On Sun, Dec 19, 2010 at 01:57:15AM +0100, Michał Mirosław wrote:
quoted text
> On Thu, Dec 16, 2010 at 11:27:29PM +0000, Ben Hutchings wrote:
> > On Wed, 2010-12-15 at 23:24 +0100, Michał Mirosław wrote:
> > > Introduce NETIF_F_RXCSUM to replace device-private flags for RX checksum
> > > offload. Integrate it with ndo_fix_features.
> > > 
> > > ethtool_op_get_rx_csum() is removed altogether as nothing in-tree uses it.
> > Not explicitly, but what about drivers that turn TX and RX checksumming
> > on and off together?  They are implicitly covered by the case which
> > you're removing:
> > [...]
> > > -	case ETHTOOL_GRXCSUM:
> > > -		rc = ethtool_get_value(dev, useraddr, ethcmd,
> > > -				       (dev->ethtool_ops->get_rx_csum ?
> > > -					dev->ethtool_ops->get_rx_csum :
> > > -					ethtool_op_get_rx_csum));
> > [...]
> I had an impression that drivers not defining get_rx_csum() had no
> RX checksumming or didn't allow it to be toggled. I'll have a closer look
> on this.

All drivers that have set_rx_csum() also define matching get_rx_csum().

Anyway, the code this is removing is broken in that it assumes TX csum
capability implying RX csum capability.

As a side note, GRO receive functions (TCPv4 and v6 - no other protocols
use it) are checking skb->ip_summed != CHECKSUM_NONE before allowing the
packet to be passed on, so it's not even necessary to force disabling GRO
when RX csum is off.

Best Regards,
Michał Mirosław
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html