Re: Problems on Ruffian ??
Peter Biechele <peter.biechele@physik.uni-freiburg.de> writes:
> Hello,
>
> we are having big trouble still with Alpha Ruffian systems having 1.5 GB. We
> were using Redhat 5.2 for a few months, but we are actually not satisfied with
> the software and the support coming with it. So we switched to Debian 2 weeks
> ago. The installation went fine and I was surprised about the great job done by
> the Debian team: Thank you !! (We are also using Debian Intel for 3 years now
> and are really satisfied).
>
> We have tried Redhat 6.0 too, but even the boot disks did not work on our
> machines.
>
> BUT as always, we ran into trouble. After a few days, I had recompiled
> the new 2.2.9 kernel for the alphas, we got kernel paging requests and the
> machines stopped working. I have read a lot since we bought our systems about
> this problem, but there seems to be nobody around trying to solve this problem
> or at least I do not know of somebody.
>
> Is there finally a solution to this problem:
It heavily depends on what kernel and compiler you use. I compiled
2.2.9 with the debian egcc and with the newest snapshot. Neigther goes
past the initial init. It just shows some unaligned traps (4 of them)
and then hangs. So what egcc version are you using?
I saw that there is a patch for 2.2.9 from Alan cox and since he has
an Alpha, he probably fixed something important for us.
> Debian 2.1 running kernel 2.2.9 I get the following:
>
> Unable to handle kernel paging request at virtual address 0018000000000040
> sort(1501): Oops 0
> pc = [<fffffc000034a260>] ps = 0000
> rp = [<fffffc0000338b20>] sp = fffffc00401b9ed8
> r0=fffffffffffffff7 r1=18000000000000 r2=fffffc003eeb3bf0 r3=fffffc00404c3020
> r8=fffffc00404c3128
> r16=fffffc005fe8eff0 r17=fffffc0040935c80 r18=0 r19=0
> r20=ffffffffffffffff r21=0 r22=fffffc00401ba000 r23=155556f2e20
> r24=8 r25=2000 r26=fffffc0000338b20 r27=fffffc000034a240
> r28=0 r29=fffffc00004bbf80 r30=fffffc00401b9ed8
> Code: a6110060 a4300098 e4200009 <a0210040> e0200004 42131410 44420411 d340000d c3e00003
> klogd: memory violation at pc=120004cf8 rp=120002578 (bad address = 00000008)
>
>
> Running kernel 2.0.36 weith the new NVR564xxx driver we get only the following:
> <sc 208(11fffff2e,0,14)>
> <sc 208(11fffff2e,0,14)>
> <sc 208(11fffff2e,0,14)>
> <sc 208(11fffff2e,0,14)>
> <sc 208(11fffff2e,0,14)>
Those are perfectly ignorable. Just some debug messages.
> <sc 208(11fffff2e,0,14)>
> <sc 208(11fffd9c0,0,5)>
> <sc 208(11fffff2e,0,14)>
> <sc 208(11ffff4b0,0,5)>
> <sc 208(11ffff4b0,0,5)>
> <sc 208(11ffff4b0,0,5)>
> <sc 208(11ffff4b0,0,5)>
> <sc 208(11ffff4b0,0,5)>
> task already on run-queue
Never got those. Could be from the NVR driver, since I didn`t use that
on my ruffian.
> Because we want to run special software, which is only available on 2.2.x
> kernels, we would like to resolve these problems.
>
> Please help !!!!
You probably have to keep trying to compile different 2.2.x or 2.3.x
kernels and see if they work. I am running a 2.2.1ac5 kernel for some
time now without any problem, but I keep trying to get a newer one
working.
May the Source be with you.
Goswin
Reply to: