Latest keyworded drivers doesn't build with >=kernel-2.6.36. /var/tmp/portage/x11-drivers/nvidia-drivers-256.53/work/kernel/nv.c:426: error: unknown field 'ioctl' specified in initializer /var/tmp/portage/x11-drivers/nvidia-drivers-256.53/work/kernel/nv.c:426: warning: initialization from incompatible pointer type make[4]: *** [/var/tmp/portage/x11-drivers/nvidia-drivers-256.53/work/kernel/nv.o] Error 1 make[3]: *** [_module_/var/tmp/portage/x11-drivers/nvidia-drivers-256.53/work/kernel] Error 2 nvidia-drivers-260.19.12 seems to be the latest stable version from nvidia but its still hardmasked as "beta". Could we fix the mask?
*** Bug 342021 has been marked as a duplicate of this bug. ***
*** Bug 342275 has been marked as a duplicate of this bug. ***
is 260 still considered beta? Gentoo portage just released *.36 into ~amd64 so people on testing can't use the driver yet. I know Ubuntu is shipping with version 260, so I wouldn't think it's still in beta.
*** Bug 342651 has been marked as a duplicate of this bug. ***
(In reply to comment #3) > is 260 still considered beta? There still are serious issues with nvidia-drivers-260.19.*, see bug #336837
since new nvidia beta drivers xbmc stopped working.
x11-drivers/nvidia-drivers-260.19.12 is masked as being beta but it's marked stable on the nvidia website. Does it still ahve major issues?
As it is stated in bug #336837, comment 33 the version 256.53 of nvidia-drivers can be compiled against 2.6.36 (gentoo-sources-)kernel with a patch that was posted to NVidia forums by artem. So with little tinkering applied for nvidia-drivers-256.53.ebuild, I am able to compile those drivers. (I don't know how they work yet, though.) The question is, would this patch (and later potential stabilizing of 256.53) be an acceptable alternative for unmasking masked drivers? Ebuild was modified as per following, shown here only for reference. The .txt extension was removed from the file posted in the forums. I'll post the real deals (ebuild and the patch) if there's any point for doing that. --- /usr/portage/x11-drivers/nvidia-drivers/nvidia-drivers-256.53.ebuild 2010-08-31 18:57:49.000000000 +0300 +++ /usr/local/portage/x11-drivers/nvidia-drivers/nvidia-drivers-256.53.ebuild 2010-11-05 17:03:37.000000000 +0200 @@ -277,6 +277,10 @@ # If greater than 2.6.5 use M= instead of SUBDIR= convert_to_m "${NV_SRC}"/Makefile.kbuild + + if kernel_is ge 2 6 36; then + epatch "${FILESDIR}"/NVIDIA-Linux-x86-256.53-linux-2.6.36.patch + fi fi }
(In reply to comment #7) > x11-drivers/nvidia-drivers-260.19.12 is masked as being beta but it's marked > stable on the nvidia website. > > Does it still ahve major issues? > opengl stuff is buggy or at least does not work as expected. In my environment app-emulation/qemu-kvm-0.13.0-r2 stops with a coredump when trying to start a virtual machine. Program terminated with signal 11, Segmentation fault. #0 0x0000003ad109483f in ?? () from //usr/lib64/opengl/nvidia/lib/libGL.so.1 (gdb) bt #0 0x0000003ad109483f in ?? () from //usr/lib64/opengl/nvidia/lib/libGL.so.1 #1 0x0000003ad1094c39 in ?? () from //usr/lib64/opengl/nvidia/lib/libGL.so.1 #2 0x0000003ad109939b in ?? () from //usr/lib64/opengl/nvidia/lib/libGL.so.1 #3 0x0000003ad109994c in ?? () from //usr/lib64/opengl/nvidia/lib/libGL.so.1 #4 0x0000003ac8c06c3a in start_thread () from /lib/libpthread.so.0 #5 0x0000003ac80d1cad in clone () from /lib/libc.so.6
(In reply to comment #7) > x11-drivers/nvidia-drivers-260.19.12 is masked as being beta but it's marked > stable on the nvidia website. > > Does it still ahve major issues? > My laptop sony vaio vpccw2s1r with gf 330m does not start xorg with x11-drivers/nvidia-drivers-260.19.12 (just black screen)
New driver version 260.19.21 was just released - might be worth a try ?
*** Bug 345269 has been marked as a duplicate of this bug. ***
(In reply to comment #7) > x11-drivers/nvidia-drivers-260.19.12 is masked as being beta but it's marked > stable on the nvidia website. > > Does it still ahve major issues? > I get something like "... nvidia.ko Device not found" with 'modprobe nvidia' on my ~amd64 system.
Created attachment 254193 [details, diff] nv-256.53_lnx-2.6.36.patch I had issues with the 260.19.12 and with gvim while using 2.6.36 kernel. (Strange if tracing the gvim process it is working) I have the 256.53 to compile with 2.6.36 , I have back-ported the config options from the 2.6.36. Now gvim is working. I tested OpenArena as well.
Created attachment 254195 [details] nvidia-drivers-256.53.ebuild A working ebuild with the nv-256.53_lnx-2.6.36.patch.
I'm happy to report that applying the patch and new ebuild has allowed me to once again use KDE on my workstation. Thank you!
*** Bug 345707 has been marked as a duplicate of this bug. ***
Created attachment 254745 [details, diff] Patch for 96.43.18 driver The same issue with x11-drivers/nvidia-drivers-96.43.18. This is patch for 260* version transposed for 96.43.18 version. Tested on x86 and Linux 2.6.36-gentoo-r1.
Hi all, gentoo-sources-2.6.36-r5 is now stable (bug 348509) on x86. Should this bug be a blocker for 2.6.36 stabilization?
(In reply to comment #19) > Hi all, > > gentoo-sources-2.6.36-r5 is now stable (bug 348509) on x86. Should this bug > be a blocker for 2.6.36 stabilization? > 2.6.35 also has some severe problems. better fix this issue I guess... Attila, please correct the patch to have the correct name! (36 not 26)
This bug is related to http://bugs.gentoo.org/show_bug.cgi?id=348865 virtualbox-modules has a similar issue, see my patch above.
(In reply to comment #19) > Hi all, > > gentoo-sources-2.6.36-r5 is now stable (bug 348509) on x86. Should this bug > be a blocker for 2.6.36 stabilization? > As someone who just downloaded the stable 2.6.36-r5 and expected stable nvidia-drivers-195.36.31 to work, I would second the notion that something should be done *soon* or alternatively block 2.6.36 as stable. I'll try the patches supplied above for now.
As I've stated in comment #8 256.53 can be compiled against 3.6.36. Now I have 3.6.36-r5 kernel running smoothly with nvidia-drivers-256.53, but it is NOT using any patches attached here. Instead the patch (mentioned in my comment) that contains 3 very small changes is used. Additionally it seems, that some of 260.*-series drivers are not hardmasked anymore so the question is, do we want working older (256.53) drivers beside new cutting-edge drivers, or not?
I have gentoo-sources-2.6.36-r5 and nvidia-drivers-260.19.29 installed (directly from portage, without any patches). They're running without any problems!
(In reply to comment #24) > I have gentoo-sources-2.6.36-r5 and nvidia-drivers-260.19.29 installed > (directly from portage, without any patches). They're running without any > problems! > I agree, 260.19.29 is the only version of the 260 series that works fine for me.
(In reply to comment #24) I can second this. I have the same setup. I also tested the corresponding nvidia-setting-260.19.29 and it seems to be working as well. The only minor thing I have found is "nvidia: module license 'NVIDIA' taints kernel" in dmesg. BTW, I'm running a nVidia Corporation G96 [GeForce 9400 GT] (rev a1).
*** Bug 349303 has been marked as a duplicate of this bug. ***
*** Bug 349616 has been marked as a duplicate of this bug. ***
This should be solved by bug 348186.