[PATCH v3 1/3] ARM: BCM5301X: initial support for the BCM5301X/BCM470X SoCs with ARM CPU

[Will Deacon]

On Tue, Jul 30, 2013 at 09:54:04PM +0100, Hauke Mehrtens wrote:
> On 07/30/2013 03:36 PM, Will Deacon wrote:
> > On Sat, Jul 27, 2013 at 08:49:16PM +0100, Arnd Bergmann wrote:
> >> On Friday 26 July 2013, Will Deacon wrote:
> >>>>> At least, we need a pretty good explanation of what exactly is causing
> >>>>> these spurious aborts before we start ignoring them unconditionally like
> >>>>> this. You're effectively masking an extremely serious error indicator with
> >>>>> this change.
> >>>>
> >>>> This fault occurs once every boot sometime early in the boot process,
> >>>> but the actual time this happens varies randomly.
> >>>
> >>> Well that's interesting in itself. It sounds like we don't know *for sure*
> >>> whether the abort is triggered by Linux. Since the abort is imprecise, the
> >>> timing will vary.
> >>
> >> It might be possible to find out the culprit if you just enter an endless loop
> >> in the early kernel boot code. If you enter the loop before Linux does something
> >> wrong, it won't crash, otherwise it will. After that, you could bisect the
> >> boot process by moving the busy loop around.
> >>
> >> If it even crashes at the point where Linux gets entered, it's a bug in the
> >> boot loader.
> > 
> > Can you give Arnd's suggestion a go please Hauke?
> 
> Hi
> 
> I just tried that, but it did not crash in the loop. :-(

Turn that frown around -- this is a useful result! It means that the abort
is likely caused by the kernel, rather than the boot loader (although we
still can't rule it out, it depends where you put the loop).

> Now, with 3.11-rc3 it crashed reproducible after Freeing unused kernel
> memory, see this:
> 
> [    0.859198] Freeing unused kernel memory: 1124K (c0260000 - c0379000)
> [    0.866298] Unhandled fault: imprecise external abort (0x1c06) at
> 0xb6f8b005
> [    0.873538] Kernel panic - not syncing: Attempted to kill init!
> exitcode=0x00000007
> 
> Do you have some more informations about this type of error?

Hmm, so that's now exploding on a user address. Is this an A9-based SoC? If
so, can you try disabling PL310 (CONFIG_CACHE_L2X0) and see if it makes a
different please?

Cheers,

Will