On Sun, Feb 23, 2003 at 01:12:17AM -0800, David S. Miller wrote: > This makes the locking much simpler. Add to this the per-cpu socket > idea I proposed and we arrive at the patch below (against 2.5.x, > easily ported to 2.4.x just remove the cpu_possible() stuff). Hmm... Is this issue (ICMP packets being dropped due to lazy locking) supposed to be fixed in current 2.4 (2.4.22pre4, for example)? I seem to still be seeing ICMP packet loss with duplextest which disappears if I bind the NIC interrupt to one of the two CPUs (e100 driver). I checked the code and the patch appears to be in there. I wonder what is causing the loss now... tcpdump on the machine being pinged shows it not responding to all of the ICMP it receives, as was happening before. Simon- - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
