Date: Thu, 11 Dec 1997 17:30:46 -0800 From: John-Mark Gurney <gurney_j@efn.org> To: Mike Smith <mike@smith.net.au> Cc: ETX-B-SL Martti Kuparinen <erakupa@kk.etx.ericsson.se>, freebsd-hackers@FreeBSD.ORG Subject: Re: panic: npxintr from nowhere Message-ID: <19971211173046.63648@hydrogen.nike.efn.org> In-Reply-To: <199712112247.JAA01956@word.smith.net.au>; from Mike Smith on Fri, Dec 12, 1997 at 09:17:07AM %2B1030 References: <199712111644.RAA00495@kk662.kk.etx.ericsson.se> <199712112247.JAA01956@word.smith.net.au>
next in thread | previous in thread | raw e-mail | index | archive | help
Mike Smith scribbled this message on Dec 12: > > What does this message mean, i.e. what's this npxintr thing? > > It's the interrupt handler for interrupts from the FPU. > > > Position 1 > > npxintr: npxproc = 0x0, curproc = 0x0, npx_exists = 1 > > panic: npxintr from nowhere > > > > Why does it crash when/before executing the function? Memory leak > > somewhere? ``netstat -m'' does not show any leak during execution. > > > > Corrupted instruction pointer? Why does it "find" the first > > printf line? > > npxintr() is being called without a valid NPX interrupt status. Are > you calling any functions indirectly? actually, now that someone else brought it up, any good way to call floating point routins from the kernel? if you do, and don't have any processes that have used floating point, will will get the above mentioned panic... quite easy to reproduce... boot kernel, run floating point code in kernel, see it panic... now boot same kernel, run systat, run floating point code in kernel, see it run normally.. :) this was from a current kernel as of a few weeks ago... -- John-Mark Gurney Modem/FAX: +1 541 683 6954 Cu Networking Live in Peace, destroy Micro$oft, support free software, run FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19971211173046.63648>