Discussion:
Coffee Lake intel GPU support on android nougat x86-rc1
(too old to reply)
YoungJoon Lee
2018-01-13 15:27:03 UTC
Permalink
Hello.

I am trying to use android nougat x86 rc1 on Coffee Lake system.

Linux kernel community said
[*] Enable alpha quality support for new Intel hardware by default
Makes it possible to use new intel GPU.

Q1. Can I use kernel branch 'kernel-4.14' source on Android nougat?
Q2. Is there any modification of code is needed for HAL using new GPU?

Best regards.
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-15 15:12:49 UTC
Permalink
Thank you for your answer. I am testing maurossi's kernel-4.15rc8. I have
to modify "./device/generic/common/tp_smapi/hdaps.c" little.

Finally I can build it.
Test result :
This build work good on Sky Lake system but can't boot on Coffee Lake
system. Android service is killed and restarted recursively. I attach
logcat and dmesg. What I should do for fix this Error?
Post by YoungJoon Lee
Hello.
I am trying to use android nougat x86 rc1 on Coffee Lake system.
Linux kernel community said
[*] Enable alpha quality support for new Intel hardware by default
Makes it possible to use new intel GPU.
Q1. Can I use kernel branch 'kernel-4.14' source on Android nougat?
cd kernel/
git fetch x86 kernel-4.14
git checkout x86/kernel-4.14
(note only 64-bit defconfig is available now)
You can also refer to Mauro's repo which
https://github.com/maurossi/linux
He also released some testing iso with new kernel.
Just search the forum for it.
Post by YoungJoon Lee
Q2. Is there any modification of code is needed for HAL using new GPU?
Mesa is probably the only one needs to be updated.
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-15 15:59:40 UTC
Permalink
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.

Best regards.
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8. I
have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee Lake
system. Android service is killed and restarted recursively. I attach
logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-15 16:25:55 UTC
Permalink
I dump dmesg before bootstrap is not complete. I reattach new dmesg log.
Post by YoungJoon Lee
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.
Best regards.
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8. I
have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee Lake
system. Android service is killed and restarted recursively. I attach
logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-15 16:32:50 UTC
Permalink
I have to change hdaps.c on ./device/generic/common/tp_smapi/

here is diff.
472d471
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
474,476d472
< #else
< static void hdaps_mousedev_poll(struct timer_list *unused)
< #endif
786d781
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
789,791d783
< #else
< timer_setup(&hdaps_timer, hdaps_mousedev_poll, 0);
< #endif
Post by YoungJoon Lee
I dump dmesg before bootstrap is not complete. I reattach new dmesg log.
Post by YoungJoon Lee
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.
Best regards.
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8. I
have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee Lake
system. Android service is killed and restarted recursively. I attach
logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
Mauro Rossi
2018-01-15 17:29:19 UTC
Permalink
If you are trying latest kernel, here is the two essential patches I'm
using for kernel 4.15:

*device/generic/common*
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337


Some essential permission settings (needed since kernel 4.14) to get wifi
working:

*system/connectivity/wificond*
https://github.com/maurossi/system_connectivity_wificond/commit/41e8ae0c075d6b70356e5942248f5814600d6743


Here are the optional ones:

optional: some hack to have Developer Settings working and possibility to
enable "Stay Awake" (if needed the last two commits)

packages/apps/Settings
https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_810_hacks_v2


optional libdrm 2.4.89
https://github.com/maurossi/drm/commits/2.4.89_android-x86

optional mesa-dev
https://github.com/maurossi/mesa/tree/17.4.0devel_w02

Mauro
Post by YoungJoon Lee
I have to change hdaps.c on ./device/generic/common/tp_smapi/
here is diff.
472d471
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
474,476d472
< #else
< static void hdaps_mousedev_poll(struct timer_list *unused)
< #endif
786d781
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
789,791d783
< #else
< timer_setup(&hdaps_timer, hdaps_mousedev_poll, 0);
< #endif
Post by YoungJoon Lee
I dump dmesg before bootstrap is not complete. I reattach new dmesg log.
Post by YoungJoon Lee
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.
Best regards.
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8. I
have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee Lake
system. Android service is killed and restarted recursively. I attach
logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
Mauro Rossi
2018-01-15 17:33:10 UTC
Permalink
..but...if you are building nougat, as mentioned in the title,
you should need only to patch device/generic/common

sorry, my list was for oreo-x86
Post by Mauro Rossi
If you are trying latest kernel, here is the two essential patches I'm
*device/generic/common*
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
Some essential permission settings (needed since kernel 4.14) to get wifi
*system/connectivity/wificond*
https://github.com/maurossi/system_connectivity_wificond/commit/41e8ae0c075d6b70356e5942248f5814600d6743
optional: some hack to have Developer Settings working and possibility to
enable "Stay Awake" (if needed the last two commits)
packages/apps/Settings
https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_810_hacks_v2
optional libdrm 2.4.89
https://github.com/maurossi/drm/commits/2.4.89_android-x86
optional mesa-dev
https://github.com/maurossi/mesa/tree/17.4.0devel_w02
Mauro
Post by YoungJoon Lee
I have to change hdaps.c on ./device/generic/common/tp_smapi/
here is diff.
472d471
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
474,476d472
< #else
< static void hdaps_mousedev_poll(struct timer_list *unused)
< #endif
786d781
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
789,791d783
< #else
< timer_setup(&hdaps_timer, hdaps_mousedev_poll, 0);
< #endif
Post by YoungJoon Lee
I dump dmesg before bootstrap is not complete. I reattach new dmesg log.
Post by YoungJoon Lee
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.
Best regards.
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8. I
have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee
Lake
Post by YoungJoon Lee
system. Android service is killed and restarted recursively. I
attach logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-15 22:27:51 UTC
Permalink
Thank you for your answers.

I also want to try oreo. Thank you very much. I will try oreo today. Best
regards.
Post by Mauro Rossi
..but...if you are building nougat, as mentioned in the title,
you should need only to patch device/generic/common
sorry, my list was for oreo-x86
Post by Mauro Rossi
If you are trying latest kernel, here is the two essential patches I'm
*device/generic/common*
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
Some essential permission settings (needed since kernel 4.14) to get wifi
*system/connectivity/wificond*
https://github.com/maurossi/system_connectivity_wificond/commit/41e8ae0c075d6b70356e5942248f5814600d6743
optional: some hack to have Developer Settings working and possibility
to enable "Stay Awake" (if needed the last two commits)
packages/apps/Settings
https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_810_hacks_v2
optional libdrm 2.4.89
https://github.com/maurossi/drm/commits/2.4.89_android-x86
optional mesa-dev
https://github.com/maurossi/mesa/tree/17.4.0devel_w02
Mauro
Post by YoungJoon Lee
I have to change hdaps.c on ./device/generic/common/tp_smapi/
here is diff.
472d471
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
474,476d472
< #else
< static void hdaps_mousedev_poll(struct timer_list *unused)
< #endif
786d781
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
789,791d783
< #else
< timer_setup(&hdaps_timer, hdaps_mousedev_poll, 0);
< #endif
Post by YoungJoon Lee
I dump dmesg before bootstrap is not complete. I reattach new dmesg log.
Post by YoungJoon Lee
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.
Best regards.
Post by YoungJoon Lee
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8.
I have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee
Lake
Post by YoungJoon Lee
system. Android service is killed and restarted recursively. I
attach logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-16 05:11:05 UTC
Permalink
One more question, Mauro Rossi.
Is -C option on android make system removed on Oreo? Which can be used on
Kernel's menuconfig?
Post by Mauro Rossi
..but...if you are building nougat, as mentioned in the title,
you should need only to patch device/generic/common
sorry, my list was for oreo-x86
Post by Mauro Rossi
If you are trying latest kernel, here is the two essential patches I'm
*device/generic/common*
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
Some essential permission settings (needed since kernel 4.14) to get wifi
*system/connectivity/wificond*
https://github.com/maurossi/system_connectivity_wificond/commit/41e8ae0c075d6b70356e5942248f5814600d6743
optional: some hack to have Developer Settings working and possibility
to enable "Stay Awake" (if needed the last two commits)
packages/apps/Settings
https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_810_hacks_v2
optional libdrm 2.4.89
https://github.com/maurossi/drm/commits/2.4.89_android-x86
optional mesa-dev
https://github.com/maurossi/mesa/tree/17.4.0devel_w02
Mauro
Post by YoungJoon Lee
I have to change hdaps.c on ./device/generic/common/tp_smapi/
here is diff.
472d471
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
474,476d472
< #else
< static void hdaps_mousedev_poll(struct timer_list *unused)
< #endif
786d781
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
789,791d783
< #else
< timer_setup(&hdaps_timer, hdaps_mousedev_poll, 0);
< #endif
Post by YoungJoon Lee
I dump dmesg before bootstrap is not complete. I reattach new dmesg log.
Post by YoungJoon Lee
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.
Best regards.
Post by YoungJoon Lee
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8.
I have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee
Lake
Post by YoungJoon Lee
system. Android service is killed and restarted recursively. I
attach logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
Mauro Rossi
2018-01-16 13:03:08 UTC
Permalink
Hi,
Post by YoungJoon Lee
One more question, Mauro Rossi.
Is -C option on android make system removed on Oreo? Which can be used on
Kernel's menuconfig?
I don't know about -C option uses, sorry, but it seams a make option to
launch a make "in another tree",
I never had to use it

When I move to recent kernel and there is a build error I do:

lunch android_x86_64-eng
cd out/target/program/x86_64/obj/kernel
make menuconfig
[enter as many times as requested]

when I have a working defconfig file I move it in kernel/arch/x86/configs
then I use the following option to build with custom config:

example:
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig

However by checking out latest kernel-4.15rc8, the default defconfig files
are already ok

In case you introduce different defconfigs, please make sure that you
have CONFIG_NO_GPU_AUTHENTICATION=y
because it's not a standard one, it should be preserved at build, but I'm
not sure it will survive a make oldconfig command.

The SIGABRT due to no EGLconfig may happen for following reasons:

- not having CONFIG_NO_GPU_AUTHENTICATION=y

- with latest mesa-dev the gamut wide colour support in mesa-dev causes
this,
This repo contains the latest patches/workarounds, courtesy of Tapani
Palli, https://github.com/maurossi/mesa/tree/17.4.0devel_w03

- failure in dlopen() of mesa EGL/gallium_dri libraries, maybe due to part
of the libraries being moved to /vendor/changes
This repo may help also for nougat-x86, in case you to build latest mesa
17.4.0devel: https://github.com/maurossi/drm/commits/2.4.89_android-x86

Mauro
Post by YoungJoon Lee
Post by Mauro Rossi
..but...if you are building nougat, as mentioned in the title,
you should need only to patch device/generic/common
sorry, my list was for oreo-x86
Post by Mauro Rossi
If you are trying latest kernel, here is the two essential patches I'm
*device/generic/common*
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
Some essential permission settings (needed since kernel 4.14) to get
*system/connectivity/wificond*
https://github.com/maurossi/system_connectivity_wificond/commit/41e8ae0c075d6b70356e5942248f5814600d6743
optional: some hack to have Developer Settings working and possibility
to enable "Stay Awake" (if needed the last two commits)
packages/apps/Settings
https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_810_hacks_v2
optional libdrm 2.4.89
https://github.com/maurossi/drm/commits/2.4.89_android-x86
optional mesa-dev
https://github.com/maurossi/mesa/tree/17.4.0devel_w02
Mauro
Post by YoungJoon Lee
I have to change hdaps.c on ./device/generic/common/tp_smapi/
here is diff.
472d471
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
474,476d472
< #else
< static void hdaps_mousedev_poll(struct timer_list *unused)
< #endif
786d781
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
789,791d783
< #else
< timer_setup(&hdaps_timer, hdaps_mousedev_poll, 0);
< #endif
Post by YoungJoon Lee
I dump dmesg before bootstrap is not complete. I reattach new dmesg log.
Post by YoungJoon Lee
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.
Best regards.
Post by YoungJoon Lee
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8.
I have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee
Lake
Post by YoungJoon Lee
system. Android service is killed and restarted recursively. I
attach logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
Mauro Rossi
2018-01-16 16:37:53 UTC
Permalink
Hi, a little errata corrige
Post by YoungJoon Lee
Hi,
One more question, Mauro Rossi.
Is -C option on android make system removed on Oreo? Which can be used on Kernel's menuconfig? 
I don't know about -C option uses, sorry, but it seams a make option to launch a make "in another tree",
I never had to use it
lunch android_x86_64-eng
cd out/target/program/x86_64/obj/kernel
make menuconfig
Correction: it is make oldconfig
Post by YoungJoon Lee
[enter as many times as requested]
when I have a working defconfig file I move it in kernel/arch/x86/configs
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
However by checking out latest kernel-4.15rc8, the default defconfig files are already ok
In case you introduce different defconfigs, please make sure that you have CONFIG_NO_GPU_AUTHENTICATION=y
because it's not a standard one, it should be preserved at build, but I'm not sure it will survive a make oldconfig command.
- not having CONFIG_NO_GPU_AUTHENTICATION=y
- with latest mesa-dev the gamut wide colour support in mesa-dev causes this,
This repo contains the latest patches/workarounds, courtesy of Tapani Palli, https://github.com/maurossi/mesa/tree/17.4.0devel_w03
- failure in dlopen() of mesa EGL/gallium_dri libraries, maybe due to part of the libraries being moved to /vendor/changes 
This repo may help also for nougat-x86, in case you to build latest mesa 17.4.0devel: https://github.com/maurossi/drm/commits/2.4.89_android-x86
Mauro
 
..but...if you are building nougat, as mentioned in the title,
you should need only to patch  device/generic/common
sorry, my list was for oreo-x86
device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
system/connectivity/wificond
https://github.com/maurossi/system_connectivity_wificond/commit/41e8ae0c075d6b70356e5942248f5814600d6743
optional:  some hack to have Developer Settings working and possibility to enable "Stay Awake" (if needed the last two commits)
packages/apps/Settings
https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_810_hacks_v2
optional libdrm 2.4.89
https://github.com/maurossi/drm/commits/2.4.89_android-x86
optional mesa-dev
https://github.com/maurossi/mesa/tree/17.4.0devel_w02
Mauro
I have to change hdaps.c on ./device/generic/common/tp_smapi/ 
here is diff. 
472d471
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
474,476d472
< #else
< static void hdaps_mousedev_poll(struct timer_list *unused)
< #endif
786d781
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
789,791d783
< #else
<       timer_setup(&hdaps_timer, hdaps_mousedev_poll, 0);
< #endif
I dump dmesg before bootstrap is not complete. I reattach new dmesg log. 
Thank you for your answer. I will boot to debug mode  and check it is loaded or not. And then I will modprobe manually if needed. 
Best regards.
Thank you for your answer. I am testing maurossi's kernel-4.15rc8. I have to
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee Lake
system. Android service is killed and restarted recursively. I attach logcat
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-17 03:28:56 UTC
Permalink
Thank you very much. I am check 3 point on your advise.


- On all of image I have built, kernel
setting CONFIG_NO_GPU_AUTHENTICATION is yes.
- mesa and drm : I am trying to use your code on my source. But not
tested yet.

I don't know detail but image working perfectly on skylake is not working
on skylake system. It is very hard to boot on coffee lake system.
Post by Mauro Rossi
Hi, a little errata corrige
Hi,
Il giorno martedì 16 gennaio 2018 06:11:06 UTC+1, YoungJoon Lee ha
One more question, Mauro Rossi.
Is -C option on android make system removed on Oreo? Which can be used
on Kernel's menuconfig?
I don't know about -C option uses, sorry, but it seams a make option to
launch a make "in another tree",
I never had to use it
lunch android_x86_64-eng
cd out/target/program/x86_64/obj/kernel
make menuconfig
Correction: it is make oldconfig
[enter as many times as requested]
when I have a working defconfig file I move it in
kernel/arch/x86/configs
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
However by checking out latest kernel-4.15rc8, the default defconfig
files are already ok
In case you introduce different defconfigs, please make sure that you
have CONFIG_NO_GPU_AUTHENTICATION=y
because it's not a standard one, it should be preserved at build, but
I'm not sure it will survive a make oldconfig command.
- not having CONFIG_NO_GPU_AUTHENTICATION=y
- with latest mesa-dev the gamut wide colour support in mesa-dev causes
this,
This repo contains the latest patches/workarounds, courtesy of Tapani
Palli, https://github.com/maurossi/mesa/tree/17.4.0devel_w03
- failure in dlopen() of mesa EGL/gallium_dri libraries, maybe due to
part of the libraries being moved to /vendor/changes
This repo may help also for nougat-x86, in case you to build latest mesa
17.4.0devel: https://github.com/maurossi/drm/commits/2.4.89_android-x86
Mauro
..but...if you are building nougat, as mentioned in the title,
you should need only to patch device/generic/common
sorry, my list was for oreo-x86
If you are trying latest kernel, here is the two essential patches I'm
device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
Some essential permission settings (needed since kernel 4.14) to get
system/connectivity/wificond
https://github.com/maurossi/system_connectivity_wificond/commit/41e8ae0c075d6b70356e5942248f5814600d6743
optional: some hack to have Developer Settings working and possibility
to enable "Stay Awake" (if needed the last two commits)
packages/apps/Settings
https://github.com/maurossi/packages_apps_Settings/commits/oreo-x86_810_hacks_v2
optional libdrm 2.4.89
https://github.com/maurossi/drm/commits/2.4.89_android-x86
optional mesa-dev
https://github.com/maurossi/mesa/tree/17.4.0devel_w02
Mauro
Il giorno lunedì 15 gennaio 2018 17:32:51 UTC+1, YoungJoon Lee ha
I have to change hdaps.c on ./device/generic/common/tp_smapi/
here is diff.
472d471
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
474,476d472
< #else
< static void hdaps_mousedev_poll(struct timer_list *unused)
< #endif
786d781
< #if LINUX_VERSION_CODE < KERNEL_VERSION(4,15,0)
789,791d783
< #else
< timer_setup(&hdaps_timer, hdaps_mousedev_poll, 0);
< #endif
I dump dmesg before bootstrap is not complete. I reattach new dmesg
log.
Thank you for your answer. I will boot to debug mode and check it is
loaded or not. And then I will modprobe manually if needed.
Best regards.
2018년 1월 16음 화요음 였전 12시 54분 34쎈 UTC+9, Chih-Wei Huang 님의 말:2018-01-15
Post by YoungJoon Lee
Thank you for your answer. I am testing maurossi's kernel-4.15rc8. I
have to
Post by YoungJoon Lee
modify "./device/generic/common/tp_smapi/hdaps.c" little.
Finally I can build it.
This build work good on Sky Lake system but can't boot on Coffee Lake
system. Android service is killed and restarted recursively. I attach
logcat
Post by YoungJoon Lee
and dmesg. What I should do for fix this Error?
I didn't see any i915 loaded message in dmesg.
What's "cat /proc/fb"?
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
spin877
2018-01-18 20:26:41 UTC
Permalink
Post by Mauro Rossi
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
I to speed up the build of the kernel I add this:

/device/generic/common/build/tasks/kernel.mk

mk_kernel := + $(hide) $(MAKE) -jX -C

X=cpu

I refer to build oreo-x86
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-20 05:20:05 UTC
Permalink
Thank you for your answer. I finally success working nougat on Coffee lake
system. I left How can I do.

1. HW : https://www.msi.com/Motherboard/Z370I-GAMING-PRO-CARBON-AC.html
2. Base : e8653d3
<https://osdn.net/projects/android-x86/scm/git/manifest/commits/e8653d3a463c6894daff37073b35b2f880e59a0e>
3. /Kernel : https://github.com/maurossi/linux/tree/kernel-4.15rc8
4. /external/mesa :
https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86
<https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86>
5. /external/llvm : https://github.com/maurossi/llvm/tree/nougat-x86_exports
6. /external/libdrm : Must use Default one.
7. code modification : /device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337

What I have tested(Working)

- Realtek ALC1220 Codec
- Realtek ALC1220 Codec Optical out :
/device/generic/common/alsa/PCH.state needed.(
https://github.com/getfeus/android-oreo-x86_android-auto/blob/master/device/generic/common/alsa/PCH.state
)
- Intel® Dual Band Wireless-AC 8265 card
- NVMe
- Intel I219-V Gigabit LAN controller
- enable_nativebridge : Only check command success. Not testing Glthread
fatal error I have reported.(
https://groups.google.com/forum/#!msg/android-x86/E8iRTfYNJ90/wbnaYCzwBwAJ;context-place=msg/android-x86/ilk69kFFcWY/WlPFjIhSAAAJ
)


What is not stable.

- During boot, ACPI error is displayed. Some kernel module crashed and
re-loaded. Every module is ACPI related of thermal sensor
related.(Mis-configured of defconfig?)
- I can't dump message from dmesg.(Before logging daemon start?) But,
Start with this string, ACPI Error: [_OSI] Namespace lookup failure,
AE_NOT_FOUND on early boot process(Prior to android device dev partition
detect). Very close to this post.
https://bugzilla.redhat.com/show_bug.cgi?id=1413342
- This message is disappeared on old kernel like 4.9.x but Coffee
lake GPU can only working on 4.13.x~. All Kernels I have tested have same
errors. But not care.(Not displayed after boot)
- I can't boot from any Oreo build.

Best regards.
Post by spin877
Post by Mauro Rossi
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
/device/generic/common/build/tasks/kernel.mk
mk_kernel := + $(hide) $(MAKE) -jX -C
X=cpu
No need to add it if you have added
-jX to make iso_img.
Submake is parallel execution automatically,
29417 pts/42 Sl+ 0:00 | \_ /mnt/cwhuang/oreo-x86/out/soong_ui
--make-mode -j24 iso_img
30617 pts/42 S+ 0:09 | \_
prebuilts/build-tools/linux-x86/bin/ninja -d keepdepfile iso_img -j 24
-f out/combined-android_x86_64.ninja -w dupbuild=err
30944 pts/42 S 0:00 | \_ /bin/sh -c /bin/bash -c
"(make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=\"
30945 pts/42 S 0:00 | \_ /bin/bash -c (make -j24
-C kernel O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=" /mnt/
30946 pts/42 S 0:00 | \_ make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE= /mnt//cwhuang
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-20 06:51:50 UTC
Permalink
Sorry. I post incorrect information.

/external/libdrm -> I use
https://github.com/maurossi/drm/tree/2.4.89_android-x86

Best regards.
Post by YoungJoon Lee
Thank you for your answer. I finally success working nougat on Coffee lake
system. I left How can I do.
1. HW : https://www.msi.com/Motherboard/Z370I-GAMING-PRO-CARBON-AC.html
2. Base : e8653d3
<https://osdn.net/projects/android-x86/scm/git/manifest/commits/e8653d3a463c6894daff37073b35b2f880e59a0e>
3. /Kernel : https://github.com/maurossi/linux/tree/kernel-4.15rc8
https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86
<https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86>
https://github.com/maurossi/llvm/tree/nougat-x86_exports
6. /external/libdrm : Must use Default one.
7. code modification : /device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
What I have tested(Working)
- Realtek ALC1220 Codec
/device/generic/common/alsa/PCH.state needed.(
https://github.com/getfeus/android-oreo-x86_android-auto/blob/master/device/generic/common/alsa/PCH.state
)
- Intel® Dual Band Wireless-AC 8265 card
- NVMe
- Intel I219-V Gigabit LAN controller
- enable_nativebridge : Only check command success. Not testing
Glthread fatal error I have reported.(
https://groups.google.com/forum/#!msg/android-x86/E8iRTfYNJ90/wbnaYCzwBwAJ;context-place=msg/android-x86/ilk69kFFcWY/WlPFjIhSAAAJ
)
What is not stable.
- During boot, ACPI error is displayed. Some kernel module crashed and
re-loaded. Every module is ACPI related of thermal sensor
related.(Mis-configured of defconfig?)
- I can't dump message from dmesg.(Before logging daemon start?) But,
Start with this string, ACPI Error: [_OSI] Namespace lookup failure,
AE_NOT_FOUND on early boot process(Prior to android device dev partition
detect). Very close to this post.
https://bugzilla.redhat.com/show_bug.cgi?id=1413342
- This message is disappeared on old kernel like 4.9.x but Coffee
lake GPU can only working on 4.13.x~. All Kernels I have tested have same
errors. But not care.(Not displayed after boot)
- I can't boot from any Oreo build.
Best regards.
Post by spin877
Post by Mauro Rossi
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
/device/generic/common/build/tasks/kernel.mk
mk_kernel := + $(hide) $(MAKE) -jX -C
X=cpu
No need to add it if you have added
-jX to make iso_img.
Submake is parallel execution automatically,
29417 pts/42 Sl+ 0:00 | \_ /mnt/cwhuang/oreo-x86/out/soong_ui
--make-mode -j24 iso_img
30617 pts/42 S+ 0:09 | \_
prebuilts/build-tools/linux-x86/bin/ninja -d keepdepfile iso_img -j 24
-f out/combined-android_x86_64.ninja -w dupbuild=err
30944 pts/42 S 0:00 | \_ /bin/sh -c /bin/bash -c
"(make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=\"
30945 pts/42 S 0:00 | \_ /bin/bash -c (make -j24
-C kernel O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=" /mnt/
30946 pts/42 S 0:00 | \_ make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE= /mnt//cwhuang
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
Mauro Rossi
2018-01-21 11:12:01 UTC
Permalink
This pattern of mesa 17.2.x + nougat-x86 libdrm working suggests
that the problem with Coffelake (gen 8) was most probably due to mesa
libraries moved to /vendor in mesa 17.3 and later
Also when I build nougat, with mesa 17.3 and later, I use x86/oreo-x86
version of external/drm_gralloc

In order to confirm I'm uploading experimental test images with mesa 17.3.3
here (3 hours to go):

https://drive.google.com/drive/folders/0B_OFHiIqgpSFMlk4bkhaMnAwbDA

Mauro
Post by YoungJoon Lee
Thank you for your answer. I finally success working nougat on Coffee lake
system. I left How can I do.
1. HW : https://www.msi.com/Motherboard/Z370I-GAMING-PRO-CARBON-AC.html
2. Base : e8653d3
<https://osdn.net/projects/android-x86/scm/git/manifest/commits/e8653d3a463c6894daff37073b35b2f880e59a0e>
3. /Kernel : https://github.com/maurossi/linux/tree/kernel-4.15rc8
https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86
<https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86>
https://github.com/maurossi/llvm/tree/nougat-x86_exports
6. /external/libdrm : Must use Default one.
7. code modification : /device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
What I have tested(Working)
- Realtek ALC1220 Codec
/device/generic/common/alsa/PCH.state needed.(
https://github.com/getfeus/android-oreo-x86_android-auto/blob/master/device/generic/common/alsa/PCH.state
)
- Intel® Dual Band Wireless-AC 8265 card
- NVMe
- Intel I219-V Gigabit LAN controller
- enable_nativebridge : Only check command success. Not testing
Glthread fatal error I have reported.(
https://groups.google.com/forum/#!msg/android-x86/E8iRTfYNJ90/wbnaYCzwBwAJ;context-place=msg/android-x86/ilk69kFFcWY/WlPFjIhSAAAJ
)
What is not stable.
- During boot, ACPI error is displayed. Some kernel module crashed and
re-loaded. Every module is ACPI related of thermal sensor
related.(Mis-configured of defconfig?)
- I can't dump message from dmesg.(Before logging daemon start?) But,
Start with this string, ACPI Error: [_OSI] Namespace lookup failure,
AE_NOT_FOUND on early boot process(Prior to android device dev partition
detect). Very close to this post.
https://bugzilla.redhat.com/show_bug.cgi?id=1413342
- This message is disappeared on old kernel like 4.9.x but Coffee
lake GPU can only working on 4.13.x~. All Kernels I have tested have same
errors. But not care.(Not displayed after boot)
- I can't boot from any Oreo build.
Best regards.
Post by spin877
Post by Mauro Rossi
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
/device/generic/common/build/tasks/kernel.mk
mk_kernel := + $(hide) $(MAKE) -jX -C
X=cpu
No need to add it if you have added
-jX to make iso_img.
Submake is parallel execution automatically,
29417 pts/42 Sl+ 0:00 | \_ /mnt/cwhuang/oreo-x86/out/soong_ui
--make-mode -j24 iso_img
30617 pts/42 S+ 0:09 | \_
prebuilts/build-tools/linux-x86/bin/ninja -d keepdepfile iso_img -j 24
-f out/combined-android_x86_64.ninja -w dupbuild=err
30944 pts/42 S 0:00 | \_ /bin/sh -c /bin/bash -c
"(make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=\"
30945 pts/42 S 0:00 | \_ /bin/bash -c (make -j24
-C kernel O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=" /mnt/
30946 pts/42 S 0:00 | \_ make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE= /mnt//cwhuang
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-21 16:18:06 UTC
Permalink
Thank you Mauro Rossi.
I will test your test image and report here.

Today I found few problems.

(Fatal) 1 : System is extremely heavy. All response is very slow(Slower
than skylake). I monitor details of top command. But it looks like there is
no process consume CPU 100%. I guess it is DVFS(Dynamic Frequency)
Issue.(Kernel message said ACPI related driver is unstable and thermal
driver also print error messages.) How can I debug performance issue? For
example, is there any way to disable power management or frequency
control(Working on maximum clock speed)? I remember that kerner setting on
defconfig's power profile is on demand.

(Not Fatal) 2 : App crashes when keyboard input pop up.(On some apps only).
I will attach logcat and left code here related to this error. It looks
like it is webview related issue. So, I will attach my device mk file which
is import opengapps.

01-21 13:13:56.677 16130 16118 F google-breakpad: -----END BREAKPAD MICRODUMP-----
01-21 13:13:56.800 15305 16118 F libc : Fatal signal 11 (SIGSEGV), code -6, fault addr 0x3bc9 in tid 16118 (Chrome_InProcGp)
01-21 13:13:56.899 16133 16133 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
01-21 13:13:56.899 16133 16133 F DEBUG : Build fingerprint: 'Android-x86/android_x86_64/x86_64:7.1.2/N2G48H/root01210301:userdebug/test-keys'
01-21 13:13:56.899 16133 16133 F DEBUG : Revision: '0'
01-21 13:13:56.899 16133 16133 F DEBUG : ABI: 'x86'
01-21 13:13:56.899 16133 16133 F DEBUG : pid: 15305, tid: 16118, name: Chrome_InProcGp >>> jp.co.yahoo.android.apps.navi <<<
01-21 13:13:56.899 16133 16133 F DEBUG : signal 11 (SIGSEGV), code -6 (SI_TKILL), fault addr 0x3bc9
01-21 13:13:56.899 16133 16133 F DEBUG : eax 00000000 ebx cf899d40 ecx 00008000 edx c0dbec58
01-21 13:13:56.899 16133 16133 F DEBUG : esi 00000000 edi 00000000
01-21 13:13:56.899 16133 16133 F DEBUG : xcs 00000023 xds 0000002b xes 0000002b xfs 0000006b xss 0000002b
01-21 13:13:56.899 16133 16133 F DEBUG : eip cf66012c ebp c0dbec78 esp c0dbe680 flags 00210282
01-21 13:13:56.899 16133 16133 F DEBUG :
01-21 13:13:56.899 16133 16133 F DEBUG : backtrace:
01-21 13:13:56.899 16133 16133 F DEBUG : #00 pc 005d812c /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #01 pc 000267b2 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #02 pc 00029bee /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #03 pc 0035e248 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #04 pc 0000ab1e /system/lib/libglapi.so
01-21 13:13:56.899 16133 16133 F DEBUG : #05 pc 00a12d91 /system/app/webview/webview.apk (offset 0x927000)


Here is mk file.
GAPPS_VARIANT := nano
PRODUCT_PACKAGES += Chrome \
Maps \
KeyboardGoogle \
GoogleTTS \
YouTube \
PixelIcons \
PixelLauncher \
Wallpapers \
PixelLauncherIcons \
GoogleServicesFramework \
GoogleLoginService



#PRODUCT_PACKAGES += Hangouts
GAPPS_EXCLUDED_PACKAGES := PrebuiltGmsCoreInstantApps \
FaceLock \
AndroidPlatformServices

GAPPS_FORCE_BROWSER_OVERRIDES := true

PRODUCT_PROPERTY_OVERRIDES := \
ro.kernel.android.gps=ttyUSB0 \
ro.kernel.android.gpsttybaud=4800

$(call inherit-product,build/target/product/languages_full.mk)

$(call inherit-product,vendor/opengapps/build/opengapps-packages.mk)

# includes the base of Android-x86 platform
$(call inherit-product,device/generic/common/x86_64.mk)

# Overrides
PRODUCT_NAME := android_x86_64
PRODUCT_BRAND := Android-x86
PRODUCT_DEVICE := x86_64
PRODUCT_MODEL := Generic Android-x86_64


(Not fatal && Solve myself) 3. Cannot adjust spdif optical out's volume. I
think it is normal and can adjust on RC1 is abnormal. I will attach code
what I have modify.
diff --git a/audio/AudioPolicyManagerBase.cpp
b/audio/AudioPolicyManagerBase.cpp
index 74ee22a..e3d16da 100644
--- a/audio/AudioPolicyManagerBase.cpp
+++ b/audio/AudioPolicyManagerBase.cpp
@@ -3248,7 +3248,7 @@ float AudioPolicyManagerBase::computeVolume(int
stream,
index != mStreams[stream].mIndexMin &&
(device == AUDIO_DEVICE_OUT_AUX_DIGITAL ||
device == AUDIO_DEVICE_OUT_DGTL_DOCK_HEADSET)) {
- return 1.0;
+ // return 1.0;
}


diff --git a/services/core/java/com/android/server/audio/AudioService.java
b/services/core/java/com/android/server/audio/AudioService.java
index c70a87c..87cabe0 100644
--- a/services/core/java/com/android/server/audio/AudioService.java
+++ b/services/core/java/com/android/server/audio/AudioService.java
@@ -520,12 +520,13 @@ public class AudioService extends IAudioService.Stub {
= new RemoteCallbackList<IAudioRoutesObserver>();

// Devices for which the volume is fixed and VolumePanel slider should
be disabled
- int mFixedVolumeDevices = AudioSystem.DEVICE_OUT_HDMI |
- AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_HDMI_ARC |
- AudioSystem.DEVICE_OUT_SPDIF |
- AudioSystem.DEVICE_OUT_AUX_LINE;
+ int mFixedVolumeDevices = 0;
+ //AudioSystem.DEVICE_OUT_HDMI |
+ //AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_HDMI_ARC |
+ //AudioSystem.DEVICE_OUT_SPDIF |
+ //AudioSystem.DEVICE_OUT_AUX_LINE;
int mFullVolumeDevices = 0;

I think this modification is not good and it is not intent of designer's.
Any idea's without modification of this code and how to adjust volume
possible?

FATAL EXCEPTION: GLThread 178 issue I have
reported(https://groups.google.com/forum/#!topic/android-x86/E8iRTfYNJ90)
is fixed. It looks like it is related to libdrm or mesa.
Post by Mauro Rossi
This pattern of mesa 17.2.x + nougat-x86 libdrm working suggests
that the problem with Coffelake (gen 8) was most probably due to mesa
libraries moved to /vendor in mesa 17.3 and later
Also when I build nougat, with mesa 17.3 and later, I use x86/oreo-x86
version of external/drm_gralloc
In order to confirm I'm uploading experimental test images with mesa
https://drive.google.com/drive/folders/0B_OFHiIqgpSFMlk4bkhaMnAwbDA
Mauro
Post by YoungJoon Lee
Thank you for your answer. I finally success working nougat on Coffee
lake system. I left How can I do.
1. HW : https://www.msi.com/Motherboard/Z370I-GAMING-PRO-CARBON-AC.html
2. Base : e8653d3
<https://osdn.net/projects/android-x86/scm/git/manifest/commits/e8653d3a463c6894daff37073b35b2f880e59a0e>
3. /Kernel : https://github.com/maurossi/linux/tree/kernel-4.15rc8
https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86
<https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86>
https://github.com/maurossi/llvm/tree/nougat-x86_exports
6. /external/libdrm : Must use Default one.
7. code modification : /device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
What I have tested(Working)
- Realtek ALC1220 Codec
/device/generic/common/alsa/PCH.state needed.(
https://github.com/getfeus/android-oreo-x86_android-auto/blob/master/device/generic/common/alsa/PCH.state
)
- Intel® Dual Band Wireless-AC 8265 card
- NVMe
- Intel I219-V Gigabit LAN controller
- enable_nativebridge : Only check command success. Not testing
Glthread fatal error I have reported.(
https://groups.google.com/forum/#!msg/android-x86/E8iRTfYNJ90/wbnaYCzwBwAJ;context-place=msg/android-x86/ilk69kFFcWY/WlPFjIhSAAAJ
)
What is not stable.
- During boot, ACPI error is displayed. Some kernel module crashed
and re-loaded. Every module is ACPI related of thermal sensor
related.(Mis-configured of defconfig?)
- I can't dump message from dmesg.(Before logging daemon start?) But,
Start with this string, ACPI Error: [_OSI] Namespace lookup failure,
AE_NOT_FOUND on early boot process(Prior to android device dev partition
detect). Very close to this post.
https://bugzilla.redhat.com/show_bug.cgi?id=1413342
- This message is disappeared on old kernel like 4.9.x but Coffee
lake GPU can only working on 4.13.x~. All Kernels I have tested have same
errors. But not care.(Not displayed after boot)
- I can't boot from any Oreo build.
Best regards.
Post by spin877
Post by Mauro Rossi
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
/device/generic/common/build/tasks/kernel.mk
mk_kernel := + $(hide) $(MAKE) -jX -C
X=cpu
No need to add it if you have added
-jX to make iso_img.
Submake is parallel execution automatically,
29417 pts/42 Sl+ 0:00 | \_ /mnt/cwhuang/oreo-x86/out/soong_ui
--make-mode -j24 iso_img
30617 pts/42 S+ 0:09 | \_
prebuilts/build-tools/linux-x86/bin/ninja -d keepdepfile iso_img -j 24
-f out/combined-android_x86_64.ninja -w dupbuild=err
30944 pts/42 S 0:00 | \_ /bin/sh -c /bin/bash -c
"(make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=\"
30945 pts/42 S 0:00 | \_ /bin/bash -c (make -j24
-C kernel O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=" /mnt/
30946 pts/42 S 0:00 | \_ make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE= /mnt//cwhuang
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-21 16:30:10 UTC
Permalink
I forgot to attach logcat file. Sorry.
Post by YoungJoon Lee
Thank you Mauro Rossi.
I will test your test image and report here.
Today I found few problems.
(Fatal) 1 : System is extremely heavy. All response is very slow(Slower
than skylake). I monitor details of top command. But it looks like there is
no process consume CPU 100%. I guess it is DVFS(Dynamic Frequency)
Issue.(Kernel message said ACPI related driver is unstable and thermal
driver also print error messages.) How can I debug performance issue? For
example, is there any way to disable power management or frequency
control(Working on maximum clock speed)? I remember that kerner setting on
defconfig's power profile is on demand.
(Not Fatal) 2 : App crashes when keyboard input pop up.(On some apps
only). I will attach logcat and left code here related to this error. It
looks like it is webview related issue. So, I will attach my device mk file
which is import opengapps.
01-21 13:13:56.677 16130 16118 F google-breakpad: -----END BREAKPAD MICRODUMP-----
01-21 13:13:56.800 15305 16118 F libc : Fatal signal 11 (SIGSEGV), code -6, fault addr 0x3bc9 in tid 16118 (Chrome_InProcGp)
01-21 13:13:56.899 16133 16133 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
01-21 13:13:56.899 16133 16133 F DEBUG : Build fingerprint: 'Android-x86/android_x86_64/x86_64:7.1.2/N2G48H/root01210301:userdebug/test-keys'
01-21 13:13:56.899 16133 16133 F DEBUG : Revision: '0'
01-21 13:13:56.899 16133 16133 F DEBUG : ABI: 'x86'
01-21 13:13:56.899 16133 16133 F DEBUG : pid: 15305, tid: 16118, name: Chrome_InProcGp >>> jp.co.yahoo.android.apps.navi <<<
01-21 13:13:56.899 16133 16133 F DEBUG : signal 11 (SIGSEGV), code -6 (SI_TKILL), fault addr 0x3bc9
01-21 13:13:56.899 16133 16133 F DEBUG : eax 00000000 ebx cf899d40 ecx 00008000 edx c0dbec58
01-21 13:13:56.899 16133 16133 F DEBUG : esi 00000000 edi 00000000
01-21 13:13:56.899 16133 16133 F DEBUG : xcs 00000023 xds 0000002b xes 0000002b xfs 0000006b xss 0000002b
01-21 13:13:56.899 16133 16133 F DEBUG : eip cf66012c ebp c0dbec78 esp c0dbe680 flags 00210282
01-21 13:13:56.899 16133 16133 F DEBUG : #00 pc 005d812c /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #01 pc 000267b2 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #02 pc 00029bee /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #03 pc 0035e248 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #04 pc 0000ab1e /system/lib/libglapi.so
01-21 13:13:56.899 16133 16133 F DEBUG : #05 pc 00a12d91 /system/app/webview/webview.apk (offset 0x927000)
Here is mk file.
GAPPS_VARIANT := nano
PRODUCT_PACKAGES += Chrome \
Maps \
KeyboardGoogle \
GoogleTTS \
YouTube \
PixelIcons \
PixelLauncher \
Wallpapers \
PixelLauncherIcons \
GoogleServicesFramework \
GoogleLoginService
#PRODUCT_PACKAGES += Hangouts
GAPPS_EXCLUDED_PACKAGES := PrebuiltGmsCoreInstantApps \
FaceLock \
AndroidPlatformServices
GAPPS_FORCE_BROWSER_OVERRIDES := true
PRODUCT_PROPERTY_OVERRIDES := \
ro.kernel.android.gps=ttyUSB0 \
ro.kernel.android.gpsttybaud=4800
$(call inherit-product,build/target/product/languages_full.mk)
$(call inherit-product,vendor/opengapps/build/opengapps-packages.mk)
# includes the base of Android-x86 platform
$(call inherit-product,device/generic/common/x86_64.mk)
# Overrides
PRODUCT_NAME := android_x86_64
PRODUCT_BRAND := Android-x86
PRODUCT_DEVICE := x86_64
PRODUCT_MODEL := Generic Android-x86_64
(Not fatal && Solve myself) 3. Cannot adjust spdif optical out's volume. I
think it is normal and can adjust on RC1 is abnormal. I will attach code
what I have modify.
diff --git a/audio/AudioPolicyManagerBase.cpp
b/audio/AudioPolicyManagerBase.cpp
index 74ee22a..e3d16da 100644
--- a/audio/AudioPolicyManagerBase.cpp
+++ b/audio/AudioPolicyManagerBase.cpp
@@ -3248,7 +3248,7 @@ float AudioPolicyManagerBase::computeVolume(int
stream,
index != mStreams[stream].mIndexMin &&
(device == AUDIO_DEVICE_OUT_AUX_DIGITAL ||
device == AUDIO_DEVICE_OUT_DGTL_DOCK_HEADSET)) {
- return 1.0;
+ // return 1.0;
}
diff --git a/services/core/java/com/android/server/audio/AudioService.java
b/services/core/java/com/android/server/audio/AudioService.java
index c70a87c..87cabe0 100644
--- a/services/core/java/com/android/server/audio/AudioService.java
+++ b/services/core/java/com/android/server/audio/AudioService.java
@@ -520,12 +520,13 @@ public class AudioService extends IAudioService.Stub {
= new RemoteCallbackList<IAudioRoutesObserver>();
// Devices for which the volume is fixed and VolumePanel slider
should be disabled
- int mFixedVolumeDevices = AudioSystem.DEVICE_OUT_HDMI |
- AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_HDMI_ARC |
- AudioSystem.DEVICE_OUT_SPDIF |
- AudioSystem.DEVICE_OUT_AUX_LINE;
+ int mFixedVolumeDevices = 0;
+ //AudioSystem.DEVICE_OUT_HDMI |
+ //AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_HDMI_ARC |
+ //AudioSystem.DEVICE_OUT_SPDIF |
+ //AudioSystem.DEVICE_OUT_AUX_LINE;
int mFullVolumeDevices = 0;
I think this modification is not good and it is not intent of designer's.
Any idea's without modification of this code and how to adjust volume
possible?
FATAL EXCEPTION: GLThread 178 issue I have reported(
https://groups.google.com/forum/#!topic/android-x86/E8iRTfYNJ90) is
fixed. It looks like it is related to libdrm or mesa.
Post by Mauro Rossi
This pattern of mesa 17.2.x + nougat-x86 libdrm working suggests
that the problem with Coffelake (gen 8) was most probably due to mesa
libraries moved to /vendor in mesa 17.3 and later
Also when I build nougat, with mesa 17.3 and later, I use x86/oreo-x86
version of external/drm_gralloc
In order to confirm I'm uploading experimental test images with mesa
https://drive.google.com/drive/folders/0B_OFHiIqgpSFMlk4bkhaMnAwbDA
Mauro
Post by YoungJoon Lee
Thank you for your answer. I finally success working nougat on Coffee
lake system. I left How can I do.
1. HW : https://www.msi.com/Motherboard/Z370I-GAMING-PRO-CARBON-AC.html
2. Base : e8653d3
<https://osdn.net/projects/android-x86/scm/git/manifest/commits/e8653d3a463c6894daff37073b35b2f880e59a0e>
3. /Kernel : https://github.com/maurossi/linux/tree/kernel-4.15rc8
https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86
<https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86>
https://github.com/maurossi/llvm/tree/nougat-x86_exports
6. /external/libdrm : Must use Default one.
7. code modification : /device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
What I have tested(Working)
- Realtek ALC1220 Codec
/device/generic/common/alsa/PCH.state needed.(
https://github.com/getfeus/android-oreo-x86_android-auto/blob/master/device/generic/common/alsa/PCH.state
)
- Intel® Dual Band Wireless-AC 8265 card
- NVMe
- Intel I219-V Gigabit LAN controller
- enable_nativebridge : Only check command success. Not testing
Glthread fatal error I have reported.(
https://groups.google.com/forum/#!msg/android-x86/E8iRTfYNJ90/wbnaYCzwBwAJ;context-place=msg/android-x86/ilk69kFFcWY/WlPFjIhSAAAJ
)
What is not stable.
- During boot, ACPI error is displayed. Some kernel module crashed
and re-loaded. Every module is ACPI related of thermal sensor
related.(Mis-configured of defconfig?)
- I can't dump message from dmesg.(Before logging daemon start?)
But, Start with this string, ACPI Error: [_OSI] Namespace lookup failure,
AE_NOT_FOUND on early boot process(Prior to android device dev partition
detect). Very close to this post.
https://bugzilla.redhat.com/show_bug.cgi?id=1413342
- This message is disappeared on old kernel like 4.9.x but Coffee
lake GPU can only working on 4.13.x~. All Kernels I have tested have same
errors. But not care.(Not displayed after boot)
- I can't boot from any Oreo build.
Best regards.
Post by spin877
Post by Mauro Rossi
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
/device/generic/common/build/tasks/kernel.mk
mk_kernel := + $(hide) $(MAKE) -jX -C
X=cpu
No need to add it if you have added
-jX to make iso_img.
Submake is parallel execution automatically,
29417 pts/42 Sl+ 0:00 | \_ /mnt/cwhuang/oreo-x86/out/soong_ui
--make-mode -j24 iso_img
30617 pts/42 S+ 0:09 | \_
prebuilts/build-tools/linux-x86/bin/ninja -d keepdepfile iso_img -j 24
-f out/combined-android_x86_64.ninja -w dupbuild=err
30944 pts/42 S 0:00 | \_ /bin/sh -c /bin/bash -c
"(make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=\"
30945 pts/42 S 0:00 | \_ /bin/bash -c (make -j24
-C kernel O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=" /mnt/
30946 pts/42 S 0:00 | \_ make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE= /mnt//cwhuang
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
Mauro Rossi
2018-01-21 16:54:13 UTC
Permalink
Post by YoungJoon Lee
Thank you Mauro Rossi.
I will test your test image and report here.
Today I found few problems.
(Fatal) 1 : System is extremely heavy. All response is very slow(Slower
than skylake). I monitor details of top command. But it looks like there is
no process consume CPU 100%. I guess it is DVFS(Dynamic Frequency)
Issue.(Kernel message said ACPI related driver is unstable and thermal
driver also print error messages.) How can I debug performance issue? For
example, is there any way to disable power management or frequency
control(Working on maximum clock speed)? I remember that kerner setting on
defconfig's power profile is on demand.
(Not Fatal) 2 : App crashes when keyboard input pop up.(On some apps
only). I will attach logcat and left code here related to this error. It
looks like it is webview related issue. So, I will attach my device mk file
which is import opengapps.
01-21 13:13:56.677 16130 16118 F google-breakpad: -----END BREAKPAD MICRODUMP-----
01-21 13:13:56.800 15305 16118 F libc : Fatal signal 11 (SIGSEGV), code -6, fault addr 0x3bc9 in tid 16118 (Chrome_InProcGp)
01-21 13:13:56.899 16133 16133 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
01-21 13:13:56.899 16133 16133 F DEBUG : Build fingerprint: 'Android-x86/android_x86_64/x86_64:7.1.2/N2G48H/root01210301:userdebug/test-keys'
01-21 13:13:56.899 16133 16133 F DEBUG : Revision: '0'
01-21 13:13:56.899 16133 16133 F DEBUG : ABI: 'x86'
01-21 13:13:56.899 16133 16133 F DEBUG : pid: 15305, tid: 16118, name: Chrome_InProcGp >>> jp.co.yahoo.android.apps.navi <<<
01-21 13:13:56.899 16133 16133 F DEBUG : signal 11 (SIGSEGV), code -6 (SI_TKILL), fault addr 0x3bc9
01-21 13:13:56.899 16133 16133 F DEBUG : eax 00000000 ebx cf899d40 ecx 00008000 edx c0dbec58
01-21 13:13:56.899 16133 16133 F DEBUG : esi 00000000 edi 00000000
01-21 13:13:56.899 16133 16133 F DEBUG : xcs 00000023 xds 0000002b xes 0000002b xfs 0000006b xss 0000002b
01-21 13:13:56.899 16133 16133 F DEBUG : eip cf66012c ebp c0dbec78 esp c0dbe680 flags 00210282
01-21 13:13:56.899 16133 16133 F DEBUG : #00 pc 005d812c /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #01 pc 000267b2 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #02 pc 00029bee /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #03 pc 0035e248 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #04 pc 0000ab1e /system/lib/libglapi.so
Here is important to find the source file/line of code causing the
segfault, by providing [???] (output of addr2line)

on source folder:

addr2line -Cfe out/product/target/x86_64/symbols/system/lib/dri/i965_dri.so
005d812c
[???]

000267b2
[???]

00029bee
[???]

0035e248

[???]


^C



addr2line -Cfe out/product/target/x86_64/symbols/system/lib/libglapi.so

0000ab1e

[???]



Same procedure required for the other SIGSEGV


01-21 12:47:21.791 3767 3959 F libc : Fatal signal 11 (SIGSEGV), code 1, fault addr 0x20c in tid 3959 (Binder:3767_4)
01-21 12:47:21.899 4090 4090 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
01-21 12:47:21.899 4090 4090 F DEBUG : Build fingerprint: 'Android-x86/android_x86_64/x86_64:7.1.2/N2G48H/root01210301:userdebug/test-keys'
01-21 12:47:21.899 4090 4090 F DEBUG : Revision: '0'
01-21 12:47:21.899 4090 4090 F DEBUG : ABI: 'x86_64'
01-21 12:47:21.899 4090 4090 F DEBUG : pid: 3767, tid: 3959, name: Binder:3767_4 >>> com.google.android.gms.unstable <<<
01-21 12:47:21.899 4090 4090 F DEBUG : signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x20c
01-21 12:47:21.899 4090 4090 F DEBUG : rax 0000000000000000 rbx 0000000000000000 rcx 00007c395843de0c rdx 00007c39591d5d94
01-21 12:47:21.899 4090 4090 F DEBUG : rsi 0000000000000000 rdi 0000000000000000
01-21 12:47:21.899 4090 4090 F DEBUG : r8 0000000000000003 r9 00007c39591d7450 r10 0000000000000000 r11 0000000000000005
01-21 12:47:21.899 4090 4090 F DEBUG : r12 00007c395846bc00 r13 00007c3972dd1f00 r14 0000000000000020 r15 0000000000000001
01-21 12:47:21.899 4090 4090 F DEBUG : cs 0000000000000033 ss 000000000000002b
01-21 12:47:21.899 4090 4090 F DEBUG : rip 00007c39587eb741 rbp 0000000000000000 rsp 00007c39591d5ca0 eflags 0000000000010246
01-21 12:47:23.574 4090 4090 F DEBUG :
01-21 12:47:23.574 4090 4090 F DEBUG : backtrace:
01-21 12:47:23.574 4090 4090 F DEBUG : #00 pc 0000000000064741 /system/lib64/dri/i965_dri.so
01-21 12:47:23.574 4090 4090 F DEBUG : #01 pc 000000000006769a /system/lib64/dri/i965_dri.so


on source folder:

addr2line -Cfe out/product/target/x86_64/symbols/system/lib64/dri/i965_dri.so

0000000000064741

[???]

000000000006769a

[???]
Post by YoungJoon Lee
01-21 13:13:56.899 16133 16133 F DEBUG : #05 pc 00a12d91 /system/app/webview/webview.apk (offset 0x927000)
Here is mk file.
GAPPS_VARIANT := nano
PRODUCT_PACKAGES += Chrome \
Maps \
KeyboardGoogle \
GoogleTTS \
YouTube \
PixelIcons \
PixelLauncher \
Wallpapers \
PixelLauncherIcons \
GoogleServicesFramework \
GoogleLoginService
#PRODUCT_PACKAGES += Hangouts
GAPPS_EXCLUDED_PACKAGES := PrebuiltGmsCoreInstantApps \
FaceLock \
AndroidPlatformServices
GAPPS_FORCE_BROWSER_OVERRIDES := true
PRODUCT_PROPERTY_OVERRIDES := \
ro.kernel.android.gps=ttyUSB0 \
ro.kernel.android.gpsttybaud=4800
$(call inherit-product,build/target/product/languages_full.mk)
$(call inherit-product,vendor/opengapps/build/opengapps-packages.mk)
# includes the base of Android-x86 platform
$(call inherit-product,device/generic/common/x86_64.mk)
# Overrides
PRODUCT_NAME := android_x86_64
PRODUCT_BRAND := Android-x86
PRODUCT_DEVICE := x86_64
PRODUCT_MODEL := Generic Android-x86_64
(Not fatal && Solve myself) 3. Cannot adjust spdif optical out's volume. I
think it is normal and can adjust on RC1 is abnormal. I will attach code
what I have modify.
diff --git a/audio/AudioPolicyManagerBase.cpp
b/audio/AudioPolicyManagerBase.cpp
index 74ee22a..e3d16da 100644
--- a/audio/AudioPolicyManagerBase.cpp
+++ b/audio/AudioPolicyManagerBase.cpp
@@ -3248,7 +3248,7 @@ float AudioPolicyManagerBase::computeVolume(int
stream,
index != mStreams[stream].mIndexMin &&
(device == AUDIO_DEVICE_OUT_AUX_DIGITAL ||
device == AUDIO_DEVICE_OUT_DGTL_DOCK_HEADSET)) {
- return 1.0;
+ // return 1.0;
}
diff --git a/services/core/java/com/android/server/audio/AudioService.java
b/services/core/java/com/android/server/audio/AudioService.java
index c70a87c..87cabe0 100644
--- a/services/core/java/com/android/server/audio/AudioService.java
+++ b/services/core/java/com/android/server/audio/AudioService.java
@@ -520,12 +520,13 @@ public class AudioService extends IAudioService.Stub {
= new RemoteCallbackList<IAudioRoutesObserver>();
// Devices for which the volume is fixed and VolumePanel slider
should be disabled
- int mFixedVolumeDevices = AudioSystem.DEVICE_OUT_HDMI |
- AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_HDMI_ARC |
- AudioSystem.DEVICE_OUT_SPDIF |
- AudioSystem.DEVICE_OUT_AUX_LINE;
+ int mFixedVolumeDevices = 0;
+ //AudioSystem.DEVICE_OUT_HDMI |
+ //AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_HDMI_ARC |
+ //AudioSystem.DEVICE_OUT_SPDIF |
+ //AudioSystem.DEVICE_OUT_AUX_LINE;
int mFullVolumeDevices = 0;
I think this modification is not good and it is not intent of designer's.
Any idea's without modification of this code and how to adjust volume
possible?
FATAL EXCEPTION: GLThread 178 issue I have reported(
https://groups.google.com/forum/#!topic/android-x86/E8iRTfYNJ90) is
fixed. It looks like it is related to libdrm or mesa.
Post by Mauro Rossi
This pattern of mesa 17.2.x + nougat-x86 libdrm working suggests
that the problem with Coffelake (gen 8) was most probably due to mesa
libraries moved to /vendor in mesa 17.3 and later
Also when I build nougat, with mesa 17.3 and later, I use x86/oreo-x86
version of external/drm_gralloc
In order to confirm I'm uploading experimental test images with mesa
https://drive.google.com/drive/folders/0B_OFHiIqgpSFMlk4bkhaMnAwbDA
Mauro
Post by YoungJoon Lee
Thank you for your answer. I finally success working nougat on Coffee
lake system. I left How can I do.
1. HW : https://www.msi.com/Motherboard/Z370I-GAMING-PRO-CARBON-AC.html
2. Base : e8653d3
<https://osdn.net/projects/android-x86/scm/git/manifest/commits/e8653d3a463c6894daff37073b35b2f880e59a0e>
3. /Kernel : https://github.com/maurossi/linux/tree/kernel-4.15rc8
https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86
<https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86>
https://github.com/maurossi/llvm/tree/nougat-x86_exports
6. /external/libdrm : Must use Default one.
7. code modification : /device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
What I have tested(Working)
- Realtek ALC1220 Codec
/device/generic/common/alsa/PCH.state needed.(
https://github.com/getfeus/android-oreo-x86_android-auto/blob/master/device/generic/common/alsa/PCH.state
)
- Intel® Dual Band Wireless-AC 8265 card
- NVMe
- Intel I219-V Gigabit LAN controller
- enable_nativebridge : Only check command success. Not testing
Glthread fatal error I have reported.(
https://groups.google.com/forum/#!msg/android-x86/E8iRTfYNJ90/wbnaYCzwBwAJ;context-place=msg/android-x86/ilk69kFFcWY/WlPFjIhSAAAJ
)
What is not stable.
- During boot, ACPI error is displayed. Some kernel module crashed
and re-loaded. Every module is ACPI related of thermal sensor
related.(Mis-configured of defconfig?)
- I can't dump message from dmesg.(Before logging daemon start?)
But, Start with this string, ACPI Error: [_OSI] Namespace lookup failure,
AE_NOT_FOUND on early boot process(Prior to android device dev partition
detect). Very close to this post.
https://bugzilla.redhat.com/show_bug.cgi?id=1413342
- This message is disappeared on old kernel like 4.9.x but Coffee
lake GPU can only working on 4.13.x~. All Kernels I have tested have same
errors. But not care.(Not displayed after boot)
- I can't boot from any Oreo build.
Best regards.
Post by spin877
Post by Mauro Rossi
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
/device/generic/common/build/tasks/kernel.mk
mk_kernel := + $(hide) $(MAKE) -jX -C
X=cpu
No need to add it if you have added
-jX to make iso_img.
Submake is parallel execution automatically,
29417 pts/42 Sl+ 0:00 | \_ /mnt/cwhuang/oreo-x86/out/soong_ui
--make-mode -j24 iso_img
30617 pts/42 S+ 0:09 | \_
prebuilts/build-tools/linux-x86/bin/ninja -d keepdepfile iso_img -j 24
-f out/combined-android_x86_64.ninja -w dupbuild=err
30944 pts/42 S 0:00 | \_ /bin/sh -c /bin/bash -c
"(make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=\"
30945 pts/42 S 0:00 | \_ /bin/bash -c (make -j24
-C kernel O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=" /mnt/
30946 pts/42 S 0:00 | \_ make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE= /mnt//cwhuang
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-22 01:32:58 UTC
Permalink
Ok, I read log file after reading your comment. It seems like i965(mesa?)
issue. Not opengapps issue.

First, I will check i965_dri.so. I don't know it is right.
***@UbuntuPV:/mnt/nfs/data/src/android-x86-nougat-x86-rc2# addr2line -Cfe
./out/target/product/x86_64/symbols/system/lib/dri/i965_dri.so
005d812c
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
get_fast_clear_rect
/proc/self/cwd/external/mesa/src/intel/blorp/blorp_clear.c:196
000267b2
do_single_blorp_clear
/proc/self/cwd/external/mesa/src/mesa/drivers/dri/i965/brw_blorp.c:859
00029bee
brw_clear
/proc/self/cwd/external/mesa/src/mesa/drivers/dri/i965/brw_clear.c:233
0035e248
clear
/proc/self/cwd/external/mesa/src/mesa/main/clear.c:221

Next, libglapi.so
***@UbuntuPV:/mnt/nfs/data/src/android-x86-nougat-x86-rc2# addr2line -Cfe
./out/target/product/x86_64/symbols/system/lib/libglapi.so
0000ab1e
shared_dispatch_stub_203
/proc/self/cwd/out/target/product/x86_64/obj_x86/SHARED_LIBRARIES/libglapi_intermediates/shared-glapi/glapi_mapi_tmp.h:18408

Next, i965_dri.so
***@UbuntuPV:/mnt/nfs/data/src/android-x86-nougat-x86-rc2# addr2line -Cfe
./out/target/product/x86_64/symbols/system/lib64/dri/i965_dri.so
0000000000064741
do_single_blorp_clear
/proc/self/cwd/external/mesa/src/mesa/drivers/dri/i965/brw_blorp.c:789
000000000006769a
brw_clear
/proc/self/cwd/external/mesa/src/mesa/drivers/dri/i965/brw_clear.c:233
I will check each lines of code.
Post by Mauro Rossi
Post by YoungJoon Lee
Thank you Mauro Rossi.
I will test your test image and report here.
Today I found few problems.
(Fatal) 1 : System is extremely heavy. All response is very slow(Slower
than skylake). I monitor details of top command. But it looks like there is
no process consume CPU 100%. I guess it is DVFS(Dynamic Frequency)
Issue.(Kernel message said ACPI related driver is unstable and thermal
driver also print error messages.) How can I debug performance issue? For
example, is there any way to disable power management or frequency
control(Working on maximum clock speed)? I remember that kerner setting on
defconfig's power profile is on demand.
(Not Fatal) 2 : App crashes when keyboard input pop up.(On some apps
only). I will attach logcat and left code here related to this error. It
looks like it is webview related issue. So, I will attach my device mk file
which is import opengapps.
01-21 13:13:56.677 16130 16118 F google-breakpad: -----END BREAKPAD MICRODUMP-----
01-21 13:13:56.800 15305 16118 F libc : Fatal signal 11 (SIGSEGV), code -6, fault addr 0x3bc9 in tid 16118 (Chrome_InProcGp)
01-21 13:13:56.899 16133 16133 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
01-21 13:13:56.899 16133 16133 F DEBUG : Build fingerprint: 'Android-x86/android_x86_64/x86_64:7.1.2/N2G48H/root01210301:userdebug/test-keys'
01-21 13:13:56.899 16133 16133 F DEBUG : Revision: '0'
01-21 13:13:56.899 16133 16133 F DEBUG : ABI: 'x86'
01-21 13:13:56.899 16133 16133 F DEBUG : pid: 15305, tid: 16118, name: Chrome_InProcGp >>> jp.co.yahoo.android.apps.navi <<<
01-21 13:13:56.899 16133 16133 F DEBUG : signal 11 (SIGSEGV), code -6 (SI_TKILL), fault addr 0x3bc9
01-21 13:13:56.899 16133 16133 F DEBUG : eax 00000000 ebx cf899d40 ecx 00008000 edx c0dbec58
01-21 13:13:56.899 16133 16133 F DEBUG : esi 00000000 edi 00000000
01-21 13:13:56.899 16133 16133 F DEBUG : xcs 00000023 xds 0000002b xes 0000002b xfs 0000006b xss 0000002b
01-21 13:13:56.899 16133 16133 F DEBUG : eip cf66012c ebp c0dbec78 esp c0dbe680 flags 00210282
01-21 13:13:56.899 16133 16133 F DEBUG : #00 pc 005d812c /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #01 pc 000267b2 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #02 pc 00029bee /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #03 pc 0035e248 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #04 pc 0000ab1e /system/lib/libglapi.so
Here is important to find the source file/line of code causing the
segfault, by providing [???] (output of addr2line)
addr2line -Cfe out/product/target/x86_64/symbols/
system/lib/dri/i965_dri.so
005d812c
[???]
000267b2
[???]
00029bee
[???]
0035e248
[???]
^C
addr2line -Cfe out/product/target/x86_64/symbols/system/lib/libglapi.so
0000ab1e
[???]
Same procedure required for the other SIGSEGV
01-21 12:47:21.791 3767 3959 F libc : Fatal signal 11 (SIGSEGV), code 1, fault addr 0x20c in tid 3959 (Binder:3767_4)
01-21 12:47:21.899 4090 4090 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
01-21 12:47:21.899 4090 4090 F DEBUG : Build fingerprint: 'Android-x86/android_x86_64/x86_64:7.1.2/N2G48H/root01210301:userdebug/test-keys'
01-21 12:47:21.899 4090 4090 F DEBUG : Revision: '0'
01-21 12:47:21.899 4090 4090 F DEBUG : ABI: 'x86_64'
01-21 12:47:21.899 4090 4090 F DEBUG : pid: 3767, tid: 3959, name: Binder:3767_4 >>> com.google.android.gms.unstable <<<
01-21 12:47:21.899 4090 4090 F DEBUG : signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x20c
01-21 12:47:21.899 4090 4090 F DEBUG : rax 0000000000000000 rbx 0000000000000000 rcx 00007c395843de0c rdx 00007c39591d5d94
01-21 12:47:21.899 4090 4090 F DEBUG : rsi 0000000000000000 rdi 0000000000000000
01-21 12:47:21.899 4090 4090 F DEBUG : r8 0000000000000003 r9 00007c39591d7450 r10 0000000000000000 r11 0000000000000005
01-21 12:47:21.899 4090 4090 F DEBUG : r12 00007c395846bc00 r13 00007c3972dd1f00 r14 0000000000000020 r15 0000000000000001
01-21 12:47:21.899 4090 4090 F DEBUG : cs 0000000000000033 ss 000000000000002b
01-21 12:47:21.899 4090 4090 F DEBUG : rip 00007c39587eb741 rbp 0000000000000000 rsp 00007c39591d5ca0 eflags 0000000000010246
01-21 12:47:23.574 4090 4090 F DEBUG : #00 pc 0000000000064741 /system/lib64/dri/i965_dri.so
01-21 12:47:23.574 4090 4090 F DEBUG : #01 pc 000000000006769a /system/lib64/dri/i965_dri.so
addr2line -Cfe out/product/target/x86_64/symbols/system/lib64/dri/i965_dri.so
0000000000064741
[???]
000000000006769a
[???]
Post by YoungJoon Lee
01-21 13:13:56.899 16133 16133 F DEBUG : #05 pc 00a12d91 /system/app/webview/webview.apk (offset 0x927000)
Here is mk file.
GAPPS_VARIANT := nano
PRODUCT_PACKAGES += Chrome \
Maps \
KeyboardGoogle \
GoogleTTS \
YouTube \
PixelIcons \
PixelLauncher \
Wallpapers \
PixelLauncherIcons \
GoogleServicesFramework \
GoogleLoginService
#PRODUCT_PACKAGES += Hangouts
GAPPS_EXCLUDED_PACKAGES := PrebuiltGmsCoreInstantApps \
FaceLock \
AndroidPlatformServices
GAPPS_FORCE_BROWSER_OVERRIDES := true
PRODUCT_PROPERTY_OVERRIDES := \
ro.kernel.android.gps=ttyUSB0 \
ro.kernel.android.gpsttybaud=4800
$(call inherit-product,build/target/product/languages_full.mk)
$(call inherit-product,vendor/opengapps/build/opengapps-packages.mk)
# includes the base of Android-x86 platform
$(call inherit-product,device/generic/common/x86_64.mk)
# Overrides
PRODUCT_NAME := android_x86_64
PRODUCT_BRAND := Android-x86
PRODUCT_DEVICE := x86_64
PRODUCT_MODEL := Generic Android-x86_64
(Not fatal && Solve myself) 3. Cannot adjust spdif optical out's volume.
I think it is normal and can adjust on RC1 is abnormal. I will attach code
what I have modify.
diff --git a/audio/AudioPolicyManagerBase.cpp
b/audio/AudioPolicyManagerBase.cpp
index 74ee22a..e3d16da 100644
--- a/audio/AudioPolicyManagerBase.cpp
+++ b/audio/AudioPolicyManagerBase.cpp
@@ -3248,7 +3248,7 @@ float AudioPolicyManagerBase::computeVolume(int
stream,
index != mStreams[stream].mIndexMin &&
(device == AUDIO_DEVICE_OUT_AUX_DIGITAL ||
device == AUDIO_DEVICE_OUT_DGTL_DOCK_HEADSET)) {
- return 1.0;
+ // return 1.0;
}
diff --git
a/services/core/java/com/android/server/audio/AudioService.java
b/services/core/java/com/android/server/audio/AudioService.java
index c70a87c..87cabe0 100644
--- a/services/core/java/com/android/server/audio/AudioService.java
+++ b/services/core/java/com/android/server/audio/AudioService.java
@@ -520,12 +520,13 @@ public class AudioService extends
IAudioService.Stub {
= new RemoteCallbackList<IAudioRoutesObserver>();
// Devices for which the volume is fixed and VolumePanel slider
should be disabled
- int mFixedVolumeDevices = AudioSystem.DEVICE_OUT_HDMI |
- AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_HDMI_ARC |
- AudioSystem.DEVICE_OUT_SPDIF |
- AudioSystem.DEVICE_OUT_AUX_LINE;
+ int mFixedVolumeDevices = 0;
+ //AudioSystem.DEVICE_OUT_HDMI |
+ //AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_HDMI_ARC |
+ //AudioSystem.DEVICE_OUT_SPDIF |
+ //AudioSystem.DEVICE_OUT_AUX_LINE;
int mFullVolumeDevices = 0;
I think this modification is not good and it is not intent of designer's.
Any idea's without modification of this code and how to adjust volume
possible?
FATAL EXCEPTION: GLThread 178 issue I have reported(
https://groups.google.com/forum/#!topic/android-x86/E8iRTfYNJ90) is
fixed. It looks like it is related to libdrm or mesa.
Post by Mauro Rossi
This pattern of mesa 17.2.x + nougat-x86 libdrm working suggests
that the problem with Coffelake (gen 8) was most probably due to mesa
libraries moved to /vendor in mesa 17.3 and later
Also when I build nougat, with mesa 17.3 and later, I use x86/oreo-x86
version of external/drm_gralloc
In order to confirm I'm uploading experimental test images with mesa
https://drive.google.com/drive/folders/0B_OFHiIqgpSFMlk4bkhaMnAwbDA
Mauro
Post by YoungJoon Lee
Thank you for your answer. I finally success working nougat on Coffee
lake system. I left How can I do.
1. HW : https://www.msi.com/Motherboard/Z370I-GAMING-PRO-CARBON-AC.html
2. Base : e8653d3
<https://osdn.net/projects/android-x86/scm/git/manifest/commits/e8653d3a463c6894daff37073b35b2f880e59a0e>
3. /Kernel : https://github.com/maurossi/linux/tree/kernel-4.15rc8
https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86
<https://github.com/maurossi/mesa/tree/17.2.0rc4_nougat-x86>
https://github.com/maurossi/llvm/tree/nougat-x86_exports
6. /external/libdrm : Must use Default one.
7. code modification : /device/generic/common
https://github.com/maurossi/device_generic_common/commit/eed772d7828990d64655370be6e0f02b9397f337
What I have tested(Working)
- Realtek ALC1220 Codec
/device/generic/common/alsa/PCH.state needed.(
https://github.com/getfeus/android-oreo-x86_android-auto/blob/master/device/generic/common/alsa/PCH.state
)
- Intel® Dual Band Wireless-AC 8265 card
- NVMe
- Intel I219-V Gigabit LAN controller
- enable_nativebridge : Only check command success. Not testing
Glthread fatal error I have reported.(
https://groups.google.com/forum/#!msg/android-x86/E8iRTfYNJ90/wbnaYCzwBwAJ;context-place=msg/android-x86/ilk69kFFcWY/WlPFjIhSAAAJ
)
What is not stable.
- During boot, ACPI error is displayed. Some kernel module crashed
and re-loaded. Every module is ACPI related of thermal sensor
related.(Mis-configured of defconfig?)
- I can't dump message from dmesg.(Before logging daemon start?)
But, Start with this string, ACPI Error: [_OSI] Namespace lookup failure,
AE_NOT_FOUND on early boot process(Prior to android device dev partition
detect). Very close to this post.
https://bugzilla.redhat.com/show_bug.cgi?id=1413342
- This message is disappeared on old kernel like 4.9.x but
Coffee lake GPU can only working on 4.13.x~. All Kernels I have tested have
same errors. But not care.(Not displayed after boot)
- I can't boot from any Oreo build.
Best regards.
Post by spin877
Post by Mauro Rossi
make -j3 iso_img TARGET_KERNEL_CONFIG=4150_64_defconfig
/device/generic/common/build/tasks/kernel.mk
mk_kernel := + $(hide) $(MAKE) -jX -C
X=cpu
No need to add it if you have added
-jX to make iso_img.
Submake is parallel execution automatically,
29417 pts/42 Sl+ 0:00 | \_ /mnt/cwhuang/oreo-x86/out/soong_ui
--make-mode -j24 iso_img
30617 pts/42 S+ 0:09 | \_
prebuilts/build-tools/linux-x86/bin/ninja -d keepdepfile iso_img -j 24
-f out/combined-android_x86_64.ninja -w dupbuild=err
30944 pts/42 S 0:00 | \_ /bin/sh -c /bin/bash -c
"(make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=\"
30945 pts/42 S 0:00 | \_ /bin/bash -c (make -j24
-C kernel O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE=" /mnt/
30946 pts/42 S 0:00 | \_ make -j24 -C kernel
O=/mnt/cwhuang/oreo-x86/out/target/product/x86_64/obj/kernel
ARCH=x86_64 CROSS_COMPILE= /mnt//cwhuang
--
Chih-Wei
Android-x86 project
http://www.android-x86.org
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
YoungJoon Lee
2018-01-22 01:34:31 UTC
Permalink
I found that mesa I have used is Not rc4. rc2 is used.
/external/mesa : https://github.com/maurossi/mesa/tree/17.2.0rc2_nougat-x86
<https://github.com/maurossi/mesa/tree/17.2.0rc2_nougat-x86>
Post by YoungJoon Lee
Ok, I read log file after reading your comment. It seems like i965(mesa?)
issue. Not opengapps issue.
First, I will check i965_dri.so. I don't know it is right.
./out/target/product/x86_64/symbols/system/lib/dri/i965_dri.so
005d812c
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
addr2line: Dwarf Error: Info pointer extends beyond end of attributes
get_fast_clear_rect
/proc/self/cwd/external/mesa/src/intel/blorp/blorp_clear.c:196
000267b2
do_single_blorp_clear
/proc/self/cwd/external/mesa/src/mesa/drivers/dri/i965/brw_blorp.c:859
00029bee
brw_clear
/proc/self/cwd/external/mesa/src/mesa/drivers/dri/i965/brw_clear.c:233
0035e248
clear
/proc/self/cwd/external/mesa/src/mesa/main/clear.c:221
Next, libglapi.so
./out/target/product/x86_64/symbols/system/lib/libglapi.so
0000ab1e
shared_dispatch_stub_203
/proc/self/cwd/out/target/product/x86_64/obj_x86/SHARED_LIBRARIES/libglapi_intermediates/shared-glapi/glapi_mapi_tmp.h:18408
Next, i965_dri.so
./out/target/product/x86_64/symbols/system/lib64/dri/i965_dri.so
0000000000064741
do_single_blorp_clear
/proc/self/cwd/external/mesa/src/mesa/drivers/dri/i965/brw_blorp.c:789
000000000006769a
brw_clear
/proc/self/cwd/external/mesa/src/mesa/drivers/dri/i965/brw_clear.c:233
I will check each lines of code.
Post by Mauro Rossi
Post by YoungJoon Lee
Thank you Mauro Rossi.
I will test your test image and report here.
Today I found few problems.
(Fatal) 1 : System is extremely heavy. All response is very slow(Slower
than skylake). I monitor details of top command. But it looks like there is
no process consume CPU 100%. I guess it is DVFS(Dynamic Frequency)
Issue.(Kernel message said ACPI related driver is unstable and thermal
driver also print error messages.) How can I debug performance issue? For
example, is there any way to disable power management or frequency
control(Working on maximum clock speed)? I remember that kerner setting on
defconfig's power profile is on demand.
(Not Fatal) 2 : App crashes when keyboard input pop up.(On some apps
only). I will attach logcat and left code here related to this error. It
looks like it is webview related issue. So, I will attach my device mk file
which is import opengapps.
01-21 13:13:56.677 16130 16118 F google-breakpad: -----END BREAKPAD MICRODUMP-----
01-21 13:13:56.800 15305 16118 F libc : Fatal signal 11 (SIGSEGV), code -6, fault addr 0x3bc9 in tid 16118 (Chrome_InProcGp)
01-21 13:13:56.899 16133 16133 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
01-21 13:13:56.899 16133 16133 F DEBUG : Build fingerprint: 'Android-x86/android_x86_64/x86_64:7.1.2/N2G48H/root01210301:userdebug/test-keys'
01-21 13:13:56.899 16133 16133 F DEBUG : Revision: '0'
01-21 13:13:56.899 16133 16133 F DEBUG : ABI: 'x86'
01-21 13:13:56.899 16133 16133 F DEBUG : pid: 15305, tid: 16118, name: Chrome_InProcGp >>> jp.co.yahoo.android.apps.navi <<<
01-21 13:13:56.899 16133 16133 F DEBUG : signal 11 (SIGSEGV), code -6 (SI_TKILL), fault addr 0x3bc9
01-21 13:13:56.899 16133 16133 F DEBUG : eax 00000000 ebx cf899d40 ecx 00008000 edx c0dbec58
01-21 13:13:56.899 16133 16133 F DEBUG : esi 00000000 edi 00000000
01-21 13:13:56.899 16133 16133 F DEBUG : xcs 00000023 xds 0000002b xes 0000002b xfs 0000006b xss 0000002b
01-21 13:13:56.899 16133 16133 F DEBUG : eip cf66012c ebp c0dbec78 esp c0dbe680 flags 00210282
01-21 13:13:56.899 16133 16133 F DEBUG : #00 pc 005d812c /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #01 pc 000267b2 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #02 pc 00029bee /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #03 pc 0035e248 /system/lib/dri/i965_dri.so
01-21 13:13:56.899 16133 16133 F DEBUG : #04 pc 0000ab1e /system/lib/libglapi.so
Here is important to find the source file/line of code causing the
segfault, by providing [???] (output of addr2line)
addr2line -Cfe out/product/target/x86_64/symbols/
system/lib/dri/i965_dri.so
005d812c
[???]
000267b2
[???]
00029bee
[???]
0035e248
[???]
^C
addr2line -Cfe out/product/target/x86_64/symbols/system/lib/libglapi.so
0000ab1e
[???]
Same procedure required for the other SIGSEGV
01-21 12:47:21.791 3767 3959 F libc : Fatal signal 11 (SIGSEGV), code 1, fault addr 0x20c in tid 3959 (Binder:3767_4)
01-21 12:47:21.899 4090 4090 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
01-21 12:47:21.899 4090 4090 F DEBUG : Build fingerprint: 'Android-x86/android_x86_64/x86_64:7.1.2/N2G48H/root01210301:userdebug/test-keys'
01-21 12:47:21.899 4090 4090 F DEBUG : Revision: '0'
01-21 12:47:21.899 4090 4090 F DEBUG : ABI: 'x86_64'
01-21 12:47:21.899 4090 4090 F DEBUG : pid: 3767, tid: 3959, name: Binder:3767_4 >>> com.google.android.gms.unstable <<<
01-21 12:47:21.899 4090 4090 F DEBUG : signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x20c
01-21 12:47:21.899 4090 4090 F DEBUG : rax 0000000000000000 rbx 0000000000000000 rcx 00007c395843de0c rdx 00007c39591d5d94
01-21 12:47:21.899 4090 4090 F DEBUG : rsi 0000000000000000 rdi 0000000000000000
01-21 12:47:21.899 4090 4090 F DEBUG : r8 0000000000000003 r9 00007c39591d7450 r10 0000000000000000 r11 0000000000000005
01-21 12:47:21.899 4090 4090 F DEBUG : r12 00007c395846bc00 r13 00007c3972dd1f00 r14 0000000000000020 r15 0000000000000001
01-21 12:47:21.899 4090 4090 F DEBUG : cs 0000000000000033 ss 000000000000002b
01-21 12:47:21.899 4090 4090 F DEBUG : rip 00007c39587eb741 rbp 0000000000000000 rsp 00007c39591d5ca0 eflags 0000000000010246
01-21 12:47:23.574 4090 4090 F DEBUG : #00 pc 0000000000064741 /system/lib64/dri/i965_dri.so
01-21 12:47:23.574 4090 4090 F DEBUG : #01 pc 000000000006769a /system/lib64/dri/i965_dri.so
addr2line -Cfe out/product/target/x86_64/symbols/system/lib64/dri/i965_dri.so
0000000000064741
[???]
000000000006769a
[???]
Post by YoungJoon Lee
01-21 13:13:56.899 16133 16133 F DEBUG : #05 pc 00a12d91 /system/app/webview/webview.apk (offset 0x927000)
Here is mk file.
GAPPS_VARIANT := nano
PRODUCT_PACKAGES += Chrome \
Maps \
KeyboardGoogle \
GoogleTTS \
YouTube \
PixelIcons \
PixelLauncher \
Wallpapers \
PixelLauncherIcons \
GoogleServicesFramework \
GoogleLoginService
#PRODUCT_PACKAGES += Hangouts
GAPPS_EXCLUDED_PACKAGES := PrebuiltGmsCoreInstantApps \
FaceLock \
AndroidPlatformServices
GAPPS_FORCE_BROWSER_OVERRIDES := true
PRODUCT_PROPERTY_OVERRIDES := \
ro.kernel.android.gps=ttyUSB0 \
ro.kernel.android.gpsttybaud=4800
$(call inherit-product,build/target/product/languages_full.mk)
$(call inherit-product,vendor/opengapps/build/opengapps-packages.mk)
# includes the base of Android-x86 platform
$(call inherit-product,device/generic/common/x86_64.mk)
# Overrides
PRODUCT_NAME := android_x86_64
PRODUCT_BRAND := Android-x86
PRODUCT_DEVICE := x86_64
PRODUCT_MODEL := Generic Android-x86_64
(Not fatal && Solve myself) 3. Cannot adjust spdif optical out's volume.
I think it is normal and can adjust on RC1 is abnormal. I will attach code
what I have modify.
diff --git a/audio/AudioPolicyManagerBase.cpp
b/audio/AudioPolicyManagerBase.cpp
index 74ee22a..e3d16da 100644
--- a/audio/AudioPolicyManagerBase.cpp
+++ b/audio/AudioPolicyManagerBase.cpp
@@ -3248,7 +3248,7 @@ float AudioPolicyManagerBase::computeVolume(int
stream,
index != mStreams[stream].mIndexMin &&
(device == AUDIO_DEVICE_OUT_AUX_DIGITAL ||
device == AUDIO_DEVICE_OUT_DGTL_DOCK_HEADSET)) {
- return 1.0;
+ // return 1.0;
}
diff --git
a/services/core/java/com/android/server/audio/AudioService.java
b/services/core/java/com/android/server/audio/AudioService.java
index c70a87c..87cabe0 100644
--- a/services/core/java/com/android/server/audio/AudioService.java
+++ b/services/core/java/com/android/server/audio/AudioService.java
@@ -520,12 +520,13 @@ public class AudioService extends
IAudioService.Stub {
= new RemoteCallbackList<IAudioRoutesObserver>();
// Devices for which the volume is fixed and VolumePanel slider
should be disabled
- int mFixedVolumeDevices = AudioSystem.DEVICE_OUT_HDMI |
- AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
- AudioSystem.DEVICE_OUT_HDMI_ARC |
- AudioSystem.DEVICE_OUT_SPDIF |
- AudioSystem.DEVICE_OUT_AUX_LINE;
+ int mFixedVolumeDevices = 0;
+ //AudioSystem.DEVICE_OUT_HDMI |
+ //AudioSystem.DEVICE_OUT_DGTL_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_ANLG_DOCK_HEADSET |
+ //AudioSystem.DEVICE_OUT_HDMI_ARC |
+ //AudioSystem.DEVICE_OUT_SPDIF |
+ //AudioSystem.DEVICE_OUT_AUX_LINE;
int mFullVolumeDevices = 0;
I think this modification is not good and it is not intent of
designer's. Any idea's without modification of this code and how to adjust
volume possible?
FATAL EXCEPTION: GLThread 178 issue I have reported(
https://groups.google.com/forum/#!topic/android-x86/E8iRTfYNJ90) is
fixed. It looks like it is related to libdrm or mesa.
Post by Mauro Rossi
This pattern of mesa 17.2.x + nougat-x86 libdrm working suggests
that the problem with Coffelake (gen 8) was most probably due to mesa
libraries moved to /vendor in mesa 17.3 and later
Also when I build nougat, with mesa 17.3 and later, I use x86/oreo-x86
version of external/drm_gralloc
In order to confirm I'm uploading experimental test images with mesa
https://drive.google.com/drive/folders/0B_OFHiIqgpSFMlk4bkhaMnAwbDA
Mauro
Post by YoungJoon Lee
Thank you for your answer. I finally success working nougat on Coffee
lake system. I left How can I do.
https://www.msi.com/Motherboard/Z370I-GAMING-PRO-CARBON-AC.html
2. Base : <a href="
https://osdn.net/projects/android-x86/scm/git/manifest/commits/e8653d3a463c6894daff37073b35b2f880e59a0e"
title="e8653d3a463c6894daff37073b35b2f880e59a0e"
style="color:rgb(0,85,128);outline:0px;font-family:"Helvetica
Neue",Helvetica,Arial,sans-serif;font-size:14.4px" rel="nofollow"
target="_blank" onmousedo
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
fguy
2018-01-21 17:10:05 UTC
Permalink
Post by Mauro Rossi
This pattern of mesa 17.2.x + nougat-x86 libdrm working suggests
that the problem with Coffelake (gen 8) was most probably due to mesa
libraries moved to /vendor in mesa 17.3 and later
Also when I build nougat, with mesa 17.3 and later, I use x86/oreo-x86
version of external/drm_gralloc
In order to confirm I'm uploading experimental test images with mesa
https://drive.google.com/drive/folders/0B_OFHiIqgpSFMlk4bkhaMnAwbDA
I tested your latest build
"nougat_x86_64_amd_dc_llvm50_mesa-17.3.3_hwc1.iso" on the skylake 6Y30.
Unfortunately, the new versions of MESA and kernel did not fix the problem
in 7.1.2 which appeared since November 2017. A number of applications and
games (antutu 3dmark, defense zone 3) using the Unity at startup lead to a
reboot of the android. This problem is not present in 8.1 and in October
assemblies 7.1.2.
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
fguy
2018-01-22 21:35:54 UTC
Permalink
Post by fguy
I tested your latest build
"nougat_x86_64_amd_dc_llvm50_mesa-17.3.3_hwc1.iso" on the skylake 6Y30.
Unfortunately, the new versions of MESA and kernel did not fix the problem
in 7.1.2 which appeared since November 2017. A number of applications and
games (antutu 3dmark, defense zone 3) using the Unity at startup lead to a
reboot of the android. This problem is not present in 8.1 and in October
assemblies 7.1.2.
This error occurs if "ro.hardware.sensor=kbd", if you assign "hdaps", then
these applications run without problems.
--
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 android-***@googlegroups.com.
Visit this group at https://groups.google.com/group/android-x86.
For more options, visit https://groups.google.com/d/optout.
Loading...