currently cbq works incorrectly for limits > 10% real link bandwidth,
and practically does not work for limits > 50% real link bandwidth.
Below are results of experiments taken on 1 Gbit link
In shaper | Actual Result
-----------+---------------
100M | 108 Mbps
200M | 244 Mbps
300M | 412 Mbps
500M | 893 Mbps
This happen because of q->now changes incorrectly in cbq_dequeue():
when it is called before real end of packet transmitting,
L2T is greater than real time delay, q_now gets an extra boost
but never compensate it.
To fix this problem we prevent change of q->now until its synchronization
with real time.
Signed-off-by: Vasily Averin <vvs@openvz.org>
Reviewed-by: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, we hold a max of sch->limit -1 number of packets instead of
sch->limit packets. Fix this off-by-one error.
Signed-off-by: Vijay Subramanian <subramanian.vijay@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
QFQ+ can select for service only 'eligible' aggregates, i.e.,
aggregates that would have started to be served also in the emulated
ideal system. As a consequence, for QFQ+ to be work conserving, at
least one of the active aggregates must be eligible when it is time to
choose the next aggregate to serve.
The set of eligible aggregates is updated through the function
qfq_update_eligible(), which does guarantee that, after its
invocation, at least one of the active aggregates is eligible.
Because of this property, this function is invoked in
qfq_deactivate_agg() to guarantee that at least one of the active
aggregates is still eligible after an aggregate has been deactivated.
In particular, the critical case is when there are other active
aggregates, but the aggregate being deactivated happens to be the only
one eligible.
However, this precaution is not needed for QFQ+ to be work conserving,
because update_eligible() is always invoked also at the beginning of
qfq_choose_next_agg(). This patch removes the additional invocation of
update_eligible() in qfq_deactivate_agg().
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Reviewed-by: Fabio Checconi <fchecconi@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
By definition of (the algorithm of) QFQ+, the system virtual time must
be pushed up only if there is no 'eligible' aggregate, i.e. no
aggregate that would have started to be served also in the ideal
system emulated by QFQ+. QFQ+ serves only eligible aggregates, hence
the aggregate currently in service is eligible. As a consequence, to
decide whether there is no eligible aggregate, QFQ+ must also check
whether there is no aggregate in service.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Reviewed-by: Fabio Checconi <fchecconi@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Aggregate budgets are computed so as to guarantee that, after an
aggregate has been selected for service, that aggregate has enough
budget to serve at least one maximum-size packet for the classes it
contains. For this reason, after a new aggregate has been selected
for service, its next packet is immediately dequeued, without any
further control.
The maximum packet size for a class, lmax, can be changed through
qfq_change_class(). In case the user sets lmax to a lower value than
the the size of some of the still-to-arrive packets, QFQ+ will
automatically push up lmax as it enqueues these packets. This
automatic push up is likely to happen with TSO/GSO.
In any case, if lmax is assigned a lower value than the size of some
of the packets already enqueued for the class, then the following
problem may occur: the size of the next packet to dequeue for the
class may happen to be larger than lmax, after the aggregate to which
the class belongs has been just selected for service. In this case,
even the budget of the aggregate, which is an unsigned value, may be
lower than the size of the next packet to dequeue. After dequeueing
this packet and subtracting its size from the budget, the latter would
wrap around.
This fix prevents the budget from wrapping around after any packet
dequeue.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Reviewed-by: Fabio Checconi <fchecconi@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If no aggregate is in service, then the function qfq_dequeue() does
not dequeue any packet. For this reason, to guarantee QFQ+ to be work
conserving, a just-activated aggregate must be set as in service
immediately if it happens to be the only active aggregate.
This is done by the function qfq_enqueue().
Unfortunately, the function qfq_add_to_agg(), used to add a class to
an aggregate, does not perform this important additional operation.
In particular, if: 1) qfq_add_to_agg() is invoked to complete the move
of a class from a source aggregate, becoming, for this move, inactive,
to a destination aggregate, becoming instead active, and 2) the
destination aggregate becomes the only active aggregate, then this
aggregate is not however set as in service. QFQ+ remains then in a
non-work-conserving state until a new invocation of qfq_enqueue()
recovers the situation.
This fix solves the problem by moving the logic for setting an
aggregate as in service directly into the function qfq_activate_agg().
Hence, from whatever point qfq_activate_aggregate() is invoked, QFQ+
remains work conserving. Since the more-complex logic of this new
version of activate_aggregate() is not necessary, in qfq_dequeue(), to
reschedule an aggregate that finishes its budget, then the aggregate
is now rescheduled by invoking directly the functions needed.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Reviewed-by: Fabio Checconi <fchecconi@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Between two invocations of make_eligible, the system virtual time may
happen to grow enough that, in its binary representation, a bit with
higher order than 31 flips. This happens especially with
TSO/GSO. Before this fix, the mask used in make_eligible was computed
as (1UL<<index_of_last_flipped_bit)-1, whose value is well defined on
a 64-bit architecture, because index_of_flipped_bit <= 63, but is in
general undefined on a 32-bit architecture if index_of_flipped_bit > 31.
The fix just replaces 1UL with 1ULL.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Reviewed-by: Fabio Checconi <fchecconi@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
QFQ+ schedules the active aggregates in a group using a bucket list
(one list per group). The bucket in which each aggregate is inserted
depends on the aggregate's timestamps, and the number
of buckets in a group is enough to accomodate the possible (range of)
values of the timestamps of all the aggregates in the group. For this
property to hold, timestamps must however be computed correctly. One
necessary condition for computing timestamps correctly is that the
number of bits dequeued for each aggregate, while the aggregate is in
service, does not exceed the maximum budget budgetmax assigned to the
aggregate.
For each aggregate, budgetmax is proportional to the number of classes
in the aggregate. If the number of classes of the aggregate is
decreased through qfq_change_class(), then budgetmax is decreased
automatically as well. Problems may occur if the aggregate is in
service when budgetmax is decreased, because the current remaining
budget of the aggregate and/or the service already received by the
aggregate may happen to be larger than the new value of budgetmax. In
this case, when the aggregate is eventually deselected and its
timestamps are updated, the aggregate may happen to have received an
amount of service larger than budgetmax. This may cause the aggregate
to be assigned a higher virtual finish time than the maximum
acceptable value for the last bucket in the bucket list of the group.
This fix introduces a cap that addresses this issue.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Reviewed-by: Fabio Checconi <fchecconi@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
I'm not sure why, but the hlist for each entry iterators were conceived
list_for_each_entry(pos, head, member)
The hlist ones were greedy and wanted an extra parameter:
hlist_for_each_entry(tpos, pos, head, member)
Why did they need an extra pos parameter? I'm not quite sure. Not only
they don't really need it, it also prevents the iterator from looking
exactly like the list iterator, which is unfortunate.
Besides the semantic patch, there was some manual work required:
- Fix up the actual hlist iterators in linux/list.h
- Fix up the declaration of other iterators based on the hlist ones.
- A very small amount of places were using the 'node' parameter, this
was modified to use 'obj->member' instead.
- Coccinelle didn't handle the hlist_for_each_entry_safe iterator
properly, so those had to be fixed up manually.
The semantic patch which is mostly the work of Peter Senna Tschudin is here:
@@
iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host;
type T;
expression a,c,d,e;
identifier b;
statement S;
@@
-T b;
<+... when != b
(
hlist_for_each_entry(a,
- b,
c, d) S
|
hlist_for_each_entry_continue(a,
- b,
c) S
|
hlist_for_each_entry_from(a,
- b,
c) S
|
hlist_for_each_entry_rcu(a,
- b,
c, d) S
|
hlist_for_each_entry_rcu_bh(a,
- b,
c, d) S
|
hlist_for_each_entry_continue_rcu_bh(a,
- b,
c) S
|
for_each_busy_worker(a, c,
- b,
d) S
|
ax25_uid_for_each(a,
- b,
c) S
|
ax25_for_each(a,
- b,
c) S
|
inet_bind_bucket_for_each(a,
- b,
c) S
|
sctp_for_each_hentry(a,
- b,
c) S
|
sk_for_each(a,
- b,
c) S
|
sk_for_each_rcu(a,
- b,
c) S
|
sk_for_each_from
-(a, b)
+(a)
S
+ sk_for_each_from(a) S
|
sk_for_each_safe(a,
- b,
c, d) S
|
sk_for_each_bound(a,
- b,
c) S
|
hlist_for_each_entry_safe(a,
- b,
c, d, e) S
|
hlist_for_each_entry_continue_rcu(a,
- b,
c) S
|
nr_neigh_for_each(a,
- b,
c) S
|
nr_neigh_for_each_safe(a,
- b,
c, d) S
|
nr_node_for_each(a,
- b,
c) S
|
nr_node_for_each_safe(a,
- b,
c, d) S
|
- for_each_gfn_sp(a, c, d, b) S
+ for_each_gfn_sp(a, c, d) S
|
- for_each_gfn_indirect_valid_sp(a, c, d, b) S
+ for_each_gfn_indirect_valid_sp(a, c, d) S
|
for_each_host(a,
- b,
c) S
|
for_each_host_safe(a,
- b,
c, d) S
|
for_each_mesh_entry(a,
- b,
c, d) S
)
...+>
[akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c]
[akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c]
[akpm@linux-foundation.org: checkpatch fixes]
[akpm@linux-foundation.org: fix warnings]
[akpm@linux-foudnation.org: redo intrusive kvm changes]
Tested-by: Peter Senna Tschudin <peter.senna@gmail.com>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
proc_net_remove is only used to remove proc entries
that under /proc/net,it's not a general function for
removing proc entries of netns. if we want to remove
some proc entries which under /proc/net/stat/, we still
need to call remove_proc_entry.
this patch use remove_proc_entry to replace proc_net_remove.
we can remove proc_net_remove after this patch.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Right now, some modules such as bonding use proc_create
to create proc entries under /proc/net/, and other modules
such as ipv4 use proc_net_fops_create.
It looks a little chaos.this patch changes all of
proc_net_fops_create to proc_create. we can remove
proc_net_fops_create after this patch.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This function will be used in next GRE_GSO patch. This patch does
not change any functionality.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Current act_police uses rate table computed by the "tc" userspace
program, which has the following issue:
The rate table has 256 entries to map packet lengths to token (time
units). With TSO sized packets, the 256 entry granularity leads to
loss/gain of rate, making the token bucket inaccurate.
Thus, instead of relying on rate table, this patch explicitly computes
the time and accounts for packet transmission times with nanosecond
granularity.
This is a followup to 56b765b79e
("htb: improved accuracy at high rates").
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
It's not used anywhere else, so move it.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current TBF uses rate table computed by the "tc" userspace program,
which has the following issue:
The rate table has 256 entries to map packet lengths to
token (time units). With TSO sized packets, the 256 entry granularity
leads to loss/gain of rate, making the token bucket inaccurate.
Thus, instead of relying on rate table, this patch explicitly computes
the time and accounts for packet transmission times with nanosecond
granularity.
This is a followup to 56b765b79e
("htb: improved accuracy at high rates").
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tbf will need to schedule watchdog in ns. No need to convert it twice.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As it is going to be used in tbf as well, push these to generic code.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These are in ns so convert from ticks to ns.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These are initialized correctly a couple of lines later in the
function.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
The bnx2x gso_type setting bug fix in 'net' conflicted with
changes in 'net-next' that broke the gso_* setting logic
out into a seperate function, which also fixes the bug in
question. Thus, use the 'net-next' version.
Signed-off-by: David S. Miller <davem@davemloft.net>
in htb_change_class() cl->buffer and cl->buffer are stored in ns.
So in dump, convert them back to psched ticks.
Note this was introduced by:
commit 56b765b79e
htb: improved accuracy at high rates
Please consider this for -net/-stable.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/intel/e1000e/ethtool.c
drivers/net/vmxnet3/vmxnet3_drv.c
drivers/net/wireless/iwlwifi/dvm/tx.c
net/ipv6/route.c
The ipv6 route.c conflict is simple, just ignore the 'net' side change
as we fixed the same problem in 'net-next' by eliminating cached
neighbours from ipv6 routes.
The e1000e conflict is an addition of a new statistic in the ethtool
code, trivial.
The vmxnet3 conflict is about one change in 'net' removing a guarding
conditional, whilst in 'net-next' we had a netdev_info() conversion.
The iwlwifi conflict is dealing with a WARN_ON() conversion in
'net-next' vs. a revert happening in 'net'.
Signed-off-by: David S. Miller <davem@davemloft.net>
The delay calculation with the rate extension introduces in v3.3 does
not properly work, if other packets are still queued for transmission.
For the delay calculation to work, both delay types (latency and delay
introduces by rate limitation) have to be handled differently. The
latency delay for a packet can overlap with the delay of other packets.
The delay introduced by the rate however is separate, and can only
start, once all other rate-introduced delays finished.
Latency delay is from same distribution for each packet, rate delay
depends on the packet size.
.: latency delay
-: rate delay
x: additional delay we have to wait since another packet is currently
transmitted
.....---- Packet 1
.....xx------ Packet 2
.....------ Packet 3
^^^^^
latency stacks
^^
rate delay doesn't stack
^^
latency stacks
-----> time
When a packet is enqueued, we first consider the latency delay. If other
packets are already queued, we can reduce the latency delay until the
last packet in the queue is send, however the latency delay cannot be
<0, since this would mean that the rate is overcommitted. The new
reference point is the time at which the last packet will be send. To
find the time, when the packet should be send, the rate introduces delay
has to be added on top of that.
Signed-off-by: Johannes Naab <jn@stusta.de>
Acked-by: Hagen Paul Pfeifer <hagen@jauu.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet pointed out that act_mirred needs to find the current net_ns,
and struct net pointer is not provided in the call chain. His original
patch made use of current->nsproxy->net_ns to find the network namespace,
but this fails to work correctly for userspace code that makes use of
netlink sockets in different network namespaces. Instead, pass the
"struct net *" down along the call chain to where it is needed.
This version removes the ifb changes as Eric has submitted that patch
separately, but is otherwise identical to the previous version.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Tested-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixed integer overflow in function htb_dequeue
Signed-off-by: Stefan Hasko <hasko.stevo@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking changes from David Miller:
1) Allow to dump, monitor, and change the bridge multicast database
using netlink. From Cong Wang.
2) RFC 5961 TCP blind data injection attack mitigation, from Eric
Dumazet.
3) Networking user namespace support from Eric W. Biederman.
4) tuntap/virtio-net multiqueue support by Jason Wang.
5) Support for checksum offload of encapsulated packets (basically,
tunneled traffic can still be checksummed by HW). From Joseph
Gasparakis.
6) Allow BPF filter access to VLAN tags, from Eric Dumazet and
Daniel Borkmann.
7) Bridge port parameters over netlink and BPDU blocking support
from Stephen Hemminger.
8) Improve data access patterns during inet socket demux by rearranging
socket layout, from Eric Dumazet.
9) TIPC protocol updates and cleanups from Ying Xue, Paul Gortmaker, and
Jon Maloy.
10) Update TCP socket hash sizing to be more in line with current day
realities. The existing heurstics were choosen a decade ago.
From Eric Dumazet.
11) Fix races, queue bloat, and excessive wakeups in ATM and
associated drivers, from Krzysztof Mazur and David Woodhouse.
12) Support DOVE (Distributed Overlay Virtual Ethernet) extensions
in VXLAN driver, from David Stevens.
13) Add "oops_only" mode to netconsole, from Amerigo Wang.
14) Support set and query of VEB/VEPA bridge mode via PF_BRIDGE, also
allow DCB netlink to work on namespaces other than the initial
namespace. From John Fastabend.
15) Support PTP in the Tigon3 driver, from Matt Carlson.
16) tun/vhost zero copy fixes and improvements, plus turn it on
by default, from Michael S. Tsirkin.
17) Support per-association statistics in SCTP, from Michele
Baldessari.
And many, many, driver updates, cleanups, and improvements. Too
numerous to mention individually.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1722 commits)
net/mlx4_en: Add support for destination MAC in steering rules
net/mlx4_en: Use generic etherdevice.h functions.
net: ethtool: Add destination MAC address to flow steering API
bridge: add support of adding and deleting mdb entries
bridge: notify mdb changes via netlink
ndisc: Unexport ndisc_{build,send}_skb().
uapi: add missing netconf.h to export list
pkt_sched: avoid requeues if possible
solos-pci: fix double-free of TX skb in DMA mode
bnx2: Fix accidental reversions.
bna: Driver Version Updated to 3.1.2.1
bna: Firmware update
bna: Add RX State
bna: Rx Page Based Allocation
bna: TX Intr Coalescing Fix
bna: Tx and Rx Optimizations
bna: Code Cleanup and Enhancements
ath9k: check pdata variable before dereferencing it
ath5k: RX timestamp is reported at end of frame
ath9k_htc: RX timestamp is reported at end of frame
...
Pull cgroup changes from Tejun Heo:
"A lot of activities on cgroup side. The big changes are focused on
making cgroup hierarchy handling saner.
- cgroup_rmdir() had peculiar semantics - it allowed cgroup
destruction to be vetoed by individual controllers and tried to
drain refcnt synchronously. The vetoing never worked properly and
caused good deal of contortions in cgroup. memcg was the last
reamining user. Michal Hocko removed the usage and cgroup_rmdir()
path has been simplified significantly. This was done in a
separate branch so that the memcg people can base further memcg
changes on top.
- The above allowed cleaning up cgroup lifecycle management and
implementation of generic cgroup iterators which are used to
improve hierarchy support.
- cgroup_freezer updated to allow migration in and out of a frozen
cgroup and handle hierarchy. If a cgroup is frozen, all descendant
cgroups are frozen.
- netcls_cgroup and netprio_cgroup updated to handle hierarchy
properly.
- Various fixes and cleanups.
- Two merge commits. One to pull in memcg and rmdir cleanups (needed
to build iterators). The other pulled in cgroup/for-3.7-fixes for
device_cgroup fixes so that further device_cgroup patches can be
stacked on top."
Fixed up a trivial conflict in mm/memcontrol.c as per Tejun (due to
commit bea8c150a7 ("memcg: fix hotplugged memory zone oops") in master
touching code close to commit 2ef37d3fe4 ("memcg: Simplify
mem_cgroup_force_empty_list error handling") in for-3.8)
* 'for-3.8' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup: (65 commits)
cgroup: update Documentation/cgroups/00-INDEX
cgroup_rm_file: don't delete the uncreated files
cgroup: remove subsystem files when remounting cgroup
cgroup: use cgroup_addrm_files() in cgroup_clear_directory()
cgroup: warn about broken hierarchies only after css_online
cgroup: list_del_init() on removed events
cgroup: fix lockdep warning for event_control
cgroup: move list add after list head initilization
netprio_cgroup: allow nesting and inherit config on cgroup creation
netprio_cgroup: implement netprio[_set]_prio() helpers
netprio_cgroup: use cgroup->id instead of cgroup_netprio_state->prioidx
netprio_cgroup: reimplement priomap expansion
netprio_cgroup: shorten variable names in extend_netdev_table()
netprio_cgroup: simplify write_priomap()
netcls_cgroup: move config inheritance to ->css_online() and remove .broken_hierarchy marking
cgroup: remove obsolete guarantee from cgroup_task_migrate.
cgroup: add cgroup->id
cgroup, cpuset: remove cgroup_subsys->post_clone()
cgroup: s/CGRP_CLONE_CHILDREN/CGRP_CPUSET_CLONE_CHILDREN/
cgroup: rename ->create/post_create/pre_destroy/destroy() to ->css_alloc/online/offline/free()
...
With BQL being deployed, we can more likely have following behavior :
We dequeue a packet from qdisc in dequeue_skb(), then we realize target
tx queue is in XOFF state in sch_direct_xmit(), and we have to hold the
skb into gso_skb for later.
This shows in stats (tc -s qdisc dev eth0) as requeues.
Problem of these requeues is that high priority packets can not be
dequeued as long as this (possibly low prio and big TSO packet) is not
removed from gso_skb.
At 1Gbps speed, a full size TSO packet is 500 us of extra latency.
In some cases, we know that all packets dequeued from a qdisc are
for a particular and known txq :
- If device is non multi queue
- For all MQ/MQPRIO slave qdiscs
This patch introduces a new qdisc flag, TCQ_F_ONETXQUEUE to mark
this capability, so that dequeue_skb() is allowed to dequeue a packet
only if the associated txq is not stopped.
This indeed reduce latencies for high prio packets (or improve fairness
with sfq/fq_codel), and almost remove qdisc 'requeues'.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch turns QFQ into QFQ+, a variant of QFQ that provides the
following two benefits: 1) QFQ+ is faster than QFQ, 2) differently
from QFQ, QFQ+ correctly schedules also non-leaves classes in a
hierarchical setting. A detailed description of QFQ+, plus a
performance comparison with DRR and QFQ, can be found in [1].
[1] P. Valente, "Reducing the Execution Time of Fair-Queueing Schedulers"
http://algo.ing.unimo.it/people/paolo/agg-sched/agg-sched.pdf
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch makes it possible to build the CAN Identifier into the kernel, even
if the CAN support is build as a module.
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
It turns out that we'll have to live with attributes which are
inherited at cgroup creation time but not affected by further updates
to the parent afterwards - such attributes are already in wide use
e.g. for cpuset.
So, there's nothing to do for netcls_cgroup for hierarchy support.
Its current behavior - inherit only during creation - is good enough.
Move config inheriting from ->css_alloc() to ->css_online() for
consistency, which doesn't change behavior at all, and remove
.broken_hierarchy marking.
Signed-off-by: Tejun Heo <tj@kernel.org>
Tested-and-Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Acked-by: David S. Miller <davem@davemloft.net>
Rename cgroup_subsys css lifetime related callbacks to better describe
what their roles are. Also, update documentation.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
- In rtnetlink_rcv_msg convert the capable(CAP_NET_ADMIN) check
to ns_capable(net->user-ns, CAP_NET_ADMIN). Allowing unprivileged
users to make netlink calls to modify their local network
namespace.
- In the rtnetlink doit methods add capable(CAP_NET_ADMIN) so
that calls that are not safe for unprivileged users are still
protected.
Later patches will remove the extra capable calls from methods
that are safe for unprivilged users.
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
Minor conflict between the BCM_CNIC define removal in net-next
and a bug fix added to net. Based upon a conflict resolution
patch posted by Stephen Rothwell.
Signed-off-by: David S. Miller <davem@davemloft.net>
If the max packet size for some class (configured through tc) is
violated by the actual size of the packets of that class, then QFQ
would not schedule classes correctly, and the data structures
implementing the bucket lists may get corrupted. This problem occurs
with TSO/GSO even if the max packet size is set to the MTU, and is,
e.g., the cause of the failure reported in [1]. Two patches have been
proposed to solve this problem in [2], one of them is a preliminary
version of this patch.
This patch addresses the above issues by: 1) setting QFQ parameters to
proper values for supporting TSO/GSO (in particular, setting the
maximum possible packet size to 64KB), 2) automatically increasing the
max packet size for a class, lmax, when a packet with a larger size
than the current value of lmax arrives.
The drawback of the first point is that the maximum weight for a class
is now limited to 4096, which is equal to 1/16 of the maximum weight
sum.
Finally, this patch also forcibly caps the timestamps of a class if
they are too high to be stored in the bucket list. This capping, taken
from QFQ+ [3], handles the unfrequent case described in the comment to
the function slot_insert.
[1] http://marc.info/?l=linux-netdev&m=134968777902077&w=2
[2] http://marc.info/?l=linux-netdev&m=135096573507936&w=2
[3] http://marc.info/?l=linux-netdev&m=134902691421670&w=2
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Tested-by: Cong Wang <amwang@redhat.com>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 56b765b79e (htb: improved accuracy at high rates)
introduced two bugs :
1) one bstats_update() was inadvertently removed from
htb_dequeue_tree(), breaking statistics/rate estimation.
2) Missing qdisc_put_rtab() calls in htb_change_class(),
leaking kernel memory, now struct htb_class no longer
retains pointers to qdisc_rate_table structs.
Since only rate is used, dont use qdisc_get_rtab() calls
copying data we ignore anyway.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Vimalkumar <j.vimal@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current HTB (and TBF) uses rate table computed by the "tc"
userspace program, which has the following issue:
The rate table has 256 entries to map packet lengths
to token (time units). With TSO sized packets, the
256 entry granularity leads to loss/gain of rate,
making the token bucket inaccurate.
Thus, instead of relying on rate table, this patch
explicitly computes the time and accounts for packet
transmission times with nanosecond granularity.
This greatly improves accuracy of HTB with a wide
range of packet sizes.
Example:
tc qdisc add dev $dev root handle 1: \
htb default 1
tc class add dev $dev classid 1:1 parent 1: \
rate 5Gbit mtu 64k
Here is an example of inaccuracy:
$ iperf -c host -t 10 -i 1
With old htb:
eth4: 34.76 Mb/s In 5827.98 Mb/s Out - 65836.0 p/s In 481273.0 p/s Out
[SUM] 9.0-10.0 sec 669 MBytes 5.61 Gbits/sec
[SUM] 0.0-10.0 sec 6.50 GBytes 5.58 Gbits/sec
With new htb:
eth4: 28.36 Mb/s In 5208.06 Mb/s Out - 53704.0 p/s In 430076.0 p/s Out
[SUM] 9.0-10.0 sec 594 MBytes 4.98 Gbits/sec
[SUM] 0.0-10.0 sec 5.80 GBytes 4.98 Gbits/sec
The bits per second on the wire is still 5200Mb/s with new HTB
because qdisc accounts for packet length using skb->len, which
is smaller than total bytes on the wire if GSO is used. But
that is for another patch regardless of how time is accounted.
Many thanks to Eric Dumazet for review and feedback.
Signed-off-by: Vimalkumar <j.vimal@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The cgroup logic part of net_cls is very similar as the one in
net_prio. Let's stream line the net_cls logic with the net_prio one.
The net_prio update logic was changed by following commit (note there
were some changes necessary later on)
commit 406a3c638c
Author: John Fastabend <john.r.fastabend@intel.com>
Date: Fri Jul 20 10:39:25 2012 +0000
net: netprio_cgroup: rework update socket logic
Instead of updating the sk_cgrp_prioidx struct field on every send
this only updates the field when a task is moved via cgroup
infrastructure.
This allows sockets that may be used by a kernel worker thread
to be managed. For example in the iscsi case today a user can
put iscsid in a netprio cgroup and control traffic will be sent
with the correct sk_cgrp_prioidx value set but as soon as data
is sent the kernel worker thread isssues a send and sk_cgrp_prioidx
is updated with the kernel worker threads value which is the
default case.
It seems more correct to only update the field when the user
explicitly sets it via control group infrastructure. This allows
the users to manage sockets that may be used with other threads.
Since classid is now updated when the task is moved between the
cgroups, we don't have to call sock_update_classid() from various
places to ensure we always using the latest classid value.
[v2: Use iterate_fd() instead of open coding]
Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Cc: Li Zefan <lizefan@huawei.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Joe Perches <joe@perches.com>
Cc: John Fastabend <john.r.fastabend@intel.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Stanislav Kinsbursky <skinsbursky@parallels.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: <netdev@vger.kernel.org>
Cc: <cgroups@vger.kernel.org>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ns_to_ktime() seems better than ktime_set() + ktime_add_ns()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking changes from David Miller:
1) GRE now works over ipv6, from Dmitry Kozlov.
2) Make SCTP more network namespace aware, from Eric Biederman.
3) TEAM driver now works with non-ethernet devices, from Jiri Pirko.
4) Make openvswitch network namespace aware, from Pravin B Shelar.
5) IPV6 NAT implementation, from Patrick McHardy.
6) Server side support for TCP Fast Open, from Jerry Chu and others.
7) Packet BPF filter supports MOD and XOR, from Eric Dumazet and Daniel
Borkmann.
8) Increate the loopback default MTU to 64K, from Eric Dumazet.
9) Use a per-task rather than per-socket page fragment allocator for
outgoing networking traffic. This benefits processes that have very
many mostly idle sockets, which is quite common.
From Eric Dumazet.
10) Use up to 32K for page fragment allocations, with fallbacks to
smaller sizes when higher order page allocations fail. Benefits are
a) less segments for driver to process b) less calls to page
allocator c) less waste of space.
From Eric Dumazet.
11) Allow GRO to be used on GRE tunnels, from Eric Dumazet.
12) VXLAN device driver, one way to handle VLAN issues such as the
limitation of 4096 VLAN IDs yet still have some level of isolation.
From Stephen Hemminger.
13) As usual there is a large boatload of driver changes, with the scale
perhaps tilted towards the wireless side this time around.
Fix up various fairly trivial conflicts, mostly caused by the user
namespace changes.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1012 commits)
hyperv: Add buffer for extended info after the RNDIS response message.
hyperv: Report actual status in receive completion packet
hyperv: Remove extra allocated space for recv_pkt_list elements
hyperv: Fix page buffer handling in rndis_filter_send_request()
hyperv: Fix the missing return value in rndis_filter_set_packet_filter()
hyperv: Fix the max_xfer_size in RNDIS initialization
vxlan: put UDP socket in correct namespace
vxlan: Depend on CONFIG_INET
sfc: Fix the reported priorities of different filter types
sfc: Remove EFX_FILTER_FLAG_RX_OVERRIDE_IP
sfc: Fix loopback self-test with separate_tx_channels=1
sfc: Fix MCDI structure field lookup
sfc: Add parentheses around use of bitfield macro arguments
sfc: Fix null function pointer in efx_sriov_channel_type
vxlan: virtual extensible lan
igmp: export symbol ip_mc_leave_group
netlink: add attributes to fdb interface
tg3: unconditionally select HWMON support when tg3 is enabled.
Revert "net: ti cpsw ethernet: allow reading phy interface mode from DT"
gre: fix sparse warning
...
Pull user namespace changes from Eric Biederman:
"This is a mostly modest set of changes to enable basic user namespace
support. This allows the code to code to compile with user namespaces
enabled and removes the assumption there is only the initial user
namespace. Everything is converted except for the most complex of the
filesystems: autofs4, 9p, afs, ceph, cifs, coda, fuse, gfs2, ncpfs,
nfs, ocfs2 and xfs as those patches need a bit more review.
The strategy is to push kuid_t and kgid_t values are far down into
subsystems and filesystems as reasonable. Leaving the make_kuid and
from_kuid operations to happen at the edge of userspace, as the values
come off the disk, and as the values come in from the network.
Letting compile type incompatible compile errors (present when user
namespaces are enabled) guide me to find the issues.
The most tricky areas have been the places where we had an implicit
union of uid and gid values and were storing them in an unsigned int.
Those places were converted into explicit unions. I made certain to
handle those places with simple trivial patches.
Out of that work I discovered we have generic interfaces for storing
quota by projid. I had never heard of the project identifiers before.
Adding full user namespace support for project identifiers accounts
for most of the code size growth in my git tree.
Ultimately there will be work to relax privlige checks from
"capable(FOO)" to "ns_capable(user_ns, FOO)" where it is safe allowing
root in a user names to do those things that today we only forbid to
non-root users because it will confuse suid root applications.
While I was pushing kuid_t and kgid_t changes deep into the audit code
I made a few other cleanups. I capitalized on the fact we process
netlink messages in the context of the message sender. I removed
usage of NETLINK_CRED, and started directly using current->tty.
Some of these patches have also made it into maintainer trees, with no
problems from identical code from different trees showing up in
linux-next.
After reading through all of this code I feel like I might be able to
win a game of kernel trivial pursuit."
Fix up some fairly trivial conflicts in netfilter uid/git logging code.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace: (107 commits)
userns: Convert the ufs filesystem to use kuid/kgid where appropriate
userns: Convert the udf filesystem to use kuid/kgid where appropriate
userns: Convert ubifs to use kuid/kgid
userns: Convert squashfs to use kuid/kgid where appropriate
userns: Convert reiserfs to use kuid and kgid where appropriate
userns: Convert jfs to use kuid/kgid where appropriate
userns: Convert jffs2 to use kuid and kgid where appropriate
userns: Convert hpfs to use kuid and kgid where appropriate
userns: Convert btrfs to use kuid/kgid where appropriate
userns: Convert bfs to use kuid/kgid where appropriate
userns: Convert affs to use kuid/kgid wherwe appropriate
userns: On alpha modify linux_to_osf_stat to use convert from kuids and kgids
userns: On ia64 deal with current_uid and current_gid being kuid and kgid
userns: On ppc convert current_uid from a kuid before printing.
userns: Convert s390 getting uid and gid system calls to use kuid and kgid
userns: Convert s390 hypfs to use kuid and kgid where appropriate
userns: Convert binder ipc to use kuids
userns: Teach security_path_chown to take kuids and kgids
userns: Add user namespace support to IMA
userns: Convert EVM to deal with kuids and kgids in it's hmac computation
...
Pull cgroup hierarchy update from Tejun Heo:
"Currently, different cgroup subsystems handle nested cgroups
completely differently. There's no consistency among subsystems and
the behaviors often are outright broken.
People at least seem to agree that the broken hierarhcy behaviors need
to be weeded out if any progress is gonna be made on this front and
that the fallouts from deprecating the broken behaviors should be
acceptable especially given that the current behaviors don't make much
sense when nested.
This patch makes cgroup emit warning messages if cgroups for
subsystems with broken hierarchy behavior are nested to prepare for
fixing them in the future. This was put in a separate branch because
more related changes were expected (didn't make it this round) and the
memory cgroup wanted to pull in this and make changes on top."
* 'for-3.7-hierarchy' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: mark subsystems with broken hierarchy support and whine if cgroups are nested for them
Pull cgroup updates from Tejun Heo:
- xattr support added. The implementation is shared with tmpfs. The
usage is restricted and intended to be used to manage per-cgroup
metadata by system software. tmpfs changes are routed through this
branch with Hugh's permission.
- cgroup subsystem ID handling simplified.
* 'for-3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: Define CGROUP_SUBSYS_COUNT according the configuration
cgroup: Assign subsystem IDs during compile time
cgroup: Do not depend on a given order when populating the subsys array
cgroup: Wrap subsystem selection macro
cgroup: Remove CGROUP_BUILTIN_SUBSYS_COUNT
cgroup: net_prio: Do not define task_netpioidx() when not selected
cgroup: net_cls: Do not define task_cls_classid() when not selected
cgroup: net_cls: Move sock_update_classid() declaration to cls_cgroup.h
cgroup: trivial fixes for Documentation/cgroups/cgroups.txt
xattr: mark variable as uninitialized to make both gcc and smatch happy
fs: add missing documentation to simple_xattr functions
cgroup: add documentation on extended attributes usage
cgroup: rename subsys_bits to subsys_mask
cgroup: add xattr support
cgroup: revise how we re-populate root directory
xattr: extract simple_xattr code from tmpfs
Conflicts:
drivers/net/team/team.c
drivers/net/usb/qmi_wwan.c
net/batman-adv/bat_iv_ogm.c
net/ipv4/fib_frontend.c
net/ipv4/route.c
net/l2tp/l2tp_netlink.c
The team, fib_frontend, route, and l2tp_netlink conflicts were simply
overlapping changes.
qmi_wwan and bat_iv_ogm were of the "use HEAD" variety.
With help from Antonio Quartulli.
Signed-off-by: David S. Miller <davem@davemloft.net>
GCC refuses to recognize that all error control flows do in fact
set err to something.
Add an explicit initialization to shut it up.
net/sched/sch_drr.c: In function ‘drr_enqueue’:
net/sched/sch_drr.c:359:11: warning: ‘err’ may be used uninitialized in this function [-Wmaybe-uninitialized]
net/sched/sch_qfq.c: In function ‘qfq_enqueue’:
net/sched/sch_qfq.c:885:11: warning: ‘err’ may be used uninitialized in this function [-Wmaybe-uninitialized]
Signed-off-by: David S. Miller <davem@davemloft.net>
We currently use a per socket order-0 page cache for tcp_sendmsg()
operations.
This page is used to build fragments for skbs.
Its done to increase probability of coalescing small write() into
single segments in skbs still in write queue (not yet sent)
But it wastes a lot of memory for applications handling many mostly
idle sockets, since each socket holds one page in sk->sk_sndmsg_page
Its also quite inefficient to build TSO 64KB packets, because we need
about 16 pages per skb on arches where PAGE_SIZE = 4096, so we hit
page allocator more than wanted.
This patch adds a per task frag allocator and uses bigger pages,
if available. An automatic fallback is done in case of memory pressure.
(up to 32768 bytes per frag, thats order-3 pages on x86)
This increases TCP stream performance by 20% on loopback device,
but also benefits on other network devices, since 8x less frags are
mapped on transmit and unmapped on tx completion. Alexander Duyck
mentioned a probable performance win on systems with IOMMU enabled.
Its possible some SG enabled hardware cant cope with bigger fragments,
but their ndo_start_xmit() should already handle this, splitting a
fragment in sub fragments, since some arches have PAGE_SIZE=65536
Successfully tested on various ethernet devices.
(ixgbe, igb, bnx2x, tg3, mellanox mlx4)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Vijay Subramanian <subramanian.vijay@gmail.com>
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Vijay Subramanian <subramanian.vijay@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If the old timestamps of a class, say cl, are stale when the class
becomes active, then QFQ may assign to cl a much higher start time
than the maximum value allowed. This may happen when QFQ assigns to
the start time of cl the finish time of a group whose classes are
characterized by a higher value of the ratio
max_class_pkt/weight_of_the_class with respect to that of
cl. Inserting a class with a too high start time into the bucket list
corrupts the data structure and may eventually lead to crashes.
This patch limits the maximum start time assigned to a class.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
net/netfilter/nfnetlink_log.c
net/netfilter/xt_LOG.c
Rather easy conflict resolution, the 'net' tree had bug fixes to make
sure we checked if a socket is a time-wait one or not and elide the
logging code if so.
Whereas on the 'net-next' side we are calculating the UID and GID from
the creds using different interfaces due to the user namespace changes
from Eric Biederman.
Signed-off-by: David S. Miller <davem@davemloft.net>
