[Info-vax] How do I make zip, unzip etc. available to all users?

[John Reagan]

On Saturday, January 9, 2016 at 10:30:54 AM UTC-5, abrsvc wrote:
> > Live patches have been around for Unix for a while with now-at-Oracle 
> > Ksplice, and similar capabilities has been added to Linux 4.   That 
> > kernel patching infrastructure is, of course, a non-trivial part of how 
> > far behind OpenVMS is, in this area.   OpenVMS can't manage to unload 
> > device drivers, for that matter.
> 
> I'm not convinced that allowing kernel level patching is a god thing with a running system.  Doesn't that provide a rather large invitation to problems? How much easier would it be to install viruses at the OS level should this be allowed?
> 
> Dan

I've looked at things like Ksplice at depth.  From a compiler perspective, it is little more than leaving some dead space in the prologue in which you can later insert an indirect call later on.

It certainly can be helpful in splicing in bug fixes.  It gets somewhat more complicated if you are splicing in a fix (or enhancement) that modifies some shared data structure but there are some tricks to help with that.

In a multi-threaded/multi-cpu environment, you have to worry about both the old and new code executing at the same time (at least for a while), etc.


As for inserting viruses, etc, I doubt it is any more risky than doing some INSTALL/REPLACE on a normal image (especially one with priv's).