Void's Forums

by **Void Main** » Mon Dec 08, 2003 11:56 am

I can't help you on the Radeon drivers since I don't have one. Does ATI have a forum like the one nVidia has? I assume you've already tried to install the latest RPM from them. I assume the ATI RPM comes with both X drivers and kernel drivers. If it's the X driver that is causing the problem you could downgrade to the version of X that comes with Red Hat 9 (ick). If it's the kernel then you could downgrade to an older kernel (ick). Best thing would be to hound ATI for drivers that work with Fedora (if they aren't already out there).

As far as the apt-get thing goes:
http://voidmain.is-a-geek.net/redhat/fe ... _have.html

by **Tux** » Mon Dec 08, 2003 12:12 pm

Did you try re-installing fglrx-glc22-4.3.0-3.2.8.i586.rpm?

by **Membrax** » Mon Dec 08, 2003 5:58 pm

Yep, I already tried to reinstall this RPM. But thanks for the advice anyway.

But, guys, I got it !
Right now I'm using the system which gave me so many troubles.

Here's the way (or at least the one which lead me to this) :

As Void Main mentioned before, do not forget to type "export CC=gcc32" in your console before proceeding to the RPM install.

Then, install thr RPM package as usual (-Uvh for some of you, -Uvh --force for the others) and actually as mentioned on ATI's driver webpage.

Some important stuff at this point :
once the RPM has been fully RPM'ed

launch the "fglrxconfig" and answer all the clear questions.
At the one asking if you want to use an external "gart" or "agp gart" you should answer "y" (yes).
This fglrxconfig file will write a new XF86Config-4 in your /etc/X11/ directory.

I clearly advise you to then delete (or preferably back-up) the original XF86Config and to rename the "XF86Config-4" in "XF86Config"

Last important point, when all this has been done, edit your modules.conf file located in your /etc/ directory.
And add this line :
options agpgart agp_try_unsupported=1

That's it.
It should run fine now ... and if it doesn't, you're then extremely close to it !

Enjoy !

by **Void Main** » Mon Dec 08, 2003 6:47 pm

Interesting, did you figure that all out on your own or did you find a nice resource out there? Probably wouldn't hurt to throw a link up of where you got information leading you to the resolution. Good work!

by **Membrax** » Tue Dec 09, 2003 6:08 am

As you imagine, I got some scratchy lines from here and there.
Gathered some tiny and unclear info from other places and had to figure out how to shake and mix all that correctly.
On top of that, needlees to mention the number of hours spent to test ... which is one of the major magical elements.

By the way, thanks for the link to the fedora reps.

by **Membrax** » Wed Dec 10, 2003 3:42 am

Hi All,

again a tough one.

I'm using the kernel version standard shipped with the 3 ISO's of FC1, ... which is 2.4.22-2215-ntlpsmp if I can recall.
I'm using gKrell to monitor the CPU usage and other stuff and, when using the so well named smp version of the kernel, only 1 CPU (actually called "CPU 0") is to be seen.

The system is an Intel 3GHz FSB800MHz Hyper-Threading-Technology which have the specific feature to be able to split himself virtually in two logical CPU's.
gKrell displays well two little windows, each stands for one CPU, but only one of those seems to be active.

Isn't there any flag to activate the effective "dual-CPU" recognition (it worked well under RH9.0)

Cheers !

by **Void Main** » Wed Dec 10, 2003 10:18 am

Do both CPUs show up in the "top" command? And do they both show up in "cat /proc/cpuinfo"?

by **Membrax** » Wed Dec 10, 2003 12:40 pm

Yes, both are seen by top.

CPU states:
cpu user nice system irq softirq iowait idle
total 5.2% 0.0% 0.8% 0.0% 0.0% 0.0% 194.0%
cpu00 5.2% 0.0% 0.8% 0.0% 0.0% 0.0% 94.0%
cpu01 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 100.0%

And cat /proc/cpuinfo sees them as well.
The output is :

processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 3.00GHz
stepping : 9
cpu MHz : 3014.562
cache size : 512 KB
physical id : 0
siblings : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid
runqueue : 0

bogomips : 6016.20

processor : 1
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 3.00GHz
stepping : 9
cpu MHz : 3014.562
cache size : 512 KB
physical id : 0
siblings : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid
runqueue : 0

bogomips : 6016.20

by **Void Main** » Wed Dec 10, 2003 2:07 pm

Hmmm, I just apt-get installed gkrellm on a dual processor RH8 box and it came right up showin both CPUs with no params required. Unfortunately I don't currently have an SMP machine running Fedora anywhere that I can test on at the moment. You say you had RH9 installed on that same machine and gkrellm worked on it with no problem? If so then it seems it would be either a kernel or a gkrellm issue. Have you tried downloading the latest gkrellm source and compiling? Might even want to use gcc32 when you do the compile.

by **Calum** » Mon Dec 15, 2003 5:49 pm

Void Main wrote:So far I'm really liking Fedora. Here's the first Fedora tip:

http://voidmain.is-a-geek.net/redhat/fe ... _have.html

Yeah, I know it's a ripoff of my other two apt-get tips but hey.

void main, its expected really, but your howto has a currently incorrect url for the apt rpm. correct one is currently http://ftp.freshrpms.net/pub/freshrpms/ ... r.i386.rpm

i have RH9 actually, and have run into a little issue trying to upgrade. here it is:

Code: Select all: [root@claudia root]# rpm -Uvh http://ftp.freshrpms.net/pub/freshrpms/fedora/linux/1/apt/apt-0.5.15cnc3-0.1.fr.i386.rpm Retrieving http://ftp.freshrpms.net/pub/freshrpms/fedora/linux/1/apt/apt-0.5.15cnc3-0.1.fr.i386.rpm warning: /var/tmp/rpm-xfer.DBYwAm: V3 DSA signature: NOKEY, key ID e42d547b error: Failed dependencies: libbeecrypt.so.6 is needed by apt-0.5.15cnc3-0.1.fr [root@claudia root]# apt-get install libbeecrypt Reading Package Lists... Done Building Dependency Tree... Done E: Couldn't find package libbeecrypt

oh well, i can find an rpm with this online, but in my experience it is wobblier and wobblier if i start installing non apt rpms to satisfy apt type dependencies :-(

i wonder if there is a better way to upgrade using apt on red hat 9 to red hat 10 (or whatever...)? any ideas?

edit: see? :

Code: Select all: [root@claudia root]# rpm -Uvh /home/calum/libbeecrypt-2.2.0-alt2.i586.rpm warning: /home/calum/libbeecrypt-2.2.0-alt2.i586.rpm: V3 DSA signature: NOKEY, key ID eac91ca0 error: Failed dependencies: /sbin/postun_ldconfig is needed by libbeecrypt-2.2.0-alt2

by **Void Main** » Mon Dec 15, 2003 7:11 pm

Dang Matthias keeps changing the filename format for his apt RPM and it screws up my auto update of that page. I fixed it until the next time the format changes (I'm not really mad at him, he's my hero). :)

Don't know where you got that "libbeecrypt" RPM but it's not from RH9 or Fedora. Just use the apt you currently have installed to do the upgrade, in the process it will upgrade apt and beecrypt and all the other packages on your system. I can't think of a reason why you would have to upgrade the apt to Fedora's before doing the upgrade. Just point to the Fedora repository and "apt-get update;apt-get dist-upgrade". Should work.

If you'll notice in one of my other threads I use the apt for the currently installed OS version and just point to the newer versions repo and upgrade away:

viewtopic.php?t=384

by **Calum** » Tue Dec 16, 2003 3:20 pm

oh ok, i figured it out fairly easily, and now have to wait only 4 1/2 hours for my upgrade (dontcha love broadband?) however i was shocked to find out it will take up ANOTHER 199MB which is kind of worrying considering how much the red hat 9 took up. i installed it on the partition i had had slack on and for some reason i had to pare down a lot of things that i hadn't had to in slack. well, maybe i'll pare down more after the upgrade. rpm makes that sort of thing easy i suppose.

by **Void Main** » Tue Dec 16, 2003 3:43 pm

199MB? I feel pinched if I don't have at least 10GB of breathing room. :) I usually work with quite a lot of data.

by **Master of Reality** » Wed Dec 17, 2003 8:06 am

i have around 300MB of breathing room in Fedora.... but i should be going on a deleting rampage now that i know how to fix the RPM freeze.

by **Calum** » Wed Dec 17, 2003 1:36 pm

well, as it happens i was almost totally stumped by the 1gig of downloads required for the rpms that are necessary for the upgrade! i have totally shunted everything around now and symlinked /var/cache/apt/archives to a directory on a partition with i hope enough space on it! i'll have to delete all the rpms afterwards! my entire hard drive on this laptop is only 11MB and i have also got win98SE installed too (for reasons too tedious to go into) (plus a few gigs of wav files, now i am putting all my music on the computer).
well, still red hat nining as we speak, network down to 19kbps now so got to wait another 6 hours for my upgrade...

Void's Forums

Fedora Core 1

FC1 smp Kernel ... HTT not detected as dual CPU ?

Who is online