go: testing: go 1.17 regression: deadlock in test suite

What version of Go are you using (go version)?

$ go version
go version go1.17 linux/amd64

Does this issue reproduce with the latest release?

yes

What operating system and processor architecture are you using (go env)?

go env Output
$ go env
GO111MODULE=""
GOARCH="amd64"
GOBIN=""
GOCACHE="/home/tbonfort/.cache/go-build"
GOENV="/home/tbonfort/.config/go/env"
GOEXE=""
GOEXPERIMENT=""
GOFLAGS=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GOINSECURE=""
GOMODCACHE="/home/tbonfort/go/pkg/mod"
GOOS="linux"
GOPATH="/home/tbonfort/go"
GOPROXY="https://proxy.golang.org,direct"
GOROOT="/usr/local/go"
GOSUMDB="sum.golang.org"
GOTMPDIR=""
GOTOOLDIR="/usr/local/go/pkg/tool/linux_amd64"
GOVCS=""
GOVERSION="go1.17"
GCCGO="gccgo"
AR="ar"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
CGO_CFLAGS="-g -O2"
CGO_CPPFLAGS=""
CGO_CXXFLAGS="-g -O2"
CGO_FFLAGS="-g -O2"
CGO_LDFLAGS="-g -O2"
PKG_CONFIG="pkg-config"
GOGCCFLAGS="-fPIC -m64 -pthread -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build2515039918=/tmp/go-build -gno-record-gcc-switches"

What did you do?

My cgo testsuite deadlocks, commenting out a seemingly random cgo call from 2 tests before the deadlock happens fixes the issue.

Failing test run:

...snip...
=== RUN   TestOpen
--- PASS: TestOpen (0.00s)
=== RUN   TestOpenUpdate
--- PASS: TestOpenUpdate (0.00s)
=== RUN   TestClosingErrors
--- PASS: TestClosingErrors (0.00s)
=== RUN   TestOpenShared
--- PASS: TestOpenShared (0.00s)
=== RUN   TestRegister
--- PASS: TestRegister (0.00s)
^Csignal: interrupt
FAIL	github.com/airbusgeo/godal	9.929s

Commenting out this line in the TestOpenShared test fixes the deadlock: https://github.com/airbusgeo/godal/blob/00980a3df723677b2d949a10eedd78ab6c5d6aee/godal_test.go#L1242

Here is a gdb stacktrace when deadlocked:

sudo gdb ~/dev/godal/godal.test 1904477
GNU gdb (Ubuntu 10.1-2ubuntu2) 10.1.90.20210411-git
Copyright (C) 2021 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /home/tbonfort/dev/godal/godal.test...
Attaching to program: /home/tbonfort/dev/godal/godal.test, process 1904477
[New LWP 1904478]
[New LWP 1904479]
[New LWP 1904480]
[New LWP 1904481]
[New LWP 1904482]
Error while reading shared library symbols for /lib/x86_64-linux-gnu/libpthread.so.0:
Cannot find user-level thread for LWP 1904482: generic error
runtime.futex () at /usr/local/go/src/runtime/sys_linux_amd64.s:520
520		MOVL	AX, ret+40(FP)
Loading Go Runtime support.
(gdb) bt
#0  runtime.futex () at /usr/local/go/src/runtime/sys_linux_amd64.s:520
#1  0x000000000043b6f6 in runtime.futexsleep (addr=0xfffffffffffffe00, val=2, ns=4670691) at /usr/local/go/src/runtime/os_linux.go:44
#2  0x00000000004155f9 in runtime.lock2 (l=0xfffffffffffffe00) at /usr/local/go/src/runtime/lock_futex.go:107
#3  0x0000000000410318 in runtime.lockWithRank (l=0x4744e3 <runtime.futex+35>, rank=<optimized out>)
    at /usr/local/go/src/runtime/lockrank_off.go:23
#4  runtime.lock (l=0x4744e3 <runtime.futex+35>) at /usr/local/go/src/runtime/lock_futex.go:48
#5  runtime.chanrecv (c=0xc0001cd340, ep=0xc00025fb67, block=true) at /usr/local/go/src/runtime/chan.go:508
#6  0x0000000000410198 in runtime.chanrecv1 (c=0xfffffffffffffe00, elem=0x2) at /usr/local/go/src/runtime/chan.go:439
#7  0x0000000000513d95 in testing.(*T).Run (t=0xc0003b41a0, name=..., f={void (testing.T *)} 0xc00025fbd8)
    at /usr/local/go/src/testing/testing.go:1307
#8  0x00000000005159ae in testing.runTests.func1 (t=0xc000283040) at /usr/local/go/src/testing/testing.go:1598
#9  0x0000000000513142 in testing.tRunner (t=0xc000283040, fn={void (testing.T *)} 0xc00025fc68)
    at /usr/local/go/src/testing/testing.go:1259
#10 0x000000000051585f in testing.runTests (matchString={void (string, string, bool *, error *)} 0xc00025fd10, 
    tests=[]testing.InternalTest = {...}, deadline=...) at /usr/local/go/src/testing/testing.go:1596
#11 0x00000000005145bd in testing.(*M).Run (m=0xc00011c800, code=0) at /usr/local/go/src/testing/testing.go:1504
#12 0x00000000009ca56b in main.main () at _testmain.go:177
(gdb) thread apply all bt

Thread 6 (LWP 1904482 "godal.test"):
#0  runtime.futex () at /usr/local/go/src/runtime/sys_linux_amd64.s:520
#1  0x000000000043b6f6 in runtime.futexsleep (addr=0xfffffffffffffe00, val=0, ns=4670691) at /usr/local/go/src/runtime/os_linux.go:44
#2  0x0000000000415947 in runtime.notesleep (n=0xfffffffffffffe00) at /usr/local/go/src/runtime/lock_futex.go:160
#3  0x00000000004461d1 in runtime.templateThread () at /usr/local/go/src/runtime/proc.go:2385
#4  0x0000000000444c93 in runtime.mstart1 () at /usr/local/go/src/runtime/proc.go:1407
#5  0x0000000000444bd9 in runtime.mstart0 () at /usr/local/go/src/runtime/proc.go:1365
#6  0x0000000000470445 in runtime.mstart () at /usr/local/go/src/runtime/asm_amd64.s:248
#7  0x0000000000474b65 in runtime.mstart () at <autogenerated>:1
#8  0x00000000009d00b2 in crosscall_amd64 () at gcc_amd64.S:40
#9  0x00007fb86affd640 in ?? ()
#10 0x0000000000000000 in ?? ()

Thread 5 (LWP 1904481 "godal.test"):
#0  runtime.epollwait () at /usr/local/go/src/runtime/sys_linux_amd64.s:666
#1  0x000000000043b45c in runtime.netpoll (delay=<optimized out>) at /usr/local/go/src/runtime/netpoll_epoll.go:127
#2  0x00000000004474d3 in runtime.findrunnable () at /usr/local/go/src/runtime/proc.go:2947
#3  0x00000000004486d9 in runtime.schedule () at /usr/local/go/src/runtime/proc.go:3367
#4  0x0000000000448c2d in runtime.park_m (gp=0xc000282340) at /usr/local/go/src/runtime/proc.go:3516
#5  0x00000000004704c3 in runtime.mcall () at /usr/local/go/src/runtime/asm_amd64.s:307
#6  0xf934f160b63ffd00 in ?? ()
#7  0x0000000000800000 in text/template/parse.(*lexer).errorf (l=0x0, format=..., args=...) at /usr/local/go/src/text/template/parse/lex.go:188
#8  text/template/parse.lexComment (l=0x0) at /usr/local/go/src/text/template/parse/lex.go:319
#9  0x0000000000000000 in ?? ()

Thread 4 (LWP 1904480 "godal.test"):
#0  runtime.futex () at /usr/local/go/src/runtime/sys_linux_amd64.s:520
--Type <RET> for more, q to quit, c to continue without paging-- 
#1  0x000000000043b6f6 in runtime.futexsleep (addr=0xfffffffffffffe00, val=0, ns=4670691) at /usr/local/go/src/runtime/os_linux.go:44
#2  0x0000000000415947 in runtime.notesleep (n=0xfffffffffffffe00) at /usr/local/go/src/runtime/lock_futex.go:160
#3  0x0000000000444d8a in runtime.mPark () at /usr/local/go/src/runtime/proc.go:1441
#4  0x00000000004462d8 in runtime.stopm () at /usr/local/go/src/runtime/proc.go:2408
#5  0x00000000004477a5 in runtime.findrunnable () at /usr/local/go/src/runtime/proc.go:2984
#6  0x00000000004486d9 in runtime.schedule () at /usr/local/go/src/runtime/proc.go:3367
#7  0x0000000000448c2d in runtime.park_m (gp=0xc0000011e0) at /usr/local/go/src/runtime/proc.go:3516
#8  0x00000000004704c3 in runtime.mcall () at /usr/local/go/src/runtime/asm_amd64.s:307
#9  0xf934f160b63ffd00 in ?? ()
#10 0x0000000000800000 in text/template/parse.(*lexer).errorf (l=0x0, format=..., args=...) at /usr/local/go/src/text/template/parse/lex.go:188
#11 text/template/parse.lexComment (l=0x0) at /usr/local/go/src/text/template/parse/lex.go:319
#12 0x0000000000000000 in ?? ()

Thread 3 (LWP 1904479 "godal.test"):
#0  runtime.futex () at /usr/local/go/src/runtime/sys_linux_amd64.s:520
#1  0x000000000043b6f6 in runtime.futexsleep (addr=0xfffffffffffffe00, val=2, ns=4670691) at /usr/local/go/src/runtime/os_linux.go:44
#2  0x00000000004155f9 in runtime.lock2 (l=0xfffffffffffffe00) at /usr/local/go/src/runtime/lock_futex.go:107
#3  0x000000000040f55a in runtime.lockWithRank (l=0x4744e3 <runtime.futex+35>, rank=<optimized out>) at /usr/local/go/src/runtime/lockrank_off.go:23
#4  runtime.lock (l=0x4744e3 <runtime.futex+35>) at /usr/local/go/src/runtime/lock_futex.go:48
#5  runtime.chansend (c=0xc0001cd340, ep=0xc000406e4f, block=true, callerpc=<optimized out>) at /usr/local/go/src/runtime/chan.go:200
#6  0x000000000040f47d in runtime.chansend1 (c=0xfffffffffffffe00, elem=0x2) at /usr/local/go/src/runtime/chan.go:143
#7  0x00000000005139e8 in testing.tRunner.func1.1 () at /usr/local/go/src/testing/testing.go:1189
#8  0x00000000005134e7 in testing.tRunner.func1 () at /usr/local/go/src/testing/testing.go:1250
#9  0x000000000051317b in testing.tRunner (t=0xc0003b41a0, fn={void (testing.T *)} 0xc000406fc8) at /usr/local/go/src/testing/testing.go:1265
#10 0x0000000000513e4a in testing.(*T).Run·dwrap·21 () at /usr/local/go/src/testing/testing.go:1306
#11 0x0000000000472701 in runtime.goexit () at /usr/local/go/src/runtime/asm_amd64.s:1581
--Type <RET> for more, q to quit, c to continue without paging--
#12 0x0000000000000000 in ?? ()

Thread 2 (LWP 1904478 "godal.test"):
#0  runtime.usleep () at /usr/local/go/src/runtime/sys_linux_amd64.s:146
#1  0x000000000044d2d1 in runtime.sysmon () at /usr/local/go/src/runtime/proc.go:5337
#2  0x0000000000444c93 in runtime.mstart1 () at /usr/local/go/src/runtime/proc.go:1407
#3  0x0000000000444bd9 in runtime.mstart0 () at /usr/local/go/src/runtime/proc.go:1365
#4  0x0000000000470445 in runtime.mstart () at /usr/local/go/src/runtime/asm_amd64.s:248
#5  0x0000000000474b65 in runtime.mstart () at <autogenerated>:1
#6  0x00000000009d00b2 in crosscall_amd64 () at gcc_amd64.S:40
#7  0x00007fb871623640 in ?? ()
#8  0x0000000000000000 in ?? ()

Thread 1 (LWP 1904477 "godal.test"):
#0  runtime.futex () at /usr/local/go/src/runtime/sys_linux_amd64.s:520
#1  0x000000000043b6f6 in runtime.futexsleep (addr=0xfffffffffffffe00, val=2, ns=4670691) at /usr/local/go/src/runtime/os_linux.go:44
#2  0x00000000004155f9 in runtime.lock2 (l=0xfffffffffffffe00) at /usr/local/go/src/runtime/lock_futex.go:107
#3  0x0000000000410318 in runtime.lockWithRank (l=0x4744e3 <runtime.futex+35>, rank=<optimized out>) at /usr/local/go/src/runtime/lockrank_off.go:23
#4  runtime.lock (l=0x4744e3 <runtime.futex+35>) at /usr/local/go/src/runtime/lock_futex.go:48
#5  runtime.chanrecv (c=0xc0001cd340, ep=0xc00025fb67, block=true) at /usr/local/go/src/runtime/chan.go:508
#6  0x0000000000410198 in runtime.chanrecv1 (c=0xfffffffffffffe00, elem=0x2) at /usr/local/go/src/runtime/chan.go:439
#7  0x0000000000513d95 in testing.(*T).Run (t=0xc0003b41a0, name=..., f={void (testing.T *)} 0xc00025fbd8) at /usr/local/go/src/testing/testing.go:1307
#8  0x00000000005159ae in testing.runTests.func1 (t=0xc000283040) at /usr/local/go/src/testing/testing.go:1598
#9  0x0000000000513142 in testing.tRunner (t=0xc000283040, fn={void (testing.T *)} 0xc00025fc68) at /usr/local/go/src/testing/testing.go:1259
#10 0x000000000051585f in testing.runTests (matchString={void (string, string, bool *, error *)} 0xc00025fd10, tests=[]testing.InternalTest = {...}, deadline=...) at /usr/local/go/src/testing/testing.go:1596
--Type <RET> for more, q to quit, c to continue without paging--
#11 0x00000000005145bd in testing.(*M).Run (m=0xc00011c800, code=0) at /usr/local/go/src/testing/testing.go:1504
#12 0x00000000009ca56b in main.main () at _testmain.go:177
(gdb) q
A debugging session is active.

	Inferior 1 [process 1904477] will be detached.

Quit anyway? (y or n) y
Detaching from program: /home/tbonfort/dev/godal/godal.test, process 1904477
[Inferior 1 (process 1904477) detached]

I have bisected the issue down to 1c590661e7d3b477662f76ead56f39567ea8345a (from @bcmills). Reverting this specific commit in master fixes the deadlock.

About this issue

  • Original URL
  • State: closed
  • Created 3 years ago
  • Comments: 21 (20 by maintainers)

Most upvoted comments

This issue can be closed as invalid. The OpenShared testcase was asking the cgo lib to populate 300 bytes of data inside a slice we had only sized to 100 bytes. I’m actually confused as to why this didn’t crash in previous go versions. Thanks @bcmills for the investigation, and sorry for erroneously dragging you into an issue that’s entirely unrelated with the bisected commit.

+1
tbonfort on Sep 23, 2021

Using libgdal 3.2.2+dfsg-3 (installed via apt) I see a test failure in TestVSIGCS under go1.16.8:

~/tmp/godal$ go1.16.8 test . -timeout=20s
ERROR 10: Pointer 'hSRS1' is NULL in 'OSRIsSame'.

ERROR 10: Pointer 'hSRS2' is NULL in 'OSRIsSame'.

ERROR 10: Pointer 'hSRS' is NULL in 'OSRIsGeographic'.

ERROR 10: Pointer 'hSRS' is NULL in 'OSRGetSemiMajor'.

ERROR 10: Pointer 'hSRS' is NULL in 'OSRGetSemiMajor'.

ERROR 10: Pointer 'hSRS' is NULL in 'OSRGetAuthorityCode'.

ERROR 10: Pointer 'hSRS' is NULL in 'OSRGetAuthorityName'.

ERROR 10: Pointer 'hSRS' is NULL in 'OSRAutoIdentifyEPSG'.

ERROR 1: PROJ: proj_get_ellipsoid: CRS has no geodetic CRS
ERROR 1: PROJ: proj_get_ellipsoid: Object is not a CRS or GeodeticReferenceFrame
ERROR 1: PROJ: proj_get_ellipsoid: CRS has no geodetic CRS
ERROR 1: PROJ: proj_get_ellipsoid: Object is not a CRS or GeodeticReferenceFrame
ERROR 1: PROJ: proj_get_ellipsoid: CRS has no geodetic CRS
ERROR 1: PROJ: proj_get_ellipsoid: Object is not a CRS or GeodeticReferenceFrame
ERROR 10: Pointer 'hGeom' is NULL in 'OGR_G_IsEmpty'.

--- FAIL: TestVSIGCS (0.22s)
    godal_test.go:3044: new reader for gs://godal-ci-data/test.tif: Get "https://storage.googleapis.com/godal-ci-data/test.tif": oauth2: cannot fetch token: 400 Bad Request
        Response: {
          "error": "invalid_grant",
          "error_description": "Bad Request"
        }
OGRCT: Wrap target at 50.
ERROR 10: Pointer 'hBand' is NULL in 'GDALGetMaskBand'.

ERROR 10: Pointer 'hBand' is NULL in 'GDALGetMaskBand'.

FAIL
FAIL    github.com/airbusgeo/godal      2.441s
FAIL

I can reproduce the reported deadlock using both go1.17 and gotip.

@bcmills I firmly believe that failing test is unrelated to the issue at hand, but is due to incorrect credential handling in another part of the code. (i.e. they’ll fail identically if you use go <1.17). The ERROR:XXX messages are expected and due to the test suite performing error-raising cases.

+1
tbonfort on Sep 22, 2021

Here is the dump:

panic: test timed out after 10s

goroutine 50 [running]:
testing.(*M).startAlarm.func1()
	/home/tbonfort/dev/thirdparty/go/src/testing/testing.go:1788 +0x8e
created by time.goFunc
	/home/tbonfort/dev/thirdparty/go/src/time/sleep.go:176 +0x32

goroutine 1 [running]:
	goroutine running on other thread; stack unavailable

goroutine 6 [select]:
go.opencensus.io/stats/view.(*worker).start(0xc0000fa280)
	/home/tbonfort/go/pkg/mod/go.opencensus.io@v0.23.0/stats/view/worker.go:276 +0xad
created by go.opencensus.io/stats/view.init.0
	/home/tbonfort/go/pkg/mod/go.opencensus.io@v0.23.0/stats/view/worker.go:34 +0x8d

goroutine 35 [running]:
	goroutine running on other thread; stack unavailable
created by testing.(*T).Run
	/home/tbonfort/dev/thirdparty/go/src/testing/testing.go:1306 +0x35f
exit status 2
FAIL	github.com/airbusgeo/godal	10.032s
+1
tbonfort on Sep 16, 2021

@bcmills , sorry, I should have posted the go backtrace in the main issue. I didn’t include it because it didn’t contain any meaningful info for me aside of messages of the like of “goroutine running on other thread; stack unavailable”. I’ll post the exact output tomorrow UTC

+1
tbonfort on Sep 16, 2021
Read more comments on GitHub
← go: testing: exclude statically unreachable code in test coverage
go: testing: panic in Init if flag.Parse has already been called →