Re: 2.3.39 compile errors on Alpha

[Alan Cox]

On Mon, 2002-09-30 at 13:25, David S. Miller wrote:
>    From: Alan Cox <[e-mail]>
>    Date: 30 Sep 2002 13:35:40 +0100
> 
>    Is this actually safe - suppose the machine has no tsc counter (eg old
>    x86 or indeed new x86 numa, speedstep using, etc). In that case
>    do_gettimeofday doesn't appear to be either IRQ safe or fast enough to
>    use in this way ?
> 
> This is how netif_rx() has worked for a long time.  ATM is just
> copying the input packet logic.
> 
> So why are you complaining now? :-)

Because I was looking over the gettimeoffset code and forgot that
gettimeofday itself takes the xtime_lock 8)

Alan