This reverts commit aaf8cdc34d.
Drivers like the ipw2100 call device_create_group when they
are initialized and device_remove_group when they are shutdown.
Moving them between namespaces deletes their sysfs groups early.
In particular the following call chain results.
netdev_unregister_kobject -> device_del -> kobject_del -> sysfs_remove_dir
With sysfs_remove_dir recursively deleting all of it's subdirectories,
and nothing adding them back.
Ouch!
Therefore we need to call something that ultimate calls sysfs_mv_dir
as that sysfs function can move sysfs directories between namespaces
without deleting their subdirectories or their contents. Allowing
us to avoid placing extra boiler plate into every driver that does
something interesting with sysfs.
Currently the function that provides that capability is device_rename.
That is the code works without nasty side effects as originally written.
So remove the misguided fix for moving devices between namespaces. The
bug in the kobject layer that inspired it has now been recognized and
fixed.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Use low order bit of skb->_skb_dst to tell dst is not refcounted.
Change _skb_dst to _skb_refdst to make sure all uses are catched.
skb_dst() returns the dst, regardless of noref bit set or not, but
with a lockdep check to make sure a noref dst is not given if current
user is not rcu protected.
New skb_dst_set_noref() helper to set an notrefcounted dst on a skb.
(with lockdep check)
skb_dst_drop() drops a reference only if skb dst was refcounted.
skb_dst_force() helper is used to force a refcount on dst, when skb
is queued and not anymore RCU protected.
Use skb_dst_force() in __sk_add_backlog(), __dev_xmit_skb() if
!IFF_XMIT_DST_RELEASE or skb enqueued on qdisc queue, in
sock_queue_rcv_skb(), in __nf_queue().
Use skb_dst_force() in dev_requeue_skb().
Note: dst_use_noref() still dirties dst, we might transform it
later to do one dirtying per jiffies.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If CONFIG_SMP=y, then we own a queue spinlock, we can avoid the atomic
test_and_set_bit() from napi_schedule_prep().
We now have same number of atomic ops per netif_rx() calls than with
pre-RPS kernel.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With RPS inclusion, skb timestamping is not consistent in RX path.
If netif_receive_skb() is used, its deferred after RPS dispatch.
If netif_rx() is used, its done before RPS dispatch.
This can give strange tcpdump timestamps results.
I think timestamping should be done as soon as possible in the receive
path, to get meaningful values (ie timestamps taken at the time packet
was delivered by NIC driver to our stack), even if NAPI already can
defer timestamping a bit (RPS can help to reduce the gap)
Tom Herbert prefer to sample timestamps after RPS dispatch. In case
sampling is expensive (HPET/acpi_pm on x86), this makes sense.
Let admins switch from one mode to another, using a new
sysctl, /proc/sys/net/core/netdev_tstamp_prequeue
Its default value (1), means timestamps are taken as soon as possible,
before backlog queueing, giving accurate timestamps.
Setting a 0 value permits to sample timestamps when processing backlog,
after RPS dispatch, to lower the load of the pre-RPS cpu.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now there's null check here and also again in the hook. Looking at bridge bits
which are simmilar, port structure is rcu_dereferenced right away in
handle_bridge and passed to hook. Looks nicer.
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduce ____napi_schedule() helper for callers in irq disabled
contexts. rps_trigger_softirq() becomes a leaf function.
Use container_of() in process_backlog() instead of accessing per_cpu
address.
Use a custom inlined version of __napi_complete() in process_backlog()
to avoid one locked instruction :
only current cpu owns and manipulates this napi,
and NAPI_STATE_SCHED is the only possible flag set on backlog.
we can use a plain write instead of clear_bit(),
and we dont need an smp_mb() memory barrier, since RPS is on,
backlog is protected by a spinlock.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case of congestion, netif_rx() frees the skb, so we must assume
dev_forward_skb() also consume skb.
Bug introduced by commit 445409602c
(veth: move loopback logic to common location)
We must change dev_forward_skb() to always consume skb, and veth to not
double free it.
Bug report : http://marc.info/?l=linux-netdev&m=127310770900442&w=3
Reported-by: Martín Ferrari <martin.ferrari@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Per cpu variable softnet_data.total was shared between IRQ and SoftIRQ context
without any protection. And enqueue_to_backlog should update the netdev_rx_stat
of the target CPU.
This patch renames softnet_data.total to softnet_data.processed: the number of
packets processed in uppper levels(IP stacks).
softnet_stat data is moved into softnet_data.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
----
include/linux/netdevice.h | 17 +++++++----------
net/core/dev.c | 26 ++++++++++++--------------
net/sched/sch_generic.c | 2 +-
3 files changed, 20 insertions(+), 25 deletions(-)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
batch skb dequeueing from softnet input_pkt_queue to reduce potential lock
contention when RPS is enabled.
Note: in the worst case, the number of packets in a softnet_data may
be double of netdev_max_backlog.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
reimplement softnet_data.output_queue as a FIFO queue to keep the
fairness among the qdiscs rescheduled.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
----
include/linux/netdevice.h | 1 +
net/core/dev.c | 22 ++++++++++++----------
2 files changed, 13 insertions(+), 10 deletions(-)
Signed-off-by: David S. Miller <davem@davemloft.net>
optimize rps_get_cpu().
don't initialize ports when we can get the ports. one memory access
for ports than two.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This way GSO packets don't get handled differently.
With help from Eric Dumazet.
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
If some skb are queued to our backlog, we are delaying IPI sending at
the end of net_rx_action(), increasing latencies. This defeats the
queueing, since we want to quickly dispatch packets to the pool of
worker cpus, then eventually deeply process our packets.
It's better to send IPI before processing our packets in upper layers,
from process_backlog().
Change the _and_disable_irq suffix to _and_enable_irq(), since we enable
local irq in net_rps_action(), sorry for the confusion.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
At this point, skb->destructor is not the original one (stored in
DEV_GSO_CB(skb)->destructor)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since netdev_chain is guarded by rtnl_lock, ASSERT_RTNL should be
present here to make sure that all callers of call_netdevice_notifiers
does the locking properly.
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case we compute a software skb->rxhash, we can generate a consistent
hash : Its value will be the same in both flow directions.
This helps some workloads, like conntracking, since the same state needs
to be accessed in both directions.
tbench + RFS + this patch gives better results than tbench with default
kernel configuration (no RPS, no RFS)
Also fixed some sparse warnings.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
struct softnet_data holds many queues, so consistent use "sd" name
instead of "queue" is better.
Adds a rps_ipi_queued() helper to cleanup enqueue_to_backlog()
Adds a _and_irq_disable suffix to net_rps_action() name, as David
suggested.
incr_input_queue_head() becomes input_queue_head_incr()
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net_rps_action() is a bit expensive on NR_CPUS=64..4096 kernels, even if
RPS is not active.
Tom Herbert used two bitmasks to hold information needed to send IPI,
but a single LIFO list seems more appropriate.
Move all RPS logic into net_rps_action() to cleanup net_rx_action() code
(remove two ifdefs)
Move rps_remote_softirq_cpus into softnet_data to share its first cache
line, filling an existing hole.
In a future patch, we could call net_rps_action() from process_backlog()
to make sure we send IPI before handling this cpu backlog.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Transmitted skb might be attached to a socket and a destructor, for
memory accounting purposes.
Traditionally, this destructor is called at tx completion time, when skb
is freed.
When tx completion is performed by another cpu than the sender, this
forces some cache lines to change ownership. XPS was an attempt to give
tx completion to initial cpu.
David idea is to call destructor right before giving skb to device (call
to ndo_start_xmit()). Because device queues are usually small, orphaning
skb before tx completion is not a big deal. Some drivers already do
this, we could do it in upper level.
There is one known exception to this early orphaning, called tx
timestamping. It needs to keep a reference to socket until device can
give a hardware or software timestamp.
This patch adds a skb_orphan_try() helper, to centralize all exceptions
to early orphaning in one spot, and use it in dev_hard_start_xmit().
"tbench 16" results on a Nehalem machine (2 X5570 @ 2.93GHz)
before: Throughput 4428.9 MB/sec 16 procs
after: Throughput 4448.14 MB/sec 16 procs
UDP should get even better results, its destructor being more complex,
since SOCK_USE_WRITE_QUEUE is not set (four atomic ops instead of one)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- There is no point to enforce a time limit in process_backlog(), since
other napi instances dont follow same rule. We can exit after only one
packet processed...
The normal quota of 64 packets per napi instance should be the norm, and
net_rx_action() already has its own time limit.
Note : /proc/net/core/dev_weight can be used to tune this 64 default
value.
- Use DEFINE_PER_CPU_ALIGNED for softnet_data definition.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch implements receive flow steering (RFS). RFS steers
received packets for layer 3 and 4 processing to the CPU where
the application for the corresponding flow is running. RFS is an
extension of Receive Packet Steering (RPS).
The basic idea of RFS is that when an application calls recvmsg
(or sendmsg) the application's running CPU is stored in a hash
table that is indexed by the connection's rxhash which is stored in
the socket structure. The rxhash is passed in skb's received on
the connection from netif_receive_skb. For each received packet,
the associated rxhash is used to look up the CPU in the hash table,
if a valid CPU is set then the packet is steered to that CPU using
the RPS mechanisms.
The convolution of the simple approach is that it would potentially
allow OOO packets. If threads are thrashing around CPUs or multiple
threads are trying to read from the same sockets, a quickly changing
CPU value in the hash table could cause rampant OOO packets--
we consider this a non-starter.
To avoid OOO packets, this solution implements two types of hash
tables: rps_sock_flow_table and rps_dev_flow_table.
rps_sock_table is a global hash table. Each entry is just a CPU
number and it is populated in recvmsg and sendmsg as described above.
This table contains the "desired" CPUs for flows.
rps_dev_flow_table is specific to each device queue. Each entry
contains a CPU and a tail queue counter. The CPU is the "current"
CPU for a matching flow. The tail queue counter holds the value
of a tail queue counter for the associated CPU's backlog queue at
the time of last enqueue for a flow matching the entry.
Each backlog queue has a queue head counter which is incremented
on dequeue, and so a queue tail counter is computed as queue head
count + queue length. When a packet is enqueued on a backlog queue,
the current value of the queue tail counter is saved in the hash
entry of the rps_dev_flow_table.
And now the trick: when selecting the CPU for RPS (get_rps_cpu)
the rps_sock_flow table and the rps_dev_flow table for the RX queue
are consulted. When the desired CPU for the flow (found in the
rps_sock_flow table) does not match the current CPU (found in the
rps_dev_flow table), the current CPU is changed to the desired CPU
if one of the following is true:
- The current CPU is unset (equal to RPS_NO_CPU)
- Current CPU is offline
- The current CPU's queue head counter >= queue tail counter in the
rps_dev_flow table. This checks if the queue tail has advanced
beyond the last packet that was enqueued using this table entry.
This guarantees that all packets queued using this entry have been
dequeued, thus preserving in order delivery.
Making each queue have its own rps_dev_flow table has two advantages:
1) the tail queue counters will be written on each receive, so
keeping the table local to interrupting CPU s good for locality. 2)
this allows lockless access to the table-- the CPU number and queue
tail counter need to be accessed together under mutual exclusion
from netif_receive_skb, we assume that this is only called from
device napi_poll which is non-reentrant.
This patch implements RFS for TCP and connected UDP sockets.
It should be usable for other flow oriented protocols.
There are two configuration parameters for RFS. The
"rps_flow_entries" kernel init parameter sets the number of
entries in the rps_sock_flow_table, the per rxqueue sysfs entry
"rps_flow_cnt" contains the number of entries in the rps_dev_flow
table for the rxqueue. Both are rounded to power of two.
The obvious benefit of RFS (over just RPS) is that it achieves
CPU locality between the receive processing for a flow and the
applications processing; this can result in increased performance
(higher pps, lower latency).
The benefits of RFS are dependent on cache hierarchy, application
load, and other factors. On simple benchmarks, we don't necessarily
see improvement and sometimes see degradation. However, for more
complex benchmarks and for applications where cache pressure is
much higher this technique seems to perform very well.
Below are some benchmark results which show the potential benfit of
this patch. The netperf test has 500 instances of netperf TCP_RR
test with 1 byte req. and resp. The RPC test is an request/response
test similar in structure to netperf RR test ith 100 threads on
each host, but does more work in userspace that netperf.
e1000e on 8 core Intel
No RFS or RPS 104K tps at 30% CPU
No RFS (best RPS config): 290K tps at 63% CPU
RFS 303K tps at 61% CPU
RPC test tps CPU% 50/90/99% usec latency Latency StdDev
No RFS/RPS 103K 48% 757/900/3185 4472.35
RPS only: 174K 73% 415/993/2468 491.66
RFS 223K 73% 379/651/1382 315.61
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When dev_pick_tx() caches tx queue_index on a socket, we must check
socket dst_entry matches skb one, or risk a crash later, as reported by
Denys Fedorysychenko, if old packets are in flight during a route
change, involving devices with different number of queues.
Bug introduced by commit a4ee3ce3
(net: Use sk_tx_queue_mapping for connected sockets)
Reported-by: Denys Fedorysychenko <nuclearcat@nuclearcat.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Paris reported netif_rx() is calling smp_processor_id() from
preemptible context, in particular when caller is
ip_dev_loopback_xmit().
RPS commit added this smp_processor_id() call, this patch makes sure
preemption is disabled. rps_get_cpus() wants rcu_read_lock() anyway, we
can dot it a bit earlier.
Reported-by: Eric Paris <eparis@redhat.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
skb_bond_should_drop() is too big to be inlined.
This patch reduces kernel text size, and its compilation time as well
(shrinking include/linux/netdevice.h)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With latest CONFIG_PROVE_RCU stuff, I felt more comfortable to make this
work.
sk->sk_dst_cache is currently protected by a rwlock (sk_dst_lock)
This rwlock is readlocked for a very small amount of time, and dst
entries are already freed after RCU grace period. This calls for RCU
again :)
This patch converts sk_dst_lock to a spinlock, and use RCU for readers.
__sk_dst_get() is supposed to be called with rcu_read_lock() or if
socket locked by user, so use appropriate rcu_dereference_check()
condition (rcu_read_lock_held() || sock_owned_by_user(sk))
This patch avoids two atomic ops per tx packet on UDP connected sockets,
for example, and permits sk_dst_lock to be much less dirtied.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Dont use netdev_warn() in dev_cap_txqueue() and get_rps_cpu() so that we
can catch following warnings without crash.
bond0.2240 received packet on queue 6, but number of RX queues is 1
bond0.2240 received packet on queue 11, but number of RX queues is 1
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As noticed by Changli Gao, we must call local_irq_enable() after
rps_unlock()
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix spin_unlock_irq which needs to be rps_unlock.
Signed-off-by: Tom Herbert <therbert@google.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Converts the list and the core manipulating with it to be the same as uc_list.
+uses two functions for adding/removing mc address (normal and "global"
variant) instead of a function parameter.
+removes dev_mcast.c completely.
+exposes netdev_hw_addr_list_* macros along with __hw_addr_* functions for
manipulation with lists on a sandbox (used in bonding and 80211 drivers)
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
+little renaming of unicast functions to be smooth with multicast ones
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Followup to commit 5acbbd428d
(net: change illegal_highdma to use dma_mask)
If dev->dev.parent is NULL, we should not try to dereference it.
Dont force inline illegal_highdma() as its pretty big now.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Robert Hancock pointed out two problems about NETIF_F_HIGHDMA:
-Many drivers only set the flag when they detect they can use 64-bit DMA,
since otherwise they could receive DMA addresses that they can't handle
(which on platforms without IOMMU/SWIOTLB support is fatal). This means that if
64-bit support isn't available, even buffers located below 4GB will get copied
unnecessarily.
-Some drivers set the flag even though they can't actually handle 64-bit DMA,
which would mean that on platforms without IOMMU/SWIOTLB they would get a DMA
mapping error if the memory they received happened to be located above 4GB.
http://lkml.org/lkml/2010/3/3/530
We can use the dma_mask if we need bouncing or not here. Then we can
safely fix drivers that misuse NETIF_F_HIGHDMA.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
keep the old behavior on SMP without rps
RPS introduces a lock operation to per cpu variable input_pkt_queue on
SMP whenever rps is enabled or not. On SMP without RPS, this lock isn't
needed at all.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
----
net/core/dev.c | 42 ++++++++++++++++++++++++++++--------------
1 file changed, 28 insertions(+), 14 deletions(-)
Signed-off-by: David S. Miller <davem@davemloft.net>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
RPS currently depends on SMP and SYSFS
Adding a CONFIG_RPS makes sense in case this requirement changes in the
future. This patch saves about 1500 bytes of kernel text in case SMP is
on but SYSFS is off.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Need to take spinlocks when dequeuing from input_pkt_queue in flush_backlog.
Also, flush_backlog can now be called directly from netdev_run_todo.
Signed-off-by: Tom Herbert <therbert@google.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use RCU to avoid RTNL use in dev_getfirstbyhwtype()
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We currently force a synchronize_net() in netdev_set_master()
This seems necessary only when a slave had a master and we dismantle it.
In the other case ("ifenslave bond0 ethO"), we dont need this long
delay.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After the type change, addresses in unicast and multicast lists wouldn't make
sense, not to mention possible different lenghts. So flush both lists here.
Note "dev_addr_discard" will be very soon replaced by "dev_mc_flush" (once
mc_list conversion will be done).
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When doing "ifenslave -d bond0 eth0", there is chance to get NULL
dereference in netif_receive_skb(), because dev->master suddenly becomes
NULL after we tested it.
We should use ACCESS_ONCE() to avoid this (or rcu_dereference())
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds the possibility to refuse the bonding type change for
other subsystems (such as for example bridge, vlan, etc.)
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch implements software receive side packet steering (RPS). RPS
distributes the load of received packet processing across multiple CPUs.
Problem statement: Protocol processing done in the NAPI context for received
packets is serialized per device queue and becomes a bottleneck under high
packet load. This substantially limits pps that can be achieved on a single
queue NIC and provides no scaling with multiple cores.
This solution queues packets early on in the receive path on the backlog queues
of other CPUs. This allows protocol processing (e.g. IP and TCP) to be
performed on packets in parallel. For each device (or each receive queue in
a multi-queue device) a mask of CPUs is set to indicate the CPUs that can
process packets. A CPU is selected on a per packet basis by hashing contents
of the packet header (e.g. the TCP or UDP 4-tuple) and using the result to index
into the CPU mask. The IPI mechanism is used to raise networking receive
softirqs between CPUs. This effectively emulates in software what a multi-queue
NIC can provide, but is generic requiring no device support.
Many devices now provide a hash over the 4-tuple on a per packet basis
(e.g. the Toeplitz hash). This patch allow drivers to set the HW reported hash
in an skb field, and that value in turn is used to index into the RPS maps.
Using the HW generated hash can avoid cache misses on the packet when
steering it to a remote CPU.
The CPU mask is set on a per device and per queue basis in the sysfs variable
/sys/class/net/<device>/queues/rx-<n>/rps_cpus. This is a set of canonical
bit maps for receive queues in the device (numbered by <n>). If a device
does not support multi-queue, a single variable is used for the device (rx-0).
Generally, we have found this technique increases pps capabilities of a single
queue device with good CPU utilization. Optimal settings for the CPU mask
seem to depend on architectures and cache hierarcy. Below are some results
running 500 instances of netperf TCP_RR test with 1 byte req. and resp.
Results show cumulative transaction rate and system CPU utilization.
e1000e on 8 core Intel
Without RPS: 108K tps at 33% CPU
With RPS: 311K tps at 64% CPU
forcedeth on 16 core AMD
Without RPS: 156K tps at 15% CPU
With RPS: 404K tps at 49% CPU
bnx2x on 16 core AMD
Without RPS 567K tps at 61% CPU (4 HW RX queues)
Without RPS 738K tps at 96% CPU (8 HW RX queues)
With RPS: 854K tps at 76% CPU (4 HW RX queues)
Caveats:
- The benefits of this patch are dependent on architecture and cache hierarchy.
Tuning the masks to get best performance is probably necessary.
- This patch adds overhead in the path for processing a single packet. In
a lightly loaded server this overhead may eliminate the advantages of
increased parallelism, and possibly cause some relative performance degradation.
We have found that masks that are cache aware (share same caches with
the interrupting CPU) mitigate much of this.
- The RPS masks can be changed dynamically, however whenever the mask is changed
this introduces the possibility of generating out of order packets. It's
probably best not change the masks too frequently.
Signed-off-by: Tom Herbert <therbert@google.com>
include/linux/netdevice.h | 32 ++++-
include/linux/skbuff.h | 3 +
net/core/dev.c | 335 +++++++++++++++++++++++++++++++++++++--------
net/core/net-sysfs.c | 225 ++++++++++++++++++++++++++++++-
net/core/skbuff.c | 2 +
5 files changed, 538 insertions(+), 59 deletions(-)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric W. Biederman