Driver for Sharp LH155BA

Driver for Sharp LH155BA

[Richard Wolf]

Hi,

I'm writing a driver for the Sharp LH155BA LCD driver.  This is a driver
designed for controlling small/medium lcd panels (up to 128 x 64 pixels).  

I'm trying to decide whether it is sensible to write a frame buffer driver
for a device of this nature.  Does the frame buffer abstraction make sense
for a display this small?  

On our target hardware it is not possible to access the display memory
directly.  Rather, all accesses to display memory and other control
registers are through a small number of io ports.  Therefore it is not
possible to map display memory.  Is this relevant?  

Thank you for your help.

Regards,

Richard Wolf
Card Access Services Pty Ltd
Level 2, 7-9 Albany Street
Crows Nest  NSW  2065
Phone: +612 9906 7209
Fax: +612 9906 7218
www.cardaccess.com.au

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Driver for Sharp LH155BA

[Antonino A. Daplas]

Richard Wolf wrote:
> Hi,
> 
> I'm writing a driver for the Sharp LH155BA LCD driver.  This is a driver
> designed for controlling small/medium lcd panels (up to 128 x 64 pixels).  
> 
> I'm trying to decide whether it is sensible to write a frame buffer driver
> for a device of this nature.  Does the frame buffer abstraction make sense
> for a display this small?  

The dimensions of the display does not matter.

> 
> On our target hardware it is not possible to access the display memory
> directly.  Rather, all accesses to display memory and other control
> registers are through a small number of io ports.  Therefore it is not
> possible to map display memory.  Is this relevant?  

Yes, if the display memory is not mappable, then it won't be usable as
a driver for X, for example.

You can however allocate a shadow buffer in system RAM which can be mmapp'ed.
Then on intervals, perhaps by interrupt or timer, update the physical screen.
Unfortunately, this is inefficient, as you may be updating the screen even if
the shadow buffer haven't changed.  But that is the best you can do without
hacking the kernel.

Another option is to just forget about mmap, and create your own fb_read and
fb_write hooks.  This is more efficient, but not all apps use this method to
update the framebuffer. The driver will be usable as a console though.

Tony


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Driver for Sharp LH155BA

[Richard Wolf]

 

> -----Original Message-----
> From: Antonino A. Daplas [mailto:adaplas@gmail.com] 
> Sent: Tuesday, 1 August 2006 10:36 AM
> To: linux-fbdev-devel@lists.sourceforge.net
> Cc: rwolf@cardaccess.com.au
> Subject: Re: [Linux-fbdev-devel] Driver for Sharp LH155BA
> 
> Richard Wolf wrote:
> > Hi,
> > 
> > I'm writing a driver for the Sharp LH155BA LCD driver.  This is a 
> > driver designed for controlling small/medium lcd panels (up 
> to 128 x 64 pixels).
> > 
> > I'm trying to decide whether it is sensible to write a frame buffer 
> > driver for a device of this nature.  Does the frame buffer 
> abstraction 
> > make sense for a display this small?
> 
> The dimensions of the display does not matter.
> 
> > 
> > On our target hardware it is not possible to access the 
> display memory 
> > directly.  Rather, all accesses to display memory and other control 
> > registers are through a small number of io ports.  
> Therefore it is not 
> > possible to map display memory.  Is this relevant?
> 
> Yes, if the display memory is not mappable, then it won't be 
> usable as a driver for X, for example.
> 
> You can however allocate a shadow buffer in system RAM which 
> can be mmapp'ed.
> Then on intervals, perhaps by interrupt or timer, update the 
> physical screen.
> Unfortunately, this is inefficient, as you may be updating 
> the screen even if the shadow buffer haven't changed.  But 
> that is the best you can do without hacking the kernel.
> 
> Another option is to just forget about mmap, and create your 
> own fb_read and fb_write hooks.  This is more efficient, but 
> not all apps use this method to update the framebuffer. The 
> driver will be usable as a console though.
> 
> Tony

Thanks Tony

Presumably if I take the shadow buffer route then I could maintain _two_
shadow buffers.  This would allow me to detect whether the user accessible
buffer had been changed and update the display only when necessary.  

Regards,

Richard


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Driver for Sharp LH155BA

[Antonino A. Daplas]

Richard Wolf wrote:
> 
> Presumably if I take the shadow buffer route then I could maintain _two_
> shadow buffers.  This would allow me to detect whether the user accessible
> buffer had been changed and update the display only when necessary.  

Yes, that will work too.

Tony

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Driver for Sharp LH155BA

[Markus Schorer]

Antonino A. Daplas wrote:
> Richard Wolf wrote:
>> Presumably if I take the shadow buffer route then I could maintain _two_
>> shadow buffers.  This would allow me to detect whether the user accessible
>> buffer had been changed and update the display only when necessary.  
> 
> Yes, that will work too.
> 

check for the epson SED15xx (not sure) driver. sounds like similar 
hardware, same approach.

markus
-- 
_______________________________________________________________________
plain GmbH        | daiserstrasse 15      | 81371 münchen
fon 089.540.149.0 | fax 089.540.149.44    | http://www.plain.de

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Driver for Sharp LH155BA

[Richard Wolf]

 

> -----Original Message-----
> From: linux-fbdev-devel-bounces@lists.sourceforge.net 
> [mailto:linux-fbdev-devel-bounces@lists.sourceforge.net] On 
> Behalf Of Markus Schorer
> Sent: Tuesday, 1 August 2006 4:58 PM
> To: linux-fbdev-devel@lists.sourceforge.net
> Subject: Re: [Linux-fbdev-devel] Driver for Sharp LH155BA
> 
> Antonino A. Daplas wrote:
> > Richard Wolf wrote:
> >> Presumably if I take the shadow buffer route then I could maintain 
> >> _two_ shadow buffers.  This would allow me to detect 
> whether the user 
> >> accessible buffer had been changed and update the display 
> only when necessary.
> > 
> > Yes, that will work too.
> > 
> 
> check for the epson SED15xx (not sure) driver. sounds like 
> similar hardware, same approach.
> 
> markus
> --

Excellent, thanks.

Richard

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV