Re: [TESTING] - NAND OMGB (1.2.3)
 

ACL - Had what appears to have been a full blown SOD this morning. Green LED but no response to any keys. I gave it ~ 10 minutes to settle (per your mdp bug instructions, even though this didn't act like that) and then hit the reset button.

Attached is the /proc/last_kmsg but I didn't notice anything obvious...

One thing probably worth mentioning, I'm running detule's new gps lib and had just finished a 30 minute navigation test about five (or less) minutes before the SOD.

Re: [TESTING] - NAND OMGB (1.2.3)
 

Thanks dood.. This continues to be a mystery since it's extremely hard to replicate. I'm going to keep the code as it is so if it does happen to me i can adb and mess around with stuff. The logs themselves dont show much so we are still in search to kill this beotch.

I'm planning either this weekend or early next week for another update.

Re: [TESTING] - NAND OMGB (1.2.3)
 

Wait, are you saying I should (might) be able to adb in and pull logs when it does this?

I figured since it was totally unresponsive my only recourse was a reset.

Next time I'll see if it responds to adb before stabbing the little red button.

Re: [TESTING] - NAND OMGB (1.2.3)
 

Originally it would panic but i made a chance that you verified last time with your logs that the phone did not die (just mdp). SO yes, in theory you can adb into the phone while this is happening.

htc code has a recovery vsyc where the panel is re-initialized if this happens. If we can replicate this bug enough then we might be able to test that fix. But the issue is really to replicate it..

Re: [TESTING] - NAND OMGB (1.2.3)
 

Same here, this kind of lock up has happend to me several times, good to know you can still adb when this happens.

I'll try to pull some logs next time.

EDIT: adb is going to be tough. I just had a lock up (solid green, no response from the screen) and when I plugged in the usb cable the phone restarted.

Re: [TESTING] - NAND OMGB (1.2.3)
 

when there is a reboot, then there was a panic. Only chance to get logs is lastkmsg. If there is no reboot and the screen is just not responding (still up just frozen), then you might be able to adb and catch the mdp bug in action.

Re: [TESTING] - NAND OMGB (1.2.3)
 

I hope this is what you are looking for...

I had what looked like the mdp lockup (screen stuck on, orange LED, totally unresponsive to keys or screen touches) a couple minutes ago, waited for the phone and sure enough it did go back to sleep (screen off, green LED). When I woke it up it had a FC message waiting for me.

I pulled last_kmsg and ran getlogs, both included in the attached ZIP.

Re: [TESTING] - NAND OMGB (1.2.3)
 

holy shish kabob !!

so when it woke up was the phone responding just fine ? I see in your logs that kernel did have the mdp bug and just stayed there for a while. It slept and when it came up the mddi resumed with no problems. If this is true, you confirmed my theory that a good nap will resolve the issue.

I'll make more changes to see if we can have a small hack to assist with this. thanks again dood..

Re: [TESTING] - NAND OMGB (1.2.3)
 

Yes, after the nap everything was "fine" (well, after I dismissed the FC dialog).

I did not reboot the phone and over two hours later it still seems to be doing fine. I figured leaving it running without the reboot might make the bug happen again so I could get you more logs. :)

Re: [TESTING] - NAND OMGB (1.2.3)
 

I dunno why your name is mbellot when it should be "the mothaf*ckinman* :hello2: