Jump to content
3992 posts in this topic

Recommended Posts

iwi3945 rev 497

please post logs

Remarks to logs:

'dmesg.txt' is empty.

The latest line of 'system.txt' right after boot was "Aug 30 21:09:00" (FYI). Then I've tried to do something with `networkSelector` but there is still no any responses except (1).

ioreg.txt

system.txt

Hi Jalavoui, here are the logs for the 497 version of iwi3945.dmg "driver". I have some observations about what is going on after installing this latest version. First, one need to clearly "uninstall" the old version because otherwise the new "driver" goes as en2 Ethernet address. Here is the way I found to do it:

a- remove the iwi3945.kext and the Extensions.kextcache and Extensions.mkext;

b- remove the full iwi3945 en1 entry in the Library/Preferences/SystemConfiguration/NetworkIntefaces.plist;

c- Delete twice (once before Restart and second after that) the PCI8086,4222 entry in the System Preferences / Network Preferences / Network Port Configuration;

d- remove with ONYX the logs (all of them);

e- Restart and don't forget to remove totaly the PCI 8086,4222 Ethernet inteface from Network Port Configuration;

f- install the new version of iwi3945.dmg then Restart and try Network Selector.

There are good news about tuning 802.11 a and g multiple chanels as you can see in the logs attached below (they were taken after launching Network Selector with option 2 "hello world..."). Bye

iwi3945 rev 498

http://code.google.com/p/iwidarwin/

 

check the logs to see what happens when using option (1) of networkselector to turn the card on/off.

it can take some time to respond because i add a timer delay.

stop posting logs without testing this - i can't use them.

 

 

moses,

 

check in ipw_associate_network():

 

priv->capability - this should be taken from the scanned network (before associating) and restored, otherwise after you try to connect to a safe net you'll fail when trying to connect to a unsecure net

 

priv->ieee->sec and priv->assoc_request - my guess you have to setup some values into this before associating

 

you should also add some validation on the rx/tx routines to ignore packets when the interface is not up - this cause some anoying bugs. maybe outputPacket() and ipw_rx().

 

try to set the debug level in iwi2200.h to define IWI_DEBUG_NORMAL

or #define IWI_DEBUG_FULL_MODE. i'm getting some bugs in the log messages but it maight help.

 

in the rx/tx we've excluded the packets without header (MBUF_PKTHDR) and the fragmented ones (with mbuf_next). i'm not shure if we need to change this for secure networks - i hope not.

 

take a look at ipw_set_hwcrypto_keys() and configu(). this will change according to the security level of the network.

 

when you upload to the download session a new developer relase remove the build folder - this way the zip will be much smaller.

 

justjoe,

the iwdarwin site is much better now - thanks :)

justjoe,

the iwdarwin site is much better now - thanks :)

My pleasure ;)

 

Let me know if you see something else you would like done differently. Also, in the source tarball, I included almost everything in the iwi2200 directory except the dmg. Will you send me a list of files/directories that definitely do not need to be included?

 

:)

iwi3945 rev 498

http://code.google.com/p/iwidarwin/

 

check the logs to see what happens when using option (1) of networkselector to turn the card on/off.

it can take some time to respond because i add a timer delay.

During that large delay my computer had a few short hang-ups with no response to mouse movement.

dmesg.txt

ioreg.txt

system.txt

During that large delay my computer had a few short hang-ups with no response to mouse movement.
can you try the on/off routine a few times and post the log?
Will you send me a list of files/directories that definitely do not need to be included?
i don't know if this is want you mean:you'll only need the files inside each main project folder (iwi2100 or iwi2200 or iwi3945 or iwi4965)for each stable version - this can be done with svn commands, but i have never try to do it.you can add this files to the tag folder, or, add a tar.gz of the folder for each stable release and a .dmg for the same release
i don't know if this is want you mean:you'll only need the files inside each main project folder (iwi2100 or iwi2200 or iwi3945 or iwi4965)for each stable version - this can be done with svn commands, but i have never try to do it.you can add this files to the tag folder, or, add a tar.gz of the folder for each stable release and a .dmg for the same release

I mean only the source files/directories which are necessary for a successful build, but there really aren't that many extras, so nevermind ;-)

I used svn commands to make the tag. That way they are 'cheap copies,' so they don't bloat the svn tree. It's really easy actually.

 

*edit* Oh, I forgot that I tagged the whole tree at r475. That's the only way Tortoise would let me do it. I am going to research tagging a specific directory using commandline or whatever. I should have it figured out before the next stable release.

moses,

 

i think we'll need to add code for the wpa_supplicant handle (check SIOCSIWAUTH in ipw2200.c)

 

 

iwi3954 rev 499

 

Card state received: HW:Off SW:On - i need to get rid of this. does anyone get a HW:ON on system.log?

i changed the delay so it will be faster this time to turn the card on/off

Hi Jalavoui, sorry for the logs uploaded last time, I didn't know you are trying to find a solution for launching the card/led. Here are the logs for the 499 version after executing option 1 on Network Selector for three times. No led activity at all and about 20 sec when I started second time under Network Selector (I think first one was turning off the card and the second was turning on again).

iwi3945 rev 500

 

the led is not supposed to work - it's not important for now since it is only usefull for blinking.

i think i find some bug in the interrupt routine.

good thing is that the rfkill=0x1 means that the card is turned on - i'll need to find a way to hack this to allow to turn on/off.

my guess is that the wirless switcher can change the rfkill value - test it if you have it.

 

iwi2100/4965

can someone post some logs?

 

 

moses,

the mbuf functions you're using in ieee80211_crypt_wep might not work well - i have some problems with them in the past. you should check them and try to replace/use the ones we've created in iwi2200.h

iwi3945 rev 500

Not sure got I clear or not - tried to press h/w switch before each launch of `networkSelector` tool, also tried to press it inside `networkSelector` (no state changes), and every time I tried to select option (1) for a few times.

 

jalavoui please give me some clear instructions to do with that stuff if you need. Cheers.

dmesg.txt

ioreg.txt

system.txt

jalavoui please give me some clear instructions to do with that stuff if you need. Cheers.

 

use the switcher until you see in logs rfkill=0x1 this is the "good" value for turning the card on. test it a few time and post the log

if the switcher don't work use option (1) of networkselector.

you should try all this with the console open and with the system.log selected so that you cann see what happens.

 

thanks.

use the switcher until you see in logs rfkill=0x1 this is the "good" value for turning the card on. test it a few time and post the log

if the switcher don't work use option (1) of networkselector.

you should try all this with the console open and with the system.log selected so that you cann see what happens.

 

thanks.

jalavoui,

My h/w switch doesn't appears in the logs, I've been trying it for many times, as within networkselector, as without it.

Then, option (1) only for the first time give me that:

kernel[0]: iwi3945: RFKILL base status: 0x1
kernel[0]: iwi3945: Trying to turn card off...

The card was turned ON after I was making last log I posted here before.

Other tries to select option (1) give me only this:

kernel[0]: iwi3945: RFKILL base status: 0x701

and rfkill stays unchanched always until I quit networkselector.

Then I pressed h/w switch, started networkselector and rfkill chanched to 0x700.

So, it looks like there is no suitable info I can give to you?

Feel free to ask something else I could do for you.

Hi Jalavoui, here are the logs of the version 500. Good news about starting the card with clear indications and exact recognition of the Intel WireLess 3945a/b/g card, but no activity. I tried several times with Network Selector, but first I tried with option 1 for starting and stoping the card.

iwi3945 rev 502

Bad luck on this - when trying to turn card on, networkSelector immideately gets 100% cpu usage and system cannot kill its process (used systemMonitor and kill -9) until computer totally hangs up. :) During second try I've made logs before system stuck but it looks like there is nothing about the problem, anyway I'll post it.

 

I have knoppix bootable cd - shall I be able to check that linux driver or it's needed linux to be installed completely?

dmesg.txt

ioreg.txt

system.txt

Hi Jalavoui, there are good improvements and also the same observations as those mentioned by Evil_Cartman, but when pressing the combination Fn+F2 (where is located the antenna), the load was diminished to normal without any hang up of the system. I am uploading the logs for several tries with Network Selector, but also some captures taken on About This Mac (one can notice the Auto Discovery - No) and also from Network Utility Info and NetStat (Inactive). Hope it helps a bit. I would like to install the Kubuntu 7.04, but takes a lot to download and install, so about those news in few days.

jalavoui,

as I told you logs are almost empty. Here's all lines of system.log (filtered by "iwi") right after last launch of networkSeleclor till it gets 100% cpu and hangs:

Sep  3 18:32:53 EvilCartman kernel[0]: iwi3945: connect
Sep  3 18:33:19 EvilCartman kernel[0]: iwi3945: RFKILL base status: 0x0
Sep  3 18:33:19 EvilCartman kernel[0]: iwi3945: Trying to turn card off...
Sep  3 18:33:19 EvilCartman kernel[0]: iwi3945: radio off CSR_UCODE_DRV_GP1 0x6 CSR_UCODE_DRV_GP2 0x0 rfkill 0x0
Sep  3 18:33:19 EvilCartman kernel[0]: iwi3945: ipw going down 
Sep  3 18:33:19 EvilCartman kernel[0]: iwi3945: stop master
Sep  3 18:33:19 EvilCartman kernel[0]: iwi3945: 0 frames on pre-allocated heap on clear.
Sep  3 18:33:24 EvilCartman kernel[0]: iwi3945: RFKILL base status: 0x0
# here I started getlogs command after waiting about 2 minutes of networkSelector's hangup
Sep  3 18:35:32 EvilCartman sudo:	 root : TTY=unknown; PWD=/; USER=root; COMMAND=/usr/bin/grep iwi3945
Sep  3 18:35:32 EvilCartman sudo:	 root : TTY=unknown; PWD=/; USER=root; COMMAND=/usr/bin/grep iwi3945
Sep  3 18:35:33 EvilCartman kernel[0]: iwi3945: getState = 1e
# here goes reboot
Sep  3 18:37:37 localhost kernel[0]: darwin_iwi3945: Ethernet address 00:13:02:33:dd:fe
Sep  3 18:37:37 localhost kernel[0]: iwi3945: ifnet_t en4 = 28e5204
Sep  3 18:37:37 localhost kernel[0]: iwi3945: ifconfig up
Sep  3 18:37:37 localhost kernel[0]: iwi3945: ifconfig going up

Guest
This topic is now closed to further replies.
×
×
  • Create New...