Android 6.0 boot up in the hardware intel Broadwell Xeon E3 server

150 views
Skip to first unread message

dong geng

unread,
Jun 26, 2016, 8:41:56 PM6/26/16
to Android-x86
when I boot up the Android 6.0 rc1 release in Intel BDW E3 server, the SurfaceFlinger process will crash, in the dmesg, I see some abnormal logs, may be that are errors. could you tell me how to fix it? thank you ver much.

in this dmesg, GMBUS is timeout and fb1 as primary device.

[ 3.092864] [drm] Initialized i915 1.6.0 20151010 for 0000:00:02.0 on minor 1
[ 3.151110] [drm] GMBUS [i915 gmbus vga] timed out, falling back to bit banging on pin 2
[ 3.183893] fbcon: inteldrmfb (fb1) is primary device
[ 3.183893] fbcon: Remapping primary device, fb1, to tty 1-63
[ 3.506879] i915 0000:00:02.0: fb1: inteldrmfb frame buffer device

Dong Geng

unread,
Jun 26, 2016, 10:03:10 PM6/26/16
to Android-x86
I attached the dmesg, could someone help me? thank you very much.
dmesg.txt
logcat.txt

Chih-Wei Huang

unread,
Jun 26, 2016, 10:09:40 PM6/26/16
to Android-x86
Due to some reasons your i915 driver
was loaded to be fb1 instead of fb0
so SurfaceFlinger can't find it:

06-24 00:48:27.753 1928 1928 E SurfaceFlinger: hwcomposer module not found
06-24 00:48:27.753 1928 1928 E SurfaceFlinger: ERROR: failed to open
framebuffer (Invalid argument), aborting

It's probably because you have multiple graphic cards?
You can try to add EXTMOD=i915
to the kernel cmdline to force
i915 to be loaded first.


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

gengdongjiu

unread,
Jun 26, 2016, 11:04:23 PM6/26/16
to andro...@googlegroups.com
Hi Chih-Wei,
   Thank you very much for your reply, I have added this parameter "EXTMOD=i915", it is same as before. it is strange, in the Android system, it  has node "/dev/graphics/fb1", but does not have "/dev/graphics/fb1, 
as you see in the dmesg, it has such log: "mgag200 0000:02:00.0: fb0: mgadrmfb frame buffer device",  so Andorid system should find the fb0.
I have tried the Android -x86 5.1-rc1 in xen server, not have such issue, seems only Android 6.0 has such issue.


--
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/tqUWMK-6m9E/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.
For more options, visit https://groups.google.com/d/optout.

Dong Geng

unread,
Jun 26, 2016, 11:04:23 PM6/26/16
to Android-x86
Hi Chih-Wei,
   Thank you very much for your reply, I have added this parameter "EXTMOD=i915", still have such issue. it is strange, in the Android system, it  has node "/dev/graphics/fb1", but does not have "/dev/graphics/fb1, 
as you see in the dmesg, it has such log: "mgag200 0000:02:00.0: fb0: mgadrmfb frame buffer device",  so Andorid system should find the fb0.
I have tried the Android -x86 5.1-rc1 in xen server, not have such issue, seems only Android 6.0 has such issue.
what is your suggestion for that? thank you a lot.

Chih-Wei Huang

unread,
Jun 26, 2016, 11:26:05 PM6/26/16
to Android-x86
2016-06-27 10:39 GMT+08:00 Dong Geng <gengd...@gmail.com>:
> Hi Chih-Wei,
> Thank you very much for your reply, I have added this parameter
> "EXTMOD=i915", still have such issue. it is strange, in the Android system,

Sorry I replied too quickly.
No, EXTMOD won't change i915's loading order.

> it has node "/dev/graphics/fb1", but does not have "/dev/graphics/fb1,
> as you see in the dmesg, it has such log: "mgag200 0000:02:00.0: fb0:
> mgadrmfb frame buffer device", so Andorid system should find the fb0.
> I have tried the Android -x86 5.1-rc1 in xen server, not have such issue,
> seems only Android 6.0 has such issue.

I think it's nothing to do with android version
but the kernel (4.0 -> 4.4).

You can try the debug mode, at the first shell, type
modprobe i915, then continue the booting.

If this works, the simplest fix is just remove
the non-working mgag200 driver:
(if you have install /system to be read-write)

rm -f `find /system/lib/modules -name mgag200.ko`

gengdongjiu

unread,
Jun 27, 2016, 12:42:58 AM6/27/16
to andro...@googlegroups.com
Hi Chih-Wei,
   Thank you very much for your reply and suggestion, now it can work now after remove the mgag200.ko

Reply all
Reply to author
Forward
0 new messages