From: Danylo Vodopianov <dvo-plv@napatech.com>
Sometimes during high flow learn performance two threads
could cause deadlock using queue and mutex.
learn_ignored variable sets when flow learn and gets
when flow removes. Hence, learn_ignore should not be protected
as far as two thread could not access to the same flow
at the same time.
Signed-off-by: Danylo Vodopianov <dvo-plv@napatech.com>
Signed-off-by: Oleksandr Kolomeiets <okl-plv@napatech.com>
---
drivers/net/ntnic/include/flow_api_engine.h | 2 +-
.../nthw/flow_api/profile_inline/flow_api_profile_inline.c | 7 +++----
2 files changed, 4 insertions(+), 5 deletions(-)
@@ -297,7 +297,7 @@ struct flow_handle {
enum flow_handle_type type;
uint32_t flm_id;
uint16_t caller_id;
- uint16_t learn_ignored;
+ RTE_ATOMIC(uint16_t) learn_ignored;
struct flow_eth_dev *dev;
struct flow_handle *next;
@@ -518,9 +518,8 @@ static void flm_mtr_read_sta_records(struct flow_eth_dev *dev, uint32_t *data, u
uint8_t port;
bool remote_caller = is_remote_caller(caller_id, &port);
- rte_spinlock_lock(&dev->ndev->mtx);
- ((struct flow_handle *)flm_h.p)->learn_ignored = 1;
- rte_spinlock_unlock(&dev->ndev->mtx);
+ rte_atomic_store_explicit(&((struct flow_handle *)flm_h.p)->learn_ignored,
+ 1, rte_memory_order_seq_cst);
struct flm_status_event_s data = {
.flow = flm_h.p,
.learn_ignore = sta_data->lis,
@@ -972,7 +971,7 @@ static int flm_flow_programming(struct flow_handle *fh, uint32_t flm_op)
if (flm_op == NT_FLM_OP_UNLEARN) {
ntnic_id_table_free_id(fh->dev->ndev->id_table_handle, flm_id);
- if (fh->learn_ignored == 1)
+ if (rte_atomic_load_explicit(&fh->learn_ignored, rte_memory_order_seq_cst) == 1)
return 0;
}