Date: Wed, 16 Aug 2006 00:10:02 +0200 From: Alexander Leidinger <Alexander@Leidinger.net> To: Steve Kargl <sgk@troutmask.apl.washington.edu> Cc: Suleiman Souhlal <ssouhlal@freebsd.org>, Kip Macy <kmacy@fsmware.com>, Julian Elischer <julian@elischer.org>, current@freebsd.org Subject: Re: HEADS-UP: starting to commit linuxolator (SoC 2006) changes... Message-ID: <20060816001002.7b3709b0@Magellan.Leidinger.net> In-Reply-To: <20060815210603.GA13060@troutmask.apl.washington.edu> References: <44E1BD03.2030402@FreeBSD.org> <20060815144625.362bf376@Magellan.Leidinger.net> <44E1C3E4.7080508@FreeBSD.org> <20060815153451.604d16f1@Magellan.Leidinger.net> <44E1E85D.5070805@FreeBSD.org> <20060815180713.6a4ee2e6@Magellan.Leidinger.net> <44E21AA2.7030301@elischer.org> <20060815193409.GA12116@troutmask.apl.washington.edu> <20060815131010.M18116@demos.bsdclusters.com> <20060815225712.5cf41e33@Magellan.Leidinger.net> <20060815210603.GA13060@troutmask.apl.washington.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
Quoting Steve Kargl <sgk@troutmask.apl.washington.edu> (Tue, 15 Aug 2006 14:06:03 -0700): > On Tue, Aug 15, 2006 at 10:57:12PM +0200, Alexander Leidinger wrote: > > Quoting Kip Macy <kmacy@fsmware.com> (Tue, 15 Aug 2006 13:10:42 -0700 (PDT)): > > > > > > > > Actually, it increases the likelihood that new bugs will increase the > > > difficulty of isolating existing bugs. > > > > > > > > > > > linux emulation on amd64 has been broken for several days. > > > > acroread and openoffice both drop core. These changes > > > > probably can't make matters worse. > > > > jhb asked for debugging infos and I tried to push people for them too. > > So far nobody helped with the necessary debug stuff. People with > > interest in getting this fixed can coordinate with jhb, it is > > independent from the SoC stuff... > > > > I've tried to get more info. Unfortunately, the core dumps > (that I've been able to get) do not reveal anything beyond > what John already knew. See Kip's quote above on why finding He also suggested to add some printf()s to the kernel. > this already elusive bug may become much more difficult. If you don't switch on the use of the new code with sysctl, you don't see a difference. acroread works on i386 as before when you don't use sysctl. The code which is known to have some issues (futexes) isn't available on amd64 (at least not the MD parts), so you will notice if acroread uses the new code or not. If it tries to use the new code for some strange reason (I would be surprised if it does), we are able to come up with a patch which disables this part very fast (actually some parts are already disabled there). Bye, Alexander. -- I CAN'T SEE DEAD PEOPLE I CAN'T SEE DEAD PEOPLE I CAN'T SEE DEAD PEOPLE I CAN'T SEE DEAD PEOPLE Bart Simpson on chalkboard in episode BABF05 http://www.Leidinger.net Alexander @ Leidinger.net: PGP ID = B0063FE7 http://www.FreeBSD.org netchild @ FreeBSD.org : PGP ID = 72077137
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060816001002.7b3709b0>