Starting from 3.6 we cache output routes for
multicasts only when using route to 224/4. For local receivers
we can set RTCF_LOCAL flag depending on the membership but
in such case we use maddr and saddr which are not caching
keys as before. Additionally, we can not use same place to
cache routes that differ in RTCF_LOCAL flag value.

	Fix it by caching only RTCF_MULTICAST entries
without RTCF_LOCAL (send-only, no loopback). As a side effect,
we avoid unneeded lookup for fnhe when not caching because
multicasts are not redirected and they do not learn PMTU.

	Thanks to Maxime Bizon for showing the caching
problems in __mkroute_output for 3.6 kernels: different
RTCF_LOCAL flag in cache can lead to wrong ip_mc_output or
ip_output call and the visible problem is that traffic can
not reach local receivers via loopback.
Reported-by: NMaxime Bizon <mbizon@freebox.fr>
Tested-by: NMaxime Bizon <mbizon@freebox.fr>
Signed-off-by: NJulian Anastasov <ja@ssi.bg>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The xfrm gc threshold value depends on ip_rt_max_size. This
value was set to INT_MAX with the routing cache removal patch,
so we start doing garbage collecting when we have INT_MAX/2
IPsec routes cached. Fix this by going back to the static
threshold of 1024 routes.
Signed-off-by: NSteffen Klassert <steffen.klassert@secunet.com>

Currently we can not flush cached pmtu/redirect informations via
the ipv4_sysctl_rtcache_flush sysctl. We need to check the rt_genid
of the old route and reset the nh exeption if the old route is
expired when we bind a new route to a nh exeption.
Signed-off-by: NSteffen Klassert <steffen.klassert@secunet.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Sparse complains about RTA_MARK which is should be host order according
to include file and usage in iproute.

net/ipv4/route.c:2223:46: warning: incorrect type in argument 3 (different base types)
net/ipv4/route.c:2223:46:    expected restricted __be32 [usertype] value
net/ipv4/route.c:2223:46:    got unsigned int [unsigned] [usertype] flowic_mark
Signed-off-by: NStephen Hemminger <shemminger@vyatta.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add flag to request that output route should be
returned with known rt_gateway, in case we want to use
it as nexthop for neighbour resolving.

	The returned route can be cached as follows:

- in NH exception: because the cached routes are not shared
	with other destinations
- in FIB NH: when using gateway because all destinations for
	NH share same gateway

	As last option, to return rt_gateway!=0 we have to
set DST_NOCACHE.
Signed-off-by: NJulian Anastasov <ja@ssi.bg>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add new flag to remember when route is via gateway.
We will use it to allow rt_gateway to contain address of
directly connected host for the cases when DST_NOCACHE is
used or when the NH exception caches per-destination route
without DST_NOCACHE flag, i.e. when routes are not used for
other destinations. By this way we force the neighbour
resolving to work with the routed destination but we
can use different address in the packet, feature needed
for IPVS-DR where original packet for virtual IP is routed
via route to real IP.
Signed-off-by: NJulian Anastasov <ja@ssi.bg>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Avoid checking nh_pcpu_rth_output in fast path,
abort fib_info creation on alloc_percpu failure.
Signed-off-by: NJulian Anastasov <ja@ssi.bg>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After "Cache input routes in fib_info nexthops" (commit
d2d68ba9) and "Elide fib_validate_source() completely when possible"
(commit 7a9bc9b8) we can not send ICMP redirects. It seems we
should not cache the RTCF_DOREDIRECT flag in nh_rth_input because
the same fib_info can be used for traffic that is not redirected,
eg. from other input devices or from sources that are not in same subnet.

	As result, we have to disable the caching of RTCF_DOREDIRECT
flag and to force source validation for the case when forwarding
traffic to the input device. If traffic comes from directly connected
source we allow redirection as it was done before both changes.

	Avoid setting RTCF_DOREDIRECT if IN_DEV_TX_REDIRECTS
is disabled, this can avoid source address validation and to
help caching the routes.

	After the change "Adjust semantics of rt->rt_gateway"
(commit f8126f1d) we should make sure our ICMP_REDIR_HOST messages
contain daddr instead of 0.0.0.0 when target is directly connected.
Signed-off-by: NJulian Anastasov <ja@ssi.bg>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We report cached pmtu values even if they are already expired.
Change this to not report these values after they are expired
and fix a race in the expire time calculation, as suggested by
Eric Dumazet.
Signed-off-by: NSteffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When a local tool like tracepath tries to send packets bigger than
the device mtu, we create a nh exeption and set the pmtu to device
mtu. The device mtu does not expire, so check if the device mtu is
smaller than the reported pmtu and don't crerate a nh exeption in
that case.
Signed-off-by: NSteffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Some protocols, like IPsec still cache routes. So we need to invalidate
the old route on pmtu events to avoid the reuse of stale routes.
We also need to update the mtu and expire time of the route if we already
use a nh exception route, otherwise we ignore newly learned pmtu values
after the first expiration.

With this patch we always invalidate or update the route on pmtu events.
Signed-off-by: NSteffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This commit prepares the use of rt_genid by both IPv4 and IPv6.
Initialization is left in IPv4 part.
Signed-off-by: NNicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We dont use jhash anymore since route cache removal,
so we can get rid of get_random_bytes() calls for rt_genid
changes.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since route cache deletion (89aef892), delay is no
more used. Remove it.
Signed-off-by: NNicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It is a frequent mistake to confuse the netlink port identifier with a
process identifier.  Try to reduce this confusion by renaming fields
that hold port identifiers portid instead of pid.

I have carefully avoided changing the structures exported to
userspace to avoid changing the userspace API.

I have successfully built an allyesconfig kernel with this change.
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: NStephen Hemminger <shemminger@vyatta.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We dont use jhash anymore since route cache removal,
so we can get rid of get_random_bytes() calls for rt_genid
changes.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since route cache deletion (89aef892), delay is no
more used. Remove it.
Signed-off-by: NNicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In ipv4_mtu there is some logic where we are testing for a non-zero value
and a timer expiration, then setting the value to zero, and then testing if
the value is zero we set it to a value based on the dst.  Instead of
bothering with the extra steps it is easier to just cleanup the logic so
that we set it to the dst based value if it is zero or if the timer has
expired.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>

Following lockdep splat was reported by Pavel Roskin :

[ 1570.586223] ===============================
[ 1570.586225] [ INFO: suspicious RCU usage. ]
[ 1570.586228] 3.6.0-rc3-wl-main #98 Not tainted
[ 1570.586229] -------------------------------
[ 1570.586231] /home/proski/src/linux/net/ipv4/route.c:645 suspicious rcu_dereference_check() usage!
[ 1570.586233]
[ 1570.586233] other info that might help us debug this:
[ 1570.586233]
[ 1570.586236]
[ 1570.586236] rcu_scheduler_active = 1, debug_locks = 0
[ 1570.586238] 2 locks held by Chrome_IOThread/4467:
[ 1570.586240]  #0:  (slock-AF_INET){+.-...}, at: [<ffffffff814f2c0c>] release_sock+0x2c/0xa0
[ 1570.586253]  #1:  (fnhe_lock){+.-...}, at: [<ffffffff815302fc>] update_or_create_fnhe+0x2c/0x270
[ 1570.586260]
[ 1570.586260] stack backtrace:
[ 1570.586263] Pid: 4467, comm: Chrome_IOThread Not tainted 3.6.0-rc3-wl-main #98
[ 1570.586265] Call Trace:
[ 1570.586271]  [<ffffffff810976ed>] lockdep_rcu_suspicious+0xfd/0x130
[ 1570.586275]  [<ffffffff8153042c>] update_or_create_fnhe+0x15c/0x270
[ 1570.586278]  [<ffffffff815305b3>] __ip_rt_update_pmtu+0x73/0xb0
[ 1570.586282]  [<ffffffff81530619>] ip_rt_update_pmtu+0x29/0x90
[ 1570.586285]  [<ffffffff815411dc>] inet_csk_update_pmtu+0x2c/0x80
[ 1570.586290]  [<ffffffff81558d1e>] tcp_v4_mtu_reduced+0x2e/0xc0
[ 1570.586293]  [<ffffffff81553bc4>] tcp_release_cb+0xa4/0xb0
[ 1570.586296]  [<ffffffff814f2c35>] release_sock+0x55/0xa0
[ 1570.586300]  [<ffffffff815442ef>] tcp_sendmsg+0x4af/0xf50
[ 1570.586305]  [<ffffffff8156fc60>] inet_sendmsg+0x120/0x230
[ 1570.586308]  [<ffffffff8156fb40>] ? inet_sk_rebuild_header+0x40/0x40
[ 1570.586312]  [<ffffffff814f4bdd>] ? sock_update_classid+0xbd/0x3b0
[ 1570.586315]  [<ffffffff814f4c50>] ? sock_update_classid+0x130/0x3b0
[ 1570.586320]  [<ffffffff814ec435>] do_sock_write+0xc5/0xe0
[ 1570.586323]  [<ffffffff814ec4a3>] sock_aio_write+0x53/0x80
[ 1570.586328]  [<ffffffff8114bc83>] do_sync_write+0xa3/0xe0
[ 1570.586332]  [<ffffffff8114c5a5>] vfs_write+0x165/0x180
[ 1570.586335]  [<ffffffff8114c805>] sys_write+0x45/0x90
[ 1570.586340]  [<ffffffff815d2722>] system_call_fastpath+0x16/0x1b
Signed-off-by: NEric Dumazet <edumazet@google.com>
Reported-by: NPavel Roskin <proski@gnu.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Multicast traffic allocates dst with DST_NOCACHE, but dst is
not inserted into rt_uncached_list.

This slowdown multicast workloads on SMP because rt_uncached_lock is
contended.

Change the test before taking the lock to actually check the dst
was inserted into rt_uncached_list.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Sylvain Munault reported following info :

 - TCP connection get "stuck" with data in send queue when doing
   "large" transfers ( like typing 'ps ax' on a ssh connection )
 - Only happens on path where the PMTU is lower than the MTU of
   the interface
 - Is not present right after boot, it only appears 10-20min after
   boot or so. (and that's inside the _same_ TCP connection, it works
   fine at first and then in the same ssh session, it'll get stuck)
 - Definitely seems related to fragments somehow since I see a router
   sending ICMP message saying fragmentation is needed.
 - Exact same setup works fine with kernel 3.5.1

Problem happens when the 10 minutes (ip_rt_mtu_expires) expiration
period is over.

ip_rt_update_pmtu() calls dst_set_expires() to rearm a new expiration,
but dst_set_expires() does nothing because dst.expires is already set.

It seems we want to set the expires field to a new value, regardless
of prior one.

With help from Julian Anastasov.
Reported-by: NSylvain Munaut <s.munaut@whatever-company.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
CC: Julian Anastasov <ja@ssi.bg>
Tested-by: NSylvain Munaut <s.munaut@whatever-company.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit caacf05e causes big drop of UDP loop back performance.
The cause of the regression is that we do not cache the local output
routes. Each time we send a datagram from unconnected UDP socket,
the kernel allocates a dst_entry and adds it to the rt_uncached_list.
It creates lock contention on the rt_uncached_lock.
Reported-by: NAlex Shi <alex.shi@intel.com>
Signed-off-by: NYan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As pointed out, there are places, that access net->loopback_dev->ifindex
and after ifindex generation is made per-net this value becomes constant
equals 1. So go ahead and introduce the LOOPBACK_IFINDEX constant and use
it where appropriate.
Signed-off-by: NPavel Emelyanov <xemul@parallels.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

performance profiles show a high cost in the IN_DEV_ROUTE_LOCALNET()
call done in ip_route_input_slow(), because of multiple dereferences,
even if cache lines are clean and available in cpu caches.

Since we already have the 'net' pointer, introduce
IN_DEV_NET_ROUTE_LOCALNET() macro avoiding two dereferences
(dev_net(in_dev->dev))

Also change the tests to use IN_DEV_NET_ROUTE_LOCALNET() only if saddr
or/and daddr are loopback addresse.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove unused includes after IP cache removal
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When a device is unregistered, we have to purge all of the
references to it that may exist in the entire system.

If a route is uncached, we currently have no way of accomplishing
this.

So create a global list that is scanned when a network device goes
down.  This mirrors the logic in net/core/dst.c's dst_ifdown().
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

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

Input path is mostly run under RCU and doesnt touch dst refcnt

But output path on forwarding or UDP workloads hits
badly dst refcount, and we have lot of false sharing, for example
in ipv4_mtu() when reading rt->rt_pmtu

Using a percpu cache for nh_rth_output gives a nice performance
increase at a small cost.

24 udpflood test on my 24 cpu machine (dummy0 output device)
(each process sends 1.000.000 udp frames, 24 processes are started)

before : 5.24 s
after : 2.06 s
For reference, time on linux-3.5 : 6.60 s
Signed-off-by: NEric Dumazet <edumazet@google.com>
Tested-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit 404e0a8b (net: ipv4: fix RCU races on dst refcounts) tried
to solve a race but added a problem at device/fib dismantle time :

We really want to call dst_free() as soon as possible, even if sockets
still have dst in their cache.
dst_release() calls in free_fib_info_rcu() are not welcomed.

Root of the problem was that now we also cache output routes (in
nh_rth_output), we must use call_rcu() instead of call_rcu_bh() in
rt_free(), because output route lookups are done in process context.

Based on feedback and initial patch from David Miller (adding another
call_rcu_bh() call in fib, but it appears it was not the right fix)

I left the inet_sk_rx_dst_set() helper and added __rcu attributes
to nh_rth_output and nh_rth_input to better document what is going on in
this code.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit c6cffba4 (ipv4: Fix input route performance regression.)
added various fatal races with dst refcounts.

crashes happen on tcp workloads if routes are added/deleted at the same
time.

The dst_free() calls from free_fib_info_rcu() are clearly racy.

We need instead regular dst refcounting (dst_release()) and make
sure dst_release() is aware of RCU grace periods :

Add DST_RCU_FREE flag so that dst_release() respects an RCU grace period
before dst destruction for cached dst

Introduce a new inet_sk_rx_dst_set() helper, using atomic_inc_not_zero()
to make sure we dont increase a zero refcount (On a dst currently
waiting an rcu grace period before destruction)

rt_cache_route() must take a reference on the new cached route, and
release it if was not able to install it.

With this patch, my machines survive various benchmarks.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

With the routing cache removal we lost the "noref" code paths on
input, and this can kill some routing workloads.

Reinstate the noref path when we hit a cached route in the FIB
nexthops.

With help from Eric Dumazet.
Reported-by: NAlexander Duyck <alexander.duyck@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit d2d68ba9 (ipv4: Cache input routes in fib_info nexthops.)
introduced rt_cache_valid() helper. It unfortunately doesn't check if
route is expired before caching it.

I noticed sk_setup_caps() was constantly called on a tcp workload.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

On input packet processing, rt->rt_iif will be zero if we should
use skb->dev->ifindex.

Since we access rt->rt_iif consistently via inet_iif(), that is
the only spot whose interpretation have to adjust.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use inet_iif() consistently, and for TCP record the input interface of
cached RX dst in inet sock.

rt->rt_iif is going to be encoded differently, so that we can
legitimately cache input routes in the FIB info more aggressively.

When the input interface is "use SKB device index" the rt->rt_iif will
be set to zero.

This forces us to move the TCP RX dst cache installation into the ipv4
specific code, and as well it should since doing the route caching for
ipv6 is pointless at the moment since it is not inspected in the ipv6
input paths yet.

Also, remove the unlikely on dst->obsolete, all ipv4 dsts have
obsolete set to a non-zero value to force invocation of the check
callback.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The last and final kernel user, ICMP address replies,
has been removed.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It's not really needed.

We only grabbed a reference to the fib_info for the sake of fib_info
local metrics.

However, fib_info objects are freed using RCU, as are therefore their
private metrics (if any).

We would have triggered a route cache flush if we eliminated a
reference to a fib_info object in the routing tables.

Therefore, any existing cached routes will first check and see that
they have been invalidated before an errant reference to these
metric values would occur.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

That is this value's only use, as a boolean to indicate whether
a route is an input route or not.

So implement it that way, using a u16 gap present in the struct
already.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Never actually used.

It was being set on output routes to the original OIF specified in the
flow key used for the lookup.

Adjust the only user, ipmr_rt_fib_lookup(), for greater correctness of
the flowi4_oif and flowi4_iif values, thanks to feedback from Julian
Anastasov.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Don't bother incrementing dst->__use and setting dst->lastuse,
they are completely pointless and just slow things down.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

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