Date: Tue, 7 Mar 1995 15:30:00 -0600 (CST) From: fredriks@mcs.com (Lars Fredriksen) To: swallace@ece.uci.edu (Steven Wallace) Cc: current@FreeBSD.org Subject: Re: How to deadlock your -current system Message-ID: <m0rm6pM-000Bk6C@mercury.mcs.com> In-Reply-To: <199503071846.AA27658@balboa.eng.uci.edu> from "Steven Wallace" at Mar 7, 95 10:46:14 am
next in thread | previous in thread | raw e-mail | index | archive | help
Steven Wallace writes: > > > This looks like the mmap() ing problem into a sparse file I reported > earlier. If you create a file and seek to someplace else without > writing data and then try to map those holes, your system is surely > to crash, as your program does. > > Also there is a problem if you try to mmap() a file with one process > and then another process mmap() that file with a different non-page > offset, your system will freeze after you quit and do some swapping > activity. > > Steven > I wonder if this is what I am seing too. This happens during a cvs update, make world, sup etc. Currently my machine at home is sitting dead whith two csh and pageout_daemon all waiting on the same wchan. The routine they are in as per the wchan is file_lock(hm that might not be the right name, I am going of my rather faulty memory, but it was not ufs_lock) on an address that looks like f2000004. Is there anything I can look at while in the kernel debugger that will help you David or any other people actively looking at this? Since I don't have access to the sources until I reboot, I wanted to offer any "dynamic" information that might help. John (he says hi from far away..), suggested that I collect all the wchans, so if there is anything else let me know before 6PM central time. Lars. -- ------------------------------------------------------------------- Lars Fredriksen fredriks@mcs.com (home) lars@fredriks.pr.mcs.net (home-home) fredriks@asiago.cs.wisc.edu
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?m0rm6pM-000Bk6C>