On Mon, Jul 27, 2015 at 2:21 PM, Tyler Baker <tyler.baker@linaro.org> wrote:

On 27 July 2015 at 09:26, Fabio Estevam <festevam@gmail.com> wrote:

On Mon, Jul 27, 2015 at 11:40 AM, Tyler Baker <tyler.baker@linaro.org> wrote:

It found "net: fec: Ensure clocks are enabled while using mdio bus" (v5) because I bisected between v4.1 and 83a5f7f22ca. I am re-running the bisection between c9805b9986ed ("Revert "net: fec: Ensure clocks are enabled while using mdio bus") and 83a5f7f22ca.

It seems that the second bisect did not work either. It points to:
# first bad commit: [75406b3b17ee5a107b361b4065ad83fe42fff5a8] tty: Convert use of __constant_htons to htons

Yes this does not appear to be the culprit, I tried the revert and the issue still remains. If I checkout [347962eb1559] "Merge remote-tracking branch 'ipmi/for-next' in -next everything works fine, however if I checkout the next merge commit [c3e5e4fa730] "Merge remote-tracking branch 'tty/tty-next'" the imx6 platforms no longer boot, due to the garbling of console messages. I'll attempt another bisection the tty/tty-next tree to see if that offers any more insight.

It seems that reverting the following commit fixes the problem:

commit e95044ba4fee93f5ea8a1a24b2d921e148503833
Author: David Jander <david@protonic.nl>
Date:   Thu Jul 2 16:29:49 2015 +0200

    tty: serial: imx.c: Reset UART before activating interrupts
   
    If the UART has been in use before this driver was loaded, IRQs might be
    active and get fired as soon as we set the handler, which will crash
    in the spin_lock_irqsave(&sport->port.lock, flags) because port.lock is
    not initialized until the port is added at the end of probe.
   
    Signed-off-by: David Jander <david@protonic.nl>
    Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Thanks,

Fabio Estevam