You are not logged in.
Pages: 1
Hello,
I just compiled a new custom kernel and the custom kernel headers and then went on to create a custom nvidia kernel module. Somehow that doesn't seem to work. Nvidia is complaining that there is something Xen-ish in my kernel:
Creating directory NVIDIA-Linux-x86_64-260.19.36-no-compat32
Verifying archive integrity... OK
Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86_64 260.19.36.....................................................................................................................
The kernel you are installing for is a Xen kernel!
The NVIDIA driver does not currently work on Xen kernels. If
you are using a stock distribution kernel, please install
a variant of this kernel without Xen support; if this is a
custom kernel, please install a standard Linux kernel. Then
try installing the NVIDIA kernel module again.
*** Failed Xen sanity check. Bailing out! ***
make[1]: *** [build-sanity-checks] Error 1
make: *** [module] Error 2
Aborting...
As far as I know there is no xen or other vm stuff in the kernel. To verify this I tried to make a new nvidia package for the standard kernel. Again the same error.
Also this makes me wonder how the PKGBUILD works since it the build function just runs:
make SYSSRC=/path/to/current/kernelmodules/build module
and there are no xen sanity checks in that script, that happens in nvidia-installer - which isn't called from the PKGBUILD....
Secondly, when doing a "make module" from the nvidia/kernel dir with the standard kernel there is a difference between the abs package and the extra package. Anyone a clue what is going on in the PKGBUILD and why the difference between the two kernel modules?
Thanks.
Last edited by Ulf (2011-02-04 22:20:51)
Offline
make SYSSRC=/path/to/current/kernelmodules/build module
^ This is what I use for my custom kernels, and then copy the resulting module and then run depmod -a. Basically, I'm doing what the PKGBUILD does, just manually. It's just one file, so I don't care that it isn't tracked by the package manager. And it works.
Last edited by Gusar (2011-02-04 18:43:33)
Offline
Yeah that seems to work but still wondering why the PKBUILD calls nvidia-installer and why there is a slight difference in size between the extra package and my own module for the same standard arch kernel and why nvidia thinks there is a kernel with Xen (it seems that it is some sort of common bug, not specific to arch but nvidia-ish).
Offline
Yeah that seems to work but still wondering why the PKBUILD calls nvidia-installer
To unpack it. From the PKGBUILD:
sh ${_pkg}.run --extract-only
Note the '--extract-only'
and why there is a slight difference in size between the extra package and my own module for the same standard arch kernel
Different cflags? Different compiler version?
and why nvidia thinks there is a kernel with Xen (it seems that it is some sort of common bug, not specific to arch but nvidia-ish).
Quite likely false detection. The installer sees something activated in the kernel and thinks it's xen related. Or something.
Offline
Hmmm compiler settings and cflags...of course
Anyway, building by hand seems to work and indeed, for one file using the old fashioned make && make install
Thanks folks.
Offline
Pages: 1