Steven Hayter <[EMAIL PROTECTED]> : [...] > It's the old version of drivers/net/ibm_emac/ibm_emac_core.c (2.6.13) > the memcpy's are near the comment 'MAL is braindead'. The new version > appearing in 2.16.15 looks much better, but doesn't look like it does > jumbo frames, hence I'm trying to tweak the old one.
Don't be shy: tweak the new one :o) [...] > Since this device doesn't appear to do that for us, would we copy the > header section to the data part of the skb and increase the the > skb_shinfo(skb)->frags[n].page_offset by the same amount. Not quite > zero-copy, but copying a few protocol headers is better than the whole > rx descriptor. > > On the right path, or still way off? On the right path but I don't know if the data part, and thus the copy of the header, is really necessary. I'd give a try to skb_reserve() on each page of the ring and assemble a skb without any data part. -- Ueimor - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
