Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some users, set hold time to zero on bridge so it always does
flooding. This is usually when using it with wireless.  The new RCU
based code changed the behaviour so the bridge would not flood for
one GC interval. This patch restores the original behaviour.
Signed-off-by: David S. Miller <davem@davemloft.net>

This can elicit ICMP message output and thus result in a
deadlock.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Andre Tomt <andre@tomt.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Harald Welte <laforge@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

In an attempt to return a locked socket ax25_get_socket() was calling
lock_sock() with a spinlock held, bad idea.  Making matters worse it's
only user is running in bottom half context resulting in a potencial
attempt to sleep in bottom half context, so fix the locking there as well.
Signed-off-by: David S. Miller <davem@davemloft.net>

netlink_autobind has always set nlk_sk(sk)->groups to zero.  This is
unnecessary because sk_alloc already zeroes the entire structure.
Since a socket can only be bound once netlink_autobind doesn't need
to zero groups at all.

This had been safe until I added mc_list.  Now it is possible for
netlink_bind to race against netlink_autobind running on the same
socket on another CPU.  The result would be a socket that's on
mc_list with groups set to zero.  This socket will be left on the
list even after it is destroyed.

The fix is to remove the zeroing in netlink_autobind.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

While replacing dst_pmtu in ip6_output I found this little gem.  In
ip6_forward we're not reloading the dst pointer after calling
xfrm6_route_forward.  So all subsequent dereferences of dst will
refer to its pre-IPsec value.

The solution is of course to refresh its value.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

I'm now cleaning up all users of dst_pmtu with the aim of replacing
dst_pmtu with dst_mtu.  I'm going to start with the ones that actually
fix bugs.

This patch fixes the length calculation in icmp_send.  As it is
we're overestimating the space available by including the space
that would be used up by IPsec encapsulation.

IPv6 doesn't have this problem since its calculation is based
on 1280 instead of the PMTU.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

When cleaning up the remaining users of dst_pmtu I noticed that
local_df wasn't being treated correctly in IPsec.  In fact, if
you socket's dst went over IPsec, local_df is essentailly ignored.

This patch fixes that.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

The patch below fixes an smp race that happens on such systems under
heavy load.
This bug was reported and solved by Steve Herrell
<steve_herrell@yahoo.ca>
Signed-off-by: Bart De Schuymer <bdschuym@pandora.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use dev_get_flags() in inet6_fill_ifinfo() to fetch interface flags
to ensure correctly reporting IFF_PROMISC and IFF_ALLMULTI flags.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

ip_tables failed to recognize IPT_RETURN because it was defined relative
to NF_MAX_VERDICT (which changed) and returned it to nf_iterate(). The
old value of IPT_RETURN matches NF_REPEAT, so the hook was called again
and again.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Cset exclude: torvalds@ppc970.osdl.org|ChangeSet|20050311080305|44342

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into suse.cz:/home/perex/bk/linux-sound/work

into suse.cz:/home/perex/bk/linux-sound/linux-sound

AC97 Codec
This patch series adds support for the WM9713/WM9714 family of AC97
codecs. This family is different from 'standard' AC97 codecs in that the
default codec power state is 'off'. i.e. performing a register reset
will power the device down.

This patch also adds better support for larger single/double channel
enumerated mixer types.
Signed-off-by: Liam Girdwood <liam.girdwood@wolfsonmicro.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Fix a goofup from the recent sort consolidation patches.

- Subtractions in comparison functions could overflow the return type for
  64-bit pointers.  Make it right again.

- Add missing size arg to ia64 extable swap function.
Signed-off-by: Matt Mackall <mpm@selenic.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This patch fixes a bug with the recently added printk-times feature.

In the case where a printk consists of only the log level (followed
subsequently by printks with more text for the same line), the printk-times
code doesn't correctly recognize the end of the string, and starts emitting
chars at the 0 byte at the end of the string.

The patch below fixes this problem.  It also adjusts the handling of
printed_len in the routine, which was affected by the printk-times feature.
Signed-off-by: Tim Bird <tim.bird@am.sony.com>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This patch updates the sungem driver. I reworked all of the PM stuff, making it
less prone to races, probably simpler to read as well, and I no longer shut the
PHY down when the interface is down so that things like laptop-net no longer
die (the gain in power consumption was minimal, not worth the pain). I also
implemented basic WOL support.

There is still something I'm not totally happy with in the locking
(explained in the comment at the beginning), basically too much locking and a
couple of places with delays in locks. I will try to improve these later on.

It also adds a fix for a MAC reset issue when the receiver gets stuck.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Randy Dunlap <rddunlap@osdl.org>
Signed-off-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Christophe Lucas <c.lucas@ifrance.com>
Signed-off-by: Domen Puncer <domen@coderock.org>
Signed-off-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
Signed-off-by: Domen Puncer <domen@coderock.org>
Signed-off-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

Moving nr_init_timers prototype to where all the other prototypes
already are.
Signed-off-by: David S. Miller <davem@davemloft.net>

ROSE wasn't verifying the ndigis argument of a new route resulting in a
minor security hole.
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: David S. Miller <davem@davemloft.net>

this patch correctly removes the pci_find_device() used by the fore200e
driver.  the __init/__exit remains a bit clunky since we need to preserve
sbus support.
Signed-off-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

Otherwise declarations for disable_irq()/enable_irq() will
not be obtained on platforms not doing CONFIG_GENERIC_HARDIRQS.
Signed-off-by: David S. Miller <davem@davemloft.net>

This check is wrong, gcc optimizes it away:

                if ((len -= sizeof(pi)) > len)
			return -EINVAL;

This could be responsible for the BUG. If len is 2 or 3 and TUN_NO_PI
isn't set it underflows. alloc_skb() allocates len + 2, which is 0 or
1 byte. skb_reserve tries to reserve 2 bytes and things explode in
skb_put.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds AGP support for the U3 northbridge used in Apple G5
machines to drivers/char/agp/uninorth-agp.c.  This patch is based on
earlier work by Jerome Glisse.  With this patch, the driver works in
both ppc32 and ppc64 kernels.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Otherwise declarations for disable_irq()/enable_irq() will
not be obtained on platforms not doing CONFIG_GENERIC_HARDIRQS.
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: David S. Miller <davem@davemloft.net>

Cacheing the current position reduces complexity from O(n^2)
to O(n).
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Low level optimization of the comparison with zero address.
On most cpu's faster to use unrolled or rather than a loop
and memcmp. This is in the fastpath of bridge forwarding.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Optimize bridge forwarding table for the fastpath of updating
an existing entry. Use RCU to find the entry and change it's time.
Fallback to normal locking for insert. This gives about a 1/3
improvement in packets forwarding per second on SMP.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>