Re: [uml-devel] When opening /tmp/uml.ctl fails, what do we get?

[Jeff D. <jd...@ad...>]

On Sun, Mar 04, 2007 at 08:13:39PM +0100, Blaisorblade wrote:
> Have you checked if other calls require that register_netdev was already 
> called?

Just did, nothing else depends on it.

> Hmm, I'm really not sure about what to do about this. However I'd leave that 
> for a future cleanup and just add a comment about this.
> And the reason for which init now returns a value is that pcap_init _can_ 
> fail, so the switch transport is not the only odd one.

Yeah, that slipped my mind.

> struct uml_net_private {
> -        int user[1];
> +        char user[0];
> };

Neat, that used to be illegal.

> Another note: before of this patch, you should please apply the attached two 
> ones.
> 
> *) The first (net-mac-check-cleanup.diff) checks the validity of assigned MAC 
> address, but to print a meaningful error message requires adding a local 
> buffer.
> *) The second (net_kern-eth_configure...) allows avoiding this local buffer by 
> moving code around.
> 
> I think I'm excessively paranoid about these two patches and about not yet 
> merging them, so please give a look and merge them.

I just dropped them in - they look OK at a first glance.

				Jeff

-- 
Work email - jdike at linux dot intel dot com