI rsyslogd: error during parsing file /etc/rsyslog.conf, on or before line 66: errors occured in file '/etc/rsyslog.conf' around line 66 [v8.2002.0 try https://www.rsyslog.com/e/2207 ]
Dec 21 21:31:16 OMNI rpc.statd[2044]: Failed to read /var/lib/nfs/state: Success
Dec 21 21:31:35 OMNI rc.docker: Plex-Media-Server: Error response from daemon: error gathering device information while adding custom device "/dev/dri": no such file or directory
Dec 21 21:31:35 OMNI rc.docker: Error: failed to start containers: Plex-Media-Server

But my cpu is working with below command:

cd /boot/config/modprobe.d

vi i915.conf

options i915 force_probe=4c8a

options i915 enable_guc=2

Do you need anymore info for this trouble shotting? please let me know

Link to comment
6 minutes ago, Lancer said:

Hi sir, not sure if anyone already reported this:

I think you are talking about the Intel GPU TOP plugin or am I wrong?

Should also work with 11th Gen CPUs fine.

Can you please send me your Diagnostics?

 

6 minutes ago, Lancer said:

Do you need anymore info for this trouble shotting? please let me know

It would be nice if you can do the following:

  1. Remove the file /boot/config/modprobe.d/i915.conf with the command 'rm -rf /boot/config/modprobe.d/i915.conf'
  2. Install the Intel GPU TOP plugin
  3. Reboot your server
  4. Pull the Diagnostics and post them here

 

Please remember if you are on 6.10.0-rc2 you have to blacklist the i915 module with the command, otherwise Intel GPU TOP won't be able to fix the wrong initialization from the module:

mkdir -p /boot/config/modprobe.d
echo "blacklist i915" > /boot/config/modprobe.d/i915.conf

 

Link to comment
  • 2 weeks later...

I bought the cheapest firewire card on Amazon which uses a Via chipset (VT6306/7/8) and it works perfectly.  Just posting in case anyone wonders about card compatibility in the future.

 

IOMMU group 12:				[1b21:1080] 03:00.0 PCI bridge: ASMedia Technology Inc. ASM1083/1085 PCIe to PCI Bridge (rev 01)
			 	[1106:3044] 04:00.0 FireWire (IEEE 1394): VIA Technologies, Inc. VT6306/7/8 [Fire II(M)] IEEE 1394 OHCI Controller (rev 80)

 

  • Like 1
Link to comment
  • ich777 changed the title to [Support] ich777 - AMD Vendor Reset, CoralTPU, hpsahba,...

Huge thanks @ich777 for the amazing effort. 

 

Nevermind this question found the answer in the description after installing the plugin :D 

"If you haven't created your own Images yet, you can do that by downloading the Docker Container 'Unraid-Kernel-Helper' from the CA App."

___ 

One simple (potentially dumb) question, when I searching the community applications I see both a plugin and doctor, which is the right one to use ? I'm looking to patch amd reset bug with this :) 

 

Thanks a ton !

kernal_helper.PNG

Edited by iskoala
found the answer to my question
Link to comment
24 minutes ago, iskoala said:

One simple (potentially dumb) question, when I searching the community applications I see both a plugin and doctor, which is the right one to use ? I'm looking to patch amd reset bug with this :)

If you want the AMD Vendor Reset Patch you have to be at least on 6.10.0 to find it in the CA App.

 

The Kernel-Helper is deprecated with v6.10.0 you only see it because you're on 6.9.2.

  • Like 1
Link to comment
6 minutes ago, ich777 said:

If you want the AMD Vendor Reset Patch you have to be at least on 6.10.0 to find it in the CA App.

 

The Kernel-Helper is deprecated with v6.10.0 you only see it because you're on 6.9.2.

 

Thanks a lot for the quick response, I will update to 6.10 instead then :). much appreciation 

  • Like 1
Link to comment

Radeon Top plugin

After a while running frigate with AMD hardware acceleration, the GPU utilisation goes to 100%. Frigate still works, unmanic still works with HW decoding, but it indicates 100% usage (added a couple screenshots). Is it actually at 100% usage, or am i running into an issue? After a reboot it comes back to normal (about 8% usage for frigate, going up to ~80% with frigate and unmanic going).

 

i have noticed some instability with frigate in general, so have just reverted back to using CPU only there. unmanic appears fine though, even when it says the GPU is at 100%

 

HW:

AMD 5600G CPU

16GB RAM

 

Unraid version 6.10 RC2

 

Thanks for the help

 

Sorry for the poor quality pictures, I was in a hurry out the door

20220104_150939.jpg

20220104_150926.jpg

Link to comment
5 hours ago, eatoff said:

Is it actually at 100% usage, or am i running into an issue?

It could be an issue but I don't think so.

 

5 hours ago, eatoff said:

i have noticed some instability with frigate in general, so have just reverted back to using CPU only there. unmanic appears fine though, even when it says the GPU is at 100%

What if you turn off HW transcoding in Frigate in general and only transcode with unmanic, does the same thing happen if yes, can you also turn on HW transcoding in Frigate and turn it off in unmanic?

I think there could be an issue with one of the two applications and not with radeontop.

 

HW transcoding will still work when it's at 100% but it should be a little slower than usual.

Link to comment
20 hours ago, ich777 said:

What if you turn off HW transcoding in Frigate in general and only transcode with unmanic, does the same thing happen if yes, can you also turn on HW transcoding in Frigate and turn it off in unmanic?

I think there could be an issue with one of the two applications and not with radeontop.

So, its been 24hrs after swtiching Frigate to CPU only, and letting plex and unmanic have at the GPU. GPU usage appears to be correct, drops to 0% when nothing is being transcoded, and shows up to 15% usage while trancoding in Plex.

 

Looks like it could be a Frigate issue. but when it was hitting 100% it needed a restart of the unraid system (not just the frigate docker) to get it to return to normal.

 

EDIT: to add, the CPU is a 5600G, integrated AMD graphics, are these drivers the latest and greatest for that?

 

Added an image of what I'm seeing with unmanic doing a transcode for a h.264 file to h.265 - Indicates GPU usage at 0% now. i was certain it showed roughly 80% usage when it was running unmanic.

radeontop.png

 

And here is a pic with both Plex transcoding and Unmanic running (just for info)

radeontop_2.png

Edited by eatoff
update
Link to comment
5 hours ago, eatoff said:

EDIT: to add, the CPU is a 5600G, integrated AMD graphics, are these drivers the latest and greatest for that?

The "drivers" are included in the Kernel so you technically don't have to install anything.

Maybe the next update from unRAID with a new Kernel version will solve the issue.

 

Also radeontop only enables the driver and installs radeontop itself so you can see the usage from your GPU nothing more.

 

But I really don't know if Frigate fully supports AMD iGPUs or better speaking APUs.

 

5 hours ago, eatoff said:

Added an image of what I'm seeing with unmanic doing a transcode for a h.264 file to h.265 - Indicates GPU usage at 0% now. i was certain it showed roughly 80% usage when it was running unmanic.

Maybe it transcodes audio first and then starts to transcode video or vice versa...?

Link to comment

How do I actually use this? The https://github.com/gnif/vendor-reset page says something about adding some kernel flags. Do I need to that (or anything else) in UNRAID to actually use the plugin?

 

For the record,

root@unraid:~# lsmod | grep reset
vendor_reset 86016 0

It does look like it installed correctly. Just unclear on whether I need to do anything else use use it.

 

Unraid 6.10 rc2, 5700xt btw if that matters.

Edited by nicus
add unraid version
Link to comment
19 hours ago, ich777 said:

The "drivers" are included in the Kernel so you technically don't have to install anything.

Maybe the next update from unRAID with a new Kernel version will solve the issue.

Ok, I'll sit tight for now i guess.

 

19 hours ago, ich777 said:

But I really don't know if Frigate fully supports AMD iGPUs or better speaking APUs.

The settings provided do work, but the CPU hit for going CPU only is so small I'm not too worried.

 

19 hours ago, ich777 said:

Maybe it transcodes audio first and then starts to transcode video or vice versa...?

No, it was when it was actually doing the video transcode. Thats all good, i just wanted to check if something wasnt working. It all appears to work ok, just doesnt report nicely. Thanks for all your help.

  • Like 1
Link to comment
14 minutes ago, cron144 said:

correct. I have connected a monitor (most of the time off), right.

Can you try to let it turned on amd see if the same happens again?

 

Also make sure that you are on the latest BIOS version.

 

Have you yet tried to enable the Kernel module by hand if it's the same because basically what the Intel GPU TOP plugin does is enable the i915 module and install the application intel_gpu_top from intel_gpu_tools

Link to comment

When monitor is turned on, it seems that no errors come up (so far). Do you know the background of this behaviour? What would be the best workaround to keep the monitor off and to have the possiblity to use the console.

 

I will do a bios update today.

 

Thanks for your hints so far.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.