Re: Raw-device-war (was Re: Sharing SCSI disks)

Randy Scott (scottr@belle.bork.com)
Tue, 1 Apr 1997 14:26:53 -0600


: > Hooks for the raw-device support should however be under kernel option
: > and should default to not to support raw-devices. I think it is not
: > wise to add straight raw-device support to kernel, because it is a
: > message to application programmers to start using their own buffer-
: > routines to optimize their application performance.

Is this a bad message to be sending? Who really cares if application
programs are using their own buffering routines? The application will
probably have a better idea of the type of usage patterns that are likely
to occur with the data than the kernel would. In addition, this leaves the
kernel's buffer cache free for data that is more likely being used by
OTHER applications in the system (for interactive users, for instance).

: mke2fs do it's work ? The discussion seems to be about
: buffered/unbuffered, and why shouldn't you supply unbuffered devices if
: someone wants them ?

Exactly...