net: sxgbe: fix error paths in sxgbe_platform_probe()

We need to use post-decrement to ensure that irq_dispose_mapping is
also called on priv->rxq[0]->irq_no; moreover, if one of the above for
loops failed already at i==0 (so we reach one of these labels with
that value of i), we'll enter an essentially infinite loop of
out-of-bounds accesses.

Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
---
David, can you take this directly? Of the three samsung people listed
by get_maintainer.pl, one address bounces and another informed me
privately that he's not actually a maintainer of this anymore.

 drivers/net/ethernet/samsung/sxgbe/sxgbe_platform.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Rasmus Villemoes <linux@rasmusvillemoes.dk> :
> We need to use post-decrement to ensure that irq_dispose_mapping is
> also called on priv->rxq[0]->irq_no; moreover, if one of the above for
> loops failed already at i==0 (so we reach one of these labels with
> that value of i), we'll enter an essentially infinite loop of
> out-of-bounds accesses.
> 
> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>

(ok, i is signed)

Reviewed-by: Francois Romieu <romieu@fr.zoreil.com>

Someone messed with my review on 2014/03/25 and got it wrong. :o/

Two years after the initial submission, there is zero change regarding use
of sxgbe_core_ops for extension or manageability. The extra indirection is
ripe for removal during next net-next.

On Sun, Mar 27 2016, Francois Romieu <romieu@fr.zoreil.com> wrote:

> Rasmus Villemoes <linux@rasmusvillemoes.dk> :
>> We need to use post-decrement to ensure that irq_dispose_mapping is
>> also called on priv->rxq[0]->irq_no; moreover, if one of the above for
>> loops failed already at i==0 (so we reach one of these labels with
>> that value of i), we'll enter an essentially infinite loop of
>> out-of-bounds accesses.
>> 
>> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
>
> (ok, i is signed)
>
> Reviewed-by: Francois Romieu <romieu@fr.zoreil.com>
>

Thanks for reviewing, but just FTR I want to point out that it doesn't
matter whether i is signed or not in

  while (i--)

However, when i is signed, there's another slightly less popular variant
which is equivalent:

  while (--i >= 0)

(a precondition for their equivalence is that i has a non-negative value
before reaching the while statement).

Neither are equivalent to the almost-always broken

  while (--i)

Rasmus

From: Francois Romieu <romieu@fr.zoreil.com>
Date: Sun, 27 Mar 2016 10:22:54 +0200

> Two years after the initial submission, there is zero change regarding use
> of sxgbe_core_ops for extension or manageability. The extra indirection is
> ripe for removal during next net-next.

I completely agree, that stuff has to go.

From: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Date: Sat, 26 Mar 2016 22:24:09 +0100

> We need to use post-decrement to ensure that irq_dispose_mapping is
> also called on priv->rxq[0]->irq_no; moreover, if one of the above for
> loops failed already at i==0 (so we reach one of these labels with
> that value of i), we'll enter an essentially infinite loop of
> out-of-bounds accesses.
> 
> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>

Applied, thanks.

> David, can you take this directly? Of the three samsung people listed
> by get_maintainer.pl, one address bounces and another informed me
> privately that he's not actually a maintainer of this anymore.

That's awesome, another pump and dump driver submission.