Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!bunyip.cc.uq.oz.au!harbinger.cc.monash.edu.au!yeshua.marcam.com!news.kei.com!sol.ctr.columbia.edu!howland.reston.ans.net!cs.utexas.edu!utah-morgan!hellgate.utah.edu!fcom.cc.utah.edu!u.cc.utah.edu!cs.weber.edu!terry From: terry@cs.weber.edu (A Wizard of Earth C) Newsgroups: comp.os.386bsd.development Subject: Re: [FreeBSD 1.0R] DMA Problems? Date: 5 Jan 1994 00:39:21 GMT Organization: Weber State University, Ogden, UT Lines: 58 Message-ID: <2gd27p$dc6@u.cc.utah.edu> References: <2fl24q$jn2@u.cc.utah.edu> <JTW.94Jan3234318@pmws.lcs.mit.edu> <jmonroyCJ3u56.1us@netcom.com> NNTP-Posting-Host: cs.weber.edu In article <jmonroyCJ3u56.1us@netcom.com> jmonroy@netcom.com (Jesus Monroy Jr) writes: >John Wroclawski (jtw@lcs.mit.edu) wrote: John> It is impossible to -probe- for a minimum DRAM refresh rate because John> the answer is not binary. If you avoid refresh for long enough, yes, John> the data will flat-out be lost. But long before that the DRAM will John> simply become much more sensitive to power supply noise, cosmic rays, John> lightning, and all of the other things that occasionally cause DRAMS John> to lose data. Unless you can duplicate all of these external events John> during your probe operation, you cannot even begin to predict when the John> RAM will fail. > I'm sorry for this reply... but > > Sagitarius (sp?) won't align with the moon for at least > four days.... Can I give you an answer then? > > What is this "cosmic ray" stuff? > Somebody please tell me if I should take this seriously? Yes, you should take it very seriously. Here, John is demonstrating a better knowledge of solid state physics than most coders who dabble in hardware drivers. Cosmic Rays are high energy particles from outer space -- you may have heard of "the solar wind", and if you have ever worked on satellite or other aerospace systems, you will probably be aware of a process called "hardening", whose intent it is to make the harware more tolerant of radiation events. A more proper term for an earthy engineer might have been "background radiation", which includes both radiated particles as decay products of matter and what cosmic radiation hits us through the magnetosphere. If you have ever taken an introductory high energy physics class and seen a gold-leaf electroscpe (a capacitive device that discharges more quickly in the presence of radiation -- it was used as a crude detector by people like Mme. Curie) and if you know anything about how dynamic RAM actuall works (using a capacitance well) or *why* it needs to be refreshed (tunneling + other leakage) in the first place, then you will know that you can not predict how a DRAM will act above and beyond the guaranteed minimal operational parameters. John knows what he is talking about with regard to noise sources, even if you have never had to consider the issues. This type of "probing" was *exactly* what I felt could not be done reliably as well -- and why I believe screwing with refresh has nothing to do with your bus-on/bus-off problems getting resolved. You may not respect the EE's who build the things, but you have to respect the physics behind it, or you will constantly run into one "mysterious" problem after another. Terry Lambert terry@cs.weber.edu --- Any opinions in this posting are my own and not those of my present or previous employers.