X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fevent%2Focteontx2%2Fotx2_tim_evdev.c;h=4c24cc8a6a1a2c55ef21273897881ae1741d9119;hb=da17b003f3648c400b20eef2db53adae6e3ee0d6;hp=5517a6bc440de7397d4670c9ac8ed606f5898fa9;hpb=464060c02fea98960115349d8eca18d82e9dc29f;p=dpdk.git

diff --git a/drivers/event/octeontx2/otx2_tim_evdev.c b/drivers/event/octeontx2/otx2_tim_evdev.c
index 5517a6bc44..4c24cc8a6a 100644
--- a/drivers/event/octeontx2/otx2_tim_evdev.c
+++ b/drivers/event/octeontx2/otx2_tim_evdev.c
@@ -11,6 +11,65 @@
 
 static struct rte_event_timer_adapter_ops otx2_tim_ops;
 
+static inline int
+tim_get_msix_offsets(void)
+{
+	struct otx2_tim_evdev *dev = tim_priv_get();
+	struct otx2_mbox *mbox = dev->mbox;
+	struct msix_offset_rsp *msix_rsp;
+	int i, rc;
+
+	/* Get TIM MSIX vector offsets */
+	otx2_mbox_alloc_msg_msix_offset(mbox);
+	rc = otx2_mbox_process_msg(mbox, (void *)&msix_rsp);
+
+	for (i = 0; i < dev->nb_rings; i++)
+		dev->tim_msixoff[i] = msix_rsp->timlf_msixoff[i];
+
+	return rc;
+}
+
+static void
+tim_set_fp_ops(struct otx2_tim_ring *tim_ring)
+{
+	uint8_t prod_flag = !tim_ring->prod_type_sp;
+
+	/* [MOD/AND] [DFB/FB] [SP][MP]*/
+	const rte_event_timer_arm_burst_t arm_burst[2][2][2][2] = {
+#define FP(_name, _f4, _f3, _f2, _f1, flags) \
+		[_f4][_f3][_f2][_f1] = otx2_tim_arm_burst_ ## _name,
+TIM_ARM_FASTPATH_MODES
+#undef FP
+	};
+
+	const rte_event_timer_arm_tmo_tick_burst_t arm_tmo_burst[2][2][2] = {
+#define FP(_name, _f3, _f2, _f1, flags) \
+		[_f3][_f2][_f1] = otx2_tim_arm_tmo_tick_burst_ ## _name,
+TIM_ARM_TMO_FASTPATH_MODES
+#undef FP
+	};
+
+	otx2_tim_ops.arm_burst =
+		arm_burst[tim_ring->enable_stats][tim_ring->optimized]
+			[tim_ring->ena_dfb][prod_flag];
+	otx2_tim_ops.arm_tmo_tick_burst =
+		arm_tmo_burst[tim_ring->enable_stats][tim_ring->optimized]
+			[tim_ring->ena_dfb];
+	otx2_tim_ops.cancel_burst = otx2_tim_timer_cancel_burst;
+}
+
+static void
+otx2_tim_ring_info_get(const struct rte_event_timer_adapter *adptr,
+		       struct rte_event_timer_adapter_info *adptr_info)
+{
+	struct otx2_tim_ring *tim_ring = adptr->data->adapter_priv;
+
+	adptr_info->max_tmo_ns = tim_ring->max_tout;
+	adptr_info->min_resolution_ns = tim_ring->tck_nsec;
+	rte_memcpy(&adptr_info->conf, &adptr->data->conf,
+		   sizeof(struct rte_event_timer_adapter_conf));
+}
+
 static void
 tim_optimze_bkt_param(struct otx2_tim_ring *tim_ring)
 {
@@ -65,6 +124,7 @@ tim_chnk_pool_create(struct otx2_tim_ring *tim_ring,
 	char pool_name[25];
 	int rc;
 
+	cache_sz /= rte_lcore_count();
 	/* Create chunk pool. */
 	if (rcfg->flags & RTE_EVENT_TIMER_ADAPTER_F_SP_PUT) {
 		mp_flags = MEMPOOL_F_SP_PUT | MEMPOOL_F_SC_GET;
@@ -79,10 +139,9 @@ tim_chnk_pool_create(struct otx2_tim_ring *tim_ring,
 		cache_sz = RTE_MEMPOOL_CACHE_MAX_SIZE;
 
 	if (!tim_ring->disable_npa) {
-		/* NPA need not have cache as free is not visible to SW */
 		tim_ring->chunk_pool = rte_mempool_create_empty(pool_name,
 				tim_ring->nb_chunks, tim_ring->chunk_sz,
-				0, 0, rte_socket_id(), mp_flags);
+				cache_sz, 0, rte_socket_id(), mp_flags);
 
 		if (tim_ring->chunk_pool == NULL) {
 			otx2_err("Unable to create chunkpool.");
@@ -195,8 +254,7 @@ otx2_tim_ring_create(struct rte_event_timer_adapter *adptr)
 	struct tim_ring_req *free_req;
 	struct tim_lf_alloc_req *req;
 	struct tim_lf_alloc_rsp *rsp;
-	uint64_t nb_timers;
-	int rc;
+	int i, rc;
 
 	if (dev == NULL)
 		return -ENODEV;
@@ -240,11 +298,24 @@ otx2_tim_ring_create(struct rte_event_timer_adapter *adptr)
 	tim_ring->tck_nsec = RTE_ALIGN_MUL_CEIL(rcfg->timer_tick_ns, 10);
 	tim_ring->max_tout = rcfg->max_tmo_ns;
 	tim_ring->nb_bkts = (tim_ring->max_tout / tim_ring->tck_nsec);
-	tim_ring->chunk_sz = OTX2_TIM_RING_DEF_CHUNK_SZ;
-	nb_timers = rcfg->nb_timers;
+	tim_ring->chunk_sz = dev->chunk_sz;
+	tim_ring->nb_timers = rcfg->nb_timers;
 	tim_ring->disable_npa = dev->disable_npa;
+	tim_ring->enable_stats = dev->enable_stats;
+
+	for (i = 0; i < dev->ring_ctl_cnt ; i++) {
+		struct otx2_tim_ctl *ring_ctl = &dev->ring_ctl_data[i];
 
-	tim_ring->nb_chunks = nb_timers / OTX2_TIM_NB_CHUNK_SLOTS(
+		if (ring_ctl->ring == tim_ring->ring_id) {
+			tim_ring->chunk_sz = ring_ctl->chunk_slots ?
+				((uint32_t)(ring_ctl->chunk_slots + 1) *
+				 OTX2_TIM_CHUNK_ALIGNMENT) : tim_ring->chunk_sz;
+			tim_ring->enable_stats = ring_ctl->enable_stats;
+			tim_ring->disable_npa = ring_ctl->disable_npa;
+		}
+	}
+
+	tim_ring->nb_chunks = tim_ring->nb_timers / OTX2_TIM_NB_CHUNK_SLOTS(
 							tim_ring->chunk_sz);
 	tim_ring->nb_chunk_slots = OTX2_TIM_NB_CHUNK_SLOTS(tim_ring->chunk_sz);
 
@@ -256,7 +327,11 @@ otx2_tim_ring_create(struct rte_event_timer_adapter *adptr)
 			tim_optimze_bkt_param(tim_ring);
 	}
 
-	tim_ring->nb_chunks = tim_ring->nb_chunks * tim_ring->nb_bkts;
+	if (tim_ring->disable_npa)
+		tim_ring->nb_chunks = tim_ring->nb_chunks * tim_ring->nb_bkts;
+	else
+		tim_ring->nb_chunks = tim_ring->nb_chunks + tim_ring->nb_bkts;
+
 	/* Create buckets. */
 	tim_ring->bkt = rte_zmalloc("otx2_tim_bucket", (tim_ring->nb_bkts) *
 				    sizeof(struct otx2_tim_bkt),
@@ -289,10 +364,27 @@ otx2_tim_ring_create(struct rte_event_timer_adapter *adptr)
 	tim_ring->base = dev->bar2 +
 		(RVU_BLOCK_ADDR_TIM << 20 | tim_ring->ring_id << 12);
 
+	rc = tim_register_irq(tim_ring->ring_id);
+	if (rc < 0)
+		goto chnk_mem_err;
+
 	otx2_write64((uint64_t)tim_ring->bkt,
 		     tim_ring->base + TIM_LF_RING_BASE);
 	otx2_write64(tim_ring->aura, tim_ring->base + TIM_LF_RING_AURA);
 
+	/* Set fastpath ops. */
+	tim_set_fp_ops(tim_ring);
+
+	/* Update SSO xae count. */
+	sso_updt_xae_cnt(sso_pmd_priv(dev->event_dev), (void *)tim_ring,
+			 RTE_EVENT_TYPE_TIMER);
+	sso_xae_reconfigure(dev->event_dev);
+
+	otx2_tim_dbg("Total memory used %"PRIu64"MB\n",
+			(uint64_t)(((tim_ring->nb_chunks * tim_ring->chunk_sz)
+			+ (tim_ring->nb_bkts * sizeof(struct otx2_tim_bkt))) /
+			BIT_ULL(20)));
+
 	return rc;
 
 chnk_mem_err:
@@ -306,6 +398,97 @@ rng_mem_err:
 	return rc;
 }
 
+static void
+otx2_tim_calibrate_start_tsc(struct otx2_tim_ring *tim_ring)
+{
+#define OTX2_TIM_CALIB_ITER	1E6
+	uint32_t real_bkt, bucket;
+	int icount, ecount = 0;
+	uint64_t bkt_cyc;
+
+	for (icount = 0; icount < OTX2_TIM_CALIB_ITER; icount++) {
+		real_bkt = otx2_read64(tim_ring->base + TIM_LF_RING_REL) >> 44;
+		bkt_cyc = rte_rdtsc();
+		bucket = (bkt_cyc - tim_ring->ring_start_cyc) /
+							tim_ring->tck_int;
+		bucket = bucket % (tim_ring->nb_bkts);
+		tim_ring->ring_start_cyc = bkt_cyc - (real_bkt *
+							tim_ring->tck_int);
+		if (bucket != real_bkt)
+			ecount++;
+	}
+	tim_ring->last_updt_cyc = bkt_cyc;
+	otx2_tim_dbg("Bucket mispredict %3.2f distance %d\n",
+		     100 - (((double)(icount - ecount) / (double)icount) * 100),
+		     bucket - real_bkt);
+}
+
+static int
+otx2_tim_ring_start(const struct rte_event_timer_adapter *adptr)
+{
+	struct otx2_tim_ring *tim_ring = adptr->data->adapter_priv;
+	struct otx2_tim_evdev *dev = tim_priv_get();
+	struct tim_enable_rsp *rsp;
+	struct tim_ring_req *req;
+	int rc;
+
+	if (dev == NULL)
+		return -ENODEV;
+
+	req = otx2_mbox_alloc_msg_tim_enable_ring(dev->mbox);
+	req->ring = tim_ring->ring_id;
+
+	rc = otx2_mbox_process_msg(dev->mbox, (void **)&rsp);
+	if (rc < 0) {
+		tim_err_desc(rc);
+		goto fail;
+	}
+#ifdef RTE_ARM_EAL_RDTSC_USE_PMU
+	uint64_t tenns_stmp, tenns_diff;
+	uint64_t pmu_stmp;
+
+	pmu_stmp = rte_rdtsc();
+	asm volatile("mrs %0, cntvct_el0" : "=r" (tenns_stmp));
+
+	tenns_diff = tenns_stmp - rsp->timestarted;
+	pmu_stmp = pmu_stmp - (NSEC2TICK(tenns_diff  * 10, rte_get_timer_hz()));
+	tim_ring->ring_start_cyc = pmu_stmp;
+#else
+	tim_ring->ring_start_cyc = rsp->timestarted;
+#endif
+	tim_ring->tck_int = NSEC2TICK(tim_ring->tck_nsec, rte_get_timer_hz());
+	tim_ring->tot_int = tim_ring->tck_int * tim_ring->nb_bkts;
+	tim_ring->fast_div = rte_reciprocal_value_u64(tim_ring->tck_int);
+
+	otx2_tim_calibrate_start_tsc(tim_ring);
+
+fail:
+	return rc;
+}
+
+static int
+otx2_tim_ring_stop(const struct rte_event_timer_adapter *adptr)
+{
+	struct otx2_tim_ring *tim_ring = adptr->data->adapter_priv;
+	struct otx2_tim_evdev *dev = tim_priv_get();
+	struct tim_ring_req *req;
+	int rc;
+
+	if (dev == NULL)
+		return -ENODEV;
+
+	req = otx2_mbox_alloc_msg_tim_disable_ring(dev->mbox);
+	req->ring = tim_ring->ring_id;
+
+	rc = otx2_mbox_process(dev->mbox);
+	if (rc < 0) {
+		tim_err_desc(rc);
+		rc = -EBUSY;
+	}
+
+	return rc;
+}
+
 static int
 otx2_tim_ring_free(struct rte_event_timer_adapter *adptr)
 {
@@ -317,6 +500,8 @@ otx2_tim_ring_free(struct rte_event_timer_adapter *adptr)
 	if (dev == NULL)
 		return -ENODEV;
 
+	tim_unregister_irq(tim_ring->ring_id);
+
 	req = otx2_mbox_alloc_msg_tim_lf_free(dev->mbox);
 	req->ring = tim_ring->ring_id;
 
@@ -333,6 +518,31 @@ otx2_tim_ring_free(struct rte_event_timer_adapter *adptr)
 	return 0;
 }
 
+static int
+otx2_tim_stats_get(const struct rte_event_timer_adapter *adapter,
+		   struct rte_event_timer_adapter_stats *stats)
+{
+	struct otx2_tim_ring *tim_ring = adapter->data->adapter_priv;
+	uint64_t bkt_cyc = rte_rdtsc() - tim_ring->ring_start_cyc;
+
+
+	stats->evtim_exp_count = __atomic_load_n(&tim_ring->arm_cnt,
+						 __ATOMIC_RELAXED);
+	stats->ev_enq_count = stats->evtim_exp_count;
+	stats->adapter_tick_count = rte_reciprocal_divide_u64(bkt_cyc,
+				&tim_ring->fast_div);
+	return 0;
+}
+
+static int
+otx2_tim_stats_reset(const struct rte_event_timer_adapter *adapter)
+{
+	struct otx2_tim_ring *tim_ring = adapter->data->adapter_priv;
+
+	__atomic_store_n(&tim_ring->arm_cnt, 0, __ATOMIC_RELAXED);
+	return 0;
+}
+
 int
 otx2_tim_caps_get(const struct rte_eventdev *evdev, uint64_t flags,
 		  uint32_t *caps,
@@ -341,11 +551,20 @@ otx2_tim_caps_get(const struct rte_eventdev *evdev, uint64_t flags,
 	struct otx2_tim_evdev *dev = tim_priv_get();
 
 	RTE_SET_USED(flags);
+
 	if (dev == NULL)
 		return -ENODEV;
 
 	otx2_tim_ops.init = otx2_tim_ring_create;
 	otx2_tim_ops.uninit = otx2_tim_ring_free;
+	otx2_tim_ops.start = otx2_tim_ring_start;
+	otx2_tim_ops.stop = otx2_tim_ring_stop;
+	otx2_tim_ops.get_info	= otx2_tim_ring_info_get;
+
+	if (dev->enable_stats) {
+		otx2_tim_ops.stats_get   = otx2_tim_stats_get;
+		otx2_tim_ops.stats_reset = otx2_tim_stats_reset;
+	}
 
 	/* Store evdev pointer for later use. */
 	dev->event_dev = (struct rte_eventdev *)(uintptr_t)evdev;
@@ -356,6 +575,89 @@ otx2_tim_caps_get(const struct rte_eventdev *evdev, uint64_t flags,
 }
 
 #define OTX2_TIM_DISABLE_NPA	"tim_disable_npa"
+#define OTX2_TIM_CHNK_SLOTS	"tim_chnk_slots"
+#define OTX2_TIM_STATS_ENA	"tim_stats_ena"
+#define OTX2_TIM_RINGS_LMT	"tim_rings_lmt"
+#define OTX2_TIM_RING_CTL	"tim_ring_ctl"
+
+static void
+tim_parse_ring_param(char *value, void *opaque)
+{
+	struct otx2_tim_evdev *dev = opaque;
+	struct otx2_tim_ctl ring_ctl = {0};
+	char *tok = strtok(value, "-");
+	struct otx2_tim_ctl *old_ptr;
+	uint16_t *val;
+
+	val = (uint16_t *)&ring_ctl;
+
+	if (!strlen(value))
+		return;
+
+	while (tok != NULL) {
+		*val = atoi(tok);
+		tok = strtok(NULL, "-");
+		val++;
+	}
+
+	if (val != (&ring_ctl.enable_stats + 1)) {
+		otx2_err(
+		"Invalid ring param expected [ring-chunk_sz-disable_npa-enable_stats]");
+		return;
+	}
+
+	dev->ring_ctl_cnt++;
+	old_ptr = dev->ring_ctl_data;
+	dev->ring_ctl_data = rte_realloc(dev->ring_ctl_data,
+					 sizeof(struct otx2_tim_ctl) *
+					 dev->ring_ctl_cnt, 0);
+	if (dev->ring_ctl_data == NULL) {
+		dev->ring_ctl_data = old_ptr;
+		dev->ring_ctl_cnt--;
+		return;
+	}
+
+	dev->ring_ctl_data[dev->ring_ctl_cnt - 1] = ring_ctl;
+}
+
+static void
+tim_parse_ring_ctl_list(const char *value, void *opaque)
+{
+	char *s = strdup(value);
+	char *start = NULL;
+	char *end = NULL;
+	char *f = s;
+
+	while (*s) {
+		if (*s == '[')
+			start = s;
+		else if (*s == ']')
+			end = s;
+
+		if (start && start < end) {
+			*end = 0;
+			tim_parse_ring_param(start + 1, opaque);
+			start = end;
+			s = end;
+		}
+		s++;
+	}
+
+	free(f);
+}
+
+static int
+tim_parse_kvargs_dict(const char *key, const char *value, void *opaque)
+{
+	RTE_SET_USED(key);
+
+	/* Dict format [ring-chunk_sz-disable_npa-enable_stats] use '-' as ','
+	 * isn't allowed. 0 represents default.
+	 */
+	tim_parse_ring_ctl_list(value, opaque);
+
+	return 0;
+}
 
 static void
 tim_parse_devargs(struct rte_devargs *devargs, struct otx2_tim_evdev *dev)
@@ -371,12 +673,23 @@ tim_parse_devargs(struct rte_devargs *devargs, struct otx2_tim_evdev *dev)
 
 	rte_kvargs_process(kvlist, OTX2_TIM_DISABLE_NPA,
 			   &parse_kvargs_flag, &dev->disable_npa);
+	rte_kvargs_process(kvlist, OTX2_TIM_CHNK_SLOTS,
+			   &parse_kvargs_value, &dev->chunk_slots);
+	rte_kvargs_process(kvlist, OTX2_TIM_STATS_ENA, &parse_kvargs_flag,
+			   &dev->enable_stats);
+	rte_kvargs_process(kvlist, OTX2_TIM_RINGS_LMT, &parse_kvargs_value,
+			   &dev->min_ring_cnt);
+	rte_kvargs_process(kvlist, OTX2_TIM_RING_CTL,
+			   &tim_parse_kvargs_dict, &dev);
+
+	rte_kvargs_free(kvlist);
 }
 
 void
 otx2_tim_init(struct rte_pci_device *pci_dev, struct otx2_dev *cmn_dev)
 {
 	struct rsrc_attach_req *atch_req;
+	struct rsrc_detach_req *dtch_req;
 	struct free_rsrcs_rsp *rsrc_cnt;
 	const struct rte_memzone *mz;
 	struct otx2_tim_evdev *dev;
@@ -407,7 +720,8 @@ otx2_tim_init(struct rte_pci_device *pci_dev, struct otx2_dev *cmn_dev)
 		goto mz_free;
 	}
 
-	dev->nb_rings = rsrc_cnt->tim;
+	dev->nb_rings = dev->min_ring_cnt ?
+		RTE_MIN(dev->min_ring_cnt, rsrc_cnt->tim) : rsrc_cnt->tim;
 
 	if (!dev->nb_rings) {
 		otx2_tim_dbg("No TIM Logical functions provisioned.");
@@ -424,8 +738,29 @@ otx2_tim_init(struct rte_pci_device *pci_dev, struct otx2_dev *cmn_dev)
 		goto mz_free;
 	}
 
+	rc = tim_get_msix_offsets();
+	if (rc < 0) {
+		otx2_err("Unable to get MSIX offsets for TIM.");
+		goto detach;
+	}
+
+	if (dev->chunk_slots &&
+	    dev->chunk_slots <= OTX2_TIM_MAX_CHUNK_SLOTS &&
+	    dev->chunk_slots >= OTX2_TIM_MIN_CHUNK_SLOTS) {
+		dev->chunk_sz = (dev->chunk_slots + 1) *
+			OTX2_TIM_CHUNK_ALIGNMENT;
+	} else {
+		dev->chunk_sz = OTX2_TIM_RING_DEF_CHUNK_SZ;
+	}
+
 	return;
 
+detach:
+	dtch_req = otx2_mbox_alloc_msg_detach_resources(dev->mbox);
+	dtch_req->partial = true;
+	dtch_req->timlfs = true;
+
+	otx2_mbox_process(dev->mbox);
 mz_free:
 	rte_memzone_free(mz);
 }