Re: [PATCH 2/2] NET: Multiqueue network device support.

Re: [PATCH 2/2] NET: Multiqueue network device support.

[David Miller]

From: Peter P.Waskiewicz Jr. <peter.p.waskiewicz.jr@intel.com>
Date: Mon, 09 Apr 2007 13:50:52 -0700

> From: Peter P. Waskiewicz Jr. <peter.p.waskiewicz.jr@intel.com>
> 
> Added an API and associated supporting routines for multiqueue network devices.
> This allows network devices supporting multiple TX queues to configure each
> queue within the netdevice and manage each queue independantly.  Changes to the
> PRIO Qdisc also allow a user to map multiple flows to individual TX queues,
> taking advantage of each queue on the device.
> 
> Signed-off-by: Peter P. Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
> Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>

This looks definitely better than the previous implementation,
myself and others will review this in more depth though.

> + 	alloc_size = (sizeof(struct net_device_subqueue) * queue_count);
> + 
> + 	p = kzalloc(alloc_size, GFP_KERNEL);
> + 	if (!p) {
> + 		printk(KERN_ERR "alloc_netdev: Unable to allocate queues.\n");
> + 		return NULL;
> + 	}
> + 
> + 	dev->egress_subqueue = (struct net_device_subqueue *)p;

No need for a cast, 'p' is "void *".

RE: [PATCH 2/2] NET: Multiqueue network device support.

[Waskiewicz Jr, Peter P]

> -----Original Message-----
> From: David Miller [mailto:davem@davemloft.net] 
> Sent: Monday, April 09, 2007 1:38 PM
> To: Waskiewicz Jr, Peter P
> Cc: netdev@vger.kernel.org; linux-kernel@vger.kernel.org; 
> jgarzik@pobox.com; cramerj; Kok, Auke-jan H; Leech, Christopher
> Subject: Re: [PATCH 2/2] NET: Multiqueue network device support.
> 
> From: Peter P.Waskiewicz Jr. <peter.p.waskiewicz.jr@intel.com>
> Date: Mon, 09 Apr 2007 13:50:52 -0700
> 
> > From: Peter P. Waskiewicz Jr. <peter.p.waskiewicz.jr@intel.com>
> > 
> > Added an API and associated supporting routines for 
> multiqueue network devices.
> > This allows network devices supporting multiple TX queues 
> to configure 
> > each queue within the netdevice and manage each queue 
> independantly.  
> > Changes to the PRIO Qdisc also allow a user to map multiple 
> flows to 
> > individual TX queues, taking advantage of each queue on the device.
> > 
> > Signed-off-by: Peter P. Waskiewicz Jr 
> > <peter.p.waskiewicz.jr@intel.com>
> > Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
> 
> This looks definitely better than the previous 
> implementation, myself and others will review this in more 
> depth though.
> 
> > + 	alloc_size = (sizeof(struct net_device_subqueue) * queue_count);
> > + 
> > + 	p = kzalloc(alloc_size, GFP_KERNEL);
> > + 	if (!p) {
> > + 		printk(KERN_ERR "alloc_netdev: Unable to 
> allocate queues.\n");
> > + 		return NULL;
> > + 	}
> > + 
> > + 	dev->egress_subqueue = (struct net_device_subqueue *)p;
> 
> No need for a cast, 'p' is "void *".

True, but the assignment for "dev" above also casts this void * to
struct net_device *:

        dev = (struct net_device *)
                (((long)p + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST);
        dev->padded = (char *)dev - (char *)p;

Shall I remove my cast?

Thanks for the feedback,

-PJ Waskiewicz

Re: [PATCH 2/2] NET: Multiqueue network device support.

[David Miller]

From: "Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@intel.com>
Date: Mon, 9 Apr 2007 13:43:11 -0700

> True, but the assignment for "dev" above also casts this void * to
> struct net_device *:
> 
>         dev = (struct net_device *)
>                 (((long)p + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST);
>         dev->padded = (char *)dev - (char *)p;
> 
> Shall I remove my cast?

I think so.  Casts tend to be red flags during patch review,
so don't add them when not necessary and let's remove the
unnecessary ones too.

RE: [PATCH 2/2] NET: Multiqueue network device support.

[Waskiewicz Jr, Peter P]

> -----Original Message-----
> From: David Miller [mailto:davem@davemloft.net] 
> Sent: Monday, April 09, 2007 1:47 PM
> To: Waskiewicz Jr, Peter P
> Cc: netdev@vger.kernel.org; linux-kernel@vger.kernel.org; 
> jgarzik@pobox.com; cramerj; Kok, Auke-jan H; Leech, Christopher
> Subject: Re: [PATCH 2/2] NET: Multiqueue network device support.
> 
> From: "Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@intel.com>
> Date: Mon, 9 Apr 2007 13:43:11 -0700
> 
> > True, but the assignment for "dev" above also casts this void * to 
> > struct net_device *:
> > 
> >         dev = (struct net_device *)
> >                 (((long)p + NETDEV_ALIGN_CONST) & 
> ~NETDEV_ALIGN_CONST);
> >         dev->padded = (char *)dev - (char *)p;
> > 
> > Shall I remove my cast?
> 
> I think so.  Casts tend to be red flags during patch review, 
> so don't add them when not necessary and let's remove the 
> unnecessary ones too.

Will do.  On a side note, I don't see my patches landing on vger again.
Did they get munched, or is majordomo lagging a bit?

Thanks,

-PJ Waskiewicz

Re: [PATCH 2/2] NET: Multiqueue network device support.

[David Miller]

From: "Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@intel.com>
Date: Mon, 9 Apr 2007 13:47:58 -0700

> Will do.  On a side note, I don't see my patches landing on vger again.
> Did they get munched, or is majordomo lagging a bit?

I forwarded you one of the bounces.

Re: [PATCH 2/2] NET: Multiqueue network device support.

[Herbert Xu]

Waskiewicz Jr, Peter P <peter.p.waskiewicz.jr@intel.com> wrote:
> 
> True, but the assignment for "dev" above also casts this void * to
> struct net_device *:
> 
>        dev = (struct net_device *)
>                (((long)p + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST);
>        dev->padded = (char *)dev - (char *)p;
> 
> Shall I remove my cast?

This is totally different as it's casting a "long", not a "void *".

Cheers,
-- 
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt