Hello, In the Message; Subject : Re: For NVIDIA 5XX driver users Message-ID : <218c2236-f4df-44ea-90b2-8487d0998f42@gmail.com> Date & Time: Tue, 8 Oct 2024 23:14:34 -0500 [DCR] == "David C. Rankin" <drankinatty@gmail.com> has written: DCR> On 10/8/24 4:17 PM, Masaru Nomiya wrote: MN> > I strongly recommend that you upgrade to kernel 6.11.2 while leaving MN> > 6.11.0 in place. MN> > I think that kernel 6.11.2 is Linus's final ultimatum for nvidia MN> > users. MN> > There is no way out now..... DCR> Ohh, do say more... DCR> I though Linus' middle-finger to Nvidia was now about DCR> development of the open-driver and wouldn't impact the legacy DCR> binary-blob drivers. Is there a new development in this long DCR> standing kerfuffle? I'm quite pessimistic...... As you also know, Linus said that NVIDIA had improved compared to 10 years ago, but the problem started with the development of kernel 6.7. However, as I reported here, I noticed the problem starting with kernel 6.10. The problem felt strange, as if something was running in the background, but when I looked at it using lxtask or Evisum (a CPU monitor provided by enlightenment), nothing was running. However, when I looked at it using the free and top commands, the memory usage only went to increase, and when I did a slightly heavy task, I was hit by a phenomenon similar to a system freeze. Furthermore, when I executed a reboot or shutdown, it only changed to a black screen. When I changed the driver to 550.120, there was no change in other phenomena, but the feeling that something was running in the background eased, so I changed the driver to 550.120 and used it with caution. A couple of days before upgrading to kernel 6.11.0, I was worried about nvidia-drm and looked into it, and found that it wasn't loaded, so I tried using kernel parameters and /etc/modprobe.d to load nvidia-drm, but it didn't work. So I looked into it further, and found that just before entering the login screen, modprobe@.service, which is included in systemd 256.6, was unloading nvidia-drm. So, after logging in, I loaded nvidia-drm using the insmod command, even though I thought it might not work, and bingo! I thought it was much better than before, but when I upgraded to kernel 6.11.0, the 550.120 driver was truncated (the nvidia-drm workaround could no longer be used). So, when I upgraded the driver to 560.35.03, I found that the nvidia-drm workaround could be used. I was relieved. However, when I updated to kernel 6.11.2 this morning, I found that I couldn't use nvidia-drm workaround, so I stayed with 6.11.0. But I'm really wondering what to do from now on... Cf. 1. Andrei's prediction In the Message; Subject : Re: Again: error with kernel 6.8.1-1 when compiling nVidia Message-ID : <940aef4e-eeb8-4cc0-955f-89bb1f8bab89@gmail.com> Date & Time: Sat, 23 Mar 2024 21:13:49 +0300 [AB] == Andrei Borzenkov <arvidjaar@gmail.com> has written: AB> On 23.03.2024 19:07, Peter Maffter wrote: PM> > I expected that someone can explain to me, why they simply dropped out PM> > DRM_UNLOCKED AB> Author: Thomas Zimmermann <tzimmermann@suse.de> AB> Date: Wed Nov 22 13:09:40 2023 +0100 AB> drm: Remove locking for legacy ioctls and DRM_UNLOCKED AB> Modern DRM drivers acquire ioctl locks by themselves. Legacy ioctls AB> for user-space mode setting used to acquire drm_global_mutex. After AB> removing the ioctl entry points, also remove the locking code. The only AB> legacy ioctl without global locking was VBLANK_WAIT, which has been AB> removed as well. Hence remove the related DRM_UNLOCKED flag. AB> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> AB> Reviewed-by: David Airlie <airlied@gmail.com> AB> Reviewed-by: Daniel Vetter <daniel@ffwll.ch> AB> Acked-by: Alex Deucher <alexander.deucher@amd.com> AB> Link: AB> https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-12-tzimme... 2. This problem report. https://www.gamingonlinux.com/2024/08/nvidia-driver-with-linux-kernel-6-10-c... NVIDIA is unable to do anything for long time..... 3. There are many cries of anguish about this issue. https://forums.developer.nvidia.com/t/560-release-feedback-discussion/300830 Best Regards. --- ┏━━┓彡 Masaru Nomiya mail-to: nomiya @ lake.dti.ne.jp ┃＼／彡 ┗━━┛ "During testing, Sakana found that its system began unexpectedly attempting to modify its own experiment code to extend the time it had to work on a problem." -- Research AI model unexpectedly attempts to modify its own code to extend runtime (ars TECHNICA) --