Home | Docs | Forums | Lists | Bugs | Planet | Store | GMN | Get Gentoo!
Not eligible to see or edit group visibility for this bug.
View Bug Activity | Format For Printing | XML | Clone This Bug
Nvidia just released a new version of their proprietary drivers. Even if it's a great version jump, renaming nvidia-drivers-100.14.23.ebuild worked fine for me and it's up and running. It compiles against zen-sources (which is latest kernel-git + additional patches and on 2.6.22 (At least an user said that here: http://forums.gentoo.org/viewtopic-t-617102.html ) so it should run fine on 2.6.23 too. Reproducible: Always Steps to Reproduce:
I want to remark that although it is a beta release, it has some significance because this is the first and currently only driver to support the 8800GT chipset (which some fools, like yours truly, bought in the belief that "8800 series" drivers would support it). I've yet to try it myself (will hopefully have the time when I get home tonight); in any case, I for one would like to see it (masked) in portage, beta or not.
Even more. It indeed fixes what is broken in the current "~stable" driver: xv overlay redirection. It's the first release since several which allows using tvtime with compiz and actually displays a picture. So beta or not, it's a real improvement against current 100.14.23, besides the support for the new card.
Works on a Core2 Duo/8800GT system. nvidia-drivers beta + Compiz Fusion is better than X on VESA, that's for sure...
A simple version bump gives you graphics, but it does not install the CUDA libraries that are now part of the driver. usr/include/cuda/*.h needs to be installed as well as usr/lib*/libcuda.so.${PV} I added # handle CUDA headers dodir /usr/include/cuda insinto /usr/include/cuda doins usr/include/cuda/*.h and # handle CUDA libs dolib.so ${usrpkglibdir}/libcuda.so.${PV} dosym libcuda.so.${PV} /usr/$(get_libdir)/libcuda.so.1 dosym libcuda.so.1 /usr/$(get_libdir)/libcuda.so to src_install-libs() in order to handle these.
What's holding this bug down?
This Driver should be added to portage as a hardmasked ebuild because of its beta status but nonethless should be added.
Created an attachment (id=137591) [details] Proposed 169.04 ebuild This is the 100.14.23 ebuild, with the cuda patches added in (including lines for textrels in libcuda). Works great for me on amd64
Just to add fuel to the fire for inclusion in portage proper, check this article out regarding XRender improvements. Its really an incredible performance gain in this area, not limited to 8000 series GPUs, but others as well. http://www.phoronix.com/scan.php?page=article&item=934&num=1 Regarding the CUDA libs, since these libs only work on 8000 series GPUs, is is wise or necessary to install them on earlier models? Many thanks, Summers
I don't think the CUDA libs will cause a problem on older cards. When CUDA initializes, it scans for available 8000 series cards and reports an error if it does not find one. Just to be sure, I will try installing these drivers with my 6000 series card later today and post again if I discover any issues.
I'm running amd64 with 2 6800 XT's on gentoo-sources 2.6.23-r3 I've been running this driver for a couple of weeks with only minor issues and overall performance and stability improvements that more than make up for the few issue's at hand (which honestly I'm not entirely sure are related to this driver anyway) I modified the 100.14.23 ebuild and added to overlay. I did not add the CUDA modifications so I can't speak as to how they work against non 8000 series. I agree this needs to be officially in portage but masked.
Everyone can agree until they're blue in the face. I do not add beta drivers to the tree. Cut and dry.
There is a problem with the ebuilds implementation of this driver, I have tried this ebuild and there are two problems specifically for gentoo. First, if too many konqueror windows get opened, x crashes and restarts. Second, if you play a full screen game like apricots, the game no longer covers the entire screen. I have this driver installed through the nvidia installer on a debian box running a gentoo sources kernel (because I like the uvesafb splash). Neither of these two problems occur on that box.
Does not compile for me on mm-sources-2.6.24_rc5-r1 Same failure with the same latest stable driver on mm-sources-2.6.24_rc5-r1 Report files will be added later.
mm-sources typically have a different kernel API then the Linus kernel so you won't ever find support for mm-sources. Additionally, non-released kernels are also unsupported.
Confirmed working on 2.6.23-gentoo-r3, however I created my own ebuild which happens to be the same as Brendan's.
169.07 (stable) has been released: http://www.nvidia.com/object/linux_display_amd64_169.07.html
*** Bug 202908 has been marked as a duplicate of this bug. ***
Adding myself to this bug, to see when 169.07 is put in.
Okay, having tested this driver with the attached 169.04 ebuild renamed to 169.07, I get continual xcrashes which require reboot, Repeatably, if you open a root konqueror window and go to the highest directory "/", click in the k icon in the top right to open a second window, it crashes it every time.
Doug, seems someone changed this original bug request for the 169.04 beta driver into a request for 169.07. 169.07 is stable and should show up us a an ebuild.
Just wanted to mention that the driver installed on Debian with the nvidia installer works very well and seems to be a big improvement.
(In reply to comment #20) > Doug, seems someone changed this original bug request for the 169.04 beta > driver into a request for 169.07. > > 169.07 is stable and should show up us a an ebuild. > I made the change.
(In reply to comment #19) > Okay, having tested this driver with the attached 169.04 ebuild > renamed to 169.07, I get continual xcrashes which require reboot, > Repeatably, if you open a root konqueror window and go to the > highest directory "/", click in the k icon in the top right to > open a second window, it crashes it every time. > Have you submitted a bug report to nvidia?
(In reply to comment #19) > Okay, having tested this driver with the attached 169.04 ebuild > renamed to 169.07, I get continual xcrashes which require reboot, > Repeatably, if you open a root konqueror window and go to the > highest directory "/", click in the k icon in the top right to > open a second window, it crashes it every time. > I tried to reproduce this bug but for me it works fine. I guess this depends on the graphic card.
Doug, I have not submitted a bug report to nvidia, reason being is that the driver works and installs perfectly on my other box that runs Debian. I believe there has to be a problem with the ebuilds implementation of the driver in Gentoo. I don't think its nvidia's problem. here is my emerge info Portage 2.1.4_rc11 (default-linux/amd64/2007.0/desktop, gcc-4.2.2, glibc-2.7-r1, 2.6.23-gentoo-r5 x86_64) ================================================================= System uname: 2.6.23-gentoo-r5 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 4200+ Timestamp of tree: Sat, 22 Dec 2007 06:30:01 +0000 app-shells/bash: 3.2_p17-r1 dev-java/java-config: 1.3.7, 2.1.3 dev-lang/python: 2.5.1-r4 dev-python/pycrypto: 2.0.1-r6 sys-apps/baselayout: 1.12.10-r5 sys-apps/sandbox: 1.2.18.1-r2 sys-devel/autoconf: 2.13, 2.61-r1 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10 sys-devel/binutils: 2.18-r1 sys-devel/gcc-config: 1.4.0-r4 sys-devel/libtool: 1.5.24 virtual/os-headers: 2.6.23-r2 ACCEPT_KEYWORDS="amd64 ~amd64" CBUILD="x86_64-pc-linux-gnu" CFLAGS="-march=native -O2 -pipe" CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/share/config" CONFIG_PROTECT_MASK="/etc/env.d /etc/env.d/java/ /etc/fonts/fonts.conf /etc/gconf /etc/revdep-rebuild /etc/splash /etc/terminfo /etc/texmf/web2c /etc/udev/rules.d" CXXFLAGS="-march=native -O2 -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="buildpkg distlocks fixpackages metadata-transfer parallel-fetch sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="http://distfiles.gentoo.org http://distro.ibiblio.org/pub/linux/distributions/gentoo" LANG="en_US.UTF-8" LC_ALL="en_US.UTF-8" MAKEOPTS="-j3" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --delete-after --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages --filter=H_**/files/digest-*" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage/miscellaneous" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="3dnow 3dnowtext S3TC X a52 aac aalib accessibility acl acpi aiglx alsa amarok amd64 apm arts artworkextra asf audiofile automount avahi avi bash-completion berkdb bitmap-fonts bittorrent branding bzip2 cairo calendar cann cdparanoia cdr cli cpudetection cracklib crypt cups curl dbus deprecated devil dga directfb discouraged dlloader doc dri dts dv dvb dvd dvdr dvdread eds effects elisp emacs emboss emerald encode esd evo examples exif ext-man ext-rtf ext-tex extraicons fam fat fax fbcon fbcondecor fbsplash ffmpeg firefox flac flash foomaticdb fortran freetype fuse gb gcj gd gdbm gdphoto2 gedit gif gimp gimpprint glitz gnome gnutls gphoto2 gpm gstreamer gtk gtk2 gtkhtml hal hddtemp hfs howl html iconv idn ieee1394 imagemagick imlib inkjar ipv6 isdnlog jack java jfs joystick jpeg kde kdehiddenvisibility kerberos lcms ldap libao libcaca libg++ lirc lm_sensors logrotate mad mail midi mikmod mime mjpeg mmx mmxext mng mono mozbranding mozcalendar mozilla mp3 mpeg mplayer mudflap mysql nas nautilus ncurses networking new-login newspr nforce2 nls nptl nptlonly nsplugin ntfs nvidia odbc offensive ogg opengl openmp oss pam parport pcre pdf pdflib perl pmu png postgres ppds pppd python qt3 qt3support qt4 quicktime rar rdesktop readline reflection regex reiser4 reiserfs samba scanner sdl seamonkey session skins slp smp sndfile snmp sound sounds speex spell spl spreadsheet sqlite sse sse2 ssl startup-notification streams subversion svg swat symlink sysfs syslog tcltk tcpd tetex theora thesaurus threads thunderbird tidy tiff timidity tk truetype truetype-fonts type1 type1-fonts udev unicode usb utempter utils v4l vcd vdr voice vorbis vram weather wmf wordperfect xattr xcb xcomposite xforms xfs xine xml xorg xpm xprint xscreensaver xv xvid zeroconf zlib" ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci emu10k1x ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0 intel8x0m maestro3 trident usb-audio via82xx via82xx-modem ymfpci" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null plug rate route share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" ELIBC="glibc" INPUT_DEVICES="keyboard mouse evdev" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" USERLAND="GNU" VIDEO_CARDS="apm ark chips cirrus cyrix dummy fbdev glint i128 i810 mach64 mga neomagic nv r128 radeon rendition s3 s3virge savage siliconmotion sis sisusb tdfx tga trident tseng v4l vesa vga via vmware voodoo" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LDFLAGS, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS My xorg.conf Section "ServerLayout" Identifier "X.org Configured" Screen 0 "Screen0" 0 0 InputDevice "Mouse0" "CorePointer" InputDevice "Keyboard0" "CoreKeyboard" EndSection Section "Files" FontPath "/usr/share/fonts/100dpi" FontPath "/usr/share/fonts/75dpi" FontPath "/usr/share/fonts/artwiz" FontPath "/usr/share/fonts/corefonts" FontPath "/usr/share/fonts/default/ghostscript" FontPath "/usr/share/fonts/liberation-fonts-ttf" FontPath "/usr/share/fonts/misc" FontPath "/usr/share/fonts/ttf-bitstream-vera" FontPath "/usr/share/fonts/Type1" FontPath "/usr/share/fonts/urw-fonts" FontPath "/usr/local/share/fonts" EndSection Section "Module" Load "glx" Load "extmod" Load "wfb" Load "xtrap" Load "vbe" Load "record" Load "dbe" Load "freetype" Load "type1" EndSection Section "ServerFlags" Option "AllowMouseOpenFail" "True" Option "Xinerama" "0" EndSection Section "InputDevice" Identifier "Keyboard0" Driver "kbd" Option "CoreKeyboard" Option "XkbRules" "xorg" Option "XkbModel" "pc104" Option "XkbOptions" "grp:toggle,grp_led:scroll" Option "XkbVariant" ",winkeys" EndSection Section "InputDevice" Identifier "Mouse0" Driver "mouse" Option "Protocol" "auto" Option "Device" "/dev/input/mice" Option "ZAxisMapping" "4 5 6 7" EndSection Section "Monitor" Identifier "Monitor0" VendorName "IBM" ModelName "IBM C190 CRT" HorizSync 30.0 - 96.0 VertRefresh 50.0 - 160.0 Option "DPMS" "False" EndSection Section "Device" Identifier "Videocard0" Driver "nvidia" VendorName "NVIDIA Corporation" BoardName "GeForce 8800 GTS" BusID "PCI:3:0:0" #Option "CustomEDID" "DFP-0:/etc/X11/edid.bin" Option "UseEDID" "False" Option "DPI" "72 x 72" Option "NoLogo" "True" Option "Coolbits" "1" Option "BackingStore" "True" EndSection Section "Screen" Identifier "Screen0" Device "Videocard0" Monitor "Monitor0" DefaultDepth 24 Option "TwinView" "0" Option "TwinViewXineramaInfoOrder" "CRT-0" Option "metamodes" "1024x768_85 +0+0; 1024x768 +0+0; 800x600 +0+0; 640x480 +0+0" SubSection "Display" Depth 24 EndSubSection EndSection Section "Extensions" Option "Composite" "Enable" EndSection My Nvidia Settings # # /home/billydv/.nvidia-settings-rc # # Configuration file for nvidia-settings - the NVIDIA X Server Settings utility # Generated on Wed Dec 19 20:24:01 2007 # # ConfigProperties: RcFileLocale = C ToolTips = Yes DisplayStatusBar = Yes SliderTextEntries = Yes IncludeDisplayNameInConfigFile = No ShowQuitDialog = No # Attributes: 0/DigitalVibrance[CRT-0]=0 0/SyncToVBlank=0 0/AllowFlipping=1 0/LogAniso=0 0/FSAA=0 0/TextureSharpen=0 0/ForceGenericCpu=0 0/GammaCorrectedAALines=0 0/CursorShadow=0 0/CursorShadowXOffset=4 0/CursorShadowYOffset=2 0/CursorShadowAlpha=64 0/CursorShadowRed=0 0/CursorShadowGreen=0 0/CursorShadowBlue=0 0/FSAAAppControlled=1 0/LogAnisoAppControlled=1 0/FSAAAppEnhanced=0 0/RedBrightness=0.000000 0/GreenBrightness=0.000000 0/BlueBrightness=0.000000 0/RedContrast=0.000000 0/GreenContrast=0.000000 0/BlueContrast=0.000000 0/RedGamma=1.000000 0/GreenGamma=1.000000 0/BlueGamma=1.000000 0/OpenGLImageSettings=1 0/XVideoTextureSyncToVBlank=0 0/XVideoSyncToDisplay=1
If anyone notices something wrong with any of my configs that may be causing the problems please let me know
PLEASE DISREGARD MY POSTS, the fact that noone else could reproduce my problem got me to testing today. My problem was a specific emerald theme. It was causing me alot of problems, probably more than just this driver problem. The current ebuild works fine renamed to 169.07. It should go into portage immediately.
The only thing I find that changes when using this driver and it only happens on an 8800gts, not with 2 other 7800gs's, is that when I play a game such as apricots, the game does not cover the entire screen but seems almost as if the screen resolution has changed
I took the attached ebuild and renamed it to 169.07 for my 8800 GTS (G92). Works. But as soon as the X server fires up the fan spins up at full speed which is very annoying... Nobody here having this problem?
The installation doesn't work for me: it ends with this error: >>> Source compiled. >>> Test phase [not enabled]: x11-drivers/nvidia-drivers-169.07 >>> Install nvidia-drivers-169.07 into /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/image/ category x11-drivers * Installing nvidia module * * ERROR: x11-drivers/nvidia-drivers-169.07 failed. * Call stack: * ebuild.sh, line 1701: Called dyn_install * ebuild.sh, line 1138: Called qa_call 'src_install' * ebuild.sh, line 44: Called src_install * nvidia-drivers-169.07.ebuild, line 233: Called die * The specific snippet of code: * [ -f "${FILESDIR}/nvidia" ] || die "nvidia missing in FILESDIR" * The die message: * nvidia missing in FILESDIR * * If you need support, post the topmost build error, and the call stack if relevant. * A complete build log is located at '/var/tmp/portage/x11-drivers/nvidia-drivers-169.07/temp/build.log'. * This ebuild is from an overlay: '/usr/local/portage/' * Does anybody have an idea to solve the problem ? Thanks for your help.
(In reply to comment #29) > I took the attached ebuild and renamed it to 169.07 for my 8800 GTS (G92). > Works. But as soon as the X server fires up the fan spins up at full speed > which is very annoying... > > Nobody here having this problem? > I have GeForce 8800 GTS (G80) and I have the same "problem".
(In reply to comment #30) > The installation doesn't work for me: it ends with this error: > > >>> Source compiled. > >>> Test phase [not enabled]: x11-drivers/nvidia-drivers-169.07 > > >>> Install nvidia-drivers-169.07 into /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/image/ category x11-drivers > * Installing nvidia module > * > * ERROR: x11-drivers/nvidia-drivers-169.07 failed. > * Call stack: > * ebuild.sh, line 1701: Called dyn_install > * ebuild.sh, line 1138: Called qa_call 'src_install' > * ebuild.sh, line 44: Called src_install > * nvidia-drivers-169.07.ebuild, line 233: Called die > * The specific snippet of code: > * [ -f "${FILESDIR}/nvidia" ] || die "nvidia missing in FILESDIR" > * The die message: > * nvidia missing in FILESDIR > * > * If you need support, post the topmost build error, and the call stack if > relevant. > * A complete build log is located at > '/var/tmp/portage/x11-drivers/nvidia-drivers-169.07/temp/build.log'. > * This ebuild is from an overlay: '/usr/local/portage/' > * > > Does anybody have an idea to solve the problem ? > Thanks for your help. > /nvidia-drivers/files/ You must copy files 09nvidia, libGL.la-r2 and nvidia from /usr/portage/x11-drivers/nvidia-drivers/files/ to the same directory in your overlay (usr/local/portage/x11-drivers/nvidia-drivers/files/).
Piotrek, the way you do this is you copy /usr/portage/x11-drivers/nvidia-drivers and place it in your own overlay in a folder marked x11-drivers. Delete everything in the nvidia-drivers folder except files, changelog, and metadata. All ebuilds should be deleted along with the manifest. Place the ebuild from this bug in the folder and replace the .04 with.07 . Then you must do ebuild /usr/local/portage/miscellaneous/x11-drivers/nvidia-drivers/nvidia-drivers-169.07.ebuild manifest replace the location with wherever you have your local overlay this will manifest the ebuild and then you can emerge it normally.
while working at adding this, please also have a look at bug #202978. The 169.07 driver also has this extra file.
I have created an ebuild for the updated nvidia-settings. It can be found here http://bugs.gentoo.org/show_bug.cgi?id=203267
Regarding this ebuild (both 169.04 and 169.07), i have an odd observation. Whenever composite is being used in any form (compiz, kompmgr, ...), Qt windows that are offscreen seem to not redraw correctly. Steps to reproduce: 1) Run X with composite enabled, desktop KDE, start a compositing manager (fe. compiz) 2) Start up konsole 3) Move konsole below the bottom of the screen (so that only first line of the terminal is visible, and rest is offscreen) 4) Type out various commands producing as much output as possible (fe. cat /var/log/messages) 5) Move konsole back onscreen, the text is drawn over itself without erasing the previous contents, producing a horrible mess. - This applies only for Qt(KDE) applications, i wasn't able to reproduce it for any similar gtk apps. - Disabling RenderAccel or generally fiddling with driver options does not seem to help anything. Disabling RenderAccel only makes the problem a bit less serious, because damage events seem to be sent to the whole window whenever something happens (So in the above steps, in 3 you need to move Konsole completely offscreen and type the commands blind), but the bug persists - Reproduced on 3 individual Gentoo boxes, each running different nvidia graphics card (7700Go, 8800GTS, quadro 140M) - Reproduced on these exact applications: kicker - when it is autohide, and you remove some tray icons when it's hidden; amarok - amarok offscreen will write the "just playing" bottom line over itself, producing a black line The way to fix this is downgrading back to 100.* drivers, or not use Composite. I am not sure whether this is a problem specific to this Gentoo ebuild, or general Nvidia problem, and i have no way to check. But considering the bug was pretty obvious in my case, and no other users have complained on nv forums yet ....
A person who's testing Second Life's Windlight client also has the same problem on a SUSE 10 install, so it's definitely the driver. Bug NVidia. This driver's best bet is hardmasked.
(In reply to comment #37) > A person who's testing Second Life's Windlight client also has the same problem > on a SUSE 10 install, so it's definitely the driver. Thanks for verification. Just posted http://www.nvnews.net/vbulletin/showthread.php?t=105174 > Bug NVidia. This driver's best bet is hardmasked. Agreed. The 100% fan bug which is another blocker has been already reported on nvnews, acknowledged, fixed, and according to zander will be released in next driver release ( http://www.nvnews.net/vbulletin/showpost.php?p=1494827&postcount=34 ). So this one will break at least 2 things that have been working already. A reason to add it hard masked is for people who hurried buying the new nvidia cards (G92).
> Regarding this ebuild (both 169.04 and 169.07), i have an odd observation. > Whenever composite is being used in any form (compiz, kompmgr, ...), Qt > windows > that are offscreen seem to not redraw correctly. > Steps to reproduce: > 1) Run X with composite enabled, desktop KDE, start a compositing manager (fe. > compiz) > 2) Start up konsole > 3) Move konsole below the bottom of the screen (so that only first line of the > terminal is visible, and rest is offscreen) > 4) Type out various commands producing as much output as possible (fe. cat > /var/log/messages) > 5) Move konsole back onscreen, the text is drawn over itself without erasing > the previous contents, producing a horrible mess. I just tried this using KDE4. And there is no bug visible concerning redrawing. So I fired up a kde3-konsole - with the bug described above... > - This applies only for Qt(KDE) applications, i wasn't able to reproduce it > for > any similar gtk apps. So this is not true for any Qt-version but only <qt-4*. With qt-3 I had some more redrawing-errors on amd64, e.g. in designer (qt-3!) when dragging the widget to design around, the whole content gets really bad messed up! This is also gone with qt-4.
(In reply to comment #36) > Regarding this ebuild (both 169.04 and 169.07), i have an odd observation. > Whenever composite is being used in any form (compiz, kompmgr, ...), Qt windows > that are offscreen seem to not redraw correctly. > > Steps to reproduce: > 1) Run X with composite enabled, desktop KDE, start a compositing manager (fe. > compiz) > 2) Start up konsole > 3) Move konsole below the bottom of the screen (so that only first line of the > terminal is visible, and rest is offscreen) > 4) Type out various commands producing as much output as possible (fe. cat > /var/log/messages) > 5) Move konsole back onscreen, the text is drawn over itself without erasing > the previous contents, producing a horrible mess. > > > - This applies only for Qt(KDE) applications, i wasn't able to reproduce it for > any similar gtk apps. > - Disabling RenderAccel or generally fiddling with driver options does not seem > to help anything. Disabling RenderAccel only makes the problem a bit less > serious, because damage events seem to be sent to the whole window whenever > something happens (So in the above steps, in 3 you need to move Konsole > completely offscreen and type the commands blind), but the bug persists > - Reproduced on 3 individual Gentoo boxes, each running different nvidia > graphics card (7700Go, 8800GTS, quadro 140M) > - Reproduced on these exact applications: kicker - when it is autohide, and you > remove some tray icons when it's hidden; amarok - amarok offscreen > will write the "just playing" bottom line over itself, producing a black line > > The way to fix this is downgrading back to 100.* drivers, or not use Composite. > > I am not sure whether this is a problem specific to this Gentoo ebuild, or > general Nvidia problem, and i have no way to check. But considering the bug was > pretty obvious in my case, and no other users have complained on nv forums yet > .... > I was experiencing the same issue with konsole not refreshing correctly. Adding the following option to the xorg.conf Device section fixed it for me: Option "UseCompositeWrapper" "true"
(In reply to comment #40) > I was experiencing the same issue with konsole not refreshing correctly. Adding > the following option to the xorg.conf Device section fixed it for me: > > Option "UseCompositeWrapper" "true" > Works for me! Looks like i overlooked one X option afterall.
Once again, I thought that changing my window decorator or the theme itself had something to do with the previous problems I had been having but no, the implementation of the driver through the ebuild must be causing some problem because other boxes with other distros do not have the same problems. This driver needs a proper ebuild created from scratch. Possible some of the patches are no longer necessary, or maybe the eselect module needs some changes.
(In reply to comment #42) > Once again, I thought that changing my window decorator or the theme itself had > something to do with the previous problems I had been having but no, the > implementation of the driver through the ebuild must be causing some problem > because other boxes with other distros do not have the same problems. This > driver needs a proper ebuild created from scratch. Possible some of the patches > are no longer necessary, or maybe the eselect module needs some changes. > At this point in time I've seen nothing wrong with the driver ebuild. It installs everything in the same fashion. It's used by thousands of people daily and it's even used by NVIDIA themselves on their Gentoo boxes. Based on the volume of reports on this bug and the fact that you're the only one experiencing this issue, I'm leaning towards a configuration bug. Whether it be improper use of eselect opengl, or a badly crafted xorg.conf, I can't tell you.
Well, I guess I will give it another try. I did find something else out last night with regards to my box, superkaramba + compiz-fusion + compiz animations plugin enabled cause problems. It was the reason why I couldn't get a login box after screensaver had run. I will now try it without superkaramba.
I dont know what it could be but the problem definitely exists. Here is my xorg.conf, if you see something wrong, please tell me. Section "ServerLayout" Identifier "X.org Configured" Screen 0 "Screen0" 0 0 InputDevice "Mouse0" "CorePointer" InputDevice "Keyboard0" "CoreKeyboard" EndSection Section "Files" FontPath "/usr/share/fonts/100dpi" FontPath "/usr/share/fonts/75dpi" FontPath "/usr/share/fonts/artwiz" FontPath "/usr/share/fonts/corefonts" FontPath "/usr/share/fonts/default/ghostscript" FontPath "/usr/share/fonts/dejavu" FontPath "/usr/share/fonts/liberation-fonts-ttf" FontPath "/usr/share/fonts/misc" FontPath "/usr/share/fonts/ttf-bitstream-vera" FontPath "/usr/share/fonts/Type1" FontPath "/usr/share/fonts/urw-fonts" FontPath "/usr/local/share/fonts" EndSection Section "Module" Load "glx" Load "extmod" Load "wfb" Load "xtrap" Load "vbe" Load "record" Load "dbe" Load "freetype" Load "type1" EndSection Section "ServerFlags" Option "AllowMouseOpenFail" "True" Option "Xinerama" "0" EndSection Section "InputDevice" Identifier "Keyboard0" Driver "kbd" Option "CoreKeyboard" Option "XkbRules" "xorg" Option "XkbModel" "pc104" Option "XkbOptions" "grp:toggle,grp_led:scroll" Option "XkbVariant" ",winkeys" EndSection Section "InputDevice" Identifier "Mouse0" Driver "mouse" Option "Protocol" "auto" Option "Device" "/dev/input/mice" Option "ZAxisMapping" "4 5 6 7" EndSection Section "Monitor" Identifier "Monitor0" VendorName "IBM" ModelName "IBM C190 CRT" HorizSync 30.0 - 96.0 VertRefresh 50.0 - 160.0 EndSection Section "Device" Identifier "Videocard0" Driver "nvidia" VendorName "NVIDIA Corporation" BoardName "GeForce 8800 GTS" Option "UseEDID" "False" Option "DPI" "72 x 72" Option "NoLogo" "True" Option "Coolbits" "1" Option "BackingStore" "True" #Option "TrippleBuffer" "True" EndSection Section "Screen" Identifier "Screen0" Device "Videocard0" Monitor "Monitor0" DefaultDepth 24 Option "TwinView" "0" Option "TwinViewXineramaInfoOrder" "CRT-0" Option "metamodes" "1024x768_85 +0+0; 1024x768 +0+0; 800x600 +0+0; 640x480 +0+0" SubSection "Display" Depth 24 EndSubSection EndSection Section "Extensions" Option "Composite" "Enable" EndSection
Created an attachment (id=139975) [details] xsession errors
Unfortunately xorg log doesnt show anything interesting about the crashes, in fact it doesnt show anything at all about the crash, its as if they have never happened
(In reply to comment #47) > Unfortunately xorg log doesnt show anything interesting about the crashes, in > fact it doesnt show anything at all about the crash, its as if they have never > happened > What does dmesg show immediately after the crash?
Sometimes a piece of (seemingly) identical hardware or combo thereof will exhibit errors that just don't crop up elsewhere. I'm not saying stop looking, but keep this in your mind as a possibility as you drive yourself crazy poking and prodding at your black box that is proprietary nvidia drivers. It is probably not the distro.
I cant get to read dmesg because I am forced to reboot from a totally blank screen. What I have tried just now is completely eliminating font and module sections from xorg.conf which seems to work, I have just opened 20-30 su konqueror windows without a freezeup. Is there a problem with just letting the default compiled in font paths and default modules load?
Okay, I found the error in messages, it shows me opening an su konqueror window and right after x crashing, heres the message Jan 3 17:56:51 Windows2 X[9440]: segfault at 00000000000001b8 rip 000000000053faf3 rsp 00007fffa6bb8e50 error 4 Jan 3 17:56:51 Windows2 NVRM: Xid (0003:00): 13, 0003 00000000 00005097 00000104 00000000 00000140 Jan 3 17:56:51 Windows2 NVRM: Xid (0003:00): 13, 0003 00000000 00005097 00000104 00000000 00000100 Jan 3 17:56:51 Windows2 kdm[7395]: X server for display :0 terminated unexpectedly This is on an amd64 but the very same thing happens on an x86 with an old socket a amd. Any Ideas?
Well now at least you know the crash is happening in X and not in an associated application. Your only chance is to get a backtrace from X when it crashes. You will probably need to then use the NVIDIA bug reporting script to gather additional useful information and forward both the backtrace and the data from the report tool to NVIDIA themselves. That being said, comment #49 is very on par for this particular issue.
Doug, without the font paths or modules specified in my xorg.conf, it is crashing only after 20 duplicate windows (in su konqueror, I hit the K icon roughly 20 times to keep producing new windows). Is it ok to leave these unspecified? For the modules it doesn't seem like a big deal because just about all the modules still get loaded, just in a different order. But what about the font paths? If I don't specify them what happens? What do I lose out on?
@Billy DeVincentis No offense intended, but it sounds like you are running a rather unstable configuration and you are cluttering up this version bump request. Can you open a different bug report instead please? Also, I'd like to add that the 169.07 driver includes powermizer support, another reason to get this ebuild into portage quickly.
I can confirm that the UseCompositeWrapper line stablizes KDE 3.x and FVWM, according to word on the Second Life Linux forum. This is on a SuSE 10 system, though. Take it for what it's worth.
I got a new dell laptop (precision M6300) and it only works with >=nvidia-drivers-169.04 . Previous versions of the nvidia drivers are leading to a complete lookup of the machine. Nvidia says that there are some problems with multi-core CPU's and newer GPU's. CPU: Intel(R) Core(TM)2 Duo CPU T7300 @ 2.00GHz GPU: Quadro FX 1600M So this driver should go to portage soon.
(In reply to comment #1) > I want to remark that although it is a beta release, it has some significance > because this is the first and currently only driver to support the 8800GT > chipset (which some fools, like yours truly, bought in the belief that "8800 > series" drivers would support it). I've yet to try it myself (will hopefully > have the time when I get home tonight); in any case, I for one would like to > see it (masked) in portage, beta or not. > I just purchased an nvidia 8800 GT and I can't get the x-server to work. I am running the 2.6.23 kernel. I have emerged both the 100.14.23 and the 100.14.19 drivers and still can't get it to work. When I run xorgcfg I get the video device not found. When I try to do all the /etc/init.d/xdm restart or even stop and start the server appears to start then goes back to the console with the move line 15 to a random number. I would really like to get this to work and I am beginning to become annoyed that there is this much problem with an nvidia release. I even went back to the 2.6.22 Kernel and still the same problems with both nvidia-drivers.
(In reply to comment #57) > (In reply to comment #1) > > I want to remark that although it is a beta release, it has some significance > > because this is the first and currently only driver to support the 8800GT > > chipset (which some fools, like yours truly, bought in the belief that "8800 > > series" drivers would support it). I've yet to try it myself (will hopefully > > have the time when I get home tonight); in any case, I for one would like to > > see it (masked) in portage, beta or not. > > > > I just purchased an nvidia 8800 GT and I can't get the x-server to work. I am > running the 2.6.23 kernel. I have emerged both the 100.14.23 and the 100.14.19 > drivers and still can't get it to work. When I run xorgcfg I get the video > device not found. When I try to do all the /etc/init.d/xdm restart or even > stop and start the server appears to start then goes back to the console with > the move line 15 to a random number. I would really like to get this to work > and I am beginning to become annoyed that there is this much problem with an > nvidia release. I even went back to the 2.6.22 Kernel and still the same > problems with both nvidia-drivers. > These drivers are too old and don't support the 8800 GT. You will need the 169.07 drivers which are currently not in portage tree.
Tim Boundy, no offense intended, I am running a testing system, it has nothing to do with why x crashes, it also crashes on a debian system, if you look at the nvidia boards you will see that although the new drivers support the new 8800gt cards, they are full of all kinds of problems, not just the fan thing, and although I think they should go into portage, I believe it should as a hardmasked ebuild. I believe the next driver nvidia comes up with will probably be better as they will have had time to fix the new problems.
(In reply to comment #59) As for the unstable configuration, I was specifically referring to compiz. Does your system still crash when you have compositing effects turned off? If not, then I suggest opening a new bug report. If it still crashes, then this bug report is as good a place as any.
The new ebuild from the tree doesn't work for me :S Gentoo-Tux julroy # emerge -avu nvidia-drivers These are the packages that would be merged, in order: Calculating dependencies... done! [ebuild U ] x11-drivers/nvidia-drivers-169.07 [100.14.19] USE="acpi gtk (-multilib)" 0 kB Total: 1 package (1 upgrade), Size of downloads: 0 kB Would you like to merge these packages? [Yes/No] >>> Verifying ebuild Manifests... >>> Emerging (1 of 1) x11-drivers/nvidia-drivers-169.07 to / * NVIDIA-Linux-x86_64-169.07-pkg2.run MD5 RMD160 SHA1 SHA256 size ;-) ... [ ok ] * checking ebuild checksums ;-) ... [ ok ] * checking auxfile checksums ;-) ... [ ok ] * checking miscfile checksums ;-) ... [ ok ] * checking NVIDIA-Linux-x86_64-169.07-pkg2.run ;-) ... [ ok ] /usr/portage/x11-drivers/nvidia-drivers/nvidia-drivers-169.07.ebuild: line 122: kernel_linux: command not found >>> Unpacking source... /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/temp/environment: line 3735: kernel_linux: command not found Creating directory NVIDIA-Linux-x86_64-169.07-pkg2 Verifying archive integrity... OK Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86_64 169.07............................................................................................................................................ * Applying NVIDIA_glx-defines.patch ... [ ok ] * Applying NVIDIA_glx-glheader.patch ... [ ok ] * Applying NVIDIA_i2c-hwmon.patch ... [ ok ] * Determining the location of the kernel source code * Found kernel source directory: * /usr/src/linux * Found kernel object directory: * /lib/modules/2.6.23-tuxonice-r6/build * Found sources for kernel version: * 2.6.23-tuxonice-r6 * Converting Makefile.kbuild to use M= instead of SUBDIRS= ... [ ok ] >>> Source unpacked. >>> Compiling source in /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/work/NVIDIA-Linux-x86_64-169.07-pkg2/usr/src/nv ... /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/temp/environment: line 3607: kernel_linux: command not found >>> Source compiled. >>> Test phase [not enabled]: x11-drivers/nvidia-drivers-169.07 >>> Install nvidia-drivers-169.07 into /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/image/ category x11-drivers dodoc: /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/work/NVIDIA-Linux-x86_64-169.07-pkg2/usr/src/nv/usr/share/doc/XF86Config.sample does not exist dodoc: /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/work/NVIDIA-Linux-x86_64-169.07-pkg2/usr/src/nv/usr/share/doc/Copyrights does not exist !!! dobin: /var/tmp/portage/x11-drivers/nvidia-drivers-169.07/work/NVIDIA-Linux-x86_64-169.07-pkg2/usr/src/nv/usr/bin/nvidia-xconfig does not exist * * ERROR: x11-drivers/nvidia-drivers-169.07 failed. * Call stack: * ebuild.sh, line 46: Called src_install * environment, line 3662: Called die * The specific snippet of code: * dobin ${NV_EXEC}/nvidia-xconfig || diefunc "$FUNCNAME" "$LINENO" "$?"; * The die message: * (no error message) * * If you need support, post the topmost build error, and the call stack if relevant. * A complete build log is located at '/var/tmp/portage/x11-drivers/nvidia-drivers-169.07/temp/build.log'. * The ebuild environment file is located at '/var/tmp/portage/x11-drivers/nvidia-drivers-169.07/temp/environment'. *
I'm disinterested in messages or posts like the one above as long as the ebuild is masked that it's being completely rewritten. It was merely committed so I could access it from another location.
(In reply to comment #62) > I'm disinterested in messages or posts like the one above as long as the ebuild > is masked that it's being completely rewritten. > > It was merely committed so I could access it from another location. Then why don't you *hard*mask it? A lot of people out there want to use ~arch nvidia-drivers.
*** Bug 206184 has been marked as a duplicate of this bug. ***
(In reply to comment #63) > (In reply to comment #62) > > I'm disinterested in messages or posts like the one above as long as the ebuild > > is masked that it's being completely rewritten. > > > > It was merely committed so I could access it from another location. > > Then why don't you *hard*mask it? A lot of people out there want to use ~arch > nvidia-drivers. > # Doug Klima <cardoe@gentoo.org> (16 Jan 2007) # Started to rewrite the ebuild when it came out # but have not had time to finish. Best to commit # it masked so at least some people can mess with it =x11-drivers/nvidia-drivers-169.07 It is hardmasked.
(In reply to comment #65) > (In reply to comment #63) > > Then why don't you *hard*mask it? A lot of people out there want to use ~arch > > nvidia-drivers. > > > > It is hardmasked. You're correct. I did another emerge --sync now, and is it indeed hardmasked *now*. An hour ago it was only marked as unstable.
Created an attachment (id=141079) [details] Quickfix(see Comment(!)) It's a quick fix I wrote - tested on ~x86 and ~amd64 It will still need a fix for the kernel_header stuff It will break vterm (!) be careful
I need these driver and they were not hard masked when I synced, i didn't unmask it, wasn't not my fault ! And I have to go through official site to have my 8800 GT to work, so please make this ebuild, I and others need them.
version 169.09 has been released: - Fixed a problem causing the fan on some GPUs to always run at full speed. - Fixed a bug that caused the X driver to crash if the X.Org GLX extension module was loaded intead of NVIDIA’s. - Improved the X driver’s awareness of the current notebook docking status. - Fixed brightness control on HP Compaq notebooks. - Fixed a bug in the Linux/i2c algorithm driver implementation that prevented core transfer types from succeeding.
I just wanted to mention that the problem I was having with x crashing when opening duplicate windows is fixed in the 169.09 version. They mention in the release that it was fixed. Renaming the ebuild in portage tree works except that there is no freebsd version sojust download the 169.07 freebsd version and rename it and put it in /usr/portage/distfiles to manifest. So far this release seems to be a real improvement.
(In reply to comment #70) > I just wanted to mention that the problem I was having with x crashing when > opening duplicate windows is fixed in the 169.09 version. They mention in the > release that it was fixed. Renaming the ebuild in portage tree works except > that there is no freebsd version sojust download the 169.07 freebsd version and > rename it and put it in /usr/portage/distfiles to manifest. So far this release > seems to be a real improvement. > 169.09 ebuild is available in Portage and was in fact available several hours before you made this comment. The lack of a FreeBSD 169.09 version was already addressed in that ebuild. Closing this bug since 169.09 is available in Portage for x86 and amd64 while 169.07 is available for x86-fbsd.
