ruby/iseq.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

3968 строки
123 KiB
C
Исходник Обычный вид История

/**********************************************************************
iseq.c -
$Author$
created at: 2006-07-11(Tue) 09:00:03 +0900
Copyright (C) 2006 Koichi Sasada
**********************************************************************/
#define RUBY_VM_INSNS_INFO 1
/* #define RUBY_MARK_FREE_DEBUG 1 */
#include "ruby/internal/config.h"
#ifdef HAVE_DLADDR
# include <dlfcn.h>
#endif
#include "eval_intern.h"
#include "gc.h"
#include "id_table.h"
#include "internal.h"
#include "internal/bits.h"
#include "internal/class.h"
#include "internal/compile.h"
#include "internal/error.h"
#include "internal/file.h"
#include "internal/hash.h"
#include "internal/parse.h"
#include "internal/sanitizers.h"
#include "internal/symbol.h"
#include "internal/thread.h"
#include "internal/variable.h"
#include "iseq.h"
#include "mjit.h"
#include "ruby/util.h"
#include "vm_core.h"
#include "vm_callinfo.h"
#include "yjit.h"
2021-09-30 11:30:04 +03:00
#include "ruby/ractor.h"
#include "builtin.h"
#include "insns.inc"
#include "insns_info.inc"
VALUE rb_cISeq;
static VALUE iseqw_new(const rb_iseq_t *iseq);
static const rb_iseq_t *iseqw_check(VALUE iseqw);
#if VM_INSN_INFO_TABLE_IMPL == 2
static struct succ_index_table *succ_index_table_create(int max_pos, int *data, int size);
static unsigned int *succ_index_table_invert(int max_pos, struct succ_index_table *sd, int size);
static int succ_index_lookup(const struct succ_index_table *sd, int x);
#endif
#define hidden_obj_p(obj) (!SPECIAL_CONST_P(obj) && !RBASIC(obj)->klass)
static inline VALUE
obj_resurrect(VALUE obj)
{
if (hidden_obj_p(obj)) {
switch (BUILTIN_TYPE(obj)) {
case T_STRING:
obj = rb_str_resurrect(obj);
break;
case T_ARRAY:
obj = rb_ary_resurrect(obj);
break;
case T_HASH:
obj = rb_hash_resurrect(obj);
break;
2020-04-08 09:13:37 +03:00
default:
break;
}
}
return obj;
}
static void
free_arena(struct iseq_compile_data_storage *cur)
{
struct iseq_compile_data_storage *next;
while (cur) {
next = cur->next;
ruby_xfree(cur);
cur = next;
}
}
static void
compile_data_free(struct iseq_compile_data *compile_data)
{
if (compile_data) {
free_arena(compile_data->node.storage_head);
free_arena(compile_data->insn.storage_head);
if (compile_data->ivar_cache_table) {
rb_id_table_free(compile_data->ivar_cache_table);
}
ruby_xfree(compile_data);
}
}
New constant caching insn: opt_getconstant_path Previously YARV bytecode implemented constant caching by having a pair of instructions, opt_getinlinecache and opt_setinlinecache, wrapping a series of getconstant calls (with putobject providing supporting arguments). This commit replaces that pattern with a new instruction, opt_getconstant_path, handling both getting/setting the inline cache and fetching the constant on a cache miss. This is implemented by storing the full constant path as a null-terminated array of IDs inside of the IC structure. idNULL is used to signal an absolute constant reference. $ ./miniruby --dump=insns -e '::Foo::Bar::Baz' == disasm: #<ISeq:<main>@-e:1 (1,0)-(1,13)> (catch: FALSE) 0000 opt_getconstant_path <ic:0 ::Foo::Bar::Baz> ( 1)[Li] 0002 leave The motivation for this is that we had increasingly found the need to disassemble the instructions between the opt_getinlinecache and opt_setinlinecache in order to determine the constant we are fetching, or otherwise store metadata. This disassembly was done: * In opt_setinlinecache, to register the IC against the constant names it is using for granular invalidation. * In rb_iseq_free, to unregister the IC from the invalidation table. * In YJIT to find the position of a opt_getinlinecache instruction to invalidate it when the cache is populated * In YJIT to register the constant names being used for invalidation. With this change we no longe need disassemly for these (in fact rb_iseq_each is now unused), as the list of constant names being referenced is held in the IC. This should also make it possible to make more optimizations in the future. This may also reduce the size of iseqs, as previously each segment required 32 bytes (on 64-bit platforms) for each constant segment. This implementation only stores one ID per-segment. There should be no significant performance change between this and the previous implementation. Previously opt_getinlinecache was a "leaf" instruction, but it included a jump (almost always to a separate cache line). Now opt_getconstant_path is a non-leaf (it may raise/autoload/call const_missing) but it does not jump. These seem to even out.
2022-08-10 20:35:48 +03:00
static void
2022-12-02 17:43:53 +03:00
remove_from_constant_cache(ID id, IC ic)
{
New constant caching insn: opt_getconstant_path Previously YARV bytecode implemented constant caching by having a pair of instructions, opt_getinlinecache and opt_setinlinecache, wrapping a series of getconstant calls (with putobject providing supporting arguments). This commit replaces that pattern with a new instruction, opt_getconstant_path, handling both getting/setting the inline cache and fetching the constant on a cache miss. This is implemented by storing the full constant path as a null-terminated array of IDs inside of the IC structure. idNULL is used to signal an absolute constant reference. $ ./miniruby --dump=insns -e '::Foo::Bar::Baz' == disasm: #<ISeq:<main>@-e:1 (1,0)-(1,13)> (catch: FALSE) 0000 opt_getconstant_path <ic:0 ::Foo::Bar::Baz> ( 1)[Li] 0002 leave The motivation for this is that we had increasingly found the need to disassemble the instructions between the opt_getinlinecache and opt_setinlinecache in order to determine the constant we are fetching, or otherwise store metadata. This disassembly was done: * In opt_setinlinecache, to register the IC against the constant names it is using for granular invalidation. * In rb_iseq_free, to unregister the IC from the invalidation table. * In YJIT to find the position of a opt_getinlinecache instruction to invalidate it when the cache is populated * In YJIT to register the constant names being used for invalidation. With this change we no longe need disassemly for these (in fact rb_iseq_each is now unused), as the list of constant names being referenced is held in the IC. This should also make it possible to make more optimizations in the future. This may also reduce the size of iseqs, as previously each segment required 32 bytes (on 64-bit platforms) for each constant segment. This implementation only stores one ID per-segment. There should be no significant performance change between this and the previous implementation. Previously opt_getinlinecache was a "leaf" instruction, but it included a jump (almost always to a separate cache line). Now opt_getconstant_path is a non-leaf (it may raise/autoload/call const_missing) but it does not jump. These seem to even out.
2022-08-10 20:35:48 +03:00
rb_vm_t *vm = GET_VM();
VALUE lookup_result;
st_data_t ic_data = (st_data_t)ic;
if (rb_id_table_lookup(vm->constant_cache, id, &lookup_result)) {
st_table *ics = (st_table *)lookup_result;
st_delete(ics, &ic_data, NULL);
if (ics->num_entries == 0) {
rb_id_table_delete(vm->constant_cache, id);
st_free_table(ics);
}
}
}
// When an ISEQ is being freed, all of its associated ICs are going to go away
New constant caching insn: opt_getconstant_path Previously YARV bytecode implemented constant caching by having a pair of instructions, opt_getinlinecache and opt_setinlinecache, wrapping a series of getconstant calls (with putobject providing supporting arguments). This commit replaces that pattern with a new instruction, opt_getconstant_path, handling both getting/setting the inline cache and fetching the constant on a cache miss. This is implemented by storing the full constant path as a null-terminated array of IDs inside of the IC structure. idNULL is used to signal an absolute constant reference. $ ./miniruby --dump=insns -e '::Foo::Bar::Baz' == disasm: #<ISeq:<main>@-e:1 (1,0)-(1,13)> (catch: FALSE) 0000 opt_getconstant_path <ic:0 ::Foo::Bar::Baz> ( 1)[Li] 0002 leave The motivation for this is that we had increasingly found the need to disassemble the instructions between the opt_getinlinecache and opt_setinlinecache in order to determine the constant we are fetching, or otherwise store metadata. This disassembly was done: * In opt_setinlinecache, to register the IC against the constant names it is using for granular invalidation. * In rb_iseq_free, to unregister the IC from the invalidation table. * In YJIT to find the position of a opt_getinlinecache instruction to invalidate it when the cache is populated * In YJIT to register the constant names being used for invalidation. With this change we no longe need disassemly for these (in fact rb_iseq_each is now unused), as the list of constant names being referenced is held in the IC. This should also make it possible to make more optimizations in the future. This may also reduce the size of iseqs, as previously each segment required 32 bytes (on 64-bit platforms) for each constant segment. This implementation only stores one ID per-segment. There should be no significant performance change between this and the previous implementation. Previously opt_getinlinecache was a "leaf" instruction, but it included a jump (almost always to a separate cache line). Now opt_getconstant_path is a non-leaf (it may raise/autoload/call const_missing) but it does not jump. These seem to even out.
2022-08-10 20:35:48 +03:00
// as well. Because of this, we need to iterate over the ICs, and clear them
// from the VM's constant cache.
static void
iseq_clear_ic_references(const rb_iseq_t *iseq)
{
// In some cases (when there is a compilation error), we end up with
// ic_size greater than 0, but no allocated is_entries buffer.
// If there's no is_entries buffer to loop through, return early.
// [Bug #19173]
if (!ISEQ_BODY(iseq)->is_entries) {
return;
}
New constant caching insn: opt_getconstant_path Previously YARV bytecode implemented constant caching by having a pair of instructions, opt_getinlinecache and opt_setinlinecache, wrapping a series of getconstant calls (with putobject providing supporting arguments). This commit replaces that pattern with a new instruction, opt_getconstant_path, handling both getting/setting the inline cache and fetching the constant on a cache miss. This is implemented by storing the full constant path as a null-terminated array of IDs inside of the IC structure. idNULL is used to signal an absolute constant reference. $ ./miniruby --dump=insns -e '::Foo::Bar::Baz' == disasm: #<ISeq:<main>@-e:1 (1,0)-(1,13)> (catch: FALSE) 0000 opt_getconstant_path <ic:0 ::Foo::Bar::Baz> ( 1)[Li] 0002 leave The motivation for this is that we had increasingly found the need to disassemble the instructions between the opt_getinlinecache and opt_setinlinecache in order to determine the constant we are fetching, or otherwise store metadata. This disassembly was done: * In opt_setinlinecache, to register the IC against the constant names it is using for granular invalidation. * In rb_iseq_free, to unregister the IC from the invalidation table. * In YJIT to find the position of a opt_getinlinecache instruction to invalidate it when the cache is populated * In YJIT to register the constant names being used for invalidation. With this change we no longe need disassemly for these (in fact rb_iseq_each is now unused), as the list of constant names being referenced is held in the IC. This should also make it possible to make more optimizations in the future. This may also reduce the size of iseqs, as previously each segment required 32 bytes (on 64-bit platforms) for each constant segment. This implementation only stores one ID per-segment. There should be no significant performance change between this and the previous implementation. Previously opt_getinlinecache was a "leaf" instruction, but it included a jump (almost always to a separate cache line). Now opt_getconstant_path is a non-leaf (it may raise/autoload/call const_missing) but it does not jump. These seem to even out.
2022-08-10 20:35:48 +03:00
for (unsigned int ic_idx = 0; ic_idx < ISEQ_BODY(iseq)->ic_size; ic_idx++) {
IC ic = &ISEQ_IS_IC_ENTRY(ISEQ_BODY(iseq), ic_idx);
// Iterate over the IC's constant path's segments and clean any references to
// the ICs out of the VM's constant cache table.
const ID *segments = ic->segments;
// It's possible that segments is NULL if we overallocated an IC but
// optimizations removed the instruction using it
if (segments == NULL)
continue;
for (int i = 0; segments[i]; i++) {
ID id = segments[i];
if (id == idNULL) continue;
remove_from_constant_cache(id, ic);
}
ruby_xfree((void *)segments);
}
}
2015-07-22 01:52:59 +03:00
void
rb_iseq_free(const rb_iseq_t *iseq)
{
RUBY_FREE_ENTER("iseq");
if (iseq && ISEQ_BODY(iseq)) {
iseq_clear_ic_references(iseq);
struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
mjit.c: merge MJIT infrastructure that allows to JIT-compile Ruby methods by generating C code and using C compiler. See the first comment of mjit.c to know what this file does. mjit.c is authored by Vladimir Makarov <vmakarov@redhat.com>. After he invented great method JIT infrastructure for MRI as MJIT, Lars Kanis <lars@greiz-reinsdorf.de> sent the patch to support MinGW in MJIT. In addition to merging it, I ported pthread to Windows native threads. Now this MJIT infrastructure can be compiled on Visual Studio. This commit simplifies mjit.c to decrease code at initial merge. For example, this commit does not provide multiple JIT threads support. We can resurrect them later if we really want them, but I wanted to minimize diff to make it easier to review this patch. `/tmp/_mjitXXX` file is renamed to `/tmp/_ruby_mjitXXX` because non-Ruby developers may not know the name "mjit" and the file name should make sure it's from Ruby and not from some harmful programs. TODO: it may be better to store this to some temporary directory which Ruby is already using by Tempfile, if it's not bad for performance. mjit.h: New. It has `mjit_exec` interface similar to `vm_exec`, which is for triggering MJIT. This drops interface for AOT compared to the original MJIT. Makefile.in: define macros to let MJIT know the path of MJIT header. Probably we can refactor this to reduce the number of macros (TODO). win32/Makefile.sub: ditto. common.mk: compile mjit.o and mjit_compile.o. Unlike original MJIT, this commit separates MJIT infrastructure and JIT compiler code as independent object files. As initial patch is NOT going to have ultra-fast JIT compiler, it's likely to replace JIT compiler, e.g. original MJIT's compiler or some future JIT impelementations which are not public now. inits.c: define MJIT module. This is added because `MJIT.enabled?` was necessary for testing. test/lib/zombie_hunter.rb: skip if `MJIT.enabled?`. Obviously this wouldn't work with current code when JIT is enabled. test/ruby/test_io.rb: skip this too. This would make no sense with MJIT. ruby.c: define MJIT CLI options. As major difference from original MJIT, "-j:l"/"--jit:llvm" are renamed to "--jit-cc" because I want to support not only gcc/clang but also cl.exe (Visual Studio) in the future. But it takes only "--jit-cc=gcc", "--jit-cc=clang" for now. And only long "--jit" options are allowed since some Ruby committers preferred it at Ruby developers Meeting on January, and some of options are renamed. This file also triggers to initialize MJIT thread and variables. eval.c: finalize MJIT worker thread and variables. test/ruby/test_rubyoptions.rb: fix number of CLI options for --jit. thread_pthread.c: change for pthread abstraction in MJIT. Prefix rb_ for functions which are used by other files. thread_win32.c: ditto, for Windows. Those pthread porting is one of major works that YARV-MJIT created, which is my fork of MJIT, in Feature 14235. thread.c: follow rb_ prefix changes vm.c: trigger MJIT call on VM invocation. Also trigger `mjit_mark` to avoid SEGV by race between JIT and GC of ISeq. The improvement was provided by wanabe <s.wanabe@gmail.com>. In JIT compiler I created and am going to add in my next commit, I found that having `mjit_exec` after `vm_loop_start:` is harmful because the JIT-ed function doesn't proceed other ISeqs on RESTORE_REGS of leave insn. Executing non-FINISH frame is unexpected for my JIT compiler and `exception_handler` triggers executions of such ISeqs. So `mjit_exec` here should be executed only when it directly comes from `vm_exec` call. `RubyVM::MJIT` module and `.enabled?` method is added so that we can skip some tests which don't expect JIT threads or compiler file descriptors. vm_insnhelper.h: trigger MJIT on method calls during VM execution. vm_core.h: add fields required for mjit.c. `bp` must be `cfp[6]` because rb_control_frame_struct is likely to be casted to another struct. The last position is the safest place to add the new field. vm_insnhelper.c: save initial value of cfp->ep as cfp->bp. This is an optimization which are done in both MJIT and YARV-MJIT. So this change is added in this commit. Calculating bp from ep is a little heavy work, so bp is kind of cache for it. iseq.c: notify ISeq GC to MJIT. We should know which iseq in MJIT queue is GCed to avoid SEGV. TODO: unload some GCed units in some safe way. gc.c: add hooks so that MJIT can wait GC, and vice versa. Simultaneous JIT and GC executions may cause SEGV and so we should synchronize them. cont.c: save continuation information in MJIT worker. As MJIT shouldn't unload JIT-ed code which is being used, MJIT wants to know full list of saved execution contexts for continuation and detect ISeqs in use. mjit_compile.c: added empty JIT compiler so that you can reuse this commit to build your own JIT compiler. This commit tries to compile ISeqs but all of them are considered as not supported in this commit. So you can't use JIT compiler in this commit yet while we added --jit option now. Patch author: Vladimir Makarov <vmakarov@redhat.com>. Contributors: Takashi Kokubun <takashikkbn@gmail.com>. wanabe <s.wanabe@gmail.com>. Lars Kanis <lars@greiz-reinsdorf.de>. Part of Feature 12589 and 14235. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@62189 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-02-04 09:58:09 +03:00
mjit_free_iseq(iseq); /* Notify MJIT */
#if USE_YJIT
Rust YJIT In December 2021, we opened an [issue] to solicit feedback regarding the porting of the YJIT codebase from C99 to Rust. There were some reservations, but this project was given the go ahead by Ruby core developers and Matz. Since then, we have successfully completed the port of YJIT to Rust. The new Rust version of YJIT has reached parity with the C version, in that it passes all the CRuby tests, is able to run all of the YJIT benchmarks, and performs similarly to the C version (because it works the same way and largely generates the same machine code). We've even incorporated some design improvements, such as a more fine-grained constant invalidation mechanism which we expect will make a big difference in Ruby on Rails applications. Because we want to be careful, YJIT is guarded behind a configure option: ```shell ./configure --enable-yjit # Build YJIT in release mode ./configure --enable-yjit=dev # Build YJIT in dev/debug mode ``` By default, YJIT does not get compiled and cargo/rustc is not required. If YJIT is built in dev mode, then `cargo` is used to fetch development dependencies, but when building in release, `cargo` is not required, only `rustc`. At the moment YJIT requires Rust 1.60.0 or newer. The YJIT command-line options remain mostly unchanged, and more details about the build process are documented in `doc/yjit/yjit.md`. The CI tests have been updated and do not take any more resources than before. The development history of the Rust port is available at the following commit for interested parties: https://github.com/Shopify/ruby/commit/1fd9573d8b4b65219f1c2407f30a0a60e537f8be Our hope is that Rust YJIT will be compiled and included as a part of system packages and compiled binaries of the Ruby 3.2 release. We do not anticipate any major problems as Rust is well supported on every platform which YJIT supports, but to make sure that this process works smoothly, we would like to reach out to those who take care of building systems packages before the 3.2 release is shipped and resolve any issues that may come up. [issue]: https://bugs.ruby-lang.org/issues/18481 Co-authored-by: Maxime Chevalier-Boisvert <maximechevalierb@gmail.com> Co-authored-by: Noah Gibbs <the.codefolio.guy@gmail.com> Co-authored-by: Kevin Newton <kddnewton@gmail.com>
2022-04-19 21:40:21 +03:00
rb_yjit_iseq_free(body->yjit_payload);
#endif
ruby_xfree((void *)body->iseq_encoded);
ruby_xfree((void *)body->insns_info.body);
if (body->insns_info.positions) ruby_xfree((void *)body->insns_info.positions);
#if VM_INSN_INFO_TABLE_IMPL == 2
if (body->insns_info.succ_index_table) ruby_xfree(body->insns_info.succ_index_table);
#endif
if (LIKELY(body->local_table != rb_iseq_shared_exc_local_tbl))
ruby_xfree((void *)body->local_table);
ruby_xfree((void *)body->is_entries);
if (body->call_data) {
ruby_xfree(body->call_data);
}
ruby_xfree((void *)body->catch_table);
ruby_xfree((void *)body->param.opt_table);
if (ISEQ_MBITS_BUFLEN(body->iseq_size) > 1 && body->mark_bits.list) {
ruby_xfree((void *)body->mark_bits.list);
}
if (body->param.keyword != NULL) {
ruby_xfree((void *)body->param.keyword->default_values);
ruby_xfree((void *)body->param.keyword);
}
compile_data_free(ISEQ_COMPILE_DATA(iseq));
if (body->outer_variables) rb_id_table_free(body->outer_variables);
ruby_xfree(body);
}
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
if (iseq && ISEQ_EXECUTABLE_P(iseq) && iseq->aux.exec.local_hooks) {
rb_hook_list_free(iseq->aux.exec.local_hooks);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
}
RUBY_FREE_LEAVE("iseq");
}
typedef VALUE iseq_value_itr_t(void *ctx, VALUE obj);
static inline void
iseq_scan_bits(unsigned int page, iseq_bits_t bits, VALUE *code, iseq_value_itr_t *func, void *data)
{
unsigned int offset;
unsigned int page_offset = (page * ISEQ_MBITS_BITLENGTH);
while (bits) {
offset = ntz_intptr(bits);
VALUE op = code[page_offset + offset];
VALUE newop = func(data, op);
if (newop != op) {
code[page_offset + offset] = newop;
if (data) {
VALUE *original_iseq = (VALUE *)data;
original_iseq[page_offset + offset] = newop;
}
}
bits &= bits - 1; // Reset Lowest Set Bit (BLSR)
}
}
static void
rb_iseq_each_value(const rb_iseq_t *iseq, iseq_value_itr_t * func, void *data)
{
unsigned int size;
VALUE *code;
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
size = body->iseq_size;
code = body->iseq_encoded;
union iseq_inline_storage_entry *is_entries = body->is_entries;
if (body->is_entries) {
// Skip iterating over ivc caches
is_entries += body->ivc_size;
// ICVARC entries
for (unsigned int i = 0; i < body->icvarc_size; i++, is_entries++) {
ICVARC icvarc = (ICVARC)is_entries;
if (icvarc->entry) {
RUBY_ASSERT(!RB_TYPE_P(icvarc->entry->class_value, T_NONE));
VALUE nv = func(data, icvarc->entry->class_value);
if (icvarc->entry->class_value != nv) {
icvarc->entry->class_value = nv;
}
}
}
// ISE entries
for (unsigned int i = 0; i < body->ise_size; i++, is_entries++) {
union iseq_inline_storage_entry *const is = (union iseq_inline_storage_entry *)is_entries;
if (is->once.value) {
VALUE nv = func(data, is->once.value);
if (is->once.value != nv) {
is->once.value = nv;
}
}
}
// IC Entries
for (unsigned int i = 0; i < body->ic_size; i++, is_entries++) {
IC ic = (IC)is_entries;
if (ic->entry) {
VALUE nv = func(data, (VALUE)ic->entry);
if ((VALUE)ic->entry != nv) {
ic->entry = (void *)nv;
}
}
}
}
// Embedded VALUEs
if (body->mark_bits.list) {
if (ISEQ_MBITS_BUFLEN(size) == 1) {
iseq_scan_bits(0, body->mark_bits.single, code, func, data);
}
else {
if (body->mark_bits.list) {
for (unsigned int i = 0; i < ISEQ_MBITS_BUFLEN(size); i++) {
iseq_bits_t bits = body->mark_bits.list[i];
iseq_scan_bits(i, bits, code, func, data);
}
}
}
}
}
static VALUE
update_each_insn_value(void *ctx, VALUE obj)
{
return rb_gc_location(obj);
}
void
rb_iseq_update_references(rb_iseq_t *iseq)
{
if (ISEQ_BODY(iseq)) {
struct rb_iseq_constant_body *body = ISEQ_BODY(iseq);
body->variable.coverage = rb_gc_location(body->variable.coverage);
body->variable.pc2branchindex = rb_gc_location(body->variable.pc2branchindex);
body->variable.script_lines = rb_gc_location(body->variable.script_lines);
body->location.label = rb_gc_location(body->location.label);
body->location.base_label = rb_gc_location(body->location.base_label);
body->location.pathobj = rb_gc_location(body->location.pathobj);
if (body->local_iseq) {
body->local_iseq = (struct rb_iseq_struct *)rb_gc_location((VALUE)body->local_iseq);
}
if (body->parent_iseq) {
body->parent_iseq = (struct rb_iseq_struct *)rb_gc_location((VALUE)body->parent_iseq);
}
`Primitive.mandatory_only?` for fast path Compare with the C methods, A built-in methods written in Ruby is slower if only mandatory parameters are given because it needs to check the argumens and fill default values for optional and keyword parameters (C methods can check the number of parameters with `argc`, so there are no overhead). Passing mandatory arguments are common (optional arguments are exceptional, in many cases) so it is important to provide the fast path for such common cases. `Primitive.mandatory_only?` is a special builtin function used with `if` expression like that: ```ruby def self.at(time, subsec = false, unit = :microsecond, in: nil) if Primitive.mandatory_only? Primitive.time_s_at1(time) else Primitive.time_s_at(time, subsec, unit, Primitive.arg!(:in)) end end ``` and it makes two ISeq, ``` def self.at(time, subsec = false, unit = :microsecond, in: nil) Primitive.time_s_at(time, subsec, unit, Primitive.arg!(:in)) end def self.at(time) Primitive.time_s_at1(time) end ``` and (2) is pointed by (1). Note that `Primitive.mandatory_only?` should be used only in a condition of an `if` statement and the `if` statement should be equal to the methdo body (you can not put any expression before and after the `if` statement). A method entry with `mandatory_only?` (`Time.at` on the above case) is marked as `iseq_overload`. When the method will be dispatch only with mandatory arguments (`Time.at(0)` for example), make another method entry with ISeq (2) as mandatory only method entry and it will be cached in an inline method cache. The idea is similar discussed in https://bugs.ruby-lang.org/issues/16254 but it only checks mandatory parameters or more, because many cases only mandatory parameters are given. If we find other cases (optional or keyword parameters are used frequently and it hurts performance), we can extend the feature.
2021-11-12 20:12:20 +03:00
if (body->mandatory_only_iseq) {
body->mandatory_only_iseq = (struct rb_iseq_struct *)rb_gc_location((VALUE)body->mandatory_only_iseq);
}
if (body->call_data) {
for (unsigned int i=0; i<body->ci_size; i++) {
struct rb_call_data *cds = body->call_data;
if (!SPECIAL_CONST_P((VALUE)cds[i].ci)) {
cds[i].ci = (struct rb_callinfo *)rb_gc_location((VALUE)cds[i].ci);
}
cds[i].cc = (struct rb_callcache *)rb_gc_location((VALUE)cds[i].cc);
}
}
VALUE *original_iseq = ISEQ_ORIGINAL_ISEQ(iseq);
rb_iseq_each_value(iseq, update_each_insn_value, (void *)original_iseq);
if (body->param.flags.has_kw && ISEQ_COMPILE_DATA(iseq) == NULL) {
int i, j;
i = body->param.keyword->required_num;
for (j = 0; i < body->param.keyword->num; i++, j++) {
VALUE obj = body->param.keyword->default_values[j];
2022-11-15 07:24:08 +03:00
if (!UNDEF_P(obj)) {
body->param.keyword->default_values[j] = rb_gc_location(obj);
}
}
}
if (body->catch_table) {
struct iseq_catch_table *table = body->catch_table;
unsigned int i;
2019-09-27 04:20:56 +03:00
for (i = 0; i < table->size; i++) {
struct iseq_catch_table_entry *entry;
entry = UNALIGNED_MEMBER_PTR(table, entries[i]);
if (entry->iseq) {
entry->iseq = (rb_iseq_t *)rb_gc_location((VALUE)entry->iseq);
}
}
}
#if USE_MJIT
mjit_update_references(iseq);
#endif
#if USE_YJIT
Rust YJIT In December 2021, we opened an [issue] to solicit feedback regarding the porting of the YJIT codebase from C99 to Rust. There were some reservations, but this project was given the go ahead by Ruby core developers and Matz. Since then, we have successfully completed the port of YJIT to Rust. The new Rust version of YJIT has reached parity with the C version, in that it passes all the CRuby tests, is able to run all of the YJIT benchmarks, and performs similarly to the C version (because it works the same way and largely generates the same machine code). We've even incorporated some design improvements, such as a more fine-grained constant invalidation mechanism which we expect will make a big difference in Ruby on Rails applications. Because we want to be careful, YJIT is guarded behind a configure option: ```shell ./configure --enable-yjit # Build YJIT in release mode ./configure --enable-yjit=dev # Build YJIT in dev/debug mode ``` By default, YJIT does not get compiled and cargo/rustc is not required. If YJIT is built in dev mode, then `cargo` is used to fetch development dependencies, but when building in release, `cargo` is not required, only `rustc`. At the moment YJIT requires Rust 1.60.0 or newer. The YJIT command-line options remain mostly unchanged, and more details about the build process are documented in `doc/yjit/yjit.md`. The CI tests have been updated and do not take any more resources than before. The development history of the Rust port is available at the following commit for interested parties: https://github.com/Shopify/ruby/commit/1fd9573d8b4b65219f1c2407f30a0a60e537f8be Our hope is that Rust YJIT will be compiled and included as a part of system packages and compiled binaries of the Ruby 3.2 release. We do not anticipate any major problems as Rust is well supported on every platform which YJIT supports, but to make sure that this process works smoothly, we would like to reach out to those who take care of building systems packages before the 3.2 release is shipped and resolve any issues that may come up. [issue]: https://bugs.ruby-lang.org/issues/18481 Co-authored-by: Maxime Chevalier-Boisvert <maximechevalierb@gmail.com> Co-authored-by: Noah Gibbs <the.codefolio.guy@gmail.com> Co-authored-by: Kevin Newton <kddnewton@gmail.com>
2022-04-19 21:40:21 +03:00
rb_yjit_iseq_update_references(body->yjit_payload);
#endif
}
}
static VALUE
each_insn_value(void *ctx, VALUE obj)
{
rb_gc_mark_movable(obj);
return obj;
}
2015-07-22 01:52:59 +03:00
void
rb_iseq_mark(const rb_iseq_t *iseq)
{
RUBY_MARK_ENTER("iseq");
RUBY_MARK_UNLESS_NULL(iseq->wrapper);
if (ISEQ_BODY(iseq)) {
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
rb_iseq_each_value(iseq, each_insn_value, NULL);
rb_gc_mark_movable(body->variable.coverage);
rb_gc_mark_movable(body->variable.pc2branchindex);
2021-10-28 19:10:32 +03:00
rb_gc_mark_movable(body->variable.script_lines);
rb_gc_mark_movable(body->location.label);
rb_gc_mark_movable(body->location.base_label);
rb_gc_mark_movable(body->location.pathobj);
`Primitive.mandatory_only?` for fast path Compare with the C methods, A built-in methods written in Ruby is slower if only mandatory parameters are given because it needs to check the argumens and fill default values for optional and keyword parameters (C methods can check the number of parameters with `argc`, so there are no overhead). Passing mandatory arguments are common (optional arguments are exceptional, in many cases) so it is important to provide the fast path for such common cases. `Primitive.mandatory_only?` is a special builtin function used with `if` expression like that: ```ruby def self.at(time, subsec = false, unit = :microsecond, in: nil) if Primitive.mandatory_only? Primitive.time_s_at1(time) else Primitive.time_s_at(time, subsec, unit, Primitive.arg!(:in)) end end ``` and it makes two ISeq, ``` def self.at(time, subsec = false, unit = :microsecond, in: nil) Primitive.time_s_at(time, subsec, unit, Primitive.arg!(:in)) end def self.at(time) Primitive.time_s_at1(time) end ``` and (2) is pointed by (1). Note that `Primitive.mandatory_only?` should be used only in a condition of an `if` statement and the `if` statement should be equal to the methdo body (you can not put any expression before and after the `if` statement). A method entry with `mandatory_only?` (`Time.at` on the above case) is marked as `iseq_overload`. When the method will be dispatch only with mandatory arguments (`Time.at(0)` for example), make another method entry with ISeq (2) as mandatory only method entry and it will be cached in an inline method cache. The idea is similar discussed in https://bugs.ruby-lang.org/issues/16254 but it only checks mandatory parameters or more, because many cases only mandatory parameters are given. If we find other cases (optional or keyword parameters are used frequently and it hurts performance), we can extend the feature.
2021-11-12 20:12:20 +03:00
RUBY_MARK_MOVABLE_UNLESS_NULL((VALUE)body->mandatory_only_iseq);
RUBY_MARK_MOVABLE_UNLESS_NULL((VALUE)body->parent_iseq);
if (body->call_data) {
struct rb_call_data *cds = (struct rb_call_data *)body->call_data;
for (unsigned int i=0; i<body->ci_size; i++) {
const struct rb_callinfo *ci = cds[i].ci;
const struct rb_callcache *cc = cds[i].cc;
if (vm_ci_markable(ci)) {
rb_gc_mark_movable((VALUE)ci);
}
if (cc) {
VM_ASSERT((cc->flags & VM_CALLCACHE_ON_STACK) == 0);
if (vm_cc_markable(cc)) {
if (!vm_cc_invalidated_p(cc)) {
rb_gc_mark_movable((VALUE)cc);
}
else {
cds[i].cc = rb_vm_empty_cc();
}
}
}
}
}
if (body->param.flags.has_kw && ISEQ_COMPILE_DATA(iseq) == NULL) {
const struct rb_iseq_param_keyword *const keyword = body->param.keyword;
int i, j;
i = keyword->required_num;
for (j = 0; i < keyword->num; i++, j++) {
VALUE obj = keyword->default_values[j];
if (!SPECIAL_CONST_P(obj)) {
rb_gc_mark_movable(obj);
}
}
}
2022-07-21 19:23:58 +03:00
if (body->catch_table) {
const struct iseq_catch_table *table = body->catch_table;
unsigned int i;
2019-09-27 04:20:56 +03:00
for (i = 0; i < table->size; i++) {
const struct iseq_catch_table_entry *entry;
entry = UNALIGNED_MEMBER_PTR(table, entries[i]);
if (entry->iseq) {
rb_gc_mark_movable((VALUE)entry->iseq);
}
}
}
#if USE_MJIT
mjit_mark_cc_entries(body);
#endif
#if USE_YJIT
Rust YJIT In December 2021, we opened an [issue] to solicit feedback regarding the porting of the YJIT codebase from C99 to Rust. There were some reservations, but this project was given the go ahead by Ruby core developers and Matz. Since then, we have successfully completed the port of YJIT to Rust. The new Rust version of YJIT has reached parity with the C version, in that it passes all the CRuby tests, is able to run all of the YJIT benchmarks, and performs similarly to the C version (because it works the same way and largely generates the same machine code). We've even incorporated some design improvements, such as a more fine-grained constant invalidation mechanism which we expect will make a big difference in Ruby on Rails applications. Because we want to be careful, YJIT is guarded behind a configure option: ```shell ./configure --enable-yjit # Build YJIT in release mode ./configure --enable-yjit=dev # Build YJIT in dev/debug mode ``` By default, YJIT does not get compiled and cargo/rustc is not required. If YJIT is built in dev mode, then `cargo` is used to fetch development dependencies, but when building in release, `cargo` is not required, only `rustc`. At the moment YJIT requires Rust 1.60.0 or newer. The YJIT command-line options remain mostly unchanged, and more details about the build process are documented in `doc/yjit/yjit.md`. The CI tests have been updated and do not take any more resources than before. The development history of the Rust port is available at the following commit for interested parties: https://github.com/Shopify/ruby/commit/1fd9573d8b4b65219f1c2407f30a0a60e537f8be Our hope is that Rust YJIT will be compiled and included as a part of system packages and compiled binaries of the Ruby 3.2 release. We do not anticipate any major problems as Rust is well supported on every platform which YJIT supports, but to make sure that this process works smoothly, we would like to reach out to those who take care of building systems packages before the 3.2 release is shipped and resolve any issues that may come up. [issue]: https://bugs.ruby-lang.org/issues/18481 Co-authored-by: Maxime Chevalier-Boisvert <maximechevalierb@gmail.com> Co-authored-by: Noah Gibbs <the.codefolio.guy@gmail.com> Co-authored-by: Kevin Newton <kddnewton@gmail.com>
2022-04-19 21:40:21 +03:00
rb_yjit_iseq_mark(body->yjit_payload);
#endif
}
if (FL_TEST_RAW((VALUE)iseq, ISEQ_NOT_LOADED_YET)) {
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
rb_gc_mark(iseq->aux.loader.obj);
}
else if (FL_TEST_RAW((VALUE)iseq, ISEQ_USE_COMPILE_DATA)) {
const struct iseq_compile_data *const compile_data = ISEQ_COMPILE_DATA(iseq);
rb_iseq_mark_insn_storage(compile_data->insn.storage_head);
RUBY_MARK_UNLESS_NULL(compile_data->err_info);
if (RTEST(compile_data->catch_table_ary)) {
rb_gc_mark(compile_data->catch_table_ary);
}
VM_ASSERT(compile_data != NULL);
}
else {
/* executable */
VM_ASSERT(ISEQ_EXECUTABLE_P(iseq));
if (iseq->aux.exec.local_hooks) {
rb_hook_list_mark(iseq->aux.exec.local_hooks);
}
}
RUBY_MARK_LEAVE("iseq");
}
static size_t
param_keyword_size(const struct rb_iseq_param_keyword *pkw)
{
size_t size = 0;
if (!pkw) return size;
size += sizeof(struct rb_iseq_param_keyword);
size += sizeof(VALUE) * (pkw->num - pkw->required_num);
return size;
}
size_t
rb_iseq_memsize(const rb_iseq_t *iseq)
{
size_t size = 0; /* struct already counted as RVALUE size */
const struct rb_iseq_constant_body *body = ISEQ_BODY(iseq);
const struct iseq_compile_data *compile_data;
/* TODO: should we count original_iseq? */
2019-07-23 10:42:20 +03:00
if (ISEQ_EXECUTABLE_P(iseq) && body) {
size += sizeof(struct rb_iseq_constant_body);
size += body->iseq_size * sizeof(VALUE);
size += body->insns_info.size * (sizeof(struct iseq_insn_info_entry) + sizeof(unsigned int));
size += body->local_table_size * sizeof(ID);
size += ISEQ_MBITS_BUFLEN(body->iseq_size) * ISEQ_MBITS_SIZE;
2019-07-23 10:42:20 +03:00
if (body->catch_table) {
size += iseq_catch_table_bytes(body->catch_table->size);
}
size += (body->param.opt_num + 1) * sizeof(VALUE);
size += param_keyword_size(body->param.keyword);
2019-07-23 10:42:20 +03:00
/* body->is_entries */
size += ISEQ_IS_SIZE(body) * sizeof(union iseq_inline_storage_entry);
if (ISEQ_BODY(iseq)->is_entries) {
/* IC entries constant segments */
for (unsigned int ic_idx = 0; ic_idx < body->ic_size; ic_idx++) {
IC ic = &ISEQ_IS_IC_ENTRY(body, ic_idx);
const ID *ids = ic->segments;
if (!ids) continue;
while (*ids++) {
size += sizeof(ID);
}
size += sizeof(ID); // null terminator
New constant caching insn: opt_getconstant_path Previously YARV bytecode implemented constant caching by having a pair of instructions, opt_getinlinecache and opt_setinlinecache, wrapping a series of getconstant calls (with putobject providing supporting arguments). This commit replaces that pattern with a new instruction, opt_getconstant_path, handling both getting/setting the inline cache and fetching the constant on a cache miss. This is implemented by storing the full constant path as a null-terminated array of IDs inside of the IC structure. idNULL is used to signal an absolute constant reference. $ ./miniruby --dump=insns -e '::Foo::Bar::Baz' == disasm: #<ISeq:<main>@-e:1 (1,0)-(1,13)> (catch: FALSE) 0000 opt_getconstant_path <ic:0 ::Foo::Bar::Baz> ( 1)[Li] 0002 leave The motivation for this is that we had increasingly found the need to disassemble the instructions between the opt_getinlinecache and opt_setinlinecache in order to determine the constant we are fetching, or otherwise store metadata. This disassembly was done: * In opt_setinlinecache, to register the IC against the constant names it is using for granular invalidation. * In rb_iseq_free, to unregister the IC from the invalidation table. * In YJIT to find the position of a opt_getinlinecache instruction to invalidate it when the cache is populated * In YJIT to register the constant names being used for invalidation. With this change we no longe need disassemly for these (in fact rb_iseq_each is now unused), as the list of constant names being referenced is held in the IC. This should also make it possible to make more optimizations in the future. This may also reduce the size of iseqs, as previously each segment required 32 bytes (on 64-bit platforms) for each constant segment. This implementation only stores one ID per-segment. There should be no significant performance change between this and the previous implementation. Previously opt_getinlinecache was a "leaf" instruction, but it included a jump (almost always to a separate cache line). Now opt_getconstant_path is a non-leaf (it may raise/autoload/call const_missing) but it does not jump. These seem to even out.
2022-08-10 20:35:48 +03:00
}
}
/* body->call_data */
size += body->ci_size * sizeof(struct rb_call_data);
// TODO: should we count imemo_callinfo?
}
compile_data = ISEQ_COMPILE_DATA(iseq);
if (compile_data) {
struct iseq_compile_data_storage *cur;
size += sizeof(struct iseq_compile_data);
cur = compile_data->node.storage_head;
while (cur) {
size += cur->size + offsetof(struct iseq_compile_data_storage, buff);
cur = cur->next;
}
}
return size;
}
struct rb_iseq_constant_body *
rb_iseq_constant_body_alloc(void)
{
struct rb_iseq_constant_body *iseq_body;
iseq_body = ZALLOC(struct rb_iseq_constant_body);
return iseq_body;
}
2015-07-22 01:52:59 +03:00
static rb_iseq_t *
iseq_alloc(void)
{
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
rb_iseq_t *iseq = iseq_imemo_alloc();
ISEQ_BODY(iseq) = rb_iseq_constant_body_alloc();
2015-07-22 01:52:59 +03:00
return iseq;
}
VALUE
rb_iseq_pathobj_new(VALUE path, VALUE realpath)
{
VALUE pathobj;
VM_ASSERT(RB_TYPE_P(path, T_STRING));
2021-10-03 16:34:45 +03:00
VM_ASSERT(NIL_P(realpath) || RB_TYPE_P(realpath, T_STRING));
if (path == realpath ||
(!NIL_P(realpath) && rb_str_cmp(path, realpath) == 0)) {
pathobj = rb_fstring(path);
}
else {
if (!NIL_P(realpath)) realpath = rb_fstring(realpath);
pathobj = rb_ary_new_from_args(2, rb_fstring(path), realpath);
rb_obj_freeze(pathobj);
}
return pathobj;
}
void
rb_iseq_pathobj_set(const rb_iseq_t *iseq, VALUE path, VALUE realpath)
{
RB_OBJ_WRITE(iseq, &ISEQ_BODY(iseq)->location.pathobj,
rb_iseq_pathobj_new(path, realpath));
}
static rb_iseq_location_t *
2022-09-25 10:45:28 +03:00
iseq_location_setup(rb_iseq_t *iseq, VALUE name, VALUE path, VALUE realpath, int first_lineno, const rb_code_location_t *code_location, const int node_id)
{
rb_iseq_location_t *loc = &ISEQ_BODY(iseq)->location;
rb_iseq_pathobj_set(iseq, path, realpath);
2015-07-22 01:52:59 +03:00
RB_OBJ_WRITE(iseq, &loc->label, name);
RB_OBJ_WRITE(iseq, &loc->base_label, name);
loc->first_lineno = first_lineno;
if (code_location) {
loc->node_id = node_id;
loc->code_location = *code_location;
}
else {
loc->code_location.beg_pos.lineno = 0;
loc->code_location.beg_pos.column = 0;
loc->code_location.end_pos.lineno = -1;
loc->code_location.end_pos.column = -1;
}
return loc;
}
static void
2015-07-22 01:52:59 +03:00
set_relation(rb_iseq_t *iseq, const rb_iseq_t *piseq)
{
struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
const VALUE type = body->type;
/* set class nest stack */
if (type == ISEQ_TYPE_TOP) {
body->local_iseq = iseq;
}
else if (type == ISEQ_TYPE_METHOD || type == ISEQ_TYPE_CLASS) {
body->local_iseq = iseq;
}
2015-07-22 01:52:59 +03:00
else if (piseq) {
body->local_iseq = ISEQ_BODY(piseq)->local_iseq;
}
2015-07-22 01:52:59 +03:00
if (piseq) {
body->parent_iseq = piseq;
}
if (type == ISEQ_TYPE_MAIN) {
body->local_iseq = iseq;
}
}
static struct iseq_compile_data_storage *
new_arena(void)
{
struct iseq_compile_data_storage * new_arena =
(struct iseq_compile_data_storage *)
ALLOC_N(char, INITIAL_ISEQ_COMPILE_DATA_STORAGE_BUFF_SIZE +
offsetof(struct iseq_compile_data_storage, buff));
new_arena->pos = 0;
new_arena->next = 0;
new_arena->size = INITIAL_ISEQ_COMPILE_DATA_STORAGE_BUFF_SIZE;
return new_arena;
}
static VALUE
prepare_iseq_build(rb_iseq_t *iseq,
2022-09-25 10:45:28 +03:00
VALUE name, VALUE path, VALUE realpath, int first_lineno, const rb_code_location_t *code_location, const int node_id,
const rb_iseq_t *parent, int isolated_depth, enum rb_iseq_type type,
VALUE script_lines, const rb_compile_option_t *option)
{
VALUE coverage = Qfalse;
VALUE err_info = Qnil;
struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
if (parent && (type == ISEQ_TYPE_MAIN || type == ISEQ_TYPE_TOP))
err_info = Qfalse;
body->type = type;
set_relation(iseq, parent);
name = rb_fstring(name);
iseq_location_setup(iseq, name, path, realpath, first_lineno, code_location, node_id);
if (iseq != body->local_iseq) {
RB_OBJ_WRITE(iseq, &body->location.base_label, ISEQ_BODY(body->local_iseq)->location.label);
}
ISEQ_COVERAGE_SET(iseq, Qnil);
ISEQ_ORIGINAL_ISEQ_CLEAR(iseq);
body->variable.flip_count = 0;
2021-09-30 11:30:04 +03:00
if (NIL_P(script_lines)) {
RB_OBJ_WRITE(iseq, &body->variable.script_lines, Qnil);
}
else {
RB_OBJ_WRITE(iseq, &body->variable.script_lines, rb_ractor_make_shareable(script_lines));
}
ISEQ_COMPILE_DATA_ALLOC(iseq);
RB_OBJ_WRITE(iseq, &ISEQ_COMPILE_DATA(iseq)->err_info, err_info);
RB_OBJ_WRITE(iseq, &ISEQ_COMPILE_DATA(iseq)->catch_table_ary, Qnil);
ISEQ_COMPILE_DATA(iseq)->node.storage_head = ISEQ_COMPILE_DATA(iseq)->node.storage_current = new_arena();
ISEQ_COMPILE_DATA(iseq)->insn.storage_head = ISEQ_COMPILE_DATA(iseq)->insn.storage_current = new_arena();
ISEQ_COMPILE_DATA(iseq)->isolated_depth = isolated_depth;
ISEQ_COMPILE_DATA(iseq)->option = option;
ISEQ_COMPILE_DATA(iseq)->ivar_cache_table = NULL;
2019-11-07 10:58:00 +03:00
ISEQ_COMPILE_DATA(iseq)->builtin_function_table = GET_VM()->builtin_function_table;
2020-10-30 06:26:59 +03:00
if (option->coverage_enabled) {
VALUE coverages = rb_get_coverages();
if (RTEST(coverages)) {
coverage = rb_hash_lookup(coverages, rb_iseq_path(iseq));
if (NIL_P(coverage)) coverage = Qfalse;
}
}
ISEQ_COVERAGE_SET(iseq, coverage);
if (coverage && ISEQ_BRANCH_COVERAGE(iseq))
ISEQ_PC2BRANCHINDEX_SET(iseq, rb_ary_hidden_new(0));
return Qtrue;
}
#if VM_CHECK_MODE > 0 && VM_INSN_INFO_TABLE_IMPL > 0
static void validate_get_insn_info(const rb_iseq_t *iseq);
#endif
void
rb_iseq_insns_info_encode_positions(const rb_iseq_t *iseq)
{
#if VM_INSN_INFO_TABLE_IMPL == 2
/* create succ_index_table */
struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
int size = body->insns_info.size;
int max_pos = body->iseq_size;
int *data = (int *)body->insns_info.positions;
if (body->insns_info.succ_index_table) ruby_xfree(body->insns_info.succ_index_table);
body->insns_info.succ_index_table = succ_index_table_create(max_pos, data, size);
#if VM_CHECK_MODE == 0
ruby_xfree(body->insns_info.positions);
body->insns_info.positions = NULL;
#endif
#endif
}
#if VM_INSN_INFO_TABLE_IMPL == 2
unsigned int *
rb_iseq_insns_info_decode_positions(const struct rb_iseq_constant_body *body)
{
int size = body->insns_info.size;
int max_pos = body->iseq_size;
struct succ_index_table *sd = body->insns_info.succ_index_table;
return succ_index_table_invert(max_pos, sd, size);
}
#endif
void
rb_iseq_init_trace(rb_iseq_t *iseq)
{
iseq->aux.exec.global_trace_events = 0;
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
if (ruby_vm_event_enabled_global_flags & ISEQ_TRACE_EVENTS) {
rb_iseq_trace_set(iseq, ruby_vm_event_enabled_global_flags & ISEQ_TRACE_EVENTS);
}
}
static VALUE
finish_iseq_build(rb_iseq_t *iseq)
{
struct iseq_compile_data *data = ISEQ_COMPILE_DATA(iseq);
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
VALUE err = data->err_info;
ISEQ_COMPILE_DATA_CLEAR(iseq);
compile_data_free(data);
#if VM_CHECK_MODE > 0 && VM_INSN_INFO_TABLE_IMPL > 0
validate_get_insn_info(iseq);
#endif
if (RTEST(err)) {
VALUE path = pathobj_path(body->location.pathobj);
if (err == Qtrue) err = rb_exc_new_cstr(rb_eSyntaxError, "compile error");
rb_funcallv(err, rb_intern("set_backtrace"), 1, &path);
rb_exc_raise(err);
}
RB_DEBUG_COUNTER_INC(iseq_num);
RB_DEBUG_COUNTER_ADD(iseq_cd_num, ISEQ_BODY(iseq)->ci_size);
rb_iseq_init_trace(iseq);
return Qtrue;
}
static rb_compile_option_t COMPILE_OPTION_DEFAULT = {
OPT_INLINE_CONST_CACHE, /* int inline_const_cache; */
OPT_PEEPHOLE_OPTIMIZATION, /* int peephole_optimization; */
OPT_TAILCALL_OPTIMIZATION, /* int tailcall_optimization */
OPT_SPECIALISED_INSTRUCTION, /* int specialized_instruction; */
OPT_OPERANDS_UNIFICATION, /* int operands_unification; */
OPT_INSTRUCTIONS_UNIFICATION, /* int instructions_unification; */
OPT_STACK_CACHING, /* int stack_caching; */
OPT_FROZEN_STRING_LITERAL,
OPT_DEBUG_FROZEN_STRING_LITERAL,
TRUE, /* coverage_enabled */
};
static const rb_compile_option_t COMPILE_OPTION_FALSE = {0};
static void
set_compile_option_from_hash(rb_compile_option_t *option, VALUE opt)
{
#define SET_COMPILE_OPTION(o, h, mem) \
{ VALUE flag = rb_hash_aref((h), ID2SYM(rb_intern(#mem))); \
if (flag == Qtrue) { (o)->mem = 1; } \
else if (flag == Qfalse) { (o)->mem = 0; } \
}
#define SET_COMPILE_OPTION_NUM(o, h, mem) \
{ VALUE num = rb_hash_aref(opt, ID2SYM(rb_intern(#mem))); \
if (!NIL_P(num)) (o)->mem = NUM2INT(num); \
}
SET_COMPILE_OPTION(option, opt, inline_const_cache);
SET_COMPILE_OPTION(option, opt, peephole_optimization);
SET_COMPILE_OPTION(option, opt, tailcall_optimization);
SET_COMPILE_OPTION(option, opt, specialized_instruction);
SET_COMPILE_OPTION(option, opt, operands_unification);
SET_COMPILE_OPTION(option, opt, instructions_unification);
SET_COMPILE_OPTION(option, opt, stack_caching);
SET_COMPILE_OPTION(option, opt, frozen_string_literal);
SET_COMPILE_OPTION(option, opt, debug_frozen_string_literal);
SET_COMPILE_OPTION(option, opt, coverage_enabled);
SET_COMPILE_OPTION_NUM(option, opt, debug_level);
#undef SET_COMPILE_OPTION
#undef SET_COMPILE_OPTION_NUM
}
static void
rb_iseq_make_compile_option(rb_compile_option_t *option, VALUE opt)
{
Check_Type(opt, T_HASH);
set_compile_option_from_hash(option, opt);
}
static void
make_compile_option(rb_compile_option_t *option, VALUE opt)
{
2021-10-03 16:34:45 +03:00
if (NIL_P(opt)) {
*option = COMPILE_OPTION_DEFAULT;
}
else if (opt == Qfalse) {
*option = COMPILE_OPTION_FALSE;
}
else if (opt == Qtrue) {
int i;
for (i = 0; i < (int)(sizeof(rb_compile_option_t) / sizeof(int)); ++i)
((int *)option)[i] = 1;
}
else if (RB_TYPE_P(opt, T_HASH)) {
*option = COMPILE_OPTION_DEFAULT;
set_compile_option_from_hash(option, opt);
}
else {
rb_raise(rb_eTypeError, "Compile option must be Hash/true/false/nil");
}
}
static VALUE
make_compile_option_value(rb_compile_option_t *option)
{
2019-10-22 01:28:41 +03:00
VALUE opt = rb_hash_new_with_size(11);
#define SET_COMPILE_OPTION(o, h, mem) \
2021-08-31 14:30:35 +03:00
rb_hash_aset((h), ID2SYM(rb_intern(#mem)), RBOOL((o)->mem))
#define SET_COMPILE_OPTION_NUM(o, h, mem) \
rb_hash_aset((h), ID2SYM(rb_intern(#mem)), INT2NUM((o)->mem))
{
SET_COMPILE_OPTION(option, opt, inline_const_cache);
SET_COMPILE_OPTION(option, opt, peephole_optimization);
SET_COMPILE_OPTION(option, opt, tailcall_optimization);
SET_COMPILE_OPTION(option, opt, specialized_instruction);
SET_COMPILE_OPTION(option, opt, operands_unification);
SET_COMPILE_OPTION(option, opt, instructions_unification);
SET_COMPILE_OPTION(option, opt, stack_caching);
SET_COMPILE_OPTION(option, opt, frozen_string_literal);
SET_COMPILE_OPTION(option, opt, debug_frozen_string_literal);
SET_COMPILE_OPTION(option, opt, coverage_enabled);
SET_COMPILE_OPTION_NUM(option, opt, debug_level);
}
#undef SET_COMPILE_OPTION
#undef SET_COMPILE_OPTION_NUM
return opt;
}
2015-07-22 01:52:59 +03:00
rb_iseq_t *
rb_iseq_new(const rb_ast_body_t *ast, VALUE name, VALUE path, VALUE realpath,
const rb_iseq_t *parent, enum rb_iseq_type type)
{
2022-09-25 10:45:28 +03:00
return rb_iseq_new_with_opt(ast, name, path, realpath, 0, parent,
0, type, &COMPILE_OPTION_DEFAULT);
}
static int
ast_line_count(const rb_ast_body_t *ast)
{
if (ast->script_lines == Qfalse) {
// this occurs when failed to parse the source code with a syntax error
return 0;
}
if (RB_TYPE_P(ast->script_lines, T_ARRAY)){
return (int)RARRAY_LEN(ast->script_lines);
}
return FIX2INT(ast->script_lines);
}
static VALUE
iseq_setup_coverage(VALUE coverages, VALUE path, const rb_ast_body_t *ast, int line_offset)
{
int line_count = line_offset + ast_line_count(ast);
if (line_count >= 0) {
int len = (rb_get_coverage_mode() & COVERAGE_TARGET_ONESHOT_LINES) ? 0 : line_count;
VALUE coverage = rb_default_coverage(len);
rb_hash_aset(coverages, path, coverage);
return coverage;
}
return Qnil;
}
2015-07-22 01:52:59 +03:00
rb_iseq_t *
rb_iseq_new_top(const rb_ast_body_t *ast, VALUE name, VALUE path, VALUE realpath, const rb_iseq_t *parent)
{
VALUE coverages = rb_get_coverages();
if (RTEST(coverages)) {
iseq_setup_coverage(coverages, path, ast, 0);
}
2022-09-25 10:45:28 +03:00
return rb_iseq_new_with_opt(ast, name, path, realpath, 0, parent, 0,
ISEQ_TYPE_TOP, &COMPILE_OPTION_DEFAULT);
}
2015-07-22 01:52:59 +03:00
rb_iseq_t *
rb_iseq_new_main(const rb_ast_body_t *ast, VALUE path, VALUE realpath, const rb_iseq_t *parent, int opt)
{
return rb_iseq_new_with_opt(ast, rb_fstring_lit("<main>"),
2022-09-25 10:45:28 +03:00
path, realpath, 0,
parent, 0, ISEQ_TYPE_MAIN, opt ? &COMPILE_OPTION_DEFAULT : &COMPILE_OPTION_FALSE);
}
rb_iseq_t *
2022-09-25 10:45:28 +03:00
rb_iseq_new_eval(const rb_ast_body_t *ast, VALUE name, VALUE path, VALUE realpath, int first_lineno, const rb_iseq_t *parent, int isolated_depth)
{
if (rb_get_coverage_mode() & COVERAGE_TARGET_EVAL) {
VALUE coverages = rb_get_coverages();
if (RTEST(coverages) && RTEST(path) && !RTEST(rb_hash_has_key(coverages, path))) {
iseq_setup_coverage(coverages, path, ast, first_lineno - 1);
}
2022-09-17 11:19:57 +03:00
}
return rb_iseq_new_with_opt(ast, name, path, realpath, first_lineno,
parent, isolated_depth, ISEQ_TYPE_EVAL, &COMPILE_OPTION_DEFAULT);
}
static inline rb_iseq_t *
iseq_translate(rb_iseq_t *iseq)
{
if (rb_respond_to(rb_cISeq, rb_intern("translate"))) {
VALUE v1 = iseqw_new(iseq);
VALUE v2 = rb_funcall(rb_cISeq, rb_intern("translate"), 1, v1);
if (v1 != v2 && CLASS_OF(v2) == rb_cISeq) {
iseq = (rb_iseq_t *)iseqw_check(v2);
}
}
return iseq;
}
2015-07-22 01:52:59 +03:00
rb_iseq_t *
rb_iseq_new_with_opt(const rb_ast_body_t *ast, VALUE name, VALUE path, VALUE realpath,
2022-09-25 10:45:28 +03:00
int first_lineno, const rb_iseq_t *parent, int isolated_depth,
enum rb_iseq_type type, const rb_compile_option_t *option)
{
const NODE *node = ast ? ast->root : 0;
/* TODO: argument check */
2015-07-22 01:52:59 +03:00
rb_iseq_t *iseq = iseq_alloc();
rb_compile_option_t new_opt;
if (option) {
new_opt = *option;
}
else {
new_opt = COMPILE_OPTION_DEFAULT;
}
if (ast && ast->compile_option) rb_iseq_make_compile_option(&new_opt, ast->compile_option);
VALUE script_lines = Qnil;
if (ast && !FIXNUM_P(ast->script_lines) && ast->script_lines) {
script_lines = ast->script_lines;
}
else if (parent) {
script_lines = ISEQ_BODY(parent)->variable.script_lines;
}
prepare_iseq_build(iseq, name, path, realpath, first_lineno, node ? &node->nd_loc : NULL, node ? nd_node_id(node) : -1,
parent, isolated_depth, type, script_lines, &new_opt);
2015-07-22 01:52:59 +03:00
rb_iseq_compile_node(iseq, node);
finish_iseq_build(iseq);
2015-07-22 01:52:59 +03:00
return iseq_translate(iseq);
}
rb_iseq_t *
rb_iseq_new_with_callback(
const struct rb_iseq_new_with_callback_callback_func * ifunc,
VALUE name, VALUE path, VALUE realpath,
2022-09-25 10:45:28 +03:00
int first_lineno, const rb_iseq_t *parent,
enum rb_iseq_type type, const rb_compile_option_t *option)
{
/* TODO: argument check */
rb_iseq_t *iseq = iseq_alloc();
if (!option) option = &COMPILE_OPTION_DEFAULT;
prepare_iseq_build(iseq, name, path, realpath, first_lineno, NULL, -1, parent, 0, type, Qnil, option);
rb_iseq_compile_callback(iseq, ifunc);
finish_iseq_build(iseq);
return iseq;
}
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
const rb_iseq_t *
rb_iseq_load_iseq(VALUE fname)
{
VALUE iseqv = rb_check_funcall(rb_cISeq, rb_intern("load_iseq"), 1, &fname);
if (!SPECIAL_CONST_P(iseqv) && RBASIC_CLASS(iseqv) == rb_cISeq) {
return iseqw_check(iseqv);
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
}
return NULL;
}
#define CHECK_ARRAY(v) rb_to_array_type(v)
#define CHECK_HASH(v) rb_to_hash_type(v)
#define CHECK_STRING(v) rb_str_to_str(v)
#define CHECK_SYMBOL(v) rb_to_symbol_type(v)
static inline VALUE CHECK_INTEGER(VALUE v) {(void)NUM2LONG(v); return v;}
static enum rb_iseq_type
iseq_type_from_sym(VALUE type)
{
const ID id_top = rb_intern("top");
const ID id_method = rb_intern("method");
const ID id_block = rb_intern("block");
const ID id_class = rb_intern("class");
const ID id_rescue = rb_intern("rescue");
const ID id_ensure = rb_intern("ensure");
const ID id_eval = rb_intern("eval");
const ID id_main = rb_intern("main");
const ID id_plain = rb_intern("plain");
/* ensure all symbols are static or pinned down before
* conversion */
const ID typeid = rb_check_id(&type);
if (typeid == id_top) return ISEQ_TYPE_TOP;
if (typeid == id_method) return ISEQ_TYPE_METHOD;
if (typeid == id_block) return ISEQ_TYPE_BLOCK;
if (typeid == id_class) return ISEQ_TYPE_CLASS;
if (typeid == id_rescue) return ISEQ_TYPE_RESCUE;
if (typeid == id_ensure) return ISEQ_TYPE_ENSURE;
if (typeid == id_eval) return ISEQ_TYPE_EVAL;
if (typeid == id_main) return ISEQ_TYPE_MAIN;
if (typeid == id_plain) return ISEQ_TYPE_PLAIN;
return (enum rb_iseq_type)-1;
}
static VALUE
2015-07-22 01:52:59 +03:00
iseq_load(VALUE data, const rb_iseq_t *parent, VALUE opt)
{
2015-07-22 01:52:59 +03:00
rb_iseq_t *iseq = iseq_alloc();
VALUE magic, version1, version2, format_type, misc;
2022-09-25 10:45:28 +03:00
VALUE name, path, realpath, code_location, node_id;
VALUE type, body, locals, params, exception;
st_data_t iseq_type;
rb_compile_option_t option;
int i = 0;
rb_code_location_t tmp_loc = { {0, 0}, {-1, -1} };
/* [magic, major_version, minor_version, format_type, misc,
* label, path, first_lineno,
* type, locals, args, exception_table, body]
*/
data = CHECK_ARRAY(data);
magic = CHECK_STRING(rb_ary_entry(data, i++));
version1 = CHECK_INTEGER(rb_ary_entry(data, i++));
version2 = CHECK_INTEGER(rb_ary_entry(data, i++));
format_type = CHECK_INTEGER(rb_ary_entry(data, i++));
misc = CHECK_HASH(rb_ary_entry(data, i++));
((void)magic, (void)version1, (void)version2, (void)format_type);
name = CHECK_STRING(rb_ary_entry(data, i++));
path = CHECK_STRING(rb_ary_entry(data, i++));
realpath = rb_ary_entry(data, i++);
realpath = NIL_P(realpath) ? Qnil : CHECK_STRING(realpath);
2022-09-25 10:45:28 +03:00
int first_lineno = RB_NUM2INT(rb_ary_entry(data, i++));
type = CHECK_SYMBOL(rb_ary_entry(data, i++));
locals = CHECK_ARRAY(rb_ary_entry(data, i++));
params = CHECK_HASH(rb_ary_entry(data, i++));
exception = CHECK_ARRAY(rb_ary_entry(data, i++));
body = CHECK_ARRAY(rb_ary_entry(data, i++));
ISEQ_BODY(iseq)->local_iseq = iseq;
iseq_type = iseq_type_from_sym(type);
if (iseq_type == (enum rb_iseq_type)-1) {
rb_raise(rb_eTypeError, "unsupported type: :%"PRIsVALUE, rb_sym2str(type));
}
node_id = rb_hash_aref(misc, ID2SYM(rb_intern("node_id")));
code_location = rb_hash_aref(misc, ID2SYM(rb_intern("code_location")));
if (RB_TYPE_P(code_location, T_ARRAY) && RARRAY_LEN(code_location) == 4) {
tmp_loc.beg_pos.lineno = NUM2INT(rb_ary_entry(code_location, 0));
tmp_loc.beg_pos.column = NUM2INT(rb_ary_entry(code_location, 1));
tmp_loc.end_pos.lineno = NUM2INT(rb_ary_entry(code_location, 2));
tmp_loc.end_pos.column = NUM2INT(rb_ary_entry(code_location, 3));
}
make_compile_option(&option, opt);
option.peephole_optimization = FALSE; /* because peephole optimization can modify original iseq */
prepare_iseq_build(iseq, name, path, realpath, first_lineno, &tmp_loc, NUM2INT(node_id),
parent, 0, (enum rb_iseq_type)iseq_type, Qnil, &option);
rb_iseq_build_from_ary(iseq, misc, locals, params, exception, body);
finish_iseq_build(iseq);
2015-07-22 01:52:59 +03:00
return iseqw_new(iseq);
}
/*
* :nodoc:
*/
static VALUE
iseq_s_load(int argc, VALUE *argv, VALUE self)
{
VALUE data, opt=Qnil;
rb_scan_args(argc, argv, "11", &data, &opt);
2015-07-22 01:52:59 +03:00
return iseq_load(data, NULL, opt);
}
VALUE
* iseq.c (rb_iseq_load): renamed from ruby_iseq_load, since it is for C extensions or the ruby core. [ruby-core:21407] Index: compile.c =================================================================== --- compile.c (revision 21649) +++ compile.c (working copy) @@ -5078,5 +5078,5 @@ iseq_build_exception(rb_iseq_t *iseq, st } else { - eiseqval = ruby_iseq_load(ptr[1], iseq->self, Qnil); + eiseqval = rb_iseq_load(ptr[1], iseq->self, Qnil); } @@ -5162,5 +5162,5 @@ iseq_build_body(rb_iseq_t *iseq, LINK_AN if (op != Qnil) { if (TYPE(op) == T_ARRAY) { - argv[j] = ruby_iseq_load(op, iseq->self, Qnil); + argv[j] = rb_iseq_load(op, iseq->self, Qnil); } else if (CLASS_OF(op) == rb_cISeq) { Index: iseq.c =================================================================== --- iseq.c (revision 21649) +++ iseq.c (working copy) @@ -448,5 +448,5 @@ iseq_s_load(int argc, VALUE *argv, VALUE VALUE -ruby_iseq_load(VALUE data, VALUE parent, VALUE opt) +rb_iseq_load(VALUE data, VALUE parent, VALUE opt) { return iseq_load(rb_cISeq, data, parent, opt); Index: iseq.h =================================================================== --- iseq.h (revision 21649) +++ iseq.h (working copy) @@ -21,5 +21,5 @@ VALUE ruby_iseq_build_from_ary(rb_iseq_t /* iseq.c */ -VALUE ruby_iseq_load(VALUE data, VALUE parent, VALUE opt); +VALUE rb_iseq_load(VALUE data, VALUE parent, VALUE opt); struct st_table *ruby_insn_make_insn_table(void); git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@21650 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2009-01-18 22:05:15 +03:00
rb_iseq_load(VALUE data, VALUE parent, VALUE opt)
{
2015-07-22 01:52:59 +03:00
return iseq_load(data, RTEST(parent) ? (rb_iseq_t *)parent : NULL, opt);
}
static rb_iseq_t *
rb_iseq_compile_with_option(VALUE src, VALUE file, VALUE realpath, VALUE line, VALUE opt)
{
2015-07-22 01:52:59 +03:00
rb_iseq_t *iseq = NULL;
rb_compile_option_t option;
#if !defined(__GNUC__) || (__GNUC__ == 4 && __GNUC_MINOR__ == 8)
# define INITIALIZED volatile /* suppress warnings by gcc 4.8 */
#else
# define INITIALIZED /* volatile */
#endif
rb_ast_t *(*parse)(VALUE vparser, VALUE fname, VALUE file, int start);
int ln;
rb_ast_t *INITIALIZED ast;
VALUE name = rb_fstring_lit("<compiled>");
/* safe results first */
make_compile_option(&option, opt);
ln = NUM2INT(line);
StringValueCStr(file);
if (RB_TYPE_P(src, T_FILE)) {
parse = rb_parser_compile_file_path;
}
else {
parse = rb_parser_compile_string_path;
StringValue(src);
}
{
const VALUE parser = rb_parser_new();
const rb_iseq_t *outer_scope = rb_iseq_new(NULL, name, name, Qnil, 0, ISEQ_TYPE_TOP);
VALUE outer_scope_v = (VALUE)outer_scope;
rb_parser_set_context(parser, outer_scope, FALSE);
RB_GC_GUARD(outer_scope_v);
ast = (*parse)(parser, file, src, ln);
}
if (!ast->body.root) {
rb_ast_dispose(ast);
rb_exc_raise(GET_EC()->errinfo);
}
else {
2022-09-25 10:45:28 +03:00
iseq = rb_iseq_new_with_opt(&ast->body, name, file, realpath, ln,
NULL, 0, ISEQ_TYPE_TOP, &option);
rb_ast_dispose(ast);
}
* iseq.c, vm_eval.c: set th->base_block properly. th->base_block is information for (a) parsing, (b) compiling and (c) setting up the frame to execute the program passed by `eval' method. For example, (1) parser need to know up-level variables to detect it is variable or method without paren. Befor (a), (b) and (c), VM set th->base_block by passed bindng (or previous frame information). After execute (a), (b) and (c), VM should clear th->base_block. However, if (a), (b) or (c) raises an exception, then th->base_block is not cleared. Problem is that the uncleared value th->balo_block is used for irrelevant iseq compilation. It causes SEGV or critical error. I tried to solve this problem: to clear them before exception, but finally I found out that it is difficult to do it (Ruby program can be run in many places). Because of this background, I set th->base_block before compiling iseq and restore it after compiling. Basically, th->base_block is dirty hack (similar to global variable) and this patch is also dirty. * bootstraptest/test_eval.rb: add a test for above. * internal.h: remove unused decl. * iseq.c (rb_iseq_compile_with_option): add base_block parameter. set th->base_block before compation and restore it after compilation. * ruby.c (require_libraries): pass 0 as base_block instead of setting th->base_block * tool/compile_prelude.rb (prelude_eval): apply above changes. * vm.c, vm_eval.c: ditto. * vm_core.h: add comments. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@36179 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2012-06-22 13:32:56 +04:00
2015-07-22 01:52:59 +03:00
return iseq;
}
2015-07-22 01:52:59 +03:00
VALUE
rb_iseq_path(const rb_iseq_t *iseq)
{
return pathobj_path(ISEQ_BODY(iseq)->location.pathobj);
}
VALUE
rb_iseq_realpath(const rb_iseq_t *iseq)
{
return pathobj_realpath(ISEQ_BODY(iseq)->location.pathobj);
2015-07-22 01:52:59 +03:00
}
VALUE
rb_iseq_absolute_path(const rb_iseq_t *iseq)
{
return rb_iseq_realpath(iseq);
2015-07-22 01:52:59 +03:00
}
int
rb_iseq_from_eval_p(const rb_iseq_t *iseq)
{
return NIL_P(rb_iseq_realpath(iseq));
}
2015-07-22 01:52:59 +03:00
VALUE
rb_iseq_label(const rb_iseq_t *iseq)
{
return ISEQ_BODY(iseq)->location.label;
2015-07-22 01:52:59 +03:00
}
VALUE
rb_iseq_base_label(const rb_iseq_t *iseq)
{
return ISEQ_BODY(iseq)->location.base_label;
2015-07-22 01:52:59 +03:00
}
VALUE
rb_iseq_first_lineno(const rb_iseq_t *iseq)
{
return RB_INT2NUM(ISEQ_BODY(iseq)->location.first_lineno);
2015-07-22 01:52:59 +03:00
}
VALUE
rb_iseq_method_name(const rb_iseq_t *iseq)
{
struct rb_iseq_constant_body *const body = ISEQ_BODY(ISEQ_BODY(iseq)->local_iseq);
2015-07-22 01:52:59 +03:00
if (body->type == ISEQ_TYPE_METHOD) {
return body->location.base_label;
2015-07-22 01:52:59 +03:00
}
else {
return Qnil;
}
}
void
rb_iseq_code_location(const rb_iseq_t *iseq, int *beg_pos_lineno, int *beg_pos_column, int *end_pos_lineno, int *end_pos_column)
{
const rb_code_location_t *loc = &ISEQ_BODY(iseq)->location.code_location;
if (beg_pos_lineno) *beg_pos_lineno = loc->beg_pos.lineno;
if (beg_pos_column) *beg_pos_column = loc->beg_pos.column;
if (end_pos_lineno) *end_pos_lineno = loc->end_pos.lineno;
if (end_pos_column) *end_pos_column = loc->end_pos.column;
}
static ID iseq_type_id(enum rb_iseq_type type);
VALUE
rb_iseq_type(const rb_iseq_t *iseq)
{
return ID2SYM(iseq_type_id(ISEQ_BODY(iseq)->type));
}
VALUE
rb_iseq_coverage(const rb_iseq_t *iseq)
{
return ISEQ_COVERAGE(iseq);
}
static int
remove_coverage_i(void *vstart, void *vend, size_t stride, void *data)
{
VALUE v = (VALUE)vstart;
for (; v != (VALUE)vend; v += stride) {
void *ptr = asan_poisoned_object_p(v);
asan_unpoison_object(v, false);
if (rb_obj_is_iseq(v)) {
rb_iseq_t *iseq = (rb_iseq_t *)v;
ISEQ_COVERAGE_SET(iseq, Qnil);
}
asan_poison_object_if(ptr, v);
}
return 0;
}
void
rb_iseq_remove_coverage_all(void)
{
rb_objspace_each_objects(remove_coverage_i, NULL);
}
2015-07-22 01:52:59 +03:00
/* define wrapper class methods (RubyVM::InstructionSequence) */
static void
iseqw_mark(void *ptr)
{
rb_gc_mark((VALUE)ptr);
}
static size_t
iseqw_memsize(const void *ptr)
{
return rb_iseq_memsize((const rb_iseq_t *)ptr);
}
static const rb_data_type_t iseqw_data_type = {
"T_IMEMO/iseq",
{iseqw_mark, NULL, iseqw_memsize,},
0, 0, RUBY_TYPED_FREE_IMMEDIATELY|RUBY_TYPED_WB_PROTECTED
};
2015-07-22 01:52:59 +03:00
static VALUE
iseqw_new(const rb_iseq_t *iseq)
{
if (iseq->wrapper) {
return iseq->wrapper;
}
else {
union { const rb_iseq_t *in; void *out; } deconst;
VALUE obj;
deconst.in = iseq;
obj = TypedData_Wrap_Struct(rb_cISeq, &iseqw_data_type, deconst.out);
RB_OBJ_WRITTEN(obj, Qundef, iseq);
/* cache a wrapper object */
RB_OBJ_WRITE((VALUE)iseq, &iseq->wrapper, obj);
RB_OBJ_FREEZE((VALUE)iseq);
return obj;
}
2015-07-22 01:52:59 +03:00
}
VALUE
rb_iseqw_new(const rb_iseq_t *iseq)
{
return iseqw_new(iseq);
}
/*
* call-seq:
* InstructionSequence.compile(source[, file[, path[, line[, options]]]]) -> iseq
* InstructionSequence.new(source[, file[, path[, line[, options]]]]) -> iseq
*
* Takes +source+, a String of Ruby code and compiles it to an
* InstructionSequence.
*
* Optionally takes +file+, +path+, and +line+ which describe the file path,
* real path and first line number of the ruby code in +source+ which are
* metadata attached to the returned +iseq+.
*
* +file+ is used for `__FILE__` and exception backtrace. +path+ is used for
* +require_relative+ base. It is recommended these should be the same full
* path.
*
* +options+, which can be +true+, +false+ or a +Hash+, is used to
* modify the default behavior of the Ruby iseq compiler.
*
* For details regarding valid compile options see ::compile_option=.
*
* RubyVM::InstructionSequence.compile("a = 1 + 2")
* #=> <RubyVM::InstructionSequence:<compiled>@<compiled>>
*
* path = "test.rb"
* RubyVM::InstructionSequence.compile(File.read(path), path, File.expand_path(path))
* #=> <RubyVM::InstructionSequence:<compiled>@test.rb:1>
*
* path = File.expand_path("test.rb")
* RubyVM::InstructionSequence.compile(File.read(path), path, path)
* #=> <RubyVM::InstructionSequence:<compiled>@/absolute/path/to/test.rb:1>
*
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_s_compile(int argc, VALUE *argv, VALUE self)
{
VALUE src, file = Qnil, path = Qnil, line = INT2FIX(1), opt = Qnil;
int i;
i = rb_scan_args(argc, argv, "1*:", &src, NULL, &opt);
if (i > 4+NIL_P(opt)) rb_error_arity(argc, 1, 5);
switch (i) {
case 5: opt = argv[--i];
case 4: line = argv[--i];
case 3: path = argv[--i];
case 2: file = argv[--i];
}
if (NIL_P(file)) file = rb_fstring_lit("<compiled>");
if (NIL_P(path)) path = file;
if (NIL_P(line)) line = INT2FIX(1);
Check_Type(path, T_STRING);
Check_Type(file, T_STRING);
return iseqw_new(rb_iseq_compile_with_option(src, file, path, line, opt));
}
/*
* call-seq:
* InstructionSequence.compile_file(file[, options]) -> iseq
*
* Takes +file+, a String with the location of a Ruby source file, reads,
* parses and compiles the file, and returns +iseq+, the compiled
* InstructionSequence with source location metadata set.
*
* Optionally takes +options+, which can be +true+, +false+ or a +Hash+, to
* modify the default behavior of the Ruby iseq compiler.
*
* For details regarding valid compile options see ::compile_option=.
*
* # /tmp/hello.rb
* puts "Hello, world!"
*
* # elsewhere
* RubyVM::InstructionSequence.compile_file("/tmp/hello.rb")
* #=> <RubyVM::InstructionSequence:<main>@/tmp/hello.rb>
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_s_compile_file(int argc, VALUE *argv, VALUE self)
{
2022-09-25 10:45:28 +03:00
VALUE file, opt = Qnil;
VALUE parser, f, exc = Qnil, ret;
rb_ast_t *ast;
rb_compile_option_t option;
int i;
i = rb_scan_args(argc, argv, "1*:", &file, NULL, &opt);
if (i > 1+NIL_P(opt)) rb_error_arity(argc, 1, 2);
switch (i) {
case 2: opt = argv[--i];
}
FilePathValue(file);
file = rb_fstring(file); /* rb_io_t->pathv gets frozen anyways */
f = rb_file_open_str(file, "r");
rb_execution_context_t *ec = GET_EC();
VALUE v = rb_vm_push_frame_fname(ec, file);
parser = rb_parser_new();
rb_parser_set_context(parser, NULL, FALSE);
ast = (rb_ast_t *)rb_parser_load_file(parser, file);
if (!ast->body.root) exc = GET_EC()->errinfo;
rb_io_close(f);
if (!ast->body.root) {
rb_ast_dispose(ast);
rb_exc_raise(exc);
}
make_compile_option(&option, opt);
2015-07-22 01:52:59 +03:00
ret = iseqw_new(rb_iseq_new_with_opt(&ast->body, rb_fstring_lit("<main>"),
file,
rb_realpath_internal(Qnil, file, 1),
2022-09-25 10:45:28 +03:00
1, NULL, 0, ISEQ_TYPE_TOP, &option));
rb_ast_dispose(ast);
rb_vm_pop_frame(ec);
RB_GC_GUARD(v);
return ret;
}
/*
* call-seq:
* InstructionSequence.compile_option = options
*
* Sets the default values for various optimizations in the Ruby iseq
* compiler.
*
* Possible values for +options+ include +true+, which enables all options,
* +false+ which disables all options, and +nil+ which leaves all options
* unchanged.
*
* You can also pass a +Hash+ of +options+ that you want to change, any
* options not present in the hash will be left unchanged.
*
* Possible option names (which are keys in +options+) which can be set to
* +true+ or +false+ include:
*
* * +:inline_const_cache+
* * +:instructions_unification+
* * +:operands_unification+
* * +:peephole_optimization+
* * +:specialized_instruction+
* * +:stack_caching+
* * +:tailcall_optimization+
*
* Additionally, +:debug_level+ can be set to an integer.
*
* These default options can be overwritten for a single run of the iseq
* compiler by passing any of the above values as the +options+ parameter to
* ::new, ::compile and ::compile_file.
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_s_compile_option_set(VALUE self, VALUE opt)
{
rb_compile_option_t option;
make_compile_option(&option, opt);
COMPILE_OPTION_DEFAULT = option;
return opt;
}
/*
* call-seq:
* InstructionSequence.compile_option -> options
*
* Returns a hash of default options used by the Ruby iseq compiler.
*
* For details, see InstructionSequence.compile_option=.
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_s_compile_option_get(VALUE self)
{
return make_compile_option_value(&COMPILE_OPTION_DEFAULT);
}
2015-07-22 01:52:59 +03:00
static const rb_iseq_t *
iseqw_check(VALUE iseqw)
{
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
rb_iseq_t *iseq = DATA_PTR(iseqw);
if (!ISEQ_BODY(iseq)) {
rb_ibf_load_iseq_complete(iseq);
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
}
2015-07-22 01:52:59 +03:00
if (!ISEQ_BODY(iseq)->location.label) {
rb_raise(rb_eTypeError, "uninitialized InstructionSequence");
}
return iseq;
}
2015-07-22 01:52:59 +03:00
const rb_iseq_t *
rb_iseqw_to_iseq(VALUE iseqw)
{
return iseqw_check(iseqw);
}
/*
* call-seq:
* iseq.eval -> obj
*
* Evaluates the instruction sequence and returns the result.
*
* RubyVM::InstructionSequence.compile("1 + 2").eval #=> 3
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_eval(VALUE self)
{
2015-07-22 01:52:59 +03:00
return rb_iseq_eval(iseqw_check(self));
}
/*
* Returns a human-readable string representation of this instruction
* sequence, including the #label and #path.
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_inspect(VALUE self)
{
2015-07-22 01:52:59 +03:00
const rb_iseq_t *iseq = iseqw_check(self);
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
VALUE klass = rb_class_name(rb_obj_class(self));
2015-07-22 01:52:59 +03:00
if (!body->location.label) {
return rb_sprintf("#<%"PRIsVALUE": uninitialized>", klass);
}
2015-07-22 01:52:59 +03:00
else {
return rb_sprintf("<%"PRIsVALUE":%"PRIsVALUE"@%"PRIsVALUE":%d>",
klass,
body->location.label, rb_iseq_path(iseq),
FIX2INT(rb_iseq_first_lineno(iseq)));
2015-07-22 01:52:59 +03:00
}
}
/*
* Returns the path of this instruction sequence.
*
* <code><compiled></code> if the iseq was evaluated from a string.
*
* For example, using irb:
*
* iseq = RubyVM::InstructionSequence.compile('num = 1 + 2')
* #=> <RubyVM::InstructionSequence:<compiled>@<compiled>>
* iseq.path
* #=> "<compiled>"
*
* Using ::compile_file:
*
* # /tmp/method.rb
* def hello
* puts "hello, world"
* end
*
* # in irb
* > iseq = RubyVM::InstructionSequence.compile_file('/tmp/method.rb')
* > iseq.path #=> /tmp/method.rb
*/
2015-07-22 01:52:59 +03:00
static VALUE
iseqw_path(VALUE self)
{
2015-07-22 01:52:59 +03:00
return rb_iseq_path(iseqw_check(self));
}
/*
* Returns the absolute path of this instruction sequence.
*
* +nil+ if the iseq was evaluated from a string.
*
* For example, using ::compile_file:
*
* # /tmp/method.rb
* def hello
* puts "hello, world"
* end
*
* # in irb
* > iseq = RubyVM::InstructionSequence.compile_file('/tmp/method.rb')
* > iseq.absolute_path #=> /tmp/method.rb
*/
2015-07-22 01:52:59 +03:00
static VALUE
iseqw_absolute_path(VALUE self)
{
return rb_iseq_realpath(iseqw_check(self));
}
/* Returns the label of this instruction sequence.
*
* <code><main></code> if it's at the top level, <code><compiled></code> if it
* was evaluated from a string.
*
* For example, using irb:
*
* iseq = RubyVM::InstructionSequence.compile('num = 1 + 2')
* #=> <RubyVM::InstructionSequence:<compiled>@<compiled>>
* iseq.label
* #=> "<compiled>"
*
* Using ::compile_file:
*
* # /tmp/method.rb
* def hello
* puts "hello, world"
* end
*
* # in irb
* > iseq = RubyVM::InstructionSequence.compile_file('/tmp/method.rb')
* > iseq.label #=> <main>
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_label(VALUE self)
{
2015-07-22 01:52:59 +03:00
return rb_iseq_label(iseqw_check(self));
}
/* Returns the base label of this instruction sequence.
*
* For example, using irb:
*
* iseq = RubyVM::InstructionSequence.compile('num = 1 + 2')
* #=> <RubyVM::InstructionSequence:<compiled>@<compiled>>
* iseq.base_label
* #=> "<compiled>"
*
* Using ::compile_file:
*
* # /tmp/method.rb
* def hello
* puts "hello, world"
* end
*
* # in irb
* > iseq = RubyVM::InstructionSequence.compile_file('/tmp/method.rb')
* > iseq.base_label #=> <main>
*/
2015-07-22 01:52:59 +03:00
static VALUE
iseqw_base_label(VALUE self)
{
2015-07-22 01:52:59 +03:00
return rb_iseq_base_label(iseqw_check(self));
}
/* Returns the number of the first source line where the instruction sequence
* was loaded from.
*
* For example, using irb:
*
* iseq = RubyVM::InstructionSequence.compile('num = 1 + 2')
* #=> <RubyVM::InstructionSequence:<compiled>@<compiled>>
* iseq.first_lineno
* #=> 1
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_first_lineno(VALUE self)
{
2015-07-22 01:52:59 +03:00
return rb_iseq_first_lineno(iseqw_check(self));
}
static VALUE iseq_data_to_ary(const rb_iseq_t *iseq);
/*
* call-seq:
* iseq.to_a -> ary
*
* Returns an Array with 14 elements representing the instruction sequence
* with the following data:
*
* [magic]
* A string identifying the data format. <b>Always
* +YARVInstructionSequence/SimpleDataFormat+.</b>
*
* [major_version]
* The major version of the instruction sequence.
*
* [minor_version]
* The minor version of the instruction sequence.
*
* [format_type]
* A number identifying the data format. <b>Always 1</b>.
*
* [misc]
* A hash containing:
*
* [+:arg_size+]
* the total number of arguments taken by the method or the block (0 if
* _iseq_ doesn't represent a method or block)
* [+:local_size+]
* the number of local variables + 1
* [+:stack_max+]
* used in calculating the stack depth at which a SystemStackError is
* thrown.
*
* [#label]
* The name of the context (block, method, class, module, etc.) that this
* instruction sequence belongs to.
*
* <code><main></code> if it's at the top level, <code><compiled></code> if
* it was evaluated from a string.
*
* [#path]
* The relative path to the Ruby file where the instruction sequence was
* loaded from.
*
* <code><compiled></code> if the iseq was evaluated from a string.
*
* [#absolute_path]
* The absolute path to the Ruby file where the instruction sequence was
* loaded from.
*
* +nil+ if the iseq was evaluated from a string.
*
* [#first_lineno]
* The number of the first source line where the instruction sequence was
* loaded from.
*
* [type]
* The type of the instruction sequence.
*
* Valid values are +:top+, +:method+, +:block+, +:class+, +:rescue+,
* +:ensure+, +:eval+, +:main+, and +plain+.
*
* [locals]
* An array containing the names of all arguments and local variables as
* symbols.
*
* [params]
* An Hash object containing parameter information.
*
* More info about these values can be found in +vm_core.h+.
*
* [catch_table]
* A list of exceptions and control flow operators (rescue, next, redo,
* break, etc.).
*
* [bytecode]
* An array of arrays containing the instruction names and operands that
* make up the body of the instruction sequence.
*
* Note that this format is MRI specific and version dependent.
*
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_to_a(VALUE self)
{
2015-07-22 01:52:59 +03:00
const rb_iseq_t *iseq = iseqw_check(self);
return iseq_data_to_ary(iseq);
}
#if VM_INSN_INFO_TABLE_IMPL == 1 /* binary search */
static const struct iseq_insn_info_entry *
get_insn_info_binary_search(const rb_iseq_t *iseq, size_t pos)
{
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
size_t size = body->insns_info.size;
const struct iseq_insn_info_entry *insns_info = body->insns_info.body;
const unsigned int *positions = body->insns_info.positions;
const int debug = 0;
if (debug) {
printf("size: %"PRIuSIZE"\n", size);
printf("insns_info[%"PRIuSIZE"]: position: %d, line: %d, pos: %"PRIuSIZE"\n",
(size_t)0, positions[0], insns_info[0].line_no, pos);
}
if (size == 0) {
return NULL;
}
else if (size == 1) {
return &insns_info[0];
}
else {
size_t l = 1, r = size - 1;
while (l <= r) {
size_t m = l + (r - l) / 2;
if (positions[m] == pos) {
return &insns_info[m];
}
if (positions[m] < pos) {
l = m + 1;
}
else {
r = m - 1;
}
}
if (l >= size) {
return &insns_info[size-1];
}
if (positions[l] > pos) {
return &insns_info[l-1];
}
return &insns_info[l];
}
}
static const struct iseq_insn_info_entry *
get_insn_info(const rb_iseq_t *iseq, size_t pos)
{
return get_insn_info_binary_search(iseq, pos);
}
#endif
#if VM_INSN_INFO_TABLE_IMPL == 2 /* succinct bitvector */
static const struct iseq_insn_info_entry *
get_insn_info_succinct_bitvector(const rb_iseq_t *iseq, size_t pos)
{
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
size_t size = body->insns_info.size;
const struct iseq_insn_info_entry *insns_info = body->insns_info.body;
const int debug = 0;
if (debug) {
#if VM_CHECK_MODE > 0
const unsigned int *positions = body->insns_info.positions;
printf("size: %"PRIuSIZE"\n", size);
printf("insns_info[%"PRIuSIZE"]: position: %d, line: %d, pos: %"PRIuSIZE"\n",
(size_t)0, positions[0], insns_info[0].line_no, pos);
#else
printf("size: %"PRIuSIZE"\n", size);
printf("insns_info[%"PRIuSIZE"]: line: %d, pos: %"PRIuSIZE"\n",
(size_t)0, insns_info[0].line_no, pos);
#endif
}
if (size == 0) {
return NULL;
}
else if (size == 1) {
return &insns_info[0];
}
else {
int index;
VM_ASSERT(body->insns_info.succ_index_table != NULL);
index = succ_index_lookup(body->insns_info.succ_index_table, (int)pos);
return &insns_info[index-1];
}
}
static const struct iseq_insn_info_entry *
get_insn_info(const rb_iseq_t *iseq, size_t pos)
{
return get_insn_info_succinct_bitvector(iseq, pos);
}
#endif
#if VM_CHECK_MODE > 0 || VM_INSN_INFO_TABLE_IMPL == 0
static const struct iseq_insn_info_entry *
get_insn_info_linear_search(const rb_iseq_t *iseq, size_t pos)
{
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
size_t i = 0, size = body->insns_info.size;
const struct iseq_insn_info_entry *insns_info = body->insns_info.body;
const unsigned int *positions = body->insns_info.positions;
const int debug = 0;
if (debug) {
printf("size: %"PRIuSIZE"\n", size);
printf("insns_info[%"PRIuSIZE"]: position: %d, line: %d, pos: %"PRIuSIZE"\n",
i, positions[i], insns_info[i].line_no, pos);
}
if (size == 0) {
return NULL;
}
else if (size == 1) {
return &insns_info[0];
}
else {
for (i=1; i<size; i++) {
if (debug) printf("insns_info[%"PRIuSIZE"]: position: %d, line: %d, pos: %"PRIuSIZE"\n",
i, positions[i], insns_info[i].line_no, pos);
2022-07-21 19:23:58 +03:00
if (positions[i] == pos) {
return &insns_info[i];
}
if (positions[i] > pos) {
return &insns_info[i-1];
}
}
}
return &insns_info[i-1];
}
#endif
#if VM_INSN_INFO_TABLE_IMPL == 0 /* linear search */
static const struct iseq_insn_info_entry *
get_insn_info(const rb_iseq_t *iseq, size_t pos)
{
return get_insn_info_linear_search(iseq, pos);
}
#endif
#if VM_CHECK_MODE > 0 && VM_INSN_INFO_TABLE_IMPL > 0
static void
validate_get_insn_info(const rb_iseq_t *iseq)
{
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
size_t i;
for (i = 0; i < body->iseq_size; i++) {
if (get_insn_info_linear_search(iseq, i) != get_insn_info(iseq, i)) {
rb_bug("validate_get_insn_info: get_insn_info_linear_search(iseq, %"PRIuSIZE") != get_insn_info(iseq, %"PRIuSIZE")", i, i);
}
}
}
#endif
unsigned int
rb_iseq_line_no(const rb_iseq_t *iseq, size_t pos)
{
const struct iseq_insn_info_entry *entry = get_insn_info(iseq, pos);
if (entry) {
return entry->line_no;
}
else {
return 0;
}
}
#ifdef USE_ISEQ_NODE_ID
int
rb_iseq_node_id(const rb_iseq_t *iseq, size_t pos)
{
const struct iseq_insn_info_entry *entry = get_insn_info(iseq, pos);
if (entry) {
return entry->node_id;
}
else {
return 0;
}
}
#endif
mjit_compile.c: merge initial JIT compiler which has been developed by Takashi Kokubun <takashikkbn@gmail> as YARV-MJIT. Many of its bugs are fixed by wanabe <s.wanabe@gmail.com>. This JIT compiler is designed to be a safe migration path to introduce JIT compiler to MRI. So this commit does not include any bytecode changes or dynamic instruction modifications, which are done in original MJIT. This commit even strips off some aggressive optimizations from YARV-MJIT, and thus it's slower than YARV-MJIT too. But it's still fairly faster than Ruby 2.5 in some benchmarks (attached below). Note that this JIT compiler passes `make test`, `make test-all`, `make test-spec` without JIT, and even with JIT. Not only it's perfectly safe with JIT disabled because it does not replace VM instructions unlike MJIT, but also with JIT enabled it stably runs Ruby applications including Rails applications. I'm expecting this version as just "initial" JIT compiler. I have many optimization ideas which are skipped for initial merging, and you may easily replace this JIT compiler with a faster one by just replacing mjit_compile.c. `mjit_compile` interface is designed for the purpose. common.mk: update dependencies for mjit_compile.c. internal.h: declare `rb_vm_insn_addr2insn` for MJIT. vm.c: exclude some definitions if `-DMJIT_HEADER` is provided to compiler. This avoids to include some functions which take a long time to compile, e.g. vm_exec_core. Some of the purpose is achieved in transform_mjit_header.rb (see `IGNORED_FUNCTIONS`) but others are manually resolved for now. Load mjit_helper.h for MJIT header. mjit_helper.h: New. This is a file used only by JIT-ed code. I'll refactor `mjit_call_cfunc` later. vm_eval.c: add some #ifdef switches to skip compiling some functions like Init_vm_eval. win32/mkexports.rb: export thread/ec functions, which are used by MJIT. include/ruby/defines.h: add MJIT_FUNC_EXPORTED macro alis to clarify that a function is exported only for MJIT. array.c: export a function used by MJIT. bignum.c: ditto. class.c: ditto. compile.c: ditto. error.c: ditto. gc.c: ditto. hash.c: ditto. iseq.c: ditto. numeric.c: ditto. object.c: ditto. proc.c: ditto. re.c: ditto. st.c: ditto. string.c: ditto. thread.c: ditto. variable.c: ditto. vm_backtrace.c: ditto. vm_insnhelper.c: ditto. vm_method.c: ditto. I would like to improve maintainability of function exports, but I believe this way is acceptable as initial merging if we clarify the new exports are for MJIT (so that we can use them as TODO list to fix) and add unit tests to detect unresolved symbols. I'll add unit tests of JIT compilations in succeeding commits. Author: Takashi Kokubun <takashikkbn@gmail.com> Contributor: wanabe <s.wanabe@gmail.com> Part of [Feature #14235] --- * Known issues * Code generated by gcc is faster than clang. The benchmark may be worse in macOS. Following benchmark result is provided by gcc w/ Linux. * Performance is decreased when Google Chrome is running * JIT can work on MinGW, but it doesn't improve performance at least in short running benchmark. * Currently it doesn't perform well with Rails. We'll try to fix this before release. --- * Benchmark reslts Benchmarked with: Intel 4.0GHz i7-4790K with 16GB memory under x86-64 Ubuntu 8 Cores - 2.0.0-p0: Ruby 2.0.0-p0 - r62186: Ruby trunk (early 2.6.0), before MJIT changes - JIT off: On this commit, but without `--jit` option - JIT on: On this commit, and with `--jit` option ** Optcarrot fps Benchmark: https://github.com/mame/optcarrot | |2.0.0-p0 |r62186 |JIT off |JIT on | |:--------|:--------|:--------|:--------|:--------| |fps |37.32 |51.46 |51.31 |58.88 | |vs 2.0.0 |1.00x |1.38x |1.37x |1.58x | ** MJIT benchmarks Benchmark: https://github.com/benchmark-driver/mjit-benchmarks (Original: https://github.com/vnmakarov/ruby/tree/rtl_mjit_branch/MJIT-benchmarks) | |2.0.0-p0 |r62186 |JIT off |JIT on | |:----------|:--------|:--------|:--------|:--------| |aread |1.00 |1.09 |1.07 |2.19 | |aref |1.00 |1.13 |1.11 |2.22 | |aset |1.00 |1.50 |1.45 |2.64 | |awrite |1.00 |1.17 |1.13 |2.20 | |call |1.00 |1.29 |1.26 |2.02 | |const2 |1.00 |1.10 |1.10 |2.19 | |const |1.00 |1.11 |1.10 |2.19 | |fannk |1.00 |1.04 |1.02 |1.00 | |fib |1.00 |1.32 |1.31 |1.84 | |ivread |1.00 |1.13 |1.12 |2.43 | |ivwrite |1.00 |1.23 |1.21 |2.40 | |mandelbrot |1.00 |1.13 |1.16 |1.28 | |meteor |1.00 |2.97 |2.92 |3.17 | |nbody |1.00 |1.17 |1.15 |1.49 | |nest-ntimes|1.00 |1.22 |1.20 |1.39 | |nest-while |1.00 |1.10 |1.10 |1.37 | |norm |1.00 |1.18 |1.16 |1.24 | |nsvb |1.00 |1.16 |1.16 |1.17 | |red-black |1.00 |1.02 |0.99 |1.12 | |sieve |1.00 |1.30 |1.28 |1.62 | |trees |1.00 |1.14 |1.13 |1.19 | |while |1.00 |1.12 |1.11 |2.41 | ** Discourse's script/bench.rb Benchmark: https://github.com/discourse/discourse/blob/v1.8.7/script/bench.rb NOTE: Rails performance was somehow a little degraded with JIT for now. We should fix this. (At least I know opt_aref is performing badly in JIT and I have an idea to fix it. Please wait for the fix.) *** JIT off Your Results: (note for timings- percentile is first, duration is second in millisecs) categories_admin: 50: 17 75: 18 90: 22 99: 29 home_admin: 50: 21 75: 21 90: 27 99: 40 topic_admin: 50: 17 75: 18 90: 22 99: 32 categories: 50: 35 75: 41 90: 43 99: 77 home: 50: 39 75: 46 90: 49 99: 95 topic: 50: 46 75: 52 90: 56 99: 101 *** JIT on Your Results: (note for timings- percentile is first, duration is second in millisecs) categories_admin: 50: 19 75: 21 90: 25 99: 33 home_admin: 50: 24 75: 26 90: 30 99: 35 topic_admin: 50: 19 75: 20 90: 25 99: 30 categories: 50: 40 75: 44 90: 48 99: 76 home: 50: 42 75: 48 90: 51 99: 89 topic: 50: 49 75: 55 90: 58 99: 99 git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@62197 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-02-04 14:22:28 +03:00
MJIT_FUNC_EXPORTED rb_event_flag_t
rb_iseq_event_flags(const rb_iseq_t *iseq, size_t pos)
{
const struct iseq_insn_info_entry *entry = get_insn_info(iseq, pos);
if (entry) {
return entry->events;
}
else {
return 0;
}
}
void
rb_iseq_clear_event_flags(const rb_iseq_t *iseq, size_t pos, rb_event_flag_t reset)
{
struct iseq_insn_info_entry *entry = (struct iseq_insn_info_entry *)get_insn_info(iseq, pos);
if (entry) {
entry->events &= ~reset;
if (!(entry->events & iseq->aux.exec.global_trace_events)) {
void rb_iseq_trace_flag_cleared(const rb_iseq_t *iseq, size_t pos);
rb_iseq_trace_flag_cleared(iseq, pos);
}
}
}
static VALUE
local_var_name(const rb_iseq_t *diseq, VALUE level, VALUE op)
{
VALUE i;
VALUE name;
ID lid;
int idx;
for (i = 0; i < level; i++) {
diseq = ISEQ_BODY(diseq)->parent_iseq;
}
idx = ISEQ_BODY(diseq)->local_table_size - (int)op - 1;
lid = ISEQ_BODY(diseq)->local_table[idx];
name = rb_id2str(lid);
if (!name) {
name = rb_str_new_cstr("?");
}
else if (!rb_str_symname_p(name)) {
name = rb_str_inspect(name);
}
else {
name = rb_str_dup(name);
}
rb_str_catf(name, "@%d", idx);
return name;
}
int rb_insn_unified_local_var_level(VALUE);
2020-10-20 13:32:10 +03:00
VALUE rb_dump_literal(VALUE lit);
VALUE
rb_insn_operand_intern(const rb_iseq_t *iseq,
VALUE insn, int op_no, VALUE op,
int len, size_t pos, const VALUE *pnop, VALUE child)
{
const char *types = insn_op_types(insn);
char type = types[op_no];
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
VALUE ret = Qundef;
switch (type) {
case TS_OFFSET: /* LONG */
ret = rb_sprintf("%"PRIdVALUE, (VALUE)(pos + len + op));
break;
case TS_NUM: /* ULONG */
if (insn == BIN(defined) && op_no == 0) {
enum defined_type deftype = (enum defined_type)op;
switch (deftype) {
case DEFINED_FUNC:
ret = rb_fstring_lit("func");
break;
case DEFINED_REF:
ret = rb_fstring_lit("ref");
break;
case DEFINED_CONST_FROM:
ret = rb_fstring_lit("constant-from");
break;
default:
ret = rb_iseq_defined_string(deftype);
break;
}
if (ret) break;
}
else if (insn == BIN(checktype) && op_no == 0) {
const char *type_str = rb_type_str((enum ruby_value_type)op);
if (type_str) {
ret = rb_str_new_cstr(type_str); break;
}
}
ret = rb_sprintf("%"PRIuVALUE, op);
break;
case TS_LINDEX:{
int level;
if (types[op_no+1] == TS_NUM && pnop) {
ret = local_var_name(iseq, *pnop, op - VM_ENV_DATA_SIZE);
}
else if ((level = rb_insn_unified_local_var_level(insn)) >= 0) {
ret = local_var_name(iseq, (VALUE)level, op - VM_ENV_DATA_SIZE);
}
else {
ret = rb_inspect(INT2FIX(op));
}
break;
}
case TS_ID: /* ID (symbol) */
ret = rb_inspect(ID2SYM(op));
break;
case TS_VALUE: /* VALUE */
op = obj_resurrect(op);
if (insn == BIN(defined) && op_no == 1 && FIXNUM_P(op)) {
/* should be DEFINED_REF */
int type = NUM2INT(op);
if (type) {
if (type & 1) {
ret = rb_sprintf(":$%c", (type >> 1));
}
else {
ret = rb_sprintf(":$%d", (type >> 1));
}
break;
}
}
2020-10-20 13:32:10 +03:00
ret = rb_dump_literal(op);
if (CLASS_OF(op) == rb_cISeq) {
if (child) {
rb_ary_push(child, op);
}
}
break;
case TS_ISEQ: /* iseq */
{
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
if (op) {
const rb_iseq_t *iseq = rb_iseq_check((rb_iseq_t *)op);
ret = ISEQ_BODY(iseq)->location.label;
if (child) {
2015-07-22 01:52:59 +03:00
rb_ary_push(child, (VALUE)iseq);
}
}
else {
ret = rb_str_new2("nil");
}
break;
}
case TS_IC:
New constant caching insn: opt_getconstant_path Previously YARV bytecode implemented constant caching by having a pair of instructions, opt_getinlinecache and opt_setinlinecache, wrapping a series of getconstant calls (with putobject providing supporting arguments). This commit replaces that pattern with a new instruction, opt_getconstant_path, handling both getting/setting the inline cache and fetching the constant on a cache miss. This is implemented by storing the full constant path as a null-terminated array of IDs inside of the IC structure. idNULL is used to signal an absolute constant reference. $ ./miniruby --dump=insns -e '::Foo::Bar::Baz' == disasm: #<ISeq:<main>@-e:1 (1,0)-(1,13)> (catch: FALSE) 0000 opt_getconstant_path <ic:0 ::Foo::Bar::Baz> ( 1)[Li] 0002 leave The motivation for this is that we had increasingly found the need to disassemble the instructions between the opt_getinlinecache and opt_setinlinecache in order to determine the constant we are fetching, or otherwise store metadata. This disassembly was done: * In opt_setinlinecache, to register the IC against the constant names it is using for granular invalidation. * In rb_iseq_free, to unregister the IC from the invalidation table. * In YJIT to find the position of a opt_getinlinecache instruction to invalidate it when the cache is populated * In YJIT to register the constant names being used for invalidation. With this change we no longe need disassemly for these (in fact rb_iseq_each is now unused), as the list of constant names being referenced is held in the IC. This should also make it possible to make more optimizations in the future. This may also reduce the size of iseqs, as previously each segment required 32 bytes (on 64-bit platforms) for each constant segment. This implementation only stores one ID per-segment. There should be no significant performance change between this and the previous implementation. Previously opt_getinlinecache was a "leaf" instruction, but it included a jump (almost always to a separate cache line). Now opt_getconstant_path is a non-leaf (it may raise/autoload/call const_missing) but it does not jump. These seem to even out.
2022-08-10 20:35:48 +03:00
{
ret = rb_sprintf("<ic:%"PRIdPTRDIFF" ", (union iseq_inline_storage_entry *)op - ISEQ_BODY(iseq)->is_entries);
const ID *segments = ((IC)op)->segments;
rb_str_cat2(ret, rb_id2name(*segments++));
while (*segments) {
rb_str_catf(ret, "::%s", rb_id2name(*segments++));
}
rb_str_cat2(ret, ">");
}
break;
case TS_IVC:
case TS_ICVARC:
case TS_ISE:
ret = rb_sprintf("<is:%"PRIdPTRDIFF">", (union iseq_inline_storage_entry *)op - ISEQ_BODY(iseq)->is_entries);
break;
case TS_CALLDATA:
{
struct rb_call_data *cd = (struct rb_call_data *)op;
const struct rb_callinfo *ci = cd->ci;
VALUE ary = rb_ary_new();
ID mid = vm_ci_mid(ci);
if (mid) {
rb_ary_push(ary, rb_sprintf("mid:%"PRIsVALUE, rb_id2str(mid)));
}
rb_ary_push(ary, rb_sprintf("argc:%d", vm_ci_argc(ci)));
if (vm_ci_flag(ci) & VM_CALL_KWARG) {
const struct rb_callinfo_kwarg *kw_args = vm_ci_kwarg(ci);
VALUE kw_ary = rb_ary_new_from_values(kw_args->keyword_len, kw_args->keywords);
rb_ary_push(ary, rb_sprintf("kw:[%"PRIsVALUE"]", rb_ary_join(kw_ary, rb_str_new2(","))));
* rewrite method/block parameter fitting logic to optimize keyword arguments/parameters and a splat argument. [Feature #10440] (Details are described in this ticket) Most of complex part is moved to vm_args.c. Now, ISeq#to_a does not catch up new instruction format. * vm_core.h: change iseq data structures. * introduce rb_call_info_kw_arg_t to represent keyword arguments. * add rb_call_info_t::kw_arg. * rename rb_iseq_t::arg_post_len to rb_iseq_t::arg_post_num. * rename rb_iseq_t::arg_keywords to arg_keyword_num. * rename rb_iseq_t::arg_keyword to rb_iseq_t::arg_keyword_bits. to represent keyword bitmap parameter index. This bitmap parameter shows that which keyword parameters are given or not given (0 for given). It is refered by `checkkeyword' instruction described bellow. * rename rb_iseq_t::arg_keyword_check to rb_iseq_t::arg_keyword_rest to represent keyword rest parameter index. * add rb_iseq_t::arg_keyword_default_values to represent default keyword values. * rename VM_CALL_ARGS_SKIP_SETUP to VM_CALL_ARGS_SIMPLE to represent (ci->flag & (SPLAT|BLOCKARG)) && ci->blockiseq == NULL && ci->kw_arg == NULL. * vm_insnhelper.c, vm_args.c: rewrite with refactoring. * rewrite splat argument code. * rewrite keyword arguments/parameters code. * merge method and block parameter fitting code into one code base. * vm.c, vm_eval.c: catch up these changes. * compile.c (new_callinfo): callinfo requires kw_arg parameter. * compile.c (compile_array_): check the last argument Hash object or not. If Hash object and all keys are Symbol literals, they are compiled to keyword arguments. * insns.def (checkkeyword): add new instruction. This instruction check the availability of corresponding keyword. For example, a method "def foo k1: 'v1'; end" is cimpiled to the following instructions. 0000 checkkeyword 2, 0 # check k1 is given. 0003 branchif 9 # if given, jump to address #9 0005 putstring "v1" 0007 setlocal_OP__WC__0 3 # k1 = 'v1' 0009 trace 8 0011 putnil 0012 trace 16 0014 leave * insns.def (opt_send_simple): removed and add new instruction "opt_send_without_block". * parse.y (new_args_tail_gen): reorder variables. Before this patch, a method "def foo(k1: 1, kr1:, k2: 2, **krest, &b)" has parameter variables "k1, kr1, k2, &b, internal_id, krest", but this patch reorders to "kr1, k1, k2, internal_id, krest, &b". (locate a block variable at last) * parse.y (vtable_pop): added. This function remove latest `n' variables from vtable. * iseq.c: catch up iseq data changes. * proc.c: ditto. * class.c (keyword_error): export as rb_keyword_error(). * common.mk: depend vm_args.c for vm.o. * hash.c (rb_hash_has_key): export. * internal.h: ditto. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48239 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2014-11-02 21:02:55 +03:00
}
if (vm_ci_flag(ci)) {
VALUE flags = rb_ary_new();
# define CALL_FLAG(n) if (vm_ci_flag(ci) & VM_CALL_##n) rb_ary_push(flags, rb_str_new2(#n))
CALL_FLAG(ARGS_SPLAT);
CALL_FLAG(ARGS_BLOCKARG);
CALL_FLAG(FCALL);
CALL_FLAG(VCALL);
CALL_FLAG(ARGS_SIMPLE);
CALL_FLAG(BLOCKISEQ);
CALL_FLAG(TAILCALL);
CALL_FLAG(SUPER);
CALL_FLAG(ZSUPER);
CALL_FLAG(KWARG);
CALL_FLAG(KW_SPLAT);
Reduce allocations for keyword argument hashes Previously, passing a keyword splat to a method always allocated a hash on the caller side, and accepting arbitrary keywords in a method allocated a separate hash on the callee side. Passing explicit keywords to a method that accepted a keyword splat did not allocate a hash on the caller side, but resulted in two hashes allocated on the callee side. This commit makes passing a single keyword splat to a method not allocate a hash on the caller side. Passing multiple keyword splats or a mix of explicit keywords and a keyword splat still generates a hash on the caller side. On the callee side, if arbitrary keywords are not accepted, it does not allocate a hash. If arbitrary keywords are accepted, it will allocate a hash, but this commit uses a callinfo flag to indicate whether the caller already allocated a hash, and if so, the callee can use the passed hash without duplicating it. So this commit should make it so that a maximum of a single hash is allocated during method calls. To set the callinfo flag appropriately, method call argument compilation checks if only a single keyword splat is given. If only one keyword splat is given, the VM_CALL_KW_SPLAT_MUT callinfo flag is not set, since in that case the keyword splat is passed directly and not mutable. If more than one splat is used, a new hash needs to be generated on the caller side, and in that case the callinfo flag is set, indicating the keyword splat is mutable by the callee. In compile_hash, used for both hash and keyword argument compilation, if compiling keyword arguments and only a single keyword splat is used, pass the argument directly. On the caller side, in vm_args.c, the callinfo flag needs to be recognized and handled. Because the keyword splat argument may not be a hash, it needs to be converted to a hash first if not. Then, unless the callinfo flag is set, the hash needs to be duplicated. The temporary copy of the callinfo flag, kw_flag, is updated if a hash was duplicated, to prevent the need to duplicate it again. If we are converting to a hash or duplicating a hash, we need to update the argument array, which can including duplicating the positional splat array if one was passed. CALLER_SETUP_ARG and a couple other places needs to be modified to handle similar issues for other types of calls. This includes fairly comprehensive tests for different ways keywords are handled internally, checking that you get equal results but that keyword splats on the caller side result in distinct objects for keyword rest parameters. Included are benchmarks for keyword argument calls. Brief results when compiled without optimization: def kw(a: 1) a end def kws(**kw) kw end h = {a: 1} kw(a: 1) # about same kw(**h) # 2.37x faster kws(a: 1) # 1.30x faster kws(**h) # 2.19x faster kw(a: 1, **h) # 1.03x slower kw(**h, **h) # about same kws(a: 1, **h) # 1.16x faster kws(**h, **h) # 1.14x faster
2020-02-24 23:05:07 +03:00
CALL_FLAG(KW_SPLAT_MUT);
CALL_FLAG(OPT_SEND); /* maybe not reachable */
rb_ary_push(ary, rb_ary_join(flags, rb_str_new2("|")));
}
ret = rb_sprintf("<calldata!%"PRIsVALUE">", rb_ary_join(ary, rb_str_new2(", ")));
}
break;
case TS_CDHASH:
ret = rb_str_new2("<cdhash>");
break;
case TS_FUNCPTR:
{
#ifdef HAVE_DLADDR
Dl_info info;
if (dladdr((void *)op, &info) && info.dli_sname) {
ret = rb_str_new_cstr(info.dli_sname);
break;
}
#endif
ret = rb_str_new2("<funcptr>");
}
break;
2019-11-07 10:58:00 +03:00
case TS_BUILTIN:
{
const struct rb_builtin_function *bf = (const struct rb_builtin_function *)op;
ret = rb_sprintf("<builtin!%s/%d>",
bf->name, bf->argc);
}
break;
default:
rb_bug("unknown operand type: %c", type);
}
return ret;
}
static VALUE
right_strip(VALUE str)
{
const char *beg = RSTRING_PTR(str), *end = RSTRING_END(str);
while (end-- > beg && *end == ' ');
rb_str_set_len(str, end - beg + 1);
return str;
}
/**
* Disassemble a instruction
* Iseq -> Iseq inspect object
*/
int
rb_iseq_disasm_insn(VALUE ret, const VALUE *code, size_t pos,
const rb_iseq_t *iseq, VALUE child)
{
VALUE insn = code[pos];
int len = insn_len(insn);
int j;
const char *types = insn_op_types(insn);
VALUE str = rb_str_new(0, 0);
const char *insn_name_buff;
insn_name_buff = insn_name(insn);
if (1) {
extern const int rb_vm_max_insn_name_size;
rb_str_catf(str, "%04"PRIuSIZE" %-*s ", pos, rb_vm_max_insn_name_size, insn_name_buff);
}
else {
rb_str_catf(str, "%04"PRIuSIZE" %-28.*s ", pos,
(int)strcspn(insn_name_buff, "_"), insn_name_buff);
}
for (j = 0; types[j]; j++) {
VALUE opstr = rb_insn_operand_intern(iseq, insn, j, code[pos + j + 1],
len, pos, &code[pos + j + 2],
child);
rb_str_concat(str, opstr);
if (types[j + 1]) {
rb_str_cat2(str, ", ");
}
}
{
unsigned int line_no = rb_iseq_line_no(iseq, pos);
unsigned int prev = pos == 0 ? 0 : rb_iseq_line_no(iseq, pos - 1);
if (line_no && line_no != prev) {
long slen = RSTRING_LEN(str);
slen = (slen > 70) ? 0 : (70 - slen);
str = rb_str_catf(str, "%*s(%4d)", (int)slen, "", line_no);
}
}
{
rb_event_flag_t events = rb_iseq_event_flags(iseq, pos);
if (events) {
str = rb_str_catf(str, "[%s%s%s%s%s%s%s%s%s%s%s]",
events & RUBY_EVENT_LINE ? "Li" : "",
events & RUBY_EVENT_CLASS ? "Cl" : "",
events & RUBY_EVENT_END ? "En" : "",
events & RUBY_EVENT_CALL ? "Ca" : "",
events & RUBY_EVENT_RETURN ? "Re" : "",
events & RUBY_EVENT_C_CALL ? "Cc" : "",
events & RUBY_EVENT_C_RETURN ? "Cr" : "",
events & RUBY_EVENT_B_CALL ? "Bc" : "",
events & RUBY_EVENT_B_RETURN ? "Br" : "",
events & RUBY_EVENT_COVERAGE_LINE ? "Cli" : "",
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
events & RUBY_EVENT_COVERAGE_BRANCH ? "Cbr" : "");
}
}
right_strip(str);
if (ret) {
rb_str_cat2(str, "\n");
rb_str_concat(ret, str);
}
else {
printf("%.*s\n", (int)RSTRING_LEN(str), RSTRING_PTR(str));
}
return len;
}
static const char *
catch_type(int type)
{
switch (type) {
case CATCH_TYPE_RESCUE:
return "rescue";
case CATCH_TYPE_ENSURE:
return "ensure";
case CATCH_TYPE_RETRY:
return "retry";
case CATCH_TYPE_BREAK:
return "break";
case CATCH_TYPE_REDO:
return "redo";
case CATCH_TYPE_NEXT:
return "next";
default:
rb_bug("unknown catch type: %d", type);
return 0;
}
}
2015-07-22 01:52:59 +03:00
static VALUE
iseq_inspect(const rb_iseq_t *iseq)
{
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
if (!body->location.label) {
2015-07-22 01:52:59 +03:00
return rb_sprintf("#<ISeq: uninitialized>");
}
else {
const rb_code_location_t *loc = &body->location.code_location;
return rb_sprintf("#<ISeq:%"PRIsVALUE"@%"PRIsVALUE":%d (%d,%d)-(%d,%d)>",
body->location.label, rb_iseq_path(iseq),
loc->beg_pos.lineno,
loc->beg_pos.lineno,
loc->beg_pos.column,
loc->end_pos.lineno,
loc->end_pos.column);
2015-07-22 01:52:59 +03:00
}
}
static const rb_data_type_t tmp_set = {
"tmpset",
{(void (*)(void *))rb_mark_set, (void (*)(void *))st_free_table, 0, 0,},
0, 0, RUBY_TYPED_FREE_IMMEDIATELY
};
static VALUE
rb_iseq_disasm_recursive(const rb_iseq_t *iseq, VALUE indent)
{
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
VALUE *code;
VALUE str = rb_str_new(0, 0);
VALUE child = rb_ary_hidden_new(3);
unsigned int size;
unsigned int i;
long l;
size_t n;
enum {header_minlen = 72};
st_table *done_iseq = 0;
VALUE done_iseq_wrapper = Qnil;
const char *indent_str;
long indent_len;
size = body->iseq_size;
indent_len = RSTRING_LEN(indent);
indent_str = RSTRING_PTR(indent);
rb_str_cat(str, indent_str, indent_len);
rb_str_cat2(str, "== disasm: ");
rb_str_append(str, iseq_inspect(iseq));
rb_str_catf(str, " (catch: %s)", body->catch_except_p ? "true" : "false");
if ((l = RSTRING_LEN(str) - indent_len) < header_minlen) {
rb_str_modify_expand(str, header_minlen - l);
memset(RSTRING_END(str), '=', header_minlen - l);
}
rb_str_cat2(str, "\n");
/* show catch table information */
if (body->catch_table) {
rb_str_cat(str, indent_str, indent_len);
rb_str_cat2(str, "== catch table\n");
}
if (body->catch_table) {
rb_str_cat_cstr(indent, "| ");
indent_str = RSTRING_PTR(indent);
for (i = 0; i < body->catch_table->size; i++) {
const struct iseq_catch_table_entry *entry =
UNALIGNED_MEMBER_PTR(body->catch_table, entries[i]);
rb_str_cat(str, indent_str, indent_len);
rb_str_catf(str,
"| catch type: %-6s st: %04d ed: %04d sp: %04d cont: %04d\n",
catch_type((int)entry->type), (int)entry->start,
(int)entry->end, (int)entry->sp, (int)entry->cont);
if (entry->iseq && !(done_iseq && st_is_member(done_iseq, (st_data_t)entry->iseq))) {
rb_str_concat(str, rb_iseq_disasm_recursive(rb_iseq_check(entry->iseq), indent));
if (!done_iseq) {
done_iseq = st_init_numtable();
done_iseq_wrapper = TypedData_Wrap_Struct(0, &tmp_set, done_iseq);
}
st_insert(done_iseq, (st_data_t)entry->iseq, (st_data_t)0);
indent_str = RSTRING_PTR(indent);
}
}
rb_str_resize(indent, indent_len);
indent_str = RSTRING_PTR(indent);
}
if (body->catch_table) {
rb_str_cat(str, indent_str, indent_len);
rb_str_cat2(str, "|-------------------------------------"
"-----------------------------------\n");
}
/* show local table information */
if (body->local_table) {
const struct rb_iseq_param_keyword *const keyword = body->param.keyword;
rb_str_cat(str, indent_str, indent_len);
rb_str_catf(str,
"local table (size: %d, argc: %d "
"[opts: %d, rest: %d, post: %d, block: %d, kw: %d@%d, kwrest: %d])\n",
body->local_table_size,
body->param.lead_num,
body->param.opt_num,
body->param.flags.has_rest ? body->param.rest_start : -1,
body->param.post_num,
body->param.flags.has_block ? body->param.block_start : -1,
body->param.flags.has_kw ? keyword->num : -1,
body->param.flags.has_kw ? keyword->required_num : -1,
body->param.flags.has_kwrest ? keyword->rest_start : -1);
2022-07-21 19:23:58 +03:00
for (i = body->local_table_size; i > 0;) {
int li = body->local_table_size - --i - 1;
long width;
VALUE name = local_var_name(iseq, 0, i);
2019-04-29 05:31:18 +03:00
char argi[0x100];
char opti[0x100];
2019-04-29 05:31:18 +03:00
opti[0] = '\0';
if (body->param.flags.has_opt) {
int argc = body->param.lead_num;
int opts = body->param.opt_num;
if (li >= argc && li < argc + opts) {
snprintf(opti, sizeof(opti), "Opt=%"PRIdVALUE,
body->param.opt_table[li - argc]);
}
}
2022-07-21 19:23:58 +03:00
snprintf(argi, sizeof(argi), "%s%s%s%s%s%s", /* arg, opts, rest, post, kwrest, block */
body->param.lead_num > li ? "Arg" : "",
opti,
(body->param.flags.has_rest && body->param.rest_start == li) ? "Rest" : "",
(body->param.flags.has_post && body->param.post_start <= li && li < body->param.post_start + body->param.post_num) ? "Post" : "",
(body->param.flags.has_kwrest && keyword->rest_start == li) ? "Kwrest" : "",
(body->param.flags.has_block && body->param.block_start == li) ? "Block" : "");
2022-07-21 19:23:58 +03:00
rb_str_cat(str, indent_str, indent_len);
rb_str_catf(str, "[%2d] ", i + 1);
width = RSTRING_LEN(str) + 11;
rb_str_append(str, name);
if (*argi) rb_str_catf(str, "<%s>", argi);
if ((width -= RSTRING_LEN(str)) > 0) rb_str_catf(str, "%*s", (int)width, "");
}
rb_str_cat_cstr(right_strip(str), "\n");
}
/* show each line */
code = rb_iseq_original_iseq(iseq);
for (n = 0; n < size;) {
rb_str_cat(str, indent_str, indent_len);
n += rb_iseq_disasm_insn(str, code, n, iseq, child);
}
for (l = 0; l < RARRAY_LEN(child); l++) {
VALUE isv = rb_ary_entry(child, l);
if (done_iseq && st_is_member(done_iseq, (st_data_t)isv)) continue;
rb_str_cat_cstr(str, "\n");
rb_str_concat(str, rb_iseq_disasm_recursive(rb_iseq_check((rb_iseq_t *)isv), indent));
indent_str = RSTRING_PTR(indent);
}
RB_GC_GUARD(done_iseq_wrapper);
return str;
}
VALUE
rb_iseq_disasm(const rb_iseq_t *iseq)
{
2020-04-15 06:17:45 +03:00
VALUE str = rb_iseq_disasm_recursive(iseq, rb_str_new(0, 0));
rb_str_resize(str, RSTRING_LEN(str));
return str;
}
/*
* Estimates the number of instance variables that will be set on
* a given `class` with the initialize method defined in
* `initialize_iseq`
*/
attr_index_t
rb_estimate_iv_count(VALUE klass, const rb_iseq_t * initialize_iseq)
{
struct rb_id_table * iv_names = rb_id_table_create(0);
for (unsigned int i = 0; i < ISEQ_BODY(initialize_iseq)->ivc_size; i++) {
IVC cache = (IVC)&ISEQ_BODY(initialize_iseq)->is_entries[i];
if (cache->iv_set_name) {
rb_id_table_insert(iv_names, cache->iv_set_name, Qtrue);
}
}
attr_index_t count = (attr_index_t)rb_id_table_size(iv_names);
VALUE superclass = rb_class_superclass(klass);
count += RCLASS_EXT(superclass)->max_iv_count;
rb_id_table_free(iv_names);
return count;
}
2015-07-22 01:52:59 +03:00
/*
* call-seq:
* iseq.disasm -> str
* iseq.disassemble -> str
*
* Returns the instruction sequence as a +String+ in human readable form.
*
* puts RubyVM::InstructionSequence.compile('1 + 2').disasm
*
* Produces:
*
* == disasm: <RubyVM::InstructionSequence:<compiled>@<compiled>>==========
* 0000 trace 1 ( 1)
* 0002 putobject 1
* 0004 putobject 2
* 0006 opt_plus <ic:1>
* 0008 leave
*/
static VALUE
iseqw_disasm(VALUE self)
{
return rb_iseq_disasm(iseqw_check(self));
}
static int
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
iseq_iterate_children(const rb_iseq_t *iseq, void (*iter_func)(const rb_iseq_t *child_iseq, void *data), void *data)
{
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
unsigned int i;
VALUE *code = rb_iseq_original_iseq(iseq);
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
const rb_iseq_t *child;
VALUE all_children = rb_obj_hide(rb_ident_hash_new());
if (body->catch_table) {
for (i = 0; i < body->catch_table->size; i++) {
const struct iseq_catch_table_entry *entry =
UNALIGNED_MEMBER_PTR(body->catch_table, entries[i]);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
child = entry->iseq;
if (child) {
2021-10-03 16:34:45 +03:00
if (NIL_P(rb_hash_aref(all_children, (VALUE)child))) {
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
rb_hash_aset(all_children, (VALUE)child, Qtrue);
(*iter_func)(child, data);
}
}
}
}
for (i=0; i<body->iseq_size;) {
VALUE insn = code[i];
int len = insn_len(insn);
const char *types = insn_op_types(insn);
int j;
for (j=0; types[j]; j++) {
switch (types[j]) {
case TS_ISEQ:
child = (const rb_iseq_t *)code[i+j+1];
if (child) {
2021-10-03 16:34:45 +03:00
if (NIL_P(rb_hash_aref(all_children, (VALUE)child))) {
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
rb_hash_aset(all_children, (VALUE)child, Qtrue);
(*iter_func)(child, data);
}
}
break;
default:
break;
}
}
i += len;
}
return (int)RHASH_SIZE(all_children);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
}
static void
yield_each_children(const rb_iseq_t *child_iseq, void *data)
{
rb_yield(iseqw_new(child_iseq));
}
/*
* call-seq:
* iseq.each_child{|child_iseq| ...} -> iseq
*
* Iterate all direct child instruction sequences.
* Iteration order is implementation/version defined
* so that people should not rely on the order.
*/
static VALUE
iseqw_each_child(VALUE self)
{
const rb_iseq_t *iseq = iseqw_check(self);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
iseq_iterate_children(iseq, yield_each_children, NULL);
return self;
}
static void
push_event_info(const rb_iseq_t *iseq, rb_event_flag_t events, int line, VALUE ary)
{
#define C(ev, cstr, l) if (events & ev) rb_ary_push(ary, rb_ary_new_from_args(2, l, ID2SYM(rb_intern(cstr))));
C(RUBY_EVENT_CLASS, "class", rb_iseq_first_lineno(iseq));
C(RUBY_EVENT_CALL, "call", rb_iseq_first_lineno(iseq));
C(RUBY_EVENT_B_CALL, "b_call", rb_iseq_first_lineno(iseq));
C(RUBY_EVENT_LINE, "line", INT2FIX(line));
C(RUBY_EVENT_END, "end", INT2FIX(line));
C(RUBY_EVENT_RETURN, "return", INT2FIX(line));
C(RUBY_EVENT_B_RETURN, "b_return", INT2FIX(line));
#undef C
}
/*
* call-seq:
* iseq.trace_points -> ary
*
* Return trace points in the instruction sequence.
* Return an array of [line, event_symbol] pair.
*/
static VALUE
iseqw_trace_points(VALUE self)
{
const rb_iseq_t *iseq = iseqw_check(self);
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
unsigned int i;
VALUE ary = rb_ary_new();
for (i=0; i<body->insns_info.size; i++) {
const struct iseq_insn_info_entry *entry = &body->insns_info.body[i];
if (entry->events) {
push_event_info(iseq, entry->events, entry->line_no, ary);
}
}
return ary;
}
/*
* Returns the instruction sequence containing the given proc or method.
*
* For example, using irb:
*
* # a proc
* > p = proc { num = 1 + 2 }
* > RubyVM::InstructionSequence.of(p)
* > #=> <RubyVM::InstructionSequence:block in irb_binding@(irb)>
*
* # for a method
* > def foo(bar); puts bar; end
* > RubyVM::InstructionSequence.of(method(:foo))
* > #=> <RubyVM::InstructionSequence:foo@(irb)>
*
* Using ::compile_file:
*
* # /tmp/iseq_of.rb
* def hello
* puts "hello, world"
* end
*
* $a_global_proc = proc { str = 'a' + 'b' }
*
* # in irb
* > require '/tmp/iseq_of.rb'
*
* # first the method hello
* > RubyVM::InstructionSequence.of(method(:hello))
* > #=> #<RubyVM::InstructionSequence:0x007fb73d7cb1d0>
*
* # then the global proc
* > RubyVM::InstructionSequence.of($a_global_proc)
* > #=> #<RubyVM::InstructionSequence:0x007fb73d7caf78>
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_s_of(VALUE klass, VALUE body)
{
2015-07-22 01:52:59 +03:00
const rb_iseq_t *iseq = NULL;
if (rb_obj_is_proc(body)) {
iseq = vm_proc_iseq(body);
2015-07-22 01:52:59 +03:00
if (!rb_obj_is_iseq((VALUE)iseq)) {
iseq = NULL;
}
}
else if (rb_obj_is_method(body)) {
iseq = rb_method_iseq(body);
}
else if (rb_typeddata_is_instance_of(body, &iseqw_data_type)) {
return body;
}
return iseq ? iseqw_new(iseq) : Qnil;
}
/*
* call-seq:
* InstructionSequence.disasm(body) -> str
* InstructionSequence.disassemble(body) -> str
*
* Takes +body+, a Method or Proc object, and returns a String with the
* human readable instructions for +body+.
*
* For a Method object:
*
* # /tmp/method.rb
* def hello
* puts "hello, world"
* end
*
* puts RubyVM::InstructionSequence.disasm(method(:hello))
*
* Produces:
*
* == disasm: <RubyVM::InstructionSequence:hello@/tmp/method.rb>============
* 0000 trace 8 ( 1)
* 0002 trace 1 ( 2)
* 0004 putself
* 0005 putstring "hello, world"
* 0007 send :puts, 1, nil, 8, <ic:0>
* 0013 trace 16 ( 3)
* 0015 leave ( 2)
*
* For a Proc:
*
* # /tmp/proc.rb
* p = proc { num = 1 + 2 }
* puts RubyVM::InstructionSequence.disasm(p)
*
* Produces:
*
* == disasm: <RubyVM::InstructionSequence:block in <main>@/tmp/proc.rb>===
* == catch table
* | catch type: redo st: 0000 ed: 0012 sp: 0000 cont: 0000
* | catch type: next st: 0000 ed: 0012 sp: 0000 cont: 0012
* |------------------------------------------------------------------------
* local table (size: 2, argc: 0 [opts: 0, rest: -1, post: 0, block: -1] s1)
* [ 2] num
* 0000 trace 1 ( 1)
* 0002 putobject 1
* 0004 putobject 2
* 0006 opt_plus <ic:1>
* 0008 dup
* 0009 setlocal num, 0
* 0012 leave
*
*/
static VALUE
2015-07-22 01:52:59 +03:00
iseqw_s_disasm(VALUE klass, VALUE body)
{
2015-07-22 01:52:59 +03:00
VALUE iseqw = iseqw_s_of(klass, body);
return NIL_P(iseqw) ? Qnil : rb_iseq_disasm(iseqw_check(iseqw));
}
const char *
* this commit is a result of refactoring. only renaming functions, moving definitions place, add/remove prototypes, deleting unused variables and removing yarv.h. This commit doesn't change any behavior of ruby/vm. * yarv.h, common.mk: remove yarv.h (contents are moved to yarvcore.h). * error.c, eval_intern.h: include yarvcore.h instead yarv.h * rename some functions: * debug.[ch]: debug_*() -> ruby_debug_*() * iseq.c: iseq_*() -> rb_iseq_*(), ruby_iseq_disasm() * iseq.c: node_name() -> ruby_node_name() * vm.c: yarv_check_redefinition_opt_method() -> rb_vm_check_redefinition_opt_method() * some refactoring with checking -Wall. * array.c: remove rb_ary_ptr() (unused) and remove unused local variables. * object.c: add a prototype of rb_mod_module_exec(). * eval_intern.h (ruby_cref): set it inline. * eval_load.c (rb_load), yarvcore.c: yarv_load() -> rb_load_internal(). * parse.y: add a prototype of rb_parse_in_eval() (in eval.c). * process.c: add a prototype of rb_thread_stop_timer_thread() (in thread.c). * thread.c: remove raw_gets() function (unused) and fix some format mismatch (format mismatchs have remained yet. this is todo). * thread.c (rb_thread_wait_fd_rw): fix typo on label name. * thread_pthread.ci: comment out codes with USE_THREAD_CACHE. * vm.c (rb_svar, rb_backref_get, rb_backref_get, rb_lastline_get, rb_lastline_set) : moved from yarvcore.c. * vm.c (yarv_init_redefined_flag): add a prototype and rename yarv_opt_method_table to vm_opt_method_table. * vm.c (rb_thread_eval): moved from yarvcore.c. * yarvcore.c: remove unused global variables and fix to use nsdr(). git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@11652 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2007-02-07 04:25:05 +03:00
ruby_node_name(int node)
{
switch (node) {
#include "node_name.inc"
default:
rb_bug("unknown node: %d", node);
return 0;
}
}
static VALUE
register_label(struct st_table *table, unsigned long idx)
{
VALUE sym = rb_str_intern(rb_sprintf("label_%lu", idx));
st_insert(table, idx, sym);
return sym;
}
static VALUE
exception_type2symbol(VALUE type)
{
ID id;
switch (type) {
case CATCH_TYPE_RESCUE: CONST_ID(id, "rescue"); break;
case CATCH_TYPE_ENSURE: CONST_ID(id, "ensure"); break;
case CATCH_TYPE_RETRY: CONST_ID(id, "retry"); break;
case CATCH_TYPE_BREAK: CONST_ID(id, "break"); break;
case CATCH_TYPE_REDO: CONST_ID(id, "redo"); break;
case CATCH_TYPE_NEXT: CONST_ID(id, "next"); break;
default:
rb_bug("unknown exception type: %d", (int)type);
}
return ID2SYM(id);
}
static int
cdhash_each(VALUE key, VALUE value, VALUE ary)
{
rb_ary_push(ary, obj_resurrect(key));
rb_ary_push(ary, value);
return ST_CONTINUE;
}
static const rb_data_type_t label_wrapper = {
"label_wrapper",
{(void (*)(void *))rb_mark_tbl, (void (*)(void *))st_free_table, 0, 0,},
0, 0, RUBY_TYPED_FREE_IMMEDIATELY
};
#define DECL_ID(name) \
static ID id_##name
#define INIT_ID(name) \
id_##name = rb_intern(#name)
static VALUE
iseq_type_id(enum rb_iseq_type type)
{
DECL_ID(top);
DECL_ID(method);
DECL_ID(block);
DECL_ID(class);
DECL_ID(rescue);
DECL_ID(ensure);
DECL_ID(eval);
DECL_ID(main);
DECL_ID(plain);
if (id_top == 0) {
INIT_ID(top);
INIT_ID(method);
INIT_ID(block);
INIT_ID(class);
INIT_ID(rescue);
INIT_ID(ensure);
INIT_ID(eval);
INIT_ID(main);
INIT_ID(plain);
}
switch (type) {
case ISEQ_TYPE_TOP: return id_top;
case ISEQ_TYPE_METHOD: return id_method;
case ISEQ_TYPE_BLOCK: return id_block;
case ISEQ_TYPE_CLASS: return id_class;
case ISEQ_TYPE_RESCUE: return id_rescue;
case ISEQ_TYPE_ENSURE: return id_ensure;
case ISEQ_TYPE_EVAL: return id_eval;
case ISEQ_TYPE_MAIN: return id_main;
case ISEQ_TYPE_PLAIN: return id_plain;
};
rb_bug("unsupported iseq type: %d", (int)type);
}
static VALUE
iseq_data_to_ary(const rb_iseq_t *iseq)
{
unsigned int i;
long l;
const struct rb_iseq_constant_body *const iseq_body = ISEQ_BODY(iseq);
const struct iseq_insn_info_entry *prev_insn_info;
unsigned int pos;
int last_line = 0;
VALUE *seq, *iseq_original;
VALUE val = rb_ary_new();
ID type; /* Symbol */
VALUE locals = rb_ary_new();
VALUE params = rb_hash_new();
VALUE body = rb_ary_new(); /* [[:insn1, ...], ...] */
VALUE nbody;
VALUE exception = rb_ary_new(); /* [[....]] */
VALUE misc = rb_hash_new();
static ID insn_syms[VM_INSTRUCTION_SIZE/2]; /* w/o-trace only */
struct st_table *labels_table = st_init_numtable();
VALUE labels_wrapper = TypedData_Wrap_Struct(0, &label_wrapper, labels_table);
if (insn_syms[0] == 0) {
int i;
for (i=0; i<numberof(insn_syms); i++) {
2019-04-23 04:19:47 +03:00
insn_syms[i] = rb_intern(insn_name(i));
}
}
/* type */
type = iseq_type_id(iseq_body->type);
/* locals */
for (i=0; i<iseq_body->local_table_size; i++) {
ID lid = iseq_body->local_table[i];
if (lid) {
if (rb_id2str(lid)) {
rb_ary_push(locals, ID2SYM(lid));
}
else { /* hidden variable from id_internal() */
rb_ary_push(locals, ULONG2NUM(iseq_body->local_table_size-i+1));
}
}
else {
rb_ary_push(locals, ID2SYM(rb_intern("#arg_rest")));
}
}
/* params */
{
const struct rb_iseq_param_keyword *const keyword = iseq_body->param.keyword;
int j;
2022-07-21 19:23:58 +03:00
if (iseq_body->param.flags.has_opt) {
int len = iseq_body->param.opt_num + 1;
VALUE arg_opt_labels = rb_ary_new2(len);
2022-07-21 19:23:58 +03:00
for (j = 0; j < len; j++) {
VALUE l = register_label(labels_table, iseq_body->param.opt_table[j]);
rb_ary_push(arg_opt_labels, l);
}
rb_hash_aset(params, ID2SYM(rb_intern("opt")), arg_opt_labels);
}
2022-07-21 19:23:58 +03:00
/* commit */
if (iseq_body->param.flags.has_lead) rb_hash_aset(params, ID2SYM(rb_intern("lead_num")), INT2FIX(iseq_body->param.lead_num));
if (iseq_body->param.flags.has_post) rb_hash_aset(params, ID2SYM(rb_intern("post_num")), INT2FIX(iseq_body->param.post_num));
if (iseq_body->param.flags.has_post) rb_hash_aset(params, ID2SYM(rb_intern("post_start")), INT2FIX(iseq_body->param.post_start));
if (iseq_body->param.flags.has_rest) rb_hash_aset(params, ID2SYM(rb_intern("rest_start")), INT2FIX(iseq_body->param.rest_start));
if (iseq_body->param.flags.has_block) rb_hash_aset(params, ID2SYM(rb_intern("block_start")), INT2FIX(iseq_body->param.block_start));
if (iseq_body->param.flags.has_kw) {
VALUE keywords = rb_ary_new();
int i, j;
for (i=0; i<keyword->required_num; i++) {
rb_ary_push(keywords, ID2SYM(keyword->table[i]));
}
for (j=0; i<keyword->num; i++, j++) {
VALUE key = rb_ary_new_from_args(1, ID2SYM(keyword->table[i]));
2022-11-15 07:24:08 +03:00
if (!UNDEF_P(keyword->default_values[j])) {
rb_ary_push(key, keyword->default_values[j]);
}
rb_ary_push(keywords, key);
}
2022-07-21 19:23:58 +03:00
rb_hash_aset(params, ID2SYM(rb_intern("kwbits")),
INT2FIX(keyword->bits_start));
rb_hash_aset(params, ID2SYM(rb_intern("keyword")), keywords);
}
if (iseq_body->param.flags.has_kwrest) rb_hash_aset(params, ID2SYM(rb_intern("kwrest")), INT2FIX(keyword->rest_start));
if (iseq_body->param.flags.ambiguous_param0) rb_hash_aset(params, ID2SYM(rb_intern("ambiguous_param0")), Qtrue);
}
/* body */
iseq_original = rb_iseq_original_iseq((rb_iseq_t *)iseq);
for (seq = iseq_original; seq < iseq_original + iseq_body->iseq_size; ) {
VALUE insn = *seq++;
int j, len = insn_len(insn);
VALUE *nseq = seq + len - 1;
VALUE ary = rb_ary_new2(len);
2019-04-23 04:19:47 +03:00
rb_ary_push(ary, ID2SYM(insn_syms[insn%numberof(insn_syms)]));
for (j=0; j<len-1; j++, seq++) {
enum ruby_insn_type_chars op_type = insn_op_type(insn, j);
2022-07-21 19:23:58 +03:00
switch (op_type) {
case TS_OFFSET: {
unsigned long idx = nseq - iseq_original + *seq;
rb_ary_push(ary, register_label(labels_table, idx));
break;
}
case TS_LINDEX:
case TS_NUM:
rb_ary_push(ary, INT2FIX(*seq));
break;
case TS_VALUE:
rb_ary_push(ary, obj_resurrect(*seq));
break;
case TS_ISEQ:
{
const rb_iseq_t *iseq = (rb_iseq_t *)*seq;
if (iseq) {
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
VALUE val = iseq_data_to_ary(rb_iseq_check(iseq));
rb_ary_push(ary, val);
}
else {
rb_ary_push(ary, Qnil);
}
}
break;
case TS_IC:
New constant caching insn: opt_getconstant_path Previously YARV bytecode implemented constant caching by having a pair of instructions, opt_getinlinecache and opt_setinlinecache, wrapping a series of getconstant calls (with putobject providing supporting arguments). This commit replaces that pattern with a new instruction, opt_getconstant_path, handling both getting/setting the inline cache and fetching the constant on a cache miss. This is implemented by storing the full constant path as a null-terminated array of IDs inside of the IC structure. idNULL is used to signal an absolute constant reference. $ ./miniruby --dump=insns -e '::Foo::Bar::Baz' == disasm: #<ISeq:<main>@-e:1 (1,0)-(1,13)> (catch: FALSE) 0000 opt_getconstant_path <ic:0 ::Foo::Bar::Baz> ( 1)[Li] 0002 leave The motivation for this is that we had increasingly found the need to disassemble the instructions between the opt_getinlinecache and opt_setinlinecache in order to determine the constant we are fetching, or otherwise store metadata. This disassembly was done: * In opt_setinlinecache, to register the IC against the constant names it is using for granular invalidation. * In rb_iseq_free, to unregister the IC from the invalidation table. * In YJIT to find the position of a opt_getinlinecache instruction to invalidate it when the cache is populated * In YJIT to register the constant names being used for invalidation. With this change we no longe need disassemly for these (in fact rb_iseq_each is now unused), as the list of constant names being referenced is held in the IC. This should also make it possible to make more optimizations in the future. This may also reduce the size of iseqs, as previously each segment required 32 bytes (on 64-bit platforms) for each constant segment. This implementation only stores one ID per-segment. There should be no significant performance change between this and the previous implementation. Previously opt_getinlinecache was a "leaf" instruction, but it included a jump (almost always to a separate cache line). Now opt_getconstant_path is a non-leaf (it may raise/autoload/call const_missing) but it does not jump. These seem to even out.
2022-08-10 20:35:48 +03:00
{
VALUE list = rb_ary_new();
const ID *ids = ((IC)*seq)->segments;
while (*ids) {
rb_ary_push(list, ID2SYM(*ids++));
}
rb_ary_push(ary, list);
}
break;
case TS_IVC:
2022-02-03 05:21:41 +03:00
case TS_ICVARC:
case TS_ISE:
{
union iseq_inline_storage_entry *is = (union iseq_inline_storage_entry *)*seq;
rb_ary_push(ary, INT2FIX(is - ISEQ_IS_ENTRY_START(ISEQ_BODY(iseq), op_type)));
}
break;
case TS_CALLDATA:
{
struct rb_call_data *cd = (struct rb_call_data *)*seq;
const struct rb_callinfo *ci = cd->ci;
VALUE e = rb_hash_new();
int argc = vm_ci_argc(ci);
ID mid = vm_ci_mid(ci);
rb_hash_aset(e, ID2SYM(rb_intern("mid")), mid ? ID2SYM(mid) : Qnil);
rb_hash_aset(e, ID2SYM(rb_intern("flag")), UINT2NUM(vm_ci_flag(ci)));
if (vm_ci_flag(ci) & VM_CALL_KWARG) {
const struct rb_callinfo_kwarg *kwarg = vm_ci_kwarg(ci);
int i;
VALUE kw = rb_ary_new2((long)kwarg->keyword_len);
argc -= kwarg->keyword_len;
for (i = 0; i < kwarg->keyword_len; i++) {
rb_ary_push(kw, kwarg->keywords[i]);
}
rb_hash_aset(e, ID2SYM(rb_intern("kw_arg")), kw);
}
2022-07-21 19:23:58 +03:00
rb_hash_aset(e, ID2SYM(rb_intern("orig_argc")),
INT2FIX(argc));
rb_ary_push(ary, e);
}
break;
case TS_ID:
rb_ary_push(ary, ID2SYM(*seq));
break;
case TS_CDHASH:
{
VALUE hash = *seq;
VALUE val = rb_ary_new();
int i;
2022-07-21 19:23:58 +03:00
rb_hash_foreach(hash, cdhash_each, val);
2022-07-21 19:23:58 +03:00
for (i=0; i<RARRAY_LEN(val); i+=2) {
VALUE pos = FIX2INT(rb_ary_entry(val, i+1));
unsigned long idx = nseq - iseq_original + pos;
2022-07-21 19:23:58 +03:00
rb_ary_store(val, i+1,
register_label(labels_table, idx));
}
rb_ary_push(ary, val);
}
break;
case TS_FUNCPTR:
{
#if SIZEOF_VALUE <= SIZEOF_LONG
VALUE val = LONG2NUM((SIGNED_VALUE)*seq);
#else
VALUE val = LL2NUM((SIGNED_VALUE)*seq);
#endif
rb_ary_push(ary, val);
}
break;
2019-11-10 08:40:38 +03:00
case TS_BUILTIN:
{
VALUE val = rb_hash_new();
#if SIZEOF_VALUE <= SIZEOF_LONG
VALUE func_ptr = LONG2NUM((SIGNED_VALUE)((RB_BUILTIN)*seq)->func_ptr);
#else
VALUE func_ptr = LL2NUM((SIGNED_VALUE)((RB_BUILTIN)*seq)->func_ptr);
#endif
rb_hash_aset(val, ID2SYM(rb_intern("func_ptr")), func_ptr);
rb_hash_aset(val, ID2SYM(rb_intern("argc")), INT2NUM(((RB_BUILTIN)*seq)->argc));
rb_hash_aset(val, ID2SYM(rb_intern("index")), INT2NUM(((RB_BUILTIN)*seq)->index));
rb_hash_aset(val, ID2SYM(rb_intern("name")), rb_str_new_cstr(((RB_BUILTIN)*seq)->name));
rb_ary_push(ary, val);
}
break;
default:
rb_bug("unknown operand: %c", insn_op_type(insn, j));
}
}
rb_ary_push(body, ary);
}
nbody = body;
/* exception */
if (iseq_body->catch_table) for (i=0; i<iseq_body->catch_table->size; i++) {
VALUE ary = rb_ary_new();
const struct iseq_catch_table_entry *entry =
UNALIGNED_MEMBER_PTR(iseq_body->catch_table, entries[i]);
rb_ary_push(ary, exception_type2symbol(entry->type));
if (entry->iseq) {
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
rb_ary_push(ary, iseq_data_to_ary(rb_iseq_check(entry->iseq)));
}
else {
rb_ary_push(ary, Qnil);
}
rb_ary_push(ary, register_label(labels_table, entry->start));
rb_ary_push(ary, register_label(labels_table, entry->end));
rb_ary_push(ary, register_label(labels_table, entry->cont));
rb_ary_push(ary, UINT2NUM(entry->sp));
rb_ary_push(exception, ary);
}
/* make body with labels and insert line number */
body = rb_ary_new();
prev_insn_info = NULL;
#ifdef USE_ISEQ_NODE_ID
VALUE node_ids = rb_ary_new();
#endif
for (l=0, pos=0; l<RARRAY_LEN(nbody); l++) {
const struct iseq_insn_info_entry *info;
VALUE ary = RARRAY_AREF(nbody, l);
st_data_t label;
if (st_lookup(labels_table, pos, &label)) {
rb_ary_push(body, (VALUE)label);
}
info = get_insn_info(iseq, pos);
#ifdef USE_ISEQ_NODE_ID
rb_ary_push(node_ids, INT2FIX(info->node_id));
#endif
if (prev_insn_info != info) {
int line = info->line_no;
rb_event_flag_t events = info->events;
if (line > 0 && last_line != line) {
rb_ary_push(body, INT2FIX(line));
last_line = line;
}
#define CHECK_EVENT(ev) if (events & ev) rb_ary_push(body, ID2SYM(rb_intern(#ev)));
CHECK_EVENT(RUBY_EVENT_LINE);
CHECK_EVENT(RUBY_EVENT_CLASS);
CHECK_EVENT(RUBY_EVENT_END);
CHECK_EVENT(RUBY_EVENT_CALL);
CHECK_EVENT(RUBY_EVENT_RETURN);
CHECK_EVENT(RUBY_EVENT_B_CALL);
CHECK_EVENT(RUBY_EVENT_B_RETURN);
#undef CHECK_EVENT
prev_insn_info = info;
}
rb_ary_push(body, ary);
pos += RARRAY_LENINT(ary); /* reject too huge data */
}
RB_GC_GUARD(nbody);
RB_GC_GUARD(labels_wrapper);
rb_hash_aset(misc, ID2SYM(rb_intern("arg_size")), INT2FIX(iseq_body->param.size));
rb_hash_aset(misc, ID2SYM(rb_intern("local_size")), INT2FIX(iseq_body->local_table_size));
rb_hash_aset(misc, ID2SYM(rb_intern("stack_max")), INT2FIX(iseq_body->stack_max));
rb_hash_aset(misc, ID2SYM(rb_intern("node_id")), INT2FIX(iseq_body->location.node_id));
rb_hash_aset(misc, ID2SYM(rb_intern("code_location")),
rb_ary_new_from_args(4,
INT2FIX(iseq_body->location.code_location.beg_pos.lineno),
INT2FIX(iseq_body->location.code_location.beg_pos.column),
INT2FIX(iseq_body->location.code_location.end_pos.lineno),
INT2FIX(iseq_body->location.code_location.end_pos.column)));
#ifdef USE_ISEQ_NODE_ID
rb_hash_aset(misc, ID2SYM(rb_intern("node_ids")), node_ids);
#endif
/*
* [:magic, :major_version, :minor_version, :format_type, :misc,
* :name, :path, :absolute_path, :start_lineno, :type, :locals, :args,
* :catch_table, :bytecode]
*/
rb_ary_push(val, rb_str_new2("YARVInstructionSequence/SimpleDataFormat"));
rb_ary_push(val, INT2FIX(ISEQ_MAJOR_VERSION)); /* major */
rb_ary_push(val, INT2FIX(ISEQ_MINOR_VERSION)); /* minor */
rb_ary_push(val, INT2FIX(1));
rb_ary_push(val, misc);
rb_ary_push(val, iseq_body->location.label);
rb_ary_push(val, rb_iseq_path(iseq));
rb_ary_push(val, rb_iseq_realpath(iseq));
rb_ary_push(val, RB_INT2NUM(iseq_body->location.first_lineno));
rb_ary_push(val, ID2SYM(type));
rb_ary_push(val, locals);
rb_ary_push(val, params);
rb_ary_push(val, exception);
rb_ary_push(val, body);
return val;
}
VALUE
rb_iseq_parameters(const rb_iseq_t *iseq, int is_proc)
{
int i, r;
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
const struct rb_iseq_param_keyword *const keyword = body->param.keyword;
VALUE a, args = rb_ary_new2(body->param.size);
ID req, opt, rest, block, key, keyrest;
#define PARAM_TYPE(type) rb_ary_push(a = rb_ary_new2(2), ID2SYM(type))
#define PARAM_ID(i) body->local_table[(i)]
#define PARAM(i, type) ( \
PARAM_TYPE(type), \
rb_id2str(PARAM_ID(i)) ? \
rb_ary_push(a, ID2SYM(PARAM_ID(i))) : \
a)
CONST_ID(req, "req");
CONST_ID(opt, "opt");
if (is_proc) {
for (i = 0; i < body->param.lead_num; i++) {
PARAM_TYPE(opt);
rb_ary_push(a, rb_id2str(PARAM_ID(i)) ? ID2SYM(PARAM_ID(i)) : Qnil);
rb_ary_push(args, a);
}
}
else {
for (i = 0; i < body->param.lead_num; i++) {
rb_ary_push(args, PARAM(i, req));
}
}
r = body->param.lead_num + body->param.opt_num;
for (; i < r; i++) {
PARAM_TYPE(opt);
if (rb_id2str(PARAM_ID(i))) {
rb_ary_push(a, ID2SYM(PARAM_ID(i)));
}
rb_ary_push(args, a);
}
if (body->param.flags.has_rest) {
CONST_ID(rest, "rest");
rb_ary_push(args, PARAM(body->param.rest_start, rest));
}
r = body->param.post_start + body->param.post_num;
if (is_proc) {
for (i = body->param.post_start; i < r; i++) {
PARAM_TYPE(opt);
rb_ary_push(a, rb_id2str(PARAM_ID(i)) ? ID2SYM(PARAM_ID(i)) : Qnil);
rb_ary_push(args, a);
}
}
else {
for (i = body->param.post_start; i < r; i++) {
rb_ary_push(args, PARAM(i, req));
}
}
if (body->param.flags.accepts_no_kwarg) {
ID nokey;
CONST_ID(nokey, "nokey");
PARAM_TYPE(nokey);
rb_ary_push(args, a);
}
if (body->param.flags.has_kw) {
i = 0;
if (keyword->required_num > 0) {
ID keyreq;
CONST_ID(keyreq, "keyreq");
for (; i < keyword->required_num; i++) {
PARAM_TYPE(keyreq);
if (rb_id2str(keyword->table[i])) {
rb_ary_push(a, ID2SYM(keyword->table[i]));
}
rb_ary_push(args, a);
}
}
CONST_ID(key, "key");
for (; i < keyword->num; i++) {
PARAM_TYPE(key);
if (rb_id2str(keyword->table[i])) {
rb_ary_push(a, ID2SYM(keyword->table[i]));
}
rb_ary_push(args, a);
}
* rewrite method/block parameter fitting logic to optimize keyword arguments/parameters and a splat argument. [Feature #10440] (Details are described in this ticket) Most of complex part is moved to vm_args.c. Now, ISeq#to_a does not catch up new instruction format. * vm_core.h: change iseq data structures. * introduce rb_call_info_kw_arg_t to represent keyword arguments. * add rb_call_info_t::kw_arg. * rename rb_iseq_t::arg_post_len to rb_iseq_t::arg_post_num. * rename rb_iseq_t::arg_keywords to arg_keyword_num. * rename rb_iseq_t::arg_keyword to rb_iseq_t::arg_keyword_bits. to represent keyword bitmap parameter index. This bitmap parameter shows that which keyword parameters are given or not given (0 for given). It is refered by `checkkeyword' instruction described bellow. * rename rb_iseq_t::arg_keyword_check to rb_iseq_t::arg_keyword_rest to represent keyword rest parameter index. * add rb_iseq_t::arg_keyword_default_values to represent default keyword values. * rename VM_CALL_ARGS_SKIP_SETUP to VM_CALL_ARGS_SIMPLE to represent (ci->flag & (SPLAT|BLOCKARG)) && ci->blockiseq == NULL && ci->kw_arg == NULL. * vm_insnhelper.c, vm_args.c: rewrite with refactoring. * rewrite splat argument code. * rewrite keyword arguments/parameters code. * merge method and block parameter fitting code into one code base. * vm.c, vm_eval.c: catch up these changes. * compile.c (new_callinfo): callinfo requires kw_arg parameter. * compile.c (compile_array_): check the last argument Hash object or not. If Hash object and all keys are Symbol literals, they are compiled to keyword arguments. * insns.def (checkkeyword): add new instruction. This instruction check the availability of corresponding keyword. For example, a method "def foo k1: 'v1'; end" is cimpiled to the following instructions. 0000 checkkeyword 2, 0 # check k1 is given. 0003 branchif 9 # if given, jump to address #9 0005 putstring "v1" 0007 setlocal_OP__WC__0 3 # k1 = 'v1' 0009 trace 8 0011 putnil 0012 trace 16 0014 leave * insns.def (opt_send_simple): removed and add new instruction "opt_send_without_block". * parse.y (new_args_tail_gen): reorder variables. Before this patch, a method "def foo(k1: 1, kr1:, k2: 2, **krest, &b)" has parameter variables "k1, kr1, k2, &b, internal_id, krest", but this patch reorders to "kr1, k1, k2, internal_id, krest, &b". (locate a block variable at last) * parse.y (vtable_pop): added. This function remove latest `n' variables from vtable. * iseq.c: catch up iseq data changes. * proc.c: ditto. * class.c (keyword_error): export as rb_keyword_error(). * common.mk: depend vm_args.c for vm.o. * hash.c (rb_hash_has_key): export. * internal.h: ditto. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48239 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2014-11-02 21:02:55 +03:00
}
if (body->param.flags.has_kwrest || body->param.flags.ruby2_keywords) {
ID param;
* rewrite method/block parameter fitting logic to optimize keyword arguments/parameters and a splat argument. [Feature #10440] (Details are described in this ticket) Most of complex part is moved to vm_args.c. Now, ISeq#to_a does not catch up new instruction format. * vm_core.h: change iseq data structures. * introduce rb_call_info_kw_arg_t to represent keyword arguments. * add rb_call_info_t::kw_arg. * rename rb_iseq_t::arg_post_len to rb_iseq_t::arg_post_num. * rename rb_iseq_t::arg_keywords to arg_keyword_num. * rename rb_iseq_t::arg_keyword to rb_iseq_t::arg_keyword_bits. to represent keyword bitmap parameter index. This bitmap parameter shows that which keyword parameters are given or not given (0 for given). It is refered by `checkkeyword' instruction described bellow. * rename rb_iseq_t::arg_keyword_check to rb_iseq_t::arg_keyword_rest to represent keyword rest parameter index. * add rb_iseq_t::arg_keyword_default_values to represent default keyword values. * rename VM_CALL_ARGS_SKIP_SETUP to VM_CALL_ARGS_SIMPLE to represent (ci->flag & (SPLAT|BLOCKARG)) && ci->blockiseq == NULL && ci->kw_arg == NULL. * vm_insnhelper.c, vm_args.c: rewrite with refactoring. * rewrite splat argument code. * rewrite keyword arguments/parameters code. * merge method and block parameter fitting code into one code base. * vm.c, vm_eval.c: catch up these changes. * compile.c (new_callinfo): callinfo requires kw_arg parameter. * compile.c (compile_array_): check the last argument Hash object or not. If Hash object and all keys are Symbol literals, they are compiled to keyword arguments. * insns.def (checkkeyword): add new instruction. This instruction check the availability of corresponding keyword. For example, a method "def foo k1: 'v1'; end" is cimpiled to the following instructions. 0000 checkkeyword 2, 0 # check k1 is given. 0003 branchif 9 # if given, jump to address #9 0005 putstring "v1" 0007 setlocal_OP__WC__0 3 # k1 = 'v1' 0009 trace 8 0011 putnil 0012 trace 16 0014 leave * insns.def (opt_send_simple): removed and add new instruction "opt_send_without_block". * parse.y (new_args_tail_gen): reorder variables. Before this patch, a method "def foo(k1: 1, kr1:, k2: 2, **krest, &b)" has parameter variables "k1, kr1, k2, &b, internal_id, krest", but this patch reorders to "kr1, k1, k2, internal_id, krest, &b". (locate a block variable at last) * parse.y (vtable_pop): added. This function remove latest `n' variables from vtable. * iseq.c: catch up iseq data changes. * proc.c: ditto. * class.c (keyword_error): export as rb_keyword_error(). * common.mk: depend vm_args.c for vm.o. * hash.c (rb_hash_has_key): export. * internal.h: ditto. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48239 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2014-11-02 21:02:55 +03:00
CONST_ID(keyrest, "keyrest");
PARAM_TYPE(keyrest);
if (body->param.flags.has_kwrest &&
rb_id2str(param = PARAM_ID(keyword->rest_start))) {
rb_ary_push(a, ID2SYM(param));
}
else if (body->param.flags.ruby2_keywords) {
rb_ary_push(a, ID2SYM(idPow));
}
rb_ary_push(args, a);
}
if (body->param.flags.has_block) {
CONST_ID(block, "block");
rb_ary_push(args, PARAM(body->param.block_start, block));
}
return args;
}
VALUE
rb_iseq_defined_string(enum defined_type type)
{
static const char expr_names[][18] = {
"nil",
"instance-variable",
"local-variable",
"global-variable",
"class variable",
"constant",
"method",
"yield",
"super",
"self",
"true",
"false",
"assignment",
"expression",
};
const char *estr;
if ((unsigned)(type - 1) >= (unsigned)numberof(expr_names)) rb_bug("unknown defined type %d", type);
estr = expr_names[type - 1];
return rb_fstring_cstr(estr);
}
/* A map from encoded_insn to insn_data: decoded insn number, its len,
* non-trace version of encoded insn, and trace version. */
static st_table *encoded_insn_data;
typedef struct insn_data_struct {
int insn;
int insn_len;
void *notrace_encoded_insn;
void *trace_encoded_insn;
} insn_data_t;
static insn_data_t insn_data[VM_INSTRUCTION_SIZE/2];
void
rb_vm_encoded_insn_data_table_init(void)
{
#if OPT_DIRECT_THREADED_CODE || OPT_CALL_THREADED_CODE
const void * const *table = rb_vm_get_insns_address_table();
#define INSN_CODE(insn) ((VALUE)table[insn])
#else
#define INSN_CODE(insn) (insn)
#endif
st_data_t insn;
encoded_insn_data = st_init_numtable_with_size(VM_INSTRUCTION_SIZE / 2);
for (insn = 0; insn < VM_INSTRUCTION_SIZE/2; insn++) {
st_data_t key1 = (st_data_t)INSN_CODE(insn);
st_data_t key2 = (st_data_t)INSN_CODE(insn + VM_INSTRUCTION_SIZE/2);
insn_data[insn].insn = (int)insn;
insn_data[insn].insn_len = insn_len(insn);
if (insn != BIN(opt_invokebuiltin_delegate_leave)) {
insn_data[insn].notrace_encoded_insn = (void *) key1;
insn_data[insn].trace_encoded_insn = (void *) key2;
}
else {
insn_data[insn].notrace_encoded_insn = (void *) INSN_CODE(BIN(opt_invokebuiltin_delegate));
insn_data[insn].trace_encoded_insn = (void *) INSN_CODE(BIN(opt_invokebuiltin_delegate) + VM_INSTRUCTION_SIZE/2);
}
st_add_direct(encoded_insn_data, key1, (st_data_t)&insn_data[insn]);
st_add_direct(encoded_insn_data, key2, (st_data_t)&insn_data[insn]);
}
}
int
rb_vm_insn_addr2insn(const void *addr)
{
st_data_t key = (st_data_t)addr;
st_data_t val;
if (st_lookup(encoded_insn_data, key, &val)) {
insn_data_t *e = (insn_data_t *)val;
return (int)e->insn;
}
rb_bug("rb_vm_insn_addr2insn: invalid insn address: %p", addr);
}
// Unlike rb_vm_insn_addr2insn, this function can return trace opcode variants.
int
rb_vm_insn_addr2opcode(const void *addr)
{
st_data_t key = (st_data_t)addr;
st_data_t val;
if (st_lookup(encoded_insn_data, key, &val)) {
insn_data_t *e = (insn_data_t *)val;
int opcode = e->insn;
if (addr == e->trace_encoded_insn) {
opcode += VM_INSTRUCTION_SIZE/2;
}
return opcode;
}
rb_bug("rb_vm_insn_addr2opcode: invalid insn address: %p", addr);
}
// Decode `ISEQ_BODY(iseq)->iseq_encoded[i]` to an insn.
int
rb_vm_insn_decode(const VALUE encoded)
{
#if OPT_DIRECT_THREADED_CODE || OPT_CALL_THREADED_CODE
int insn = rb_vm_insn_addr2insn((void *)encoded);
#else
int insn = (int)encoded;
#endif
return insn;
}
static inline int
encoded_iseq_trace_instrument(VALUE *iseq_encoded_insn, rb_event_flag_t turnon, bool remain_current_trace)
{
st_data_t key = (st_data_t)*iseq_encoded_insn;
st_data_t val;
if (st_lookup(encoded_insn_data, key, &val)) {
insn_data_t *e = (insn_data_t *)val;
if (remain_current_trace && key == (st_data_t)e->trace_encoded_insn) {
turnon = 1;
}
*iseq_encoded_insn = (VALUE) (turnon ? e->trace_encoded_insn : e->notrace_encoded_insn);
return e->insn_len;
}
rb_bug("trace_instrument: invalid insn address: %p", (void *)*iseq_encoded_insn);
}
void
rb_iseq_trace_flag_cleared(const rb_iseq_t *iseq, size_t pos)
{
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
VALUE *iseq_encoded = (VALUE *)body->iseq_encoded;
encoded_iseq_trace_instrument(&iseq_encoded[pos], 0, false);
}
// We need to fire call events on instructions with b_call events if the block
// is running as a method. So, if we are listening for call events, then
// instructions that have b_call events need to become trace variants.
// Use this function when making decisions about recompiling to trace variants.
static inline rb_event_flag_t
add_bmethod_events(rb_event_flag_t events)
{
if (events & RUBY_EVENT_CALL) {
events |= RUBY_EVENT_B_CALL;
}
if (events & RUBY_EVENT_RETURN) {
events |= RUBY_EVENT_B_RETURN;
}
return events;
}
// Note, to support call/return events for bmethods, turnon_event can have more events than tpval.
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
static int
iseq_add_local_tracepoint(const rb_iseq_t *iseq, rb_event_flag_t turnon_events, VALUE tpval, unsigned int target_line)
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
{
unsigned int pc;
int n = 0;
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
VALUE *iseq_encoded = (VALUE *)body->iseq_encoded;
VM_ASSERT(ISEQ_EXECUTABLE_P(iseq));
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
for (pc=0; pc<body->iseq_size;) {
const struct iseq_insn_info_entry *entry = get_insn_info(iseq, pc);
rb_event_flag_t pc_events = entry->events;
rb_event_flag_t target_events = turnon_events;
unsigned int line = (int)entry->line_no;
if (target_line == 0 || target_line == line) {
/* ok */
}
else {
target_events &= ~RUBY_EVENT_LINE;
}
if (pc_events & target_events) {
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
n++;
}
pc += encoded_iseq_trace_instrument(&iseq_encoded[pc], pc_events & (target_events | iseq->aux.exec.global_trace_events), true);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
}
if (n > 0) {
if (iseq->aux.exec.local_hooks == NULL) {
((rb_iseq_t *)iseq)->aux.exec.local_hooks = RB_ZALLOC(rb_hook_list_t);
iseq->aux.exec.local_hooks->is_local = true;
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
}
rb_hook_list_connect_tracepoint((VALUE)iseq, iseq->aux.exec.local_hooks, tpval, target_line);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
}
return n;
}
struct trace_set_local_events_struct {
rb_event_flag_t turnon_events;
VALUE tpval;
unsigned int target_line;
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
int n;
};
static void
iseq_add_local_tracepoint_i(const rb_iseq_t *iseq, void *p)
{
struct trace_set_local_events_struct *data = (struct trace_set_local_events_struct *)p;
data->n += iseq_add_local_tracepoint(iseq, data->turnon_events, data->tpval, data->target_line);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
iseq_iterate_children(iseq, iseq_add_local_tracepoint_i, p);
}
int
rb_iseq_add_local_tracepoint_recursively(const rb_iseq_t *iseq, rb_event_flag_t turnon_events, VALUE tpval, unsigned int target_line, bool target_bmethod)
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
{
struct trace_set_local_events_struct data;
if (target_bmethod) {
turnon_events = add_bmethod_events(turnon_events);
}
data.turnon_events = turnon_events;
data.tpval = tpval;
data.target_line = target_line;
data.n = 0;
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
iseq_add_local_tracepoint_i(iseq, (void *)&data);
if (0) rb_funcall(Qnil, rb_intern("puts"), 1, rb_iseq_disasm(iseq)); /* for debug */
return data.n;
}
static int
iseq_remove_local_tracepoint(const rb_iseq_t *iseq, VALUE tpval)
{
int n = 0;
if (iseq->aux.exec.local_hooks) {
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
unsigned int pc;
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
VALUE *iseq_encoded = (VALUE *)body->iseq_encoded;
rb_event_flag_t local_events = 0;
rb_hook_list_remove_tracepoint(iseq->aux.exec.local_hooks, tpval);
local_events = iseq->aux.exec.local_hooks->events;
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
if (local_events == 0) {
rb_hook_list_free(iseq->aux.exec.local_hooks);
((rb_iseq_t *)iseq)->aux.exec.local_hooks = NULL;
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
}
local_events = add_bmethod_events(local_events);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
for (pc = 0; pc<body->iseq_size;) {
rb_event_flag_t pc_events = rb_iseq_event_flags(iseq, pc);
pc += encoded_iseq_trace_instrument(&iseq_encoded[pc], pc_events & (local_events | iseq->aux.exec.global_trace_events), false);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
}
}
return n;
}
struct trace_clear_local_events_struct {
VALUE tpval;
int n;
};
static void
iseq_remove_local_tracepoint_i(const rb_iseq_t *iseq, void *p)
{
struct trace_clear_local_events_struct *data = (struct trace_clear_local_events_struct *)p;
data->n += iseq_remove_local_tracepoint(iseq, data->tpval);
iseq_iterate_children(iseq, iseq_remove_local_tracepoint_i, p);
}
int
rb_iseq_remove_local_tracepoint_recursively(const rb_iseq_t *iseq, VALUE tpval)
{
struct trace_clear_local_events_struct data;
data.tpval = tpval;
data.n = 0;
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
iseq_remove_local_tracepoint_i(iseq, (void *)&data);
return data.n;
}
void
rb_iseq_trace_set(const rb_iseq_t *iseq, rb_event_flag_t turnon_events)
{
if (iseq->aux.exec.global_trace_events == turnon_events) {
return;
}
if (!ISEQ_EXECUTABLE_P(iseq)) {
/* this is building ISeq */
return;
}
else {
unsigned int pc;
const struct rb_iseq_constant_body *const body = ISEQ_BODY(iseq);
VALUE *iseq_encoded = (VALUE *)body->iseq_encoded;
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
rb_event_flag_t enabled_events;
rb_event_flag_t local_events = iseq->aux.exec.local_hooks ? iseq->aux.exec.local_hooks->events : 0;
((rb_iseq_t *)iseq)->aux.exec.global_trace_events = turnon_events;
enabled_events = add_bmethod_events(turnon_events | local_events);
Support targetting TracePoint [Feature #15289] * vm_trace.c (rb_tracepoint_enable_for_target): support targetting TracePoint. [Feature #15289] Tragetting TracePoint is only enabled on specified method, proc and so on, example: `tp.enable(target: code)`. `code` should be consisted of InstructionSeuqnece (iseq) (RubyVM::InstructionSeuqnece.of(code) should not return nil) If code is a tree of iseq, TracePoint is enabled on all of iseqs in a tree. Enabled tragetting TracePoints can not enabled again with and without target. * vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks` to store local hooks. `rb_iseq_t::aux::trace_events` is renamed to `global_trace_events` to contrast with `local_hooks`. * vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running` to represent how many Threads/Fibers are used this list. If this field is 0, nobody using this hooks and we can delete it. This is why we can remove code from cont.c. * vm_core.h (rb_vm_t): because of above change, we can eliminate `rb_vm_t::trace_running` field. Also renamed from `rb_vm_t::event_hooks` to `global_hooks`. * vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed from `ruby_vm_event_enabled_flags. * vm_core.h, vm.c (ruby_vm_event_local_num): added to count enabled targetting TracePoints. * vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts hook list. * vm_core.h (rb_vm_global_hooks): added for convinience. * method.h (rb_method_bmethod_t): added to maintain Proc and `rb_hook_list_t` for bmethod (defined by define_method). * prelude.rb (TracePoint#enable): extracet a keyword parameter (because it is easy than writing in C). It calls `TracePoint#__enable` internal method written in C. * vm_insnhelper.c (vm_trace): check also iseq->local_hooks. * vm.c (invoke_bmethod): check def->body.bmethod.hooks. * vm.c (hook_before_rewind): check iseq->local_hooks and def->body.bmethod.hooks before rewind by exception. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-26 21:16:39 +03:00
for (pc=0; pc<body->iseq_size;) {
rb_event_flag_t pc_events = rb_iseq_event_flags(iseq, pc);
pc += encoded_iseq_trace_instrument(&iseq_encoded[pc], pc_events & enabled_events, true);
}
}
}
bool rb_vm_call_ivar_attrset_p(const vm_call_handler ch);
void rb_vm_cc_general(const struct rb_callcache *cc);
static int
clear_attr_ccs_i(void *vstart, void *vend, size_t stride, void *data)
{
VALUE v = (VALUE)vstart;
for (; v != (VALUE)vend; v += stride) {
void *ptr = asan_poisoned_object_p(v);
asan_unpoison_object(v, false);
if (imemo_type_p(v, imemo_callcache) && rb_vm_call_ivar_attrset_p(((const struct rb_callcache *)v)->call_)) {
rb_vm_cc_general((struct rb_callcache *)v);
}
asan_poison_object_if(ptr, v);
}
return 0;
}
void
rb_clear_attr_ccs(void)
{
rb_objspace_each_objects(clear_attr_ccs_i, NULL);
}
static int
trace_set_i(void *vstart, void *vend, size_t stride, void *data)
{
rb_event_flag_t turnon_events = *(rb_event_flag_t *)data;
VALUE v = (VALUE)vstart;
for (; v != (VALUE)vend; v += stride) {
void *ptr = asan_poisoned_object_p(v);
asan_unpoison_object(v, false);
if (rb_obj_is_iseq(v)) {
rb_iseq_trace_set(rb_iseq_check((rb_iseq_t *)v), turnon_events);
}
else if (imemo_type_p(v, imemo_callcache) && rb_vm_call_ivar_attrset_p(((const struct rb_callcache *)v)->call_)) {
rb_vm_cc_general((struct rb_callcache *)v);
}
asan_poison_object_if(ptr, v);
}
return 0;
}
void
rb_iseq_trace_set_all(rb_event_flag_t turnon_events)
{
rb_objspace_each_objects(trace_set_i, &turnon_events);
}
VALUE
rb_iseqw_local_variables(VALUE iseqval)
{
return rb_iseq_local_variables(iseqw_check(iseqval));
}
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
/*
* call-seq:
* iseq.to_binary(extra_data = nil) -> binary str
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
*
* Returns serialized iseq binary format data as a String object.
* A corresponding iseq object is created by
* RubyVM::InstructionSequence.load_from_binary() method.
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
*
* String extra_data will be saved with binary data.
* You can access this data with
* RubyVM::InstructionSequence.load_from_binary_extra_data(binary).
*
* Note that the translated binary data is not portable.
* You can not move this binary data to another machine.
* You can not use the binary data which is created by another
* version/another architecture of Ruby.
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
*/
static VALUE
iseqw_to_binary(int argc, VALUE *argv, VALUE self)
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
{
VALUE opt = !rb_check_arity(argc, 0, 1) ? Qnil : argv[0];
return rb_iseq_ibf_dump(iseqw_check(self), opt);
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
}
/*
* call-seq:
* RubyVM::InstructionSequence.load_from_binary(binary) -> iseq
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
*
* Load an iseq object from binary format String object
* created by RubyVM::InstructionSequence.to_binary.
*
* This loader does not have a verifier, so that loading broken/modified
* binary causes critical problem.
*
* You should not load binary data provided by others.
* You should use binary data translated by yourself.
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
*/
static VALUE
iseqw_s_load_from_binary(VALUE self, VALUE str)
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
{
return iseqw_new(rb_iseq_ibf_load(str));
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
}
/*
* call-seq:
* RubyVM::InstructionSequence.load_from_binary_extra_data(binary) -> str
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
*
* Load extra data embed into binary format String object.
*/
static VALUE
iseqw_s_load_from_binary_extra_data(VALUE self, VALUE str)
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
{
return rb_iseq_ibf_load_extra_data(str);
* introduce new ISeq binary format serializer/de-serializer and a pre-compilation/runtime loader sample. [Feature #11788] * iseq.c: add new methods: * RubyVM::InstructionSequence#to_binary_format(extra_data = nil) * RubyVM::InstructionSequence.from_binary_format(binary) * RubyVM::InstructionSequence.from_binary_format_extra_data(binary) * compile.c: implement body of this new feature. * load.c (rb_load_internal0), iseq.c (rb_iseq_load_iseq): call RubyVM::InstructionSequence.load_iseq(fname) with loading script name if this method is defined. We can return any ISeq object as a result value. Otherwise loading will be continue as usual. This interface is not matured and is not extensible. So that we don't guarantee the future compatibility of this method. Basically, you should'nt use this method. * iseq.h: move ISEQ_MAJOR/MINOR_VERSION (and some definitions) from iseq.c. * encoding.c (rb_data_is_encoding), internal.h: added. * vm_core.h: add several supports for lazy load. * add USE_LAZY_LOAD macro to specify enable or disable of this feature. * add several fields to rb_iseq_t. * introduce new macro rb_iseq_check(). * insns.def: some check for lazy loading feature. * vm_insnhelper.c: ditto. * proc.c: ditto. * vm.c: ditto. * test/lib/iseq_loader_checker.rb: enabled iff suitable environment variables are provided. * test/runner.rb: enable lib/iseq_loader_checker.rb. * sample/iseq_loader.rb: add sample compiler and loader. $ ruby sample/iseq_loader.rb [dir] will compile all ruby scripts in [dir]. With default setting, this compile creates *.rb.yarb files in same directory of target .rb scripts. $ ruby -r sample/iseq_loader.rb [app] will run with enable to load compiled binary data. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52949 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2015-12-08 16:58:50 +03:00
}
#if VM_INSN_INFO_TABLE_IMPL == 2
/* An implementation of succinct bit-vector for insn_info table.
*
* A succinct bit-vector is a small and efficient data structure that provides
* a bit-vector augmented with an index for O(1) rank operation:
*
* rank(bv, n): the number of 1's within a range from index 0 to index n
*
* This can be used to lookup insn_info table from PC.
* For example, consider the following iseq and insn_info_table:
*
* iseq insn_info_table
* PC insn+operand position lineno event
* 0: insn1 0: 1 [Li]
* 2: insn2 2: 2 [Li] <= (A)
* 5: insn3 8: 3 [Li] <= (B)
* 8: insn4
*
* In this case, a succinct bit-vector whose indexes 0, 2, 8 is "1" and
* other indexes is "0", i.e., "101000001", is created.
* To lookup the lineno of insn2, calculate rank("10100001", 2) = 2, so
* the line (A) is the entry in question.
* To lookup the lineno of insn4, calculate rank("10100001", 8) = 3, so
* the line (B) is the entry in question.
*
* A naive implementation of succinct bit-vector works really well
* not only for large size but also for small size. However, it has
* tiny overhead for very small size. So, this implementation consist
* of two parts: one part is the "immediate" table that keeps rank result
* as a raw table, and the other part is a normal succinct bit-vector.
*/
#define IMMEDIATE_TABLE_SIZE 54 /* a multiple of 9, and < 128 */
struct succ_index_table {
uint64_t imm_part[IMMEDIATE_TABLE_SIZE / 9];
struct succ_dict_block {
unsigned int rank;
uint64_t small_block_ranks; /* 9 bits * 7 = 63 bits */
uint64_t bits[512/64];
} succ_part[FLEX_ARY_LEN];
};
#define imm_block_rank_set(v, i, r) (v) |= (uint64_t)(r) << (7 * (i))
#define imm_block_rank_get(v, i) (((int)((v) >> ((i) * 7))) & 0x7f)
#define small_block_rank_set(v, i, r) (v) |= (uint64_t)(r) << (9 * ((i) - 1))
#define small_block_rank_get(v, i) ((i) == 0 ? 0 : (((int)((v) >> (((i) - 1) * 9))) & 0x1ff))
static struct succ_index_table *
succ_index_table_create(int max_pos, int *data, int size)
{
const int imm_size = (max_pos < IMMEDIATE_TABLE_SIZE ? max_pos + 8 : IMMEDIATE_TABLE_SIZE) / 9;
const int succ_size = (max_pos < IMMEDIATE_TABLE_SIZE ? 0 : (max_pos - IMMEDIATE_TABLE_SIZE + 511)) / 512;
struct succ_index_table *sd =
rb_xcalloc_mul_add_mul(
imm_size, sizeof(uint64_t),
succ_size, sizeof(struct succ_dict_block));
int i, j, k, r;
r = 0;
for (j = 0; j < imm_size; j++) {
for (i = 0; i < 9; i++) {
if (r < size && data[r] == j * 9 + i) r++;
imm_block_rank_set(sd->imm_part[j], i, r);
}
}
for (k = 0; k < succ_size; k++) {
struct succ_dict_block *sd_block = &sd->succ_part[k];
int small_rank = 0;
sd_block->rank = r;
for (j = 0; j < 8; j++) {
uint64_t bits = 0;
if (j) small_block_rank_set(sd_block->small_block_ranks, j, small_rank);
for (i = 0; i < 64; i++) {
if (r < size && data[r] == k * 512 + j * 64 + i + IMMEDIATE_TABLE_SIZE) {
bits |= ((uint64_t)1) << i;
r++;
}
}
sd_block->bits[j] = bits;
small_rank += rb_popcount64(bits);
}
}
return sd;
}
static unsigned int *
succ_index_table_invert(int max_pos, struct succ_index_table *sd, int size)
{
const int imm_size = (max_pos < IMMEDIATE_TABLE_SIZE ? max_pos + 8 : IMMEDIATE_TABLE_SIZE) / 9;
const int succ_size = (max_pos < IMMEDIATE_TABLE_SIZE ? 0 : (max_pos - IMMEDIATE_TABLE_SIZE + 511)) / 512;
unsigned int *positions = ALLOC_N(unsigned int, size), *p;
int i, j, k, r = -1;
p = positions;
for (j = 0; j < imm_size; j++) {
for (i = 0; i < 9; i++) {
int nr = imm_block_rank_get(sd->imm_part[j], i);
if (r != nr) *p++ = j * 9 + i;
r = nr;
}
}
for (k = 0; k < succ_size; k++) {
for (j = 0; j < 8; j++) {
for (i = 0; i < 64; i++) {
if (sd->succ_part[k].bits[j] & (((uint64_t)1) << i)) {
*p++ = k * 512 + j * 64 + i + IMMEDIATE_TABLE_SIZE;
}
}
}
}
return positions;
}
static int
succ_index_lookup(const struct succ_index_table *sd, int x)
{
if (x < IMMEDIATE_TABLE_SIZE) {
const int i = x / 9;
const int j = x % 9;
return imm_block_rank_get(sd->imm_part[i], j);
}
else {
const int block_index = (x - IMMEDIATE_TABLE_SIZE) / 512;
const struct succ_dict_block *block = &sd->succ_part[block_index];
const int block_bit_index = (x - IMMEDIATE_TABLE_SIZE) % 512;
const int small_block_index = block_bit_index / 64;
const int small_block_popcount = small_block_rank_get(block->small_block_ranks, small_block_index);
const int popcnt = rb_popcount64(block->bits[small_block_index] << (63 - block_bit_index % 64));
2022-07-21 19:23:58 +03:00
return block->rank + small_block_popcount + popcnt;
}
}
#endif
/*
* call-seq:
* iseq.script_lines -> array or nil
*
* It returns recorded script lines if it is availalble.
* The script lines are not limited to the iseq range, but
* are entire lines of the source file.
*
* Note that this is an API for ruby internal use, debugging,
* and research. Do not use this for any other purpose.
* The compatibility is not guaranteed.
*/
static VALUE
iseqw_script_lines(VALUE self)
{
const rb_iseq_t *iseq = iseqw_check(self);
return ISEQ_BODY(iseq)->variable.script_lines;
}
/*
* Document-class: RubyVM::InstructionSequence
*
* The InstructionSequence class represents a compiled sequence of
* instructions for the Virtual Machine used in MRI. Not all implementations of Ruby
* may implement this class, and for the implementations that implement it,
* the methods defined and behavior of the methods can change in any version.
*
* With it, you can get a handle to the instructions that make up a method or
* a proc, compile strings of Ruby code down to VM instructions, and
* disassemble instruction sequences to strings for easy inspection. It is
* mostly useful if you want to learn how YARV works, but it also lets
* you control various settings for the Ruby iseq compiler.
*
* You can find the source for the VM instructions in +insns.def+ in the Ruby
* source.
*
* The instruction sequence results will almost certainly change as Ruby
* changes, so example output in this documentation may be different from what
* you see.
*
* Of course, this class is MRI specific.
*/
void
Init_ISeq(void)
{
/* declare ::RubyVM::InstructionSequence */
rb_cISeq = rb_define_class_under(rb_cRubyVM, "InstructionSequence", rb_cObject);
rb_undef_alloc_func(rb_cISeq);
2015-07-22 01:52:59 +03:00
rb_define_method(rb_cISeq, "inspect", iseqw_inspect, 0);
rb_define_method(rb_cISeq, "disasm", iseqw_disasm, 0);
rb_define_method(rb_cISeq, "disassemble", iseqw_disasm, 0);
rb_define_method(rb_cISeq, "to_a", iseqw_to_a, 0);
rb_define_method(rb_cISeq, "eval", iseqw_eval, 0);
rb_define_method(rb_cISeq, "to_binary", iseqw_to_binary, -1);
rb_define_singleton_method(rb_cISeq, "load_from_binary", iseqw_s_load_from_binary, 1);
rb_define_singleton_method(rb_cISeq, "load_from_binary_extra_data", iseqw_s_load_from_binary_extra_data, 1);
/* location APIs */
2015-07-22 01:52:59 +03:00
rb_define_method(rb_cISeq, "path", iseqw_path, 0);
rb_define_method(rb_cISeq, "absolute_path", iseqw_absolute_path, 0);
rb_define_method(rb_cISeq, "label", iseqw_label, 0);
rb_define_method(rb_cISeq, "base_label", iseqw_base_label, 0);
rb_define_method(rb_cISeq, "first_lineno", iseqw_first_lineno, 0);
rb_define_method(rb_cISeq, "trace_points", iseqw_trace_points, 0);
rb_define_method(rb_cISeq, "each_child", iseqw_each_child, 0);
#if 0 /* TBD */
2015-07-22 01:52:59 +03:00
rb_define_private_method(rb_cISeq, "marshal_dump", iseqw_marshal_dump, 0);
rb_define_private_method(rb_cISeq, "marshal_load", iseqw_marshal_load, 1);
/* disable this feature because there is no verifier. */
rb_define_singleton_method(rb_cISeq, "load", iseq_s_load, -1);
#endif
(void)iseq_s_load;
2015-07-22 01:52:59 +03:00
rb_define_singleton_method(rb_cISeq, "compile", iseqw_s_compile, -1);
rb_define_singleton_method(rb_cISeq, "new", iseqw_s_compile, -1);
rb_define_singleton_method(rb_cISeq, "compile_file", iseqw_s_compile_file, -1);
rb_define_singleton_method(rb_cISeq, "compile_option", iseqw_s_compile_option_get, 0);
rb_define_singleton_method(rb_cISeq, "compile_option=", iseqw_s_compile_option_set, 1);
rb_define_singleton_method(rb_cISeq, "disasm", iseqw_s_disasm, 1);
rb_define_singleton_method(rb_cISeq, "disassemble", iseqw_s_disasm, 1);
rb_define_singleton_method(rb_cISeq, "of", iseqw_s_of, 1);
// script lines
rb_define_method(rb_cISeq, "script_lines", iseqw_script_lines, 0);
rb_undef_method(CLASS_OF(rb_cISeq), "translate");
rb_undef_method(CLASS_OF(rb_cISeq), "load_iseq");
}