Jerry/bird
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Filter: Converted static global variables to a filter_state struct.

Browse source 
The static filter state was messy and blocked the planned parallel
execution of filters. Anyway, this will be also slower as the state
structure must be passed almost everywhere with us.
This commit is contained in:
Jan Maria Matejka 2018-02-27 15:39:39 +01:00 committed by Maria Matejka
parent ae294cc2d0
commit a946317fab
1 changed files with 96 additions and 93 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

189
filter/filter.c
View file

@ -50,6 +50,16 @@
#define CMP_ERROR 999 #define CMP_ERROR 999
/* Internal filter state, to be allocated on stack when executing filters */
struct filter_state {
struct rte **rte;
struct rta *old_rta;
struct ea_list **eattrs;
struct linpool *pool;
struct buffer buf;
int flags;
};
void (*bt_assert_hook)(int result, struct f_inst *assert); void (*bt_assert_hook)(int result, struct f_inst *assert);
static struct adata undef_adata; /* adata of length 0 used for undefined */ static struct adata undef_adata; /* adata of length 0 used for undefined */
@ -536,65 +546,59 @@ val_format(struct f_val v, buffer *buf)
} }
} }
static struct rte **f_rte;
static struct rta *f_old_rta;
static struct ea_list **f_eattrs;
static struct linpool *f_pool;
static struct buffer f_buf;
static int f_flags;
static inline void f_cache_eattrs(void) static inline void f_cache_eattrs(struct filter_state *fs)
{ {
f_eattrs = &((*f_rte)->attrs->eattrs); fs->eattrs = &((*fs->rte)->attrs->eattrs);
} }
static inline void f_rte_cow(void) static inline void f_rte_cow(struct filter_state *fs)
{ {
if (!((*f_rte)->flags & REF_COW)) if (!((*fs->rte)->flags & REF_COW))
return; return;
*f_rte = rte_do_cow(*f_rte); *fs->rte = rte_cow(*fs->rte);
} }
/* /*
* rta_cow - prepare rta for modification by filter * rta_cow - prepare rta for modification by filter
*/ */
static void static void
f_rta_cow(void) f_rta_cow(struct filter_state *fs)
{ {
if (!rta_is_cached((*f_rte)->attrs)) if (!rta_is_cached((*fs->rte)->attrs))
return; return;
/* Prepare to modify rte */ /* Prepare to modify rte */
f_rte_cow(); f_rte_cow(fs);
/* Store old rta to free it later, it stores reference from rte_cow() */ /* Store old rta to free it later, it stores reference from rte_cow() */
f_old_rta = (*f_rte)->attrs; fs->old_rta = (*fs->rte)->attrs;
/* /*
* Get shallow copy of rta. Fields eattrs and nexthops of rta are shared * Get shallow copy of rta. Fields eattrs and nexthops of rta are shared
* with f_old_rta (they will be copied when the cached rta will be obtained * with fs->old_rta (they will be copied when the cached rta will be obtained
* at the end of f_run()), also the lock of hostentry is inherited (we * at the end of f_run()), also the lock of hostentry is inherited (we
* suppose hostentry is not changed by filters). * suppose hostentry is not changed by filters).
*/ */
(*f_rte)->attrs = rta_do_cow((*f_rte)->attrs, f_pool); (*fs->rte)->attrs = rta_do_cow((*fs->rte)->attrs, fs->pool);
/* Re-cache the ea_list */ /* Re-cache the ea_list */
f_cache_eattrs(); f_cache_eattrs(fs);
} }
static char * static char *
val_format_str(struct f_val v) { val_format_str(struct filter_state *fs, struct f_val v) {
buffer b; buffer b;
LOG_BUFFER_INIT(b); LOG_BUFFER_INIT(b);
val_format(v, &b); val_format(v, &b);
return lp_strdup(f_pool, b.start); return lp_strdup(fs->pool, b.start);
} }
static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS; static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
#define runtime(fmt, ...) do { \ #define runtime(fmt, ...) do { \
if (!(f_flags & FF_SILENT)) \ if (!(fs->flags & FF_SILENT)) \
log_rl(&rl_runtime_err, L_ERR "filters, line %d: " fmt, what->lineno, ##__VA_ARGS__); \ log_rl(&rl_runtime_err, L_ERR "filters, line %d: " fmt, what->lineno, ##__VA_ARGS__); \
res.type = T_RETURN; \ res.type = T_RETURN; \
res.val.i = F_ERROR; \ res.val.i = F_ERROR; \
@ -609,21 +613,22 @@ static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
n, f_instruction_name(what->fi_code), t, v##n.type); n, f_instruction_name(what->fi_code), t, v##n.type);
#define INTERPRET(val, what_) \ #define INTERPRET(val, what_) \
val = interpret(what_); \ val = interpret(fs, what_); \
if (val.type & T_RETURN) \ if (val.type & T_RETURN) \
return val; return val;
#define ACCESS_RTE \ #define ACCESS_RTE \
do { if (!f_rte) runtime("No route to access"); } while (0) do { if (!fs->rte) runtime("No route to access"); } while (0)
#define ACCESS_EATTRS \ #define ACCESS_EATTRS \
do { if (!f_eattrs) f_cache_eattrs(); } while (0) do { if (!fs->eattrs) f_cache_eattrs(fs); } while (0)
#define BITFIELD_MASK(what) \ #define BITFIELD_MASK(what) \
(1u << (what->a2.i >> 24)) (1u << (what->a2.i >> 24))
/** /**
* interpret * interpret
* @fs: filter state
* @what: filter to interpret * @what: filter to interpret
* *
* Interpret given tree of filter instructions. This is core function * Interpret given tree of filter instructions. This is core function
@ -640,7 +645,7 @@ static struct tbf rl_runtime_err = TBF_DEFAULT_LOG_LIMITS;
* memory managment. * memory managment.
*/ */
static struct f_val static struct f_val
interpret(struct f_inst *what) interpret(struct filter_state *fs, struct f_inst *what)
{ {
struct symbol *sym; struct symbol *sym;
struct f_val v1, v2, v3, res = { .type = T_VOID }, *vp; struct f_val v1, v2, v3, res = { .type = T_VOID }, *vp;
@ -761,7 +766,7 @@ interpret(struct f_inst *what)
struct f_path_mask *tt = what->a1.p, *vbegin, **vv = &vbegin; struct f_path_mask *tt = what->a1.p, *vbegin, **vv = &vbegin;
while (tt) { while (tt) {
*vv = lp_alloc(f_pool, sizeof(struct f_path_mask)); *vv = lp_alloc(fs->pool, sizeof(struct f_path_mask));
if (tt->kind == PM_ASN_EXPR) { if (tt->kind == PM_ASN_EXPR) {
struct f_val res; struct f_val res;
INTERPRET(res, (struct f_inst *) tt->val); INTERPRET(res, (struct f_inst *) tt->val);
@ -895,7 +900,7 @@ interpret(struct f_inst *what)
break; break;
case FI_PRINT: case FI_PRINT:
ARG_ANY(1); ARG_ANY(1);
val_format(v1, &f_buf); val_format(v1, &fs->buf);
break; break;
case FI_CONDITION: /* ? has really strange error value, so we can implement if ... else nicely :-) */ case FI_CONDITION: /* ? has really strange error value, so we can implement if ... else nicely :-) */
ARG(1, T_BOOL); ARG(1, T_BOOL);
@ -912,8 +917,8 @@ interpret(struct f_inst *what)
case FI_PRINT_AND_DIE: case FI_PRINT_AND_DIE:
ARG_ANY(1); ARG_ANY(1);
if ((what->a2.i == F_NOP || (what->a2.i != F_NONL && what->a1.p)) && if ((what->a2.i == F_NOP || (what->a2.i != F_NONL && what->a1.p)) &&
!(f_flags & FF_SILENT)) !(fs->flags & FF_SILENT))
log_commit(*L_INFO, &f_buf); log_commit(*L_INFO, &fs->buf);
switch (what->a2.i) { switch (what->a2.i) {
case F_QUITBIRD: case F_QUITBIRD:
@ -935,14 +940,14 @@ interpret(struct f_inst *what)
case FI_RTA_GET: /* rta access */ case FI_RTA_GET: /* rta access */
{ {
ACCESS_RTE; ACCESS_RTE;
struct rta *rta = (*f_rte)->attrs; struct rta *rta = (*fs->rte)->attrs;
res.type = what->aux; res.type = what->aux;
switch (what->a2.i) switch (what->a2.i)
{ {
case SA_FROM: res.val.ip = rta->from; break; case SA_FROM: res.val.ip = rta->from; break;
case SA_GW: res.val.ip = rta->nh.gw; break; case SA_GW: res.val.ip = rta->nh.gw; break;
case SA_NET: res.val.net = (*f_rte)->net->n.addr; break; case SA_NET: res.val.net = (*fs->rte)->net->n.addr; break;
case SA_PROTO: res.val.s = rta->src->proto->name; break; case SA_PROTO: res.val.s = rta->src->proto->name; break;
case SA_SOURCE: res.val.i = rta->source; break; case SA_SOURCE: res.val.i = rta->source; break;
case SA_SCOPE: res.val.i = rta->scope; break; case SA_SCOPE: res.val.i = rta->scope; break;
@ -961,9 +966,9 @@ interpret(struct f_inst *what)
if (what->aux != v1.type) if (what->aux != v1.type)
runtime( "Attempt to set static attribute to incompatible type" ); runtime( "Attempt to set static attribute to incompatible type" );
f_rta_cow(); f_rta_cow(fs);
{ {
struct rta *rta = (*f_rte)->attrs; struct rta *rta = (*fs->rte)->attrs;
switch (what->a2.i) switch (what->a2.i)
{ {
@ -1027,7 +1032,7 @@ interpret(struct f_inst *what)
{ {
u16 code = what->a2.i; u16 code = what->a2.i;
int f_type = what->aux >> 8; int f_type = what->aux >> 8;
eattr *e = ea_find(*f_eattrs, code); eattr *e = ea_find(*fs->eattrs, code);
if (!e) { if (!e) {
/* A special case: undefined as_path looks like empty as_path */ /* A special case: undefined as_path looks like empty as_path */
@ -1114,7 +1119,7 @@ interpret(struct f_inst *what)
ACCESS_EATTRS; ACCESS_EATTRS;
ARG_ANY(1); ARG_ANY(1);
{ {
struct ea_list *l = lp_alloc(f_pool, sizeof(struct ea_list) + sizeof(eattr)); struct ea_list *l = lp_alloc(fs->pool, sizeof(struct ea_list) + sizeof(eattr));
u16 code = what->a2.i; u16 code = what->a2.i;
int f_type = what->aux >> 8; int f_type = what->aux >> 8;
@ -1151,7 +1156,7 @@ interpret(struct f_inst *what)
if (v1.type != T_IP) if (v1.type != T_IP)
runtime( "Setting ip attribute to non-ip value" ); runtime( "Setting ip attribute to non-ip value" );
int len = sizeof(ip_addr); int len = sizeof(ip_addr);
struct adata *ad = lp_alloc(f_pool, sizeof(struct adata) + len); struct adata *ad = lp_alloc(fs->pool, sizeof(struct adata) + len);
ad->length = len; ad->length = len;
(* (ip_addr *) ad->data) = v1.val.ip; (* (ip_addr *) ad->data) = v1.val.ip;
l->attrs[0].u.ptr = ad; l->attrs[0].u.ptr = ad;
@ -1166,7 +1171,7 @@ interpret(struct f_inst *what)
runtime( "Setting bit in bitfield attribute to non-bool value" ); runtime( "Setting bit in bitfield attribute to non-bool value" );
{ {
/* First, we have to find the old value */ /* First, we have to find the old value */
eattr *e = ea_find(*f_eattrs, code); eattr *e = ea_find(*fs->eattrs, code);
u32 data = e ? e->u.data : 0; u32 data = e ? e->u.data : 0;
if (v1.val.i) if (v1.val.i)
@ -1198,23 +1203,23 @@ interpret(struct f_inst *what)
default: bug("Unknown type in e,S"); default: bug("Unknown type in e,S");
} }
f_rta_cow(); f_rta_cow(fs);
l->next = *f_eattrs; l->next = *fs->eattrs;
*f_eattrs = l; *fs->eattrs = l;
} }
break; break;
case FI_PREF_GET: case FI_PREF_GET:
ACCESS_RTE; ACCESS_RTE;
res.type = T_INT; res.type = T_INT;
res.val.i = (*f_rte)->pref; res.val.i = (*fs->rte)->pref;
break; break;
case FI_PREF_SET: case FI_PREF_SET:
ACCESS_RTE; ACCESS_RTE;
ARG(1,T_INT); ARG(1,T_INT);
if (v1.val.i > 0xFFFF) if (v1.val.i > 0xFFFF)
runtime( "Setting preference value out of bounds" ); runtime( "Setting preference value out of bounds" );
f_rte_cow(); f_rte_cow(fs);
(*f_rte)->pref = v1.val.i; (*fs->rte)->pref = v1.val.i;
break; break;
case FI_LENGTH: /* Get length of */ case FI_LENGTH: /* Get length of */
ARG_ANY(1); ARG_ANY(1);
@ -1235,7 +1240,7 @@ interpret(struct f_inst *what)
{ {
net_addr_ip6_sadr *net = (void *) v1.val.net; net_addr_ip6_sadr *net = (void *) v1.val.net;
net_addr *src = lp_alloc(f_pool, sizeof(net_addr_ip6)); net_addr *src = lp_alloc(fs->pool, sizeof(net_addr_ip6));
net_fill_ip6(src, net->src_prefix, net->src_pxlen); net_fill_ip6(src, net->src_prefix, net->src_pxlen);
res.type = T_NET; res.type = T_NET;
@ -1304,7 +1309,7 @@ interpret(struct f_inst *what)
return res; return res;
case FI_CALL: /* CALL: this is special: if T_RETURN and returning some value, mask it out */ case FI_CALL: /* CALL: this is special: if T_RETURN and returning some value, mask it out */
ARG_ANY(1); ARG_ANY(1);
res = interpret(what->a2.p); res = interpret(fs, what->a2.p);
if (res.type == T_RETURN) if (res.type == T_RETURN)
return res; return res;
res.type &= ~T_RETURN; res.type &= ~T_RETURN;
@ -1342,14 +1347,14 @@ interpret(struct f_inst *what)
case FI_EMPTY: /* Create empty attribute */ case FI_EMPTY: /* Create empty attribute */
res.type = what->aux; res.type = what->aux;
res.val.ad = adata_empty(f_pool, 0); res.val.ad = adata_empty(fs->pool, 0);
break; break;
case FI_PATH_PREPEND: /* Path prepend */ case FI_PATH_PREPEND: /* Path prepend */
ARG(1, T_PATH); ARG(1, T_PATH);
ARG(2, T_INT); ARG(2, T_INT);
res.type = T_PATH; res.type = T_PATH;
res.val.ad = as_path_prepend(f_pool, v1.val.ad, v2.val.i); res.val.ad = as_path_prepend(fs->pool, v1.val.ad, v2.val.i);
break; break;
case FI_CLIST_ADD_DEL: /* (Extended) Community list add or delete */ case FI_CLIST_ADD_DEL: /* (Extended) Community list add or delete */
@ -1380,7 +1385,7 @@ interpret(struct f_inst *what)
runtime("Can't filter integer"); runtime("Can't filter integer");
res.type = T_PATH; res.type = T_PATH;
res.val.ad = as_path_filter(f_pool, v1.val.ad, set, key, pos); res.val.ad = as_path_filter(fs->pool, v1.val.ad, set, key, pos);
} }
else if (v1.type == T_CLIST) else if (v1.type == T_CLIST)
{ {
@ -1408,22 +1413,22 @@ interpret(struct f_inst *what)
if (arg_set == 1) if (arg_set == 1)
runtime("Can't add set"); runtime("Can't add set");
else if (!arg_set) else if (!arg_set)
res.val.ad = int_set_add(f_pool, v1.val.ad, n); res.val.ad = int_set_add(fs->pool, v1.val.ad, n);
else else
res.val.ad = int_set_union(f_pool, v1.val.ad, v2.val.ad); res.val.ad = int_set_union(fs->pool, v1.val.ad, v2.val.ad);
break; break;
case 'd': case 'd':
if (!arg_set) if (!arg_set)
res.val.ad = int_set_del(f_pool, v1.val.ad, n); res.val.ad = int_set_del(fs->pool, v1.val.ad, n);
else else
res.val.ad = clist_filter(f_pool, v1.val.ad, v2, 0); res.val.ad = clist_filter(fs->pool, v1.val.ad, v2, 0);
break; break;
case 'f': case 'f':
if (!arg_set) if (!arg_set)
runtime("Can't filter pair"); runtime("Can't filter pair");
res.val.ad = clist_filter(f_pool, v1.val.ad, v2, 1); res.val.ad = clist_filter(fs->pool, v1.val.ad, v2, 1);
break; break;
default: default:
@ -1450,22 +1455,22 @@ interpret(struct f_inst *what)
if (arg_set == 1) if (arg_set == 1)
runtime("Can't add set"); runtime("Can't add set");
else if (!arg_set) else if (!arg_set)
res.val.ad = ec_set_add(f_pool, v1.val.ad, v2.val.ec); res.val.ad = ec_set_add(fs->pool, v1.val.ad, v2.val.ec);
else else
res.val.ad = ec_set_union(f_pool, v1.val.ad, v2.val.ad); res.val.ad = ec_set_union(fs->pool, v1.val.ad, v2.val.ad);
break; break;
case 'd': case 'd':
if (!arg_set) if (!arg_set)
res.val.ad = ec_set_del(f_pool, v1.val.ad, v2.val.ec); res.val.ad = ec_set_del(fs->pool, v1.val.ad, v2.val.ec);
else else
res.val.ad = eclist_filter(f_pool, v1.val.ad, v2, 0); res.val.ad = eclist_filter(fs->pool, v1.val.ad, v2, 0);
break; break;
case 'f': case 'f':
if (!arg_set) if (!arg_set)
runtime("Can't filter ec"); runtime("Can't filter ec");
res.val.ad = eclist_filter(f_pool, v1.val.ad, v2, 1); res.val.ad = eclist_filter(fs->pool, v1.val.ad, v2, 1);
break; break;
default: default:
@ -1492,22 +1497,22 @@ interpret(struct f_inst *what)
if (arg_set == 1) if (arg_set == 1)
runtime("Can't add set"); runtime("Can't add set");
else if (!arg_set) else if (!arg_set)
res.val.ad = lc_set_add(f_pool, v1.val.ad, v2.val.lc); res.val.ad = lc_set_add(fs->pool, v1.val.ad, v2.val.lc);
else else
res.val.ad = lc_set_union(f_pool, v1.val.ad, v2.val.ad); res.val.ad = lc_set_union(fs->pool, v1.val.ad, v2.val.ad);
break; break;
case 'd': case 'd':
if (!arg_set) if (!arg_set)
res.val.ad = lc_set_del(f_pool, v1.val.ad, v2.val.lc); res.val.ad = lc_set_del(fs->pool, v1.val.ad, v2.val.lc);
else else
res.val.ad = lclist_filter(f_pool, v1.val.ad, v2, 0); res.val.ad = lclist_filter(fs->pool, v1.val.ad, v2, 0);
break; break;
case 'f': case 'f':
if (!arg_set) if (!arg_set)
runtime("Can't filter lc"); runtime("Can't filter lc");
res.val.ad = lclist_filter(f_pool, v1.val.ad, v2, 1); res.val.ad = lclist_filter(fs->pool, v1.val.ad, v2, 1);
break; break;
default: default:
@ -1531,11 +1536,11 @@ interpret(struct f_inst *what)
{ {
ACCESS_RTE; ACCESS_RTE;
ACCESS_EATTRS; ACCESS_EATTRS;
v1.val.net = (*f_rte)->net->n.addr; v1.val.net = (*fs->rte)->net->n.addr;
/* We ignore temporary attributes, probably not a problem here */ /* We ignore temporary attributes, probably not a problem here */
/* 0x02 is a value of BA_AS_PATH, we don't want to include BGP headers */ /* 0x02 is a value of BA_AS_PATH, we don't want to include BGP headers */
eattr *e = ea_find(*f_eattrs, EA_CODE(PROTOCOL_BGP, 0x02)); eattr *e = ea_find(*fs->eattrs, EA_CODE(PROTOCOL_BGP, 0x02));
if (!e || ((e->type & EAF_TYPE_MASK) != EAF_TYPE_AS_PATH)) if (!e || ((e->type & EAF_TYPE_MASK) != EAF_TYPE_AS_PATH))
runtime("Missing AS_PATH attribute"); runtime("Missing AS_PATH attribute");
@ -1563,7 +1568,7 @@ interpret(struct f_inst *what)
ARG_ANY(1); ARG_ANY(1);
res.type = T_STRING; res.type = T_STRING;
res.val.s = val_format_str(v1); res.val.s = val_format_str(fs, v1);
break; break;
case FI_ASSERT: /* Birdtest Assert */ case FI_ASSERT: /* Birdtest Assert */
@ -1765,38 +1770,38 @@ f_run(struct filter *filter, struct rte **rte, struct linpool *tmp_pool, int fla
int rte_cow = ((*rte)->flags & REF_COW); int rte_cow = ((*rte)->flags & REF_COW);
DBG( "Running filter `%s'...", filter->name ); DBG( "Running filter `%s'...", filter->name );
f_rte = rte; struct filter_state fs = {
f_eattrs = NULL; .rte = rte,
f_old_rta = NULL; .pool = tmp_pool,
f_pool = tmp_pool; .flags = flags,
f_flags = flags; };
LOG_BUFFER_INIT(f_buf); LOG_BUFFER_INIT(fs.buf);
struct f_val res = interpret(filter->root); struct f_val res = interpret(&fs, filter->root);
if (f_old_rta) { if (fs.old_rta) {
/* /*
* Cached rta was modified and f_rte contains now an uncached one, * Cached rta was modified and fs->rte contains now an uncached one,
* sharing some part with the cached one. The cached rta should * sharing some part with the cached one. The cached rta should
* be freed (if rte was originally COW, f_old_rta is a clone * be freed (if rte was originally COW, fs->old_rta is a clone
* obtained during rte_cow()). * obtained during rte_cow()).
* *
* This also implements the exception mentioned in f_run() * This also implements the exception mentioned in f_run()
* description. The reason for this is that rta reuses parts of * description. The reason for this is that rta reuses parts of
* f_old_rta, and these may be freed during rta_free(f_old_rta). * fs->old_rta, and these may be freed during rta_free(fs->old_rta).
* This is not the problem if rte was COW, because original rte * This is not the problem if rte was COW, because original rte
* also holds the same rta. * also holds the same rta.
*/ */
if (!rte_cow) if (!rte_cow)
(*f_rte)->attrs = rta_lookup((*f_rte)->attrs); (*fs.rte)->attrs = rta_lookup((*fs.rte)->attrs);
rta_free(f_old_rta); rta_free(fs.old_rta);
} }
if (res.type != T_RETURN) { if (res.type != T_RETURN) {
if (!(f_flags & FF_SILENT)) if (!(fs.flags & FF_SILENT))
log_rl(&rl_runtime_err, L_ERR "Filter %s did not return accept nor reject. Make up your mind", filter->name); log_rl(&rl_runtime_err, L_ERR "Filter %s did not return accept nor reject. Make up your mind", filter->name);
return F_ERROR; return F_ERROR;
} }
@ -1810,16 +1815,15 @@ struct f_val
f_eval_rte(struct f_inst *expr, struct rte **rte, struct linpool *tmp_pool) f_eval_rte(struct f_inst *expr, struct rte **rte, struct linpool *tmp_pool)
{ {
f_rte = rte; struct filter_state fs = {
f_eattrs = NULL; .rte = rte,
f_old_rta = NULL; .pool = tmp_pool,
f_pool = tmp_pool; };
f_flags = 0;
LOG_BUFFER_INIT(f_buf); LOG_BUFFER_INIT(fs.buf);
/* Note that in this function we assume that rte->attrs is private / uncached */ /* Note that in this function we assume that rte->attrs is private / uncached */
struct f_val res = interpret(expr); struct f_val res = interpret(&fs, expr);
return res; return res;
} }
@ -1827,14 +1831,13 @@ f_eval_rte(struct f_inst *expr, struct rte **rte, struct linpool *tmp_pool)
struct f_val struct f_val
f_eval(struct f_inst *expr, struct linpool *tmp_pool) f_eval(struct f_inst *expr, struct linpool *tmp_pool)
{ {
f_flags = 0; struct filter_state fs = {
f_eattrs = NULL; .pool = tmp_pool,
f_rte = NULL; };
f_pool = tmp_pool;
LOG_BUFFER_INIT(f_buf); LOG_BUFFER_INIT(fs.buf);
return interpret(expr); return interpret(&fs, expr);
} }
uint uint