Re: [android-x86-devel] Re: Android-x86 8.0 boot to home

[Chih-Wei Huang]

Hi all,
The full oreo-x86 branch is pushed.
Get it as usual:

repo init -b oreo-x86 -u git://git.osdn.net/gitroot/android-x86/manifest
repo sync --no-tags --no-clone-bundle

There are many issues to be solved.
I just list some of them:

* Unable to run init.sh in init.*.rc
* Some apps crash on booting
* Mesa not work
* Wifi, BT, camera, sensors, usb not work. But
Ethernet works. Audio should work if you
manually initialize the sound card.
* No superuser


--
Chih-Wei
Android-x86 project
http://www.android-x86.org

[fguy]

* Unable to run init.sh in init.*.rc 

maybe as in fugu

    exec - root root system -- /system/bin/init.fugu.countrycode.sh

[youling 257]

please give 32bit test iso 

[Chih-Wei Huang]

Not that easy.
I have already tried it.

There are more deeper reason to prevent
it from working. Need to dig into the code
to understand why...

[fguy]

Over the past week, we have made significant progress - the init.sh file is launched, a usb flash is connected, the development menu is launched, and the supersu runs. Still not working mesa, wifi, camera and sound. From the wishes I would like to return the old layout of the system.sfs from system.img.

[Michael Goffioul]

I've tried the new branch on my 2 test hardware and in both cases boot hangs shortly after switching to the frame buffer, it doesn't get to the boot logo. My 2 hardware are Atom D525 and Baytrail. I can't get any useful debug log, because the screen is cleared and I only get the cursor in the top-left corner. Attached keyboard does not work. Even in debug mode, the 2nd and 3rd console becomes unaccessible once the screen it cleared. I also try to set debug.sf.nobootanimation=1, but it doesn't change anything.

Michael.

On Sun, Sep 17, 2017 at 3:48 AM, fguy <fguy...@gmail.com> wrote:

Over the past week, we have made significant progress - the init.sh file is launched, a usb flash is connected, the development menu is launched, and the supersu runs. Still not working mesa, wifi, camera and sound. From the wishes I would like to return the old layout of the system.sfs from system.img.

--
You received this message because you are subscribed to the Google Groups "Android-x86" group.
To unsubscribe from this group and stop receiving emails from it, send an email to android-x86+unsubscribe@googlegroups.com.
To post to this group, send email to andro...@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.

For more options, visit https://groups.google.com/d/optout.

[fguy]

that would be 8.0 now loaded, you need to fix the file init.sh

function do_init()

{

init_misc

init_hal_audio

init_hal_bluetooth

init_hal_camera

init_hal_gps

# init_hal_gralloc

init_hal_hwcomposer

init_hal_lights

init_hal_power

# init_hal_sensors

# init_tscal

init_ril

post_init

}

[Michael Goffioul]

The same build doesn't boot into VirtualBox either, zygote keeps aborting. Note that my build is 32-bits only, not sure whether it makes any difference.

One possible relevant log line (when run in VirtualBox):

F zygote  : jni_internal.cc :508] JNI FatalError called: frameworks/base/core/jni/com_android_internal_os_Zygote.cpp:511: Unable to construct file descriptor table.

Michael.

[fguy]

The same build doesn't boot into VirtualBox either, zygote keeps aborting. Note that my build is 32-bits only, not sure whether it makes any difference.

One possible relevant log line (when run in VirtualBox):

F zygote  : jni_internal.cc :508] JNI FatalError called: frameworks/base/core/jni/com_android_internal_os_Zygote.cpp:511: Unable to construct file descriptor table. 

 I loaded 8.0 on vmware 12 - it only works in live with the key nomodeset

[Michael Goffioul]

On Sun, Sep 17, 2017 at 5:13 PM, Michael Goffioul <michael....@gmail.com> wrote:

The same build doesn't boot into VirtualBox either, zygote keeps aborting. Note that my build is 32-bits only, not sure whether it makes any difference.

One possible relevant log line (when run in VirtualBox):

F zygote  : jni_internal.cc :508] JNI FatalError called: frameworks/base/core/jni/com_android_internal_os_Zygote.cpp:511: Unable to construct file descriptor table.

This is actually preceeded with this:

E zygote  : Not whitelisted /android/sys/kernel/debug/tracing/trace_marker

According to frameworks/base/core/jni/fd_utils.cpp it should be whitelisted, except for the prefix "/android"

Michael.

[Michael Goffioul]

That does solve the problem on VirtualBox.

[fguy]

понедельник, 18 сентября 2017 г., 0:25:13 UTC+3 пользователь Michael Goffioul написал:

The patch https://osdn.net/projects/android-x86/scm/git/frameworks-base/commits/bca7b1f81f20f96ef1eeb2fd54a4eba7baa3e50e#fd-core/jni/fd_utils-inl.h from 7.1 is not used in 8.0 - I do not know the reason 

[Mauro Rossi]

frameworks/base is not complete yet,

experimental porting of bca7b1f81f20f96ef1eeb2fd54a4eba7baa3e50e to oreo-x86

is in the attachment.

M.

[Chih-Wei Huang]

2017-09-17 15:48 GMT+08:00 fguy <fguy...@gmail.com>:
> Over the past week, we have made significant progress - the init.sh file is
> launched, a usb flash is connected, the development menu is launched, and
> the supersu runs. Still not working mesa, wifi, camera and sound. From the

If you have any bugfix or improvement, please submit patches.

The sound should work now (at least works for me).

> wishes I would like to return the old layout of the system.sfs from
> system.img.

What did you mean?

[fguy]

> wishes I would like to return the old layout of the system.sfs from
> system.img.

What did you mean?

If unsquashfs "system.sfs" from 8.0 then we get the folder with the files and not the file "system.img" as in 7.1 

[Chih-Wei Huang]

Yes. I think it's better.
Why do you want the old layout?

[fguy]

Yes. I think it's better.
Why do you want the old layout?

 I install the android in a folder in the form of img files and unpacking in "systems.img" allows it to modify for normal operation of supersu and other modifications.

[youling 257]

where better ? system.sfs read-only file system，if system.sfs and data.img on NTFS partition，how to modify system ? 

even ext4 partition，i perfer like use system.img and data folder

在 2017年9月18日星期一 UTC+8下午1:37:27，Chih-Wei Huang写道：

[Chih-Wei Huang]

You can just copy the system folder and modify it.
Note this only works when installed to Linux filesystem
(i.e., ext4 or equivalent).
If you copy system folder to ntfs/vfat, you may
encounter some permission issue at runtime.

The reason we put system.img in system.sfs is
in the beginning of the project, I was unable to
create a squashfs with correct file permissions
that Android needs. Therefore I had to indirectly
create system.sfs by compressing system.img.

Now AOSP has officially supported squashfs
so we can create system.sfs directly with
correct file permissions. There is no reason to
use old layout which compromises the
performance.

[fguy]

You can just copy the system folder and modify it.
Note this only works when installed to Linux filesystem
(i.e., ext4 or equivalent).
If you copy system folder to ntfs/vfat, you may
encounter some permission issue at runtime.

I usually install (copy) a folder on the drive "C:" with NTFS. For startup I use the bootloader "refind" with the USB stick. It allows you to use the touchscreen to select the bootable OS. This installation option makes it easy to test various assemblies of android. For myself, I returned your patches for assembly through "system.img", but it may be necessary for other users. 

[fguy]

If you have any bugfix or improvement, please submit patches.

I collect with your patches. To integrate SuperSU I use https://github.com/BlissRoms-x86/vendor-supersu with files from SuperSU 2.82-SR3. Unfortunately, after updating the binary files, the android does not load, but the su works without updating.

 

The sound should work now (at least works for me).

 I have problems with the sound on your core 4.9 and in 7.1.2-r33. Perhaps in 8.0 they are even more exacerbated. For 7.1, I use the 4.11 kernel from maurossi, but Android 8.0 does not load it. I also tried to collect 8.0 from the old MESA 17.1.17 nougat-x86 - it just did not change.

[fguy]

frameworks/base is not complete yet,

experimental porting of bca7b1f81f20f96ef1eeb2fd54a4eba7baa3e50e to oreo-x86

is in the attachment.

Unfortunately this patch did not create any noticeable changes. 

[youling 257]

do you remember your 0001-DO-NOT-MERGE-temporary-workaround-for-Fds-whitelist-.patch ?

在 2017年9月18日星期一 UTC+8上午6:48:16，Mauro Rossi写道：

[Mauro Rossi]

Please be sure you have rebuilt the related code, I don't know why at first make I had no apparent effect,

btw the patch only allows to have /data/log.txt when booting in DEBUG

Mauro

[Michael Goffioul]

That worked for me. After applying this patch, and disabling HLA sensors in init.sh, I could get the build to boot LIVE in VirtualBox. No need to set nomodeset.

I didn't run exhaustive test, but I noticed Settings is crashing on start, because it can find the service wifip2p.

Michael.

[dgdn]

built with only the init.sh changes mentioned in this thread....

boots fine for me BUT get
Bluetooth has stopped
Hardware Collector has stopped
Process system has stopped

after killing them, everything comes up ok....

using included browser, internet works fine....

video/sound doesn't work when going to a website....

BUT video/sound works fine when going to youtube.com....

tested both x86_64 + x86 + x86 legacy(64 bit kernel 32 bit user)

on  two different systems....intel i3 with 24G mem + intel core duo with 3G mem

looking at asus/fugu shows lots of changes for oreo.....

new hal stuff.....https://source.android.com/devices/graphics/implement-hwc

new security stuff....

people should be patient as there are only so many hours in a day and
not many devoting exclusive time to oreo porting.....!!!

[dgdn]

opps....wrong pointer given in my previous reply for new hal stuff....should be

https://source.android.com/devices/architecture/configstore/

[Faheem Ahmed]

SEEMS YOU NEED TO GIVE IT MORE TIME PLUS WORK HARD!  AND ALSO ADD REMIX OS INSTALLER LINK TO WEBSITE FOR NONTECHIES

On Sunday, September 10, 2017 at 11:29:41 AM UTC+5, Chih-Wei Huang wrote:

Hi all,
The full oreo-x86 branch is pushed.
Get it as usual:

repo init -b oreo-x86 -u git://git.osdn.net/gitroot/android-x86/manifest
repo sync --no-tags --no-clone-bundle

There are many issues to be solved.
I just list some of them:

* Unable to run init.sh in init.*.rc

* Some apps crash on booting
* Mesa not work
* Wifi, BT, camera, sensors, usb not work. But
  Ethernet works. Audio should work if you
  manually initialize the sound card.
* No superuser

[fguy]

* Mesa not work 

I tested the android intel iA Oreo. It works MESA on skylake 6Y30. Also there is a sound but very quiet, it sees a microsd and a usb flash drive.

 

* usb not work

 I noticed that your assembly Oreo sees the USB flash drive when "androidboot.hardware=android_x86_64" is added to the kernel boot line.

[Chih-Wei Huang]

2017-10-01 17:42 GMT+08:00 fguy <fguy...@gmail.com>:
>
>> * usb not work
>
> I noticed that your assembly Oreo sees the USB flash drive when
> "androidboot.hardware=android_x86_64" is added to the kernel boot line.

Yes... in 8.0 fs_mgr use ro.boot.hardware.
This is an incompatible change which is bad.
Indeed I already have a fix but forgot to push.
Now it's in.

[Mauro Rossi]

To be precise, the submitted patch is necessary to boot in DEBUG mode,
so it is proposed to be applied to frameworks/base in oreo-x86 branch
M.

[Mauro Rossi]

Il giorno domenica 1 ottobre 2017 11:42:54 UTC+2, fguy ha scritto:
> * Mesa not work 
>
>
>
> I tested the android intel iA Oreo. It works MESA on skylake 6Y30.

Hi,
Is it correct that CrOS, Android-IA are both using minigbm as gralloc implementation and CrOS/AOSP will come ti a point It will be mantained as common library?

Could someone pinpoint the changes needed to try minigbm(drm_hwcomposer if necessary) with Intel GPUs?

Thanks

[fguy]

Yes... in 8.0 fs_mgr use ro.boot.hardware.
This is an incompatible change which is bad.
Indeed I already have a fix but forgot to push.
Now it's in.

it worked, and what are the prospects for Mesa and Wifi in 8.0?

I ran into another interesting effect with the 4.9.31 kernel - if I load 7.1 from the system.sfs, then the sound works fine, and if from the systems.img, there is no sound after boot. This effect is absent on the newer kernel 4.11 and 4.14-rc2 from Mauro Rossi.

[fguy]

Is it correct that CrOS, Android-IA are both using minigbm as gralloc implementation and CrOS/AOSP will come ti a point It will be mantained as common library?

Could someone pinpoint the changes needed to try minigbm(drm_hwcomposer if necessary) with Intel GPUs?

Did you get to run MESA via minigbm on 8.0 x64? 

[Mauro Rossi]

First attempt with minigbm on  intel i965 did not work, error: E cannot get drm fd

[1] https://github.com/maurossi/minigbm

[2] https://github.com/maurossi/device_generic_common/tree/minigbm_tests

[Lambdadroid]

On Thu, Oct 5, 2017 at 8:34 PM, Mauro Rossi <issor...@gmail.com> wrote:
> First attempt with minigbm on intel i965 did not work, error: E cannot get
> drm fd
>
> [1] https://github.com/maurossi/minigbm
> [2] https://github.com/maurossi/device_generic_common/tree/minigbm_tests

There are a few things that will cause problems here (all of these are
afaik, please correct me if I'm wrong with anything):

1) That version only has framebuffers implemented for Gralloc 1.0,
which requires additional patches in Mesa. I have an older version
that also implemented it for Gralloc 0, but I dropped it because I
wasn't using it anymore. Let me know if I should look it up.

Patch: "AndroidIA: Adds support for gralloc1.0 and uses lockflex for
gralloc1.0": https://github.com/me176c-dev/external_mesa/commit/1b8581b86216f1caec808b0d6720ec36340c9abc

2) Minigbm doesn't implement GRALLOC_MODULE_PERFORM_GET_DRM_FD that is
used by Mesa to get the DRM file descriptor. If you decide to use
Gralloc 0, you can probably easily patch it into minigbm, otherwise
Chromium/Android-IA decided to open render nodes separately from Mesa
instead.

Patch: "CHROMIUM: egl/android: Support opening render nodes from
within EGL": https://github.com/me176c-dev/external_mesa/commit/79ddc2252d9120e98d7abe16b9ae95b204c78935

3) On Oreo, you may need to set the Graphics allocator service to
passthrough mode (against the recommendation of Google). Setting it to
binderized mode caused problems with the framebuffer for me (because
that is always passthrough).

Chromium/Android-IA use additional patches to remove the compile
dependency on drm_gralloc entirely. I'm not sure if you can use these
because gbm_gralloc is likely using the same interface. Technically,
they shouldn't be needed though:

https://github.com/me176c-dev/external_mesa/commit/ec681adec424b822da5997fe5da00fc9a31ecabc
https://github.com/me176c-dev/external_mesa/commit/a620a1f8a729dd926c1a7f00c4c71041eae9526b
https://github.com/me176c-dev/external_mesa/commit/0223f8446bb8783511312dd14e16422dbf2ffbb7

[fguy]

I do not know the android in such subtleties, but it seems to me that we need to take a few more components from Intel:

https://github.com/intel/external-libdrm

https://github.com/intel/external-mesa

https://github.com/intel/external-vulkan_hal

The necessary settings for the assembly can be found here:

https://github.com/android-ia/device-androidia-mixins/tree/o_master/groups/graphics/android_ia

[Lambdadroid]

On Fri, Oct 6, 2017 at 5:31 PM, fguy <fguy...@gmail.com> wrote:
> I do not know the android in such holes, but it seems to me that we need to
> take a number of other components from Intel:

> https://github.com/intel/external-libdrm
> https://github.com/intel/external-mesa
> https://github.com/intel/external-vulkan_hal
>
> The necessary settings for the assembly can be found here:
> https://github.com/android-ia/device-androidia-mixins/tree/o_master/groups/graphics/android_ia

Well, this is were the patches I just linked were taken from. I got
most of the patches I use from Android-IA.

[dgdn]

Oreo Boot To Home on start-up - oct30

    disabled intelfbdrm and all is ok........no other changes to get it to boot to home for me...

Boot Oreo x-86, x86_64 to home from live USB....A hint maybe for Developers

(full sync on oct 29th)

--------------------------------------------------------------------------------------------------------

........disabled intelfbdrm and all is ok........

(running on intel i3-3225 CPU @ 3.30GHz × 4, Graphics Intel® Ivybridge Desktop, 24GB mem, asus P8Z77-M motherboard with x86_64 build)

(running on intel Core2 Duo E7500 @ 2.93Ghz, Graphics Intel G41 Express chipset, 3GB mem, Asus P5G41-M motherboard with x86 build)

following change in init.sh worked for me.....leave it to the experts to figure out why

function init_hal_gralloc()

{

case "$(cat /proc/fb | head -1)" in

*virtiodrmfb)

if [ "$HWACCEL" != "0" ]; then

set_property ro.hardware.hwcomposer drm

set_property ro.hardware.gralloc gbm

fi

;;

- 0*inteldrmfb|0*radeondrmfb|0*nouveaufb|0*svgadrmfb|0*amdgpudrmfb)

+ 0*radeondrmfb|0*nouveaufb|0*svgadrmfb|0*amdgpudrmfb)

if [ "$HWACCEL" != "0" ]; then

.

.

.

 browser works, video/audio plays fine in browser(chrome added), both from youtube & video clips on cnn web site...

 Still other issues such as bluetooth app crashes at startup, processes system crashes at startup but those

 both would crash just running vesa mode before the added change....just dismiss them and continue....so FWIW

On Wednesday, September 20, 2017 at 6:53:18 PM UTC-4, dgdn wrote:

[Chih-Wei Huang]

2017-11-01 1:17 GMT+08:00 dgdn <rbg...@gmail.com>:
> Oreo Boot To Home on start-up - oct30
>
> disabled intelfbdrm and all is ok........no other changes to get it to
> boot to home for me...
>
> Boot Oreo x-86, x86_64 to home from live USB....A hint maybe for Developers
> (full sync on oct 29th)
> --------------------------------------------------------------------------------------------------------
>
> ........disabled intelfbdrm and all is ok........
>
> (running on intel i3-3225 CPU @ 3.30GHz × 4, Graphics Intel® Ivybridge
> Desktop, 24GB mem, asus P8Z77-M motherboard with x86_64 build)
> (running on intel Core2 Duo E7500 @ 2.93Ghz, Graphics Intel G41 Express
> chipset, 3GB mem, Asus P5G41-M motherboard with x86 build)
>
> following change in init.sh worked for me.....leave it to the experts to
> figure out why

Known issue. The mesa (or gralloc.drm) is still broken.
You can just run with cmdline HWACCEL=0.
(equivalent to your hack)

> function init_hal_gralloc()
> {
> case "$(cat /proc/fb | head -1)" in
> *virtiodrmfb)
> if [ "$HWACCEL" != "0" ]; then
> set_property ro.hardware.hwcomposer drm
> set_property ro.hardware.gralloc gbm
> fi
> ;;
> - 0*inteldrmfb|0*radeondrmfb|0*nouveaufb|0*svgadrmfb|0*amdgpudrmfb)
> + 0*radeondrmfb|0*nouveaufb|0*svgadrmfb|0*amdgpudrmfb)
> if [ "$HWACCEL" != "0" ]; then
> .
> .
> .
>
> browser works, video/audio plays fine in browser(chrome added), both from
> youtube & video clips on cnn web site...
>
> Still other issues such as bluetooth app crashes at startup, processes
> system crashes at startup but those
> both would crash just running vesa mode before the added change....just
> dismiss them and continue....so FWIW

[673536]

ya but oreo-8.0_r4 boots to home fine for me.....many changes from r4 to present r17 and as you had said before...

"people should be patient as there are only so many hours in a day and

not many devoting exclusive time to oreo porting.....!!!"

[fguy]

Congratulations, your latest patches are launching the Mesa

воскресенье, 10 сентября 2017 г., 9:29:41 UTC+3 пользователь Chih-Wei Huang написал:

[Martin]

Finally!

As for now - oreo is stable enough to use it daily(for me).. after few small fixes. Thanks to all devs!

 

[Tomasz Witke]

Does vmwgfx should work now on oreo? I'm getting:

GRALLOC-DRM: unsupported driver: vmwgfx

On VirtualBox boot without nomodeset, but probably fallback to software renderer, because I'm getting no acceleration.

[Carpe Noctem]

A link for 32bits build for my testing will be great!

[Mauro Rossi]

I'm seeying a similar issue with amdgpudrmfb after having enforced  si_support in the kernel itself.

 GRALLOC-DRM: unsupported driver: amdgpu

There is  some problem with gralloc_drm_pipe in oreo-x86

Mauro

[Mauro Rossi]

In order to avoid the wifip2p issue, 

a dirty hack for packages/apps/Settings  is in the attachment,

for testing purposes only 

BTW another problem raises when opening Developer options

11-19 22:10:30.941  7278  7278 D SubSettings: Launching fragment com.android.settings.development.DevelopmentSettings

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: No service published for: persistent_data_block

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: android.os.ServiceManager$ServiceNotFoundException: No service published for: persistent_data_block

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.os.ServiceManager.getServiceOrThrow(ServiceManager.java:75)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.SystemServiceRegistry$69.createService(SystemServiceRegistry.java:745)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.SystemServiceRegistry$69.createService(SystemServiceRegistry.java:743)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.SystemServiceRegistry$StaticServiceFetcher.getService(SystemServiceRegistry.java:953)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.SystemServiceRegistry.getSystemService(SystemServiceRegistry.java:880)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.ContextImpl.getSystemService(ContextImpl.java:1651)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.view.ContextThemeWrapper.getSystemService(ContextThemeWrapper.java:171)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.Activity.getSystemService(Activity.java:5903)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at com.android.settings.development.DevelopmentSettings.onCreate(DevelopmentSettings.java:386)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.Fragment.performCreate(Fragment.java:2593)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.FragmentManagerImpl.moveToState(FragmentManager.java:1234)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.FragmentManagerImpl.addAddedFragments(FragmentManager.java:2415)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.FragmentManagerImpl.executeOpsTogether(FragmentManager.java:2194)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.FragmentManagerImpl.removeRedundantOperationsAndExecute(FragmentManager.java:2148)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.FragmentManagerImpl.execPendingActions(FragmentManager.java:2049)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.FragmentManagerImpl.executePendingTransactions(FragmentManager.java:798)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at com.android.settings.SettingsActivity.switchToFragment(SettingsActivity.java:768)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at com.android.settings.SettingsActivity.launchSettingFragment(SettingsActivity.java:432)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at com.android.settings.SettingsActivity.onCreate(SettingsActivity.java:345)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.Activity.performCreate(Activity.java:6985)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.Instrumentation.callActivityOnCreate(Instrumentation.java:1213)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:2770)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.ActivityThread.handleLaunchActivity(ActivityThread.java:2892)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.ActivityThread.-wrap11(Unknown Source:0)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.ActivityThread$H.handleMessage(ActivityThread.java:1593)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.os.Handler.dispatchMessage(Handler.java:105)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.os.Looper.loop(Looper.java:164)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at android.app.ActivityThread.main(ActivityThread.java:6541)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at java.lang.reflect.Method.invoke(Native Method)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at com.android.internal.os.Zygote$MethodAndArgsCaller.run(Zygote.java:240)

11-19 22:10:30.944  7278  7278 E SystemServiceRegistry: at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:767)

11-19 22:10:30.945  2838  4423 W ActivityManager:   Force finishing activity com.android.settings/.SubSettings

[youling 257]

did you boot success on Bay trail ?

在 2017年9月18日星期一 UTC+8上午3:30:17，Michael Goffioul写道：

I've tried the new branch on my 2 test hardware and in both cases boot hangs shortly after switching to the frame buffer, it doesn't get to the boot logo. My 2 hardware are Atom D525 and Baytrail. I can't get any useful debug log, because the screen is cleared and I only get the cursor in the top-left corner. Attached keyboard does not work. Even in debug mode, the 2nd and 3rd console becomes unaccessible once the screen it cleared. I also try to set debug.sf.nobootanimation=1, but it doesn't change anything.

Michael.

On Sun, Sep 17, 2017 at 3:48 AM, fguy <fguy...@gmail.com> wrote:

Over the past week, we have made significant progress - the init.sh file is launched, a usb flash is connected, the development menu is launched, and the supersu runs. Still not working mesa, wifi, camera and sound. From the wishes I would like to return the old layout of the system.sfs from system.img.

--
You received this message because you are subscribed to the Google Groups "Android-x86" group.
To unsubscribe from this group and stop receiving emails from it, send an email to android-x86...@googlegroups.com.
To post to this group, send email to andro...@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.

For more options, visit https://groups.google.com/d/optout.

[fguy]

file wpa_supplicant in oreo is now here system\vendor\bin\hw\

I changed the path in the init file, but it still did not help launch the wifi

понедельник, 20 ноября 2017 г., 2:46:44 UTC+3 пользователь Mauro Rossi написал:

[Mauro Rossi]

Il giorno lunedì 20 novembre 2017 12:07:18 UTC+1, fguy ha scritto:

file wpa_supplicant in oreo is now here system\vendor\bin\hw\

I changed the path in the init file, but it still did not help launch the wifi

I also tried to launch p2p_supplicant, as seen in "device" projects bits of other Android based Oreo releases 

Lineage OS 15.0, android-ia and me176c project

https://github.com/me176c-dev/device_asus_me176c/tree/oreo

I have to say that that I'm just doing trial and error (or lucky hack)

as I lack basics about Android services and Oreo is particularly nasty too.

What I'm doing is to to check if AOSP did some changes in android_8.0.0_r34 or master in the .java code triggering the wifip2p and persistent_data_block errors (Settings has stopped),

but I really do not understand why not having Wifi Display, wifip2p or not having Wifi at all

should cause all those fatal errors errors instead of graceful message on screen or simply wifi, wfd greyed out.

To me it seams like, you don't have a single feature/service and you do not open Settings at all.

So picky...

Mauro

[Chih-Wei Huang]

2017-11-20 19:07 GMT+08:00 fguy <fguy...@gmail.com>:
> file wpa_supplicant in oreo is now here system\vendor\bin\hw\
> I changed the path in the init file, but it still did not help launch the
> wifi

I have fixed the wifi function.
Please sync to get it.

[fguy]

вторник, 21 ноября 2017 г., 12:01:48 UTC+3 пользователь Chih-Wei Huang написал:

I have fixed the wifi function.
Please sync to get it.

I rebuilt with your patch and kernel 4.14, but the wifi still does not work 

[Chih-Wei Huang]

2017-11-22 16:23 GMT+08:00 fguy <fguy...@gmail.com>:
> вторник, 21 ноября 2017 г., 12:01:48 UTC+3 пользователь Chih-Wei Huang
> написал:
>>
>> I have fixed the wifi function.
>> Please sync to get it.

I pushed a wrong (incomplete) branch.
Sorry. Please sync again.

[fguy]

I pushed a wrong (incomplete) branch.
Sorry. Please sync again.

In the new version of wifi also does not work.

[Gerard Dosaigües]

Can you share this iso build to test it in other devices?

[673536]

Enable Wifi for Android 8.0

commit | commitdiff | tree

Chih-Wei Huang [Tue, 21 Nov 2017 01:43:43 -0500 (14:43 +0800)] 

Enable Wifi for Android 8.0

* Change wpa_supplicant path to /vendor

* Add wificond

* Allow wpa_supplicant to access /dev/rfkill

* Add HIDL interface android.hardware.wifi.supplicant

* Add wpa_supplicant_overlay.conf to disable P2P

  (a bug of android.hardware.wifi.supplicant?)

missing maybe ???

BOARD_WPA_SUPPLICANT_PRIVATE_LIB ?= private_lib_driver_cmd

WIFI_DRIVER_MODULE_PATH ?= path to driver module

[Chih-Wei Huang]

2017-11-23 6:24 GMT+08:00 673536 <dec6...@gmail.com>:
> Enable Wifi for Android 8.0
> commit | commitdiff | tree
> Chih-Wei Huang [Tue, 21 Nov 2017 01:43:43 -0500 (14:43 +0800)]
> Enable Wifi for Android 8.0
>
> * Change wpa_supplicant path to /vendor
> * Add wificond
> * Allow wpa_supplicant to access /dev/rfkill
> * Add HIDL interface android.hardware.wifi.supplicant
> * Add wpa_supplicant_overlay.conf to disable P2P
> (a bug of android.hardware.wifi.supplicant?)
>
>
> missing maybe ???
>
> BOARD_WPA_SUPPLICANT_PRIVATE_LIB ?= private_lib_driver_cmd
> WIFI_DRIVER_MODULE_PATH ?= path to driver module

We don't need that.

> On Wednesday, November 22, 2017 at 1:10:53 PM UTC-5, fguy wrote:
>>>
>>> I pushed a wrong (incomplete) branch.
>>> Sorry. Please sync again.
>>>
>> In the new version of wifi also does not work.

Then please collect some logs, including dmesg.

[fguy]

четверг, 23 ноября 2017 г., 5:12:03 UTC+3 пользователь Chih-Wei Huang написал:

Then please collect some logs, including dmesg.

8.0 + k4.9 - there is a wifi, the sound is not always

8.0 + k4.14 - there is sound and there is no wifi

7.1.2 + k4.14 - there is sound and wifi

[Mauro Rossi]

Hi fguy,

just to double check,

with stock oreo-x86 build with kernel 4.9, I still see the "Settings has stopped" error if I do not apply the Settings hack to skip getting the status of wifip2p and wfd.

Is it the same for you? I guess no, but then I do not understand why I am affected

Are you able to enter in Developers menu' in the settings in the three cases you mentioned?

Mauro

[fguy]

суббота, 25 ноября 2017 г., 13:21:18 UTC+3 пользователь Mauro Rossi написал:

just to double check,

with stock oreo-x86 build with kernel 4.9, I still see the "Settings has stopped" error if I do not apply the Settings hack to skip getting the status of wifip2p and wfd.

Is it the same for you? I guess no, but then I do not understand why I am affected

Are you able to enter in Developers menu' in the settings in the three cases you mentioned?

I collect the Settings application from AOSP and I have no errors when I enter the developer menu and the WiFi settings. I also returned to the old way of assembling the system.sfs file from the systems.img (as in 7.1.2).

Android 8.0 contains still many errors:

- after downloading, the message "applications "system" does not respond";

- closes with an error Keyboard when switching to a digital layout (123);

- does not show the video, there is only sound in the YouTube and gallery. 

[fguy]

Unfortunately, further testing of the android 7.1.2 with the 4.14 kernel showed unstable work of the wifi. I checked both versions of the 4.14 kernel - from the android x86 and from Mauro Rossi.

[Chih-Wei Huang]

It's very strange.
I have the same error in the log but
my Settings and Developer option won't crash.

My log:
https://drive.google.com/open?id=1ntJnB5eW2ZEwDm1840BdEUxCSUXcRT8o

[Mauro Rossi]

In the attachment logs of wifi not firing up on VAIO (I'm using ethernet)

Just for testing and to have the possibility to enable "stay awake" via GUI in developers menu,

I'm using the two hacks in this branch, but only because I don't know how to solve

and I don't even know if there is a problem:

https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_wfd_hack

M.

[Chih-Wei Huang]

2017-12-02 22:29 GMT+08:00 Mauro Rossi <issor...@gmail.com>:
> In the attachment logs of wifi not firing up on VAIO (I'm using ethernet)
>
> Just for testing and to have the possibility to enable "stay awake" via GUI
> in developers menu,
> I'm using the two hacks in this branch, but only because I don't know how to
> solve
> and I don't even know if there is a problem:
>
> https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_wfd_hack

I have applied a patch from AOSP master branch
to fix no wifip2p service issue.

[Mauro Rossi]

Hi,

after repo sync and rebuild I'm still affected by Setting crash when opening,
this implies that WiFi cannot be enabled,
in order to avoid the problem I still need [1]

[1] https://github.com/maurossi/packages_apps_Settings/commit/663f0f21cb8302023b32ec1c15d41b993eca016f

while to optionally avoid crash when opening Developers sub-menu
I use [2]

[2] https://github.com/maurossi/packages_apps_Settings/commit/9999c0685d24b408947137f03a7e5f940fb017cf

These two (abominable) hacks are currently working quite well,
but I use them just for testing purposes.
Mauro