Messages in this thread | | | Date | Wed, 23 Feb 2000 14:43:31 +1100 | From | Richard Gooch <> | Subject | Re: [proposal] /dev/debug to host kcore (and others in the future) |
| |
H. Peter Anvin writes: > Richard Gooch wrote: > > > > H. Peter Anvin writes: > > > The memory devices would be inadvisable to move, since they are > > > "well known" device locations. ALL Unices have them. > > > > Yes, I know. That's why the patch I sent to Linus didn't do that ;-) > > But perhaps /dev/kmem, /dev/mem, /dev/port, /dev/random and > > /dev/urandom could be moved, since only Linux-specific tools would > > need to be modified. > > > > It's really only /dev/full, /dev/null and /dev/zero that are "well > > known". > > ALL Unices have /dev/mem, /dev/kmem, and several non-Linuxes have > /dev/random and /dev/urandom. Furthermore, these names are already > used by a large number of tools.
Yeah, but each Unix has it's own specialised tools which use these files. We could (if we wanted) have a planned migration.
> There really is nothing to gain by removing them, as they are > one-of-a-kind devices. It's rather different with things that can > exist in N duplicates.
Yeah, as long as the "one-of-a-kind" devices don't proliferate. Which has been known to happen (one look a /dev reveals this). Well, at least now we can stop the rot. In my role as "gatekeeper", I won't allow any new (non-directory) entries in the top-level of devfs unless there's an awfully compelling reason for it.
Regards,
Richard.... Permanent: rgooch@atnf.csiro.au Current: rgooch@ras.ucalgary.ca
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |