535, 545, 550 fail to build with GCC14

[flukejones]

NVIDIA Open GPU Kernel Modules Version

535, 545, 550

Operating System and Version

Fedora rawhide (fedora 40 testing)

Kernel Release

6.6, 6.7

Please confirm you are running a stable release kernel (e.g. not a -rc). We do not accept bug reports for unreleased kernels.

• I am running on a stable kernel release.

Build Command

NVIDIA<version>.run

Terminal output/Build Log

conftest4394.c: In function 'conftest_drm_gem_object_vmap_has_map_arg':
conftest4394.c:24:46: error: passing argument 2 of 'obj->funcs->vmap' from incompatible pointer type [-Wincompatible-pointer-types]
   24 |                 return obj->funcs->vmap(obj, map);
      |                                              ^~~
      |                                              |
      |                                              struct dma_buf_map *
conftest4394.c:24:46: note: expected 'struct iosys_map *' but argument is of type 'struct dma_buf_map *'

More Info

The bug is in the conftest.sh

See https://forums.developer.nvidia.com/t/nvidia-modules-build-failure-with-upcoming-gcc-14-and-recent-kernels-due-to-misfiring-conftest-sh-test-heads-up/279072 for more info as it affect open plus closed driver

[aritger]

Thanks for the pointer to that forum post. I've filed NVIDIA internal bug 4478534 to track it. Hopefully we can resolve this soon.

[scaronni]

550.76 and 550.78 build fine on my Fedora systems with GCC 14.

[flukejones]

550.78 is failing with:

   /home/luke/Downloads/NVIDIA-Linux-x86_64-550.78/kernel-open/nvidia/os-mlock.c: In function 'nv_follow_pfn':
   /home/luke/Downloads/NVIDIA-Linux-x86_64-550.78/kernel-open/nvidia/os-mlock.c:42:12: error: implicit declaration of function 'follow_pfn'; did you mean 'follow_pte'? [-Wimplicit-function-declaration]
      42 |     return follow_pfn(vma, address, pfn);
         |            ^~~~~~~~~~
         |            follow_pte

kernel 6.9.0

[flukejones]

Related: https://lore.kernel.org/all/20240324234542.2038726-1-hch@lst.de/T/#mdeb04975c27076d6830f96c2041f50ef1d2554fc

follow_pfn is removed.

[aritger]

Thanks for the pointer on follow_pfn. We're aware of that and it will be fixed in a future release.

I also think that is unrelated to the originally reported problem here (interaction with -Werror=incompatible-pointer-types).

[flukejones]

Closing as fixed recently