K brb, Im sure, it didnt even finish building the kext. Thanks in advance for helpin out, Reppin Halifax, NS here

For the bug fixes (3945) under Linux I found these links:

http://www.mail-archive.com/git-commits-he...g/msg33809.html
and this:
http://www.mail-archive.com/git-commits-he...g/msg41153.html
Hope it helps

One change is
u32 count = 0; to u32 count = 8;
The other hack is more complicate: http://git.kernel.org/?p=linux/kernel/git/...94ae316c6b79a03

I wasnt able to make a kext, I looked in that directory you mentioned. here are two screenies

The Error


The warnings

thats odd... because the warning you got was for the installer script not getting a password to run a command as root... i am using xcode 3.0 in leopard 10.5.1 and even though i get the same error myself it still builds the kext, just to be clear the /build/debug/ directory will be within the directory that contains the .xcodeproj file... for you that would be the /users/doctadre/desktop/latest wireless kext/build/debug/

Ah, stupid me got the files mixed up. Alright So Ive got the build, but Im pretty sure something is wrong according the initial part of the logs. My logs do not look the same as they did with the chun nan kexts before I Added another 15 gigs to my tiny leopard partition for the xcode. Anyway, here is the iwi3945.kext rel952 for those who want to test. Thanks alot for helping out. Going to look over my old logs now to compare initial log

This is the line that is different
May 3 19:05:08 doctadres-mac-pro Terminal[342]: _CFGetHostUUIDString: unable to determine UUID for host. Error: 35

The chun nan kexts should have fixed this

@ Doctadre

i am pretty sure that the
May 3 19:05:08 doctadres-mac-pro Terminal[342]: _CFGetHostUUIDString: unable to determine UUID for host. Error: 35
line has nothing to do with the kext itself as it appears before you load the kext

perhaps you are correct. In my old logs I used to get this line

Apr 28 22:01:40 doctadres-mac-pro kernel[0]: IOPCCard info: iwl3945: Detected Intel PRO/Wireless 3945ABG Network Connection

lookin good jala. signing out for tonite i'll be available for testing/hacking tomorrow morning. Glad ive got xcode now

Here are my 3945 logs, one after loading, another after reboot.
Hope they help somehow

Click to view attachment
Click to view attachment

the driver can't hack the assoc like this
when it gets a probe resp it should auth/assoc
let it scan and check the changes in system.log
i think this is easy to fix in code - can someone try to hack?

Cxddev,
can you compare the iwi tx auth/assoc packet against the linux
i think there's a bug in iwi tx code
or better
set linux debug to 0xffffffff
this way we can see what the driver does when it get a beacon/probe response
can someone post this for 3945?

iwi3945 rel 953
iwi4965 rel 954

febelus
build with full debug and post system.log (put #define IWI_DEBUG_FULL_MODE in iwi2200.h)

iwi2200 problem!!!!



i've leopard 10.5.1 installed on sony vaio vgna-497xp all is ok except wireless.
dual boot with windows xp.
wireless switch installed on xp side.
Wireless state :
"on" wireless switch xp side
"on" wireless external button

i've installed iwi2200 driver as http://forum.insanelymac.com/index.php?showtopic=71779, kernek recognize my wifi card and discovery wifi network but when i try to connect to us the log report this:

iwi2200: radio on 0x50000 = 0x10000
May 4 13:21:08 fabio-angelettis-mac-pro kernel[0]: iwi2200: geography 2 = ZZD
May 4 13:21:16 fabio-angelettis-mac-pro kernel[0]: iwi2200: network cap 0x421 flags 0x402
May 4 13:21:16 fabio-angelettis-mac-pro kernel[0]: iwi2200: AUthenticated: 'paperetto' 00:14:a9:6a:cc:30
May 4 13:21:16 fabio-angelettis-mac-pro kernel[0]: iwi2200: associated: 'paperetto' 00:14:a9:6a:cc:30
May 4 13:21:16 fabio-angelettis-mac-pro kernel[0]: iwi2200: E Firmware error detected. Restarting.
May 4 13:21:17 fabio-angelettis-mac-pro kernel[0]: iwi2200: geography 2 = ZZD
May 4 13:21:18 fabio-angelettis-mac-pro kernel[0]: iwi2200: network cap 0x421 flags 0x402
May 4 13:21:18 fabio-angelettis-mac-pro kernel[0]: iwi2200: AUthenticated: 'paperetto' 00:14:a9:6a:cc:30
May 4 13:21:18 fabio-angelettis-mac-pro kernel[0]: iwi2200: associated: 'paperetto' 00:14:a9:6a:cc:30
May 4 13:21:20 fabio-angelettis-mac-pro kernel[0]: iwi2200: E Firmware error detected. Restarting.
May 4 13:21:20 fabio-angelettis-mac-pro kernel[0]: iwi2200: geography 2 = ZZD
May 4 13:21:24 fabio-angelettis-mac-pro kernel[0]: iwi2200: network cap 0x421 flags 0x402
May 4 13:21:24 fabio-angelettis-mac-pro kernel[0]: iwi2200: E Firmware error detected. Restarting.
May 4 13:21:24 fabio-angelettis-mac-pro kernel[0]: iwi2200: geography 2 = ZZD
May 4 13:21:28 fabio-angelettis-mac-pro kernel[0]: iwi2200: network cap 0x421 flags 0x402
May 4 13:21:29 fabio-angelettis-mac-pro kernel[0]: iwi2200: E Firmware error detected. Restarting.
May 4 13:21:29 fabio-angelettis-mac-pro kernel[0]: iwi2200: geography 2 = ZZD

and continue until i disable the wifi card.

what's the problem????
i've try 910 and 760 release and all have the same problem!!!!!

thx

fabio

waiting on some kext to post logs

Alright jala, here are my iwi3945 logs. What hacks can I try for you? Here is that line that you referenced to cxddev:

May 4 08:54:47 doctadres-mac-pro kernel[0]: iwl3945: U iwl3945_mac_tx dev->xmit(342 bytes) at rate 0xffffffff
May 4 08:54:47 doctadres-mac-pro kernel[0]: error iwl3945: ERROR: No TX rate available.

Coldest Temperature Ive seen yet.

May 4 08:56:41 doctadres-mac-pro kernel[0]: iwl3945: U iwl3945_hw_reg_txpower_get_temperature Temperature: 46

Also attached is the 3945 kext

iwi3945 rel 953 -> 2 reboots

very nice log - maybe all it takes is a few reboots

can you explain this:
iwl3945_add_station Add STA ID 0: 00:16:ca:f5:49:71
iwl3945_add_station Add STA ID 0: 00:14:d1:31:95:ca
iwl3945_add_station Add STA ID 0: 00:16:ca:f5:49:70

are this nets near you?

i'm suspecting that the driver don't scan all channels - please check this

but after you reboot you get timeouts and the driver stop working

theory for iwi3945
if the driver gets a _pmPowerState on them it's ready to work
let's try it

iwi3945 rel 955


iwi4965 rel 956
this is another way to scan


anyone to post logs?

@nordix, your logs are different than mine

@jala, whats wrong with my logs?

I have at least 3 public wifi networks in range of my card and my router ASUS WL-500G is on. Will try more reboots and post results within 30 minutes. Wait a bit, please

I think this is a good thing? 954logs + kext

iwl3945_add_station Add STA ID 0: 00:1e:2a:72:51:ec
report if you see in logs ALL scaned nets near you

Well, 3 reboots
1st - with kext
Then I delete kext and -> 2nd reboot
Then I install it again -> 3rd reboot

It works without timeouts only after manual replacement and only before next reboot.

Update: Maybe, simple kextunload-kextload will give the same result. I'll try it.

How can I check to be sure? There is only one connection near me. I think this one is it. The numbers line up, but the letters are different

report the mac address you have in iwi3945 and the ones you know are right

nordix
i don't understand your log
is it from last release?
the kextunload don't work - never use it

maybe the driver only works
after booting mac os x
and manual load with kextload

Jalavoui, it's from the last release.

Manual kextunload gives me freeze, but after reboot I have good logs. So, only complete remove-install of the kext can give me normal results.

I have to go, be back in an hour or so...

Wireless Mac address: 00-19-D2-BE-49-86
Bluetooth Mac address: 00-1A--6B-2D-DF-23

iwi 3945 955

Manual load, first 7 reboots, frozen on init_channel_map
Then restored the iosleep hack, reboot, no errors, no timeouts, just the jiffies error.
Second reboot after this, got ucode errors.
Post log from sucessfull reboot.

should i try this?:

try this in all statments
while (time_after(priv->next_scan_jiffies, jiffies)==1) {IOSleep(10);}
IWL_DELAY_NEXT_SCAN, jiffies)) {IOSleep(10);}

them try this
while (time_after(priv->next_scan_jiffies, jiffies)==0) {IOSleep(10);}
IWL_DELAY_NEXT_SCAN, jiffies)) {IOSleep(10);}

The scanned bbsid are correct, all are the real mac adresses from that stations.
But seems is just scanning channel 1. NO other mac adresses from other channels.

Yep iosleep is correct, but i will try to make it lower, is really huge i think.

...do a scan in windows,linux etc
copy to a text file the mac address of the scanned nets
use iwi3945 and check in log for the sta mac address it finds
if iwi mac address are wrong the driver will never work


this hack is already enabled
i'll enable the iosleep and commit to svn
looks like the driver is finally stable
i'll add the other way of scanning for testing
can you check if the iwi scaned mac addresses are correct?
does iwi3945 scan ALL nets? or just some
pere - the isleep(6000) is correct?

iwi3945 rel 957

iwi4965 rel 958
add the pwoer hack for testing

What I got in the build 954: 00:1e:2a:72:51:ec
mac addy scan in windows with scan range of 192.168.0.1 up to 999:

192.168.0.1 <No reverse response > 00:13:46:A2:21:72 D-Link Corporation
192.168.0.100 drew 00:12:17:71:FE:20 Cisco-Linksys, LLC

Here is the iwi3945rel957, still trying to get a good screen

I give up, It all happens too fast for me to get a good screen, 10+ times now. Near Immediate KP. Anyone know how to get logs upon restart?

can anyone atach iwi4965.kext please ?

edit: nvm doc you have 3945 didnt notice :x

3945 debian syslog with debug=0xffffffff

this log helps a lot
thanks

the iwi drivers must be very good:
firmware->size = sizeof(ipw3945_ucode_raw); //149652;//crappy
May 4 20:01:50 puter kernel: iwl3945: U iwl3945_read_ucode Got firmware 'iwlwifi-3945-1.ucode' file (149652 bytes) from disk
the bug starts here - but - it works

cxddev,
to use HZ=250 you'll need to change qeue_xx, jiffies and others
are you shure this is the best way?

here is the 4965 r958 kext
logs on the way

edit:
Here are the logs... There are 4 reboots, second one -x boot, i get a kp every time

Here's my alternative version logs. Incidentally, I also changed to the latest firmware.

I seem to get errors in the boot and intialise code but this is after it reports alive, so I really don't understand that. Odd, that they should choose to check the ucode after it's already started up ok.

I don't know if I can help much with the code discussions, I'm usually on the other side of the hardware/software divide so my knowledge of soft stuff like drivers is limited. ;-)

iwi3945 _ rel.955

no kp as i compiled.
and seems to be stable for long time.
but:
i tryed to scan with networkselector, compiled as well, but it seems doesn't scan.
i also tried to put a static address instead of dhcp, and as soon as i apply, it KPs.
this is done with eth0 off.
is this normal?

here the log.

what is the 'other way' to scan that you mentioned?

there's still a lot of work to make this work
if you manage to find a way to make the driver make the scan on all channels
and add new stations maybe it works
i'll try to help on this...

looks like the power hack don't work
fix for the kp in scan call

iwi3945 rel 959
iwi4965 rel 960

How bout the latest 957 rel, do you get a kp?

don't post old logs

Cxddev,
maybe you can figure the solution for this
i'm trying to make the driver scan all channels
looks like old way don't work
can you check it?

nordix,pere
i think the code is not the same for the timeouts fix
can you check - i might need to revert some changes

again logs coming but for now here is the kext
4965 r960

edit: Here are the logs... 3 reboots... second -x, kp every time

OOPS, I have only 957 log, KP's... Will try 959, wait...

Is that why you enter with ieee80211_sta_req_scan now?
I think ieee80211_open works ok? it asks for sta_scan anyway later...

I thought it wasn't scanning all channels either but then realised there were big gaps in the time stamps and broken log messages so I think it is covering all channels, but it's not being reported (as usual).

I tried putting IOSleep in between to fix this but it KPs as it's all IRQ routines I think? Need another way of checking perhaps? Add to an array of scanned channels? I'll see what I can do...

iwi3945 rel 959 -> only freezes (several reboots)

BTW, mac address of "station 0" in my log 953 is real

what bug?

wrote wrong code...
fixed

iwi3945 rel 961
iwi4965 rel 962


you can check the scan channels in logs in this line:
_commit_rxon
_mac_config enter to channel

Sousuke,
can you bring any light to this?

iwi3945 rel 961 -> immediate KP after kext installation, then only freezes

4965 r962

get kp still
here is the log

the kext is attached

i'll try to put back the code...

same here I get a kp as well

the code is ok
post the kp picture

Well, I have nothing to post, as my KP's occur only after blue screen loading (I mean, loading and freezing). I reboot with -v -f always.

I have try to make it kp, but just got frozen, then i booted with the hardware switch turned off, and i could get a kp,
the last line printed before the kp was the code line 5185 on compatibility.cpp, related to this:

int ieee80211_hw_config(struct ieee80211_local *local)
{
IM_HERE_NOW();
struct ieee80211_hw_mode *mode;
struct ieee80211_channel *chan;
int ret = 0;

if (local->sta_scanning) {
chan = local->scan_channel;
mode = local->scan_hw_mode;
} else {
chan = local->oper_channel;
mode = local->oper_hw_mode;

Sorry, no pic.
}

thanks pere - i am expecting this to work

iwi3945 rel 963
iwi4965 rel 964

to build iwi3945.dmg

read 1st post
open terminal
cd to iwi3945 folder...
sudo rm iwi3945.dmg
sudo xcodebuild iwi3945.xcodeproj
this instruction should be add to 1st post - this fix the error while build in xcode for non root acounts
you don't need the dmg. xcode will place iwi3945.kext in system/library/extensions
reboot and goto sys prefs -nets to add en1 if not already

Here are my shots of KP's