runtime: m called on g0 stack in tracer

[gopherbot]

#!watchflakes
default <- pkg ~ "trace" && log ~ `mcall called on m->g0 stack`

Issue created automatically to collect these failures.

Example (log):

=== RUN   TestTraceStacks/Stress
    trace_test.go:613: stderr: fatal error: runtime: mcall called on m->g0 stack
        
        runtime stack:
        runtime.throw({0x133054?, 0x605c8?})
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/panic.go:1089 +0x34 fp=0xf855437ee770 sp=0xf855437ee740 pc=0x7e924
        runtime.badmcall(0x4000102c40?)
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/proc.go:573 +0x28 fp=0xf855437ee790 sp=0xf855437ee770 pc=0x4c9d8
        runtime.badmcall(0x221280)
        	<autogenerated>:1 +0x14 fp=0xf855437ee7b0 sp=0xf855437ee790 pc=0x87304
...
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/mgc.go:1431 +0xdc fp=0x400017cfb0 sp=0x400017cf10 pc=0x2d5cc
        runtime.gcBgMarkStartWorkers.gowrap1()
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/mgc.go:1347 +0x28 fp=0x400017cfd0 sp=0x400017cfb0 pc=0x2d4b8
        runtime.goexit({})
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/asm_arm64.s:1266 +0x4 fp=0x400017cfd0 sp=0x400017cfd0 pc=0x85764
        created by runtime.gcBgMarkStartWorkers in goroutine 21
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/mgc.go:1347 +0x140
        exit status 2
    trace_test.go:615: exit status 1
--- FAIL: TestTraceStacks/Stress (0.32s)

— watchflakes

[gopherbot]

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "internal/trace" && test == "TestTraceStacks/Stress"

2025-04-18 21:13 gotip-linux-arm64-longtest go@28fd9fa8 internal/trace.TestTraceStacks/Stress (log)

=== RUN   TestTraceStacks/Stress
    trace_test.go:613: stderr: fatal error: runtime: mcall called on m->g0 stack
        
        runtime stack:
        runtime.throw({0x133054?, 0x605c8?})
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/panic.go:1089 +0x34 fp=0xf855437ee770 sp=0xf855437ee740 pc=0x7e924
        runtime.badmcall(0x4000102c40?)
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/proc.go:573 +0x28 fp=0xf855437ee790 sp=0xf855437ee770 pc=0x4c9d8
        runtime.badmcall(0x221280)
        	<autogenerated>:1 +0x14 fp=0xf855437ee7b0 sp=0xf855437ee790 pc=0x87304
...
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/mgc.go:1431 +0xdc fp=0x400017cfb0 sp=0x400017cf10 pc=0x2d5cc
        runtime.gcBgMarkStartWorkers.gowrap1()
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/mgc.go:1347 +0x28 fp=0x400017cfd0 sp=0x400017cfb0 pc=0x2d4b8
        runtime.goexit({})
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/asm_arm64.s:1266 +0x4 fp=0x400017cfd0 sp=0x400017cfd0 pc=0x85764
        created by runtime.gcBgMarkStartWorkers in goroutine 21
        	/home/swarming/.swarming/w/ir/x/w/goroot/src/runtime/mgc.go:1347 +0x140
        exit status 2
    trace_test.go:615: exit status 1
--- FAIL: TestTraceStacks/Stress (0.32s)

— watchflakes

[gopherbot]

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "internal/trace" && test == "TestTraceStacks/Stress"

2025-05-09 00:40 gotip-linux-386-longtest go@0c8615be internal/trace.TestTraceStacks/Stress (log)

=== RUN   TestTraceStacks/Stress
    exec.go:213: test timed out while running command: /home/swarming/.swarming/w/ir/x/w/goroot/bin/go run testdata/testprog/stacks.go
    trace_test.go:615: signal: killed
--- FAIL: TestTraceStacks/Stress (789.46s)

— watchflakes

[gopherbot]

Change https://go.dev/cl/680315 mentions this issue: runtime: don't do a direct G handoff in semrelease on systemstack