qemu/sparc64 hang, glib2 maybe

[Greg Troxel <gdt%lexort.com@localhost>]

Patrick Welche <prlw1%cam.ac.uk@localhost> writes:

> On Sun, Dec 29, 2019 at 09:28:52PM -0500, Greg Troxel wrote:
>> Frédéric Fauberteau <triaxx%netbsd.org@localhost> writes:
>> 
>> > Module Name:	pkgsrc
>> > Committed By:	triaxx
>> > Date:		Sun Dec 29 19:40:56 UTC 2019
>> >
>> > Modified Files:
>> > 	pkgsrc/devel/glib2: Makefile Makefile.common PLIST distinfo
>> > 	pkgsrc/devel/glib2/patches: patch-glib_gmain.c patch-glib_gtypes.h
>> > 	    patch-glib_meson.build patch-glib_tests_hash.c
>> > Added Files:
>> > 	pkgsrc/devel/glib2/patches: patch-glib_genviron.c patch-glib_gspawn.c
>> >
>> > Log Message:
>> > glib2: update to 2.62.4
>> 
>> Note that glib2 is still in freeze.   Can you explain how this update
>> relates to the qemu sparc64 issues?
>
> Having look at it yesterday, if you take qemu 4.2.0, and an ancient
> glib2 2.56.2nb4, sparc64 still hangs. This is not the same as the
> original PR 54310, and it appears that glib2 is blameless.

So you mean, more or less, that

  the latest qemu is troubled, even with old glib2.

  you theorize that the current difficulty is a regression in recent
  qemu, and that there is no evidence linking it to glib?

I guess it would be helpful then for someone to run older versions of
qemu with the current glib2, to try to confirm qemu regression theory.

I'll start building a few things to try to help diagnose.