Re: kern/51531 (Recent networking regression affecting installs)

[Andreas Gustafsson <gson%gson.org@localhost>]

The following reply was made to PR kern/51531; it has been noted by GNATS.

From: Andreas Gustafsson <gson%gson.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: kern/51531 (Recent networking regression affecting installs)
Date: Tue, 28 Sep 2021 20:51:53 +0300

 This five-year-old issue affecting sk(4) is still present in -current
 as of source date 2021.09.28.11.38.07.
 
 This time I happened to be using a kernel built with "options DEBUG",
 and this enabled an additional kernel message that might be helpful in
 tracking down the cause of the problem.
 
 As before, an attempt to install NetBSD over the network using an
 sk(4) interface failed.  The INSTALL kernel successfully booted over
 TFTP, but when sysinst (re-)configured the interface and attempted to
 download the first install set over HTTP, the connection timed out.
 
 To diagnose the failure, "ifconfig -a" was run from the shell more
 than ten minutes after sysinst had terminated, and even though
 "ifconfig -a" is supposed to only display things, not change them, and
 there had been no recent changes to the link state, it had the side
 effect of triggering the kernel debug message "sk0: link state UP (was
 DOWN)":
 
   # ifconfig -a
   sk0: link state UP (was DOWN)
   lo0: flags=0x8049<UP,LOOPBACK,RUNNING,MULTICAST> mtu 33624
           inet 127.0.0.1/8 flags 0x0
           inet6 ::1/128 flags 0x20<NODAD>
           inet6 fe80::1%lo0/64 flags 0x0 scopeid 0x1
   sk0: flags=0x8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
           ec_capabilities=5<VLAN_MTU,JUMBO_MTU>
           ec_enabled=0
           address: 00:17:31:16:dd:fa
           media: Ethernet autoselect (1000baseT full-duplex)
           status: active
           inet 10.169.3.2/24 broadcast 10.169.3.255 flags 0x1<TENTATIVE>
           inet6 fe80::217:31ff:fe16:ddfa%sk0/64 flags 0x2<TENTATIVE> scopeid 0x2
 
 To me this looks like sk(4) is failing to update the link state when
 it actually changes, and instead updating it much later at the point
 when "ifconfig -a" is run.
 -- 
 Andreas Gustafsson, gson%gson.org@localhost