ipw3945 wireless drivers

[zappala]

beatryder, I also had a similar problem with a version of this driver, and I fixed it by removing ipw3945d from the default runlevel and adding it to the boot runlevel instead. This gave it enough time to startup before net.eth1 started.

[Joebel]

beatryder, I had the same or very similar problem and solved it by adding a little sleep in /etc/init.d/ipw3945d right before the end of start().

Code: Select all

start() {
        check
        ebegin "Starting ipw3945d"
        chown ipw3945d /sys/bus/pci/drivers/ipw3945/00*/cmd
        chmod a-w,u+rw /sys/bus/pci/drivers/ipw3945/00*/cmd
        start-stop-daemon --start --exec /sbin/ipw3945d --pidfile ${PIDFILE} -- \
                --pid-file=${PIDFILE} ${ARGS}
        sleep 0.5   # <--- here
        eend ${?}
}

[post=3832423]Here's[/post] some more nfo.

Hi, I'm the guy who's having the same problem as beatryder. I read that thread to, and tried this solution. Alas, to no avail.

I even tried "sleep 1.5" --> nothing.

Thanx for the suggestion, though. Very much appreciated.

Joebel

[Joebel]

beatryder, I also had a similar problem with a version of this driver, and I fixed it by removing ipw3945d from the default runlevel and adding it to the boot runlevel instead. This gave it enough time to startup before net.eth1 started.

That was a good idea. Seems that it doesn't work for all systems: it doesn't solve it on my system, unfortunately.

Thanks for the suggestion though.

Joebel

[morbus]

So did actually any of you guys try the still highly experimental iwlwifi driver for the ipw3945 chipset? It's developed by Intel and does not require a binay-only daemon (They moved the critical parts to the firmware I think). Nevertheless it requires the devicescape wlan stack which is not yet in the mainline kernel

Link: http://bughost.org/iwlwifi/

[Lloeki]

This seems highly interesting. I'll take a look at it once I get some time on my hands.
May I suggest creating a separate thread similar to this one once we get some progress/reports, so that things are not messed up on this thread?

details from planet.gentoo.org:

The other large topic was the merge of the new wireless stack (d80211) which was donated to the community by Devicescape. The stack has many more features than the current in-kernel ieee80211/softmac stack, and improves consistency between drivers. We are now much closer to getting d80211 merged into mainline Linux, but there are still some significant technical issues to be resolved.

Before the conference, several people asked me if I could raise issues with the suckiness of Intel’s ipw3945 driver offering in that it relies on this awkward closed-source userspace daemon. Fortunately Intel have found a way to rewrite their firmware so that the regulatory domain control is no longer at the driver level, therefore a fully open-source driver with no binary userspace part has been developed (iwlwifi). Although Intel have not yet announced it, the source code has been released here. This rewritten driver uses the d80211 stack mentioned above, so it’s not possible just to compile it against mainline Linux at the moment. The iwlwifi website includes instructions for how to check out the wireless-dev tree where d80211 is being held.

and from kerneltrap.org:

Quick glance at the code shows that iwlwifi is based on ipw3945 driver and is still a bit complex compared to OpenBSD wpi driver.

EDIT: d80211 patched and built without issues against kernel suspend2-2.6.19-r1, microcode copied in /lib/firmware, driver compiled seemingly without issues, loading it causes a bad lock (keyboard gets stuck, mouse still moves, after some minutes, things (namely, X) start crashing). conclusion: for now, iwlwifi is HIGHLY UNSAFE to use.

[slackthumbz]

Hi, I've been having a strange issue with my ipw3945 lately. I'm able to install and have the ipw3945d starting on boot (had to edit the init script and remove the --pidfile option because the ipw3945d thought that it was being passed to the iface daemon rather than the start-stop daemon and crashed) however I can run Kismet fine and it detects networks but when I try to dhcpcd to a network I always get timeouts. Also when I manually set the essid (iwconfig eth1 essid "Whatever" I've found that then running iwconfig without args shows that the last character is missing from the essid so it would be "Whateve"...

Anyone had anything similar?

[beatryder]

beatryder, I also had a similar problem with a version of this driver, and I fixed it by removing ipw3945d from the default runlevel and adding it to the boot runlevel instead. This gave it enough time to startup before net.eth1 started.

That was a good idea. Seems that it doesn't work for all systems: it doesn't solve it on my system, unfortunately.

Thanks for the suggestion though.

Joebel

This solution did not work for me either.

As for the idea of downgrading to <=udev-098 that is not really an option to me. I dont understand why I should downgrade from a *stable* package?

[VinzC]

I dont understand why I should downgrade from a *stable* package?

Not because it is unstable but because its *functionnality* has deeply changed. As I said >=udev-098 is incompatible with coldplug: you have to remove it. Plug and play works in a completely different manner as of udev-098.

[VinzC]

This seems highly interesting. I'll take a look at it once I get some time on my hands.
May I suggest creating a separate thread similar to this one once we get some progress/reports, so that things are not messed up on this thread?

[topic=539015]Done[/topic]

[beatryder]

I dont understand why I should downgrade from a *stable* package?

Not because it is unstable but because its *functionnality* has deeply changed. As I said >=udev-098 is incompatible with coldplug: you have to remove it. Plug and play works in a completely different manner as of udev-098.

I don't understand then. I have been using udev <= 098 with the ipw3945-1.1.13 and everything worked just fine. Now it has been removed from portage, and I am hosed.

I understand why you say the older udev is the one that should be used. Surely there has to be a better solution.

[VinzC]

To me that's the whole plug'n'play thing that got hosed. I don't understand why udev got released though neither blacklisting nor module insert/remove instructions worked. But that's another story.

I must admit I expect much from einit or whatever comes first.

[sLoDkI]

Hi all.
Sorry for complaining about things maybe explained here, but there is lots of posts to go through, so i could miss sth.
I'm trying get my ipw3945 on dell e1405 to work for a few days, but still no success. Could someone take a look? Everytime i compile all stuff (patched or not) when all things gone fine i finally don't have wireless inteface visible/accessible. I've tried solutions with previous kernels - the same. Actually i have 2.6.20 + ipw3945-1.1.3_for_2.6.20.patch with in-kernel ieee80211 stack. My results:

Code: Select all

[root@inspirebox /usr/src/linux] dmesg | grep ipw
ipw3945: Intel(R) PRO/Wireless 3945 Network Connection driver for Linux, 1.1.3mpr
ipw3945: Copyright(c) 2003-2006 Intel Corporation
ipw3945: Intel(R) PRO/Wireless 3945 Network Connection driver for Linux, 1.1.3mpr
ipw3945: Copyright(c) 2003-2006 Intel Corporation
[root@inspirebox /usr/src/linux] dmesg | grep ieee
ieee80211: 802.11 data/management/control stack, git-1.1.13
ieee80211: Copyright (C) 2004-2005 Intel Corporation <jketreno@linux.intel.com>
ieee80211_crypt: registered algorithm 'NULL'
ieee80211_crypt: registered algorithm 'WEP'
ieee80211_crypt: registered algorithm 'CCMP'
ieee80211_crypt: registered algorithm 'TKIP'
ieee1394: Host added: ID:BUS[0-00:1023]  GUID[424fc0001805ed41]

[root@inspirebox /usr/src/linux] emerge -av ipw3945d ipw3945-ucode ipw3945 ieee80211
[ebuild   R   ] net-wireless/ipw3945d-1.7.22-r4  0 kB
[ebuild   R   ] net-wireless/ipw3945-ucode-1.14.2  0 kB
[ebuild   R   ] net-wireless/ipw3945-1.2.0  USE="-debug" 0 kB
[ebuild  N    ] net-wireless/ieee80211-1.2.16  USE="-debug" 0 kB

[root@inspirebox /usr/src/linux] /etc/init.d/ipw3945d restart
 * Stopping ipw3945d ...                                                                                                                               [ ok ] 
 * Starting ipw3945d ...
chown: nie ma dostêpu do `/sys/bus/pci/drivers/ipw3945/00*/cmd': Nie ma takiego pliku ani katalogu
chmod: nie ma dostêpu do `/sys/bus/pci/drivers/ipw3945/00*/cmd': Nie ma takiego pliku ani katalogu 

[root@inspirebox /usr/src/linux] lsmod
Module                  Size  Used by
ipw3945               112480  0
i915                   21056  3
ohci1394               31824  0
ieee1394               83892  1 ohci1394
uhci_hcd               20812  0
ehci_hcd               28364  0

[root@inspirebox /usr/src/linux] iwconfig
lo        no wireless extensions.
eth0      no wireless extensions.

[root@inspirebox /usr/src/linux] ls -la /sys/bus/pci/drivers/ipw3945
razem 0
drwxr-xr-x  2 root root    0 lut 12 15:37 .
drwxr-xr-x 17 root root    0 lut  8 21:18 ..
--w-------  1 root root 4096 lut 12 15:37 bind
-rw-r--r--  1 root root 4096 lut 12 15:37 debug_level
lrwxrwxrwx  1 root root    0 lut 12 15:37 module -> ../../../../module/ipw3945
--w-------  1 root root 4096 lut 12 15:37 new_id
--w-------  1 root root 4096 lut 12 15:37 unbind

[root@inspirebox /usr/src/linux] ls -la ..| grep linux
lrwxrwxrwx  1 root root       14 lut  8 14:04 linux -> ./linux-2.6.20

There's no /sys/bus/pci/drivers/ipw3945/00*/cmd and no eth* pointing to wifi device.
Some hints i can check?
Actually i don't know which way to choose.. prev. kernels / external ipw / ndsiwrapper / or try to find solution for this one.

I've cleaned out modules, and did other things you've discussed here, the only thing i don't make was the script /etc/init.d/ipw3945d - but i thing it will be a problem when the card runs.

Code: Select all

[root@inspirebox /usr/src/linux] uname -a
Linux inspirebox 2.6.20 #1 SMP Thu Feb 8 12:00:50 Local time zone must be set--see zic m i686 Intel(R) Core(TM)2 CPU         T5600  @ 1.83GHz GenuineIntel GNU/Linux

[sLoDkI]

Ok, I realised I have ipw3945-1.2.0 emerged and ipw3945-1.1.3 from kernel, so I've done emerge -C 1.2.0 version, rm -rf /lib/modules/2.6.20/, rebuild and reinstall kernel modules - still no efect: module loaded, no /sys/bus/pci/drivers/ipw3945/00*/cmd files and no wireless interface.

[Lloeki]

I can't recall for /sys/.../cmd stuff, but the eth device sure won't show up until ipw3945d is started.

[sLoDkI]

I know:) When I coment the chown and chmod lines out in /etc/init.d/ipw3945d - there is no errors while startng it, but (as always) no eth interface available.

[rmh3093]

Ok, I realised I have ipw3945-1.2.0 emerged and ipw3945-1.1.3 from kernel, so I've done emerge -C 1.2.0 version, rm -rf /lib/modules/2.6.20/, rebuild and reinstall kernel modules - still no efect: module loaded, no /sys/bus/pci/drivers/ipw3945/00*/cmd files and no wireless interface.

after you modprobe ipw3945 and start the daemon, type lsmod.... if the used by column is 0 for the ipw3945 module then that means there was probably a firmware error... type dmesg and look for any ipw3945 errors, i bet it says it couldnt load the firmware

[beatryder]

For all those interested in the older modules

Get them here:

http://neucode.org/gentoo/ipw3945.tar.bz2

[sLoDkI]

after you modprobe ipw3945 and start the daemon, type lsmod.... if the used by column is 0 for the ipw3945 module then that means there was probably a firmware error... type dmesg and look for any ipw3945 errors, i bet it says it couldnt load the firmware

The main problem is that there are no errors. Take a look:

Code: Select all

[root@inspirebox /usr/src/linux] lsmod
Module                  Size  Used by
i915                   21056  3
ohci1394               31824  0
ieee1394               83892  1 ohci1394
ehci_hcd               28364  0
uhci_hcd               20812  0

[root@inspirebox /usr/src/linux] modprobe ipw3945
 * Starting ipw3945d ...                   [OK]

[root@inspirebox /usr/src/linux] /etc/init.d/ipw3945d start
 * WARNING:  ipw3945d has already been started.

[root@inspirebox /usr/src/linux] lsmod | grep ipw
ipw3945               112480  0

[root@inspirebox /usr/src/linux] dmesg | grep ipw
ipw3945: Intel(R) PRO/Wireless 3945 Network Connection driver for Linux, 1.1.3mpr
ipw3945: Copyright(c) 2003-2006 Intel Corporation

[root@inspirebox /usr/src/linux] tail /var/log/messages
Feb 14 10:02:58 inspirebox ipw3945: Intel(R) PRO/Wireless 3945 Network Connection driver for Linux, 1.1.3mpr
Feb 14 10:02:58 inspirebox ipw3945: Copyright(c) 2003-2006 Intel Corporation
Feb 14 10:03:25 inspirebox rc-scripts: WARNING:  ipw3945d has already been started.

Everytime I modprobe ipw3945 - it automaticly starts ipw3945d daemon. I have commented out the chmod and chown lines in /etc/init.d/ipw3945d beacuse it always says:

Code: Select all

 [root@inspirebox /usr/src/linux] /etc/init.d/ipw3945d restart
 * Caching service dependencies ...                                                                                        [ ok ] 
 * Stopping ipw3945d ...                                                                                                        [ ok ] 
 * Starting ipw3945d ...
chown: cannot access `/sys/bus/pci/drivers/ipw3945/00*/cmd': No such file or directory
chmod: cannot access `/sys/bus/pci/drivers/ipw3945/00*/cmd': No such file or directory            [ ok ]

My daemon script:

Code: Select all

#!/sbin/runscript
# Copyright 1999-2006 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
# $Header: /var/cvsroot/gentoo-x86/net-wireless/ipw3945d/files/ipw3945d-init.d,v 1.4 2006/12/22 10:11:26 phreak Exp $

PIDFILE=/var/run/ipw3945d/ipw3945d.pid

depend() {
   before net
}

check() {
   # Let's check if the pidfile is still present.
   if [ -f "${PIDFILE}" ] ; then
      eerror "The pidfile ($PIDFILE) is still present."
      eerror "Please check that the daemon isn't running!"
      return 1
   fi
}

start() {
   check
   ebegin "Starting ipw3945d"
#  chown ipw3945d /sys/bus/pci/drivers/ipw3945/00*/cmd
#  chmod a-w,u+rw /sys/bus/pci/drivers/ipw3945/00*/cmd
   start-stop-daemon --start --exec /sbin/ipw3945d --pidfile ${PIDFILE} -- \
      --pid-file=${PIDFILE} ${ARGS}
   sleep 0.5
   eend ${?}
}

stop() {
   ebegin "Stopping ipw3945d"
   start-stop-daemon --stop --exec /sbin/ipw3945d --pidfile ${PIDFILE}
   eend ${?}
}

[Joebel]

For all those interested in the older modules

Get them here:

http://neucode.org/gentoo/ipw3945.tar.bz2

They work, although when under heavy load (sftp and the likes) the ipw3945 stutters and dies..

Remains weird though: having a completely working network setup--> portage deletes those ebuilds, introduces new ones (now even "stable")--> no working setup.

[beatryder]

For all those interested in the older modules

Get them here:

http://neucode.org/gentoo/ipw3945.tar.bz2

They work, although when under heavy load (sftp and the likes) the ipw3945 stutters and dies..

Remains weird though: having a completely working network setup--> portage deletes those ebuilds, introduces new ones (now even "stable")--> no working setup.

Yeah, that pissed me off too. I actually have considered changing distros now. The only thing that has kept me around is the fact that I know it won't be much better if I go elsewhere and I know gentoo so well now that it would more work to switch.

[rmh3093]

idk why you people have such and issue getting this driver to work...

[Lloeki]

although when under heavy load (sftp and the likes) the ipw3945 stutters and dies..

look in dmesg for TKIP replays and mickael mic digest failure.

[beatryder]

idk why you people have such and issue getting this driver to work...

Neither do we, the only solution posed to me was to downgrade udev, which is obviously not the problem or the old drivers would not work either.

[VinzC]

Everytime I modprobe ipw3945 - it automaticly starts ipw3945d daemon.

If you see the daemon run automatically once the module is loaded then you don't need ipw3945d init script because the latter was provided only for those cases where the insert/remove instructions in /etc/modules.d/ipw3945d are ignored by the system. So remove ipw3945d (emerge -C ipw3945d) and install the stable ipw3945 package. (I'd expect troubles running the daemon twice.)

EDIT: As I posted later on , to my knowledge package ipw3945d with the init script is in the ~ARCH branch while the deamon without the init script in in the stable branch. So just emerge the stable ipw3945d.

And don't forget to rc-update -d ipw3945d...

[VinzC]

They work, although when under heavy load (sftp and the likes) the ipw3945 stutters and dies..

Remains weird though: having a completely working network setup--> portage deletes those ebuilds, introduces new ones (now even "stable")--> no working setup.

Yeah, that pissed me off too. I actually have considered changing distros now. The only thing that has kept me around is the fact that I know it won't be much better if I go elsewhere and I know gentoo so well now that it would more work to switch.

Rather drastic a solution . Once I ran into such troubles I completely removed the kernel source tree and modules and re-did a fresh and clean compile with teh appropriate patches. Never had a glitch ever since.