<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>
Date   : Mon, 07 May 1984 20:40:00 PDT
From   : ACB.TYM@OFFICE-2.ARPA
Subject: BIOS buffer flush

A small contribution.  Most well behaved programs close output files before 
asking for a disk change or before returning via either warm boot or to
the CCP.
 As a result the last operation on a disk is usually a directory write
(type 1).
 If a directory write is treated as an immediate operation, the write buffer
is 
always flushed when a file is closed.  There are a very small number of 
exceptions (programs that "patch" a sector and then don't close the file, 
utilities that use the BIOS directly (copy, format)) and these are best

Changed to make their last write a type 1  write.  I have a BIOS that manages 
the buffers this way and I have never fouled a directory because of failing to 
write the directory or last sector.  Further there is no unneeded disk io at 
console read time.

For those awaiting a summary of the results of my 5 inch disk controller 
question... there were no responses.  I am thinking of the "Little Board"
as my 
solution. It seems cheaper than most s100 disk controller boards and will
serve 
as a disk copy machine quite nicely.  Anyone with experience?
<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>