3D Vision bug in 530.30.02

Hi,

Is this the place for bug reports? (The NVidia customer service did point there.)

Linux system with Quadro 4000 (or stronger) is able to produce OpenGL Quad-Buffered active stereo and with X11 option “Stereo” set to value “10” sends the syncronization signal to 3D Vision active shutter glasses via 3D Vision IR emitter. When X11 starts and NVidia’s driver is loaded, X11 logs show that stereo mode is requested, that monitor that supports 3D Vision, and that 3D Vision IR emitter is detected.

Up to NVidia driver version 525.85.12 the emitter does send the IR signal to the glasses when an application renders 3D stereo content.

With NVidia driver version 530.30.02 the emitter does not send the IR signal. This is probably a regression in the driver.

The operating system is AlmaLinux 9 (i.e. essentially RHEL 9) with MATE desktop environment from EPEL repository.
The card is Quadro RTX A4000. The drivers are from the NVidia yum repository.

The current workaround (and proof that version driver is the source of issue) was to downgrade back to the 525 series of driver.

nvidia-bug-report.log.gz (344.7 KB)
Finally got around to generate the bug report.

Haven’t figured out how to supply the ‘-logverbose’ to the X11 yet – systemd something, not startx.

Was there any resolution to this issue? I’m running into the same problem under Ubuntu Linux 24.04 with a Quadro P4000 graphics card. If I install the 535 version of the NVIDIA driver, the 3D stereo emitter does not activate. If I revert to the 470 version of the NVIDIA driver, then the emitter activates without any problem.

Hi,

I forgot how I got into the forum and practically never read my gmail.

Anyway, yes, after another bug report via email I did got reply from NVidia. They did spot the issue and a fix should be in proprietary driver version 570.

to 29. elok. 2024 klo 17.17 stevewt67ok via NVIDIA Developer Forums (notifications@nvidia.discoursemail.com) kirjoitti:

Thanks for responding! It’s good to hear that there will be a fix in driver version 570.

Seems still not fixed yet, even with version 580. Have no idea how this could happen that NVIDIA has 3D vision in the driver but does not function well. My applications did well with version 418 and before but not running good since version 470 (not tested the middle ones from 418 to 470, probably will do testing, since I have no choice). All versions 470-525 have a same problem with my application and seems 530 fixed it but stop sending signal of emitter as you know. I hope I could keep version 418 for ever but the most recent application I am using is updating to newer Linux kernels that does not support version 418 anymore. I really want to borrow the words from Linus Torvalds, but still, there are some good guys that keep the 3D vision function in the driver secretly under-knees. Still cannot understand why NVIDIA would stop such a good product of 3D vision. It is amazing especially to share with people together on the same screen. And also I could check something else easily, like reading the printed instructions. Can you do that with AR/VR? That will be a terrible idea for sure, even Apple VR.