Arch Linux User Repository

@anonfunc: No reason, just a leftover of a copy-paste action from linux-lts. :-) Fixed. Note, that on pentium4/i686/i486 there is no documentation built currently (that's the sed-fu at the end of the PKGBUILD when CARCH is 32-bittish).

Is there a specific reason to limit python-sphinx to <6.0.0 ? I just build the package successfully with python-sphinx 6.1.3 with no apparent problems.

/lib/modules/5.15.98-1-lts515/build/ exists. I get:

make
make V=1 CONFIG_MODULE_SIG= CONFIG_MODULE_SIG_ALL= -C /lib/modules/5.15.98-1-lts515/build M=/usr/src/vboxhost-7.0.6_OSE/vboxdrv SRCROOT=/usr/src/vboxhost-7.0.6_OSE/vboxdrv -j2 modules
make -f ./scripts/Makefile.build obj=/usr/src/vboxhost-7.0.6_OSE/vboxdrv \
single-build= \
need-builtin=1 need-modorder=1
make[2]: *** No rule to make target '/usr/src/vboxhost-7.0.6_OSE/vboxdrv/linux/SUPDrv-linux.o', needed by '/usr/src/vboxhost-7.0.6_OSE/vboxdrv/vboxdrv.o'.  Stop.
make[1]: *** [Makefile:1903: /usr/src/vboxhost-7.0.6_OSE/vboxdrv] Error 2
make: *** [/usr/src/vboxhost-7.0.6_OSE/vboxdrv/Makefile-footer.gmk:133: vboxdrv] Error 2

CONFIG_DRM_SIMPLEDRM=y and CONFIG_DEBUG_INFO_BTF=y are in all config files. The problem with dkms seems to be something deeper with the mechanism..

I think they are related. And I was wondering if it has anything to do with any flag that must be enabled at compile time. I found this piece of information that talks about kernel parameters to install nvidia-dkms https://wiki.archlinux.org/title/NVIDIA#Custom_kernel

Yeah, I noticed that too, I tried do build virtualbox-dkms, and it failed. Not sure, if that's related to your Nvidia issue though..

Hi! I can't install nvidia-dkms using this kernel. I am getting the error: Error! Bad return status for module build on kernel: 5.15.96-2-lts515 (x86_64) The error is only on this kernal and happens also for nvidia-vulkan-dkms. Any idea?

I'm having the same problem with the broken dependency being reported by yay. However this is probably not a typo: It looks like the _get_gcc_version_from_config function isn't run properly when generating the .SRCINFO file. It works when invoking makepkg directly, since it doesn't read the broken file.

This dependency problem looks like a typo:

$ yay -S linux-lts515 linux-lts515-headers
 -> Could not find all required packages:
    gcc-libs= (Wanted by: linux-lts515-headers)

Is this accurate?

$ yay -Sii linux-lts515-headers | awk '/Depends On/'
Depends On                    : pahole  gcc-libs=

Hmmm...now the build is failing (original release built fine):

==> Finished making: linux-lts515 5.15.95-2 (Tue Feb 28 15:53:52 2023)
==> Cleaning up...
loading packages...
resolving dependencies...
warning: cannot resolve "gcc-libs=", a dependency of "linux-lts515-headers"
:: The following package cannot be upgraded due to unresolvable dependencies:
      linux-lts515-headers

:: Do you want to skip the above package for this upgrade? [y/N] 
error: failed to prepare transaction (could not satisfy dependencies)
:: unable to satisfy dependency 'gcc-libs=' required by linux-lts515-headers