Thank you very much....
-
After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.
Root [ROM] CM7 TG-Reloaded (Final: 03-25)
- Thread starter Whyzor
- Start date
-
- Tags
- motorola triumph
Thank you very much....
buonadomenica
Newbie
just upgraded to the latest 02-02 rom and have the same crash under CM7 settings - input. Everything else works fine. I did not have that in the previous roms. Not a big deal as I don't really use that setting.
thank you for the continued work on this rom. I could never go back to stock - ugh!!
BD
thank you for the continued work on this rom. I could never go back to stock - ugh!!
BD
f4lloutdiablo
Lurker
thank you! ill go ahead and try this out
Whyzor
Android Expert
I saw he pulled from my source changes, but don't think he built one yet. So you can either sync with one of our sources & build or use my pre-built 01-31 kernel in post #2.
bobloadmire
Android Enthusiast
Hey Wyzor is there any reason this rom would have issues hooping APs? my friend has a Atrix running CM7 and it hops fine, but mine disconnects and reconnects to every AP and won't continue the session.
Chairshot215
Android Expert
Whyzor,
Have been on a build from source kick and was wondering if inserting your kernel would be as easy as using 7zip to delete the boot.img from my build and then adding the image from your Rom? Is there anything else that would need to be done or is there a specific process of doing what I just described?
Have been on a build from source kick and was wondering if inserting your kernel would be as easy as using 7zip to delete the boot.img from my build and then adding the image from your Rom? Is there anything else that would need to be done or is there a specific process of doing what I just described?
Whyzor
Android Expert
What is hooping APs? I don't have problems connecting to one router at a time.
It should if it's a CM7 build, the boot.img contains the the kernel 'zImage', and the ramdisk. Ramdisk should be same across CM7 on the MT. If you want to put the binary in your source tree & compile with my kernel, then you have to extract the zImage from the boot.img (google for tools & how to do that), then place it as file: ~/android/device/motorola/triumph/kernel
Bug reports like this are NOT helpful at all. What were you doing at the time? Watching youtube for half an hour? On a phone call? All would seem like normal expected behavior.
I have encountered the 'event' wakelock hanging a few rare times. Normally I see 2 active event wakelocks by 'cat /proc/wakelocks'. They seem to increment when the screen is turned off & on. And the timeout count matches the total count. I suspect a few times when the timeout is 1 less than total, a system interrupt happened (A forced close, app crash?) of some sort. And the wakelock didn't expire and was left hanging.
It's hard to find the cause of these bugs without more detailed reports of what was running, & what happened that lead to it. On a related note. the kernel wakelock feature was something Google wanted to put into Linux, but kernel developers didn't like it because of possible issues like this. Here's some more background on it for those curious.
Wakelocks and the embedded problem [LWN.net]
bobloadmire
Android Enthusiast
At our work we have a wireless network called "AHPublic" and it has probably 50 Access Points (APs) On most phone when one goes out of range, there is another in range and instead of disconnecting it migrates to the closer AP. The MT with this CM7 build does not, but stock 2.2 did. Is that clear? Thanks for the help!
Whyzor
Android Expert
Hmm, TG's wifi hacks probably doesn't support this mode, stock may be using a different wifi module/driver than CM7. Keep in mind the hardware interaction level is all reverse engineered without documentation.
bobloadmire
Android Enthusiast
ok cool, so probably not gonna happen?
hamfactorial
Newbie
Whyzor:
I've figured out what is going on with the wifi wakelocks and JuiceDefender. When the user is managing wi-fi directly, everything works as intended. Turning on wi-fi loads the libra driver and connects as normal. Turning off wi-fi unloads the driver and disconnects, no problem.
Things get more complex when the screen turns on and off. With libra loaded and the screen off, the btwlan_early_suspend() routine runs, creating the wlan_lock wakelock and keeping wifi active as intended.
Upon turning the screen on, the btwlan_late_resume() routine runs, destroying the wlan_lock wakelock. This is all good!
When JD runs, the normal logic breaks down since it operates on the BT and wifi interfaces after the screen has already turned off, bypassing the custom wakelock logic that you've programmed into drivers/power/bq275x0_battery.c
With JD, the libra driver is unloaded maybe 5 seconds after the screen turns off and the interface is determined idle. Since your btwlan_early_suspend() routine has created a wakelock after screen off, it remains there until the screen turns on again, despite whatever JD does to the interface in the meantime.
I'm not an expert C coder, but it looks like the wakelock routines are being event driven. Is that the case? If so, the kernel needs some hook to be informed that the wifi has been deactivated in some alternate way.
I've figured out what is going on with the wifi wakelocks and JuiceDefender. When the user is managing wi-fi directly, everything works as intended. Turning on wi-fi loads the libra driver and connects as normal. Turning off wi-fi unloads the driver and disconnects, no problem.
Things get more complex when the screen turns on and off. With libra loaded and the screen off, the btwlan_early_suspend() routine runs, creating the wlan_lock wakelock and keeping wifi active as intended.
Upon turning the screen on, the btwlan_late_resume() routine runs, destroying the wlan_lock wakelock. This is all good!
When JD runs, the normal logic breaks down since it operates on the BT and wifi interfaces after the screen has already turned off, bypassing the custom wakelock logic that you've programmed into drivers/power/bq275x0_battery.c
With JD, the libra driver is unloaded maybe 5 seconds after the screen turns off and the interface is determined idle. Since your btwlan_early_suspend() routine has created a wakelock after screen off, it remains there until the screen turns on again, despite whatever JD does to the interface in the meantime.
I'm not an expert C coder, but it looks like the wakelock routines are being event driven. Is that the case? If so, the kernel needs some hook to be informed that the wifi has been deactivated in some alternate way.
Whyzor
Android Expert
Unless another dev steps up and fixes it, no.
If JD can properly disable the wifi module after early_suspend() runs during screen-off, a big *IF*. I guess then there could be a way to tell the kernel that's in deep sleep mode to poll every once in a while to check if wifi is loaded, if it's not, release the wakelock. I'd have to dig around to see if such a function exists & where.
This is all a lot of work just to get an automated way of turning off the wifi with JD (which I don't use). The current manual toggle is very simple already. Not a high priority fix, maybe when I have more time to play with things.
lpnalda
Member
Can you answer a question for me? I'm Visually impaired, and want to know, does CyanogenMod7 offer settings for changin the system wide android font? And if so, how far can it go up in point size? And bold? Yes, I've heard that 2.3 has font size settings in display, but I've seen 2.3 phones and they have no such feature, and it's not in documentation either. I'm a root newbie, and I won't try this stuff if it doesn't help me by giving what I want. I've also been told that Pico TTS will not be affected. Is this true? Thanks in advance, and yes I do think what you do is cool, and want to be a part of it.
radiotecha1234
Newbie
First, I wanted to say thank you again to Whyzor who's 2-02 build is awesome. I have the best battery life ever and the phone is so responsive. I am glad I kept it.
I spent some time yesterday updating my Froyo backup in case I need it for HDMI. I was quite shocked by how slow it seems to me now that I have been on CM7 for several months. I am NEVER going back to Froyo daily!
I spent some time yesterday updating my Froyo backup in case I need it for HDMI. I was quite shocked by how slow it seems to me now that I have been on CM7 for several months. I am NEVER going back to Froyo daily!
thangfish
Android Expert
+1
Same thing happened to me.
Get so used to this, you really forget how it was at first.
Whyzor
Android Expert
Yes, there's an app called "Dev Tools" (part of CM7, which is built on top of Gingerbread) that allows system wide font size changes, it has normal, large, extra large, extremely large. Extremely large may make some text spill over the normal size allowed for each of the buttons. It basically affects all the typical Android settings & interface "White text with black background" areas. Apps have to have their own font size setting because they don't obey the global OS setting (this is probably where ICS's enhancement is). I use Opera Mobile browser and it has font size settings.
As for the install instructions, it's fairly straightforward, just have to pay attention to the details and follow them carefully. So far nobody has reported messing up their MT so bad that it's unrecoverable. Many people have gone through these instructions without problem:
http://androidforums.com/triumph-al...stalling-cwm-custom-rom-motorola-triumph.html
Yeah, seriously. I am getting about 5% per hour even with reasonable usage. Before CM7 I was getting 15% per hour or so. It really has made the phone go from cool to phenomenal. 
Chairshot215
Android Expert
Very new to these things but could JD turning wifi off after the phone has become idle (screen off) and not before have anything to do with the issue?
Edit: Never mind, just now understood what you had said above.
frdmspoplr
Well-Known Member
Is there a way to map the recent apps menu to the search key. I can't remember if it's in the CM7 input settings because of the bug in the current build. If its not in there, is there an app that can do it?
Whyzor
Android Expert
I don't think there exists a solution to do it. I personally would like faster access to the recent apps menu instead of the hold-home-key too. Even filed an enhancement tracker for Cyanogenmod, so far not much interest.
sharpe351
Well-Known Member
does that work? I've never known about the long press home key thing. Have to try it. Isn't there an app like a button remapper on the market?
frdmspoplr
Well-Known Member
Yeah, it's definitely pretty annoying having the hold the home button every time I wanna multitask. Would it be possible to modify the time you need to hold it like you did with the back button? Obviously this could potentially lead to more accidental pushes, which could annoy people, but it's a thought.
You learn something new every day
Yeah, it's similar to the multitask button un ICS. The only app i can find in the market right now allows you to assign various functions to the search key, but recent apps is not an option. I wonder if google even allowed access to the function. I'd write the developer, but it hasn't been updated in over a year, so I'm doubting it's still being actively developed.
TheBritton
Resident Galaxy Cat
This ROM is pretty awesome. Approaching nine hours on battery and still have just under 75% left.
