event/octeontx: improve dequeue performance

switch tag wait is a costly operation as it may
translate to IOB read if core swtag cache is not updated.
Do tag switch wait only when there is a tag request on
the same hardware work slot.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

diff --git a/drivers/event/octeontx/ssovf_worker.c b/drivers/event/octeontx/ssovf_worker.c
index fcb5f31..4839f06 100644 (file)
--- a/drivers/event/octeontx/ssovf_worker.c
+++ b/drivers/event/octeontx/ssovf_worker.c
@@ -126,9 +126,9 @@ ssows_deq(void *port, struct rte_event *ev, uint64_t timeout_ticks)
 
        RTE_SET_USED(timeout_ticks);
 
-       ssows_swtag_wait(ws);
        if (ws->swtag_req) {
                ws->swtag_req = 0;
+               ssows_swtag_wait(ws);
                return 1;
        } else {
                return ssows_get_work(ws, ev);
@@ -142,9 +142,9 @@ ssows_deq_timeout(void *port, struct rte_event *ev, uint64_t timeout_ticks)
        uint64_t iter;
        uint16_t ret = 1;
 
-       ssows_swtag_wait(ws);
        if (ws->swtag_req) {
                ws->swtag_req = 0;
+               ssows_swtag_wait(ws);
        } else {
                ret = ssows_get_work(ws, ev);
                for (iter = 1; iter < timeout_ticks && (ret == 0); iter++)