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


Recommended Posts

Hi!

 

I would love to get this rmmr patch running on my Gen8 Microserver, I am currently running 6.9.2 so cannot see the kernal helper in the app section, would it be visible if I changed to 6.10.0-rc3?  there are no other options in the update os branch?

 

The prerequisite is 6.9.0 RC 2 and up I would have thought that would have included 6.9.2?

 

The purpose for me is to pass through a 4 port  network card

 

Many thanks

Edited by mbc0
Link to comment
3 hours ago, mbc0 said:

I would love to get this rmmr patch running on my Gen8 Microserver, I am currently running 6.9.2 so cannot see the kernal helper in the app section, would it be visible if I changed to 6.10.0-rc3?

I think you didn‘t read the changelog:

Quote

[rc2] Added so-called "add-relaxable-rmrr-5_8_and_up.patch" modified for our kernel
https://github.com/kiler129/relax-intel-rmrr/blob/master/patches/add-relaxable-rmrr-5_8_and_up.patch

Thanks to @ich777 for pointing this out.


The RMRR patch is integrated directly into Unraid since RC2 and you only have to add this:

intel_iommu=relax_rmrr

to your syslinux.conf and it will work right OOB since RC2.

Link to comment
10 minutes ago, MrLondon said:

I have not blacklisted anything, do you believe that it would work if I blacklist it with Alder Lake?

It is not loading the driver correctly.

 

what is the output for cat /boot/config/modprobe.d/i915.conf should be

 

root@computenode:~# cat /boot/config/modprobe.d/i915.conf 
blacklist i915
root@computenode:~# 

 

Mar 11 17:15:33 Tower kernel: i915 0000:00:02.0: Your graphics device 4692 is not properly supported by the driver in this
Mar 11 17:15:33 Tower kernel: kernel version. To force driver probe anyway, use i915.force_probe=4692
Mar 11 17:15:33 Tower kernel: module parameter or CONFIG_DRM_I915_FORCE_PROBE=4692 configuration option,

Edited by SimonF
Link to comment
1 hour ago, MrLondon said:

I have not blacklisted anything, do you believe that it would work if I blacklist it with Alder Lake?

Blacklist the driver like @SimonF mentioned and Intel-GPU-TOP will take care of the probe.

 

Don't forget to reboot after you've blacklisted the driver and maybe post the Diagnostics afterwards.

 

To blacklist the driver run this from an Unraid terminal:

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

@ich777

 

Thank you for your reply 🙂

 

I cannot see the changelog link?  Maybe this will help me and others if this topic is modified to day it is not in use?

 

I went from this post

 

to this post titled New & Current Method

 

 

On this post there is no mention of a change? it is read as it is titled as current method

 

Thank you for letting me know I just need to add intel_iommu=relax_rmrr to my syslinux.cfg but if I click the link to github that you provided the syslinux.cfg looks completely different to mine? do I just put intel_iommu=relax_rmrr at the end? I cannot find any documentation for this?

 

This is my syslinux.cfg

 

default menu.c32
menu title Lime Technology, Inc.
prompt 0
timeout 50
label Unraid OS
  menu default
  kernel /bzimage
  append pcie_acs_override=downstream initrd=/bzroot
label Unraid OS GUI Mode
  kernel /bzimage
  append pcie_acs_override=downstream initrd=/bzroot,/bzroot-gui
label Unraid OS Safe Mode (no plugins, no GUI)
  kernel /bzimage
  append initrd=/bzroot unraidsafemode
label Unraid OS GUI Safe Mode (no plugins)
  kernel /bzimage
  append initrd=/bzroot,/bzroot-gui unraidsafemode
label Memtest86+
  kernel /memtest
 

 

Thank you so much for your time!

 

 

 

 

 

Link to comment
56 minutes ago, mbc0 said:

I cannot see the changelog link?

Because I've quoted the part from the RC3 Changelog where it says that it's now part of Unraid itself, anyways here is the link: Click

 

57 minutes ago, mbc0 said:

I went from this post

Those two posts are now also deprecated because it's part of Unraid itself.

 

Go to the Unraid "Main" tab -> click on your USB Boot device (blue text "flash") -> at the green Syslinux configuration add this (just add it if you have any other things there):

image.thumb.png.6d97b7d68e0b203e4eac1481942ad77d.png

 

Click on the bottom on "Apply" and reboot your server.

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

Because I've quoted the part from the RC3 Changelog where it says that it's now part of Unraid itself, anyways here is the link: Click

 

Those two posts are now also deprecated because it's part of Unraid itself.

 

Go to the Unraid "Main" tab -> click on your USB Boot device (blue text "flash") -> at the green Syslinux configuration add this (just add it if you have any other things there):

image.thumb.png.6d97b7d68e0b203e4eac1481942ad77d.png

 

Click on the bottom on "Apply" and reboot your server.

 

Thank you, that makes sense now 🙂

 

Unfortunately I still have the problem

 

internal error: qemu unexpectedly closed the monitor: 2022-03-12T11:53:40.135099Z qemu-system-x86_64: -device vfio-pci,host=0000:07:00.2,id=hostdev0,bus=pci.3,addr=0x0: vfio 0000:07:00.2: failed to setup container for group 13: Failed to set iommu for container: Operation not permitted

 

It does this for groups 12-14 (for some reason I cannot select group 11?) 

 

image.thumb.png.6a047d5136828ba641fa9375968b0060.png

 

What is annoying is I did this years ago successfully with the same hardware and I have resurrected the microserver.

 

I have entered the command as shown

 

image.thumb.png.20b5b0321297139d35288c096e8d2ae5.png

 

Any ideas what could be the problem?

 

Thank you

 

 

 

 

 

tower-diagnostics-20220312-1158.zip

Link to comment
1 hour ago, mbc0 said:

It does this for groups 12-14 (for some reason I cannot select group 11?) 

I think because 11 is used for Unraid…

 

Is this a quad gigabit card? I would recommend that you bind all NICs that you want to use to VFIO and then try to pass them through.

Link to comment
2 hours ago, ich777 said:

I think because 11 is used for Unraid…

 

Is this a quad gigabit card? I would recommend that you bind all NICs that you want to use to VFIO and then try to pass them through.

 

Hi, yes it is a quad gigabit card, I have tried singularly passing through ports 12-14 (11 is greyed out) 

 

I have removed the card and now the 2 ports on the Gen8 are showing in group 11 is that what you mean? does unraid HAVE to use 11?

 

When you say bind all the NICS that I want to pass through on the 4 port card, do you mean within unraid network settings?

 

Thank you 

 

This is the list with the Quad NIC removed

 

image.thumb.png.8abcd53a15263f969ee05c714eb0b44d.png

 

 

This is with the QUAD NIC installed

 

image.thumb.png.88be1eda28e3adfca6672a940af2809f.png

Edited by mbc0
Link to comment
2 hours ago, mbc0 said:

does unraid HAVE to use 11?

Of course not.

 

2 hours ago, mbc0 said:

When you say bind all the NICS that I want to pass through on the 4 port card, do you mean within unraid network settings?

Tick the boxes at System Devices and at the bottom click „Bind on Reboot to VFIO“ and reboot afterwards.

 

This question would be better suired in the VM sub forums how to bind cards, RMRR is working from what I see.

Link to comment
1 hour ago, ich777 said:

Of course not.

 

Tick the boxes at System Devices and at the bottom click „Bind on Reboot to VFIO“ and reboot afterwards.

 

This question would be better suired in the VM sub forums how to bind cards, RMRR is working from what I see.

 

OK, thank you anyway, I did what you said and all 4 ports are now tickable and available in the VM but I still get an error when starting the VM

 

image.thumb.png.db6173228d38b417c16b6d4ec2d74405.png

 

image.png.f2da4271d3b22da11e188ea3b108f4ae.png

 

image.png.cfd75701ea1a5ced82133bfaf4b807dc.png

 

 

Link to comment
11 hours ago, dhstsw said:

Since RC3 the GPU isn't resetting correctly in VMs (it did with with the AMD Vendor Reset plugin in RC2 without any problem).

This is really weird because nothing else changed under the hood from what I know.

I will take a look at the build logs from the plugin but from what I know it should work just fine, keep in mind this patch is only a workaround and not a real solution to the issue...

Link to comment
50 minutes ago, mbc0 said:

Thanks, I appreciate it! so many hours on this now, would love to know what is wrong.

From what I see in your log from your VM RMRR isn't the issue here, please try to create a post in the VM support subforums, I'm not really the expert when it comes to passing through things to VMs... :/

Link to comment
Just now, ich777 said:

From what I see in your log from your VM RMRR isn't the issue here, please try to create a post in the VM support subforums, I'm not really the expert when it comes to passing through things to VMs... :/

 

Thank you, I created a post in VM support yesterday on your advice, many thanks for your time!

  • Like 1
Link to comment
4 hours ago, ich777 said:

This is really weird because nothing else changed under the hood from what I know.

I will take a look at the build logs from the plugin but from what I know it should work just fine, keep in mind this patch is only a workaround and not a real solution to the issue...


Thanks, let me know if you find anything.
Meanwhile i'm going back to RC2, can't use my VMs as it is.

C.

Link to comment
54 minutes ago, dhstsw said:

Thanks, let me know if you find anything.
Meanwhile i'm going back to RC2, can't use my VMs as it is.

I‘ve gone through the build log and everything seema fine.

 

I‘ve now read through your first post, please enable the iGPU and set it to the primary graphics device so that Unraid can use it for the console output.

Link to comment
28 minutes ago, ich777 said:

I‘ve gone through the build log and everything seema fine.

 

I‘ve now read through your first post, please enable the iGPU and set it to the primary graphics device so that Unraid can use it for the console output.

Not possible on my system (i only have one tv connected to the server, i can't keep switching hdmi cable everytime i have to run a VM. Unless it's for debugging purposes).
Beside, it has always worked in RC2 (and still is, i went back to it) with the iGPU disabled (?).

If needed for debugging i'll do it; i'll need to install RC3 again tho.

Thanks.

Edited by dhstsw
Link to comment
2 minutes ago, dhstsw said:

Not possible on my system (i only have one tv connected to the server, i can't keep switching hdmi cable everytime i have to run a VM.

But you don,t have to switch it, or why do you want to switch it?

You can also buy a HDMI Dummy plug for the iGPU.

 

The VM will start fine if you‘ve assigned the dGPU wirhout switching the HDMI cable.

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.