Hello,
With te new kernel 3.2.5-3.fc16.x86_64 I cannot start the graphics mode!
X.Org X Server 1.11.4 Release Date: 2012-01-27 [ 161.446] X Protocol Version 11, Revision 0 [ 161.446] Build Operating System: x86-01 2.6.32-220.4.1.el6.x86_64 [ 161.446] Current Operating System: Linux Homere 3.2.5-3.fc16.x86_64 #1 SMP Thu Feb 9 01:24:38 UTC 2012 x86_64 [ 161.446] Kernel command line: BOOT_IMAGE=/boot/vmlinuz-3.2.5-3.fc16.x86_64 root=UUID=e0d44568-5e7b-4c91-921f-ab1130406253 ro rd.md=0 rd.lvm=0 rd.dm=0 KEYTABLE=us SYSFONT=latarcyrheb-sun16 rd.luks=0 LANG=en_US.UTF-8 [ 161.446] Build Date: 08 February 2012 10:54:57PM [ 161.446] Build ID: xorg-x11-server 1.11.4-1.fc16 [ 161.447] Current version of pixman: 0.22.2 [ 161.447] Before reporting problems, check http://wiki.x.org to make sure that you have the latest version. [ 161.447] Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. [ 161.448] (==) Log file: "/var/log/Xorg.0.log", Time: Wed Feb 15 12:13:46 2012 [ 161.449] (==) Using config directory: "/etc/X11/xorg.conf.d" [ 161.449] (==) Using system config directory "/usr/share/X11/xorg.conf.d" [ 161.449] (==) No Layout section. Using the first Screen section. [ 161.449] (**) |-->Screen "Default Screen" (0) [ 161.449] (**) | |-->Monitor "VGA-1" [ 161.449] (==) Automatically adding devices [ 161.449] (==) Automatically enabling devices [ 161.449] (==) FontPath set to: catalogue:/etc/X11/fontpath.d, built-ins [ 161.449] (**) ModulePath set to "/usr/lib64/xorg/modules/extensions/nvidia,/usr/lib64/xorg/modules" [ 161.449] (II) The server relies on udev to provide the list of input devices. If no devices become available, reconfigure udev or disable AutoAddDevices. [ 161.449] (II) Loader magic: 0x7caac0 [ 161.449] (II) Module ABI versions: [ 161.449] X.Org ANSI C Emulation: 0.4 [ 161.449] X.Org Video Driver: 11.0 [ 161.449] X.Org XInput driver : 13.0 [ 161.449] X.Org Server Extension : 6.0 [ 161.450] (--) PCI:*(0:1:0:0) 10de:0a65:0000:0000 rev 162, Mem @ 0xfd000000/16777216, 0xd0000000/268435456, 0xce000000/33554432, I/O @ 0x0000d800/128, BIOS @ 0x????????/524288 [ 161.450] (II) LoadModule: "extmod" [ 161.450] (II) Loading /usr/lib64/xorg/modules/extensions/libextmod.so [ 161.451] (II) Module extmod: vendor="X.Org Foundation" [ 161.451] compiled for 1.11.4, module version = 1.0.0 [ 161.451] Module class: X.Org Server Extension [ 161.451] ABI class: X.Org Server Extension, version 6.0 [ 161.451] (II) Loading extension SELinux [ 161.451] (II) Loading extension MIT-SCREEN-SAVER [ 161.451] (II) Loading extension XFree86-VidModeExtension [ 161.451] (II) Loading extension XFree86-DGA [ 161.451] (II) Loading extension DPMS [ 161.451] (II) Loading extension XVideo [ 161.451] (II) Loading extension XVideo-MotionCompensation [ 161.451] (II) Loading extension X-Resource [ 161.451] (II) LoadModule: "dbe" [ 161.451] (II) Loading /usr/lib64/xorg/modules/extensions/libdbe.so [ 161.451] (II) Module dbe: vendor="X.Org Foundation" [ 161.451] compiled for 1.11.4, module version = 1.0.0 [ 161.451] Module class: X.Org Server Extension [ 161.451] ABI class: X.Org Server Extension, version 6.0 [ 161.451] (II) Loading extension DOUBLE-BUFFER [ 161.451] (II) LoadModule: "glx" [ 161.451] (II) Loading /usr/lib64/xorg/modules/extensions/nvidia/libglx.so [ 161.454] (II) Module glx: vendor="NVIDIA Corporation" [ 161.454] compiled for 4.0.2, module version = 1.0.0 [ 161.454] Module class: X.Org Server Extension [ 161.454] (II) NVIDIA GLX Module 290.10 Wed Nov 16 18:01:24 PST 2011 [ 161.454] (II) Loading extension GLX [ 161.454] (II) LoadModule: "record" [ 161.455] (II) Loading /usr/lib64/xorg/modules/extensions/librecord.so [ 161.455] (II) Module record: vendor="X.Org Foundation" [ 161.455] compiled for 1.11.4, module version = 1.13.0 [ 161.455] Module class: X.Org Server Extension [ 161.455] ABI class: X.Org Server Extension, version 6.0 [ 161.455] (II) Loading extension RECORD [ 161.455] (II) LoadModule: "dri" [ 161.455] (II) Loading /usr/lib64/xorg/modules/extensions/libdri.so [ 161.455] (II) Module dri: vendor="X.Org Foundation" [ 161.455] compiled for 1.11.4, module version = 1.0.0 [ 161.455] ABI class: X.Org Server Extension, version 6.0 [ 161.455] (II) Loading extension XFree86-DRI [ 161.455] (II) LoadModule: "dri2" [ 161.455] (II) Loading /usr/lib64/xorg/modules/extensions/libdri2.so [ 161.455] (II) Module dri2: vendor="X.Org Foundation" [ 161.456] compiled for 1.11.4, module version = 1.2.0 [ 161.456] ABI class: X.Org Server Extension, version 6.0 [ 161.456] (II) Loading extension DRI2 [ 161.456] (==) Matched nouveau as autoconfigured driver 0 [ 161.456] (==) Matched vesa as autoconfigured driver 1 [ 161.456] (==) Matched fbdev as autoconfigured driver 2 [ 161.456] (==) Assigned the driver to the xf86ConfigLayout [ 161.456] (II) LoadModule: "nouveau" [ 161.456] (II) Loading /usr/lib64/xorg/modules/drivers/nouveau_drv.so [ 161.456] (II) Module nouveau: vendor="X.Org Foundation" [ 161.456] compiled for 1.10.99.902, module version = 0.0.16 [ 161.456] Module class: X.Org Video Driver [ 161.456] ABI class: X.Org Video Driver, version 11.0 [ 161.456] (II) LoadModule: "vesa" [ 161.456] (II) Loading /usr/lib64/xorg/modules/drivers/vesa_drv.so [ 161.456] (II) Module vesa: vendor="X.Org Foundation" [ 161.456] compiled for 1.10.99.902, module version = 2.3.0 [ 161.456] Module class: X.Org Video Driver [ 161.456] ABI class: X.Org Video Driver, version 11.0 [ 161.456] (II) LoadModule: "fbdev" [ 161.456] (II) Loading /usr/lib64/xorg/modules/drivers/fbdev_drv.so [ 161.456] (II) Module fbdev: vendor="X.Org Foundation" [ 161.456] compiled for 1.10.99.902, module version = 0.4.2 [ 161.456] ABI class: X.Org Video Driver, version 11.0 [ 161.456] (II) NOUVEAU driver [ 161.456] (II) NOUVEAU driver for NVIDIA chipset families : [ 161.456] RIVA TNT (NV04) [ 161.456] RIVA TNT2 (NV05) [ 161.456] GeForce 256 (NV10) [ 161.456] GeForce 2 (NV11, NV15) [ 161.456] GeForce 4MX (NV17, NV18) [ 161.456] GeForce 3 (NV20) [ 161.456] GeForce 4Ti (NV25, NV28) [ 161.456] GeForce FX (NV3x) [ 161.456] GeForce 6 (NV4x) [ 161.456] GeForce 7 (G7x) [ 161.456] GeForce 8 (G8x) [ 161.456] GeForce GTX 200 (NVA0) [ 161.456] GeForce GTX 400 (NVC0) [ 161.456] (II) VESA: driver for VESA chipsets: vesa [ 161.457] (II) FBDEV: driver for framebuffer: fbdev [ 161.457] (--) using VT number 2
[ 161.460] drmOpenDevice: node name is /dev/dri/card0 [ 161.470] drmOpenByBusid: Searching for BusID pci:0000:01:00.0 [ 161.470] drmOpenDevice: node name is /dev/dri/card0 [ 161.476] drmOpenByBusid: drmOpenMinor returns -1 [ 161.476] drmOpenDevice: node name is /dev/dri/card1 [ 161.482] drmOpenByBusid: drmOpenMinor returns -1 [ 161.482] drmOpenDevice: node name is /dev/dri/card2 [ 161.488] drmOpenByBusid: drmOpenMinor returns -1 [ 161.488] drmOpenDevice: node name is /dev/dri/card3 [ 161.495] drmOpenByBusid: drmOpenMinor returns -1 [ 161.495] drmOpenDevice: node name is /dev/dri/card4 [ 161.501] drmOpenByBusid: drmOpenMinor returns -1 [ 161.501] drmOpenDevice: node name is /dev/dri/card5 [ 161.507] drmOpenByBusid: drmOpenMinor returns -1 [ 161.507] drmOpenDevice: node name is /dev/dri/card6 [ 161.513] drmOpenByBusid: drmOpenMinor returns -1 [ 161.513] drmOpenDevice: node name is /dev/dri/card7 [ 161.519] drmOpenByBusid: drmOpenMinor returns -1 [ 161.519] drmOpenDevice: node name is /dev/dri/card8 [ 161.525] drmOpenByBusid: drmOpenMinor returns -1 [ 161.525] drmOpenDevice: node name is /dev/dri/card9 [ 161.532] drmOpenByBusid: drmOpenMinor returns -1 [ 161.532] drmOpenDevice: node name is /dev/dri/card10 [ 161.538] drmOpenByBusid: drmOpenMinor returns -1 [ 161.538] drmOpenDevice: node name is /dev/dri/card11 [ 161.544] drmOpenByBusid: drmOpenMinor returns -1 [ 161.544] drmOpenDevice: node name is /dev/dri/card12 [ 161.550] drmOpenByBusid: drmOpenMinor returns -1 [ 161.551] drmOpenDevice: node name is /dev/dri/card13 [ 161.557] drmOpenByBusid: drmOpenMinor returns -1 [ 161.557] drmOpenDevice: node name is /dev/dri/card14 [ 161.563] drmOpenByBusid: drmOpenMinor returns -1 [ 161.563] drmOpenDevice: node name is /dev/dri/card15 [ 161.569] drmOpenByBusid: drmOpenMinor returns -1 [ 161.569] drmOpenDevice: node name is /dev/dri/card0 [ 161.580] drmOpenDevice: node name is /dev/dri/card0 [ 161.586] drmOpenDevice: node name is /dev/dri/card1 [ 161.593] drmOpenDevice: node name is /dev/dri/card2 [ 161.599] drmOpenDevice: node name is /dev/dri/card3 [ 161.605] drmOpenDevice: node name is /dev/dri/card4 [ 161.611] drmOpenDevice: node name is /dev/dri/card5 [ 161.617] drmOpenDevice: node name is /dev/dri/card6 [ 161.624] drmOpenDevice: node name is /dev/dri/card7 [ 161.630] drmOpenDevice: node name is /dev/dri/card8 [ 161.636] drmOpenDevice: node name is /dev/dri/card9 [ 161.642] drmOpenDevice: node name is /dev/dri/card10 [ 161.648] drmOpenDevice: node name is /dev/dri/card11 [ 161.655] drmOpenDevice: node name is /dev/dri/card12 [ 161.661] drmOpenDevice: node name is /dev/dri/card13 [ 161.667] drmOpenDevice: node name is /dev/dri/card14 [ 161.673] drmOpenDevice: node name is /dev/dri/card15 [ 161.680] (EE) [drm] failed to open device [ 161.680] (II) Loading /usr/lib64/xorg/modules/drivers/vesa_drv.so [ 161.680] Refusing to touch device with a bound kernel driver [ 161.680] (WW) Falling back to old probe method for vesa [ 161.680] (II) Loading sub module "fbdevhw" [ 161.680] (II) LoadModule: "fbdevhw" [ 161.681] (II) Loading /usr/lib64/xorg/modules/libfbdevhw.so [ 161.681] (II) Module fbdevhw: vendor="X.Org Foundation" [ 161.681] compiled for 1.11.4, module version = 0.0.2 [ 161.681] ABI class: X.Org Video Driver, version 11.0 [ 161.681] (II) Loading /usr/lib64/xorg/modules/drivers/fbdev_drv.so [ 161.681] (II) Loading /usr/lib64/xorg/modules/libfbdevhw.so [ 161.692] (EE) open /dev/fb0: No such device [ 161.692] (WW) Falling back to old probe method for fbdev [ 161.692] (II) UnloadModule: "vesa" [ 161.692] (II) Unloading vesa [ 161.692] (II) UnloadModule: "fbdev" [ 161.692] (II) Unloading fbdev [ 161.692] (II) UnloadModule: "fbdevhw" [ 161.692] (II) Unloading fbdevhw [ 161.692] (EE) Screen(s) found, but none have a usable configuration. [ 161.692] Fatal server error: [ 161.692] no screens found [ 161.692] Please consult the Fedora Project support at http://wiki.x.org for help. [ 161.692] Please also check the log file at "/var/log/Xorg.0.log" for additional information. [ 161.692] [ 161.694] Server terminated with error (1). Closing log file.
On 15/02/12 11:21, Patrick Dupre wrote:
Hello,
With te new kernel 3.2.5-3.fc16.x86_64 I cannot start the graphics mode!
You have a mix of nVidia and nouveau drivers, and probably need a nVidia kmod that matches your new kernel.
See recent posts about 'no more nvidia' for hints.
John P
On Wednesday 15 February 2012 11:21:49 Patrick Dupre wrote:
Hello,
With te new kernel 3.2.5-3.fc16.x86_64 I cannot start the graphics mode!
[ 161.451] (II) LoadModule: "glx" [ 161.451] (II) Loading /usr/lib64/xorg/modules/extensions/nvidia/libglx.so [ 161.454] (II) Module glx: vendor="NVIDIA Corporation" [ 161.454] (II) NVIDIA GLX Module 290.10 Wed Nov 16 18:01:24 PST 2011
[snip]
[ 161.456] (II) LoadModule: "nouveau" [ 161.456] (II) Loading /usr/lib64/xorg/modules/drivers/nouveau_drv.so [ 161.456] (II) Module nouveau: vendor="X.Org Foundation"
You have both nVidia binary blob driver and Nouveau driver installed at the same time. This is well-known to Not Work. Either remove the binary blob and let Nouveau drive your hardware, or remove nouveau and let the binary blob drive your hardware. The world is too small for both of them... ;-)
HTH, :-) Marko
Thank for your email.
But could you tell me exactly what I should do?
I removed xorg-x11-drv-nouveau (and xorg-x11-drivers), but I did not get any improvements (see below)
Then I even cannot get back with the 1920x1080 resolution even with the previous kernel. So reinstalled the removed packages.
X.Org X Server 1.11.4 Release Date: 2012-01-27 [ 30.939] X Protocol Version 11, Revision 0 [ 30.939] Build Operating System: x86-01 2.6.32-220.4.1.el6.x86_64 [ 30.939] Current Operating System: Linux Homere 3.2.5-3.fc16.x86_64 #1 SMP Thu Feb 9 01:24:38 UTC 2012 x86_64 [ 30.939] Kernel command line: BOOT_IMAGE=/boot/vmlinuz-3.2.5-3.fc16.x86_64 root=UUID=e0d44568-5e7b-4c91-921f-ab1130406253 ro rd.md=0 rd.lvm=0 rd.dm=0 KEYTABLE=us SYSFONT=latarcyrheb-sun16 rd.luks=0 LANG=en_US.UTF-8 [ 30.940] Build Date: 08 February 2012 10:54:57PM [ 30.940] Build ID: xorg-x11-server 1.11.4-1.fc16 [ 30.940] Current version of pixman: 0.22.2 [ 30.940] Before reporting problems, check http://wiki.x.org to make sure that you have the latest version. [ 30.941] Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. [ 30.942] (==) Log file: "/var/log/Xorg.0.log", Time: Wed Feb 15 20:17:36 2012 [ 31.007] (==) Using config directory: "/etc/X11/xorg.conf.d" [ 31.007] (==) Using system config directory "/usr/share/X11/xorg.conf.d" [ 31.048] (==) No Layout section. Using the first Screen section. [ 31.048] (**) |-->Screen "Default Screen" (0) [ 31.048] (**) | |-->Monitor "VGA-1" [ 31.048] (==) Automatically adding devices [ 31.048] (==) Automatically enabling devices [ 31.056] (==) FontPath set to: catalogue:/etc/X11/fontpath.d, built-ins [ 31.056] (**) ModulePath set to "/usr/lib64/xorg/modules/extensions/nvidia,/usr/lib64/xorg/modules" [ 31.056] (II) The server relies on udev to provide the list of input devices. If no devices become available, reconfigure udev or disable AutoAddDevices. [ 31.056] (II) Loader magic: 0x7caac0 [ 31.056] (II) Module ABI versions: [ 31.056] X.Org ANSI C Emulation: 0.4 [ 31.056] X.Org Video Driver: 11.0 [ 31.056] X.Org XInput driver : 13.0 [ 31.056] X.Org Server Extension : 6.0 [ 31.058] (--) PCI:*(0:1:0:0) 10de:0a65:0000:0000 rev 162, Mem @ 0xfd000000/16777216, 0xd0000000/268435456, 0xce000000/33554432, I/O @ 0x0000d800/128, BIOS @ 0x????????/524288 [ 31.058] (II) LoadModule: "extmod" [ 31.072] (II) Loading /usr/lib64/xorg/modules/extensions/libextmod.so [ 31.082] (II) Module extmod: vendor="X.Org Foundation" [ 31.082] compiled for 1.11.4, module version = 1.0.0 [ 31.082] Module class: X.Org Server Extension [ 31.082] ABI class: X.Org Server Extension, version 6.0 [ 31.082] (II) Loading extension SELinux [ 31.082] (II) Loading extension MIT-SCREEN-SAVER [ 31.082] (II) Loading extension XFree86-VidModeExtension [ 31.082] (II) Loading extension XFree86-DGA [ 31.082] (II) Loading extension DPMS [ 31.082] (II) Loading extension XVideo [ 31.082] (II) Loading extension XVideo-MotionCompensation [ 31.082] (II) Loading extension X-Resource [ 31.082] (II) LoadModule: "dbe" [ 31.083] (II) Loading /usr/lib64/xorg/modules/extensions/libdbe.so [ 31.084] (II) Module dbe: vendor="X.Org Foundation" [ 31.084] compiled for 1.11.4, module version = 1.0.0 [ 31.084] Module class: X.Org Server Extension [ 31.084] ABI class: X.Org Server Extension, version 6.0 [ 31.084] (II) Loading extension DOUBLE-BUFFER [ 31.084] (II) LoadModule: "glx" [ 31.084] (II) Loading /usr/lib64/xorg/modules/extensions/nvidia/libglx.so [ 31.391] (II) Module glx: vendor="NVIDIA Corporation" [ 31.403] compiled for 4.0.2, module version = 1.0.0 [ 31.403] Module class: X.Org Server Extension [ 31.403] (II) NVIDIA GLX Module 290.10 Wed Nov 16 18:01:24 PST 2011 [ 31.403] (II) Loading extension GLX [ 31.403] (II) LoadModule: "record" [ 31.404] (II) Loading /usr/lib64/xorg/modules/extensions/librecord.so [ 31.410] (II) Module record: vendor="X.Org Foundation" [ 31.410] compiled for 1.11.4, module version = 1.13.0 [ 31.410] Module class: X.Org Server Extension [ 31.410] ABI class: X.Org Server Extension, version 6.0 [ 31.410] (II) Loading extension RECORD [ 31.410] (II) LoadModule: "dri" [ 31.411] (II) Loading /usr/lib64/xorg/modules/extensions/libdri.so [ 31.437] (II) Module dri: vendor="X.Org Foundation" [ 31.437] compiled for 1.11.4, module version = 1.0.0 [ 31.437] ABI class: X.Org Server Extension, version 6.0 [ 31.437] (II) Loading extension XFree86-DRI [ 31.437] (II) LoadModule: "dri2" [ 31.438] (II) Loading /usr/lib64/xorg/modules/extensions/libdri2.so [ 31.445] (II) Module dri2: vendor="X.Org Foundation" [ 31.445] compiled for 1.11.4, module version = 1.2.0 [ 31.445] ABI class: X.Org Server Extension, version 6.0 [ 31.445] (II) Loading extension DRI2 [ 31.445] (==) Matched nouveau as autoconfigured driver 0 [ 31.445] (==) Matched vesa as autoconfigured driver 1 [ 31.445] (==) Matched fbdev as autoconfigured driver 2 [ 31.445] (==) Assigned the driver to the xf86ConfigLayout [ 31.445] (II) LoadModule: "nouveau" [ 31.447] (WW) Warning, couldn't open module nouveau [ 31.447] (II) UnloadModule: "nouveau" [ 31.447] (II) Unloading nouveau [ 31.447] (EE) Failed to load module "nouveau" (module does not exist, 0) [ 31.447] (II) LoadModule: "vesa" [ 31.448] (II) Loading /usr/lib64/xorg/modules/drivers/vesa_drv.so [ 31.455] (II) Module vesa: vendor="X.Org Foundation" [ 31.455] compiled for 1.10.99.902, module version = 2.3.0 [ 31.455] Module class: X.Org Video Driver [ 31.455] ABI class: X.Org Video Driver, version 11.0 [ 31.455] (II) LoadModule: "fbdev" [ 31.456] (II) Loading /usr/lib64/xorg/modules/drivers/fbdev_drv.so [ 31.466] (II) Module fbdev: vendor="X.Org Foundation" [ 31.466] compiled for 1.10.99.902, module version = 0.4.2 [ 31.466] ABI class: X.Org Video Driver, version 11.0 [ 31.466] (II) VESA: driver for VESA chipsets: vesa [ 31.466] (II) FBDEV: driver for framebuffer: fbdev [ 31.466] (--) using VT number 2
[ 31.471] (II) Loading /usr/lib64/xorg/modules/drivers/vesa_drv.so [ 31.471] Refusing to touch device with a bound kernel driver [ 31.471] (WW) Falling back to old probe method for vesa [ 31.471] (II) Loading sub module "fbdevhw" [ 31.471] (II) LoadModule: "fbdevhw" [ 31.472] (II) Loading /usr/lib64/xorg/modules/libfbdevhw.so [ 31.482] (II) Module fbdevhw: vendor="X.Org Foundation" [ 31.482] compiled for 1.11.4, module version = 0.0.2 [ 31.482] ABI class: X.Org Video Driver, version 11.0 [ 31.482] (II) Loading /usr/lib64/xorg/modules/drivers/fbdev_drv.so [ 31.482] (II) Loading /usr/lib64/xorg/modules/libfbdevhw.so [ 31.487] (EE) open /dev/fb0: No such device [ 31.487] (WW) Falling back to old probe method for fbdev [ 31.487] (II) UnloadModule: "vesa" [ 31.487] (II) Unloading vesa [ 31.487] (II) UnloadModule: "fbdev" [ 31.487] (II) Unloading fbdev [ 31.487] (II) UnloadModule: "fbdevhw" [ 31.487] (II) Unloading fbdevhw [ 31.487] (EE) Screen(s) found, but none have a usable configuration. [ 31.487] Fatal server error: [ 31.487] no screens found [ 31.487] Please consult the Fedora Project support at http://wiki.x.org for help. [ 31.487] Please also check the log file at "/var/log/Xorg.0.log" for additional information. [ 31.487] [ 31.499] Server terminated with error (1). Closing log file.
On Wednesday 15 February 2012 11:21:49 Patrick Dupre wrote: Hello,
With te new kernel 3.2.5-3.fc16.x86_64 I cannot start the graphics mode!
[ 161.451] (II) LoadModule: "glx" [ 161.451] (II) Loading /usr/lib64/xorg/modules/extensions/nvidia/libglx.so [ 161.454] (II) Module glx: vendor="NVIDIA Corporation" [ 161.454] (II) NVIDIA GLX Module 290.10 Wed Nov 16 18:01:24 PST 2011
[snip]
[ 161.456] (II) LoadModule: "nouveau" [ 161.456] (II) Loading /usr/lib64/xorg/modules/drivers/nouveau_drv.so [ 161.456] (II) Module nouveau: vendor="X.Org Foundation"
You have both nVidia binary blob driver and Nouveau driver installed at the same time. This is well-known to Not Work. Either remove the binary blob and let Nouveau drive your hardware, or remove nouveau and let the binary blob drive your hardware. The world is too small for both of them... ;-)
HTH, :-) Marko
On Wednesday 15 February 2012 20:08:22 Patrick Dupre wrote:
But could you tell me exactly what I should do?
I removed xorg-x11-drv-nouveau (and xorg-x11-drivers), but I did not get any improvements (see below)
Then I even cannot get back with the 1920x1080 resolution even with the previous kernel. So reinstalled the removed packages.
If you want to use nVidia's binary blob, then in general you should:
(a) enable rpmfusion repositories (you probably have done that already), and (b) do a "yum install akmod-nvidia".
The latter will blacklist nouveau driver, rebuild your initrd image so that nouveau doesn't get loaded on reboot, build the nVidia binary blob against your running kernel (preferably the latest one), and set up xorg.conf appropriately for nVidia driver to work. After that you must reboot for the whole thing to take effect.
Note, though, that this is generic advice. Since it appears you have been removing/installing packages randomly, something else might be broken as well. If the above procedure doesn't give you expected results, post the latest /var/log/Xorg.0.log again so that we can troubleshoot it further.
HTH, :-) Marko
On 02/15/2012 12:38 PM, Marko Vojinovic wrote:
(b) do a "yum install akmod-nvidia".
Back when I installed akmod-nvidia, it needed either kernel-devel or kernel-headers, if not both. (It's been several years, so my memory's a tad fuzzy.) However, for whatever reason, it didn't list them as dependencies, which meant that you just had to know about them to get it working right. Is that still true, or has that been corrected?
On Wednesday 15 February 2012 12:54:30 Joe Zeff wrote:
On 02/15/2012 12:38 PM, Marko Vojinovic wrote:
(b) do a "yum install akmod-nvidia".
Back when I installed akmod-nvidia, it needed either kernel-devel or kernel-headers, if not both. (It's been several years, so my memory's a tad fuzzy.) However, for whatever reason, it didn't list them as dependencies, which meant that you just had to know about them to get it working right. Is that still true, or has that been corrected?
I don't know, but I assume yum will pull in any and all dependencies required to make it work, including the kernel header packages and the development toolchain for compilation.
On all machines where I had akmod-nvidia installed, I usually also had all that other stuff already installed beforehand, so I am not sure if akmod would drag it in if it weren't already there.
But even if it doesn't work automatically, the OP can always do it like this:
yum install kernel-headers kernel-devel yum groupinstall "Development Tools" yum install akmod-nvidia
That would probably cover all cases. :-) Also, IIRC, the OP had already installed the nVidia's .run package which also needs all that stuff, so it's likely he already has it all installed.
Best, :-) Marko
On 15 February 2012 22:37, Marko Vojinovic vvmarko@gmail.com wrote:
On Wednesday 15 February 2012 12:54:30 Joe Zeff wrote:
On 02/15/2012 12:38 PM, Marko Vojinovic wrote:
(b) do a "yum install akmod-nvidia".
Back when I installed akmod-nvidia, it needed either kernel-devel or kernel-headers, if not both. (It's been several years, so my memory's a tad fuzzy.) However, for whatever reason, it didn't list them as dependencies, which meant that you just had to know about them to get it working right. Is that still true, or has that been corrected?
I don't know, but I assume yum will pull in any and all dependencies required to make it work, including the kernel header packages and the development toolchain for compilation.
On all machines where I had akmod-nvidia installed, I usually also had all that other stuff already installed beforehand, so I am not sure if akmod would drag it in if it weren't already there.
But even if it doesn't work automatically, the OP can always do it like this:
yum install kernel-headers kernel-devel yum groupinstall "Development Tools" yum install akmod-nvidia
That would probably cover all cases. :-) Also, IIRC, the OP had already installed the nVidia's .run package which also needs all that stuff, so it's likely he already has it all installed.
$ rpm -qR akmods|grep devel kernel-devel
...so yes it's listed as a dependency and should get pulled in by yum.
On 02/15/2012 02:37 PM, Marko Vojinovic wrote:
I don't know, but I assume yum will pull in any and all dependencies required to make it work, including the kernel header packages and the development toolchain for compilation.
That's true if and only if they're properly listed. Back when I installed them, they weren't, so yum had no way of knowing that they were needed.
On Wed, 15 Feb 2012, Marko Vojinovic wrote:
On Wednesday 15 February 2012 12:54:30 Joe Zeff wrote: On 02/15/2012 12:38 PM, Marko Vojinovic wrote:
(b) do a "yum install akmod-nvidia".
Back when I installed akmod-nvidia, it needed either kernel-devel or kernel-headers, if not both. (It's been several years, so my memory's a tad fuzzy.) However, for whatever reason, it didn't list them as dependencies, which meant that you just had to know about them to get it working right. Is that still true, or has that been corrected?
I don't know, but I assume yum will pull in any and all dependencies required to make it work, including the kernel header packages and the development toolchain for compilation.
On all machines where I had akmod-nvidia installed, I usually also had all that other stuff already installed beforehand, so I am not sure if akmod would drag it in if it weren't already there.
But even if it doesn't work automatically, the OP can always do it like this:
yum install kernel-headers kernel-devel yum groupinstall "Development Tools" yum install akmod-nvidia
THis is OK
That would probably cover all cases. :-) Also, IIRC, the OP had already installed the nVidia's .run package which also needs all that stuff, so it's likely he already has it all installed.
Best, :-) Marko
On 02/17/2012 02:21 PM, Patrick Dupre wrote:
I don't know, but I assume yum will pull in any and all dependencies required to make it work, including the kernel header packages and the development toolchain for compilation.
Only if they're *listed* as dependencies. Back when I first installed akmod, back around F9, they weren't properly listed in the rpm, so you just had to know what to install to make them work. As I've never had to install them again, I didn't know if the problem still existed or not.
(Sorry for the delay in responding. I'm at a convention, and this is the first time I've been able to get caught up on email.)
Thank for your email.
Before I do any thing, I can see that I have kmod-nvidia-3.2.3-2.fc16.x86_64-290.10-1.fc16.12.x86_64 and kmod-nvidia-3.2.5-3.fc16.x86_64-290.10-1.fc16.13.x86_64
Should I remove the first one?
Thank.
On Wednesday 15 February 2012 20:08:22 Patrick Dupre wrote: But could you tell me exactly what I should do?
I removed xorg-x11-drv-nouveau (and xorg-x11-drivers), but I did not get any improvements (see below)
Then I even cannot get back with the 1920x1080 resolution even with the previous kernel. So reinstalled the removed packages.
If you want to use nVidia's binary blob, then in general you should:
(a) enable rpmfusion repositories (you probably have done that already), and (b) do a "yum install akmod-nvidia".
The latter will blacklist nouveau driver, rebuild your initrd image so that nouveau doesn't get loaded on reboot, build the nVidia binary blob against your running kernel (preferably the latest one), and set up xorg.conf appropriately for nVidia driver to work. After that you must reboot for the whole thing to take effect.
Note, though, that this is generic advice. Since it appears you have been removing/installing packages randomly, something else might be broken as well. If the above procedure doesn't give you expected results, post the latest /var/log/Xorg.0.log again so that we can troubleshoot it further.
HTH, :-) Marko
On Thursday 16 February 2012 11:55:43 Patrick Dupre wrote:
Before I do any thing, I can see that I have kmod-nvidia-3.2.3-2.fc16.x86_64-290.10-1.fc16.12.x86_64 and kmod-nvidia-3.2.5-3.fc16.x86_64-290.10-1.fc16.13.x86_64
Should I remove the first one?
The first kmod-nvidia is there for the kernel 3.2.3-2, while the second one is for kernel 3.2.5-3. I guess you can keep both packages if you plan to use the corresponding kernels.
Best, :-) Marko
On 16/02/12 11:55, Patrick Dupre wrote:
Thank for your email.
Before I do any thing, I can see that I have kmod-nvidia-3.2.3-2.fc16.x86_64-290.10-1.fc16.12.x86_64 and kmod-nvidia-3.2.5-3.fc16.x86_64-290.10-1.fc16.13.x86_64
Should I remove the first one?
Thank.
Hi: I use the ATrpms repo and so don't have first-hand experience of fusion, but your first post showed you have the 3.2.5-3 SMP kernel. You probably need a plain nvidia 290.10 package to match that, and a matching kmod. I had assumed the problem was that that had not been built when you tried the upgrade, but now you seem to have it, unless there's a SMP variant. If you were running with the 3.2.3-2 kmod package before, the one you have now should be ok; the one that you suggest removing won't be any use and the answer would be Yes.
I hope you don't mind me asking why you are using F16 in what sounds as if it ought to be a production environment, with no-one on site to help you. You will need to upgrade at least once a year: F16 isn't really a stable platform. I don't expect that you want to move to CentOS_6 or SL6 now, but they are expected to have RH support for 7 years and they are really quite competent.
Good luck,
John P
On Thu, 16 Feb 2012, John Pilkington wrote:
On 16/02/12 11:55, Patrick Dupre wrote:
Thank for your email.
Before I do any thing, I can see that I have kmod-nvidia-3.2.3-2.fc16.x86_64-290.10-1.fc16.12.x86_64 and kmod-nvidia-3.2.5-3.fc16.x86_64-290.10-1.fc16.13.x86_64
Should I remove the first one?
Thank.
Hi: I use the ATrpms repo and so don't have first-hand experience of fusion, but your first post showed you have the 3.2.5-3 SMP kernel. You probably need a plain nvidia 290.10 package to match that, and a matching kmod. I had assumed the problem was that that had not been built when you tried the upgrade, but now you seem to have it, unless there's a SMP variant. If you were running with the 3.2.3-2 kmod package before, the one you have now should be ok; the one that you suggest removing won't be any use and the answer would be Yes.
Sorry, but what is the SMP? I do not see any SMP stuff! kernel-3.2.5-3.fc16.x86_64 and kernel-3.2.3-2.fc16.x86_64 are both installed as well as kernel-devel-3.2.3-2.fc16.x86_64 and kernel-devel-3.2.5-3.fc16.x86_64, however, only kernel-headers-3.2.5-3.fc16.x86_64 is installed (no kernel-headers-3.2.2.3-2.fc16.x86_64)
I hope you don't mind me asking why you are using F16 in what sounds as if it ought to be a production environment, with no-one on site to help you. You will need to upgrade at least once a year: F16 isn't really a stable platform. I don't expect that you want to move to CentOS_6 or SL6 now, but they are expected to have RH support for 7 years and they are really quite competent.
Good luck,
John P
On Wednesday 15 February 2012 20:08:22 Patrick Dupre wrote: But could you tell me exactly what I should do?
I removed xorg-x11-drv-nouveau (and xorg-x11-drivers), but I did not get any improvements (see below)
Then I even cannot get back with the 1920x1080 resolution even with the previous kernel. So reinstalled the removed packages.
If you want to use nVidia's binary blob, then in general you should:
(a) enable rpmfusion repositories (you probably have done that already), and (b) do a "yum install akmod-nvidia".
The latter will blacklist nouveau driver, rebuild your initrd image so that nouveau doesn't get loaded on reboot, build the nVidia binary blob against your running kernel (preferably the latest one), and set up xorg.conf appropriately for nVidia driver to work. After that you must reboot for the whole thing to take effect.
Do you mean that I should recompile a kernel? I used to do so a long time ago, but with the advent of fedora, I did not have to do so, and the users should not have to recompile kernels.
To come back to the issue. I can run the graphics with the kernel 3,2,3-2 and if it does not work with the new kernel 3,2,5-3, I can only think that the new kernel has not been build up properly. However, when I compare the 2 files /usr/src/kernels/3.2.3-2.fc16.x86_64/.config and /usr/src/kernels/3.2.5-3.fc16.x86_64/.config they are identical. Then I can deduce that is is probably a modules issues. lsmod does seem to show that the module nvidia.ko is loaded tcp_lp 12663 0 fuse 77538 5 lockd 84578 0 nf_conntrack_ipv4 14622 8 nf_defrag_ipv4 12673 1 nf_conntrack_ipv4 ip6t_REJECT 12939 2 nf_conntrack_ipv6 14290 8 nf_defrag_ipv6 18139 1 nf_conntrack_ipv6 xt_state 12578 16 nf_conntrack 82331 3 nf_conntrack_ipv4,nf_conntrack_ipv6,xt_state ip6table_filter 12815 1 ip6_tables 26976 1 ip6table_filter be2iscsi 72362 0 iscsi_boot_sysfs 15641 1 be2iscsi bnx2i 54521 0 cnic 58676 1 bnx2i uio 19067 1 cnic cxgb4i 32909 0 cxgb4 102888 1 cxgb4i cxgb3i 32972 0 libcxgbi 56477 2 cxgb4i,cxgb3i cxgb3 155412 1 cxgb3i mdio 13398 1 cxgb3 ib_iser 38073 0 rdma_cm 41898 1 ib_iser ib_cm 41692 1 rdma_cm iw_cm 18176 1 rdma_cm ib_sa 28407 2 rdma_cm,ib_cm ib_mad 46392 2 ib_cm,ib_sa ib_core 73791 6 ib_iser,rdma_cm,ib_cm,iw_cm,ib_sa,ib_mad ib_addr 13748 1 rdma_cm iscsi_tcp 18333 0 libiscsi_tcp 23970 4 cxgb4i,cxgb3i,libcxgbi,iscsi_tcp libiscsi 50527 8 be2iscsi,bnx2i,cxgb4i,cxgb3i,libcxgbi,ib_iser,iscsi_tcp,libiscsi_tcp scsi_transport_iscsi 51823 8 be2iscsi,bnx2i,libcxgbi,ib_iser,iscsi_tcp,libiscsi snd_hda_codec_hdmi 36277 4 joydev 17412 0 snd_hda_codec_via 46559 1 snd_hda_intel 33276 3 snd_hda_codec 114615 3 snd_hda_codec_hdmi,snd_hda_codec_via,snd_hda_intel snd_hwdep 17611 1 snd_hda_codec snd_seq 64807 0 snd_seq_device 14129 1 snd_seq snd_pcm 97100 4 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec ppdev 17635 0 parport_pc 28082 0 parport 44515 2 ppdev,parport_pc snd_timer 28815 2 snd_seq,snd_pcm edac_core 51619 0 snd 74425 14 snd_hda_codec_hdmi,snd_hda_codec_via,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_seq,snd_seq_device,snd_pcm,snd_timer edac_mce_amd 22882 0 sp5100_tco 13697 0 soundcore 14484 1 snd serio_raw 13371 0 r8169 60789 0 mii 13527 1 r8169 snd_page_alloc 18101 2 snd_hda_intel,snd_pcm i2c_piix4 13694 0 microcode 23240 0 k10temp 13119 0 uinput 17606 0 sunrpc 235041 2 lockd binfmt_misc 17431 1 ata_generic 12899 0 pata_acpi 13027 0 pata_atiixp 13197 0 nouveau 763073 2 ttm 68749 1 nouveau drm_kms_helper 40141 1 nouveau drm 225935 4 nouveau,ttm,drm_kms_helper i2c_algo_bit 13156 1 nouveau i2c_core 37955 5 i2c_piix4,nouveau,drm_kms_helper,drm,i2c_algo_bit mxm_wmi 12823 1 nouveau wmi 18697 1 mxm_wmi video 18932 1 nouveau
Could you please clarify the situation?
Thank.
Note, though, that this is generic advice. Since it appears you have been removing/installing packages randomly, something else might be broken as well. If the above procedure doesn't give you expected results, post the latest /var/log/Xorg.0.log again so that we can troubleshoot it further.
HTH, :-) Marko
On Friday 17 February 2012 22:18:58 Patrick Dupre wrote:
On Wednesday 15 February 2012 20:08:22 Patrick Dupre wrote: But could you tell me exactly what I should do?
I removed xorg-x11-drv-nouveau (and xorg-x11-drivers), but I did not get any improvements (see below)
Then I even cannot get back with the 1920x1080 resolution even with the previous kernel. So reinstalled the removed packages.
If you want to use nVidia's binary blob, then in general you should:
(a) enable rpmfusion repositories (you probably have done that already), and (b) do a "yum install akmod-nvidia".
The latter will blacklist nouveau driver, rebuild your initrd image so that nouveau doesn't get loaded on reboot, build the nVidia binary blob against your running kernel (preferably the latest one), and set up xorg.conf appropriately for nVidia driver to work. After that you must reboot for the whole thing to take effect.
Do you mean that I should recompile a kernel?
No, there is no need to recompile the whole kernel. Just the graphics driver, which is a kernel module.
To come back to the issue. I can run the graphics with the kernel 3,2,3-2 and if it does not work with the new kernel 3,2,5-3, I can only think that the new kernel has not been build up properly. However, when I compare the 2 files /usr/src/kernels/3.2.3-2.fc16.x86_64/.config and /usr/src/kernels/3.2.5-3.fc16.x86_64/.config they are identical. Then I can deduce that is is probably a modules issues.
Precisely. The kernel configs have nothing to do with the graphics driver. The graphics driver is a kernel module, and apparently you have one for the 3.2.3-2 kernel, while you don't have it for the 3.2.5-3 kernel.
lsmod does seem to show that the module nvidia.ko is loaded
No, it doesn't, the word "nvidia" doesn't appear anywhere in the list. See below for an explanation.
tcp_lp 12663 0 fuse 77538 5 lockd 84578 0 nf_conntrack_ipv4 14622 8 nf_defrag_ipv4 12673 1 nf_conntrack_ipv4 ip6t_REJECT 12939 2 nf_conntrack_ipv6 14290 8 nf_defrag_ipv6 18139 1 nf_conntrack_ipv6 xt_state 12578 16 nf_conntrack 82331 3 nf_conntrack_ipv4,nf_conntrack_ipv6,xt_state ip6table_filter 12815 1 ip6_tables 26976 1 ip6table_filter be2iscsi 72362 0 iscsi_boot_sysfs 15641 1 be2iscsi bnx2i 54521 0 cnic 58676 1 bnx2i uio 19067 1 cnic cxgb4i 32909 0 cxgb4 102888 1 cxgb4i cxgb3i 32972 0 libcxgbi 56477 2 cxgb4i,cxgb3i cxgb3 155412 1 cxgb3i mdio 13398 1 cxgb3 ib_iser 38073 0 rdma_cm 41898 1 ib_iser ib_cm 41692 1 rdma_cm iw_cm 18176 1 rdma_cm ib_sa 28407 2 rdma_cm,ib_cm ib_mad 46392 2 ib_cm,ib_sa ib_core 73791 6 ib_iser,rdma_cm,ib_cm,iw_cm,ib_sa,ib_mad ib_addr 13748 1 rdma_cm iscsi_tcp 18333 0 libiscsi_tcp 23970 4 cxgb4i,cxgb3i,libcxgbi,iscsi_tcp libiscsi 50527 8 be2iscsi,bnx2i,cxgb4i,cxgb3i,libcxgbi,ib_iser,iscsi_tcp,libiscsi_tcp scsi_transport_iscsi 51823 8 be2iscsi,bnx2i,libcxgbi,ib_iser,iscsi_tcp,libiscsi snd_hda_codec_hdmi 36277 4 joydev 17412 0 snd_hda_codec_via 46559 1 snd_hda_intel 33276 3 snd_hda_codec 114615 3 snd_hda_codec_hdmi,snd_hda_codec_via,snd_hda_intel snd_hwdep 17611 1 snd_hda_codec snd_seq 64807 0 snd_seq_device 14129 1 snd_seq snd_pcm 97100 4 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec ppdev 17635 0 parport_pc 28082 0 parport 44515 2 ppdev,parport_pc snd_timer 28815 2 snd_seq,snd_pcm edac_core 51619 0 snd 74425 14 snd_hda_codec_hdmi,snd_hda_codec_via,snd_hda_intel,snd_hda_codec,snd_hwdep,s nd_seq,snd_seq_device,snd_pcm,snd_timer edac_mce_amd 22882 0 sp5100_tco 13697 0 soundcore 14484 1 snd serio_raw 13371 0 r8169 60789 0 mii 13527 1 r8169 snd_page_alloc 18101 2 snd_hda_intel,snd_pcm i2c_piix4 13694 0 microcode 23240 0 k10temp 13119 0 uinput 17606 0 sunrpc 235041 2 lockd binfmt_misc 17431 1 ata_generic 12899 0 pata_acpi 13027 0 pata_atiixp 13197 0 nouveau 763073 2 ttm 68749 1 nouveau drm_kms_helper 40141 1 nouveau drm 225935 4 nouveau,ttm,drm_kms_helper i2c_algo_bit 13156 1 nouveau i2c_core 37955 5 i2c_piix4,nouveau,drm_kms_helper,drm,i2c_algo_bit mxm_wmi 12823 1 nouveau wmi 18697 1 mxm_wmi video 18932 1 nouveau
Could you please clarify the situation?
Ok, this is how it all works. The driver for your nVidia graphics hardware is a kernel module. The module version has to match the kernel version, so you need one driver module for each kernel. Technically speaking, the module needs to be built "against the appropriate kernel", typically while that kernel is running.
There are two different graphics drivers for nVidia cards. One is called "nouveau", and is an open-source driver provided by Fedora out-of-the-box. The other is called "nvidia", and is a closed-source driver provided by nVidia (also commonly called "binary blob"). The nvidia driver mostly works better than nouveau, since nouveau is still under development.
Nouveau and nvidia modules cannot coexist. If nouveau is to work properly, nvidia driver must be uninstalled from the machine. If nvidia is to work properly, nouveau needs to be blacklisted and unistalled from the initrd. Since the Fedora's kernel automatically loads nouveau on boot (if it is not blacklisted), one needs to reboot the machine in order to switch from nouveau to nvidia (or vice-versa).
The nvidia driver can be obtained in two different ways. One way is to download the appropriate .run file from nVidia website and run it. This is the *wrong* way, since it doesn't play nice with Fedora, overwrites some system files, and is a pain to remove once installed. The other way is to install (via yum) the package called akmod-nvidia, which is provided by the rpmfusion repository. This is the *correct* way, since that package was designed precisely for Fedora, doesn't overwrite system files and can be easily removed. The driver itself (the nvidia kernel module) is the same as the one provided by nVidia website --- the difference is in the details of the installation procedure.
For your situation, I suggest the following:
(1) Run "yum update". There is a new kernel available (3.2.6-3), and you should install it.
(2) Run "mv /etc/X11/xorg.conf /etc/X11/xorg.conf-backup". This will rename the xorg.conf file (if you have it), and let X run with default configuration (i.e. without an xorg.conf file) after a reboot.
(3) Boot into the new kernel. Since you have just installed it, and since there is no xorg.conf, the nouveau driver should be activated. Maybe it will have problems running on your graphics card (it shouldn't, but...), and maybe you will not have high resolution, or even X might not start at all. That is ok, it doesn't matter, as long as you do the next step.
(4) Run "yum install akmod-nvidia". Let it install. It will provide the nvidia driver matching your new kernel, it will write a new xorg.conf, it will blacklist nouveau and remove it from initrd of your kernel. It might take some time to complete.
(5) Reboot the machine again into the latest kernel. This is necessary in order to change from nouveau to nvidia driver.
If all is well, the nvidia driver should be loaded and X should work perfectly, high resolution and all. If it doesn't work, post the contents of /var/log/Xorg.0.log file back to us, so that we can see why it doesn't work and tell you how to fix it.
I hope this was clear enough. Just follow the steps (1) - (5), and it should start working again.
Note that once you have installed akmod-nvidia, in the future when a new kernel gets installed with an update, everything will happen automatically and you do not need to do anything in order to get the nvidia driver for that new kernel. The above (1)-(5) procedure should be done only once, and after that it will work automatically for every new kernel.
HTH, :-) Marko
On 02/17/2012 10:46 PM, Marko Vojinovic wrote:
The nvidia driver can be obtained in two different ways. One way is to download the appropriate .run file from nVidia website and run it. This is the*wrong* way, since it doesn't play nice with Fedora, overwrites some system files, and is a pain to remove once installed. The other way is to install (via yum) the package called akmod-nvidia, which is provided by the rpmfusion repository.
Third way: install kmod-nvidia, from the same repo. This has the advantage that it's pre-built and needs no development packages or other baggage and the disadvantage that it sometimes lags behind the new kernel by a day or so, in which case you simply don't reboot into the new kernel until it does.
Dear Marko,
Thank you very much for your long email. See below my comments.
On Friday 17 February 2012 22:18:58 Patrick Dupre wrote:
On Wednesday 15 February 2012 20:08:22 Patrick Dupre wrote: But could you tell me exactly what I should do?
I removed xorg-x11-drv-nouveau (and xorg-x11-drivers), but I did not get any improvements (see below)
Then I even cannot get back with the 1920x1080 resolution even with the previous kernel. So reinstalled the removed packages.
If you want to use nVidia's binary blob, then in general you should:
(a) enable rpmfusion repositories (you probably have done that already), and (b) do a "yum install akmod-nvidia".
The latter will blacklist nouveau driver, rebuild your initrd image so that nouveau doesn't get loaded on reboot, build the nVidia binary blob against your running kernel (preferably the latest one), and set up xorg.conf appropriately for nVidia driver to work. After that you must reboot for the whole thing to take effect.
Do you mean that I should recompile a kernel?
No, there is no need to recompile the whole kernel. Just the graphics driver, which is a kernel module.
To come back to the issue. I can run the graphics with the kernel 3,2,3-2 and if it does not work with the new kernel 3,2,5-3, I can only think that the new kernel has not been build up properly. However, when I compare the 2 files /usr/src/kernels/3.2.3-2.fc16.x86_64/.config and /usr/src/kernels/3.2.5-3.fc16.x86_64/.config they are identical. Then I can deduce that is is probably a modules issues.
Precisely. The kernel configs have nothing to do with the graphics driver. The graphics driver is a kernel module, and apparently you have one for the 3.2.3-2 kernel, while you don't have it for the 3.2.5-3 kernel.
lsmod does seem to show that the module nvidia.ko is loaded
No, it doesn't, the word "nvidia" doesn't appear anywhere in the list. See below for an explanation.
tcp_lp 12663 0 fuse 77538 5 lockd 84578 0 nf_conntrack_ipv4 14622 8 nf_defrag_ipv4 12673 1 nf_conntrack_ipv4 ip6t_REJECT 12939 2 nf_conntrack_ipv6 14290 8 nf_defrag_ipv6 18139 1 nf_conntrack_ipv6 xt_state 12578 16 nf_conntrack 82331 3 nf_conntrack_ipv4,nf_conntrack_ipv6,xt_state ip6table_filter 12815 1 ip6_tables 26976 1 ip6table_filter be2iscsi 72362 0 iscsi_boot_sysfs 15641 1 be2iscsi bnx2i 54521 0 cnic 58676 1 bnx2i uio 19067 1 cnic cxgb4i 32909 0 cxgb4 102888 1 cxgb4i cxgb3i 32972 0 libcxgbi 56477 2 cxgb4i,cxgb3i cxgb3 155412 1 cxgb3i mdio 13398 1 cxgb3 ib_iser 38073 0 rdma_cm 41898 1 ib_iser ib_cm 41692 1 rdma_cm iw_cm 18176 1 rdma_cm ib_sa 28407 2 rdma_cm,ib_cm ib_mad 46392 2 ib_cm,ib_sa ib_core 73791 6 ib_iser,rdma_cm,ib_cm,iw_cm,ib_sa,ib_mad ib_addr 13748 1 rdma_cm iscsi_tcp 18333 0 libiscsi_tcp 23970 4 cxgb4i,cxgb3i,libcxgbi,iscsi_tcp libiscsi 50527 8 be2iscsi,bnx2i,cxgb4i,cxgb3i,libcxgbi,ib_iser,iscsi_tcp,libiscsi_tcp scsi_transport_iscsi 51823 8 be2iscsi,bnx2i,libcxgbi,ib_iser,iscsi_tcp,libiscsi snd_hda_codec_hdmi 36277 4 joydev 17412 0 snd_hda_codec_via 46559 1 snd_hda_intel 33276 3 snd_hda_codec 114615 3 snd_hda_codec_hdmi,snd_hda_codec_via,snd_hda_intel snd_hwdep 17611 1 snd_hda_codec snd_seq 64807 0 snd_seq_device 14129 1 snd_seq snd_pcm 97100 4 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec ppdev 17635 0 parport_pc 28082 0 parport 44515 2 ppdev,parport_pc snd_timer 28815 2 snd_seq,snd_pcm edac_core 51619 0 snd 74425 14 snd_hda_codec_hdmi,snd_hda_codec_via,snd_hda_intel,snd_hda_codec,snd_hwdep,s nd_seq,snd_seq_device,snd_pcm,snd_timer edac_mce_amd 22882 0 sp5100_tco 13697 0 soundcore 14484 1 snd serio_raw 13371 0 r8169 60789 0 mii 13527 1 r8169 snd_page_alloc 18101 2 snd_hda_intel,snd_pcm i2c_piix4 13694 0 microcode 23240 0 k10temp 13119 0 uinput 17606 0 sunrpc 235041 2 lockd binfmt_misc 17431 1 ata_generic 12899 0 pata_acpi 13027 0 pata_atiixp 13197 0 nouveau 763073 2 ttm 68749 1 nouveau drm_kms_helper 40141 1 nouveau drm 225935 4 nouveau,ttm,drm_kms_helper i2c_algo_bit 13156 1 nouveau i2c_core 37955 5 i2c_piix4,nouveau,drm_kms_helper,drm,i2c_algo_bit mxm_wmi 12823 1 nouveau wmi 18697 1 mxm_wmi video 18932 1 nouveau
Could you please clarify the situation?
Ok, this is how it all works. The driver for your nVidia graphics hardware is a kernel module. The module version has to match the kernel version, so you need one driver module for each kernel. Technically speaking, the module needs to be built "against the appropriate kernel", typically while that kernel is running.
There are two different graphics drivers for nVidia cards. One is called "nouveau", and is an open-source driver provided by Fedora out-of-the-box. The other is called "nvidia", and is a closed-source driver provided by nVidia (also commonly called "binary blob"). The nvidia driver mostly works better than nouveau, since nouveau is still under development.
Nouveau and nvidia modules cannot coexist. If nouveau is to work properly, nvidia driver must be uninstalled from the machine. If nvidia is to work properly, nouveau needs to be blacklisted and unistalled from the initrd. Since the Fedora's kernel automatically loads nouveau on boot (if it is not blacklisted), one needs to reboot the machine in order to switch from nouveau to nvidia (or vice-versa).
The nvidia driver can be obtained in two different ways. One way is to download the appropriate .run file from nVidia website and run it. This is the *wrong* way, since it doesn't play nice with Fedora, overwrites some system files, and is a pain to remove once installed. The other way is to install (via yum) the package called akmod-nvidia, which is provided by the rpmfusion repository. This is the *correct* way, since that package was designed precisely for Fedora, doesn't overwrite system files and can be easily removed. The driver itself (the nvidia kernel module) is the same as the one provided by nVidia website --- the difference is in the details of the installation procedure.
For your situation, I suggest the following:
(1) Run "yum update". There is a new kernel available (3.2.6-3), and you should install it.
(2) Run "mv /etc/X11/xorg.conf /etc/X11/xorg.conf-backup". This will rename the xorg.conf file (if you have it), and let X run with default configuration (i.e. without an xorg.conf file) after a reboot.
======================================================================= This is probably were my problem is. I removed the /etc/X11/xorg.conf because if it is here I can never start my graphic mode with the rigth resolution. So I use /etc/X11/xorg.conf.d with the following files:
00-nvidia.conf 00-system-setup-keyboard.conf 40-monitor.conf 50-screen.conf (see attached file)
How is is connectred to nouveau and nvdia?
For now I have: xorg-x11-drv-nouveau-0.0.16-27.20110720gitb806e3f.fc16.x86_64 and
akmod-nvidia-290.10-1.fc16.4.x86_64 kmod-nvidia-3.2.7-1.fc16.x86_64-290.10-1.fc16.15.x86_64 xorg-x11-drv-nvidia-290.10-1.fc16.x86_64 kmod-nvidia-3.2.3-2.fc16.x86_64-290.10-1.fc16.12.x86_64 kmod-nvidia-3.2.6-3.fc16.x86_64-290.10-1.fc16.14.x86_64 xorg-x11-drv-nvidia-libs-290.10-1.fc16.x86_64 nvidia-xconfig-1.0-11.fc16.x86_64
Should I remove one of them? After we will have clarify this point, I will eventually go into your procedure. But my understanding is that I should start nouveau since there is no xorg.conf, but when I build such a file, then the graphics is wrong!!
Thank again for your help.
=========================================================================
(3) Boot into the new kernel. Since you have just installed it, and since there is no xorg.conf, the nouveau driver should be activated. Maybe it will have problems running on your graphics card (it shouldn't, but...), and maybe you will not have high resolution, or even X might not start at all. That is ok, it doesn't matter, as long as you do the next step.
(4) Run "yum install akmod-nvidia". Let it install. It will provide the nvidia driver matching your new kernel, it will write a new xorg.conf, it will blacklist nouveau and remove it from initrd of your kernel. It might take some time to complete.
(5) Reboot the machine again into the latest kernel. This is necessary in order to change from nouveau to nvidia driver.
If all is well, the nvidia driver should be loaded and X should work perfectly, high resolution and all. If it doesn't work, post the contents of /var/log/Xorg.0.log file back to us, so that we can see why it doesn't work and tell you how to fix it.
I hope this was clear enough. Just follow the steps (1) - (5), and it should start working again.
Note that once you have installed akmod-nvidia, in the future when a new kernel gets installed with an update, everything will happen automatically and you do not need to do anything in order to get the nvidia driver for that new kernel. The above (1)-(5) procedure should be done only once, and after that it will work automatically for every new kernel.
HTH, :-) Marko
On Friday 24 February 2012 22:55:09 Patrick Dupre wrote:
(2) Run "mv /etc/X11/xorg.conf /etc/X11/xorg.conf-backup". This will rename the xorg.conf file (if you have it), and let X run with default configuration (i.e. without an xorg.conf file) after a reboot.
=======================================================================
This is probably were my problem is. I removed the /etc/X11/xorg.conf because if it is here I can never start my graphic mode with the rigth resolution. So I use /etc/X11/xorg.conf.d with the following files:
00-nvidia.conf 00-system-setup-keyboard.conf 40-monitor.conf 50-screen.conf (see attached file)
How is is connectred to nouveau and nvdia?
Well, obviously, you want to (re)move these files as well. The idea of removing xorg.conf is to start with a clean configuration. You want to remove any and all changes you have made to xorg.conf and xorg.conf.d/ directory. Delete/rename the xorg.conf file, and make sure that the directory xorg.conf.d/ is empty.
Once we fix the loading of the proper driver, we can re-add the configuration in order to fix your resolution problems. Also, I could bet that your resolution will work out-of-the-box, once the nvidia or nouveau driver is properly in place.
Currently, you have nouveau loading, and the 00-nvidia.conf file is making it fail (remember, never both drivers at the same time!). Your system then falls back to the default generic vesa driver, which doesn't support resolutions beyond 1024x768. That is why your resolution doesn't work properly.
So, remove the xorg.conf and empty the xorg.conf.d, and then do all the steps as I explained in the previous e-mail.
For now I have: xorg-x11-drv-nouveau-0.0.16-27.20110720gitb806e3f.fc16.x86_64 and
akmod-nvidia-290.10-1.fc16.4.x86_64 kmod-nvidia-3.2.7-1.fc16.x86_64-290.10-1.fc16.15.x86_64 xorg-x11-drv-nvidia-290.10-1.fc16.x86_64 kmod-nvidia-3.2.3-2.fc16.x86_64-290.10-1.fc16.12.x86_64 kmod-nvidia-3.2.6-3.fc16.x86_64-290.10-1.fc16.14.x86_64 xorg-x11-drv-nvidia-libs-290.10-1.fc16.x86_64 nvidia-xconfig-1.0-11.fc16.x86_64
Should I remove one of them?
Ok, it would be best to do a
yum remove akmod-nvidia kmod-nvidia xorg-x11-drv-nvidia nvidia-xconfig
We will re-add these as necessary later. Remember, remove *everything* nvidia- related. All rpm packages and any configuration files. You want to start with a clean, default Fedora configuration for X.
After we will have clarify this point, I will eventually go into your procedure. But my understanding is that I should start nouveau since there is no xorg.conf, but when I build such a file, then the graphics is wrong!!
The nouveau driver, if present on your system, will be loaded by the *kernel*, during boot procedure, long before xorg.conf is looked at. The xorg.conf file has nothing to do with loading nouveau.
If instead you want to use nvidia driver (and you do), then nouveau would be removed, the kernel would boot with the generic vesa driver, then it would read the (nvidia-created) xorg.conf, and then load the nvidia driver when X gets started.
That is how it works. Nouveau and nvidia drivers work quite differently, get loaded at different times, one doesn't need an xorg.conf and the other does need it, etc.
So again, please --- if you want to fix your problem, you need to remove any and all traces of nvidia software from your system, remove any custom X configuration files, perform a "yum update", reboot into the latest kernel, do a "yum install akmod-nvidia", and reboot into the same kernel again.
That should be enough to make your graphics work properly. If it still doesn't work, send us the /var/log/xorg.0.log so that we can see what else is failing.
HTH, :-) Marko
Hello Marko,
Congratulations, I worked exactly like you said. Thank again.
On Friday 24 February 2012 22:55:09 Patrick Dupre wrote: (2) Run "mv /etc/X11/xorg.conf /etc/X11/xorg.conf-backup". This will rename the xorg.conf file (if you have it), and let X run with default configuration (i.e. without an xorg.conf file) after a reboot.
=======================================================================
This is probably were my problem is. I removed the /etc/X11/xorg.conf because if it is here I can never start my graphic mode with the rigth resolution. So I use /etc/X11/xorg.conf.d with the following files:
00-nvidia.conf 00-system-setup-keyboard.conf 40-monitor.conf 50-screen.conf (see attached file)
How is is connectred to nouveau and nvdia?
Well, obviously, you want to (re)move these files as well. The idea of removing xorg.conf is to start with a clean configuration. You want to remove any and all changes you have made to xorg.conf and xorg.conf.d/ directory. Delete/rename the xorg.conf file, and make sure that the directory xorg.conf.d/ is empty.
Once we fix the loading of the proper driver, we can re-add the configuration in order to fix your resolution problems. Also, I could bet that your resolution will work out-of-the-box, once the nvidia or nouveau driver is properly in place.
Currently, you have nouveau loading, and the 00-nvidia.conf file is making it fail (remember, never both drivers at the same time!). Your system then falls back to the default generic vesa driver, which doesn't support resolutions beyond 1024x768. That is why your resolution doesn't work properly.
So, remove the xorg.conf and empty the xorg.conf.d, and then do all the steps as I explained in the previous e-mail.
For now I have: xorg-x11-drv-nouveau-0.0.16-27.20110720gitb806e3f.fc16.x86_64 and
akmod-nvidia-290.10-1.fc16.4.x86_64 kmod-nvidia-3.2.7-1.fc16.x86_64-290.10-1.fc16.15.x86_64 xorg-x11-drv-nvidia-290.10-1.fc16.x86_64 kmod-nvidia-3.2.3-2.fc16.x86_64-290.10-1.fc16.12.x86_64 kmod-nvidia-3.2.6-3.fc16.x86_64-290.10-1.fc16.14.x86_64 xorg-x11-drv-nvidia-libs-290.10-1.fc16.x86_64 nvidia-xconfig-1.0-11.fc16.x86_64
Should I remove one of them?
Ok, it would be best to do a
yum remove akmod-nvidia kmod-nvidia xorg-x11-drv-nvidia nvidia-xconfig
We will re-add these as necessary later. Remember, remove *everything* nvidia- related. All rpm packages and any configuration files. You want to start with a clean, default Fedora configuration for X.
After we will have clarify this point, I will eventually go into your procedure. But my understanding is that I should start nouveau since there is no xorg.conf, but when I build such a file, then the graphics is wrong!!
The nouveau driver, if present on your system, will be loaded by the *kernel*, during boot procedure, long before xorg.conf is looked at. The xorg.conf file has nothing to do with loading nouveau.
If instead you want to use nvidia driver (and you do), then nouveau would be removed, the kernel would boot with the generic vesa driver, then it would read the (nvidia-created) xorg.conf, and then load the nvidia driver when X gets started.
That is how it works. Nouveau and nvidia drivers work quite differently, get loaded at different times, one doesn't need an xorg.conf and the other does need it, etc.
So again, please --- if you want to fix your problem, you need to remove any and all traces of nvidia software from your system, remove any custom X configuration files, perform a "yum update", reboot into the latest kernel, do a "yum install akmod-nvidia", and reboot into the same kernel again.
That should be enough to make your graphics work properly. If it still doesn't work, send us the /var/log/xorg.0.log so that we can see what else is failing.
HTH, :-) Marko
On Sat, 18 Feb 2012, Marko Vojinovic wrote:
On Friday 17 February 2012 22:18:58 Patrick Dupre wrote:
On Wednesday 15 February 2012 20:08:22 Patrick Dupre wrote: But could you tell me exactly what I should do?
I removed xorg-x11-drv-nouveau (and xorg-x11-drivers), but I did not get any improvements (see below)
Then I even cannot get back with the 1920x1080 resolution even with the previous kernel. So reinstalled the removed packages.
If you want to use nVidia's binary blob, then in general you should:
(a) enable rpmfusion repositories (you probably have done that already), and (b) do a "yum install akmod-nvidia".
The latter will blacklist nouveau driver, rebuild your initrd image so that nouveau doesn't get loaded on reboot, build the nVidia binary blob against your running kernel (preferably the latest one), and set up xorg.conf appropriately for nVidia driver to work. After that you must reboot for the whole thing to take effect.
Do you mean that I should recompile a kernel?
No, there is no need to recompile the whole kernel. Just the graphics driver, which is a kernel module.
To come back to the issue. I can run the graphics with the kernel 3,2,3-2 and if it does not work with the new kernel 3,2,5-3, I can only think that the new kernel has not been build up properly. However, when I compare the 2 files /usr/src/kernels/3.2.3-2.fc16.x86_64/.config and /usr/src/kernels/3.2.5-3.fc16.x86_64/.config they are identical. Then I can deduce that is is probably a modules issues.
Precisely. The kernel configs have nothing to do with the graphics driver. The graphics driver is a kernel module, and apparently you have one for the 3.2.3-2 kernel, while you don't have it for the 3.2.5-3 kernel.
lsmod does seem to show that the module nvidia.ko is loaded
No, it doesn't, the word "nvidia" doesn't appear anywhere in the list. See below for an explanation.
tcp_lp 12663 0 fuse 77538 5 lockd 84578 0 nf_conntrack_ipv4 14622 8 nf_defrag_ipv4 12673 1 nf_conntrack_ipv4 ip6t_REJECT 12939 2 nf_conntrack_ipv6 14290 8 nf_defrag_ipv6 18139 1 nf_conntrack_ipv6 xt_state 12578 16 nf_conntrack 82331 3 nf_conntrack_ipv4,nf_conntrack_ipv6,xt_state ip6table_filter 12815 1 ip6_tables 26976 1 ip6table_filter be2iscsi 72362 0 iscsi_boot_sysfs 15641 1 be2iscsi bnx2i 54521 0 cnic 58676 1 bnx2i uio 19067 1 cnic cxgb4i 32909 0 cxgb4 102888 1 cxgb4i cxgb3i 32972 0 libcxgbi 56477 2 cxgb4i,cxgb3i cxgb3 155412 1 cxgb3i mdio 13398 1 cxgb3 ib_iser 38073 0 rdma_cm 41898 1 ib_iser ib_cm 41692 1 rdma_cm iw_cm 18176 1 rdma_cm ib_sa 28407 2 rdma_cm,ib_cm ib_mad 46392 2 ib_cm,ib_sa ib_core 73791 6 ib_iser,rdma_cm,ib_cm,iw_cm,ib_sa,ib_mad ib_addr 13748 1 rdma_cm iscsi_tcp 18333 0 libiscsi_tcp 23970 4 cxgb4i,cxgb3i,libcxgbi,iscsi_tcp libiscsi 50527 8 be2iscsi,bnx2i,cxgb4i,cxgb3i,libcxgbi,ib_iser,iscsi_tcp,libiscsi_tcp scsi_transport_iscsi 51823 8 be2iscsi,bnx2i,libcxgbi,ib_iser,iscsi_tcp,libiscsi snd_hda_codec_hdmi 36277 4 joydev 17412 0 snd_hda_codec_via 46559 1 snd_hda_intel 33276 3 snd_hda_codec 114615 3 snd_hda_codec_hdmi,snd_hda_codec_via,snd_hda_intel snd_hwdep 17611 1 snd_hda_codec snd_seq 64807 0 snd_seq_device 14129 1 snd_seq snd_pcm 97100 4 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec ppdev 17635 0 parport_pc 28082 0 parport 44515 2 ppdev,parport_pc snd_timer 28815 2 snd_seq,snd_pcm edac_core 51619 0 snd 74425 14 snd_hda_codec_hdmi,snd_hda_codec_via,snd_hda_intel,snd_hda_codec,snd_hwdep,s nd_seq,snd_seq_device,snd_pcm,snd_timer edac_mce_amd 22882 0 sp5100_tco 13697 0 soundcore 14484 1 snd serio_raw 13371 0 r8169 60789 0 mii 13527 1 r8169 snd_page_alloc 18101 2 snd_hda_intel,snd_pcm i2c_piix4 13694 0 microcode 23240 0 k10temp 13119 0 uinput 17606 0 sunrpc 235041 2 lockd binfmt_misc 17431 1 ata_generic 12899 0 pata_acpi 13027 0 pata_atiixp 13197 0 nouveau 763073 2 ttm 68749 1 nouveau drm_kms_helper 40141 1 nouveau drm 225935 4 nouveau,ttm,drm_kms_helper i2c_algo_bit 13156 1 nouveau i2c_core 37955 5 i2c_piix4,nouveau,drm_kms_helper,drm,i2c_algo_bit mxm_wmi 12823 1 nouveau wmi 18697 1 mxm_wmi video 18932 1 nouveau
Could you please clarify the situation?
Ok, this is how it all works. The driver for your nVidia graphics hardware is a kernel module. The module version has to match the kernel version, so you need one driver module for each kernel. Technically speaking, the module needs to be built "against the appropriate kernel", typically while that kernel is running.
There are two different graphics drivers for nVidia cards. One is called "nouveau", and is an open-source driver provided by Fedora out-of-the-box. The other is called "nvidia", and is a closed-source driver provided by nVidia (also commonly called "binary blob"). The nvidia driver mostly works better than nouveau, since nouveau is still under development.
Nouveau and nvidia modules cannot coexist. If nouveau is to work properly, nvidia driver must be uninstalled from the machine. If nvidia is to work properly, nouveau needs to be blacklisted and unistalled from the initrd. Since the Fedora's kernel automatically loads nouveau on boot (if it is not blacklisted), one needs to reboot the machine in order to switch from nouveau to nvidia (or vice-versa).
The nvidia driver can be obtained in two different ways. One way is to download the appropriate .run file from nVidia website and run it. This is the *wrong* way, since it doesn't play nice with Fedora, overwrites some system files, and is a pain to remove once installed. The other way is to install (via yum) the package called akmod-nvidia, which is provided by the rpmfusion repository. This is the *correct* way, since that package was designed precisely for Fedora, doesn't overwrite system files and can be easily removed. The driver itself (the nvidia kernel module) is the same as the one provided by nVidia website --- the difference is in the details of the installation procedure.
For your situation, I suggest the following:
(1) Run "yum update". There is a new kernel available (3.2.6-3), and you should install it.
Done
(2) Run "mv /etc/X11/xorg.conf /etc/X11/xorg.conf-backup". This will rename the xorg.conf file (if you have it), and let X run with default configuration (i.e. without an xorg.conf file) after a reboot.
Evey time that I create a xorg.conf file I cannot start the graphics. The only way was to create a xorg.conf.d directory (see attached file) Is it because I am running nouveau and not the nvidia package?
(3) Boot into the new kernel. Since you have just installed it, and since there is no xorg.conf, the nouveau driver should be activated. Maybe it will have problems running on your graphics card (it shouldn't, but...), and maybe you will not have high resolution, or even X might not start at all. That is ok, it doesn't matter, as long as you do the next step.
(4) Run "yum install akmod-nvidia". Let it install. It will provide the nvidia driver matching your new kernel, it will write a new xorg.conf, it will blacklist nouveau and remove it from initrd of your kernel. It might take some time to complete.
It is already installed, should I keep it or remove it?
Thank for your help.
(5) Reboot the machine again into the latest kernel. This is necessary in order to change from nouveau to nvidia driver.
If all is well, the nvidia driver should be loaded and X should work perfectly, high resolution and all. If it doesn't work, post the contents of /var/log/Xorg.0.log file back to us, so that we can see why it doesn't work and tell you how to fix it.
I hope this was clear enough. Just follow the steps (1) - (5), and it should start working again.
Note that once you have installed akmod-nvidia, in the future when a new kernel gets installed with an update, everything will happen automatically and you do not need to do anything in order to get the nvidia driver for that new kernel. The above (1)-(5) procedure should be done only once, and after that it will work automatically for every new kernel.
HTH, :-) Marko