Help Does rooting fix the sporadic rebooting issue (caused by secure wifi)?

[ZZDMAN]

Hello,

I recently "upgraded" from an Optimus V to an Optimus Elite and so far, I hate it.

Aside from the screen having significantly worse color reproduction, and the device detecting my 8gb sd card as 1gb, I am suffering from the more serious issue of random reboots.

I cannot post links (new to the forums) but if you google "Optimus Elite reboots a lot" you'll find a thread at androidcentral which talks about this exact issue. It is apparently caused by the wifi trying to reconnect when waking the device from sleep. I've had the device for 1 day and have experienced 6 reboots (1 disabled my alarm clock, almost missed class )

I live on a college campus and not using a secure wifi connection is out of the question as far as I'm concerned. Does anyone know if rooting the phone and installing a custom rom (Harmonia) would solve the issue? If not, is there another known way of solving the issue?

Thanks in advance

 

[giantpune]

No, rooting the phone won't fix the issue. And installing any of our existing roms wont fix the issue. Rooting the phone will allow you to fix the issue though. It will give you the freedom to read, write, delete, and replace all the hundreds of individual programs that make up the operating system. It will still take somebody actually tracking down the exact cause of the bug and fixing it.

Next time it crashes and reboots, you can pull the /data/dontpanic/last_kmsg and see if you have a kernel panic logged at the end. If you do, that will tell you a really good idea about which driver crashed.

 

[ZZDMAN]

No, rooting the phone won't fix the issue. And installing any of our existing roms wont fix the issue. Rooting the phone will allow you to fix the issue though. It will give you the freedom to read, write, delete, and replace all the hundreds of individual programs that make up the operating system. It will still take somebody actually tracking down the exact cause of the bug and fixing it.

Next time it crashes and reboots, you can pull the /data/dontpanic/last_kmsg and see if you have a kernel panic logged at the end. If you do, that will tell you a really good idea about which driver crashed.

If I had the programming knowledge required to fix the issue, believe me I would.
Judging by the wording of your post, I somewhat doubt that this issue will ever get resolved. Nonetheless, I will post the log file if it exists (as well as take a stab at identifying the culprit on my own).

Also note that I just emailed LG. My expectations are low but it's worth a shot.

Thank you

 

[ZZDMAN]

Update:

It seems that I cannot access /data/dontpanic/last_kmsg without rooting the phone.

I would like to wait for an update from LG before rooting the phone and potentially voiding my warranty.

I am sure that others are suffering from this issue as well. Would anyone please be so kind as to upload the crash dump?

 

[giantpune]

I think you may be doing it wrong then. You dont need root to read, or even to write, that file. You just need a less crappy file browser/text editor.

 

[ZZDMAN]

After trying out several file managers to no avail, I used an app called "log this" to grab the kernel log. Here's a link:

kernel_log.doc download - 2shared

Also, here is the ridiculous reply I got from LG:

------Reply to Your Inquiry-------


Until there is a patch to fix this I know now how to avoid this from continuing aside from disabling the secure wifi connections...

I am not aware of thie having yet been reported at least not to LG as a serious issue requiring a Software Update, but until then I can only advise performing a Factory Reset according to the Manual to perhaps reset the device back to its original Factory Settings, and then chekcing for any changes.

I hope this information helps and should you have ANY future concerns or questions, then please don't hesitate to email us again or PREFERABLY call LG toll free at 1-800-243-0000 anytime, day or night to speak with a Live Representative!

Thank you again for contacting LG Electronics, it was my pleasure to serve you, many thanks for being an LG Customer, and be blessed!

D. Stanley, Jr.
e-mail Correspondence, LGEAI

Hope that helps

 

[giantpune]

It looks like this is the cause of the crash -

[ 7558.446899] Internal error: Unable to handle kernel NULL dereference at virtual address 0000000c
...
[ 7558.447418] LR is at csrRoamCallCallback+0x94/0x9c [wlan]

The bad news is that it is inside the proprietary kernel module WCN1314_rf.ko, which they didn't give us source code to. That means we cant fix the source code and recompile it, which would be the clean and proper way to do it. What you can do, though, is a patch. You could either patch the kernel module and replace it on your phone, or create some hocus-pocus that will patch the module every time it gets loaded (whenever you turn on wifi).





If you know somebody that can reliably reproduce this crash, and has their phone rooted, you can try this to fix it. Here is the function pointed to by their LR. It only contains 1 branch-link which looks like it only goes somewhere to log something, and is otherwise nonfunctional. Have them either replace the "BL smsLog" with a "nop" or just jump over that block completely by changing "BEQ loc_26BF0" to just "B loc_26BF0".

.text:00026B8C
.text:00026B8C                 EXPORT csrRoamCallCallback
.text:00026B8C csrRoamCallCallback                     ; CODE XREF: csrScanForSSID+2D4p
.text:00026B8C                                         ; csrScanHandleSearchForSSID+12Cp ...
.text:00026B8C
.text:00026B8C var_28          = -0x28
.text:00026B8C var_24          = -0x24
.text:00026B8C arg_0           =  0
.text:00026B8C arg_4           =  4
.text:00026B8C
.text:00026B8C                 STMFD           SP!, {R0,R1,R4-R10,LR}
.text:00026B90                 LDR             R9, [SP,#0x28+arg_0]
.text:00026B94                 MOV             R6, R1
.text:00026B98                 MOV             R1, #0x4BD4
.text:00026B9C                 SUBS            R7, R2, #0
.text:00026BA0                 MOVNE           R7, #1
.text:00026BA4                 MOV             R4, R2
.text:00026BA8                 LDR             R1, [R0,R1]
.text:00026BAC                 CMP             R9, #7
.text:00026BB0                 MOVNE           R2, #0
.text:00026BB4                 ANDEQ           R2, R7, #1
.text:00026BB8                 MOV             R5, #0x628
.text:00026BBC                 CMP             R2, #0
.text:00026BC0                 MOV             R10, R3
.text:00026BC4                 LDR             R8, [SP,#0x28+arg_4]
.text:00026BC8                 MLA             R5, R5, R6, R1
.text:00026BCC                 BEQ             loc_26BF0
.text:00026BD0                 LDR             R12, [R4,#0x28]
.text:00026BD4                 MOV             R1, #3
.text:00026BD8                 LDR             R2, =0xA554
.text:00026BDC                 MOV             R3, R8
.text:00026BE0                 STR             R12, [SP,#0x28+var_28]
.text:00026BE4                 LDR             R12, [R4,#0x2C]
.text:00026BE8                 STR             R12, [SP,#0x28+var_24]
.text:00026BEC                 BL              smsLog
.text:00026BF0
.text:00026BF0 loc_26BF0                               ; CODE XREF: csrRoamCallCallback+40j
.text:00026BF0                 LDR             R0, [R5,#8]
.text:00026BF4                 CMP             R0, #0
.text:00026BF8                 BEQ             locret_26C20
.text:00026BFC                 CMP             R7, #0
.text:00026C00                 MOV             R1, R4
.text:00026C04                 STRNEB          R6, [R4,#0x34]
.text:00026C08                 MOV             R2, R10
.text:00026C0C                 STR             R8, [SP,#0x28+var_28]
.text:00026C10                 MOV             R3, R9
.text:00026C14                 LDR             R0, [R5,#0xC]
.text:00026C18                 MOV             LR, PC
.text:00026C1C                 LDR             PC, [R5,#8]
.text:00026C20
.text:00026C20 locret_26C20                            ; CODE XREF: csrRoamCallCallback+6Cj
.text:00026C20                 LDMFD           SP!, {R2-R10,PC}
.text:00026C20 ; End of function csrRoamCallCallback
.text:00026C20
.text:00026C20 ; ---------------------------------------------------------------------------
.text:00026C24 dword_26C24     DCD 0xA554              ; DATA XREF: csrRoamCallCallback+4Cr
.text:00026C28
.text:00026C28 ; =============== S U B R O U T I N E =======================================

EDIT:
If you can dump your /proc/kallsyms and upload it somewhere, it would help to match the disassembly up with the addresses in your crash log so you would know exactly which line is causing the crash.

 

[ZZDMAN]

I unfortunately don't know of anyone who can reproduce the issue. Honestly, I'm amazed that this thread isn't flooded with people having the same issue; I guess WPA2-enterprise isn't as commonly used as I thought (or perhaps it is something more specific).

In any case, I've uploaded the kallsyms file as per your request:
kallsyms download - 2shared

If i don't hear anything from LG by this weekend, I'll root the phone so that I can test the code you posted. Please take into consideration that you may have to dumb some stuff down for me, I have yet to achieve pro status ;-)

Thanks!

 

[giantpune]

Matching up you kallsyms + the panic log + the disassembled driver, it look like it is crashing at "00026C1C LDR PC, [R5,#8]". They check if that is NULL, which it isnt, so it passes that check, even though its still an invalid address (0x0000000C). You'd have to work your way backward and see where they set that value to find the real issue.

 

[jmfolcik]

Lots of people have this problem just dont know what causes it I'm sure. I isolated it to the wifi after my phone only rebooted at work and had VMU replace it with the same model for it to continue doing the same thing. I first thought it was just the signal in the area causing a glitch on bad hardware.

I would have it happen on WPA Personal on DD-WRT D-Link DIR615-E3 with builds from 2012, and at work on our Cisco enterprise APs that we enter in login credentials to connect in the wifi settings. I think its just a crappy driver, which this phone was full of. Koreans really screwed off while designing this one.

 

[ZZDMAN]

Giantpune:
Thank you for taking the time to diagnose the problem. As previously stated, I don't have half of the technical knowledge required to take on such a task. If you (or any of the other developers on this site) would like to take this on as a challenge, I would greatly appreciate it.

Seeing as I am somehow one of the only people disturbed by this issue, I am considering simply ebaying the phone and going back to my slow (but stable) optimus v. Side note: I'm very uncomfortable ebaying a phone I know has an inherent issue. Anyone have any advice on what information I should put in my ebay listing?

jmfolcik:
You betcha they screwed up, big time. I did a decent amount of research on the phone before buying it and, honestly, barely found any info on this glaring issue. I am still shocked that this phone somehow passed quality control and is being sold in mass quantities. I can't tell you which router I am trying to connect to as it is a college campus network. I can, however, tell you that it is a WPA2-enterprise connection, 802.1x EAP (PEAP).

 

[giantpune]

Just so you know, BestBuy had these phones for $50 brand new a couple weeks ago, and they aren't selling them with a big note under them that they crash using a certain type of WiFi. I wouldn't expect to get very much off your ebay listing.

But, if you do decide to put up a listing, PM me a link. I was looking at getting a spare one of these to tinker around with, so I'll probably throw up a bid.