Re: [MiNT] GlueSTiK and mem protection

[Scott Bigham <dsb@cs.duke.edu>]

On Mon, 7 Jun 1999, Frank Naumann wrote:

> I analyzed the existing gluestik and the design is completly wrong.

That's a judgment call, surely.

> First, on all places you break memory protection because private kernel 
> memory is allocated.

Okay, this much I'll concede.  In my defense, I assumed that since it
was writing directly into user memory anyway, there was pretty much no
hope of getting it to run under memory protection.

> Then the complete xdd is loaded in global(!) memory. 

Only part of it really needed to be in global memory, of course, namely
the functions in the STiK-esque function table; but AFAIK it's not
possible to do a split-mem thing like that with the program text.
I mean, sure, I could have had the STiK-table functions in a TSR and
called the device driver from there, but that would have been just one
more separate thing to install.  The separate resolver was bad enough.

> The data flow looks horrible:
> 
> user program -> STiK cookie -> get direct entry points to kernel code -> 
> jump to these functions in user mode content (the function looks like a 
> normal function call)

This much is exactly like the STiK data flow, yes?

> -> these functions made system calls to the kernel 
> -> the gluestik.xdd only redirect all calls directly to the Sockets 
> (through internal kernel calls).

...er, well, yes, that was the whole point of the exercise, wasn't it?

> DNS resolving is much more complicated.

(shrug) I never said I liked having to do it that way.

> And at least there is an enourmous calling overhead.

No more so than other device drivers, AFAICT.

> I see no reason for a device driver.

The original reason for switching to a device driver IIRC was to try to
fix instability problems in the earlier TSR versions.

> The only thing that the device driver do is to redirect all calls to
> the sockets.

?  The only thing the TSR did was redirect calls to the sockets.  Why
does it matter what the thing doing the redirecting is called?

> That's why I rewritten GlueSTiK completly at the weekend.

Eek!

> GlueSTiK is now a normal TSR that is placed in global memory (but in
> user space!).

Hmm.  How does your code compare to pre-3.1 versions?

> Scott, what are the copyright conditions of GlueSTiK? I doesn't found 
> anything in the readme about this. GNU? Public Domain?

Actually, I'm quite prepared to let you take over the project if you
want.  I haven't really been in a position to be able to do anything
with it for a while now.  As for copyright... (checks documentation)
Hmm.  Apparently I never settled on anything.  I'd say GPL.

						-sbigham