runtime: cgo stuck because of go signal handler went into dead loop

[cocktail828]

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

$ go version
go1.19

Does this issue reproduce with the latest release?

Yes. It can be reproduce on go version 1.19

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

containered CentOS 7

go env Output

GO111MODULE="on"
GOARCH="amd64"
GOBIN=""
GOCACHE="/root/.cache/go-build"
GOENV="/root/.config/go/env"
GOEXE=""
GOEXPERIMENT=""
GOFLAGS=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GOINSECURE=""
GOMODCACHE="/root/go/pkg/mod"
GONOPROXY=""
GONOSUMDB=""
GOOS="linux"
GOPATH="/root/go:/data/home/xqzhu11/go"
GOPRIVATE=""
GOPROXY="https://goproxy.cn,direct"
GOROOT="/usr/local/go"
GOSUMDB="sum.golang.org"
GOTMPDIR=""
GOTOOLDIR="/usr/local/go/pkg/tool/linux_amd64"
GOVCS=""
GOVERSION="go1.19"
GCCGO="gccgo"
GOAMD64="v1"
AR="ar"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
GOMOD="/dev/null"
GOWORK=""
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 -Wl,--no-gc-sections -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build1022167488=/tmp/go-build -gno-record-gcc-switches"

The following is what I see.
Backtrace with gdb:

Thread 271 (Thread 0x7f634dfff700 (LWP 223)):
#0  runtime.usleep () at /usr/local/go/src/runtime/sys_linux_amd64.s:140
#1  0x0000000000452c66 in runtime.raisebadsignal (sig=11, c=0x7f634df8e290) at /usr/local/go/src/runtime/signal_unix.go:939
#2  0x00000000004532ec in runtime.badsignal (sig=11, c=0x7f634df8e290) at /usr/local/go/src/runtime/signal_unix.go:1054
#3  0x0000000000451885 in runtime.sigtrampgo (sig=11, info=0x7f634df8e430, ctx=0x7f634df8e300) at /usr/local/go/src/runtime/signal_unix.go:461
#4  0x0000000000471e46 in runtime.sigtramp () at /usr/local/go/src/runtime/sys_linux_amd64.s:359
#5  <signal handler called>
#6  0x00007f695e284f78 in std::basic_string<char, std::char_traits<char>, std::allocator<char> >::basic_string(std::string const&) () from /lib64/libstdc++.so.6
#7  0x00007f69152ae7bf in FANetBuilder::BuildWordLevelNetNew (this=this@entry=0x7f59e6295440, pWordsID=pWordsID@entry=0x7f59c40640cc, wordCount=6, pFANodeWordArr=0x7f59e62954a8, nWordNodeCount=nWordNodeCount@entry=@0x7f634df8e9a8: 0) at FANetBuilder.cpp:169
...

Strace syscalls: As you can see, SIGSEGV is raised again and again...

[pid 108096] --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x51ebcdf0} ---
[pid 108096] rt_sigprocmask(SIG_SETMASK, NULL, ~[KILL STOP], 8) = 0
[pid 108096] rt_sigprocmask(SIG_SETMASK, ~[], NULL, 8) = 0
[pid 108096] sigaltstack(NULL, {ss_sp=0, ss_flags=SS_DISABLE, ss_size=0}) = 0
[pid 108096] sigaltstack({ss_sp=0xc0272c0000, ss_flags=0, ss_size=32768}, NULL) = 0
[pid 108096] rt_sigprocmask(SIG_SETMASK, ~[HUP INT QUIT ILL TRAP ABRT BUS FPE KILL SEGV TERM STKFLT CHLD STOP URG PROF SYS RTMIN RT_1 RT_2], NULL, 8) = 0
[pid 108096] gettid()                   = 223
[pid 108096] rt_sigprocmask(SIG_UNBLOCK, [SEGV], NULL, 8) = 0
[pid 108096] rt_sigaction(SIGSEGV, {SIG_DFL, ~[], SA_RESTORER|SA_STACK|SA_RESTART|SA_SIGINFO, 0x7f695ef1d630}, NULL, 8) = 0
[pid 108096] getpid()                   = 1
[pid 108096] gettid()                   = 223
[pid 108096] tgkill(1, 223, SIGSEGV)    = 0
[pid 108096] --- SIGSEGV {si_signo=SIGSEGV, si_code=SI_TKILL, si_pid=1, si_uid=0} ---
[pid 108096] nanosleep({0, 1000000}, NULL) = 0
[pid 108096] rt_sigaction(SIGSEGV, {0x471e80, ~[], SA_RESTORER|SA_STACK|SA_RESTART|SA_SIGINFO, 0x7f695ef1d630}, NULL, 8) = 0
[pid 108096] rt_sigprocmask(SIG_SETMASK, ~[], NULL, 8) = 0
[pid 108096] sigaltstack({ss_sp=0, ss_flags=SS_DISABLE, ss_size=0}, NULL) = 0
[pid 108096] rt_sigprocmask(SIG_SETMASK, ~[KILL STOP], NULL, 8) = 0
[pid 108096] rt_sigreturn()             = 0
[pid 108096] --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x51ebcdf0} ---
[pid 108096] rt_sigprocmask(SIG_SETMASK, NULL, ~[KILL STOP], 8) = 0
[pid 108096] rt_sigprocmask(SIG_SETMASK, ~[], NULL, 8) = 0
[pid 108096] sigaltstack(NULL, {ss_sp=0, ss_flags=SS_DISABLE, ss_size=0}) = 0
[pid 108096] sigaltstack({ss_sp=0xc0272c0000, ss_flags=0, ss_size=32768}, NULL) = 0
[pid 108096] rt_sigprocmask(SIG_SETMASK, ~[HUP INT QUIT ILL TRAP ABRT BUS FPE KILL SEGV TERM STKFLT CHLD STOP URG PROF SYS RTMIN RT_1 RT_2], NULL, 8) = 0
[pid 108096] gettid()                   = 223
[pid 108096] rt_sigprocmask(SIG_UNBLOCK, [SEGV], NULL, 8) = 0
[pid 108096] rt_sigaction(SIGSEGV, {SIG_DFL, ~[], SA_RESTORER|SA_STACK|SA_RESTART|SA_SIGINFO, 0x7f695ef1d630}, NULL, 8) = 0
[pid 108096] getpid()                   = 1
[pid 108096] gettid()                   = 223
[pid 108096] tgkill(1, 223, SIGSEGV)    = 0
[pid 108096] --- SIGSEGV {si_signo=SIGSEGV, si_code=SI_TKILL, si_pid=1, si_uid=0} ---
...

What did you expect to see?

Panic.

What did you see instead?

It stucked, no crash or panic.

[cocktail828]

@ianlancetaylor can you help me, thanks very much.

[tarikkilic]

We faced of the same issue. But when go test -msan command run, and error occured like this:

gcc: error: unrecognized argument to -fsanitize= option: 'memory'

It happens sporadically.

[cocktail828]

@tarikkilic GCC does not support -fsanitize=memory. Try use clang instead.

[mknyszek]

Am I understanding correctly that the issue is occurring in Go 1.10.3? That version has not been supported for a very long time. Can you reproduce the issue with the latest release of Go?

CC @golang/runtime

[cocktail828]

Yep. The issue occurs on product enviroment. We are trying recompile the docker images but it may take several days to update all services nodes and reproduce(if not fixed) the issue.
Thanks for your reply.

[joedian]

@cocktail828 can you confirm if the issue is still happening after recompiling?

[cocktail828]

@cocktail828 can you confirm if the issue is still happening after recompiling?

Sorry for late response. It does not happened for now. I will keep trace the issue. If it does not hanppen for 2 weeks I will close the issue myself.

[cocktail828]

@joedian Sorry to interrupt you. The issue reproduce again(go 1.19) after big traffic. I have update this issue with the latest info.