Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

unique: Fatal errors (found bad pointer in Go heap, found pointer to free object) and memory corruption #69643

Closed
connorszczepaniak-wk opened this issue Sep 26, 2024 · 6 comments
Labels
compiler/runtime Issues related to the Go compiler and/or runtime. NeedsInvestigation Someone must examine and confirm this is a valid issue and not a duplicate of an existing one. WaitingForInfo Issue is not actionable because of missing required information, which needs to be provided.
Milestone

Comments

@connorszczepaniak-wk
Copy link

connorszczepaniak-wk commented Sep 26, 2024

Go version

go version go1.23.1 darwin/arm64

Output of go env in your module/workspace:

GO111MODULE=''
GOARCH='amd64'
GOBIN=''
GOCACHE='/root/.cache/go-build'
GOENV='/root/.config/go/env'
GOEXE=''
GOEXPERIMENT=''
GOFLAGS=''
GOHOSTARCH='amd64'
GOHOSTOS='linux'
GOINSECURE=''
GOMODCACHE='/go/pkg/mod'
GONOPROXY=''
GONOSUMDB=''
GOOS='linux'
GOPATH='/go'
GOPRIVATE=''
GOPROXY='https://proxy.golang.org,direct'
GOROOT='/usr/local/go'
GOSUMDB='sum.golang.org'
GOTMPDIR=''
GOTOOLCHAIN='local'
GOTOOLDIR='/usr/local/go/pkg/tool/linux_amd64'
GOVCS=''
GOVERSION='go1.23.1'
GODEBUG=''
GOTELEMETRY='local'
GOTELEMETRYDIR='/root/.config/go/telemetry'
GCCGO='gccgo'
GOAMD64='v1'
AR='ar'
CC='gcc'
CXX='g++'
CGO_ENABLED='1'
GOMOD='/go/src/github.com/my/repo/go.mod'
GOWORK=''
CGO_CFLAGS='-O2 -g'
CGO_CPPFLAGS=''
CGO_CXXFLAGS='-O2 -g'
CGO_FFLAGS='-O2 -g'
CGO_LDFLAGS='-O2 -g'
PKG_CONFIG='pkg-config'
GOGCCFLAGS='-fPIC -m64 -pthread -Wl,--no-gc-sections -fmessage-length=0 -ffile-prefix-map=/tmp/go-build3070165830=/tmp/go-build -gno-record-gcc-switches'

What did you do?

In our deployed application, we started using the unique package to intern some strings. It's hard to provide a complete example because of the size of our application and the potential usage graph of the unique.Handles would be difficult to compute. We also haven't been able to repro with a more trivial example yet. The gist of it it this:

type resource struct {
	target unique.Handle[string]
}

type identifierKindA unique.Handle[string]

type identifierKindB string

func (r *resource) getIDA() identifierKindA {
	return identifierKindA(r.target)
}

func (r *resource) getIDB() identifierKindB {
	return identifierKindB(r.target.Value())
}

getIDA and getIDB are getting called in many disparate places, and potentially across many goroutines if that's important.

What did you see happen?

We had a few different fatal errors occur. We also saw memory corruption; getting the value from the handle seems to have returned a string from elsewhere in the program that could never have been an input when creating this particular type.

fatal error: found pointer to free object

Stack Trace
goroutine 15 gp=0xc000d80540 m=13 mp=0xc000300e08 [running]:
runtime.throw({0x556fda3?, 0xc0de54d710?})
	/usr/local/go/src/runtime/panic.go:1067 +0x48 fp=0xc04d702b60 sp=0xc04d702b30 pc=0x473e08
runtime.(*mspan).reportZombies(0x7fa5c5a60688)
	/usr/local/go/src/runtime/mgcsweep.go:890 +0x2ea fp=0xc04d702be0 sp=0xc04d702b60 pc=0x428aea
runtime.(*sweepLocked).sweep(0x56cb850?, 0x0)
	/usr/local/go/src/runtime/mgcsweep.go:658 +0xb54 fp=0xc04d702d00 sp=0xc04d702be0 pc=0x428134
runtime.(*mspan).ensureSwept(0xc0d53cf3c8?)
	/usr/local/go/src/runtime/mgcsweep.go:474 +0xc5 fp=0xc04d702d38 sp=0xc04d702d00 pc=0x427565
internal/weak.runtime_makeStrongFromWeak(0xc0ddeb1f30)
	/usr/local/go/src/runtime/mheap.go:2069 +0xa9 fp=0xc04d702d58 sp=0xc04d702d38 pc=0x472a69
internal/weak.Pointer[...].Strong(...)
	/usr/local/go/src/internal/weak/pointer.go:74
unique.addUniqueMap[...].func1.1({0xc17f6cab40})
	/usr/local/go/src/unique/handle.go:130 +0x39 fp=0xc04d702da8 sp=0xc04d702d58 pc=0x167de99
internal/concurrent.(*HashTrieMap[...]).iter(0x60824c0, 0xc0d53cf360, 0xc04d702f48)
	/usr/local/go/src/internal/concurrent/hashtriemap.go:298 +0xe3 fp=0xc04d702de8 sp=0xc04d702da8 pc=0x167c163
internal/concurrent.(*HashTrieMap[...]).iter(0x60824c0, 0xc041930280, 0xc04d702f48)
	/usr/local/go/src/internal/concurrent/hashtriemap.go:291 +0x65 fp=0xc04d702e28 sp=0xc04d702de8 pc=0x167c0e5
internal/concurrent.(*HashTrieMap[...]).iter(0x60824c0, 0xc03f937860, 0xc04d702f48)
	/usr/local/go/src/internal/concurrent/hashtriemap.go:291 +0x65 fp=0xc04d702e68 sp=0xc04d702e28 pc=0x167c0e5
internal/concurrent.(*HashTrieMap[...]).iter(0x60824c0, 0xc048c73a40, 0xc04d702f48)
	/usr/local/go/src/internal/concurrent/hashtriemap.go:291 +0x65 fp=0xc04d702ea8 sp=0xc04d702e68 pc=0x167c0e5
internal/concurrent.(*HashTrieMap[...]).iter(0x60824c0, 0xc03fe74be0, 0xc04d702f48)
	/usr/local/go/src/internal/concurrent/hashtriemap.go:291 +0x65 fp=0xc04d702ee8 sp=0xc04d702ea8 pc=0x167c0e5
internal/concurrent.(*HashTrieMap[...]).iter(0x60824c0, 0xc040b94f00, 0xc04d702f48)
	/usr/local/go/src/internal/concurrent/hashtriemap.go:291 +0x65 fp=0xc04d702f28 sp=0xc04d702ee8 pc=0x167c0e5
unique.addUniqueMap[...]).All.2(...)
	/usr/local/go/src/internal/concurrent/hashtriemap.go:280
unique.addUniqueMap[...].func1()
	/usr/local/go/src/unique/handle.go:129 +0x45 fp=0xc04d702f70 sp=0xc04d702f28 pc=0x167de45
unique.registerCleanup.func1()
	/usr/local/go/src/unique/handle.go:157 +0xd0 fp=0xc04d702fb8 sp=0xc04d702f70 pc=0x5cb170
runtime.unique_runtime_registerUniqueMapCleanup.func1(...)
	/usr/local/go/src/runtime/mgc.go:1733
runtime.unique_runtime_registerUniqueMapCleanup.gowrap1()
	/usr/local/go/src/runtime/mgc.go:1735 +0x39 fp=0xc04d702fe0 sp=0xc04d702fb8 pc=0x41e7b9
runtime.goexit({})
	/usr/local/go/src/runtime/asm_amd64.s:1700 +0x1 fp=0xc04d702fe8 sp=0xc04d702fe0 pc=0x47c461
created by unique.runtime_registerUniqueMapCleanup in goroutine 1
	/usr/local/go/src/runtime/mgc.go:1730 +0x96
[originating from goroutine 1]:
runtime.systemstack_switch(...)
	/usr/local/go/src/runtime/asm_amd64.s:479 +0x8
runtime.newproc(...)
	/usr/local/go/src/runtime/proc.go:4977 +0x3f
unique.runtime_registerUniqueMapCleanup(...)
	/usr/local/go/src/runtime/mgc.go:1736 +0x96
unique.registerCleanup(...)
	/usr/local/go/src/unique/handle.go:169 +0x1a
sync.(*Once).doSlow(...)
	/usr/local/go/src/sync/once.go:75 +0xb4
sync.(*Once).Do(...)
	/usr/local/go/src/sync/once.go:69 +0x19
unique.Make[...](...)
	/usr/local/go/src/unique/handle.go:40 +0x8d
net/netip.init(...)
	/usr/local/go/src/net/netip/netip.go:70 +0x25
runtime.doInit1(...)
	/usr/local/go/src/runtime/proc.go:7287 +0xe8
runtime.doInit(...)
	/usr/local/go/src/runtime/proc.go:7256 +0x345
runtime.main(...)
	/usr/local/go/src/runtime/proc.go:254 +0x22e
runtime.goexit(...)
	/usr/local/go/src/runtime/asm_amd64.s:1701 +0x1

fatal error: found bad pointer in Go heap (incorrect use of unsafe or cgo?)

Stack Trace
runtime.throw({0x5656b0e?, 0x6?})
	/usr/local/go/src/runtime/panic.go:1067 +0x48 fp=0xc00021fdf8 sp=0xc00021fdc8 pc=0x473e08
runtime.badPointer(0x7f0b918ec6e8, 0xc1683e6f00, 0xc53fe46000, 0x86f8)
	/usr/local/go/src/runtime/mbitmap.go:1247 +0x165 fp=0xc00021fe48 sp=0xc00021fdf8 pc=0x4162e5
runtime.findObject(0xc686822bd0?, 0xc1fc048a20?, 0x1?)
	/usr/local/go/src/runtime/mbitmap.go:1299 +0xa6 fp=0xc00021fe80 sp=0xc00021fe48 pc=0x472526
runtime.scanobject(0xc0000d1750?, 0xc0000d1750)
	/usr/local/go/src/runtime/mgcmark.go:1464 +0x14c fp=0xc00021ff10 sp=0xc00021fe80 pc=0x42204c
runtime.gcDrain(0xc0000d1750, 0x2)
	/usr/local/go/src/runtime/mgcmark.go:1230 +0x1f4 fp=0xc00021ff78 sp=0xc00021ff10 pc=0x421914
runtime.gcDrainMarkWorkerDedicated(...)
	/usr/local/go/src/runtime/mgcmark.go:1112
runtime.gcBgMarkWorker.func2()
	/usr/local/go/src/runtime/mgc.go:1455 +0x14a fp=0xc00021ffc8 sp=0xc00021ff78 pc=0x41dfca
runtime.systemstack(0x0)
	/usr/local/go/src/runtime/asm_amd64.s:514 +0x4a fp=0xc00021ffd8 sp=0xc00021ffc8 pc=0x47a62a

What did you expect to see?

No fatal errors and no memory corruption.

@connorszczepaniak-wk
Copy link
Author

connorszczepaniak-wk commented Sep 26, 2024

It looks like this issue might have been the same: #69210

That seems like maybe there will be a fix in 1.23.2. Can someone confirm that this issue could have caused the memory corruption that we were seeing?

@MikeMitchellWebDev
Copy link
Contributor

@connorszczepaniak-wk you can apply the fix now and test it out yourself https://go-review.googlesource.com/c/go/+/610696

@mknyszek
Copy link
Contributor

@connorszczepaniak-wk Yeah, that's almost certainly it. And indeed, it's fixed at tip and the fix is already on the Go 1.23 release branch. It will be in the next minor release.

It may be worth giving the patch a try as @MikeMitchellWebDev suggests, to confirm it does resolve the issue for you. Leaving this issue open for that reason, for now.

Apologies for the breakage.

@mknyszek mknyszek added NeedsInvestigation Someone must examine and confirm this is a valid issue and not a duplicate of an existing one. compiler/runtime Issues related to the Go compiler and/or runtime. WaitingForInfo Issue is not actionable because of missing required information, which needs to be provided. labels Sep 27, 2024
@mknyszek mknyszek added this to the Go1.24 milestone Sep 27, 2024
@connorszczepaniak-wk
Copy link
Author

Thanks for the quick reply; I think this may be a bit tricky for us to test out ahead of it being included in a proper patch release (it's a bit unclear how I'd apply a patch to the stdlib that we use to build in a docker image without making some significant changes to our build process), but we could test it once 1.23.2 is out to confirm that we don't see the same issue.

@seankhliao seankhliao added WaitingForInfo Issue is not actionable because of missing required information, which needs to be provided. and removed WaitingForInfo Issue is not actionable because of missing required information, which needs to be provided. labels Oct 1, 2024
@gopherbot
Copy link
Contributor

Timed out in state WaitingForInfo. Closing.

(I am just a bot, though. Please speak up if this is a mistake or you have the requested information.)

@gopherbot gopherbot closed this as not planned Won't fix, can't repro, duplicate, stale Nov 1, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
compiler/runtime Issues related to the Go compiler and/or runtime. NeedsInvestigation Someone must examine and confirm this is a valid issue and not a duplicate of an existing one. WaitingForInfo Issue is not actionable because of missing required information, which needs to be provided.
Projects
Development

No branches or pull requests

6 participants