This bug is pernicious... It occurs when kernel output (printk's) are normally disabled because of the call to
mxc_early_uart_console_disable() in drivers/tty/serial/mxc_uart_early.c.
We were able to narrow in on the problem using the attached patch.
With the patch, it was pretty easy to see where the problem occurred:
We didn't have p8 formatted because mkfs.ext4 doesn't seem to like
8 partitions or more on our Ubuntu Maverick machine (and we don't
know precisely what should go into the /device mount point).
Removing the mount statement from init.freescale.rc allowed us to
boot every time.
That said, this isn't the root cause. It appears that something else is
causing the system to get into a non-recoverable sleep.
This bug is pernicious... It occurs when kernel output (printk's) are normally disabled because of the call to uart_console_ disable( ) in drivers/ tty/serial/ mxc_uart_ early.c.
mxc_early_
We were able to narrow in on the problem using the attached patch.
With the patch, it was pretty easy to see where the problem occurred:
EXT4-fs (mmcblk0p8): VFS: Can't find ext4 filesystem
.... stalled here until hard-key press
We didn't have p8 formatted because mkfs.ext4 doesn't seem to like
8 partitions or more on our Ubuntu Maverick machine (and we don't
know precisely what should go into the /device mount point).
Removing the mount statement from init.freescale.rc allowed us to
boot every time.
That said, this isn't the root cause. It appears that something else is
causing the system to get into a non-recoverable sleep.