Kernel 2.6.32 from Samsung Acclaim R880 on Spica

[cadavre]

Follow: https://github.com/OpenEtna/android_...roid-lg-2.6.32

They're porting Android to non-Android device - they really need to port everything from the beginning... Etna i using similar components like akm8973, maybe more...

[tom3q]

A small update on my "dancing with 2.6.32 kernel from Acclaim" .

Any chance that your source tree is available somewhere to take a look at it?

[gotenks]

Any chance that your source tree is available somewhere to take a look at it?

Actually they are only on my vm disk . Unfortunately I am not familiar with those "source control things" like git, so I do not have any public repository. However when I find enough time to cleanup my sources from lots of debug messages added by me, and my "trial&test code", I'll try to learn git and I'll put my sources to publicy available repo.

[gotenks]

Finally I figured out what's going on with this "new I2C driver model" and this allowed me to fix detection of few more Spica devices.

1. There are no errors in writing to QT5480 registers, but it hangs on request_irq() function call. I am out of ideas why this happens . To allow me to boot Spica in this case, I've changed touchscreen driver to melfas. Although it does not work, it serves me as a "plumbing" allowing me to run my Eclair and to see what's going on.

2. AK8973B is properly recognised - hopefully it works.

3. Accelerometer does not work, although it is properly recognised (but here I am not sure).

3. Sound works. At least I can hear it .

4. Vibrator works. Not too strong but works.

Here is screenshot of kernel version (I hope it's visible, because an image is on a free host).

[gotenks]

Guys, please, keep in mind that this is a development thread .

M

All posts which were not related to development have been deleted as requested by the developer with no dis-respect to the original posters. Please keep this thread as clean as possible to allow the devs to go through the thread with as less posts as possible. Once the kernel is usable and released a discussion thread will be created. Thanks for understanding

[gotenks]

UPDATE : a phone modem is working . I've ported dpram+multipdp from Spica, because original Acclaim's dpram is simply unusable due to the fact that Acclaim uses completly different method of phone modem initialisation. But the next major problem - QT5480 driver - seems to be unsolvable by me currently .

[tom3q]

UPDATE : a phone modem is working . I've ported dpram+multipdp from Spica, because original Acclaim's dpram is simply unusable due to the fact that Acclaim uses completly different method of phone modem initialisation. But the next major problem - QT5480 driver - seems to be unsolvable by me currently .

I see that request_irq is called with 0 (NULL) as the last parameter. Could you try to pass aDevice instead of 0?
The same has to be passed as the last parameter of free_irq.

[gotenks]

I see that request_irq is called with 0 (NULL) as the last parameter. Could you try to pass aDevice instead of 0?
The same has to be passed as the last parameter of free_irq.

Ok, I'll try tommorow night, because my daughter plans "a murder" on me

[gotenks]

I see that request_irq is called with 0 (NULL) as the last parameter. Could you try to pass aDevice instead of 0?
The same has to be passed as the last parameter of free_irq.

I tried your suggestion but it didn't help unfortunately. However ... see below ...

But the next major problem - QT5480 driver - seems to be unsolvable by me currently .

Actually I was able to find the problem . I was wrong claiming that it is in request_irq() func. Well, it was trigged by calling that func, but the problem itself was in an interrupt handler routine. This routine used disable_irq() call, instead of disable_irq_nosync() (well known behavior of kernels starting from 2.6.31) .

So ... QT5480 driver is working (hurah !!)... strange . I mean, generally it works, but if you put a phone into a sleep mode and wake it up later, the screen is "upside-down" . I guess it's a known bug/behavior from previous kernels, so I am pretty sure our "old kernel hackers" can suggest a solution . Thanks in advance.

But now ... I am going to look at accelerometer, because it doesn't work. There is a message in a logcat that it can't be opened. Doesn't really know what it means, ... but why not have a look at it .

---------- Post added at 06:47 PM ---------- Previous post was at 06:46 PM ----------

So ... QT5480 driver is working (hurah !!)... strange . I mean, generally it works, but if you put a phone into a sleep mode and wake it up later, the screen is "upside-down"

Of course, it is not a QT5480 problem, but I guess, a screen handling (lcd driver) .


PS. MODERATORS : WHY DOES THIS STUPID ENGINE JOIN POSTS TOGETHER ?! IT'S USELESS !!

[cadavre]

Can you paste logcat for accelerometer? : )