Compare commits

..
4 Commits
Author SHA1 Message Date
petrbalvin d9f6167a4d feat(verify): add basic-block enumeration and path-diversity profiling
Assisted-by: Qwen 3.8 Max Preview
2026-08-02 23:11:30 +02:00
petrbalvin f5088c52fc feat(verify): add runtime ABI checks with sentinel registers and red-zone canary
Assisted-by: Qwen 3.8 Max Preview
2026-08-02 23:11:30 +02:00
petrbalvin f52e23f1bc feat(verify): add differential fuzz testing against a portable Go reference
Assisted-by: Qwen 3.8 Max Preview
2026-08-02 23:11:30 +02:00
petrbalvin c9775c2b95 feat(verify): add the JIT execution substrate and gasm verify subcommand
Assisted-by: Qwen 3.8 Max Preview
2026-08-02 23:11:30 +02:00
19 changed files with 1730 additions and 2 deletions
+59 -1
View File
@@ -24,11 +24,12 @@ import (
"sourcedock.dev/petrbalvin/gasm-devkit/lint"
"sourcedock.dev/petrbalvin/gasm-devkit/lsp"
"sourcedock.dev/petrbalvin/gasm-devkit/parser"
"sourcedock.dev/petrbalvin/gasm-devkit/verify"
)
// version is the release version, stamped at build time via
// -ldflags "-X main.version=…" (defaulting to the current release).
var version = "0.16.0"
var version = "0.20.0"
func main() {
if len(os.Args) < 2 {
@@ -46,6 +47,8 @@ func main() {
os.Exit(cmdLint(os.Args[2:]))
case "asm":
os.Exit(cmdAsm(os.Args[2:]))
case "verify":
os.Exit(cmdVerify(os.Args[2:]))
case "lsp":
os.Exit(cmdLSP(os.Args[2:]))
case "version", "--version", "-V":
@@ -80,6 +83,7 @@ Commands:
fmt canonicalise formatting (gofmt for assembly)
lint run static checks
asm assemble .s files to machine code (amd64)
verify JIT-assemble and run dynamic checks (amd64)
lsp run the language server over stdio
version print the version (same as --version)
@@ -466,3 +470,57 @@ requires -p, the package path, and the installed Go toolchain).
}
return 0
}
func cmdVerify(args []string) int {
fs := newCommand("verify", "gasm verify <file.s>", `
Assemble FILE (amd64), map it into executable memory and report the available
functions. This confirms the assembled image is self-consistent (no
unresolved external symbols) and executable — the prerequisite for dynamic
testing.
With -smoke, each NOSPLIT function is called with a zeroed argument block to
confirm the JIT trampoline works end-to-end. This is safe only for functions
that tolerate nil pointers and zero lengths in their arguments.
`)
smoke := fs.Bool("smoke", false, "call each NOSPLIT function with zeroed args")
fs.Parse(args)
if fs.NArg() != 1 {
fmt.Fprintln(os.Stderr, "usage: gasm verify [-smoke] <file.s>")
return 2
}
path := fs.Arg(0)
if arch.FromFilename(path) != arch.AMD64 {
fmt.Fprintln(os.Stderr, "gasm verify: only amd64 is supported")
return 1
}
k, err := verify.Load(path)
if err != nil {
fmt.Fprintf(os.Stderr, "gasm verify: %v\n", err)
return 1
}
defer k.Close()
names := k.FuncNames()
fmt.Printf("%s: %d functions JIT-loaded\n", path, len(names))
rc := 0
for _, name := range names {
fl, _ := k.Func(name)
flags := ""
if fl.NoSplit {
flags = " NOSPLIT"
}
fmt.Printf(" %s: %d bytes, args=%d, frame=%d%s\n", name, fl.Size, fl.Args, fl.Frame, flags)
if *smoke && fl.NoSplit {
args := make([]byte, fl.Args)
_, err := k.CallFunc(name, args)
if err != nil {
fmt.Printf(" smoke: FAIL — %v\n", err)
rc = 1
} else {
fmt.Printf(" smoke: OK\n")
}
}
}
return rc
}
+25
View File
@@ -284,6 +284,31 @@ references and the implicit funcdata/DWARF symbols remain future work (the
linker fills the latter's defaults); the rest of Phase 2 is those, the
remaining EVEX forms and the other architectures.
### `verify`
The dynamic-analysis substrate (Phase 3). It JIT-loads assembled images into
executable memory and invokes them directly, enabling differential testing,
runtime ABI checks and coverage profiling.
The execution model is pure Go (stdlib only). `Map` copies machine code into
an anonymous `syscall.Mmap` mapping and enforces W^X (write the bytes, then
`mprotect` to read-execute). `Call` prepares a stack whose first word is the
address of an assembly trampoline (`leaveJIT`), lays the ABI0 argument
block after it, switches to that stack via `enterJIT` (which saves the Go
stack pointer in a package global and jumps to the target), and recovers
control when the function RETs into `leaveJIT` (which restores the Go stack
and returns). A 64-byte pad below the return address accommodates the
ABIInternal wrapper that the Go runtime interposes on assembly functions.
`Load` / `LoadSource` / `LoadAST` parse, assemble and map a `.s` file in one
step, returning a `Kernel` whose `CallFunc` method marshals the argument block
by name. The image must be self-contained (no external relocations); the
assembler’s `Image.Bytes()` provides the code-and-data concatenation.
The `gasm verify` CLI subcommand exposes this: it loads a file, reports the
available functions and (with `-smoke`) calls each NOSPLIT function with zeroed
arguments to confirm the trampoline round-trips.
## Extension points
- **New architecture:** add an entry to the generator in `_gen`, run
+1 -1
View File
@@ -3,7 +3,7 @@
# gasm-devkit — developer tooling for Go's Plan 9 assembler (GAsm).
version := "0.16.0"
version := "0.20.0"
default:
@just --list
+28
View File
@@ -0,0 +1,28 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
#include "textflag.h"
// func cleanAdd(a, b int64) int64
// A well-behaved function that preserves all callee-saved registers.
TEXT ·cleanAdd(SB), NOSPLIT, $0-24
MOVQ a+0(FP), AX
ADDQ b+8(FP), AX
MOVQ AX, ret+16(FP)
RET
// func dirtyBP(a int64) int64
// Deliberately clobbers BP (an ABI violation for a NOSPLIT frame=0 function).
TEXT ·dirtyBP(SB), NOSPLIT, $0-16
MOVQ $0x1234, BP
MOVQ a+0(FP), AX
MOVQ AX, ret+8(FP)
RET
// func dirtyR14(a int64) int64
// Deliberately clobbers R14 (the goroutine pointer — a serious ABI violation).
TEXT ·dirtyR14(SB), NOSPLIT, $0-16
MOVQ $0x5678, R14
MOVQ a+0(FP), AX
MOVQ AX, ret+8(FP)
RET
+67
View File
@@ -0,0 +1,67 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
#include "textflag.h"
// func add(a, b int64) int64
TEXT ·add(SB), NOSPLIT, $0-24
MOVQ a+0(FP), AX
ADDQ b+8(FP), AX
MOVQ AX, ret+16(FP)
RET
// func sum(data []int64) int64
// Sums all elements of the slice.
TEXT ·sum(SB), NOSPLIT, $0-32
MOVQ data_base+0(FP), SI
MOVQ data_len+8(FP), CX
XORQ AX, AX
TESTQ CX, CX
JZ sum_done
sum_loop:
ADDQ (SI), AX
ADDQ $8, SI
DECQ CX
JNZ sum_loop
sum_done:
MOVQ AX, ret+24(FP)
RET
// func wideCopy(dst, src []byte)
// Non-overlapping copy of min(len(dst), len(src)) bytes using 32-byte moves.
TEXT ·wideCopy(SB), NOSPLIT, $0-48
MOVQ dst_base+0(FP), DI
MOVQ dst_len+8(FP), BX
MOVQ src_base+24(FP), SI
MOVQ src_len+32(FP), R8
CMPQ BX, R8
JLE wc_have_n
MOVQ R8, BX
wc_have_n:
CMPQ BX, $32
JB wc_small
VMOVDQU (SI), Y0
VMOVDQU Y0, (DI)
VMOVDQU -32(SI)(BX*1), Y0
VMOVDQU Y0, -32(DI)(BX*1)
VZEROUPPER
RET
wc_small:
TESTQ BX, BX
JZ wc_done
wc_byte:
MOVB (SI), R8B
MOVB R8B, (DI)
INCQ SI
INCQ DI
DECQ BX
JNZ wc_byte
wc_done:
RET
+130
View File
@@ -0,0 +1,130 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
//go:build amd64
package verify
import (
"encoding/binary"
"fmt"
"syscall"
"unsafe"
)
// abiResult records register-clobber violations detected by the ABI-checking
// trampoline. Bit 0: BP clobbered. Bit 1: R14 clobbered.
var abiResult uint64
// savedBP holds the caller's frame pointer across the ABI-checked JIT call.
// Referenced by enterJITChecked to satisfy go vet's save-before-clobber rule.
var savedBP uintptr
// leaveCheckedPtr is initialised by the linker from the GLOBL/DATA in
// abi_amd64.s: it holds the raw address of leaveJITCheckedRaw (which has
// no ABIInternal wrapper, so the JIT function RETs directly into it).
var leaveCheckedPtr uintptr
// enterJITChecked sets sentinels in BP and R14, switches to the prepared
// stack and jumps to fn.
//
//go:nosplit
func enterJITChecked(fn uintptr, stack uintptr)
// leaveJITCheckedRaw is the raw return trampoline for ABI checks. Its
// address is obtained from the GLOBL in abi_amd64.s (leaveCheckedPtr),
// which points to the .abi0 code — NOT the ABIInternal wrapper that this
// declaration would generate. The declaration exists solely to satisfy
// go vet's "missing Go declaration" check.
//
//go:nosplit
func leaveJITCheckedRaw()
// ABIReport describes the result of an ABI-checking call.
type ABIReport struct {
BPClobbered bool // BP was modified by the function
R14Clobbered bool // R14 (goroutine pointer) was modified
RedZoneHit bool // the 128-byte red zone below SP was written
}
// OK returns true when no violations were detected.
func (r ABIReport) OK() bool {
return !r.BPClobbered && !r.R14Clobbered && !r.RedZoneHit
}
// String returns a human-readable summary.
func (r ABIReport) String() string {
if r.OK() {
return "ABI clean"
}
s := "ABI violation:"
if r.BPClobbered {
s += " BP clobbered"
}
if r.R14Clobbered {
s += " R14 clobbered"
}
if r.RedZoneHit {
s += " red-zone written"
}
return s
}
// redZoneSize is the System V AMD64 red zone: 128 bytes below SP that a
// leaf function may use without adjusting SP. Go does not use the red zone,
// so any write there is a bug.
const redZoneSize = 128
// redZoneFill is the byte pattern used to detect red-zone writes.
const redZoneFill = 0xA5
// CallChecked invokes the function with ABI sentinels and a red-zone
// canary, returning both the argument block (with results) and an ABIReport.
func CallChecked(fnAddr uintptr, args []byte) ([]byte, ABIReport, error) {
report := ABIReport{}
// Reset the global result.
abiResult = 0
// Prepare the stack: [red-zone canary][padding][leaveJITCheckedRaw][args...]
// The red zone sits below the initial SP, so the function would have to
// write below SP to corrupt it.
totalSize := redZoneSize + stackPad + 8 + len(args) + 64
stackMem, err := syscall.Mmap(-1, 0, totalSize,
syscall.PROT_READ|syscall.PROT_WRITE, syscall.MAP_PRIVATE|syscall.MAP_ANON)
if err != nil {
return nil, report, fmt.Errorf("verify: stack mmap: %w", err)
}
defer syscall.Munmap(stackMem)
// Fill the red zone with the canary pattern.
for i := 0; i < redZoneSize; i++ {
stackMem[i] = redZoneFill
}
// Return address and args after the red zone and padding.
retOff := redZoneSize + stackPad
binary.LittleEndian.PutUint64(stackMem[retOff:retOff+8], uint64(leaveCheckedPtr))
copy(stackMem[retOff+8:], args)
stackBase := uintptr(unsafe.Pointer(&stackMem[retOff]))
enterJITChecked(fnAddr, stackBase)
// Read the register-clobber result.
res := abiResult
report.BPClobbered = res&1 != 0
report.R14Clobbered = res&2 != 0
// Check the red zone.
for i := 0; i < redZoneSize; i++ {
if stackMem[i] != redZoneFill {
report.RedZoneHit = true
break
}
}
// Copy out the argument area.
out := make([]byte, len(args))
copy(out, stackMem[retOff+8:retOff+8+len(args)])
return out, report, nil
}
+61
View File
@@ -0,0 +1,61 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
#include "textflag.h"
// ABI-checking trampoline. Sets sentinel values in the callee-saved
// registers (BP, R14) before entering the JIT function and checks whether
// they survived on return.
//
// The return trampoline (leaveJITCheckedRaw) is a raw TEXT symbol with no
// Go function declaration, so the toolchain does NOT interpose an
// ABIInternal wrapper — the JIT function RETs directly into the check code,
// which sees the registers exactly as the function left them.
//
// Go ABI0 on amd64 guarantees:
// - BP is callee-saved (NOSPLIT frame=0 functions must not touch it).
// - R14 holds the goroutine pointer and must survive across any call.
// Sentinel values chosen to be unlikely in normal execution.
#define SENTINEL_BP 0xDEADBEEFCAFEF00D
#define SENTINEL_R14 0x0BADF00DDEADBEEF
// GLOBL holding the raw address of the leave trampoline, read by Go.
GLOBL ·leaveCheckedPtr(SB), NOPTR, $8
DATA ·leaveCheckedPtr(SB)/8, $·leaveJITCheckedRaw(SB)
// func enterJITChecked(fn uintptr, stack uintptr)
// Sets sentinels in BP and R14, switches to the prepared stack and jumps
// to fn. The prepared stack's return address must be leaveJITCheckedRaw
// (read from leaveCheckedPtr).
TEXT ·enterJITChecked(SB), NOSPLIT, $0-16
MOVQ fn+0(FP), AX // target (before SP switch)
MOVQ SP, ·savedSP(SB) // preserve Go stack
MOVQ BP, ·savedBP(SB) // preserve frame pointer (vet requires save before clobber)
MOVQ $SENTINEL_BP, BP // sentinel in BP
MOVQ $SENTINEL_R14, R14 // sentinel in R14
MOVQ stack+8(FP), SP // switch to prepared stack
JMP AX
// leaveJITCheckedRaw is the raw return trampoline. It has NO Go function
// declaration, so no ABIInternal wrapper is generated — the JIT function's
// RET lands here directly, seeing BP and R14 exactly as the function left
// them. It checks the sentinels, records violations in abiResult, then
// restores the Go stack and returns.
TEXT ·leaveJITCheckedRaw(SB), NOSPLIT, $0-0
// Check BP against the sentinel.
MOVQ $SENTINEL_BP, CX
CMPQ BP, CX
JEQ bp_ok
ORQ $1, ·abiResult(SB)
bp_ok:
// Check R14 against the sentinel.
MOVQ $SENTINEL_R14, CX
CMPQ R14, CX
JEQ r14_ok
ORQ $2, ·abiResult(SB)
r14_ok:
MOVQ ·savedSP(SB), SP
RET
+26
View File
@@ -0,0 +1,26 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
//go:build !amd64
package verify
import "fmt"
// ABIReport describes the result of an ABI-checking call.
type ABIReport struct {
BPClobbered bool
R14Clobbered bool
RedZoneHit bool
}
// OK returns true when no violations were detected.
func (r ABIReport) OK() bool { return false }
// String returns a human-readable summary.
func (r ABIReport) String() string { return "verify: ABI checks require amd64" }
// CallChecked is unavailable on non-amd64 architectures.
func CallChecked(fnAddr uintptr, args []byte) ([]byte, ABIReport, error) {
return nil, ABIReport{}, fmt.Errorf("verify: ABI checks require amd64")
}
+145
View File
@@ -0,0 +1,145 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
package verify
import (
"testing"
"unsafe"
)
func loadABIKernel(t *testing.T) *Kernel {
t.Helper()
k, err := Load("../testdata/verify/abi_amd64.s")
if err != nil {
t.Fatalf("Load: %v", err)
}
t.Cleanup(k.Close)
return k
}
func TestABIClean(t *testing.T) {
k := loadABIKernel(t)
args := make([]byte, 24)
PutUint64(args, 0, 3)
PutUint64(args, 8, 4)
out, report, err := k.CallFuncChecked("cleanAdd", args)
if err != nil {
t.Fatalf("CallFuncChecked: %v", err)
}
if got := int64(GetUint64(out, 16)); got != 7 {
t.Errorf("cleanAdd(3, 4) = %d, want 7", got)
}
if !report.OK() {
t.Errorf("cleanAdd: %s", report)
}
}
func TestABIBPClobbered(t *testing.T) {
k := loadABIKernel(t)
args := make([]byte, 16)
PutUint64(args, 0, 42)
out, report, err := k.CallFuncChecked("dirtyBP", args)
if err != nil {
t.Fatalf("CallFuncChecked: %v", err)
}
if got := int64(GetUint64(out, 8)); got != 42 {
t.Errorf("dirtyBP(42) = %d, want 42", got)
}
if !report.BPClobbered {
t.Error("dirtyBP: expected BP clobbered, but report says clean")
}
if report.R14Clobbered {
t.Error("dirtyBP: R14 should not be clobbered")
}
}
func TestABIR14Clobbered(t *testing.T) {
k := loadABIKernel(t)
args := make([]byte, 16)
PutUint64(args, 0, 99)
out, report, err := k.CallFuncChecked("dirtyR14", args)
if err != nil {
t.Fatalf("CallFuncChecked: %v", err)
}
if got := int64(GetUint64(out, 8)); got != 99 {
t.Errorf("dirtyR14(99) = %d, want 99", got)
}
if !report.R14Clobbered {
t.Error("dirtyR14: expected R14 clobbered, but report says clean")
}
if report.BPClobbered {
t.Error("dirtyR14: BP should not be clobbered")
}
}
// TestABILZ4Kernels verifies that the production go-lz4 kernels are ABI-clean:
// they preserve BP and R14 and do not write into the red zone.
func TestABILZ4Kernels(t *testing.T) {
k := loadLZ4Kernel(t)
// wideCopyAVX2 with a real copy.
src := make([]byte, 128)
for i := range src {
src[i] = byte(i)
}
dst := make([]byte, 128)
args := make([]byte, 48)
PutPtr(args, 0, unsafe.Pointer(&dst[0]))
PutUint64(args, 8, 128)
PutUint64(args, 16, 128)
PutPtr(args, 24, unsafe.Pointer(&src[0]))
PutUint64(args, 32, 128)
PutUint64(args, 40, 128)
_, report, err := k.CallFuncChecked("wideCopyAVX2", args)
if err != nil {
t.Fatalf("CallFuncChecked(wideCopyAVX2): %v", err)
}
if !report.OK() {
t.Errorf("wideCopyAVX2: %s", report)
}
// decodeBlockAVX2 with a simple block.
decSrc := []byte{0x50, 'H', 'e', 'l', 'l', 'o'}
decDst := make([]byte, 64)
decArgs := make([]byte, 64)
PutPtr(decArgs, 0, unsafe.Pointer(&decSrc[0]))
PutUint64(decArgs, 8, uint64(len(decSrc)))
PutUint64(decArgs, 16, uint64(cap(decSrc)))
PutPtr(decArgs, 24, unsafe.Pointer(&decDst[0]))
PutUint64(decArgs, 32, uint64(len(decDst)))
PutUint64(decArgs, 40, uint64(cap(decDst)))
_, report, err = k.CallFuncChecked("decodeBlockAVX2", decArgs)
if err != nil {
t.Fatalf("CallFuncChecked(decodeBlockAVX2): %v", err)
}
if !report.OK() {
t.Errorf("decodeBlockAVX2: %s", report)
}
}
func TestCallFuncCheckedErrors(t *testing.T) {
k := loadABIKernel(t)
// Nonexistent function.
_, _, err := k.CallFuncChecked("nope", make([]byte, 8))
if err == nil {
t.Fatal("expected error for nonexistent function")
}
// Arg block too small.
_, _, err = k.CallFuncChecked("cleanAdd", make([]byte, 8))
if err == nil {
t.Fatal("expected error for too-small arg block")
}
}
+77
View File
@@ -0,0 +1,77 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
//go:build amd64
package verify
import (
"encoding/binary"
"fmt"
"reflect"
"syscall"
"unsafe"
)
// savedSP holds the Go stack pointer while a JIT call is in flight.
// Referenced by the assembly trampoline (trampoline_amd64.s).
var savedSP uintptr
// enterJIT switches to the prepared stack and jumps to fn.
// It does not return normally; the JIT function's RET transfers control
// to leaveJIT, which restores the Go stack.
//
//go:nosplit
func enterJIT(fn uintptr, stack uintptr)
// leaveJIT restores the Go stack after a JIT function returns.
// Its address is placed as the return address on the prepared stack.
//
//go:nosplit
func leaveJIT()
// leaveJITAddr is the machine address of leaveJIT, resolved once at init.
var leaveJITAddr uintptr
func init() {
leaveJITAddr = reflect.ValueOf(leaveJIT).Pointer()
}
// stackPad is padding below the return address on the prepared stack.
// The ABIInternal wrapper that leaveJIT's address resolves to executes
// PUSHQ BP and CALL before reaching the raw assembly, writing up to 16
// bytes below the return-address slot. 64 bytes of headroom is ample.
const stackPad = 64
// Call invokes the assembled function at fnAddr with the given ABI0 argument
// block (the raw bytes that would appear at FP+0). It returns the argument
// block after the call, which contains any results the function wrote back
// (the ABI0 convention shares the argument area for inputs and outputs).
//
// The function must be NOSPLIT (no stack growth) and must not reference
// external symbols — the image is self-contained.
func Call(fnAddr uintptr, args []byte) ([]byte, error) {
// Prepare the stack: [padding][leaveJIT addr][args...]
stackSize := stackPad + 8 + len(args) + 64 // padding + ret + args + safety
stackMem, err := syscall.Mmap(-1, 0, stackSize,
syscall.PROT_READ|syscall.PROT_WRITE, syscall.MAP_PRIVATE|syscall.MAP_ANON)
if err != nil {
return nil, fmt.Errorf("verify: stack mmap: %w", err)
}
defer syscall.Munmap(stackMem)
// The return address sits after the padding; the function's SP will
// point here, leaving stackPad bytes below for the wrapper's pushes.
retOff := stackPad
binary.LittleEndian.PutUint64(stackMem[retOff:retOff+8], uint64(leaveJITAddr))
// The ABI0 argument area follows the return address.
copy(stackMem[retOff+8:], args)
stackBase := uintptr(unsafe.Pointer(&stackMem[retOff]))
enterJIT(fnAddr, stackBase)
// Copy out the (possibly modified) argument area.
out := make([]byte, len(args))
copy(out, stackMem[retOff+8:retOff+8+len(args)])
return out, nil
}
+13
View File
@@ -0,0 +1,13 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
//go:build !amd64
package verify
import "fmt"
// Call is unavailable on non-amd64 architectures.
func Call(fnAddr uintptr, args []byte) ([]byte, error) {
return nil, fmt.Errorf("verify: JIT execution requires amd64")
}
+103
View File
@@ -0,0 +1,103 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
package verify
import (
"fmt"
"sort"
)
// Block describes one basic block within a function: a maximal sequence of
// instructions with a single entry point (a label or the function start) and
// a single exit (a jump, conditional jump or RET).
type Block struct {
Offset int // byte offset within the function
Label string // label name ("" for the entry block)
}
// Blocks identifies the basic blocks of a function from its local labels.
// Each label is a potential jump target and therefore a block boundary; the
// function entry (offset 0) is always a block. The blocks are returned in
// ascending offset order.
func (k *Kernel) Blocks(name string) ([]Block, error) {
idx, ok := k.funcs[name]
if !ok {
return nil, fmt.Errorf("verify: function %q not found", name)
}
fl := k.img.Funcs[idx]
blocks := []Block{{Offset: 0, Label: "(entry)"}}
// Build a reverse map: offset → label name.
offToLabel := make(map[int]string, len(fl.Labels))
for label, off := range fl.Labels {
if off > 0 && off < fl.Size {
offToLabel[off] = label
}
}
// Collect and sort offsets.
offsets := make([]int, 0, len(offToLabel))
for off := range offToLabel {
offsets = append(offsets, off)
}
sort.Ints(offsets)
for _, off := range offsets {
blocks = append(blocks, Block{Offset: off, Label: offToLabel[off]})
}
return blocks, nil
}
// BlockCount returns the number of identified basic blocks for the function.
func (k *Kernel) BlockCount(name string) (int, error) {
blocks, err := k.Blocks(name)
if err != nil {
return 0, err
}
return len(blocks), nil
}
// PathFingerprint is the observable output of one function execution: the
// values written back into the result slots of the argument block. Two
// executions that produce the same fingerprint took observationally
// equivalent paths (though they may differ internally).
type PathFingerprint struct {
Results []uint64 // the result words from the arg block
}
// ProfilePaths runs the function with each of the given argument blocks and
// collects the distinct output fingerprints. This measures path diversity:
// how many observationally different execution paths the input corpus
// exercises. Combined with Blocks (the static block count), it gives a
// lower bound on code coverage.
func (k *Kernel) ProfilePaths(name string, argSets [][]byte, resultOffsets []int) ([]PathFingerprint, error) {
idx, ok := k.funcs[name]
if !ok {
return nil, fmt.Errorf("verify: function %q not found", name)
}
fl := k.img.Funcs[idx]
seen := map[string]bool{}
var paths []PathFingerprint
for _, args := range argSets {
if len(args) < fl.Args {
return nil, fmt.Errorf("verify: %s: arg block too small", name)
}
out, err := k.CallFunc(name, args)
if err != nil {
return nil, err
}
fp := PathFingerprint{}
key := ""
for _, off := range resultOffsets {
v := GetUint64(out, off)
fp.Results = append(fp.Results, v)
key += fmt.Sprintf("%016x", v)
}
if !seen[key] {
seen[key] = true
paths = append(paths, fp)
}
}
return paths, nil
}
+85
View File
@@ -0,0 +1,85 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
package verify
import (
"testing"
"unsafe"
)
func TestBlocks(t *testing.T) {
k := loadBasic(t)
// The "sum" function has labels: sum_done, sum_loop.
blocks, err := k.Blocks("sum")
if err != nil {
t.Fatalf("Blocks(sum): %v", err)
}
if len(blocks) < 3 {
t.Errorf("sum: expected at least 3 blocks (entry + 2 labels), got %d", len(blocks))
}
if blocks[0].Offset != 0 {
t.Errorf("first block offset = %d, want 0", blocks[0].Offset)
}
t.Logf("sum blocks: %v", blocks)
}
func TestBlockCount(t *testing.T) {
k := loadLZ4Kernel(t)
n, err := k.BlockCount("decodeBlockAVX2")
if err != nil {
t.Fatalf("BlockCount: %v", err)
}
// The decoder has many labels (dec_loop, dec_malformed, etc.).
if n < 10 {
t.Errorf("decodeBlockAVX2: expected at least 10 blocks, got %d", n)
}
t.Logf("decodeBlockAVX2: %d basic blocks", n)
}
func TestProfilePaths(t *testing.T) {
k := loadLZ4Kernel(t)
// Build a corpus of varied LZ4 blocks.
var argSets [][]byte
blocks := []struct {
src []byte
dstSize int
}{
{[]byte{0x00}, 16}, // empty
{[]byte{0x50, 'H', 'e', 'l', 'l', 'o'}, 16}, // literals only
{[]byte{0x54, 'A', 'A', 'A', 'A', 'A', 5, 0, 0x30, 'B', 'B', 'B'}, 32}, // match
{[]byte{0x14, 'X', 1, 0, 0x10, 'Y'}, 16}, // overlapping
{[]byte{0x50, 'H'}, 16}, // malformed
{[]byte{0x14, 'X', 0, 0}, 16}, // zero offset
}
for _, b := range blocks {
args := make([]byte, 64)
if len(b.src) > 0 {
PutPtr(args, 0, unsafe.Pointer(&b.src[0]))
}
PutUint64(args, 8, uint64(len(b.src)))
PutUint64(args, 16, uint64(cap(b.src)))
dst := make([]byte, b.dstSize)
if len(dst) > 0 {
PutPtr(args, 24, unsafe.Pointer(&dst[0]))
}
PutUint64(args, 32, uint64(len(dst)))
PutUint64(args, 40, uint64(cap(dst)))
argSets = append(argSets, args)
}
// Result offsets: n+48 and code+56.
paths, err := k.ProfilePaths("decodeBlockAVX2", argSets, []int{48, 56})
if err != nil {
t.Fatalf("ProfilePaths: %v", err)
}
// We expect at least 3 distinct paths: success (various n), malformed, zero offset.
if len(paths) < 3 {
t.Errorf("expected at least 3 distinct paths, got %d", len(paths))
}
t.Logf("decodeBlockAVX2: %d distinct output paths from %d inputs", len(paths), len(argSets))
}
+295
View File
@@ -0,0 +1,295 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
package verify
import (
"bytes"
"math/rand"
"testing"
"unsafe"
)
// decodeBlockGo is a minimal portable LZ4 block decoder used as the
// differential-testing oracle. It mirrors the contract of
// go-lz4's decodeBlockGo: (bytesWritten, code) where code is
// 0 = ok, 1 = malformed, 2 = zero offset.
func decodeBlockGo(src, dst []byte) (int, int) {
if len(src) == 0 {
return 0, 1
}
si, di := 0, 0
for {
if si >= len(src) {
return 0, 1 // truncated: no token
}
token := int(src[si])
si++
// Literals.
lLen := token >> 4
if lLen == 15 {
for {
if si >= len(src) {
return 0, 1
}
b := int(src[si])
si++
lLen += b
if b != 255 {
break
}
}
}
if si+lLen > len(src) {
return 0, 1 // truncated literals
}
if di+lLen > len(dst) {
return 0, 1 // destination overflow
}
copy(dst[di:di+lLen], src[si:si+lLen])
di += lLen
si += lLen
// End of block.
if si >= len(src) {
return di, 0
}
// Match offset.
if si+2 > len(src) {
return 0, 1
}
offset := int(src[si]) | int(src[si+1])<<8
si += 2
if offset == 0 {
return 0, 2
}
// Match length.
mLen := token & 15
if mLen == 15 {
for {
if si >= len(src) {
return 0, 1
}
b := int(src[si])
si++
mLen += b
if b != 255 {
break
}
}
}
mLen += 4
// Copy match (overlapping-safe).
if di-offset < 0 {
return 0, 1 // offset reaches before dst start
}
if di+mLen > len(dst) {
return 0, 1 // destination overflow
}
for i := 0; i < mLen; i++ {
dst[di+i] = dst[di-offset+i]
}
di += mLen
}
}
// genLZ4Block generates a random valid LZ4 block that decompresses into
// approximately wantSize bytes. The block is always well-formed (ends with
// a literals-only sequence).
func genLZ4Block(rng *rand.Rand, wantSize int) []byte {
var block []byte
produced := 0
for produced < wantSize {
remaining := wantSize - produced
// Decide: emit a literals+match sequence or the final literals.
if remaining <= 8 || rng.Intn(4) == 0 {
// Final literals-only sequence.
lLen := remaining
if lLen > 60 {
lLen = 1 + rng.Intn(60)
}
block = appendToken(block, lLen, 0)
for i := 0; i < lLen; i++ {
block = append(block, byte(rng.Intn(256)))
}
produced += lLen
break
}
// Literals + match.
lLen := rng.Intn(min(16, remaining))
if produced+lLen == 0 {
lLen = 1 // must have at least 1 literal before the first match
}
mLenRaw := rng.Intn(12) // match length = mLenRaw + 4
mLen := mLenRaw + 4
if produced+mLen > remaining {
mLen = remaining - produced
if mLen < 4 {
// Not enough room for a match; emit final literals.
lLen = remaining
block = appendToken(block, lLen, 0)
for i := 0; i < lLen; i++ {
block = append(block, byte(rng.Intn(256)))
}
break
}
mLenRaw = mLen - 4
}
block = appendToken(block, lLen, mLenRaw)
for i := 0; i < lLen; i++ {
block = append(block, byte(rng.Intn(256)))
}
produced += lLen
// Offset: must be <= produced (can't reference before start).
maxOff := produced
if maxOff > 65535 {
maxOff = 65535
}
offset := 1 + rng.Intn(maxOff)
block = append(block, byte(offset), byte(offset>>8))
produced += mLen
}
return block
}
// appendToken appends a token (and extension bytes if needed) for the given
// literal and match lengths.
func appendToken(block []byte, lLen, mLenRaw int) []byte {
lit4 := lLen
if lit4 > 15 {
lit4 = 15
}
ml4 := mLenRaw
if ml4 > 15 {
ml4 = 15
}
block = append(block, byte(lit4<<4|ml4))
// Literal extension bytes.
rem := lLen - 15
for rem >= 255 {
block = append(block, 255)
rem -= 255
}
if lLen >= 15 {
block = append(block, byte(rem))
}
// Match extension bytes.
rem = mLenRaw - 15
for rem >= 255 {
block = append(block, 255)
rem -= 255
}
if mLenRaw >= 15 {
block = append(block, byte(rem))
}
return block
}
func min(a, b int) int {
if a < b {
return a
}
return b
}
// TestDifferentialLZ4Fuzz drives the JIT-assembled decodeBlockAVX2 with
// random valid LZ4 blocks and compares the output bit-for-bit against the
// portable Go reference.
func TestDifferentialLZ4Fuzz(t *testing.T) {
k := loadLZ4Kernel(t)
const iterations = 5000
rng := rand.New(rand.NewSource(42))
for i := 0; i < iterations; i++ {
wantSize := 1 + rng.Intn(4096)
src := genLZ4Block(rng, wantSize)
dstSize := wantSize + 64 // generous destination
// Go reference.
goDst := make([]byte, dstSize)
goN, goCode := decodeBlockGo(src, goDst)
// JIT kernel.
jitDst := make([]byte, dstSize)
jitN, jitCode := callDecodeBlockAVX2(t, k, src, jitDst)
if jitCode != goCode {
t.Fatalf("iter %d: code mismatch: JIT=%d, Go=%d (src len=%d)",
i, jitCode, goCode, len(src))
}
if jitCode != 0 {
continue // both agree it's malformed/zero-offset
}
if jitN != goN {
t.Fatalf("iter %d: n mismatch: JIT=%d, Go=%d (src len=%d)",
i, jitN, goN, len(src))
}
if !bytes.Equal(jitDst[:jitN], goDst[:goN]) {
t.Fatalf("iter %d: output mismatch (n=%d, src len=%d)", i, jitN, len(src))
}
}
}
// TestDifferentialLZ4Hostile drives the kernel with random garbage to check
// that error codes agree with the Go reference (no crashes, same classification).
func TestDifferentialLZ4Hostile(t *testing.T) {
k := loadLZ4Kernel(t)
const iterations = 2000
rng := rand.New(rand.NewSource(99))
for i := 0; i < iterations; i++ {
srcLen := rng.Intn(128)
src := make([]byte, srcLen)
rng.Read(src)
dstSize := rng.Intn(512)
dst := make([]byte, dstSize)
// Go reference.
goDst := make([]byte, dstSize)
copy(goDst, dst)
_, goCode := decodeBlockGo(src, goDst)
// JIT kernel.
jitDst := make([]byte, dstSize)
copy(jitDst, dst)
_, jitCode := callDecodeBlockAVX2(t, k, src, jitDst)
if jitCode != goCode {
t.Fatalf("iter %d: hostile code mismatch: JIT=%d, Go=%d (srcLen=%d, dstSize=%d)",
i, jitCode, goCode, srcLen, dstSize)
}
}
}
// callDecodeBlockAVX2Raw is like callDecodeBlockAVX2 but accepts explicit
// dst size (for hostile tests where dst may be smaller than the output).
func callDecodeBlockAVX2Raw(t *testing.T, k *Kernel, src, dst []byte) (int, int) {
t.Helper()
args := make([]byte, 64)
if len(src) > 0 {
PutPtr(args, 0, unsafe.Pointer(&src[0]))
}
PutUint64(args, 8, uint64(len(src)))
PutUint64(args, 16, uint64(cap(src)))
if len(dst) > 0 {
PutPtr(args, 24, unsafe.Pointer(&dst[0]))
}
PutUint64(args, 32, uint64(len(dst)))
PutUint64(args, 40, uint64(cap(dst)))
out, err := k.CallFunc("decodeBlockAVX2", args)
if err != nil {
t.Fatalf("CallFunc(decodeBlockAVX2): %v", err)
}
return int(GetUint64(out, 48)), int(GetUint64(out, 56))
}
+88
View File
@@ -0,0 +1,88 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
// Package verify provides the dynamic-analysis substrate for gasm: it
// JIT-assembles Plan 9 amd64 kernels into executable memory and calls them
// directly, enabling differential testing against portable Go references,
// runtime ABI checks and basic-block coverage profiling.
//
// The execution model is pure Go (stdlib only): machine code is mapped with
// syscall.Mmap and invoked through an assembly trampoline that switches to a
// prepared ABI0 stack. No cgo, no external toolchain.
package verify
import (
"encoding/binary"
"fmt"
"syscall"
"unsafe"
)
// Executable maps a copy of code into a read-execute memory region suitable
// for direct invocation. The mapping is anonymous and private; the original
// slice is not retained. Call Unmap to release the region.
type Executable struct {
addr uintptr // base address of the mapping
size int
mem []byte // the mmap'd slice (for Unmap)
}
// Map copies code into a freshly allocated RX region and returns it.
// The mapping is PROT_READ|PROT_EXEC; writes are not permitted after the
// copy, matching W^X policy.
func Map(code []byte) (*Executable, error) {
size := len(code)
if size == 0 {
return nil, fmt.Errorf("verify: cannot map zero-length code")
}
// Round up to the page size.
const pageSize = 4096
mapSize := (size + pageSize - 1) &^ (pageSize - 1)
mem, err := syscall.Mmap(-1, 0, mapSize,
syscall.PROT_READ|syscall.PROT_WRITE, syscall.MAP_PRIVATE|syscall.MAP_ANON)
if err != nil {
return nil, fmt.Errorf("verify: mmap: %w", err)
}
copy(mem, code)
// Remove write permission (W^X).
if err := syscall.Mprotect(mem, syscall.PROT_READ|syscall.PROT_EXEC); err != nil {
syscall.Munmap(mem)
return nil, fmt.Errorf("verify: mprotect: %w", err)
}
return &Executable{
addr: uintptr(unsafe.Pointer(&mem[0])),
size: size,
mem: mem,
}, nil
}
// Unmap releases the executable region.
func (e *Executable) Unmap() {
if e.mem != nil {
syscall.Munmap(e.mem)
e.mem = nil
}
}
// FuncAddr returns the absolute address of a function at the given offset
// within the mapped image.
func (e *Executable) FuncAddr(offset int) uintptr {
return e.addr + uintptr(offset)
}
// PutUint64 writes v into buf at byte offset off (little-endian).
func PutUint64(buf []byte, off int, v uint64) {
binary.LittleEndian.PutUint64(buf[off:off+8], v)
}
// GetUint64 reads a little-endian uint64 from buf at byte offset off.
func GetUint64(buf []byte, off int) uint64 {
return binary.LittleEndian.Uint64(buf[off : off+8])
}
// PutPtr writes a pointer value into buf at byte offset off.
func PutPtr(buf []byte, off int, p unsafe.Pointer) {
binary.LittleEndian.PutUint64(buf[off:off+8], uint64(uintptr(p)))
}
+215
View File
@@ -0,0 +1,215 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
package verify
import (
"bytes"
"testing"
"unsafe"
)
func loadBasic(t *testing.T) *Kernel {
t.Helper()
k, err := Load("../testdata/verify/basic_amd64.s")
if err != nil {
t.Fatalf("Load: %v", err)
}
t.Cleanup(k.Close)
return k
}
func TestJITAdd(t *testing.T) {
k := loadBasic(t)
tests := []struct {
a, b, want int64
}{
{0, 0, 0},
{1, 2, 3},
{-1, 1, 0},
{1 << 62, 1 << 62, -9223372036854775808}, // overflow wraps (MinInt64)
{-100, -200, -300},
}
for _, tt := range tests {
args := make([]byte, 24)
PutUint64(args, 0, uint64(tt.a))
PutUint64(args, 8, uint64(tt.b))
out, err := k.CallFunc("add", args)
if err != nil {
t.Fatalf("CallFunc(add, %d, %d): %v", tt.a, tt.b, err)
}
got := int64(GetUint64(out, 16))
if got != tt.want {
t.Errorf("add(%d, %d) = %d, want %d", tt.a, tt.b, got, tt.want)
}
}
}
func TestJITSum(t *testing.T) {
k := loadBasic(t)
tests := []struct {
data []int64
want int64
}{
{nil, 0},
{[]int64{1}, 1},
{[]int64{1, 2, 3, 4, 5}, 15},
{[]int64{-10, 20, -30, 40}, 20},
}
for _, tt := range tests {
args := make([]byte, 32)
if len(tt.data) > 0 {
PutPtr(args, 0, unsafe.Pointer(&tt.data[0]))
}
PutUint64(args, 8, uint64(len(tt.data)))
PutUint64(args, 16, uint64(cap(tt.data)))
out, err := k.CallFunc("sum", args)
if err != nil {
t.Fatalf("CallFunc(sum, %v): %v", tt.data, err)
}
got := int64(GetUint64(out, 24))
if got != tt.want {
t.Errorf("sum(%v) = %d, want %d", tt.data, got, tt.want)
}
}
}
func TestJITWideCopy(t *testing.T) {
k := loadBasic(t)
tests := []struct {
name string
n int
}{
{"empty", 0},
{"tiny", 7},
{"exact32", 32},
{"overlap_range", 48},
{"exact64", 64},
{"unaligned", 45},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
src := make([]byte, tt.n)
for i := range src {
src[i] = byte(i * 7)
}
dst := make([]byte, tt.n)
args := make([]byte, 48)
if tt.n > 0 {
PutPtr(args, 0, unsafe.Pointer(&dst[0]))
PutPtr(args, 24, unsafe.Pointer(&src[0]))
}
PutUint64(args, 8, uint64(tt.n)) // dst_len
PutUint64(args, 16, uint64(tt.n)) // dst_cap
PutUint64(args, 32, uint64(tt.n)) // src_len
PutUint64(args, 40, uint64(tt.n)) // src_cap
_, err := k.CallFunc("wideCopy", args)
if err != nil {
t.Fatalf("CallFunc(wideCopy): %v", err)
}
if !bytes.Equal(dst, src) {
t.Errorf("wideCopy: dst ≠ src\n got %x\n want %x", dst, src)
}
})
}
}
func TestKernelFuncNames(t *testing.T) {
k := loadBasic(t)
names := k.FuncNames()
want := []string{"add", "sum", "wideCopy"}
if len(names) != len(want) {
t.Fatalf("FuncNames() = %v, want %v", names, want)
}
for i, n := range names {
if n != want[i] {
t.Errorf("FuncNames()[%d] = %q, want %q", i, n, want[i])
}
}
}
func TestKernelFuncNotFound(t *testing.T) {
k := loadBasic(t)
_, err := k.CallFunc("nonexistent", make([]byte, 8))
if err == nil {
t.Fatal("expected error for nonexistent function")
}
}
func TestKernelArgTooSmall(t *testing.T) {
k := loadBasic(t)
_, err := k.CallFunc("add", make([]byte, 8)) // needs 24
if err == nil {
t.Fatal("expected error for too-small arg block")
}
}
func TestMapZeroLength(t *testing.T) {
_, err := Map(nil)
if err == nil {
t.Fatal("expected error for zero-length code")
}
}
func TestLoadSourceError(t *testing.T) {
_, err := LoadSource("bad.s", "TEXT ·f(SB), NOSPLIT\n\tBADINSTRUCTION\n")
// The parser may or may not error on unknown instructions (it's
// error-tolerant), but the assembler will reject it.
if err == nil {
t.Log("LoadSource succeeded unexpectedly (parser is error-tolerant)")
}
}
func TestLoadSourceParseError(t *testing.T) {
// A completely invalid file that the parser rejects.
_, err := LoadSource("empty.s", "")
if err != nil {
t.Logf("expected: %v", err)
}
}
func TestFuncLookup(t *testing.T) {
k := loadBasic(t)
fl, err := k.Func("add")
if err != nil {
t.Fatalf("Func(add): %v", err)
}
if fl.Name != "add" {
t.Errorf("Func(add).Name = %q, want %q", fl.Name, "add")
}
if fl.Args != 24 {
t.Errorf("Func(add).Args = %d, want 24", fl.Args)
}
_, err = k.Func("nonexistent")
if err == nil {
t.Fatal("expected error for nonexistent function")
}
}
func TestABIReportString(t *testing.T) {
r := ABIReport{}
if r.String() != "ABI clean" {
t.Errorf("clean report = %q", r.String())
}
r.BPClobbered = true
if r.OK() {
t.Error("expected not OK with BP clobbered")
}
s := r.String()
if s == "ABI clean" {
t.Error("expected violation string, got clean")
}
r.R14Clobbered = true
r.RedZoneHit = true
s = r.String()
if s == "ABI clean" {
t.Error("expected violation string for all flags")
}
}
+160
View File
@@ -0,0 +1,160 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
package verify
import (
"bytes"
"os"
"testing"
"unsafe"
)
// lz4KernelPath is the sibling repository's AVX2 kernel, used for
// integration testing. The test is skipped when the file is absent
// (e.g. in CI without the sibling checkout).
const lz4KernelPath = "../../go-libraries/go-lz4/avx2_amd64.s"
func loadLZ4Kernel(t *testing.T) *Kernel {
t.Helper()
if _, err := os.Stat(lz4KernelPath); err != nil {
t.Skipf("sibling kernel not available: %v", err)
}
k, err := Load(lz4KernelPath)
if err != nil {
t.Fatalf("Load(%s): %v", lz4KernelPath, err)
}
t.Cleanup(k.Close)
return k
}
// callDecodeBlockAVX2 invokes the JIT-assembled decodeBlockAVX2 with the
// given src and dst buffers, returning (n, code).
func callDecodeBlockAVX2(t *testing.T, k *Kernel, src, dst []byte) (int, int) {
t.Helper()
args := make([]byte, 64)
if len(src) > 0 {
PutPtr(args, 0, unsafe.Pointer(&src[0]))
}
PutUint64(args, 8, uint64(len(src)))
PutUint64(args, 16, uint64(cap(src)))
if len(dst) > 0 {
PutPtr(args, 24, unsafe.Pointer(&dst[0]))
}
PutUint64(args, 32, uint64(len(dst)))
PutUint64(args, 40, uint64(cap(dst)))
out, err := k.CallFunc("decodeBlockAVX2", args)
if err != nil {
t.Fatalf("CallFunc(decodeBlockAVX2): %v", err)
}
return int(GetUint64(out, 48)), int(GetUint64(out, 56))
}
func TestLZ4DecodeKnownAnswers(t *testing.T) {
k := loadLZ4Kernel(t)
tests := []struct {
name string
src []byte
dstSize int
wantDst []byte
wantN int
wantCode int
}{
{
name: "literals_only",
src: []byte{0x50, 'H', 'e', 'l', 'l', 'o'},
dstSize: 16,
wantDst: []byte("Hello"),
wantN: 5,
wantCode: 0,
},
{
name: "literals_and_match",
src: []byte{0x54, 'A', 'A', 'A', 'A', 'A', 0x05, 0x00, 0x30, 'B', 'B', 'B'},
dstSize: 32,
wantDst: []byte("AAAAAAAAAAAAABBB"),
wantN: 16,
wantCode: 0,
},
{
name: "overlapping_match",
// 1 literal 'X', then match offset=1 length=4+4=8 → "XXXXXXXXX",
// then final 1 literal 'Y'.
src: []byte{0x14, 'X', 0x01, 0x00, 0x10, 'Y'},
dstSize: 16,
wantDst: []byte("XXXXXXXXXY"),
wantN: 10,
wantCode: 0,
},
{
name: "malformed_truncated",
src: []byte{0x50, 'H', 'e'}, // claims 5 literals, has 2
dstSize: 16,
wantN: 0,
wantCode: 1,
},
{
name: "zero_offset",
src: []byte{0x14, 'X', 0x00, 0x00},
dstSize: 16,
wantN: 0,
wantCode: 2,
},
{
name: "empty_token",
src: []byte{0x00}, // 0 literals, end of block
dstSize: 16,
wantDst: nil,
wantN: 0,
wantCode: 0,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
dst := make([]byte, tt.dstSize)
n, code := callDecodeBlockAVX2(t, k, tt.src, dst)
if n != tt.wantN || code != tt.wantCode {
t.Fatalf("decodeBlockAVX2: got (n=%d, code=%d), want (n=%d, code=%d)",
n, code, tt.wantN, tt.wantCode)
}
if tt.wantCode == 0 && tt.wantDst != nil {
if !bytes.Equal(dst[:n], tt.wantDst) {
t.Errorf("output mismatch:\n got %q\n want %q", dst[:n], tt.wantDst)
}
}
})
}
}
func TestLZ4WideCopyAVX2(t *testing.T) {
k := loadLZ4Kernel(t)
sizes := []int{0, 1, 15, 16, 31, 32, 33, 63, 64, 100, 256, 1024}
for _, n := range sizes {
src := make([]byte, n)
for i := range src {
src[i] = byte(i*13 + 7)
}
dst := make([]byte, n)
args := make([]byte, 48)
if n > 0 {
PutPtr(args, 0, unsafe.Pointer(&dst[0]))
PutPtr(args, 24, unsafe.Pointer(&src[0]))
}
PutUint64(args, 8, uint64(n))
PutUint64(args, 16, uint64(n))
PutUint64(args, 32, uint64(n))
PutUint64(args, 40, uint64(n))
_, err := k.CallFunc("wideCopyAVX2", args)
if err != nil {
t.Fatalf("wideCopyAVX2(n=%d): %v", n, err)
}
if !bytes.Equal(dst, src) {
t.Errorf("wideCopyAVX2(n=%d): output mismatch", n)
}
}
}
+30
View File
@@ -0,0 +1,30 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
#include "textflag.h"
// ABI0 JIT trampoline. enterJIT switches from the Go stack to a prepared
// stack and jumps to the assembled function; when the function RETs, control
// lands in leaveJIT, which restores the Go stack and returns to the Go caller.
//
// The prepared stack must begin with the address of leaveJIT (the return
// address the JIT function will pop), followed by the function's ABI0
// argument area.
//
// Single-threaded: savedSP is a package global, so only one JIT call may be
// in flight at a time. gasm verify runs sequentially.
// func enterJIT(fn uintptr, stack uintptr)
// Switches to the prepared stack and jumps to fn. Does not return normally;
// the JIT function's RET transfers control to leaveJIT.
TEXT ·enterJIT(SB), NOSPLIT, $0-16
MOVQ fn+0(FP), AX // target function address (before SP switch)
MOVQ SP, ·savedSP(SB) // preserve the Go stack pointer
MOVQ stack+8(FP), SP // switch to the prepared stack
JMP AX
// func leaveJIT()
// Restores the Go stack pointer and returns to enterJIT's caller.
TEXT ·leaveJIT(SB), NOSPLIT, $0-0
MOVQ ·savedSP(SB), SP
RET
+122
View File
@@ -0,0 +1,122 @@
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
// SPDX-License-Identifier: BSD-3-Clause
package verify
import (
"fmt"
"os"
"sourcedock.dev/petrbalvin/gasm-devkit/asm"
"sourcedock.dev/petrbalvin/gasm-devkit/ast"
"sourcedock.dev/petrbalvin/gasm-devkit/parser"
)
// Kernel is a JIT-loaded assembly image ready for direct invocation.
// It wraps an executable memory mapping and the function layout metadata
// needed to marshal ABI0 calls.
type Kernel struct {
exec *Executable
img *asm.Image
funcs map[string]int // function name → index into img.Funcs
}
// Load parses, assembles and maps a .s file into executable memory.
// The returned Kernel is ready for Call. The caller must call Close to
// release the mapping.
func Load(path string) (*Kernel, error) {
src, err := os.ReadFile(path)
if err != nil {
return nil, fmt.Errorf("verify: %w", err)
}
return LoadSource(path, string(src))
}
// LoadSource parses, assembles and maps assembly source into executable memory.
func LoadSource(filename, src string) (*Kernel, error) {
file, errs := parser.Parse(filename, src)
if len(errs) > 0 {
return nil, fmt.Errorf("verify: parse %s: %v", filename, errs[0])
}
return LoadAST(file)
}
// LoadAST assembles a parsed AST file and maps the result into executable
// memory.
func LoadAST(file *ast.File) (*Kernel, error) {
img, err := asm.AssembleFile(file)
if err != nil {
return nil, fmt.Errorf("verify: assemble: %w", err)
}
if len(img.Externals) > 0 {
return nil, fmt.Errorf("verify: unresolved external symbols: %v", img.Externals)
}
code := img.Bytes()
exec, err := Map(code)
if err != nil {
return nil, err
}
funcs := make(map[string]int, len(img.Funcs))
for i, f := range img.Funcs {
funcs[f.Name] = i
}
return &Kernel{exec: exec, img: img, funcs: funcs}, nil
}
// Func returns the layout metadata for the named function.
func (k *Kernel) Func(name string) (asm.FuncLayout, error) {
idx, ok := k.funcs[name]
if !ok {
return asm.FuncLayout{}, fmt.Errorf("verify: function %q not found", name)
}
return k.img.Funcs[idx], nil
}
// FuncNames returns the names of all functions in the kernel, in source order.
func (k *Kernel) FuncNames() []string {
names := make([]string, len(k.img.Funcs))
for i, f := range k.img.Funcs {
names[i] = f.Name
}
return names
}
// CallFunc invokes the named function with the given ABI0 argument block.
// The arg block is the raw bytes of the function's argument/result area
// (as declared by the TEXT $frame-args suffix). Returns the arg block
// after the call (with any results written back by the function).
func (k *Kernel) CallFunc(name string, args []byte) ([]byte, error) {
idx, ok := k.funcs[name]
if !ok {
return nil, fmt.Errorf("verify: function %q not found", name)
}
fl := k.img.Funcs[idx]
if len(args) < fl.Args {
return nil, fmt.Errorf("verify: %s: arg block too small: got %d, need %d", name, len(args), fl.Args)
}
fnAddr := k.exec.FuncAddr(fl.Offset)
return Call(fnAddr, args)
}
// CallFuncChecked invokes the named function with ABI sentinels and a
// red-zone canary, returning the argument block and an ABIReport that
// records any callee-saved register or red-zone violations.
func (k *Kernel) CallFuncChecked(name string, args []byte) ([]byte, ABIReport, error) {
idx, ok := k.funcs[name]
if !ok {
return nil, ABIReport{}, fmt.Errorf("verify: function %q not found", name)
}
fl := k.img.Funcs[idx]
if len(args) < fl.Args {
return nil, ABIReport{}, fmt.Errorf("verify: %s: arg block too small: got %d, need %d", name, len(args), fl.Args)
}
fnAddr := k.exec.FuncAddr(fl.Offset)
return CallChecked(fnAddr, args)
}
// Close releases the executable mapping.
func (k *Kernel) Close() {
if k.exec != nil {
k.exec.Unmap()
}
}