nanogui: question about customized frame buffer driver


Previous by date: 10 Sep 2007 22:55:31 +0100 Re: Touch screen issues, David Leake
Next by date: 10 Sep 2007 22:55:31 +0100 Re: Corrupted Packet Nano-X, Greg Haerr
Previous in thread:
Next in thread:

Subject: question about customized frame buffer driver
From: "Allen Yang" ####@####.####
Date: 10 Sep 2007 22:55:31 +0100
Message-Id: <33B07717EB1E594F934D63BE3BDA5A3FA206D9@itexch01.nucomm.local>

Hi,

Unlike other LCD controllers, which use a part of "system memory" as its
"Display memory", my system is interfaced with a LCD panel (which has
LCD controller, LCD and Display memory). It's only connected to CPU by
Data Bus, NO address bus. The LCD panel's Display memory is accessible
only through "data register" of LCD controller. I have to write the
"address" of Display RAM first to the "data register" and then the
"data" to the "data register". In this case, the normal *mmap* done by
fbmem.c would not work here.

I am currently writing the frame buffer for the above LCD in uClinux
2.6.17. Basically I got some display on the LCD.

However, I still have some questions:

(1) When I kill the nano-X process, I got the following message, and
system dies.

# ****************munmap of non-mmaped memory by process 146 (nano-X):
40cc0000

What's the meaning of this message?

In the above, process 146 is the Nano-X process. 0x40cc0000 is starting
address of my frame buffer.


(2) I implemented a special command "FBIO_UPDATE_SCREEN" in function
lcdfb_ioctl(). In this command, I just send all the data in my frame
buffer to the LCD RAM.

In my Nano-X application, first I call Nano-X APIs (GrText()...) to
prepare the LCD display, then I call "GrFlush()". Then I call my own
lcdfb_ioctl() with command "FBIO_UPDATE_SCREEN". I thought the LCD
should display what I just prepared.

For example, I want to display the following screens:
SCREEN 1 ->	SCREEN 2 ->	SCREEN 3 ->	SCREEN 4 ->	SCREEN
5...

LCD always displays the previous SCREEN instead of the current SCREEN.

How could this happen?

Does "GrFlush()" send Nano-X buffered message into my LCD frame buffer?


Any suggestions will be highly appreciated!

Regards,

Allen







The information contained in this email and attachments to this email are the proprietary and confidential property 
of Nucomm, Inc.  The information is provided in strict confidence and shall not be reproduced, copied, or
used (partially or wholly) in any manner without prior, express written authorization of Nucomm, Inc.


Previous by date: 10 Sep 2007 22:55:31 +0100 Re: Touch screen issues, David Leake
Next by date: 10 Sep 2007 22:55:31 +0100 Re: Corrupted Packet Nano-X, Greg Haerr
Previous in thread:
Next in thread:


Powered by ezmlm-browse 0.20.