commit d2d68ba9fe (ipv4: Cache input routes in fib_info nexthops.)
introduced a regression for forwarding.
This was hard to reproduce but the symptom was that packets were
delivered to local host instead of being forwarded.
David suggested to add fib_type to fib_info so that we dont
inadvertently share same fib_info for different purposes.
With help from Julian Anastasov who provided very helpful
hints, reproduced here :
<quote>
Can it be a problem related to fib_info reuse
from different routes. For example, when local IP address
is created for subnet we have:
broadcast 192.168.0.255 dev DEV proto kernel scope link src
192.168.0.1
192.168.0.0/24 dev DEV proto kernel scope link src 192.168.0.1
local 192.168.0.1 dev DEV proto kernel scope host src 192.168.0.1
The "dev DEV proto kernel scope link src 192.168.0.1" is
a reused fib_info structure where we put cached routes.
The result can be same fib_info for 192.168.0.255 and
192.168.0.0/24. RTN_BROADCAST is cached only for input
routes. Incoming broadcast to 192.168.0.255 can be cached
and can cause problems for traffic forwarded to 192.168.0.0/24.
So, this patch should solve the problem because it
separates the broadcast from unicast traffic.
And the ip_route_input_slow caching will work for
local and broadcast input routes (above routes 1 and 3) just
because they differ in scope and use different fib_info.
</quote>
Many thanks to Chris Clayton for his patience and help.
Reported-by: Chris Clayton <chris2553@googlemail.com>
Bisected-by: Chris Clayton <chris2553@googlemail.com>
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Julian Anastasov <ja@ssi.bg>
Tested-by: Chris Clayton <chris2553@googlemail.com>
Signed-off-by: David 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: Eric Dumazet <edumazet@google.com>
Tested-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 404e0a8b6a (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: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Caching input routes is slightly simpler than output routes, since we
don't need to be concerned with nexthop exceptions. (locally
destined, and routed packets, never trigger PMTU events or redirects
that will be processed by us).
However, we have to elide caching for the DIRECTSRC and non-zero itag
cases.
Signed-off-by: David S. Miller <davem@davemloft.net>
If we have an output route that lacks nexthop exceptions, we can cache
it in the FIB info nexthop.
Such routes will have DST_HOST cleared because such routes refer to a
family of destinations, rather than just one.
The sequence of the handling of exceptions during route lookup is
adjusted to make the logic work properly.
Before we allocate the route, we lookup the exception.
Then we know if we will cache this route or not, and therefore whether
DST_HOST should be set on the allocated route.
Then we use DST_HOST to key off whether we should store the resulting
route, during rt_set_nexthop(), in the FIB nexthop cache.
With help from Eric Dumazet.
Signed-off-by: David S. Miller <davem@davemloft.net>
Use global seqlock for the nh_exceptions. Call
fnhe_oldest with the right hash chain. Correct the diff
value for dst_set_expires.
v2: after suggestions from Eric Dumazet:
* get rid of spin lock fnhe_lock, rearrange update_or_create_fnhe
* continue daddr search in rt_bind_exception
v3:
* remove the daddr check before seqlock in rt_bind_exception
* restart lookup in rt_bind_exception on detected seqlock change,
as suggested by David Miller
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
In a regime where we have subnetted route entries, we need a way to
store persistent storage about destination specific learned values
such as redirects and PMTU values.
This is implemented here via nexthop exceptions.
The initial implementation is a 2048 entry hash table with relaiming
starting at chain length 5. A more sophisticated scheme can be
devised if that proves necessary.
Signed-off-by: David S. Miller <davem@davemloft.net>
We only use it to fetch the rule's tclassid, so just store the
tclassid there instead.
This also decreases the size of fib_result by a full 8 bytes on
64-bit. On 32-bits it's a wash.
Signed-off-by: David S. Miller <davem@davemloft.net>
All paths assume, when CONFIG_IP_MULTIPLE_TABLES is enabled, that any
successful call to fib_lookup() will initialize the fib_result->r
value to something.
We violated that expectation in the new fib_lookup() fast path.
Reported-by: Or Gerlitz <ogerlitz@mellanox.com>
Tested-by: Eric Dumazet <eric.dumazet@gmail.com>
Tested-by: Greg Rose <gregory.v.rose@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If the user hasn't actually installed any custom rules, or fiddled
with the default ones, don't go through the whole FIB rules layer.
It's just pure overhead.
Instead do what we do with CONFIG_IP_MULTIPLE_TABLES disabled, check
the individual tables by hand, one by one.
Also, move fib_num_tclassid_users into the ipv4 network namespace.
Signed-off-by: David S. Miller <davem@davemloft.net>
If rpfilter is off (or the SKB has an IPSEC path) and there are not
tclassid users, we don't have to do anything at all when
fib_validate_source() is invoked besides setting the itag to zero.
We monitor tclassid uses with a counter (modified only under RTNL and
marked __read_mostly) and we protect the fib_validate_source() real
work with a test against this counter and whether rpfilter is to be
done.
Having a way to know whether we need no tclassid processing or not
also opens the door for future optimized rpfilter algorithms that do
not perform full FIB lookups.
Signed-off-by: David S. Miller <davem@davemloft.net>
Checking for in_dev being NULL is pointless.
In fact, all of our callers have in_dev precomputed already,
so just pass it in and remove the NULL checking.
Signed-off-by: David S. Miller <davem@davemloft.net>
The specific destination is the host we direct unicast replies to.
Usually this is the original packet source address, but if we are
responding to a multicast or broadcast packet we have to use something
different.
Specifically we must use the source address we would use if we were to
send a packet to the unicast source of the original packet.
The routing cache precomputes this value, but we want to remove that
precomputation because it creates a hard dependency on the expensive
rpfilter source address validation which we'd like to make cheaper.
There are only three places where this matters:
1) ICMP replies.
2) pktinfo CMSG
3) IP options
Now there will be no real users of rt->rt_spec_dst and we can simply
remove it altogether.
Signed-off-by: David S. Miller <davem@davemloft.net>
Use of "unsigned int" is preferred to bare "unsigned" in net tree.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
fib_select_default() is a complete NOP, and completely pointless
to invoke, when we have no more than 1 default route installed.
And this is far and away the common case.
So remember how many prefixlen==0 routes we have in the routing
table, and elide the call when we have no more than one of those.
This cuts output route creation time by 157 cycles on Niagara2+.
In order to add the new int to fib_table, we have to correct the type
of ->tb_data[] to unsigned long, otherwise the private area will be
unaligned on 64-bit systems.
Signed-off-by: David S. Miller <davem@davemloft.net>
Reviewed-by: Eric Dumazet <eric.dumazet@gmail.com>
This makes sk_buff available for other use in fib_validate_source().
Signed-off-by: Michael Smith <msmith@cbnco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move the scope value out of the fib alias entries and into fib_info,
so that we always use the correct scope when recomputing the nexthop
cached source address.
Reported-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
Any operation that:
1) Brings up an interface
2) Adds an IP address to an interface
3) Deletes an IP address from an interface
can potentially invalidate the nh_saddr value, requiring
it to be recomputed.
Perform the recomputation lazily using a generation ID.
Reported-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
To start doing these conversions, we need to add some temporary
flow4_* macros which will eventually go away when all the protocol
code paths are changed to work on AF specific flowi objects.
Signed-off-by: David S. Miller <davem@davemloft.net>
We have to use cfg->fc_scope not the final nh_scope value.
Reported-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet
David S. Miller
Eric Dumazet
Julian Anastasov
Michael Smith