diff options
author | Aaron Patterson <[email protected]> | 2024-04-15 10:48:53 -0700 |
---|---|---|
committer | Aaron Patterson <[email protected]> | 2024-06-18 09:28:25 -0700 |
commit | cdf33ed5f37f9649c482c3ba1d245f0d80ac01ce (patch) | |
tree | 1f8ef823ad788ee2c57d2a438f04e2ba7b1800e7 /yjit/src | |
parent | 921f22e563d6372d9f853f87d48b11c479126da9 (diff) |
Optimized forwarding callers and callees
This patch optimizes forwarding callers and callees. It only optimizes methods that only take `...` as their parameter, and then pass `...` to other calls.
Calls it optimizes look like this:
```ruby
def bar(a) = a
def foo(...) = bar(...) # optimized
foo(123)
```
```ruby
def bar(a) = a
def foo(...) = bar(1, 2, ...) # optimized
foo(123)
```
```ruby
def bar(*a) = a
def foo(...)
list = [1, 2]
bar(*list, ...) # optimized
end
foo(123)
```
All variants of the above but using `super` are also optimized, including a bare super like this:
```ruby
def foo(...)
super
end
```
This patch eliminates intermediate allocations made when calling methods that accept `...`.
We can observe allocation elimination like this:
```ruby
def m
x = GC.stat(:total_allocated_objects)
yield
GC.stat(:total_allocated_objects) - x
end
def bar(a) = a
def foo(...) = bar(...)
def test
m { foo(123) }
end
test
p test # allocates 1 object on master, but 0 objects with this patch
```
```ruby
def bar(a, b:) = a + b
def foo(...) = bar(...)
def test
m { foo(1, b: 2) }
end
test
p test # allocates 2 objects on master, but 0 objects with this patch
```
How does it work?
-----------------
This patch works by using a dynamic stack size when passing forwarded parameters to callees.
The caller's info object (known as the "CI") contains the stack size of the
parameters, so we pass the CI object itself as a parameter to the callee.
When forwarding parameters, the forwarding ISeq uses the caller's CI to determine how much stack to copy, then copies the caller's stack before calling the callee.
The CI at the forwarded call site is adjusted using information from the caller's CI.
I think this description is kind of confusing, so let's walk through an example with code.
```ruby
def delegatee(a, b) = a + b
def delegator(...)
delegatee(...) # CI2 (FORWARDING)
end
def caller
delegator(1, 2) # CI1 (argc: 2)
end
```
Before we call the delegator method, the stack looks like this:
```
Executing Line | Code | Stack
---------------+---------------------------------------+--------
1| def delegatee(a, b) = a + b | self
2| | 1
3| def delegator(...) | 2
4| # |
5| delegatee(...) # CI2 (FORWARDING) |
6| end |
7| |
8| def caller |
-> 9| delegator(1, 2) # CI1 (argc: 2) |
10| end |
```
The ISeq for `delegator` is tagged as "forwardable", so when `caller` calls in
to `delegator`, it writes `CI1` on to the stack as a local variable for the
`delegator` method. The `delegator` method has a special local called `...`
that holds the caller's CI object.
Here is the ISeq disasm fo `delegator`:
```
== disasm: #<ISeq:delegator@-e:1 (1,0)-(1,39)>
local table (size: 1, argc: 0 [opts: 0, rest: -1, post: 0, block: -1, kw: -1@-1, kwrest: -1])
[ 1] "..."@0
0000 putself ( 1)[LiCa]
0001 getlocal_WC_0 "..."@0
0003 send <calldata!mid:delegatee, argc:0, FCALL|FORWARDING>, nil
0006 leave [Re]
```
The local called `...` will contain the caller's CI: CI1.
Here is the stack when we enter `delegator`:
```
Executing Line | Code | Stack
---------------+---------------------------------------+--------
1| def delegatee(a, b) = a + b | self
2| | 1
3| def delegator(...) | 2
-> 4| # | CI1 (argc: 2)
5| delegatee(...) # CI2 (FORWARDING) | cref_or_me
6| end | specval
7| | type
8| def caller |
9| delegator(1, 2) # CI1 (argc: 2) |
10| end |
```
The CI at `delegatee` on line 5 is tagged as "FORWARDING", so it knows to
memcopy the caller's stack before calling `delegatee`. In this case, it will
memcopy self, 1, and 2 to the stack before calling `delegatee`. It knows how much
memory to copy from the caller because `CI1` contains stack size information
(argc: 2).
Before executing the `send` instruction, we push `...` on the stack. The
`send` instruction pops `...`, and because it is tagged with `FORWARDING`, it
knows to memcopy (using the information in the CI it just popped):
```
== disasm: #<ISeq:delegator@-e:1 (1,0)-(1,39)>
local table (size: 1, argc: 0 [opts: 0, rest: -1, post: 0, block: -1, kw: -1@-1, kwrest: -1])
[ 1] "..."@0
0000 putself ( 1)[LiCa]
0001 getlocal_WC_0 "..."@0
0003 send <calldata!mid:delegatee, argc:0, FCALL|FORWARDING>, nil
0006 leave [Re]
```
Instruction 001 puts the caller's CI on the stack. `send` is tagged with
FORWARDING, so it reads the CI and _copies_ the callers stack to this stack:
```
Executing Line | Code | Stack
---------------+---------------------------------------+--------
1| def delegatee(a, b) = a + b | self
2| | 1
3| def delegator(...) | 2
4| # | CI1 (argc: 2)
-> 5| delegatee(...) # CI2 (FORWARDING) | cref_or_me
6| end | specval
7| | type
8| def caller | self
9| delegator(1, 2) # CI1 (argc: 2) | 1
10| end | 2
```
The "FORWARDING" call site combines information from CI1 with CI2 in order
to support passing other values in addition to the `...` value, as well as
perfectly forward splat args, kwargs, etc.
Since we're able to copy the stack from `caller` in to `delegator`'s stack, we
can avoid allocating objects.
I want to do this to eliminate object allocations for delegate methods.
My long term goal is to implement `Class#new` in Ruby and it uses `...`.
I was able to implement `Class#new` in Ruby
[here](https://github.com/ruby/ruby/pull/9289).
If we adopt the technique in this patch, then we can optimize allocating
objects that take keyword parameters for `initialize`.
For example, this code will allocate 2 objects: one for `SomeObject`, and one
for the kwargs:
```ruby
SomeObject.new(foo: 1)
```
If we combine this technique, plus implement `Class#new` in Ruby, then we can
reduce allocations for this common operation.
Co-Authored-By: John Hawthorn <[email protected]>
Co-Authored-By: Alan Wu <[email protected]>
Diffstat (limited to 'yjit/src')
-rw-r--r-- | yjit/src/codegen.rs | 75 | ||||
-rw-r--r-- | yjit/src/cruby.rs | 1 | ||||
-rw-r--r-- | yjit/src/cruby_bindings.inc.rs | 4 | ||||
-rw-r--r-- | yjit/src/stats.rs | 3 |
4 files changed, 61 insertions, 22 deletions
diff --git a/yjit/src/codegen.rs b/yjit/src/codegen.rs index 755e64c244..26234aec41 100644 --- a/yjit/src/codegen.rs +++ b/yjit/src/codegen.rs @@ -7118,6 +7118,8 @@ fn gen_send_iseq( let kw_splat = flags & VM_CALL_KW_SPLAT != 0; let splat_call = flags & VM_CALL_ARGS_SPLAT != 0; + let forwarding_call = unsafe { rb_get_iseq_flags_forwardable(iseq) }; + // For computing offsets to callee locals let num_params = unsafe { get_iseq_body_param_size(iseq) as i32 }; let num_locals = unsafe { get_iseq_body_local_table_size(iseq) as i32 }; @@ -7160,9 +7162,15 @@ fn gen_send_iseq( exit_if_supplying_kw_and_has_no_kw(asm, supplying_kws, doing_kw_call)?; exit_if_supplying_kws_and_accept_no_kwargs(asm, supplying_kws, iseq)?; exit_if_doing_kw_and_splat(asm, doing_kw_call, flags)?; - exit_if_wrong_number_arguments(asm, arg_setup_block, opts_filled, flags, opt_num, iseq_has_rest)?; + if !forwarding_call { + exit_if_wrong_number_arguments(asm, arg_setup_block, opts_filled, flags, opt_num, iseq_has_rest)?; + } exit_if_doing_kw_and_opts_missing(asm, doing_kw_call, opts_missing)?; exit_if_has_rest_and_optional_and_block(asm, iseq_has_rest, opt_num, iseq, block_arg)?; + if forwarding_call && flags & VM_CALL_OPT_SEND != 0 { + gen_counter_incr(asm, Counter::send_iseq_send_forwarding); + return None; + } let block_arg_type = exit_if_unsupported_block_arg_type(jit, asm, block_arg)?; // Bail if we can't drop extra arguments for a yield by just popping them @@ -7671,25 +7679,26 @@ fn gen_send_iseq( } } - // Nil-initialize missing optional parameters - nil_fill( - "nil-initialize missing optionals", - { - let begin = -argc + required_num + opts_filled; - let end = -argc + required_num + opt_num; + if !forwarding_call { + // Nil-initialize missing optional parameters + nil_fill( + "nil-initialize missing optionals", + { + let begin = -argc + required_num + opts_filled; + let end = -argc + required_num + opt_num; - begin..end - }, - asm - ); - // Nil-initialize the block parameter. It's the last parameter local - if iseq_has_block_param { - let block_param = asm.ctx.sp_opnd(-argc + num_params - 1); - asm.store(block_param, Qnil.into()); - } - // Nil-initialize non-parameter locals - nil_fill( - "nil-initialize locals", + begin..end + }, + asm + ); + // Nil-initialize the block parameter. It's the last parameter local + if iseq_has_block_param { + let block_param = asm.ctx.sp_opnd(-argc + num_params - 1); + asm.store(block_param, Qnil.into()); + } + // Nil-initialize non-parameter locals + nil_fill( + "nil-initialize locals", { let begin = -argc + num_params; let end = -argc + num_locals; @@ -7697,7 +7706,13 @@ fn gen_send_iseq( begin..end }, asm - ); + ); + } + + if forwarding_call { + assert_eq!(1, num_params); + asm.mov(asm.stack_opnd(-1), VALUE(ci as usize).into()); + } // Points to the receiver operand on the stack unless a captured environment is used let recv = match captured_opnd { @@ -7716,7 +7731,13 @@ fn gen_send_iseq( jit_save_pc(jit, asm); // Adjust the callee's stack pointer - let callee_sp = asm.lea(asm.ctx.sp_opnd(-argc + num_locals + VM_ENV_DATA_SIZE as i32)); + let callee_sp = if forwarding_call { + let offs = num_locals + VM_ENV_DATA_SIZE as i32; + asm.lea(asm.ctx.sp_opnd(offs)) + } else { + let offs = -argc + num_locals + VM_ENV_DATA_SIZE as i32; + asm.lea(asm.ctx.sp_opnd(offs)) + }; let specval = if let Some(prev_ep) = prev_ep { // We've already side-exited if the callee expects a block, so we @@ -8519,6 +8540,14 @@ fn gen_send_general( return Some(EndBlock); } + let ci_flags = unsafe { vm_ci_flag(ci) }; + + // Dynamic stack layout. No good way to support without inlining. + if ci_flags & VM_CALL_FORWARDING != 0 { + gen_counter_incr(asm, Counter::send_iseq_forwarding); + return None; + } + let recv_idx = argc + if flags & VM_CALL_ARGS_BLOCKARG != 0 { 1 } else { 0 }; let comptime_recv = jit.peek_at_stack(&asm.ctx, recv_idx as isize); let comptime_recv_klass = comptime_recv.class_of(); @@ -9286,6 +9315,10 @@ fn gen_invokesuper_specialized( gen_counter_incr(asm, Counter::invokesuper_kw_splat); return None; } + if ci_flags & VM_CALL_FORWARDING != 0 { + gen_counter_incr(asm, Counter::invokesuper_forwarding); + return None; + } // Ensure we haven't rebound this method onto an incompatible class. // In the interpreter we try to avoid making this check by performing some diff --git a/yjit/src/cruby.rs b/yjit/src/cruby.rs index 53586cb4f4..b069137664 100644 --- a/yjit/src/cruby.rs +++ b/yjit/src/cruby.rs @@ -714,6 +714,7 @@ mod manual_defs { pub const VM_CALL_ARGS_SIMPLE: u32 = 1 << VM_CALL_ARGS_SIMPLE_bit; pub const VM_CALL_ARGS_SPLAT: u32 = 1 << VM_CALL_ARGS_SPLAT_bit; pub const VM_CALL_ARGS_BLOCKARG: u32 = 1 << VM_CALL_ARGS_BLOCKARG_bit; + pub const VM_CALL_FORWARDING: u32 = 1 << VM_CALL_FORWARDING_bit; pub const VM_CALL_FCALL: u32 = 1 << VM_CALL_FCALL_bit; pub const VM_CALL_KWARG: u32 = 1 << VM_CALL_KWARG_bit; pub const VM_CALL_KW_SPLAT: u32 = 1 << VM_CALL_KW_SPLAT_bit; diff --git a/yjit/src/cruby_bindings.inc.rs b/yjit/src/cruby_bindings.inc.rs index 1f128f5e7e..994fef28b7 100644 --- a/yjit/src/cruby_bindings.inc.rs +++ b/yjit/src/cruby_bindings.inc.rs @@ -675,7 +675,8 @@ pub const VM_CALL_ZSUPER_bit: vm_call_flag_bits = 9; pub const VM_CALL_OPT_SEND_bit: vm_call_flag_bits = 10; pub const VM_CALL_KW_SPLAT_MUT_bit: vm_call_flag_bits = 11; pub const VM_CALL_ARGS_SPLAT_MUT_bit: vm_call_flag_bits = 12; -pub const VM_CALL__END: vm_call_flag_bits = 13; +pub const VM_CALL_FORWARDING_bit: vm_call_flag_bits = 13; +pub const VM_CALL__END: vm_call_flag_bits = 14; pub type vm_call_flag_bits = u32; #[repr(C)] pub struct rb_callinfo_kwarg { @@ -1184,6 +1185,7 @@ extern "C" { pub fn rb_get_iseq_flags_has_block(iseq: *const rb_iseq_t) -> bool; pub fn rb_get_iseq_flags_ambiguous_param0(iseq: *const rb_iseq_t) -> bool; pub fn rb_get_iseq_flags_accepts_no_kwarg(iseq: *const rb_iseq_t) -> bool; + pub fn rb_get_iseq_flags_forwardable(iseq: *const rb_iseq_t) -> bool; pub fn rb_get_iseq_body_param_keyword( iseq: *const rb_iseq_t, ) -> *const rb_seq_param_keyword_struct; diff --git a/yjit/src/stats.rs b/yjit/src/stats.rs index 244ccfd55f..8df5bd7ee3 100644 --- a/yjit/src/stats.rs +++ b/yjit/src/stats.rs @@ -378,6 +378,7 @@ make_counters! { send_iseq_block_arg_type, send_iseq_clobbering_block_arg, send_iseq_complex_discard_extras, + send_iseq_forwarding, send_iseq_leaf_builtin_block_arg_block_param, send_iseq_kw_splat_non_nil, send_iseq_kwargs_mismatch, @@ -385,6 +386,7 @@ make_counters! { send_iseq_has_no_kw, send_iseq_accepts_no_kwarg, send_iseq_materialized_block, + send_iseq_send_forwarding, send_iseq_splat_not_array, send_iseq_splat_with_kw, send_iseq_missing_optional_kw, @@ -414,6 +416,7 @@ make_counters! { send_optimized_block_arg, invokesuper_defined_class_mismatch, + invokesuper_forwarding, invokesuper_kw_splat, invokesuper_kwarg, invokesuper_megamorphic, |