Le 20/09/2010 19:25, Curt Howland wrote: > Hi. Up to date Squeeze, compiling the latest 2.6.36-rc4 kernel. > > Last time the problem was compiling the kernel at all, which is > working just fine now thank you Debian-User. > > fakeroot make-kpkg --append-to-version -curt1.0 --initrd kernel_image > > creates the .deb file just fine, it installs fine, but when I run it > and try to get the VirtualBox and Nvidia drivers to compile, this is > what I get: > > from vbox-installer.log > > ========== > Makefile:170: *** Error: /usr/src/linux (version 2.6.36-rc4) does not > match the current kernel (version 2.6.36-rc4-curt1.0). Stop. > ========== > > The thing is, there is no such error with 2.6.35.4-curt1.0, nor when > it was 35-rc4-curt1.0. > > So why would it be choking on the custom extension now? I used > _exactly_ the same compile command, listed above. > > > from /var/log/nvidia-installer.log > > =========== > In file included from include/linux/kernel.h:17, > from include/linux/sched.h:54, > from include/linux/utsname.h:35, > from > /tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel/nv-linu > x.h:27, > from > /tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel/nv.c:13 > : > include/linux/bitops.h: In function ‘hweight_long’: > include/linux/bitops.h:49: warning: signed and unsigned type in conditional > expression > In file included from > /usr/src/linux-2.6.36-rc4/arch/x86/include/asm/uaccess > .h:571, > from include/linux/poll.h:14, > from > /tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel/nv-linu > x.h:84, > from > /tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel/nv.c:13 > : > /usr/src/linux-2.6.36-rc4/arch/x86/include/asm/uaccess_32.h: In function ‘ > copy_from_user’: > /usr/src/linux-2.6.36-rc4/arch/x86/include/asm/uaccess_32.h:209: warning: > co > mparison between signed and unsigned integer expressions > /tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel/nv.c: At top level: > /tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel/nv.c:426: error: unknown > fiel > d ‘ioctl’ specified in initializer > /tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel/nv.c:426: warning: > initializa > tion from incompatible pointer type > make[4]: *** [/tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel/nv.o] Error 1 > make[3]: *** [_module_/tmp/selfgz2500/NVIDIA-Linux-x86-256.53/kernel] Error > 2 > make[2]: *** [sub-make] Error 2 > NVIDIA: left KBUILD. > nvidia.ko failed to build! > make[1]: *** [module] Error 1 > make: *** [module] Error 2 > -> Error. > ERROR: Unable to build the NVIDIA kernel module. > ========== > > Which I guess means I have to wait for Nvidia to fix a signed vs > unsigned change in the kernel source. > > Can folks suggest if it would be worth while reporting this to Nvidia? > > Curt- > >
Hi, I build my kernels with kernel-package, currently a "2.6.35.4-vanilla64" and have no problem. I build the headers at the same time because some installers will look for it even if the full sources are available (NVidia .run was one of those in my late experience). I now use nvidia (now dkms-ready) from debian Sid and have no problem, VirtualBox from Oracle/Sun repo is doing fine too here. I always check that the links "build" and "source" in /lib/modules/$(uname -r) are properly set up. Hope it helps. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c97b945.3040...@googlemail.com
