[Q] Wifi Error

[thegogetter-]

How do I exactly do this? Where do I find the dhd.ko file?

 

[dLo GSR]

How do I exactly do this? Where do I find the dhd.ko file?

if you open up any kernel CWM, it's in system/lib/module

 

[thegogetter-]

Did it work for you? I'm still lost. Can I do it w a root explorer?

 

[gold_sprinter]

I too am having the same wifi/Bluetooth problems. My phone reboots constantly, and have recently discovered that my camera is not working either. I have tried multiple kernels, ROMS, firmwares and various combinations of each.

I am simply looking to see if I am able to get back to my original Stock Bell firmware. I have gone through numerous threads trying to download the SBF files to flash in RSD. Unfortunately most links are dead, or have been removed.

I am now unable to flash in CWM as the zip file download from here has rendered my phone unbootable (bootloop), also cannot enter recovery mode. All I get is an android, triangle and exclamation mark.

My question is....Does anybody have a working link to download the French SBF and the Bell SBF files so I can return to my stock firmware. If this doesn't get my phone working again, at least it will be back to stock adn can send it away for warranty.



Thanks is advance

 

[dude772]

Wifi/ bluetooth on one of my atrix also.

That youtube trick didn't help me at all.

 

[TheTinkerer]

My Wifi and BT have been gone for weeks now - I figured out what cokesprite meant by his post, so I'm in the process of changing kernels to the faux123's "enhanced stock kernel" from this thread for Gingerbread.

I'll post back here if I am successful, and give detailed instructions.

---------- Post added at 04:41 PM ---------- Previous post was at 03:49 PM ----------

Failure. I still get the ERROR message when attempting to turn on Wifi.

Steps I took:

1. Downloaded new kernel,
2. Moved kernel to my SD card
3. Rebooted to CWM
4. Installed kernel ZIP
5. Rebooted
6. Tried to enable WiFi - still get the error
7. Unzipped kernel .zip on my PC, copy to C:\Android (where I keep adb and fastboot), renamed to dhd.ko_new
8. Backed up my old file - "adb pull /system/lib/dhd.ko"
9. Renamed old file to dhd.ko_old
10. Remounted as read/write "adb remount -rw"
11. Pushed new dhd.ko to device - "adb push dhd.ko_new /system/lib/dhd.ko"
12. Rebooted device
13. Tried to enable WiFi
14. Cried

What's next?

 

[dhionisis]

youtube

i do not know you people butt i had the same problem ,wifi error and no bluetooth and guess what i did what exactly he did on the video and i power on the phone without the cover and the phone is back to normal with one exemption that i did that few times and it seams to work with active sim card on the phone.this is not a joke.i do not know what the battery has to do with the wifi but it works.

 

[djkmito]

unfortunately I add to you
my wifi and bluetooth doesnt work too, my rom CM9 and kernel is from jokersax

 

[eksempel_123]

I bought the phone from ebay, a while ago. And now this happened to me as well (on stock!)
However, Motorola have the best support ever. Even tho I am not ATT customer or live in the US, they said my device was under warranty and I could send it to a repair shop in my country. The phone is bootloader unlocked, so I might actually have to pay for the repairs tho :/

 

[djkmito]

Have anyone here definitive answer?
Is it hardware or software issue?

 

[simbill]

its a hardware issue. Had the problem, and it got struck in a reboot cycle. I had to let the phone completely die (to the point it would NOT power on at all) for the reboot loop to cease. I opened my phone, and the metal piece going over and around the "battery" looking piece on the upper back left is broken. Hence the broken wifi/bt. its the chip. shot. I dont dare try and turn wifi or bt on ever again. Its a shame, had to upgrade my grandfathered 2gb data plan to the 3gb because without wifi im nearing my cap month to month. waiting for the Galaxy Note to drop a little more and ill pick one up.

---------- Post added at 06:15 PM ---------- Previous post was at 05:36 PM ----------

hardware

 

[Dcskeete]

Hi, I just got the same Wi-Fi error issue.
International atrix, enhanced stock kernel, and 4.9.141 updated.

The root cause is dhd.ko is not matching with the kernel, cannot be probed successfully.
From the faux123's thread which published kernels,
along with the boot.img, there is dhd.ko also packed in the zip package.
Just push it to /system/lib, and reboot, then Wi-Fi can be activated.

How do you "just push it to/system/lib"

 

[mvniekerk]

Worked - once, twice.

1. Did the detach-battery spiel
2. Wiped Dalvik cache
3. Ran Fix-Permissions

Upon boot-up the WiFi/BT didn't work.
Then I rebooted, then it worked.
Downloaded some stuff.
After putting the phone down, it stopped working.
Power cycled (on->off->on, not a reboot), it worked.
Let it go to sleep, stop working.

Stopped working ever since. Random reboots on phone calls now as well, even with WiFi/BT disabled.

 

[mvniekerk]

More analysis

[ 31.028143] sdhci_tegra_wlan_detect Detecting WLAN
[ 31.028164] mot_wifi_set_carddetect: Nobody to notify
[ 31.029391]
[ 31.029395] Dongle Host Driver, version 4.218.248.32
[ 31.029400] Compiled in /tank/jokersax/cm9/kernel/moto/olympus/wifi-module/wlan/osrc/open-src/src/dhd/linux/../.. on May 7 2012 at 11:40:09
[ 31.328158] mmc1: error -110 whilst initialising SDIO card

WARNING: at /tank/jokersax/cm9/kernel/moto/olympus/lib/kobject.c:595 kobject_put+0x50/0x60()
[ 214.627572] kobject: '<NULL>' (0003da74): is not initialized, yet kobject_put() is being called.
[ 214.627581] Modules linked in: dhd(+)
[ 214.627591] Backtrace:
[ 214.627612] [<c0040010>] (dump_backtrace+0x0/0x10c) from [<c05b68c8>] (dump_stack+0x18/0x1c)
[ 214.627622] r7:c031c6c4 r6:c071fb48 r5:00000253 r4:e83a1e98
[ 214.627647] [<c05b68b0>] (dump_stack+0x0/0x1c) from [<c00bc5b8>] (warn_slowpath_common+0x50/0x68)
[ 214.627663] [<c00bc568>] (warn_slowpath_common+0x0/0x68) from [<c00bc61c>] (warn_slowpath_fmt+0x30/0x38)
[ 214.627673] r7:bf081000 r6:00000000 r5:bf076570 r4:0003da74
[ 214.627692] [<c00bc5ec>] (warn_slowpath_fmt+0x0/0x38) from [<c031c6c4>] (kobject_put+0x50/0x60)
[ 214.627701] r3:00000000 r2:c071fb80
[ 214.627720] [<c031c674>] (kobject_put+0x0/0x60) from [<c03787f0>] (put_driver+0x14/0x18)
[ 214.627729] r4:e98390c0
[ 214.627742] [<c03787dc>] (put_driver+0x0/0x18) from [<c0378938>] (driver_register+0x144/0x164)
[ 214.627757] [<c03787f4>] (driver_register+0x0/0x164) from [<c037994c>] (platform_driver_register+0x4c/0x60)
[ 214.627767] r9:00000000 r8:00000000 r7:bf081000 r6:00000000 r5:e83a0000
[ 214.627782] r4:bf078890
[ 214.627826] [<c0379900>] (platform_driver_register+0x0/0x60) from [<bf049fa4>] (wifi_add_dev+0x24/0x40 [dhd])
[ 214.627870] [<bf049f80>] (wifi_add_dev+0x0/0x40 [dhd]) from [<bf0810bc>] (init_module+0xbc/0x238 [dhd])
[ 214.627902] [<bf081000>] (init_module+0x0/0x238 [dhd]) from [<c003b5c8>] (do_one_initcall+0x40/0x1bc)
[ 214.627912] r5:e83a0000 r4:c07e0c00
[ 214.627932] [<c003b588>] (do_one_initcall+0x0/0x1bc) from [<c00f3880>] (sys_init_module+0xcc/0x200)
[ 214.627941] r9:e83a0000 r8:c003c364 r7:00046bd8 r6:00000000 r5:bf076748
[ 214.627957] r4:5c126008
[ 214.627970] [<c00f37b4>] (sys_init_module+0x0/0x200) from <c003c1e0>] (ret_fast_syscall+0x0/0x2c)
[ 214.627979] r7:00000080 r6:5e0fdb34 r5:5c126008 r4:405ad0f0
[ 214.627995] ---[ end trace 2416c978061dd5a5 ]---
[ 214.628002] Error: Driver 'bcm4329_wlan' is already registered, aborting...

Now I had some experience with WiFi kernel driver development (iMx51 + CSR based WiFi/BT combo chip).
The startup sequence on kernel level will probably go like this:
1) Kernel startup
2) ARM CPU specifics startup
3) Platform specific device setup (setup clocks)
3) Platform driver dependency registration (like, on SDIO 1 there is a WiFI chip that needs driver X to work)
4) Sub-platform driver startup (SPI, SDIO, UART, I2C)
I think this is where we fail first - Error -110 means timeout. Each MMC / SDIO chip has a MMC specific startup sequence, then depending what was given on number 3, it jumps over to SDIO (SDIO is a superset of MMC).
When our BT/WiFi chip is put into SDIO mode, the SDIO commands fail with a time out.

Possible reasons for it failing:
1) Incorrect voltages
2) Skewed clock
3) Busted chip.
4) We talk to it too fast
5) Kernel is messed up

Now, seeing that (through weird loops, check my previous post) I could get the unit running two more times, I do not think it is 3.

Why it could be 1:
Voltage rails needs to start up at the correct sequence for some chips. Like some chips will specify an operating range of between 1.8V and 3.3V. That means if it finds 5V on it, it will shut itself off. Sometimes some chips needs an initial 3V, then after 10 seconds it can be put into 1.8V (in-chip startup). Sometimes, it needs 10seconds to stabilize before you can talk to it (in our case via SDIO).
Why it isn't 1:
I've tried a lot of other kernels. It is all based of the ATT code, and they do the voltage rail setup correctly. I've even reverted back to stock code - no dice.

Why it could be 2
By removing the small cell battery, the drain on the battery was lifted. It could then store up enough current to put itself into the correct clock rate when it was put back.
Why it isn't 2
? Don't know. More needs to be known of the function of this battery. Obviously it is driving some RTC (real time clock). RTC's normally work at 32MHz, (32768, easilly divideable by 2, 4, 8 etc).
Does the Broadcom chip need this RTC? Isn't its clock lines driven by the Tegra2?

Why it could be 4
If the RTC is compromised (number 2) then init will suffer. If the Tegra's clock is not derived from the battery and the Broadcom's is, then there will be a disparity between the 2 units chatting together, as SDIO communication is clock based and not latch based (like SPI). If I'm chatting to you at 32MHz and you are 5% off, then we will not be communicating (if you are an SDIO master and I'm an SDIO slave).
Why it isn't
No idea.

Why it could be 5
Bunch of kernel errors stating un-inited modules pointing to null.

Why it isn't 5
Stock kernels also have the same problem

So, sorry folks, this issue points to a hardware problem.
Our biggest answer will be at the following:
1) What does that battery do?
2) Why is the SDIO subsystem having timeouts?

That's all for now. I'm going to cross-post it on the developer's page in hope that someone will give notice.

 

[thegogetter-]

i think its software issue. i got my phone returned back from moto and it was working fine. as soon as i unlocked it, it went error. i wiped cache and davlik, went to a different house to catch wifi and it worked. came home and it said error, reset my router, worked all night. drove around this morning and it went error again, tried wiping and its not working anymore.

im on cm7rc2

edit. so it did one of those random reboots again, upon rebooting, it connected wifi again automatically, stayed on for about a min then went error.

 

[mvniekerk]

Yeah, I'm still gunning for hardware.

I've made a CM9 build, with the WiFi yanked out. I'm testing it, no restarts the past two hours. Will post it tomorrow, I'm thinking of naming it "The Sawed Off Hand Edition".

 

[mvniekerk]

http://forum.xda-developers.com/showthread.php?p=26163379

Check there for the Sawned Off Hand Edition

 

[runekock]

Thanks to mvniekerk for the great analysis.

My guess is that the chip goes out-of-spec as it gets older.

The RTC-battery has nothing to do with it. But the chip sometimes works for a short while, maybe especially if it is cold -- so while we fiddle with the battery, the phone cools off, and may work a bit.

It just might be possible to make a software work-around even if it is a hardware error. Perhaps changing the voltage or increasing timeout. But difficult...

Anyway, the reboots are likely a kernel bug. I would expect the phone to keep running even with wifi dead.

 

[mvniekerk]

A good test if the battery (and the RTC it is powering) has something to do with it is to:
1. Find the voltage of the battery
2. Disconnect it
3. Put on a power supply at that voltage
4. Boot up the phone
5. Check if WiFi still freaks out.

The phone reboots because of a kernel panic on a null pointer due to an unititialized driver (due to an interrupt from the SDIO card).

A good case for "it is the chip" - people that bought the phone new off the shelves at AT&T also got the same problem.

 

[djkmito]

New rom is not required because if I doesn't turn on wifi then not start reboot loop and device working normally.

I wanted send it to motorola but I have unlocked bootloader and rooted so my warranty is falled or no?