Date: Mon, 20 Jan 1997 08:05:35 +0200 From: Mark Murray <mark@grondar.za> To: Peter Wemm <peter@spinner.dialix.com> Cc: Bill Paul <wpaul@freefall.freebsd.org>, mark@grondar.za (Mark Murray), current@freebsd.org, peter@freebsd.org, dyson@freebsd.org Subject: Re: VM bogon? Was: Re: NIS breakage Message-ID: <199701200605.IAA15320@grackle.grondar.za>
next in thread | raw e-mail | index | archive | help
Peter Wemm wrote: > This is looking like a manifestation of a VM system problem.. :-( We've > had a few PR's now where people report that 'vi crashes after being left > idle for 5 minutes', and so on. It all seems to be with select() or read() > etc specifying *valid* addresses that work on one time arount a loop with > ktrace, and a short while later get an EFAULT on a perfectly valid address. > Having accept() fail with an EFAULT is a new one that I've not seen > before. In this case, the two variables being copyout()'ed are on the > stack. John???? > > BTW, are these failure cases happening on P5 or P6's? Perhaps the > fpu-based "fast" P5 copyin/copyout are to blame here? Mark, if you are > seeing this on a P5 machine and can reproduce this easily, try looking up > the option for NPX_DISABLE_I586_OPTIMIZED_COPYIO in npx.c and see if you > can figure out how to configure npx's flags to disable this floating point. > I think it's "device npx0 ..... flags 4" or something like that. Perhaps > "flags 7" might be best to eliminate the possibility. This is happening on an AMD386sx/40. :-( M -- Mark Murray PGP key fingerprint = 80 36 6E 40 83 D6 8A 36 This .sig is umop ap!sdn. BC 06 EA 0E 7A F2 CE CE
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199701200605.IAA15320>