runtime: "morestack on g0" in debugcall on android-amd64

[prattmic]

Seen once on gomote with all.bash:

fatal: morestack on g0
trap in unknown function runtime.abort
SIGTRAP: trace trap
PC=0x7e7838bba342 m=11 sigcode=128

goroutine 0 [idle]:
runtime.abort()
        /workdir/go/src/runtime/asm_amd64.s:1059 +0x2 fp=0x7e780fafd410 sp=0x7e780fafd408 pc=0x7e7838bba342
runtime.morestack()
        /workdir/go/src/runtime/asm_amd64.s:538 +0x22 fp=0x7e780fafd418 sp=0x7e780fafd410 pc=0x7e7838bb8322

goroutine 23000 [running, locked to thread]:
runtime.systemstack_switch()
        /workdir/go/src/runtime/asm_amd64.s:459 fp=0xc0024ba620 sp=0xc0024ba618 pc=0x7e7838bb8260
runtime.debugCallCheck(0x7e7838e0bce5)
        /workdir/go/src/runtime/debugcall.go:42 +0x85 fp=0xc0024ba660 sp=0xc0024ba620 pc=0x7e7838b4c765
runtime.debugCallCheck(0x7e7838e0bce5)
        <autogenerated>:1 +0x27 fp=0xc0024ba678 sp=0xc0024ba660 pc=0x7e7838bbca07
runtime.debugCallV2()
        /workdir/go/src/runtime/asm_amd64.s:1833 +0x89 fp=0xc0024ba720 sp=0xc0024ba678 pc=0x7e7838bba789
runtime_test.debugCallWorker2(0xc000c3bb00?, 0xc0000343d8?)
        /workdir/go/src/runtime/debug_test.go:94 +0x5 fp=0xc0024ba728 sp=0xc0024ba720 pc=0x7e7838e0bce5
runtime_test.debugCallWorker(0xc000ef84e0?, 0x3e7?, 0xc000034380?)
        /workdir/go/src/runtime/debug_test.go:79 +0x8b fp=0xc0024ba7b8 sp=0xc0024ba728 pc=0x7e7838e0bc0b
runtime_test.startDebugCallWorker.func2()
        /workdir/go/src/runtime/debug_test.go:59 +0x2e fp=0xc0024ba7e0 sp=0xc0024ba7b8 pc=0x7e7838e0bb4e
runtime.goexit()
        /workdir/go/src/runtime/asm_amd64.s:1594 +0x1 fp=0xc0024ba7e8 sp=0xc0024ba7e0 pc=0x7e7838bba501
created by runtime_test.startDebugCallWorker
        /workdir/go/src/runtime/debug_test.go:59 +0xee

This looks like morestack right on entry to this closure, which is odd because we should have plenty of stack.

cc @golang/runtime @mknyszek

[cherrymui]

Yeah, that call is made on the system stack. From the stack trace it seems it successfully switched to the g0 stack. Maybe the stack bounds of g0 is wrong?