os/signal: unrecognized failures

[gopherbot]

#!watchflakes
default <- pkg == "os/signal" && test == ""

Issue created automatically to collect these failures.

Example (log):

SIGSEGV: segmentation violation
PC=0x3a1b8c m=10 sigcode=1 addr=0x3a1b8c

goroutine 0 gp=0xc000082e00 m=10 mp=0xc000081808 [idle]:
runtime.fpTracebackPCs(...)
	/tmp/workdir/go/src/runtime/trace2stack.go:255
runtime.traceStack(0xc000082700?, 0x200000004?, 0x4d)
	/tmp/workdir/go/src/runtime/trace2stack.go:109 +0x28c fp=0x2f4250db0 sp=0x2f4250948 pc=0x3a1b8c
runtime.traceAdvance.func1.1()
	/tmp/workdir/go/src/runtime/trace2.go:370 +0xbd fp=0x2f4250e08 sp=0x2f4250db0 pc=0x39d43d
...
r11    0xc000032508
r12    0x2f42509c8
r13    0xc000033b00
r14    0xc000082e00
r15    0x4
rip    0x3a1b8c
rflags 0x10202
cs     0x2b
fs     0x0
gs     0x0

— watchflakes

[gopherbot]

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "os/signal" && test == ""

2024-04-08 20:51 openbsd-amd64-72 go@8008998b os/signal (log)

SIGSEGV: segmentation violation
PC=0x3a1b8c m=10 sigcode=1 addr=0x3a1b8c

goroutine 0 gp=0xc000082e00 m=10 mp=0xc000081808 [idle]:
runtime.fpTracebackPCs(...)
	/tmp/workdir/go/src/runtime/trace2stack.go:255
runtime.traceStack(0xc000082700?, 0x200000004?, 0x4d)
	/tmp/workdir/go/src/runtime/trace2stack.go:109 +0x28c fp=0x2f4250db0 sp=0x2f4250948 pc=0x3a1b8c
runtime.traceAdvance.func1.1()
	/tmp/workdir/go/src/runtime/trace2.go:370 +0xbd fp=0x2f4250e08 sp=0x2f4250db0 pc=0x39d43d
...
r11    0xc000032508
r12    0x2f42509c8
r13    0xc000033b00
r14    0xc000082e00
r15    0x4
rip    0x3a1b8c
rflags 0x10202
cs     0x2b
fs     0x0
gs     0x0

— watchflakes

[ianlancetaylor]

From the build dashboard this problem seems to have started since the stack containing https://go.dev/cl/567076 was submitted. Could that be related? @mknyszek

[gopherbot]

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "os/signal" && test == ""

2024-04-10 19:41 openbsd-amd64-72 go@5b5d6f87 os/signal (log)

SIGSEGV: segmentation violation
PC=0x3a1dac m=7 sigcode=1 addr=0x3a1dac

goroutine 0 gp=0xc000082540 m=7 mp=0xc000080808 [idle]:
runtime.fpTracebackPCs(...)
	/tmp/workdir/go/src/runtime/trace2stack.go:248
runtime.traceStack(0xc000082700?, 0x200000004?, 0x7)
	/tmp/workdir/go/src/runtime/trace2stack.go:109 +0x28c fp=0x23029d7f0 sp=0x23029d388 pc=0x3a1dac
runtime.traceAdvance.func1.1()
	/tmp/workdir/go/src/runtime/trace2.go:370 +0xbd fp=0x23029d848 sp=0x23029d7f0 pc=0x39d6bd
...
r11    0x140
r12    0x23029d408
r13    0xffffffffffffffff
r14    0xc000082540
r15    0x4
rip    0x3a1dac
rflags 0x10202
cs     0x2b
fs     0x0
gs     0x0

— watchflakes

[mknyszek]

Thanks. I'll look into it. At first glance, it looks like there's a place on this platform where a goroutine can stop and frame pointers aren't preserved.

[gopherbot]

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "os/signal" && test == ""

2024-04-10 22:22 solaris-amd64-oraclerel go@7b3c3804 os/signal (log)

SIGSEGV: segmentation violation
PC=0x461c8c m=8 sigcode=1 addr=0x461c8c

goroutine 0 gp=0xc00018c1c0 m=8 mp=0xc000188008 [idle]:
runtime.fpTracebackPCs(...)
	/tmp/workdir-host-solaris-oracle-amd64-oraclerel/go/src/runtime/trace2stack.go:248
runtime.traceStack(0x7fff75ffff90?, 0xc000188008?, 0x7fff75ffff68)
	/tmp/workdir-host-solaris-oracle-amd64-oraclerel/go/src/runtime/trace2stack.go:109 +0x28c fp=0x7fff75ffff38 sp=0x7fff75fffad0 pc=0x461c8c
runtime: g 0: unexpected return pc for runtime.traceStack called from 0x7fffbf5d3a40
stack: frame={sp:0x7fff75fffad0, fp:0x7fff75ffff38} stack=[0x7fff75efffc0,0x7fff75ffffc0)
...
r11    0x140
r12    0x7fff75fffb50
r13    0xffffffffffffffff
r14    0xc00018c1c0
r15    0x4
rip    0x461c8c
rflags 0x10206
cs     0x53
fs     0x0
gs     0x0

— watchflakes

[gopherbot]

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "os/signal" && test == ""

2024-04-12 14:07 openbsd-amd64-72 go@5c20d5fa os/signal (log)

SIGSEGV: segmentation violation
PC=0x3a20ac m=8 sigcode=1 addr=0x3a20ac

goroutine 0 gp=0xc0001841c0 m=8 mp=0xc000182008 [idle]:
runtime.fpTracebackPCs(...)
	/tmp/workdir/go/src/runtime/trace2stack.go:248
runtime.traceStack(0xc000082700?, 0x200000004?, 0x49)
	/tmp/workdir/go/src/runtime/trace2stack.go:109 +0x28c fp=0x2190b38c0 sp=0x2190b3458 pc=0x3a20ac
runtime.traceAdvance.func1.1()
	/tmp/workdir/go/src/runtime/trace2.go:370 +0xbd fp=0x2190b3918 sp=0x2190b38c0 pc=0x39d9bd
...
r11    0x140
r12    0x2190b34d8
r13    0xffffffffffffffff
r14    0xc0001841c0
r15    0x4
rip    0x3a20ac
rflags 0x10202
cs     0x2b
fs     0x0
gs     0x0

— watchflakes

[gopherbot]

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "os/signal" && test == ""

2024-04-16 17:46 openbsd-amd64-72 go@7a0e2db1 os/signal (log)

SIGSEGV: segmentation violation
PC=0x3a672c m=13 sigcode=1 addr=0x3a672c

goroutine 0 gp=0xc00013c540 m=13 mp=0xc000183008 [idle]:
runtime.fpTracebackPCs(...)
	/tmp/workdir/go/src/runtime/tracestack.go:246
runtime.traceStack(0xc000082700?, 0x200000004?, 0x2e)
	/tmp/workdir/go/src/runtime/tracestack.go:107 +0x28c fp=0x2931363f0 sp=0x293135f88 pc=0x3a672c
runtime.traceAdvance.func1.1()
	/tmp/workdir/go/src/runtime/trace.go:368 +0xbd fp=0x293136448 sp=0x2931363f0 pc=0x39e0bd
...
r11    0x140
r12    0x293136008
r13    0xffffffffffffffff
r14    0xc00013c540
r15    0x4
rip    0x3a672c
rflags 0x10202
cs     0x2b
fs     0x0
gs     0x0

— watchflakes

[nazar256]

We have been experiencing very similar issues in production since 13th of August with both go 1.22.6 and 1.23.0. It happens at least few times a day crashing our services. The issue also looks very similar to this.

[ianlancetaylor]

Ping @mknyszek . Is this still an open issue? Thanks.