331.20 on 3.13-rc1 kernel

They’ve changed back ACPI_HANDLE() to DEVICE_ACPI_HANDLE() in the final 3.13 release, the 2nd patch should no longer be used or it won’t compile.

Are you sure? I checked the commit logs and there is no evidence that they did that.

I may be wrong actually, but it doesn’t compile anymore with the patch I mentioned. They did some changes to ACPI_HANDLE:

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=82023bb7f75b0052f40d3e74169d191c3e4e6286

3.13 is now stable. So - how can I install nvidia on it? Is this possible or not possible today? I do need 3.13, because there are some usefull ssd disk support improvements.

Everything works, expect nvidia (as always, I’m really not suprised). Please help??

I used lwfinger’s patch from post #6 and it works on 3.13 stable on Fedora 20. (with --no-unified-memory)
Couldn’t just copy/paste from this forum though, looks like the formatting is bad.

I need to compile on Debian with 3.13 aptosid kernel. What is no-unified-memory? Also, will this patch work with nvidia 331.38?

I compiled 3.13 fine on ubuntu using xorg-edgers Nvidia 331.38-0ubuntu1~xedgers~saucy1
no patching needed.

lwfingers patch + gun26’s correction worked for me on Funtoo 3.13 64 bit, 331.38 nvidia driver. Note the formatting is slightly different for the patch on 331.38 (location in file). So I will, attributing all to lwfinger and gun26, paste it here again.

— /work/kernel2/nv-acpi.c 2014-01-08 18:49:24.000000000 -0800
+++ /work/kernel/nv-acpi.c 2014-01-22 16:08:38.794722023 -0800
@@ -303,7 +303,10 @@

 if (pNvAcpiObject->notify_handler_installed)
 {

+#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 13, 0)

  •   /* beginning with 3.13, acpi_remove_notify_handler() waits for events to finish */
       NV_ACPI_OS_WAIT_EVENTS_COMPLETE();
    

+#endif

     // remove event notifier
     status = acpi_remove_notify_handler(device->handle, ACPI_DEVICE_NOTIFY, nv_acpi_event);

This patch works for both nvidia.ko and uvm module. [url]http://cvs.rpmfusion.org/viewvc/rpms/nvidia-kmod/devel/nvidia_3.13_kernel.patch?revision=1.2&root=nonfree&view=markup[/url]

Tested on openSUSE 13.1 x64 with 3.13 kernel.

There is a newer patch

http://cvs.rpmfusion.org/viewvc/rpms/nvidia-kmod/devel/nvidia_3.13_kernel.patch?revision=1.3&root=nonfree&view=markup

Thanks Leigh.

I have applied the patches mentioned in

http://cvs.rpmfusion.org/viewvc/rpms/nvidia-kmod/devel/nvidia_3.13_kernel.patch?revision=1.3&root=nonfree&view=markup

However, I still have the samme issues mentioned earlier ("cannot copy vbios to system memory). I suspect I am doing something wrong in applying the patches to the driver interface. Can someone help me specifying the exact steps I need to take?

I did the following

  • extract the nvidia installer (331.38)
  • apply pathches to kernel/nv-acpi.c and kernel/uvm/nvidia_uvm_linux.h
  • run the ./nvidia-installer -n

I am at a loss what is going wrong. I am running pure 3.13 kernel compiled for debian wheezy. I thought it could be because of precompiled kernel interface but this cannot be the issue as long as you specify the -n option?

Also I wonder if anyone has any overview of where all the installed files from the archive go.

The patches in this thread are not enough for some setups:
[url][Solved] UX32VD: "Unable to copy VBIOS to system memory" - Linux - NVIDIA Developer Forums

GitHub issue thread:
[url]https://github.com/Bumblebee-Project/bbswitch/issues/80[/url]

I apparently had some extra issues because I was running EFI + no CSM