TW - note: gcc 14 break Nvidia driver build - still working on patch for 390/470
TW users, Just a note, gcc 14 will break the nvidia driver module build. The discussion and initial workaround/patch is discussed in https://forums.developer.nvidia.com/t/nvidia-modules-build-failure-with-upco... I've not got a patch for the 470 or 390 drivers worked out yet -- if you do, post a link, but I think we can port the fixes in the above discussion to those versions as well. The CFLAGS default change also impacts the build. If I work out a patch or get one from the Arch list, I'll drop a link here. -- David C. Rankin, J.D.,P.E.
On Thursday 09 May 2024, David C. Rankin wrote:
TW users,
Just a note, gcc 14 will break the nvidia driver module build. The discussion and initial workaround/patch is discussed in https://forums.developer.nvidia.com/t/nvidia-modules-build-failure-with- upcoming-gcc-14-and-recent-kernels-due-to-misfiring-conftest-sh-test-head s-up/279072
I've not got a patch for the 470 or 390 drivers worked out yet -- if you do, post a link, but I think we can port the fixes in the above discussion to those versions as well. The CFLAGS default change also impacts the build.
If I work out a patch or get one from the Arch list, I'll drop a link here.
Thanks for the warning. If I don't update the kernel and the Nvidia driver (by zypper lock), will it be OK to allow the dup to gcc 14 to proceed? (I normally have both of these locked and only dup including them when I have time to deal with any fallout.) Michael
On 5/8/24 21:33, Michael Hamilton wrote:
Thanks for the warning.
If I don't update the kernel and the Nvidia driver (by zypper lock), will it be OK to allow the dup to gcc 14 to proceed?
(I normally have both of these locked and only dup including them when I have time to deal with any fallout.)
Michael
This is one I'd just not update for a week or two while monitoring the list and whatever links you can find on SUSE's gcc-14 implementation. When you can confirm a patch, then update at that point. Of course if there is another xz issue that pops up between now and then -- you likely will have to update at that point if there are any security implications for how you use your box. Looks like a patch to conftest.sh has been done for the 390.xx driver. I think there will be more needed, but I could be wrong. I'll test and report back. The link is this: https://bbs.archlinux.org/viewtopic.php?pid=2170078#p2170078 -- David C. Rankin, J.D.,P.E.
On 5/8/24 21:33, Michael Hamilton wrote:
Thanks for the warning.
If I don't update the kernel and the Nvidia driver (by zypper lock), will it be OK to allow the dup to gcc 14 to proceed?
(I normally have both of these locked and only dup including them when I have time to deal with any fallout.)
Michael
Sorry for using the same post, since I get the list via gmail, I don't see my replies. The patch WORKS for the 390.xx driver and the 6.8.9 kernel with gcc-14. I used: $ cat gcc-14.patch diff --git a/kernel/conftest.sh b/kernel/conftest.sh index 093704c..a9db899 100755 --- a/kernel/conftest.sh +++ b/kernel/conftest.sh @@ -153,7 +153,8 @@ test_headers() { build_cflags() { BASE_CFLAGS="-O2 -D__KERNEL__ \ -DKBUILD_BASENAME=\"#conftest$$\" -DKBUILD_MODNAME=\"#conftest$$\" \ --nostdinc -isystem $ISYSTEM" +-nostdinc -isystem $ISYSTEM \ +-Wno-implicit-function-declaration -Wno-strict-prototypes -Wno-incompatible-pointer-types" if [ "$OUTPUT" != "$SOURCES" ]; then OUTPUT_CFLAGS="-I$OUTPUT/include2 -I$OUTPUT/include" On Arch I built the kernel modules and opencl (the nvidia-utils-390xx package) for both the 6.8.9 kernel and the LTS kernel 6.6.30 and had no issues whatsoever. Original patch link again: https://bbs.archlinux.org/viewtopic.php?pid=2170078#p2170078 Hope it helps! -- David C. Rankin, J.D.,P.E.
participants (2)
-
David C. Rankin -
Michael Hamilton