Jerry/bird
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
bird/filter/interpret.m4

75 lines
2 KiB
Text
Raw Normal View History

Filter refactoring: The instructions are converted to the switch body by M4
2018-12-20 23:25:54 +08:00
m4_divert(-1)m4_dnl
#
# BIRD -- Generator of Filter Instructions
#
# (c) 2018 Maria Matejka <mq@jmq.cz>
#
# Can be freely distributed and used under the terms of the GNU GPL.
#
# Common aliases
m4_define(DNL, `m4_dnl')
Filter refactoring: dropped the recursion from the interpreter This is a major change of how the filters are interpreted. If everything works how it should, it should not affect you unless you are hacking the filters themselves. Anyway, this change should make a huge improvement in the filter performance as previous benchmarks showed that our major problem lies in the recursion itself. There are also some changes in nest and protocols, related mostly to spreading const declarations throughout the whole BIRD and also to refactored dynamic attribute definitions. The need of these came up during the whole work and it is too difficult to split out these not-so-related changes.
2018-12-27 21:26:11 +08:00
m4_define(INST, `break; case $1:
m4_ifelse(eval($2 > 0), `if (vstk.cnt < $2) runtime("Stack underflow");', `')
vstk.cnt -= $2;
')
Filter refactoring: The instructions are converted to the switch body by M4
2018-12-20 23:25:54 +08:00
Filter refactoring: dropped the recursion from the interpreter This is a major change of how the filters are interpreted. If everything works how it should, it should not affect you unless you are hacking the filters themselves. Anyway, this change should make a huge improvement in the filter performance as previous benchmarks showed that our major problem lies in the recursion itself. There are also some changes in nest and protocols, related mostly to spreading const declarations throughout the whole BIRD and also to refactored dynamic attribute definitions. The need of these came up during the whole work and it is too difficult to split out these not-so-related changes.
2018-12-27 21:26:11 +08:00
m4_define(ARG, `if (v$1.type != $2) runtime("Argument $1 of instruction %s must be of type $2, got 0x%02x", f_instruction_name(what->fi_code), v$1.type)')
m4_define(RESULT_OK, `vstk.cnt++')
m4_define(RESULT, `RESULT_VAL([[ (struct f_val) { .type = $1, .val.$2 = $3 } ]])')
m4_define(RESULT_VAL, `do { res = $1; RESULT_OK; } while (0)')
m4_define(LINEX, `do {
estk.item[estk.cnt].pos = 0;
estk.item[estk.cnt].line = $1;
estk.item[estk.cnt].ventry = vstk.cnt;
estk.item[estk.cnt].emask = 0;
estk.cnt++;
} while (0)')
m4_define(LINE, `do {
if (what->lines[$2]) {
estk.item[estk.cnt].pos = 0;
estk.item[estk.cnt].line = what->lines[$2];
estk.item[estk.cnt].ventry = vstk.cnt;
estk.item[estk.cnt].emask = 0;
estk.cnt++;
}
} while (0)')
m4_define(LINEP, LINE($1, $2))
m4_define(ARG_ANY, `')
m4_define(SYMBOL, `const struct symbol *sym = what->sym')
Filter: refactoring of instruction constructors
2019-01-21 16:17:54 +08:00
m4_define(VALI, `res = what->val')
m4_define(VALP, `res = what->val')
m4_define(VAR, `res = *what->vp')
Filter refactoring: dropped the recursion from the interpreter This is a major change of how the filters are interpreted. If everything works how it should, it should not affect you unless you are hacking the filters themselves. Anyway, this change should make a huge improvement in the filter performance as previous benchmarks showed that our major problem lies in the recursion itself. There are also some changes in nest and protocols, related mostly to spreading const declarations throughout the whole BIRD and also to refactored dynamic attribute definitions. The need of these came up during the whole work and it is too difficult to split out these not-so-related changes.
2018-12-27 21:26:11 +08:00
m4_define(FRET, `enum filter_return fret = what->fret')
m4_define(ECS, `enum ec_subtype ecs = what->ecs')
m4_define(RTC, `struct rtable *table = what->rtc->table')
m4_define(STATIC_ATTR, `struct f_static_attr sa = what->sa')
m4_define(DYNAMIC_ATTR, `struct f_dynamic_attr da = what->da')
Filter: refactoring of instruction constructors
2019-01-21 16:17:54 +08:00
m4_define(TREE, `')
m4_define(STRING, `')
m4_define(COUNT, `')
Filter refactoring: dropped the recursion from the interpreter This is a major change of how the filters are interpreted. If everything works how it should, it should not affect you unless you are hacking the filters themselves. Anyway, this change should make a huge improvement in the filter performance as previous benchmarks showed that our major problem lies in the recursion itself. There are also some changes in nest and protocols, related mostly to spreading const declarations throughout the whole BIRD and also to refactored dynamic attribute definitions. The need of these came up during the whole work and it is too difficult to split out these not-so-related changes.
2018-12-27 21:26:11 +08:00
m4_define(POSTFIXIFY, `')
m4_define(LINE_SIZE, `')
m4_define(SAME, `')
Filter: refactoring of instruction constructors
2019-01-21 16:17:54 +08:00
m4_define(STRUCT, `')
m4_define(NEW, `')
Filter refactoring: dropped the recursion from the interpreter This is a major change of how the filters are interpreted. If everything works how it should, it should not affect you unless you are hacking the filters themselves. Anyway, this change should make a huge improvement in the filter performance as previous benchmarks showed that our major problem lies in the recursion itself. There are also some changes in nest and protocols, related mostly to spreading const declarations throughout the whole BIRD and also to refactored dynamic attribute definitions. The need of these came up during the whole work and it is too difficult to split out these not-so-related changes.
2018-12-27 21:26:11 +08:00
m4_m4wrap(`
m4_divert(0)DNL
case FI_NOP: bug("This shall not happen");
m4_undivert(1)
break; default: bug( "Unknown instruction %d (%c)", what->fi_code, what->fi_code & 0xff);
')
m4_divert(1)
Filter refactoring: The instructions are converted to the switch body by M4
2018-12-20 23:25:54 +08:00
m4_changequote([[,]])
Filter refactoring: dropped the recursion from the interpreter This is a major change of how the filters are interpreted. If everything works how it should, it should not affect you unless you are hacking the filters themselves. Anyway, this change should make a huge improvement in the filter performance as previous benchmarks showed that our major problem lies in the recursion itself. There are also some changes in nest and protocols, related mostly to spreading const declarations throughout the whole BIRD and also to refactored dynamic attribute definitions. The need of these came up during the whole work and it is too difficult to split out these not-so-related changes.
2018-12-27 21:26:11 +08:00
Reference in a new issue Copy permalink