[dpdk-dev] [PATCH] rte_mbuf: bulk allocation and freeing functions + unittest

2015-03-17 Thread Vadim Suraev
>>I don't understand the "assumes refcnt has been already decremented". >I changed to 'assumes refcnt equals 0' 1. I changed to 'assumes refcnt equals 1' 2. I have a doubt about manipulating refcnt in this function. Should it be the only check/assert or should it be responsibility of the caller?

[dpdk-dev] [PATCH] rte_mbuf: bulk allocation and freeing functions + unittest

2015-03-17 Thread Vadim Suraev
Hi, Olivier, >I don't understand the "assumes refcnt has been already decremented". I changed to 'assumes refcnt equals 0' >Adding this function is not a problem today because it is the free > function associated to rte_pktmbuf_bulk_raw_alloc(). >However, I think that the 'raw' alloc/free funct

[dpdk-dev] [PATCH] rte_mbuf: bulk allocation and freeing functions + unittest

2015-03-16 Thread Olivier MATZ
Hi Vadim, Please see some comments below. On 03/13/2015 11:14 AM, vadim.suraev at gmail.com wrote: > From: "vadim.suraev at gmail.com" > > - an API function to allocate a bulk of rte_mbufs > - an API function to free a bulk of rte_mbufs > - an API function to free a chained rte_mbuf > - unittes

[dpdk-dev] [PATCH] rte_mbuf: bulk allocation and freeing functions + unittest

2015-03-13 Thread vadim.sur...@gmail.com
From: "vadim.suraev at gmail.com" - an API function to allocate a bulk of rte_mbufs - an API function to free a bulk of rte_mbufs - an API function to free a chained rte_mbuf - unittest for aboce Signed-off-by: vadim.suraev at gmail.com --- app/test/test_mbuf.c | 73 +++