We introduce a simple variable window congestion control for links.
The algorithm is inspired by the Reno algorithm, covering both 'slow
start', 'congestion avoidance', and 'fast recovery' modes.
- We introduce hard lower and upper window limits per link, still
different and configurable per bearer type.
- We introduce a 'slow start theshold' variable, initially set to
the maximum window size.
- We let a link start at the minimum congestion window, i.e. in slow
start mode, and then let is grow rapidly (+1 per rceived ACK) until
it reaches the slow start threshold and enters congestion avoidance
mode.
- In congestion avoidance mode we increment the congestion window for
each window-size number of acked packets, up to a possible maximum
equal to the configured maximum window.
- For each non-duplicate NACK received, we drop back to fast recovery
mode, by setting the both the slow start threshold to and the
congestion window to (current_congestion_window / 2).
- If the timeout handler finds that the transmit queue has not moved
since the previous timeout, it drops the link back to slow start
and forces a probe containing the last sent sequence number to the
sent to the peer, so that this can discover the stale situation.
This change does in reality have effect only on unicast ethernet
transport, as we have seen that there is no room whatsoever for
increasing the window max size for the UDP bearer.
For now, we also choose to keep the limits for the broadcast link
unchanged and equal.
This algorithm seems to give a 50-100% throughput improvement for
messages larger than MTU.
Suggested-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The TIPC_MEDIA_ADDR_SIZE and TIPC_MEDIA_ADDR_OFFSET names
are misleading, as they actually define the size and offset of
the whole media info field and not the address part. This patch
does not have any functional changes.
Signed-off-by: Erik Hugne <erik.hugne@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TIPC currently handles two media specific addresses: Ethernet MAC
addresses and InfiniBand addresses. Those are kept in three different
formats:
1) A "raw" format as obtained from the device. This format is known
only by the media specific adapter code in eth_media.c and
ib_media.c.
2) A "generic" internal format, in the form of struct tipc_media_addr,
which can be referenced and passed around by the generic media-
unaware code.
3) A serialized version of the latter, to be conveyed in neighbor
discovery messages.
Conversion between the three formats can only be done by the media
specific code, so we have function pointers for this purpose in
struct tipc_media. Here, the media adapters can install their own
conversion functions at startup.
We now introduce a new such function, 'raw2addr()', whose purpose
is to convert from format 1 to format 2 above. We also try to as far
as possible uniform commenting, variable names and usage of these
functions, with the purpose of making them more comprehensible.
We can now also remove the function tipc_l2_media_addr_set(), whose
job is done better by the new function.
Finally, we expand the field for serialized addresses (format 3)
in discovery messages from 20 to 32 bytes. This is permitted
according to the spec, and reduces the risk of problems when we
add new media in the future.
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently TIPC supports two L2 media types, Ethernet and Infiniband.
Because both these media are accessed through the common net_device API,
several functions in the two media adaptation files turn out to be
fully or almost identical, leading to unnecessary code duplication.
In this commit we extract this common code from the two media files
and move them to the generic bearer.c. Additionally, we change
the function names to reflect their real role: to access L2 media,
irrespective of type.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Patrick McHardy <kaber@trash.net>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, registering a TIPC stack handler in the network device layer
is done twice, once for Ethernet (eth_media) and Infiniband (ib_media)
repectively. But, as this registration is not media specific, we can
avoid some code duplication by moving the registering function to
the generic bearer layer, to the file bearer.c, and call it only once.
The same is true for the network device event notifier.
As a side effect, the two workqueues we are using for for setting up/
cleaning up media can now be eliminated. Furthermore, the array for
storing the specific media type structs, media_array[], can be entirely
deleted.
Note that the eth_started and ib_started flags were removed during the
code relocation. There is now only one call to bearer_setup and
bearer_cleanup, and these can logically not race against each other.
Despite its size, this cleanup work incurs no functional changes in TIPC.
In particular, it should be noted that the sequence ordering of received
packets is unaffected by this change, since packet reception never was
subject to any work queue handling in the first place.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TIPC is currently using the field 'af_packet_priv' in struct net_device
as a handle to find the bearer instance associated to the given network
device. But, by doing so it is blocking other networking cleanups, such
as the one discussed here:
http://patchwork.ozlabs.org/patch/178044/
This commit removes this usage from TIPC. Instead, we introduce a new
field, 'tipc_ptr', to the net_device structure, to serve this purpose.
When TIPC bearer is enabled, the bearer object is associated to
'tipc_ptr'. When a TIPC packet arrives in the recv_msg() upcall
from a networking device, the bearer object can now be obtained from
'tipc_ptr'. When a bearer is disabled, the bearer object is detached
from its underlying network device by setting 'tipc_ptr' to NULL.
Additionally, an RCU lock is used to protect the new pointer.
Henceforth, the existing tipc_net_lock is used in write mode to
serialize write accesses to this pointer, while the new RCU lock is
applied on the read side to ensure that the pointer is 100% valid
within its wrapped area for all readers.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Patrick McHardy <kaber@trash.net>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Communication media types are abstracted through the struct 'tipc_media',
one per media type. These structs are allocated statically inside their
respective media file.
Furthermore, in order to be able to reach all instances from a central
location, we keep a static array with pointers to these structs. This
array is currently initialized at runtime, under protection of
tipc_net_lock. However, since the contents of the array itself never
changes after initialization, we can just as well initialize it at
compile time and make it 'const', at the same time making it obvious
that no lock protection is needed here.
This commit makes the array constant and removes the redundant lock
protection.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
struct 'tipc_bearer' is a generic representation of the underlying
media type, and exists in a one-to-one relationship to each interface
TIPC is using. The struct contains a 'blocked' flag that mirrors the
operational and execution state of the represented interface, and is
updated through notification calls from the latter. The users of
tipc_bearer are checking this flag before each attempt to send a
packet via the interface.
This state mirroring serves no purpose in the current code base. TIPC
links will not discover a media failure any faster through this
mechanism, and in reality the flag only adds overhead at packet
sending and reception.
Furthermore, the fact that the flag needs to be protected by a spinlock
aggregated into tipc_bearer has turned out to cause a serious and
completely unnecessary deadlock problem.
CPU0 CPU1
---- ----
Time 0: bearer_disable() link_timeout()
Time 1: spin_lock_bh(&b_ptr->lock) tipc_link_push_queue()
Time 2: tipc_link_delete() tipc_bearer_blocked(b_ptr)
Time 3: k_cancel_timer(&req->timer) spin_lock_bh(&b_ptr->lock)
Time 4: del_timer_sync(&req->timer)
I.e., del_timer_sync() on CPU0 never returns, because the timer handler
on CPU1 is waiting for the bearer lock.
We eliminate the 'blocked' flag from struct tipc_bearer, along with all
tests on this flag. This not only resolves the deadlock, but also
simplifies and speeds up the data path execution of TIPC. It also fits
well into our ongoing effort to make the locking policy simpler and
more manageable.
An effect of this change is that we can get rid of functions such as
tipc_bearer_blocked(), tipc_continue() and tipc_block_bearer().
We replace the latter with a new function, tipc_reset_bearer(), which
resets all links associated to the bearer immediately after an
interface goes down.
A user might notice one slight change in link behaviour after this
change. When an interface goes down, (e.g. through a NETDEV_DOWN
event) all attached links will be reset immediately, instead of
leaving it to each link to detect the failure through a timer-driven
mechanism. We consider this an improvement, and see no obvious risks
with the new behavior.
Signed-off-by: Erik Hugne <erik.hugne@ericsson.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Paul Gortmaker <Paul.Gortmaker@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, rcv_msg() always returns zero on a packet delivery upcall
from net_device.
To make its behavior more compliant with the way this API should be
used, we change this to let it return NET_RX_SUCCESS (which is zero
anyway) when it is able to handle the packet, and NET_RX_DROP otherwise.
The latter does not imply any functional change, it only enables the
driver to keep more accurate statistics about the fate of delivered
packets.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tipc_block_bearer() currently takes a bearer name (const char*)
as argument. This requires the function to make a lookup to find
the pointer to the corresponding bearer struct. In the current
code base this is not necessary, since the only two callers
(tipc_continue(),recv_notification()) already have validated
copies of this pointer, and hence can pass it directly in the
function call.
We change tipc_block_bearer() to directly take struct tipc_bearer*
as argument instead.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TIPC 'bearer' exists as an abstract concept, while 'media'
is deemed a specific implementation of a bearer, such as Ethernet
or Infiniband media. When a component inside TIPC wants to control
a specific media, it only needs to access the generic bearer API
to achieve this. However, in the current media implementations,
the 'bearer' name is also extensively used in media specific
function and variable names.
This may create confusion, so we choose to replace the term 'bearer'
with 'media' in all function names, variable names, and prefixes
where this is what really is meant.
Note that this change is cosmetic only, and no runtime behaviour
changes are made here.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of passing each byte by stack let's use nice specifier for that.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Convert enable_bearer() to RCU locking with dev_get_by_name().
Based on a similar changeset in commit 840a185d ["aoe: remove
dev_base_lock use from aoecmd_cfg_pkts()"] -- quoting that:
"dev_base_lock is the legacy way to lock the device list,
and is planned to disappear. (writers hold RTNL, readers
hold RCU lock)"
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
So far, only net_device * could be passed along with netdevice notifier
event. This patch provides a possibility to pass custom structure
able to provide info that event listener needs to know.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
v2->v3: fix typo on simeth
shortened dev_getter
shortened notifier_info struct name
v1->v2: fix notifier_call parameter in call_netdevice_notifier()
Signed-off-by: David S. Miller <davem@davemloft.net>
Add InfiniBand media type based on the ethernet media type.
The only real difference is that in case of InfiniBand, we need the entire
20 bytes of space reserved for media addresses, so the TIPC media type ID is
not explicitly stored in the packet payload.
Sample output of tipc-config:
# tipc-config -v -addr -netid -nt=all -p -m -b -n -ls
node address: <10.1.4>
current network id: 4711
Type Lower Upper Port Identity Publication Scope
0 167776257 167776257 <10.1.1:1855512577> 1855512578 cluster
167776260 167776260 <10.1.4:1216454657> 1216454658 zone
1 1 1 <10.1.4:1216479235> 1216479236 node
Ports:
1216479235: bound to {1,1}
1216454657: bound to {0,167776260}
Media:
eth
ib
Bearers:
ib:ib0
Nodes known:
<10.1.1>: up
Link <broadcast-link>
Window:20 packets
RX packets:0 fragments:0/0 bundles:0/0
TX packets:0 fragments:0/0 bundles:0/0
RX naks:0 defs:0 dups:0
TX naks:0 acks:0 dups:0
Congestion bearer:0 link:0 Send queue max:0 avg:0
Link <10.1.4:ib0-10.1.1:ib0>
ACTIVE MTU:2044 Priority:10 Tolerance:1500 ms Window:50 packets
RX packets:80 fragments:0/0 bundles:0/0
TX packets:40 fragments:0/0 bundles:0/0
TX profile sample:22 packets average:54 octets
0-64:100% -256:0% -1024:0% -4096:0% -16384:0% -32768:0% -66000:0%
RX states:410 probes:213 naks:0 defs:0 dups:0
TX states:410 probes:197 naks:0 acks:0 dups:0
Congestion bearer:0 link:0 Send queue max:1 avg:0
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Maloy
Erik Hugne
Ying Xue
Andy Shevchenko
Jiri Pirko
Patrick McHardy