Re: [suse-amd64] PCMCIA [Was: X.org, USB optical mouse and SuSE 9.3 AMD64 : bad behaviour at 3 (components) !
Hello
Well, I think that I have a good new about solving the problem. It seems that
the kernel requires the following parameter : noapic.
The laptop has been "working" since last evening .... and this morning it is
still alive with a mouse and sound device working fine :-) !!!!!
So it must be tested on a longer period in order to validate the option. I
think that it is on the right way :-)
The system behaviour when the problem occured suggested me that it was related
to interruptions => apic. And it was proposed on a "64 bits linux" forum to
use "noapic" for some distros (Mandrake for example).
And this morning, after reinstallation of the nVidia driver (8178) I got no
full freeze during 15 minutes !! Wonderfull, 2 problmes solved with noapic !
So, I suggest you try this option "noapic" at boot. You will tell me if your
USB problem has disappeared :-).
Question : why is "noapic" required for kernel >=2.6.11 and not for 2.6.8 and
before ? Like you, I am not a specialist of the Linux kernel :-).
I can see that your laptop is very similar, with the main chipset nForce3 150.
Almost same laptop .... same problem. About PCMCIA, it was very hard to get
it working perfectly. I tried several values for range and port in
config.opts file. In fact, this is WinXP that helped me to find the right
values : I took the values given by XP and put it in the config.opts file :
bingo !! It works : modem card and network FA411 card. Hereafter my config
file. With 32 bits kernel, it is easier :-) : no tuning !
Regards
Pascal
My confiq.opts file (tested with SuSE 9.2, SuSE 9.3, SuSE 9.0 and SuSE 9.1) :
# This are the official ports to use from pcmcia-cs, but ...
# include port 0x100-0x4ff, port 0x800-0x8ff, port 0xc00-0xcff
# ... ports 0x810-0x81f hurt on some DELL machines and ...
# ... ports 0x3b0-0x3df hurt on some FSC machines.
#include port 0x100-0x4ff
#include port 0x800-0x8ff
#include port 0xc00-0xcff
#include port 0x3000-0x7fff
# according to lspci -vv for Cardbus
#include port 0x0000-0xafff
#
#include port 0xb000-0xbfff
#include port 0xc000-0xcfff
# avec ceci, carte reseau FA411 montee en eth1
#include port 0xd000-0xffff
# d'apres WinXP :
include port 0xcc00-0xccff
include port 0xd000-0xd0ff
include port 0xd400-0xd4ff
include port 0xdc00-0xdcff
# d'apres lspci -vv decompose en 4 gammes
#include port 0xb000-0xb7ff
#include port 0xb800-0xbfff
#include port 0xc000-0xc7ff
#include port 0xc800-0xcfff
# according to lspci -vv for Cardbus
#include memory 0xf8a00000-0xfd1ff000
# plage memoire relative au pont PCI (secondary=02)
#include memory 0xf8a00000-0xfeafffff
# d'apres WinXP :
#include 0xfa200000-0xfa200fff
#include 0xfb200000-0xfd1fffff
#include 0xfd200000-0xfd200fff
#include 0xfdaf5000-0xfeaf4fff
#include 0xfeaf5000-0xfeaf5fff
#include 0xfeaf6000-0xfeaf6fff
#include 0xfeaf7000-0xfeaf7fff
#include 0xfeafc000-0xfeafcfff
include memory 0xfa200000-0xfeafcfff
# decomposition en 2 => ne marche pas : carte non visible
#include memory 0xf8a00000-0xfa1ff000
#include memory 0xfba00000-0xfd1ff000
# decomposition en 4 => ne marche pas : carte non visible
#include memory 0xf8a00000-0xf95ff000
#include memory 0xf9600000-0xfa1ff000
#include memory 0xfba00000-0xfc5ff000
#include memory 0xfc600000-0xfd1ff000
#
# High port numbers do not always work...
# include port 0x1000-0x17ff
# Extra port range for IBM Token Ring
#include port 0xa00-0xaff
# Resources we should not use, even if they appear to be available
# First built-in serial port
exclude irq 4
# Second built-in serial port
#exclude irq 3
# First built-in parallel port
exclude irq 7
=====================================================
De: Constantine 'Gus' Fantanas
patheve2 wrote:
Hello
Well, I think that I have a good new about solving the problem. It seems that the kernel requires the following parameter : noapic. The laptop has been "working" since last evening .... and this morning it is still alive with a mouse and sound device working fine :-) !!!!! So it must be tested on a longer period in order to validate the option. I think that it is on the right way :-) The system behaviour when the problem occured suggested me that it was related to interruptions => apic. And it was proposed on a "64 bits linux" forum to use "noapic" for some distros (Mandrake for example). And this morning, after reinstallation of the nVidia driver (8178) I got no full freeze during 15 minutes !! Wonderfull, 2 problmes solved with noapic ! So, I suggest you try this option "noapic" at boot. You will tell me if your USB problem has disappeared :-). Question : why is "noapic" required for kernel >=2.6.11 and not for 2.6.8 and before ? Like you, I am not a specialist of the Linux kernel :-).
I can see that your laptop is very similar, with the main chipset nForce3 150. Almost same laptop .... same problem. About PCMCIA, it was very hard to get it working perfectly. I tried several values for range and port in config.opts file. In fact, this is WinXP that helped me to find the right values : I took the values given by XP and put it in the config.opts file : bingo !! It works : modem card and network FA411 card. Hereafter my config file. With 32 bits kernel, it is easier :-) : no tuning !
Regards
Pascal
Hi, Pascal. Concerning the Presario R3000-series and its twin brother Pavillion 5000z-series, I think that the APIC problems had been resolved by the time 64-bit SuSE 9.3 was out. I remember that a year and a half ago we had to cope with kernel messages about many lost ticks, drivers hogging interrupts, random lockups, unresponsive keyboard, and a crazy clock that for some was too fast and for others too slow (I experienced a very fast clock and all the rest). I vaguely remember people running into other problems after booting with the 'noapic' kernel directive (does your speed governor work with 'noapic'?). Newer BIOSes (is your BIOS up to date, by the way?) and updates in the kernels eliminated all these problems. Some people suggested booting with the kernel directive 'acpi_skip_timer_override = 1'. I don't use it any more and I don't see any problems in my sytem. If I remember correctly, people from SuSE (e.g. Andi Kleen) and AMD, who monitor this board, were actively involved in fixing the APIC problems, at least in the Presario R3000/Pavillion 5000z-series AMD64 laptops. Hats off to them! Things have come a long way since the fall of 2004! On the hardware side, I have not been able to successfully resume from suspension or hibernation (upon resume, I lose my keyboard and display, but the computer responds to pings) and on the software side I still have problems with sound mixing of outputs of 64-bit applications with outputs of 32-bit applications. But overall the progress of 64-bit SuSE from the fall of 2004 to spring of 2005 was nothing but spectacular. Regards, Gus Fantanas
participants (2)
-
Constantine 'Gus' Fantanas
-
patheve2