On Thu, Mar 24, 2011 at 01:12:02AM +0100, Witold Baryluk wrote: > Please fix this, as it is very simple change to add nanosleep or something.
While adding an arbitrary delay would be easy, what would you have the ѕpecific delay be? Consider the difference in ping's behavior between handling locally sourced ICMP_FRAG_NEEDED and similar messages coming from remote hosts. > > IMHO also icmp_seq should be incremented each time. Again, consider what would happen if we did this every time we received a ICMP_FRAG_NEEDED, and the message originated with a remote host. I think incrementing icmp_ѕeq should be done whenever an icmp packet is emitted, which, as you observed, does not happen if we violate the local link's mtu. noah
signature.asc
Description: Digital signature
