justan0th3rstud3nt
/
garble
mirror of https://github.com/burrowers/garble
1
0
Fork 0
Code Issues 1 Projects Releases Wiki Activity
You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
master
release-v0.12
ci-test
literalRecheck
v0.12.1
v0.12.0
v0.11.0
v0.10.1
v0.10.0
v0.9.3
v0.9.2
v0.9.1
v0.9.0
v0.8.0
v0.7.2
v0.7.1
v0.7.0
v0.6.0
v0.5.1
v0.5.0
v0.4.0
v0.3.0
v0.2.0
v0.1.0
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'c41f026cd6'
${ noResults }
garble/go_std_tables.go

203 lines
8.0 KiB
Go
Raw Normal View History Unescape Escape

generate go_std_tables.go in its entirety No more having to manually run the script and adapting it to Go code.
1 year ago
// Code generated by scripts/gen-go-std-tables.sh; DO NOT EDIT.
drop Go 1.21 and start using go/version Needing to awkwardly treat Go versions as if they were semver is no longer necessary thanks to go/version being in Go 1.22.0 now.
3 months ago
// Generated from Go version go1.22.0.
generate go_std_tables.go in its entirety No more having to manually run the script and adapting it to Go code.
1 year ago
split runtimeAndDeps and runtimeLinknamed into a separate Go file The next commit will start generating these via //go:generate, so this first change keeps the diffs easier to review.
1 year ago
package main
var runtimeAndDeps = map[string]bool{
"internal/goarch": true,
"unsafe": true,
"internal/abi": true,
"internal/cpu": true,
"internal/bytealg": true,
fix support for go1.22rc1 In early December, a new internal package linknamed from runtime was introduced, internal/chacha8rand. Re-generate the tables. Note that due to the same group of CLs and refactors, math/rand and net are no longer linknamed from runtime in Go 1.22. They are still in Go 1.21, so keep those entries around for now. We can remove math/rand/v2, as it doesn't yet exist in 1.21. Fixes #820.
5 months ago
"internal/chacha8rand": true,
generate go_std_tables.go in its entirety No more having to manually run the script and adapting it to Go code.
1 year ago
"internal/coverage/rtcov": true,
update to Go tip from April 28th A couple of new packages in runtimeAndDeps, and go list's Package.DepsErrors may now include package build errors which we want to ignore as we would print those as duplicates.
1 year ago
"internal/godebugs": true,
split runtimeAndDeps and runtimeLinknamed into a separate Go file The next commit will start generating these via //go:generate, so this first change keeps the diffs easier to review.
1 year ago
"internal/goexperiment": true,
"internal/goos": true,
"runtime/internal/atomic": true,
"runtime/internal/math": true,
"runtime/internal/sys": true,
"runtime/internal/syscall": true,
"runtime": true,
}
var runtimeLinknamed = []string{
"arena",
"crypto/internal/boring",
"crypto/internal/boring/bcache",
"crypto/internal/boring/fipstls",
"crypto/x509/internal/macos",
"internal/godebug",
"internal/poll",
"internal/reflectlite",
"internal/syscall/unix",
re-generate go_std_tables.go with Go master Two new packages linknamed with the runtime package, one new intrinsic function, and one that is being removed in Go 1.22 but we want to keep around as long as we support Go 1.21. Also note that, since math/rand/v2 simply does not exist until Go 1.22, we need to adjust appendListedPackages to not fail on older versions.
6 months ago
"internal/syscall/windows",
update to Go tip from April 28th A couple of new packages in runtimeAndDeps, and go list's Package.DepsErrors may now include package build errors which we want to ignore as we would print those as duplicates.
1 year ago
"maps",
split runtimeAndDeps and runtimeLinknamed into a separate Go file The next commit will start generating these via //go:generate, so this first change keeps the diffs easier to review.
1 year ago
"os",
"os/signal",
"plugin",
"reflect",
"runtime/coverage",
"runtime/debug",
"runtime/metrics",
"runtime/pprof",
"runtime/trace",
"sync",
"sync/atomic",
"syscall",
"syscall/js",
"time",
drop Go 1.21 and start using go/version Needing to awkwardly treat Go versions as if they were semver is no longer necessary thanks to go/version being in Go 1.22.0 now.
3 months ago
// The net package linknames to the runtime, not the other way around.
// TODO: support this automatically via our script.
fix support for go1.22rc1 In early December, a new internal package linknamed from runtime was introduced, internal/chacha8rand. Re-generate the tables. Note that due to the same group of CLs and refactors, math/rand and net are no longer linknamed from runtime in Go 1.22. They are still in Go 1.21, so keep those entries around for now. We can remove math/rand/v2, as it doesn't yet exist in 1.21. Fixes #820.
5 months ago
"net",
split runtimeAndDeps and runtimeLinknamed into a separate Go file The next commit will start generating these via //go:generate, so this first change keeps the diffs easier to review.
1 year ago
}
avoid breaking intrinsics when obfuscating names We obfuscate import paths as well as their declared names. The compiler treats some packages and APIs in special ways, and the way it detects those is by looking at import paths and names. In the past, we have avoided obfuscating some names like embed.FS or reflect.Value.MethodByName for this reason. Otherwise, go:embed or the linker's deadcode elimination might be broken. This matching by path and name also happens with compiler intrinsics. Intrinsics allow the compiler to rewrite some standard library calls with small and efficient assembly, depending on the target GOARCH. For example, math/bits.TrailingZeros32 gets replaced with ssa.OpCtz32, which on amd64 may result in using the TZCNTL instruction. We never noticed that we were breaking many of these intrinsics. The intrinsics for funcs declared in the runtime and its dependencies still worked properly, as we do not obfuscate those packages yet. However, for other packages like math/bits and sync/atomic, the intrinsics were being entirely disabled due to obfuscated names. Skipping intrinsics is particularly bad for performance, and it also leads to slightly larger binaries: │ old │ new │ │ bin-B │ bin-B vs base │ Build-16 5.450Mi ± ∞ ¹ 5.333Mi ± ∞ ¹ -2.15% (p=0.029 n=4) Finally, the main reason we noticed that intrinsics were broken is that apparently GOARCH=mips fails to link without them, as some symbols end up being not defined at all. This patch fixes builds for the MIPS family of architectures. Rather than building and linking all of std for every GOARCH, test that intrinsics work by asking the compiler to print which intrinsics are being applied, and checking that math/bits gets them. This fix is relatively unfortunate, as it means we stop obfuscating about 120 function names and a handful of package paths. However, fixing builds and intrinsics is much more important. We can figure out better ways to deal with intrinsics in the future. Fixes #646.
1 year ago
var compilerIntrinsicsPkgs = map[string]bool{
"math": true,
"math/big": true,
"math/bits": true,
"runtime": true,
"runtime/internal/atomic": true,
"runtime/internal/math": true,
"runtime/internal/sys": true,
"sync": true,
"sync/atomic": true,
}
var compilerIntrinsicsFuncs = map[string]bool{
"math.Abs": true,
"math/big.mulWW": true,
"math/bits.Add": true,
"math/bits.Add64": true,
"math/bits.Div": true,
"math/bits.Div64": true,
"math/bits.Len": true,
"math/bits.Len16": true,
"math/bits.Len32": true,
"math/bits.Len64": true,
"math/bits.Len8": true,
"math/bits.Mul": true,
"math/bits.Mul64": true,
"math/bits.OnesCount": true,
"math/bits.OnesCount16": true,
"math/bits.OnesCount32": true,
"math/bits.OnesCount64": true,
"math/bits.OnesCount8": true,
"math/bits.Reverse": true,
"math/bits.Reverse16": true,
"math/bits.Reverse32": true,
"math/bits.Reverse64": true,
"math/bits.Reverse8": true,
CI: test against a newer Go tip While here, update staticcheck.
1 year ago
"math/bits.ReverseBytes16": true,
avoid breaking intrinsics when obfuscating names We obfuscate import paths as well as their declared names. The compiler treats some packages and APIs in special ways, and the way it detects those is by looking at import paths and names. In the past, we have avoided obfuscating some names like embed.FS or reflect.Value.MethodByName for this reason. Otherwise, go:embed or the linker's deadcode elimination might be broken. This matching by path and name also happens with compiler intrinsics. Intrinsics allow the compiler to rewrite some standard library calls with small and efficient assembly, depending on the target GOARCH. For example, math/bits.TrailingZeros32 gets replaced with ssa.OpCtz32, which on amd64 may result in using the TZCNTL instruction. We never noticed that we were breaking many of these intrinsics. The intrinsics for funcs declared in the runtime and its dependencies still worked properly, as we do not obfuscate those packages yet. However, for other packages like math/bits and sync/atomic, the intrinsics were being entirely disabled due to obfuscated names. Skipping intrinsics is particularly bad for performance, and it also leads to slightly larger binaries: │ old │ new │ │ bin-B │ bin-B vs base │ Build-16 5.450Mi ± ∞ ¹ 5.333Mi ± ∞ ¹ -2.15% (p=0.029 n=4) Finally, the main reason we noticed that intrinsics were broken is that apparently GOARCH=mips fails to link without them, as some symbols end up being not defined at all. This patch fixes builds for the MIPS family of architectures. Rather than building and linking all of std for every GOARCH, test that intrinsics work by asking the compiler to print which intrinsics are being applied, and checking that math/bits gets them. This fix is relatively unfortunate, as it means we stop obfuscating about 120 function names and a handful of package paths. However, fixing builds and intrinsics is much more important. We can figure out better ways to deal with intrinsics in the future. Fixes #646.
1 year ago
"math/bits.ReverseBytes32": true,
"math/bits.ReverseBytes64": true,
"math/bits.RotateLeft": true,
"math/bits.RotateLeft16": true,
"math/bits.RotateLeft32": true,
"math/bits.RotateLeft64": true,
"math/bits.RotateLeft8": true,
"math/bits.Sub": true,
"math/bits.Sub64": true,
"math/bits.TrailingZeros16": true,
"math/bits.TrailingZeros32": true,
"math/bits.TrailingZeros64": true,
"math/bits.TrailingZeros8": true,
"math.Ceil": true,
"math.Copysign": true,
"math.Floor": true,
"math.FMA": true,
"math.Round": true,
"math.RoundToEven": true,
"math.sqrt": true,
"math.Trunc": true,
"runtime/internal/atomic.And": true,
"runtime/internal/atomic.And8": true,
"runtime/internal/atomic.Cas": true,
"runtime/internal/atomic.Cas64": true,
"runtime/internal/atomic.Casint32": true,
"runtime/internal/atomic.Casint64": true,
"runtime/internal/atomic.Casp1": true,
"runtime/internal/atomic.CasRel": true,
"runtime/internal/atomic.Casuintptr": true,
"runtime/internal/atomic.Load": true,
"runtime/internal/atomic.Load64": true,
"runtime/internal/atomic.Load8": true,
"runtime/internal/atomic.LoadAcq": true,
"runtime/internal/atomic.LoadAcq64": true,
"runtime/internal/atomic.LoadAcquintptr": true,
"runtime/internal/atomic.Loadint32": true,
"runtime/internal/atomic.Loadint64": true,
"runtime/internal/atomic.Loadp": true,
"runtime/internal/atomic.Loaduint": true,
"runtime/internal/atomic.Loaduintptr": true,
"runtime/internal/atomic.Or": true,
"runtime/internal/atomic.Or8": true,
"runtime/internal/atomic.Store": true,
"runtime/internal/atomic.Store64": true,
"runtime/internal/atomic.Store8": true,
"runtime/internal/atomic.Storeint32": true,
"runtime/internal/atomic.Storeint64": true,
"runtime/internal/atomic.StorepNoWB": true,
"runtime/internal/atomic.StoreRel": true,
"runtime/internal/atomic.StoreRel64": true,
"runtime/internal/atomic.StoreReluintptr": true,
"runtime/internal/atomic.Storeuintptr": true,
"runtime/internal/atomic.Xadd": true,
"runtime/internal/atomic.Xadd64": true,
"runtime/internal/atomic.Xaddint32": true,
"runtime/internal/atomic.Xaddint64": true,
"runtime/internal/atomic.Xadduintptr": true,
"runtime/internal/atomic.Xchg": true,
"runtime/internal/atomic.Xchg64": true,
"runtime/internal/atomic.Xchgint32": true,
"runtime/internal/atomic.Xchgint64": true,
"runtime/internal/atomic.Xchguintptr": true,
re-generate go_std_tables.go with Go master Two new packages linknamed with the runtime package, one new intrinsic function, and one that is being removed in Go 1.22 but we want to keep around as long as we support Go 1.21. Also note that, since math/rand/v2 simply does not exist until Go 1.22, we need to adjust appendListedPackages to not fail on older versions.
6 months ago
"runtime/internal/math.Add64": true,
avoid breaking intrinsics when obfuscating names We obfuscate import paths as well as their declared names. The compiler treats some packages and APIs in special ways, and the way it detects those is by looking at import paths and names. In the past, we have avoided obfuscating some names like embed.FS or reflect.Value.MethodByName for this reason. Otherwise, go:embed or the linker's deadcode elimination might be broken. This matching by path and name also happens with compiler intrinsics. Intrinsics allow the compiler to rewrite some standard library calls with small and efficient assembly, depending on the target GOARCH. For example, math/bits.TrailingZeros32 gets replaced with ssa.OpCtz32, which on amd64 may result in using the TZCNTL instruction. We never noticed that we were breaking many of these intrinsics. The intrinsics for funcs declared in the runtime and its dependencies still worked properly, as we do not obfuscate those packages yet. However, for other packages like math/bits and sync/atomic, the intrinsics were being entirely disabled due to obfuscated names. Skipping intrinsics is particularly bad for performance, and it also leads to slightly larger binaries: │ old │ new │ │ bin-B │ bin-B vs base │ Build-16 5.450Mi ± ∞ ¹ 5.333Mi ± ∞ ¹ -2.15% (p=0.029 n=4) Finally, the main reason we noticed that intrinsics were broken is that apparently GOARCH=mips fails to link without them, as some symbols end up being not defined at all. This patch fixes builds for the MIPS family of architectures. Rather than building and linking all of std for every GOARCH, test that intrinsics work by asking the compiler to print which intrinsics are being applied, and checking that math/bits gets them. This fix is relatively unfortunate, as it means we stop obfuscating about 120 function names and a handful of package paths. However, fixing builds and intrinsics is much more important. We can figure out better ways to deal with intrinsics in the future. Fixes #646.
1 year ago
"runtime/internal/math.Mul64": true,
"runtime/internal/math.MulUintptr": true,
"runtime/internal/sys.Bswap32": true,
"runtime/internal/sys.Bswap64": true,
"runtime/internal/sys.Len64": true,
"runtime/internal/sys.Len8": true,
"runtime/internal/sys.OnesCount64": true,
"runtime/internal/sys.Prefetch": true,
"runtime/internal/sys.PrefetchStreamed": true,
"runtime/internal/sys.TrailingZeros32": true,
"runtime/internal/sys.TrailingZeros64": true,
"runtime/internal/sys.TrailingZeros8": true,
"runtime.publicationBarrier": true,
"sync/atomic.AddInt32": true,
"sync/atomic.AddInt64": true,
"sync/atomic.AddUint32": true,
"sync/atomic.AddUint64": true,
"sync/atomic.AddUintptr": true,
"sync/atomic.CompareAndSwapInt32": true,
"sync/atomic.CompareAndSwapInt64": true,
"sync/atomic.CompareAndSwapUint32": true,
"sync/atomic.CompareAndSwapUint64": true,
"sync/atomic.CompareAndSwapUintptr": true,
"sync/atomic.LoadInt32": true,
"sync/atomic.LoadInt64": true,
"sync/atomic.LoadPointer": true,
"sync/atomic.LoadUint32": true,
"sync/atomic.LoadUint64": true,
"sync/atomic.LoadUintptr": true,
"sync/atomic.StoreInt32": true,
"sync/atomic.StoreInt64": true,
"sync/atomic.StoreUint32": true,
"sync/atomic.StoreUint64": true,
"sync/atomic.StoreUintptr": true,
"sync/atomic.SwapInt32": true,
"sync/atomic.SwapInt64": true,
"sync/atomic.SwapUint32": true,
"sync/atomic.SwapUint64": true,
"sync/atomic.SwapUintptr": true,
"sync.runtime_LoadAcquintptr": true,
"sync.runtime_StoreReluintptr": true,
}
rework reflection detection with ssa (#732) This is significantly more robust, than the ast based detection and can record very complex cases of indirect parameter reflection. Fixes #554
1 year ago
var reflectSkipPkg = map[string]bool{
"fmt": true,
}