Intel Wireless 2100, 2200bg, 2915bg, 3945abg, 4965agn

[Ruud]

Hey all, i wanted to thank you all for your help getting this driver working. Jalavoui you are doing great work overhere.

 

I wanted to tell that I can no longer help testing. Today i received a atheros 5008 mini pcix at my work and this one works very well so i replaced my 4965.

 

I wish everyone good luck with this project.

[nordix]

iwi3945 rel 924 -> no KP, but strange and "lazy" work.

 

Jalavoui, what's there in the "modified" folder?

system.txt

[myself108]

here is my log for the 4965 v925... no kernel panic, but log also shows no lines that show scanning

system.txt

[pere]

iwi3945 rel 924

 

No ucode errors on boot

No timeouts.

Got en1 not running.

 

To all of you, dont get weak now, lets get close to Javalui, he is the only hero!!

boot.txt

[jalavoui]

good to see less bugs - i'll post a new scan call. this is called in enable() if there's no previous errors in driver

 

iwi3945 rel 926

iwi4965 rel 927

 

don't worry i won't give up

[adamisnt]

I am at work right now, and I don't think there are any wireless networks in range. Therefor, I am not getting a lot of the scan/reply code that I see in pere's logs.

 

Is it pointless for me to test with no networks in range?

[myself108]

here is my log for 4965r927

 

it now has lines showing the driver scanning, and i also get a scan_complete notification

 

Edit: I attached the compiled kext for testing

system.txt

iwi4965.kext.zip

[adamisnt]

iwi4965 rel 927

 

Lots of stuff in this one. I see it scanning (note: there are no wifi networks in range as far as I know)

 

Haven't seen this before:

May 2 11:47:53 adam-notebook kernel[0]: last_scan_jiffies

 

Nice:

May 2 11:48:23 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_handle r = 119, i = 118, SCAN_COMPLETE_NOTIFICATION, 0x84

May 2 11:48:23 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_scan_complete_notif Scan complete: 11 channels (TSF 0x001F782A:00000000) - 1

May 2 11:48:23 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_scan_complete_notif Scan pass on 2.4GHz took 70ms

 

Keep it up, Jalavoui

systemlog_927.txt

[i_Geek]

Nice:

May 2 11:48:23 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_handle r = 119, i = 118, SCAN_COMPLETE_NOTIFICATION, 0x84

May 2 11:48:23 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_scan_complete_notif Scan complete: 11 channels (TSF 0x001F782A:00000000) - 1

May 2 11:48:23 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_scan_complete_notif Scan pass on 2.4GHz took 70ms

 

isn`t this good ?!

[pere]

iwi 3945 rel926

 

Cant make the card start without the error:

May 2 21:12:55 Pere kernel[0]: error iwl3945: Error sending REPLY_RXON: time out after 500ms.

May 2 21:12:55 Pere kernel[0]: error iwl3945: Error setting new configuration (-60).

 

But i have found new things even the log is not correct.

 

May 2 21:12:59 Pere kernel[0]: iwl3945: U iwl3945_mac_config_interface bssid: 00:00:00:00:00:00

May 2 21:12:59 Pere kernel[0]: iwl3945: U iwl3945_mac_config_interface leave - scanning

May 2 21:12:59 Pere kernel[0]: scanend stations

May 2 21:12:59 Pere kernel[0]: 1

May 2 21:12:59 Pere kernel[0]: 2

May 2 21:12:59 Pere kernel[0]: 3

 

Im still trying other iosleep values.

[nordix]

iwi3945 rel 926 -> no KP, same results as from Pere's log

system.txt

[jalavoui]

i hope pere can fix the cmd timeouts

in iwi4965 there's no problem with that

the driver scans (maybe with wrong bands)

the problem with association starts with

ieee80211_rx_mgmt_beacon()

this can be the gateway to start associating to a network

 

i'll try to hack the driver to associate when it gets a call to ieee80211_rx_mgmt_probe_resp()

 

the goal is to do this and stop the scan loop

 

let's try it...

 

iwi3945 rel 928 - only works if you don't get cmd timeouts

iwi4965 rel 929

[adamisnt]

iwi4965 rel 927

 

Rebooted my machine again, got some different results. Wanted to note the differences.

 

This time around it is scanning a/b/g networks as opposed to just 802.11a.

 

Didn't get this last time:

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode Got firmware 'iwlwifi-4965-1.ucode' file (2d5a4d bytes) from disk

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode f/w package hdr ucode version = 0x42c0114

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode f/w package hdr runtime inst size = 92720

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode f/w package hdr runtime data size = 40960

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode f/w package hdr init inst size = 10312

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode f/w package hdr init data size = 40960

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode f/w package hdr boot inst size = 788

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode Copying (but not loading) uCode instr len 16A30d

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode uCode instr buf vaddr = 0x0x345ef000, paddr = 0x05731000

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode Copying (but not loading) uCode data len A000dVGA: family specific matching fails

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode Copying (but not loading) init instr len 2848d

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode Copying (but not loading) init data len A000d

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_read_ucode Copying (but not loading) boot instr len 314d

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_hw_nic_init HW Revision ID = 0x61

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_hw_card_show_info 4965ABGN HW Version 0.0.209

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_hw_card_show_info 4965ABGN PBA Number 5593006

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_hw_nic_init SW RF KILL supported in EEPROM.

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_hw_nic_init HW RF KILL supported in EEPROM.

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_load_bsm Begin load bsm

May 2 13:00:19 adam-notebook kernel[0]: iwl4965: U iwl4965_verify_bsm Begin verify bsm

 

Also noticed this. It appears in my other r927 log as well.

 

May 2 13:00:29 adam-notebook kernel[0]: ieee80211_sta_rx_scan @ /Users/Adam/Desktop/root/compatibility.cpp:3465

May 2 13:00:29 adam-notebook kernel[0]: ieee80211_rx_mgmt_beacon @ /Users/Adam/Desktop/root/compatibility.cpp:3418

May 2 13:00:29 adam-notebook kernel[0]: ieee80211_rx_bss_info @ /Users/Adam/Desktop/root/compatibility.cpp:3061

May 2 13:00:29 adam-notebook kernel[0]: IOPCCard debug: en1: RX Beacon from 00:18:71:5f:13:f6 to ff:ff:ff:ff:ff:ff

systemlog_927_2.txt

[adamisnt]

iwi4965 rel 929

 

Wow. Lots of new stuff here. Looks very promising.

 

May 2 14:01:19 adam-notebook kernel[0]: iwl4965: U iwl4965_tx_queue_set_status Activate AC Queue 0 on AC 3

May 2 14:01:19 adam-notebook kernel[0]: iwl4965: U iwl4965_tx_queue_set_status Activate AC Queue 1 on AC 2

May 2 14:01:19 adam-notebook kernel[0]: iwl4965: U iwl4965_tx_queue_set_status Activate AC Queue 2 on AC 1

etc

 

May 2 14:01:19 adam-notebook kernel[0]: iwl4965: U iwl4965_init_channel_map Initializing regulatory info from EEPROM

May 2 14:01:19 adam-notebook kernel[0]: iwl4965: U iwl4965_init_channel_map Parsing data for 56 channels.

May 2 14:01:19 adam-notebook kernel[0]: iwl4965: U iwl4965_init_channel_map Ch. 1 [2.4GHz] IBSS ACTIVE WIDE (0x2b 14dBm): Ad-Hoc supported

May 2 14:01:19 adam-notebook kernel[0]: iwl4965: U iwl4965_init_channel_map Ch. 2 [2.4GHz] IBSS ACTIVE WIDE (0x2b 14dBm): Ad-Hoc supported

etc

 

 

May 2 14:01:19 adam-notebook kernel[0]: iwl4965: U iwl4965_init_channel_map Ch. 3 [2.4GHz] IBSS ACTIVE WIDE (0x2b 14dBm): Ad-Hoc supported

 

SCAN_COMPLETE for 2.4Ghz AND 5.2 Ghz

May 2 14:01:55 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_scan_complete_notif Scan complete: 13 channels (TSF 0x0035DB4D:00000000) - 1

May 2 14:01:55 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_scan_complete_notif Scan pass on 5.2GHz took 1470ms

May 2 14:01:55 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_scan_complete_notif Setting scan to off

May 2 14:01:55 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_scan_complete_notif Scan took 1530ms

May 2 14:01:55 adam-notebook kernel[0]: iwl4965: U iwl4965_bg_scan_completed SCAN complete scan

May 2 14:01:55 adam-notebook kernel[0]: ieee80211_scan_completed

 

Get the 'scan to off', but scanning still loops as far as I can tell.

 

5 minutes later I got this: (not in log)

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_bssid_match @ /Users/Adam/Desktop/root/compatibility.cpp:1416

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_invoke_rx_handlers @ /Users/Adam/Desktop/root/compatibility.cpp:1065

May 2 14:07:51 adam-notebook kernel[0]: __ieee80211_invoke_rx_handlers @ /Users/Adam/Desktop/root/compatibility.cpp:1035

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_h_if_stats @ /Users/Adam/Desktop/root/compatibility.cpp:4338

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_h_monitor @ /Users/Adam/Desktop/root/compatibility.cpp:4346

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_h_passive_scan @ /Users/Adam/Desktop/root/compatibility.cpp:4353

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_sta_rx_scan @ /Users/Adam/Desktop/root/compatibility.cpp:3477

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_mgmt_beacon @ /Users/Adam/Desktop/root/compatibility.cpp:3430

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_bss_info @ /Users/Adam/Desktop/root/compatibility.cpp:3061

May 2 14:07:51 adam-notebook kernel[0]: IOPCCard debug: en1: RX Beacon from 00:18:71:5f:13:f6 to ff:ff:ff:ff:ff:ff

May 2 14:07:51 adam-notebook kernel[0]: ieee802_11_parse_elems @ /Users/Adam/Desktop/root/compatibility.cpp:2846

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_bss_get @ /Users/Adam/Desktop/root/compatibility.cpp:2958

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_bss_add @ /Users/Adam/Desktop/root/compatibility.cpp:2989

May 2 14:07:51 adam-notebook kernel[0]: __ieee80211_rx_bss_hash_add @ /Users/Adam/Desktop/root/compatibility.cpp:2979

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_bss_put @ /Users/Adam/Desktop/root/compatibility.cpp:3042

May 2 14:07:51 adam-notebook kernel[0]: __ieee80211_rx_bss_hash_del @ /Users/Adam/Desktop/root/compatibility.cpp:3012

May 2 14:07:51 adam-notebook kernel[0]: ieee80211_rx_bss_free @ /Users/Adam/Desktop/root/compatibility.cpp:3032

systemlog_929.txt

[myself108]

4965 r 929 log

 

still seems to be looping the scan

system.txt

[jalavoui]

iwl4965_mac_config_interface bssid: 00:18:71:5f:13:f6

need a litle hack on this

wait...

[nordix]

iwi3945 rel 928 -> first time freeze (shot below), then ok, but still have 500ms timeout. How can I make the time for waiting RXON_REPLY larger, more than 500ms? Maybe, 500ms is not enough?

system.txt

[jalavoui]

let's move on

 

iwi3945 rel 930

iwi4965 rel 931

 

there's still some things to fix - the driver needs to add the network we want to associate to the stations table

you can see that in logs when you get a iwl4965_mac_config_interface

 

nordix,

don't mess with the code

 

pere,

have you find a fix?

 

looks like the driver works diferent in linux

the scans/associate requests are made out of the driver

i'm trying to make this work inside iwi3945/4965 - as i've done for iwi2200

so if this works the driver will associate to the first non secure network it finds

if i manage to do this i know nobody will complain :mellow:

[pere]

I have been playing with the previous version here:

 

static int iwl3945_send_cmd_sync(struct iwl3945_priv *priv, struct iwl3945_host_cmd *cmd)

{

int cmd_idx;

int ret;

static atomic_t entry = ATOMIC_INIT(0); /* reentrance protection */

 

BUG_ON(cmd->meta.flags & CMD_ASYNC);

 

If i change 0 to 1, the it says there is already a command send... maybe there is one command doenst unqueue?,

and that causes the next command dont load?

 

Going to check this version.

[myself108]

log for 4965 r931

Shows adding to the stations table

system.txt

[jalavoui]

I have been playing with the previous version here:

 

static int iwl3945_send_cmd_sync(struct iwl3945_priv *priv, struct iwl3945_host_cmd *cmd)

 

this is why i disabled the atomic_xchg on previous releases - maybe you can try it

 

 

iwi4965

iwl4965_add_station_flags Add STA ID 0: 00:1c:f0:40:1c:22

nice - this means station 0 is in the table

maybe i can use networkselector to allow user to choose the net to associate

meanwhile i'll try a easy hack

[adamisnt]

iwi4965 rel 931

 

First:

May 2 14:58:31 adam-notebook kernel[0]: iwl4965: U iwl4965_mac_config_interface enter: interface id 592703520

May 2 14:58:31 adam-notebook kernel[0]: iwl4965: U iwl4965_mac_config_interface bssid: 00:00:00:00:00:00

May 2 14:58:31 adam-notebook kernel[0]: iwl4965: U iwl4965_commit_rxon Sending RXON

May 2 14:58:31 adam-notebook kernel[0]: * without RXON_FILTER_ASSOC_MSK

May 2 14:58:31 adam-notebook kernel[0]: * channel = 1

May 2 14:58:31 adam-notebook kernel[0]: * bssid = 00:00:00:00:00:00

 

Then:

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_mac_config_interface enter: interface id 592703520

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_mac_config_interface bssid: 00:18:71:5f:13:f6

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_commit_rxon Sending RXON

May 2 14:59:21 adam-notebook kernel[0]: * without RXON_FILTER_ASSOC_MSK

May 2 14:59:21 adam-notebook kernel[0]: * channel = 1

May 2 14:59:21 adam-notebook kernel[0]: * bssid = 00:18:71:5f:13:f6

 

Finally:

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_add_station_flags Add STA ID 0: 00:18:71:5f:13:f6

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_enqueue_hcmd Sending command REPLY_ADD_STA (#18), seq: 0x0493, 72 bytes at 147[19]:4

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_handle r = 168, i = 167, REPLY_ADD_STA, 0x18

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_rx_reply_add_sta Received REPLY_ADD_STA: 0x01

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_send_cmd_sync Wait REPLY_ADD_STA Condition: 49ms left

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_send_add_station REPLY_ADD_STA PASSED

May 2 14:59:21 adam-notebook kernel[0]: iwl4965: U iwl4965_add_station_flags Add STA ID 0: 00:18:71:5f:13:f6

 

It looks like it is looping adding the station

systemlog_931.txt

[nordix]

iwi3945 rel 930, first time - freeze, then norm, but still timeout

system.txt

[x4hu]

you're getting awesome logs, this driver is almost working thats like 80% i would say.. great work jalavoui **clap**

[pere]

Iwi 3945 rel 930

 

I have manage to start the card without the rx on timeout changing iosleep here:

 

if ( pmPCICapPtr )

{

fPCIDevice->saveDeviceState();

fPCIDevice->configWrite16( kPCIPMCSR, 0x8000 );

IOSleep(100); // wait for internal reset completion

fPCIDevice->restoreDeviceState();

 

Before was 10.

But now i dont see on my log what i saw in the last 3 versions....

 

Nordix, try it and post log. So we can compare if the hack is correct.

reboot.txt

iwi3945.kext.zip