U32 classifier and new hashing support

[Adrian Ban]

Another reason that I've build the rtnl_u32_set_flags function is that 
in the rtnl_u32_set_flags is setting U32_ATTR_SELECTOR flag which I 
guess is not required (I don't remember anymore because the patch I've 
made in November 2011) to the terminal class.
You can take a look in the iproute2/tc command and check this.

Best regards,
Adrian Ban


On 05/25/2012 02:08 PM, Adrian Ban wrote:
> Hi,
>
> I really don't know which is the difference between the terminal class 
> and a normal one, but obvious there is a difference when you using the 
> iproute2/tc command to view the filters and classes which will show 
> you an "*" somewhere in the output.
> For this reason I've took a look into the iproute2/tc command and 
> found this TC_U32_TERMINAL flag that is set for leaf classes.
>
> If I remember correctly I got some issues with the matching of the 
> class when I the class was not set as terminal and I've applied the 
> SFQ qdisc.
>
> Hope it helps.
>
> Best regards,
> Adrian Ban
>
> On 05/25/2012 02:42 AM, A C wrote:
>> I recently pulled the git tree and saw the new u32 hashing support and
>> had a quick question about it.  Is there a reason
>> rtnl_u32_set_hashmask/rtnl_u32_set_cls_terminal are adding new keys to
>> the selector?  I don't quite understand why rtnl_u32_set_cls_terminal
>> needs to be different than rtnl_u32_set_flags(..., TC_U32_TERMINAL).
>>
>> On a side note, perhaps it would useful to include a note about the
>> importance of TC_U32_TERMINAL in the docs for u32 classid?  For the
>> most part, the docs specify everything needed to get a filter up and
>> working, but it was non-obvious to me without looking at
>> iproute2/kernel that this flag is required for the classid to go
>> anywhere.
>>
>> _______________________________________________
>> libnl mailing list
>> libnl at lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/libnl
>
>
> _______________________________________________
> libnl mailing list
> libnl at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/libnl