Suse 9.0 and nvidia problem. Black screen with kernel 2.4.21
hi, I have been trying to get Suse 9.0 to work with nvidia (5950Ultra) card for the past 3 days without any success, I seek your help. First of all, my hardwares: Asus S8V with 1gig of ram (512x2) AMD Athlon-64 3400+ Asus nvidia 5950Ultra, 256MB WD120gb LiteOn DVD/CDRW Combo The basic Suse 9.0 installed without any problem (though with some minor stability problem). The onboard ethernet (3com 940) is recognized with the standard kernel as sk98lin, which works just fine. I proceeded with installation of Suse 9.0 with "nv" driver, which works as well. I then performed YOU to update everything, including kernel 2.4.21-203. I then performed the following before I installed the nvidia driver: cd /usr/src/linux make cloneconfig && make dep When Installation the nvidia driver ( NVIDIA-Linux-x86-1.0-5332-pkg0.run), the kernel-modules compiled just fine. In fact, the installation process went finished without any error message. However, when I tried to use SaX2 to configuration my X using the following command: sax2 -m 0=nvidia I get a black screen. X locked up completely and I lose the keyboard. I had to remotely login and reboot the machine instead. After some search (google is your friend) I found that problem is NOT isolated. It had something to do with how AGP is handled with kernel 2.4. So far, I have successfully got nvidia's driver to work by disabling AGP: by adding the following line in my /etc/X11/XF86Config: Option "NvAGP" "0" # or 1, doesn't matter By doing this, when X starts I get the nvidia's splash screen. Running glxgear I'll get something like 500FPS, which is pathetic (in comparison, my Athlon XP 2400+ with Ti4200 gets 6500FPS), but still better than then 150FPS of the "nv" driver. Setting NvAGP to 2 will results in locking up X. My question is, is there anyway to get AGP/nvidia to work with kernel 2.4? My (limited) understanding is that one has to use amd64_agp instead of AGPGART for nvidia's driver to work, but amd64_agp is a feature that is only available in kernel 2.6. I have tried to compile kernel 2.6.4 (standard from kernel.org) but ran into the (another) famous problem of "unable to mount VFS" problem, which trashed the partition and hence the existing 2.4 (Suse 9.0) installation. Any help is very much appreciated, Elvis _________________________________________________________________ MSN Premium with Virus Guard and Firewall* from McAfee® Security : 2 months FREE* http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=http://hotmail.com/enca&HL=Market_MSNIS_Taglines
On Fri, 19 Mar 2004 21:04:09 +0000
"Elvis Chen"
First of all, my hardwares:
Asus S8V with 1gig of ram (512x2)
That has the nvidia chipset, hasn't it?
After some search (google is your friend) I found that problem is NOT isolated. It had something to do with how AGP is handled with kernel 2.4. So far, I have successfully got nvidia's driver to work by disabling AGP: by adding the following line in my /etc/X11/XF86Config:
Option "NvAGP" "0" # or 1, doesn't matter
By doing this, when X starts I get the nvidia's splash screen. Running glxgear I'll get something like 500FPS, which is pathetic (in comparison, my Athlon XP 2400+ with Ti4200 gets 6500FPS), but still better than then 150FPS of the "nv" driver.
Setting NvAGP to 2 will results in locking up X.
My question is, is there anyway to get AGP/nvidia to work with kernel 2.4? My (limited) understanding is that one has to use amd64_agp instead of AGPGART for nvidia's driver to work, but amd64_agp is a feature that is only
2.4 has an AMD64 AGP driver builtin too, it just has a different name. You could try the latest 9.0 update kernel, it had a few AGP fixes.
available in kernel 2.6. I have tried to compile kernel 2.6.4 (standard from kernel.org) but ran into the (another) famous problem of "unable to mount VFS" problem, which trashed the partition and hence the existing 2.4 (Suse 9.0) installation.
Trashed the partition? normally that message shouldn't write anything to disk. -Andi
I think the V would be Via. Boards with an N are nvidia. Mike On Fri, 19 Mar 2004, Andi Kleen wrote:
Date: Fri, 19 Mar 2004 22:21:24 +0100 From: Andi Kleen
To: Elvis Chen Cc: suse-amd64@suse.com Subject: Re: [suse-amd64] Suse 9.0 and nvidia problem. Black screen with kernel 2.4.21 On Fri, 19 Mar 2004 21:04:09 +0000 "Elvis Chen"
wrote: First of all, my hardwares:
Asus S8V with 1gig of ram (512x2)
That has the nvidia chipset, hasn't it?
After some search (google is your friend) I found that problem is NOT isolated. It had something to do with how AGP is handled with kernel 2.4. So far, I have successfully got nvidia's driver to work by disabling AGP: by adding the following line in my /etc/X11/XF86Config:
Option "NvAGP" "0" # or 1, doesn't matter
By doing this, when X starts I get the nvidia's splash screen. Running glxgear I'll get something like 500FPS, which is pathetic (in comparison, my Athlon XP 2400+ with Ti4200 gets 6500FPS), but still better than then 150FPS of the "nv" driver.
Setting NvAGP to 2 will results in locking up X.
My question is, is there anyway to get AGP/nvidia to work with kernel 2.4? My (limited) understanding is that one has to use amd64_agp instead of AGPGART for nvidia's driver to work, but amd64_agp is a feature that is only
2.4 has an AMD64 AGP driver builtin too, it just has a different name.
You could try the latest 9.0 update kernel, it had a few AGP fixes.
available in kernel 2.6. I have tried to compile kernel 2.6.4 (standard from kernel.org) but ran into the (another) famous problem of "unable to mount VFS" problem, which trashed the partition and hence the existing 2.4 (Suse 9.0) installation.
Trashed the partition? normally that message shouldn't write anything to disk.
-Andi
I have been having one heck of a postfix with imap and ssl to work on SLES 8 for AMD64. This is our only server at the current time, so it is fireall, app server and email server. It has plenty of power, so that is not an issue. So, looking through the README's I found /usr/share/doc/howto/en/html/Postfix-Cyrus-Web-cyradm-HOWTO which seemed to suggest that postfix with cyrus was the best approach. Being as we only has a small number of email users, going to a MySQL database seemed like overkill. Plus these instructions build most everything from source. With SLES 8 for AMD64 binary RPMs is it possible to get postfix and cyrus working with TLS/SSL for imap and pop? If so where could I find better HOWTO's. I have spent a lot of time reading mailing lists but so far no success. Otherwise, would it make sense to obtain the SuSE 9.0 for AMD64 source RPM's for postfix, cyrus-imap and cyrus-sasl2 and compile those on SLES for AMD64? I have read that postfix 1.x will not work with saslauthd. Thank you - Richard BTW, here is the error I am currently getting using the SLES rpms: rnmixon@kingfish:~> imtest -m shadow -a rnmixon kingfish S: * OK kingfish.FromLtoJSoftware.com Cyrus IMAP4 v2.1.16 server ready C: C01 CAPABILITY S: * CAPABILITY IMAP4 IMAP4rev1 ACL QUOTA LITERAL+ MAILBOX-REFERRALS NAMESPACE UIDPLUS ID NO_ATOMIC_RENAME UNSELECT CHILDREN MULTIAPPEND SORT THREAD=ORDEREDSUBJECT THREAD=REFERENCES IDLE STARTTLS AUTH=GSSAPI AUTH=CRAM-MD5 AUTH=DIGEST-MD5 X-NETSCAPE S: C01 OK Completed Authentication failed. no mechanism available Security strength factor: 0 When I try and connect from a Windoze machine with Outlook here is what I get: Mar 19 07:30:28 kingfish master[32466]: about to exec /usr/lib/cyrus/bin/imapd Mar 19 07:30:28 kingfish imaps[32466]: executed Mar 19 07:30:28 kingfish imapd[32466]: accepted connection Mar 19 07:30:29 kingfish imapd[32466]: starttls: TLSv1 with cipher RC4-MD5 (128/128 bits reused) no authentication Mar 19 07:30:29 kingfish imapd[32466]: OTP unavailable because can't read/write key database /etc/opiekeys: No such file or directory Mar 19 07:30:29 kingfish saslauthd[31370]: AUTHFAIL: user=rnmixon service=imap realm= Mar 19 07:30:29 kingfish imapd[32466]: badlogin: vdsl-130-13-89-148.phnx.uswest.net[130.13.89.148] plaintext rnmixon SASL(-13): authentication failure: checkpass failed Mar 19 07:30:45 kingfish master[31525]: process 32439 exited, status 0
On Friday 19 March 2004 22:04, Elvis Chen wrote:
First of all, my hardwares:
Asus S8V with 1gig of ram (512x2) AMD Athlon-64 3400+ Asus nvidia 5950Ultra, 256MB WD120gb LiteOn DVD/CDRW Combo The "S8V" is new to me?; I only know K8V of ASUS for socket 754
I then performed the following before I installed the nvidia driver:
cd /usr/src/linux make cloneconfig && make dep
When Installation the nvidia driver ( NVIDIA-Linux-x86-1.0-5332-pkg0.run), the kernel-modules compiled just fine. In fact, the installation process went finished without any error message. However, when I tried to use SaX2 to configuration my X using the following command:
sax2 -m 0=nvidia Although I am using 2.6.3 kernel, I started with a working default XF86Config; After building/installing the NVIDIA package I only changed the XF86Config Driver line to using the nvidia driver and then it worked; (after modprobing nvidia). My card is a Geforce2 clone and it runs in AGP 4x mode. So maybe do a diff of the XF86Configs to find a clue.
"unable to mount VFS" problem, which trashed the partition and hence the existing 2.4 (Suse 9.0) installation. I've had the same strange mess-up of the root XFS partition the first day I installed from DVD, after accessing the partition with 32-bit SuSE 9.0. BUT I now use 2x 512MB Infineons instead of Corsair (the board acepted only 1 module! ); It's very stable now.
/Henk
Hello, I am running SLES 8 AM64 and want to run spamassasin. I could not find a binary RPM so I am trying to compile the SuSE 9.0 AMD64 source RPM: spamassassin-2.55-65.src.rpm First, is there a binary RPM that I have overlooked? Second, I am getting the following error when I try and build from the source RPM: rpm -bb /usr/src/packages/SPECS/spamassassin.spec ... + make DESTDIR=/var/tmp/spamassassin-2.55-build install_vendor inst_cfs Manifying blib/man1/sa-learn.1 Manifying blib/man1/spamassassin.1 Use of uninitialized value in -d at /usr/lib/perl5/5.8.0/File/Path.pm line 141. fileparse(): need a valid pathname at /usr/lib/perl5/5.8.0/File/Path.pm line 142 make: *** [pure_vendor_install] Error 255 Bad exit status from /var/tmp/rpm-tmp.66819 (%install) kingfish:/usr/src/packages # Is this possible to fix? Thank you - Richard
participants (5)
-
Andi Kleen
-
Elvis Chen
-
Henk Slager
-
mbrown
-
Richard Mixon (qwest)