Re: Non-bootstrapping Go, COMPAT options

[christos%astron.com@localhost (Christos Zoulas)]

In article <20180113.234925.1513924779772085164.oshimaya%yagoto-urayama.jp@localhost>,
Yasushi Oshima  <oshima-ya%yagoto-urayama.jp@localhost> wrote:
>Thu, 11 Jan 2018 23:43:39 +0900 (JST)
>I wrote:
>
>> And more, in newer kernel(ex. 20180108 for netbsd-8),
>> another problem appears,
>> During to build go1.9.2, sometimes a kernel hangup occurs.
>> I have not examined anything about this yet.
>
>There is the ddb backtrace in this case by latest netbsd-8 kernel
>from daily snapshot:
> http://nycdn.netbsd.org/pub/NetBSD-daily/netbsd-8/201801120300Z/evbarm-earmv6hf/binary/kernel/netbsd-RPI.bin.gz
>
>ddb log when hangup from serial console:
> https://gist.github.com/oshimaya/00c3d5e17b7b17ce5e78bc2b46680bc9#file-armv6_go1-9-2_build_hangup_ddb-log
>
>This shows go process wants to exit but anything block it,
>sys_exit ->  exit1 -> exit_lwps -> lwp_wait -> cv_wait_sig -> sleepq_block...
>
>This happen before building a bootstrap for go1.9.2,
>So this 'go' process will be go1.4.
>
>The thread handling of my Go1.4 port for NetBSD/earm may be
>something wrong, but the hangup the kernel is important problem,
>I think.
>In my opinion, any userland process should not be able to
>hangup whole system.

You need: cvs diff -r1.190 -r 1.191 kern_lwp.c 
I've asked for a pullup.

christos