[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: HDB style locks (was "kermit")

>>> device name has upper case letters.  System V Release 4 UUCP forms the
>>> lock file name using the major and minor device numbers rather than
>>> the device name (this is pretty sensible if you think about it).
>>> FSSTND decided against the SCO and SVR4 implementations for various
>>> reasons.

Robert Sanders wrote:

>> That's a shame.  It is a VERY sensible format, and the benefits are
>> probably worth the porting effort.

Charlie Brady <charlieb@tplrd.tpl.oz.au> writes:

> I agree with Robert - the SVR4 implementation is the only correct way to
> lock devices - by major and minor device numbers. Other methods only lock
> files, which don't have a 1-1 relationship with the underlying device.
> What reasons were there for rejecting the SVR4 implementation?

(Didn't I explain this before?)

a) There are proposed plans to change the device numbering scheme in
   Linux eventually (that might conceivably affect this).

b) Some devices share the same device, but have different major
   numbers (i.e., /dev/cua0 and /dev/ttyS0).

c) It breaks a *lot* of programs.  Most can handle traditional HDB
   style locks as described in the FSSTND.  (Note there is a typo in
   the FSSTND in this section, "padded with zeroes" should be "padded
   with spaces".  This has been fixed for the next version.)

There are other reasons, but those are the big ones.  SVR4 sounds
neat, but it is really overkill and it brings problems with it.


Daniel Quinlan  <quinlan@spectrum.cs.bucknell.edu>

Reply to: