Re: [net-next 30/40] ixgbe: add support for nfc addition and removal of filters

[Alexander Duyck <alexander.h.duyck@intel.com>]

On 06/07/2011 02:24 PM, Ben Hutchings wrote:
> On Tue, 2011-06-07 at 05:33 -0700, Jeff Kirsher wrote:
>> From: Alexander Duyck<alexander.h.duyck@intel.com>
>>
>> This change is meant to allow for nfc to insert and remove filters in order
>> to test the ethtool interface which includes it's own rules manager.
>>
>> Signed-off-by: Alexander Duyck<alexander.h.duyck@intel.com>
>> Tested-by: Ross Brattain<ross.b.brattain@intel.com>
>> Signed-off-by: Jeff Kirsher<jeffrey.t.kirsher@intel.com>
>> ---
>>   drivers/net/ixgbe/ixgbe_ethtool.c |  245 +++++++++++++++++++++++++++++++++++++
>>   drivers/net/ixgbe/ixgbe_main.c    |   45 +++++++
>>   2 files changed, 290 insertions(+), 0 deletions(-)
>>
>> diff --git a/drivers/net/ixgbe/ixgbe_ethtool.c b/drivers/net/ixgbe/ixgbe_ethtool.c
>> index 2c70363..f37c9b8 100644
>> --- a/drivers/net/ixgbe/ixgbe_ethtool.c
>> +++ b/drivers/net/ixgbe/ixgbe_ethtool.c
> [...]
>> +	/* Copy input into formatted structures */
>> +	input->filter.formatted.src_ip[0] = fsp->h_u.tcp_ip4_spec.ip4src;
>> +	mask.formatted.src_ip[0] = fsp->m_u.tcp_ip4_spec.ip4src;
>> +	input->filter.formatted.dst_ip[0] = fsp->h_u.tcp_ip4_spec.ip4dst;
>> +	mask.formatted.dst_ip[0] = fsp->m_u.tcp_ip4_spec.ip4dst;
>> +	input->filter.formatted.src_port = fsp->h_u.tcp_ip4_spec.psrc;
>> +	mask.formatted.src_port = fsp->m_u.tcp_ip4_spec.psrc;
>> +	input->filter.formatted.dst_port = fsp->h_u.tcp_ip4_spec.pdst;
>> +	mask.formatted.dst_port = fsp->m_u.tcp_ip4_spec.pdst;
>> +
>> +	if (fsp->flow_type&  FLOW_EXT) {
>> +		input->filter.formatted.vm_pool =
>> +				(unsigned char)ntohl(fsp->h_ext.data[1]);
>> +		mask.formatted.vm_pool =
>> +				(unsigned char)ntohl(fsp->m_ext.data[1]);
>> +		input->filter.formatted.vlan_id = fsp->h_ext.vlan_tci;
>> +		mask.formatted.vlan_id = fsp->m_ext.vlan_tci;
>> +		input->filter.formatted.flex_bytes =
>> +						fsp->h_ext.vlan_etype;
>> +		mask.formatted.flex_bytes = fsp->m_ext.vlan_etype;
>> +	}
> [...]
>
> Sure you don't need any byte-swapping for the IP, port, and VLAN tag?
>
> Also is the 'vlan_id' field used to match the whole VLAN/priority tag or
> only the VID?  If it only matches the VID then you need to check that
> the priority and CF bits are not set in the given mask.
>
> Ben.
I'm pretty sure that is all correct.  The input filter for the Flow 
Director hashes has everything in big-endian byte order.  Most of the 
input fields in the hardware for the perfect filters also accept things 
in big-endian byte order.  So it is in my best interest to just leave 
things in that byte order since that way I have to do as few byte swaps 
as possible.

The VLAN field can include the VLAN priority.  There is a check of the 
input masks done that verifies that the valid VLAN masks can be 0xEFFF, 
0xE000, 0x0FFF, and 0x0000.  This mask is anded against the input data 
when setting up the filter so there isn't any way that a input with a CF 
bit would be valid.

Thanks,

Alex