Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!psuvax1!uwm.edu!zaphod.mps.ohio-state.edu!mips!daver!bungi.com!news From: culberts@hplwbc.hpl.hp.com (Bruce Culbertson) Newsgroups: comp.sys.nsc.32k Subject: Minix bug fix Message-ID: <9007231749.AA00495@hplwbc.hpl.hp.com> Date: 23 Jul 90 17:49:42 GMT Sender: news@daver.bungi.com Lines: 128 Approved: news@daver.bungi.com This is a PC532 Minix bug fix. John Connen plugged 8 megabytes into his pc532 and Minix ceased to run. Klib32k.s has a routine, mem_size(), which does a write, read, and compare in successive pages to find the end of RAM. Unfortunately, I didn't notice that the address decoding causes the RAM to re-appear multiple times all the way up to 128M. Minix thought it had found 128 megabytes of RAM and this caused it considerable trouble later on. John reports that the following fix makes Minix happy with either 4M or 8M. Lets hope this works with 16M and 32M, also. Kermit has been working reliably for me lately. (I do not know why I had trouble earlier.) You might try Kermit as an alternate method of loading your disk. Bruce ------------------------- cut ------------------------------- CRC's 06990 4691 /tmp/tmp (old) 56041 5196 klib32k.s (new) ------------------------- cut ------------------------------- begin 666 klib32k.s.cdif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end