Merge pull request #28 from pjbgf/optimisations

goasm: Implement amd64 optimised code

.github/workflows/build.yml

@@ -29,5 +29,7 @@ jobs:
         uses: actions/setup-go@6edd4406fa81c3da01a34fa6f6343087c207a568 # v3.5.0
         with:
           go-version: ${{ matrix.go-version }}
+      - name: Run Verify
+        run: make verify
       - name: Run Cross Build
         run: make cross-build

Makefile

@@ -30,3 +30,11 @@ build-nocgo:
 
 # Run cross-compilation to assure supported architectures.
 cross-build: build-arm build-arm64 build-nocgo
+
+generate:
+	go run sha1cdblock_amd64_asm.go -out sha1cdblock_amd64.s
+	sed -i 's;&\samd64;&\n// +build !noasm,gc,amd64;g' sha1cdblock_amd64.s
+
+verify: generate
+	git diff --exit-code
+	go vet ./...

go.mod

@@ -1,3 +1,10 @@
 module github.com/pjbgf/sha1cd
 
-go 1.15
+go 1.19
+
+require (
+	github.com/mmcloughlin/avo v0.5.0 // indirect
+	golang.org/x/mod v0.6.0 // indirect
+	golang.org/x/sys v0.1.0 // indirect
+	golang.org/x/tools v0.2.0 // indirect
+)

go.sum

@@ -0,0 +1,37 @@
+github.com/mmcloughlin/avo v0.5.0 h1:nAco9/aI9Lg2kiuROBY6BhCI/z0t5jEvJfjWbL8qXLU=
+github.com/mmcloughlin/avo v0.5.0/go.mod h1:ChHFdoV7ql95Wi7vuq2YT1bwCJqiWdZrQ1im3VujLYM=
+github.com/yuin/goldmark v1.4.13/go.mod h1:6yULJ656Px+3vBD8DxQVa3kxgyrAnzto9xy5taEt/CY=
+golang.org/x/arch v0.1.0/go.mod h1:5om86z9Hs0C8fWVUuoMHwpExlXzs5Tkyp9hOrfG7pp8=
+golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
+golang.org/x/crypto v0.0.0-20210921155107-089bfa567519/go.mod h1:GvvjBRRGRdwPK5ydBHafDWAxML/pGHZbMvKqRZ5+Abc=
+golang.org/x/crypto v0.1.0/go.mod h1:RecgLatLF4+eUMCP1PoPZQb+cVrJcOPbHkTkbkB9sbw=
+golang.org/x/mod v0.6.0-dev.0.20220419223038-86c51ed26bb4/go.mod h1:jJ57K6gSWd91VN4djpZkiMVwK6gcyfeH4XE8wZrZaV4=
+golang.org/x/mod v0.6.0 h1:b9gGHsz9/HhJ3HF5DHQytPpuwocVTChQJK3AvoLRD5I=
+golang.org/x/mod v0.6.0/go.mod h1:4mET923SAdbXp2ki8ey+zGs1SLqsuM2Y0uvdZR/fUNI=
+golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
+golang.org/x/net v0.0.0-20220722155237-a158d28d115b/go.mod h1:XRhObCWvk6IyKnWLug+ECip1KBveYUHfp+8e9klMJ9c=
+golang.org/x/net v0.1.0/go.mod h1:Cx3nUiGt4eDBEyega/BKRp+/AlGL8hYe7U9odMt2Cco=
+golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20220722155255-886fb9371eb4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210615035016-665e8c7367d1/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20220520151302-bc2c85ada10a/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20220722155257-8c9f86f7a55f/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.1.0 h1:kunALQeHf1/185U1i0GOB/fy1IPRDDpuoOOqRReG57U=
+golang.org/x/sys v0.1.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
+golang.org/x/term v0.0.0-20210927222741-03fcf44c2211/go.mod h1:jbD1KX2456YbFQfuXm/mYQcufACuNUgVhRMnK/tPxf8=
+golang.org/x/term v0.1.0/go.mod h1:jbD1KX2456YbFQfuXm/mYQcufACuNUgVhRMnK/tPxf8=
+golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
+golang.org/x/text v0.3.7/go.mod h1:u+2+/6zg+i71rQMx5EYifcz6MCKuco9NR6JIITiCfzQ=
+golang.org/x/text v0.4.0/go.mod h1:mrYo+phRRbMaCq/xk9113O4dZlRixOauAjOtrjsXDZ8=
+golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.1.12/go.mod h1:hNGJHUnrk76NpqgfD5Aqm5Crs+Hm0VOH/i9J2+nxYbc=
+golang.org/x/tools v0.2.0 h1:G6AHpWxTMGY1KyEYoAQ5WTtIekUUvDNjan3ugu60JvE=
+golang.org/x/tools v0.2.0/go.mod h1:y4OqIKeOV/fWJetJ8bXPU1sEVniLMIyDAZWeHdV+NTA=
+golang.org/x/xerrors v0.0.0-20190717185122-a985d3407aa7/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+rsc.io/pdf v0.1.1/go.mod h1:n8OzWcQ6Sp37PL01nO98y4iUCRdTGarVfzxY20ICaU4=

internal/const.go

@@ -33,6 +33,10 @@ const (
 	// SHA1 processes the input data in chunks. Each chunk contains 64 bytes.
 	Chunk = 64
 
+	// The number of pre-step compression state to store.
+	// Currently there are 3 pre-step compression states required: 0, 58, 65.
+	PreStepState = 3
+
 	Magic         = "shacd\x01"
 	MarshaledSize = len(Magic) + 5*4 + Chunk + 8
 )

sha1cd.go

@@ -38,7 +38,8 @@ type digest struct {
 	len uint64
 
 	// col defines whether a collision has been found.
-	col bool
+	col       bool
+	blockFunc func(dig *digest, p []byte)
 }
 
 func (d *digest) MarshalBinary() ([]byte, error) {
@@ -127,6 +128,17 @@ func (d *digest) Reset() {
 func New() hash.Hash {
 	d := new(digest)
 
+	d.blockFunc = block
+	d.Reset()
+	return d
+}
+
+// NewGeneric is equivalent to New but uses the Go generic implementation,
+// avoiding any processor-specific optimizations.
+func NewGeneric() hash.Hash {
+	d := new(digest)
+
+	d.blockFunc = blockGeneric
 	d.Reset()
 	return d
 }
@@ -146,14 +158,14 @@ func (d *digest) Write(p []byte) (nn int, err error) {
 		n := copy(d.x[d.nx:], p)
 		d.nx += n
 		if d.nx == shared.Chunk {
-			block(d, d.x[:])
+			d.blockFunc(d, d.x[:])
 			d.nx = 0
 		}
 		p = p[n:]
 	}
 	if len(p) >= shared.Chunk {
 		n := len(p) &^ (shared.Chunk - 1)
-		block(d, p[:n])
+		d.blockFunc(d, p[:n])
 		p = p[n:]
 	}
 	if len(p) > 0 {

sha1cd_test.go

@@ -193,16 +193,25 @@ func TestLargeHashes(t *testing.T) {
 }
 
 func TestAllocations(t *testing.T) {
+	t.Run("generic", func(t *testing.T) {
+		testAllocations(NewGeneric(), t)
+	})
+
+	t.Run("native", func(t *testing.T) {
+		testAllocations(New(), t)
+	})
+}
+
+func testAllocations(h hash.Hash, t *testing.T) {
 	in := []byte("hello, world!")
 	out := make([]byte, 0, Size)
-	h := New()
 	n := int(testing.AllocsPerRun(10, func() {
 		h.Reset()
 		h.Write(in)
 		out = h.Sum(out[:0])
 	}))
 
-	if n > 0 {
-		t.Errorf("allocs = %d, want < 1", n)
+	if n > 2 {
+		t.Errorf("allocs = %d, want < 3", n)
 	}
 }

sha1cdblock_amd64.go

@@ -0,0 +1,50 @@
+//go:build !noasm && gc && amd64
+// +build !noasm,gc,amd64
+
+package sha1cd
+
+import (
+	"math"
+	"unsafe"
+
+	shared "github.com/pjbgf/sha1cd/internal"
+)
+
+type sliceHeader struct {
+	base uintptr
+	len  int
+	cap  int
+}
+
+// blockAMD64 hashes the message p into the current state in dig.
+// Both m1 and cs are used to store intermediate results which are used by the collision detection logic.
+//
+//go:noescape
+func blockAMD64(dig *digest, p sliceHeader, m1 []uint32, cs [][5]uint32)
+
+func block(dig *digest, p []byte) {
+	m1 := [shared.Rounds]uint32{}
+	cs := [shared.PreStepState][shared.WordBuffers]uint32{}
+
+	for len(p) >= shared.Chunk {
+		// Only send a block to be processed, as the collission detection
+		// works on a block by block basis.
+		ips := sliceHeader{
+			base: uintptr(unsafe.Pointer(&p[0])),
+			len:  int(math.Min(float64(len(p)), float64(shared.Chunk))),
+			cap:  shared.Chunk,
+		}
+
+		blockAMD64(dig, ips, m1[:], cs[:])
+
+		col := checkCollision(m1, cs, dig.h)
+		if col {
+			dig.col = true
+
+			blockAMD64(dig, ips, m1[:], cs[:])
+			blockAMD64(dig, ips, m1[:], cs[:])
+		}
+
+		p = p[shared.Chunk:]
+	}
+}