On Sat, Jan 08, 2005 at 11:07:07AM +0800, Xin LI wrote: > On Fri, Jan 07, 2005 at 04:55:40PM -0800, Steve Kargl wrote: > > On Sat, Jan 08, 2005 at 01:11:40AM +0100, Ivan Voras wrote: > > > It's been a while now and (judging from this list at least), people are > > > not complaining about ULE, so maybe (with re_at_ approval) the fix & > > > supporting infrastructure could be brought to RELENG_5? > > > > > > > That's not a good idea. I can lock up ULE+PREEMPTION on > > a dual amd64 system within 30 minutes. The system is > > runnnig X11 and normally firefox (or other threaded > > app) and it simply freezes. No panic. No keyboard > > response. Nothing. > > This is observed in pre-5.3RELEASE CURRENT, but I thought Jeff has > already fixed it in -CURRENT. I don't have dual amd64 system to do > such experiments, are you sure you are talking about the latest > 6-CURRENT, or just 5.3? > Yes, I'm talking about 6-CURRENT. My last kernel/world build is 04 Jan 05. I can only invoke the problem with X11 running, and I currently don't have a serial console on this system. I'm guessing that I hit a deadlock (or livelock). -- SteveReceived on Sat Jan 08 2005 - 02:44:27 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:25 UTC