[comp.unix.wizards] Do multi-sector disk transfers need to be atomic?

mangler@cit-vax.Caltech.Edu (System Mangler) (01/04/87)

I'm thinking of changing the way 4.3bsd hp.c does its bad block
replacement, to allow other I/O requests to be satisfied in between
retries and revectors.	This means that multi-sector transfers
will not necessarily be atomic.  Is there anything that depends
upon disk transfers being atomic?

Don Speck   speck@vlsi.caltech.edu  {seismo,rutgers,ames}!cit-vax!speck