Re: Non-bootstrapping Go, COMPAT options

[Harold Gutch <logix%foobar.franken.de@localhost>]

Hi,

On Thu, Jan 18, 2018 at 12:44:50AM +0900, Yasushi Oshima wrote:
> Hello Martin,
> 
> Sat, 13 Jan 2018 17:15:49 +0100
> Martin Husemann <martin%duskware.de@localhost> wrote:
> 
> > Could you file a PR and provide a test binary for this case?
> 
> Ok, I'll send-pr later, wait a little more. 
> 
> A test binary can be make by building pkgsrc/lang/go14 on NetBSD/earmv6hf.
> 
> Theare is a binary for this case in my environment,
> http://www.yagoto-urayama.jp/~oshimaya/netbsd/files/goarm/original/go_bootstrap.gz
> 
> When exec this go_bootstrap on NetBSD/earmv6hf-el 8.0_BETA with RPI,
> It will be lock system, ping OK, but no response from remote by ssh. 

Ah, this explains the "unexplainable reboots" I have seen when
building Go 1.2.2 (the oldest I could find) on a Pi 1 running 7.1.

I have a watchdog (in user tickle mode) configured, that explains why
it doesn't just crash for me but instead reboots - I don't have a
screen or a serial console attached, therefore I so far hadn't been
able to check the console to see what exactly happens when it crashes.

But all this is consistent, I just gave it another try and indeed it
stopped responding to pings on the (userland) tun interface pretty
much immediately, while it continues to respond to pings to the
ethernet IP, at least for a few seconds until the watchdog timer kicked
in.

All this is on an almost vanilla Go 1.2.2, merely make.bash was edited,
changing two occurrences of "==" to "=" (which wouldn't make a
difference to the build process either way).


  Harold