[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-usb
Subject:    Re: [PATCH 2/2] cdc_ncm: Admit multicast traffic
From:       Miguel_Rodríguez_Pérez <miguel () det ! uvigo ! gal>
Date:       2018-06-30 14:22:15
Message-ID: 66cb1f07-20b2-5f9b-db64-4d94a4c184fa () det ! uvigo ! gal
[Download RAW message or body]

On 30/06/18 14:17, Bjørn Mork wrote:
> Miguel Rodríguez Pérez <miguel@det.uvigo.gal> writes:
> 
>> +static void cdc_ncm_update_filter(struct usbnet *dev)
>> +{
>> +       struct cdc_ncm_ctx *ctx = (struct cdc_ncm_ctx *)dev->data[0];
>> +	u8 iface_no = ctx->control->cur_altsetting->desc.bInterfaceNumber;
>> +	struct net_device *net = dev->net;
> 
> Why not change usbnet_cdc_update_filter to get the interface number from
> dev->intf instead? Then you can export it and reuse it here instead of
> creating a copy.  And it will also work for any other usbnet minidriver.

Well, I just wanted to limit my changes to the minimum, as I lack
experience with kernel development. Where should I declare
usbnet_cdc_update then? Would linux/usb/cdc.h the right place? Is it
enough to do EXPORT_GPL_SYMBOL(usbnet_cdc_update_filter) in cdc_ether.c
for it to be loaded automatically from cdc_ncm.c?
> 
> I.e. change this:
> 
> static void usbnet_cdc_update_filter(struct usbnet *dev)
> {
> 	struct cdc_state	*info = (void *) &dev->data;
> 	struct usb_interface	*intf = info->control;
> 
> into:
> 
> static void usbnet_cdc_update_filter(struct usbnet *dev)
> {
> 	struct usb_interface	*intf = dev->intf;
> 
> 
> or simply use dev->intf to deref directly into an interface number, like
> you do in your version.  The local 'intf' variable doesn't do much
> good here.
> 
Ok, done.
> 
> 
>> +
>> +	u16 cdc_filter = USB_CDC_PACKET_TYPE_DIRECTED
>> +			| USB_CDC_PACKET_TYPE_BROADCAST;
>> +
>> +	/* filtering on the device is an optional feature and not worth
>> +	 * the hassle so we just roughly care about snooping and if any
>> +	 * multicast is requested, we take every multicast
>> +	 */
>> +	if (net->flags & IFF_PROMISC)
>> +		cdc_filter |= USB_CDC_PACKET_TYPE_PROMISCUOUS;
>> +	if (!netdev_mc_empty(net) || (net->flags & IFF_ALLMULTI))
>> +		cdc_filter |= USB_CDC_PACKET_TYPE_ALL_MULTICAST;
>> +
>> +	usbnet_write_cmd(dev,
> 
> This is a nice change. It should be probably done in
> usbnet_cdc_update_filter in any case.  Unless there is some reason it
> doesn't alreay use usbnet_write_cmd?
I don't know of any, so I'll keep it changed.
> 
>> +			USB_CDC_SET_ETHERNET_PACKET_FILTER,
>> +			USB_TYPE_CLASS | USB_DIR_OUT | USB_RECIP_INTERFACE,
>> +			cdc_filter,
>> +			iface_no,
>> +			NULL,
>> +			0);
>> +}
>> +
>>  static const struct ethtool_ops cdc_ncm_ethtool_ops = {
>>  	.get_link          = usbnet_get_link,
>>  	.nway_reset        = usbnet_nway_reset,
>> @@ -1652,6 +1679,7 @@ static const struct driver_info cdc_ncm_info = {
>>  	.status = cdc_ncm_status,
>>  	.rx_fixup = cdc_ncm_rx_fixup,
>>  	.tx_fixup = cdc_ncm_tx_fixup,
>> +	.set_rx_mode = cdc_ncm_update_filter,
>>  };
>>
>>  /* Same as cdc_ncm_info, but with FLAG_WWAN */
> 
> 
> As the comment indicates:  There are more 'struct driver_info' variants
> here.  Please add .set_rx_mode to all of them, unless you have a reason
> not to?
I simply lack access to the hardware to test and I didn't want to break
anybody else's hardware. In any case, the change seems quite safe, so I
modified them all too.

I will for your comments before sending the updated version of the patches.
> 
> 
> Bjørn
> 

-- 
Miguel Rodríguez Pérez
Laboratorio de Redes
EE Telecomunicación – Universidade de Vigo

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
[prev in list] [next in list] [prev in thread] [next in thread]