Re: port-sparc/32052 (tsleep() does not work on SMP sparc while shutting down)

[Hauke Fath <hauke%Espresso.Rhein-Neckar.DE@localhost>]

The following reply was made to PR port-sparc/32052; it has been noted by GNATS.

From: Hauke Fath <hauke%Espresso.Rhein-Neckar.DE@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: port-sparc-maintainer%NetBSD.org@localhost
Subject: Re: port-sparc/32052 (tsleep() does not work on SMP sparc while
 shutting down)
Date: Tue, 1 Mar 2011 20:51:38 +0100

 At 10:48 Uhr +0000 1.3.2011, mrg%NetBSD.org@localhost wrote:
 >netbsd/sparc never called hardclock on non-boot cpus before semi-recent
 >times, and this was the source of some of the sparc smp problems for a
 >while.  i wonder if this is the problem here.  are you able to test
 >either a patched netbsd-5 or a -current kernel?  i may also be able
 >to provide a patched netbsd-4 kernel.  thanks.
 
 Hi,
 
 thanks for looking at the issue. It's been a while... I ended up getting an
 external wlan access point in place of the pcmcia card, and the machine has
 morphed into an SS20 in the meantime.
 
 I can drop in the nell, and try a -current kernel; in fact, I have a
 sort-of-current installation around from testing the waters, since it looks
 like netbsd-5 on sparc-mp is never going to happen, and things from a newer
 version of pf(4) to wapbl look tempting. Fine-grained locking changes have
 broken appletalk support on many ethernet interfaces, though - not le(4)
 but hme(4).
 
 OTOH, I'd definitely be interested in netbsd-4 patches, since that would be
 more of a drop-in replacement, pf compatible and all. Plus, it's the last
 usable release on sparc-mp. Just tell me if the patches are much of an
 effort.
 
        hauke
 
 --
 "It's never straight up and down"     (DEVO)