Currently, the flow information that is matched for tunnels and
the tunnel data passed around with packets is the same. However,
as additional information is added this is not necessarily desirable,
as in the case of pointers.
This adds a new structure for tunnel metadata which currently contains
only the existing struct. This change is purely internal to the kernel
since the current OVS_KEY_ATTR_IPV4_TUNNEL is simply a compressed version
of OVS_KEY_ATTR_TUNNEL that is translated at flow setup.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Pravin B Shelar <pshelar@nicira.com>
In queue_userspace_packet(), the ovs_nla_put_flow return value is
not checked. This is fine as long as key_attr_size() returns the
correct value. In case it does not, the current code may corrupt buffer
memory. Add a run time assertion catch this case to avoid silent
failure.
Reported-by: Ben Pfaff <blp@nicira.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>
The key_attr_size() was not updated when RECIRC_ID and DP_HASH
key fields are added to support recircualtion. This patch fixes it.
Reported-by: Ben Pfaff <blp@nicira.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>
VMware-BZ: 1266214
Current datapath limits the number of times same packet can loop
through action execution to avoid blowing out the kernel stack.
Recirculation also adds to action execution count, but does not use
the same amount of stack compare to other services, such as IPsec.
This patch introduces the concept of stack cost. Recirculation has a
stack cost of 1 while other services have stack cost of 4. Datapath
packet process can accommodate packets that need both services and
recirculation as long as the total stack cost does not exceed the max
stack cost. Packets exceed the limit will be treated as looped packets
and dropped.
The max stack cost is set to allow up to 4 regular services, plus up
to 3 recirculation. The behavior of packets do not recirculate does
not change.
Signed-off-by: Andy Zhou <azhou@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
On every packet OVS needs to lookup flow-table with every mask
until it finds a match. The packet flow-key is first masked
with mask in the list and then the masked key is looked up in
flow-table. Therefore number of masks can affect packet
processing performance.
Following patch adds mask index to mask cache from last
pakcet lookup in same flow. Index of mask is stored in
this cache. This cache is searched by 5 tuple hash (skb rxhash).
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Thomas Graf <tgraf@redhat.com>
Ths simplifies flow-table-destroy API. This change is required
for following patches.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Thomas Graf <tgraf@redhat.com>
This patch removes the new_vport() forward declaration in datapath.c
as it is not needed.
Signed-off-by: Rami Rosen <ramirose@gmail.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
In order to allow handlers directly read upcalls from datapath,
we need to support per-handler netlink socket for each vport in
datapath. This commit makes this happen. Also, it is guaranteed
to be backward compatible with previous branch.
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Thomas Graf <tgraf@redhat.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>
This is the ported version of commit 36d5fe6a with the same name from net-next.
Apart from the small datapath.c changes it adjust the compat layer files as
well. This is the original commit message:
"skb_zerocopy can copy elements of the frags array between skbs, but it doesn't
orphan them. Also, it doesn't handle errors, so this patch takes care of that
as well, and modify the callers accordingly. skb_tx_error() is also added to
the callers so they will signal the failed delivery towards the creator of the
skb."
Signed-off-by: Zoltan Kiss <zoltan.kiss@citrix.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Flow statistics need to take into account the TCP flags from the packet
currently being processed (in 'key'), not the TCP flags matched by the
flow found in the kernel flow table (in 'flow').
This bug made the Open vSwitch userspace fin_timeout action have no effect
in many cases.
Bug #1219516.
Reported-by: Len Gao <leng@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Add support for building the in-tree kernel datapath for
Linux kernels up to 3.13. There were some changes in the
netlink area which required adding new compatibility code
for this layer. Also, some new per-cpu stats initialization
code was added.
Based on patch from Kyle Mestery.
Signed-off-by: Kyle Mestery <mestery@noironetworks.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Kyle Mestery <mestery@noironetworks.com>
Following patch will be easier to reason about with separate
ovs_flow_cmd_new() and ovs_flow_cmd_set() functions.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
ovs_flow_cmd_del() now allocates reply (if needed) after the flow has
already been removed from the flow table. If the reply allocation
fails, a netlink error is signaled with netlink_set_err(), as is
already done in ovs_flow_cmd_new_or_set() in the similar situation.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Reduce and clarify locking requirements for ovs_flow_cmd_alloc_info(),
ovs_flow_cmd_fill_info() and ovs_flow_cmd_build_info().
A datapath pointer is available only when holding a lock. Change
ovs_flow_cmd_fill_info() and ovs_flow_cmd_build_info() to take a
dp_ifindex directly, rather than a datapath pointer that is then
(only) used to get the dp_ifindex. This is useful, since the
dp_ifindex is available even when the datapath pointer is not, both
before and after taking a lock, which makes further critical section
reduction possible.
Make ovs_flow_cmd_alloc_info() take an 'acts' argument instead a
'flow' pointer. This allows some future patches to do the allocation
before acquiring the flow pointer.
The locking requirements after this patch are:
ovs_flow_cmd_alloc_info(): May be called without locking, must not be
called while holding the RCU read lock (due to memory allocation).
If 'acts' belong to a flow in the flow table, however, then the
caller must hold ovs_mutex.
ovs_flow_cmd_fill_info(): Either ovs_mutex or RCU read lock must be held.
ovs_flow_cmd_build_info(): This calls both of the above, so the caller
must hold ovs_mutex.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Move most memory allocations away from the ovs_mutex critical
sections. vport allocations still happen while the lock is taken, as
changing that would require major refactoring. Also, vports are
created very rarely so it should not matter.
Change ovs_dp_cmd_get() now only takes the rcu_read_lock(), rather
than ovs_lock(), as nothing need to be changed. This was done by
ovs_vport_cmd_get() already.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Use netlink_has_listeners() and NLM_F_ECHO flag to determine if a
reply is needed or not for OVS_FLOW_CMD_NEW, OVS_FLOW_CMD_SET, or
OVS_FLOW_CMD_DEL. Currently, OVS userspace does not request a reply
for OVS_FLOW_CMD_NEW, but usually does for OVS_FLOW_CMD_DEL, as stats
may have changed.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Remove unnecessary locking from functions that are always called with
appropriate locking.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Thomas Graf <tgraf@redhat.com>
Flow SET can accept an empty set of actions, with the intended
semantics of leaving existing actions unmodified. This seems to have
been brokin after OVS 1.7, as we have assigned the flow's actions
pointer to NULL in this case, but we never check for the NULL pointer
later on. This patch restores the intended behavior and documents it
in the include/linux/openvswitch.h.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Minimize padding in sw_flow_key and move 'tp' top the main struct.
These changes simplify code when accessing the transport port numbers
and the tcp flags, and makes the sw_flow_key 8 bytes smaller on 64-bit
systems (128->120 bytes). These changes also make the keys for IPv4
packets to fit in one cache line.
There is a valid concern for safety of packing the struct
ovs_key_ipv4_tunnel, as it would be possible to take the address of
the tun_id member as a __be64 * which could result in unaligned access
in some systems. However:
- sw_flow_key itself is 64-bit aligned, so the tun_id within is always
64-bit aligned.
- We never make arrays of ovs_key_ipv4_tunnel (which would force every
second tun_key to be misaligned).
- We never take the address of the tun_id in to a __be64 *.
- Whereever we use struct ovs_key_ipv4_tunnel outside the sw_flow_key,
it is in stack (on tunnel input functions), where compiler has full
control of the alignment.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
The 5-tuple optimization becomes unnecessary with a later per-NUMA
node stats patch. Remove it first to make the changes easier to
grasp.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
This fixes crash when userspace does "ovs-dpctl add-dp dev" where dev is
existing non-dp netdevice.
Introduced by:
commit 94358dcffb
"openvswitch: Drop user features if old user space attempted to create datapath"
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Jesse Gross <jesse@nicira.com>
ovs_vport_cmd_dump() did rcu_read_lock() only after getting the
datapath, which could have been deleted in between. Resolved by
taking rcu_read_lock() before the get_dp() call.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
With subfacets, we'd expect megaflow updates message to carry
the original micro flow. If not, EINVAL is returned and kernel
logs an error message. Now that the user space subfacet layer is
removed, it is expected that flow updates can arrive with a
micro flow other than the original. Change the return code to
EEXIST and remove the kernel error log message.
Reported-by: Ben Pfaff <blp@nicira.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>
commit c58cc9a460 (datapath: Allow user space to
announce ability to accept unaligned Netlink messages) introduced
OVS_DP_ATTR_USER_FEATURES netlink attribute in datapath responses,
but the attribute size was not taken into account in ovs_dp_cmd_msg_size().
Signed-off-by: Daniele Di Proietto <daniele.di.proietto@gmail.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Both mega flow mask's reference counter and per flow table mask list
should only be accessed when holding ovs_mutex() lock. However
this is not true with ovs_flow_table_flush(). The patch fixes this bug.
Reported-by: Joe Stringer <joestringer@nicira.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>
While the zerocopy method is correctly omitted if user space
does not support unaligned Netlink messages. The attribute is
still not padded correctly as skb_zerocopy() will not ensure
padding and the attribute size is no longer pre calculated
though nla_reserve() which ensured padding previously.
This patch applies appropriate padding if a linear data copy
was performed in skb_zerocopy().
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Zoltan Kiss <zoltan.kiss@citrix.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
The copy & csum optimization is no longer present with zerocopy
enabled. Compute the checksum in skb_gso_segment() directly by
dropping the HW CSUM capability from the features passed in.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Use of skb_zerocopy() can avoid the expensive call to memcpy()
when copying the packet data into the Netlink skb. Completes
checksum through skb_checksum_help() if not already done in
GSO segmentation.
Zerocopy is only performed if user space supported unaligned
Netlink messages. memory mapped netlink i/o is preferred over
zerocopy if it is set up.
Cost of upcall is significantly reduced from:
+ 7.48% vhost-8471 [k] memcpy
+ 5.57% ovs-vswitchd [k] memcpy
+ 2.81% vhost-8471 [k] csum_partial_copy_generic
to:
+ 5.72% ovs-vswitchd [k] memcpy
+ 3.32% vhost-5153 [k] memcpy
+ 0.68% vhost-5153 [k] skb_zerocopy
(megaflows disabled)
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Allows removing the net and dp_ifindex argument and simplify the
code.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Drop user features if an outdated user space instance that does not
understand the concept of user_features attempted to create a new
datapath.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Reviewed-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Use percpu allocator for stats due to objection to stats array.
But percpu allocator is not designed for high churn allocation/
deallcation. so we need to avoid allocating percpu flow for
short lived flows. One cheaper way to detect flow is by checking
if 5-tuple used in RSS are masked or not. if any one of them is
masked, flow is likely shared across CPU where percpu stat
should be more scalable. And that flow should be relatively
long lived flow.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Flow lookup can happen either in packet processing context or userspace
context but it was annotated as requiring RCU read lock to be held. This
also allows OVS mutex to be held without causing warnings.
Reported-by: Justin Pettit <jpettit@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Reviewed-by: Thomas Graf <tgraf@redhat.com>
API changes only for code readability. No functional chnages.
This patch removes the underscored version. Added a new API ovs_flow_tbl_lookup_stats() that
returns the n_mask_hits.
Reported by: Ben Pfaff <blp@nicira.com>
Reviewed-by: Thomas Graf <tgraf@redhat.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>
Widen TCP flags handling from 7 bits (uint8_t) to 12 bits (uint16_t).
The kernel interface remains at 8 bits, which makes no functional
difference now, as none of the higher bits is currently of interest
to the userspace.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
With mega flow implementation ovs flow can be shared between
multiple CPUs which makes stats updates highly contended
operation. Following patch allocates separate stats for each
CPU to make stats update scalable.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Collect mega flow mask stats. ovs-dpctl show command can be used to
display them.
Signed-off-by: Andy Zhou <azhou@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Hides mega-flow implementation in flow_table.c rather than
datapath.c.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
ovs-flow rehash does not touch mega flow list. Following patch
moves it dp struct datapath. Avoid one extra indirection for
accessing mega-flow list head on every packet receive.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Over the time datapath.c and flow.c has became pretty large files.
Following patch restructures functionality of component into three
different components:
flow.c: contains flow extract.
flow_netlink.c: netlink flow api.
flow_table.c: flow table api.
Diffstat is showing wrong count. This patch mostly restructures code
without changing logic.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
OVS has its own workq implementation for coupe of reasons. first
was to avoid system freeze due to ovs-flow rehash softlockup.
We have moved out rehash from workq, So this problem does not exist.
second was related bugs in kernel workq implementation in pre-2.6.32
kernel. But we have dropped support for older kernel.
So there is no reason to keep ovs-workq around. Following patch
removes it.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Rehashing in ovs-workqueue can cause ovs-mutex lock contentions
in case of heavy flow setups where both needs ovs-mutex. So by
moving rehashing to flow-setup we can eliminate contention.
This also simplify ovs locking and reduces dependence on
workqueue.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
