cat: rom: Input/output error
dazWiLLiE opened this issue · comments
I followed your tutorial, and got the "cat: rom: Input/output error" error. Installed the "VFIO-PCI Config" plugin, marked both nvidia devices, saved config, rebooted and still get this error..
Script location: /tmp/user.scripts/tmpScripts/dump vbios/script
Note that closing this window will abort the execution of this script
You have selected this device to dump the vbios from
08:00.0 VGA compatible controller: NVIDIA Corporation GK208B [GeForce GT 710] (rev a1)
This does look like a valid GPU to me. Continuing .........
Checking if location to put vbios file exists
Vbios folder already exists
I will try and dump the vbios without disconnecting and reconnecting the GPU
This normally only works if the GPU is NOT the Primary or the only GPU
I will check the vbios at the end. If it seems wrong I will then retry after disconnecting the GPU
Defining temp vm with gpu attached
Domain dumpvbios defined from /tmp/dumpvbios.xml
Starting the temp vm to allow dump
Domain dumpvbios started
Waiting for a few seconds .....
Stopping the temp vm
Domain dumpvbios destroyed
Removing the temp vm
Domain dumpvbios has been undefined
Okay dumping vbios file named gt 710.rom to the location /mnt/user/isos/vbios/
cat: rom: Input/output error
Um.... somethings gone wrong and I couldn't dump the vbios for some reason
Sometimes when this happens all we need to do to fix this is 'stub' or 'bind to the vfio' the gpu and reboot the server
This can be done in Unraid 6.8.3 with the use of the vfio config plugin or if you are on Unraid 6.9 or above it can be done
directly from the gui in Tools/System Devices .....So please do this and run the script again
Edit: I should mention that once I saw output on the screen connected to the card, it said something about not finding a file or something.
Edit2: I increased the sleep time in "startstopvm()" and could see the error messages:
Booting from Harddisk...
Boot failed: could not read the boot disk"
Booting from Floppy...
Boot failed: could not read the boot disk"
No bootable device
I am having the same problem FYI.
I had also received the cat: rom: Input/output error when dumping my RX550 vbios. I had to actually start a Windows VM that had the RX550 already passed through to it (no vbios specified). I then shut down that VM and ran the script. Which worked.
I had found the following in my syslog when the vbios dump script was run and failed.
kernel: pci 0000:01:00.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0xffff
Searching for that brought me to https://www.redhat.com/archives/vfio-users/2019-March/msg00004.html
His experience was exactly the same as mine.
I also am having the same problem. I tried what @0v3rcome mentioned (started a vm with GPU passed through, no Vbios, stopped it), and still see the same error unfortunately.
Adding this to the syslinux config (this can be easily accomplished through the Unraid UI), rebooting, then running the script worked for me:
vfio-pci.disable_idle_d3=1
You'll want to remove that and reboot once again after you've successfully dumped your vbios. Thanks @0v3rcome for the link to that thread.
I have the same error: On 6.10.0rc2 w/ nvidia 2080 super founders
Script location: /tmp/user.scripts/tmpScripts/dump-vbios/script
Note that closing this window will abort the execution of this script
You have selected this device to dump the vbios from
01:00.0 VGA compatible controller: NVIDIA Corporation TU104 [GeForce RTX 2080 SUPER] (rev a1)
This does look like a valid GPU to me. Continuing .........
Checking if location to put vbios file exists
Vbios folder already exists
I will try and dump the vbios without disconnecting and reconnecting the GPU
This normally only works if the GPU is NOT the Primary or the only GPU
I will check the vbios at the end. If it seems wrong I will then retry after disconnecting the GPU
Defining temp vm with gpu attached
Domain 'dumpvbios' defined from /tmp/dumpvbios.xml
Starting the temp vm to allow dump
Domain 'dumpvbios' started
Waiting for a few seconds .....
Stopping the temp vm
Domain 'dumpvbios' destroyed
Removing the temp vm
Domain 'dumpvbios' has been undefined
Okay dumping vbios file named rtx-2080-super.rom to the location /mnt/user/isos/vbios/
cat: rom: Input/output error
Um.... somethings gone wrong and I couldn't dump the vbios for some reason
Sometimes when this happens all we need to do to fix this is 'stub' or 'bind to the vfio' the gpu and reboot the server
This can be done in Unraid 6.8.3 with the use of the vfio config plugin or if you are on Unraid 6.9 or above it can be done
directly from the gui in Tools/System Devices .....So please do this and run the script again
I tried binding the card to vfio, appending the vfio-pci.disable_idle_d3=1 line above in main -> flash -> global configuration (with reboot), and tried setting forcereset from no to yes like in #7 (comment)
It always ends up with the same cat: rom: Input/output error. Any ideas?
I was running into this error a lot, I followed all of the suggestions and no dice. What actually worked for me was trying one of the vbios from the techpowerup db, editing the hex file. Trying it out. That didn't work but when I shut down the VM and tried again, I got a 76kb vbios which worked - sorta. It at least allowed me to get into Windows but after doing all of the driver installs for the graphics card, I shut down the VM and ran the script again after going into System Information and binding the graphics card. (MAKE SURE THE VM ISN'T SET TO AUTOSTART), reboot after binding and presto, worked fine and got my 980 vbios.
I am having the same issue with this device: Alder Lake-N [UHD Graphics]
Tried everything mentioned but end up either with:
❯ echo 1 > /sys/bus/pci/devices/0000:00:02.0/enable
echo: write error: device or resource busyOr:
❯ cat /sys/bus/pci/devices/0000:00:02.0/rom
cat: '/sys/bus/pci/devices/0000:00:02.0/rom': Input/output error
@bolzerrr I too got the same error but what I did is followed spaceinvaders tutorial to get the vbios without dumping
https://www.youtube.com/watch?v=1IP-h9IKof0&ab_channel=SpaceinvaderOne
I found that having device bound to any driver actually hinders the VBIOS dumping,
use following commands to unbind and then rebind to vfio-pci driver (where 0000:02:00.0 is device address):
echo '0000:02:00.0' > /sys/bus/pci/drivers/vfio-pci/unbind
echo '0000:02:00.0' > /sys/bus/pci/drivers/vfio-pci/bind
@0xE1 Thanks! I'll try it out later.