[RECOVERY] Working recovery with UI glitches

[xaueious]

I hope this invigorates development again.

I just got my S7 and pushed a bunch of system files, including stuff from old Liquid ROMs. We also now have the ability to build parts off AOSP.

I guess we are missing a user workable Clockwork recovery to do user backups and to help deploy them to more people? It at least needs nandroid?

 

[buzzman]

xaueious- Congrats in receiving your hs7. This forum is starting to get very exciting with all the breakthroughs going on.

 

[SikYou]

I hope this invigorates development again.

I just got my S7 and pushed a bunch of system files, including stuff from old Liquid ROMs. We also now have the ability to build parts off AOSP.

I guess we are missing a user workable Clockwork recovery to do user backups and to help deploy them to more people? It at least needs nandroid?

Yea, so far the recovery is not usable. Well it works but the GUI is no good. An updated kernel should solve the recovery port problems and move us a few steps toward custom ROMs and all the other goodies.

**Edit- Just tried clockwork with the Jan 6 kernel build and I got the same mirrored results.

 

[xaueious]

We need to be able to build a working one then

 

[SikYou]

We need to be able to build a working one then

That's the plan A few of us have been working on it. Check out the dead irc, maybe you'll catch someone working on it

Sent from my nonsensikal froyo

 

[xaueious]

I've been trying to just repack the image from my repack-H.pl script. I haven't gotten it to boot even though unpack works. Is there some sort of trick to it?

I'm just trying to replace boot.img right now...

Is there some magic with mkbootimg, the base address, or something else?

nvm. There is a trick.

use same base address as Nexus one and it will boot

 

[SikYou]

I've been trying to just repack the image from my repack-H.pl script. I haven't gotten it to boot even though unpack works. Is there some sort of trick to it?

I'm just trying to replace boot.img right now...

Is there some magic with mkbootimg, the base address, or something else?

nvm. There is a trick.

use same base address as Nexus one and it will boot

What is the base address?


Sent from my S7

 

[xaueious]

Just realized that you can use fastboot to boot recovery and not wipe, but if you flash it and use reboot recovery at any time, it will wipe

 

[SikYou]

Just realized that you can use fastboot to boot recovery and not wipe, but if you flash it and use reboot recovery at any time, it will wipe

Yep. This problem should be fixed with a minor change to init.rc and a cwr port.

I think you have provided the missing piece to compiling the kernel but I wont be able to test build for a few days.

Sent from my nonsensikal froyo

 

[babybearsg]

Hi guys,

i'm having some issues with adding framebuffer support.
Does anyone know if the mddi toshiba drivers compiled during make plays a part in the S7 screen control?

Anyone knows what screen type the S7 uses anyway?

I've tried to define the fb as 2:
struct msm_panel_info *pinfo;
pinfo->fb_num = 2;

I compiled and flashed recovery.
got 4 bloody split screens instead of the 2.

I think the huawei driver is defective.
Should i spend some time rewriting the driver or am i headed in the wrong direction?

 

[xaueious]

Not that I know the answer, but...

I got 4 screens instead of the 2 booting perivarlura's recovery as well...

 

[SikYou]

I have ported the recovery multiple times and have had both the mirrored and double mirrored screen. I have yet to identify the exact change that causes the 4 vs 2 screen images.

Sent from my S7

 

[babybearsg]

hi,

is there a way to disable frame buffering in cwr?
does amon-ra have this issue?

i looked at the kernel for other huawei phones like the ascend and the video drivers seem to be very different.
Do you have the source for your ported cwr i could have a look at?

thanks!

 

[SikYou]

Sorry, I don't have any of the cwr port stuff left. I broke a bunch of stuff in my CM source and had to clean everything and sync, of course I didn't have anything backed up ;(

I suspect that Amon recovery would have the same problem as cwr but I have not tried a port.

The huawei kernel is very poorly written and the drivers are no exception. I have tried half a million changes to the kernel and I just can't get anything done.

 

[xaueious]

Isn't the framebuffer stuff actually in msm_fb.c?

Huawei modified the driver here. We have codeaurora to refer to: https://www.codeaurora.org/gitweb/q...fs/heads/eclair_chocolate;hb=eclair_chocolate

Display Drivers | Android Open Source

I finally understand why you guys have been saying it's a framebuffer problem. Not that I have any idea how to fix it though:

Troubleshooting
Both of the following problems have a similar cause:

Number keys: In the dialer application, when a number key is pressed to dial a phone number, the number doesn't display on the screen until after the next number has been pressed.
Arrow keys: When an arrow key is pressed, the desired icon doesn't get highlighted. For example, if you browse through icons in the Applications menu, you might notice that icons aren't highlighted as expected when you use the arrow key to navigate between options.
Both problems are caused by an incorrect implementation of the frame buffer's page flipping. Key events are captured, but the graphical interface appears to drop every other frame.

Or someone has to fix the graphics under /bootable/recovery/miniui?