![](https://seccdn.libravatar.org/avatar/d09bcfa0e73bd3815a4a73cd8d7e87bb.jpg?s=120&d=mm&r=g)
On 16-Feb-98 Bodo Bauer wrote:
Ted Harding wrote:
I think it makes just as much sense to write to /dev/zero as to /dev/null. The differences
I don't think so. /dev/null is for writing and /dev/zero for reading of useless data.
is in reading. If setting idiosyncratic permissions is likely to cause several programs to misbehave, then please leave them as 666!
But anyway, may be you are right and we should turn the permissions to 666. I don't think it's a security hole to have it writable for everyone.
Ciao, BB
I've now had a look at some UNIX and Linux systems around here. SunOS has 666
on /dev/zero; an older Red HAT, a current Debian, and a current FreeBSD all
also have 666. So far, S.u.S.E is the only exception!
However, the keeper of the local Debian (Owen Leblanc of fame) was emphatic that
666 is not a good idea for /dev/zero since "writing to /dev/zero is not defined
in the standards and so it doesn't make sense" (though I have yet to check
POSIX or Linux FSSTND). On that basis his view also was that WABI is doing the
wrong thing.
In the case of undefined effect, writing to /dev/zero *could* have
unpredictable results; but presumably a system that allows 666 (e.g. SunOS)
would have made provision so that this did not cause trouble. Linux too?
Best wishes to all,
Ted.
--------------------------------------------------------------------
E-Mail: Ted Harding