Jukka Larja (roskakori2006.delete@this.gmail.com) on March 17, 2021 3:42 am wrote:
> dmcq (dmcq.delete@this.fano.co.uk) on March 16, 2021 4:06 pm wrote:
> > Jukka Larja (roskakori2006.delete@this.gmail.com) on March 16, 2021 11:22 am wrote:
> > > dmcq (dmcq.delete@this.fano.co.uk) on March 16, 2021 9:17 am wrote:
> > >
> > > > However then I go on to
> > > > wondering what kind of thing would mean that we didn't have megabytes of code in the kernel and yet could
> > > > do things as efficiently or better and support most current programs without disruptive changes? As far
> > > > as I can see if you want to avoid copying lots of data it seems capabilities are the way - that brings
> > > > other problems of course but I think they would be easier to manage and contain.
> > >
> > > Do you honestly think OS kernels are the most relevant (or even very relevant) security issue
> > > we are dealing with? I'm not an expert, but seems to me that just about all newsworthy security
> > > issues (that aren't about hardware) are not kernel related. Personally I'm mostly worried about
> > > stuff that will happily run with user privileges, e.g. inside a browser or email client.
> > >
> > > -JLarja
> >
> > Ah the good old appeal to worse problems.
>
> Hey, I'm not telling you not go and develop a micro-kernel OS. But it sounds to me like you think not having
> one is a big problem that should be fixed ASAP. That with micro-kernel OS some major part of security problems
> would be gone. Maybe I just misunderstood you (your writing is pretty hard to comprehend for large parts).
>
> -JLarja

Yes I think we should be doing that but I think it needs hardware support and is a long term job. I don't see that it conflicts with other work much and I think it would make some tasks easier and more efficient at user level too.
I'm sorry you find difficulty with what I write. Others have said that and I do try but succeeding is a different thing unfortunately.