A lightweight WebAssembly runtime that is fast, secure, and standards-compliant
winch(aarch64): v128 min, max, and avgr ops (#13939)
Daniel Macovei599ffde0069a
diff --git a/tests/disas/winch/aarch64/v128_ops/i16x8_avgr_u.wat b/tests/disas/winch/aarch64/v128_ops/i16x8_avgr_u.watnew file mode 100644index 0000000000..f89d56ada1--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i16x8_avgr_u.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i16x8.avgr_u (v128.const i16x8 1 2 0xFFFF 0xFFFF 0 1 0x8000 0x7FFF) (v128.const i16x8 1 2 0xFFFF 0xFFFF 0 1 0x8000 0x7FFF)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; urhadd v1.8h, v1.8h, v0.8h+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x01, 0x00, 0x02, 0x00+;; 84: .byte 0xff, 0xff, 0xff, 0xff+;; 88: .byte 0x00, 0x00, 0x01, 0x00+;; 8c: .byte 0x00, 0x80, 0xff, 0x7fdiff --git a/tests/disas/winch/aarch64/v128_ops/i16x8_max_s.wat b/tests/disas/winch/aarch64/v128_ops/i16x8_max_s.watnew file mode 100644index 0000000000..cb2263acc9--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i16x8_max_s.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i16x8.max_s (v128.const i16x8 0x8000 -1 1 0 12345 -12345 32767 255) (v128.const i16x8 0x8000 -1 1 0 12345 -12345 32767 255)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; smax v1.8h, v0.8h, v1.8h+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x00, 0x80, 0xff, 0xff+;; 84: udf #1+;; 88: .byte 0x39, 0x30, 0xc7, 0xcf+;; 8c: .byte 0xff, 0x7f, 0xff, 0x00diff --git a/tests/disas/winch/aarch64/v128_ops/i16x8_max_u.wat b/tests/disas/winch/aarch64/v128_ops/i16x8_max_u.watnew file mode 100644index 0000000000..eb3f729478--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i16x8_max_u.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i16x8.max_u (v128.const i16x8 0x8000 -1 1 0 12345 -12345 32767 255) (v128.const i16x8 0x8000 -1 1 0 12345 -12345 32767 255)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; umax v1.8h, v0.8h, v1.8h+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x00, 0x80, 0xff, 0xff+;; 84: udf #1+;; 88: .byte 0x39, 0x30, 0xc7, 0xcf+;; 8c: .byte 0xff, 0x7f, 0xff, 0x00diff --git a/tests/disas/winch/aarch64/v128_ops/i16x8_min_s.wat b/tests/disas/winch/aarch64/v128_ops/i16x8_min_s.watnew file mode 100644index 0000000000..f176b53263--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i16x8_min_s.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i16x8.min_s (v128.const i16x8 0x8000 -1 1 0 12345 -12345 32767 255) (v128.const i16x8 0x8000 -1 1 0 12345 -12345 32767 255)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; smin v1.8h, v0.8h, v1.8h+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x00, 0x80, 0xff, 0xff+;; 84: udf #1+;; 88: .byte 0x39, 0x30, 0xc7, 0xcf+;; 8c: .byte 0xff, 0x7f, 0xff, 0x00diff --git a/tests/disas/winch/aarch64/v128_ops/i16x8_min_u.wat b/tests/disas/winch/aarch64/v128_ops/i16x8_min_u.watnew file mode 100644index 0000000000..04ce16e9c9--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i16x8_min_u.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i16x8.min_u (v128.const i16x8 0x8000 -1 1 0 12345 -12345 32767 255) (v128.const i16x8 0x8000 -1 1 0 12345 -12345 32767 255)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; umin v1.8h, v0.8h, v1.8h+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x00, 0x80, 0xff, 0xff+;; 84: udf #1+;; 88: .byte 0x39, 0x30, 0xc7, 0xcf+;; 8c: .byte 0xff, 0x7f, 0xff, 0x00diff --git a/tests/disas/winch/aarch64/v128_ops/i32x4_max_s.wat b/tests/disas/winch/aarch64/v128_ops/i32x4_max_s.watnew file mode 100644index 0000000000..f64807d01e--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i32x4_max_s.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i32x4.max_s (v128.const i32x4 0x80000000 -1 1 0) (v128.const i32x4 0x80000000 -1 1 0)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; smax v1.4s, v0.4s, v1.4s+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x00, 0x00, 0x00, 0x80+;; 84: .byte 0xff, 0xff, 0xff, 0xff+;; 88: udf #1+;; 8c: udf #0diff --git a/tests/disas/winch/aarch64/v128_ops/i32x4_max_u.wat b/tests/disas/winch/aarch64/v128_ops/i32x4_max_u.watnew file mode 100644index 0000000000..bc9b20b9cc--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i32x4_max_u.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i32x4.max_u (v128.const i32x4 0x80000000 -1 1 0) (v128.const i32x4 0x80000000 -1 1 0)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; umax v1.4s, v0.4s, v1.4s+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x00, 0x00, 0x00, 0x80+;; 84: .byte 0xff, 0xff, 0xff, 0xff+;; 88: udf #1+;; 8c: udf #0diff --git a/tests/disas/winch/aarch64/v128_ops/i32x4_min_s.wat b/tests/disas/winch/aarch64/v128_ops/i32x4_min_s.watnew file mode 100644index 0000000000..2a12ccf502--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i32x4_min_s.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i32x4.min_s (v128.const i32x4 0x80000000 -1 1 0) (v128.const i32x4 0x80000000 -1 1 0)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; smin v1.4s, v0.4s, v1.4s+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x00, 0x00, 0x00, 0x80+;; 84: .byte 0xff, 0xff, 0xff, 0xff+;; 88: udf #1+;; 8c: udf #0diff --git a/tests/disas/winch/aarch64/v128_ops/i32x4_min_u.wat b/tests/disas/winch/aarch64/v128_ops/i32x4_min_u.watnew file mode 100644index 0000000000..38619091f1--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i32x4_min_u.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i32x4.min_u (v128.const i32x4 0x80000000 -1 1 0) (v128.const i32x4 0x80000000 -1 1 0)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; umin v1.4s, v0.4s, v1.4s+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x00, 0x00, 0x00, 0x80+;; 84: .byte 0xff, 0xff, 0xff, 0xff+;; 88: udf #1+;; 8c: udf #0diff --git a/tests/disas/winch/aarch64/v128_ops/i8x16_avgr_u.wat b/tests/disas/winch/aarch64/v128_ops/i8x16_avgr_u.watnew file mode 100644index 0000000000..f3b3a3949d--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i8x16_avgr_u.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i8x16.avgr_u (v128.const i8x16 1 2 255 255 0 1 128 127 3 4 5 6 7 8 9 10) (v128.const i8x16 1 2 255 255 0 1 128 127 3 4 5 6 7 8 9 10)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; urhadd v1.16b, v1.16b, v0.16b+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x01, 0x02, 0xff, 0xff+;; 84: .byte 0x00, 0x01, 0x80, 0x7f+;; 88: .byte 0x03, 0x04, 0x05, 0x06+;; 8c: and w7, w0, w9, lsl #2diff --git a/tests/disas/winch/aarch64/v128_ops/i8x16_max_s.wat b/tests/disas/winch/aarch64/v128_ops/i8x16_max_s.watnew file mode 100644index 0000000000..8c62e63b85--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i8x16_max_s.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i8x16.max_s (v128.const i8x16 128 255 1 0 2 254 127 3 4 5 6 7 8 9 10 11) (v128.const i8x16 128 255 1 0 2 254 127 3 4 5 6 7 8 9 10 11)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; smax v1.16b, v0.16b, v1.16b+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x80, 0xff, 0x01, 0x00+;; 84: .byte 0x02, 0xfe, 0x7f, 0x03+;; 88: .byte 0x04, 0x05, 0x06, 0x07+;; 8c: add w8, w8, w10, lsl #2diff --git a/tests/disas/winch/aarch64/v128_ops/i8x16_max_u.wat b/tests/disas/winch/aarch64/v128_ops/i8x16_max_u.watnew file mode 100644index 0000000000..5d7e621b0c--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i8x16_max_u.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i8x16.max_u (v128.const i8x16 128 255 1 0 2 254 127 3 4 5 6 7 8 9 10 11) (v128.const i8x16 128 255 1 0 2 254 127 3 4 5 6 7 8 9 10 11)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; umax v1.16b, v0.16b, v1.16b+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x80, 0xff, 0x01, 0x00+;; 84: .byte 0x02, 0xfe, 0x7f, 0x03+;; 88: .byte 0x04, 0x05, 0x06, 0x07+;; 8c: add w8, w8, w10, lsl #2diff --git a/tests/disas/winch/aarch64/v128_ops/i8x16_min_s.wat b/tests/disas/winch/aarch64/v128_ops/i8x16_min_s.watnew file mode 100644index 0000000000..05681e77d1--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i8x16_min_s.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i8x16.min_s (v128.const i8x16 128 255 1 0 2 254 127 3 4 5 6 7 8 9 10 11) (v128.const i8x16 128 255 1 0 2 254 127 3 4 5 6 7 8 9 10 11)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; smin v1.16b, v0.16b, v1.16b+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x80, 0xff, 0x01, 0x00+;; 84: .byte 0x02, 0xfe, 0x7f, 0x03+;; 88: .byte 0x04, 0x05, 0x06, 0x07+;; 8c: add w8, w8, w10, lsl #2diff --git a/tests/disas/winch/aarch64/v128_ops/i8x16_min_u.wat b/tests/disas/winch/aarch64/v128_ops/i8x16_min_u.watnew file mode 100644index 0000000000..effb4ad599--- /dev/null+++ b/tests/disas/winch/aarch64/v128_ops/i8x16_min_u.wat@@ -0,0 +1,45 @@+;;! target = "aarch64"+;;! test = "winch"++(module+ (memory 1)+ (func (export "run")+ (v128.store (i32.const 0)+ (i8x16.min_u (v128.const i8x16 128 255 1 0 2 254 127 3 4 5 6 7 8 9 10 11) (v128.const i8x16 128 255 1 0 2 254 127 3 4 5 6 7 8 9 10 11)))))+;; wasm[0]::function[0]:+;; stp x29, x30, [sp, #-0x10]!+;; mov x29, sp+;; str x28, [sp, #-0x10]!+;; mov x28, sp+;; ldur x16, [x0, #8]+;; ldur x16, [x16, #0x18]+;; mov x17, #0+;; movk x17, #0x10+;; add x16, x16, x17+;; cmp sp, x16+;; b.lo #0x74+;; 2c: mov x9, x0+;; sub x28, x28, #0x10+;; mov sp, x28+;; stur x0, [x28, #8]+;; stur x1, [x28]+;; ldr q0, #0x80+;; ldr q1, #0x80+;; umin v1.16b, v0.16b, v1.16b+;; mov x0, #0+;; ldur x1, [x9, #0x38]+;; add x1, x1, w0, uxtw+;; stur q1, [x1]+;; add x28, x28, #0x10+;; mov sp, x28+;; mov sp, x28+;; ldr x28, [sp], #0x10+;; ldp x29, x30, [sp], #0x10+;; ret+;; 74: udf #0xc11f+;; 78: udf #0+;; 7c: udf #0+;; 80: .byte 0x80, 0xff, 0x01, 0x00+;; 84: .byte 0x02, 0xfe, 0x7f, 0x03+;; 88: .byte 0x04, 0x05, 0x06, 0x07+;; 8c: add w8, w8, w10, lsl #2diff --git a/winch/codegen/src/isa/aarch64/masm.rs b/winch/codegen/src/isa/aarch64/masm.rsindex 26eb092fcc..f66873b92a 100644--- a/winch/codegen/src/isa/aarch64/masm.rs+++ b/winch/codegen/src/isa/aarch64/masm.rs@@ -1951,22 +1951,46 @@ impl Masm for MacroAssembler {fn v128_min(&mut self,- _src1: Reg,- _src2: Reg,- _dst: WritableReg,- _kind: V128MinKind,+ src1: Reg,+ src2: Reg,+ dst: WritableReg,+ kind: V128MinKind,) -> Result<()> {- Err(format_err!(CodeGenError::unimplemented_masm_instruction()))+ let (op, size) = match kind {+ V128MinKind::F32x4 | V128MinKind::F64x2 => {+ bail!(CodeGenError::unimplemented_masm_instruction())+ }+ V128MinKind::I8x16S => (VecALUOp::Smin, VectorSize::Size8x16),+ V128MinKind::I8x16U => (VecALUOp::Umin, VectorSize::Size8x16),+ V128MinKind::I16x8S => (VecALUOp::Smin, VectorSize::Size16x8),+ V128MinKind::I16x8U => (VecALUOp::Umin, VectorSize::Size16x8),+ V128MinKind::I32x4S => (VecALUOp::Smin, VectorSize::Size32x4),+ V128MinKind::I32x4U => (VecALUOp::Umin, VectorSize::Size32x4),+ };+ self.asm.vec_rrr(op, src1, src2, dst, size);+ Ok(())}fn v128_max(&mut self,- _src1: Reg,- _src2: Reg,- _dst: WritableReg,- _kind: V128MaxKind,+ src1: Reg,+ src2: Reg,+ dst: WritableReg,+ kind: V128MaxKind,) -> Result<()> {- Err(format_err!(CodeGenError::unimplemented_masm_instruction()))+ let (op, size) = match kind {+ V128MaxKind::F32x4 | V128MaxKind::F64x2 => {+ bail!(CodeGenError::unimplemented_masm_instruction())+ }+ V128MaxKind::I8x16S => (VecALUOp::Smax, VectorSize::Size8x16),+ V128MaxKind::I8x16U => (VecALUOp::Umax, VectorSize::Size8x16),+ V128MaxKind::I16x8S => (VecALUOp::Smax, VectorSize::Size16x8),+ V128MaxKind::I16x8U => (VecALUOp::Umax, VectorSize::Size16x8),+ V128MaxKind::I32x4S => (VecALUOp::Smax, VectorSize::Size32x4),+ V128MaxKind::I32x4U => (VecALUOp::Umax, VectorSize::Size32x4),+ };+ self.asm.vec_rrr(op, src1, src2, dst, size);+ Ok(())}fn v128_extmul(@@ -2025,14 +2049,15 @@ impl Masm for MacroAssembler {Ok(())}- fn v128_avgr(- &mut self,- _lhs: Reg,- _rhs: Reg,- _dst: WritableReg,- _size: OperandSize,- ) -> Result<()> {- bail!(CodeGenError::unimplemented_masm_instruction())+ fn v128_avgr(&mut self, lhs: Reg, rhs: Reg, dst: WritableReg, size: OperandSize) -> Result<()> {+ self.asm.vec_rrr(+ VecALUOp::Urhadd,+ lhs,+ rhs,+ dst,+ VectorSize::from_lane_size(size.into(), true),+ );+ Ok(())}fn v128_div(