[Lcdproc] Please test: hd44780 backlight update
bsdfan at nurfuerspam.de
Tue Feb 24 22:12:22 UTC 2009
Markus Dolze wrote:
> After analyzing Sascha's problem with 40x4 winamp-wired display I had
> a closer look how the backlight is used.
> I found that:
> 1. In render.c the backlight state is determined.
> 2. drivers.c passed the BL state to all drivers, if supported
> 3. The hd44780 passed the BL to the connection type, if supported
> 4. The conntection type sets the hardware accordingly.
> Besides a possible problem in 3. because the hardware is asked to set
> the backlight off if the user specified no backlight being available
> in the config file, the whole chain is executed with every screen
> Looking at how other drivers handle this I found some implement an
> "incremental" backlight change (only update hardware if the the state
> has really changed) while others do not.
> I decided to implement this "incremental update" for the hd44780
> driver as well. Patch attached.
> As this applies to all connection types, I ask developers and users
> with hd44780 type displays and backlight to test the impact of this
> change. I already tested with lcd2usb and winamp (but without
> switchable backlight) and found no problems so far.
> Please also test what happens if "Backlight=off" is set in LCDd.conf.
Commited this patch to CVS.
More information about the LCDproc