|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |
To: linux-net@vger.rutgers.edu Date: Tue, 14 Feb 1995 19:02:33 +0100 (MEZ) From: "17 Linux Adm." <marekm@i17unixb.ists.pwr.wroc.pl> In-Reply-To: <m0rdNGA-0001ktC@titan.westfalen.de> from "Johannes Stille" at Feb 11, 95 08:13:00 pm Johannes Stille wrote: > What do you mean by "network adresses"? The host part of the address is > zero? Are these addresses from subnets of your network or from other > networks? The problem addresses end with something like .32 or .128 (not .0). They are from other (sometimes very distant) networks. I know that at least some of them are really network addresses (not hosts) with netmasks like 255.255.255.224. > >Most of these routes are redundant (the gateway is the same as the > >default one) and maybe they shouldn't be added at all? > > In this case, the gateway should not send NET REDIRECT messages. But the gateway has no way to know it is default for me. There is another gateway which could be default for me as well. What about an extra check? Even if some other system is broken, we could fix it by not adding redundant routes. > Just because of the possible problems with subnetting, it is officially > recommended to treat NET REDIRECT as HOST REDIRECT. And this is what > Linux does. But the comment around line 360 of linux/net/inet/icmp.c says we should use ICMP_ADDRESS to get the subnet mask - but we don't. This is exactly what I am talking about... BTW, the first part of my original message (about the acked < copied tcp bug) seems much less interesting... :) Seriously, what I can do to help finding the bug? (For example, I might make that printk display some extra information, values of some variables, etc. - and report results here in a week or two, when the bug manifests itself again). This system is quite heavily loaded - good for testing... Regards, -- Marek Michalkiewicz <marekm@i17linuxa.ists.pwr.wroc.pl>
|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |