dev_queue_xmit enqueue's a skb and calls qdisc_run which
dequeue's the skb and xmits it. In most cases, the skb that
is enqueue'd is the same one that is dequeue'd (unless the
queue gets stopped or multiple cpu's write to the same queue
and ends in a race with qdisc_run). For default qdiscs, we
can remove the redundant enqueue/dequeue and simply xmit the
skb since the default qdisc is work-conserving.
The patch uses a new flag - TCQ_F_CAN_BYPASS to identify the
default fast queue. The controversial part of the patch is
incrementing qlen when a skb is requeued - this is to avoid
checks like the second line below:
+ } else if ((q->flags & TCQ_F_CAN_BYPASS) && !qdisc_qlen(q) &&
>> !q->gso_skb &&
+ !test_and_set_bit(__QDISC_STATE_RUNNING, &q->state)) {
Results of a 2 hour testing for multiple netperf sessions (1,
2, 4, 8, 12 sessions on a 4 cpu system-X). The BW numbers are
aggregate Mb/s across iterations tested with this version on
System-X boxes with Chelsio 10gbps cards:
----------------------------------
Size | ORG BW NEW BW |
----------------------------------
128K | 156964 159381 |
256K | 158650 162042 |
----------------------------------
Changes from ver1:
1. Move sch_direct_xmit declaration from sch_generic.h to
pkt_sched.h
2. Update qdisc basic statistics for direct xmit path.
3. Set qlen to zero in qdisc_reset.
4. Changed some function names to more meaningful ones.
Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
String literals are constant, and usually, we can also tag the array
of pointers const too, moving it to the .rodata section.
Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is a path when an assetion in dev_unicast_sync() appears.
igmp6_group_added -> dev_mc_add -> __dev_set_rx_mode ->
-> vlan_dev_set_rx_mode -> dev_unicast_sync
Therefore we cannot protect this list with rtnl. This patch restores the
original protecting this list with spinlock.
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Tested-by: Meelis Roos <mroos@linux.ee>
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to make cfg80211/nl80211 aware of network namespaces,
we have to do the following things:
* del_virtual_intf method takes an interface index rather
than a netdev pointer - simply change this
* nl80211 uses init_net a lot, it changes to use the sender's
network namespace
* scan requests use the interface index, hold a netdev pointer
and reference instead
* we want a wiphy and its associated virtual interfaces to be
in one netns together, so
- we need to be able to change ns for a given interface, so
export dev_change_net_namespace()
- for each virtual interface set the NETIF_F_NETNS_LOCAL
flag, and clear that flag only when the wiphy changes ns,
to disallow breaking this invariant
* when a network namespace goes away, we need to reparent the
wiphy to init_net
* cfg80211 users that support creating virtual interfaces must
create them in the wiphy's namespace, currently this affects
only mac80211
The end result is that you can now switch an entire wiphy into
a different network namespace with the new command
iw phy#<idx> set netns <pid>
and all virtual interfaces will follow (or the operation fails).
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
For mac80211, with the master netdev removal, we need to be
able to sync a multicast address list onto another list that
is not tracked within a netdev, so we need access to the
functions doing that.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This patch converts the remaining occurences of raw return values to their
symbolic counterparts in ndo_start_xmit() functions that were missed by the
previous automatic conversion.
Additionally code that assumed the symbolic value of NETDEV_TX_OK to be zero
is changed to explicitly use NETDEV_TX_OK.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
When NAPI is disabled while we're in net_rx_action, we end up
calling __napi_complete without flushing GRO packets. This is
a bug as it would cause the GRO packets to linger, of course it
also literally BUGs to catch error like this :)
This patch changes it to napi_complete, with the obligatory IRQ
reenabling. This should be safe because we've only just disabled
IRQs and it does not materially affect the test conditions in
between.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to get the tun driver to account packets, we need to be
able to receive packets with destructors set. To be on the safe
side, I added an skb_orphan call for all protocols by default since
some of them (IP in particular) cannot handle receiving packets
destructors properly.
Now it seems that at least one protocol (CAN) expects to be able
to pass skb->sk through the rx path without getting clobbered.
So this patch attempts to fix this properly by moving the skb_orphan
call to where it's actually needed. In particular, I've added it
to skb_set_owner_[rw] which is what most users of skb->destructor
call.
This is actually an improvement for tun too since it means that
we only give back the amount charged to the socket when the skb
is passed to another socket that will also be charged accordingly.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Tested-by: Oliver Hartkopp <olver@hartkopp.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is inspired by patch recently posted by Johannes Berg. Basically what
my patch does is to group list and a count of addresses into newly introduced
structure netdev_hw_addr_list. This brings us two benefits:
1) struct net_device becames a bit nicer.
2) in the future there will be a possibility to operate with lists independently
on netdevices (with exporting right functions).
I wanted to introduce this patch before I'll post a multicast lists conversion.
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
drivers/net/bnx2.c | 4 +-
drivers/net/e1000/e1000_main.c | 4 +-
drivers/net/ixgbe/ixgbe_main.c | 6 +-
drivers/net/mv643xx_eth.c | 2 +-
drivers/net/niu.c | 4 +-
drivers/net/virtio_net.c | 10 ++--
drivers/s390/net/qeth_l2_main.c | 2 +-
include/linux/netdevice.h | 17 +++--
net/core/dev.c | 130 ++++++++++++++++++--------------------
9 files changed, 89 insertions(+), 90 deletions(-)
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch changes FDB entry check for ATM LANE bridge integration.
There's no point in holding a FDB entry around SKB building.
br_fdb_get()/br_fdb_put() pair are changed into single br_fdb_test_addr()
hook that checks if the addr has FDB entry pointing to other port
to the one the request arrived on.
FDB entry refcounting is removed as it's not used anywhere else.
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit b00055aacd " [NET] core: add
RFC2863 operstate" defined new interface flag values. Its
documentation specified that these flags could be accessed from user
space via SIOCGIFFLAGS. However, this does not work because the new
flags do not fit in that ioctl's argument width.
Change the documentation to match the code's behavior. Also change
the source to explicitly show the truncation. This _should_ have no
effect on executable code, and did not with gcc 4.2.4 generating x86
code.
A new ioctl could be defined to return all interface flags to user
space. However, since this has been broken for three years with no
one complaining, there doesn't seem much need. They are still
accessible via netlink.
Reported-by: "Fredrik Arnerup" <fredrik.arnerup@edgeware.tv>
Signed-off-by: John Dykstra <john.dykstra1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
IEEE 802.15.4 stack requires several constants to be defined/adjusted.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Sergey Lapin <slapin@ossfans.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit f001fde5ea
(net: introduce a list of device addresses dev_addr_list (v6))
added one regression Vegard Nossum found in its testings.
With kmemcheck help, Vegard found some uninitialized memory
was read and reported to user, potentialy leaking kernel data.
( thread can be found on http://lkml.org/lkml/2009/5/30/177 )
dev_addr_init() incorrectly uses sizeof() operator. We were
initializing one byte instead of MAX_ADDR_LEN bytes.
Reported-by: Vegard Nossum <vegard.nossum@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
NETDEV_UP is called after the device is set UP, but sometimes
it is useful to be able to veto the device UP. Introduce a
new NETDEV_PRE_UP notifier that can be used for exactly this.
The first use case will be cfg80211 denying interfaces to be
set UP if the device is known to be rfkill'ed.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Define three accessors to get/set dst attached to a skb
struct dst_entry *skb_dst(const struct sk_buff *skb)
void skb_dst_set(struct sk_buff *skb, struct dst_entry *dst)
void skb_dst_drop(struct sk_buff *skb)
This one should replace occurrences of :
dst_release(skb->dst)
skb->dst = NULL;
Delete skb->dst field
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch converts unicast address list to standard list_head using
previously introduced struct netdev_hw_addr. It also relaxes the
locking. Original spinlock (still used for multicast addresses) is not
needed and is no longer used for a protection of this list. All
reading and writing takes place under rtnl (with no changes).
I also removed a possibility to specify the length of the address
while adding or deleting unicast address. It's always dev->addr_len.
The convertion touched especially e1000 and ixgbe codes when the
change is not so trivial.
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
drivers/net/bnx2.c | 13 +--
drivers/net/e1000/e1000_main.c | 24 +++--
drivers/net/ixgbe/ixgbe_common.c | 14 ++--
drivers/net/ixgbe/ixgbe_common.h | 4 +-
drivers/net/ixgbe/ixgbe_main.c | 6 +-
drivers/net/ixgbe/ixgbe_type.h | 4 +-
drivers/net/macvlan.c | 11 +-
drivers/net/mv643xx_eth.c | 11 +-
drivers/net/niu.c | 7 +-
drivers/net/virtio_net.c | 7 +-
drivers/s390/net/qeth_l2_main.c | 6 +-
drivers/scsi/fcoe/fcoe.c | 16 ++--
include/linux/netdevice.h | 18 ++--
net/8021q/vlan.c | 4 +-
net/8021q/vlan_dev.c | 10 +-
net/core/dev.c | 195 +++++++++++++++++++++++++++-----------
net/dsa/slave.c | 10 +-
net/packet/af_packet.c | 4 +-
18 files changed, 227 insertions(+), 137 deletions(-)
Signed-off-by: David S. Miller <davem@davemloft.net>
Use ALIGN() and PTR_ALIGN() macros instead of handcoding them.
Get rid of NETDEV_ALIGN_CONST ugly define
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As we know the only packets which need the final pskb_may_pull
are completely non-linear, and have all the required bits in
frag0, we can perform a straight memcpy instead of going through
pskb_may_pull and doing skb_copy_bits.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
For the overwhelming majority of cases, skb_gro_header's return
value cannot be NULL. Yet we must check it because of its current
form. This patch splits it up into multiple functions in order
to avoid this.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
By caching frag0_len, we can avoid checking both frag0 and the
length separately in skb_gro_header. This helps as skb_gro_header
is called four times per packet which amounts to a few million
times at 10Gb/s.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently skb_gro_header is used for packets which put the hardware
header in skb->data with the rest in frags. Since the drivers that
need this optimisation all provide completely non-linear packets,
we can gain extra optimisations by only performing the frag0
optimisation for completely non-linear packets.
In particular, we can simply test frag0 (instead of skb_headlen)
to see whether the optimisation is in force.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The function skb_gro_header is called four times per packet which
quickly adds up at 10Gb/s. This patch inlines it to allow better
optimisations.
Some architectures perform multiplication for page_address, which
is done by each skb_gro_header invocation. This patch caches that
value in skb->cb to avoid the unnecessary multiplications.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
We would like to get rid of netdev->trans_start = jiffies; that about all net
drivers have to use in their start_xmit() function, and use txq->trans_start
instead.
This can be done generically in core network, as suggested by David.
Some devices, (particularly loopback) dont need trans_start update, because
they dont have transmit watchdog. We could add a new device flag, or rely
on fact that txq->tran_start can be updated is txq->xmit_lock_owner is
different than -1. Use a helper function to hide our choice.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All drivers are already converted to new net_device_ops API
and nobody uses old API anymore.
Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Patch to add the ability to detect drops in hardware interfaces via dropwatch.
Adds a tracepoint to net_rx_action to signal everytime a napi instance is
polled. The dropmon code then periodically checks to see if the rx_frames
counter has changed, and if so, adds a drop notification to the netlink
protocol, using the reserved all-0's vector to indicate the drop location was in
hardware, rather than somewhere in the code.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
include/linux/net_dropmon.h | 8 ++
include/trace/napi.h | 11 +++
net/core/dev.c | 5 +
net/core/drop_monitor.c | 124 ++++++++++++++++++++++++++++++++++++++++++--
net/core/net-traces.c | 4 +
net/core/netpoll.c | 2
6 files changed, 149 insertions(+), 5 deletions(-)
Signed-off-by: David S. Miller <davem@davemloft.net>
One point of contention in high network loads is the dst_release() performed
when a transmited skb is freed. This is because NIC tx completion calls
dev_kree_skb() long after original call to dev_queue_xmit(skb).
CPU cache is cold and the atomic op in dst_release() stalls. On SMP, this is
quite visible if one CPU is 100% handling softirqs for a network device,
since dst_clone() is done by other cpus, involving cache line ping pongs.
It seems right place to release dst is in dev_hard_start_xmit(), for most
devices but ones that are virtual, and some exceptions.
David Miller suggested to define a new device flag, set in alloc_netdev_mq()
(so that most devices set it at init time), and carefuly unset in devices
which dont want a NULL skb->dst in their ndo_start_xmit().
List of devices that must clear this flag is :
- loopback device, because it calls netif_rx() and quoting Patrick :
"ip_route_input() doesn't accept loopback addresses, so loopback packets
already need to have a dst_entry attached."
- appletalk/ipddp.c : needs skb->dst in its xmit function
- And all devices that call again dev_queue_xmit() from their xmit function
(as some classifiers need skb->dst) : bonding, vlan, macvlan, eql, ifb, hdlc_fr
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
offsetof(struct net_device, features)=0x44
offsetof(struct net_device, stats.tx_packets)=0x54
offsetof(struct net_device, stats.tx_bytes)=0x5c
offsetof(struct net_device, stats.tx_dropped)=0x6c
Network drivers that touch dev->stats.tx_packets/stats.tx_bytes in their
tx path can slow down SMP operations, since they dirty a cache line
that should stay shared (dev->features is needed in rx and tx paths)
We could move away stats field in net_device but it wont help that much.
(Two cache lines dirtied in tx path, we can do one only)
Better solution is to add tx_packets/tx_bytes/tx_dropped in struct
netdev_queue because this structure is already touched in tx path and
counters updates will then be free (no increase in size)
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add missed checking of dev_addr_init return value in alloc_netdev_mq.
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
net/core/dev.c | 15 ++++++++++++---
1 files changed, 12 insertions(+), 3 deletions(-)
Signed-off-by: David S. Miller <davem@davemloft.net>
v5 -> v6 (current):
-removed so far unused static functions
-corrected dev_addr_del_multiple to call del instead of add
v4 -> v5:
-added device address type (suggested by davem)
-removed refcounting (better to have simplier code then safe potentially few
bytes)
v3 -> v4:
-changed kzalloc to kmalloc in __hw_addr_add_ii()
-ASSERT_RTNL() avoided in dev_addr_flush() and dev_addr_init()
v2 -> v3:
-removed unnecessary rcu read locking
-moved dev_addr_flush() calling to ensure no null dereference of dev_addr
v1 -> v2:
-added forgotten ASSERT_RTNL to dev_addr_init and dev_addr_flush
-removed unnecessary rcu_read locking in dev_addr_init
-use compare_ether_addr_64bits instead of compare_ether_addr
-use L1_CACHE_BYTES as size for allocating struct netdev_hw_addr
-use call_rcu instead of rcu_synchronize
-moved is_etherdev_addr into __KERNEL__ ifdef
This patch introduces a new list in struct net_device and brings a set of
functions to handle the work with device address list. The list is a replacement
for the original dev_addr field and because in some situations there is need to
carry several device addresses with the net device. To be backward compatible,
dev_addr is made to point to the first member of the list so original drivers
sees no difference.
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Based almost entirely upon a patch by Eric Dumazet.
The common case is to have num-tx-queues <= num_rx_queues
and even if num_tx_queues is larger it will not be significantly
larger.
Therefore, a subtraction loop is always going to be faster than
modulus.
Signed-off-by: David S. Miller <davem@davemloft.net>
When skb_rx_queue_recorded() is true, we dont want to use jash distribution
as the device driver exactly told us which queue was selected at RX time.
jhash makes a statistical shuffle, but this wont work with 8 static inputs.
Later improvements would be to compute reciprocal value of real_num_tx_queues
to avoid a divide here. But this computation should be done once,
when real_num_tx_queues is set. This needs a separate patch, and a new
field in struct net_device.
Reported-by: Andrew Dickinson <andrew@whydna.net>
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The skb_gro_* code fails to handle the case where a header starts
in the linear area but ends in the frags area. Since the goal
of skb_gro_* is to optimise the case of completely non-linear
packets, we can simply bail out if we have anything in the linear
area.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This loop over fragments in napi_fraginfo_skb() was "interesting".
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Alex Sidorenko reported:
"while experimenting with 'netem' we have found some strange behaviour. It
seemed that ingress delay as measured by 'ping' command shows up on some
hosts but not on others.
After some investigation I have found that the problem is that skbuff->tstamp
field value depends on whether there are any packet sniffers enabled. That
is:
- if any ptype_all handler is registered, the tstamp field is as expected
- if there are no ptype_all handlers, the tstamp field does not show the delay"
This patch prevents unnecessary update of tstamp in dev_queue_xmit_nit()
on ingress path (with act_mirred) adding a check, so minimal overhead on
the fast path, but only when sniffers etc. are active.
Since netem at ingress seems to logically emulate a network before a host,
tstamp is zeroed to trigger the update and pretend delays are from the
outside.
Reported-by: Alex Sidorenko <alexandre.sidorenko@hp.com>
Tested-by: Alex Sidorenko <alexandre.sidorenko@hp.com>
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It turns out that copying a 16-byte area at ~800k times a second
can be really expensive :) This patch redesigns the frags GRO
interface to avoid copying that area twice.
The two disciples of the frags interface have been converted.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since everybody has been focusing on baremetal GRO performance
no one noticed when I added a bug that zapped gso_size for all
GRO packets. This only gets picked up when you forward the skb
out of an interface.
Thanks to Mark Wagner for noticing this bug when testing kvm.
Reported-by: Mark Wagner <mwagner@redhat.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently netif_device_attach/detach are only stopping one queue. They
should be starting and stopping all the queues on a given device.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
GRO assumes that there is a one-to-one relationship between NAPI
structure and network device. Some devices like sky2 share multiple
devices on a single interrupt so only have one NAPI handler. Rather than
split GRO from NAPI, just have GRO assume if device changes that
it is a different flow.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When I fixed the GRO crash in the legacy receive path I used
napi_complete to replace __napi_complete. Unfortunately they're
not the same when NETPOLL is enabled, which may result in us
not calling __napi_complete at all.
What's more, we really do need to keep the __napi_complete call
within the IRQ-off section since in theory an IRQ can occur in
between and fill up the backlog to the maximum, causing us to
lock up.
Since we can't seem to find a fix that works properly right now,
this patch reverts all the GRO support from the netif_rx path.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is no gain using prefetch() in dev_hard_start_xmit(), since
we already had to read ops->ndo_select_queue pointer in dev_pick_tx(),
and both pointers are probably located in the same cache line.
This prefetch call slows down fast path because of a stall in address
computation.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some minor changes to queue hashing:
1. Use const on accessor functions
2. Export skb_tx_hash for use in drivers (see ixgbe)
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
