Kernel 3.18 is ready for developer testing

[Chih-Wei Huang]

Hi,
I have pushed the android-3.18 branch of kernel
based on the vanilla kernel 3.18 to the git server.
To test it, fetch it from the server and checkout it:

cd kernel
git fetch x86 android-3.18
git checkout x86/android-3.18
cd ..

Then build the iso_img as usual.

m -jX iso_img

You may also build the 64-bit version:
(with 32-bit userspace)

touch bionic/libc/kernel/uapi/linux/binder.h
m -jX iso_img TARGET_KERNEL_ARCH=x86_64 \
TARGET_USES_64_BIT_BINDER=true

It is only build-able with the lollipop-x86 branch now.
Later I'll update the kitkat-x86 kernel building rules
to align with lollipop-x86.

Especially thanks to pstglia who made this possible.


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

[pstglia]

Thank you very much for this! After trying to merge 3.16 and Android 3.14 kernels through "git rebase", I know how hard this work is. 

Plus, I saw the patches you pushed to fix lot of problems

It's really amazing the job you've been doing all these years with Android-x86

Regards,

pstglia

ps: For many people, you are Santa Claus today :)

[rbg]

thanks for the update...

[rbg]

re:

touch bionic/libc/kernel/uapi/linux/binder.h 
m -jX iso_img TARGET_KERNEL_ARCH=x86_64 \ 
TARGET_USES_64_BIT_BINDER=true 

missing android-x86_64_defconfig....??

On Thursday, December 25, 2014 2:38:54 PM UTC-5, Chih-Wei Huang wrote:

[Chih-Wei Huang]

2014-12-26 12:05 GMT+08:00 rbg <rbg.hp...@gmail.com>:
> re:
> touch bionic/libc/kernel/uapi/linux/binder.h
> m -jX iso_img TARGET_KERNEL_ARCH=x86_64 \
> TARGET_USES_64_BIT_BINDER=true
>
> missing android-x86_64_defconfig....??

Oh, I pushed the wrong branch.
Too sleepy...

Now corrected.
Sorry for that.

[Mauro Rossi]

Chih-Wei thanks a lot for brand new kernel to you and pstglia

I'm trying to build x86_64 iso, with mesa 9.2.0devel modded for 64 bit, at this point the only thing missing is to enable ASM optimizations for i915 and i965.

Question1: is TARGET_KERNEL_ARCH=x86_64 mandatory when giving the following command (without 64 bit binder, just 32 bit one)?

make -j5 iso_img TARGET_PRODUCT=android_x86_64 TARGET_KERNEL_CONFIG=my_x86_64_defconfig

Question2: TARGET_USES_64_BIT_BINDER=true in my undestanding this will enable binder64, will it also enable zygote64?

Thanks

[Chih-Wei Huang]

2014-12-26 18:14 GMT+08:00 Mauro Rossi <issor...@gmail.com>:
> Chih-Wei thanks a lot for brand new kernel to you and pstglia
>
> I'm trying to build x86_64 iso, with mesa 9.2.0devel modded for 64 bit, at
> this point the only thing missing is to enable ASM optimizations for i915
> and i965.
> Question1: is TARGET_KERNEL_ARCH=x86_64 mandatory when giving the following
> command (without 64 bit binder, just 32 bit one)?
>
> make -j5 iso_img TARGET_PRODUCT=android_x86_64
> TARGET_KERNEL_CONFIG=my_x86_64_defconfig

You don't need to specify TARGET_KERNEL_ARCH=x86_64
to build target android_x86_64.
It's set automatically.
(64-bit userspace requires 64-bit kernel, of course)

> Question2: TARGET_USES_64_BIT_BINDER=true in my undestanding this will
> enable binder64, will it also enable zygote64?

No.
It just specifies the binder version to be used.
64-bit kernel requires binder version 8
(so TARGET_USES_64_BIT_BINDER=true)
I'll push another change to force
TARGET_USES_64_BIT_BINDER=true
if 64-bit kernel is built.

32-bit kernel can use binder version 7 or 8 so
TARGET_USES_64_BIT_BINDER can be true or false.
It just needs to meet your kernel config.
(CONFIG_ANDROID_BINDER_IPC_32BIT)

[Vaidotas]

I must be doing something wrong, because after checkout command I get an error message.

$ git fetch x86 android-3.18
From http://git.android-x86.org/kernel/common
 * branch            android-3.18 -> FETCH_HEAD
$ git checkout x86/android-3.18
error: pathspec 'x86/android-3.18' did not match any file(s) known to git.

[Vaidotas]

Looks like I needed to do a $git fetch first.

[Chih-Wei Huang]

2014-12-27 1:29 GMT+08:00 Vaidotas <jos...@gmail.com>:
> I must be doing something wrong, because after checkout command I get an
> error message.
>
> $ git fetch x86 android-3.18
> From http://git.android-x86.org/kernel/common
> * branch android-3.18 -> FETCH_HEAD
> $ git checkout x86/android-3.18
> error: pathspec 'x86/android-3.18' did not match any file(s) known to git.

Try

git checkout FETCH_HEAD

This is because your git version is too old so
it didn't create the remote branch after fetching.

[sarma...@yahoo.com]

Are the changes you made for the ASUS T100 image you made included in this branch?

Thanks for all your great work!

Eric

[György Pejović]

Android-x86 5.0.2 kernel 3.18

https://docs.google.com/uc?id=0B73McXwkxx69V0xXSmw3X2Ezdk0&export=download

[Chih-Wei Huang]

2014-12-27 10:27 GMT+08:00 sarman_1998 via Android-x86
<andro...@googlegroups.com>:

> Are the changes you made for the ASUS T100 image you made included in this
> branch?

Yep, I think so.

[Chih-Wei Huang]

2014-12-26 3:38 GMT+08:00 Chih-Wei Huang <cwh...@android-x86.org>:
> It is only build-able with the lollipop-x86 branch now.
> Later I'll update the kitkat-x86 kernel building rules
> to align with lollipop-x86.

The kitkat-x86 branch is updated to be compatible
with kernel 3.18. You can build it the same as lollipop-x86 now.
Unfortunately the 64-bit kernel is not compatible with kitkat-x86
(since kitkat only supports binder version 7)
So you can only build the 32-bit kernel with kitkat-x86.

[Mashnoon Ibtesum]

Hi! Could you build a 3.18 iso with the gapps ? (Pretty please with sprinkle on top?)

[György Pejović]

android-x86/out/target/product/x86/system/app

android-x86/out/target/product/x86/system/priv-app

You add + https://drive.google.com/open?id=0B73McXwkxx69WGxRaEpSaUpoeUE&authuser=0

[György Pejović]

5.0.2, kernel 3.18, +gapps

https://drive.google.com/file/d/0B73McXwkxx69blZXY3p2cnlRQUk/view?usp=sharing

[George Turner]

Still getting 972 error from play store when installed to a drive by itself. Format ext3, yes for grub, system RW.

[Mashnoon Ibtesum]

Really appreciate the build, but unfortunately when I try to sign in to play store/gmail/youtube playservice would crash. Everytime!

On Saturday, December 27, 2014 8:05:32 PM UTC-5, György Pejović wrote:

[György Pejović]

Sorry, I don't tested. Tonight try to improve.

[bk chaw]

Google play services FC everytime when i try to run internet browser,gmail,google account setup, etc... On lollipop

On Sunday, December 28, 2014, György Pejović <truet...@gmail.com> wrote:

Sorry, I don't tested. Tonight try to improve.

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

[Alessandro Bertolero]

Thank you very much for this!!

[monstercameron]

Have you included support for amd gcn gpus?

On Thursday, December 25, 2014 11:38:54 AM UTC-8, Chih-Wei Huang wrote:

[rbg]

kitkat-x86 with Kernel 3.18, for TESTING PURPOSES ONLY, available here:

   https://drive.google.com/file/d/0B2evBKvIpdYkdTZDbjV0UGJFVVE/view?usp=sharing

help developers, my system don't work...doesn't help anyone....

identify your system, provide logcat / dmesg / bugreport logs if

possible....

don't know how...use the android-x86 forums and look it up,

many others have done it....

Help the developers help you....

[Fajar A. Nugraha]

On Tue, Dec 30, 2014 at 10:21 AM, rbg <rbg.hp...@gmail.com> wrote:
> kitkat-x86 with Kernel 3.18, for TESTING PURPOSES ONLY, available here:
>
>
> https://drive.google.com/file/d/0B2evBKvIpdYkdTZDbjV0UGJFVVE/view?usp=sharing
>
>
> help developers, my system don't work...doesn't help anyone....
>
> identify your system, provide logcat / dmesg / bugreport logs if
> possible....

logcat from Asus T100 boot here: http://pastebin.com/ZKk7z40k

boots quite fast, but wifi doesn't work, so not much test can be done.
Can wifi be enabled without having to compile the kernel manually
(e.g. by downloading a firmware)?

--
Fajar

[bk chaw]

Thanks ! Been waiting for this

--

[Andrew Stubbs]

I have successfully managed to build the lollipop branch with this kernel, and it run's successfully on my Toshiba WT10-A tablet, however the Wifi does not appear to work. I can turn it on, however no access points appear.

I have located the source code for my wifi card (RTL8723BS) https://github.com/hadess/rtl8723bs - However, i am unsure how to merge the code into the android source to compile this driver into Android.

Any help?

[pstglia]

Hi Andrew

I have located the source code for my wifi card (RTL8723BS) https://github.com/hadess/rtl8723bs - However, i am unsure how to merge the code into the android source to compile this driver into Android.

 

You can do something like this:

1) Get hadess's git and copy entire directory to kernel/drivers/net/wireless:

$ ANDROID_SRC_DIR/kernel/drivers/net/wireless/rtl8723as$ ls -l

-rwxr-xr-x 1 paulo paulo   360 Dec 30 02:04 Kconfig*

-rwxr-xr-x 1 paulo paulo 42101 Dec 30 02:04 Makefile*

drwxr-xr-x 3 paulo paulo  4096 Dec 30 02:04 core/

drwxr-xr-x 7 paulo paulo  4096 Dec 30 02:04 hal/

drwxr-xr-x 4 paulo paulo 12288 Dec 30 02:04 include/

drwxr-xr-x 3 paulo paulo  4096 Dec 30 02:04 os_dep/

2) Under kernel/drivers/net/wireless, edit "Kconfig" file, adding the line marked with + signal (do not include the plus sign when pasting):

source "drivers/net/wireless/rtl818x/Kconfig"

+ source "drivers/net/wireless/rtl8723bs/Kconfig"

3) Go to Android root directory and type these to access kernel menuconfig:

$ . build/envsetup.sh
$ lunch android_x86-eng
$ make -C kernel O=$OUT/obj/kernel ARCH=x86 menuconfig

4) Inside kernel menuconfig, navigate to Device Drivers -> Network Device Support -> Wireless LAN and enable these 2 options using spacebar ( [M] to compile as module ; [*] to be built-in kernel img ):

Realtek 8723A SDIO or SPI WiFi

Realtek 8723B SDIO or SPI WiFi

Enable SDIO interface for Realtek

5) Save your new config and exit menuconfig

6) Delete these files/directories to ensure they will be built with your changes:

cd YOUR_ANDROID_ROOT_SRC

rm -f out/target/product/x86/kernel out/target/product/x86/system.sfs out/target/product/x86/obj/PACKAGING/systemimage_intermediates/system.img  

rm -rf out/target/product/x86/system/lib/modules

7) build as usual

Note: You can also edit .config and add options directly on it. It's your choice

Hope this helps

[Andrew Stubbs]

I completed all these steps successfully as a module, however the rtl8723bs code is not compiling. I've double, and triple checked the steps.

Android is building OK, just not the new driver.

[pstglia]

I completed all these steps successfully as a module, however the rtl8723bs code is not compiling. I've double, and triple checked the steps.

Android is building OK, just not the new driver.

A) Confirm if the config options were included to your .config file:

cd YOUR_ANDROID_SRC

egrep "8723[AB]S|WLAN_SDIO" out/target/product/x86/obj/kernel/.config

Should return this:

CONFIG_RTL8723AS=m

CONFIG_RTL8723BS=m

CONFIG_WLAN_SDIO=m

B) Confirming 1, try removing all the files and directories inside out/target/product/x86/obj/kernel ( keep there just .config file ). Remove all the other files/directories mentioned before on step 6 and try building again

Regards,

pstglia

 

[rbg]

>You may also build the 64-bit version: 
>(with 32-bit userspace) 

lollipop-x86-5.0.2_r1 with Kernel 3.18-x64, for TESTING PURPOSES ONLY, available here:

  

   https://drive.google.com/file/d/0B2evBKvIpdYkcm9EOVREejZSRGs/view?usp=sharing

help developers, my system don't work...doesn't help anyone....

identify:

---------

what build your running: lp-x86-5.0.2_r1 with kernel 3.18-x64

your system: dell xyz / asus xyz / etc.

issue: wifi / screen / etc. 

provide logcat / dmesg / bugreport logs if possible....

don't know how...use the android-x86 forums and look it up,

many others have done it....

Help the developers help you....

On Thursday, December 25, 2014 2:38:54 PM UTC-5, Chih-Wei Huang wrote:

Hi,
I have pushed the android-3.18 branch of kernel
based on the vanilla kernel 3.18 to the git server.
To test it, fetch it from the server and checkout it:

cd kernel
git fetch x86 android-3.18
git checkout x86/android-3.18
cd ..

Then build the iso_img as usual.

m -jX iso_img

You may also build the 64-bit version:
(with 32-bit userspace)

touch bionic/libc/kernel/uapi/linux/binder.h
m -jX iso_img TARGET_KERNEL_ARCH=x86_64 \
TARGET_USES_64_BIT_BINDER=true

It is only build-able with the lollipop-x86 branch now.
Later I'll update the kitkat-x86 kernel building rules
to align with lollipop-x86.

Especially thanks to pstglia who made this possible.

[Fajar A. Nugraha]

What is the difference between this one and the one you publish
yesterday (android_x86-KK-KRNL318-20141229.iso)? Is it 64bit vs 32bit
userspace?

--
Fajar

[rbg]

the one from yesterday is KITKAT with a standard 32 bit kernel, this one is Lollipop with with the 64 bit kernel as explained by Chih-Wei

[rbg]

your wifi doesn't work because the driver isn't loading....maybe a firmware issue???

[Carbon Jiao]

Thanks for your build!

While it cannot boot to Android system in my tablet, http://paste.ubuntu.com/9649616/ is the cpuinfo, and http://paste.ubuntu.com/9649620/ is the kmsg,

could you help to check how to settle the problems?

Thanks a lot!

Carbon  < carbonjiao # gmail dot com >

在 2014年12月30日星期二UTC+8上午11时21分05秒，rbg写道：

[Miray Cakin]

If you download brcmfmac43241b4-sdio.txt and copy it to /lib/firmware/brcm/ you will get wifi to work..

But with 3.18 it still hanging sometimes...

[Andrew Stubbs]

I can confirm the presence of the lines in the .config file, I can also confirm all build files from the out directory were removed.

I even did a clean, and fresh build however the driver is not building.

I can only assume something is missing preventing the driver from building?

[Andrew Stubbs]

Ok, i got a little closer. I added the following line to kernel/drivers/net/wireless/Makefile

obj-$(CONFIG_RTL8723BS)         += rtl8723bs/

and the driver now compiles.

[pstglia]

obj-$(CONFIG_RTL8723BS)         += rtl8723bs/

Nicely done! I forgot this

 

and the driver starts to compile, however errors out with

  CHK     include/config/kernel.release

  GEN     ./Makefile

  CHK     include/generated/uapi/linux/version.h

  Using /mnt/android-x86/kernel as source for kernel

  CHK     include/generated/utsrelease.h

  CALL    /mnt/android-x86/kernel/scripts/checksyscalls.sh

  CHK     include/generated/compile.h

  CC [M]  drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.o

/mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c: In function ‘rtw_cfg80211_rx_action_p2p’:

/mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c:3805:2: error: too many arguments to function ‘cfg80211_rx_mgmt’

  rtw_cfg80211_rx_mgmt(padapter, freq, 0, pmgmt_frame, frame_len, GFP_ATOMIC);

  ^

In file included from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/osdep_service_linux.h:57:0,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/osdep_service.h:36,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/drv_types.h:32,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c:22:

/mnt/android-x86/kernel/include/net/cfg80211.h:4470:6: note: declared here

 bool cfg80211_rx_mgmt(struct wireless_dev *wdev, int freq, int sig_dbm,

      ^

/mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c: In function ‘rtw_cfg80211_rx_p2p_action_public’:

/mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c:3828:2: error: too many arguments to function ‘cfg80211_rx_mgmt’

  rtw_cfg80211_rx_mgmt(padapter, freq, 0, pmgmt_frame, frame_len, GFP_ATOMIC);

  ^

In file included from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/osdep_service_linux.h:57:0,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/osdep_service.h:36,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/drv_types.h:32,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c:22:

/mnt/android-x86/kernel/include/net/cfg80211.h:4470:6: note: declared here

 bool cfg80211_rx_mgmt(struct wireless_dev *wdev, int freq, int sig_dbm,

      ^

/mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c: In function ‘rtw_cfg80211_rx_action’:

/mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c:3854:2: error: too many arguments to function ‘cfg80211_rx_mgmt’

  rtw_cfg80211_rx_mgmt(adapter, freq, 0, frame, frame_len, GFP_ATOMIC);

  ^

In file included from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/osdep_service_linux.h:57:0,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/osdep_service.h:36,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/include/drv_types.h:32,

                 from /mnt/android-x86/kernel/drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.c:22:

/mnt/android-x86/kernel/include/net/cfg80211.h:4470:6: note: declared here

 bool cfg80211_rx_mgmt(struct wireless_dev *wdev, int freq, int sig_dbm,

      ^

make[6]: *** [drivers/net/wireless/rtl8723bs/os_dep/ioctl_cfg80211.o] Error 1

make[5]: *** [drivers/net/wireless/rtl8723bs] Error 2

make[5]: *** Waiting for unfinished jobs....

make[4]: *** [drivers/net/wireless] Error 2

make[3]: *** [drivers/net] Error 2

make[3]: *** Waiting for unfinished jobs....

make[2]: *** [drivers] Error 2

make[2]: Leaving directory `/mnt/android-x86/out/target/product/x86/obj/kernel'

make[1]: *** [sub-make] Error 2

make[1]: Leaving directory `/mnt/android-x86/kernel'

make: *** [out/target/product/x86/kernel] Error 2

Seems related to this patch:

http://www.spinics.net/lists/linux-wireless/msg125660.html

A parameter was removed from cfg80211_rx_mgmt function (gfp_t gfp)

Try removing the last parameter from every call to "cfg80211_rx_mgmt" and try again.

Regards,

pstglia

 

[Chih-Wei Huang]

2014-12-30 11:21 GMT+08:00 rbg <rbg.hp...@gmail.com>:
> kitkat-x86 with Kernel 3.18, for TESTING PURPOSES ONLY, available here:
>
> https://drive.google.com/file/d/0B2evBKvIpdYkdTZDbjV0UGJFVVE/view?usp=sharing
>
> help developers, my system don't work...doesn't help anyone....
>
> identify your system, provide logcat / dmesg / bugreport logs if
> possible....

Anyone one has Dell Venue 8 Pro or other similar BayTrail-based tablet,
could you please test the iso and show me the result of the commands?

cat /sys/class/dmi/id/uevent

lsmod


Thanks!

[Fajar A. Nugraha]

Does Asus T100 qualify as "similar baytrail"? 

cat and lsmod result attached (tried copy-pasting it, but android terminal emulator is not copy-paste friendly).

Additional info:

wifi:

- works after I installed it to disk with system rw (doesn't work on live). No need to add any additional firmware.

- kinda flakey. It works when its on from the start, or when it's off but then you turn it on. It doesn't work when you turn it off and then on.

- connection speed is good, which is better then on ubuntu and t100-specific kernel (https://drive.google.com/folderview?id=0B4s5KNXf2Z36MlhrQVplZC14UWM&usp=sharing&tid=0B9C1WK1FQhjfcXNrbzN6djQzajg ,  slow to the point of unusable)

other things that work:

- youtube

- battery status

what doesn't work:

- bluetooth

- physical buttons (which works on ubuntu with the above kernel)

- scroll using touchpad

- other things that are known to be broken (i.e. suspend/resume using power button)

Sometimes it also freezes, or seems to be under heavy load (slow to response)

-- 

Fajar

 

[Vaidotas]

Lenovo Miix 2 8 tablet. Log files attached from this particular image. Video fails to work, even with nomodeset, but that works in lollipop.

[Andrew Stubbs]

Removing of the parameter appeared to work. The wifi now turns on, and displays neighboring access points. However it does not appear to last forever.

When you select your AP and enter the password, it saves it but does not connect. When you reselect, and click connect, it just sits as saved.

If you refresh the AP List, none appear. Looks like the wifi initially works, then fails.

[Jocky Wilson]

I tested with my Lenovo Thinkpad Tablet 8. Wifi and BT don't work, probably because the broadcom device registers as a usb device instead of as a sdio device (see attached dmesg). Sound also doesn't work. In order to get wifi to work i dumped the broadcom nvram from efivard using the fedlet linux. I also attached that file.I had to rebuild system.img with that file included in /lib/firmware/brcm I also replaced the wifi main driver file by the attached one which I extracted from the win8.1 driver package. So far I never got wifi to work with any of the android and linux images I tried on my thinkpad tablet.It appears the sdio wifi device depends on some other device.

Keep up the great work guys!

/JockyW

PS: On my Voyo A1 mini, which is also a baytrail device with intel hd graphics, the display garbles as soon as android starts when it should show the Android splash. The only OS I can get to work on the Voyo is Ubuntu 14.04 with a 3.13.0-32-generic kernel. All other images (fedlet, android, etc.) have the same garbled graphics issue upon start.

[Jocky Wilson]

I posted a reply 1.5 hr ago, but why doesn't it appear? In the left pane it says I posted to the group....

[lewis donofrio]

if this is page for 12-27-2014 kern 3.18 with gapps, I'm still getting 972 on apps store updates (or probably any apk install) this is on my hp notebook 8470p hard drive install.....help!

[zeethe...@gmail.com]

I tried the two commands on my DVP8.. but don't know how to save it to a file.. and upload it here.. Can someone pls tell me how to save it ? it’d be helpful

and the display is totally distorted.. but touchscreen works.. could press some options through the distorted display.

Sent from Windows Mail

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

[pstglia]

Have you looked dmesg/logcat outputs? Maybe we can have a clue what's going on there

[Andrew Stubbs]

I'll be honest, I'm not 100% sure what I'm looking at with the log files, but here they are

[Edgardo Gho]

I have a similar problem.

If I use kernel 3.18 the SSID scan returns 0 Beacons:

D/wpa_supplicant( 1832): wlan0: Event SCAN_STARTED (49) received

D/wpa_supplicant( 1832): wlan0: Own scan request started a scan in 0.000042 seconds

I/wpa_supplicant( 1832): wlan0: CTRL-EVENT-SCAN-STARTED 

D/wpa_supplicant( 1832): CTRL_IFACE monitor sent successfully to /data/misc/wifi/sockets/wpa_ctrl_1745-2\x00

D/wpa_supplicant( 1832): wlan0: Control interface command 'SCAN TYPE=ONLY'

D/wpa_supplicant( 1832): Ongoing scan action - reject new request

D/wpa_supplicant( 1832): RTM_NEWLINK: ifi_index=4 ifname=wlan0 wext ifi_flags=0x1003 ([UP])

D/wpa_supplicant( 1832): nl80211: Event message available

D/wpa_supplicant( 1832): nl80211: Drv Event 34 (NL80211_CMD_NEW_SCAN_RESULTS) received for wlan0

D/wpa_supplicant( 1832): wlan0: nl80211: New scan results available

D/wpa_supplicant( 1832): nl80211: Scan probed for SSID ''

D/wpa_supplicant( 1832): nl80211: Scan included frequencies: 2412 2417 2422 2427 2432 2437 2442 2447 2452 2457 2462 2467 2472

D/wpa_supplicant( 1832): wlan0: Event SCAN_RESULTS (3) received

D/wpa_supplicant( 1832): wlan0: Scan completed in 0.984934 seconds

D/wpa_supplicant( 1832): nl80211: Received scan results (0 BSSes)

D/wpa_supplicant( 1832): wlan0: BSS: Start scan result update 40

D/wpa_supplicant( 1832): BSS: last_scan_res_used=0/0

D/wpa_supplicant( 1832): wlan0: Scan-only results received

D/wpa_supplicant( 1832): CTRL_IFACE monitor sent successfully to /data/misc/wifi/sockets/wpa_ctrl_1745-2\x00

D/wpa_supplicant( 1832): wlan0: Radio work 'scan'@0xb6e18430 done in 1.346841 seconds

D/wpa_supplicant( 1832): wlan0: Control interface command 'BSS RANGE=0- MASK=0x21987'

D/wpa_supplicant( 1832): wlan0: Control interface command 'SCAN TYPE=ONLY'

D/wpa_supplicant( 1832): wlan0: Setting scan request: 0.000000 sec

D/wpa_supplicant( 1832): wlan0: Starting AP scan for wildcard SSID

D/wpa_supplicant( 1832): WPS: Building WPS IE for Probe Request

D/wpa_supplicant( 1832): WPS:  * Version (hardcoded 0x10)

D/wpa_supplicant( 1832): WPS:  * Request Type

D/wpa_supplicant( 1832): WPS:  * Config Methods (4288)

D/wpa_supplicant( 1832): WPS:  * UUID-E

D/wpa_supplicant( 1832): WPS:  * Primary Device Type

D/wpa_supplicant( 1832): WPS:  * RF Bands (1)

D/wpa_supplicant( 1832): WPS:  * Association State

D/wpa_supplicant( 1832): WPS:  * Configuration Error (0)

D/wpa_supplicant( 1832): WPS:  * Device Password ID (0)

D/wpa_supplicant( 1832): WPS:  * Manufacturer

D/wpa_supplicant( 1832): WPS:  * Model Name

D/wpa_supplicant( 1832): WPS:  * Model Number

D/wpa_supplicant( 1832): WPS:  * Device Name

D/wpa_supplicant( 1832): WPS:  * Version2 (0x20)

D/wpa_supplicant( 1832): P2P: * P2P IE header

D/wpa_supplicant( 1832): P2P: * Capability dev=21 group=00

D/wpa_supplicant( 1832): P2P: * Listen Channel: Regulatory Class 81 Channel 1

D/wpa_supplicant( 1832): wlan0: Add radio work 'scan'@0xb6e18430

D/wpa_supplicant( 1832): wlan0: First radio work item in the queue - schedule start immediately

D/wpa_supplicant( 1832): wlan0: Starting radio work 'scan'@0xb6e18430 after 0.000030 second wait

D/wpa_supplicant( 1832): wlan0: nl80211: scan request

D/wpa_supplicant( 1832): nl80211: Scan SSID - hexdump(len=0): [NULL]

D/wpa_supplicant( 1832): nl80211: Scan extra IEs - hexdump(len=185): dd a4 00 50 f2 04 10 4a 00 01 10 10 3a 00 01 00 10 08 00 02 42 88 10 47 00 10 88 51 71 a5 2f 49 ...

D/wpa_supplicant( 1832): Scan requested (ret=0) - scan timeout 30 seconds

D/wpa_supplicant( 1832): nl80211: Event message available

D/wpa_supplicant( 1832): nl80211: Drv Event 33 (NL80211_CMD_TRIGGER_SCAN) received for wlan0

D/wpa_supplicant( 1832): wlan0: nl80211: Scan trigger

D/wpa_supplicant( 1832): wlan0: Event SCAN_STARTED (49) received

If I use 3.10 it works.

I use a realtek card and as far as I can tell the driver loads properly. It turns on, but above you can see no scans results.

[Edgardo Gho]

I found the problem on my wifi card:

https://bugzilla.kernel.org/show_bug.cgi?id=88951

Kernel 3.18.1 comes with a fix for the problem of no networks on the SSID scan.

I backported the 3 changes mentioned on the bug and I was able to compile and have wifi working.

I guess when the project gets updated to 3.18.1 this will be fix for everyone. As for now, seems to work by just applying the patches.

[pstglia]

Hi Andrew, I had missed your reply, sorry

I saw some occurrences of this msg on dmesg output:

<0>[  248.319520] RTL871X:  fw read cmd failed..

This is print inside FillH2CCmd8723B function:

========================================================================
      //pay attention to if  race condition happened in  H2C cmd setting.
        do{
                h2c_box_num = pHalData->LastHMEBoxNum;

                if(!_is_fw_read_cmd_down(padapter, h2c_box_num)){
                        DBG_8192C(" fw read cmd failed...\n");
#ifdef DBG_CHECK_FW_PS_STATE
                        DBG_871X("MAC_1C0=%08x, MAC_1C4=%08x, MAC_1C8=%08x, MAC_1CC=%08x\n", rtw_read32(padapter, 0x1c0), rtw_read32(padapter, 0x1c4)
                                , rtw_read32(padapter, 0x1c8), rtw_read32(padapter, 0x1cc));
#endif //DBG_CHECK_FW_PS_STATE
                        //DBG_8192C(" 0x1c0: 0x%8x\n", rtw_read32(padapter, 0x1c0));
                        //DBG_8192C(" 0x1c4: 0x%8x\n", rtw_read32(padapter, 0x1c4));
                        goto exit;
                }
========================================================================

To be very honest, I have no idea this function and "_is_fw_read_cmd_down" is supposed to (fw reading?)
This appears to be part of Realtek original code, as many drivers share the same function call/ debugging msg.

I tried to search more info on web. On fedlet page, I saw someone say that they could with networks using WEP encryption and Open networks (both with bugs - not that stable), but no WPA.
Can't hurt to try :)

[Chih-Wei Huang]

2015-01-08 2:31 GMT+08:00 Edgardo Gho <edgar...@gmail.com>:
> I found the problem on my wifi card:
> https://bugzilla.kernel.org/show_bug.cgi?id=88951
> Kernel 3.18.1 comes with a fix for the problem of no networks on the SSID
> scan.
> I backported the 3 changes mentioned on the bug and I was able to compile
> and have wifi working.
> I guess when the project gets updated to 3.18.1 this will be fix for
> everyone. As for now, seems to work by just applying the patches.

Good to know that.
Thank you for the info.

[rbg]

did the patch yesterday...3.18 patched to 3.18.1 fine

ref: https://www.kernel.org/pub/linux/kernel/v3.x/patch-3.18.1.xz

[Jocky Wilson]

Yes, the patch works fine and I could build a x86 kk image with kernel 3.18.1 booting up nicely on my baytrail tablets, a Lenovo Thinkpad Tablet 8 and a Voyo A1 Mini.
Unfortunately, wifi still doesn't work. My Voyo A1 mini has a RTK8723BS SDIO which isn't detected. There are no devices shown in /sys/bus/sdio/devices
When I boot Fedlet (== Fedora with Adam Williamson's Baytrail patches) which uses kernel 3.18.0 the RTK8723BS sdio device is detected and shows as mmc1:0001:1 directory under /sys/bus/sdio/devices

Is SDIO support missing in the kernel config? I used git checkout x86/android-3.18 and applied this 3.18.1 patch https://www.kernel.org/pub/linux/kernel/v3.x/patch-3.18.1.xz

Cheers
JockyW

[Chih-Wei Huang]

2015-01-08 21:32 GMT+08:00 Jocky Wilson <jocky...@gmail.com>:
> Yes, the patch works fine and I could build a x86 kk image with kernel
> 3.18.1 booting up nicely on my baytrail tablets, a Lenovo Thinkpad Tablet 8
> and a Voyo A1 Mini.
> Unfortunately, wifi still doesn't work. My Voyo A1 mini has a RTK8723BS SDIO
> which isn't detected. There are no devices shown in /sys/bus/sdio/devices
> When I boot Fedlet (== Fedora with Adam Williamson's Baytrail patches) which

Most kernel patches in 20141209-fedlet release
have been applied to the 4.4-r2 release.
(except the backlight patch which has some issues)

> uses kernel 3.18.0 the RTK8723BS sdio device is detected and shows as
> mmc1:0001:1 directory under /sys/bus/sdio/devices
>
> Is SDIO support missing in the kernel config? I used git checkout

What config does it need?
You may compare the kernel configs between
android-x86 and fedlet.

[Andrew Stubbs]

Thanks pstglia for looking into this. I'll have a look around regarding the specific errors and see if I can find anything.

I've also noted a few new commits to the drivers git, one specifically for the 3.18 kernel, so I'll sync my code tonight and rebuild to see what happens.

[Andrew Stubbs]

Ok, I synced the new commits, and now the driver loads and works. It does not always connect first time, but persistence is key.

I may start a new thread shortly with a download link

[Jocky Wilson]

hi Andrew, which commit version of https://github.com/hadess/rtl8723bs did you use?
The current version is apparently not working in android, see: https://github.com/hadess/rtl8723bs/issues/6

Cheers,
JockyW

[Andrew Stubbs]

Hi Jocky,

I just downloaded the master branch, cleared all build files and rebuilt 5.0.2 with 3.18 kernel - When it boots, it run's like a charm. Connecting to the WiFi initially is a game of change.. Goes connecting... then back to saved about 5 times. Then it connects and all run's well. Google Play store work's fine, installed a few games which ran without issue (other than no sound)

Any one want's to try,

https://drive.google.com/file/d/0B7iqrsHkLwSqOW5uYlJfeGxnQ28/view?usp=sharing

[Jocky Wilson]

Hi Andrew,

On Thursday, January 8, 2015 at 10:08:09 PM UTC+1, Andrew Stubbs wrote:

Hi Jocky,

I just downloaded the master branch, cleared all build files and rebuilt 5.0.2 with 3.18 kernel - When it boots, it run's like a charm. Connecting to the WiFi initially is a game of change.. Goes connecting... then back to saved about 5 times. Then it connects and all run's well. Google Play store work's fine, installed a few games which ran without issue (other than no sound)

Hi Andrew,

The master branch has changed a lot in the past couple of days. You can find your version in the file .git/logs/HEAD, the commit version is usually the first 7 digits of the hash value, e.g. 3dd8594d if the hash is 3dd8594d850502fb76d0170209ef24d93e372753
 

 

Any one want's to try,

https://drive.google.com/file/d/0B7iqrsHkLwSqOW5uYlJfeGxnQ28/view?usp=sharing

Thanks for your iso. I tried it out on my Voyo which also has a 8723BS SDIO. It boots up fine to the home screen, but unfortunately, wifi doesn't work. 

Cheers,
JockyW

[Andrew Stubbs]

Hi. I'm unable to get the commit details at the moment. But I downloaded the branch about 3 hours ago.

[Jocky Wilson]

I merged various settings from the Fedlet kernel config. Unfortunately still no sdio devices listed on the Voyo A1 mini tablet.
In the kernel log there is an entry:
mmc1: SDHCI controller on ACPI [INT33BB:00] using ADMA
When I insmod 8723bs.ko then the driver loads without problems, but no wifi.

On the Lenovo Thinkpad Tablet 8, the broadcom sdio wifi device (brcmfmac) is found and does show up under /sys/bus/sdio/devices, but wifi doesn't start
After insmod brcmfmac.ko I see "mmc1: queueing unknown CIS tuple 0x80 (2 bytes)" errors, followed by "byt-rt5640 baytrail-pcm-audio not registered", "byt-rt5640 requests probe deferral", brcmfmac loads, followed by "brcm_sdio_htclk: HT Avail timeout (1000000): clkctl 0x50" errors. It seems the broadcom wifi device is somehow coupled with the sound device?!

Not so easy to tackle these issues :(

[Mashnoon Ibtesum]

Hi andrew, thanks for the build but Im getting black screen on my AMD A8. Any idea cuz other 3.18 L build works fine but play services crashed on them every time?

[Miker1029]

Hey rbg,

Just Downloaded And Installed 5.0.2 -r2 x64, Running like a Bad A$$, Wifi, Sound working on my Dell 1750, Question though, Got A way to Put GAPPS on Here, Atleast the Play Store? Or Is that Gonna Cause An Issue?

Or Maybe Update from the 5.0 Alpha? Dunno If that would work with the Kernel Change....

Any Ideas?

Thanks

Mike

[oscar bg]

It's very nice to see android x86 progressing fast with lollipop 64bit kernel builds but what about 64 bit user space lollipop builds?

it's possible? what's left to do?

thanks..

On Thursday, December 25, 2014 at 8:38:54 PM UTC+1, Chih-Wei Huang wrote:

Hi,
I have pushed the android-3.18 branch of kernel
based on the vanilla kernel 3.18 to the git server.
To test it, fetch it from the server and checkout it:

cd kernel
git fetch x86 android-3.18
git checkout x86/android-3.18
cd ..

Then build the iso_img as usual.

m -jX iso_img

You may also build the 64-bit version:
(with 32-bit userspace)

touch bionic/libc/kernel/uapi/linux/binder.h
m -jX iso_img TARGET_KERNEL_ARCH=x86_64 \
TARGET_USES_64_BIT_BINDER=true

It is only build-able with the lollipop-x86 branch now.
Later I'll update the kitkat-x86 kernel building rules
to align with lollipop-x86.

Especially thanks to pstglia who made this possible.

[Greg McGee]

Just finished a little experiment with todays 4.4.4 build and the gapps extracted from 4.4R1 and 4.4R2 isos

Both gapps act funky and have issues (google keyboard inop, must skip through setup and turn it off) maps FC)

.. but if you use the gapps included in 4.4r1, they UPDATE, and everything starts working well, eventually.
(still have to work your way through turning off google keyboard until updated etc, bt works after updates and a reboot)

Using the gapps included on 4.4R2 they mostly update, but one of the core ones will not update(provisioning?), and neither will maps.

The installs using the R2 included gapps also hang randomly.

[Greg McGee]

Will 64 bit userspace builds help anything?

A 64 bit kernel allows full memory space, and one can install a 64 bit Linux in a chroot, but I can't think of too many Android apps that can really need/use 64 bit address space.

...Or am I missing something?

[Jason Mayfield-Lewis]

x86 has twice the registers in x86-64 Long mode, hence it actually performs better. Does ART in Lollipop support AMD64? If so it would definitely be worth it (unlike the Nexus 9 reviewers have been making a fuss about, which gains nothing and only has 2GB of RAM!)

[fuzzy7k]

I tried lollipop on an AMD Turion(tm) X2 Ultra Dual-Core Mobile ZM-80 and got the following fatal error.

 F libc    : Fatal signal 4 (SIGILL), code 2, fault addr 0xb56ae795 in tid 1892 (main)

Tried again with 3.18 to see if the SSSE3 kernel patch would fix the problem, but it did not. I'm not sure what else would be causing a SIGILL other than SSSE3, as the chip has SSE3. But, that would mean that the SSSE3 emulation patch to enable houdini would be pointless if the system couldn't boot in the first place. So, I'm lost.

[Mike Ross]

--

It's great to see you back....Just say'n

--
Mike
Android - X86 Forums

[Chih-Wei Huang]

If I remember correctly, Robert, the author of
the SSSE3 emulation patch said he only
emulates the SSSE3 instructions used
by houdini.

Another possible reason is, Robert's patch
was developed on kernel 3.10 before.
It may not be correctly ported to 3.18.
(I just applied it and fixed some minor conflicts)

Hope Robert can give better comment.

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

[fuzzy7k]

Ok, and I'm sure that patch was written for KK, which does boot without SSSE3.

So, that patch probably just needs expanded upon.

[rbg]

lollipop-x86-5.0.2_r1 with Linux Stable Kernel 3.18.5 for TESTING PURPOSES ONLY, available here:

   https://drive.google.com/folderview?id=0B2evBKvIpdYkfkptbGV3UGpNNnpSb2UwNzVXbUUxSV9ES0g1UEZ6bC1vN3BfMF9XZVhZaDg&usp=sharing

On Wednesday, December 31, 2014 at 12:47:44 AM UTC-5, rbg wrote:

>You may also build the 64-bit version: 
>(with 32-bit userspace) 

lollipop-x86-5.0.2_r1 with Kernel 3.18-x64, for TESTING PURPOSES ONLY, available here:

  

   https://drive.google.com/file/d/0B2evBKvIpdYkcm9EOVREejZSRGs/view?usp=sharing

help developers, my system don't work...doesn't help anyone....

identify:

---------

what build your running: lp-x86-5.0.2_r1 with kernel 3.18-x64

your system: dell xyz / asus xyz / etc.

issue: wifi / screen / etc. 

provide logcat / dmesg / bugreport logs if possible....

don't know how...use the android-x86 forums and look it up,

many others have done it....

Help the developers help you....

[Mario Holzinger]

Hi all!

May I ask for some help regarding the 3.18 kernel?

building the kernel out of the box for x86 is working without build errors but when booting on my Microsoft Surface Pro 2 I get a bootloop

bootscreen shows like a normal boot procedure but befor getting to the android boot logo the last line : SysRq emegency remount R/O ==> leads to a reboot.

building the complete source for X86_64 I got building errors like glabi_mapi_tmp.h (Mesa stuff) not found,...

Would be very interesting how you solve this issues

And yes, its a fresh and unmodified source :-D

regards

Mario

[Mario Holzinger]

OK must be my little knowledge about git.

Using gitg under my Ubuntu environment to change back to 3.10 and return to 3.18.2 now it boots (x86 build) :-)

[Mario Holzinger]

great kernel is updated to 3.18.8

it can be built out of the box and first tests seems to be ok

[Forcen89]

Hi Jocky,
do u have any stable version (i mean with working touch, battery, wi-fi, sound and may be sdcard) of android x86 for a1 mini? doesnt metter 4.4 or 5.1 or cyan or something like that.

[Jocky Wilson]

On Monday, April 25, 2016 at 3:21:44 AM UTC+2, Forcen89 wrote:

Hi Jocky,
do u have any stable version (i mean with working touch, battery, wi-fi, sound and may be sdcard) of android x86 for a1 mini? doesnt metter 4.4 or 5.1 or cyan or something like that.

Hi, I'm using the stock 4.4 version which was published together with a dual-boot firmware (bios) by Voyo. It works perfect for me, but I understand some a1 users had issues after they flashed the firmware, like no screen output iirc.
If time allows I want to build a android-x86 or cyan based on the stock 64bit kernel/modules version.

/JockyW

[Forcen89 .]

I have 9v 2a version (on intel z3735d) and this version didn't have dual boot bios. I'm tested Android x86 4.4 rc5 and 5.1 aaaaand as u know there lots of troubles (like Wi-Fi and etc).

25 апр. 2016 г. 10:28 пользователь "Jocky Wilson" <jocky...@gmail.com> написал:

--
You received this message because you are subscribed to a topic in the Google Groups "Android-x86" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/android-x86/RX2qNkRJ0UI/unsubscribe.
To unsubscribe from this group and all its topics, send an email to android-x86...@googlegroups.com.

To post to this group, send email to andro...@googlegroups.com.

Visit this group at https://groups.google.com/group/android-x86.