Unexpected/shorter CS when interfacing LCD controller

[Mike Wolfram <mike.wolfram@xxxxxx>]

Hi,

on our custom AT92RM9200 board we have a SED1335 LCD controller on CS5. After 
a long time of experiments we are now quite sure to have the correct values 
we have to program into the SMC registers. While most of our devices work 
fine we have some which show a different/unexpected behaviour.

The Linux driver we use for the SED1335 writes all 9600 bytes for the graphics 
mode at once. For framebuffer support we allocated some memory which is 
copied every 100ms byte by byte to the LCD controllers memory. On some boards 
we can see from time to time that the display has an offset of 3/4 of the 
screen. Some have it more, others less often. Looking on the signals we found 
out that the chip select is in that cases much shorter than in other write 
operations.

Is there anybody who has similar experiences? Might there be anything else in 
the Linux kernel which tries to access CS5. CF support is disabled in the 
kernel.

--
Regards,
Mike Wolfram


-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php