[Development][ROM] Stock FROYO + 2.6.32 kernel

[Guest El Nino9]

im trying to look at the ril data...all i have found out probably things that you already know...

.plt:0000CF6C ; Needed Library 'liblog.so'

.plt:0000CF6C ; Needed Library 'libcutils.so'

.plt:0000CF6C ; Needed Library 'libutils.so'

.plt:0000CF6C ; Needed Library 'libril.so'

.plt:0000CF6C ; Needed Library 'liboncrpc.so'

.plt:0000CF6C ; Needed Library 'libcm.so'

.plt:0000CF6C ; Needed Library 'libmmgsdilib.so'

.plt:0000CF6C ; Needed Library 'libgsdi_exp.so'

.plt:0000CF6C ; Needed Library 'libgstk_exp.so'

.plt:0000CF6C ; Needed Library 'libwms.so'

.plt:0000CF6C ; Needed Library 'libnv.so'

.plt:0000CF6C ; Needed Library 'libwmsts.so'

.plt:0000CF6C ; Needed Library 'libril-qcril-hook-oem.so'

.plt:0000CF6C ; Needed Library 'libdss.so'

.plt:0000CF6C ; Needed Library 'libqmi.so'

.plt:0000CF6C ; Needed Library 'libc.so'

.plt:0000CF6C ; Needed Library 'libstdc++.so'

.plt:0000CF6C ; Needed Library 'libm.so'

.plt:0000CF6C ; Shared Name 'libril-qc-1.so'

Thats when i opend libril-qc-q-1.so of the tmobile 2.1 rom using IDA (trial version)

also a question, is the stock rom a port from a CDMA phone?

[Guest pier11]

also a question, is the stock rom a port from a CDMA phone?

nope, u8150 is a GSM phone...

Edited January 22, 2012 by pier11

[Guest dr.flo]

pier11, you asked me to upload my version of this experiment with ascend froyo rom. Here it is: mini_ascend (mini because shrinked, rom, not mini like in pulse mini).

Be warned: during boot, about 5sec of loud noise coming out of the loudspeaker is waiting for you...

I took this rom upload as base. Note that the .32 kernel inside this mini_ascend image is very old now since it was some weeks ago where I composed this rom.

I replaced original launcher2.apk (6 mb) with zeam launcher (400k) to save space. I don't remember exactly what else was needed to get this working, but I know it wasn't much.

@readers: this is not a functional rom! don't download...

[Guest pier11]

pier11, you asked me to upload my version of this experiment with ascend froyo rom...

thanks a lot,

I based my unsuccessfull ascend port on official froyo rom from metroPCS site, as I remember...

[Guest skykilla]

pier11 I took these 2 files from cm6 AudioFilter.csv and AudioPreProcess.csv and have inserted in system\etc The sound has earned

[Guest dr.flo]

pier11 I took these 2 files from cm6 AudioFilter.csv and AudioPreProcess.csv and have inserted in system\etc The sound has earned

Are you saying that audio works on "rom_u8150-FR.01.zip" if you insert these two files? That would be great...

EDIT: confirmed! Great advice! I would expect this to also fix audio in ascend experiment...

Edited February 4, 2012 by dr.flo

[Guest pier11]

pier11 I took these 2 files from cm6 AudioFilter.csv and AudioPreProcess.csv and have inserted in system\etc The sound has earned

wow!

would never thought it could be solved this way. I tried to replace audio libs, but only thing I got is complete silence :)

yet to test myself, but awesome, thanks man!

[Guest dr.flo]

this motivates me to look at that rom again ;)

Audio works. I'll try updating the kernel to our latest version. Maybe video will also work then... Then my cam wrapper will work.

A small tip: changing density from 120 to 160 makes screen resolution look as it should on our display resolution.

[Guest skykilla]

communication to start on 32 kernel it would be healthy

Edited February 4, 2012 by skykilla

[Guest pier11]

this motivates me to look at that rom again ;)

Audio works. I'll try updating the kernel to our latest version. Maybe video will also work then... Then my cam wrapper will work.

I noticed, the proposed files are nearly identical to what is shipped with Magic's official FROYO.

While what is shipped with our Eclair is identical to u8150 Froyo and doesn't work. Strange...

[Guest dr.flo]

side note: the same trick makes sound work on the ascend rom. The noise during boot turns into a metropcs jingle.

Well, this could have been expected, so no breaking news here.

One thing, though: I noticed that with the older (aka non-compat) version of kernel, I initially composed mini_ascend with, the keypad leds are functional. With the current (compat) kernel, they aren't.

This means we could use ascend's led libs if we wanted to revert led driver in compat kernel. As ascend has functional - inofficial - CM7 port, there should even be the sources for it (not just the binary, which would be enough).

Edited February 4, 2012 by dr.flo

[Guest pier11]

are you saying that on ascend rom keypad leds work, but on u8150 they don't? Good find! We want source version of the lib ideally...

[Guest dr.flo]

are you saying that on ascend rom keypad leds work, but on u8150 they don't? Good find! We want source version of the lib ideally...

Exactly. Unfortunately we might have no sources. "Isaacj87" had the ascend device tree, but it was deleted from his github. The ascend device tree in the CM repo isn't complete.

[Guest pier11]

Exactly. Unfortunately we might have no sources. "Isaacj87" had the ascend device tree, but it was deleted from his github. The ascend device tree in the CM repo isn't complete.

this was a fork of isaacj87's android_device_huawei_ascend tree at some point of time:

https://github.com/y...e_huawei_ascend

https://github.com/h...ts/gingerbread/

EDIT:

actually many of the below trees contain isaac's work up until June 12 2011.

https://github.com/y...network/members

Edited February 4, 2012 by pier11

[Guest dr.flo]

this just proves I'm a github noob :)

But in extract_files, I see -> lib/hw/lights.msm7k.so <- which means they didn't rebuild it from source. But, since it is named msm7k and not device specific, the sources might simply be those in hardware/msm7k/liblights

[Guest pier11]

might.

I also noticed in LG KH5200(aka GW620) semi-official froyo rom, that they don't have ANY audio config files in /etc folder. If you are still on ascend rom, could you please try to:

- remove AudioPreProcess.csv and see if there is still sound (after reboot);

- completely remove both files and check the same way...

[Guest dr.flo]

- remove AudioPreProcess.csv and see if there is still sound (after reboot);

- completely remove both files and check the same way...

Sound working in both cases (did reboot; also did ls to verify file was really deleted)

[Guest pier11]

Thanks!

I made the first experiment on cm72 -but just renamed the file. Sound still was there after reboot.

Looks like some sort of equalizer settings. Somehow they got messed in our unusual combination.

Edited February 4, 2012 by pier11

[Guest dr.flo]

Regarding RIL I found stuff like this

## RIL related stuffbin/port-bridgebin/qmuxdlib/libauth.solib/libcm.solib/libdiag.solib/libdll.solib/libdsm.solib/libdss.solib/libgsdi_exp.solib/libgstk_exp.solib/libmmgsdilib.solib/libnv.solib/liboem_rapi.solib/liboncrpc.solib/libpbmlib.solib/libqmi.solib/libqueue.solib/libuim.solib/libril.solib/libril-qc-1.solib/libwms.solib/libwmsts.so

Not our device but all these proprietary files also exist on our eclair rom, and... all these are pulled from the phone and put into CM6 or CM7.

So maybe, when you did a replace *ril* experiment to get RIL working, you just didn't replace enough...

edit: sorry, line breaks disappeared in code environment. quite unreadable now. just look at the link.

Edited February 4, 2012 by dr.flo

[Guest pier11]

We need to re-experiment. Also might makes sense to try to borrow the same libs from Magic froyo (hope it is GSM), or from Hero, or gw620.

Edited February 4, 2012 by pier11

[Guest pier11]

Playing with u8150 froyo rom:

1. Fixed sound by renaming AudioFilter.csv as BAK.AudioFilter.csv.BAK

2. Replaced lights.msm7k.so from the one from Ascend. Did not help to keypad light. It's the same lib in fact - checked md5. Solution looks like in some other place.

3. I helped Android to load and unload wifi kernel module - ar6000.ko - by replacing libhardware_legacy.so from our CM6. Rationale. Now it refuses to scan network. But progress is obvious. Before that it was even unable to load wifi module.

3.1 Also set ro.config.wifi_chip_is_bcm=false

4. Display density: I just removed ro.sf.lcd_density=120. Resolution is back to traditional for Pulse now. (though I somewhat liked 120).

[Guest dr.flo]

3. I helped Android to load and unload wifi kernel module - ar6000.ko - by replacing libhardware_legacy.so from our CM6. Rationale. Now it refuses to scan network. But progress is obvious. Before that it was even unable to load wifi module.

Good idea! Now I'm quite sure the remaining problems reside in wpa_supplicant binary.

Just go to system/bin in shell and execute it. It closes with a segfault. I don't think this is normal.

This wpa_supplicant (a v6) might be tied too closely to broadcom wifi chipset.

Replacing wpa_supplicant with the one from CM7 also doesn't work, but the error is different. It doesn't support "wext" driver, instead it supports "awext". Is looks like some v5 vs v6 issue.

There must be some config for this. Board_config.mk (CM7) has

BOARD_WPA_SUPPLICANT_DRIVER := WEXT for u8150 and

BOARD_WPA_SUPPLICANT_DRIVER := AWEXT for u8220

we should find out where this gets used. And than exchange this file in addition. I believe ar6000.ko needs a matched (v5??) wpa_supplicant. When I played with different wpa_supplicants from several roms (trying to fix wifi reconnect in CM6 with kernel .32 *) I once also reached the state where wifi would activate but not scan for networks...

*) cm6-clean shows same wifi issues as cm7 when running on kernel .32 (ar6000.ko replaced; wifi works initially, but no reconnect after wifi sleep timeout)

4. Display density: I just removed ro.sf.lcd_density=120. Resolution is back to traditional for Pulse now. (though I somewhat liked 120).

Then maybe density 140 would be perfect for you. You might like "LCDDensity for Root" app (market) once we have a stable rom..

Edited February 5, 2012 by dr.flo

[Guest dr.flo]

This is for sure related to our problem. It was from TomGs original attempt to port froyo (without cm6) to pulse. Initially, wifi wasn't working.

I did:

- exchange libhardware_legacy.so

- exchange wpa_supplicant

- edit init.rc in ramdisk in a way like in posted link

result:

E/WifiHW ( 1268): Unable to open connection to supplicant on "eth0": No such file or directory

I/wpa_supplicant( 2092): CTRL-EVENT-STATE-CHANGE id=-1 state=0

E/wpa_supplicant( 2092): Failed to disable WPA in the driver.

E/WifiHW ( 1268): Supplicant not running, cannot connect

E/WifiHW ( 1268): Supplicant not running, cannot connect

E/WifiHW ( 1268): Supplicant not running, cannot connect

[Guest pier11]

couple of thoughts on wifi.

1. When I replaced only libhardware_legacy.so and enabled wifi via android wifi settings, iwconfig command for some reason did not show eth0 at all (I'd notice as it prints much additional info if wireless extension is available on interface). Contrary, if just insmod ar6000.ko manually, eth0 with wireless extension appears just fine (checked with iwconfig).

2. Communication between libhardware_legacy.so and wpa_supplicatn is happening via a socket. It is configured in different place on u81x0 roms vs. (u82x0 and CMx). But I think the location is configured in wpa_supplicant.conf and not hardcoded in either parties (?)

Here is overview of how wifi functioning in android, Tom used...

Edited February 5, 2012 by pier11

[Guest pier11]

*) cm6-clean shows same wifi issues as cm7 when running on kernel .32 (ar6000.ko replaced; wifi works initially, but no reconnect after wifi sleep timeout)

I believe the problem resides in .32 kernel with regard to wifi integration. Particular place is drivers/mmc/host/msm_sdcc.c

There happens integration of wifi module as SDIO card into kernel. It worth (re)studying the same .29 kernel file, and changes huawei made there vs. code aurora original version (I can post a diff if needed).

As just a workaround people propose to set wifi to never sleep in android wifi settings. Doesn't solve the issue, but might temporarily help to someone.

Edited February 5, 2012 by pier11