This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

More on serial port issues.

I'm still stuck moving ttyO2 to ttyO1 as console/login on my custom board. It's getting a tad weird.

Console output during kernel boot works a treat, no problems. Getty reports for duty at the end of Inittab asking for the root login. That's where it ends. No input accepted.

However, during my experimentation with some other issues I got it working - as long as:

a) I disable generic frame buffer support (.config)

b) I query fbset -i TWICE in the rcS script. (it reports no fb0 found).

c) I also move getty to before the rcS scipt executes (i.e. getty is not at the end of the script but the start).

In the above case getty starts early while output from the scripts continues on the serial port for a couple 100 chars. Then, miraculously, ttyO1 accepts input, I can log on and everything works fine. Change something a little so a few more or less characters get sent to the console after getty starts and it no longer accepts input, as before.

Everything points to a wacky omap serial driver (omap_serial.c). I removed all hadshaking, just in case - no difference.

My question: I want to stop and restart the driver after the kernel has loaded, before I execute getty (I'm hoping reinitializing the diver will fix it). So far I have not succeeded in this simple task. How do you stop and restart the ttyO1 driver ? I tried making a node for it in /dev but what is the Major device number - I can't find anything in the sources ?

Rainier

  • Problem solved.

    The RS232 driver on the board can be partly disabled via an enable line that happened to be wired to SPI1_CS2. UBOOT had this line in the correct polarity while sometime after Linux boot the line flipped, disabling RX. This flip only happens if the SHARP-LQ43 panel is specified in the .Config and this driver is loaded.