Firmware mess?

208 views
Skip to first unread message

fuzzy7k

unread,
Feb 12, 2011, 3:03:36 PM2/12/11
to Android-x86
I have some questions about how firmware is loaded with android, and
would ultimately like to ask, "is there a better way"? I've got two
devices that require firmware radeon.ko and b43.ko. All newer radeon
chips require appropriate firmware and let's face it, there are a lot
of wireless chips that require firmware. So, I would venture to say
that this issue is a holdout for a good number of devices. Now, these
drivers work on earlier kernels in a normal linux environment. I'm not
trying to offend anyone here, I know this is not a normal environment.
Without further adieu, lets see if we can't improve this.

I've seen Chih-Wei state that request_firmware is handled by init, but
init is not run until after "exit twice" in debug mode(i.e. after the
modules have loaded). Therefore, the best chance of getting it to work
is running in normal mode to minimize the chance of timeout. This is a
bit of a headache in gingerbread as there is no terminal in normal
mode. (If someone knows how to enable it in normal mode, please
share.)

What role does init have in request_firmware? If the modules are
loaded by the kernel, shouldn't it be responsible?

I noticed in debug mode that there is no /lib/firmware link like there
is for /lib/modules. Adding it didn't help, which stands to reason if
init is responsible.

Is there a way to set things up so that the kernel can reliably get
it's own firmware? Or, is there a way to prevent/delay the loading of
modules that require firmware?

That's it, I know it was kind of scatterbrain but if you can comment
on a part of it or all of it, please do.

-Kyle

Mushishi

unread,
Feb 12, 2011, 7:40:19 PM2/12/11
to andro...@googlegroups.com
Well ALT + F1 will give you consol access in normal boot and ALT+F7 will bring you back to the android gui just like on normal linux.

Chih-Wei Huang

unread,
Feb 13, 2011, 2:18:33 AM2/13/11
to andro...@googlegroups.com
I didn't get the point you talked.
There is no problem to load firmware in android-x86,
either in froyo-x86 or gingerbread-x86.
The firmware should be put into /system/lib/firmware.

In debug mode you have to enter init in 60sec,
or request_firmware will timeout.
Remember debug mode is only for debugging purpose.
I don't consider it is a bug.

2011/2/13 fuzzy7k <kva...@gmail.com>:

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

fuzzy7k

unread,
Feb 13, 2011, 12:25:11 PM2/13/11
to Android-x86
Maybe it's just my stuff then. I struggle to understand why this stuff
works in Fedora/Gentoo, but not in android. The firmware is there.
I've taken it from my running/working gentoo install and placed in
exactly where you say. I search in these discussions and on the issues
page and find others with firmware loading issues but no solutions
that work. What I'm driving at is, where is the disconnect? If the
device doesn't work, "it's not supported". I'm trying to figure out
where the support is missing. It's not on the stock linux kernel. Is
it the user space or is it modifications that have been made to the
android kernel? Has anyone been able to load radeon ucode? I assumed
that my b43 device was a firmware issue due to the dmesg in froyo, but
the switch from froyo to gingerbread seems to have improved the
situation for this device. It's still not working though, so I'll have
to dig further to see why not. I'm going to try some more stuff today
and report back if I uncover anything new.

> Well ALT + F1 will give you consol access in normal boot and ALT+F7 will
> bring you back to the android gui just like on normal linux.

not in gingerbread. Like Chih-Wei say's, it's not a bug, it's
obviously by design. But, when you have to use normal mode for
something to work, and it doesn't, it make troubleshooting a headache
(e.g. I can't look at the dmesg in normal mode.) I was just hoping
someone knew of a quick and dirty way to re-enable it on my local
build.



On Feb 13, 2:18 am, Chih-Wei Huang <cwhu...@android-x86.org> wrote:
> I didn't get the point you talked.
> There is no problem to load firmware in android-x86,
> either in froyo-x86 or gingerbread-x86.
> The firmware should be put into /system/lib/firmware.
>
> In debug mode you have to enter init in 60sec,
> or request_firmware will timeout.
> Remember debug mode is only for debugging purpose.
> I don't consider it is a bug.
>
> 2011/2/13 fuzzy7k <kvan...@gmail.com>:

Arobase40

unread,
Feb 13, 2011, 6:03:02 PM2/13/11
to Android-x86
Like Chih-Wei , I don't understand what is your real problem. I built
an Android froyo version with Radeon module on for my MSI X600 based
on Radeon HD 4330 and I had no problem with firmware.

I can' talk about b43 or bc43 legacy modules or STA version as I don't
use them yet...

May be you should look at a Ubuntu functionnal distribution (or any
other one which has this firmware) to see what and where the firmware
is ?
I guess you'd simply need to copy it from a Linux disro to the
appropriate Android source code directory. Somewhere in /kernel/
firmware...

At building time, this firmware should be place in the correct
place.... ????

As I will have to create a specific Android build with one of these
drivers later on, I guess I 'll investigate this way...

Chac

Chih-Wei Huang

unread,
Feb 13, 2011, 8:18:50 PM2/13/11
to andro...@googlegroups.com
2011/2/14 fuzzy7k <kva...@gmail.com>:

>> Well ALT + F1 will give you consol access in normal boot and ALT+F7 will
>> bring you back to the android gui just like on normal linux.
>
> not in gingerbread. Like Chih-Wei say's, it's not a bug, it's
> obviously by design. But, when you have to use normal mode for
> something to work, and it doesn't, it make troubleshooting a headache
> (e.g. I can't look at the dmesg in normal mode.) I was just hoping
> someone knew of a quick and dirty way to re-enable it on my local
> build.

Sorry, what not work did you say?
Alt - F1 or dmesg?
All must work in gingerbread-x86,
either normal or debug mode.
If not, your build is broken.

Now I doubt what source tree you're using.
Please state clearly how you get the source,
which branch you use and any modification
you've done.

fuzzy7k

unread,
Feb 13, 2011, 8:36:59 PM2/13/11
to Android-x86
Chac, you give me hope, that is hope that I am just doing something
wrong. Did you build your kernel with modesetting enabled by default,
and did dmesg indicate that the microcode loaded properly? I don't
have any problem if I don't use modesetting by default and if it is
enabled then KMS will work, but acceleration is disabled and
everything is extremely slow at the high resolution.

> May be you should look at a Ubuntu functionnal distribution (or any
> other one which has this firmware) to see what and where the firmware
> is ?

The standard place for firmware is /lib/firmware. I have placed it in
out/system/lib/firmware as per instructions.

> I guess you'd simply need to copy it from a Linux disro to the
> appropriate Android source code directory. Somewhere in /kernel/
> firmware...

I don't believe the build process looks at firmware when building, at
lease this wasn't necessary on Gentoo. I chose gentoo because one,
that is what I use, and two, it allows me to use the stock linux
kernel from git. I am using vanilla-sources-2.6.37.

Chih-Whei, after you mentioned about timeout I remember something
about this in the kernel documentation. I changed the timeout to 600s
to make sure I wasn't missing my window. What I found is that the boot
process hangs while trying to load the firmware. It doesn't continue
until the timeout has passed. This means that init never gets a
chance. Plugging and prodding I found /sys/class/firmware/radeon_cl.0/
loading was "0" and as soon as I echo 1 > loading, KMS would
immediately kick in with dmesg showing failure to load firmware. I
don't know if this is a driver bug, or the kernel claiming that this
video module is too important to continue without. I'm not sure what
triggers the init, but if it is not triggered until all kernel modules
are loaded then this would present a definite problem. I could not
test increasing the timeout in normal mode.

In Gentoo, I realize that KMS is not enabled until after init starts
( by init I mean * Starting ... [ok]"). I see there is no break in
"drm" section of dmesg over there either. Also, I am not using any
ramdisk.

snippet of the dmesg for android
...
[ 398.480975] [drm] Loading RS780 Microcode
[ 600.860040] INFO: task modprobe:989 blocked for more than 120
seconds.
[ 600.860101] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs"
disables this message.
[ 600.860171] modprobe D f5c05bc0 0 989 950
0x00000000
[ 600.860315] f58f1ccc 00000082 c12e9200 f5c05bc0 00000000 f58f1c40
c1026fc2 c1483bc0
[ 600.860703] c1483bc0 c1044944 f58e1b20 f5884960 f58e1d98 f5d05bc0
00000001 f58f1cdc
[ 600.861089] f58e1b20 f58f1c94 c12deae7 f58e1d9c f58e1d98 c14b8a40
f58f1c74 c1030b77
[ 600.861476] Call Trace:
[ 600.861539] [<c1026fc2>] ? check_preempt_curr+0x2c/0x69
[ 600.861602] [<c1044944>] ? wq_worker_waking_up+0xb/0x2f
[ 600.861664] [<c12deae7>] ? schedule_timeout+0x18/0x169
[ 600.861725] [<c1030b77>] ? wake_up_process+0xf/0x11
[ 600.861784] [<c10422a8>] ? wake_up_worker+0x19/0x1b
[ 600.861843] [<c1042318>] ? insert_work+0x6e/0x73
[ 600.861904] [<c102802d>] ? __wake_up+0x31/0x3b
[ 600.861962] [<c12deae7>] schedule_timeout+0x18/0x169
[ 600.862026] [<c12dde86>] wait_for_common+0x98/0xee
[ 600.862085] [<c1030b4f>] ? default_wake_function+0x0/0xd
[ 600.862146] [<c12ddf5d>] wait_for_completion+0x12/0x14
[ 600.862207] [<c11d52ce>] _request_firmware+0x221/0x2c8
[ 600.862270] [<c11d53e1>] request_firmware+0xc/0xf
[ 600.862414] [<f83ecd35>] r600_init_microcode+0x1c5/0x2d7 [radeon]
[ 600.862582] [<f83ef202>] r600_startup+0x36/0x8dc [radeon]
[ 600.862742] [<f83bb8c4>] ? radeon_gart_table_vram_alloc+0x30/0x38
[radeon]
[ 600.862909] [<f83efd16>] r600_init+0x1ac/0x232 [radeon]
[ 600.863063] [<f83a9b5f>] radeon_device_init+0x337/0x3ad [radeon]
[ 600.863217] [<f83a91df>] ? radeon_vga_set_decode+0x0/0x20 [radeon]
[ 600.863375] [<f83aae19>] radeon_driver_load_kms+0xb4/0xff [radeon]
[ 600.863484] [<f82526b2>] drm_get_pci_dev+0x126/0x208 [drm]
[ 600.863613] [<f8414a7a>] radeon_pci_probe+0x7b/0x191 [radeon]
[ 600.863709] [<c1162188>] local_pci_probe+0xe/0x10
[ 600.863768] [<c11627e8>] pci_device_probe+0x43/0x66
[ 600.863829] [<c11cfebb>] driver_probe_device+0x81/0x115
[ 600.863890] [<c11cff92>] __driver_attach+0x43/0x5f
[ 600.863949] [<c11cf8b0>] bus_for_each_dev+0x3d/0x67
[ 600.864009] [<c11cfd73>] driver_attach+0x14/0x16
[ 600.864068] [<c11cff4f>] ? __driver_attach+0x0/0x5f
[ 600.864128] [<c11cf30f>] bus_add_driver+0xa2/0x1d4
[ 600.864188] [<c11d01d7>] driver_register+0x8b/0xeb
[ 600.864247] [<c11629bb>] __pci_register_driver+0x38/0x91
[ 600.864319] [<f82527c4>] drm_pci_init+0x30/0x88 [drm]
[ 600.864388] [<f824d3e3>] drm_init+0x4a/0x4f [drm]
[ 600.864510] [<f84430ac>] radeon_init+0xac/0xae [radeon]
[ 600.864604] [<c100115e>] do_one_initcall+0x71/0x113
[ 600.864724] [<f8443000>] ? radeon_init+0x0/0xae [radeon]
[ 600.864821] [<c10596a4>] sys_init_module+0x77/0x197
[ 600.864881] [<c10028d0>] sysenter_do_call+0x12/0x26
[ 1042.400631] r600_cp: Failed to load firmware "radeon/RS780_me.bin"
[ 1042.400708] [drm:r600_startup] *ERROR* Failed to load firmware!
[ 1042.400771] radeon 0000:01:05.0: disabling GPU acceleration
[ 1042.401856] radeon 0000:01:05.0: f5881200 unpin not necessary
[ 1042.401952] radeon 0000:01:05.0: f5881200 unpin not necessary
...

Chih-Wei Huang

unread,
Feb 14, 2011, 5:37:45 AM2/14/11
to andro...@googlegroups.com
OK, I just noticed some firmwares
are loaded by request_firmware, while
some are loaded by request_firmware_nowait.
The later can be loaded asynchronously,
but the prior can't.

That means we can't always reply on android init
to load firmware in the bootstrap script.

Let me think how to solve that.

fuzzy7k

unread,
Feb 14, 2011, 7:42:41 PM2/14/11
to Android-x86
mm, I missed this yesterday....

> Sorry, what not work did you say?
> Alt - F1 or dmesg?

Alt - F1 switches to vt1, but there is no prompt, thus no way to run
dmesg. Or, is there an app for that? (jk, although I don't doubt that
there is one.) vt2 - 6 are disabled. Also, in debug mode I have no
prompt on vt1, only vt 2 & 3. Maybe I do have a broken build, in which
case it should clear itself up next time I wipe my tree and resync.

> Now I doubt what source tree you're using.
> Please state clearly how you get the source,

Just to clarify, I am using the genuine gingerbread-x86 branch, kernel
and all. In the base directory I have several hidden files
(e.g. .buildspec, .firmware.tar.gz, .o -> out/... ). This includes a
script that: rm -rf ./* the base directory, then repo sync, then make
-j5 usb_img. From there I copy the kernel, initrd.img, ramdisk.img, &
system.sfs to an android folder and boot as per instructions in the
advanced section of the "Get Source" page. I first run without any of
my own modifications, including to the kernel config. This gives me a
chance to see what works and what doesn't "out of box". Then I edit
the kernel config, run, then add firmware, run ....

> OK, I just noticed some firmwares
> are loaded by request_firmware, while
> some are loaded by request_firmware_nowait.

Sounds like you've pinned down the issue. I don't know how other
distro's do it but just to kick some things your way, I believe fedora
includes modules & firmware in there initrd, as they start plymouth
( graphical boot ) very early. On the other end of the spectrum, I do
not use any ramdisk on my gentoo install with the stock linux kernel.



On Feb 13, 8:18 pm, Chih-Wei Huang <cwhu...@android-x86.org> wrote:
> 2011/2/14 fuzzy7k <kvan...@gmail.com>:

Chih-Wei Huang

unread,
Feb 14, 2011, 8:29:09 PM2/14/11
to andro...@googlegroups.com
2011/2/15 fuzzy7k <kva...@gmail.com>:

> Alt - F1 switches to vt1, but there is no prompt, thus no way to run
> dmesg. Or, is there an app for that? (jk, although I don't doubt that
> there is one.) vt2 - 6 are disabled. Also, in debug mode I have no
> prompt on vt1, only vt 2 & 3. Maybe I do have a broken build, in which
> case it should clear itself up next time I wipe my tree and resync.

The console shell is started by the
service in init.rc:

service console /system/bin/sh
console
disabled
# user shell
group log

Check if you have /system/bin/sh
and it is executable.

Except the console, you may also run the
terminal emulator from Dev Tools -> Terminal Emulator
(it needs /system/bin/sh, too)

Ikem Krueger

unread,
Feb 16, 2011, 5:05:03 AM2/16/11
to andro...@googlegroups.com
> 2011/2/14 Arobase40 <arob...@gmail.com>:

> I built an Android froyo version with Radeon module

> for my MSI X600 based on Radeon HD 4330
> and I had no problem with firmware.

Maybe you upload them, so he can download them?

--
LXDE is looking for a new maintainer of LXPanel. So if you are a
developer, get in contact with us. :)

fuzzy7k

unread,
Feb 23, 2011, 10:45:54 PM2/23/11
to Android-x86
Chih-Wei, I figured it out!

mdev can do it, there are just two small pieces to the puzzle out of
place.

CONFIG_UEVENT_HELPER_PATH="/sbin/mdev"

and

/lib/firmware

more details here http://code.google.com/p/android-x86/issues/detail?id=183


On Feb 16, 5:05 am, Ikem Krueger <ikem.krue...@googlemail.com> wrote:
> > 2011/2/14 Arobase40 <arobas...@gmail.com>:

Arobase40

unread,
Feb 23, 2011, 11:53:21 PM2/23/11
to Android-x86
Can you at least give a try to my first Androix-x86 beta for the Acer
1825PTZ, as it was generic enough to support my MSI-X600 with Radeon
4330 (with kms enabled) ?

I would like to understand the difference with what you describe
here... ^^

Here is the link again :

http://www.google.com/url?sa=D&q=http://rapidshare.com/files/446220918/Acer-1825ptz_laptop.iso

Thx

fuzzy7k

unread,
Feb 24, 2011, 12:17:52 PM2/24/11
to Android-x86
I tried it, but it wouldn't boot, kept getting a kernel panic flashing
caps and all. I tried a lot of different things, including pulling the
kernel, initrd.img, ramdisk.img, & system.sfs and booting them just
like I do with my builds, only changing generic_x86 to asus_laptop in
the grub line.

You say KMS worked, but what about the microcode? KMS would work for
me, but it wouldn't load the microcode.

On Feb 23, 11:53 pm, Arobase40 <arobas...@gmail.com> wrote:
> Can you at least give a try to my first Androix-x86 beta for the Acer
> 1825PTZ, as it was generic enough to support my MSI-X600 with Radeon
> 4330 (with kms enabled) ?
>
> I would like to understand the difference with what you describe
> here... ^^
>
> Here is the link again :
>
> http://www.google.com/url?sa=D&q=http://rapidshare.com/files/44622091...

Arobase40

unread,
Feb 24, 2011, 6:47:11 PM2/24/11
to Android-x86
Weird... it works quite fine with my MSI X600 which is compiling at
present time.

I just tested the same build on an Asus M51Vr that just came back from
repair. It also has a ATI/AMD Radeon GPU not as powerfull as yours but
still. It works.
But the Radeon Module is not loaded though...

How do I know about the microcode, I havn't seen anything with dmesg
even on my MSI X600 ???

fuzzy7k

unread,
Feb 24, 2011, 7:13:17 PM2/24/11
to Android-x86
This is what my dmesg looks like

failing:
...
[ 398.480975] [drm] Loading RS780 Microcode
[ 1042.400631] r600_cp: Failed to load firmware "radeon/RS780_me.bin"
[ 1042.400708] [drm:r600_startup] *ERROR* Failed to load firmware!
[ 1042.400771] radeon 0000:01:05.0: disabling GPU acceleration
[ 1042.401856] radeon 0000:01:05.0: f5881200 unpin not necessary
[ 1042.401952] radeon 0000:01:05.0: f5881200 unpin not necessary
...

working:
...
[ 4.985488] [drm] Loading RS780 Microcode
[ 5.011210] radeon 0000:01:05.0: WB enabled
[ 5.042822] [drm] ring test succeeded in 1 usecs
[ 5.042946] [drm] radeon: ib pool ready.
[ 5.043014] [drm] ib test succeeded in 0 usecs
...

Older radeon cards do not require anything other than what is in the
stock kernel. Newer cards do. Currently CONFIG_DRM_RADEON_KMS=n in the
default android config. From the help text, it checking this y
actually builds a different driver.

│ This is a completely new driver. It's only part of the existing
drm
│ for compatibility reasons. It requires an entirely different
graphics
│ stack above it and works very differently from the old drm stack.

fuzzy7k

unread,
Feb 24, 2011, 7:17:15 PM2/24/11
to Android-x86
I'll try going back and using it the old way to see if performance is
any better.

fuzzy7k

unread,
Feb 24, 2011, 8:11:27 PM2/24/11
to Android-x86
Now I remember why I was trying to use the new driver, I can not get
KMS working with the old driver. I append radeon.modeset=1 and the log
still says using userspace modesetting when loading the radeon module.
Also, it's not autoloaded and when I add it to 0-autoload it still
uses uvesafb.

Arobase40

unread,
Feb 24, 2011, 8:38:01 PM2/24/11
to Android-x86
And again, if my memory is not deficient, Chih Wei said in an old post
dated november 2010, that the use of modeset is not needed...

http://groups.google.com/group/android-x86/tree/browse_frm/thread/5fbea680f53ab257/245da81121910a17?_done=%2Fgroup%2Fandroid-x86%2Fbrowse_frm%2Fthread%2F5fbea680f53ab257%2F245da81121910a17%3Ftvc%3D1%26&tvc=1

It was about Intel GPU but I think the logic is just about the same
with Radeon GPU :

Replace : CONFIG_DRM_I915_KMS=y with the equivalent for Radeon. that's
what I did intuitively with my first beta version before I found out
this post, and that 's may be the reason it works fine with my MSI-
X600.

:D

Chih-Wei Huang

unread,
Feb 25, 2011, 3:14:01 AM2/25/11
to andro...@googlegroups.com
2011/2/24 fuzzy7k <kva...@gmail.com>:

> Chih-Wei, I figured it out!
> mdev can do it, there are just two small pieces to the puzzle out of
> place.

Thank you for the hint.
I've pushed a fix in the newinstaller dir/.

> CONFIG_UEVENT_HELPER_PATH="/sbin/mdev"

You don't need this since it is configured at runtime.

fuzzy7k

unread,
Feb 25, 2011, 10:04:52 AM2/25/11
to Android-x86
> Thank you for the hint.
> I've pushed  a fix in the newinstaller dir/.

Glade to help

> > CONFIG_UEVENT_HELPER_PATH="/sbin/mdev"
>
> You don't need this since it is configured at runtime.

I did. Creating a /lib/firmware link at the first debug prompt (before
the drivers are loaded) was the first thing I did a couple weeks ago.
It wasn't until I started doing research on how the other guys do it,
and what the android process is that I discovered mdev. I also saw
some stuff in the mdev source about being invoked as hotplug, so I
tried creating an /sbin/hotplug link along with the /lib/firmware link
and CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug". That doesn't work
either.

I've tried this every which way, I can only get it to work with the
combo I first mentioned. If you've made other changes that you're not
mentioning, let me know and I'll do some more testing once they've
come through. As a thought, I've been seeing major differences between
normal mode and debug mode, (e.g. wacom not loading in normal mode,
but loading in debug mode; meanwhile b43 working in normal mode, but
not working in debug mode. And, the issue with the radeon module,
http://code.google.com/p/android-x86/issues/detail?id=183 ) so, maybe
your statement, CONFIG_UEVENT_HELPER_PATH="/sbin/mdev" is not needed,
is correct during normal mode, but not debug mode...?

Let me know if you need me to do any more testing.

-Kyle

On Feb 25, 3:14 am, Chih-Wei Huang <cwhu...@android-x86.org> wrote:
> 2011/2/24 fuzzy7k <kvan...@gmail.com>:

fuzzy7k

unread,
Feb 25, 2011, 11:24:57 AM2/25/11
to Android-x86
No, still doesn't work in normal mode. I'm guessing the runtime
configuration doesn't kick in until after it is needed by the kernel
to load the firmware with request_firmware .

On Feb 25, 10:04 am, fuzzy7k <kvan...@gmail.com> wrote:
> > Thank you for the hint.
> > I've pushed  a fix in the newinstaller dir/.
>
> Glade to help
>
> > > CONFIG_UEVENT_HELPER_PATH="/sbin/mdev"
>
> > You don't need this since it is configured at runtime.
>
> I did. Creating a /lib/firmware link at the first debug prompt (before
> the drivers are loaded) was the first thing I did a couple weeks ago.
> It wasn't until I started doing research on how the other guys do it,
> and what the android process is that I discovered mdev. I also saw
> some stuff in the mdev source about being invoked as hotplug, so I
> tried creating an /sbin/hotplug link along with the /lib/firmware link
> and CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug". That doesn't work
> either.
>
> I've tried this every which way, I can only get it to work with the
> combo I first mentioned. If you've made other changes that you're not
> mentioning, let me know and I'll do some more testing once they've
> come through. As a thought, I've been seeing major differences between
> normal mode and debug mode, (e.g. wacom not loading in normal mode,
> but loading in debug mode; meanwhile b43 working in normal mode, but
> not working in debug mode. And, the issue with the radeon module,http://code.google.com/p/android-x86/issues/detail?id=183) so, maybe

fuzzy7k

unread,
May 15, 2013, 11:49:08 AM5/15/13
to andro...@googlegroups.com
Ancient thread, the radeon firmware loading issue has been solved for a long time. Sounds like you need to disable the one so that radeon is used, but I have no experience with multi-GPU hardware. Please start a new thread.

On Monday, May 13, 2013 1:30:51 PM UTC-4, ShaHoang wrote:
And how about the driver ? You know that I have a dream Androidx86 will have a source supporting many hardwares for PC like as MacOSx86 done. I installed Androidx86 but it only works in GUI mode with my onboard GPU - GMA950. I don't know how to use the ATI Radeon X1550 (PCIe). So I couldn't use 100% my hardware and made me unsatisfied :(
Reply all
Reply to author
Forward
0 new messages