Hello, the new media-video/nvidia-kernel-1.0.6106 driver + glx part lets my system freeze therefore please mask it. I'm using 2.6.7 kernel but had the same problem after I switched back to the latest 2.4. It's always the same. Everything works fine but after a few seconds watching tv with motv (xawtv, and I heared kwintv too) will freeze my system. Nothing works, even the magic sysrq will not help and I cound not find anything in my syslog. I thing switching channels with the mouse speeds it up but this could be just by chance :) Sometimes while I'm switching tabs in firefox my mouse+kbd hangs but the system seems still to works. However tvtime works fine but my 500mhz cpu drops some frames :/ I'm using now nvidia-kernel-1.0.5336-r4 and nvidia-glx-1.0.5336-r2, kernel 2.6.7 and everything is fine. Reproducible: Always Steps to Reproduce: 1. emerge nvidia-kernel/glx 1.0.6106, xawtv-3.92 2. Host bridge: VIA Technologies, Inc. VT8363/8365 [KT133/KM133] (rev 03) + AMD tbird 3. start motv switch some channels Actual Results: system will freeze Expected Results: not freeze the whole system Portage 2.0.50-r9 (default-x86-1.4, gcc-3.3.4, glibc-2.3.4.20040619-r0, 2.6.7-hardened-r3) ================================================================= System uname: 2.6.7-hardened-r3 i686 AMD Athlon(tm) Processor Gentoo Base System version 1.5.1 Autoconf: sys-devel/autoconf-2.59-r4 Automake: sys-devel/automake-1.8.5-r1 ACCEPT_KEYWORDS="x86 ~x86" AUTOCLEAN="yes" CFLAGS="-march=athlon-tbird -O3 -pipe -fomit-frame-pointer" CHOST="i686-pc-linux-gnu" COMPILER="gcc3" CONFIG_PROTECT="/etc /opt/glftpd/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3.2/share/config /usr/kde/3/share/config /usr/share/config /usr/share/texmf/dvipdfm/config/ /usr/share/texmf/dvips/config/ /usr/share/texmf/tex/generic/config/ /usr/share/texmf/tex/platex/config/ /usr/share/texmf/xdvi/ /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-march=athlon-tbird -O2 -pipe -fomit-frame-pointer" DISTDIR="/usr/portage/distfiles" FEATURES="autoaddcvs ccache sandbox" GENTOO_MIRRORS="http://sunsite.cnlab-switch.ch/ftp/mirror/gentoo/ http://ftp.lug.ro/gentoo ftp://ftp.solnet.ch/mirror/Gentoo" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/portage.local" SYNC="rsync://rsync.de.gentoo.org/gentoo-portage" USE="3dnow X aalib alsa apm arts avi berkdb clamav crypt cups directfb dvd encode esd faad fbcon flash foomaticdb gdbm gif gtk gtkhtml guile imlib ipv6 java jpeg kde libg++ libwww mad mikmod mmx motif mpeg ncurses nls oggvorbis opengl oss pam pdflib perl png python qt quicktime readline samba sdl slang spell ssl tcltk tcpd tetex tiff truetype vim-with-x x86
Please test this with 6111 which just went into cvs. Report back on success or failures. Thanks.
system freezed with 6111, too
Try different values with the NvAGP option in xorg.conf Via chipsets are nortorious for nvidia problems...
Created attachment 38475 [details, diff] Patch to nvidia kernel package Hello, I have experienced frequent lockups (sometimes after only a few minutes) with the nvidia driver under X on a Abit KG7 (VIA chipset). The symptoms were that the mouse still moved but keyboard etc. were dead and top showed 99.8% CPU usage for X. Since I have patched the package (manual extract and build) with the attached file (which apparently was created by a developer at nvidia, but don't hold me liable for this), the machine has been running under x (with opengl screensaver active) for two days straight, while without it, it made it two hours tops. The patch works for both 6106 and 6111 (no failed hunks). Hope this is of use. /martin
martin that patch just went into cvs for 6111 Can I get some feedback off everyone as to how the new 6111 (ie with the enable_pci patch) works in regards to this bug... please :)
Further more I _cannot_ reproduce this with tvtime, xawtv, or (most importantly) motv. Using 6111 with the enable pci patch, and xorg-x11-6.8.0 (or 6.7.0-r2 for that matter)
Sebastian, please provide feedback
sorry for the late reply... nv_enable_pci.patch seems to be martin's patch and allready in portage. I tested nvidia-kernel-1.0.6111.ebuild the result was sad: starting motv and switching channels with my mouse resulted in a freeze withhin 5 seconds. Switching channels via lirc delayed the freeze: after like 7-8 minutes the machine did not response and I did nothing.
Hello, sorry that the patch is not working; apparently I was just unusually lucky, because my system is also freezing all the time now. For two days I have been trying to get to a stable configuration with kernels 2.6.7 and 2.6.8, but without any success. I believe something went sour between versions 2.6.5 and 2.6.7. For one thing, the newer drivers will not build with 2.6.5 and the older drivers won't build with 2.6.7+. All the suggested remedies (disable ACPI, disable APIC, mem=nopentium, load agpgart as module, do not use agpgart but nvagp, ping the box from another machine...) really did nothing except postpone freezing for a few minutes at the most. The best setup I could get was 2.6.5-r1, with drivers 4496-r3. That works like a charm, no hangs, no freezes. It's a bit sad though, not to be able to use the newest kernel. Another indication is that the kernel module size went from around 1.7M (2.6.5) to something like 4.8M (6111 on 2.6.8). Again I apologize for submitting a patch that didn't work; I hope the config I specified might work for you, too. Note: specifying a kernel source dir didn't work all the well either. I always reset /usr/src/linux to the tested kernel version and booted into that kernel before emerging nvidia-kernel and nvidia-glx. Cheers /martin
I seem to be having freeze problems too. I have tried the NvAgp options and using the AGP from both nvidia and the kernel. However I only see problems when using flightgear, and only after using it for at least 15 - 20 min. After that closing the program causes the computer to freeze with "caps lock" and "scroll lock" on the keyboard blinking. I also have a VIA KX 133 chipset. any ideas?
Hello again, I couldn't let this issue ride and fiddled a bit more. Additional info: With my KT266A (VIA) chipset the problem does occur only with AGP-Cards that do not match exactly my motherboards AGP (2x/4x) port i.e. what I mean by this: It does not work with a Ti4200 8x (4x/8x) even when explicitly specifying 4x in BIOS (no matter AGP Fast Write on/off) It does not work with a FX5700LE (2x/4x/8x). It works perfectly with 3 different Ti4200 (1x/2x/4x; 1 VTD,2 no frills). Apparently there are conditions when the different specifications of AGP 2.0/3.0 become a problem. Sometimes with the Ti4200 8x it would work on the first start of X and hang on the next (could be reproduced several times). Sometimes it would hang on the first try already. With the 3 Ti4200 it always worked (with/without pci_enable.patch, nvagp and agpgart). I tried all of the cards in windows as well; there were no hangs with the two troublesome cards, but plenty of artifact activity in normal mode and blank screens in directx. So the windows driver appears a bit more stable but I really think it might be a hardware compatibility issue. /martin
That doesn't make a whole lot of sense...? This is definatly a problem that has showed up after newer versions of the driver have come about. Before using Gentoo I used FreeBSD and had this card working without any problems. Even after moving to Gentoo nearly 8 months ago it worked great, somewhere after many updates and not running flightgear, I finaly had some time on my hands where I was running flightgear again, then my freezing problems showed up. I wish I understood some of the BIOS settings so I could make some intelligent decisions regarding how to set up the card, but how come this has not come up before?
One more thing to try: What I found out with the nonworking Ti4200 is that the driver sets the clocks to the settings for 3D/Performance mode. So in console mode, the card is running 250/333 and works, in X it is running 250/513 (use nvclock -i) and freezes. When I put it to 245/400 (nvclock -n 245 -m 400) it seems to work. Hmm, I'll give this some more testing.
I have been playing around in BIOS, and this is what I found, Trying to be safe by disabling the fast AGP writes seems to be a large contribitor to the stability problems I have been having. After enabling that I have not yet had a crash even after running flightgear for more than an hour (It was pretty much guarenteed to crash after 30 min). The above post about timing my have something to do with that.
Ok I just tried the 2.6.9 kernel from gentoo-dev-sources and I have yet to have a hang up, I have been playing flightgear for over 4 hours (cross country flight) It makes me wonder if the driver is geared to the newer kernel.
Closing then.
