Re: small ndis fix

[Andrew Atrens <atrens@xxxxxxxxxxxxxxxxxx>]

Matthew Dillon wrote:

> 
> :
> :On Wed, Nov 17, 2004 at 09:02:54AM -0500, Andrew Atrens wrote:
> :> 
> :> Straightforward bug fix on error path. Noticed while reviewing the ndis
> :> code ...
> :
> :Even better, the checks are always true, because malloc with M_WAITOK
> :can't return NULL :)
> :
> :Joerg
> 
>     I'm going to leave the (fixed) code in for the moment as a reminder
>     in case we need to make the allocation M_NOWAIT in the future.  I

I tend to agree. Might be better to be robust than to depend on
control-coupled behaviour.

>     doubt we ever will, though, since windoz drivers probably just assume
>     that the allocations always succeed, so maybe its overkill to leave it
>     in.

I suspect that ndis will still churn a bit as it's level of robustness isn't
perfect from what I've experienced up to now ... On my Dell laptop I get
seemingly random GPFs in the windows binary part of the driver, and the
provisioning (ifconfig) side is very shaky with my LinkSys WL565 A/B/G
card...  Still, I think it's a pretty cool piece of software :) :)

Andrew.