Re: [PATCH net-2.6.24] remove bogus qeth type check

[David Miller <davem@davemloft.net>]

From: Stephen Hemminger <shemminger@linux-foundation.org>
Date: Thu, 11 Oct 2007 13:24:58 -0700

> Not sure why this crept in with hard-header-ops change.
> Just kill it. Need to start doing cross-compiles for s390...
> 
> Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
> 
> --- a/drivers/s390/net/qeth_main.c	2007-10-10 08:26:48.000000000 -0700
> +++ b/drivers/s390/net/qeth_main.c	2007-10-11 13:21:17.000000000 -0700
> @@ -6581,9 +6581,6 @@ qeth_hard_header_parse(const struct sk_b
>  	const struct qeth_card *card;
>  	const struct ethhdr *eth;
>  
> -	if (dev->type != ARPHRD_IEEE802_TR)
> -		return 0;
> -
>  	card = qeth_get_card_from_dev(skb->dev);
>  	if (card->options.layer2)
>  		goto haveheader;

Stephen this change is not right.

In order to preserve the semantics that existed before your
headerops changes, this test must be there.  From the original
qeth_main.c code:

	if (qeth_get_netdev_flags(card) & IFF_NOARP) {
		dev->rebuild_header = NULL;
		dev->hard_header = NULL;
		dev->header_cache_update = NULL;
		dev->hard_header_cache = NULL;
	}
 ...
	if (card->options.fake_ll &&
		(qeth_get_netdev_flags(card) & IFF_NOARP))
			dev->hard_header = qeth_fake_header;
	if (dev->type == ARPHRD_IEEE802_TR)
		dev->hard_header_parse = NULL;
	else
		dev->hard_header_parse = qeth_hard_header_parse;

With the way you set up headerops in the driver, the only way
to preserve the ARPHRD_IEEE802_TR guard, is to keep the
test that you added to qeth_hard_header_parse().

Andrew, please go back to using the first patch you wrote to
fix the build, alternative resync with net-2.6.git as that
version of the build fix is there too.

Thans.