[dpdk-users] rte_pktmbuf_alloc stuck forever in dpdk 18.05.1

[Manish Kumar]

Hello Team,

I have been trying dpdk 18.05.1 on a small project which was first written
for dpdk 16.11 release.
It is a primary secondary process architecture where primary process is
reading from the NIC and then sending it to secondary process where I am
calling rte_pktmbuf_alloc for a mempool which was created in the primary
process. In dpdk 16.11 eveything is working fine, however in dpdk 18.05.1 ,
as soon as rte_pktmbuf_alloc API is called, it gets stuck in stack_dequeue
function and keeps in that state forever.  Can anybody help in identifying
what could be causing it.

#0 0x00007f7ef232d9e0 in stack_dequeue ()
from /usr/lib64/librte_mempool_stack.so.1.1
#1 0x00000000004110de in rte_mempool_ops_dequeue_bulk (n=33,
obj_table=0x7f4f0c735450, mp=0x7f4f0c6fef00)
at
/home/manishk/dpdk-latest/x86_64-native-linuxapp-gcc/include/rte_mempool.h:679
#2 __mempool_generic_get (cache=0x7f4f0c735440, n=1,
obj_table=0x7f7ee9d41b18, mp=0x7f4f0c6fef00)
at
/home/manishk/dpdk-latest/x86_64-native-linuxapp-gcc/include/rte_mempool.h:1495
#3 rte_mempool_generic_get (cache=0x7f4f0c735440, n=1,
obj_table=0x7f7ee9d41b18, mp=0x7f4f0c6fef00)
at
/home/manishk/dpdk-latest/x86_64-native-linuxapp-gcc/include/rte_mempool.h:1558
#4 rte_mempool_get_bulk (n=1, obj_table=0x7f7ee9d41b18, mp=0x7f4f0c6fef00)
at
/home/manishk/dpdk-latest/x86_64-native-linuxapp-gcc/include/rte_mempool.h:1591
#5 rte_mempool_get (obj_p=0x7f7ee9d41b18, mp=0x7f4f0c6fef00)
at
/home/manishk/dpdk-latest/x86_64-native-linuxapp-gcc/include/rte_mempool.h:1617
#6 rte_mbuf_raw_alloc (mp=0x7f4f0c6fef00)
at
/home/manishk/dpdk-latest/x86_64-native-linuxapp-gcc/include/rte_mbuf.h:992
#7 rte_pktmbuf_alloc (mp=0x7f4f0c6fef00)
at
/home/manishk/dpdk-latest/x86_64-native-linuxapp-gcc/include/rte_mbuf.h:1253


Thank you so much in advance.

Thanks
Manish