We watched a rte panic of mbuf_autotest in our qualcomm arm64 server due
to a possible race condition. To fix this race condition, rmb() is needed to add
between the 2 loads.
Already fuctionally tested on the machines as follows:
- on X86
- on arm64
---
Changelog
V5: split it into 2 patchset due to the milestone concerns, this is the 1st one
V4: split into small patches
V3: arch specific implementation for enqueue/dequeue barrier
V2: let users choose whether using load_acquire/store_release
V1: rte_smp_rmb() between 2 loads
Jia He (4):
eal/arm64: remove the braces {} for dmb() and dsb()
ring: guarantee load/load order in enqueue and dequeue
ring: introduce new header file to include common functions
ring: introduce new header file to support C11 memory model
config/common_armv8a_linuxapp | 2 +
.../common/include/arch/arm/rte_atomic_64.h | 4 +-
lib/librte_eventdev/rte_event_ring.h | 6 +-
lib/librte_ring/Makefile | 4 +-
lib/librte_ring/rte_ring.h | 161 ++---------------
lib/librte_ring/rte_ring_c11_mem.h | 185 ++++++++++++++++++++
lib/librte_ring/rte_ring_generic.h | 194 +++++++++++++++++++++
7 files changed, 404 insertions(+), 152 deletions(-)
create mode 100644 lib/librte_ring/rte_ring_c11_mem.h
create mode 100644 lib/librte_ring/rte_ring_generic.h
--
2.7.4
