Dale Grover wrote: > Is this approach useful? > http://linuxjunk.blogspot.com/2009/01/beagleboard-gpio-input-driverless.html > Well, I took the SD card into work, where I have an SD card reader on my desktop computer, and hacked the passwords on Matt's Debian install. Then, I had problems getting the computer on my network. I've never gotten good results with DHCP, so I just use static IPs here. But I couldn't seem to turn off the darn DHCP client daemon on the Beagle. Well, after more research, and giving up entirely on the graphical network configure tool, which just can't set up a system properly WITHOUT DHCP, I finally got it on the network.
The program you gave the link to shows great promise, but may need a fair bit of work. It DOES compile, so Matt's Debian install has all the required libraries and includes. The program as written locks up the Beagle completely, which is no surprise, as it takes over a whole I/O bank. This program may have been set up for a Beagle running entirely off the POP memories on the OMAP chip, and not for a system running out of an SD card. I'll try to get some more info on what configuration this program works on, as it IS supposed to be for the Beagle board. So, on to figuring out a LOT more about the OMAP chip. Ummm, did you know the Technical Ref Manual for the OMAP3530 is 3517 pages long? That is NOT a typo - Three THOUSAND five Hundred pages! YIKES!!!!! Although the OMAP has 192 I/O pins, the Beagle board only brings out a VERY small number of them for general use. That is kind of a problem. Also, they bring out a wild mess of them, with very few numerically adjacent pins. There is no complete aligned byte worth of bits brought out. There is slightly more than one NON-aligned byte available. I wanted an aligned byte so that I could do a single byte write (not even sure that is permitted) to the I/O port register and have it show up a full byte in correct bt order. What I will have to do is take the parallel port byte and shift it over a couple bits and do a 16-bit write to the port. This is still a lot better than having to mask each bit from a source register to the correct I/O port bit. ARM CPU's much prefer a linear string of instructions over loops and conditional branches. Jon ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
