[ICS] [CM9] [4.0.4] CyanogenMod 9 for the Huawei Ascend G300 [ROM]

[Guest EviL FazZ]

Hey Dazzozo I've been using this great value Huawei G300 for about a month now and enjoying it daily ;)

My main and current background is on the HTC HD2, which recently failed, which I thought would be replaced with something else through my insurance, but ended being repaired... lame. Anywho, the G300 I'm also using will be my sexy other half's main phone, but only as soon as I'm happy with it's stability and functionality. I've currently got a random build of iCS 4.0.4 on it, but damn it's made this phone almost unusable <_<

I've tried all sorts with trying to make it smoother, responsive and with the best experience, but like us all, time isn't always that spare :unsure:

I do believe iCS can work far better on the G300 and I simply want to help... if I can... and if you need any help? I'm a huge fan and supporter of CyanogenMod and still have CM7 on my HD2 (which I have a feeling I'll be using for the next few months until I can save some pennies for something else).

I've tried setting up the ability to create ROM's on my Windows OS, but just not sure I'm doing it right as it never finishes to install... long download of about 6 hours, which never finishes?!

Anywho, as I have my HD2 back (need rooting again), I'll mainly using that again and meanwhile can do some serious G300 testing for you, if you need, as it'll just be sitting here now, until I get it fully working. Everything currently works on it using (a slightly custom ROM of) Infusion, but no offence to the hard work the builder put into, but it's not a good ROM. Memory usage is terrible, apps close that are in usage without asking, on daily use it can take up to 10 seconds just to open the dialling pad, trying to answer an incoming call isn't always possible because by the time it actually shows you the option to take the call, it's already gone to voicemail. Anyway, I'm not here to moan, but just state some obvious faults... and there are A LOT more :(

I will admit that haven't actually tried your ROM yet, but hopefully after the weekend (busy with Olympic partying and Closing Ceremony) I will. I'm only cautious because of the lack of WiFi, Bluetooth and the camera. All features I use daily :)

The main reason I'm saying ALL this is your developing for CyanogenMod9 on the G300. This is outstanding to me and will support soon after I'm happy with the test. I've been randomly monitoring your Tweet's which all sound good, just wished you used Google+, but each to there own I guess :P

Meanwhile... enjoy the games (if you're watching... not that any of us in the UK can actually avoid it anyway) and see you again on here after the weekend. Christian O'Connell has to currently play Spandau Ballet's Gold, 26 times on Monday... LMAO!!! :D

[Guest Dazzozo]

I'm not developing CM9 anymore as there's no point when Jelly Bean is what ICS should've been and CM10 is just CM9 but better.

This has become more of a CM10 dev discussion thread.

Edit: from the guy who's "busy" himself: https://github.com/i...m9_device_c8650

[Guest EviL FazZ]

Thanks for letting me know Dazzozo ;)

I will keep my eyes out for CM10 builds, but have a feeling it's going to be a long wait until it's really working :huh:

[Guest Dazzozo]

<4>[1088, Thread-104] [  160.072068] dhd_module_init: Enter

<4>[1088, Thread-104] [  160.072369] ## wifi_probe

<4>[1088, Thread-104] [  160.072381] wifi_set_power = 1

<3>[1088, Thread-104] [  160.254721] bcm_wifi_set_power: wifi power successed to pull up

<4>[1088, Thread-104] [  160.254751] dhd_enable_mmchost_polling: /sys/devices/platform/msm_sdcc.2/polling

<4>[1088, Thread-104] [  160.254884] set_polling : no need to enable polling for slot 2 (as host->pdev_id)

<4>[1088, Thread-104] [  160.266506]

<4>[1088, Thread-104] [  160.266516] Dongle Host Driver, version 5.90.125.120.2

<6>[1067, kworker/u:4] [  160.540696] mmc2: new high speed SDIO card at address0001

<4>[1067, kworker/u:4] [  160.542958] dhd_customer_oob_irq_map oob gpio is : 48;wifi slot is :2

<4>[1067, kworker/u:4] [  160.543071] alloc static buf at dc540000!

<4>[1067, kworker/u:4] [  160.543846] dhd_common_init:fw_path = /data/misc/wifi/load/firmware.bin

<4>[1067, kworker/u:4] [  160.548743] DHD: dongle ram size is set to 294912(orig 294912)

<4>[1067, kworker/u:4] [  160.565029] Broadcom Dongle Host Driver: register interface [eth0] MAC: 00:90:4c:11:22:33

<4>[1088, Thread-104] [  160.566781] wifi_set_power = 0

<3>[1088, Thread-104] [  160.567801] bcm_wifi_set_power: wifi power successed to pull down

<4>[1088, Thread-104] [  160.567813] =========== WLAN placed in RESET ========

<4>[1096, wpa_supplicant] [  160.650649]

<4>[1096, wpa_supplicant] [  160.650664] Dongle Host Driver, version 5.90.125.120.2

<4>[1096, wpa_supplicant] [  160.650681] wl_android_wifi_on in

<4>[1096, wpa_supplicant] [  160.650693] wifi_set_power = 1

<3>[1096, wpa_supplicant] [  160.953276] bcm_wifi_set_power: wifi power successed to pull up

<4>[1096, wpa_supplicant] [  160.953304] =========== WLAN going back to live  ========

<4>[1096, wpa_supplicant] [  160.953318] sdio_reset_comm():

<4>[1096, wpa_supplicant] [  161.134833] dhd_preinit_ioctls: use MAC address innvram 00:90:4c:ce:43:30

<4>[1096, wpa_supplicant] [  161.134863] Firmware up: op_mode=1, Broadcom Dongle Host Driver mac=00:90:4c:ce:43:30

<4>[1096, wpa_supplicant] [  161.828021] wl_android_wifi_off in

<3>[1096, wpa_supplicant] [  161.830189] bcm_wifi_set_power: wifi power successed to pull down

<4>[1096, wpa_supplicant] [  161.830208] =========== WLAN placed in RESET ========

<4>[1096, wpa_supplicant] [  161.832369] do_exit: exit code=65280

<4>[1096, wpa_supplicant] [  161.834036] wpa_supplicant used greatest stack depth: 4768 bytes left

[/CODE]


Boom.


Should be trivial now to get Android to like it... he says. :P


Also, this means we don't need to hack Android anymore! The new kernel and new Wi-Fi system introduced with the 3.0 kernel means no hacks to vold or libhardware_legacy.


Edit: gj huawei "/home/l81003604/U8818/build/Version_U8818_B926/androidcode/"


Oops :P

Edited August 12, 2012 by Dazzozo

[Guest Cyda]

Great work Daz. :)

[Guest unaszplodrmann]

They very thought of Broadcom wifi drivers for Linux makes me cringe, but alas thse days are gone... thank f*ck.

[Guest Dazzozo]

- snip - CloudFlare managed to cause me to double post

[Guest Dazzozo]

The kernel is totally fine with the Wi-Fi now. Inserts and removes cleanly. There is a bit of weirdness to figure out with cfg80211.ko and dhd_4330.ko.

The former is required and must be loaded before you load the second one, and the second one actually enables the Wi-Fi, otherwise you get missing symbols.

It seems Huawei's attitude is "we don't understand the standards, lets make a binary wrapper so we can use the same ROM on about 10 devices". wlan_detect in this case, in the stock ROMs.

cfg80211 is built with the kernel and I might just see if there's a way of building it in to the kernel instead of as a module. Two modules is pretty pointless. dhd_4330 is totally closed, as expected.

At the moment the rest of the Wi-Fi stuff is entirely in Android. Softap fails with:

E/SoftapController(  114): SIOCGIPRIV failed: -1

E/SoftapController(  114): Softap fwReload - failed: -1

'5 softap fwreload eth0 STA' failed with '400 5 Softap operation failed (Operation not supported on transport endpoint)'

[/CODE]


...and then wpa_supplicant ultimately dies like every Wi-Fi issue ever.

[Guest Dazzozo]

https://github.com/D...6d99bd2b8d93398

:)

Working on the overclock implementation now, I'll make it so the defaults are the stock speeds.

Edit: by the way, that commit above damages the stock ROMs a little bit, the stock ROMs will still expect cfg80211 to be a module, and will try to load it. Theoretically it shouldn't prevent Wi-Fi from working, as wlan_detect in the stock ROMs will try to load the module, realise the contents of the module are already there, and fail to do so.

Ultimately no harm should be caused and everything should still work but I thought I should explain that.

[Guest Jekle]

Is there a chance you can release a cm10 build that your testing your OC kernel on?

[Guest Dazzozo]

Just decided to whack in the early sample of ARMv7 Adreno 2xx Jelly Bean drivers from: https://developer.qualcomm.com/mobile-development/mobile-technologies/gaming-graphics-optimization-adreno/tools-and-resources

Let's see what happens now. :P

[Guest anything-but]

There`s a name for people like you dazz and it ends in tease! lol

[Guest vladax10]

if you make CM10 to work,there still will be same isues like in CM9 (BT and camera),isnt it??

[Guest Dazzozo]

Oh wow that was not a mistake! That was the best thing I've ever done, this thing runs so smooth now. No hacks either, I just chucked out a ton of hacks in the device repo.

if you make CM10 to work,there still will be same isues like in CM9 (BT and camera),isnt it??

Well surely the ultimate goal of CM9 was to fix those?

The problem was Jelly Bean released at a bad time for this device, in the middle of ICS upgrades.

CM10 will initially have those broken too, yes, as we never had them working. The devs want everything working too, believe it or not.

[Guest unaszplodrmann]

Oh wow that was not a mistake! That was the best thing I've ever done, this thing runs so smooth now. No hacks either, I just chucked out a ton of hacks in the device repo.

Blimey, you beauty! Definitely the most positive news I've read n the G300 front. Sounds like Huawei would do well to skip ICS :lol:

[Guest Davidoff59]

To think that dev appeared all but dead for the G300,what a turn around

[Guest ramitos]

Oh wow that was not a mistake! That was the best thing I've ever done, this thing runs so smooth now. No hacks either, I just chucked out a ton of hacks in the device repo.

So... If I get it right, you implemented the GPU driver for Jelly Bean which means GPU acceleration?

[Guest Dazzozo]

So... If I get it right, you implemented the GPU driver for Jelly Bean which means GPU acceleration?

Correct.

I/Adreno200-EGL(  452): <qeglDrvAPI_eglInitialize:294>: EGL 1.4 QUALCOMM build: AU_LINUX_ANDROID_JB.04.01.01.00.036_msm8960_JB_CL2644550_release_AU (CL2644550)

I/Adreno200-EGL(  452): Build Date: 07/31/12 Tue

I/Adreno200-EGL(  452): Local Branch:

I/Adreno200-EGL(  452): Remote Branch: quic/master

I/Adreno200-EGL(  452): Local Patches: NONE

I/Adreno200-EGL(  452): Reconstruct Branch: AU_LINUX_ANDROID_JB.04.01.01.00.036 +  NOTHING

D/OpenGLRenderer(  452): Enabling debug mode 0

[/CODE]

[Guest ramitos]

Correct.

Awesome. Great work. I definitely need to donate

Wasn't the aosp camera.app not working because the lack of hw acceleration?

[Guest Dazzozo]

Not sure, I didn't try the AOSP Camera app but it's believable. Huawei would've made their own anyway. :P

Trying to get vsync and triple buffering in the kernel now so we can be totally ~Project Butter~.

Edit: Triple buffering is there already! Just vsync then...

[Guest Davidoff59]

Is that 1.5ghz still stable after all your testing

[Guest tcpaulh]

Daz, you beauty :)

[Guest Dazzozo]

I haven't had 1.5 GHz crash once yet. While that sounds surprising, remember this did come from another device.

I sniped someone talking about the 7x27A in #cyanogenmod-dev and that's where it started.

[Guest carstenheuer]

@ Daz

small mutivation ;)

[Guest anything-but]

I foresee the G300 becoming a new currency,lol

Awesome work,i`m sure you`ll have many a guinea pig ready to help test (me included!). :D