This is a discussion on possible X bug w/ nVidia 96.43.05 driver +slack 12? within the Slackware Linux Support forums, part of the Unix Operating Systems category; --> after my recent crash course in Afrikaans I'm into the following: Asus CUV4X P3-1000 1250mb ram Geforce4 MX440-SE I'm ...
| |||||||
| Register | FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| after my recent crash course in Afrikaans I'm into the following: Asus CUV4X P3-1000 1250mb ram Geforce4 MX440-SE I'm using the latest driver for this card from nVidia 96.43.05 released on Jan 29,08. It installs nicely, image is crisp, glxgears shows just under 1000fps, nvidia controls work BUT... when I end my kde session, the screen goes to black, stays black and the keyboard goes dead...no current, lights or response...I have to reset the machine in order to get back to work. I thought that perhaps disabling the nVidia splash logo would do it, but no joy. The nVidia forum suggests disabling DRI to deal with instability but I see that the install program has done that already. There are errors in both the Xorg and kdm logs that I can't post because I'm not at that machine. The kdm log finishes with something to the effect of: <QImage: yadda yadda ...is a null image> repeated twice. Any SWAG's or suggestions? Nothing of this sort shows up at linuxquestions or the nvidia forums. Googling tells me I can get the best prices on "X fails to reset" at Bizrait and eBuy...:-( |
| |||
| On Tue, 26 Feb 2008 08:04:40 -0500, mr.b wrote: > after my recent crash course in Afrikaans I'm into the following: > > Asus CUV4X > P3-1000 > 1250mb ram > Geforce4 MX440-SE > > I'm using the latest driver for this card from nVidia 96.43.05 released on > Jan 29,08. It installs nicely, image is crisp, glxgears shows just under > 1000fps, nvidia controls work BUT... when I end my kde session, the screen > goes to black, stays black and the keyboard goes dead...no > current, lights or response...I have to reset the machine in order to get > back to work. > > I thought that perhaps disabling the nVidia splash logo would do it, but > no joy. The nVidia forum suggests disabling DRI to deal with > instability but I see that the install program has done that already. > There are errors in both the Xorg and kdm logs that I can't post because > I'm not at that machine. The kdm log finishes with something to the effect > of: > > <QImage: yadda yadda ...is a null image> repeated twice. > > Any SWAG's or suggestions? Nothing of this sort shows up at > linuxquestions or the nvidia forums. Googling tells me I can get the best > prices on "X fails to reset" at Bizrait and eBuy...:-( > I know you are using nVidia's proprietary driver, and as a WAG, perhaps the latest X fixes your problems. I fixed a similar problem by doing that. I upgraded X to the version in -current because the default i915 drivers in Slack 12 were locking the machine. Be careful if going down this road, because of possible library mismatches. So far, AFAIK, the libraries in -current are "close enough" matches with "12". I suppose that is subject to change. BTW, that is good performance on fairly old hardware. -- Douglas Mayne |
| |||
| "mr.b" <mist@b.com> wrote: > I'm using the latest driver for this card from nVidia 96.43.05 released > on Jan 29,08. Ouch, you have tainted your kernel. > It installs nicely, image is crisp, glxgears shows just > under 1000fps, nvidia controls work BUT... when I end my kde session, > the screen goes to black, stays black and the keyboard goes dead...no > current, lights or response... This sounds like a bug to me. Bugs are fixed by developers. > I have to reset the machine in order to get back to work. Are you sure? Are you able to ping the machine from another networked machine? Are you able to login by ssh? If you are able to login by ssh only the console is dead and this is probably a bug in X or any of the drivers that you use in X. If you are unable to login by ssh your entire kernel has locked. No kernel developer will try to and probably would not be able to fix that bug as you have a tainted kernel. > I thought that perhaps disabling the nVidia splash logo would do it, but > no joy. Your bug seems related to X and replacing the driver might really be a good idea. With the opensource nv driver you will sacrifice DRI but you will get full control of your machine as you have access to the source code of all drivers that are run and opensource developers will be able and willing to help you. > The nVidia forum suggests disabling DRI If running the nVidia driver with disabled DRI is your choice you might just as well use the nv driver. regards Henrik -- The address in the header is only to prevent spam. My real address is: hc3(at)poolhem.se Examples of addresses which go to spammers: root@localhost postmaster@localhost |
| |||
| On Tue, 26 Feb 2008 21:59:15 +0100, Henrik Carlqvist wrote: > "mr.b" <mist@b.com> wrote: >> I'm using the latest driver for this card from nVidia 96.43.05 released >> on Jan 29,08. > > Ouch, you have tainted your kernel. in fact dmesg even informs me that the kernel has been tainted >> It installs nicely, image is crisp, glxgears shows just under 1000fps, >> nvidia controls work BUT... when I end my kde session, the screen goes >> to black, stays black and the keyboard goes dead...no current, lights or >> response... > > This sounds like a bug to me. Bugs are fixed by developers. > >> I have to reset the machine in order to get back to work. > > Are you sure? Are you able to ping the machine from another networked > machine? Are you able to login by ssh? > > If you are able to login by ssh only the console is dead and this is > probably a bug in X or any of the drivers that you use in X. thank you Henrik, your observations are spot on > If you are unable to login by ssh your entire kernel has locked. No kernel > developer will try to and probably would not be able to fix that bug as > you have a tainted kernel. > >> I thought that perhaps disabling the nVidia splash logo would do it, but >> no joy. > > Your bug seems related to X and replacing the driver might really be a > good idea. With the opensource nv driver you will sacrifice DRI but you > will get full control of your machine as you have access to the source > code of all drivers that are run and opensource developers will be able > and willing to help you. > >> The nVidia forum suggests disabling DRI > > If running the nVidia driver with disabled DRI is your choice you might > just as well use the nv driver. > > regards Henrik I appreciate your comments about open source |
| |||
| On Feb 26, 4:49 pm, Douglas Mayne <d...@sl12.localnet> wrote: > I know you are using nVidia's proprietary driver Does it really correct to apply the word proprietary to driver? In my opinion, driver is not a kind of software, but rather a hardware component, loadable, extractable and replaceable but hardware. Companies/vendors that produce hardware devices without drivers are in fact producing something as cars without engines, useless scrap. nvidia is not the only player on the market of graphic chips, and the company does not want to play poker having their cards open, I mean to open the driver sources under GPL. I do not see any problem here. --- Bogdan |
| |||
| Henrik Carlqvist wrote: > Ouch, you have tainted your kernel. Not to argue "religious" issues, but I've always hated the use of "tainted" in this context. My kernel contains an unsupported module. It may even be politically incorrect. But it is not diseased or corrupted. (I do not wish to argue this issue. I respect Henrik's position.) <snip some good advice and some open source advocacy> >> The nVidia forum suggests disabling DRI Because the standard X DRI module conflicts with the nVidia driver and because the standard DRI module is not needed by the nVidia driver for direct rendering. > If running the nVidia driver with disabled DRI is your choice you might > just as well use the nv driver. Not true. the nVidia driver does direct rendering without using X's DRI. -- Old Man Playing with the ODE will make you go blind. |
| |||
| On Thu, 28 Feb 2008 10:46:37 -0500, Old Man wrote: > Henrik Carlqvist wrote: > >> Ouch, you have tainted your kernel. > > Not to argue "religious" issues, but I've always hated the use of > "tainted" in this context. My kernel contains an unsupported module. It > may even be politically incorrect. But it is not diseased or corrupted. > (I do not wish to argue this issue. I respect Henrik's position.) > > <snip some good advice and some open source advocacy> > >>> The nVidia forum suggests disabling DRI > > Because the standard X DRI module conflicts with the nVidia driver and > because the standard DRI module is not needed by the nVidia driver for > direct rendering. > >> If running the nVidia driver with disabled DRI is your choice you might >> just as well use the nv driver. > > Not true. the nVidia driver does direct rendering without using X's DRI. yes and with Henrik's suggestion of using the nv driver, I'm able to end my session and have the login screen reset, but I've taken a pretty serious performance hit. glxgears has dropped from just about 1000fps with the nvidia driver to 150fps with the nv driver...not that it's that important...I don't spend any quality time enjoying any of the currently popular killing games that just demand so-o-o much video performance |
| |||
| (B.Yanchitsky@gmail.com) writes: > On Feb 26, 4:49 pm, Douglas Mayne <d...@sl12.localnet> wrote: >> I know you are using nVidia's proprietary driver > > Does it really correct to apply the word proprietary to driver? In my > opinion, driver is not a kind of software, but rather a hardware > component, loadable, extractable and replaceable but hardware. > Companies/vendors that produce hardware devices without drivers are in > fact producing something as cars without engines, useless scrap. > nvidia is not the only player on the market of graphic chips, and the > company does not want to play poker having their cards open, I mean to > open the driver sources under GPL. I do not see any problem here. > Companies that supply hardware without drivers either release the information about the hardware, or have a product that nobody can use. There are companies that have decided that information about the hardware is "proprietary". You have to be someone special, likely also having to sign something and maybe even pay upfront, to get the information. These same companies may often ignore that there is a world beyond Windows. Or, at the very most, they will release their own software, closed source and with all the restrictions they can find, so that people using a popular operating system that isn't Windows can use their hardware. But the company hasn't changed, they still don't release information about the hardware and they won't release the source code. They are just as "proprietary" as they ever were. Michael |
| |||
| B.Yanchitsky@gmail.com wrote: > Does it really correct to apply the word proprietary to driver? In my > opinion, driver is not a kind of software, but rather a hardware > component, loadable, extractable and replaceable but hardware. Do you really mean that it is possible to download hardware by ftp? I think that most people agree that drivers are software. There is also something called firmware. Firmware is not a driver with instructions that run in your computers main CPU. Firmware is software or data used in hardware. The firmware could be in a ROM on a card or it could be loaded by the driver to the card. Firmware is usually not considered to be hardware or software, it is only called firmware. > Companies/vendors that produce hardware devices without drivers are in > fact producing something as cars without engines, useless scrap. nvidia > is not the only player on the market of graphic chips, and the company > does not want to play poker having their cards open, I mean to open the > driver sources under GPL. I do not see any problem here. The problem is that sometimes software is broken, this is known as bugs. Mr.b is suffering from a bug that completely locks his machine. Who is able to fix a bug? Any skilled programmer with access to the source code. Who has the complete source code? In this case only nVidia. It doesn't matter if the fault is in the opensource Linux kernel source code or if it is in the nVidia driver. Only those with access to the complete source code will be able to reproduce and analyze the bug to fix it. Only nVidia is able to help mr.b. Do you think that nVidia cares? regards Henrik -- The address in the header is only to prevent spam. My real address is: hc3(at)poolhem.se Examples of addresses which go to spammers: root@localhost postmaster@localhost |
| ||||
| mr.b wrote: > after my recent crash course in Afrikaans I'm into the following: > > Asus CUV4X > P3-1000 > 1250mb ram > Geforce4 MX440-SE > > I'm using the latest driver for this card from nVidia 96.43.05 released on > Jan 29,08. It installs nicely, image is crisp, glxgears shows just under [snip] Don't they have a separate version of the driver for older hardware? I know they used to at some point. Also, possibly try finding the setting in kdm that restarts X after each session. Or perhaps, start at "init 3" then do startx, then end session, then startx again and see if that works. If so, then probably the kdm session will do the trick. I'm late for work so can't give details on any of these, I suppose you're familiar with google though. -- Posted via a free Usenet account from http://www.teranews.com |
| Thread Tools | |
| Display Modes | |
|
|