Hello Ludovic,

On Tue, Jul 28, 2015 at 03:12:53PM +0200, Ludovic Desroches wrote:

On Tue, Jul 28, 2015 at 02:48:09PM +0200, Linus Walleij wrote:

quoted

On Tue, Jul 28, 2015 at 9:48 AM, Ludovic Desroches
[off-list ref] wrote:

quoted

From: David Dueck <redacted>

Not all gpio banks are necessarily enabled, in the current code this can
lead to null pointer dereferences.

[   51.130000] Unable to handle kernel NULL pointer dereference at virtual address 00000058

(...)

quoted

Signed-off-by: David Dueck <redacted>
Acked-by: Ludovic Desroches <redacted>
Acked-by: Alexandre Belloni <redacted>
CC: Nicolas Ferre <redacted>
CC: Boris Brezillon <redacted>
CC: linux-arm-kernel at lists.infradead.org
CC: linux-kernel at vger.kernel.org
---

This patch fixes a oops in the kernel because of a NULL pointer in a table.
Having a NULL pointer in this table is the normal behavior if a PIO controller
is not enabled. So this fix is not a quick and dirty hack, it's usual to skip
an entry from a table if it is not filled.

Fair enough, better too many checks than too few.

Is this a regression to v4.2 that should go to stable or v4.3 material?

Yes it is a regression from v4.0, it applies well on v4.0.9

Fixes: a0b957f306fa ("pinctrl: at91: allow to have disabled gpio bank")
Cc: stable at vger.kernel.org # 4.0

a0b957f306fa have a stable tag up to 3.18, should this patch inherit a 
stable tag up to 3.18 instead of only up to 4.0 ?

Sylvain