Re: [RFC PATCH arm: initial TI-Nspire support]

From: Daniel Tang
Date: Tue Apr 09 2013 - 02:00:10 EST



On 09/04/2013, at 5:16 AM, Fabian Vogt <fabian@xxxxxxxxxxxxxx> wrote:

> The latest kernel it seems to get stuck at
> console_lock() in register_framebuffer (drivers/video/fbmem.c:1655)
> if the LCD-controller is enabled. (Early printk and serial console works fine)
> CONFIG_NO_HZ is not activated, it works completely.
> Could this be a kernel bug or is this an issue with our timer configuration?

Even with the CONFIG_NO_HZ option disabled, I've found that probing of our LCD hardware sometimes still fail with an error of -2.

Depending on how I arrange the device tree or the contents of the boot arguments or, as said, the CONFIG_NO_HZ option, I get different behaviours ranging from:
* Getting a -2 error during device probe (possibly due to a clock not being found).
* LCD hardware is found but boot hangs right after the driver is initialised.
* LCD hardware is apparently initialised correctly and text appears onscreen but user mode init hangs and I never get to a shell prompt.

This smells of a race condition that occurs before the clcd driver even gets to initialise. Some bug in our timer code? A OF or console bug perhaps? Though, I'm a bit hesitant to blame it on a kernel subsystem bug until it's clear that our code isn't at fault.

Any suggestions?--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/