I've slightly re-purposed this thread since the other one seems to be dead
Bootmenu v0.6.0 test2
Bootmenu v0.6.1
http://forum.xda-developers.com/showpost.php?p=13474771&postcount=35
Installation:
http://forum.xda-developers.com/showpost.php?p=13527738&postcount=40
I'm not making new 2nd-init updaters (it's been integrated in full ROMs which I'm not releasing here), and I will not update 2nd-init alone anymore.
But I will update Bootmenu here instead when I can.
----------------------------------------------------------------------------------------------------------
so i've kind of integrated dexter's 2nd-init and the korean guy's bootmenu into one flashable zip file.
big thanks to all the devs for making this possible
NOTE:
it's based on dexter_nlb's method which uses only shell scripts, it should work on many other devices in principal, ms2/d2/dx etc. of course you need to modify the .rc and other defy specific files.
this method only relys on 2nd-init by @Skrilax_CZ, it's compatible but does not rely on bootmenu.
starting v2 you can actually remove bootmenu binary while still able to boot into bootstrap. in fact v3.2 contains nothing from the "bootmenu" method (otherwise identical to v3.1)
i do not seek to optimize your system, all parameters changed are to give you a clear evidence that all the parts are indeed functioning, so quit whinging and mod the scripts yourself
just for show casing purpose, it includes the following features
-extra cpufreq governors, smartass by default
-lowmemorykiller value changed
-init.d enabled, including a undervolting script @800/50, 600/40, 300/30
all *.rc taken from 3.4.2-164 release, so i suppose anything close to that will do
usage:
0. clean up any other hijacking other than logwrapper hijaking!! including the oringinal one by koreans
1. root & install clockwork mod recovery
2. flash the zip attached
3. remove clockwork recovery if you wish, DON'T use "install recovery" option, it will mess up 2nd-init.
4. 2nd-init will load automatically, press vol-down while blue led is on to enter bootmenu
NOTE: when applying zip patch, sometimes it will fail with "resource busy", enable adbd from bootmenu
before enter cmw will fix that.
v2
-updated to cmw 2.5.1.8
-remove prior installation by default
-normal "boot recovery" or rather "/data/.recovery_mode" should work now,
just don't use "install recovery"
-adbd will start shortly after 2nd-init, while starting either cmw or bootmenu
-300s delay for overclock script (compatible with any 2.2.x kernel sans chinese one)
v3
-replaced overclock with sysctl to be more compatible
-remove changes made by bootstrap
-included a special version of bootstrap which should overide the default one (kb7sqi's version)
-boot to bootstrap via a blank file named /sdcard/cwm_boot enabled
-boot to bootstrap will turn off the blue led now
v3 removed due to faulty cwm replacer
v4
-added windows installer, it's both installable via cwm or windows/adb (windows installer does not work on some setup)
-updated to newer bootmenu
-you can completely bypass bootmenu without losing 2ndinit by place a blank file called /sdcard/boot_bypass, boot into recovery changed to /sdcard/boot_cwm)
-2nd-init option (which is in fact 3rd boot) from bootmenu should work
-bootstrap cwm will be properly removed
v5
-updated bootmenu to 0.4.5
-removed windows installer
-removed most changes in *.rc,
now it only loads init.d and changes 1 pair of lowmemorykiller parameters
IT'S NOT QUITE FINISHED, BUT GOOD ENOUGH FOR NOW, USE AT YOUR OWN RISK
Nice. Even i'm not a dev will try this and report any problems =)
Thanks
Will it 2nd-init by default? Is it possible to set it to 2nd-init by default? Is it possible to enable adb from start with default.prop?
after flashing this zip i'was unable to boot.
At the point where the blue led should light up the phone turns off(it doesn't light up).
I had the "normal" custom bootmenu already installed. maybe that was the problem.
It would make more sense to have zip which can be installed with adb, since bootmenu already has recovery, and messing different hijacks can lead to bad consequences. Great effort still
ill test it with stock rom and stock bootloader.
astearon said:
ill test it with stock rom and stock bootloader.
Click to expand...
Click to collapse
are you testing the 2nd-init? Anyone is testing?
Simplestas said:
It would make more sense to have zip which can be installed with adb, since bootmenu already has recovery, and messing different hijacks can lead to bad consequences. Great effort still
Click to expand...
Click to collapse
i'm using logwrapper hijacking instead of mount_ext3. so yes they don't work together.
Simplestas said:
Will it 2nd-init by default? Is it possible to set it to 2nd-init by default? Is it possible to enable adb from start with default.prop?
Click to expand...
Click to collapse
i don't see why you can't, in fact i did enable 2nd-init and adbd by default, commands both
from init.rc (init.d) and init.mapphone_umts.rc (all other mods) are properly executed
Simplestas said:
It would make more sense to have zip which can be installed with adb, since bootmenu already has recovery, and messing different hijacks can lead to bad consequences. Great effort still
Click to expand...
Click to collapse
true, but it takes a lot more time to write and follow the instructions than this
and i suppose anyone who find it useful can turn it into adb commands in a few minutes.
plus i wasn't trying to explaining anything but merely showcasing the possibility
I've installed 2nd-init not as default (just copied 2nd-init, sysinit, busybox and /etc/rootfs) and i must say it works. However, despite having persist.service.adb.enable=1 adb does not start until gui (or usb service?) so it's not useful yet.
I think it's better to wait bootmenu enhanced to support settings for default init, etc. This zip is too quick&dirty and it has overclock.ko tailored to specific kernel address so it will bootloop for many.
Simplestas said:
I've installed 2nd-init not as default (just copied 2nd-init, sysinit, busybox and /etc/rootfs) and i must say it works. However, despite having persist.service.adb.enable=1 adb does not start until gui (or usb service?) so it's not useful yet.
I think it's better to wait bootmenu enhanced to support settings for default init, etc. This zip is too quick&dirty and it has overclock.ko tailored to specific kernel address so it will bootloop for many.
Click to expand...
Click to collapse
why exactly did you need adb enabled automatically? you can manually enable it in bootmenu.
and the overclock.ko address is tailored for ALL 2.2.x kernels as does setvsel with an exception of the chinese one, so why would you care? and any one with even slightest knowledge on this topic could rectify it by simply removing the script.
i did say this is just a showcase and it's not finished, didn't i?
Simplestas said:
I think it's better to wait bootmenu enhanced to support settings for default init, etc. This zip is too quick&dirty and it has overclock.ko tailored to specific kernel address so it will bootloop for many.
Click to expand...
Click to collapse
most of the draw backs fixed.
i still don't understand why would you wait for something close-sourced for the moment, untested, while we have a simple script which can already do the trick, rely on nothing but plain shell script and tested on dexter's mod for months now?
racca said:
most of the draw backs fixed.
i still don't understand why would you wait for something close-sourced for the moment, untested, while we have a simple script which can already do the trick, rely on nothing but plain shell script and tested on dexter's mod for months now?
Click to expand...
Click to collapse
I just need to test the 2nd-Init. If works, it is everything okay
2nd-init works fine, but if I choose Start ADB daemon, it writes something like - can't start 2nd-init, read only file system and continues booting normally with mismatched screen and after booting, the system works somehow weird, so restart is needed. I am using 3.4.2-145.
peetr_ said:
2nd-init works fine, but if I choose Start ADB daemon, it writes something like - can't start 2nd-init, read only file system and continues booting normally with mismatched screen and after booting, the system works somehow weird, so restart is needed. I am using 3.4.2-145.
Click to expand...
Click to collapse
both adb and 2nd-init are started by default, 2nd-init from bootmenu will NOT work properly. because that'll be actually 3rd-init
Ok. But ADB is not working for me and I am affraid to turn it on in settings, if it's needed. Because I had the boot loop with the Korean bootmenu. Do I have to turn on USB Debugging?
Working! awsome! Couldnt do it with the other thread because couldnt push files with adb to /root/bootmenu. But flashing this zip with cwm did the trick.
Sent from my MB525 using XDA App
peetr_ said:
Ok. But ADB is not working for me and I am affraid to turn it on in settings, if it's needed. Because I had the boot loop with the Korean bootmenu. Do I have to turn on USB Debugging?
Click to expand...
Click to collapse
adb works regardless of your system settings (well until 2nd init.rc kicks in at least), so i don't see why you need to turn off debugging.
i can't say for sure what went wrong with your system, but i assure you it will work on all clean 3.4.x systems
So....this means that we can have litle AOSP taste??
So now that 2nd init is working can I dream with Cyanogen Mod ported to Defy, just like the Milestone?
I've tried several times and searched around but just want to know is this update rootable and if so how
I believe some people reported it was possible by flashing the 2.3.5 through odin, then using odin to flash infusion 1.1 kernel which includes root script. then you can either use the infusion (RUXKG) kernel, or if you want the 2.3.5 stock kernel you can flash it back and the root should remain.
all 2.3.3+ roms need to be rooted on insecure boot (custom kernel) or pre rooted. rage against the cage was being used to push malware to android phones and that exploit has been locked out. so your choices are to use magic iso to push su and superuser.apk to /system/xbin and /system/app respectively in factoryfs.rfs from the odin/heimdall package and flash in odin/heimdall or to use odin or heimdall to flash the zimage from any one of the custom rogers kenrels
if you do this i would use entropys kernel as the wifi works do to initramfs mods pulled from the 2.3.5 kernel. problems with entropys kernel include no dock audio sound redirect (audio over usb) due to a bug with the voodoo port and no hdmi. infusion has audio over usb but no wifi or hdmi
flashing back to the factory 2.3.5 kernel will restore hdmi and work correctly but have no recovery (use kernel flasher from the market and keep a backup of the custom kernel you choose) and it doesnt have the custom tweaks (obviously) like init.d support, jhash update, memory mods, vm management mods, and voodoo so it is all a compromise till sources are posted. but removing the custom kernel will leave the phone rooted.
Dani897 said:
all 2.3.3+ roms need to be rooted on insecure boot (custom kernel) or pre rooted. rage against the cage was being used to push malware to android phones and that exploit has been locked out. so your choices are to use magic iso to push su and superuser.apk to /system/xbin and /system/app respectively in factoryfs.rfs from the odin/heimdall package and flash in odin/heimdall or to use odin or heimdall to flash the zimage from any one of the custom rogers kenrels
if you do this i would use entropys kernel as the wifi works do to initramfs mods pulled from the 2.3.5 kernel. problems with entropys kernel include no dock audio sound redirect (audio over usb) due to a bug with the voodoo port and no hdmi. infusion has audio over usb but no wifi or hdmi
flashing back to the factory 2.3.5 kernel will restore hdmi and work correctly but have no recovery (use kernel flasher from the market and keep a backup of the custom kernel you choose) and it doesnt have the custom tweaks (obviously) like init.d support, jhash update, memory mods, vm management mods, and voodoo so it is all a compromise till sources are posted. but removing the custom kernel will leave the phone rooted.
Click to expand...
Click to collapse
Wait, wifi DOES work in UCKI3/I4 with my kernel? I saw multiple reports that it still didn't. (e.g. my experiment failed.)
I may try tonight, now that I own two devices I can be a little more adventurous with one of them.
I am working on a custom jb4.2.2 rom for the A13 SOC and wanted to add some features, ROOT By default - I added the su Binary and the SuperSU.apk however it does not work by default. If you boot into CWM Recovery then Reboot the device it offers to fix Root, when I select yes it proceeds and works fine... I would Rather make the ROM Rooted without having to do that any Ideas On what Im doing wrong ?
I am also interested in adding features to the ROM automatic Zipalign would be desirable as well as the ability to use a swap / zram. I am using a Stock JB 3.0.8+ Kernel and in at least one prior iteration I could use zram and activate swap... I would Like to make those features user selectable.
I currently have init.d working in the rom set in the init.rc using busybox run-parts. I do not have run-parts installed by default, is there a simple way to include the full busybox install as well ?
Assistance on any of these questions is appreciated.
I wanted a working init.d to mod some stock Interactive Gov settings in AospExtended ROM so I took the clever universal init.d enabler created by @_alexndr and modified it for a slightly different outcome.
The untouched version ends up using /system/bin/debuggerd as the host file to launch init.d support. This works perfectly well except debuggered is called at the post-fs-data point in the boot process.
This is too early for some init.d settings to stick as they can be overwritten later in the boot process by stock settings. I found /system/bin/mpdecision is called on property:sys.boot_completed=1 making it a much better candidate to use for our phone. So the original script has been modified to use mpdecision instead of debuggerd as the host file to start init.d
I also removed some lines automatically setting up a 99SuperSUDaemon init.d script since we may have magisk instead
All the magic is done in update-binary for those wondering how this works.
Tested working on AospExtended V4.2
Edit: to completely uninstall - delete /system/bin/mpdecision and rename /system/bin/mpdecision_real to mpdecision. Too easy!
Edit, edit: Some custom kernels disable mpdecision by renaming the same file that is being used here. That will prevent correct operation but can be easily overcome by first letting the custom kernel rename mpdecision (to mpdecision.bak) then make another (empty) file called mpdecision so it will become the host to launch init.d functionality. The key to correct operation is that the init.qcom.rc file still has the line that starts mpdecision.
Edit, edit, edit: What is wrong with the built in sysinit script to enable init.d you might well say.......2 problems: it requires busybox and it runs too early in the boot process (on post-fs-data)
Hope this comes in useful.