netdev
[Top] [All Lists]

Re: [PATCH] kill rtnl_exlock stubs

To: Stephen Hemminger <shemminger@xxxxxxxx>
Subject: Re: [PATCH] kill rtnl_exlock stubs
From: Jamal Hadi Salim <hadi@xxxxxxxx>
Date: 29 Jul 2004 09:16:28 -0400
Cc: "David S. Miller" <davem@xxxxxxxxxx>, netdev@xxxxxxxxxxx, Alexey <kuznet@xxxxxxxxxxxxx>
In-reply-to: <20040727095057.78c7419c@xxxxxxxxxxxxxxxxxxxxx>
Organization: ZNYX Networks
References: <20040727095057.78c7419c@xxxxxxxxxxxxxxxxxxxxx>
Reply-to: hadi@xxxxxxxx
Sender: netdev-bounce@xxxxxxxxxxx
On Tue, 2004-07-27 at 12:50, Stephen Hemminger wrote:

> @@ -404,13 +402,8 @@
>               return -1;
>       }
>  
> -     if (kind != 2) {
> -             if (rtnl_exlock_nowait()) {
> -                     *errp = 0;
> -                     return -1;
> -             }
> +     if (kind != 2) 
>               exclusive = 1;
> -     }
>  
>       memset(&rta, 0, sizeof(rta));
>  
> @@ -439,14 +432,10 @@
>               goto err_inval;
>       err = link->doit(skb, nlh, (void *)&rta);
>  
> -     if (exclusive)
> -             rtnl_exunlock();
>       *errp = err;
>       return err;
>  
>  err_inval:
> -     if (exclusive)
> -             rtnl_exunlock();
>       *errp = -EINVAL;
>       return -1;
>  }

This piece is the only one i would worry about.
I dont remember the historical reasoning for rtnl_ex* I know its not
useful as is right now. OTOH, if it was useful then the above code would
have meant something. Dave/Alexey?

cheers,
jamal




<Prev in Thread] Current Thread [Next in Thread>