Instead of using tp->q, use block to get the net pointer.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Store net pointer in the block structure. Along the way, introduce
qdisc_net helper which allows to easily obtain net pointer for
qdisc instance.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Prepare for removal of tp->q and store Qdisc pointer in the block
structure.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch makes a slight tweak to mqprio in order to bring the
classid values used back in line with what is used for mq. The general idea
is to reserve values :ffe0 - :ffef to identify hardware traffic classes
normally reported via dev->num_tc. By doing this we can maintain a
consistent behavior with mq for classid where :1 - :ffdf will represent a
physical qdisc mapped onto a Tx queue represented by classid - 1, and the
traffic classes will be mapped onto a known subset of classid values
reserved for our virtual qdiscs.
Note I reserved the range from :fff0 - :ffff since this way we might be
able to reuse these classid values with clsact and ingress which would mean
that for mq, mqprio, ingress, and clsact we should be able to maintain a
similar classid layout.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Jesus Sanchez-Palencia <jesus.sanchez-palencia@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jeff Kirsher says:
====================
40GbE Intel Wired LAN Driver Updates 2017-10-13
This series contains updates to mqprio and i40e.
Amritha introduces a new hardware offload mode in tc/mqprio where the TCs,
the queue configurations and bandwidth rate limits are offloaded to the
hardware. The existing mqprio framework is extended to configure the queue
counts and layout and also added support for rate limiting. This is
achieved through new netlink attributes for the 'mode' option which takes
values such as 'dcb' (default) and 'channel' and a 'shaper' option for
QoS attributes such as bandwidth rate limits in hw mode 1. Legacy devices
can fall back to the existing setup supporting hw mode 1 without these
additional options where only the TCs are offloaded and then the 'mode'
and 'shaper' options defaults to DCB support. The i40e driver enables the
new mqprio hardware offload mechanism factoring the TCs, queue
configuration and bandwidth rates by creating HW channel VSIs.
In this new mode, the priority to traffic class mapping and the user
specified queue ranges are used to configure the traffic class when the
'mode' option is set to 'channel'. This is achieved by creating HW
channels(VSI). A new channel is created for each of the traffic class
configuration offloaded via mqprio framework except for the first TC (TC0)
which is for the main VSI. TC0 for the main VSI is also reconfigured as
per user provided queue parameters. Finally, bandwidth rate limits are set
on these traffic classes through the shaper attribute by sending these
rates in addition to the number of TCs and the queue configurations.
Colin Ian King makes an array of constant values "constant".
Alan fixes and issue where on some firmware versions, we were failing to
actually fill out the phy_types which caused ethtool to not report any
link types. Also hardened against a potentially malicious VF by not
letting the VF to reset itself after requesting to change the number of
queues (via ethtool), let the PF reset the VF to institute the requested
changes.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Apparently ife_meta_id2name() is only called when
CONFIG_MODULES is defined.
This fixes:
net/sched/act_ife.c:251:20: warning: ‘ife_meta_id2name’ defined but not used [-Wunused-function]
static const char *ife_meta_id2name(u32 metaid)
^~~~~~~~~~~~~~~~
Fixes: d3f24ba895 ("net sched actions: fix module auto-loading")
Cc: Roman Mashak <mrv@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The offload types currently supported in mqprio are 0 (no offload) and
1 (offload only TCs) by setting these values for the 'hw' option. If
offloads are supported by setting the 'hw' option to 1, the default
offload mode is 'dcb' where only the TC values are offloaded to the
device. This patch introduces a new hardware offload mode called
'channel' with 'hw' set to 1 in mqprio which makes full use of the
mqprio options, the TCs, the queue configurations and the QoS parameters
for the TCs. This is achieved through a new netlink attribute for the
'mode' option which takes values such as 'dcb' (default) and 'channel'.
The 'channel' mode also supports QoS attributes for traffic class such as
minimum and maximum values for bandwidth rate limits.
This patch enables configuring additional HW shaper attributes associated
with a traffic class. Currently the shaper for bandwidth rate limiting is
supported which takes options such as minimum and maximum bandwidth rates
and are offloaded to the hardware in the 'channel' mode. The min and max
limits for bandwidth rates are provided by the user along with the TCs
and the queue configurations when creating the mqprio qdisc. The interface
can be extended to support new HW shapers in future through the 'shaper'
attribute.
Introduces a new data structure 'tc_mqprio_qopt_offload' for offloading
mqprio queue options and use this to be shared between the kernel and
device driver. This contains a copy of the existing data structure
for mqprio queue options. This new data structure can be extended when
adding new attributes for traffic class such as mode, shaper, shaper
parameters (bandwidth rate limits). The existing data structure for mqprio
queue options will be shared between the kernel and userspace.
Example:
queues 4@0 4@4 hw 1 mode channel shaper bw_rlimit\
min_rate 1Gbit 2Gbit max_rate 4Gbit 5Gbit
To dump the bandwidth rates:
qdisc mqprio 804a: root tc 2 map 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0
queues:(0:3) (4:7)
mode:channel
shaper:bw_rlimit min_rate:1Gbit 2Gbit max_rate:4Gbit 5Gbit
Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch changes the parameter updating via RCU and not protected by a
spinlock anymore. This reduce the time that the spinlock is being held.
Signed-off-by: Alexander Aring <aring@mojatatu.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch migrates the current counter handling which is protected by a
spinlock to a per-cpu counter handling. This reduce the time where the
spinlock is being held.
Signed-off-by: Alexander Aring <aring@mojatatu.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds the check of the two possible ife handlings encode
and decode to the init callback. The decode value is for usability
aspect and used in userspace code only. The current code offers encode
else decode only. This patch avoids any other option than this.
Signed-off-by: Alexander Aring <aring@mojatatu.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Macro __stringify_1() can stringify a macro argument, however IFE_META_*
are enums, so they never expand, however request_module expects an integer
in IFE module name, so as a result it always fails to auto-load.
Fixes: ef6980b6be ("introduce IFE action")
Signed-off-by: Roman Mashak <mrv@mojatatu.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Make style of module alias name consistent with other subsystems in kernel,
for example net devices.
Fixes: 084e2f6566 ("Support to encoding decoding skb mark on IFE action")
Fixes: 200e10f469 ("Support to encoding decoding skb prio on IFE action")
Fixes: 408fbc22ef ("net sched ife action: Introduce skb tcindex metadata encap decap")
Signed-off-by: Roman Mashak <mrv@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The helper and the struct field ares no longer used by any code,
so remove them.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The only user of cls_flower->egress_dev is mlx5. So do the conversion
there alongside with the code originating the call in cls_flower
function fl_hw_replace_filter to the newly introduced egress device
callback infrastucture.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduce infrastructure that allows drivers to register callbacks that
are called whenever tc would offload inserted rule and specified device
acts as tc action egress device.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Return dev directly, NULL if not possible. That is enough.
Makes no sense to pass struct net * to get_dev op, as there is only one
net possible, the one the action was created in. So just store it in
mirred priv and use directly.
Rename the mirred op callback function.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geeralize private netem_rb_to_skb()
TCP rtx queue will soon be converted to rb-tree,
so we will need skb_rbtree_walk() helpers.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move dissection of tunnel info from the flower classifier to the flow
dissector where all other dissection occurs. This should not have any
behavioural affect on other users of the flow dissector.
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The assignment of -EINVAL to variable ret is redundant as it
is being overwritten on the following error exit paths or
to the return value from the following call to basic_set_parms.
Fix this up by removing it. Cleans up clang warning message:
net/sched/cls_basic.c:185:2: warning: Value stored to 'err' is never read
Fixes: 1d8134fea2 ("net_sched: use idr to allocate basic filter handles")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of calling u32_lookup_ht() in a loop to find
a unused handle, just switch to idr API to allocate
new handles. u32 filters are special as the handle
could contain a hash table id and a key id, so we
need two IDR to allocate each of them.
Cc: Chris Mi <chrism@mellanox.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of calling basic_get() in a loop to find
a unused handle, just switch to idr API to allocate
new handles.
Cc: Chris Mi <chrism@mellanox.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of calling cls_bpf_get() in a loop to find
a unused handle, just switch to idr API to allocate
new handles.
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Chris Mi <chrism@mellanox.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Just do the rename into bpf_compute_data_pointers() as we'll add
one more pointer here to recompute.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While running TCP tests involving netem storing millions of packets,
I had the idea to speed up tfifo_reset() and did experiments.
I tried the rbtree_postorder_for_each_entry_safe() method that is
used in skb_rbtree_purge() but discovered it was slower than the
current tfifo_reset() method.
I measured time taken to release skbs with three occupation levels :
10^4, 10^5 and 10^6 skbs with three methods :
1) (current 'naive' method)
while ((p = rb_first(&q->t_root))) {
struct sk_buff *skb = netem_rb_to_skb(p);
rb_erase(p, &q->t_root);
rtnl_kfree_skbs(skb, skb);
}
2) Use rb_next() instead of rb_first() in the loop :
p = rb_first(&q->t_root);
while (p) {
struct sk_buff *skb = netem_rb_to_skb(p);
p = rb_next(p);
rb_erase(&skb->rbnode, &q->t_root);
rtnl_kfree_skbs(skb, skb);
}
3) "optimized" method using rbtree_postorder_for_each_entry_safe()
struct sk_buff *skb, *next;
rbtree_postorder_for_each_entry_safe(skb, next,
&q->t_root, rbnode) {
rtnl_kfree_skbs(skb, skb);
}
q->t_root = RB_ROOT;
Results :
method_1:while (rb_first()) rb_erase() 10000 skbs in 690378 ns (69 ns per skb)
method_2:rb_first; while (p) { p = rb_next(p); ...} 10000 skbs in 541846 ns (54 ns per skb)
method_3:rbtree_postorder_for_each_entry_safe() 10000 skbs in 868307 ns (86 ns per skb)
method_1:while (rb_first()) rb_erase() 99996 skbs in 7804021 ns (78 ns per skb)
method_2:rb_first; while (p) { p = rb_next(p); ...} 100000 skbs in 5942456 ns (59 ns per skb)
method_3:rbtree_postorder_for_each_entry_safe() 100000 skbs in 11584940 ns (115 ns per skb)
method_1:while (rb_first()) rb_erase() 1000000 skbs in 108577838 ns (108 ns per skb)
method_2:rb_first; while (p) { p = rb_next(p); ...} 1000000 skbs in 82619635 ns (82 ns per skb)
method_3:rbtree_postorder_for_each_entry_safe() 1000000 skbs in 127328743 ns (127 ns per skb)
Method 2) is simply faster, probably because it maintains a smaller
working size set.
Note that this is the method we use in tcp_ofo_queue() already.
I will also change skb_rbtree_purge() in a second patch.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes: c15ab236d6 ("net/sched: Change cls_flower to use IDR")
Cc: Chris Mi <chrism@mellanox.com>
Cc: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If real-time or fair-share curves are enabled in hfsc_change_class()
class isn't inserted into rb-trees yet. Thus init_ed() and init_vf()
must be called in place of update_ed() and update_vf().
Remove isn't required because for now curves cannot be disabled.
Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
SKB stored in qdisc->gso_skb also counted into backlog.
Some qdiscs don't reset backlog to zero in ->reset(),
for example sfq just dequeue and free all queued skb.
Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Fixes: 2ccccf5fb4 ("net_sched: update hierarchical backlog too")
Signed-off-by: David S. Miller <davem@davemloft.net>
gen estimator has been rewritten in commit 1c0d32fde5
("net_sched: gen_estimator: complete rewrite of rate estimators"),
the caller no longer needs to wait for a grace period. So this
patch gets rid of it.
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
skb->rbnode shares space with skb->next, skb->prev and skb->tstamp
Current uses (TCP receive ofo queue and netem) need to save/restore
tstamp, while skb->dev is either NULL (TCP) or a constant for a given
queue (netem).
Since we plan using an RB tree for TCP retransmit queue to speedup SACK
processing with large BDP, this patch exchanges skb->dev and
skb->tstamp.
This saves some overhead in both TCP and netem.
v2: removes the swtstamp field from struct tcp_skb_cb
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Soheil Hassas Yeganeh <soheil@google.com>
Cc: Wei Wang <weiwan@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
HTB qdisc overlimits counter is properly increased, but we have no per
class counter, meaning it is difficult to diagnose HTB problems.
This patch adds this counter, visible in "tc -s class show dev eth0",
with current iproute2.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Denys Fedoryshchenko <nuclearcat@nuclearcat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pointer tcm is being initialized and is never read, it is only being used
to determine the size of struct tcmsg. Clean this up by removing
variable tcm and explicitly using the sizeof struct tcmsg rather than *tcm.
Cleans up clang warning:
warning: Value stored to 'tcm' during its initialization is never read
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
this script, edited from Linux Advanced Routing and Traffic Control guide
tc q a dev en0 root handle 1: htb default a
tc c a dev en0 parent 1: classid 1:1 htb rate 6mbit burst 15k
tc c a dev en0 parent 1:1 classid 1:a htb rate 5mbit ceil 6mbit burst 15k
tc c a dev en0 parent 1:1 classid 1:b htb rate 1mbit ceil 6mbit burst 15k
tc f a dev en0 parent 1:0 prio 1 $clsname $clsargs classid 1:b
ping $address -c1
tc -s c s dev en0
classifies traffic to 1:b or 1:a, depending on whether the packet matches
or not the pattern $clsargs of filter $clsname. However, when $clsname is
'matchall', a systematic crash can be observed in htb_classify(). HTB and
classful qdiscs don't assign initial value to struct tcf_result, but then
they expect it to contain valid values after filters have been run. Thus,
current 'matchall' ignores the TCA_MATCHALL_CLASSID attribute, configured
by user, and makes HTB (and classful qdiscs) dereference random pointers.
By assigning head->res to *res in mall_classify(), before the actions are
invoked, we fix this crash and enable TCA_MATCHALL_CLASSID functionality,
that had no effect on 'matchall' classifier since its first introduction.
BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=1460213
Reported-by: Jiri Benc <jbenc@redhat.com>
Fixes: b87f7936a9 ("net/sched: introduce Match-all classifier")
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Acked-by: Yotam Gigi <yotamg@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Recent commit d7fb60b9ca ("net_sched: get rid of tcfa_rcu") removed
freeing in call_rcu, which changed already existing hard-to-hit
race condition into 100% hit:
[ 598.599825] BUG: unable to handle kernel NULL pointer dereference at 0000000000000030
[ 598.607782] IP: tcf_action_destroy+0xc0/0x140
Or:
[ 40.858924] BUG: unable to handle kernel NULL pointer dereference at 0000000000000030
[ 40.862840] IP: tcf_generic_walker+0x534/0x820
Fix this by storing the ops and use them directly for module_put call.
Fixes: a85a970af2 ("net_sched: move tc_action into tcf_common")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As pointed out by Jiri, there is still a race condition between
tcf_block_put() and tcf_chain_destroy() in a RCU callback. There
is no way to make it correct without proper locking or synchronization,
because both operate on a shared list.
Locking is hard, because the only lock we can pick here is a spinlock,
however, in tc_dump_tfilter() we iterate this list with a sleeping
function called (tcf_chain_dump()), which makes using a lock to protect
chain_list almost impossible.
Jiri suggested the idea of holding a refcnt before flushing, this works
because it guarantees us there would be no parallel tcf_chain_destroy()
during the loop, therefore the race condition is gone. But we have to
be very careful with proper synchronization with RCU callbacks.
Suggested-by: Jiri Pirko <jiri@mellanox.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes the following ugliness of tc filter chain refcnt:
a) tp proto should hold a refcnt to the chain too. This significantly
simplifies the logic.
b) Chain 0 is no longer special, it is created with refcnt=1 like any
other chains. All the ugliness in tcf_chain_put() can be gone!
c) No need to handle the flushing oddly, because block still holds
chain 0, it can not be released, this guarantees block is the last
user.
d) The race condition with RCU callbacks is easier to handle with just
a rcu_barrier(). Much easier to understand, nothing to hide. Thanks
to the previous patch. Please see also the comments in code.
e) Make the code understandable by humans, much less error-prone.
Fixes: 744a4cf63e ("net: sched: fix use after free when tcf_chain_destroy is called multiple times")
Fixes: 5bc1701881 ("net: sched: introduce multichain support for filters")
Cc: Jiri Pirko <jiri@mellanox.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
gen estimator has been rewritten in commit 1c0d32fde5
("net_sched: gen_estimator: complete rewrite of rate estimators"),
the caller is no longer needed to wait for a grace period.
So this patch gets rid of it.
This also completely closes a race condition between action free
path and filter chain add/remove path for the following patch.
Because otherwise the nested RCU callback can't be caught by
rcu_barrier().
Please see also the comments in code.
Cc: Jiri Pirko <jiri@mellanox.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes sparse warning about pointer in gen_handle:
net/sched/cls_rsvp.h:392:40: warning: Using plain integer as NULL pointer
Fixes: 8113c09567 ("net_sched: use void pointer for filter handle")
Signed-off-by: Josh Hunt <johunt@akamai.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There's a memleak happening for chain 0. The thing is, chain 0 needs to
be always present, not created on demand. Therefore tcf_block_get upon
creation of block calls the tcf_chain_create function directly. The
chain is created with refcnt == 1, which is not correct in this case and
causes the memleak. So move the refcnt increment into tcf_chain_get
function even for the case when chain needs to be created.
Reported-by: Jakub Kicinski <kubakici@wp.pl>
Fixes: 5bc1701881 ("net: sched: introduce multichain support for filters")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Tested-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The commit 520ac30f45 ("net_sched: drop packets after root qdisc lock
is released) made a big change of tc for performance. There are two points
left in sch_prio and sch_qfq which are not changed with that commit. Now
enhance them now with __qdisc_drop.
Signed-off-by: Gao Feng <gfree.wind@vip.163.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TC filters when used as classifiers are bound to TC classes.
However, there is a hidden difference when adding them in different
orders:
1. If we add tc classes before its filters, everything is fine.
Logically, the classes exist before we specify their ID's in
filters, it is easy to bind them together, just as in the current
code base.
2. If we add tc filters before the tc classes they bind, we have to
do dynamic lookup in fast path. What's worse, this happens all
the time not just once, because on fast path tcf_result is passed
on stack, there is no way to propagate back to the one in tc filters.
This hidden difference hurts performance silently if we have many tc
classes in hierarchy.
This patch intends to close this gap by doing the reverse binding when
we create a new class, in this case we can actually search all the
filters in its parent, match and fixup by classid. And because
tcf_result is specific to each type of tc filter, we have to introduce
a new ops for each filter to tell how to bind the class.
Note, we still can NOT totally get rid of those class lookup in
->enqueue() because cgroup and flow filters have no way to determine
the classid at setup time, they still have to go through dynamic lookup.
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently only a memory allocation failure can lead to this, so let's
initialize the timer first.
Fixes: 6529eaba33 ("net: sched: introduce tcf block infractructure")
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It is very unlikely to happen but the backlogs memory allocation
could fail and will free q->flows, but then ->destroy() will free
q->flows too. For correctness remove the first free and let ->destroy
clean up.
Fixes: 87b60cfacf ("net_sched: fix error recovery at qdisc creation")
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Depending on where ->init fails we can get a null pointer deref due to
uninitialized hires timer (watchdog) or a double free of the qdisc hash
because it is already freed by ->destroy().
Fixes: 8d55373875 ("net/sched/hfsc: allocate tcf block for hfsc root class")
Fixes: 87b60cfacf ("net_sched: fix error recovery at qdisc creation")
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
