New ZTE 32 kernel

[Guest RuiADF]

Here is another BOOT.IMG for CM base on the same source but I have refine the defconfig.features as follow:

1. overclocking

2. EXT4 , CIFS and TUN support.

3. Fix sensor issue.

4. improve browsing with default ADW desktop [much smoother]

5. usb and wifi tethering is working

6. remove debug option that have no necessary

CM7boot_0624v3.zip

md5 :e580ba6263a0c31c83805a3c947d9c5d

I have the proximity sensor issue on CM7 ROM.

How can I test this Kernel???

[Guest burstlam]

What adw do with the kernel ?

How do you have corrected the bugs ?

Guess you have used Tom_g patches

You have probably contaminated the new source

If it is the case this kernel worth nothing ; just continue to use the CM7 kernel

contaminated? Why did u enable the overclock ability then?

basically zte source is a 2G VM SPLIT kernel.

if u did not contaminate it , How should we make it compatible to run on CM?

the 2G option merged today is incomplete.

this new source originally serve the X880, a new version recently listed in China.

using the BCM chips for wifi and blueooth.

so this source is some how incompatible with the original v880.

and there are many option that should leave close or enabled.

and some of the drivers are even broken (ie OLED driver...). I just replace it with the older version from ZTE.

saying contaminate is nonsense to me, for ZTE mess up the source too much.

leaving too much debug options enabled that slow down the machine.

Edited July 1, 2011 by burstlam

[Guest Victor von Zeppelin]

Burstlam, top work. You too, Pheonix Silver.

A question for both of you, is there a way the edge of screen detection could be built back in, like was added to the default kernel relatively recently? Means I can't push the notification bar back up easily.

Also, specifically at Burstlam (probably) ,is there anything to speed up web browsing in yours? Its the one I'm using, and modaco feels silky smooth

Edited July 1, 2011 by Victor von Zeppelin

[Guest burstlam]

Burstlam, top work. You too, Pheonix Silver.

A question for both of you, is there a way the edge of screen detection could be built back in, like was added to the default kernel relatively recently? Means I can't push the notification bar back up easily.

Also, specifically at Burstlam (probably) ,is there anything to speed up web browsing in yours? Its the one I'm using, and modaco feels silky smooth

Not sure as the input driver has been change that even the touchscreen fix of no detection is not implemented is this update source.

I am using an oled Blade and for me the patch of screen edge detection behaves abnormally with screen freezes ocasionally.

So I am using self build version and dropped this patch since official merge before this comes up, so far it works normally for me.

For surfing speed. In addition to the wifi driver of ZTE update, I tweaked the TCP buffer size in the ramdisk as well.

Edited July 2, 2011 by burstlam

[Guest KapTmaH]

official kernel_2.6.32_blade

http://support.zte.com.cn/support/uploads/..._2.6.7z.001.zip

http://support.zte.com.cn/support/uploads/..._2.6.7z.002.zip

http://support.zte.com.cn/support/uploads/..._2.6.7z.003.zip

http://support.zte.com.cn/support/uploads/..._2.6.7z.004.zip

http://support.zte.com.cn/support/uploads/..._2.6.7z.005.zip

http://support.zte.com.cn/support/uploads/..._2.6.7z.006.zip

[Guest bykumza]

I ask you to show the screen version of the kernel. (Written using the google translator)

Ps. Sorry for my bad English.

Edited July 2, 2011 by bykumza

[Guest hecatae]

contaminated? Why did u enable the overclock ability then?

basically zte source is a 2G VM SPLIT kernel.

if u did not contaminate it , How should we make it compatible to run on CM?

the 2G option merged today is incomplete.

this new source originally serve the X880, a new version recently listed in China.

using the BCM chips for wifi and blueooth.

so this source is some how incompatible with the original v880.

and there are many option that should leave close or enabled.

and some of the drivers are even broken (ie OLED driver...). I just replace it with the older version from ZTE.

saying contaminate is nonsense to me, for ZTE mess up the source too much.

leaving too much debug options enabled that slow down the machine.

because overclock ability is allowed by cyanogenmod using SetCPU

you can manually make the kernel build as 3g vm split, but it will only allow 256mb ram in the standard configuration

X880 does not use BCM chips, we have a stock rom for the X880 and it works fine on the blade.

this source is not incompatible, you just have to know what to enable when you make config

OLED driver is not broken, you just need to enable it, by replacing it with the driver from the previous source you have contaminated the source

[Guest k0zmic]

I have the proximity sensor issue on CM7 ROM.

How can I test this Kernel???

If you have the Proximity Sensor issue please star and attach a logcat with a comment of what happens here (if you haven't already): http://code.google.com/p/cyanogenmod/issue...tars%20Priority

A logcat can be taken by alogcat from the Market.

Open alogcat

Dial a number

Wait for the screen to return

Go back to alogcat

Save

Take off your SD Card (it's in the alogcat folder)

Upload

Thank you.

Edited July 2, 2011 by k0zmic

[Guest burstlam]

because overclock ability is allowed by cyanogenmod using SetCPU

you can manually make the kernel build as 3g vm split, but it will only allow 256mb ram in the standard configuration

X880 does not use BCM chips, we have a stock rom for the X880 and it works fine on the blade.

this source is not incompatible, you just have to know what to enable when you make config

OLED driver is not broken, you just need to enable it, by replacing it with the driver from the previous source you have contaminated the source

To me, even manual editing to the source means the source being contaminated. Unless u touch nothing.

so I think it is a nonsense.

N880 (x880 is a typo mistake) in china is different from those outside, using the OLED display and the BCM chipset.

I don't know why, by the default config. the OLED is enabled. but it breaks.

As this source served primary for V9. I suspect whether this OLED driver is a ready to release version for the Blade.

(noted that the OLED option is not enabled with V9 config)

I don't have the ability to fix it. So I choose myself to replace it by the old one.

Edited July 2, 2011 by burstlam

[Guest wbaw]

Good work Burstlam, do you have a github?

[Guest t0mm13b]

Look ....

Who ever mentioned ADW feeling snappier - pffftttt

Lets look at this in objective manner -

stop contaminating the sources and putting in patches etc into the newer source tree... there's half of stuff in there that is not in the old .32 sources and by contaminating it - the result would be confusion over as to what's been put in etc...and possibly make the kernel worse ... quite frankly there's a lot more in there that surprised me such as code comments which was omitted in the old .32 source, also blade's supporting dependency on other boards which appears to be far more complete than the old .32 source.... there's too many questions about it...

so stop right there producing a kernel for now....do it for yourself but do not share it as everyone is going to be under the impression its the same as the old .32 kernel currently in cm7.....

[Guest wbaw]

WTF? Do you own the Linux kernel now, t0mm13b? What gives you the right to tell Burstlam what he can & can't do with it? It's this kind of bullshit that puts people off doing anything to improve what we've got.

Edited July 2, 2011 by wbaw

[Guest t0mm13b]

WTF? Do you own the Linux kernel now, t0mm13b? What gives you the right to tell Burstlam what he can & can't do with it? It's this kind of bullshit that puts people off doing anything to improve what we've got.

Hang on - you misread and overreact

the sources seems to be better, but by putting in overclock for example halves the memory which Phoenix discovered the bug....

All I'm saying is, by putting in other patches into it could end up making the kernel back to square one like what's currently in place

Edited July 2, 2011 by t0mm13b

[Guest will8578]

the memory bug was in the first kernel uploaded by phoenix and the kernel burstlam put up fixed it

[Guest Pelemane]

the memory bug was in the first kernel uploaded by phoenix and the kernel burstlam put up fixed it

Yes burstlam fixed it by putting some stuff from old kernel into it.

We should try to fix new kernel, not to mix new and old kernels.

[Guest hedgepigdaniel]

Reading this thread, one could be forgiven for thinking it was impossible to extract the original ZTE source again if any modifications have unwanted side effects. Modifying source code is not like eating a cake.

[Guest Phoenix Silver]

To me, even manual editing to the source means the source being contaminated. Unless u touch nothing.

so I think it is a nonsense.

N880 (x880 is a typo mistake) in china is different from those outside, using the OLED display and the BCM chipset.

I don't know why, by the default config. the OLED is enabled. but it breaks.

As this source served primary for V9. I suspect whether this OLED driver is a ready to release version for the Blade.

(noted that the OLED option is not enabled with V9 config)

I don't have the ability to fix it. So I choose myself to replace it by the old one.

Burstlam you have made a excellent work congrats

All i would want to say is applying AUTO patchs is a very bad way because they'll probably corrupt the code

You have edited manually the code so you have made an excellent work

All the community appreciate your help :)

Cheers

Alice

[Guest wbaw]

Reading this thread, one could be forgiven for thinking it was impossible to extract the original ZTE source again if any modifications have unwanted side effects. Modifying source code is not like eating a cake.

I know, it's stupid. You can have the source code, but you mustn't change it! Don't dare to try fixing any bugs! Only Pheonix Silver & t0mm13b are allowed to touch it!

Isn't this what version control software, like git is supposed to sort out? Just put it on a github & let us see what patches you've added.

Edited July 2, 2011 by wbaw

[Guest Phoenix Silver]

I know, it's stupid. You can have the source code, but you mustn't change it! Don't dare to try fixing any bugs! Only Pheonix Silver & t0mm13b are allowed to touch it!

Isn't this what version control software, like git is supposed to sort out? Just put it on a github & let us see what patches you've added.

Have you read my post above or you just like to add fuel to the fire ?

[Guest Mushroom_Lord]

Nice work here Phoenix Silver :)

Good luck getting RAM to work properly :rolleyes:

Does this make scrolling Smoother? or is that hidden in like an Xml in framework-res or something... (could we change the math in the Cm7 one from something else to make scrolling smoother?)

[Guest johnsmithx]

Have you read my post above or you just like to add fuel to the fire ?

He is just making a legitimate request. You both have made some changes in the source codes and you both distributed a compiled version. Since the code is GPL, you are obliged to release your source code. None of you did that so far.

I really don't think that asking for something that is not only totally usual and common in opensource community, but what is also guaranteed by the law (yes, the license is a legal contract you are bound to comply with from the moment you started using the GPL licensed work) is "adding fuel to the fire".

From http://www.gnu.org/licenses/gpl.html:

"..if you distribute copies of such a program, whether gratis or for a fee, you must pass on to the recipients the same freedoms that you received. You must make sure that they, too, receive or can get the source code."

I believe many people definitely appreciate your work and we all want you to keep it that way. You are doing it for free, you are doing it for the community. No doubt it's a good thing. It's really great. Even if this particular work wasn't the most useful on the world, it's still very good that you are being involved and interested and you are willing to help and make things better, not only for yourself but for benefit of others as well. No one wants to discourage you in any way, please don't take my words that way. I don't even think those people who made such comments here as to urging to not change the kernel, not mix "old" and "new" etc., really did mean to discourage you either, they meant it well to protect people from confusion (the question is whether people really need such protection, whether they really need someone else to decide what they should or should not get).

But please, if someone is asking for the source code of your changes in GPL work, just give it to them. Making "patch -ru kernel.orig kernel.changed > mypatch.diff" is a question of just a few seconds.

Edited July 2, 2011 by johnsmithx

[Guest burstlam]

Yes burstlam fixed it by putting some stuff from old kernel into it.

We should try to fix new kernel, not to mix new and old kernels.

the memory fault of 256mb is a wrong setting of the page numbers and io address

and it is snappier because I have disabled those useless debug options in defconfig that slow down the machine.

the only thing I replaced is the oled driver that breaks the build.

[Guest Redflake]

WTF? Do you own the Linux kernel now, t0mm13b? What gives you the right to tell Burstlam what he can & can't do with it? It's this kind of bullshit that puts people off doing anything to improve what we've got.

+1

[Guest unrandomsam]

Hang on - you misread and overreact

the sources seems to be better, but by putting in overclock for example halves the memory which Phoenix discovered the bug....

All I'm saying is, by putting in other patches into it could end up making the kernel back to square one like what's currently in place

Having the memory is not caused by putting in overclock.

[Guest will8578]

what we do need is something like the kernel that franco in the lg optimus one threads has done, he pushes a kernel out what seems like everyday, the kernel he made from 32 kernel is something amazing and now he's started on 35 kernel, you should see the list of things he's added, the lg op one has same specs as blade, with his kernel and another blokes rom, noiejn or something like that, they afe getting something like 2000+ scores in quadrent, i know not to base things on that, but its still truely amazing

link to his thread:

http://forum.xda-developers.com/showthread.php?t=1049518

My daughter has this on her phone, she doesnt realise how lucky she is lol