cmd/loopvec-toolexec: rewrite default builds on darwin, ios, android and windows

.github/workflows/test.yml

@@ -61,8 +61,13 @@ jobs:
       - name: Test tsvc (amd64, SIMD)
         run: make tsvc-test
 
-      - name: Install qemu-user-static
-        run: sudo apt-get update && sudo apt-get install -y qemu-user-static
+      # ubuntu-26.04 has no qemu-user-static: qemu-user ships the static
+      # emulators there, without the -static names the Makefile runs.
+      - name: Install qemu-user
+        run: |
+          sudo apt-get update && sudo apt-get install -y qemu-user
+          command -v qemu-aarch64-static || sudo ln -s "$(command -v qemu-aarch64)" /usr/local/bin/qemu-aarch64-static
+          qemu-aarch64-static --version
 
       - name: Test tsvc (arm64, qemu)
         run: make tsvc-test-qemu-arm64

BUGS.md

@@ -2,11 +2,11 @@
 
 loopvec rewrites code to use `GOEXPERIMENT=simd`, which is new, so most of what it hits
 is in the Go toolchain. This file lists the Go issues that affect loopvec. "tip" means
-gotip, "1.27.1" is the release toolchain loopvec is tested against.
+gotip, "1.27.3" is the release toolchain loopvec is tested against.
 
 | Issue | Status | Affects | loopvec's answer |
 |---|---|---|---|
-| [#80657](https://github.com/golang/go/issues/80657) simd in a method: internal compiler error | fixed on tip (CL 839405), still crashes on 1.27.1 | `-methods` | methods are skipped unless `-methods` is passed |
+| [#80657](https://github.com/golang/go/issues/80657) simd in a method: internal compiler error | fixed on tip (CL 839405), still crashes on 1.27.3 | `-methods` | methods are skipped unless `-methods` is passed |
 | [#80835](https://github.com/golang/go/issues/80835) legacy SSE encodings in functions using simd cause AVX-SSE transition penalties | open, fix pending (CLs 825145, 825146, in tip) | speed of any scalar float code that runs after a rewritten loop, in the same function or another | none; the portable `simd` package has no `ClearAVXUpperBits` |
 | [#81844](https://github.com/golang/go/issues/81844) blank identifier parameters rejected | fixed on tip (CL 841145) | every rewrite of a function with a `_` parameter | the generated file renames them `_p0`, `_p1`, ... |
 | [#81845](https://github.com/golang/go/issues/81845) calls left unresolved when inlined into a package initializer | closed as a duplicate of [#80689](https://github.com/golang/go/issues/80689), which is open | loops in package-level `var x = func() {...}()` | not rewritten (the generated initializer fails to link), see `package_initializer_no_rewrite.txt` and `toolexec_package_initializer.txt` |
@@ -19,7 +19,7 @@ gotip, "1.27.1" is the release toolchain loopvec is tested against.
 
 A function with a receiver that contains simd code failed to compile with
 `internal compiler error: missing Types entry`. loopvec does not rewrite loops in methods
-by default; `-methods` turns it on. The fix is on tip. On 1.27.1 the crash is still there,
+by default; `-methods` turns it on. The fix is on tip. On 1.27.3 the crash is still there,
 and `flag_methods_ice.txt` pins it, so that test starts failing when the pinned toolchain
 gets the fix. Reported by someone else; the fix is [CL 839405](https://go.dev/cl/839405).
 

Makefile

@@ -1,4 +1,4 @@
-export GOTOOLCHAIN ?= go1.27.1
+export GOTOOLCHAIN ?= go1.27.3
 
 .DEFAULT_GOAL := build
 

cmd/loopvec-toolexec/compile.go

@@ -34,10 +34,30 @@ type compileArgs struct {
 	importcfg int   // index in args of the importcfg path, or -1
 	files     []int // indexes in args of the Go source files
 	// unsupported is set for build modes whose objects the nested simd build
-	// would not match (race, msan, asan, shared, dynlink, -trimpath).
+	// would not match (race, msan, asan, shared, dynlink, -trimpath). -shared
+	// is also what the go command passes for a position-independent
+	// executable, the default build mode on some targets (see defaultPIE);
+	// there the nested build uses the same default and matches.
 	unsupported bool
 }
 
+// defaultPIE reports whether the target builds position-independent
+// executables by default (internal/platform.DefaultPIE), so that the go command
+// passes -shared to every compile of a default build. The go command sets GOOS
+// for the tools it runs. On windows -race turns the default off, but -race is
+// unsupported here on its own.
+func defaultPIE() bool {
+	goos := os.Getenv("GOOS")
+	if goos == "" {
+		goos = runtime.GOOS
+	}
+	switch goos {
+	case "android", "darwin", "ios", "windows":
+		return true
+	}
+	return false
+}
+
 func parseCompileArgs(args []string) compileArgs {
 	c := compileArgs{importcfg: -1}
 	first := len(args)
@@ -59,8 +79,10 @@ func parseCompileArgs(args []string) compileArgs {
 			c.lang = args[i+1]
 		case strings.HasPrefix(arg, "-lang="):
 			c.lang = strings.TrimPrefix(arg, "-lang=")
-		case arg == "-race", arg == "-msan", arg == "-asan", arg == "-shared", arg == "-dynlink":
+		case arg == "-race", arg == "-msan", arg == "-asan", arg == "-dynlink":
 			c.unsupported = true
+		case arg == "-shared":
+			c.unsupported = !defaultPIE()
 		case arg == "-trimpath" && hasValue && strings.Contains(args[i+1], ";"):
 			c.unsupported = true
 		}

go.mod

@@ -3,7 +3,7 @@ module github.com/mauri870/loopvec
 go 1.27.1
 
 require (
-	golang.org/x/tools v0.51.0
+	golang.org/x/tools v0.52.1-0.20261030190145-0f433775b36d
 	rsc.io/script v0.0.2
 )
 
@@ -207,10 +207,10 @@ require (
 	go.uber.org/zap v1.27.0 // indirect
 	go.yaml.in/yaml/v3 v3.0.5 // indirect
 	golang.org/x/exp/typeparams v0.0.0-20260811152304-ee035b5b010f // indirect
-	golang.org/x/mod v0.41.0 // indirect
-	golang.org/x/sync v0.23.0 // indirect
-	golang.org/x/sys v0.48.0 // indirect
-	golang.org/x/telemetry v0.0.0-20260924152758-ed294f943157 // indirect
+	golang.org/x/mod v0.42.0 // indirect
+	golang.org/x/sync v0.24.0 // indirect
+	golang.org/x/sys v0.49.0 // indirect
+	golang.org/x/telemetry v0.0.0-20261009151841-97e8ea77aa2a // indirect
 	golang.org/x/text v0.41.0 // indirect
 	google.golang.org/protobuf v1.36.12 // indirect
 	gopkg.in/ini.v1 v1.67.0 // indirect

go.sum

@@ -688,8 +688,8 @@ golang.org/x/mod v0.4.1/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
 golang.org/x/mod v0.6.0-dev.0.20220106191415-9b9b3d81d5e3/go.mod h1:3p9vT2HGsQu2K1YbXdKPJLVgG5VJdoTa1poYQBtP1AY=
 golang.org/x/mod v0.6.0-dev.0.20220419223038-86c51ed26bb4/go.mod h1:jJ57K6gSWd91VN4djpZkiMVwK6gcyfeH4XE8wZrZaV4=
 golang.org/x/mod v0.8.0/go.mod h1:iBbtSCu2XBx23ZKBPSOrRkjjQPZFPuis4dIYUhu/chs=
-golang.org/x/mod v0.41.0 h1:qJmnOUb4YB+FsEuM3HcWucdZASCPGhsX6uljO6pog0c=
-golang.org/x/mod v0.41.0/go.mod h1:Ek9pY8RKWXwsWvd3rQiHYtMqkjSUV+s1Rj7j4H5Ur6o=
+golang.org/x/mod v0.42.0 h1:ICeL3t2L30ljSPHHxDnR4fqVcBt/zR2zJAb9juavmQg=
+golang.org/x/mod v0.42.0/go.mod h1:u5fzIrRInkL8MaO1cbRkUHT64v303YIT5KEqpNr7jRs=
 golang.org/x/net v0.0.0-20180724234803-3673e40ba225/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
 golang.org/x/net v0.0.0-20180826012351-8a410e7b638d/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
 golang.org/x/net v0.0.0-20181114220301-adae6a3d119a/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
@@ -724,8 +724,8 @@ golang.org/x/net v0.0.0-20210525063256-abc453219eb5/go.mod h1:9nx3DQGgdP8bBQD5qx
 golang.org/x/net v0.0.0-20211015210444-4f30a5c0130f/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
 golang.org/x/net v0.0.0-20220722155237-a158d28d115b/go.mod h1:XRhObCWvk6IyKnWLug+ECip1KBveYUHfp+8e9klMJ9c=
 golang.org/x/net v0.6.0/go.mod h1:2Tu9+aMcznHK/AK1HMvgo6xiTLG5rD5rZLDS+rp2Bjs=
-golang.org/x/net v0.59.0 h1:5zfYln+w5XCxwrnMMJPufRgNoXEaGxl0wo5GqPXyues=
-golang.org/x/net v0.59.0/go.mod h1:2DA/G1UfVbCpQPeWTmMPGY7Cs2PkBkwu743bVX5PIVg=
+golang.org/x/net v0.61.0 h1:RR+6j/BTBCrggOLugxa+66R8FvoN6MHlzDlZIbN9YFw=
+golang.org/x/net v0.61.0/go.mod h1:WgDCOTH8iwtB66T4/GfI94AHo/hhdb7Hah9NDseQfiI=
 golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
 golang.org/x/oauth2 v0.0.0-20190226205417-e64efc72b421/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
 golang.org/x/oauth2 v0.0.0-20190604053449-0f29369cfe45/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
@@ -745,8 +745,8 @@ golang.org/x/sync v0.0.0-20201207232520-09787c993a3a/go.mod h1:RxMgew5VJxzue5/jJ
 golang.org/x/sync v0.0.0-20210220032951-036812b2e83c/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20220722155255-886fb9371eb4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.1.0/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.23.0 h1:KameEIfc1IkluZyXWLn39Wd4tURc6GbCiISGiZm2bQk=
-golang.org/x/sync v0.23.0/go.mod h1:sUUOizhqBxiL6pEWpqNLUiaJn1ShEbZ6BBqskPbjZm0=
+golang.org/x/sync v0.24.0 h1:iqiDrGoDgyuvAKrhLQthUEzU/R378N78gXbBjxed7oA=
+golang.org/x/sync v0.24.0/go.mod h1:sUUOizhqBxiL6pEWpqNLUiaJn1ShEbZ6BBqskPbjZm0=
 golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20180905080454-ebe1bf3edb33/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20181116152217-5ac8a444bdc5/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
@@ -790,10 +790,10 @@ golang.org/x/sys v0.0.0-20220114195835-da31bd327af9/go.mod h1:oPkhp1MJrh7nUepCBc
 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 v0.5.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.48.0 h1:bbX/i/6MgT9BVLM9RT1thmxL04yeTAhbEz4SyadbXoo=
-golang.org/x/sys v0.48.0/go.mod h1:hNLxWAXmnKAxqDtdwIYC4bM9oQPEecfsnNMuSxOs3og=
-golang.org/x/telemetry v0.0.0-20260924152758-ed294f943157 h1:DftxASHKl6JolFPqoN+2WrFO53JvkHwRXvFqf8gsivA=
-golang.org/x/telemetry v0.0.0-20260924152758-ed294f943157/go.mod h1:i+ivNqjDnTF3WTElsdk5g9V5DTSBYgdNo7xTU9SDwYA=
+golang.org/x/sys v0.49.0 h1:XbzkgYJHdqh/8m2Uu0W/dQv8nktxx4BFHp1M0gROTXA=
+golang.org/x/sys v0.49.0/go.mod h1:hNLxWAXmnKAxqDtdwIYC4bM9oQPEecfsnNMuSxOs3og=
+golang.org/x/telemetry v0.0.0-20261009151841-97e8ea77aa2a h1:wnveMBKHwx8/nIb8WYT5P9vtefdXOvIWxzycCkc5n6Q=
+golang.org/x/telemetry v0.0.0-20261009151841-97e8ea77aa2a/go.mod h1:i+ivNqjDnTF3WTElsdk5g9V5DTSBYgdNo7xTU9SDwYA=
 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 v0.5.0/go.mod h1:jMB1sMXY+tzblOD4FWmEbocvup2/aLOaQEp7JmGp78k=
@@ -857,8 +857,8 @@ golang.org/x/tools v0.1.1-0.20210302220138-2ac05c832e1a/go.mod h1:9bzcO0MWcOuT0t
 golang.org/x/tools v0.1.10/go.mod h1:Uh6Zz+xoGYZom868N8YTex3t7RhtHDBrE8Gzo9bV56E=
 golang.org/x/tools v0.1.12/go.mod h1:hNGJHUnrk76NpqgfD5Aqm5Crs+Hm0VOH/i9J2+nxYbc=
 golang.org/x/tools v0.6.0/go.mod h1:Xwgl3UAJ/d3gWutnCtw505GrjyAbvKui8lOU390QaIU=
-golang.org/x/tools v0.51.0 h1:k4Xc/1Om9jwkBJBo4NVLMSARBoWtK10mx+W5BnXCeAI=
-golang.org/x/tools v0.51.0/go.mod h1:9eEncMayCV6zRMGhR5eZEC2iBx98qWcF1HZ9Z7wJOoA=
+golang.org/x/tools v0.52.1-0.20261030190145-0f433775b36d h1:F5ot9G6RM6pgx3TTdKF6IZa2nocit4tcsrVQlq4Zh5w=
+golang.org/x/tools v0.52.1-0.20261030190145-0f433775b36d/go.mod h1:ItcP5UNaVGG+Yc4eIYmaFLNJ4soEopkx5Y2d4lcEurs=
 golang.org/x/tools/go/expect v0.1.1-deprecated h1:jpBZDwmgPhXsKZC6WhL20P4b/wmnpsEAGHaNy0n/rJM=
 golang.org/x/tools/go/expect v0.1.1-deprecated/go.mod h1:eihoPOH+FgIqa3FpoTwguz/bVUSGBlGQU67vpBeOrBY=
 golang.org/x/tools/go/packages/packagestest v0.1.1-deprecated h1:1h2MnaIAIXISqTFKdENegdpAgUXz6NrPEsbIeWaBRvM=

testdata/toolexec_buildmode_pie.txt

@@ -0,0 +1,64 @@
+# The go command compiles a position-independent executable with -shared. It is the
+# default build mode on darwin, ios, android and windows, where the nested simd build
+# is one too, so a default build is rewritten there. Asking for one where it is not
+# the default (-buildmode=pie on linux) still leaves the package alone: the nested
+# build would not match it.
+
+# Reference: the scalar program.
+exec go run .
+cmp stdout want.out
+
+env GOEXPERIMENT=simd
+env GOCACHE=$WORK/gocache
+env LOOPVEC_TOOLEXEC_LOG=$WORK/log
+env LOOPVEC_TOOLEXEC_LOG_PKGS=example.com/app/vec
+
+[GOOS:darwin] exec go build -toolexec=$LOOPVEC_TOOLEXEC -o app .
+[GOOS:darwin] cmp log rewrote.want
+
+[GOOS:linux] env LOOPVEC_TOOLEXEC_DEBUG=1
+[GOOS:linux] exec go build -buildmode=pie -toolexec=$LOOPVEC_TOOLEXEC -o app .
+[GOOS:linux] cmp log skipped.want
+
+[GOOS:darwin] exec ./app
+[GOOS:darwin] cmp stdout want.out
+[GOOS:linux] exec ./app
+[GOOS:linux] cmp stdout want.out
+
+-- go.mod --
+module example.com/app
+
+go 1.27.1
+-- vec/vec.go --
+package vec
+
+func AddInt32s(dst, a, b []int32) {
+	for i := range dst {
+		dst[i] = a[i] + b[i]
+	}
+}
+-- main.go --
+package main
+
+import (
+	"fmt"
+
+	"example.com/app/vec"
+)
+
+func main() {
+	dst := make([]int32, 21)
+	a := make([]int32, 21)
+	b := make([]int32, 21)
+	for i := range a {
+		a[i], b[i] = int32(i), int32(2*i)
+	}
+	vec.AddInt32s(dst, a, b)
+	fmt.Println(dst)
+}
+-- want.out --
+[0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60]
+-- rewrote.want --
+rewrote 1 loop(s) in example.com/app/vec
+-- skipped.want --
+skipped example.com/app/vec: simd unavailable or unsupported build mode

toolexec.md

@@ -28,7 +28,9 @@ Packages are left alone when:
   `os`, `strconv`, `sync`, `sync/atomic`, `math`, and more). Rewritten code
   imports `simd`, so rewriting these would create an import cycle.
 - They use cgo, or the build uses `-race`, `-msan`, `-asan`, `-shared`,
-  `-dynlink` or `-trimpath`.
+  `-dynlink` or `-trimpath`. `-shared` is accepted where a position-independent
+  executable is the default build (darwin, ios, android, windows), since the
+  nested simd build is one there too.
 - They are cryptographic: `crypto/...` and `golang.org/x/crypto/...`, vendored
   copies included. Constant-time code is written so that its running time does
   not depend on secret data, and a rewrite adds a runtime overlap check, a length