Tim,

First, let me say that all of your work is awesome.  Thanks for doing
all of this.  It was certainly a pleasant surprise.

In your brk patch, I think the zero function on physical memory is not
the right thing to do.  You want to use a memory port and call
memsetBlob on that port.  Similar to how you called writeBlob in your
time() patch.  Is there some reason to think that this wouldn't work?
I'll admit that brk could be a special case, but if it's not, it's
better to use the port interface.

I haven't managed to read through all of your patches, but as the
resident anal coder, I'd like to request a few style fixes.

1) Can you make sure that you wrap lines at 80 columns?
2) I know it's probably not in our style guide, but can you make it so
namespaces don't increase the indent?  I know there's a lot of code in
the tree that is indented because of the namespace, but it's really
just a waste to do that.
3) the function name in a function definition should start on a new
line (even if the definition is within a class declaration).  e.g.
    inline void
    foo()
4) if, while, for, etc. should have a space before the paren since
they're not functions.

I realize that you copied a lot of existing code that got some of this
wrong, but it'd be nice not to add new problems to the codebase.  I
hope you don't feel this is too big a burden.

Finally, one last thing.  Can you create at least a hello world
regression test?  More would of course be appreciated, but at least
one would be nice.

Are you using mercurial queues?  My guess is that you are, but if not,
you should.

This is all great!  I'm out of town for a while, so I may not be able
to respond to questions very promptly.  Hopefully others can continue
to chime in.

Thanks,

  Nate
_______________________________________________
m5-dev mailing list
m5-dev@m5sim.org
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to