You are not logged in.
Hi,
I was suggested a while back to use the 550xx branch of driver as the current driver I think it could have been 570 or 580 version kept kernel panicking my machine.
01:00.0 VGA compatible controller: NVIDIA Corporation GA107GLM [RTX A2000 8GB Laptop GPU] (rev a1)
Subsystem: Lenovo Device 22f8
Kernel driver in use: nvidia
Kernel modules: nouveau, nvidia_drm, nvidiaI see that there's a 590 branch out.
Has anyone got any good results with this?
I found I needed to rebuild my machine every few days after kernel panic and the system would somehow send many library files situated in /lib to Lost+Found/
I'm trying to run Davinci Resolve Studio - downloaded from BlackMagick website and not AUR and so far managed to get it to the point where at least it will launch by downgrading OpenCL.
This isn't a Resolve posting so no need to move to AUR section but all I'm trying to find out is, will the newer 590 branch work with my system stably and does it have working CUDA?
Offline
That card appears to be from the ampere family and uses a GA 107 / NV 177 chipset .
is this a laptop ?
Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.
clean chroot building not flexible enough ?
Try clean chroot manager by graysky
Offline
That card appears to be from the ampere family and uses a GA 107 / NV 177 chipset .
is this a laptop ?
Yes! It's a Lenovo P1 Gen 5
so far I have installed the 590xx or "Open" branch and the system seems to be stable. DaVinci Resolve works again which is good news, I haven't tested with other things like Steam and my drone sims but for now at least I have a GUI at correct resolution.
Offline
There is a forum post from a user with a laptop with rtx 3070 (also ampere) where the nvidia firmware had to be disabled (only possible with proprietary driver) .
I haven't been able to find other problem reports.
Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.
clean chroot building not flexible enough ?
Try clean chroot manager by graysky
Offline
Hi, I came across your post about the RTX A2000 8GB on your ThinkPad P1 Gen 5 — I have the exact same machine and GPU (subsystem 17AA:22F8) but my GPU's VBIOS is corrupted (Code 43, BIOS Unknown in GPU-Z, 0MB VRAM reported). Since your card is working, would you be willing to dump your VBIOS and share it?
On Linux it's one command:
sudo nvflash --save my_vbios.rom
(Linux NVFlash binary: https://www.techpowerup.com/download/nvidia-nvflash/)
It would genuinely save my laptop. Happy to share more details if useful. Thanks for considering it!
Offline
Hi, I came across your post about the RTX A2000 8GB on your ThinkPad P1 Gen 5 — I have the exact same machine and GPU (subsystem 17AA:22F8) but my GPU's VBIOS is corrupted (Code 43, BIOS Unknown in GPU-Z, 0MB VRAM reported). Since your card is working, would you be willing to dump your VBIOS and share it?
On Linux it's one command:
sudo nvflash --save my_vbios.rom
(Linux NVFlash binary: https://www.techpowerup.com/download/nvidia-nvflash/)It would genuinely save my laptop. Happy to share more details if useful. Thanks for considering it!
Standby... I am currently trying to install nvflash once it's done I'll let you know ;-)
Offline
Hi JohnnySSH,
Thanks so much for getting back to me and for being willing to help — I really appreciate it. Just checking in to see how the install is going, and thought I'd share the exact steps in case it helps speed things up.
Machine details:
- Model: ThinkPad P1 Gen 5
- GPU: NVIDIA RTX A2000 8GB Laptop, Device ID 10DE 25BA, Subsystem 17AA 22F8
If you're on Linux (since your forum post mentioned the nvidia driver on what sounds like an Arch setup), there's a dedicated Linux build on the same TechPowerUp page:
1. Download nvflash_5_867_linux.zip from https://www.techpowerup.com/download/nvidia-nvflash/
2. Extract it — there will be folders for different architectures (x86, x64, aarch64, ppc64)
In Ubuntu Terminal, once that folder is on your Desktop
3. Use the x64 folder, since your P1 Gen 5 is x86-64:
cd ~/Desktop ----- Enter
find . -iname "nvflash*" ----- Enter
cd x64 ----- Enter
chmod +x nvflash ----- Enter
sudo ./nvflash –list ----- Enter
sudo ./nvflash --save current_vbios.rom ----- Enter
Either way, the file will save in the same folder you're working from — you can just zip it up and email it over, or share it however's easiest for you.
I just want to be upfront: I'm genuinely relying on this. I've already tried five different NVIDIA driver branches, every BIOS/Secure Boot setting Lenovo support suggested, and NVFlash itself on both Windows and a Linux live USB — every read attempt fails identically with "a system restart might be required," which points to the VBIOS chip not responding at all. A working dump from your machine is genuinely the most realistic path I have left to get my GPU running again, short of a full hardware chip-level repair. So whatever you're able to do, even if it takes a bit of time, would mean a lot.
Thanks again — happy to answer any questions if something doesn't work as expected.
Best,
John
Offline
Hi John,
I'm currently using yay to build from AUR. It's been going since yesterday. I set it to basically upgrade the few packages I have installed in addition.
Currently it's trying to build inav-configurator which I use for drones and has been running the "post package hook" since well daytime yesterday??
It might be that it takes a while. Been trying to upgrade LLVM19 on FreeBSD for a week now :-(
Some packages just take forever I guess ??
Offline
Hi John,
I'm currently using yay to build from AUR. It's been going since yesterday. I set it to basically upgrade the few packages I have installed in addition.
Currently it's trying to build inav-configurator which I use for drones and has been running the "post package hook" since well daytime yesterday??
It might be that it takes a while. Been trying to upgrade LLVM19 on FreeBSD for a week now :-(
Some packages just take forever I guess ??
Thank you Johnny,
I will be waiting for the GPU vbios when you are done.
Offline
JohnnySSH wrote:Hi John,
I'm currently using yay to build from AUR. It's been going since yesterday. I set it to basically upgrade the few packages I have installed in addition.
Currently it's trying to build inav-configurator which I use for drones and has been running the "post package hook" since well daytime yesterday??
It might be that it takes a while. Been trying to upgrade LLVM19 on FreeBSD for a week now :-(
Some packages just take forever I guess ??
Thank you Johnny,
I will be waiting for the GPU vbios when you are done.
Uh I stopped the build as I don't really want to make you wait for too long as I know how frustrating it is to not have a working machine.
Just running again with inav excluded. I'll build it later.... Cross fingers in a few hours it should all be ready ;-)
Apologies!!
Offline
Ok got it installed:
nvflash --save my_vbios.rom
NVIDIA Firmware Update Utility (Version 5.867.0)
Copyright (C) 1993-2024, NVIDIA Corporation. All rights reserved.
Device:00:01:00:00=10DE:25BA:17AA:22F8 GPU
ERROR: mmap(): /dev/mem[ Base addrres = 0xaf000000, size = 0x01000000]
Attempt to map physical memory failed.Offline
Ok got it installed:
nvflash --save my_vbios.rom NVIDIA Firmware Update Utility (Version 5.867.0) Copyright (C) 1993-2024, NVIDIA Corporation. All rights reserved. Device:00:01:00:00=10DE:25BA:17AA:22F8 GPU ERROR: mmap(): /dev/mem[ Base addrres = 0xaf000000, size = 0x01000000] Attempt to map physical memory failed.
That's a great sign — your GPU itself is fine, this is just a Linux permission issue blocking direct memory access (This is a permissions/access issue, not a hardware fault on your end — your GPU works fine, this is just /dev/mem access being blocked). A couple of things to try:
Run with sudo if you weren't already:
Code:
sudo nvflash --save my_vbios.rom
If still blocked, check if iomem=relaxed needs to be added as a kernel boot parameter (some distros restrict /dev/mem access by default via CONFIG_STRICT_DEVMEM). Try booting with this kernel parameter added (edit GRUB or your bootloader):
Code:
iomem=relaxed
Then retry the nvflash command.
Thank you so much
Last edited by johnchampion (2026-06-21 13:47:51)
Offline
Same issue sadly :-(
No sudo problems
It could be that I need to reboot as I updated the system hence kernel and GPU but that's a problem as I have a FreeBSD @Ports update going on and even after many weeks I'm still only on 148 of 824
If you can wait till at least llvm builds which I have no clue how long that's gona be then I can stop that, run it from a different machine and reboot this one?
PS. you were wondering about new driver stability and I can say that it's probably more stable then the old one. I don't get kernel panics any more which is a great sign!
Last edited by JohnnySSH (2026-06-21 13:53:15)
Offline
Same issue sadly :-(
No sudo problems
It could be that I need to reboot as I updated the system hence kernel and GPU but that's a problem as I have a FreeBSD @Ports update going on and even after many weeks I'm still only on 148 of 824
If you can wait till at least llvm builds which I have no clue how long that's gona be then I can stop that, run it from a different machine and reboot this one?
PS. you were wondering about new driver stability and I can say that it's probably more stable then the old one. I don't get kernel panics any more which is a great sign!
No problem at all, please don't interrupt your FreeBSD build for this — I know how painful a stalled @Ports update can be, especially with llvm in the queue. I'll wait, genuinely.
I just want to say thank you. You're the only lead I have that's actually gotten this far — every other avenue (Lenovo support, TechPowerUp's database, NVFlash on my own machine across Windows and Linux) has hit a dead end. The fact that you're willing to take the time to help a stranger fix a laptop means a lot.
Whenever it's convenient — next week,— just let me know and I'll be ready.
And good to hear about the driver stability improvement, that's genuinely useful to know. Glad the kernel panics are gone.
Thanks again, really.
Offline
Greetings Johnny,
Believing you are doing great.
Just a reminder.
Thank You.
John
Offline
Greetings Johnny,
Believing you are doing great.
Just a reminder.
Thank You.
John
Managed to reboot.... not intentionally lol my wifi disconnected so the @ports rebuild got interrupted grrrr :-(
Unfortunately I'm still seeing the same thing?
NVIDIA Firmware Update Utility (Version 5.867.0)
Copyright (C) 1993-2024, NVIDIA Corporation. All rights reserved.
Device:00:01:00:00=10DE:25BA:17AA:22F8 GPU
ERROR: mmap(): /dev/mem[ Base addrres = 0xaf000000, size = 0x01000000]
Attempt to map physical memory failed.Oh actually I found this:
https://www.reddit.com/r/linux_gaming/c … _gpu_bios/
To read the ROM, your graphics card driver must NOT be loaded at all (nvflash will refuse to run if it notices such a state). Basically, if you're at your desktop environment via any driver (NVIDIA, nouveau or nova), then you're screwed. You will have to reboot without an active driver.I'll try without GPU initialization and see if that works?
Offline
Ok... I found a way to grab it
Boot from usb key then arch-chroot and run command
How do you want me to share it with you?
Offline
Ok... I found a way to grab it
Boot from usb key then arch-chroot and run command
How do you want me to share it with you?
Greeting Johnny,
Kindly use my email: redacted
or let me know any other way that is more convinent for you.
Really Appriciated.
Moderator Note
I've removed the email address.
Both of you allow sending an email to you through the forum email system.
Use the email link under a username to send someone an email without making your email address publicly visible.
Last edited by Lone_Wolf (2026-07-03 08:30:16)
Offline