James wrote:


>>From the looks of these error messages Civilme was (as usual) on the
> money.  They are paging errors.  Some people just seem to have all the
> answers. *grin*  Should this be used with K-6 as well?
> 
> James

Well I'm not complaining about *that* ! hehe... but he is not a total 
magician, I actually wrote that I had a paging error in my first post.


Nicolas:

 >>
I experienced EXACTLY the same thing, with my Mdk 8.2 just installed. I 
have
a Duron 900 and a motherboard MSI 6330 Lite Edition 3 (VIA KT133A chipset)
and a 3Dfx Voodoo 3 2000.
 >>

Well you haven't even got an  Nvidia board, so this kinds of rules out Nvidia funky 
binary drivers...

Note that I was using Mandrake 8.1, with the 2.4.8-26mdk kernel 
(recompiled for Athlon arch) I only upgraded to 2.4.18 to see if the 
problem would go away. No luck. This also answers etharp question, I had 
freezes both with a Mandrake patched kernel and the latest stock Linus 
kernel.
Speaking about Athlon arch, I heard that compiling the kernel with an 
Athlon arch would cause problems, not the opposite; can someone clarify 
that?
About downgrading, using a 2.4.3 LM8.0 kernel does not look very 
attractive to me, it will probably break compatibility with some things 
(like ext3). A 2.2.x kernel is not an option...

Goshko:
 >>
Hey all,

well let me put few cents on top. I do have dual Intel(R) Xeon(TM) CPU
1.70GHz stepping 02 DELL machine with 1G of memory, and I do experience
the same random freezings like the "AMD" users using Mandrake 8.1 kernel:
Linux version 2.4.8-26mdkenterprise. Thought it's a X system, so I
totally disabled it, but this doesn't solve the problem. I was thinking
to upgrade to newer kernel/system version, but after the last e-mails
looks like I have to downgrade...
 >>
Rodolfo:
 >>
Well, I'm realizing that instability is much more extended than I 
thought. I've got a 1.5 GHz P-IV (Intel D845WN mobo)which is totally 
unusable under 8.1 because of random full lock-ups. I've tried 
everything, from changing memory modules, or switching kernels to 
disable usb, with no avail. I'd try another distros to see what happens, 
but their installation programs complain about the partitioning table 
Mandrake created, and cannot solve that without too much tinkering. A 
pity...

 >>


Now we have the freezes spreading to Intel architetures too... a note on 
the DiskDrake problem, I see a HUGE number of people who has had 
problems such as losing their window$ partitions with it. I thought they 
were just too green and had done smthing silly, but then some quite 
experienced friend of mine also had this problem, and now this... 
incidentally I never used DiskDrake to repartition anything, I used 
cfdisk. Lucky me. Mandrake guys should take a deep look at DiskDrake.

George:

 >>

I'll be interested in hearing / seeing what you come up with on this.  If this does 
work, I'd be curious to find out if you try turning off the video accelleration, but 
leave the mem setting at the default.
>>


There's is no such an option , disabling hadware acceleration (at least 
not in MCC 0.70 which comes in LM8.1). I believe that what you are 
referring to is using Xfree 3.x.x, which has open souce "nv" drivers 
both with and without acceleration. But I use Xfree 4.1.0, with the 
Nvidia binary drivers "nvidia". If that is the case, i.e, you are using 
"nv" drivers, then they *are* known to be very unstable with 3DAccel. If 
you want to use the open source drivers than you should really not use 
acceleration. You can look at the /etc/X11/XF86Config-4 (or XF86Config, 
if using Xfree3) to know for sure.

George
 >>
Also, I recall reading an article that mentioned a change in the memory 
management in the 2.4.x kernels. (This was several months back: an 
interview with Linus or one of the other kernel developers in Linux 
Journal or Linux Magazine..or other similar magazine.)
 >>

I believe you are referring to the so called "VM war"; Linus wanted a 
new VirtualMemory management model while Cox (and van Riel also I think) 
would rather improve the old one. Linus prevailed and Andrea Arcangeli's 
(right spelling?) model was implemented. This was shortly before Tosatti 
was chosen as the new kernel maintainer, if I recall it correctly.

Nicolas:
 >>
That's how I knew FOR SURE it has nothing to
do with X, but with the kernel.
 >>

While I can know for SURE I'm presently very inclined to agree with you. 
Kernel 2.4 has been a very problematic series unfortunately. Remember 
that 2.4.15 kernel which had severe bugs that would corrupt the filesystem?


Collecting pieces, the advices for people having trouble are:

-For Athlon systems:
                -try mem=nopentium (BTW, Nvidia recommends this in their README)
                -Try noathlon (can someone comment on this option?)

                -Try noapic
-For Intel systems:
                -try noapic

-For everyone:
                -downgrade your kernel... :^(



Wooky

-- 
-- 
shinjiteiru shinjirareru,
korekara aruku kono michi wo!

kimi ga iru yo, boku ga iru yo
sore ijou nani mo iranai.
umareta imi ,sagasu yori mo
ima ikiteru koto kanjite,
kotae yori mo, daiji na mono
hitotsu hitotsu mitsuketeiku...


Want to buy your Pack or Services from MandrakeSoft? 
Go to http://www.mandrakestore.com

Reply via email to