Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

Alright, just tried this again, put a brand new clean FRX04 build w/ the latest rootfs that was just released last night and this kernel, let it boot, and just sit for a few, at which time, without me touching it, it booted back to winmo. Before I do anything further (like run the freaking phone over with my truck...) do you want me to attempt to get a log on it or will one have not even been created as I'm not sure exactly how far the process went?

Edit: I changed a couple of things in the startup.txt and played around with it a bit and giving it another shot. Maybe I should try with the actual package first, then after it boots, go back to winmo, swap in the new rootfs and test kernel?

Edit part 2: Still reboots back to winmo and reboots back before I can get a log or anything. forget it, I just threw it at my neighbor, I'm going back to a freaking flip phone... lol. I'm gonna try booting just FRX04 out of the box and then make changes with rootfs and kernel.

So far so good with clean FRX04...

Re: [TESTING] Updated 2/2/11 New Kernel
 

I'd love to help you out. I have downloaded your latest kernel and it's running fine so far, seems quicker than previously used kernel. I can't speak for stability yet as I've been running it for all of 5 minutes.

I have a RHOD500 but don't know how to get you my dmesg log.

Re: [TESTING] Updated 2/2/11 New Kernel
 

I'm already thinking I know what it'll say, but the more logs the merrier. Do you have adb installed? Droid Explorer is the easiest way to get it working in Windows.

If not, there's an app on the market called "OSMonitor". You can pull dmesg and logcat with this app AFAIK. You then have the added fun of trying to get the log off of the phone - hence the reason I prefer adb, makes everything so much easier IMHO.

There's apps that can help you get stuff off the phone as well, like the DropBox app..

Re: [TESTING] Updated 2/2/11 New Kernel
 

I'm going to assume you're running Windows, if not let us know.
Install guide for ADB HERE: [SOLVED] ADB Setup for Windows - xda-developers
Set up adb, at the point where you've verified you can run adb, run the following command:
This will create a file called dmesg.txt in whatever folder command prompt says you are in. Attach that file here or upload the contents to Pastebin.com - #1 paste tool since 2002!

EDIT: arrrghhh beat me to it

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

k so if an incoming call comes in while the screen is off, it will ring and such, and then give me my gv vm notification but the screen will not come on, or even have the backlight kick on. if i wait for it to go to sleep then the screen will come back up. i have to wait to get another call to duplicate this, can i pull a dmesg once i get the screen to come back up for this event or do i need to reboot?

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

I guess I'm missing something here.

So when the phone is sleeping... it wakes on a phone call.

When the phone isn't sleeping, it doesn't wake for the phone call? I haven't really had any failed wakes on incoming calls, unless the processor was being slaughtered by some other process (not only system_server can cause issues, just that system_server issues would run without stopping).

We have the same phone, what build are you running? Overclocking? Need... more... info...

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

It sounds like tacotoy is saying the phone rings, but the screen never comes on with the slide to answer, so he can't answer the call.

The only time I've experienced that is when system_server was running rampant. If you haven't already I'd try dropping the newly fixed rootfs on the system and seeing if you get any better results.

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

@arrrghhh: Pastbin dmesg from rhod500

Edit: now maybe someone can explain to me what panel detection is. ;)

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

[ 1.206878] Panel type detected: 14 <---- thanks!

Means you have the EID panel, which the vast majority of our phones do. Seems a few RHOD100's have a different panel type, AUO. Which means the commands to wake and collapse that panel are going to be slightly different - and if the proper panel is not detected in the kernel, then the phone will have problems ;).

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

Well I'm glad I could participate. :) So the "panel" that this references is the actual display panel (or its controller)?

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

Fish, seems like you have been getting a lot of resets to WinMo issue lately, with Tiad8's build and FRX04. Hmmm... I had that issue the first time I tried FRX04 with 1253 OC'ed on first boot but it works fine after the first boot. I'm currently using FRX04, with 2/9 rootfs, and this kernel OCed at 614Mhz and it's pretty solid. Got the panel 14 type in my dmesg. When my phone goes to sleep, I still have to press the button several times to really wake it up. I'm using an app called nolock because I can't stand the slider to unlock screen that occurs everytime my phone wakes up; saves me a couple precious seconds so I can use my phone right away :-D.

arrrghhh, besides ADB, DroidExplorer or the other debugging methods you guys are using, is there a way we can set up a telnetd or sshd daemon with which we log into? The builtin terminal is good but the lack of some keys (>, |, tab, etc) makes it hard to play with (I guess I can use manekineko's rootfs with the added key functions).

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

The most effective way to guarantee a good wake is to touch the screen after you've hit the power button and waiting about a second.

If you prefer the END=HOME button layout check out F22's ROOTFS over on XDA, it also adds in the missing keys for effectively running shell commands.

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

Thanks, I'll keep that in mind. Does F22's rootfs have the system_server bug fix in it?

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

Indeed it does, he just updated it last night.

There is ssh into the phone, search for 'dropbear'.

Plus, you can get a full-featured shell using adb... just need to enable it. should do it.

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

you got it right, when the phone is sleeping it will ring, the screen will never come up but it will vibrate and ring. if i get a call while i am using the phone ( i e screen is on) then i can answer it just fine.

running the latest panel reint, and mannieko's modded rootfs with FRX04. i oc, but i have been oc'ing since it was available. this issue started once i loaded the most recent panel reint kernel. ill boot back to winmo and see if that makes a difference


edit: after booting back and then rebooting with no oc it works fine for a bit, but if i get a bunch of programs running, close them down ( ie eat up available memory) then it doesn't work, so seems like there is a memory leak. so over time as programs autoload then it prevents the phone from waking to answer the phone. oc'ing just seems to exacerbate the issue

Do a "ps" sometime, you'll see there's already a telnetd running, listening on your usb0 interface. It's built in to busybox and always started by the init script so it's always there.

Sent from my MSM using Tapatalk

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

Well if the phone is working hard, waking the panel isn't so easy for it.

However, it seems you are experiencing many more 'failed wakes' than anyone else - I hardly ever get failed wakes anymore, and unfortunately the problem is in the framebuffer. Not sure where they're at on fixing that piece, but WisTilt2 said his panel init code was basically done. Now we just need all the other pieces to play nice...

I'm glad you got it somewhat working I was going to suggest maybe a hard reset on the winmo side, I remember reading something awhile ago about a similiar problem and that's how the guy fixed it I donno just my 2 cents Good Luck

Sent from my FROYO X using Tapatalk

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

Well, I finally installed ADB. Thinking it was only available with the Android SDK, I downloaded the whole Android SDK and boy, the packages it downloaded took forever :( (my speed is fast but the site it was retrieving packages from is slow); I had this set up on my laptop. Then I vaguely remembered that DroidExplorer already has ADB ... (it was installed on my desktop) and that was a lot easier to set up... duh! Oh well, I guess I will learn how to develop apps for Android now.

I tried telneting into the phone via the DOS prompt, using the command "adb shell" but the screen is messed up (it has extra terminal control characters). What is everyone using for telneting or sshing? putty.exe?

arrrghhh, sorry for being off topic, but since you're actively following almost every XDAndroid thread on XDA, I'll ask: have you tested viruscrazy's latest camera library/kernel? I can only get his latest libcamera.so (2/7 - not the 5MP AF one) to work with his 1/23 kernel; his latest kernel (1/28 ) or (2/7 - the 5MP AF one) gives me a PINK/PURPLE screen after the capture. I posted my dmesg log on XDA for him to check it out.

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

1) I don't use telnet or ssh to get into the phone. Just usb connection...

2) I get the same issue with his newest kernel... I saw that log you posted, I need to get him one as well just haven't had the time to mess with it...

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

First off thanks for all the hard work, you guys are awesome.

I am testing the Gingerbread on my Rhod500. And I noticed that I don't have a single issue with the panel turning on when I use the top button, but if I use the end key (which is the power button on Rhod500) I get the wakeup issue. And when I use the latest kernel from this thread I get the same issue. Is it possible that the Sprint and Verizon versions have them mapped differently? I didn't do much searching to be honest, but I'm pretty sure I'm onto something here.. (feeling kinda smart)

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

arrrghhh, I've been using this with the FRX04 Froyo bundle you posted yesterday. Seems to work fine most of the times. However, I have a couple times in which the phone went to sleep (blinking green and orange) and when I try to wake it, I had to press the power button multiple times for like 30 seconds; sometimes, I would need to slide out the keyboard to wake it up; almost seemed like I had a SOD but I eventually get it to wake with more power button presses or keyboard sliding outs. This occurs when I was using an app (like tapatalk, sms, browser) and left the phone there to go do something else; when I came back it was in deep sleep; when I press the power button, it turned and stayed orange and seemed unresponsive as I mentioned above.

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

Yup, I believe acl said it was a vsync issue with the framebuffer... I think I need to get him to clean up his framebuffer code, as I think it's also causing the problem with not being able to calibrate the screen on newer kernels...

He's swamped in NAND right now, so not sure when I can get him to work on this... but it seems the timing on the framebuffer needs some work. Thanks for the report!

Re: [TESTING] Updated 2/9/11 New Kernel - Proper Panel Detect!!
 

I've posted a new kernel - this is for everyone (with a RHOD), especially RHOD100/AUO users who had issues with the last kernels, or with the newest GIT kernel 1258. My phone has been waking basically instantaneously, which is incredible. Let me know how it goes guys!

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

I just switched over to this new kernel and boom (doh!!!), got a reboot to WinMo. This happened after I accidentally enabled airplane mode, so I disabled it and data still didn't work. So I decided to add a second APN and while entering the APN name using the soft keyboard, it booted back to WinMo. Attached is the log.

Got this line in the log:
[ 817.262115] kernel BUG at drivers/char/tty_ldisc.c:199!

After which, the kernel crash dump data appears.

Distribution used: FRX04 Official Bundle, OCed at 700Mhz. I'm using ADW Launcher. SIDE NOTE: I was using Zeam Launcher but kept getting FCs on Zeam after my TP2 went into deep sleep using the last 1253 Wistilt2 test kernel you posted; so I decided to use ADW Launcher. Not a big fan of default Launcher2.

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

Interesting... I saw the infamous line right after that kernel BUG line. I'll shoot this log to WisTilt2, see what he thinks. Thanks!

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

OK, pretty solid now. Been running it all morning (some emailing, some tapatalk, some texting, some other stuff) without issue. With this kernel, the led stays solid green when in deep sleep; it's not blinking green and orange like the previous test kernel, is that correct?

Hmmmm... my battery indicator always seems to be near "full" (even when it's really at 45%), didn't see this on the previous kernel.

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

Battery indicator needs some work, perhaps WisTilt2 has started to work on it and some of it slipped into this kernel... not sure on that.

He pulled 1258 as a base, which set the sleep LED back to solid green. He didn't change it back.

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

Got me another reboot to WinMo (while filling out a form with soft keyboard under Opera Mini). See attached log (looks like a IP stack/ppp network issue).

I actually had a SOD earlier (green LED wouldn't wake up no matter what I did) when my battery was less than 10% (this was the value when I pulled the battery and boot back to WinMo). Unfortunately, I wasn't able to find the SYSTEM_LAST_KMSG for that SOD :(

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

Got another new kernel from WisTilt2. A lot of AUO folks were having problems, I'm hoping those are now squashed with this new kernel.

On my RHOD400 I do still see occasional times where the phone doesn't wake on its own without touching the panel, and WisTilt2 is pretty confident at this stage that the issue is on the .35 framebuffer - timing issues with the vsync basically.

With that said, this new kernel should be better. Until we can sort out the issue with the fb, they won't entirely go away...

So test this kernel out guys, let me know how it goes!

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

i am having this same problem too arrrghhh. Have you tested this new kernel, is it anybetter?

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

Of course, I'm on it right now. That was my feedback on this kernel... this is meant for testing, so test away!

I have my feedback, the whole point of this thread is to get feedback from others, thanks!

Re: [TESTING] Updated 2/16/11 New Kernel - Wake Issues/AUO Issues!!
 

Can't fairly compare battery life from my last kernel, but did have two delays waking up. The 1st delayed 20 seconds before waking up, don't know if I touched the panel; the second took about 10. Hopefully it's the timing issue.

So far so good. Been running for about 6 hrs with light usage and 45% battery depletion. I'm OCing at 614mHz. With deep sleep and I havent't touched the phone for a while, I usually double click the power button and it wakes up 90% of the times. On sleep where I haven't touched for a short period, it wakes up with one click. Since the "double click" works really well, I now have a habit of double clicking/pushing the power button to wake it up.

Recursion, try double clicking on your power button when waking up the phone.

Sent from my MSM using Tapatalk

ndno: I didn't know there was a deep sleep or a related doubleclick. I'll try it out...thanks! Btw, which OC are you using?

Sent from my Droid-Pimped TP2 using Tapatalk

Yeah, sometimes when it's in deep sleep (solid green led), I can't wake it up with just a push of the power button; I found that if I push it twice, most of the times I'll wake up the phone successfully.

For OC, I use 614mhz (acpuclock.oc_freq_khz=614400). It's a nice balance between stock (528) and the super fast 700+. Stock is too sluggish for me to have a nice experience; 700+ is great but drains the battery faster. If battery life is not a concern for you, use the 700 or more to get a wonderful smooth experience; I've used it and it's pretty amazing... there may be some stability issues but the kernel is getting better and better each day so it's more stable than before.

Update: I've been using this kernel for 8 hrs now without SOD or restart to winmo. Mainly using tapatalk, a few texts, a few calls, reading the news app, running the stock quotes app, and watching a couple youtube videos. Pretty solid thus far :)

Update again: this kernel is snappy! Everything seems smoother. Got quadrant scores in the 585-605 range with 614Mhz OC.

Hey arrrghhh, is wistilt2's kernel source available for public downloading? Does he have a branch off the linux-msm tree, like viruscrazy?

Sent from my MSM using Tapatalk

Ok, just tried the doubleclick & I see what you mean. 1 usually wakes it up over here but sometimes it takes 2.

Thanks for the input on OC...acpuclock is it then. I'll check it out!

I'm finding this build very smooth too, with battery life being the only less than awesome item. I am spoiled by Vin's work on the winmo side...so, I have about 50% drain in 3-4 hours of medium use. But I did use GPS and data.

I also noticed that when charging, the battery seems hotter than I'm used to. Is this typical?

Re: [TESTING] Updated 2/18/11 New Kernel - Cont'd Wake Issues/AUO Issues!!
 

Nope, he only has a local repo.

Yea, GPS/data does kill the battery. We also desperately need a fixed battery meter, the more I try to get accurate battery readings the more I notice how off the dang meter really is.

As for charging, I'm not sure why it would be hotter with this kernel. I have noticed the battery does get very hot when charging in Android, doesn't happen in WinMo. WisTilt2 and Jonpry are rewriting the battery code, only good things can come from that.


Glad to see you guys are having such good reports from this newest kernel! Hope to narrow down and squash the remaining failed panel wakes. Thanks again for the reports guys!

Update: very good battery life with this kernel. Went to sleep last night and awoke by my tp2 alarm this morning: 8 hrs and only 15% usage. I don't disable data at all. Recursion, I disable gps on mine so that's maybe the reason I get better battery life on mine. Yeah, you're not going to get the same battery performance under Android (at least not yet) as that using Vin's ROM. It's getting a lot better so I have been using Vin's Neo half the day and FRX04 with this kernel half the day.

Recursion is right, it gets pretty warm while charging if the phone isn't in sleep mode when I start to charge it. If I force it to sleep (push the power button) while charging, it's not warm. It's almost as though it won't go into deep sleep automatically if the charging is happening unless I force it with a push of the power button.

Very happy with this test kernel thus far. :)

Sent from my FRX04ed TP2

Re: [TESTING] Updated 2/18/11 New Kernel - Cont'd Wake Issues/AUO Issues!!
 

I was comparing this to WinMo, sorry about any confusion. Yeah, the heat from charging may be related to the battery code. Last night's test gave me this: as soon as Android told me battery was 100% charged, I rebooted to check in WinMo, which told me charge was 80%. Also, I noticed that Android didn't stop charging when the battery level got to 100%, it took an extra 15 mins or so.

I've got GPS and App Sync disabled, but I've been noticing weird LED behavior today: it'll go green (sleep), then quickly (1/10 sec) toggle between orange and green about a dozen times before stopping on either one. I did OC, which I'm about to remove to see if things change.

8 hours-->15% is good then. Okay, my baseline was Vin's NEO, which is clearly an unfair comparison at this stage of FRX04's evolution! :evil4:

Yeah, GPS & auto-sync are disabled over here. I'm going to fully charge under WinMo & test. Will test battery life on a sleeping TP2 with good signal strength, and GPS and Sync disabled. Oh, the heat...like I mentioned above, I'd guess the battery code has something to do with that.

Do either of you guys have a native Android device? If so, how hot does it get when charging?

Edit: Just checked again. Connected to PC via USB. Droid says ... "charged" && WinMo says 80%.