• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root [UPDATED 8/8/12]Breath of fresh air....an update

I gave a taste of jellybean to the marquee and now I'm back looking at the spectrum. Been looking at the hang/reboot all day with no luck.

I downloaded the spectrum kernel source and compiled a fresh kernel. Then I tossed in all the vendor libs from my phone (instead of the ones d2a was using). Still had the same reboot issue.

Then I removed pretty much everything from the init rc-files that I could. I only left the basic stuff -- mkdirs, mounts, and such -- and the following services: ueventd, adbd, servicemanager, surfaceflinger, zygote. Still rebooted. But when I took out servicemanager, it stopped rebooting and continuously complained "Waiting for service SurfaceFlinger".

Hmm ok, maybe that's a clue. So next I ran servicemanager under strace and looked at the output. Nothing really obvious, it doesn't get a chance to do much except open /dev/binder and issue a bunch of ioctls. But of course the reboot may not allow all of the strace output to be written, so perhaps the real issue is not showing up.

Tomorrow I think I'll run surfaceflinger under strace to see if that has any clues...

What a PITA.
 
In case anyone cares, the msm fbcon works when CONFIG_FRAMEBUFFER and friends are enabled, and "console=tty0" is passed on the kernel commandline. I was hoping this would show a panic message or something. Unfortunately, it doesn't show anything more than what is already available in /proc/kmsg.

Might be a nice geeky addition to the boot sequence, though. :)
 
Clue: disabling CONFIG_LGE_VIBRATOR and CONFIG_ANDROID_VIBRATOR stops the reboot loop.

I'm still running with my skeleton init and lack of dbus is causing the system to fail to start. Will start adding things back in momentarily...
 
Playing around a bit more with GL stuff. I tossed in the GL libs from the CM9 build. Looks like they might require a newer KGSL kernel driver...

I/Adreno ( 244): ioctl code 0x40140931 (IOCTL_KGSL_TIMESTAMP_EVENT) failed: errno 22 Invalid argument

Next I think I'm going to look at the p930 CM9 kernel.
 
For posterity...

Linux banner for latest CM9 nightly build says:

Linux version 3.0.8-cyanogenmod+ (pinky-ghost@cyanogenmod) (gcc version 4.4.3 (GCC) ) #1 SMP PREEMPT Sun Sep 16 14:34:32 PDT 2012

That would probably be the android-msm-3.0 branch in CM's lge-kernel-iproj.
 
Wow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wow, I guess that should do it :)
 
badrsj said:
Wow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wowWow wow wow, I guess that should do it :)
Click to expand...


Remember, its still gna be awhile till a stable rom. We have to worry about next is hardware stability and radio cinflicts
 
Yes that is likely next and hopefully our developers be able to work around that. I am happy that for the longest time the device would not even boot to a lock screen - this with good and hard work by developers into the device not allowing further work. So hopefully that milestone has passed :)
 
Crazy ass shit yo.. This will probably succeed before VZW drops ICS..

Always had phones with huge dev following this is the first time Ive been in a smaller "phone community" (is it called a "phone community"??), but yo guess what im gettin at is before I was kind of used to havin crazy options, roms, mods, and other shit so I took it for granted this really kinda brings me around to realizing how much effort is put in on these things..

Than again the spec is a pain in the dick (not innuendo, pls dont ban) phone to work on from what I gather..

Either way big ups guys, Im down to help how i can jus hit me up..
 
You must log in or register to reply here.
Back
Top Bottom