|
|
|
|
/* Copyright (C) 2013-2014 Open Information Security Foundation
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
*
|
|
|
|
|
* You can copy, redistribute or modify this Program under the terms of
|
|
|
|
|
* the GNU General Public License version 2 as published by the Free
|
|
|
|
|
* Software Foundation.
|
|
|
|
|
*
|
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
|
*
|
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
|
* version 2 along with this program; if not, write to the Free Software
|
|
|
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
|
|
|
|
|
* 02110-1301, USA.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* \file
|
|
|
|
|
*
|
|
|
|
|
* \author Anoop Saldanha <anoopsaldanha@gmail.com>
|
|
|
|
|
* \author Ken Steele <suricata@tilera.com>
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#ifndef __UTIL_MPM_AC_TILE__H__
|
|
|
|
|
#define __UTIL_MPM_AC_TILE__H__
|
|
|
|
|
|
|
|
|
|
typedef struct SCACTilePatternList_ {
|
|
|
|
|
uint8_t *cs;
|
|
|
|
|
uint16_t patlen;
|
|
|
|
|
|
|
|
|
|
uint16_t offset;
|
|
|
|
|
uint16_t depth;
|
|
|
|
|
|
In AC-Tile, convert from using pids for indexing to pattern index
Use an MPM specific pattern index, which is simply an index starting
at zero and incremented for each pattern added to the MPM, rather than
the externally provided Pattern ID (pid), since that can be much
larger than the number of patterns. The Pattern ID is shared across at
MPMs. For example, an MPM with one pattern with pid=8000 would result
in a max_pid of 8000, so the pid_pat_list would have 8000 entries.
The pid_pat_list[] is replaced by a array of pattern indexes. The PID is
moved to the SCACTilePatternList as a single value. The PatternList is
also indexed by the Pattern Index.
max_pat_id is no longer needed and mpm_ctx->pattern_cnt is used instead.
The local bitarray is then also indexed by pattern index instead of PID, making
it much smaller. The local bit array sets a bit for each pattern found
for this MPM. It is only kept during one MPM search (stack allocated).
One note, the local bit array is checked first and if the pattern has already
been found, it will stop checking, but count a match. This could result in
over counting matches of case-sensitve matches, since following case-insensitive
matches will also be counted. For example, finding "Foo" in "foo Foo foo" would
report finding "Foo" 2 times, mis-counting the third word as "Foo".
11 years ago
|
|
|
/* Pattern Id */
|
|
|
|
|
uint32_t pid;
|
|
|
|
|
|
|
|
|
|
/* sid(s) for this pattern */
|
|
|
|
|
uint32_t sids_size;
|
|
|
|
|
SigIntId *sids;
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
} SCACTilePatternList;
|
|
|
|
|
|
|
|
|
|
typedef struct SCACTileOutputTable_ {
|
In AC-Tile, convert from using pids for indexing to pattern index
Use an MPM specific pattern index, which is simply an index starting
at zero and incremented for each pattern added to the MPM, rather than
the externally provided Pattern ID (pid), since that can be much
larger than the number of patterns. The Pattern ID is shared across at
MPMs. For example, an MPM with one pattern with pid=8000 would result
in a max_pid of 8000, so the pid_pat_list would have 8000 entries.
The pid_pat_list[] is replaced by a array of pattern indexes. The PID is
moved to the SCACTilePatternList as a single value. The PatternList is
also indexed by the Pattern Index.
max_pat_id is no longer needed and mpm_ctx->pattern_cnt is used instead.
The local bitarray is then also indexed by pattern index instead of PID, making
it much smaller. The local bit array sets a bit for each pattern found
for this MPM. It is only kept during one MPM search (stack allocated).
One note, the local bit array is checked first and if the pattern has already
been found, it will stop checking, but count a match. This could result in
over counting matches of case-sensitve matches, since following case-insensitive
matches will also be counted. For example, finding "Foo" in "foo Foo foo" would
report finding "Foo" 2 times, mis-counting the third word as "Foo".
11 years ago
|
|
|
/* list of pattern indexes */
|
|
|
|
|
MpmPatternIndex *patterns;
|
|
|
|
|
/* number of entries in pattern list */
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
uint32_t no_of_entries;
|
|
|
|
|
} SCACTileOutputTable;
|
|
|
|
|
|
|
|
|
|
struct SCACTileSearchCtx_;
|
|
|
|
|
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
/* Reordered for Tilera cache */
|
|
|
|
|
typedef struct SCACTileCtx_ {
|
|
|
|
|
|
|
|
|
|
/* Convert input character to matching alphabet */
|
|
|
|
|
uint8_t translate_table[256];
|
|
|
|
|
|
Add 8-bit states to ac-tile
When running with sgh-mpm-context: full, many more MPMs are created
(16K) and many are small. If they have less than 128 states, they only
need 1 byte for the next state instead of 2 bytes, cutting the size of
the next-state table in half. This reduces total memory usage.
Since that makes 3 different state sizes (1, 2 and 4 bytes), rather
than going from 2 copies of the code to create the MPM to 3, I
factored out the code that fills the next-state table into three
functions so that all the other code could be the same.
The search function is now parameterize for 8-bit and 16-bit state
sizes and alphabet sizes 8, 16, 32, 64, 128 and 256.
12 years ago
|
|
|
/* The all important memory hungry state_table.
|
|
|
|
|
* The size of each next-state is determined by bytes_per_state.
|
|
|
|
|
*/
|
|
|
|
|
void *state_table;
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
|
|
|
|
|
/* Specialized search function based on size of data in delta
|
|
|
|
|
* tables. The alphabet size determines address shifting and the
|
|
|
|
|
* number of states could make the next state could be 16 bits or
|
|
|
|
|
* 32 bits.
|
|
|
|
|
*/
|
|
|
|
|
uint32_t (*Search)(const struct SCACTileSearchCtx_ *ctx, struct MpmThreadCtx_ *,
|
|
|
|
|
PrefilterRuleStore *, const uint8_t *, uint16_t);
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
|
Add 8-bit states to ac-tile
When running with sgh-mpm-context: full, many more MPMs are created
(16K) and many are small. If they have less than 128 states, they only
need 1 byte for the next state instead of 2 bytes, cutting the size of
the next-state table in half. This reduces total memory usage.
Since that makes 3 different state sizes (1, 2 and 4 bytes), rather
than going from 2 copies of the code to create the MPM to 3, I
factored out the code that fills the next-state table into three
functions so that all the other code could be the same.
The search function is now parameterize for 8-bit and 16-bit state
sizes and alphabet sizes 8, 16, 32, 64, 128 and 256.
12 years ago
|
|
|
/* Function to set the next state based on size of next state
|
|
|
|
|
* (bytes_per_state).
|
|
|
|
|
*/
|
|
|
|
|
void (*SetNextState)(struct SCACTileCtx_ *ctx, int state, int aa,
|
Add 8-bit states to ac-tile
When running with sgh-mpm-context: full, many more MPMs are created
(16K) and many are small. If they have less than 128 states, they only
need 1 byte for the next state instead of 2 bytes, cutting the size of
the next-state table in half. This reduces total memory usage.
Since that makes 3 different state sizes (1, 2 and 4 bytes), rather
than going from 2 copies of the code to create the MPM to 3, I
factored out the code that fills the next-state table into three
functions so that all the other code could be the same.
The search function is now parameterize for 8-bit and 16-bit state
sizes and alphabet sizes 8, 16, 32, 64, 128 and 256.
12 years ago
|
|
|
int new_state, int outputs);
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
|
In AC-Tile, convert from using pids for indexing to pattern index
Use an MPM specific pattern index, which is simply an index starting
at zero and incremented for each pattern added to the MPM, rather than
the externally provided Pattern ID (pid), since that can be much
larger than the number of patterns. The Pattern ID is shared across at
MPMs. For example, an MPM with one pattern with pid=8000 would result
in a max_pid of 8000, so the pid_pat_list would have 8000 entries.
The pid_pat_list[] is replaced by a array of pattern indexes. The PID is
moved to the SCACTilePatternList as a single value. The PatternList is
also indexed by the Pattern Index.
max_pat_id is no longer needed and mpm_ctx->pattern_cnt is used instead.
The local bitarray is then also indexed by pattern index instead of PID, making
it much smaller. The local bit array sets a bit for each pattern found
for this MPM. It is only kept during one MPM search (stack allocated).
One note, the local bit array is checked first and if the pattern has already
been found, it will stop checking, but count a match. This could result in
over counting matches of case-sensitve matches, since following case-insensitive
matches will also be counted. For example, finding "Foo" in "foo Foo foo" would
report finding "Foo" 2 times, mis-counting the third word as "Foo".
11 years ago
|
|
|
/* List of patterns that match for this state. Indexed by State Number */
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
SCACTileOutputTable *output_table;
|
In AC-Tile, convert from using pids for indexing to pattern index
Use an MPM specific pattern index, which is simply an index starting
at zero and incremented for each pattern added to the MPM, rather than
the externally provided Pattern ID (pid), since that can be much
larger than the number of patterns. The Pattern ID is shared across at
MPMs. For example, an MPM with one pattern with pid=8000 would result
in a max_pid of 8000, so the pid_pat_list would have 8000 entries.
The pid_pat_list[] is replaced by a array of pattern indexes. The PID is
moved to the SCACTilePatternList as a single value. The PatternList is
also indexed by the Pattern Index.
max_pat_id is no longer needed and mpm_ctx->pattern_cnt is used instead.
The local bitarray is then also indexed by pattern index instead of PID, making
it much smaller. The local bit array sets a bit for each pattern found
for this MPM. It is only kept during one MPM search (stack allocated).
One note, the local bit array is checked first and if the pattern has already
been found, it will stop checking, but count a match. This could result in
over counting matches of case-sensitve matches, since following case-insensitive
matches will also be counted. For example, finding "Foo" in "foo Foo foo" would
report finding "Foo" 2 times, mis-counting the third word as "Foo".
11 years ago
|
|
|
/* Indexed by MpmPatternIndex */
|
|
|
|
|
SCACTilePatternList *pattern_list;
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
|
|
|
|
|
/* pattern arrays. We need this only during the goto table
|
|
|
|
|
creation phase */
|
|
|
|
|
MpmPattern **parray;
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
|
|
|
|
|
/* goto_table, failure table and output table. Needed to create
|
|
|
|
|
* state_table. Will be freed, once we have created the
|
|
|
|
|
* state_table */
|
|
|
|
|
int32_t (*goto_table)[256];
|
|
|
|
|
int32_t *failure_table;
|
|
|
|
|
|
Add 8-bit states to ac-tile
When running with sgh-mpm-context: full, many more MPMs are created
(16K) and many are small. If they have less than 128 states, they only
need 1 byte for the next state instead of 2 bytes, cutting the size of
the next-state table in half. This reduces total memory usage.
Since that makes 3 different state sizes (1, 2 and 4 bytes), rather
than going from 2 copies of the code to create the MPM to 3, I
factored out the code that fills the next-state table into three
functions so that all the other code could be the same.
The search function is now parameterize for 8-bit and 16-bit state
sizes and alphabet sizes 8, 16, 32, 64, 128 and 256.
12 years ago
|
|
|
/* Number of states used by ac-tile */
|
|
|
|
|
uint32_t state_count;
|
|
|
|
|
/* Number of states allocated for ac-tile. */
|
|
|
|
|
uint32_t allocated_state_count;
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
|
|
|
|
|
uint32_t alpha_hist[256];
|
Add 8-bit states to ac-tile
When running with sgh-mpm-context: full, many more MPMs are created
(16K) and many are small. If they have less than 128 states, they only
need 1 byte for the next state instead of 2 bytes, cutting the size of
the next-state table in half. This reduces total memory usage.
Since that makes 3 different state sizes (1, 2 and 4 bytes), rather
than going from 2 copies of the code to create the MPM to 3, I
factored out the code that fills the next-state table into three
functions so that all the other code could be the same.
The search function is now parameterize for 8-bit and 16-bit state
sizes and alphabet sizes 8, 16, 32, 64, 128 and 256.
12 years ago
|
|
|
/* Number of characters in the compressed alphabet. */
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
uint16_t alphabet_size;
|
Add 8-bit states to ac-tile
When running with sgh-mpm-context: full, many more MPMs are created
(16K) and many are small. If they have less than 128 states, they only
need 1 byte for the next state instead of 2 bytes, cutting the size of
the next-state table in half. This reduces total memory usage.
Since that makes 3 different state sizes (1, 2 and 4 bytes), rather
than going from 2 copies of the code to create the MPM to 3, I
factored out the code that fills the next-state table into three
functions so that all the other code could be the same.
The search function is now parameterize for 8-bit and 16-bit state
sizes and alphabet sizes 8, 16, 32, 64, 128 and 256.
12 years ago
|
|
|
/* Space used to store compressed alphabet is the next
|
|
|
|
|
* larger or equal power-of-2.
|
|
|
|
|
*/
|
|
|
|
|
uint16_t alphabet_storage;
|
|
|
|
|
|
|
|
|
|
/* How many bytes are used to store the next state. */
|
|
|
|
|
uint8_t bytes_per_state;
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
|
|
|
|
|
} SCACTileCtx;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/* Only the stuff used at search time. This
|
|
|
|
|
* structure is created after all the patterns are added.
|
|
|
|
|
*/
|
|
|
|
|
typedef struct SCACTileSearchCtx_ {
|
|
|
|
|
|
|
|
|
|
/* Specialized search function based on size of data in delta
|
|
|
|
|
* tables. The alphabet size determines address shifting and the
|
|
|
|
|
* number of states could make the next state could be 16 bits or
|
|
|
|
|
* 32 bits.
|
|
|
|
|
*/
|
|
|
|
|
uint32_t (*Search)(const struct SCACTileSearchCtx_ *ctx, struct MpmThreadCtx_ *,
|
|
|
|
|
PrefilterRuleStore *, const uint8_t *, uint16_t);
|
|
|
|
|
|
|
|
|
|
/* Convert input character to matching alphabet */
|
|
|
|
|
uint8_t translate_table[256];
|
|
|
|
|
|
|
|
|
|
/* the all important memory hungry state_table */
|
Add 8-bit states to ac-tile
When running with sgh-mpm-context: full, many more MPMs are created
(16K) and many are small. If they have less than 128 states, they only
need 1 byte for the next state instead of 2 bytes, cutting the size of
the next-state table in half. This reduces total memory usage.
Since that makes 3 different state sizes (1, 2 and 4 bytes), rather
than going from 2 copies of the code to create the MPM to 3, I
factored out the code that fills the next-state table into three
functions so that all the other code could be the same.
The search function is now parameterize for 8-bit and 16-bit state
sizes and alphabet sizes 8, 16, 32, 64, 128 and 256.
12 years ago
|
|
|
void *state_table;
|
|
|
|
|
|
In AC-Tile, convert from using pids for indexing to pattern index
Use an MPM specific pattern index, which is simply an index starting
at zero and incremented for each pattern added to the MPM, rather than
the externally provided Pattern ID (pid), since that can be much
larger than the number of patterns. The Pattern ID is shared across at
MPMs. For example, an MPM with one pattern with pid=8000 would result
in a max_pid of 8000, so the pid_pat_list would have 8000 entries.
The pid_pat_list[] is replaced by a array of pattern indexes. The PID is
moved to the SCACTilePatternList as a single value. The PatternList is
also indexed by the Pattern Index.
max_pat_id is no longer needed and mpm_ctx->pattern_cnt is used instead.
The local bitarray is then also indexed by pattern index instead of PID, making
it much smaller. The local bit array sets a bit for each pattern found
for this MPM. It is only kept during one MPM search (stack allocated).
One note, the local bit array is checked first and if the pattern has already
been found, it will stop checking, but count a match. This could result in
over counting matches of case-sensitve matches, since following case-insensitive
matches will also be counted. For example, finding "Foo" in "foo Foo foo" would
report finding "Foo" 2 times, mis-counting the third word as "Foo".
11 years ago
|
|
|
/* List of patterns that match for this state. Indexed by State Number */
|
|
|
|
|
SCACTileOutputTable *output_table;
|
In AC-Tile, convert from using pids for indexing to pattern index
Use an MPM specific pattern index, which is simply an index starting
at zero and incremented for each pattern added to the MPM, rather than
the externally provided Pattern ID (pid), since that can be much
larger than the number of patterns. The Pattern ID is shared across at
MPMs. For example, an MPM with one pattern with pid=8000 would result
in a max_pid of 8000, so the pid_pat_list would have 8000 entries.
The pid_pat_list[] is replaced by a array of pattern indexes. The PID is
moved to the SCACTilePatternList as a single value. The PatternList is
also indexed by the Pattern Index.
max_pat_id is no longer needed and mpm_ctx->pattern_cnt is used instead.
The local bitarray is then also indexed by pattern index instead of PID, making
it much smaller. The local bit array sets a bit for each pattern found
for this MPM. It is only kept during one MPM search (stack allocated).
One note, the local bit array is checked first and if the pattern has already
been found, it will stop checking, but count a match. This could result in
over counting matches of case-sensitve matches, since following case-insensitive
matches will also be counted. For example, finding "Foo" in "foo Foo foo" would
report finding "Foo" 2 times, mis-counting the third word as "Foo".
11 years ago
|
|
|
SCACTilePatternList *pattern_list;
|
|
|
|
|
|
|
|
|
|
/* Number of bytes in the array of bits. One bit per pattern in this MPM. */
|
|
|
|
|
uint32_t mpm_bitarray_size;
|
|
|
|
|
|
|
|
|
|
/* Number of states used by ac-tile */
|
|
|
|
|
uint32_t state_count;
|
|
|
|
|
|
|
|
|
|
uint32_t pattern_cnt;
|
|
|
|
|
|
Add 8-bit states to ac-tile
When running with sgh-mpm-context: full, many more MPMs are created
(16K) and many are small. If they have less than 128 states, they only
need 1 byte for the next state instead of 2 bytes, cutting the size of
the next-state table in half. This reduces total memory usage.
Since that makes 3 different state sizes (1, 2 and 4 bytes), rather
than going from 2 copies of the code to create the MPM to 3, I
factored out the code that fills the next-state table into three
functions so that all the other code could be the same.
The search function is now parameterize for 8-bit and 16-bit state
sizes and alphabet sizes 8, 16, 32, 64, 128 and 256.
12 years ago
|
|
|
/* MPM Creation data, only used at initialization. */
|
|
|
|
|
SCACTileCtx *init_ctx;
|
|
|
|
|
|
|
|
|
|
} SCACTileSearchCtx;
|
|
|
|
|
|
|
|
|
|
|
New Multi-pattern matcher, ac-tile, optimized for Tile architecture.
Aho-Corasick mpm optimized for Tilera Tile-Gx architecture. Based on the
util-mpm-ac.c code base. The primary optimizations are:
1) Matching function used Tilera specific instructions.
2) Alphabet compression to reduce delta table size to increase cache
utilization and performance.
The basic observation is that not all 256 ASCII characters are used by
the set of multiple patterns in a group for which a DFA is
created. The first reason is that Suricata's pattern matching is
case-insensitive, so all uppercase characters are converted to
lowercase, leaving a hole of 26 characters in the
alphabet. Previously, this hole was simply left in the middle of the
alphabet and thus in the generated Next State (delta) tables.
A new, smaller, alphabet is created using a translation table of 256
bytes per mpm group. Previously, there was one global translation
table for converting upper case to lowercase.
Additional, unused characters are found by creating a histogram of all
the characters in all the patterns. Then all the characters with zero
counts are mapped to one character (0) in the new alphabet. Since
These characters appear in no pattern, they can all be mapped to a
single character and still result in the same matches being
found. Zero was chosen for the value in the new alphabet since this
"character" is more likely to appear in the input. The unused
character always results in the next state being state zero, but that
fact is not currently used by the code, since special casing takes
additional instructions.
The characters that do appear in some pattern are mapped to
consecutive characters in the new alphabet, starting at 1. This
results in a dense packing of next state values in the delta tables
and additionally can allow for a smaller number of columns in that
table, thus using less memory and better packing into the cache. The
size of the new alphabet is the number of used characters plus 1 for
the unused catch-all character.
The alphabet size is rounded up to the next larger power-of-2 so that
multiplication by the alphabet size can be done with a shift. It
might be possible to use a multiply instruction, so that the exact
alphabet size could be used, which would further reduce the size of
the delta tables, increase cache density and not require the
specialized search functions. The multiply would likely add 1 cycle to
the inner search loop.
Since the multiply by alphabet-size is cleverly merged with a mask
instruction (in the SINDEX macro), specialized versions of the
SCACSearch function are generated for alphabet sizes 256, 128, 64, 32
and 16. This is done by including the file util-mpm-ac-small.c
multiple times with a redefined SINDEX macro. A function pointer is
then stored in the mpm context for the search function. For alpha bit
sizes of 8 or smaller, the number of states usually small, so the DFA
is already very small, so there is little difference using the 16
state search function.
The SCACSearch function is also specialized by the size of the value
stored in the next state (delta) tables, either 16-bits or 32-bits.
This removes a conditional inside the Search function. That
conditional is only called once, but doesn't hurt to remove
it. 16-bits are used for up to 32K states, with the sign bit set for
states with matches.
Future optimization:
The state-has-match values is only needed per state, not per next
state, so checking the next-state sign bit could be replaced with
reading a different value, at the cost of an additional load, but
increasing the 16-bit next state span to 64K.
Since the order of the characters in the new alphabet doesn't matter,
the new alphabet could be sorted by the frequency of the characters in
the expected input stream for that multi-pattern matcher. This would
group more frequent characters into the same cache lines, thus
increasing the probability of reusing a cache-line.
All the next state values for each state live in their own set of
cache-lines. With power-of-two sizes alphabets, these don't overlap.
So either 32 or 16 character's next states are loaded in each cache
line load. If the alphabet size is not an exact power-of-2, then the
last cache-line is not completely full and up to 31*2 bytes of that
line could be wasted per state.
The next state table could be transposed, so that all the next states
for a specific character are stored sequentially, this could be better
if some characters, for example the unused character, are much more
frequent.
12 years ago
|
|
|
typedef struct SCACTileThreadCtx_ {
|
|
|
|
|
/* the total calls we make to the search function */
|
|
|
|
|
uint32_t total_calls;
|
|
|
|
|
/* the total patterns that we ended up matching against */
|
|
|
|
|
uint64_t total_matches;
|
|
|
|
|
} SCACTileThreadCtx;
|
|
|
|
|
|
|
|
|
|
void MpmACTileRegister(void);
|
|
|
|
|
|
|
|
|
|
#endif /* __UTIL_MPM_AC_TILE__H__ */
|
