Add a "colocated" flag to symbol references. (#298)

This adds a "colocated" flag to function and symbolic global variables which
indicates that they are defined along with the current function, so they can
use PC-relative addressing.

This also changes the function decl syntax; the name now always precedes the
signature, and the "function" keyword is no longer included.

cranelift/filetests/isa/riscv/binary32.cton

@@ -4,7 +4,7 @@ isa riscv
 
 function %RV32I(i32 link [%x1]) -> i32 link [%x1] {
     sig0 = ()
-    fn0 = function %foo()
+    fn0 = %foo()
 
 ebb0(v9999: i32):
     [-,%x10]            v1 = iconst.i32 1

cranelift/filetests/isa/riscv/legalize-abi.cton

@@ -17,8 +17,8 @@ ebb0(v0: i64):
 }
 
 function %split_call_arg(i32) {
-    fn1 = function %foo(i64)
-    fn2 = function %foo(i32, i64)
+    fn1 = %foo(i64)
+    fn2 = %foo(i32, i64)
 ebb0(v0: i32):
     v1 = uextend.i64 v0
     call fn1(v1)
@@ -30,7 +30,7 @@ ebb0(v0: i32):
 }
 
 function %split_ret_val() {
-    fn1 = function %foo() -> i64
+    fn1 = %foo() -> i64
 ebb0:
     v1 = call fn1()
     ; check: ebb0($(link=$V): i32):
@@ -45,7 +45,7 @@ ebb1(v10: i64):
 
 ; First return value is fine, second one is expanded.
 function %split_ret_val2() {
-    fn1 = function %foo() -> i32, i64
+    fn1 = %foo() -> i32, i64
 ebb0:
     v1, v2 = call fn1()
     ; check: ebb0($(link=$V): i32):
@@ -70,7 +70,7 @@ ebb0(v1: i8, v2: i8, v3: i8):
 
 ; Function produces single return value, still need to copy.
 function %ext_ret_val() {
-    fn1 = function %foo() -> i8 sext
+    fn1 = %foo() -> i8 sext
 ebb0:
     v1 = call fn1()
     ; check: ebb0($V: i32):
@@ -124,7 +124,7 @@ ebb0(v0: i32, v1: f32x2):
 ; Call a function that takes arguments on the stack.
 function %stack_args(i32) {
     ; check: $(ss0=$SS) = outgoing_arg 4
-    fn1 = function %foo(i64, i64, i64, i64, i32)
+    fn1 = %foo(i64, i64, i64, i64, i32)
 ebb0(v0: i32):
     v1 = iconst.i64 1
     call fn1(v1, v1, v1, v1, v0)

cranelift/filetests/isa/riscv/parse-encoding.cton

@@ -27,9 +27,9 @@ function %parse_encoding(i32 [%x5]) -> i32 [%x10] {
     ; check: sig5 = () -> f32 [0] system_v
 
     ; function + signature
-    fn0 = function %bar(i32 [%x10]) -> b1 [%x10] system_v
+    fn0 = %bar(i32 [%x10]) -> b1 [%x10] system_v
     ; check: sig6 = (i32 [%x10]) -> b1 [%x10] system_v
-    ; nextln: fn0 = sig6 %bar
+    ; nextln: fn0 = %bar sig6
 
 ebb0(v0: i32):
     return v0

cranelift/filetests/isa/riscv/verify-encoding.cton

@@ -2,7 +2,7 @@ test verifier
 isa riscv
 
 function %RV32I(i32 link [%x1]) -> i32 link [%x1] {
-    fn0 = function %foo()
+    fn0 = %foo()
 
 ebb0(v9999: i32):
     ; iconst.i32 needs legalizing, so it should throw a
@@ -11,7 +11,7 @@ ebb0(v9999: i32):
 }
 
 function %RV32I(i32 link [%x1]) -> i32 link [%x1] {
-    fn0 = function %foo()
+    fn0 = %foo()
 
 ebb0(v9999: i32):
     v1 = iconst.i32 1

cranelift/filetests/isa/x86/abi64.cton

@@ -21,7 +21,7 @@ ebb0:
 
 function %pass_stack_int64(i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64 vmctx) spiderwasm {
     sig0 = (i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64, i64 vmctx) spiderwasm
-    fn0 = sig0 u0:0
+    fn0 = u0:0 sig0
 
 ebb0(v0: i64, v1: i64, v2: i64, v3: i64, v4: i64, v5: i64, v6: i64, v7: i64, v8: i64, v9: i64, v10: i64, v11: i64, v12: i64, v13: i64, v14: i64, v15: i64, v16: i64, v17: i64, v18: i64, v19: i64, v20: i64):
     call fn0(v0, v1, v2, v3, v4, v5, v6, v7, v8, v9, v10, v11, v12, v13, v14, v15, v16, v17, v18, v19, v20)

cranelift/filetests/isa/x86/allones_funcaddrs32.cton

@@ -12,7 +12,7 @@ isa x86 haswell
 ; Tests from binary32.cton affected by allones_funcaddrs.
 function %I32() {
     sig0 = ()
-    fn0 = function %foo()
+    fn0 = %foo()
 
 ebb0:
 

cranelift/filetests/isa/x86/allones_funcaddrs64.cton

@@ -13,7 +13,7 @@ isa x86 haswell
 ; Tests from binary64.cton affected by allones_funcaddrs.
 function %I64() {
     sig0 = ()
-    fn0 = function %foo()
+    fn0 = %foo()
 
 ebb0:
 

cranelift/filetests/isa/x86/binary32.cton

@@ -10,7 +10,7 @@ isa x86 haswell
 
 function %I32() {
     sig0 = ()
-    fn0 = function %foo()
+    fn0 = %foo()
 
     gv0 = globalsym %some_gv
 

cranelift/filetests/isa/x86/binary64-pic.cton

@@ -13,9 +13,11 @@ isa x86 haswell
 ; Tests for i64 instructions.
 function %I64() {
     sig0 = ()
-    fn0 = function %foo()
+    fn0 = %foo()
+    fn1 = colocated %bar()
 
     gv0 = globalsym %some_gv
+    gv1 = globalsym colocated %some_gv
 
     ; Use incoming_arg stack slots because they won't be relocated by the frame
     ; layout.
@@ -29,6 +31,9 @@ ebb0:
     ; asm: call foo@PLT
     call fn0()                                  ; bin: e8 PLTRel4(%foo-4) 00000000
 
+    ; asm: call foo
+    call fn1()                                  ; bin: e8 PCRel4(%bar-4) 00000000
+
     ; asm: mov 0x0(%rip), %rax
     [-,%rax]            v0 = func_addr.i64 fn0        ; bin: 48 8b 05 GOTPCRel4(%foo-4) 00000000
     ; asm: mov 0x0(%rip), %rsi
@@ -50,5 +55,12 @@ ebb0:
     ; asm: mov 0x0(%rip), %r10
     [-,%r10]            v5 = globalsym_addr.i64 gv0    ; bin: 4c 8b 15 GOTPCRel4(%some_gv-4) 00000000
 
+    ; asm: lea 0x0(%rip), %rcx
+    [-,%rcx]            v6 = globalsym_addr.i64 gv1    ; bin: 48 8d 0d PCRel4(%some_gv-4) 00000000
+    ; asm: lea 0x0(%rip), %rsi
+    [-,%rsi]            v7 = globalsym_addr.i64 gv1    ; bin: 48 8d 35 PCRel4(%some_gv-4) 00000000
+    ; asm: lea 0x0(%rip), %r10
+    [-,%r10]            v8 = globalsym_addr.i64 gv1    ; bin: 4c 8d 15 PCRel4(%some_gv-4) 00000000
+
     return
 }

cranelift/filetests/isa/x86/binary64.cton

@@ -12,7 +12,7 @@ isa x86 haswell
 ; Tests for i64 instructions.
 function %I64() {
     sig0 = ()
-    fn0 = function %foo()
+    fn0 = %foo()
 
     gv0 = globalsym %some_gv
 
@@ -473,12 +473,6 @@ ebb0:
     ; asm: movzbq %dl, %rsi
     [-,%rsi]             v351 = bint.i64 v301   ; bin: 0f b6 f2
 
-    ; TODO: x86-64 can't encode a direct call to an arbitrary 64-bit address in
-    ; a single instruction. When we add a concept of colocated definitions, this
-    ; test can be re-enabled.
-    ; disabled: asm: call foo
-    ; disabled: call fn0()                                  ; bin: e8 PCRel4(%foo-4) 00000000
-
     ; asm: movabsq $0, %rcx
     [-,%rcx]            v400 = func_addr.i64 fn0        ; bin: 48 b9 Abs8(%foo) 0000000000000000
     ; asm: movabsq $0, %rsi
@@ -713,7 +707,7 @@ ebb0:
 ; be done by an instruction shrinking pass.
 function %I32() {
     sig0 = ()
-    fn0 = function %foo()
+    fn0 = %foo()
 
     ss0 = incoming_arg 8, offset 0
     ss1 = incoming_arg 1024, offset -1024

cranelift/filetests/isa/x86/legalize-libcall.cton

@@ -12,5 +12,5 @@ ebb0(v0: f32):
 }
 ; check: function %floor(f32 [%xmm0]) -> f32 [%xmm0] system_v {
 ; check: sig0 = (f32) -> f32 system_v
-; check: fn0 = sig0 %FloorF32
+; check: fn0 = %FloorF32 sig0
 ; check: v1 = call fn0(v0)

cranelift/filetests/isa/x86/prologue-epilogue.cton

@@ -45,7 +45,7 @@ ebb0:
 ; A function performing a call.
 
 function %call() {
-    fn0 = function %foo()
+    fn0 = %foo()
 
 ebb0:
     call fn0()
@@ -55,7 +55,7 @@ ebb0:
 ; check: function %call(i64 fp [%rbp]) -> i64 fp [%rbp] system_v {
 ; nextln:     ss0 = incoming_arg 16, offset -16
 ; nextln:     sig0 = () system_v
-; nextln:     fn0 = sig0 %foo
+; nextln:     fn0 = %foo sig0
 ; nextln: 
 ; nextln: ebb0(v0: i64 [%rbp]):
 ; nextln:     x86_push v0

cranelift/filetests/parser/call.cton

@@ -26,22 +26,22 @@ ebb1:
 function %signatures() {
     sig10 = ()
     sig11 = (i32, f64) -> i32, b1 spiderwasm
-    fn5 = sig11 %foo
-    fn8 = function %bar(i32) -> b1
+    fn5 = %foo sig11
+    fn8 = %bar(i32) -> b1
 }
 ; sameln: function %signatures() system_v {
 ; check:      sig10 = () system_v
 ; check:      sig11 = (i32, f64) -> i32, b1 spiderwasm
 ; check:      sig12 = (i32) -> b1 system_v
 ; not:        fn0
-; check:      fn5 = sig11 %foo
-; check:      fn8 = sig12 %bar
+; check:      fn5 = %foo sig11
+; check:      fn8 = %bar sig12
 ; check:  }
 
 function %direct() {
-    fn0 = function %none()
-    fn1 = function %one() -> i32
-    fn2 = function %two() -> i32, f32
+    fn0 = %none()
+    fn1 = %one() -> i32
+    fn2 = %two() -> i32, f32
 
 ebb0:
     call fn0()
@@ -72,7 +72,7 @@ ebb0(v0: i64):
 
 function %long_call() {
     sig0 = ()
-    fn0 = sig0 %none
+    fn0 = %none sig0
 
 ebb0:
     v0 = func_addr.i32 fn0

cranelift/filetests/regalloc/basic.cton

@@ -53,7 +53,7 @@ ebb1(v10: i32):
 
 ; Pass an EBB argument as a function argument.
 function %callebb(i32, i32) -> i32 {
-    fn0 = function %foo(i32) -> i32
+    fn0 = %foo(i32) -> i32
 
 ebb0(v1: i32, v2: i32):
     brnz v1, ebb1(v1)
@@ -66,7 +66,7 @@ ebb1(v10: i32):
 
 ; Pass an EBB argument as a jump argument.
 function %jumpebb(i32, i32) -> i32 {
-    fn0 = function %foo(i32) -> i32
+    fn0 = %foo(i32) -> i32
 
 ebb0(v1: i32, v2: i32):
     brnz v1, ebb1(v1, v2)

cranelift/filetests/regalloc/coalescing-207.cton

@@ -11,9 +11,9 @@ function %pr207(i64 vmctx, i32, i32) -> i32 system_v {
     sig0 = (i64 vmctx, i32, i32) -> i32 system_v
     sig1 = (i64 vmctx, i32, i32, i32) -> i32 system_v
     sig2 = (i64 vmctx, i32, i32, i32) -> i32 system_v
-    fn0 = sig0 u0:2
-    fn1 = sig1 u0:0
-    fn2 = sig2 u0:1
+    fn0 = u0:2 sig0
+    fn1 = u0:0 sig1
+    fn2 = u0:1 sig2
 
 ebb0(v0: i64, v1: i32, v2: i32):
     v3 = iconst.i32 0
@@ -1038,7 +1038,7 @@ function %musl(f64 [%xmm0], i64 vmctx [%rdi]) -> f64 [%xmm0] system_v {
     gv0 = vmctx
     heap0 = static gv0, min 0, bound 0x0001_0000_0000, guard 0x8000_0000
     sig0 = (f64 [%xmm0], i32 [%rdi], i64 vmctx [%rsi]) -> f64 [%xmm0] system_v
-    fn0 = sig0 u0:517
+    fn0 = u0:517 sig0
 
 ebb0(v0: f64, v1: i64):
     v3 = iconst.i64 0

cranelift/filetests/regalloc/reload-208.cton

@@ -16,8 +16,8 @@ function %pr208(i64 vmctx [%rdi]) system_v {
     heap0 = static gv0, min 0, bound 0x5000, guard 0x0040_0000
     sig0 = (i64 vmctx [%rdi]) -> i32 [%rax] system_v
     sig1 = (i64 vmctx [%rdi], i32 [%rsi]) system_v
-    fn0 = sig0 u0:1
-    fn1 = sig1 u0:3
+    fn0 = u0:1 sig0
+    fn1 = u0:3 sig1
 
 ebb0(v0: i64):
     v1 = iconst.i32 0

cranelift/filetests/regalloc/reload.cton

@@ -5,7 +5,7 @@ isa riscv enable_e
 
 ; Check that we can handle a function return value that got spilled.
 function %spill_return() -> i32 {
-    fn0 = function %foo() -> i32 system_v
+    fn0 = %foo() -> i32 system_v
 
 ebb0:
     v0 = call fn0()

cranelift/filetests/regalloc/spill.cton

@@ -70,7 +70,7 @@ ebb0(v1: i32):
 
 ; All values live across a call must be spilled
 function %across_call(i32) {
-    fn0 = function %foo(i32)
+    fn0 = %foo(i32)
 ebb0(v1: i32):
     ; check: v1 = spill
     call fn0(v1)
@@ -83,7 +83,7 @@ ebb0(v1: i32):
 
 ; The same value used for two function arguments.
 function %doubleuse(i32) {
-    fn0 = function %xx(i32, i32)
+    fn0 = %xx(i32, i32)
 ebb0(v0: i32):
     ; check: $(c=$V) = copy v0
     call fn0(v0, v0)

cranelift/filetests/verifier/type_check.cton

@@ -42,14 +42,14 @@ function %type_mismatch_controlling_variable() {
 }
 
 function %fn_call_too_few_args() {
-    fn2 = function %great_fn(i32, f32)
+    fn2 = %great_fn(i32, f32)
     ebb0:
         call fn2() ; error: mismatched argument count, got 0, expected 2
         return
 }
 
 function %fn_call_too_many_args() {
-    fn5 = function %best_fn()
+    fn5 = %best_fn()
     ebb0:
         v0 = iconst.i64 56
         v1 = f32const 0.0