Forum switched to read-only as of 2020/06/01
|Latest product reviews|
OK, some more progress!
Using a 3.0.36+ kernel that properly includes SELinux as described here:
I can boot and connect fine with "adb shell" using "Oma_RK30_Pipo_S1s_KK_4.4_v1.0", but not using "Oma_RK31_Chuwi_V88_KK_4.4.2_v2.3_gen1 loader", which still hangs displaying the boot framebuffer. Next, I tried to flash the boot.img from the Pipo 4.4 build into the Chuwi 4.4.2 ROM, and - success! - the 4.4.2 ROM also boots fine.
Unfortunately, as soon as the "Welcome" screen appears (requiring to select the language), I am stuck, as the touchscreen does not yet work. Some seconds later, adb all of a sudden gets disconnected.
I have attached a dmesg log and a logcat log shortly before adb hangs - see below.
Can you please help resolving any of these questions:
- What could be the difference between the Chuwi 4.4.2 boot.img and the Pipo 4.4 boot.img that causes the hang with the Chuwi boot files?
- Do you have any idea why adb gets disconnected shortly after the Welcome screen appears on first boot? Is this expected, and I need to switch USB debugging back on via the Settings app once I get touchscreen to work properly?
- What do I need to change the Chuwi ROM to use landscape (tablet) rather than portrait (phone) mode? Is there also a "tablet UI" patch for KitKat?
Many thanks one more time!
Would it be possible that the following change causes the hang with the Chuwi boot.img?
--- boot_pipo/init.rk30board.rc 2015-04-01 21:38:21.889597942 +0200
+++ boot_chuwi/init.rk30board.rc 2015-04-01 18:38:23.383864739 +0200
@@ -10,9 +10,10 @@
- insmod /system/lib/modules/mali.ko mali_dvfs=50,100,133,160,200,266,400 mali_init_clock=50
+ insmod /system/lib/modules/mali.ko mali_dvfs=133,160,200,266,400,600,798 mali_init_clock=50
EDIT: Nope - switching to slower clock speeds did not cure the hang...
@Oma/Jochen (sorry for my impatience in case you are on Easter vacation!):
I still need your help with regards to my questions in #121.
Also, do you see anything in the logs as attached to #123 that could explain why the display goes black after the framebuffer (but only with the Chuwi boot.img)?
I'm still experimenting with making kernel drivers/options more similar to the stock TrekStor 3.0.8+ kernel, but so far, have not succeeded in getting the display and touchscreen to work properly, even though the correct drivers are included in my kernel:
Unfortunately, in the 3.0.36+ dmesg log in #123, I do NOT see the "GTP-INFO" startup messages from the touchscreen driver that are present in 3.0.8+, where they look like the following:
<4>[ 5.555485] <<-GTP-INFO->> GTP driver installing...
<4>[ 5.560512] <<-GTP-INFO->> GTP Driver Version: V1.8<2013/06/08>
<4>[ 5.568525] <<-GTP-INFO->> GTP Driver Built@21:25:39, Mar 4 2015
<4>[ 5.574649] <<-GTP-INFO->> GTP I2C Address: 0x5d
<4>[ 5.715059] <<-GTP-INFO->> driver send config
<6>[ 5.743715] input: Goodix Capacitive TouchScreen as /devices/virtual/input/input3
What can be the reason for this? How is the touchscreen initialized?
Thanks a million & best regards,
could you upload your current kernel source for the 3.0.36+ to github? I have to look at the driver code, perhaps I can help you.
Regarding the TabletUI, perhaps you can find something in the build.prop.
sorry it took that long for me to upload the source to github. As a complete newbie to git, I needed to sort out two major issues first in order to be able to successfully push: First, I was running into
(which basically means that using the https protocol from Ubuntu to github does not work due to a bug in GnuTLS and you need to revert to using ssh) and then with ssh I still needed to find out on increasing git client-side ssh timeouts (as my ADSL upload is real slow):
Quite a disillusioning first-time git experience...
The repository is finally online now:
So the most critical questions/issues with regards to the 3.0.36+ kernel (kernel config in arch/arm/configs/st10216-2a_defconfig) currently are:
1) Why does the display go black after the frame buffer with Chuwi boot.img?
2) Why does adb hang after some seconds if I use the Pipo boot.img?
3) Why does the tochpad not work even though there is a display with Pipo boot.img?
4) Why is
<<-GTP-INFO->> GTP driver installing...
the only dmesg line written by the touchpad driver (as opposed to the correct drive startup documented in #124 for 3.0.8+)?
I have now put several days of research into this, but without any progress - so I am thinking about abandoning this project and staying with the CM10.1 build for the Ventos...
Thanks a million in advance for your kind help!
one step after the other..
And you can not believe how many months(!) we needed for many homebrew kernels to work, so this is normal or... welcome to the homebrew scene.
It is always not easy to say something about a boot.img you did not build at your own. So regarding the different boot.img's Oma should help, I'm only a kernel guy.
I hope I could have some time in the next days to look at your kernel and perhaps I find something regarding the touch.
Touch and LCD are two different things, if LCD is working, this has nothing todo with the touch.
some limited progress:
When (only) replacing the init.rc within the Chuwi boot.img by the one from the Pipo boot.img, I can make the display work with the Chuwi image as well, but at the expense of adb hanging some very few seconds after the Welcome screen is displayed.
When looking at the differences between these two init.rc files, there are only three minor (!?) differences:
- The Chuwi version writes a value for kernel parameter "ping_group_range"
- The Chuwi version has the "vold" service disabled, while it is enabled in the Pipo version
- The Chuwi version has an additional service "daemonsu"
--- boot_pipo/init.rc 1970-01-01 01:00:00.000000000 +0100
+++ boot_chuwi/init.rc 1970-01-01 01:00:00.000000000 +0100
@@ -131,6 +131,10 @@
# This is needed by any process that uses socket tagging.
chmod 0644 /dev/xt_qtaguid
+# ICMPv6 group id range from 0~0x7fffffff for passing cts loopback ping
+ write /proc/sys/net/ipv4/ping_group_range "0 2147483647"
# Create location for fs_mgr to store abbreviated output from filesystem
# checker programs.
mkdir /dev/fscklogs 0770 root system
@@ -469,6 +473,7 @@
socket vold stream 0660 root mount
ioprio be 2
service netd /system/bin/netd
@@ -605,3 +610,8 @@
service whtest /system/bin/whtest.sh
+service daemonsu /system/xbin/daemonsu --auto-daemon
+ class core
+ user root
+ group root
While at the first glance, I don't see why any of those should cause the differences in behaviour:
- Screen does properly display (Pipo) vs. stays blank black (Chuwi)
- adbd hangs some few seconds after displaying the Welcome screen (Pipo) vs. adbd works fine (Chuwi)
Will try to track this down further during the next couple of days...
Thanks & best regards,
OK, so it looks like the fact whether the "vold" service is disabled or not determines the behaviour as described above:
- vold disabled: Display stays blank black, but adbd continues to work fine
- vold enabled: Display shows Welcome screen, but adbd hangs shortly after the Welcome screen is displayed
Do you have an idea why this might be the case? Obviously I need both: adbd and the display...
Thanks again & best regards,
daemonsu is for the root service.
Regarding the adb, unpack both boot.img. Then make a grep for adb in both. There should be many different entries for adb and adbd in:
file_contexts, init.rockchip.rc, init.rk30board.usb.rc, property.contexts, init.usb.rc, init.rc, default.prop and uevent.rc.
Perhaps it helps to switch the sepolicy file.
Hello Jochen, Oma, all,
any news from your side regarding my questions in #126 and #129?
I am completely stuck - I had already unpacked (and successfully repacked) both boot.img versions (Pipo and Chuwi), and I've also done a lot of experiments with selinux and vold configurations, but I don't get anywhere...
I still completely unable to figure out why adb - and seemingly the whole kernel - hangs once vold is started, and I don't see any helpful error messages at all until the time of the hang. Also, I have no clues why I don't end up at the Welcome screen (the display stays black such that I don't see anything on the screen) when the vold service is disabled, but in this case, I don't see a hang (and also no signs of any issues in the logs)...
If you have any ideas about what might be going on, or what I should try to change inside the boot images, please let me know - I am at my wits' end and now indeed think about giving up with this...
Thanks & best regards
Hi Awl, "vold" itself may not be the cause of your problems..
This line should be present in "init.hardware.rc" , at the end of "on post-fs-data" section, and commented out in "init.rc"
# Set indication (checked by vold) that we have finished this action
setprop vold.post_fs_data_done 1
re adb: Try adding this to /system/build.prop, (set perms to 0644)
The following users say thank you to ri9000 for this useful post:awl
thanks, but unfortunately both your proposals are already included:
I do have "setprop vold.post_fs_data_done 1" in init.rk30board.rc at the proper place, and also in /system/build.prop, I even have "persist.sys.usb.config=mtp,adb"...
Any other ideas?
Thanks again & best regards,
Most Users Ever Online: 749
Currently Browsing this Page:
Devices in use: Desktop (56), Phone (8)
Guest Posters: 43
Moderators: globula_neagra, exelletor, JochenKauz, Oma7144, cracktech