Arch Linux User Repository

Please don't flag this package out-of-date unless a new version has been released by NVIDIA.

The DKMS package guidelines are explicit that linux-headers should not be a dependency of any DKMS package.

As a concrete example of why including that as a hard dependency is a bad idea, what happens when linux is not an installed kernel?

PACKAGE NEEDS LONG TERM MAINTAINER

I have adopted the 390xx packages to keep them secure. I do not run any 390xx hardware so will not notice any breakages and cannot test any changes.

Until such time as someone else steps up to maintain these packages - ideally someone who actually has 390xx hardware - I have to rely on you to tell me what changes are needed.

Don't expect a response if you post only "this doesn't work", and do not email me to complain about the package not working!

Use this forum thread for discussion: https://bbs.archlinux.org/viewtopic.php?pid=1946926

A binary package is also available in my kernel-lts unofficial user repository.

I hadn't read this thread, so I fixed it myself, with this patch:

--- a/kernel/Kbuild
+++ b/kernel/Kbuild
@@ -63,7 +63,7 @@

 EXTRA_CFLAGS += -I$(src)/common/inc
 EXTRA_CFLAGS += -I$(src)
-EXTRA_CFLAGS += -Wall -MD $(DEFINES) $(INCLUDES) -Wsign-compare -Wno-cast-qual -Wno-error
+EXTRA_CFLAGS += -Wall -MD $(DEFINES) $(INCLUDES) -Wsign-compare -Wno-cast-qual -Wno-error -fpermissive
 EXTRA_CFLAGS += -D__KERNEL__ -DMODULE -DNVRM -DNV_VERSION_STRING=\"390.157\" -Wno-unused-function -Wuninitialized -fno-strict-aliasing -mno-red-zone -mcmodel=kernel -DNV_UVM_ENABLE -Wno-sign-compare -Wno-format-extra-args
 EXTRA_CFLAGS += $(call cc-option,-Werror=undef,)
 EXTRA_CFLAGS += -DNV_SPECTRE_V2=$(NV_SPECTRE_V2)
@@ -101,7 +101,9 @@
 NV_CONFTEST_CMD := /bin/sh $(NV_CONFTEST_SCRIPT) \
  "$(CC)" "$(HOST_CC)" $(ARCH) $(NV_KERNEL_SOURCES) $(NV_KERNEL_OUTPUT)

-NV_CONFTEST_CFLAGS := $(shell $(NV_CONFTEST_CMD) build_cflags)
+NV_CFLAGS_FROM_CONFTEST := $(shell $(NV_CONFTEST_CMD) build_cflags)
+
+NV_CONFTEST_CFLAGS = $(NV_CFLAGS_FROM_CONFTEST) $(EXTRA_CFLAGS) -fno-pie

 NV_CONFTEST_COMPILE_TEST_HEADERS := $(obj)/conftest/macros.h
 NV_CONFTEST_COMPILE_TEST_HEADERS += $(obj)/conftest/functions.h

I think this is more proper, GCC itself recommends using -fpermissive if there is need for the legacy behavior: https://gcc.gnu.org/gcc-14/porting_to.html

The change about NV_CONFTEST_CFLAGS I copied from the nvidia-470xx-utils file. They are mostly identical, but this one were missing that extra bit that forwards the extra flags to the conftest script.

Thank you @aruncveli - that patch does work. I thought it would take much more than that? But I just built the package with the patch (applied - patch -Np2 -i ../../gcc-14.patch) and both LTS and the normal kernel modules built fine. This patch works.

For the build issue on 6.8.9-arch1-2, this patch worked for me.

If someone else can confirm too, maybe it can be added to this repo.

Thanks.

WOW - 6.8.9-Arch1 was fine, 6.8.9-Arch2 broke the build (I suspect this is not the kernel, but a related package - likely util-linux).

Error! Bad return status for module build on kernel: 6.8.9-arch1-2 (x86_64) Consult /var/lib/dkms/nvidia/390.157/build/make.log for more information. ==> WARNING: `dkms install --no-depmod nvidia/390.157 -k 6.8.9-arch1-2' exited 10

Details are "error: #error wait_on_bit_lock() conftest failed!", "error: initialization of ‘void *’ from ‘int’ makes pointer from integer without a cast", and "error: implicit declaration of function ‘ioremap_nocache’;"

I'll start snooping, anybody else have a patch, please post it.

It's GCC. Not sure what changed, but Doug Newgard ID'd gcc as the culprit.

Looks like we will have to cobble a patch from the info in 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 unless somebody has a link to an already create patch for the 470 driver we can draw from.

Arch linux-6.8.1 was released today and I can report a seamless update and kernel module rebuild via dkms. The patch works flawlessly. Thanks to all!

@vnctdj - Thanks for kernel-6.8.patch, just in time for the 6.8.0 linux and -zen kernels released today, 11 March! All working correctly here.

Cheers

Thank you @vnctdj! I'll give it a shot as soon as the 6.8 kernel comes out. If all future kernel updates go as well as the past 2 -- my stress level over kernel updates for this driver can drop a notch or two. Thank you again for the effort.

(NOTE: patch and rebuild works fine. Just add the patch under the kernel-6.6.patch add the b2sum to the list (don't forget to bump the pkgrel) and makepkg -Cs)

It works for 6.8.0-rc7-1-mainline. Thanks for your effort.

I've just added a patch for the 6.8 kernel inspired by the following patch by Joan Bruguera: https://gist.github.com/joanbm/24f4d4f4ec69f0c37038a6cc9d132b43

I hope it will works when 6.8 will release!

It's that time again...

Torvalds announced no need for an rc-8, so the 6.8 kernel will be released next week and will show up on Arch shortly thereafter. Anybody have any idea on what, if any, patch will be needed for 6.8?