[oe-issues] [Bug 490] tosa screen corruption ("vertical yellow/blue lines")

[bugzilla-daemon at tinman.treke.net]

http://bugs.openembedded.org/show_bug.cgi?id=490





------- Comment #7 from hector at marcansoft.com  2006-10-21 20:17 -------
The previous Python program demonstrates the problem and solution. Pixels
_MUST_ be written to completely, 16 bits at a time. Writing one half of a pixel
and then the other half screws it up. ipkg install python-core to run it.

The corruption itself can be seen by analyzing a framebuffer dump. Basically,
the first byte is partly or fully corrupted, which includes the blue channel
and the 3 LSBs of the green channel.

There is a pattern to the corruption. Some vertical lines(columns) are (almost)
always fully columns of every 8-column block). There is random corruption (some
bits corrupted) on the others.

I believe the framebuffer is implemented as a direct memory map in the kernel,
so a fix there would be hard to implement. Most likely, this will have to be
fixed in the apps themselves. I am not sure if this is the only cause of the
bug, but I haven't been able to reproduce it in test programs by writing a full
16-bits at a time.


-- 
Configure bugmail: http://bugs.openembedded.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.