Re: ATF tests hanging

[Andreas Gustafsson <gson%gson.org@localhost>]

Christos,

You wrote:
> |   2017.12.16.12.40.30 pgoyette src/share/man/man9/module.9 1.39
> |   2017.12.16.14.45.25 christos src/tests/lib/libc/sys/t_ptrace_wait.c 1.14
> |   2017.12.16.14.45.25 christos src/tests/lib/libc/sys/t_ptrace_x86_wait.h 1.3
> | 
> | the i386 tests on b5 started hanging at the lib/libc/sys/t_ptrace_wait:resume1
> | test case.  Then, after these commits:
> 
> This is marked as fail, will timeout. Is the timeout broken on i386? 

Other test cases do time out as expected.  I don't know if the issue
is i386 specific since it got masked by the t_spwanattr hang before
the testbed got around to testing any other architecture.

> BTW, this is not much different than the code that was there before...

Just a guess: the line "ATF_REQUIRE(0 && "In order to get reliable
failure, abort")" was removed; perhaps the code after that line was
not previously executed, but is now executed and causes the kernel to
hang.

Looks like I already reported this same hang a couple of weeks ago,
and Kamil responded saying he would disable the test:

  http://mail-index.netbsd.org/current-users/2017/12/04/msg032841.html
  http://mail-index.netbsd.org/current-users/2017/12/04/msg032842.html

> Ok, I understand this one, should be fixed!

OK, presumably that will make it go back to hanging in
t_ptrace_wait:resume1 again.
-- 
Andreas Gustafsson, gson%gson.org@localhost