shaba/lvm2
1
0
Fork 0
mirror of git://sourceware.org/git/lvm2.git synced 2024-12-22 17:35:59 +03:00
Code Releases Activity
f6524657fa
lvm2/lib/regex/matcher.c

350 lines
6.7 KiB
C
Raw Normal View History

o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
/*
* Copyright (C) 2001 Sistina Software (UK) Limited.
*
o LGPL list implementation
2001-10-31 15:47:01 +03:00
* This file is released under the LGPL.
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
*/
#include "matcher.h"
#include "parse_rx.h"
#include "log.h"
#include "ttree.h"
#include "bitset.h"
#include <string.h>
#include <stdio.h>
#include <assert.h>
struct dfa_state {
int final;
struct dfa_state *lookup[256];
};
struct state_queue {
struct dfa_state *s;
bitset_t bits;
struct state_queue *next;
};
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
struct matcher { /* Instance variables for the lexer */
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
struct dfa_state *start;
int num_nodes, nodes_entered;
struct rx_node **nodes;
struct pool *scratch, *mem;
};
#define TARGET_TRANS '\0'
static int _count_nodes(struct rx_node *rx)
{
int r = 1;
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (rx->left)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
r += _count_nodes(rx->left);
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (rx->right)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
r += _count_nodes(rx->right);
return r;
}
static void _fill_table(struct matcher *m, struct rx_node *rx)
{
assert((rx->type != OR) || (rx->left && rx->right));
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (rx->left)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
_fill_table(m, rx->left);
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (rx->right)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
_fill_table(m, rx->right);
m->nodes[m->nodes_entered++] = rx;
}
static void _create_bitsets(struct matcher *m)
{
int i;
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
for (i = 0; i < m->num_nodes; i++) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
struct rx_node *n = m->nodes[i];
n->firstpos = bitset_create(m->scratch, m->num_nodes);
n->lastpos = bitset_create(m->scratch, m->num_nodes);
n->followpos = bitset_create(m->scratch, m->num_nodes);
}
}
static void _calc_functions(struct matcher *m)
{
int i, j, final = 1;
struct rx_node *rx, *c1, *c2;
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
for (i = 0; i < m->num_nodes; i++) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
rx = m->nodes[i];
c1 = rx->left;
c2 = rx->right;
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (bit(rx->charset, TARGET_TRANS))
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
rx->final = final++;
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
switch (rx->type) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
case CAT:
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (c1->nullable)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
bit_union(rx->firstpos,
c1->firstpos, c2->firstpos);
else
bit_copy(rx->firstpos, c1->firstpos);
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (c2->nullable)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
bit_union(rx->lastpos,
c1->lastpos, c2->lastpos);
else
bit_copy(rx->lastpos, c2->lastpos);
rx->nullable = c1->nullable && c2->nullable;
break;
case PLUS:
bit_copy(rx->firstpos, c1->firstpos);
bit_copy(rx->lastpos, c1->lastpos);
rx->nullable = c1->nullable;
break;
case OR:
bit_union(rx->firstpos, c1->firstpos, c2->firstpos);
bit_union(rx->lastpos, c1->lastpos, c2->lastpos);
rx->nullable = c1->nullable || c2->nullable;
break;
case QUEST:
case STAR:
bit_copy(rx->firstpos, c1->firstpos);
bit_copy(rx->lastpos, c1->lastpos);
rx->nullable = 1;
break;
case CHARSET:
bit_set(rx->firstpos, i);
bit_set(rx->lastpos, i);
rx->nullable = 0;
break;
default:
Tweak some error message levels.
2002-01-28 00:30:47 +03:00
log_error("Internal error: Unknown calc node type");
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
}
/*
* followpos has it's own switch
* because PLUS and STAR do the
* same thing.
*/
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
switch (rx->type) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
case CAT:
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
for (j = 0; j < m->num_nodes; j++) {
if (bit(c1->lastpos, j)) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
struct rx_node *n = m->nodes[j];
bit_union(n->followpos,
n->followpos, c2->firstpos);
}
}
break;
case PLUS:
case STAR:
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
for (j = 0; j < m->num_nodes; j++) {
if (bit(rx->lastpos, j)) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
struct rx_node *n = m->nodes[j];
bit_union(n->followpos,
n->followpos, rx->firstpos);
}
}
break;
}
}
}
static inline struct dfa_state *_create_dfa_state(struct pool *mem)
{
return pool_zalloc(mem, sizeof(struct dfa_state));
}
static struct state_queue *_create_state_queue(struct pool *mem,
struct dfa_state *dfa,
bitset_t bits)
{
struct state_queue *r = pool_alloc(mem, sizeof(*r));
if (!r) {
stack;
return NULL;
}
r->s = dfa;
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
r->bits = bitset_create(mem, bits[0]); /* first element is the size */
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
bit_copy(r->bits, bits);
r->next = 0;
return r;
}
static int _calc_states(struct matcher *m, struct rx_node *rx)
{
int iwidth = (m->num_nodes / BITS_PER_INT) + 1;
struct ttree *tt = ttree_create(m->scratch, iwidth);
struct state_queue *h, *t, *tmp;
struct dfa_state *dfa, *ldfa;
int i, a, set_bits = 0, count = 0;
bitset_t bs = bitset_create(m->scratch, m->num_nodes), dfa_bits;
if (!tt) {
stack;
return 0;
}
if (!bs) {
stack;
return 0;
}
/* create first state */
dfa = _create_dfa_state(m->mem);
m->start = dfa;
o Filter for the dev cache that takes values from config file: devices { # first match is final, eg. /dev/ide/cdrom # get's rejected due to the first pattern filter=["r/cdrom/", # don't touch the music ! "a/hd[a-d][0-9]+/", "a/ide/", "a/sd/", "a/md/", "a|loop/[0-9]+|", # accept devfs style loop back "r/loop/", # and reject old style "a/dasd/", "a/dac960/", "a/nbd/", "a/ida/", "a/cciss/", "a/ubd/", "r/.*/"] # reject all others } Alasdair this is ready to roll into the tools now.
2001-10-19 22:20:37 +04:00
ttree_insert(tt, rx->firstpos + 1, dfa);
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
/* prime the queue */
h = t = _create_state_queue(m->scratch, dfa, rx->firstpos);
o Filter for the dev cache that takes values from config file: devices { # first match is final, eg. /dev/ide/cdrom # get's rejected due to the first pattern filter=["r/cdrom/", # don't touch the music ! "a/hd[a-d][0-9]+/", "a/ide/", "a/sd/", "a/md/", "a|loop/[0-9]+|", # accept devfs style loop back "r/loop/", # and reject old style "a/dasd/", "a/dac960/", "a/nbd/", "a/ida/", "a/cciss/", "a/ubd/", "r/.*/"] # reject all others } Alasdair this is ready to roll into the tools now.
2001-10-19 22:20:37 +04:00
while (h) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
/* pop state off front of the queue */
dfa = h->s;
dfa_bits = h->bits;
h = h->next;
/* iterate through all the inputs for this state */
bit_clear_all(bs);
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
for (a = 0; a < 256; a++) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
/* iterate through all the states in firstpos */
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
for (i = bit_get_first(dfa_bits);
i >= 0; i = bit_get_next(dfa_bits, i)) {
if (bit(m->nodes[i]->charset, a)) {
if (a == TARGET_TRANS)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
dfa->final = m->nodes[i]->final;
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
bit_union(bs, bs,
m->nodes[i]->followpos);
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
set_bits = 1;
}
}
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (set_bits) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
ldfa = ttree_lookup(tt, bs + 1);
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (!ldfa) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
/* push */
ldfa = _create_dfa_state(m->mem);
ttree_insert(tt, bs + 1, ldfa);
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
tmp =
_create_state_queue(m->scratch,
ldfa, bs);
if (!h)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
h = t = tmp;
else {
t->next = tmp;
t = tmp;
}
count++;
}
dfa->lookup[a] = ldfa;
set_bits = 0;
bit_clear_all(bs);
}
}
}
o Filter for the dev cache that takes values from config file: devices { # first match is final, eg. /dev/ide/cdrom # get's rejected due to the first pattern filter=["r/cdrom/", # don't touch the music ! "a/hd[a-d][0-9]+/", "a/ide/", "a/sd/", "a/md/", "a|loop/[0-9]+|", # accept devfs style loop back "r/loop/", # and reject old style "a/dasd/", "a/dac960/", "a/nbd/", "a/ida/", "a/cciss/", "a/ubd/", "r/.*/"] # reject all others } Alasdair this is ready to roll into the tools now.
2001-10-19 22:20:37 +04:00
log_debug("Matcher built with %d dfa states", count);
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
return 1;
}
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
struct matcher *matcher_create(struct pool *mem, const char **patterns, int num)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
{
char *all, *ptr;
int i, len = 0;
struct rx_node *rx;
struct pool *scratch = pool_create(10 * 1024);
struct matcher *m;
if (!scratch) {
stack;
return NULL;
}
if (!(m = pool_alloc(mem, sizeof(*m)))) {
stack;
return NULL;
}
memset(m, 0, sizeof(*m));
/* join the regexps together, delimiting with zero */
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
for (i = 0; i < num; i++)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
len += strlen(patterns[i]) + 8;
ptr = all = pool_alloc(scratch, len + 1);
if (!all) {
stack;
goto bad;
}
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
for (i = 0; i < num; i++) {
ptr += sprintf(ptr, "(.*(%s)%c)", patterns[i], TARGET_TRANS);
if (i < (num - 1))
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
*ptr++ = '|';
}
/* parse this expression */
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
if (!(rx = rx_parse_tok(scratch, all, ptr))) {
Tweak some error message levels.
2002-01-28 00:30:47 +03:00
log_error("Couldn't parse regex");
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
goto bad;
}
m->mem = mem;
m->scratch = scratch;
m->num_nodes = _count_nodes(rx);
m->nodes = pool_alloc(scratch, sizeof(*m->nodes) * m->num_nodes);
if (!m->nodes) {
stack;
goto bad;
}
_fill_table(m, rx);
_create_bitsets(m);
_calc_functions(m);
_calc_states(m, rx);
pool_destroy(scratch);
m->scratch = NULL;
return m;
Merge with text format branch. Lots of changes/very little testing so far => there'll be bugs! Use 'vgcreate -M text' to create a volume group with its metadata stored in text files. Text format metadata changes should be reasonably atomic, with a (basic) automatic recovery mechanism if the system crashes while a change is in progress. Add a metadata section to lvm.conf to specify multiple directories if you want (recommended) to keep multiple copies of the metadata (eg on different filesystems). e.g. metadata { dirs = ["/etc/lvm/metadata1","/usr/local/lvm/metadata2"] } Plenty of refinements still in the pipeline.
2002-04-24 22:20:51 +04:00
bad:
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
pool_destroy(scratch);
pool_destroy(mem);
return NULL;
}
o Filter for the dev cache that takes values from config file: devices { # first match is final, eg. /dev/ide/cdrom # get's rejected due to the first pattern filter=["r/cdrom/", # don't touch the music ! "a/hd[a-d][0-9]+/", "a/ide/", "a/sd/", "a/md/", "a|loop/[0-9]+|", # accept devfs style loop back "r/loop/", # and reject old style "a/dasd/", "a/dac960/", "a/nbd/", "a/ida/", "a/cciss/", "a/ubd/", "r/.*/"] # reject all others } Alasdair this is ready to roll into the tools now.
2001-10-19 22:20:37 +04:00
int matcher_run(struct matcher *m, const char *b)
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
{
struct dfa_state *cs = m->start;
o Filter for the dev cache that takes values from config file: devices { # first match is final, eg. /dev/ide/cdrom # get's rejected due to the first pattern filter=["r/cdrom/", # don't touch the music ! "a/hd[a-d][0-9]+/", "a/ide/", "a/sd/", "a/md/", "a|loop/[0-9]+|", # accept devfs style loop back "r/loop/", # and reject old style "a/dasd/", "a/dac960/", "a/nbd/", "a/ida/", "a/cciss/", "a/ubd/", "r/.*/"] # reject all others } Alasdair this is ready to roll into the tools now.
2001-10-19 22:20:37 +04:00
int r = 0;
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
o Filter for the dev cache that takes values from config file: devices { # first match is final, eg. /dev/ide/cdrom # get's rejected due to the first pattern filter=["r/cdrom/", # don't touch the music ! "a/hd[a-d][0-9]+/", "a/ide/", "a/sd/", "a/md/", "a|loop/[0-9]+|", # accept devfs style loop back "r/loop/", # and reject old style "a/dasd/", "a/dac960/", "a/nbd/", "a/ida/", "a/cciss/", "a/ubd/", "r/.*/"] # reject all others } Alasdair this is ready to roll into the tools now.
2001-10-19 22:20:37 +04:00
for (; *b; b++) {
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
if (!(cs = cs->lookup[(int) (unsigned char) *b]))
break;
o tidying
2001-10-21 14:24:10 +04:00
if (cs->final && (cs->final > r))
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
r = cs->final;
}
/* subtract 1 to get back to zero index */
o Filter for the dev cache that takes values from config file: devices { # first match is final, eg. /dev/ide/cdrom # get's rejected due to the first pattern filter=["r/cdrom/", # don't touch the music ! "a/hd[a-d][0-9]+/", "a/ide/", "a/sd/", "a/md/", "a|loop/[0-9]+|", # accept devfs style loop back "r/loop/", # and reject old style "a/dasd/", "a/dac960/", "a/nbd/", "a/ida/", "a/cciss/", "a/ubd/", "r/.*/"] # reject all others } Alasdair this is ready to roll into the tools now.
2001-10-19 22:20:37 +04:00
return r - 1;
o First pass at the regex code. lib/regex/matcher takes an array of regex's and builds a *very* efficient engine that will tell you which regex a string matches with only a single pass through the string. To be used in the config file when specifying devices. o Anchor's aren't supported yet (^ and $) but that won't take long. o Also when we get some realistic config files we may want to consider adding an extra level of indirection to the dfa state in order to compress the table. It all depends on how large typical tables get.
2001-10-19 18:36:57 +04:00
}
Copy Permalink