Re: usb flash drive removal (Re: Desktop NetBSD needs your help)

[Antti Kantee <pooka%netbsd.org@localhost>]

On Tue Feb 10 2009 at 20:52:07 +0100, Michael van Elst wrote:
> On Tue, Feb 10, 2009 at 06:04:21PM +0200, Antti Kantee wrote:
> > On Tue Feb 10 2009 at 07:11:22 +0100, Michael van Elst wrote:
> > > You know, that's not necessarily true. Nothing prevents you from
> > > handling such failures in a graceful manner.
> > 
> > Nothing prevents NetBSD from being bug-free.  chip-chop-chip
> 
> And that's an argument against graceful error handling?
> 
> > There is a great difference between what is and what should be.
> 
> Apparently I am asking for what should be.

I fear we will be unable to do it correctly.  In theory it is nice, but I
worry that the practical complexity will make the system less functional.
But as I always say: I'm happy to be proven wrong.

I added very simple fault injection to the rump block device code.
Anyone wanting to explore the effects deeper can at least scratch the
surface pretty easily now.

> > > And
> > > that's independent of kernel vs. userland implementation.
> > 
> > There should be no separate user and kernel implementations for kernel
> > services.
> 
> Says who? :)

Quoting you: "Antti Kantee wrote".  That would make it me ;)