You are not logged in.
I have an issue with the discrete NVIDIA GPU (NVIDIA RTX A3000 Mobile) in my ThinkPad P15 Gen 2i laptop, which I discribed some time ago on this forum:
https://bbs.archlinux.org/viewtopic.php … 7#p2208797
Since then, nothing has changed. The solution is still to add the kernel parameter nvidia.NVreg_EnableGpuFirmware=0. But now the binary driver version has been removed from the repositories and I am stuck on driver version 580.
Could you advise on how to keep the driver version up to date in this situation? Perhaps there is still something that can be done to properly fix this issue.
I am also not sure whether NVIDIA is even aware of this problem. Maybe it would make sense to report it on an NVIDIA forum? I get the impression that this issue is often dismissed as an incorrect driver installation rather than a real bug.
Offline
I updated the BIOS using fwupd, and this completely solved all my issues.
It might be worth mentioning this in the original post as well (even though it’s quite old) because I’m sure this would have helped back then too
Offline
I’m sure this would have helped back then too
Did that BIOS version exist in 2024?
Also the nvidia-open drivers and nvidia provided firmware (GSP) have considerably improved since then.
Either way
I updated the BIOS using fwupd, and this completely solved all my issues.
You might want to record to which exact BIOS version and then please always remember to mark resolved threads by editing your initial posts subject - so others will know that there's no task left, but maybe a solution to find.
Thanks.
Online
I made a mistake, and in fact the problem persists. It's just that sometimes the video card can boot correctly after restarting with the nvidia-open595 driver, which misled me.
Can you tell me why there is no binary driver version 595 in AUR? I can't use nvidia-open, my card supports new driver versions, but only with the NVreg_EnableGpuFirmware flag=0. Neither the official repository nor the AUR has the latest version of the binary driver.
Here, just in case, I'll attach a successful launch report from nvidia-open595: https://paste.c-net.org/PriestsDrawers
And unsuccessful: https://paste.c-net.org/GooeyDeaver
Offline
Apr 07 00:46:51 green kernel: NVRM: loading NVIDIA UNIX Open Kernel Module for x86_64 595.58.03 Release Build (root@)
Apr 07 00:46:53 green kernel: NVRM: GPU0 nvAssertOkFailedNoLog: Assertion failed: Failure: Generic Error [NV_ERR_GENERIC] (0x0000FFFF) returned from RPC_HDR->rpc_result @ kernel_gsp.c:6068
Apr 07 00:46:53 green kernel: NVRM: GPU0 nvAssertOkFailedNoLog: Assertion failed: Failure: Generic Error [NV_ERR_GENERIC] (0x0000FFFF) returned from kgspWaitForRmInitDone(pGpu, pKernelGsp) @ kernel_gsp_tu102.c:569
Apr 07 00:46:53 green kernel: NVRM: GPU0 _kgspBootGspRm: unexpected WPR2 already up, cannot proceed with booting GSP
Apr 07 00:46:53 green kernel: NVRM: GPU0 _kgspBootGspRm: (the GPU is likely in a bad state and may need to be reset)
Apr 07 00:46:53 green kernel: NVRM: GPU0 RmInitAdapter: Cannot initialize GSP firmware RM
Apr 07 00:46:53 green kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x62:0x40:2168)
Apr 07 00:46:53 green kernel: NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0https://github.com/NVIDIA/open-gpu-kern … issues/965 - seems to relate to hybrid graphics systems… can you try to move the i915 drivers out of the initramfs? (kms hook or explicit MODULES item)
Can you tell me why there is no binary driver version 595 in AUR?
Maintainer decided to switch - the main reason is likely that moving to 590xx, GPUs not covered by the open driver dropped into legacy anyway and the latest nvidia GPU aren't supported by the binary driver any more.
https://archlinux.org/news/nvidia-590-d … l-modules/
Online
I added the i915 to MODULES in mkinitcpio.conf. Executed mkinitcpio -P and rebooted. Nothing has changed. I think I'll mark the topic as Solved and continue waiting for it to be fixed. Now I will continue to use the driver from AUR.
It might be worth reporting the problem to lenovo or nvidia. After all, this problem has been going on for over a year
Last edited by r0bo7 (2026-04-07 18:34:46)
Offline
No, the idea was to move i915 *out* of the initramfs. Might be there because of the kms hook, but nividia isn't in the initramfs, this seems to be a hybrid graphics problem and a race condition.
Online