faust93
994539e94a
Bring multirom to M86
2016-12-18 14:13:41 +03:00
Zhuowei Zhang
42d2dcb3cf
trampoline: Add new pre/post encrypt hooks, fake system properties library
...
These are needed on Snapdragon 810 where an extra daemon (qseecomd) must
be started prior to decryption, and where the decryption library polls
a system property to make sure that daemon is running
Also fix missing fonts log spam in trampoline_encmnt by symlinking the one available font
2016-02-10 20:19:16 -08:00
Zhuowei Zhang
4d02af6e26
Fix encryption support build in Omni 6.0 tree
2016-02-10 20:19:16 -08:00
HashBang
20a0889dac
fix build with android-6.0 tree
...
Change-Id: Ia1126b186f77875868e417efeb121ba894580aa0
2016-02-10 20:19:16 -08:00
Vojtech Bocek
cae487184e
Restore build in omni 6.0 tree
2016-01-10 21:43:44 +01:00
Captain Throwback
0f1c246dec
Initialize adb via usb_function_switch
2015-06-01 20:05:47 -04:00
Vojtech Bocek
0ff3e9988e
[t26] Backport some ueventd fixes
...
e2d63af Move sprintf to snprintf.
d1d9602 init: refactor firmware loading locations into table
d1d9602 init: refactor firmware loading locations into table
1225a61 init: consider all platform devices for find_platform_device
2015-05-29 21:00:47 +02:00
Vojtech Bocek
21ec010d5c
[t25] Properly clean-up in trampoline (fixes M preview boot)
2015-05-29 00:19:59 +02:00
Vojtech Bocek
df3d66a473
Revert "[t22] Don't move /main_init to /init before executing it"
...
This reverts commit d3147e465c .
Conflicts:
version.h
SELinux in Android M preview needs it to be called /init.
2015-05-29 00:18:38 +02:00
KINGbabasula
55600f4acc
Use a macro for unused variables
...
Change-Id: I491bea842274029feb965b71c3a62b5953c2ce5c
2015-03-23 14:57:20 +01:00
KINGbabasula
5f73cc6c99
Silence warnings from unused variables
...
Change-Id: I4bebc7947770a26467f3435a4249fa34c94c266b
2015-03-22 10:38:36 +01:00
Vojtech Bocek
d3147e465c
[t22] Don't move /main_init to /init before executing it
...
* This causes "Device or resource busy" when remounting rootfs as RO
on hammerhead with encryption and thus causing bootloop because...
why the fuck exactly? Only on this device and only with encryption?
I don't care at this point, already took several hours to debug,
hopefuly SELinux won't mind.
2015-03-11 02:17:08 +01:00
Vojtech Bocek
222af62472
[t19] Improve cleanup after decryption in trampoline
2015-03-07 15:30:45 +01:00
Vojtech Bocek
6b14068b8c
Fix some log messages missing newline at the end
2015-03-07 15:27:26 +01:00
Vojtech Bocek
30f5316b7b
Reboot when trampoline can't mount /data on second boot
...
* On second boot, we have secondary ROM's kernel running and don't
want to boot the primary ROM with it.
2015-03-06 21:05:23 +01:00
Vojtech Bocek
d6e8ff40b4
Try to mount /data first before trying encryption
...
* This is the same way Android does it.
2015-03-05 18:33:10 +01:00
Vojtech Bocek
1cb2f48e15
Support for last_kmsg in pstore filesystem
2015-03-04 14:39:38 +01:00
Vojtech Bocek
56a68961bd
Properly clean-up encryption mounts & files
2015-03-04 01:04:08 +01:00
Vojtech Bocek
c7aee95d1f
Remove ADB debugging code
2015-03-03 19:47:32 +01:00
Vojtech Bocek
dcbea313ba
Add "close on exec" flag to all open() calls
2015-03-03 19:46:19 +01:00
Vojtech Bocek
b6529781d7
Fix some missing newlines in trampoline logging
2015-03-03 19:40:17 +01:00
Vojtech Bocek
ed2860ee3b
Initialize /vendor and mount /firmware for trampoline_encmnt
2015-03-03 19:31:43 +01:00
Vojtech Bocek
e835b7173d
Support adb with functionfs
2015-03-03 19:30:36 +01:00
Vojtech Bocek
f7f91aaa3b
Backport "ueventd: UEVENT_MSG_LEN changed to 2048" to devices.c
...
* ...from 1048 in compliance with the BUFFER_LEN in the kobject_uevent in kernel.
Commit: c603720aed304fdd36a5815f81eb22bd4d9968b7
2015-03-03 19:28:50 +01:00
Vojtech Bocek
d220c03f44
Backport "HACK: use /firmware/image as a 3rd choice" to devices.c
...
Commmit: 029d44e6aef334d2f0f03cbc9a5609cddc1b23e3
2015-03-03 19:25:39 +01:00
Vojtech Bocek
e75092e73b
Support device encryption with passphrase
2015-03-02 20:58:39 +01:00
Vojtech Bocek
87cfd90c01
Support device encryption with PIN
2015-03-01 22:30:53 +01:00
Vojtech Bocek
34b206a730
[t17] Support device encryption with default password
2015-03-01 19:43:46 +01:00
Vojtech Bocek
959b76d08a
Move bootimg injecting to the lib, trampoline now has --inject=boot.img
2015-02-26 15:30:18 +01:00
Vojtech Bocek
4f48c4bd80
Debugging stuff, copy libs to the ZIP, misc
2015-02-19 12:54:41 +01:00
Vojtech Bocek
7db8f34697
Move drawing, input and utils code into separate library
2015-02-16 21:01:06 +01:00
Vojtech Bocek
2449b5b27e
Add trampoline_encmnt
2015-02-15 21:11:22 +01:00
Vojtech Bocek
3bc4f1511f
[t16] Fix crash in fallback mounting code in trampoline
2015-02-13 17:21:48 +01:00
Vojtech Bocek
b08a83e0e0
Print error when multirom crashes
2015-02-08 18:46:57 +01:00
Vojtech Bocek
33b44d71aa
[m31][t15] Add support for multiple entries per path in fstab
2015-02-08 18:45:32 +01:00
Vojtech Bocek
e0c48a1132
Restore build in 5.0 tree
2015-01-05 00:34:10 +01:00
Vojtech Bocek
d6fcafc36a
Fix tags in logs
2014-08-20 16:31:00 +02:00
Vojtech Bocek
c3e3432133
Add bunch of O_CLOEXEC
2014-08-06 20:04:06 +02:00
Vojtech Bocek
3229eac1b0
Fix mounting of radio.img on Android L preview
...
* Workaround SELinux stuff
* Use fstab "class" for changing the fstab instead of separate code
2014-08-06 17:48:38 +02:00
Vojtech Bocek
4a606638ad
Add SHELL environment variable to adb, makes using gdb easier
2014-07-02 18:49:00 +02:00
Vojtech Bocek
1789c084dd
[t14] mv /main_init -> /init before running it to make SELinux happier
2014-06-04 19:11:54 +02:00
Vojtech Bocek
67c611afb7
[t13] If /realdata mount with fstab fails, try to mount with defaults
...
* busybox does it the same way if you specify -t auto
2014-05-14 15:30:41 +02:00
Vojtech Bocek
ae4c3d3fcd
trampoline: remove unneded run_multirom_bin() and use run_cmd() instead
2014-03-04 20:42:36 +01:00
Vojtech Bocek
0a3df551a3
Backport "ueventd: fix a busy loop while reading uevents"
...
commit 1d4e86c44589b3a97ca0113493c2e569c3aabcc6
Author: Amir Goldstein <amir@cellrox.com >
Date: Sun, 10 Nov 2013 15:36:58 +0200
ueventd: fix a busy loop while reading uevents
Under certain conditions, poll() may raise the POLLERR
flag along with POLLIN, in which case the check for
(ufd.revents == POLLIN) results in an endless busy loop.
The following fix was applied to
hardware/libhardware_legacy/uevent/uevent.c
to fix a similar bug:
commit 3aabb260ceef10377c31c9e45fb239247f5cfeba
Author: Mathias Agopian <mathias@google.com >
Date: Mon Oct 1 14:53:18 2012 -0700
fix a typo in uevent_next_eventi
Bug: 7114973
Change-Id: I15a4c714b59aeb1d02db00517d70b5f0e5ab22c2
Applying the same fix for two more poll loops in init
and ueventd.
Change-Id: I50693f6d3c904992ac4b8a9a14a83c7106e6b9e0
2014-02-01 18:31:45 +01:00
Vojtech Bocek
d3b2fb89ba
[t12] Init /dev/fuse in devices.c instead of device-specific inits
2014-01-11 09:04:11 +01:00
Vojtech Bocek
c2d29a88b5
Move containers to standalone file
2014-01-02 17:25:05 +01:00
Vojtech Bocek
e5f658cd04
Redirect STDIO of execvd' processes to /dev/null
...
* Fixes 30s freeze on dup2() in Android's init binary after it is
started by trampoline if adb in trampoline is enabled
* I think the timeout is on /dev/console, which is wired
to UART on hammerhead. dup2 tries to flush the file but
UART driver won't accept immediately since nothing is connected
to the UART lines.
2013-12-09 20:20:54 +01:00
Vojtech Bocek
b6e870793f
Add tramp_hook_before_device_init hook, move hooks to their own header
2013-12-09 18:53:02 +01:00
Vojtech Bocek
0fb82a3c41
fflush(stdout); after printing version for '-v'
...
* When built in omni tree, version doesn't get to multirom/MultiROMMgr
who request it via invoking -v argument
2013-12-07 21:13:12 +01:00
Vojtech Bocek
75d89b832c
trampoline: improve logging
2013-12-07 01:07:04 +01:00