[Pipo-Project] M7Pro kernel

[d33]

There are very numerous differences for this file between your (multi tablet) copy and the one from Pipo; and they are not just ifdefine here and there, so I wonder if I should not try to just inject in your kernel tree Pipo's M7Pro file (which is supposed to be fully OK and production quality)... any comment on that ?

Yes, don't. :-)

Simply incorporate the changes with defines where model specific.

I understand you made two changes. Did you test them serially? Perhaps only one change is required?

[edit] I just looked at the Pipo supplied source for M7 and checked the backlight parameters. It seems they were as I had them:

.min_brightness = 20,
.max_brightness = 150,
.brightness_mode =BRIGHTNESS_MODE_CONIC,

[d33]

My attempt to fix - problem is I've been messing with the U8 so may have broken bits that were working

Give it a try please

[TriA]

d33, with this kernel only middle 1/3 (approximately) of scale is useful:
from 0 to 1/3 - backlight is completely off,
from 2/3 to full - backlight is at maximum

Tested GPS and Bluetooth with 1.32.2 and 1.32.3
GPS seems ok (it is able to fix position), while BT is not...
At first, BT cannot be switched on if the tablet was booted with Wi-Fi off.
Next, I was able to create a BT pairing with Nokia C5-00 (Symbian OS), but no data is going over BT, and I cannot even break the pairing (no reaction).
There is no problem with BT on Riley ROM kernel

[d33]

Thanks TriA can you report the ROM image you are using?

[d33]

I may have not got that build right

I'll look into BT but others have reported all well?

[raxy]

dee,
regarding the pipo kernel tree in the m7p SDK,
I added a + to kernel_local_release in .config
but kernel.img still unbootable...
I have had a small hope..scheise !

[TriA]

I use test kernels with Riley ROM 3.4.2

BT is essential for me, since I have Wi-Fi-only tablet, so I use 3G via BT tethering with the good old Nokia

[raxy]

results for 1.32.4:
between 20% and full scale of the ruler : all seem to be quite progressive
at full scale: brightness does not seem to be maxed out, but this is just my impression
I did not dare to go lower than 20% of the scale because the backlight begun to turn completely off at 20% and I was afraid to loose control like before 1.32.2

And automatic brightness cannot be turned on.

yes, pipo's M7p sdk had the values you put in your 1.32.2
so I think the difference comes from another file !

[d33]

I suspect you will find 1.32.4 fixes the issue. The problem arose in the backlight driver. I'm reasonably sure I fixed it now. I'll push my amends when it is confirmed working.
TriA perhaps you are willing to test again?

[raxy]

lowest brightness management and auto still cry for a 1.32.5

[TriA]

With 1.32.4 I am getting exactly the same results as Raxy:
backlight is off at first 1/5 of scale, the rest of scale is ok, auto brightness doesn't work (?) - i switched it on, then screen gone black and I wasn't able to get it back

[d33]

lowest brightness management and auto still cry for a 1.32.5

I'm not sure I follow you. You said that you are too scared to test 1.32.4 at low brightness so what difference will another version make?

Hopefully TriA or Randy will test 1.32.4 properly

[raxy]

backlight completely turning off should be avoided, even at the leftmost position of the ruler, because if you remove your finger from the screen, then you are dead and will have to reflash.; min value of 60 in the c file shall do the job
And post 296 says also that auto brightness is broken, it is not only me.
Randy did not double check those things yet.

[d33]

Understood. So 1.32.5 with some PWM fixes and the minimum backlight value raised to 40

[d33]

I'll push the code so that you can mess with the minimum yourself Raxy

[edit] Push complete