On 20/01/15 6:34 pm, Bill Fischofer wrote:
The idea behind this is to allow incoming packets to be sorted into
different pools based on their lengths for storage efficiency. For
example, you might have three pools of sizes 256 bytes, 1500 bytes,
and 9200 bytes and sort arriving packets into them based on their
lengths. So each arriving packet winds up being stored as a single
segment in an appropriately-sized pool.
Additional minor point to note,
If an implementation cannot support ODP_PMR_LEN the same can be returned
to the application when the application queries the supported PMR terms
using the API odp_pmr_terms_cap() which returns a bit mask for each of
the supported odp_pmr_term_e
Regards,
Bala
On Tue, Jan 20, 2015 at 2:54 AM, Radu-Andrei Bulie
<[email protected] <mailto:[email protected]>> wrote:
Hi,
In classification there is a field named –/ODP_PMR_LEN//./ Is it
used for normal classification as the other fields or it has the
purpose of selecting or
changing the buffer pool? (what I mean to say is that for example
this field is used to classify frames and then send them to a
specific buffer pool with buffers
of specific length). Why would one classify based on this field?
(as I understood this field doesn’t mean packet len, but frame length)
Regards,
Radu
_______________________________________________
lng-odp mailing list
[email protected] <mailto:[email protected]>
http://lists.linaro.org/mailman/listinfo/lng-odp
_______________________________________________
lng-odp mailing list
[email protected]
http://lists.linaro.org/mailman/listinfo/lng-odp
_______________________________________________
lng-odp mailing list
[email protected]
http://lists.linaro.org/mailman/listinfo/lng-odp
