[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <87fxoyal9j.fsf@natisbad.org>
Date: Thu, 21 Aug 2008 09:18:48 +0200
From: arno@...isbad.org (Arnaud Ebalard)
To: netdev@...r.kernel.org
Cc: Al Viro <viro@...iv.linux.org.uk>,
David Miller <davem@...emloft.net>
Subject: [bisected] Weird sysctl regression
Hi,
While updating some IPv6-related patches for net-2.6, I hit the
following issue on a test laptop (G4-based 12" powerbook). After
compiling current net-2.6 kernel on another ppc device (G4-based 15"
powerbook), I have the same issue. Note that:
- it never shows up on any of the x86 devices I have.
- hardware of both ppc devices are similar (bcm4306, sungem, ...)
i.e. they do not only have endianness in common (and i had no sparc
device to test if it is an endianness issue or a coincidence)
Usually, I have the following on those devices:
arno@...it:~$ ls /proc/sys/net/
bridge core ipv4 ipv6 netfilter nf_conntrack_max unix
arno@...it:~$ ls /proc/sys/net/ipv6
bindv6only conf icmp ip6frag_high_thresh ip6frag_low_thresh ip6frag_secret_interval ip6frag_time ip6_queue_maxlen mld_max_msf neigh route
But with current net-2.6, I have:
arno@...it:~$ ls /proc/sys/net/
bridge core ipv4 ipv6 ipv6 netfilter nf_conntrack_max unix
arno@...it:~$ ls /proc/sys/net/ipv6/
ip6_queue_maxlen
i.e. ipv6 is there twice (sometimes even three times), and most of the
keys I expect to find here have vanished.
I bisected it to 9043476f726802f4b00c96d0c4f418dde48d1304 ([PATCH]
sanitize proc_sysctl) which is in the middle of a set of fs/sysctl
related patches.
My kernel config is attached and I can do additional tests if you need
(today or on sunday; I may be unresponsive tomorrow and the day after).
Cheers,
a+
View attachment "dot-config.txt" of type "text/plain" (53825 bytes)
Powered by blists - more mailing lists