Bugzilla – Bug 129933
acpid or powersaved disable ipw2200 if power is (un-)plugged
Last modified: 2005-11-17 12:34:41 UTC
I was not sure, if this is a network or kernel bug, so please re-assign, if I chose the wrong category. I have a Samsung P35 II 1600 with Intel Centrino chipset. To be able to use the WLAN-button I need to load the asus_acpi module. According to the logs, the module is loaded without problems. However, it does not work until 5 minutes after the notebook booted into KDE and the WLAN gets disabled, if I un-plug the power. Here is how I can reproduce the issue: I start my notebook with WLAN disabled, it boots into KDE. asus_acpi is loaded and a script for EVENT_OTHER registered. Yet pressing the WLAN-button does not do anything. I have to wait for 2 minutes, then I get some powersave-warrnings (18:22 - in the logs). The WLAN-button still does not work. I have to wait another 3 minutes till the WLAN-button triggers the event (18:25 - 0000005e). The network-card works as expected. However, when I unplug the power the WLAN-button is triggered somehow, although I did not press it! (18:27:56) It might be co-incidence, but it happens at the same time kpowersave switches from the plug-icon to the battery. This is pretty bad, especially for me, as the cisco_vpn client stalls the whole system, if the network-device it uses is removed while a connection is active. Another thing I noticed, if I restart the notebook with WLAN enabled, it gets enabled (button's LED is turned on) again when powersavd and acpid are loaded during boot. Yet some moments later it gets disabled again by some mysterious action. It happens in the moment the warn-log displays: kernel: ieee80211_crypt_wep: module not supported by Novell, setting U taint flag this is executed. I will attach the logs that I get after KDE has booted. warn, messages and acpid in one file. I can also attach the boot.mdg, if it is needed.
Created attachment 55049 [details] acpi, messages and warn logs
Just a small additional note that might help. The trigger for this bug must have been introduced after SuSE 9.3, as the issue never came up with that version. The WLAN-button worked instantly and WLAN was not disabled when the power was (un-)plugged.
Just had a quick look, I expect a bug in the script or in acpi module? Could some of you have a look at that and take the Bug Holger/Timo?
Sven, you put a script to /usr/lib/powersave/scripts/ and registered it with the event other in /etc/sysconfig/powersave/events, right? If so, you did not read file:///usr/share/doc/packages/powersave/powersave_manual.html#Scripts ;-) Your script has to return to the daemon when it has finished through the script proxy located at /usr/lib/powersave/scripts/powersave_script_return. Please try this and tell us whether you still encounter both problems.
It's true that my script needed some updates, yet it is not responsible for the bug. I'll attach it anyway for you to have a look. When booting the notebook, /sys/bus/pci/drivers/ipw2200/0000:02:02.0/rf_kill is always set to 2, so that a WLAN connection cannot be established (radio off). I think that the default should be 0 in order for radio to be on. After booting, the WLAN-button is still unresponsive for about 7 minutes, even after that it is not as responsive as it should be. ct 21 20:32:03 pc192s logger: event [thermal_zone THRM 00000081 00000000] Oct 21 20:32:06 pc192s SuSEfirewall2: batch committing... Oct 21 20:32:09 pc192s SuSEfirewall2: Firewall rules successfully set Oct 21 20:32:14 pc192s kernel: ieee80211_crypt_wep: module not supported by Novell, setting U taint flag. Oct 21 20:32:14 pc192s kernel: ieee80211_crypt: registered algorithm 'WEP' Oct 21 20:32:33 pc192s hp: unable to open /var/run/hpiod.port: No such file or directory: prnt/hpijs/hplip_api.c 75 Oct 21 20:32:33 pc192s su: (to root) aphrodite on /dev/pts/3 Oct 21 20:33:26 pc192s su: (to root) aphrodite on /dev/pts/2 Oct 21 20:33:41 pc192s su: (to root) aphrodite on /dev/pts/5 Oct 21 20:34:06 pc192s su: (to root) aphrodite on /dev/pts/6 Oct 21 20:41:29 pc192s logger: event [hotkey ATKD 0000005e 00000000] Oct 21 20:41:29 pc192s logger: acpi_hotkeys_Samsung_P35 WLAN-card configured as is coming up. Oct 21 20:41:31 pc192s kernel: ipw2200: failed to send CARD_DISABLE command Un-plugging power triggers the WLAN-button event, although the button itself was not pressed, and hence disables WLAN. I remeber that Oct 21 20:46:05 pc192s logger: event [hotkey ATKD 00000027 00000000] seems to trigger 5f as well (the button). I do not know what 27 is though. If I plug the power back in, WLAN is turned on again and the brightness increased, after that WLAN is turned off again. Power un-plugged /sys/bus/pci/drivers/ipw2200/0000:02:02.0/rf_kill == 0: Oct 21 20:46:04 pc192s logger: event [hotkey ATKD 0000002e 00000000] Oct 21 20:46:04 pc192s logger: event [hotkey ATKD 0000002d 00000000] Oct 21 20:46:04 pc192s logger: event [hotkey ATKD 0000002c 00000000] Oct 21 20:46:04 pc192s logger: event [hotkey ATKD 0000002b 00000000] Oct 21 20:46:05 pc192s logger: event [hotkey ATKD 0000002a 00000000] Oct 21 20:46:05 pc192s logger: event [hotkey ATKD 00000029 00000000] Oct 21 20:46:05 pc192s logger: event [hotkey ATKD 00000028 00000000] Oct 21 20:46:05 pc192s logger: event [hotkey ATKD 00000027 00000000] Oct 21 20:46:33 pc192s logger: event [hotkey ATKD 0000005f 00000000] Oct 21 20:46:33 pc192s logger: acpi_hotkeys_Samsung_P35 WLAN-card configured as eth1 is going down. Oct 21 20:46:33 pc192s SuSEfirewall2: Warning: ip6tables does not support state matching. Extended IPv6 support disabled. Oct 21 20:46:33 pc192s SuSEfirewall2: Setting up rules from /etc/sysconfig/SuSEfirewall2 ... Oct 21 20:46:33 pc192s SuSEfirewall2: batch committing... Oct 21 20:46:33 pc192s SuSEfirewall2: Firewall rules successfully set Oct 21 20:46:35 pc192s kernel: ipw2200: failed to send CARD_DISABLE command Oct 21 20:46:35 pc192s kernel: ACPI: PCI interrupt for device 0000:02:02.0 disabled Oct 21 20:46:35 pc192s kernel: ieee80211_crypt: unregistered algorithm 'WEP' Oct 21 20:46:35 pc192s kernel: ieee80211_crypt: unregistered algorithm 'NULL' (deinit) Oct 21 20:46:37 pc192s SuSEfirewall2: Warning: ip6tables does not support state matching. Extended IPv6 support disabled. Oct 21 20:46:37 pc192s SuSEfirewall2: Setting up rules from /etc/sysconfig/SuSEfirewall2 ... Oct 21 20:46:37 pc192s SuSEfirewall2: batch committing... Oct 21 20:46:37 pc192s SuSEfirewall2: Firewall rules successfully set Power re-plugged: Oct 21 20:48:53 pc192s logger: event [hotkey ATKD 0000005e 00000001] Oct 21 20:48:53 pc192s logger: acpi_hotkeys_Samsung_P35 WLAN-card configured as eth1 is coming up. Oct 21 20:48:53 pc192s kernel: ieee80211_crypt: module not supported by Novell, setting U taint flag. Oct 21 20:48:53 pc192s kernel: ieee80211_crypt: registered algorithm 'NULL' Oct 21 20:48:53 pc192s kernel: ieee80211: module not supported by Novell, setting U taint flag. Oct 21 20:48:53 pc192s kernel: ieee80211: 802.11 data/management/control stack, 1.0.3 Oct 21 20:48:53 pc192s kernel: ieee80211: Copyright (C) 2004-2005 Intel Corporation <jketreno@linux.intel.com> Oct 21 20:48:53 pc192s kernel: ipw2200: module not supported by Novell, setting U taint flag. Oct 21 20:48:53 pc192s kernel: ipw2200: Intel(R) PRO/Wireless 2200/2915 Network Driver, 1.0.6 Oct 21 20:48:53 pc192s kernel: ipw2200: Copyright(c) 2003-2004 Intel Corporation Oct 21 20:48:53 pc192s kernel: ACPI: PCI Interrupt 0000:02:02.0[A] -> Link [LNKB] -> GSI 5 (level, low) -> IRQ 5 Oct 21 20:48:53 pc192s kernel: ipw2200: Detected Intel PRO/Wireless 2200BG Network Connection Oct 21 20:48:53 pc192s kernel: ieee80211_crypt_wep: module not supported by Novell, setting U taint flag. Oct 21 20:48:53 pc192s kernel: ieee80211_crypt: registered algorithm 'WEP' Oct 21 20:48:53 pc192s SuSEfirewall2: Warning: ip6tables does not support state matching. Extended IPv6 support disabled. Oct 21 20:48:53 pc192s SuSEfirewall2: Setting up rules from /etc/sysconfig/SuSEfirewall2 ... Oct 21 20:48:53 pc192s SuSEfirewall2: batch committing... Oct 21 20:48:53 pc192s SuSEfirewall2: Firewall rules successfully set Oct 21 20:48:54 pc192s SuSEfirewall2: Warning: ip6tables does not support state matching. Extended IPv6 support disabled. Oct 21 20:48:54 pc192s SuSEfirewall2: Setting up rules from /etc/sysconfig/SuSEfirewall2 ... Oct 21 20:48:54 pc192s SuSEfirewall2: batch committing... Oct 21 20:48:54 pc192s SuSEfirewall2: Firewall rules successfully set Oct 21 20:48:58 pc192s SuSEfirewall2: Warning: ip6tables does not support state matching. Extended IPv6 support disabled. Oct 21 20:48:58 pc192s SuSEfirewall2: Setting up rules from /etc/sysconfig/SuSEfirewall2 ... Oct 21 20:48:58 pc192s SuSEfirewall2: batch committing... Oct 21 20:48:58 pc192s SuSEfirewall2: Firewall rules successfully set Oct 21 20:49:02 pc192s SuSEfirewall2: Warning: ip6tables does not support state matching. Extended IPv6 support disabled. Oct 21 20:49:03 pc192s SuSEfirewall2: Setting up rules from /etc/sysconfig/SuSEfirewall2 ... Oct 21 20:49:03 pc192s SuSEfirewall2: batch committing... Oct 21 20:49:03 pc192s SuSEfirewall2: Firewall rules successfully set Oct 21 20:49:06 pc192s kernel: ACPI-1048: *** Warning: Failed to acquire semaphore[dff54820|1|0], AE_TIME Oct 21 20:49:06 pc192s kernel: ACPI-1048: *** Warning: Failed to acquire semaphore[dff54820|1|0], AE_TIME Oct 21 20:49:06 pc192s logger: event [hotkey ATKD 00000018 00000000] Oct 21 20:49:06 pc192s logger: event [hotkey ATKD 00000019 00000000] Oct 21 20:49:06 pc192s logger: event [hotkey ATKD 0000001a 00000000] Oct 21 20:49:07 pc192s logger: event [hotkey ATKD 0000001b 00000000] Oct 21 20:49:07 pc192s logger: event [hotkey ATKD 0000001c 00000000] Oct 21 20:49:07 pc192s logger: event [hotkey ATKD 0000001d 00000000] Oct 21 20:49:07 pc192s logger: event [hotkey ATKD 0000001e 00000000] Oct 21 20:49:08 pc192s logger: event [hotkey ATKD 0000001f 0000000f] Oct 21 20:49:13 pc192s kernel: eth1: no IPv6 routers present WLAN is off and /sys/bus/pci/drivers/ipw2200/0000:02:02.0/rf_kill is == 2.
Created attachment 55114 [details] ACPI script I am not sure why Encryption Key is set to off, if I ifup the device just once.
I just realised that Oct 21 21:02:11 pc192s logger: event [hotkey ATKD 0000005f 00000001] is also triggered randomly. :( Maybe it is some battery event? [Fri Oct 21 20:53:44 2005] BEGIN HANDLER MESSAGES [Fri Oct 21 20:53:44 2005] END HANDLER MESSAGES [Fri Oct 21 20:53:44 2005] action exited with status 0 [Fri Oct 21 20:53:44 2005] completed event "hotkey ATKD 0000005e 00000002" ^^^^ That was me This came out of the blue. [Fri Oct 21 21:02:11 2005] received event "hotkey ATKD 0000005f 00000001" [Fri Oct 21 21:02:11 2005] notifying client 6461[0:0] [Fri Oct 21 21:02:11 2005] notifying client 6540[0:0] [Fri Oct 21 21:02:11 2005] executing action "logger "event [hotkey ATKD 0000005f 00000001]"" [Fri Oct 21 21:02:11 2005] BEGIN HANDLER MESSAGES [Fri Oct 21 21:02:11 2005] END HANDLER MESSAGES [Fri Oct 21 21:02:11 2005] action exited with status 0 [Fri Oct 21 21:02:11 2005] completed event "hotkey ATKD 0000005f 00000001" Why is the number after 5f 000..01 if I used that button before, is there maybe some event that resets the counter? Normally the mentioned number increases by one, if I trigger a certain event again.
Is there any other information I can provide in order for this bug to get fixed?
I found some more info concerning the WLAN-button delay and WLAN-off on un-pluggin the power! "http://sourceforge.net/forum/forum.php?thread_id=1351917&forum_id=278158" It seems that the events are delayed by acpi_asus and that plugging the power flushes the "buffer" That is why WLAN is turned on/off when power is un-plugged or sometimes out of the blue, i.e. after a long delay.
Hello, I don't think this bug is caused by powersave scripts. I have a Samsung P30 and have been seeing erratic behavior of the ACPI buttons/hotkeys under SuSE 10.0. After some testing, I've found the following: - boot into single user mode - unload all unused modules - modprobe the acpi modules: ac battery button fan processor thermal asus_acpi - cat /proc/acpi/events Now, when pushing the Buttons (lid,power,sleep), keys (Fn+F2,F4,F5,F6,F10,Up,Down,Left,Right), and hotkeys ('i','mail','check','wlan') _not_ _faster_ than maybe 1 per second, all the right events are generated. Hitting the buttons too fast causes some events to go unprocessed. After that, one still gets one line from /proc/acpi/events for every button push, however, the line corresponds to a button which has been pushed some time before. Example: I start by pushing Fn+F2 (Battery, ID 61) 20-30 times, then go to pushing Fn+F6 (Mute, ID 32). hotkey ATKD 00000061 000000b3 <- Fn+F2 pushed hotkey ATKD 00000061 000000b4 <- Fn+F2 pushed hotkey ATKD 00000061 000000b5 <- Fn+F6 pushed ! hotkey ATKD 00000061 000000b6 <- Fn+F6 pushed hotkey ATKD 00000061 000000b7 <- Fn+F6 pushed hotkey ATKD 00000061 000000b8 <- Fn+F6 pushed hotkey ATKD 00000061 000000b9 <- Fn+F6 pushed hotkey ATKD 00000061 000000ba <- Fn+F6 pushed hotkey ATKD 00000061 000000bb <- Fn+F6 pushed hotkey ATKD 00000032 00000007 <- Fn+F6 pushed ! hotkey ATKD 00000032 00000008 <- Fn+F6 pushed hotkey ATKD 00000032 00000009 <- Fn+F6 pushed ... I have seen delays of 2, 4, 6 and 8 button pushes. As I keep pushing the buttons, the delay quickly grows to 8 and stays there. Interestingly, any kind of ACPI-event advances this 'FIFO' in the same way as above (including the AC, BATTEY, CPU (unplug/restore wall power)). Also, 'button' and 'asus_acpi' seem to be fighting over the 'sleep' button. Depending on prior button press sequence and the order in which the two modules get loaded, I've seen 'button/sleep SLPB' and 'hotkey ATKD' events generated by the same sleep-button alternatively. Beside the maximum delay of 8, the behavior is hard to reproduce. An uneducated guess: interrupts getting dropped/io-adress misalignment? I tried the vanilla version of the module (0.29 vs. 0.29-SUSE), it complies but segfaults and oopses when modprobed. The module worked perfectly under 9.3. As it controls the rf-kill switch, this is kind of a show stopper for me. Any suggestions? Does it make sense to try and fix the DSDT? I'd be happy to do the testing of any potential fixes. P.S.: On both 9.3 and 10.0, dmesg shows only one warning when the ACPI-modules are loaded: 'ACPI-0496: *** Warning: Invalid PBLK length [7]'
Thank god I am not the only one seeing this! :) I compiled asus_acpi CVS-version, which works with SuSE 10, the delay between booting and the WLAN-button actually working is a bit shorter now, the other problems remain. kill-switch is on and the buttons do not work as expected. The thing that bugs me most is that it worked in previous SuSE versions. So changing the DSDT should not make a difference. The only thing I changed in the one of the P35 is the 07 to 06 change for the CPU, which did not change anything regarding this issue. Which asus_acpi version was shipped with 9.3, maybe that one still works on 10?
This delay is not caused by asus_acpi but the Linux ACPI implementation. Thus, using the version we shipped with 9.3 will work but show the same odd behaviour as the version we ship with 10.0 :/
So is all of this caused by one app, or are these issues in fact several bugs? I can see the connection between the delays and un-plugging power flushing the "buffer", thus disabling WLAN. But what about the rf_kill switch, is that also due to the delay?
(In reply to comment #12) Sounds menacing :( Did the Linux ACPI stuff change that much between the 9.3 (2.6.11.4-21.9-default) and 10.0 (2.6.13-15-default), though?
(In reply to comment #13) That's what surprises me. The _actions_ (back light on/off, brightness up/down, rf-kill + WLAN LED, even the fan and CPU switching) are delayed as well, such that they stay in sync with the delivery of the corresponding events. I thought, they are executed in hardware by the ACPI-controller without the intervention of the kernel. Something like: you push the button, the ACPI-controller turns off the back light, and then it (the controller) tells the kernel (through an interrrupt) that the event occured. Apparently, either the controller just tells the kernel and the kernel then turns off the back light, or the controller does not execute the action untill the kernel has 'reaped' the event. Esoterisme.
(In reply to comment #13) A possible workaround is to leave the keys alone and use the files in /proc/acpi/asus/. You might be able to turn the radio on by echo 1 > /proc/acpi/asus/wled. Analogously for the backlight (lcd) and brightness (brn).
(In reply to comment #12) I've just tested it with the vanilla kernel 2.6.14 and the _SuSE_ driver from /usr/src/linux-2.6.13-15/drivers/acpi/asus_acpi.c and it works! Alas, some other things don't :(
What other things? Would it not be possible to backport the acpi-changes from .14 to the .13-SuSE kernel?
I have discovered the same error on my Samsung P30 and SuSE 10.0 (delayed keypresses and turning off WLAN by unplugging the power connect). In addition to what is already reported I have found out that this error can't be in the powersaved or acpid programms/scripts. How I checked this: I stopped both powersaved and acpid, loaded the asus_acpi module and started a cat on /proc/acpi/event: here I could discovered the same behaviour as Andreas reported. The key presses are delayed by a multiple of 2 presses. I have tried the same with SuSE 9.1 and there the output in /proc/acpi/event was as it should. Changing the DSDT didn't help. I discovered the same behaviour with the original and with the patched DSDT. Are there any plans when a newer kernel shall be provided for SuSE 10.0? Has anyone tested this behaviour with 10.1?
Does anyone of you use WEP? Because it seems that at least for me, if the ieee80211_crypt_wep module is loaded after booting, it disables the WLAN (button). Do not aks me how that can be, I just see that as soon as the entry on konsole 10 for the WEP-module comes up, the WLAN-button is disabled.
(In reply to comment #20) I use WEP. I had a look at console 10 when I booted my notebook today: - first module ieee80211_crypt_wep was loaded - then module ieee80211 - then modules ip6_tables, ip_tables and ip_conntrack - then module ieee80211_crypt_wep once more After that the WLAN button was lighted and WLAN was up and running!
With the kernel from bug 132114, I do not get the delay after booting. Yet it still happens that something sends the disable acpi-hotkey and thus the wlan-card is disabled. It cannot be due to some previously pressed buttons that are delayed, as the WLAN was up after booting without me pressing any button and went of a few minutes after. In the acpid log I see the hotke-event, as if the button was pressed, I do also see that 3 minutes later ipw2200 writes to log that it could not send the card_disable command but I have absolutely no clue where the ghost-button-event comes from.
Unfortunately my statement was wrong. There is still a delay after booting. I found out that I had "" set for EVENT_OTHER and changed it to "ignore", so which prog/script handles turning off the WLAN, if I have not only set none in EVENT_OTHER but also set it to ignore, i.e. ignore the WLAN-event and not do anything? When I have a low battery-status, the WLAN-button does not seem to work at all. Is there any progress on this bug or should I give up on SuSE 10 and switch back to 9.3?
Sven, would you mind giving it a try with a recent kernel of the day?
cc'ing Thomas
Yes this ACPI event buffering is a known problem. If all the bad issues are coming from the ACPI event buffering this is a duplicate of #61106. I mark it as a duplicate now, if there should still be other bugs causing your problems please reopen if they still exist and I could fix the event buffering. It is also known that this problem is fixed in 2.6.14 kernels. So you might want to use a kotd (not 10.0, a kernel of the day for i386/x86_64), if all other devices still work as expected you might be able to live with that. The big problem is that these changes came into the kernel with a lot whitespace and tab corrections and the patch as some ten thousand lines long ... As it works for mainline, this is not the highest priority for me, but as so much people are affected ... I do what I can. *** This bug has been marked as a duplicate of 61106 ***
I too would be intrested in trying the kernel of the day. Where do I get it?
ftp.suse.com/pub/people/mantel/kotd/i386/HEAD/kernel-default.i586.rpm
After installing the kotd, I get an error when loading the ipw2200 module. It is something like Unknown parameter: hwcrypto. I have not modified anything in /etc/modprobe.d/ipw2200.
I've installed kernel-default-2.6.14.2-20051112230006 & Co. The asus_acpi.ko in this package segfaults when modprobed. I've compiled the driver from the standard kernel (kernel-source-2.6.13-15). With this version the hotkeys work correctly. When trying to start WLAN, iwconfig complains about version missmatch of wireless extensions and missing io-controls and connecting fails:-( Are there also 'wireless-tools of the day'?
I just read on the ipw2100 irc-channel that the module before 1.0.6 and maybe even that one, sometimes restart when traffic is too high. Maybe that is another reason for WLAN being turned off.
comment #30: The asus_acpi.ko in this package segfaults when modprobed. Timo, anything we fixed there and has not been sent upstream, or is this a new bug?
Yes, I guess so. Since we're talking about a Samusung P35 it is possible that our driver is better (read: fixes we provide are not upstream). Sven, could you please put acpi_asus.c into the KOTD source and recompile this module? You should be able to load this module without oops afterwards.
To avoid confusion: Please compile the version of asus_acpi.c we're shipping with SL10.0 with the KOTD you tried.
How can I compile just that module? Last time I compiled modules using make modules, it took a very long time, because all modules where compiled. I tried to enter the dir, in that case it was usb and tried make modules, but there was no rule found. For me, the WLAN-button worked perfectly with the kotd, it worked right after booting and switched without delay on every press.
Maybe I got it wrong. The kernel package does not replace the previous asus_acpi module? I had compiled asus_acpi from its CVS but that was before installing the kernel that fixed the snd_usb bug and before I installed the kotd for this bug, so I should be using the module from the kotd and if it would segfault for me, the WLAN button would not switch LED, would it not?
Have a look at README.SUSE in /usr/src/linux (line 188). Follow steps 1-3 and then * change into $SRC_DIR/drivers/acpi * run make -C /usr/src/linux M=$(pwd) Then simply: insmod ./acpi_asus.ko
(In reply to comment #36) > Maybe I got it wrong. The kernel package does not replace the previous > asus_acpi module? I had compiled asus_acpi from its CVS Which CVS? I think the ASUS sf.net CVS does not have the fix for the Samsung P35 included. Please do not use the CVS driver but the one shipped with the kernel source package of SL10.0.
Ok, I'll do so. I used the sf.net CVS version some days before 0.29 got released, because in the forum it was mentioned that it might fix the delay problem.
Perfect. I really do not want to see someone step from 10.0 back to the dark ages (9.3) ;-) IMHO, the driver can not fix the delay since it is caused by the ACPI subsystem. Thus, using the upstream acpi_asus.c driver does not help.
Just to avoid confusion. I have reported (Comment #30) that asus_acpi in the KOTD segfaults on my *P30*. I don't know, if there is any difference between P30 and P35 as far as the hotkeys are concerned. Maybe not, since the driver recognizes the P35 as a P30 and uses the same model_data struct for both. The driver in the KOTD source (2.6.14.2-20051112230006) has ASUS_ACPI_VERSION "0.29" and segfaults on my machine as mentioned. I have taken the asus_acpi.c from SL9.3 ("0.29-SUSE", the diff is substantual) and compiled it in the KOTD tree. This version loads and works correctly. BTW, the KOTD is compiled with gcc-4.1 and SL10.0 has gcc-4.0 installed, so I had to manually change the version magic string in the driver's binary to make it load. To summarize: - A way to get the hotkeys on the Samsung P3x-series working under SL10.0 (and thus to reliably operate the rf-switch) is to use the KOTD (2.6.14+) and the asus_acpi.c from SL9.3. - I cannot connect to my WLAN router using WPA, because the KOTD uses a newer version of the 'Wireless Extensions' than the wireless-tools in SL10.0 resulting in 'unknown ioctl' errors. Question: Is there a newer version of the wireless-tools at SUSE which I can simply download and install :-)
Joe, could you please comment on the wirless-tools issue?
The KOTD has newer wireless extensions, that's right, but it's the least problem here. Kernel 2.6.14 comes with ipw2200 version 1.0.0, which has no WPA support at all. The only possibility I see here is to download and compile the latest version of ipw2200 from ipw2200.sourceforge.net.
> Kernel 2.6.14 comes with ipw2200 version 1.0.0 Is that the reason, why I cannot load the module, i.e.it does not know the parameter hwcrypto in version 1.0.0?
Yes, at least partly. It does not know about the module parameters in /etc/modprobe.d/ipw2200. You could remove this file, of course. But you also need an older firmware.
KOTD then means that: - I have to compile my own ipw-driver. - I do not get the security updates via YOU. - I have to compile the asus_acpi module. - my USB-headset does not work anymore, not even the LED blinks. I hope it is comprehensible, if I'd rather switch back to 9.3 than do all this till 10.1 is released.