ATI Radeon Xpress 200M

[Bogdan Adrian Gecse]

Hello! I have a HP Compaq Presario v2000 (v2410us variant with ATI Radeon Xpress 200M video card).

When I try to boot Android-x86 4.4, it hangs after "Detecting Android-x86". The problem looks like to be the video card driver.

My question is: Is ATI Radeon Xpress 200M supported on Android-x86? 

I need to mention that it starts in Vesa mode :)

[walter]

Have you tried the image from this post ?
https://groups.google.com/d/msg/android-x86/84d-brSsUSc/8zesF-lKowgJ

[Bogdan Adrian Gecse]

Yes, but I will try again now.

On Android-x86.org, there are Release notes for 4.4RC2, but it's not available to download... Maybe this new release it's working :D

[Bogdan Adrian Gecse]

Here is another strange thing - when booting from 4.4 RC1 release from website and from the image from that post, it just hangs after "Detecting Android-x86".

While using some "nightly" release, built on 18 May, it goes a little further than that, to the "A N D R O I D" message. After 1-2 seconds, the "A N D R O I D" message goes away, and there remains only a blinking underline. If I choose Vesa mode, it will show the graphical UI.

Also, I can't boot in Vesa mode from 4.4 RC1 release from website and from that image from that post... I don't know how to export logcat, if you tell me how, I will do it :)

[pstglia]

Hi, are you trying to boot from flashdrive? Have you tried to install it in your harddisk and boot?

I think when "Detecting Android x86" is shown, it is trying to determine where Android installion/ kernel image is. After this the answer is "found at /dev/sdX"

I had a similar problem, but in my case I was pointing the wrong path in boot config ( i use lilo ). Fixing the path solved my problem

You can write logcat ouput mounting some hdd partition or flashdrive. In my case, I save to harddisk executing something like this:

mkdir /system/data/foo
busybox mount /dev/block/sda5 /system/data/foo
logcat -f /system/data/foo/output-logcat.txt

[Bogdan Adrian Gecse]

It says "Detecting Android-x86..." then it shows the "found" message and nothing else. This happens with RC2 too. As I said, with a "nightly" downloaded from another website, built on May 18, after the "found" message, it shows the "A N D R O I D" message...

[pstglia]

More info will be needed. Can you see some suggesting error message on logcat or dmesg? Could you post them?

Cheers,

pstglia

[Bogdan Adrian Gecse]

Please tell me where do I need to enter those commands for saving logcat? Start in debug mode? 

[pstglia]

Press control+alt+f1 to enter console mode. Then type the comands to save logcat (change device name according to your environment. In my case, sda5 is a partition of my hard disk)

You can log using "live cd" mode. 

[Bogdan]

Ok, I read something on forums before, but they said just ALT+F1. Now, when should I press ctrl+alt+f1? After "Detecting android x86"? When loading kernel? Sorry for too many questions, I just don't know too many things about saving logcat :)

[pstglia]

I use to press ctrl+alt+f1 because of Xorg/Linux.... But alt+f1 is enough 

Press it the point your machine hangs. When you get the prompt, type the cmds. 

In case you can't enable the prompt, reboot the machine an try to enter DEBUG mode. Type  "exit" when required until it tries to load android. Then, press ALT+F1 to get prompt. If you have no response, try ALT+F2

If you have prompt, but find difficult mounting your hard disk partition or flash drive (the cmds I posted), try at least to do this and take a note of it (write it down or take a shot):

logcat -d | grep -iE "gralloc|drm|radeon" | tail -40

Cheers,

pstglia

[Bogdan]

I can't get to the prompt with both RC1 and RC2 releases, I just don't know what it's happening... It just shows "Detecting" then "found at", the resolution changes to the display resolution (which is 1280x768), and nothing else. Pressing ALT+F1 or ALT+F2 does nothing. I don't think this problem is related to the video card...

Anyway, I exported a logcat from the "nightly" build, built on 18 May. I attached the file.

[rbg]

try this...

https://drive.google.com/folderview?id=0B2evBKvIpdYkSkVBWkxMSWtZWk0&usp=sharing

[Bogdan]

With this build I can get past the "Detection" phase, to the "A N D R O I D" message, but no graphic UI. The logcat looks the same as the one from the nightly build.

[pstglia]

At some points your log shows this msg:

E/gralloc ( 1298): Unsupported format 8

It's telling that your current depth (bpp) is set to 8, which is not supported by gralloc (see hardware/libhardware/modules/gralloc/framebuffer.cpp, function fb_device_open):

            if (m->info.bits_per_pixel == 32) {
                format = (m->info.red.offset == 16) ? HAL_PIXEL_FORMAT_BGRA_8888
                       : (m->info.red.offset == 24) ? HAL_PIXEL_FORMAT_RGBA_8888
                       : HAL_PIXEL_FORMAT_RGBX_8888;
            } else if (m->info.bits_per_pixel == 16) {
                format = HAL_PIXEL_FORMAT_RGB_565;
            } else {
                ALOGE("Unsupported format %d", m->info.bits_per_pixel);
                return -EINVAL;
            }

As result, device couldn't be set. This probably is the cause your SurfaceFlinger is crashing and you can't complete startup

We have to figure out why this is being set with 8 by default. Can you post your dmesg output?

Try appending this kernel parameter (on grub menu, select DEBUG entry, press "e" to edit, then press e again under kernel /..."):

video=radeonfb:1280x768-16@60

Also, u can try to change your depth manually with fbset. But in debug mode and after the 1st exit cmd issued type this:

fbset -depth 16

Cheers,

pstglia

[Bogdan]

"fbset" is not working in debug mode, it says "/dev/fb0 not found". I will try again, maybe I did something wrong. Anyway, I attached another logcat, from debug mode.

[Bogdan]

I read your message again, and I entered "fbset" after first "exit" command :D

I deleted "vga=788" parameter, because with that, Android won't use maximum resolution, but 800x600 and no Radeon driver. I also added that kernel parameter.

All goes well, after first exit I enter fbset command, exit again, then the screen turns black, no access to prompt on both ALT+F1 and F2. It's like something is infront of console prompt. Maybe the driver tries to load, but another problem appeared, I don't know. I tried to blindly type "mkdir", "busybox mount", "logcat", but no luck.

Why I say "it's like something is infront of console prompt"? I tried ALT+F2 and logcat commands (blindly). Waited a moment, pressing CTRL+C, unmounted second usb flash (which I use for logcat), then I tried typing "reboot" and ENTER - laptop still on, so it was not working. I pressed ALT+F1, then typed reboot there, and laptop restarted.

After this, I booted again, used "fbset" after first exit, and after second exit, when the screen goes black, I tried "mkdir". The screen stays black, but sometimes it switches fast to the console prompt, then back at the black screen. At the console, I can hardly see (because everything happens very fast) the command I entered (mkdir), and an blinking underline under the command, like the command is still executing or something. I can enter other commands, but they don't execute, like "reboot".

[Bogdan]

Sorry for multiple answers. I saved 2 more logcats. One is from debug mode, with "vga=788" switch removed from GRUB - where Android chooses 8 bit graphics.

Because if I set "fbset" to 16, I can't get any logcat because of the problem mentioned above, I tried to set "fbset" to 32. The screen still goes black, but I can see a blinking underline at the bottom of the screen, and if I press enter I get the "root_x86#" prompt.

And a question - how can I export all debug messages? Logcat saves another info, which I can't see, but it does not save all debug messages... (I think i found something, I will try)

[Bogdan]

Here is the dmesg - Debug mode with vga=788 removed.

[pstglia]

Hi again,

could you execute this please? (after 1st exit on debug mode - also remove vga=788 param):

cat /proc/fb

According to your dmesg output, you're using radeonfb and not radeondrmfb. I'd like to confirm it

About logcat, you can query entire log in /data/log.txt

Thanks

[Bogdan]

It shows "0 ATI Radeon 5955".

[pstglia]

Let's chech which video driver is loaded.
Under debug mode (again, remove vga=788) after the 1st exit type this:
lsmod

Check if radeon is listed. If not, do this:
1) reboot your machine
2) enter debug mode, and *before* 1st exit type:
modprobe radeon
3) type lsmod and check if the driver was loaded.
4) if radeon was loaded, execute this and check if radeondrmfb is returned:
cat /proc/fb
5) case radeondrmfb is returned, type exit and continue booting

[Bogdan]

Okay.

Under debug mode, after first exit, radeon appeared in lsmod. I exported lsmod into a .txt.

I did cat /proc/fb and there it was ATI RADEON 5955.

I rebooted, and before first exit, I entered the "modprobe" command. The resolution instantly changed to 1280x768.

When I typed lsmod, there was the same "radeon" but with a different size. I exported this lsmod too.

When I entered cat /proc/fb, there was no more ATI RADEON 5955, but "radeondrmfb".

I typed exit to continue booting, and surprise, Android logo showed up, and it successfully booted. I saved logcat and dmesg if you need.

As a side note, I think that my CX-20468 sound card is not supported :))

[pstglia]

This confirms what I suspected. Your system was using "fb_ddc + radeonfb" instead of radeon + radeondrmfb for graphics. As consequence, drm gralloc (direct rendering/hardware accel) wasn't being enabled in init.sh (it's only enabled when radeondrmfb is in use, which depends on radeon driver).

This avoided displaying graphics

# device/generic/x86/init.sh

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

0*inteldrmfb|0*radeondrmfb) set_property hal.gralloc drm set_drm_mode

To auto load "drivers" (kernel modules), it queries all modules listed in "/sys/bus/*/devices/*/modalias" files and check if there's a equivalent kernel module in /lib/modules/`uname -r`/modules.alias. If found, the driver is loaded (see bootable/newinstaller/initrd/scripts/0-auto-detect).

In your case, or your device is not being listed on "/sys/bus/*/devices/*/modalias" or doesn't match any of the aliases in /lib/modules/`uname -r`/modules.alias.

If you load it manually (as you did) it loads radeondrmfb, enable gralloc drm and works correctly. As workaround, you can create a new ISO forcing radeon module to be loaded.

Suggestion: edit bootable/newinstaller/initrd/init file and add this before the function load_modules:

modprobe radeon

About your sound, try loading "snd-hda-intel" module using the same procedure you used to boot (type modprobe snd-hda-intel before 1st exit)

Cheers,

pstglia

[Chih-Wei Huang]

Probably disabling CONFIG_FB_RADEON would help.
It may be automatically loaded first and take the framebuffer
so radeondrm can't be loaded.

Bogdan, if you know how to build from source,
try to remove CONFIG_FB_RADEON=m
in kernel/arch/x86/configs/android-x86_defconfig
and rebuild it.

> --
> 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.



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

[Bogdan]

Unfortunately, I never built anything from source, and I have only this laptop (1.6GHz single-core, 1GB ram). I guess it will take days to build....

About the sound card, which is Conexant CX-20468, I found something interesting, an ALSA driver package, with "improved support for Conexant chipsets". Maybe can be integrated within the next Android release.

[rbg]

test builds kernel has been updated as follows:

 CONFIG_FB_RADEON=m

changed to: # CONFIG_FB_RADEON is not set

available here: https://drive.google.com/folderview?id=0B2evBKvIpdYkSkVBWkxMSWtZWk0&usp=sharing

[Bogdan]

Yes, I tried the new built and it loads correctly without any intervention from me!

About the sound card, no, snd-hda-intel does not work... I think because it's not an Intel sound card, but a Conexant one. Here is the Alsa driver with "improved support for Conexant chipsets" for Linux which I've found: http://www.linuxant.com/alsa-driver/

marți, 20 mai 2014, 03:40:08 UTC+3, Bogdan a scris:

Hello! I have a HP Compaq Presario v2000 (v2410us variant with ATI Radeon Xpress 200M video card).

When I try to boot Android-x86 4.4, it hangs after "Detecting Android-x86". The problem looks like to be the video card driver.

My question is: Is ATI Radeon Xpress 200M supported on Android-x86? 

I need to mention that it starts in Vesa mode :)

[Bogdan]

Found something interesting. The whole Alsa driver with support for Conexant chipsets is just a patch, no file to integrate into the build or anything. For example, here: http://www.linuxant.com/alsa-driver/alsa-driver-1.0.23-1.patch

[pstglia]

Hi Bogdan,

This weekend I borrowed my friend's notebook. It's a Compaq Pressario M2000, Sempron  Mobile Processor and ATI Express X200 like yours). I think they are almost the same, except for the processor (your machine uses Turion, correct?)

The same issue as yours: no sound.

After some research on web, discovered snd_atiixp (the sound driver for this card) set sound rate to 48000Hz by default. However, all playbacks sent to tinyalsa (a interface between alsa and kernel) set rate to 44100. This produces an error under external/tinyalsa/pcm.c (function pcm_open - ioctl call):

  if (ioctl(pcm->fd, SNDRV_PCM_IOCTL_HW_PARAMS, &params)) {

        oops(pcm, errno, "cannot set hw params");

        goto fail_close;

    }

I included the parameter (ac97_clock=44100) loading this driver and I could have sound working. The only problem I couldn't solve is that it is played a bit faster than expected. Maybe someone else here can help

This probably will work on your machine. If you want to check:

1) Boot under debug mode

2) Before 1st exit, load the module setting the parameter:

modprobe snd_atiixp ac97_clock=44100

3) Continue booting

Hope this helps

Ps: Also had problems with Broadcom wi-fi driver ( wl proprietary driver ). I had to replace it by b43 and blacklisted wl. Also, had to include proprietary fw, as described here:

http://wireless.kernel.org/en/users/Drivers/b43

[pstglia]

I created this debug ISO for my friend (owner of that Pressario I mentioned). In case you want it...

https://drive.google.com/file/d/0BxO6THtB865fUUF6Q3R4YUljQzA/edit?usp=sharing

[Bogdan]

I will try now by typing that before first exit. I have no problems with Wi-Fi driver :)

[Abu Al]

Bonjour How Iso Androidx86 created. Is it possible to create Iso medium for ati x1270 (RS690). I do not think I have the skills to realized the project. Can you help me ? Google translation