The nx9420 notebook is claimed to be certified for Novell Linux 9. The HP whitepaper on Linux for mobile devices recommends the setting APCI=off for the installation. Still, the Novell installer does not recognize the hard disk, so the installation fails. Which setup procedure should be followed?
Note: If you are the author of this question and wish to assign points to any of the answers, please login first.For more information on assigning points ,click
here
buckle your seatbelt. I do not care what HP may say, Novell (Suse) 9.x is not ready for SATA hard drives out of the box. Get yourself Suse 10.0 or 10.1 and read up on installing Linux on a SATA drive.
I think I am on the way to success using the brand new openSuSE 10.1 . I'll post a summary when everything is done. The SuSE 10.0 distribution didn't work (I only tried 'out of the box', no special drivers/kernels)
I have the same problem with the nx 9420. The Suse 10.0 doesn't work either, so I see no other way around other than reconstructing the kernel to incorporate the SATA driver. The problem is, how do I do that when I can't even install the OS?!? (no Harddisk=no installation!) please help!!
NLD 9 may function when installed with SP1 (I remember a similar problem with installing Suse 9.2 on sata system: solved booting from SP1 disk downloaded from Suse site.). Anyway NLD 9 SP3 is available for download from Novell site. Suse 10.0 may work if you install booting with "noapic nolapic" options. (probed with another HP notebook) Suse 10.1 is working out of the box! (except for sppecial keybord keys, minor speestep issues and the tipical nightmare installing ATI driver)
Actually, I'm still not ready with the Linux setup on my nx9420. I want to install a dual boot WinXP / SuSE10.1, but I can't get PartitionMagic8 to repartition my filesystem (aborts with failure without any effective action). The SuSE partitioner also failed. Any hints on that issue?
Mine is actually working with full sata support and without any kernel boot parameter. Maybe disabling SATA access is necessary only to resize windows partition... I don't know as I directly format the entire drive (I use Windows XP Pro with VMware 5.5 Workstation for legacy applications). I suppose you have allready "defragmented" your windows partition. Please see http://www.novell.com/documentation/suse10/index.html?page=/documentation/suse10/startguide/data/sec_i_yast2_proposal.html for how to proceed.
Lots of good suggestions, unfortunately at least in my case, none work. I've tried: - ACPI off during installation - SATA in setup disabled - "noapic nolapic" options during booting/ installation
Thanks for the support and suggestions, but has anyone really carried out a successful Linux installation on an nx9420? Everything I've tried results in the same thing, no Harddisk detected!
Phillippe you mentioned "(I use Windows XP Pro with VMware 5.5 Workstation for legacy applications)". Could you elaborate on that a bit please? I also would like to run Win XP Pro and (if possible) Suse Linux on the Notebook. If thats possible in conjunction with VMWare then so be it - I just have to know how!!
As I explained before, I'm answering you with Suse 10.1 on my nx9420. I've tried both the downloaded 32bits DVD and the box DVD. The installation went without any problem. Maybe my success reside in having deleted all the original Windows and HP partitions. Don't forget to make the restore DVDs if you ever wish to reinstall Windows...
If I understand correctly the problem resides in Linux being unable to shrink the windows partition. It is very important to read the Novell paper indicated in previous post as Windows XP may write system files at the end of its partition resulting in an unshrinkable partition.
If you go with deleting the windows partition, I don't think you'll have any problem with Suse 10.1 (I haven't tried any other version with this machine)
VMware workstation works just right with Suse 10.1 as host and WinXP Pro as a guest. The only problem I have seen is that the time goes speedier than real. This is a known problem easily fixed (I have not yet fixed it because of lack of time and interest).
The install is a bit tricky if you never do it before. Download the VMware Workstation trial from VMware site. Do a Google search on "vmware any any updates" and download the file. You have to install Suse with kernel sources, GCC and GCC++ Go to /usr/src/linux and do (as root) "make cloneconfig" and "make modules_prepare". From now you can do "vmware-install" after having 'dezipped' vmware. When the install program tells you if you want to run "vmware-config" say no and run the vmware-any-any-update install program. That's all!
Thats exactly what I needed! Thanks very much Phillippe! As soon as I have everything up and running, I shall post the results on this page. Merci' Boucoup!!
I finally succeeded with the Linux installation on the nx9420. Here's the report:
1.) Used Paragon Partition Manager 2005 to shrink the opriginal XP partition. Note: All other partitioning tools (see above) failed! 2.) Installed SuSE 10.1 out of the box, no special parameters, usual partitioning (all on a logical partition). Installed nearly package. Mind to select kernel sources and such, I think the ATI driver needs them (?). This gave me a working installation with X running with max. 800x600 resolution. 3.) Installed the ATI proprietary driver for notebooks availlable through https://support.ati.com/ics/support/default.asp?deptID=894&task=knowledge&folderID=27 with standard options.
It's a dual boot installation, with Linux and the original XP to choose from in the GRUB bootloader menu. If you follow my procedure, be sure to do *all* the partitioning in one step, *before* the new (linux) bootloader is installed! since the partitioning tool works at boottime, this can be problematic in the other case. I didn't realize this at first try, but got everything working again by restoring the original (windows-only) MBR using the YaST bootloader configuration (on Linux)
Philippe, you mentioned speedstep issues. I notice, that the processors won't "accelerate" beyond 1333MHz on my 1830MHz machine, even though I'm heavily loading them with tough mathematics (two simultanous Computational Fluid Dynamics task). Is there any "switch" to manually adjust the processor speed?
Yes, you could use cpufreq-set to change the max cpu frequency the governor "ondemand" may select.
Actually this is not working for me as I have made a lot of changes (kernel, etc) but used to work with the stock install (no sure now...).
You have to type:
cpufreq-set --cpu 0 --max 1833000
cpufreq-set --cpu 1 --max 1833000
I have reported this bug to https://bugzilla.novell.com/show_bug.cgi?id=177963 .
Actually they think it may be a problem with the bios or BIOS <--> OS communication problem. Try to report too: you'll have to register to Novell Bugzilla to do this...
I have run into a problem when installing SUSE 10.1 on a nx9420. Some of the posts in this thread apply to my situation, so I decided to try and get some help here.
I want to keep Windows and for this reason I have repartitioned the HD before installing Linux from the SUSE 10.1 DVD. After the installation of the software packages the computer reboots, but there is a problem with GRUB and the screen with the OS selection is not displayed. Instead all I have on the screen is:
GRUB =A|=A|
I have booted from the SUSE 10.1 DVD again to try to repair GRUB, but then I get the error message
Error 21: Selected disk does not exist
Now it seems that the HD is not recognized by the BIOS, so I modified the Native SATA Mode in the BIOS (i.e. Active / Non Active), with no success: I have the same trouble with GRUB. Perhaps this is the problem Frederik mentions in one of his messages.
Maybe one possibility is to recover the original MBR, but I overwrote it and so this is no longer possible.
Anybody has any suggestions? Should I forget the two installed OSs and start from scratch?
I have no idea why you run into this problem. Maybe your partition software have messed with the partition table and GRUB is lost with it. The SuSE YAST partition manager is able to resize a Windows partition if it is in 'good conditions' (defragmented), so this is the preferred way. I don't know your skills, so I'd recommend (If google give no usefull answers) to reinstall Windows with the stock install procedure (beware you'll lost ALL your data!) and to reinstall SuSE after that. Let SuSE resize your partition. If resizing gives you an error, cancel the install and defrag the windows partition and retry (See http://www.novell.com/documentation/suse10/index.html?page=/documentation/suse10/startguide/data/sec_i_yast2_proposal.html for more details on the procedure. This is the most costly in time but simplest solution as we allready know SuSE 10.1 is working with this machine. ¡Courage!
I'm sorry, my english is terrible!... By "stock install", I mean to use the original installation/recover procedure from HP. You can use it when booting, just after the BIOS post. Or if you allready have created the recovery CDs/DVDs from the installed Windows, just boot with the first one. The only problem I see with it, is that your windows install will be loaded with a lot of software. Second, it may be that the partitions created this way gives some problems with resizing... You can also use a "normal" Windows CD install, but you'll have problems with windows not seeing the SATA drive. Disable Sata in BIOS or if you have an external floppy, use the SATA driver from HP. Anyway, there's a lot of threads just on this in this forum. If you go with the "standart" install from MS, you'll be able to specify the partition size it will use, so no need to resize...
did anyone get suspend2ram working on nx9420 under linux? I use SuSE10.1, everything works out of the box except for suspend2ram. After a resume, upon almost any command (e.g. "ls") I obtain "Input/Output-Error" messages. The error seems to be similar as those described in http://www.thinkwiki.org/wiki/Problems_with_SATA_and_Linux However, SuSe10.1 uses kernel 2.6.16 with a all SATA kernel patches (I should assume)... Does anyone have an idea or a solution to this problem?
I can't access the other bug, no. 177963, although I am registered and logged into Novell's bugzilla I got an "You are not authorized to access bug #177963" error... Could you tell me what this bug is about?
I believe the "Patch against SATA-resume problem with T60" won't help here because these patches are already in the SuSE-10.1 kernel, or so it looks at least...
Bug 177963 present the same symptoms as bug 179702. This is the bios on the X60 which were limiting max cpufreq. Changing bios option 'battery mode' from 'powersave' to 'automatic' solved the problem. I indicate you this workaround on the T60, because it seems like the T60, even with the kernel patches applied, is not working properly. For the nx9420 we'll have to wait...
my nx 9420 is working quite well with ubuntu linux but I have a strange problem. When I've used linux, the next time I boot, it takes nearly 30 seconds to get through the bios. First I only see a blinking curser for 10 to 15 seconds and then for 10 to 15 seconds the bios with the HP logo. Tried 3 linux distributions, 2 nx 9420 and 2 bios versions (from march and mai).
I have exactly the same problem with SuSE 10.1. If I unplug the AC adaptor AND remove the battery (i.e. remove all power sources), the boot behavior is back to normal. Maybe this also has something to do with the broken ACPI...?
Mine is a little slow too to get through the bios, but a lot speedier than yours :-) No, I have no idea why this is happening, you may try to contact HP support. Possibly an ACPI problem... Please could you tell me if you've found any of the problems described in bug 179702 (see previous posts) with Ubuntu. Basically, it is related with acpi: impossible to run at max cpufreq and battery status not changing if connecting or disconnecting AC plug.... Which kernel are you running? any patch?
Thank you for your reply. @Peter Great, it works for me too with unplugging, but I doubt that the support knows what to do. Removing battery and power supply is not a real solution.
@Philippe Sadly, I have the same cpu freq problem with the ubuntu and the current version of the 686 ubuntu kernel.
One comment about the cpu frequency: In my case, /proc/cpuinfo does in fact show 1333 MHz, but "powersave -r" returns 1828.785767 MHz. A quote from the powersave man-page:
"--calc-CPU-frequency -r: prints out the current CPU frequency of the machine using the time stamp counter (tsc) register This should be a simi lar value as in /proc/cpuinfo, however the value there could be wrong if you encounter a kernel bug. If you are fiddling around with SpeedStep or PowerNow you should rely on this value; This currently only works on some single processor machines (i686, x86_64)"
I've submitted your post to Novell for bug 179702 (You know the link!). Maybe it helps them. For now, it seems they're doing a lot of work and maybe we could see a solution in a relativaly short time ... :-))
there is another problem which might be ACPI related, too. If I start linux and then reboot, the laptop is in a "bad state", as described by Siegfried earlier. Besides taking longer for the HP boot logo, the power state is broken in the "bad state", too. This means, that if I boot while the laptop is plugged into AC, /proc/acpi/ac_adapter/C1B7/state will show "on-line". But this doesn't change even if I unplug the AC adapter. Similarly, if the laptop runs on battery while booting, the battery charging state won't change later on, even if I let the laptop run on battery for a long time or if I plug in the AC adapter.
All this happens only if the laptop is in the "bad state". After unpluggin all power sources, the behavior is "normal", until the next reboot.
B.t.w., after rebooting from Windows, the "bad state" hasn't appeared on my system, yet.
Could you guys check whether this happens on your systems, too?
These problems are known and are being working on. I have the same behaviour on mine...
Also, have you tried to update your BIOS to v. F11. If my BIOS boot is slow, it is not SO slow as on your machines. This is the only explanation I can think of for this difference.
I have BIOS v. F11. From pressing the power button until the appearance of the GRUB menu it takes 10 sec. in the "good state" and 35 sec. in the "bad sate". Not really serious, just a hint that something's wrong...
I have Suse 10.1 installed on my nx9420. I didn't have any problems with SATA, dual boot, or stand-alone.
But what about wireless? I installed the ipw3945 regulatory daemon wpa_supplicant and wpa_supplicant gui wireless tools (suse)
I uninstalled the ipw firmware as it says it's for the 2200 wlan card.
The ipw3945 interface is detected in yast and I can see it when I run lspci. I also can see the module seems to be loaded when I run lsmod.
What's bad is that I had it working before with 10.1 when I had a dual boot, but I don't remember doing anything different than what I'm doing now. What am I forgetting?
Man, this weird. I installed the ipw-firmware and now wireless is working, but I had it installed before and it didn't work. So I uninstalled it, tried again, no go, posted on this forum, shut down the laptop and came home. After I read your post, reinstalled the firmware and now it works...
if you type "rpm -q --scripts ipw-firmware" you'll see that the installation of ipw-firmware usually also triggers the load of module "ipw3945". The only thing I can think of why reinstalling ipw-firmware helps, is that the module is not loaded automatically. Please try "modprobe ipw3945" after a reboot and check whether this resolves the problem. If yes, execute "yast2 lan" and select Edit -> Advanced -> Hardware Details. Check whether "module name" is set to "ipw3945".
And, it does list the ipw3945 under module name in Yast. I should probably reboot, see if it is there and then do the rmmod and modprobe. I don't know what that would tell me, but I'm curious.
* The "long boot problem" as well as the problem about the battery status not being updated seem to be fixed with this kernel.
* /proc/cpuinfo still shows the wrong frequency.
* suspend2ram now results into a hang after resume and a command which needs disk access ("ls" is o.k., but "du" not). I assume this is still the SATA problem.
same here with a basic nx9420 (core duo t2300) running FC5 (packaged 2.6.17 from updates repo): - long boot with cursor flashing before splash screen - cpufreq scaling active up to 1.33 GHz (BIOS F.11) - *no* CPU power management with F.12 (no working ACPI power profiles)
I found something out: If you turn off the computer by pressing the power button, then it doesn't end up in a "bad state". One way to achieve that the computer is not switched off after calling "halt" or "shutdown" is to add a line at line 53 of /etc/init.d/halt (the line after the esac) as follows:
command="halt" (this overrides the previous value of "halt -p")
Then the computer won't be switched off after "halt" and just ends up in a halting state. So you can switch it off manually and it will remain in a "good state" for the next boot.
It seems that the kernel-command which actually turns the computer off (probably some ACPI command) is responsible for the "bad state"...
Philippe, do you think this might be interesting for the guys at Novell who are working on the ACPI bug?
Yes I think it may be interesting you submit your findings about the suser-jengelh kernel, halt, etc... About the bug, it turns out it's a bios bug, so HP is working on it since 28/06/2006. Since then, we have no return on their work... I join the DSDT patch from Thomas Renninger. Please, use it only with the F11 bios and don't forget to remove it if you upgrade your bios. It correct the max cpu frequency: now up to 1833MHz. I've not posted it before, because I expect a new bios out in a short time...
Have you tried the 2.6.18 rc2 suser-jengelh kernel? I think I'll install it next week-end.
fglrx-version: 8.26.18 (most recent one) I don't remember the error for sure. While running fglrx-kernel-build.sh some compile error occured (a real compile error). I thought I wouldn't be able to fix it in a reasonable time, so I removed everything again (after all ist's only an rc-kernel and I'd rather waith for the released one).
I'll try the DSDT soon, didn't have time, yet.
B.t.w., I found out that it's not the 2.6.17-kernel where the "bad state" does not occur, but it seems to have something to do with the patches applied by suser-jengelh. I just tried a vanilla 2.6.17 kernel, and there the "bad state" still occurs.
I already contacted suser-jengelh. Maybe he can find out which patch changes this behaviour.
>modprobe speedstep_centrino FATAL: Error inserting speedstep_centrino (/lib/modules/2.6.16.13-4-smp/kernel/arch/i386/kernel/cpu/cpufreq/speedstep-centrino.ko): No such device
Therefore, no speedstepping with that DSDT at all!
(I used the SUSE stock kernel for these experiments)
Here are more news about the "bad state problem", in which the BIOS boot takes so long and the ACPI battery update doesn't work:
This problem does not occur if "psmouse" is compiled as a module and if this module is unloaded before the computer is rebooted. The reason why I never experienced the problem with the suser-jengelh-kernel is because there "psmouse" is in fact compiled as a module which I never loaded because I only used an external USB mouse. You need the module only for the internal touchpad.
So for all who suffer from this problem and still can't figure out themselves how to solve it, here is a straigh-forward work-around:
1) install the kernel-sources from SUSE 2) in /usr/src/linux do "make cloneconfig" 3) in the files /usr/src/linux/.config replace the line "CONFIG_MOUSE_PS2=y" by "CONFIG_MOUSE_PS2=m" 4) compile and install the kernel (make; make modules_install; make install) 5) in the file /etc/sysconfig/kernel" add the value "psmouse" to the variable MODULES_LOADED_ON_BOOT 6) to the file /etc/init.d/halt.local add the line "rmmod psmouse"
If you install the suser-jengelh kernel and you want to use the touchpad just do steps 5) and 6).
I don't understand your problem with speedstep-centrino: mine is working perfectly with original kernel... I suppose your bios is F11? if so, I have no idea. Any error when loading the DSDT?
I have to say too that I have not the problem with "the bad state problem": booting is not specially slow as you and others are mentioning...
Perhaps we have two different revision of the same hardware?
If you want we could try to compare the two machine and see why yours is not working?
1) DSDT - I found the reason: In BIOS I disabled "native SATA" because this allows suspend2ram to work partially (see https://bugzilla.novell.com/show_bug.cgi?id=187955). I didn't say anything here about this solution, because I didn't test it thoroughly yet, and there are still some problems. I planned to write about my experiences later... However, now i enabled "native SATA" again and surprise, with the new DSDT speedstep_centrion can be loaded. So why does the DSDT override not work correctly if native SATA is disabled? For SATA I use the modules "ahci" and "piix" in the initrd, for IDE I use "ata_piix" instead. So this could be a reason.
In fact using your DSDT in IDE mode, gives some additional ACPI errormessages. I attach the output of "dmesg | grep -i acpi" to this mail, so you can have a look if you want. The error messages which are not present inthe SATA configuration start in line 84.
2) About the "bad state problem": From your messages regarding bug 179702 on the novell site, it seems that you too sometimes had the problem that the battery state was not updated. This effect vanished when your batter exhausted... this looks quite similar to the "bad state problem". Could you try the following, so that we can check whether the behaviour is different on our computers? Use the default SUSE kernel + default initrd. Turn off the computer and remove AC adapter + battery. Then turn it on again and watch for the following:
1.) Just before the BIOS menu appears (where you have the choice of entering the BIOS with F10 etc.), is there a blinking cursor in the upper left corner of the screen? If there is a cursor, how long does it take until the boot menu appears? 2.) After linux is booted, insert/remove the battery. Is the battery status being updated? Then reboot, but don't remove the AC adaptor and do 1.) and 2.) again. Are there any differences?
For me, after the first boot, there is no cursor and the battery status is being updated correctly. After the 2nd boot there is a cursor for > 10 seconds and the battery status is not being updated at all.
I have an nx7400, and I also have the bad state problem when I switch off the computer. I tried with Ubuntu dapper 2.6.15-26 kernel, edgy kernel 2.6.17-5 (ubuntu's development release), Fedora Core 5 with kernel 2.6.15, and the latest 2.6.17........... With all of them the bug occurs. Now reading your posts I will try to compile the psmouse as module. When my laptop is in the bad state, also the keyboard doesn't work (in addition to the 1.33ghz clock problem, and the slow bios post). The only way is remove battery and ac adapter.
I will send a support mail to hp.. How I can do it?
I have a nc6320 and I also experience the "bad state problem". I use windowsxpsp2/opensuse10.1 dual boot. How can I email hp with this problem? I can't seem to find a proper connection details. Or is it possible to get them to hear about this through here?
The problem is the psmouse driver. In OpenSuse is compiled built-in so you can't unload it before reboot or shutdown. To fix it you should download kernel sources and recompile the kernel with psmouse as module. The you must find the way to add a 'modprobe -r psmouse' command in the shutdown script. I've done that easily on Ubuntu, in Suse it should be similar..
As I've seen are many the models affected by this bug. HP should take care about this....
Hi all, is there anything bad with BIOS F.12, I'm trying to get my CPU to go up to the max frequency of 1833 and it can't, it stays at 1333 and I need that 500Mhz, else I would have stayed with my previous P-M 1500.
Can anyone help me out on this, I am a bit confused at the DSDT someone posted here also.
The DSDT correct a defect in the bios which do not allow speed > 1333mHz. This DSDT should ONLY be used with bios F11, so please don't use it with F12 or other.
To install: (you have to be root to do this) 1. Copy the DSDT file to /etc/acpi and eventually rename it to DSDT.aml. 2. Edit /etc/sysconfig/kernel and change the value ACPI_DSDT to point to the file you copied in the previous step. ACPI_DSDT="/etc/acpi/DSDT.aml" 3. Create a new initrd by running the script mkinitrd. This might be dangerous so watch out. 4. Reboot
To my knowledge, all the bios up to now have this defect, ... and others! Maybe next version? (due 06/08/15)
No one from HP is monitoring this? There is no response from them for now in the bug I open with Novell bugzilla. Only solutions are from the very kind people of Novell/Suse and this threads. And it seems there are a lot of models from HP affected by these bugs: all new models?
I decided to try out F.12 to see how it behaved and it is the solution for me right now, it doesn't let me scale the freq. down, but it keeps it at the highest freq. possible (1.83Ghz) so it solves my problem.
Thanks for the fast response.
I guess since we are in August answers from big companies such HP are slower, but hopefuly this will get corrected soon.
I own the nx7400 - unfortunately with the 'bad state' Problem. But: As Peter said, the 'bad state' problem is gone if i unload psmouse within the shutdown/reboot script of my UbuntuLinux. (No recompiling is necessary, psmouse as module is standard in the ubuntu kernel). Also the bad state is to be blame for the faulty Battery/ac adapter detection. And only when the notebook is in the 'good state' i can manually set the cpu-speed to 1.66 Ghz (instead of 1.33 Ghz) with the "cpufrequtils" (man cpufreq-set).
Philippe: I haven't tried DSDT, since I didn't have F.11 for download.
With F.12 it seems more stable then the previous F.02 and a bit faster on loading times (maybe is the CPU in Max speed). I removed the psmouse on the reboot and shutdown and there is no bad state now.
The problem now is that with F.12 there is no CPU Freq. management, giving me always 1.83Ghz, which causes de base to get a bit hot and the battery to run down in 1h40m , compared to the previous 2h40m, one hour less.
But since I need the cpu to have the max power available I have to stick to this BIOS for now, just hope HP sends out a new BIOS soon that fixes this so I can have what I bought working 100%.
could anyone post the BIOS F.11? I don't have the file either, and if the next BIOS comes out I'd like to be sure that I can revert to F.11. Is it actually possible to downgrade a BIOS? Or does the BIOS installer complain about that?
I don't know whether posting the old BIOS is o.k. with HP, but maybe someone can open an anonymous account and then post the BIOS. Sadly enough, HP doesn't seem to care about this forum and the problems discussed therein anyway... so in the best case posting the BIOS draws their attention to this matter (which seems to affect several other HP laptops as well).
Yes, only four days to wait! I suppose that if HP removed it from the bios download list they have a reason, so I imagine I can't post it and make it public. But I think it may be OK in a private correspondance... So if someone really need it, send your email to adminNO at SPAMphdserver.com and I'll search for it on my disk.
Could someone report me what is the max measured frequency of the cpu on windows. I'd like the input be with this info: - Cpu model - Max theorical cpu freq - Max measured cpu freq - Bios and windows version
Thanks to all.
Philippe.
* This info is for the people at Novell bugzilla who are working on the problem.
I transmit you 'as it is' the last response HP Support people send to me.
They were insisting that the machine has no problem with Novel NLD9 because it is certified. When I told them that SLED 10 is also certified (see http://developer.novell.com/yes/85095.htm) and we are having the problem with it too, they respond:
Thank you for contacting HewlettâPackard's Commercial Solutions Center.
This is in response to your eâmail regarding the HP nx9420 notebook.
We would like to thank you for sharing the link with us.
Further, Linux is developed by Novell and customers need to contact Novell for any issues with Linux. We apologize for the inconvenience that this has caused to you.
HP provides Linux software downloads as a courtesy to our customers who are using the Linux operating system. These software downloads and any related documentation are not supported by HP Customer Support and are provided without warranties of any kind. The Linux software provided under software downloads is the extent of Linux software that HP will provide for these products. HP cannot ensure the compatibility, quality, or performance of this software, and HP will not necessarily provide maintenance or updates. HP does not endorse any specific distribution of Linux.
Please eâmail us if you need any further assistance and we will be glad to help. --------------------------------------------
For anyone who is having the "bad state" problem, please make the machine having again the "good state" by removing the battery and power cord. It seems like when the machine is in "bad state", it never goes in good state again even if you apply the psmouse trick.
I have a nc6320 with t2400. In windows the max cpu is 1833 MHz the lowest is 1000 MHz. Those are the values in linux too if I boot into linux without the bad state problem (shutdown the machine and remove battery. then start the machine and linux)
Did anyone try the F.14 Bios on a hp nx9420 or some other dual core centrino. When I did an update to F.12 module speedstep_centrino could no longer be loaded, so i downgraded to BIOS F.02 again. Using hp nx9420 SuSE 10.1, kernel 2.6.16.21-0.13-smp compiled with psmouse as module.
Bios F.17 has been uploaded by HP. It is similar to bios F.11: cpu freq is working again, but max cpu speed is not attainable. I hope we could have a new DSDT in a short time because the solution is similar to Bios F11.
Suse people has not found a solution for bios F14: they didn't found why there is no power management.
The nx9420 and the nw9440 share the same Bios. Therefore I've got the same problems on my nw9440: with Bios version newer newer than F.11 the cpu freq settings are broken, but with F.17 it works again.
Max cpu speed settings was never the problem as the maximum speed for the daemon can be set to 2.1GHz by cpufreq-set -u 2100000 This is the max freq of the T2600 Duo Core running in my nw9440. But there are still one error and one warning in the Bios, ckecked by iasl -d DSDT.aml.
Philippe, this one might be for you. nx9420, T2400 CPU, Bios F.17, SuSE 10.1, at the moment without some customized ACPI_DSDT, stock kernel 2.6.16.21-0.25-smp recompiled with psmouse as module, installed rpm cpufrequtils 0.4.13. Effects: - No boot-delay (psmouse) - frequency scaling up to 1833 Mhz (only when forced to use 1833). When the system comes up, you will notice that cpufreq-info tells you a possible speed of 1.83 GHz, 1.33 GHz, 1000 MHz on both cpus, but is only using a strategy of using userspace scaling in between 1000 MHz und 1.33 GHz.
Use 'cpufreq-set -c 0 -u 1833000 ; cpufreq-set -c 1 -u 1833000' and you are done. I made a boot-script that loads just after gpm. Maybe it should even be loaded later.
This one is taken when compiling kernel on just one cpu (defaults). ap@hpnx:~> cat /proc/cpuinfo | grep "cpu MHz" cpu MHz : 1833.000 cpu MHz : 1000.000
Yes, echo 1833000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq echo 1833000 > /sys/devices/system/cpu/cpu1/cpufreq/scaling_max_freq is working perfectly. How could I have the script to start automatically with Suse 10.1?
One problem remain: sometimes one fan start and don't stop (you have to switch off the machine), meanwhile acpi say all the fans are off! Any idea?
From what I've tried I can't have it done automatically, acpi always takes control of it even in init time if I have a script that just runs at the end of every other scripts, so I just did a script that I run under sudo and do this after I get into Gnome.
The thing is that ACPI is reporting the max speed to be 1333000 and not 1833000, making it not working with the max speed.
Also, if you don't kill psmouse as you should it won't let you even set the max speed to 1833000 when restart (after the slowdown at boot).
One more thing, this seems to be a bit more unstable BIOS then the previous ones, since I already had some quirks with it, but nothing messy I must say.
At least its getting there, I just hope that the developers of the BIOS read this and get this feedback to correct the thing.
- Sound will crash after a while and I'm unable to remove the hda driver and restore it again, have to reboot to get sound working again. - Wifi driver gets totally scrambled sometimes, need to turn it off and on again for it to work as it should. - Power consumptions remains the same even if both CPUs are at 1Ghz, the total battery life is the same, either if both are at 1Ghz or both are at 1.83Ghz, making it pretty useless.
Basically I will revert back to F.14 again for now and hope for a new better BIOS to get out soon, since this one is really crappy.
About the script auto start, I have tried to integrate it within init.d/rc.d (suse 10.1) but I had problems with execution order: when it start, dir /sys/devices/system/cpu/cpu1/cpufreq/scaling_max_freq does not exist yet...
Exept for the fan problem, I have not had any of the problem you mention yet. For the speed limit problem, I have noticed that if you do not load speedstep-centrino the cpu is operating at 1833MHz (I was trying to use acpi-cpufreq, but it does not load). So I think the culprit is speedstep-centrino??? I have recompiled the kernel to load psmouse as a module, so no problem at all with "bad state" thing. As for the hda driver, I listening to the radio with kmplayer all day long without any problem, And it is still working after several days, but I have not tried to suspend the machine yet, No problem with wifi: I'm using it right now since this morning. Power consumption is much better now: with bios F.14 (cpufreq=1833MHz) battery last less than 2hours without wifi. Bios F.17, (1000MHz < cpufreq < 1833MHz), Battery last 3h30 without wifi and +/-2h15 with wifi.
>About the script auto start, I have tried to integrate it within init.d/rc.d (suse 10.1) but I had problems with execution order: when it start, dir /sys/devices/system/cpu/cpu1/cpufreq/scaling_max_freq does not exist yet...
I think it might be the same problem here.
> For the speed limit problem, I have noticed that if you do not load speedstep-centrino the cpu is operating at 1833MHz (I was trying to use acpi-cpufreq, but it does not load). > So I think the culprit is speedstep-centrino???
With F.17 I can get speed to have three steps but only after running the commands I mentioned earlier. Else only two steps work, the two lower ones
> I have recompiled the kernel to load psmouse as a module, so no problem at all with "bad state" thing.
In Ubuntu this is the default, so no problem with this, but if the kernel locks and I need a hard reboot it will be in bad state and there is no way to set the max cpu, have to reboot again to have it working ok
> As for the hda driver, I listening to the radio with kmplayer all day long without any problem, And it is still working after several days, but I have not tried to suspend the machine yet,
This is sort of random, right now I had been listening to some videos for 4 hours and no problem. But during the afternoon it just went mute for no reason, it said the sound was ok and everything was as it should, but no sound came from the speakers. This hasn't happened before with the F.14 . But I'm still trying to figure what might be causing this.
> No problem with wifi: I'm using it right now since this morning.
I'm having trouble with networks that have a lower strenght, with F.14 I could use them with no problems, with F.17 it just drops the connections really often and its barely usable. With decent strenght networks there are no problems.
> Power consumption is much better now: with bios F.14 (cpufreq=1833MHz) battery last less than 2hours without wifi. Bios F.17, (1000MHz < cpufreq < 1833MHz), Battery last 3h30 without wifi and +/-2h15 with wifi.
I was hoping to see a bigger increase, mine lasts 1:40 minutes with F.14 and 1:50 minutes with F.17 , using wifi, I was hoping for much more time, not just 10 minutes.
> What distribution do you use?
Ubuntu 6.06, but will install 6.10 in a week or so.
Does anybody else also have problems with cdrecord? I use SuSE 10.1 and dvd burning works fine but cd burning doesn't work , I've tried both cdrecord and nero. If someone thinks he could help I would paste the error log.
Are you using the notebook in ahci mode or normal (in the bios there is a setting called Sata native mode)? I had problems in normal mode, as the piix_ata module grabbed also parallel ata ports.
In order to set the maximum CPU frequency automatically (with F.17) on SuSE 10.1, I changed the script /etc/init.d/powersaved.
I added the following two lines just after the line starting with "startproc". A cleaner way would probably to add a new init-script, but I am too lazy.
I'd just like to let everybody know that suspend to RAM finally works with the newest kernel, 2.6.19-rc6. I have tried it with bios F.11 and F.17 using SuSE 10.1 and a vanilla 2.6.19-rc6 kernel. The SATA no-wakeup problem is definitely gone. Also, I don't experience any of the problems people usually have with ati adapters, with and without the ati fglrx driver.
Do you have noticed any keyboard problem after resume from suspend to ram? I've an nx7400, and I'm using 2.6.19-rc6. In 70% of resumes, keyboard doesn't work. I've enabled debugging in the i8042 driver on kernel command line. It lost the keyboard interrupt. I noticed that with ethernet lan up, keyboard works well after suspend, I don't know why. Using irqpoll kernel option, and reconnecting the kbd device using sysfs it works, but irqpoll is weird. I got many irq with no data after resume, using irqpoll. 2.6.19-rc6 with irqpoll neither boots (2.6.19-rc5 worked).
Now I've compiled i8042 as module. Unloading it before resume and loading it after makes keyboard work reliably on resume all the times. In windows after suspend to ram I noticed that keyboard is reenabled after few seconds, maybe it does a similar thing. May someone from hp can confirm me this? In my page http://emisca.altervista.org/nx7400 I described all my tests and configurations...
Emilio, I have no keyboard problems at all. i8042 is compiled into my kernel.
When I was experimenting with IDE-mode and suspend2ram (using an older kernel), I also experienced problems, where the internal keyboard wouldn't work anymore after a suspend2ram. (My usb-keyboard still worked.) I remember that it seemed as if unloading all alsasound modules snd-* before a s2ram would help.
I think it's a bad interaction between pci devices irq and ps2 (keyboard and touchpad) interaction. Perhaps the reason could be the same as the "bad state" problem........
I am in the process of getting 2.6.19 compiled, but I am running in some kernel panic while rebooting with my fresh 2.6.19. I think that I've missed some critial modules to be loaded into the kernel.. Could you please post you .config file in your /usr/src/linux-2.6.19 ?
Many thanks in adavance...
P.S. Lots some more hairs with this <beep> no-keyboard bug thing.
This is a common kernel compilation error. You have set some important drivers as module (needed for boot) but you have not made an initramfs/initrd. The module needed is mostly ahci and it's dependencies, and the filesystem module you use (ext3 for example). You have two chances: - compile built-in the drivers you need - figure out on how to build an initramfs image
In debian/ubuntu kernel-package creates the initrd/initramfs image without any big effort. Perhaps for other distributions you must search google...
I have upgraded bios to f.08 and i still have problems with speedstepping.
It return this error: "FATAL: Error inserting acpi_cpufreq (/lib/modules/2.6.18-1.2798.fc6/kernel/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.ko): No such device"
i checked /sys/devices/system/cpu/cpu0/ and there isnt any files of dirs which has in name cpu_freq or sth like that... :(
Do you have any deas how to fix it?
In prev bios it worked(i dont remember what version it was - f.07 i think).
For those who are interested: I installed opensuse 10.2 on my nx9420 today. Interestingly, although kernel version is 2.6.18, suspend to ram works (probably they used the recent libata patches).
Main problem: I needed to manually downgrade xorg from v. 7.2 to v. 7.1, because the ati fglrx-driver does not work with 7.2 (next time, i try to find a laptop with nvidia), and somehow i could not set up the right resolution with the framebuffer device.
Did anyone else try the new ati fglrx-driver? With X.org 7.2 (as delivered for example with opensuse 10.2) one has to use this fglrx version. And under X.org 7.2 and with framebuffer device I can't even set the resolution to 1680x1050. The problem with fglrx 8.32.5 is that it breaks suspend2ram. I.e., on resume the computer crashes. It's not only that the screen doesn't wake up, it's a real hang.
Does anyone have X.org 7.2 running? Or any knowledge about how to fix the suspend2ram with fglrx 8.32.5? Even framebuffer would be o.k. for me, if I only could suspend and resume...
At http://en.opensuse.org/ATI_Driver there is a section on "openSUSE Linux 10.2 - fglrx 8.32.5". Maybe it give you some hints... I have not tried it as I'm still with SLED 10...
The new fglrx driver works fine with my suse 10.2 regarding resolution and 3d accelaration but I have a new bug with it: when I use the driver, the notebook heats up and a minute later the fan is running very loud. This hasn't happened with older distros.
A simple workaround is to set the clock to a lower rate with the command: aticonfig --set-powerstate 1 (to check on which level you are: aticonfig --list-powerstates)
I'm hopping between distros here... I want to run Opensuse 10.2 but if I disconnect my AC the laptop just shuts down. Does anyone else have this problem?
I think this forum is better than Novell Bugzilla to treat this problem. Have you installed acpiw? In /battery/C1B2/info I have: design capacity: 4178mAh last full capacity: 4178mAh In /battery/C1B2/state I have: remaining capacity: 4038mAh With that, KPowersave give me 1:40h remaining (shorter with opensuse 10.2 than with SLED10 2h30) and I have no problem if I quit the mains.
Also, what is your Bios version? From you bugzilla post I deduce your nx9420 is in "good state"? What version of kPowerSave are you running (mine is 0.7.1)? Since 0.7.0 (Suse 10.2) kPowerSave is using HAL and NOT powersave deamon... perhaps something related to your problem... (i.e. bad install of HAL,..)
Not yet... If I have some time tomorrow, I'll try it and let you know. It's a very interesting version as it permit to enable the VT capabilities of the cpu for use in XEN or other virtualisation technologies.
I have red your opinion and I don't find it very helpful.
I have installed and used Linux on probably 7 laptops within the last ten years, among them several IBM notebooks. And let me tell you this: IBM is not necessarily better than HP (I have no experience with Lenovo). Sometimes it works, sometimes it doesn't. Only a small change in hardware (even within one series of laptops) can make your life hard.
I have an HP nx9600 running SuSE 10.1 extremely well. Power Management, accelerated graphics etc., everything great. On my nx9420 it took half a year until I got suspend2ram working, but this was not HP's problem, but the linux driver was not ready for SATA suspend. And now, especially due to the help in this forum, almost everything works fine.
One problem is still the ATI graphics card, as ATI's driver is bad (e.g., suspend2ram is broken again with newer drivers). But does IBM or Lenovo use different graphic cards? No! Have a look at thinkwiki and you will find out about all the problems users have with their Thinkpads!
The problems you experience in your Novell bug 234475 seem to be unique to your laptop, as there are no other reports about similar problems. Also you say yourself, that it may be hardware related. So you say HP has bad hardware support. Maybe, but is IBM/Lenovo better? I once had problems with a Thinkpad, where the battery created a high pitch noise and did not hold charge. I sent it in, they broke the USB port and wanted to charge me for replacing the mother board because of the broken USB port. Luckily, my company was taking care of that.
One thing I agree with you is this: The nx9420 does not (or rather did not) deserve a Linux certification. But I think the problem is with Novell which allows HP to claim that the laptop is certified.
Let's face it: hardware providers don't really care about Linux. But IBM did not care much more than HP. In my experience, before you buy a laptop, you have to find out about all the hardware in it and if you buy a laptop which has come on the market only recently, chances are that something does not work.
Anyhow, we should blame all the hardware providers for not supporting Linux very well. And first of all we should blame ATI for their bad Linux graphics drivers and for not allowing open source graphics drivers to be developed. (I would probably buy HP again, but I would try not to buy a laptop with an ATI card.) Apparently you had bad luck with your laptop, as in the beginning some things didn't work with linux and also probably it suffered from a hardware defect - tough luck. But just blaming one brand and glorifying another seems not very fair.
Anyhow, I wish you better luck with your next laptop. I'm quite happy with my nx9420, although I whish the graphics card problem would be resolved...
I really feel like HP don't care. What are the problems regarding the ATI card? The bogus temp readings I get are with SuSE's Powersaved. As I said, I'm still awaiting the new motherboard and battery and then I'll test further. But I'm giving this laptop away to a colleague of mine. The Lenovo I'm getting comes with an Nvidia card.
I've tried everything damn bios since F.11 and what is taking HP so long? they were quick to release their Vista drivers yesterday weren't they? There's just no urgency for us Linux users. Unless you want to run some commercial Unix like HP UX then they might care and sell you some expensive server.
Both parties are to blame. I bought this laptop based on those.
I'm confused how you say yours works? Does your battery status update as it should? I doubt HP even bother to read this post. What's the point of it all? We're just wasting our time on here. They're not going to release a bios that works.
Michael, it is understandable that you get frustrated if you don't get Linux properly running on your laptop. I don't know how much experience you have with Linux and laptops. But from my experience, in the beginning there are almost always some kind of hardware problems with Linux on laptops (due to missing support of the hardware providers). Unless you spend a lot of time and effort in them, it is hard to get all of them resolved, especially if you are rather "new" to Linux.
In order to bring some objectivity in the discussion, let me tell you what works with my nx9420 and why it works. As far as I can judge from Novell bugzilla and this forum, the status should be the same on all nx9420 laptops, where the hardware is not broken (as it may be the case with yours).
Maybe other readers find this helpful as well, as it is sometimes hard to find all the information from the different forums and bugzilla entries.
This refers to BIOS version F.17.
(1) Battery status + "bad state" problem. Yes, my battery status gets updated and the bad state problem is resolved. (In fact, I don't understand why you still have this problem). The bad state problem could be resolved in the beginning by compiling your own kernel with psmouse as a module and then removing this module before reboot or halt. For opensuse 10.1 and 10.2 users, there is no need for that anymore since the most recent kernels available from software.opensuse.org fixes this issue. This issue has also been marked as resolved in the corresponding Novell bugzilla entry: https://bugzilla.novell.com/show_bug.cgi?id=179702 Apparently it has also been resolved with the most recent BIOS F.19 (but I haven't verified this).
(2) CPU frequency scaling. Before BIOS F.17 there were some problems with this which are unrelated to the "bad state" problem. This is fixed with the new BIOS (and there is an ACPI fix described in this forum for older BIOS as well). You have to do a little hand tuning, though. This forum and many other nx9420 related sites, especially Emilios excellent(!) nx7400 site (link somewhere in this forum) explain how to make sure that maximum CPU frequency can be reached. Also, the recent opensuse kernels (see above) fix this problem completely.
(3) Suspend to RAM. This didn't work in the beginning because of insufficient SATA suspend support in pre 2.6.19 kernels. Clearly we can't blame HP for this. With the 2.6.19 kernel or e.g. the patched opensuse 10.2 kernel (which is still 2.6.18), this problem does not exist anymore and the machine suspends and resumes fine. However, one needs the closed source ATI driver in order to wake up the screen properly during resume. With opensuse 10.1, Xorg 6.9 (I think), and kernel 2.6.19, fglrx 8.29.6, this works without problems. However, the newer ATI fglrx driver (since 8.31.5) break suspend to RAM on this laptop. Since more recent drivers are needed for Xorg 7.2, which is installed with opensuse 10.2, no suspend to RAM is possible anymore with distro unless you downgrade Xorg by hand to an older version which is kind of difficult) and install the older fglrx drivers. Note that there is a way of getting suspend to RAM to work if you don't care about dri (if someone cares, I can describe it). But again, we should blame ATI and not HP for this.
Other than the "I want DRI + suspend to RAM" issue I am not aware of anything that does not work (I haven't tried the card reader or the fingerprint scanner). Even the DRI issue could be fixed if cared enough to downgrade to Xorg 6.9.
To summarize: If you use the right setup, everything works, unless your machine is somehow broken.
Having this in mind, have a look at one of the "linux on laptos" sites, e.g. http://www.linux-on-laptops.com Take a couple of random descriptions for laptos and see for how many of them everything works "out of the box". There are probably some, but almost all laptops need some adjustments and for some of them certain things (especially ACPI related functions and suspend2RAM) won't work reliably even after a lot of effort.
One final point: Would you be willing to pay twice as much for a laptop if Linux would be as well supported as Windows? The problem is, there are just not enough Linux users out there and developing and testing drivers is expensive. So if they don't put as much effort in supporting Linux as they do for Windows, it is understandable. But what I really don't like is if the open source community puts a lot of effort in developing their own drivers and companies like ATI put a lot of effort in stopping this...
I am rather experienced in Linux. If you were to google my name you would see that I wrote custom slmodem scripts for the Smartlink modem drivers. I've also written rpms for open source projects like Xchm, gSTM and rezlooks GTK2 engine.
I've been using Linux now for +- 6 years. I also work as a networking administrator support an 8 node VPN around South Africa. Each node has a Win2k server and Linux firewall on site.
I guess I feel stupid as I received my new battery and motherboard yesterday and all problems are fixed now. Perhaps running other distros without the patched kernel ruined the hardware? I somehow believe these patches came a but late but hopefully other people don't suffer the same problem as me.
You will see that I closed my bug report at Novel and apologised. I've just been burnt too many times. I still believe the PDF documents produced regarding the NX 8220 and NX 9420 are lies. One should rather be honest and show that there are bugs, perhaps give the user the links to the bugs so that one can stay informed. While we are the minority it's just nice to know that "we value linux users". Afterall, I paid just as much as any other Windows user would pay for this laptop.
I have not botherred with suspend or hibernate to be honest. I think I should have this laptop for another week or two before my new one arives.
Sorry for the negativity, guess I will remove my laptop page.
I guess it depends on how much effort you are willing to make. If you don't care about suspend2ram, it is actually quite easy:
(1) install opensuse 10.2 (2) install kernel-default-2.6.18.5-181. This can be found somewhere in the repository "Kernel:" at software.opensuse.org. (3) install the ATI fglrx driver (explanations how to do that can be found in the opensuse wiki).
For me, everything except suspend2ram worked after this (I am omitting some details as which packages to install for WLAN, partitioning of the hard drive etc. - feel free to ask if you don't know or if something doesn't work as expected).
If you need suspend2ram, it gets a little bit more complicated, but I can explain it if you are interested.
I've been reading the post and, sincerelly, what I'm looking for is something like this: I'm thinking on installing a Linux in my nx9420 but, which one? Can be OpenSuse 10.2 a "estable" candidate? Can someone tell me so?
Actually, looks like the bad state problem is not only still there, but it's even worse. I have the F.19 BIOS and get into the "Bad State" ACPI problem after I put my ATI radeon card into powerstate 1 (to save energy). After that, no matter what I do with psmouse, I get into "Bad State" after reboot. When I boot Linux in "Bad State" I get terribly long delays on every ACPI subsystem access (also when the kernel itself boots); sometimes I can't power on the laptop (only the "charging" LED is lit, power button does nothing); so I need to remove the battery. Sometimes (as with the "charging" LED problem) I have to remove the battery for a few minutes.
Hope some HP engineers look into this thread and can do some work. Feel free to contact me directly in case you need more info.
I run Ubuntu Edgy 6.10 on my nx9420. I've almost got everything to work, enough for me to continue with Ubuntu at least! However, some issues are still unsolved.
I still experience the "bad state" issue although I have upgraded to the newest BIOS F 19. As I have, only recently, installed Ubuntu on this machine I will have to solve the problems using the work-arounds described in this thread.
There are three issues with the "bad state" on my machine:
1) delayed BIOS loading. 2) battery status is recorded correct initially and then does not change. "Status missing" in Gnome power-manager. 3) Volume control through "soft-keys" is diabled and mic and headphone jacks do not work.
I run dual boot and these issues are solved after a reboot from windows and removing battery and power cord.
Beryl/Xgl with ATI's newest prop. driver seems to work ok. I am not a gamer though... The only problem I experience is a strange background after login and until my own desktop background loads. I have removed both Bery and Gnome splash.
Hibernate does not work. Have not tried Suspend yet. Suspect it will not work either.
I seem to have the same heat issue with the new fglrx driver as you have described. I am running Ubuntu with the 2.6.20.9 kernel, x.org 7.2 and the 8.33.6 fglrx driver.
The ATI card loads at powerstate 2 and leads to a thermal readout of TZ2 around 60 c. Changing the powerstate to 1 reduces temp of TZ2 when idling to around 46-47 c. I am unsure whether these readouts are correct or not but the relative values help me understand that changing the powerstate actually makes a great difference.
I added the changing of the powerstate to 1 as a command in the session manager, so that I do not have to run this command manually after every start.
Another interesting fact is that if TZ2 drops under 45 c the fan completely shuts down and I can enjoy silence! However, I cannot get under 46-47 c unless I lift my computer off the desk surface and allow air to flow under the computer.
I have also noticed that the reading of TZ5 follows the fan speed, in steps 0, 25 and 40.
My thermal readouts are generated using lm-sensors in Ubuntu and the sensors applet in a panel.
My question is whether the GPU scales up when neccessary or if setting a powerstate makes the frequency permanent?
Just to remind you that fglrx 8.34.8 has been out for some time now. It fixes suspend2ram. Don't know whether it makes any difference temperature-wise.
You can see on linux kernel mailing list that they have made patches for psmouse to correct the bad state problem. They are included in vanilla kernel 2.6.21-rc2. Perhaps ubuntu folks maybe have backported them to 2.6.20... I haven't tried those.
However, WLAN does not work. Suse did not detect the WLAN during install. @Peter Paul: Peter, you have offered to explain the WLAN installation on request. Could you please do this for me? I have no idea how to make WLAN work with openSuse on my nx9420. I would be grateful for some hint. Please explain at dummy level; I am a Linux newbie (electronics engineer, previous Unix experience was in 1986-88). Many thanks...
first some background: the laptop has an Intel 3945ABG wlan adapter. You need three things: 1) firmware 2) kernel module 3) userspace daemon.
openSUSE provides all these as packages, and if I remember right, all you have to do is install these. For 1) this should be "ipw-firmware" for 2) "ipw3945-kmp-default" for 3) "ipw3945d"
It's quite a while ago that I used the openSUSE packaged version of the driver (I currently install everything by hand, see http://ipw3945.sourceforge.net/ ).
Anyhow, try to install the packages I mentioned, and then run "yast lan".
If it doesn't work, try "modprobe ipw3945" and check /var/log/messages, and also whether the userspace daemon got loaded ("ps ax | grep ipw3945d")
I wanted to respond only after having some results, so it took a while (I had a tough prj deadline in between).
Here is my current status in ipw3945 on nx9420: - The procedure in ipw3945.sourceforge.net seemed too complicated for me. I am lacking background knowledge for that. - The recent openSuse 10.2 DVD obviously has already the required packages: ipw3945d-1.7.18-29 ipw-firmware-7-31 - Instead of "ipw3945-kmp-default" I found wlan-kmp-default-1_2.6.18.2_34-16 - I installed them using yast, on a fresh Suse 10.2 installation. Moreover, I installed kinternet-0.75-5. - Hooray: The WLAN works! (still without security features) I had to disable all WPA and WEP security settings on the router and in the config.
So I think I am on a good path, and hope to have solved the WPA inability soon. Thanks again for your help!
If you have hints about the WPA settings, I would be grateful. Btw., I would also be glad to find some interactive Linux tool that allows to detect active WLAN stations and to select the one to connect to. Currently I am using a second laptop running Windows for this task.
This might be a little off topic as I'm not running SuSE of any flavour, but I'd like to share my recent experiences with the long-time problems of the nx9420.
* Bad state, as stated before didn't properly go away after upgrading to F.19, however seems to have a gone a few kernel releases later, around 2.6.20, I'd say. Currently I'm running the most recent BIOS (the one after F.19) with kernel 2.6.22.1 and all seems to be perfect.
* Video card temperature. This has been written here before most probably but just in case:
aticonfig --set-powerstate 1
will set the card in low powermode resulting in only one fan working in the whole machine and at the lowest setting (under normal (read "no") load).
* Wireless problems: I have fiddled with many versions of ipw3945 along with it's regulatory daemon and the microcode and never got rid of errors and random disconnects. When kernel 2.6.22 came out a few days ago I installed the iwlwifi driver (http://intellinuxwireless.org/) and it works perfectly. I go to bed, get up in the morning and my TCP connections are still alive. Happy-Happy, Joy-Joy. And there hideous daemon is gone, too.
NetworkManager has known problems with WPA on some network cards (including this one). So you might want to try switching to the "traditional method" for your network connection (configure with yast --lan).
Printable version
