This this the counterpart of 8aa7b526dc ("openvswitch: handle DNAT
tuple collision") for act_ct. From that commit changelog:
"""
With multiple DNAT rules it's possible that after destination
translation the resulting tuples collide.
...
Netfilter handles this case by allocating a null binding for SNAT at
egress by default. Perform the same operation in openvswitch for DNAT
if no explicit SNAT is requested by the user and allocate a null binding
for SNAT for packets in the "original" direction.
"""
Fixes: 95219afbb9 ("act_ct: support asymmetric conntrack")
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The commit ae81feb733 ("sch_htb: fix null pointer dereference
on a null new_q") fixes a NULL pointer dereference bug, but it
is not correct.
Because htb_graft_helper properly handles the case when new_q
is NULL, and after the previous patch by skipping this call
which creates an inconsistency : dev_queue->qdisc will still
point to the old qdisc, but cl->parent->leaf.q will point to
the new one (which will be noop_qdisc, because new_q was NULL).
The code is based on an assumption that these two pointers are
the same, so it can lead to refcount leaks.
The correct fix is to add a NULL pointer check to protect
qdisc_refcount_inc inside htb_parent_to_leaf_offload.
Fixes: ae81feb733 ("sch_htb: fix null pointer dereference on a null new_q")
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Suggested-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix current behavior of skipping template allocation in case the
ct action is in zone 0.
Skipping the allocation may cause the datapath ct code to ignore the
entire ct action with all its attributes (commit, nat) in case the ct
action in zone 0 was preceded by a ct clear action.
The ct clear action sets the ct_state to untracked and resets the
skb->_nfct pointer. Under these conditions and without an allocated
ct template, the skb->_nfct pointer will remain NULL which will
cause the tc ct action handler to exit without handling commit and nat
actions, if such exist.
For example, the following rule in OVS dp:
recirc_id(0x2),ct_state(+new-est-rel-rpl+trk),ct_label(0/0x1), \
in_port(eth0),actions:ct_clear,ct(commit,nat(src=10.11.0.12)), \
recirc(0x37a)
Will result in act_ct skipping the commit and nat actions in zone 0.
The change removes the skipping of template allocation for zone 0 and
treats it the same as any other zone.
Fixes: b57dc7c13e ("net/sched: Introduce action ct")
Signed-off-by: Ariel Levkovich <lariel@nvidia.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Link: https://lore.kernel.org/r/20210526170110.54864-1-lariel@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently established connections are not offloaded if the filter has a
"ct commit" action. This behavior will not offload connections of the
following scenario:
$ tc_filter add dev $DEV ingress protocol ip prio 1 flower \
ct_state -trk \
action ct commit action goto chain 1
$ tc_filter add dev $DEV ingress protocol ip chain 1 prio 1 flower \
action mirred egress redirect dev $DEV2
$ tc_filter add dev $DEV2 ingress protocol ip prio 1 flower \
action ct commit action goto chain 1
$ tc_filter add dev $DEV2 ingress protocol ip prio 1 chain 1 flower \
ct_state +trk+est \
action mirred egress redirect dev $DEV
Offload established connections, regardless of the commit flag.
Fixes: 46475bb20f ("net/sched: act_ct: Software offload of established flows")
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Paul Blakey <paulb@nvidia.com>
Link: https://lore.kernel.org/r/1622029449-27060-1-git-send-email-paulb@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
the following script:
# tc qdisc add dev eth0 handle 0x1 root fq_pie flows 2
# tc qdisc add dev eth0 clsact
# tc filter add dev eth0 egress matchall action skbedit priority 0x10002
# ping 192.0.2.2 -I eth0 -c2 -w1 -q
produces the following splat:
BUG: KASAN: slab-out-of-bounds in fq_pie_qdisc_enqueue+0x1314/0x19d0 [sch_fq_pie]
Read of size 4 at addr ffff888171306924 by task ping/942
CPU: 3 PID: 942 Comm: ping Not tainted 5.12.0+ #441
Hardware name: Red Hat KVM, BIOS 1.11.1-4.module+el8.1.0+4066+0f1aadab 04/01/2014
Call Trace:
dump_stack+0x92/0xc1
print_address_description.constprop.7+0x1a/0x150
kasan_report.cold.13+0x7f/0x111
fq_pie_qdisc_enqueue+0x1314/0x19d0 [sch_fq_pie]
__dev_queue_xmit+0x1034/0x2b10
ip_finish_output2+0xc62/0x2120
__ip_finish_output+0x553/0xea0
ip_output+0x1ca/0x4d0
ip_send_skb+0x37/0xa0
raw_sendmsg+0x1c4b/0x2d00
sock_sendmsg+0xdb/0x110
__sys_sendto+0x1d7/0x2b0
__x64_sys_sendto+0xdd/0x1b0
do_syscall_64+0x3c/0x80
entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x7fe69735c3eb
Code: 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 f3 0f 1e fa 48 8d 05 75 42 2c 00 41 89 ca 8b 00 85 c0 75 14 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 75 c3 0f 1f 40 00 41 57 4d 89 c7 41 56 41 89
RSP: 002b:00007fff06d7fb38 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
RAX: ffffffffffffffda RBX: 000055e961413700 RCX: 00007fe69735c3eb
RDX: 0000000000000040 RSI: 000055e961413700 RDI: 0000000000000003
RBP: 0000000000000040 R08: 000055e961410500 R09: 0000000000000010
R10: 0000000000000000 R11: 0000000000000246 R12: 00007fff06d81260
R13: 00007fff06d7fb40 R14: 00007fff06d7fc30 R15: 000055e96140f0a0
Allocated by task 917:
kasan_save_stack+0x19/0x40
__kasan_kmalloc+0x7f/0xa0
__kmalloc_node+0x139/0x280
fq_pie_init+0x555/0x8e8 [sch_fq_pie]
qdisc_create+0x407/0x11b0
tc_modify_qdisc+0x3c2/0x17e0
rtnetlink_rcv_msg+0x346/0x8e0
netlink_rcv_skb+0x120/0x380
netlink_unicast+0x439/0x630
netlink_sendmsg+0x719/0xbf0
sock_sendmsg+0xe2/0x110
____sys_sendmsg+0x5ba/0x890
___sys_sendmsg+0xe9/0x160
__sys_sendmsg+0xd3/0x170
do_syscall_64+0x3c/0x80
entry_SYSCALL_64_after_hwframe+0x44/0xae
The buggy address belongs to the object at ffff888171306800
which belongs to the cache kmalloc-256 of size 256
The buggy address is located 36 bytes to the right of
256-byte region [ffff888171306800, ffff888171306900)
The buggy address belongs to the page:
page:00000000bcfb624e refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x171306
head:00000000bcfb624e order:1 compound_mapcount:0
flags: 0x17ffffc0010200(slab|head|node=0|zone=2|lastcpupid=0x1fffff)
raw: 0017ffffc0010200 dead000000000100 dead000000000122 ffff888100042b40
raw: 0000000000000000 0000000000100010 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected
Memory state around the buggy address:
ffff888171306800: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ffff888171306880: 00 00 00 00 00 00 00 00 00 00 00 00 fc fc fc fc
>ffff888171306900: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
^
ffff888171306980: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff888171306a00: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
fix fq_pie traffic path to avoid selecting 'q->flows + q->flows_cnt' as a
valid flow: it's an address beyond the allocated memory.
Fixes: ec97ecf1eb ("net: sched: add Flow Queue PIE packet scheduler")
CC: stable@vger.kernel.org
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
the patch that fixed an endless loop in_fq_pie_init() was not considering
that 65535 is a valid class id. The correct bugfix for this infinite loop
is to change 'idx' to become an u32, like Colin proposed in the past [1].
Fix this as follows:
- restore 65536 as maximum possible values of 'flows_cnt'
- use u32 'idx' when iterating on 'q->flows'
- fix the TDC selftest
This reverts commit bb2f930d6d.
[1] https://lore.kernel.org/netdev/20210407163808.499027-1-colin.king@canonical.com/
CC: Colin Ian King <colin.king@canonical.com>
CC: stable@vger.kernel.org
Fixes: bb2f930d6d ("net/sched: fix infinite loop in sch_fq_pie")
Fixes: ec97ecf1eb ("net: sched: add Flow Queue PIE packet scheduler")
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The netdev qeueue might be stopped when byte queue limit has
reached or tx hw ring is full, net_tx_action() may still be
rescheduled if STATE_MISSED is set, which consumes unnecessary
cpu without dequeuing and transmiting any skb because the
netdev queue is stopped, see qdisc_run_end().
This patch fixes it by checking the netdev queue state before
calling qdisc_run() and clearing STATE_MISSED if netdev queue is
stopped during qdisc_run(), the net_tx_action() is rescheduled
again when netdev qeueue is restarted, see netif_tx_wake_queue().
As there is time window between netif_xmit_frozen_or_stopped()
checking and STATE_MISSED clearing, between which STATE_MISSED
may set by net_tx_action() scheduled by netif_tx_wake_queue(),
so set the STATE_MISSED again if netdev queue is restarted.
Fixes: 6b3ba9146f ("net: sched: allow qdiscs to handle locking")
Reported-by: Michal Kubecek <mkubecek@suse.cz>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently qdisc_run() checks the STATE_DEACTIVATED of lockless
qdisc before calling __qdisc_run(), which ultimately clear the
STATE_MISSED when all the skb is dequeued. If STATE_DEACTIVATED
is set before clearing STATE_MISSED, there may be rescheduling
of net_tx_action() at the end of qdisc_run_end(), see below:
CPU0(net_tx_atcion) CPU1(__dev_xmit_skb) CPU2(dev_deactivate)
. . .
. set STATE_MISSED .
. __netif_schedule() .
. . set STATE_DEACTIVATED
. . qdisc_reset()
. . .
.<--------------- . synchronize_net()
clear __QDISC_STATE_SCHED | . .
. | . .
. | . some_qdisc_is_busy()
. | . return *false*
. | . .
test STATE_DEACTIVATED | . .
__qdisc_run() *not* called | . .
. | . .
test STATE_MISS | . .
__netif_schedule()--------| . .
. . .
. . .
__qdisc_run() is not called by net_tx_atcion() in CPU0 because
CPU2 has set STATE_DEACTIVATED flag during dev_deactivate(), and
STATE_MISSED is only cleared in __qdisc_run(), __netif_schedule
is called at the end of qdisc_run_end(), causing tx action
rescheduling problem.
qdisc_run() called by net_tx_action() runs in the softirq context,
which should has the same semantic as the qdisc_run() called by
__dev_xmit_skb() protected by rcu_read_lock_bh(). And there is a
synchronize_net() between STATE_DEACTIVATED flag being set and
qdisc_reset()/some_qdisc_is_busy in dev_deactivate(), we can safely
bail out for the deactived lockless qdisc in net_tx_action(), and
qdisc_reset() will reset all skb not dequeued yet.
So add the rcu_read_lock() explicitly to protect the qdisc_run()
and do the STATE_DEACTIVATED checking in net_tx_action() before
calling qdisc_run_begin(). Another option is to do the checking in
the qdisc_run_end(), but it will add unnecessary overhead for
non-tx_action case, because __dev_queue_xmit() will not see qdisc
with STATE_DEACTIVATED after synchronize_net(), the qdisc with
STATE_DEACTIVATED can only be seen by net_tx_action() because of
__netif_schedule().
The STATE_DEACTIVATED checking in qdisc_run() is to avoid race
between net_tx_action() and qdisc_reset(), see:
commit d518d2ed86 ("net/sched: fix race between deactivation
and dequeue for NOLOCK qdisc"). As the bailout added above for
deactived lockless qdisc in net_tx_action() provides better
protection for the race without calling qdisc_run() at all, so
remove the STATE_DEACTIVATED checking in qdisc_run().
After qdisc_reset(), there is no skb in qdisc to be dequeued, so
clear the STATE_MISSED in dev_reset_queue() too.
Fixes: 6b3ba9146f ("net: sched: allow qdiscs to handle locking")
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
V8: Clearing STATE_MISSED before calling __netif_schedule() has
avoid the endless rescheduling problem, but there may still
be a unnecessary rescheduling, so adjust the commit log.
Signed-off-by: David S. Miller <davem@davemloft.net>
Lockless qdisc has below concurrent problem:
cpu0 cpu1
. .
q->enqueue .
. .
qdisc_run_begin() .
. .
dequeue_skb() .
. .
sch_direct_xmit() .
. .
. q->enqueue
. qdisc_run_begin()
. return and do nothing
. .
qdisc_run_end() .
cpu1 enqueue a skb without calling __qdisc_run() because cpu0
has not released the lock yet and spin_trylock() return false
for cpu1 in qdisc_run_begin(), and cpu0 do not see the skb
enqueued by cpu1 when calling dequeue_skb() because cpu1 may
enqueue the skb after cpu0 calling dequeue_skb() and before
cpu0 calling qdisc_run_end().
Lockless qdisc has below another concurrent problem when
tx_action is involved:
cpu0(serving tx_action) cpu1 cpu2
. . .
. q->enqueue .
. qdisc_run_begin() .
. dequeue_skb() .
. . q->enqueue
. . .
. sch_direct_xmit() .
. . qdisc_run_begin()
. . return and do nothing
. . .
clear __QDISC_STATE_SCHED . .
qdisc_run_begin() . .
return and do nothing . .
. . .
. qdisc_run_end() .
This patch fixes the above data race by:
1. If the first spin_trylock() return false and STATE_MISSED is
not set, set STATE_MISSED and retry another spin_trylock() in
case other CPU may not see STATE_MISSED after it releases the
lock.
2. reschedule if STATE_MISSED is set after the lock is released
at the end of qdisc_run_end().
For tx_action case, STATE_MISSED is also set when cpu1 is at the
end if qdisc_run_end(), so tx_action will be rescheduled again
to dequeue the skb enqueued by cpu2.
Clear STATE_MISSED before retrying a dequeuing when dequeuing
returns NULL in order to reduce the overhead of the second
spin_trylock() and __netif_schedule() calling.
Also clear the STATE_MISSED before calling __netif_schedule()
at the end of qdisc_run_end() to avoid doing another round of
dequeuing in the pfifo_fast_dequeue().
The performance impact of this patch, tested using pktgen and
dummy netdev with pfifo_fast qdisc attached:
threads without+this_patch with+this_patch delta
1 2.61Mpps 2.60Mpps -0.3%
2 3.97Mpps 3.82Mpps -3.7%
4 5.62Mpps 5.59Mpps -0.5%
8 2.78Mpps 2.77Mpps -0.3%
16 2.22Mpps 2.22Mpps -0.0%
Fixes: 6b3ba9146f ("net: sched: allow qdiscs to handle locking")
Acked-by: Jakub Kicinski <kuba@kernel.org>
Tested-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The assignment is not being used and redundant.
The check for null is redundant as nf_conntrack_put() also
checks this.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Paul Blakey <paulb@nvidia.com>
Link: https://lore.kernel.org/r/20210428060532.3330974-1-roid@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There is a reproducible sequence from the userland that will trigger a WARN_ON()
condition in taprio_get_start_time, which causes kernel to panic if configured
as "panic_on_warn". Catch this condition in parse_taprio_schedule to
prevent this condition.
Reported as bug on syzkaller:
https://syzkaller.appspot.com/bug?extid=d50710fd0873a9c6b40c
Reported-by: syzbot+d50710fd0873a9c6b40c@syzkaller.appspotmail.com
Signed-off-by: Du Cheng <ducheng2@gmail.com>
Acked-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
MAINTAINERS
- keep Chandrasekar
drivers/net/ethernet/mellanox/mlx5/core/en_main.c
- simple fix + trust the code re-added to param.c in -next is fine
include/linux/bpf.h
- trivial
include/linux/ethtool.h
- trivial, fix kdoc while at it
include/linux/skmsg.h
- move to relevant place in tcp.c, comment re-wrapped
net/core/skmsg.c
- add the sk = sk // sk = NULL around calls
net/tipc/crypto.c
- trivial
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
With recent changes that separated action module load from action
initialization tcf_action_init() function error handling code was modified
to manually release the loaded modules if loading/initialization of any
further action in same batch failed. For the case when all modules
successfully loaded and some of the actions were initialized before one of
them failed in init handler. In this case for all previous actions the
module will be released twice by the error handler: First time by the loop
that manually calls module_put() for all ops, and second time by the action
destroy code that puts the module after destroying the action.
Reproduction:
$ sudo tc actions add action simple sdata \"2\" index 2
$ sudo tc actions add action simple sdata \"1\" index 1 \
action simple sdata \"2\" index 2
RTNETLINK answers: File exists
We have an error talking to the kernel
$ sudo tc actions ls action simple
total acts 1
action order 0: Simple <"2">
index 2 ref 1 bind 0
$ sudo tc actions flush action simple
$ sudo tc actions ls action simple
$ sudo tc actions add action simple sdata \"2\" index 2
Error: Failed to load TC action module.
We have an error talking to the kernel
$ lsmod | grep simple
act_simple 20480 -1
Fix the issue by modifying module reference counting handling in action
initialization code:
- Get module reference in tcf_idr_create() and put it in tcf_idr_release()
instead of taking over the reference held by the caller.
- Modify users of tcf_action_init_1() to always release the module
reference which they obtain before calling init function instead of
assuming that created action takes over the reference.
- Finally, modify tcf_action_init_1() to not release the module reference
when overwriting existing action as this is no longer necessary since both
upper and lower layers obtain and manage their own module references
independently.
Fixes: d349f99768 ("net_sched: fix RTNL deadlock again caused by request_module()")
Suggested-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Action init code increments reference counter when it changes an action.
This is the desired behavior for cls API which needs to obtain action
reference for every classifier that points to action. However, act API just
needs to change the action and releases the reference before returning.
This sequence breaks when the requested action doesn't exist, which causes
act API init code to create new action with specified index, but action is
still released before returning and is deleted (unless it was referenced
concurrently by cls API).
Reproduction:
$ sudo tc actions ls action gact
$ sudo tc actions change action gact drop index 1
$ sudo tc actions ls action gact
Extend tcf_action_init() to accept 'init_res' array and initialize it with
action->ops->init() result. In tcf_action_add() remove pointers to created
actions from actions array before passing it to tcf_action_put_many().
Fixes: cae422f379 ("net: sched: use reference counting action init")
Reported-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 6855e8213e.
Following commit in series fixes the issue without introducing regression
in error rollback of tcf_action_destroy().
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The 'unlocked_driver_cb' struct field in 'bo' is not being initialized
in tcf_block_offload_init(). The uninitialized 'unlocked_driver_cb'
will be used when calling unlocked_driver_cb(). So initialize 'bo' to
zero to avoid the issue.
Addresses-Coverity: ("Uninitialized scalar variable")
Fixes: 0fdcf78d59 ("net: use flow_indr_dev_setup_offload()")
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sch_htb: fix null pointer dereference on a null new_q
Currently if new_q is null, the null new_q pointer will be
dereference when 'q->offload' is true. Fix this by adding
a braces around htb_parent_to_leaf_offload() to avoid it.
Addresses-Coverity: ("Dereference after null check")
Fixes: d03b195b5a ("sch_htb: Hierarchical QoS hardware offload")
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, action creation using ACT API in replace mode is buggy.
When invoking for non-existent action index 42,
tc action replace action bpf obj foo.o sec <xyz> index 42
kernel creates the action, fills up the netlink response, and then just
deletes the action after notifying userspace.
tc action show action bpf
doesn't list the action.
This happens due to the following sequence when ovr = 1 (replace mode)
is enabled:
tcf_idr_check_alloc is used to atomically check and either obtain
reference for existing action at index, or reserve the index slot using
a dummy entry (ERR_PTR(-EBUSY)).
This is necessary as pointers to these actions will be held after
dropping the idrinfo lock, so bumping the reference count is necessary
as we need to insert the actions, and notify userspace by dumping their
attributes. Finally, we drop the reference we took using the
tcf_action_put_many call in tcf_action_add. However, for the case where
a new action is created due to free index, its refcount remains one.
This when paired with the put_many call leads to the kernel setting up
the action, notifying userspace of its creation, and then tearing it
down. For existing actions, the refcount is still held so they remain
unaffected.
Fortunately due to rtnl_lock serialization requirement, such an action
with refcount == 1 will not be concurrently deleted by anything else, at
best CLS API can move its refcount up and down by binding to it after it
has been published from tcf_idr_insert_many. Since refcount is atleast
one until put_many call, CLS API cannot delete it. Also __tcf_action_put
release path already ensures deterministic outcome (either new action
will be created or existing action will be reused in case CLS API tries
to bind to action concurrently) due to idr lock serialization.
We fix this by making refcount of newly created actions as 2 in ACT API
replace mode. A relaxed store will suffice as visibility is ensured only
after the tcf_idr_insert_many call.
Note that in case of creation or overwriting using CLS API only (i.e.
bind = 1), overwriting existing action object is not allowed, and any
such request is silently ignored (without error).
The refcount bump that occurs in tcf_idr_check_alloc call there for
existing action will pair with tcf_exts_destroy call made from the
owner module for the same action. In case of action creation, there
is no existing action, so no tcf_exts_destroy callback happens.
This means no code changes for CLS API.
Fixes: cae422f379 ("net: sched: use reference counting action init")
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
s/procdure/procedure/
s/maintanance/maintenance/
Signed-off-by: Bhaskar Chowdhury <unixbhaskar@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Invalid detection works with two distinct moments: act_ct tries to find
a conntrack entry and set post_ct true, indicating that that was
attempted. Then, when flow dissector tries to dissect CT info and no
entry is there, it knows that it was tried and no entry was found, and
synthesizes/sets
key->ct_state = TCA_FLOWER_KEY_CT_FLAGS_TRACKED |
TCA_FLOWER_KEY_CT_FLAGS_INVALID;
mimicing what OVS does.
OVS has this a bit more streamlined, as it recomputes the key after
trying to find a conntrack entry for it.
Issue here is, when we have 'tc action ct clear', it didn't clear
post_ct, causing a subsequent match on 'ct_state -trk' to fail, due to
the above. The fix, thus, is to clear it.
Reproducer rules:
tc filter add dev enp130s0f0np0_0 ingress prio 1 chain 0 \
protocol ip flower ip_proto tcp ct_state -trk \
action ct zone 1 pipe \
action goto chain 2
tc filter add dev enp130s0f0np0_0 ingress prio 1 chain 2 \
protocol ip flower \
action ct clear pipe \
action goto chain 4
tc filter add dev enp130s0f0np0_0 ingress prio 1 chain 4 \
protocol ip flower ct_state -trk \
action mirred egress redirect dev enp130s0f1np1_0
With the fix, the 3rd rule matches, like it does with OVS kernel
datapath.
Fixes: 7baf2429a1 ("net/sched: cls_flower add CT_FLAGS_INVALID flag support")
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: wenxu <wenxu@ucloud.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
The existing code is functionally correct: iproute2 parses the ip_flags
argument for tc-flower and really packs it as big endian into the
TCA_FLOWER_KEY_FLAGS netlink attribute. But there is a problem in the
fact that W=1 builds complain:
net/sched/cls_flower.c:1047:15: warning: cast to restricted __be32
This is because we should use the dedicated helper for obtaining a
__be32 pointer to the netlink attribute, not a u32 one. This ensures
type correctness for be32_to_cpu.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A make W=1 build complains that:
net/sched/cls_flower.c:214:20: warning: cast from restricted __be16
net/sched/cls_flower.c:214:20: warning: incorrect type in argument 1 (different base types)
net/sched/cls_flower.c:214:20: expected unsigned short [usertype] val
net/sched/cls_flower.c:214:20: got restricted __be16 [usertype] dst
This is because we use htons on struct flow_dissector_key_ports members
src and dst, which are defined as __be16, so they are already in network
byte order, not host. The byte swap function for the other direction
should have been used.
Because htons and ntohs do the same thing (either both swap, or none
does), this change has no functional effect except to silence the
warnings.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When using short intervals e.g. below one millisecond, large packets won't be
transmitted at all. The software implementations checks whether the packet can
be fit into the remaining interval. Therefore, it takes the packet length and
the transmission speed into account. That is correct.
However, for large packets it may be that the transmission time exceeds the
interval resulting in no packet transmission. The same situation works fine with
hardware offloading applied.
The problem has been observed with the following schedule and iperf3:
|tc qdisc replace dev lan1 parent root handle 100 taprio \
| num_tc 8 \
| map 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 \
| queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \
| base-time $base \
| sched-entry S 0x40 500000 \
| sched-entry S 0xbf 500000 \
| clockid CLOCK_TAI \
| flags 0x00
[...]
|root@tsn:~# iperf3 -c 192.168.2.105
|Connecting to host 192.168.2.105, port 5201
|[ 5] local 192.168.2.121 port 52610 connected to 192.168.2.105 port 5201
|[ ID] Interval Transfer Bitrate Retr Cwnd
|[ 5] 0.00-1.00 sec 45.2 KBytes 370 Kbits/sec 0 1.41 KBytes
|[ 5] 1.00-2.00 sec 0.00 Bytes 0.00 bits/sec 0 1.41 KBytes
After debugging, it seems that the packet length stored in the SKB is about
7000-8000 bytes. Using a 100 Mbit/s link the transmission time is about 600us
which larger than the interval of 500us.
Therefore, segment the SKB into smaller chunks if the packet is too big. This
yields similar results than the hardware offload:
|root@tsn:~# iperf3 -c 192.168.2.105
|Connecting to host 192.168.2.105, port 5201
|- - - - - - - - - - - - - - - - - - - - - - - - -
|[ ID] Interval Transfer Bitrate Retr
|[ 5] 0.00-10.00 sec 48.9 MBytes 41.0 Mbits/sec 0 sender
|[ 5] 0.00-10.02 sec 48.7 MBytes 40.7 Mbits/sec receiver
Furthermore, the segmentation can be skipped for the full offload case, as the
driver or the hardware is expected to handle this.
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
The ct_state validate should not only check the mask bit and also
check mask_bit & key_bit..
For the +new+est case example, The 'new' and 'est' bits should be
set in both state_mask and state flags. Or the -new-est case also
will be reject by kernel.
When Openvswitch with two flows
ct_state=+trk+new,action=commit,forward
ct_state=+trk+est,action=forward
A packet go through the kernel and the contrack state is invalid,
The ct_state will be +trk-inv. Upcall to the ovs-vswitchd, the
finally dp action will be drop with -new-est+trk.
Fixes: 1bcc51ac07 ("net/sched: cls_flower: Reject invalid ct_state flags rules")
Fixes: 3aed8b6333 ("net/sched: cls_flower: validate ct_state for invalid and reply flags")
Signed-off-by: wenxu <wenxu@ucloud.cn>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When openvswitch conntrack offload with act_ct action. The first rule
do conntrack in the act_ct in tc subsystem. And miss the next rule in
the tc and fallback to the ovs datapath but miss set post_ct flag
which will lead the ct_state_key with -trk flag.
Fixes: 7baf2429a1 ("net/sched: cls_flower add CT_FLAGS_INVALID flag support")
Signed-off-by: wenxu <wenxu@ucloud.cn>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, callers of psample_sample_packet() pass three metadata
attributes: Ingress port, egress port and truncated size. Subsequent
patches are going to add more attributes (e.g., egress queue occupancy),
which also need an indication whether they are valid or not.
Encapsulate packet metadata in a struct in order to keep the number of
arguments reasonable.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow a policer action to enforce a rate-limit based on packets-per-second,
configurable using a packet-per-second rate and burst parameters.
e.g.
tc filter add dev tap1 parent ffff: u32 match \
u32 0 0 police pkts_rate 3000 pkts_burst 1000
Testing was unable to uncover a performance impact of this change on
existing features.
Signed-off-by: Baowen Zheng <baowen.zheng@corigine.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Louis Peens <louis.peens@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow flow_offload API to configure packet-per-second policing using rate
and burst parameters.
Dummy implementations of tcf_police_rate_pkt_ps() and
tcf_police_burst_pkt() are supplied which return 0, the unconfigured state.
This is to facilitate splitting the offload, driver, and TC code portion of
this feature into separate patches with the aim of providing a logical flow
for review. And the implementation of these helpers will be filled out by a
follow-up patch.
Signed-off-by: Xingfeng Hu <xingfeng.hu@corigine.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Louis Peens <louis.peens@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
htb_init may fail to do the offload if it's not supported or if a
runtime error happens when allocating direct qdiscs. In those cases
TC_HTB_CREATE command is not sent to the driver, however, htb_destroy
gets called anyway and attempts to send TC_HTB_DESTROY.
It shouldn't happen, because the driver didn't receive TC_HTB_CREATE,
and also because the driver may not support ndo_setup_tc at all, while
q->offload is true, and htb_destroy mistakenly thinks the offload is
supported. Trying to call ndo_setup_tc in the latter case will lead to a
NULL pointer dereference.
This commit fixes the issues with htb_destroy by deferring assignment of
q->offload until after the TC_HTB_CREATE command. The necessary cleanup
of the offload entities is already done in htb_init.
Reported-by: syzbot+b53a709f04722ca12a3c@syzkaller.appspotmail.com
Fixes: d03b195b5a ("sch_htb: Hierarchical QoS hardware offload")
Suggested-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
htb_select_queue assumes it's always the offload mode, and it ends up in
calling ndo_setup_tc without any checks. It may lead to a NULL pointer
dereference if ndo_setup_tc is not implemented, or to an error returned
from the driver, which will prevent attaching qdiscs to HTB classes in
the non-offload mode.
This commit fixes the bug by adding the missing check to
htb_select_queue. In the non-offload mode it will return sch->dev_queue,
mimicking tc_modify_qdisc's behavior for the case where select_queue is
not implemented.
Reported-by: syzbot+b53a709f04722ca12a3c@syzkaller.appspotmail.com
Fixes: d03b195b5a ("sch_htb: Hierarchical QoS hardware offload")
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Implement this callback in order to get the offloaded stats added to the
kernel stats.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is a follow up of commit ea32746953 ("net: sched: avoid
duplicates in qdisc dump") which has fixed the issue only for the qdisc
dump.
The duplicate printing also occurs when dumping the classes via
tc class show dev eth0
Fixes: 59cc1f61f0 ("net: sched: convert qdisc linked list to hashtable")
Signed-off-by: Maximilian Heyne <mheyne@amazon.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add invalid and reply flags validate in the fl_validate_ct_state.
This makes the checking complete if compared to ovs'
validate_ct_state().
Signed-off-by: wenxu <wenxu@ucloud.cn>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Link: https://lore.kernel.org/r/1614064315-364-1-git-send-email-wenxu@ucloud.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reject the unsupported and invalid ct_state flags of cls flower rules.
Fixes: e0ace68af2 ("net/sched: cls_flower: Add matching on conntrack info")
Signed-off-by: wenxu <wenxu@ucloud.cn>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Give offloading drivers the direction of the offloaded ct flow,
this will be used for matches on direction (ct_state +/-rpl).
Signed-off-by: Paul Blakey <paulb@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This commit adds support for statistics of offloaded HTB. Bytes and
packets counters for leaf and inner nodes are supported, the values are
taken from per-queue qdiscs, and the numbers that the user sees should
have the same behavior as the software (non-offloaded) HTB.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
HTB doesn't scale well because of contention on a single lock, and it
also consumes CPU. This patch adds support for offloading HTB to
hardware that supports hierarchical rate limiting.
In the offload mode, HTB passes control commands to the driver using
ndo_setup_tc. The driver has to replicate the whole hierarchy of classes
and their settings (rate, ceil) in the NIC. Every modification of the
HTB tree caused by the admin results in ndo_setup_tc being called.
After this setup, the HTB algorithm is done completely in the NIC. An SQ
(send queue) is created for every leaf class and attached to the
hierarchy, so that the NIC can calculate and obey aggregated rate
limits, too. In the future, it can be changed, so that multiple SQs will
back a single leaf class.
ndo_select_queue is responsible for selecting the right queue that
serves the traffic class of each packet.
The data path works as follows: a packet is classified by clsact, the
driver selects a hardware queue according to its class, and the packet
is enqueued into this queue's qdisc.
This solution addresses two main problems of scaling HTB:
1. Contention by flow classification. Currently the filters are attached
to the HTB instance as follows:
# tc filter add dev eth0 parent 1:0 protocol ip flower dst_port 80
classid 1:10
It's possible to move classification to clsact egress hook, which is
thread-safe and lock-free:
# tc filter add dev eth0 egress protocol ip flower dst_port 80
action skbedit priority 1:10
This way classification still happens in software, but the lock
contention is eliminated, and it happens before selecting the TX queue,
allowing the driver to translate the class to the corresponding hardware
queue in ndo_select_queue.
Note that this is already compatible with non-offloaded HTB and doesn't
require changes to the kernel nor iproute2.
2. Contention by handling packets. HTB is not multi-queue, it attaches
to a whole net device, and handling of all packets takes the same lock.
When HTB is offloaded, it registers itself as a multi-queue qdisc,
similarly to mq: HTB is attached to the netdev, and each queue has its
own qdisc.
Some features of HTB may be not supported by some particular hardware,
for example, the maximum number of classes may be limited, the
granularity of rate and ceil parameters may be different, etc. - so, the
offload is not enabled by default, a new parameter is used to enable it:
# tc qdisc replace dev eth0 root handle 1: htb offload
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In a following commit, sch_htb will start using extack in the delete
class operation to pass hardware errors in offload mode. This commit
prepares for that by adding the extack parameter to this callback and
converting usage of the existing qdiscs.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
