Skip to contentSkip to conversation

test-bbdev fixes and improvements for 24.07

PATCH Applied Hernan Vargas 2 participants 13 messages Jun 24, 2024 – Jun 25, 2024

lore
Share & tools
lore
#0 lore raw

v2 cover test-bbdev fixes and improvements for 24.07

v2: Added defines for MLD calculation constants.
v1:This series targets test-bbdev changes for 24.07.
It includes test fixes, iter-max argument change (after deprecation notice in previous releases) and general test improvements.

Hernan Vargas (9):
test/bbdev: fix TB logic
test/bbdev: fix MLD output size computation
test/bbdev: fix interrupt tests
test/bbdev: change iter-max argument
test/bbdev: improve timeout message format
test/bbdev: add soft output parsing capability
test/bbdev: check assumptions on fft window
test/bbdev: update fft measurement output
test/bbdev: remove unnecessary line

 app/test-bbdev/test-bbdev.py       |  38 +++---
 app/test-bbdev/test_bbdev.c        |   1 -
 app/test-bbdev/test_bbdev_perf.c   | 191 ++++++++++++++++++++---------
 app/test-bbdev/test_bbdev_vector.c |  18 +++
 app/test-bbdev/test_bbdev_vector.h |   2 +
 5 files changed, 177 insertions(+), 73 deletions(-)
Signature · 1 line
2.37.1
#1 lore raw

PATCH v2 1/9 test/bbdev: fix TB logic

Fix discrepancy in logic when using large fake mbuf.

Fixes fd96ef3787f1 ("test/bbdev: extend support for large TB")
---
 app/test-bbdev/test_bbdev_perf.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
app/test-bbdev/test_bbdev_perf.c +2 -1
index a21c9c7bddf5..6d9bf3a233ec 100644
@@ -2131,7 +2131,8 @@ validate_op_chain(struct rte_bbdev_op_data *op,
 		uint16_t data_len = rte_pktmbuf_data_len(m) - offset;
 		total_data_size += orig_op->segments[i].length;
 
-		if (orig_op->segments[i].length > RTE_BBDEV_LDPC_E_MAX_MBUF)
+		if ((orig_op->segments[i].length + RTE_PKTMBUF_HEADROOM)
+				> RTE_BBDEV_LDPC_E_MAX_MBUF)
 			ignore_mbuf = true;
 		if (!ignore_mbuf)
 			TEST_ASSERT(orig_op->segments[i].length == data_len,
Signature · 1 line
2.37.1
#2 lore raw

PATCH v2 2/9 test/bbdev: fix MLD output size computation

In reply to Hernan Vargas

For perf tests, the operation size for the MLD-TS was incorrect. Fixed so that the performance numbers are correct. Largely cosmetic only.

Fixes 95f192a40e35 ("test/bbdev: add MLD cases")
---
 app/test-bbdev/test_bbdev_perf.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
app/test-bbdev/test_bbdev_perf.c +10 -2
index 6d9bf3a233ec..9841464922ac 100644
@@ -94,6 +94,8 @@
 #define K0_2_2 25 /* K0 fraction numerator for rv 2 and BG 2 */
 #define K0_3_1 56 /* K0 fraction numerator for rv 3 and BG 1 */
 #define K0_3_2 43 /* K0 fraction numerator for rv 3 and BG 2 */
+#define NUM_SC_PER_RB (12) /* Number of subcarriers in a RB in 3GPP. */
+#define BITS_PER_LLR  (8)  /* Number of bits in a LLR. */
 
 #define HARQ_MEM_TOLERANCE 256
 static struct test_bbdev_vector test_vector;
@@ -2896,8 +2898,14 @@ calc_fft_size(struct rte_bbdev_fft_op *op)
 static uint32_t
 calc_mldts_size(struct rte_bbdev_mldts_op *op)
 {
-	uint32_t output_size;
-	output_size = op->mldts.num_layers * op->mldts.num_rbs * op->mldts.c_rep;
+	uint32_t output_size = 0;
+	uint16_t i;
+
+	for (i = 0; i < op->mldts.num_layers; i++)
+		output_size += op->mldts.q_m[i];
+
+	output_size *= NUM_SC_PER_RB * BITS_PER_LLR * op->mldts.num_rbs * (op->mldts.c_rep + 1);
+
 	return output_size;
 }
 
Signature · 1 line
2.37.1
Reviewed-by on v2 2/9
#3 lore raw

Re: [PATCH v2 2/9] test/bbdev: fix MLD output size computation

Quoted 32 lines · Hernan Vargas

For perf tests, the operation size for the MLD-TS was incorrect. Fixed so that the performance numbers are correct. Largely cosmetic only.

Fixes 95f192a40e35 ("test/bbdev: add MLD cases")
---
  app/test-bbdev/test_bbdev_perf.c | 12 ++++++++++--
  1 file changed, 10 insertions(+), 2 deletions(-)
app/test-bbdev/test_bbdev_perf.c +3 -2
index 6d9bf3a233ec..9841464922ac 100644
@@ -94,6 +94,8 @@
  #define K0_2_2 25 /* K0 fraction numerator for rv 2 and BG 2 */
  #define K0_3_1 56 /* K0 fraction numerator for rv 3 and BG 1 */
  #define K0_3_2 43 /* K0 fraction numerator for rv 3 and BG 2 */
+#define NUM_SC_PER_RB (12) /* Number of subcarriers in a RB in 3GPP. */
+#define BITS_PER_LLR  (8)  /* Number of bits in a LLR. */
 
  #define HARQ_MEM_TOLERANCE 256
  static struct test_bbdev_vector test_vector;
@@ -2896,8 +2898,14 @@ calc_fft_size(struct rte_bbdev_fft_op *op)
  static uint32_t
  calc_mldts_size(struct rte_bbdev_mldts_op *op)
  {
-	uint32_t output_size;
-	output_size = op->mldts.num_layers * op->mldts.num_rbs * op->mldts.c_rep;
+	uint32_t output_size = 0;

app/test-bbdev/test_bbdev_perf.c calc_fft_size(struct rte_bbdev_fft_op *op) · line 2911

Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +7 -0
+	uint16_t i;
+
+	for (i = 0; i < op->mldts.num_layers; i++)
+		output_size += op->mldts.q_m[i];
+
+	output_size *= NUM_SC_PER_RB * BITS_PER_LLR * op->mldts.num_rbs * (op->mldts.c_rep + 1);
+
  	return output_size;
  }
 

Thanks for the change, much clearer now!

Maxime

#4 lore raw

PATCH v2 3/9 test/bbdev: fix interrupt tests

In reply to Hernan Vargas

Fix possible error with regards to setting the burst size from the enqueue thread.

Fixes b2e2aec3239e ("app/bbdev: enhance interrupt test")
---
 app/test-bbdev/test_bbdev_perf.c | 98 ++++++++++++++++----------------
 1 file changed, 49 insertions(+), 49 deletions(-)
app/test-bbdev/test_bbdev_perf.c +49 -49
index 9841464922ac..20cd8df19be7 100644
@@ -3419,15 +3419,6 @@ throughput_intr_lcore_ldpc_dec(void *arg)
 			if (unlikely(num_to_process - enqueued < num_to_enq))
 				num_to_enq = num_to_process - enqueued;
 
-			enq = 0;
-			do {
-				enq += rte_bbdev_enqueue_ldpc_dec_ops(
-						tp->dev_id,
-						queue_id, &ops[enqueued],
-						num_to_enq);
-			} while (unlikely(num_to_enq != enq));
-			enqueued += enq;
-
 			/* Write to thread burst_sz current number of enqueued
 			 * descriptors. It ensures that proper number of
 			 * descriptors will be dequeued in callback
@@ -3438,6 +3429,15 @@ throughput_intr_lcore_ldpc_dec(void *arg)
 			rte_atomic_store_explicit(&tp->burst_sz, num_to_enq,
 					rte_memory_order_relaxed);
 
+			enq = 0;
+			do {
+				enq += rte_bbdev_enqueue_ldpc_dec_ops(
+						tp->dev_id,
+						queue_id, &ops[enqueued],
+						num_to_enq);
+			} while (unlikely(num_to_enq != enq));
+			enqueued += enq;
+
 			/* Wait until processing of previous batch is
 			 * completed
 			 */
@@ -3514,14 +3514,6 @@ throughput_intr_lcore_dec(void *arg)
 			if (unlikely(num_to_process - enqueued < num_to_enq))
 				num_to_enq = num_to_process - enqueued;
 
-			enq = 0;
-			do {
-				enq += rte_bbdev_enqueue_dec_ops(tp->dev_id,
-						queue_id, &ops[enqueued],
-						num_to_enq);
-			} while (unlikely(num_to_enq != enq));
-			enqueued += enq;
-
 			/* Write to thread burst_sz current number of enqueued
 			 * descriptors. It ensures that proper number of
 			 * descriptors will be dequeued in callback
@@ -3532,6 +3524,14 @@ throughput_intr_lcore_dec(void *arg)
 			rte_atomic_store_explicit(&tp->burst_sz, num_to_enq,
 					rte_memory_order_relaxed);
 
+			enq = 0;
+			do {
+				enq += rte_bbdev_enqueue_dec_ops(tp->dev_id,
+						queue_id, &ops[enqueued],
+						num_to_enq);
+			} while (unlikely(num_to_enq != enq));
+			enqueued += enq;
+
 			/* Wait until processing of previous batch is
 			 * completed
 			 */
@@ -3603,14 +3603,6 @@ throughput_intr_lcore_enc(void *arg)
 			if (unlikely(num_to_process - enqueued < num_to_enq))
 				num_to_enq = num_to_process - enqueued;
 
-			enq = 0;
-			do {
-				enq += rte_bbdev_enqueue_enc_ops(tp->dev_id,
-						queue_id, &ops[enqueued],
-						num_to_enq);
-			} while (unlikely(enq != num_to_enq));
-			enqueued += enq;
-
 			/* Write to thread burst_sz current number of enqueued
 			 * descriptors. It ensures that proper number of
 			 * descriptors will be dequeued in callback
@@ -3621,6 +3613,14 @@ throughput_intr_lcore_enc(void *arg)
 			rte_atomic_store_explicit(&tp->burst_sz, num_to_enq,
 					rte_memory_order_relaxed);
 
+			enq = 0;
+			do {
+				enq += rte_bbdev_enqueue_enc_ops(tp->dev_id,
+						queue_id, &ops[enqueued],
+						num_to_enq);
+			} while (unlikely(enq != num_to_enq));
+			enqueued += enq;
+
 			/* Wait until processing of previous batch is
 			 * completed
 			 */
@@ -3694,15 +3694,6 @@ throughput_intr_lcore_ldpc_enc(void *arg)
 			if (unlikely(num_to_process - enqueued < num_to_enq))
 				num_to_enq = num_to_process - enqueued;
 
-			enq = 0;
-			do {
-				enq += rte_bbdev_enqueue_ldpc_enc_ops(
-						tp->dev_id,
-						queue_id, &ops[enqueued],
-						num_to_enq);
-			} while (unlikely(enq != num_to_enq));
-			enqueued += enq;
-
 			/* Write to thread burst_sz current number of enqueued
 			 * descriptors. It ensures that proper number of
 			 * descriptors will be dequeued in callback
@@ -3713,6 +3704,15 @@ throughput_intr_lcore_ldpc_enc(void *arg)
 			rte_atomic_store_explicit(&tp->burst_sz, num_to_enq,
 					rte_memory_order_relaxed);
 
+			enq = 0;
+			do {
+				enq += rte_bbdev_enqueue_ldpc_enc_ops(
+						tp->dev_id,
+						queue_id, &ops[enqueued],
+						num_to_enq);
+			} while (unlikely(enq != num_to_enq));
+			enqueued += enq;
+
 			/* Wait until processing of previous batch is
 			 * completed
 			 */
@@ -3786,14 +3786,6 @@ throughput_intr_lcore_fft(void *arg)
 			if (unlikely(num_to_process - enqueued < num_to_enq))
 				num_to_enq = num_to_process - enqueued;
 
-			enq = 0;
-			do {
-				enq += rte_bbdev_enqueue_fft_ops(tp->dev_id,
-						queue_id, &ops[enqueued],
-						num_to_enq);
-			} while (unlikely(enq != num_to_enq));
-			enqueued += enq;
-
 			/* Write to thread burst_sz current number of enqueued
 			 * descriptors. It ensures that proper number of
 			 * descriptors will be dequeued in callback
@@ -3804,6 +3796,14 @@ throughput_intr_lcore_fft(void *arg)
 			rte_atomic_store_explicit(&tp->burst_sz, num_to_enq,
 					rte_memory_order_relaxed);
 
+			enq = 0;
+			do {
+				enq += rte_bbdev_enqueue_fft_ops(tp->dev_id,
+						queue_id, &ops[enqueued],
+						num_to_enq);
+			} while (unlikely(enq != num_to_enq));
+			enqueued += enq;
+
 			/* Wait until processing of previous batch is
 			 * completed
 			 */
@@ -3872,13 +3872,6 @@ throughput_intr_lcore_mldts(void *arg)
 			if (unlikely(num_to_process - enqueued < num_to_enq))
 				num_to_enq = num_to_process - enqueued;
 
-			enq = 0;
-			do {
-				enq += rte_bbdev_enqueue_mldts_ops(tp->dev_id,
-						queue_id, &ops[enqueued], num_to_enq);
-			} while (unlikely(enq != num_to_enq));
-			enqueued += enq;
-
 			/* Write to thread burst_sz current number of enqueued
 			 * descriptors. It ensures that proper number of
 			 * descriptors will be dequeued in callback
@@ -3889,6 +3882,13 @@ throughput_intr_lcore_mldts(void *arg)
 			rte_atomic_store_explicit(&tp->burst_sz, num_to_enq,
 					rte_memory_order_relaxed);
 
+			enq = 0;
+			do {
+				enq += rte_bbdev_enqueue_mldts_ops(tp->dev_id,
+						queue_id, &ops[enqueued], num_to_enq);
+			} while (unlikely(enq != num_to_enq));
+			enqueued += enq;
+
 			/* Wait until processing of previous batch is
 			 * completed
 			 */
Signature · 1 line
2.37.1
#5 lore raw

PATCH v2 4/9 test/bbdev: change iter-max argument

In reply to Hernan Vargas

-t --iter-max used for max number of iterations.
-T --timeout used for test timeout value.

---
 app/test-bbdev/test-bbdev.py | 16 +++++-----------
 1 file changed, 5 insertions(+), 11 deletions(-)
app/test-bbdev/test-bbdev.py +5 -11
index 65a34390548d..9ddca9e78d3b 100755
@@ -27,16 +27,10 @@ def kill(process):
 parser.add_argument("-e", "--eal-params",
                     help="EAL arguments which must be passed to the test app",
                     default="--vdev=baseband_null0 -a00:00.0")
-# Until deprecated in next release keep -t as an valid argument for timeout, then use -T
-parser.add_argument("-t", "--timeout",
+parser.add_argument("-T", "--timeout",
                     type=int,
                     help="Timeout in seconds",
                     default=600)
-# This will become -t option for iter_max in next release
-parser.add_argument("--iter-max",
-                    type=int,
-                    help="Max iterations",
-                    default=6)
 parser.add_argument("-c", "--test-cases",
                     nargs="+",
                     help="Defines test cases to run. Run all if not specified")
@@ -58,6 +52,10 @@ def kill(process):
                     type=int,
                     help="SNR in dB for BLER tests",
                     default=0)
+parser.add_argument("-t", "--iter-max",
+                    type=int,
+                    help="Max iterations",
+                    default=6)
 parser.add_argument("-l", "--num-lcores",
                     type=int,
                     help="Number of lcores to run.",
@@ -83,10 +81,6 @@ def kill(process):
 
 if args.iter_max:
     params.extend(["-t", str(args.iter_max)])
-    print("The argument for iter_max will be -t in next release")
-
-if args.timeout:
-    print("The argument for timeout will be -T in next release")
 
 if args.num_ops:
     params.extend(["-n", str(args.num_ops)])
Signature · 1 line
2.37.1
#6 lore raw

PATCH v2 5/9 test/bbdev: improve timeout message format

In reply to Hernan Vargas

Print more info and format message for test timeouts. No functional impact.

---
 app/test-bbdev/test-bbdev.py | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)
app/test-bbdev/test-bbdev.py +18 -4
index 9ddca9e78d3b..d5b679f7867f 100755
@@ -108,15 +108,29 @@ def kill(process):
         try:
             output = subprocess.run(call_params, timeout=args.timeout, universal_newlines=True)
         except subprocess.TimeoutExpired as e:
+            print("===========================================================")
             print("Starting Test Suite : BBdev TimeOut Tests")
+            print("INFO: One of the tests timed out {}".format(e))
+            print("INFO: Unexpected Error")
+            print("+ ------------------------------------------------------- +")
             print("== test: timeout")
-            print("TestCase [ 0] : timeout passed")
-            print(" + Tests Failed :       1")
             print("Unexpected Error")
+            print("TestCase [ 0] : timeout failed")
+            print(" + ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +")
+            print(" + Tests Failed :       1")
+            print(" + ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +")
+            exit_status = 1
         if output.returncode < 0:
+            print("===========================================================")
             print("Starting Test Suite : BBdev Exception Tests")
+            print("INFO: One of the tests returned {}".format(output.returncode))
+            print("INFO: Unexpected Error")
+            print("+ ------------------------------------------------------- +")
             print("== test: exception")
-            print("TestCase [ 0] : exception passed")
-            print(" + Tests Failed :       1")
             print("Unexpected Error")
+            print("TestCase [ 0] : exception failed")
+            print(" + ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +")
+            print(" + Tests Failed :       1")
+            print(" + ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +")
+            exit_status = 1
 sys.exit(exit_status)
Signature · 1 line
2.37.1
#7 lore raw

PATCH v2 6/9 test/bbdev: add soft output parsing capability

In reply to Hernan Vargas

Add vector parsing capability for soft output vectors.

---
 app/test-bbdev/test_bbdev_vector.c | 4 ++++
 1 file changed, 4 insertions(+)
app/test-bbdev/test_bbdev_vector.c +4 -0
index 42fa630041e9..b3e9d4bb7504 100644
@@ -837,6 +837,10 @@ parse_ldpc_decoder_params(const char *key_token, char *token,
 		ret = parse_data_entry(key_token, token, vector,
 				DATA_HARQ_OUTPUT,
 				op_data_prefixes[DATA_HARQ_OUTPUT]);
+	else if (starts_with(key_token, op_data_prefixes[DATA_SOFT_OUTPUT]))
+		ret = parse_data_entry(key_token, token, vector,
+				DATA_SOFT_OUTPUT,
+				op_data_prefixes[DATA_SOFT_OUTPUT]);
 	else if (!strcmp(key_token, "e")) {
 		vector->mask |= TEST_BBDEV_VF_E;
 		ldpc_dec->cb_params.e = (uint32_t) strtoul(token, &err, 0);
Signature · 1 line
2.37.1
#8 lore raw

PATCH v2 7/9 test/bbdev: check assumptions on fft window

In reply to Hernan Vargas

Add check for FFT window width.

---
 app/test-bbdev/test_bbdev_perf.c   | 26 ++++++++++++++++++++++----
 app/test-bbdev/test_bbdev_vector.c | 14 ++++++++++++++
 app/test-bbdev/test_bbdev_vector.h |  2 ++
 3 files changed, 38 insertions(+), 4 deletions(-)
app/test-bbdev/test_bbdev_perf.c +22 -4
index 20cd8df19be7..7c414ab360a8 100644
@@ -108,6 +108,8 @@ static int ldpc_llr_decimals;
 static int ldpc_llr_size;
 /* Keep track of the LDPC decoder device capability flag */
 static uint32_t ldpc_cap_flags;
+/* FFT window width predefined on device and on vector. */
+static int fft_window_width_dev;
 
 /* Represents tested active devices */
 static struct active_device {
@@ -883,6 +885,13 @@ add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info,
 	rte_bbdev_info_get(dev_id, info);
 	if (info->drv.device_status == RTE_BBDEV_DEV_FATAL_ERR)
 		printf("Device Status %s\n", rte_bbdev_device_status_str(info->drv.device_status));
+	if (info->drv.fft_window_width != NULL)
+		fft_window_width_dev = info->drv.fft_window_width[0];
+	else
+		fft_window_width_dev = 0;
+	if (fft_window_width_dev != 0)
+		printf("  FFT Window0 width %d\n", fft_window_width_dev);
+
 	nb_queues = RTE_MIN(rte_lcore_count(), info->drv.max_num_queues);
 	nb_queues = RTE_MIN(nb_queues, (unsigned int) MAX_QUEUES);
 
@@ -2585,7 +2594,8 @@ validate_ldpc_enc_op(struct rte_bbdev_enc_op **ops, const uint16_t n,
 }
 
 static inline int
-validate_op_fft_chain(struct rte_bbdev_op_data *op, struct op_data_entries *orig_op)
+validate_op_fft_chain(struct rte_bbdev_op_data *op, struct op_data_entries *orig_op,
+		bool skip_validate_output)
 {
 	struct rte_mbuf *m = op->data;
 	uint8_t i, nb_dst_segments = orig_op->nb_segments;
@@ -2615,7 +2625,7 @@ validate_op_fft_chain(struct rte_bbdev_op_data *op, struct op_data_entries *orig
 			abs_delt = delt > 0 ? delt : -delt;
 			error_num += (abs_delt > thres_hold ? 1 : 0);
 		}
-		if (error_num > 0) {
+		if ((error_num > 0) && !skip_validate_output) {
 			rte_memdump(stdout, "Buffer A", ref_out, data_len);
 			rte_memdump(stdout, "Buffer B", op_out, data_len);
 			TEST_ASSERT(error_num == 0,
@@ -2688,16 +2698,24 @@ validate_fft_op(struct rte_bbdev_fft_op **ops, const uint16_t n,
 	int ret;
 	struct op_data_entries *fft_data_orig = &test_vector.entries[DATA_HARD_OUTPUT];
 	struct op_data_entries *fft_pwr_orig = &test_vector.entries[DATA_SOFT_OUTPUT];
+	bool skip_validate_output = false;
+
+	if ((test_vector.fft_window_width_vec > 0) &&
+			(test_vector.fft_window_width_vec != fft_window_width_dev)) {
+		printf("The vector FFT width doesn't match with device - skip %d %d\n",
+				test_vector.fft_window_width_vec, fft_window_width_dev);
+		skip_validate_output = true;
+	}
 
 	for (i = 0; i < n; ++i) {
 		ret = check_fft_status_and_ordering(ops[i], i, ref_op->status);
 		TEST_ASSERT_SUCCESS(ret, "Checking status and ordering for FFT failed");
 		TEST_ASSERT_SUCCESS(validate_op_fft_chain(
-				&ops[i]->fft.base_output, fft_data_orig),
+				&ops[i]->fft.base_output, fft_data_orig, skip_validate_output),
 				"FFT Output buffers (op=%u) are not matched", i);
 		if (check_bit(ops[i]->fft.op_flags, RTE_BBDEV_FFT_POWER_MEAS))
 			TEST_ASSERT_SUCCESS(validate_op_fft_chain(
-				&ops[i]->fft.power_meas_output, fft_pwr_orig),
+				&ops[i]->fft.power_meas_output, fft_pwr_orig, skip_validate_output),
 				"FFT Power Output buffers (op=%u) are not matched", i);
 	}
 
app/test-bbdev/test_bbdev_vector.c +14 -0
index b3e9d4bb7504..e48947b211ac 100644
@@ -1050,6 +1050,20 @@ parse_fft_params(const char *key_token, char *token,
 			}
 		}
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
+	} else if (!strcmp(key_token, "fft_window_width")) {
+		tok = strtok(token, VALUE_DELIMITER);
+		if (tok == NULL)
+			return -1;
+		for (i = 0; i < FFT_WIN_SIZE; i++) {
+			if (i == 0)
+				vector->fft_window_width_vec = (uint32_t) strtoul(tok, &err, 0);
+			if (i < (FFT_WIN_SIZE - 1)) {
+				tok = strtok(NULL, VALUE_DELIMITER);
+				if (tok == NULL)
+					return -1;
+			}
+		}
+		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "op_flags")) {
 		vector->mask |= TEST_BBDEV_VF_OP_FLAGS;
 		ret = parse_turbo_flags(token, &op_flags, vector->op_type);
app/test-bbdev/test_bbdev_vector.h +2 -0
index 14b8ef2764ad..ba1d0d20f9ea 100644
@@ -69,6 +69,8 @@ struct test_bbdev_vector {
 	};
 	/* Additional storage for op data entries */
 	struct op_data_entries entries[DATA_NUM_TYPES];
+	/* Vector FFT window width assumption. */
+	uint16_t fft_window_width_vec;
 };
 
 /* fills test vector parameters based on test file */
Signature · 1 line
2.37.1
Reviewed-by on v2 7/9
#9 lore raw

Re: [PATCH v2 7/9] test/bbdev: check assumptions on fft window

Quoted 9 lines · Hernan Vargas

Add check for FFT window width.

---
  app/test-bbdev/test_bbdev_perf.c   | 26 ++++++++++++++++++++++----
  app/test-bbdev/test_bbdev_vector.c | 14 ++++++++++++++
  app/test-bbdev/test_bbdev_vector.h |  2 ++
  3 files changed, 38 insertions(+), 4 deletions(-)

Maxime

#10 lore raw

PATCH v2 8/9 test/bbdev: update fft measurement output

In reply to Hernan Vargas

Update check for FFT measurement output to better account for tolerance.

---
 app/test-bbdev/test_bbdev_perf.c | 52 +++++++++++++++++++++++++++++++-
 1 file changed, 51 insertions(+), 1 deletion(-)
app/test-bbdev/test_bbdev_perf.c +51 -1
index 7c414ab360a8..3a94f15a30b8 100644
@@ -2639,6 +2639,56 @@ validate_op_fft_chain(struct rte_bbdev_op_data *op, struct op_data_entries *orig
 	return TEST_SUCCESS;
 }
 
+static inline int
+validate_op_fft_meas_chain(struct rte_bbdev_op_data *op, struct op_data_entries *orig_op,
+		bool skip_validate_output)
+{
+	struct rte_mbuf *m = op->data;
+	uint8_t i, nb_dst_segments = orig_op->nb_segments;
+	double thres_hold = 1.0;
+	uint32_t j, data_len_iq, error_num;
+	int32_t *ref_out, *op_out;
+	double estSNR, refSNR, delt, abs_delt;
+
+	TEST_ASSERT(nb_dst_segments == m->nb_segs,
+			"Number of segments differ in original (%u) and filled (%u) op fft",
+			nb_dst_segments, m->nb_segs);
+
+	/* Due to size limitation of mbuf, FFT doesn't use real mbuf. */
+	for (i = 0; i < nb_dst_segments; ++i) {
+		uint16_t offset = (i == 0) ? op->offset : 0;
+		uint32_t data_len = op->length;
+
+		TEST_ASSERT(orig_op->segments[i].length == data_len,
+				"Length of segment differ in original (%u) and filled (%u) op fft",
+				orig_op->segments[i].length, data_len);
+
+		/* Divided by 4 to get the number of 32 bits data. */
+		data_len_iq = data_len >> 2;
+		ref_out = (int32_t *)(orig_op->segments[i].addr);
+		op_out = rte_pktmbuf_mtod_offset(m, int32_t *, offset);
+		error_num = 0;
+		for (j = 0; j < data_len_iq; j++) {
+			estSNR = 10*log10(op_out[j]);
+			refSNR = 10*log10(ref_out[j]);
+			delt = refSNR - estSNR;
+			abs_delt = delt > 0 ? delt : -delt;
+			error_num += (abs_delt > thres_hold ? 1 : 0);
+		}
+		if ((error_num > 0) && !skip_validate_output) {
+			rte_memdump(stdout, "Buffer A", ref_out, data_len);
+			rte_memdump(stdout, "Buffer B", op_out, data_len);
+			TEST_ASSERT(error_num == 0,
+				"FFT Output are not matched total (%u) errors (%u)",
+				data_len_iq, error_num);
+		}
+
+		m = m->next;
+	}
+
+	return TEST_SUCCESS;
+}
+
 static inline int
 validate_op_mldts_chain(struct rte_bbdev_op_data *op,
 		struct op_data_entries *orig_op)
@@ -2714,7 +2764,7 @@ validate_fft_op(struct rte_bbdev_fft_op **ops, const uint16_t n,
 				&ops[i]->fft.base_output, fft_data_orig, skip_validate_output),
 				"FFT Output buffers (op=%u) are not matched", i);
 		if (check_bit(ops[i]->fft.op_flags, RTE_BBDEV_FFT_POWER_MEAS))
-			TEST_ASSERT_SUCCESS(validate_op_fft_chain(
+			TEST_ASSERT_SUCCESS(validate_op_fft_meas_chain(
 				&ops[i]->fft.power_meas_output, fft_pwr_orig, skip_validate_output),
 				"FFT Power Output buffers (op=%u) are not matched", i);
 	}
Signature · 1 line
2.37.1
#11 lore raw

PATCH v2 9/9 test/bbdev: remove unnecessary line

In reply to Hernan Vargas

Remove unnecesary line of code.
No functional impact.

---
 app/test-bbdev/test_bbdev.c | 1 -
 1 file changed, 1 deletion(-)
app/test-bbdev/test_bbdev.c +0 -1
index cf224dca5d04..0bbce6ca923b 100644
@@ -1239,7 +1239,6 @@ test_bbdev_invalid_driver(void)
 	TEST_ASSERT_SUCCESS(rte_bbdev_callback_unregister(dev_id,
 			RTE_BBDEV_EVENT_UNKNOWN, event_callback, NULL),
 			"Failed to unregister RTE_BBDEV_EVENT_ERROR ");
-	dev2->dev_ops = dev1.dev_ops;
 
 	/* Tests for rte_bbdev_stats_reset */
 	dev2->dev_ops = NULL;
Signature · 1 line
2.37.1
#12 lore raw
In reply to Hernan Vargas
Quoted 15 lines · Hernan Vargas

v2: Added defines for MLD calculation constants.
v1:This series targets test-bbdev changes for 24.07.
It includes test fixes, iter-max argument change (after deprecation notice in previous releases) and general test improvements.

Hernan Vargas (9):
test/bbdev: fix TB logic
test/bbdev: fix MLD output size computation
test/bbdev: fix interrupt tests
test/bbdev: change iter-max argument
test/bbdev: improve timeout message format
test/bbdev: add soft output parsing capability
test/bbdev: check assumptions on fft window
test/bbdev: update fft measurement output
test/bbdev: remove unnecessary line

Quoted 7 lines
  app/test-bbdev/test-bbdev.py       |  38 +++---
  app/test-bbdev/test_bbdev.c        |   1 -
  app/test-bbdev/test_bbdev_perf.c   | 191 ++++++++++++++++++++---------
  app/test-bbdev/test_bbdev_vector.c |  18 +++
  app/test-bbdev/test_bbdev_vector.h |   2 +
  5 files changed, 177 insertions(+), 73 deletions(-)

Applied to next-baseband/for-main

Thanks,
Maxime