Some versions of Centos 5.6 backport the flag IFF_BRIDGE_PORT
without the associated rx_handler changes, so this changes to
use a version check since we really don't care about the actual
symbol.
Reported-by: Srinivasan Ramasubramanian <vrsrini@gmail.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Currently ovs is using device stats for Linux devices and count them
itself in other situations. This leads to overlap with hardware stats,
inconsistencies, etc. It's much better to just always count the packets
flowing through the switch and let userspace do any merging that it wants.
Following patch removes vport->get_stats() interface. vport-stat is changed
to use new `struct ovs_vport_stat` rather than rtnl_link_stats64.
Definitions of rtnl_link_stats64 is removed from OVS. dipf_port->stat is also
removed as aggregate stats are only available at netdev layer.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Currently the kernel automatically sets the MTU of any internal
interfaces to the minimum of all attached interfaces because the Linux
bridge does this. Userspace can do this with more knowledge and
flexibility.
Feature #7323
Signed-off-by: Justin Pettit <jpettit@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Currently OVS uses its own hashing implmentation for hash tables
which has some problems, e.g. error case on deletion code.
Following patch replaces that with hlist based hash table which is
consistent with other kernel hash tables. As Jesse suggested, flex-array
is used for allocating hash buckets, So that we can have large
hash-table without large contiguous kernel memory.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Currently the kernel vlan actions mirror those used by OpenFlow 1.0.
i.e. MODIFY and STRIP. More flexible approach is to have an action to
push a tag and pop a tag off, so that it can handle multiple levels of vlan
tags. Plus it aligns with newer version of OpenFlow.
As this patch replaces MODIFY with PUSH semantic, action
mapping done in userpace is fixed accordingly.
GSO handling for multiple levels of vlan tags is also added as
Jesse suggested before.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Remove iflink from vport interface. iflink is not used anywhere in
OVS. So there is not need to have iflink as vport attribute.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Whenever a port is added to the datapath, LRO is automatically disabled.
In the future, we may want to enable LRO in some circumstances, so have
userspace disable LRO through the ethtool ioctls.
As part of this change, the MTU and LRO checks are moved to
netdev-vport's send(), which is where they're actually needed.
Feature #6810
Signed-off-by: Justin Pettit <jpettit@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
The prefix "ODP_*" is not overly descriptive in the context of the
larger Linux tree. This commit changes the prefix to "OVS_*" for the
userpace to kernel interactions. The userspace libraries still use
"ODP_" in many of their interfaces since it is more descriptive in the
OVS oeuvre.
Feature #6904
Signed-off-by: Justin Pettit <jpettit@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
It's possible to trace kfree_skb() call sites to find out where
packets are getting dropped. Situations where kfree_skb() does
not actually indicate an error adds additional noise, so use
consume_skb() instead to avoid tracing non-errors.
Suggested-by: Ben Pfaff <blp@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Older kernels (those before 2.6.22) rely on implicit assumptions
to determine checksum offloading status. These assumptions tend
to break down when doing switching because it sits in the middle
of the transmit and receive path. Newer kernels deal with this
problem by adding more explicit information about how to checksum.
This replicates that behavior by mirroring the state from newer
kernels in private OVS storage on the kernels that lack it. On
ingress and egress we then map that state onto the appropriate
location for the given kernel and can consistently manipulate it
within OVS. Some of this was already done for the checksum type
but this makes it more robust and expands it to the checksum start
and offset as well.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Starting in 2.6.37 we have our own flag for identifying net_devices
as being attached to OVS. However, it's possible to receive packets
before this flag has been applied, resulting in a NULL vport when
processing the packet. This checks to make sure that the vport is
valid instead of crashing.
Bug #5675
Reported-by: Brad Hall <brad@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
netdev_rx_handler_register() changed the type of the skb argument to the
callback function as well as the return type. Special-case
netdev_frame_hook() to do the right thing on 2.6.39 and later.
Signed-off-by: Andrew Evans <aevans@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Some Linux network drivers support a feature called "VLAN acceleration",
associated with a data structure called a "vlan_group". A vlan_group is,
abstractly, a dictionary that maps from a VLAN ID (in the range 0...4095)
to a VLAN device, that is, a Linux network device associated with a
particular VLAN, e.g. "eth0.9" for VLAN 9 on eth0.
Some drivers that support VLAN acceleration have bugs that fall roughly
into the following categories:
* Some NICs strip VLAN tags on receive if no vlan_group is registered,
so that the tag is completely lost.
* Some drivers size their receive buffers based on whether a vlan_group
is enabled, meaning that a maximum size packet with a VLAN tag will
not fit if a vlan_group is not configured.
* On transmit some drivers expect that VLAN acceleration will be used
if it is available (which can only be done if a vlan_group is
configured). In these cases, the driver may fail to parse the packet
and correctly setup checksum offloading and/or TSO.
The correct long term solution is to fix these driver bugs. To cope until
then, we have prepared a patch to the Linux kernel network stack that works
around these problems. This commit adds support for the workaround
implemented by that patch.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Support for offloading over vlans wasn't introduced until 2.6.26,
so do full software emulation on kernels before that when dealing
with vlan packets.
Reported-by: Justin Pettit <jpettit@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
We currently perform GSO on packets before adding a vlan tag,
which is reliable but hurts performance. Even NICs that support
TSO on vlan tagged packets typically expect vlan acceleration to
be used. Before 2.6.37 we can't use vlan acceleration and must
place the tag in the packet itself, which is risky when used with
TSO. However, if the driver is known to work with internally
tagged packets and TSO this exposes a module parameter to enable it.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Using the kernel vlan acceleration has a number of benefits:
it enables hardware tagging, allows usage of TSO and checksum
offloading, and is generally easier to manipulate. This switches
the vlan actions to use skb->vlan_tci field for any necessary
changes. In places that do not support vlan acceleration in a way
that we can use (in particular kernels before 2.6.37) we perform
any necessary conversions, such as tagging and GSO before the
packet leaves Open vSwitch.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Kernels prior to 2.6.27 did not have a vlan_tci field in struct
sk_buff for vlan acceleration. It's very convenient to use this
field for manipulating vlan tags, so we would like to use it as
the primary mechanism. To enable this, this commit adds similar
infrastructure to the OVS_CB on the kernels that need it and a
set of functions to use the correct location.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
I plan to make the vport type part of the standard header stuck on each
Netlink message related to a vport. As such, it is more convenient to use
an integer than a string. In addition, by being fundamentally different
from strings, using an integer may reduce the confusion we've had in the
past over the differences in userspace and kernel names for network device
and vport types.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
I had completely forgotten that we had a top-level compat.h and compat26.h.
It's better to distribute their contents to individual compat headers, so
this commit does so and deletes them.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Using rcu_dereference() makes lockdep complain if rcu_read_lock
is not held. This is OK if the update side lock is held. This
adds checks to see if RTNL lock is held, if that is also a
correct form of protection. Alternately, it enforces that RTNL
must be held.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Starting in 2.6.37 we have our own unique identifier to be able
to find ports attached to OVS. Take advantage of it to avoid
ugly workarounds.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Checksum offloading has changed quite a bit across different kernel
and Xen versions. Since it is part of the skb data structure it is
unfortunately difficult to separate out into compatibility code.
This consolidates all of the checksum code in one place which makes
it easier read and remove as we prepare for upstreaming. On newer
kernels it also puts everything in inline functions, eliminating the
need to run through the compat code or make extra function calls.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
These steps are sequentially in lockstep, so we might as well combine them.
This expands the region over which the vport_lock is held. I didn't
carefully verify that this was OK.
This also eliminates the synchronize_rcu() call from destruction of tunnel
vports, since they didn't appear to me to need it.
It should be possible to eliminate the synchronize_rcu() from the netdev,
patch, and internal_dev vports, but this commit does not do that.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
After the previous commit, which changed the datapath to always create and
attach a vport at the same time, and to always detach and delete a vport
at the same time, there is no longer any real distinction between a dp_port
and a vport. This commit, therefore, merges the two together to simplify
code. It might even improve performance, although I have not checked.
I wasn't sure at first whether the merged structure should be "struct
dp_port" or "struct vport". I went with the latter since the "v" prefix
sounds cool.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Upcoming commits will keep needing to pass more information to the vport
'create' member function. It's annoying to have to modify a dozen pieces
of code every time just to do this, so this commit encapsulates all of the
parameters in a new struct and passes that instead.
Acked-by: Jesse Gross <jesse@nicira.com>
vport_ops, tunnel_ops, and ethtool_ops should not change at runtime.
Therefore, mark them as const to keep them out of the hotpath and to
prevent them from getting trampled.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
We currently call skb_reset_mac_header() in a few places when a
packet is received. However, this is not needed because flow_extract()
will set all of the protocol headers during parsing and nothing needs
the packet headers before that time.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Linux 2.6.35 added struct rtnl_link_stats64, which as a set of 64-bit
network device counters is what the OVS datapath needs. We might as well
use it instead of our own.
This commit moves the if_link.h compat header from datapath/ into the
top-level include/ directory so that it is visible both to kernel and
userspace code.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
dev_get_stats() is documented to return its output argument, so there's no
need to maintain a separate pointer variable.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Build tested (only) only on i386 for 2.6.26, 2.6.29, 2.6.33, 2.6.34,
and 2.6.36, and on x86-64 for 2.6.31 and 2.6.36.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
This adds compatibility with a series kernel changesets that
introduces 64bit statistics. The final changeset (to date) being
"net: Document that dev_get_stats() returns the given pointer".
The relevant changesets were added between 2.6.35 and 2.6.36-rc1.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Jesse Gross <jesse@nicira.com>
This adds compatibility with kernel changeset
"bridge: use rx_handler_data pointer to store net_bridge_port pointer"
which was added between 2.6.35 and 2.6.36-rc1.
With this change it is now safe to (attempt to) insert both bridge and
datapath with newer (>=2.6.36) kernels, although whichever is inserted
second will fail to initialise on the call to netdev_rx_handler_register()
Signed-off-by: Simon Horman <horms@verge.net.au>
[Jesse: fixed merge conflicts in vport-netdev.c and netdevice.h]
Signed-off-by: Jesse Gross <jesse@nicira.com>
Although not strictly necessary, this will make this
function more consistent when compatibility for 2.6.36 is added.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Jesse Gross <jesse@nicira.com>
For kernels that have netdev_rx_handler_register() (>=2.6.35),
duplicate netdevs are detected by netdev_rx_handler_register().
So by adding duplicate detection to the netdev_rx_handler_register()
compatibility code the explicit check in netdev_create() can be removed.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Jesse Gross <jesse@nicira.com>
This adds compatibility with kernel changeset
of changeset "net: add rx_handler data pointer"
and thus "net: replace hooks in __netif_receive_skb V5",
which were added between 2.6.35 and 2.6.36-rc1
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Jesse Gross <jesse@nicira.com>
On vport ingress we already check for shared SKBs but then later
warn in several other places. In a similar vein, we check every
packet to see if it is LRO but only certain vports can produce
these packets. Remove and consolidate checks to the places where
they are needed.
In some places we would put the return type on the same line as
the rest of the function definition and other places we wouldn't.
Reformat everything to match kernel style.
Linux devices store stats in counters the size of a machine word,
which are rapidly overflowed on a 32-bit machine. In this
situation we now use the vport stats layer, which always uses 64-
bit stats. On 64-bit machines we continue to use the normal
Linux stats to avoid the extra overhead of counting twice.
Since vport implementations have no header files they needed to be
declared as extern before being used. They are currently declared
in vport.c but this isn't safe because the compiler will silently
accept it if the type is incorrect. This moves those declarations
into vport.h, which is included by all implementations and will
cause errors about conflicting types if there is a mismatch.
Enables checksum offloading, scatter/gather, and TSO on internal
devices. While these optimizations were not previously enabled on
internal ports we already could receive these types of packets from
Xen guests. This has the obvious performance benefits when these
packets can be passed directly to hardware.
There is also a more subtle benefit for GRE on Xen. GRE packets
pass through OVS twice - once before encapsulation and once after
encapsulation, moving through an internal device in the process.
If it is a SG packet (as is common on Xen), a copy was necessary
to linearize for the internal device. However, Xen uses the
memory allocator to track packets so when the original packet is
freed after the copy netback notifies the guest that the packet
has been sent, despite the fact that it is actually sitting in the
transmit queue. The guest then sends packets as fast as the CPU
can handle, overflowing the transmit queue. By enabling SG on
the internal device, we avoid the copy and keep the accounting
correct.
In certain circumstances this patch can decrease performance for
TCP. TCP has its own mechanism for tracking in-flight packets
and therefore does not benefit from the corrected socket accounting.
However, certain NICs do not like SG when it is not being used for
TSO (these packets can no longer be handled by TSO after GRE
encapsulation). These NICs presumably enable SG even though they
can't handle it well because TSO requires SG.
Tested controllers (all 1G):
Marvell 88E8053 (large performance hit)
Broadcom BCM5721 (small performance hit)
Intel 82571EB (no change)
Currently the datapath directly accesses devices through their
Linux functions. Obviously this doesn't work for virtual devices
that are not backed by an actual Linux device. This creates a
new virtual port layer which handles all interaction with devices.
The existing support for Linux devices was then implemented on top
of this layer as two device types. It splits out and renames dp_dev
to internal_dev. There were several places where datapath devices
had to handled in a special manner and this cleans that up by putting
all the special casing in a single location.
