Jump to content
PaulOBrien

Advent Vega kernel source code now available!

Recommended Posts

Not tried them yet. We were using, as far as i know, blobs extracted, probably from the picasso ROM... The ones in the ICS image have a rotation bug (in fact, seems they don't properly implement rotation of big textures in opengles, and that is causing the rotation bug)... With a bit of luck, the new ICS image will fix that opengles bug and will be usable on JB...

I have extracted them and there are available on my github https://github.com/D...iba_betelgeuse. The rotation problem does not exist in them and everything else also seems to work fine...

The rotation problem is fixed, but video decoding is broken:

E/OMXCodec( 268): failed to allocate node OMX.Nvidia.h264.decode

I/OMXCodec( 268): [OMX.google.h264.decoder] AVC profile = 100 (High), level = 31

I/OMXCodec( 268): [OMX.google.h264.decoder] video dimensions are 320 x 240

I/OMXCodec( 268): [OMX.google.h264.decoder] Crop rect is 320 x 240 @ (0, 0)

E/SoftAVC ( 268): Decoder failed: -2

E/OMXCodec( 268): [OMX.google.h264.decoder] ERROR(0x80001001, -1007)

Edited by DerArtem

Share this post


Link to post
Share on other sites

I have extracted them and there are available on my github https://github.com/D...iba_betelgeuse. The rotation problem does not exist in them and everything else also seems to work fine...

The rotation problem is fixed, but video decoding is broken:

E/OMXCodec( 268): failed to allocate node OMX.Nvidia.h264.decode

I/OMXCodec( 268): [OMX.google.h264.decoder] AVC profile = 100 (High), level = 31

I/OMXCodec( 268): [OMX.google.h264.decoder] video dimensions are 320 x 240

I/OMXCodec( 268): [OMX.google.h264.decoder] Crop rect is 320 x 240 @ (0, 0)

E/SoftAVC ( 268): Decoder failed: -2

E/OMXCodec( 268): [OMX.google.h264.decoder] ERROR(0x80001001, -1007)

Maybe another function renamed in libmedia ? ... I've already posted an script that is able to figure out if some function imported from nvidia libs is missing... I will test those libs myself when i get home, later ;)

Share this post


Link to post
Share on other sites

Maybe another function renamed in libmedia ? ... I've already posted an script that is able to figure out if some function imported from nvidia libs is missing... I will test those libs myself when i get home, later ;)

The libs are in their own branch now: https://github.com/D.../ics-21.05.2012

Maybe I have also missed to copy something...

Edited by DerArtem

Share this post


Link to post
Share on other sites

The libs are in their own branch now: https://github.com/D.../ics-21.05.2012

Maybe I have also missed to copy something...

Using the attached patch to the framework, the video and audio acceleration using those libraries from the latest ICS image for Ventana will work... Yes, libnvomx.so was also using another function that was renamed from ICS to JB.

Those latest nvidia propietary libs work pretty well :)

Just for reference, the extracted libs: http://www.mediafire...gi5qi08v7xguu55

edit1: I found out that at line 84, in nvidia.mk, there are trailing spaces that cause errors. No trailing spaces are allowed in that file ;)

frameworks.rar

Edited by ejtagle

Share this post


Link to post
Share on other sites

Using the attached patch to the framework, the video and audio acceleration using those libraries from the latest ICS image for Ventana will work... Yes, libnvomx.so was also using another function that was renamed from ICS to JB.

Those latest nvidia propietary libs work pretty well :)

Just for reference, the extracted libs: http://www.mediafire...gi5qi08v7xguu55

edit1: I found out that at line 84, in nvidia.mk, there are trailing spaces that cause errors. No trailing spaces are allowed in that file ;)

Updated these files and using a new build with these. What are the expected improvements with these new libs in normal tablet use?

I noticed Google Search has updated to Google Now and voice/sound returned from searches is very clear.

Share this post


Link to post
Share on other sites

Posting the latest RIL... Voice seems to work,,,

- Outgoing call ends abruptly.

- Incoming call does not show in screen/ui. I can see it in logcat.

- no 3g data. connection ended after successful auth errors in logs.

Cannot check if there is voice or not due to above. It seems this version is not hanging the modem/ril. Logs attached.

308ril-full.txt

308ril-radio.txt

getprop.txt

Share this post


Link to post
Share on other sites

Using the attached patch to the framework, the video and audio acceleration using those libraries from the latest ICS image for Ventana will work... Yes, libnvomx.so was also using another function that was renamed from ICS to JB.

Those latest nvidia propietary libs work pretty well :)

Just for reference, the extracted libs: http://www.mediafire...gi5qi08v7xguu55

edit1: I found out that at line 84, in nvidia.mk, there are trailing spaces that cause errors. No trailing spaces are allowed in that file ;)

No I am again confused. The libs you have uploaded - are this the libs I have extracted from ventana? Have you modyfied something or just rar'ed and uploaded them?

Share this post


Link to post
Share on other sites

No I am again confused. The libs you have uploaded - are this the libs I have extracted from ventana? Have you modyfied something or just rar'ed and uploaded them?

Yes, the uploaded files are the ones extracted from the latest ventana ics image, as provided by nvidia...

The patches to the framework are the required ones to make those ICS libraries work in JB

Share this post


Link to post
Share on other sites

Attached my current version of RIL. Don't bother testing it, I will test it in a few hours. I am just posting just to keep everyone updated.

All the problems on RIL not responding, 3G modem not responding after receiving a voice call or after placing a voice call, are caused by our attempt to support voice calls. Huawei modem have probed to be very picky on accessing voice data. Timing is critical, exact order of reads and writes is critical, the exact amount of data to read and write is critical... Even, if you read or write voice data at improper times, the modem does not return anything, but, also, manages to block the linux thread accessing it.

That is why it has taken so long to make it work. Add to all of that that android itself has very different requirements on audio data compared to the Huawei modem, forcing us to implement a selfadapting resampler to match both... But, i guess we have reached what it can be done on voice support. Every attempt to improve it or implement more clever algorithms have resulted in worse audio quality... I think i have tried more than 100 variations, and the one posted now is the one that works best.

Regarding data connections, it is a ip-up permission thing the problem. Keep in mind that pppd is not executed as root... I haven't investigated why the pppd daemon does not execute the ip-up script... If anyone can find the reason, please do it, as i don't have a means to test it myself..

huawei-generic-aug30-v2.rar

Share this post


Link to post
Share on other sites

- Outgoing call ends abruptly.

- Incoming call does not show in screen/ui. I can see it in logcat.

- no 3g data. connection ended after successful auth errors in logs.

Cannot check if there is voice or not due to above. It seems this version is not hanging the modem/ril. Logs attached.

Those logs seem not to be from the latest huawei ril.. To test new rils, what i did was to modify build.prop so it looks for llibhuaweigeneric-ril.so in /system/vendor/lib. And i adb push the new ril there, then do a stop ril-daemon and then a start ril-daemon and i can try a new rill without too much hassle...

Share this post


Link to post
Share on other sites

Latest RIL version

-Voice working. Does not lock the RIL

Regarding data connection,

/etc/ppp/ip-up does need to be chmod 0755

And must be accessible by the 'radio' user....

For some reason, and I cannot figure out why, I cannot get the results you seem to be getting since the last 2 days. Ditto, with the latest version. Maybe have got some files mixed up.

Maybe Scanno can try to update his github with your latest changes and I can sync it from there. Sorry. :-((

Share this post


Link to post
Share on other sites

Hello,

are you on kernel rel-15r7 ?

I have ported it to the folio but I am getting I2C transfer problems. Looks like a clocking problem.

Did you had to adjust any clocking things or some pinmux stuff to get it working?

Share this post


Link to post
Share on other sites
Hello,

are you on kernel rel-15r7 ?

I have ported it to the folio but I am getting I2C transfer problems. Looks like a clocking problem.

Did you had to adjust any clocking things or some pinmux stuff to get it working?

Have a look at my github: https://github.com/scanno/rel-15r7

That is the kernel that we are using now in VegaBean it includes all the changes needed to make it work on the P10AN01. Use the jb branch.

Share this post


Link to post
Share on other sites

Have a look at my github: https://github.com/scanno/rel-15r7

That is the kernel that we are using now in VegaBean it includes all the changes needed to make it work on the P10AN01. Use the jb branch.

Yes, I have seen that. But it has no update histrory to compare the changes to the previous kernel.

Did you had to change something in the pinmux or the clock from the previous 3.1 kernel?

The "old" 3.1 kernel (android-tegra-nv-3.1) is working just fine on the Folio 100. But with rel-15r7 I have some kind of I2C problems and also the device reboots on it's own from time to time.

I think that it is a clocking or pinmux problem, but I already tried out a lot of different stuff with no luck...

Share this post


Link to post
Share on other sites

Yes, I have seen that. But it has no update histrory to compare the changes to the previous kernel.

Did you had to change something in the pinmux or the clock from the previous 3.1 kernel?

The "old" 3.1 kernel (android-tegra-nv-3.1) is working just fine on the Folio 100. But with rel-15r7 I have some kind of I2C problems and also the device reboots on it's own from time to time.

I think that it is a clocking or pinmux problem, but I already tried out a lot of different stuff with no luck...

Eduardo made changes to i2c if i recall. You could do a diff between the original rel-15r7 sources and the one from my github. You can see the changes. You could also read back in the forum because Eduardo has posted his patches to the rel-15r7 here. My github just contains the rel-15r7 kernel and his changes applied to them.

I guess it is best to have a look at Eduardo's latest complete patchset to the rel-15r7 sources.

Edited by Scanno

Share this post


Link to post
Share on other sites

Hello,

are you on kernel rel-15r7 ?

I have ported it to the folio but I am getting I2C transfer problems. Looks like a clocking problem.

Did you had to adjust any clocking things or some pinmux stuff to get it working?

No, i did not touch pinmux settings. But, i have touched clock settings. In the rel15r7 the less you touch the clock initialization tables, the better. Folio100 is based on harmony. The exact same clock table that i used for p10an01 should work for you... The idea in rel15r7 is that when you specify an initial clock table, you are forcing that setting and restricting the kernel ability to change it. That causes severe problems on i2c and audio drivers...

Also, take a look at the i2c tegra controller initialization structures (tegra_i2c_platform_data).. Compare with harmony and you will see that they added some gpio data to it.. You need to add that to the structs...

And, if Folio100 is based on ventana, I'd suggest to look at my initial port of rel15r7 to p10an10 (aka Advent vega 2), posted on this forum, but in its own thread... It will give you a very good starting point... ;)

Share this post


Link to post
Share on other sites

Latest RIL version

-Voice working. Does not lock the RIL

Regarding data connection,

/etc/ppp/ip-up does need to be chmod 0755

And must be accessible by the 'radio' user....

Tried again with the latest RIL (31-8), etc.rar, framework.rar provided recently. Please see attached logs. Unable to detect any change in almost all the recently reported 3g data/voice call parameters. Also, unable to replicate the report on tabletroms jellybean thread stating there was voice on a received phone call.

Is it due to different hardware that we may all be having in our vega-alternate tablets that is giving different results to each one of us?

109-full.txt

109-radio.txt

getprop.txt

Edited by univall

Share this post


Link to post
Share on other sites

Hello everybody!

Dont want to waste your valuable time, my entry is only to be sure You all know that there is an ICS test rom out from LG ( on XDA in their thread related to the OX2 ) which appoints to the SU660 ( korean version of OX2 ), there is also the full KDZ link, very huge, so I wouldn't post the link to waste your time.

Regards.

Alessandro

Edited by Alexzandro

Share this post


Link to post
Share on other sites

Just tried to compile your ril lib in Cyanogemod 10:

target thumb C: libhuaweigeneric-ril <= hardware/huawei-generic-ril/requestdatahandler.c

Export includes file: hardware/huawei-generic-ril/Android.mk -- out/target/product/betelgeuse/obj/SHARED_LIBRARIES/libhuaweigeneric-ril_intermediates/export_includes

In file included from hardware/huawei-generic-ril/requestdatahandler.c:105:0:

hardware/ril/libril/ril_commands.h:65:59: error: 'responseStringsNetworks' undeclared here (not in a function)

hardware/ril/libril/ril_commands.h:65:5: warning: missing initializer [-Wmissing-field-initializers]

hardware/ril/libril/ril_commands.h:65:5: warning: (near initialization for 's_commandInfo[48].responseFunction') [-Wmissing-field-initializers]

Edited by DerArtem

Share this post


Link to post
Share on other sites

Hello Everyone

There is jellybean official rom which support external huawei 3G dongle . please check this . maybe is that possible to use libs

Download

you can find other version here

Also some device like "Pipu U1" have the same ability to use huawei external 3g modem in jellybean,

Edited by arya farjand

Share this post


Link to post
Share on other sites

Just tried to compile your ril lib in Cyanogemod 10:

target thumb C: libhuaweigeneric-ril <= hardware/huawei-generic-ril/requestdatahandler.c

Export includes file: hardware/huawei-generic-ril/Android.mk -- out/target/product/betelgeuse/obj/SHARED_LIBRARIES/libhuaweigeneric-ril_intermediates/export_includes

In file included from hardware/huawei-generic-ril/requestdatahandler.c:105:0:

hardware/ril/libril/ril_commands.h:65:59: error: 'responseStringsNetworks' undeclared here (not in a function)

hardware/ril/libril/ril_commands.h:65:5: warning: missing initializer [-Wmissing-field-initializers]

hardware/ril/libril/ril_commands.h:65:5: warning: (near initialization for 's_commandInfo[48].responseFunction') [-Wmissing-field-initializers]

I have seen this error before, and it makes me believe that either CM10 is using an older RIL or a newer RIL compared to JB AOSP.. ...

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×

Important Information

By using this site, you agree to our Terms of Use.