Separate rule versioning to lib/versions.h to make it easier to use
versioning for other data types.
Signed-off-by: Jarno Rajahalme <jarno@ovn.org>
Acked-by: Ben Pfaff <blp@ovn.org>
Use the new ccmap type instead of cmap for staged lookup indices to
fix the problem with slow removal of rules with large number of
duplicates. This was problematic especially when many rules shared
the same match in packet metadata (e.g., a port number, but nothing
else), causing a large number of duplicates to be inserted into the
staged lookup index. ccmap only keeps the count of inserted (hash)
values, so duplicates do not add any performance penalty.
Reported-by: Alok Kumar Maurya <alok-kumar.maurya@hpe.com>
Signed-off-by: Jarno Rajahalme <jarno@ovn.org>
Acked-by: Ben Pfaff <blp@ovn.org>
Addition of table versioning exposed struct cls_rule member
'cls_match' to RCU readers and made it possible for 'cls_match' become
NULL while being accessed by an RCU reader, but we failed to check for
this condition. This may have resulted in NULL pointer dereference
and ovs-vswitchd crash.
Fix this by making the 'cls_match' member an RCU pointer and checking
the value whenever it potentially read by an RCU reader. In these
instances we use ovsrcu_get(), whereas functions accessible only by
the exclusive writers use ovsrcu_get_protected() and do not need to
check the result.
VMware-BZ: 1643642
Fixes: 2b7b1427 ("classifier: Support table versioning")
Signed-off-by: Jarno Rajahalme <jarno@ovn.org>
Classifier partitions allowed skipping subtables when if was known
from the flow's metadata field that the subtable cannot possibly
match. This functionality was later implemented in a more general
fashion by staged lookup, where the first stage also covers the
metadata field, among the rest of the non-packet fields in the struct
flow. While in theory skipping a subtable on the basis of the
metadata field alone could produce more effective wildcards, on the
basis of our testsuite coverage it does not seem to be the case, as
removing the partitioning feature did not result in any test failures.
Removing the partitioning feature makes classifier lookups roughly 20%
faster when a wildcard mask is not needed, and roughly 10% faster when
a wildcard mask is needed, as tested with the test-classifier
benchmark with one lookup thread.
Found by profiling with 'perf'.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Struct miniflow is now sometimes used just as a map. Define a new
struct flowmap for that purpose. The flowmap is defined as an array of
maps, and it is automatically sized according to the size of struct
flow, so it will be easier to maintain in the future.
It would have been tempting to use the existing struct bitmap for this
purpose. The main reason this is not feasible at the moment is that
some flowmap algorithms are simpler when it can be assumed that no
struct flow member requires more bits than can fit to a single map
unit. The tunnel member already requires more than 32 bits, so the map
unit needs to be 64 bits wide.
Performance critical algorithms enumerate the flowmap array units
explicitly, as it is easier for the compiler to optimize, compared to
the normal iterator. Without this optimization a classifier lookup
without wildcard masks would be about 25% slower.
With this more general (and maintainable) algorithm the classifier
lookups are about 5% slower, when the struct flow actually becomes big
enough to require a second map. This negates the performance gained
in the "Pre-compute stage masks" patch earlier in the series.
Requested-by: Ben Pfaff <blp@nicira.com>
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
This makes stage mask computation happen only when a subtable is
inserted and allows simplification of the main lookup function.
Classifier benchmark shows that this speeds up the classification
(with wildcards) about 5%.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
minimask_hash() can be simplified as each value is known to be non-zero.
Move miniflow_hash() into test-classifier.c as miniflow_hash__() as it
is no longer needed elsewhere.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Joe Stringer <joestringer@nicira.com>
Use two maps in miniflow to allow for expansion of struct flow past
512 bytes. We now have one map for tunnel related fields, and another
for the rest of the packet metadata and actual packet header fields.
This split has the benefit that for non-tunneled packets the overhead
should be minimal.
Some miniflow utilities now exist in two variants, new ones operating
over all the data, and the old ones operating only on a single 64-bit
map at a time. The old ones require doubling of code but should
execute faster, so those are used in the datapath and classifier's
lookup path.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
MSVC does not like zero sized arrays in structs. Hence, remove the
'values' member from struct miniflow and add back the getters
miniflow_values() and miniflow_get_values().
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Now that performance critical code already inlines miniflows and
minimasks, we can simplify struct miniflow by always dynamically
allocating miniflows and minimasks to the correct size. This changes
the struct minimatch to always contain pointers to its miniflow and
minimask.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
After all, there are some cases in which both the insertion version
and removal version of a rule need to be considered. This makes the
cls_match a bit bigger, but makes classifier versioning much simpler
to understand.
Also, avoid using type larger than int in an enum, as it is not
portable C.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
The traversal of the list of identical rules from the lookup threads
is fragile if the list head is removed during the list traversal.
This patch simplifies the implementation of that list by making the
list NULL terminated, singly linked RCU-protected list. By having the
NULL at the end there is no longer a possiblity of missing the point
when the list wraps around. This is significant when there can be
multiple elements with the same priority in the list.
This change also decreases the size of the struct cls_match back
pre-'visibility' attribute size.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
This patch allows classifier rules to become visible and invisible in
specific versions. A 'version' is defined as a positive monotonically
increasing integer, which never wraps around.
The new 'visibility' attribute replaces the prior 'to_be_removed' and
'visible' attributes.
When versioning is not used, the 'version' parameter should be passed
as 'CLS_MIN_VERSION' when creating rules, and 'CLS_MAX_VERSION' when
looking up flows.
This feature enables the support for atomic OpenFlow bundles without
significant performance penalty on 64-bit systems. There is a
performance decrease in 32-bit systems due to 64-bit atomics used.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
This makes it possible to tentatively add flows to the classifier
without the datapath seeing them.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
A "conjunctive match" allows higher-level matches in the flow table, such
as set membership matches, without causing a cross-product explosion for
multidimensional matches. Please refer to the documentation that this
commit adds to ovs-ofctl(8) for a better explanation, including an example.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>
So far the compressed flow data in struct miniflow has been in 32-bit
words with a 63-bit map, allowing for a maximum size of struct flow of
252 bytes. With the forthcoming Geneve options this is not sufficient
any more.
This patch solves the problem by changing the miniflow data to 64-bit
words, doubling the flow max size to 504 bytes. Since the word size
is doubled, there is some loss in compression efficiency. To counter
this some of the flow fields have been reordered to keep related
fields together (e.g., the source and destination IP addresses share
the same 64-bit word).
This change should speed up flow data processing on 64-bit CPUs, which
may help counterbalance the impact of making the struct flow bigger in
the future.
Classifier lookup stage boundaries are also changed to 64-bit
alignment, as the current algorithm depends on each miniflow word to
not be split between ranges. This has resulted in new padding (part
of the 'mpls_lse' field).
The 'dp_hash' field is also moved to packet metadata to eliminate
otherwise needed padding there. This allows the L4 to fit into one
64-bit word, and also makes matches on 'dp_hash' more efficient as
misses can be found already on stage 1.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Add support for adding 64-bit words to hashes. This will be used by
subsequent patches.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
This silences the following sparse warning which appears to be a bug
in sparse:
lib/classifier-private.h:210:45: warning: call with no type!
lib/classifier-private.h:179:45: warning: call with no type!
Can be reverted when sparse has been fixed.
Cc: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Thomas Graf <tgraf@noironetworks.com>
Reviewed-by: Flavio Leitner <fbl@redhat.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Use MAP_FOR_EACH_INDEX to improve readability when there is no
apparent cost for it.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Almost all classifier users already exclude concurrent modifications,
or are single-threaded, hence the classifier internal mutex can be
removed. Due to this change, ovs-router.c and tnl-ports.c need new
mutexes, which are added.
As noted by Ben in review, ovs_router_flush() should also free the
entries it removes from the classifier. It now calls
ovsrcu_postpone() to that effect.
Suggested-by: Ben Pfaff <blp@nicira.com>
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Previously, accurate iteration required writers to be excluded during
iteration. This patch adds an rculist to struct cls_subtable, and a
corresponding list node to struct cls_rule, which makes iteration more
straightforward, and allows the iterators to remain ignorant of the
internals of the cls_match. This new list allows iteration of rules
in the classifier by traversing the RCU-friendly subtables vector, and
the rculist of rules in each subtable.
Classifier modifications may be performed concurrently, but whether or
not the concurrent iterator sees those changes depends on the timing
of change. More specifically, an concurrent iterator:
- May or may not see a rule that is being inserted or removed.
- Will see either the new or the old version of a rule that is replaced.
- Will see all the other rules (that are not being modified).
Finally, The subtable's rculist also allows to make
classifier_rule_overlaps() lockless, which this patch also does.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
There is no point in adding duplicate information into prefix tries.
Also, since the lower-priority duplicate rules are not visible to
lookups, they do not need to be in staged lookup indices directly
either (the head rule is).
Finally, now that cmap operations return the number of elements in the
cmap, subtable's 'n_rules' member is not needed any more.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Some struct cls_match and cls_subtable fields were already documented
of being const. Make them const and use CONST_CAST where appropriate
to initialize them.
This will help catch future errors modifying those fields after
initialization.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
The list of identical, but lower priority rules is not currently used
in classifier lookup. A later patch introducing conjunctive matches
needs to access the list during lookups, so we must make the list RCU.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
OpenFlow has priorities in the 16-bit unsigned range, from 0 to 65535.
In the classifier, it is sometimes useful to be able to have values below
and above this range. With the 'unsigned int' type used for priorities
until now, there were no values below the range, so some code worked
around it by converting priorities to 64-bit signed integers. This didn't
seem so great to me given that a plain 'int' also had the needed range.
This commit therefore changes the type used for priorities to int.
The interesting parts of this change are in pvector.h and classifier.c,
where one can see the elimination of the use of int64_t.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>
tests/test-classifier.c used to include lib/classifier.c to gain
access to the internal data structures and some utility functions.
This was confusing, so this patch splits the relevant groups of
classifier internal definations to a new file
(lib/classifier-private.h), which is included by both lib/classifier.c
and tests/test-classifier.c. Other use of the new file is
discouraged.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
