[NEED ASSISTANCE] Wifi N 5GHz mac80211 driver ini's

[derenma]

Hello all,

Quick background before I get into the gory details.. I am a linux person by nature, and a extreme nub when it comes to phone hacks, but have a moderate-to-high knowledge of building and hacking hardware. Long story short, I have about 20 different projects at the moment and don't have time to learn the quirks and file structure of android.

I have found it quite annoying that my Razr Maxx does not connect to my 5ghz/N network, and have done some prelimnary research, and it seems that the WL1285C / WL7 should support N on that high band.

Looking at the OMAP info in the OMAPpedia forums, it seems the kernel has drivers written for the 5ghz WL7. If the chip actually _supports_ 5GHz, i dunno as I cannot find the raw datasheet for the WL7 1285c

Searching around the net, and my phone, it seems that the Razr /system/etc/wlan_fem.ini is just a quickie hack of the [ source ini ] from Texas Instruments. Basically, its just missing the 5ghz configs for the mac80211 kernel module.

It seems like this is a simple hack, no? Additionally, I would try this out if I could unmount the /system directory on my phone... any quick instructions on how I can modify those contents?

Regards,
~ma

EDIT: Clarification:

derpderp. mount/remount for /system directory. regardless, i have found the ini file, but just need to get the module rebuilt i think..

EDIT: Additional data:

I should be able to rebuild the wifi firmware config on the device, but was hoping someone has a list of all the process that use the wl kernel modules...

[derenma]

Bump... with current progress. 5GHz band is enabled as far as i can tell. I will soon discover if its the modules itself that have band disabled...

For you hardcore people, there is a good guide here on what I am doing: http://omappedia.org/wiki/Android_Mac80211#WLAN_Calibration_Instruction

The 1285C on the RAZR actually uses a hybrid of files from the 1217 and 1251 (it seems that drivers and firmware are intermixed) as shown in the /system/lib/modules and /system/etc/wifi directories. Sheesh.

c:\adb shell

#>export TARGET_FW_DIR=/system/etc/firmware/ti-connectivity
#>export TARGET_NVS_FILE=$TARGET_FW_DIR/wl1271-nvs.bin
#>export WL12xx_MODULE=/system/lib/modules/wl12xx_sdio.ko
#>export TARGET_INI_FILE=/system/etc/wifi/wlan_fem.ini

#>ifconfig wlan0 down

**[turn off wifi radio in GUI]

#>rmmod wl12xx_sdio
#>cp /system/etc/firmware/ti-connectivity/wl1271-nvs.bin /system/etc/firmware/ti-connectivity/wl1271-nvs.bin.old

#>rm /system/etc/firmware/ti-connectivity/wl1271-nvs.bin

#>calibrator plt autocalibrate wlan0 $WL12xx_MODULE $TARGET_INI_FILE $TARGET_NVS_FILE

WARNING: 2 FEMS for 1 bands with autofem off looks like a strange configuration
Using nvs version 2.1
Calibrate /system/etc/firmware/ti-connectivity/wl1271-nvs.bin
Writing calibration data to /system/etc/firmware/ti-connectivity/wl1271-nvs.bin
WARNING: No MAC specified. Using random MAC!Writing mac address (null) to file /system/etc/firmware/ti-connectivity/wl1271-nvs.bin
Calibration done. FEM1 has 2 bands. AutoFEM is off. Resulting nvs is /system/etc/firmware/ti-connectivity/wl1271-nvs.bin

#>reboot

[Reboot]

[FoxKat]

First, took care of your misplaced post. Second, and please don't misunderstand, but what benefit will you gain by having N on the phone? It seems to me that the maximum communication speed of the existing hardware (the phone itself) can't even reach the maximums provided by B or G, but maybe I'm wrong. Also since N is 5GHz, in contrast to B & G which is 2.4GHz, the range will be less due to the shorter wavelength.

[derenma]

First, took care of your misplaced post. Second, and please don't misunderstand, but what benefit will you gain by having N on the phone?"

This seemed like the last thing that hasn't been 'fixed' with this phone. I have read an abundance of posts (like, i dunno, _three_ tops. ) on the interwebs that only have 5ghz networks or live in zones where the 2.4(a/n) spectrum is extremely crowded. Just because a person changes the channel, does not necessarily mean the connection will be interference free. There are so many combinations of problems out there, I can't even begin to explain all the reasons some one might want to change frequencies.

My personal reasons are so OCD I don't want to get into it at the moment.

Also since N is 5GHz, in contrast to B & G which is 2.4GHz, the range will be less due to the shorter bandwidth.

Taking into account our commercial-grade hardware, you are correct for the wrong reason. Range is actually less because transmission power is limited. The higher spectrum (and wider channels, I might add) simply give us more options.

There is one more important reason, and this should be the mantra of all any mod sites on the net: Its just because we [cleaned] can and we love doing it!

---

Philosophical stuffs aside for a second, if I could just get my hands on the TI datasheet for the 1285C (and some super hi-res pics of the mainboard) I could put this to rest for once and for all. In theory, both channels could use the same antenna and outputs. The balun and the filter cap should be exposed, so that might give a good hint as to how it is constructed. (ie: one or two antennas).

[FoxKat]

I knew there had to be SOME good reasons (besides the OCD - LOL! ) Yes, the 2.4G band is definitely over-saturated in concentrated residential areas, so that makes sense. As for the explanation or clarification of the range difference on the N protocol (i.e. 5GHz), thanks for that. I had always assumed it was a frequency penetration or reflection issue - that being shorter bandwidth lacking the ability to penetrate harder wall materials, not simply a lower power that was the culpret.

[derenma]

By Jove, I think i got it with the firmware settings update. My problem now is that VZ disabled the band selection menu in the advanced wifi menu. There was a person who did this with a Galaxy somthing-or-another over here: [APP] Wifi Band Switcher - Switch between 2.4GHz and 5.0GHz (ROOT required) - xda-developers . I decompiled that app and saw that he is actually just making API calls to WifiManager.setFrequencyBand() and WifiManager.isDualBandSupported() .

My other option is to create a full android kernel-and-friends build environment and build the iw tools and components for the RAZR and set the band through that suite. Not very time effective if it doesn't work.

I really don't feel like reverting back to a stock ICS at the moment _just_ for that one menu option. I am 90% sure that those API calls check some random internal kernel flags that are set at compile time. I could be wrong, but who knows... I think I could manually set the flags in /system/proc/sys if I knew what they were.

Anyone know any app developers that would be willing to take a stab at this? I can go over to those 'other' forums and bug them, but since I am here and spent so much time typing.... .. .. . . .