Return to BSD News archive
Path: sserve!manuel!munnari.oz.au!spool.mu.edu!agate!soda.berkeley.edu!wjolitz From: wjolitz@soda.berkeley.edu (William F. Jolitz) Newsgroups: comp.unix.bsd Subject: Re: 386BSD hangs Date: 6 Sep 1992 00:27:52 GMT Organization: U.C. Berkeley, CS Undergraduate Association Lines: 42 Message-ID: <18bja8INNjt2@agate.berkeley.edu> References: <ZCO6TOH@desert.in-berlin.de> <david.715668854@mlb.geomechanics.csiro.au> NNTP-Posting-Host: soda.berkeley.edu In article <david.715668854@mlb.geomechanics.csiro.au> david@mlb.geomechanics.csiro.au (David Le Blanc) writes: >hell@desert.in-berlin.de (Steffen Hellwig) writes: > >>Hallo world. >>I've some problems running 386BSD on my 386 40MHZ with an ADAPTEC 1542A >>I've installed 386BSD sucsessfully on a Fujutsu M2624FA SCSI Disk and every >>thinks works fine. >>But sometimes, most often during heavy disk-io, the machine hangs without >>any comments. > >>Anyone else noticed that strange thing? > >I have had that happening. It usually happens when I do two large things >such as run GCC twice while under Xwindows.. The swapping will halt, and >the machine will go quiet. And stay that way. It most likely you have too little memory and it's paging itself into a deficit. The current page replacement algorithm isn't up to the task for systems with 4 MB RAM or less. The page replacement algorithm is being reworked (it's horrible) to operate correctly, but it's slow going. I suggest that in the meantime if you wish to run very memory intensive operations (such as windows and gcc), you install more RAM. I do much more intensive operations than these, but on 8 MB, 16 MB, and 20 MB systems, and do not experience this problem. BTW, if you use 8MB+ systems, make sure you either use a patched and recompiled kernel for the 8MB stability bug, or insert and recompile the fix yourself. Otherwise, you may get silent hangs for a completely different reason. >Also, I have had problems where while swapping the IDE drive (which has >the swap partition, I have /usr files on a 'D:' SCSI drive (adaptec 1542b)) >will pause, with the activity light on solid for maybe five seconds, then >resume. Any ideas? Sounds like a bug in the SCSI driver...The pause is because it gets retried and eventually gets through. The SCSI driver does not obtain the status at the appropriate time and causes an operation to be aborted and retried... until it works. Lynne.