[ROM][GEN2]CyanogenMod 7 (Android 2.3.7)

[Guest sej7278]

interestingly enough the new kernel seems to turn off usb mass storage when you eject the device from your computer - you don't need to click the orange "turn off" button. as soon as you eject it, it goes back to the green "turn on" button.

just a thought - will the new proximity sensor calibration app (zte blade settings) work with the new kernel drivers?

also the gps seems more accurate - like it knows i'm in my house rather than up the street and across the road like it usually does, could also be part of the maps 5.1 update i guess.

ping still working with the screen off, goes off after 15mins and doesn't reconnect :(

Edited September 9, 2011 by sej7278

[Guest sej7278]

but new kernel is merged.... you can try:

http://www.the-jedi..../downloads/cm7/

there is 2635 zip

thanks to sej...

no that's not merged, that's just a test build based on gerrit/github plus some patches from git and me. its just to get some early testing done.

you're welcome to try it (forgot to mention it!)

Edited September 9, 2011 by sej7278

[Guest hlubepet]

no that's not merged, that's just a test build based on gerrit/github plus some patches from git and me. its just to get some early testing done.

you're welcome to try it (forgot to mention it!)

yes, Ive flashed it on way to work.... 4 people were watching me WHAT THE HELL IM DOING TO MY PHONE :-D Your test build seems to be great (for a testing build) :) good job!

[Guest samjam]

yes, Ive flashed it on way to work.... 4 people were watching me WHAT THE HELL IM DOING TO MY PHONE :-D Your test build seems to be great (for a testing build) :) good job!

You must work in a techie part of town.

While travelling to work I used to play bubblets on a pre-release of the first MS Smartphone, the SPV1, and no-body cared. Ignorant philistines.

All those 80's High-School movies when the geek finally gets acceptance are a big lie; fortunately Napoleon Dynamite came along to neutralise all the lies... :-)

However these days you were lucky you weren't arrested for whatever a copy saw on the sc-fi channel the previous week.

[Guest hlubepet]

You must work in a techie part of town.

While travelling to work I used to play bubblets on a pre-release of the first MS Smartphone, the SPV1, and no-body cared. Ignorant philistines.

All those 80's High-School movies when the geek finally gets acceptance are a big lie; fortunately Napoleon Dynamite came along to neutralise all the lies... :-)

However these days you were lucky you weren't arrested for whatever a copy saw on the sc-fi channel the previous week.

but I live in czech republic... the terrible country.. everybody is taking care of things whats not of their bussiness... sadly

[Guest dadashi]

Ive been to tschech a couple of times recently, only thing I realy remember are the young girls waiting to be picked up on the motorway to germany...sorry if I got a poor impression...

actualy I was there a couple of times in the 80s when it was czechoslovakia...back then the girls were big soviet mother types doing mans work..

Edited September 9, 2011 by dadashi

[Guest androrphin]

interestingly enough the new kernel seems to turn off usb mass storage when you eject the device from your computer - you don't need to click the orange "turn off" button. as soon as you eject it, it goes back to the green "turn on" button.

just a thought - will the new proximity sensor calibration app (zte blade settings) work with the new kernel drivers?

also the gps seems more accurate - like it knows i'm in my house rather than up the street and across the road like it usually does, could also be part of the maps 5.1 update i guess.

ping still working with the screen off, goes off after 15mins and doesn't reconnect :(

Rather than load the whole ROM, could I just extract the boot.img from your 2.6.35 ROM and fastboot it onto my Blade ?

[Guest Benoe]

I hope to have the new kernel merged early next week. There are still some features I haven't tested, but everything I have tested is working at the moment.

If you are working on the new kernel source would you please check if you could add RDS support to the FM radio driver?

I know it is not the most important thing to do, but knowing that our device is capable of something, what is hidden by the software is not good.

Andorko made some progress in it, he even published his kernel sources (altough it is for 2.6.32) here: link

He also modified the fm radio app to display those rds messages in that topic, maybe we could use that or I don't know if CM is capable to display rds if the device has the function officially supported.

Edited September 9, 2011 by Benoe

[Guest wbaw]

If you are working on the new kernel source would you please check if you could add RDS support to the FM radio driver?

I know it is not the most important thing to do, but knowing that our device is capable of something, what is hidden by the software is not good.

Andorko made some progress in it, he even published his kernel sources (altough it is for 2.6.32) here: link

He also modified the fm radio app to display those rds messages in that topic, maybe we could use that or I don't know if CM is capable to display rds if the device has the function officially supported.

Burstlam added the patch to his .35 kernel & it works on zte roms - https://github.com/burstlam/kernel-35/commit/7bffbd152242b6a5e65be897c9ca66fcdd0fd8d2 I'm not sure if the radio app works differently on cm7 though.

[Guest targetbsp]

debug.composition.type=gpu

That line in build.prop makes the scrolling faster in zte-gingerbread, without that it's slower than cm7. I guess it's using the gpu for the ui composition, which makes it faster. It was buggy in the leaked version, it spammed logcat with lots of errors, but with the updated libsurfaceflinger.so from the skate it's less buggy & doesn't spam logcat.

It could take more than the new kernel to make it work in cm7.

This line seems to work without error (visual or logcat) on sej's build of CM7 with the new kernel. Whether it has any effect or not I don't know. I don't think so but I'm not a very good judge of 2d framerates!

Edited September 9, 2011 by targetbsp

[Guest targetbsp]

just a thought - will the new proximity sensor calibration app (zte blade settings) work with the new kernel drivers?

It worked for me. And indeed, needed to be done. :)

[Guest Benoe]

I'm not sure if the radio app works differently on cm7 though.

The question is what radio app Burstlam is using in his rom?

Andorko's app is a modified MIUI FM radio app, wich (the original) is in cm7.

[Guest fldc]

debug.composition.type=gpu

That line in build.prop makes the scrolling faster in zte-gingerbread, without that it's slower than cm7. I guess it's using the gpu for the ui composition, which makes it faster. It was buggy in the leaked version, it spammed logcat with lots of errors, but with the updated libsurfaceflinger.so from the skate it's less buggy & doesn't spam logcat.

It could take more than the new kernel to make it work in cm7.

right, ok, thank you, but no, that doesn't seem to have much of an effect on cm7 unfortunately :D

[Guest CaptainSpectacular]

What's about the scroling speed with the new kernel ? Some improvements ?

Thanks.

[Guest fldc]

What's about the scroling speed with the new kernel ? Some improvements ?

Thanks.

look one post up ;)

[Guest CaptainSpectacular]

look one post up ;)

I'm becoming blind sorry, thanks. ;)

[Guest Tom G]

interestingly enough the new kernel seems to turn off usb mass storage when you eject the device from your computer - you don't need to click the orange "turn off" button. as soon as you eject it, it goes back to the green "turn on" button.

just a thought - will the new proximity sensor calibration app (zte blade settings) work with the new kernel drivers?

also the gps seems more accurate - like it knows i'm in my house rather than up the street and across the road like it usually does, could also be part of the maps 5.1 update i guess.

ping still working with the screen off, goes off after 15mins and doesn't reconnect :(

proximity calibration is still the same (the ioctl is the same), but the values are totally different, so for anyone that has used the calibration on the old kernel they will need to re-run it after upgrading. No changes are needed to the app, the sensor just needs to be recalibrated again.

Rather than load the whole ROM, could I just extract the boot.img from your 2.6.35 ROM and fastboot it onto my Blade ?

The patch contains changes to components not in the boot image, the one with the biggest impact will be the sensor lib. If you only replace the boot image your light and proximity sensors will now work.

If you are working on the new kernel source would you please check if you could add RDS support to the FM radio driver?

I know it is not the most important thing to do, but knowing that our device is capable of something, what is hidden by the software is not good.

Andorko made some progress in it, he even published his kernel sources (altough it is for 2.6.32) here: link

He also modified the fm radio app to display those rds messages in that topic, maybe we could use that or I don't know if CM is capable to display rds if the device has the function officially supported.

The CM7 FM app doesn't currently support RDS and I don't think Andarko's app will work on CM7 (I think the FM interface is different), so adding RDS to the kernel at the moment isn't useful. Also, the way Andarko did some of it isn't very tidy (not suggesting his code is bad, he had to make it work with the interface provided by the ZTE prebuilt roms which doesn't leave many options). I will look at adding RDS to the CM7 app some time in the future. I also want to redesign the ui of the FM app (I don't like the current one) so when I do that I will look at adding RDS and a few other things.

Edited September 9, 2011 by Tom G

[Guest Benoe]

The CM7 FM app doesn't currently support RDS and I don't think Andarko's app will work on CM7 (I think the FM interface is different)

When I last tried his app it was really not working. But he also used the same MIUI Fm app as a base, as CM7 has. (at least the gui looked like that, with some extra field for rds information :) )

Edited September 9, 2011 by Benoe

[Guest Tom G]

I'm trying to reproduce a shutdown I experienced twice with the new kernel. It happened after I played some angry birds and after trying to wake my phone from sleep It's been off. While I'm trying to reproduce it, can someone tell how was I supposed to get a decent logcat.

$ adb logcat > logcat.txt

was it?

Thank you. Don't bother providing a log.

I wanted to see if the old overheating bug was still a problem before patching it, and it looks like it is. This bug looks like a firmware issue, so I wasn't expecting a kernel update to fix it, but wanted to check before including the patch. I'll get a new kernel up sometime in the next day to fix it. This bug cannot be reproduced on either of my blades so I rely on others for feedback (Jacob doesn't get it either).

Edited September 9, 2011 by Tom G

[Guest Tom G]

When I last tried his app it was really not working. But he also used the same MIUI Fm app as a base, as CM7 has. (at least the gui looked like that, with some extra field for rds information :) )

The difference will be in the calls to the framework. Andarko uses the interface used by ZTE, while the interface in CM7 will probably be based on either AOSP or CAF. I think ZTE originally added FM support when there was no support in the standard framework, so they would have made their own interface and I think they have carried that same interface through froyo and gingerbread (I think gingerbread is the first AOSP version with an FM interface).

Edited September 9, 2011 by Tom G

[Guest waby38]

no that's not merged, that's just a test build based on gerrit/github plus some patches from git and me. its just to get some early testing done.

you're welcome to try it (forgot to mention it!)

I've just tested your 179++ 2.6.35 on my blade. (http://www.the-jedi..../downloads/cm7/)

- good: it work, and my GFX / textures issue are gone (see my previous post about that. It's probably due to your new testing library)

- bad: no GPS signal :( !!!!! (kernel related ?) GPS is working as usual !!

I also tested your 179++ (with the stable 2.6.32)

- GFX / textures bad (even with your testing library ?) / Seem to be kernel related ??

What about you ?

Edited September 9, 2011 by waby38

[Guest sej7278]

I've just tested your 179++ 2.6.35 on my blade. (http://www.the-jedi..../downloads/cm7/)

- good: it work, and my GFX / textures issue are gone (see my previous post about that. It's probably due to your new testing library)

- bad: no GPS signal :( !!!!! (kernel related ?)

I also tested your 179++ (with the stable 2.6.32)

- GFX / textures bad (even with your testing library ?) / Seem to be kernel related ??

What about you ?

n179plus (2.6.32) doesn't use the new libraries. the proprietary files are taken from my phone after i tpt'd it to gen2.

n179-2635 (2.6.35) uses the new libraries and kernel. i found getting a GPS fix is very slow but once its locked it seems more accurate than with the "old" kernel/libs.

edit: 2nd time around i went outside and gps fix was a few seconds that's all (igo and maps). note to self, when trying to get a gps signal do not stay inside a house with 3ft think stone walls.

Edited September 9, 2011 by sej7278

[Guest bhf]

n179-2635 (2.6.35) uses the new libraries. i found getting a GPS fix is very slow but once its locked it seems more accurate than with the "old" kernel/libs.

I always use faster fix from market and select my region. Since N175 I have not one GPS related reboot. With the new kernel I got a fix within 10 seconds. Although the accuracy seems a little off. I check GPS with GPS Test and the accuracy just tripled. With the old kernel I got accuracy about 10 to 15 feet, now it's up to 20 Meters. Everything else is fine.

Should this new kernel boost the performance? Because my anTuTu benchmarks are still about the same, all around 1480 to 1530 points, depending how well my sd card is working. :)

Edited September 9, 2011 by bhf

[Guest sej7278]

proximity calibration is still the same (the ioctl is the same), but the values are totally different, so for anyone that has used the calibration on the old kernel they will need to re-run it after upgrading. No changes are needed to the app, the sensor just needs to be recalibrated again.

bugger, it was working until i ran the calibration app, now i can't get the screen to go back on during a call :(

deleted prox_data.txt and rebooted, now the screen won't go off during a call. re-run the app and it goes off but won't come back on! couple of results from prox_cal:

# /system/bin/prox_cal -d

prox_theshold_hi = 4786

prox_theshold_lo = 3829

prox_int_time = 238

prox_adc_time = 238

prox_wait_time = 246

prox_intr_filter = 0

prox_config = 0

prox_pulse_cnt = 8

prox_gain = 32


# /system/bin/prox_cal -d

prox_theshold_hi = 462

prox_theshold_lo = 369

prox_int_time = 246

prox_adc_time = 255

prox_wait_time = 255

prox_intr_filter = 0

prox_config = 0

prox_pulse_cnt = 16

prox_gain = 32


# dmesg|tail

<2>[09-09 15:41:36.360000] [4: events/0]TAOS: als_interrupt =905

<2>[09-09 15:41:38.750000] [944: prox_cal][taos]prox on while light on

<2>[09-09 15:41:38.760000] [944: prox_cal][taos]TAOS_IOCTL_PROX_ON

<2>[09-09 15:41:38.880000] [4: events/0]TAOS: prox_interrupt =1263, distance=5

<2>[09-09 15:41:42.170000] [944: prox_cal]<3>taos prox_cal_threshold_hi=5185,prox_cal_threshold_lo=4148

<2>[09-09 15:41:42.230000] [4: events/0]TAOS: als_interrupt =86

<2>[09-09 15:41:42.230000] [944: prox_cal][taos]prox off while light still on

<2>[09-09 15:41:42.240000] [944: prox_cal][taos]TAOS_IOCTL_PROX_OFF

<2>[09-09 15:41:42.610000] [4: events/0]TAOS: als_interrupt =841

<2>[09-09 15:41:45.250000] [4: events/0]TAOS: als_interrupt =312

# proxcal -d

proxcal: not found

# /system/bin/prox_cal -d

prox_theshold_hi = 5185

prox_theshold_lo = 4148

prox_int_time = 238

prox_adc_time = 238

prox_wait_time = 246

prox_intr_filter = 0

prox_config = 0

prox_pulse_cnt = 8

prox_gain = 32

# root@greivous:~# 

oddly enough screen on/off works fine if its plugged into usb. without usb i have to push the power button to get the screen back on during a call.

nandroid didn't fix it.

Edited September 9, 2011 by sej7278

[Guest waby38]

n179plus (2.6.32) doesn't use the new libraries. the proprietary files are taken from my phone after i tpt'd it to gen2.

n179-2635 (2.6.35) uses the new libraries and kernel. i found getting a GPS fix is very slow but once its locked it seems more accurate than with the "old" kernel/libs.

Sorry for the bad noise about n179-2635, GPS is working perfectly. I was in a building during my first test (no GPS signal)

GFX / texture issue are alway raised by this game. Sometime also with Angry Bird Rio (less often)

This is fixed with your n179-2635, but still present with your n179++

First slash screen is ok, then the second is bad (nothing on the screen), and in the game, we never see the mouse !