Intel Wireless 2100, 2200bg, 2915bg, 3945abg, 4965agn, IWIDarwin drivers for mac os x 10.4 and 10.5 *under development* Options

[pere]

IWI 3945 rel841

Javalui! I think you passed the freeze reboot issue...
Check this Reboot log, havent start as before, but now i got a fully init without errors (i think).

Needed several reboots to get it.
Continue testing next version.

Attached File(s)

 Succesfull_init_reboot.txt ( 563.59K ) Number of downloads: 18
 manual_load.txt ( 38.82K ) Number of downloads: 6

 

[Ruud]

[attachment=26613:system.txt]

iwi4965 rel 842
scan without kp

mlazy,
the driver isn't working - all cmds timeout
while you get this the driver fail.
try to change this
queue_work(priv->workqueue, &priv->up); // maybe kp here
to
//queue_work(priv->workqueue, &priv->up); // maybe kp here

no kp in this test, but i have seen a thread error


i dont understand why im not getting first load log lines.... im using kextload

Attached File(s)

 system.txt ( 10.85K ) Number of downloads: 10

 

[jalavoui]

this is for iwi3945
this are old firmware version
if you get cmd timeout errors try with this files
mlazy - hope this works for you - your card is very hard to work
if none work delete and update from svn latest

pere,
iwl3945_mac_open Wait START_ALIVE Condition: 0ms left\
Apr 25 22:39:40 Pere kernel[0]: iwl3945: U iwl3945_mac_open Wait for START_ALIVE timeout after 2000ms.\
Apr 25 22:39:40 Pere kernel[0]: iwl3945: U iwl3945_mac_open leave - failed\
this means you can try to hack the timeout - can you try?

looks like your card start to send cmds after reboot - maybe this work for mlazy
i think you can try to add this before ret=...
IOSLeep(1000);//or other value 500, 2000
ret = UCODE_READY_TIMEOUT;
the goal is to give time to iwl3945_alive_start() completes

Attached File(s)

 ipw3945.ucode4.h.zip ( 100.72K ) Number of downloads: 24
 ipw3945.ucode3.h.zip ( 99.3K ) Number of downloads: 13

 

[Ruud]

logs are attached a few posts up

it looks like that when the scan is starting, a thread is created and this is causing an error

[myself108]

4965 r842 scans w/o kp

no scan_complete notification in logs

Attached File(s)

 system.txt ( 570.99K ) Number of downloads: 9

 

[pmorcos]

if none work delete and update from svn latest

When you say download from svn latest are you talking about the whole install package or just certain files from your list of files in the version trunk?
Can someone write a short "How to" so others can get involved in the testing.
It seems the more eyes you have on this the better.
thx
pmorcos

[pere]

this is for iwi3945
this are old firmware version
if you get cmd timeout errors try with this files
mlazy - hope this works for you - your card is very hard to work
if none work delete and update from svn latest

pere,
iwl3945_mac_open Wait START_ALIVE Condition: 0ms left\
Apr 25 22:39:40 Pere kernel[0]: iwl3945: U iwl3945_mac_open Wait for START_ALIVE timeout after 2000ms.\
Apr 25 22:39:40 Pere kernel[0]: iwl3945: U iwl3945_mac_open leave - failed\
this means you can try to hack the timeout - can you try?

looks like your card start to send cmds after reboot - maybe this work for mlazy

Yep. Where can i find that lines?
EDIT: Found

[jalavoui]

Ruud,
you're right - i fixed this in iwi3945
going to fix for 4965

pere,mlazy
in xcode -> find -> find in project

after all Ruud as find a new bug - hope this hack fixes it
iwi3945 rel 843
iwi4965 rel 844

i'm not shore about this hack...


iwi4965/3945
search in xcode for IOLog("1st scan\n");
there are several ways to hack this -> try some
the goal is to get a scan_complete msg in logs without cmd timeouts

i can't help much for today
good luck

[pere]

Ruud,
you're right - i fixed this in iwi3945
going to fix for 4965

pere,mlazy
in xcode -> find -> find in project

after all Ruud as find a new bug - hope this hack fixes it
iwi3945 rel 843
iwi4965 rel 844

i'm not shore about this hack...

Used a value of 500 for iosleep and got this log, in not sure whether it says its ok...but seems.

Attached File(s)

 reboot_log.txt ( 20.01K ) Number of downloads: 24

 

[Ruud]

oke.. somthing bad happend... my osx partition is corrupted somehow.. I have to reinstall.

Oke, tomorrow evening I'll reinstall and continue testing, but for now i can't help anymore


good luck jalavoui see you tomorrow

[Gabik123]

pere, would u be willing to write a quick command for automatically downloading iwi4965 like the one u have for 3945? I would really appreciate it

[revolt_team]

Used a value of 500 for iosleep and got this log, in not sure whether it says its ok...but seems.

Apr 25 23:27:36 Pere kernel[0]: iwl3945: U iwl3945_init_channel_map Ch. 56 [5.2GHz] RADAR WIDE DFS (0xb1 15dBm): Ad-Hoc not supported\



the name RADAR WIDE DFS is ur acces point ?

[pere]

pere, would u be willing to write a quick command for automatically downloading iwi4965 like the one u have for 3945? I would really appreciate it

Just download mine some post ago and change the adress for download... i dont know the ones for iwi 4965... :i Remember to "Save all" on file menu to avoid permissions problems on the file after change the download adress (dont save or save as). My post is for newbies like me, so sure you will be like fish on the water...
http://forum.insanelymac.com/index.php?s=&...st&p=721057

Apr 25 23:27:36 Pere kernel[0]: iwl3945: U iwl3945_init_channel_map Ch. 56 [5.2GHz] RADAR WIDE DFS (0xb1 15dBm): Ad-Hoc not supported\ the name RADAR WIDE DFS is ur acces point ?

Nop. Maybe Javalui is trying to contact the martians! XD

[myself108]

Here is my log for the 4965 r844

HACKED IOSLEEP in compatibility.cpp to 1000 and got a SCAN_COMPLETE_NOTIFICATION in log

I realize there are lines missing at the beginning, and i will test again to try get them

Eric

Attached File(s)

 system.txt ( 136.62K ) Number of downloads: 31

 

[mprewitt]

Guys, I have the Xcode 3.x and when i do the build (using pere's scripts to download the project & such).. I get 26 'warning' errors...
Many of them relate to the variables being defined but not used, or incorrect types...
Is this ok?

I am testing 844 now and will see how it goes at boot.Mark

Well, I got nothing.. I tried Network selector and it keeps giving me the
Could not get id for kernel control. 2
logout.
I assume this might be because it is not at root, so will try again.. Will also try to boot into single user mode and try...

Ok, I booted into single user mode, and my iwi3945kext isn't loading...

This means I screwed something up in the install, and now I have to go repair the disk as about every 3rd or 4th time I go single user mode, it screws up my boot blocks and I have to boot up the CD to repair the MBR.
sigh
Can someone list out the exact testing procedure for Leopard?
I used pere's where i went to 'Network' and it didn't find my wifi.. but I want to make sure that this is the correct test procedure.
Mark

[jalavoui]

iwi4965
can you get the scan_complete without the iosleep hack? this was only for iwi3945
try to find a way to make the scan loop

iwi3945
looks like iosleep fix the ucode delay error
try to hack the scan call

iwi3945 rel 845
iwi4965 rel 846
just a way to call a scan - need some hacking

[Roal2488]

i think this one is a nasty log either way her it is

Attached File(s)

 system.rtf.zip ( 28.89K ) Number of downloads: 2

 

[mlazy]

Thanks for the alternate ucode files. v4 and the original in the project both seem to work best for me. I'll keep testing. I know my card is a pain in the @SS.

[pere]

Yesterday night tried the Iosleep with several values, 500, 1000, 1500, 2000, the last value makes the microcode works 100% of the times, and i got this:
1st scan
not ready for 1st scan

Which is the way to hack this?

IOLog("1st scan\n");
if (res==0)
{
ieee80211_sta_start_scan(dev, (u8*)"<hidden>", sizeof("<hidden>"));
//local->ops->hw_scan(local_to_hw(local),(u8*)"<hidden>", sizeof("<hidden>"));
}
else
IOLog("not ready for 1st scan\n");

Tell me what to do and ill do it.

[BoboBear]

Been testing 843.

initially it gave lots of cmd timeout and ucode errors.

I increased the IOSleep for iwl3945_send_cmd_sync

ret = HOST_COMPLETE_TIMEOUT;
while(test_bit(STATUS_HCMD_ACTIVE, &priv->status)) {
IOSleep(100); //increased from 1
ret--;
if(ret==0)
break;

and iwl3945_mac_open

ret = UCODE_READY_TIMEOUT;
while(!(test_bit(STATUS_READY, &priv->status))) {
IOSleep(10); //increased from1
ret--;
if(ret==0)
break;
}

This seems to have helped. -x boot now complete scans although I still get an error with a normal boot.

 system.txt ( 892.16K ) Number of downloads: 6
 ioreg.txt ( 29.41K ) Number of downloads: 0
 dmesg.txt ( 4K ) Number of downloads: 2


I'll test 845 next