test/bbdev: changes for 23.03
v2 cover test/bbdev: changes for 23.03
v2: Made changes requested during review. Added 3 commits.
v1: Upstreaming bbdev-test changes for 23.03.
Hernan Vargas (16):
test/bbdev: fix seg fault for non supported HARQ len test/bbdev: extend HARQ tolerance test/bbdev: refactor TB throughput report test/bbdev: add timeout for latency tests test/bbdev: enable early termination for validation test/bbdev: report device status in test-bbdev test/bbdev: test start/stop bbdev API test/bbdev: add support for BLER for 4G test/bbdev: extend support for large TB test/bbdev: adjustment for soft output test/bbdev: expose warning counters test/bbdev: remove check for invalid opaque data test/bbdev: remove iteration count check test/bbdev: use mbuf reset function test/bbdev: remove max iteration from vectors test/bbdev: remove iter count from bler test
app/test-bbdev/test_bbdev_perf.c | 400 +++++++++++++----- app/test-bbdev/test_bbdev_vector.c | 14 - app/test-bbdev/test_bbdev_vector.h | 1 - .../test_vectors/ldpc_dec_HARQ_1_0.data | 3 - .../test_vectors/ldpc_dec_HARQ_1_1.data | 3 - .../test_vectors/ldpc_dec_HARQ_1_2.data | 3 - .../test_vectors/ldpc_dec_v11835.data | 3 - .../test_vectors/ldpc_dec_v2342_drop.data | 3 - .../test_vectors/ldpc_dec_v7813.data | 3 - .../test_vectors/ldpc_dec_v8480.data | 3 - .../test_vectors/ldpc_dec_v8568.data | 3 - .../test_vectors/ldpc_dec_v9503.data | 3 - ...turbo_dec_c1_k40_r0_e17280_sbd_negllr.data | 3 - ..._r0_e10376_crc24b_sbd_negllr_high_snr.data | 3 - ...4_r0_e10376_crc24b_sbd_negllr_low_snr.data | 3 - .../turbo_dec_c1_k6144_r0_e34560_posllr.data | 3 - ...rbo_dec_c1_k6144_r0_e34560_sbd_negllr.data | 3 - ...rbo_dec_c1_k6144_r0_e34560_sbd_posllr.data | 3 - ...c_c2_k3136_r0_e4920_sbd_negllr_crc24b.data | 3 - 19 files changed, 301 insertions(+), 162 deletions(-)
Signature · 1 line
2.37.1
PATCH v2 1/16 test/bbdev: fix seg fault for non supported HARQ len
Fix segmentation fault happening in corner case in test-bbdev. This fault could happen when running some specific vectors which size are not supported by the PMD.
--- app/test-bbdev/test_bbdev_perf.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
app/test-bbdev/test_bbdev_perf.c +3 -3
@@ -84,7 +84,7 @@/* Increment for next code block in external HARQ memory */ #define HARQ_INCR 32768/* Headroom for filler LLRs insertion in HARQ buffer */ -#define FILLER_HEADROOM 1024 +#define FILLER_HEADROOM 2048/* Constants from K0 computation from 3GPP 38.212 Table 5.4.2.1-2 */ #define N_ZC_1 66/* N = 66 Zc for BG 1 */ #define N_ZC_2 50/* N = 50 Zc for BG 2 */
@@ -2111,9 +2111,9 @@ validate_op_harq_chain(struct rte_bbdev_op_data *op, ops_ld->n_filler;if (data_len > deRmOutSize) data_len = deRmOutSize; -if (data_len > orig_op->segments[i].length) - data_len = orig_op->segments[i].length; } +if (data_len > orig_op->segments[i].length) + data_len = orig_op->segments[i].length;/* * HARQ output can have minor differences * due to integer representation and related scaling
Signature · 1 line
2.37.1
Re: [PATCH v2 01/16] test/bbdev: fix seg fault for non supported HARQ len
Quoted 27 lines · Hernan Vargas
Fix segmentation fault happening in corner case in test-bbdev. This fault could happen when running some specific vectors which size are not supported by the PMD.
Fixes 335c11fd276 ("app/bbdev: support HARQ validation")Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)app/test-bbdev/test_bbdev_perf.c +1
-1@@ -84,7 +84,7 @@/* Increment for next code block in external HARQ memory */ #define HARQ_INCR 32768/* Headroom for filler LLRs insertion in HARQ buffer */ -#define FILLER_HEADROOM 1024 +#define FILLER_HEADROOM 2048/* Constants from K0 computation from 3GPP 38.212 Table 5.4.2.1-2 */ #define N_ZC_1 66/* N = 66 Zc for BG 1 */ #define N_ZC_2 50/* N = 50 Zc for BG 2 */ @@ -2111,9 +2111,9 @@ validate_op_harq_chain(struct rte_bbdev_op_data *op, ops_ld->n_filler;
app/test-bbdev/test_bbdev_perf.c validate_op_harq_chain(struct rte_bbdev_op_data *op, · line 2119
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +2
-2if (data_len > deRmOutSize) data_len = deRmOutSize; -if (data_len > orig_op->segments[i].length) - data_len = orig_op->segments[i].length; } +if (data_len > orig_op->segments[i].length) + data_len = orig_op->segments[i].length;/* * HARQ output can have minor differences * due to integer representation and related scaling
Thanks,
Maxime
PATCH v2 2/16 test/bbdev: extend HARQ tolerance
In reply to Hernan VargasHARQ memory implementation could have different size assumptions. Extending HARQ tolerance to cover different implementations.
--- app/test-bbdev/test_bbdev_perf.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
app/test-bbdev/test_bbdev_perf.c +8 -3
@@ -95,6 +95,7 @@#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 HARQ_MEM_TOLERANCE 256static struct test_bbdev_vector test_vector;/* Switch between PMD and Interrupt for throughput TC */
@@ -2090,13 +2091,17 @@ validate_op_harq_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; -TEST_ASSERT (orig_op->segments[i].length < - (uint32_t )(data_len +64 ), +TEST_ASSERT (orig_op->segments[i].length < (uint32_t )(data_len +HARQ_MEM_TOLERANCE ),"Length of segment differ in original (%u) and filled (%u) op" , orig_op->segments[i].length, data_len); harq_orig = (int8_t *) orig_op->segments[i].addr; harq_out = rte_pktmbuf_mtod_offset(m, int8_t *, offset); +/* Cannot compare HARQ output data for such cases */ +if ((ldpc_llr_decimals >1 ) && ((ops_ld->op_flags &RTE_BBDEV_LDPC_LLR_COMPRESSION ) + || (ops_ld->op_flags &RTE_BBDEV_LDPC_HARQ_6BIT_COMPRESSION ))) +break ; + if (!(ldpc_cap_flags & RTE_BBDEV_LDPC_INTERNAL_HARQ_MEMORY_FILLERS ) || (ops_ld->op_flags &
@@ -2172,7 +2177,7 @@ validate_op_harq_chain(struct rte_bbdev_op_data *op,/* Validate total mbuf pkt length */ uint32_t pkt_len =rte_pktmbuf_pkt_len (op->data) - op->offset; -TEST_ASSERT (total_data_size < pkt_len +64 , +TEST_ASSERT (total_data_size < pkt_len +HARQ_MEM_TOLERANCE ,"Length of data differ in original (%u) and filled (%u) op" , total_data_size, pkt_len);
Signature · 1 line
2.37.1
Re: [PATCH v2 02/16] test/bbdev: extend HARQ tolerance
Quoted 44 lines · Hernan Vargas
HARQ memory implementation could have different size assumptions. Extending HARQ tolerance to cover different implementations.
Fixes 335c11fd276 ("app/bbdev: support HARQ validation")Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)app/test-bbdev/test_bbdev_perf.c +7
-2@@ -95,6 +95,7 @@#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 HARQ_MEM_TOLERANCE 256static struct test_bbdev_vector test_vector;/* Switch between PMD and Interrupt for throughput TC */ @@ -2090,13 +2091,17 @@ validate_op_harq_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; -TEST_ASSERT (orig_op->segments[i].length < - (uint32_t )(data_len +64 ), +TEST_ASSERT (orig_op->segments[i].length < (uint32_t )(data_len +HARQ_MEM_TOLERANCE ),"Length of segment differ in original (%u) and filled (%u) op" , orig_op->segments[i].length, data_len); harq_orig = (int8_t *) orig_op->segments[i].addr; harq_out = rte_pktmbuf_mtod_offset(m, int8_t *, offset); +/* Cannot compare HARQ output data for such cases */ +if ((ldpc_llr_decimals >1 ) && ((ops_ld->op_flags &RTE_BBDEV_LDPC_LLR_COMPRESSION ) + || (ops_ld->op_flags &RTE_BBDEV_LDPC_HARQ_6BIT_COMPRESSION ))) +break ; + if (!(ldpc_cap_flags & RTE_BBDEV_LDPC_INTERNAL_HARQ_MEMORY_FILLERS ) || (ops_ld->op_flags &
app/test-bbdev/test_bbdev_perf.c validate_op_harq_chain(struct rte_bbdev_op_data *op, · line 2183
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c +1
-1@@ -2172,7 +2177,7 @@ validate_op_harq_chain(struct rte_bbdev_op_data *op,/* Validate total mbuf pkt length */ uint32_t pkt_len =rte_pktmbuf_pkt_len (op->data) - op->offset; -TEST_ASSERT (total_data_size < pkt_len +64 , +TEST_ASSERT (total_data_size < pkt_len +HARQ_MEM_TOLERANCE ,"Length of data differ in original (%u) and filled (%u) op" , total_data_size, pkt_len);
Thanks,
Maxime
PATCH v2 3/16 test/bbdev: refactor TB throughput report
In reply to Hernan VargasRefactor calculation for tb_size.
No functional impact.
--- app/test-bbdev/test_bbdev_perf.c | 38 ++++++++++++++------------------ 1 file changed, 16 insertions(+), 22 deletions(-)
app/test-bbdev/test_bbdev_perf.c +16 -22
@@ -2580,19 +2580,16 @@ calc_dec_TB_size(struct rte_bbdev_dec_op *op)static uint32_t calc_ldpc_dec_TB_size (struct rte_bbdev_dec_op *op) { -uint8_t i; -uint32_t c, r, tb_size =0 ; +uint8_t num_cbs =0 ; +uint32_t tb_size =0 ;uint16_t sys_cols = (op->ldpc_dec.basegraph ==1 ) ?22 :10 ; -if (op->ldpc_dec.code_block_mode ==RTE_BBDEV_CODE_BLOCK ) { - tb_size = sys_cols * op->ldpc_dec.z_c - op->ldpc_dec.n_filler; - }else { - c = op->ldpc_dec.tb_params.c; - r = op->ldpc_dec.tb_params.r; -for (i =0 ; i < c-r; i++) - tb_size += sys_cols * op->ldpc_dec.z_c - - op->ldpc_dec.n_filler; - } +if (op->ldpc_dec.code_block_mode ==RTE_BBDEV_CODE_BLOCK ) + num_cbs =1 ; +else + num_cbs = op->ldpc_dec.tb_params.c - op->ldpc_dec.tb_params.r; + + tb_size = (sys_cols * op->ldpc_dec.z_c - op->ldpc_dec.n_filler) * num_cbs;return tb_size; }
@@ -2618,19 +2615,16 @@ calc_enc_TB_size(struct rte_bbdev_enc_op *op)static uint32_t calc_ldpc_enc_TB_size (struct rte_bbdev_enc_op *op) { -uint8_t i; -uint32_t c, r, tb_size =0 ; +uint8_t num_cbs =0 ; +uint32_t tb_size =0 ;uint16_t sys_cols = (op->ldpc_enc.basegraph ==1 ) ?22 :10 ; -if (op->ldpc_enc.code_block_mode ==RTE_BBDEV_CODE_BLOCK ) { - tb_size = sys_cols * op->ldpc_enc.z_c - op->ldpc_enc.n_filler; - }else { - c = op->turbo_enc.tb_params.c; - r = op->turbo_enc.tb_params.r; -for (i =0 ; i < c-r; i++) - tb_size += sys_cols * op->ldpc_enc.z_c - - op->ldpc_enc.n_filler; - } +if (op->ldpc_enc.code_block_mode ==RTE_BBDEV_CODE_BLOCK ) + num_cbs =1 ; +else + num_cbs = op->ldpc_enc.tb_params.c - op->ldpc_enc.tb_params.r; + + tb_size = (sys_cols * op->ldpc_enc.z_c - op->ldpc_enc.n_filler) * num_cbs;return tb_size; }
Signature · 1 line
2.37.1
Re: [PATCH v2 03/16] test/bbdev: refactor TB throughput report
Quoted 8 lines · Hernan Vargas
Refactor calculation for
tb_size.
No functional impact.Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 38 ++++++++++++++------------------ 1 file changed, 16 insertions(+), 22 deletions(-)
Thanks,
Maxime
PATCH v2 4/16 test/bbdev: add timeout for latency tests
In reply to Hernan VargasAdd a timeout to force exit the latency tests in case dequeue never happens.
--- app/test-bbdev/test_bbdev_perf.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)
app/test-bbdev/test_bbdev_perf.c +19 -5
@@ -26,6 +26,7 @@#define MAX_QUEUES RTE_MAX_LCORE#define TEST_REPETITIONS 100 +#define TIME_OUT_POLL 1e8#define WAIT_OFFLOAD_US 1000#ifdef RTE_BASEBAND_FPGA_LTE_FEC
@@ -4546,6 +4547,7 @@ latency_test_ldpc_dec(struct rte_mempool *mempool,for (i =0 , dequeued =0 ; dequeued < num_to_process; ++i) {uint16_t enq =0 , deq =0 ; +uint32_t time_out =0 ;bool first_time = true; last_time =0 ;
@@ -4597,7 +4599,8 @@ latency_test_ldpc_dec(struct rte_mempool *mempool, last_time =rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time);
@@ -4606,7 +4609,12 @@ latency_test_ldpc_dec(struct rte_mempool *mempool,if (extDdr)retrieve_harq_ddr (dev_id, queue_id, ops_enq, burst_sz); -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info); +TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret =validate_ldpc_dec_op (ops_deq, burst_sz, ref_op, vector_mask);TEST_ASSERT_SUCCESS (ret,"Validation failed!" );
@@ -4632,6 +4640,7 @@ latency_test_enc(struct rte_mempool *mempool,for (i =0 , dequeued =0 ; dequeued < num_to_process; ++i) {uint16_t enq =0 , deq =0 ; +uint32_t time_out =0 ;bool first_time = true; last_time =0 ;
@@ -4667,13 +4676,18 @@ latency_test_enc(struct rte_mempool *mempool, last_time +=rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time); *total_time += last_time; - -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info); +TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret =validate_enc_op (ops_deq, burst_sz, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }
Signature · 1 line
2.37.1
Re: [PATCH v2 04/16] test/bbdev: add timeout for latency tests
Quoted 39 lines · Hernan Vargas
Add a timeout to force exit the latency tests in case dequeue never happens.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)app/test-bbdev/test_bbdev_perf.c +4
-1@@ -26,6 +26,7 @@#define MAX_QUEUES RTE_MAX_LCORE#define TEST_REPETITIONS 100 +#define TIME_OUT_POLL 1e8#define WAIT_OFFLOAD_US 1000#ifdef RTE_BASEBAND_FPGA_LTE_FEC @@ -4546,6 +4547,7 @@ latency_test_ldpc_dec(struct rte_mempool *mempool,for (i =0 , dequeued =0 ; dequeued < num_to_process; ++i) {uint16_t enq =0 , deq =0 ; +uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4597,7 +4599,8 @@ latency_test_ldpc_dec(struct rte_mempool *mempool, last_time =rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time);
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool, · line 4615
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c +4
-1@@ -4606,7 +4609,12 @@ latency_test_ldpc_dec(struct rte_mempool *mempool,if (extDdr)retrieve_harq_ddr (dev_id, queue_id, ops_enq, burst_sz); -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);
What is the point of calling rte_bbdev_info_get() here and below? info is not used afterwards.
Quoted 20 lines
app/test-bbdev/test_bbdev_perf.c +5
-1+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret =validate_ldpc_dec_op (ops_deq, burst_sz, ref_op, vector_mask);TEST_ASSERT_SUCCESS (ret,"Validation failed!" );@@ -4632,6 +4640,7 @@ latency_test_enc(struct rte_mempool *mempool,for (i =0 , dequeued =0 ; dequeued < num_to_process; ++i) {uint16_t enq =0 , deq =0 ; +uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4667,13 +4676,18 @@ latency_test_enc(struct rte_mempool *mempool, last_time +=rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL ));
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool *mempool, · line 4688
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-2*max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time); *total_time += last_time; - -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);
Same here.
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret =validate_enc_op (ops_deq, burst_sz, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }
RE: [PATCH v2 04/16] test/bbdev: add timeout for latency tests
Hi Maxime,
Quoted 59 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Monday, February 20, 2023 10:33 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsapp/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool
Quoted 31 lines · Hernan Vargas
Add a timeout to force exit the latency tests in case dequeue never happens.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)app/test-bbdev/test_bbdev_perf.c +2
-0@@ -26,6 +26,7 @@#define MAX_QUEUES RTE_MAX_LCORE#define TEST_REPETITIONS 100 +#define TIME_OUT_POLL 1e8#define WAIT_OFFLOAD_US 1000#ifdef RTE_BASEBAND_FPGA_LTE_FEC @@ -4546,6 +4547,7 @@ latency_test_ldpc_dec(struct rte_mempool *mempool, for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) {uint16_t enq =0 , deq =0 ; +uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4597,7 +4599,8 @@ latency_test_ldpc_dec(struct rte_mempool*mempool,
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 9 lines
app/test-bbdev/test_bbdev_perf.c +2
-1last_time =rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time = RTE_MIN(*min_time, last_time); @@ -4606,7+4609,12 @@
Quoted 3 lines
latency_test_ldpc_dec(struct
rte_mempool*mempool,
if (extDdr)
retrieve_harq_ddr(dev_id,queue_id,ops_enq,
burst_sz);Quoted 1 line
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool, · line 4615
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +4
-1-if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);What is the point of calling
rte_bbdev_info_get()here and below? info is not used afterwards.
The reason for calling this function is to check the device status and if there is something wrong the PMD would display it to standard output.
Quoted 42 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 4 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret = validate_ldpc_dec_op(ops_deq, burst_sz,
ref_op,Quoted 2 lines
vector_mask);
TEST_ASSERT_SUCCESS(ret, "Validation failed!"); @@-4632,6
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool
Quoted 9 lines
+4640,7 @@ latency_test_enc(struct
rte_mempool*mempool,for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) { uint16_t enq = 0, deq = 0;app/test-bbdev/test_bbdev_perf.c +1
-0+uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4667,13 +4676,18 @@ latency_test_enc(struct rte_mempool*mempool,
Quoted 6 lines
app/test-bbdev/test_bbdev_perf.c +2
-1last_time +=rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL ));app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool *mempool, · line 4688
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-2*max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time); *total_time += last_time; - -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);Same here.
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret =validate_enc_op (ops_deq, burst_sz, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }
Re: [PATCH v2 04/16] test/bbdev: add timeout for latency tests
Quoted 70 lines · Hernan
Hi Maxime,
Quoted 59 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Monday, February 20, 2023 10:33 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsapp/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool
Quoted 31 lines · Hernan Vargas
Add a timeout to force exit the latency tests in case dequeue never happens.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)app/test-bbdev/test_bbdev_perf.c +2
-0@@ -26,6 +26,7 @@#define MAX_QUEUES RTE_MAX_LCORE#define TEST_REPETITIONS 100 +#define TIME_OUT_POLL 1e8#define WAIT_OFFLOAD_US 1000#ifdef RTE_BASEBAND_FPGA_LTE_FEC @@ -4546,6 +4547,7 @@ latency_test_ldpc_dec(struct rte_mempool *mempool, for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) {uint16_t enq =0 , deq =0 ; +uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4597,7 +4599,8 @@ latency_test_ldpc_dec(struct rte_mempool*mempool,
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 9 lines
app/test-bbdev/test_bbdev_perf.c +2
-1last_time =rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time = RTE_MIN(*min_time, last_time); @@ -4606,7+4609,12 @@
Quoted 3 lines
latency_test_ldpc_dec(struct
rte_mempool*mempool,
if (extDdr)
retrieve_harq_ddr(dev_id,queue_id,ops_enq,
burst_sz);Quoted 1 line
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool, · line 4615
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +4
-1-if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);What is the point of calling
rte_bbdev_info_get()here and below? info is not used afterwards.
Quoted 1 line
The reason for calling this function is to check the device status and if there is something wrong the PMD would display it to standard output.
What kind of info exactly, I don't see much meaningful logs in rte_bbdev_info_get() except printing error if dev_info == NULL.
Regards,
Maxime
Quoted 43 lines
Quoted 42 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 4 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret = validate_ldpc_dec_op(ops_deq, burst_sz,
ref_op,Quoted 2 lines
vector_mask);
TEST_ASSERT_SUCCESS(ret, "Validation failed!"); @@-4632,6
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool
Quoted 9 lines
+4640,7 @@ latency_test_enc(struct
rte_mempool*mempool,for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) { uint16_t enq = 0, deq = 0;app/test-bbdev/test_bbdev_perf.c +1
-0+uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4667,13 +4676,18 @@ latency_test_enc(struct rte_mempool*mempool,
Quoted 6 lines
app/test-bbdev/test_bbdev_perf.c +2
-1last_time +=rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL ));app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool *mempool, · line 4688
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-2*max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time); *total_time += last_time; - -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);Same here.
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret =validate_enc_op (ops_deq, burst_sz, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }
RE: [PATCH v2 04/16] test/bbdev: add timeout for latency tests
Hi Maxime,
Quoted 83 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Thursday, February 23, 2023 2:32 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsQuoted 71 lines · Hernan
Hi Maxime,
Quoted 68 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Monday, February 20, 2023 10:33 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsapp/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool
Quoted 31 lines · Hernan Vargas
Add a timeout to force exit the latency tests in case dequeue never happens.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)app/test-bbdev/test_bbdev_perf.c +2
-0@@ -26,6 +26,7 @@#define MAX_QUEUES RTE_MAX_LCORE#define TEST_REPETITIONS 100 +#define TIME_OUT_POLL 1e8#define WAIT_OFFLOAD_US 1000#ifdef RTE_BASEBAND_FPGA_LTE_FEC @@ -4546,6 +4547,7 @@ latency_test_ldpc_dec(struct rte_mempool *mempool, for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) {uint16_t enq =0 , deq =0 ; +uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4597,7 +4599,8 @@ latency_test_ldpc_dec(struct rte_mempool*mempool,
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 9 lines
app/test-bbdev/test_bbdev_perf.c +2
-1last_time =rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time = RTE_MIN(*min_time, last_time); @@ -4606,7+4609,12 @@
Quoted 3 lines
latency_test_ldpc_dec(struct
rte_mempool*mempool,
if (extDdr)
retrieve_harq_ddr(dev_id,queue_id,ops_enq,
burst_sz);app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool, · line 4615
Quoted 6 lines
app/test-bbdev/test_bbdev_perf.c +4
-1-if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);What is the point of calling
rte_bbdev_info_get()here and below? info is not used afterwards.
Quoted 5 lines
Quoted 1 line
The reason for calling this function is to check the device status and if there
is something wrong the PMD would display it to standard output.
What kind of info exactly, I don't see much meaningful logs in
rte_bbdev_info_get()except printing error ifdev_info== NULL.
rte_bbdev_info_get() calls the device's info_get function that is specified in the PMD. For example, for ACC100, acc100_dev_info_get() gets called to check the device status.
Thanks,
Hernan
Quoted 47 lines
Regards,
MaximeQuoted 43 lines
Quoted 42 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 4 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret = validate_ldpc_dec_op(ops_deq, burst_sz,
ref_op,Quoted 2 lines
vector_mask);
TEST_ASSERT_SUCCESS(ret, "Validation failed!"); @@-4632,6
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool
Quoted 9 lines
+4640,7 @@ latency_test_enc(struct
rte_mempool*mempool,for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) { uint16_t enq = 0, deq = 0;app/test-bbdev/test_bbdev_perf.c +1
-0+uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4667,13 +4676,18 @@ latency_test_enc(struct rte_mempool*mempool,
Quoted 6 lines
app/test-bbdev/test_bbdev_perf.c +2
-1last_time +=rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL ));app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool *mempool, · line 4688
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-2*max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time); *total_time += last_time; - -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);Same here.
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret =validate_enc_op (ops_deq, burst_sz, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }
Re: [PATCH v2 04/16] test/bbdev: add timeout for latency tests
Quoted 85 lines · Hernan
Hi Maxime,
Quoted 83 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Thursday, February 23, 2023 2:32 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsQuoted 71 lines · Hernan
Hi Maxime,
Quoted 68 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Monday, February 20, 2023 10:33 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsapp/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool
Quoted 31 lines · Hernan Vargas
Add a timeout to force exit the latency tests in case dequeue never happens.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)app/test-bbdev/test_bbdev_perf.c +2
-0@@ -26,6 +26,7 @@#define MAX_QUEUES RTE_MAX_LCORE#define TEST_REPETITIONS 100 +#define TIME_OUT_POLL 1e8#define WAIT_OFFLOAD_US 1000#ifdef RTE_BASEBAND_FPGA_LTE_FEC @@ -4546,6 +4547,7 @@ latency_test_ldpc_dec(struct rte_mempool *mempool, for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) {uint16_t enq =0 , deq =0 ; +uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4597,7 +4599,8 @@ latency_test_ldpc_dec(struct rte_mempool*mempool,
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 9 lines
app/test-bbdev/test_bbdev_perf.c +2
-1last_time =rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time = RTE_MIN(*min_time, last_time); @@ -4606,7+4609,12 @@
Quoted 3 lines
latency_test_ldpc_dec(struct
rte_mempool*mempool,
if (extDdr)
retrieve_harq_ddr(dev_id,queue_id,ops_enq,
burst_sz);app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool, · line 4615
Quoted 6 lines
app/test-bbdev/test_bbdev_perf.c +4
-1-if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);What is the point of calling
rte_bbdev_info_get()here and below? info is not used afterwards.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
The reason for calling this function is to check the device status and if there
is something wrong the PMD would display it to standard output.
What kind of info exactly, I don't see much meaningful logs in
rte_bbdev_info_get()except printing error ifdev_info== NULL.
rte_bbdev_info_get()calls the device'sinfo_getfunction that is specified in the PMD. For example, for ACC100,acc100_dev_info_get()gets called to check the device status.
Ok, I looked at this function and it might be more relevant to mention vrb1 than acc100, because acc100 does not support device status:
/* Check the status of device */ dev_info->device_status = RTE_BBDEV_DEV_NOT_SUPPORTED;
Also, this dev_info->device_status field is set by all the drivers but never read (neither by the driver, nor library nor apps).
So if this is the only reason rte_bbdev_info_get() is called, that is quite useless. Am I missing someting?
Thanks,
Maxime
Quoted 50 lines
Thanks,
HernanQuoted 47 lines
Regards,
MaximeQuoted 43 lines
Quoted 42 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 4 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret = validate_ldpc_dec_op(ops_deq, burst_sz,
ref_op,Quoted 2 lines
vector_mask);
TEST_ASSERT_SUCCESS(ret, "Validation failed!"); @@-4632,6
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool
Quoted 9 lines
+4640,7 @@ latency_test_enc(struct
rte_mempool*mempool,for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) { uint16_t enq = 0, deq = 0;app/test-bbdev/test_bbdev_perf.c +1
-0+uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4667,13 +4676,18 @@ latency_test_enc(struct rte_mempool*mempool,
Quoted 6 lines
app/test-bbdev/test_bbdev_perf.c +2
-1last_time +=rte_rdtsc_precise () - start_time; first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL ));app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool *mempool, · line 4688
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-2*max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time); *total_time += last_time; - -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);Same here.
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret =validate_enc_op (ops_deq, burst_sz, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }
RE: [PATCH v2 04/16] test/bbdev: add timeout for latency tests
Hi Maxime,
Quoted 114 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Monday, February 27, 2023 1:45 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsQuoted 62 lines · Hernan
Hi Maxime,
Quoted 60 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Thursday, February 23, 2023 2:32 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsQuoted 47 lines · Hernan
Hi Maxime,
Quoted 45 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Monday, February 20, 2023 10:33 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsapp/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool
Quoted 30 lines · Hernan Vargas
Add a timeout to force exit the latency tests in case dequeue never happens.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)app/test-bbdev/test_bbdev_perf.c +1
-0@@ -26,6 +26,7 @@#define MAX_QUEUES RTE_MAX_LCORE#define TEST_REPETITIONS 100 +#define TIME_OUT_POLL 1e8#define WAIT_OFFLOAD_US 1000 #ifdef RTE_BASEBAND_FPGA_LTE_FEC @@ -4546,6 +4547,7 @@latency_test_ldpc_dec(struct
rte_mempool*mempool,for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) { uint16_t enq = 0, deq = 0;app/test-bbdev/test_bbdev_perf.c +1
-0+uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4597,7 +4599,8 @@ latency_test_ldpc_dec(struct rte_mempool*mempool,
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0last_time = rte_rdtsc_precise() -
start_time;Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c +2
-1first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time = RTE_MIN(*min_time, last_time); @@ -4606,7
Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 14 lines
+4609,12 @@
Quoted 3 lines
latency_test_ldpc_dec(struct
rte_mempool*mempool,
if (extDdr)
retrieve_harq_ddr(dev_id,queue_id,ops_enq,
burst_sz);app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool, · line 4615
Quoted 6 lines
app/test-bbdev/test_bbdev_perf.c +4
-1-if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);What is the point of calling
rte_bbdev_info_get()here and below? info is not used afterwards.Quoted 8 lines
Quoted 6 lines
Quoted 2 lines
The reason for calling this function is to check the device status and if there
is something wrong the PMD would display it to standard output.
What kind of info exactly, I don't see much meaningful logs in
rte_bbdev_info_get()except printing error ifdev_info== NULL.
rte_bbdev_info_get()calls the device'sinfo_getfunction that is specified inthe PMD.
Quoted 1 line
For example, for ACC100,
acc100_dev_info_get()gets called to check thedevice status.
Ok, I looked at this function and it might be more relevant to mention vrb1 than acc100, because acc100 does not support device status:
Quoted 8 lines
/* Check the status of device */ dev_info->device_status = RTE_BBDEV_DEV_NOT_SUPPORTED;Also, this
dev_info->device_statusfield is set by all the drivers but never read (neither by the driver, nor library nor apps).So if this is the only reason
rte_bbdev_info_get()is called, that is quite useless. Am I missing someting?
Note the device status is printed in the bbdev-test app in that other patchset in the series https://patches.dpdk.org/project/dpdk/patch/[email protected]/
Also calling that function then exercise interaction with underlying HW and PF driver which is valuable and the expected usecase.
This is indeed relevant from vrb1 in term of current intel PMD implementation, but the bbdev-test is HW agnostic.
Let us know if unclear,
Thanks,
Nic
Quoted 58 lines
Thanks,
MaximeQuoted 13 lines
Thanks,
HernanQuoted 11 lines
Regards,
MaximeQuoted 7 lines
Quoted 7 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 4 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret = validate_ldpc_dec_op(ops_deq, burst_sz,
ref_op,Quoted 2 lines
vector_mask); TEST_ASSERT_SUCCESS(ret, "Validationfailed!"); @@
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
-4632,6
Quoted 8 lines
+4640,7 @@ latency_test_enc(struct
rte_mempool*mempool,for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) { uint16_t enq = 0, deq = 0;app/test-bbdev/test_bbdev_perf.c +1
-0+uint32_t time_out =0 ;bool first_time = true; last_time =0 ;Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0@@ -4667,13 +4676,18 @@ latency_test_enc(struct rte_mempool*mempool,
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0last_time += rte_rdtsc_precise() -
start_time;Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +2
-1first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL ));app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool *mempool, · line 4688
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-2*max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time); *total_time += last_time; - -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);Same here.
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool *mempool,
Quoted 3 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret = validate_enc_op(ops_deq, burst_sz,
ref_op);Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool · line 4692
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0TEST_ASSERT_SUCCESS(ret, "Validationfailed!");
Quoted 3 lines
Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0}
Maxime Coquelin
Reviewed-byRe: [PATCH v2 04/16] test/bbdev: add timeout for latency tests
Quoted 119 lines · Nicolas
Hi Maxime,
Quoted 117 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Monday, February 27, 2023 1:45 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsQuoted 62 lines · Hernan
Hi Maxime,
Quoted 60 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Thursday, February 23, 2023 2:32 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsQuoted 47 lines · Hernan
Hi Maxime,
Quoted 45 lines
-----Original Message-----
From: Maxime Coquelin <[email protected]>
Sent: Monday, February 20, 2023 10:33 AM
To: Vargas, Hernan <[email protected]>; [email protected]; [email protected]; Rix, Tom <[email protected]>
Cc: Chautru, Nicolas <[email protected]>; Zhang, Qi Z <[email protected]>
Subject: Re: [PATCH v2 04/16] test/bbdev: add timeout for latency testsapp/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool
Quoted 30 lines · Hernan Vargas
Add a timeout to force exit the latency tests in case dequeue never happens.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)app/test-bbdev/test_bbdev_perf.c +1
-0@@ -26,6 +26,7 @@#define MAX_QUEUES RTE_MAX_LCORE#define TEST_REPETITIONS 100 +#define TIME_OUT_POLL 1e8#define WAIT_OFFLOAD_US 1000 #ifdef RTE_BASEBAND_FPGA_LTE_FEC @@ -4546,6 +4547,7 @@latency_test_ldpc_dec(struct
rte_mempool*mempool,for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) { uint16_t enq = 0, deq = 0;app/test-bbdev/test_bbdev_perf.c +1
-0+uint32_t time_out =0 ;bool first_time = true; last_time =0 ;@@ -4597,7 +4599,8 @@ latency_test_ldpc_dec(struct rte_mempool*mempool,
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0last_time = rte_rdtsc_precise() -
start_time;Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c latency_test_ldpc_dec(struct rte_mempool · line 4606
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c +2
-1first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL )); *max_time =RTE_MAX (*max_time, last_time); *min_time = RTE_MIN(*min_time, last_time); @@ -4606,7
Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 14 lines
+4609,12 @@
Quoted 3 lines
latency_test_ldpc_dec(struct
rte_mempool*mempool,
if (extDdr)
retrieve_harq_ddr(dev_id,queue_id,ops_enq,
burst_sz);app/test-bbdev/test_bbdev_perf.c
Quoted 6 lines
app/test-bbdev/test_bbdev_perf.c +4
-1-if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);What is the point of calling
rte_bbdev_info_get()here and below? info is not used afterwards.Quoted 8 lines
Quoted 6 lines
Quoted 2 lines
The reason for calling this function is to check the device status and if there
is something wrong the PMD would display it to standard output.
What kind of info exactly, I don't see much meaningful logs in
rte_bbdev_info_get()except printing error ifdev_info== NULL.
rte_bbdev_info_get()calls the device'sinfo_getfunction that is specified inthe PMD.
Quoted 1 line
For example, for ACC100,
acc100_dev_info_get()gets called to check thedevice status.
Ok, I looked at this function and it might be more relevant to mention vrb1 than acc100, because acc100 does not support device status:
/* Check the status of device */ dev_info->device_status = RTE_BBDEV_DEV_NOT_SUPPORTED;
Quoted 7 lines
Quoted 5 lines
Also, this
dev_info->device_statusfield is set by all the drivers but never read (neither by the driver, nor library nor apps).So if this is the only reason
rte_bbdev_info_get()is called, that is quite useless. Am I missing someting?Note the device status is printed in the bbdev-test app in that other patchset in the series https://patches.dpdk.org/project/dpdk/patch/[email protected]/
Not related to this patch, but maybe you should use an assert in the patch you link above. Indeed, what is the point of continuing the init if the device status is "FATAL_ERR"?
Quoted 3 lines
Also calling that function then exercise interaction with underlying HW and PF driver which is valuable and the expected usecase. This is indeed relevant from vrb1 in term of current intel PMD implementation, but the bbdev-test is HW agnostic. Let us know if unclear,
OK, that's clearer.
For this patch:
Thanks,
Maxime
Quoted 63 lines
Thanks,
NicQuoted 58 lines
Thanks,
MaximeQuoted 13 lines
Thanks,
HernanQuoted 11 lines
Regards,
MaximeQuoted 7 lines
Quoted 7 lines
app/test-bbdev/test_bbdev_perf.c
Quoted 4 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret = validate_ldpc_dec_op(ops_deq, burst_sz,
ref_op,Quoted 2 lines
vector_mask); TEST_ASSERT_SUCCESS(ret, "Validationfailed!"); @@
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
-4632,6
Quoted 8 lines
+4640,7 @@ latency_test_enc(struct
rte_mempool*mempool,for (i = 0, dequeued = 0; dequeued < num_to_process; ++i) { uint16_t enq = 0, deq = 0;app/test-bbdev/test_bbdev_perf.c +1
-0+uint32_t time_out =0 ;bool first_time = true; last_time =0 ;Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0@@ -4667,13 +4676,18 @@ latency_test_enc(struct rte_mempool*mempool,
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0last_time += rte_rdtsc_precise() -
start_time;Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 5 lines
app/test-bbdev/test_bbdev_perf.c +2
-1first_time = false; } - }while (unlikely (burst_sz != deq)); + time_out++; + }while ((burst_sz != deq) && (time_out <TIME_OUT_POLL ));app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool · line 4688
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-2*max_time =RTE_MAX (*max_time, last_time); *min_time =RTE_MIN (*min_time, last_time); *total_time += last_time; - -if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { +if (burst_sz != deq) { +struct rte_bbdev_info info; + ret =TEST_FAILED ; +rte_bbdev_info_get (dev_id, &info);Same here.
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool · line 4691
Quoted 3 lines
app/test-bbdev/test_bbdev_perf.c +2
-0+TEST_ASSERT_SUCCESS (ret,"Dequeue timeout!" ); + }else if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { ret = validate_enc_op(ops_deq, burst_sz,
ref_op);Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c latency_test_enc(struct rte_mempool · line 4692
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0TEST_ASSERT_SUCCESS(ret, "Validationfailed!");
Quoted 3 lines
Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line
app/test-bbdev/test_bbdev_perf.c +0
-0}
PATCH v2 5/16 test/bbdev: enable early termination for validation
In reply to Hernan VargasEnhancement to enable early termination for validation tests if the device supports it.
--- app/test-bbdev/test_bbdev_perf.c | 30 +++++++++++++++++++----------- 1 file changed, 19 insertions(+), 11 deletions(-)
app/test-bbdev/test_bbdev_perf.c +19 -11
@@ -3775,11 +3775,11 @@ throughput_pmd_lcore_ldpc_dec(void *arg)TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_ops);/* For throughput tests we need to disable early termination */ -if (check_bit (ref_op->ldpc_dec.op_flags, -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) - ref_op->ldpc_dec.op_flags -= -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; +if (check_bit (ref_op->ldpc_dec.op_flags,RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) + ref_op->ldpc_dec.op_flags -=RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; + ref_op->ldpc_dec.iter_max =get_iter_max (); +/* Since ET is disabled, the expected iter_count is iter_max */ ref_op->ldpc_dec.iter_count = ref_op->ldpc_dec.iter_max;if (test_vector.op_type !=RTE_BBDEV_OP_NONE )
@@ -4465,7 +4465,7 @@ latency_test_dec(struct rte_mempool *mempool, struct test_buffers *bufs, struct rte_bbdev_dec_op *ref_op, int vector_mask, uint16_t dev_id, uint16_t queue_id, const uint16_t num_to_process, uint16_t burst_sz, - uint64_t *total_time, uint64_t *min_time, uint64_t *max_time) + uint64_t *total_time, uint64_t *min_time, uint64_t *max_time, bool disable_et) { int ret = TEST_SUCCESS;uint16_t i, j, dequeued;
@@ -4481,8 +4481,14 @@ latency_test_dec(struct rte_mempool *mempool, burst_sz = num_to_process - dequeued; ret =rte_bbdev_dec_op_alloc_bulk (mempool, ops_enq, burst_sz); -TEST_ASSERT_SUCCESS (ret, -"rte_bbdev_dec_op_alloc_bulk() failed" ); +TEST_ASSERT_SUCCESS (ret,"rte_bbdev_dec_op_alloc_bulk() failed" ); + + ref_op->turbo_dec.iter_max =get_iter_max (); +/* For validation tests we want to enable early termination */ +if (!disable_et && !check_bit (ref_op->turbo_dec.op_flags, +RTE_BBDEV_TURBO_EARLY_TERMINATION )) + ref_op->turbo_dec.op_flags |=RTE_BBDEV_TURBO_EARLY_TERMINATION ; +if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_dec_op(ops_enq, burst_sz, dequeued, bufs->inputs,
@@ -4561,10 +4567,12 @@ latency_test_ldpc_dec(struct rte_mempool *mempool,/* For latency tests we need to disable early termination */ if (disable_et &&check_bit (ref_op->ldpc_dec.op_flags,RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) - ref_op->ldpc_dec.op_flags -= -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; + ref_op->ldpc_dec.op_flags -=RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; + ref_op->ldpc_dec.iter_max =get_iter_max (); - ref_op->ldpc_dec.iter_count = ref_op->ldpc_dec.iter_max; +/* When ET is disabled, the expected iter_count is iter_max */ +if (disable_et) + ref_op->ldpc_dec.iter_count = ref_op->ldpc_dec.iter_max;if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_ldpc_dec_op(ops_enq, burst_sz, dequeued,
@@ -4873,7 +4881,7 @@ validation_latency_test(struct active_device *ad, iter =latency_test_dec (op_params->mp, bufs, op_params->ref_dec_op, op_params->vector_mask, ad->dev_id, queue_id, num_to_process, - burst_sz, &total_time, &min_time, &max_time); + burst_sz, &total_time, &min_time, &max_time, latency_flag); else if (op_type == RTE_BBDEV_OP_LDPC_ENC) iter = latency_test_ldpc_enc(op_params->mp, bufs, op_params->ref_enc_op, ad->dev_id, queue_id,
Signature · 1 line
2.37.1
Re: [PATCH v2 05/16] test/bbdev: enable early termination for validation
Quoted 71 lines · Hernan Vargas
Enhancement to enable early termination for validation tests if the device supports it.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 30 +++++++++++++++++++----------- 1 file changed, 19 insertions(+), 11 deletions(-)app/test-bbdev/test_bbdev_perf.c +18
-10@@ -3775,11 +3775,11 @@ throughput_pmd_lcore_ldpc_dec(void *arg)TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_ops);/* For throughput tests we need to disable early termination */ -if (check_bit (ref_op->ldpc_dec.op_flags, -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) - ref_op->ldpc_dec.op_flags -= -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; +if (check_bit (ref_op->ldpc_dec.op_flags,RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) + ref_op->ldpc_dec.op_flags -=RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; + ref_op->ldpc_dec.iter_max =get_iter_max (); +/* Since ET is disabled, the expected iter_count is iter_max */ ref_op->ldpc_dec.iter_count = ref_op->ldpc_dec.iter_max;if (test_vector.op_type !=RTE_BBDEV_OP_NONE )@@ -4465,7 +4465,7 @@ latency_test_dec(struct rte_mempool *mempool, struct test_buffers *bufs, struct rte_bbdev_dec_op *ref_op, int vector_mask, uint16_t dev_id, uint16_t queue_id, const uint16_t num_to_process, uint16_t burst_sz, - uint64_t *total_time, uint64_t *min_time, uint64_t *max_time) + uint64_t *total_time, uint64_t *min_time, uint64_t *max_time, bool disable_et) { int ret = TEST_SUCCESS;uint16_t i, j, dequeued;@@ -4481,8 +4481,14 @@ latency_test_dec(struct rte_mempool *mempool, burst_sz = num_to_process - dequeued; ret =rte_bbdev_dec_op_alloc_bulk (mempool, ops_enq, burst_sz); -TEST_ASSERT_SUCCESS (ret, -"rte_bbdev_dec_op_alloc_bulk() failed" ); +TEST_ASSERT_SUCCESS (ret,"rte_bbdev_dec_op_alloc_bulk() failed" ); + + ref_op->turbo_dec.iter_max =get_iter_max (); +/* For validation tests we want to enable early termination */ +if (!disable_et && !check_bit (ref_op->turbo_dec.op_flags, +RTE_BBDEV_TURBO_EARLY_TERMINATION )) + ref_op->turbo_dec.op_flags |=RTE_BBDEV_TURBO_EARLY_TERMINATION ; +if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_dec_op(ops_enq, burst_sz, dequeued, bufs->inputs,@@ -4561,10 +4567,12 @@ latency_test_ldpc_dec(struct rte_mempool *mempool,/* For latency tests we need to disable early termination */ if (disable_et &&check_bit (ref_op->ldpc_dec.op_flags,RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) - ref_op->ldpc_dec.op_flags -= -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; + ref_op->ldpc_dec.op_flags -=RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; + ref_op->ldpc_dec.iter_max =get_iter_max (); - ref_op->ldpc_dec.iter_count = ref_op->ldpc_dec.iter_max; +/* When ET is disabled, the expected iter_count is iter_max */ +if (disable_et) + ref_op->ldpc_dec.iter_count = ref_op->ldpc_dec.iter_max;if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_ldpc_dec_op(ops_enq, burst_sz, dequeued,
app/test-bbdev/test_bbdev_perf.c validation_latency_test(struct active_device *ad, · line 4887
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c +1
-1@@ -4873,7 +4881,7 @@ validation_latency_test(struct active_device *ad, iter =latency_test_dec (op_params->mp, bufs, op_params->ref_dec_op, op_params->vector_mask, ad->dev_id, queue_id, num_to_process, - burst_sz, &total_time, &min_time, &max_time); + burst_sz, &total_time, &min_time, &max_time, latency_flag); else if (op_type == RTE_BBDEV_OP_LDPC_ENC) iter = latency_test_ldpc_enc(op_params->mp, bufs, op_params->ref_enc_op, ad->dev_id, queue_id,
Thanks,
Maxime
PATCH v2 6/16 test/bbdev: report device status in test-bbdev
In reply to Hernan VargasNo functional impact.
--- app/test-bbdev/test_bbdev_perf.c | 2 ++ 1 file changed, 2 insertions(+)
app/test-bbdev/test_bbdev_perf.c +2 -0
@@ -850,6 +850,8 @@ add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info,#endif /* Let's refresh this now this is configured */ 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)); nb_queues =RTE_MIN (rte_lcore_count (), info->drv.max_num_queues); nb_queues =RTE_MIN (nb_queues, (unsigned int )MAX_QUEUES );
Signature · 1 line
2.37.1
Re: [PATCH v2 06/16] test/bbdev: report device status in test-bbdev
Quoted 12 lines · Hernan Vargas
No functional impact.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 2 ++ 1 file changed, 2 insertions(+)app/test-bbdev/test_bbdev_perf.c +0
-0
app/test-bbdev/test_bbdev_perf.c add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info, · line 857
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c +2
-0@@ -850,6 +850,8 @@ add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info,#endif /* Let's refresh this now this is configured */ 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)); nb_queues =RTE_MIN (rte_lcore_count (), info->drv.max_num_queues); nb_queues =RTE_MIN (nb_queues, (unsigned int )MAX_QUEUES );
Thanks,
Maxime
PATCH v2 7/16 test/bbdev: test start/stop bbdev API
In reply to Hernan VargasAdd a call to queue start and queue stop specifically for testing the bbdev API.
--- app/test-bbdev/test_bbdev_perf.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
app/test-bbdev/test_bbdev_perf.c +4 -2
@@ -890,6 +890,7 @@ add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info, ret =rte_bbdev_queue_configure (ad->dev_id, queue_id, &qconf); } +rte_bbdev_queue_start (ad->dev_id, queue_id);if (ret !=0 ) {printf ("All queues on dev %u allocated: %u\n" , dev_id, queue_id);
@@ -898,8 +899,8 @@ add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info, ad->queue_ids[queue_id] = queue_id; }TEST_ASSERT (queue_id !=0 , -"ERROR Failed to configure any queues on dev %u" , - dev_id); +"ERROR Failed to configure any queues on dev %u\n" +"\tthe device may not support or have been configured" , dev_id); ad->nb_queues = queue_id;set_avail_op (ad, op_type);
@@ -3846,6 +3847,7 @@ throughput_pmd_lcore_ldpc_dec(void *arg)TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); } +rte_bbdev_queue_stop (tp->dev_id, queue_id);rte_bbdev_dec_op_free_bulk (ops_enq, num_ops);double tb_len_bits =calc_ldpc_dec_TB_size (ref_op);
Signature · 1 line
2.37.1
Re: [PATCH v2 07/16] test/bbdev: test start/stop bbdev API
Quoted 13 lines · Hernan Vargas
Add a call to queue start and queue stop specifically for testing the bbdev API.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)app/test-bbdev/test_bbdev_perf.c +0
-0
app/test-bbdev/test_bbdev_perf.c add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info, · line 893
Quoted 4 lines
app/test-bbdev/test_bbdev_perf.c +1
-0@@ -890,6 +890,7 @@ add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info, ret =rte_bbdev_queue_configure (ad->dev_id, queue_id, &qconf); } +rte_bbdev_queue_start (ad->dev_id, queue_id);
You may want to check the return value, that would be useful since the goal is to test the API.
Quoted 15 lines
app/test-bbdev/test_bbdev_perf.c +2
-2if (ret !=0 ) {printf ("All queues on dev %u allocated: %u\n" , dev_id, queue_id);@@ -898,8 +899,8 @@ add_bbdev_dev(uint8_t dev_id, struct rte_bbdev_info *info, ad->queue_ids[queue_id] = queue_id; }TEST_ASSERT (queue_id !=0 , -"ERROR Failed to configure any queues on dev %u" , - dev_id); +"ERROR Failed to configure any queues on dev %u\n" +"\tthe device may not support or have been configured" , dev_id); ad->nb_queues = queue_id;set_avail_op (ad, op_type);
app/test-bbdev/test_bbdev_perf.c throughput_pmd_lcore_ldpc_dec(void *arg) · line 3850
Quoted 4 lines
app/test-bbdev/test_bbdev_perf.c +1
-0@@ -3846,6 +3847,7 @@ throughput_pmd_lcore_ldpc_dec(void *arg)TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); } +rte_bbdev_queue_stop (tp->dev_id, queue_id);
Ditto.
Quoted 3 lines
app/test-bbdev/test_bbdev_perf.c +0
-0rte_bbdev_dec_op_free_bulk (ops_enq, num_ops);double tb_len_bits =calc_ldpc_dec_TB_size (ref_op);
PATCH v2 8/16 test/bbdev: add support for BLER for 4G
In reply to Hernan VargasNew feature to add BLER support for 4G in test-bbdev.
--- app/test-bbdev/test_bbdev_perf.c | 183 ++++++++++++++++++++++++++++++- 1 file changed, 182 insertions(+), 1 deletion(-)
app/test-bbdev/test_bbdev_perf.c +182 -1
@@ -1757,6 +1757,30 @@ gen_qm2_llr(int8_t *llrs, uint32_t j, double N0, double llr_max) llrs[j] = (int8_t ) b; } +/* Simple LLR generation assuming AWGN and QPSK */ +static void +gen_turbo_llr (int8_t *llrs,uint32_t j,double N0 ,double llr_max) +{ +double b, b1, n; +double coeff =2.0 *sqrt (N0 ); + +/* Ignore in vectors null LLRs not to be saturated */ +if (llrs[j] ==0 ) +return ; + +/* Note don't change sign here */ + n =randn (j %2 ); + b1 = ((llrs[j] >0 ?2.0 :-2.0 ) + + coeff * n) /N0 ; + b = b1 * (1 <<4 ); + b =round (b); +if (b > llr_max) + b = llr_max; +if (b < -llr_max) + b = -llr_max; + llrs[j] = (int8_t ) b; +} +/* Generate LLR for a given SNR */ static void generate_llr_input(uint16_t n, struct rte_bbdev_op_data *inputs,
@@ -1792,6 +1816,27 @@ generate_llr_input(uint16_t n, struct rte_bbdev_op_data *inputs, } } +/* Generate LLR for turbo decoder for a given SNR */ +static void +generate_turbo_llr_input (uint16_t n,struct rte_bbdev_op_data *inputs, +struct rte_bbdev_dec_op *ref_op) +{ +struct rte_mbuf *m; +uint32_t i, j, range; +double N0 , llr_max; + + llr_max =127 ; + range = ref_op->turbo_dec.input.length; +N0 =1.0 /pow (10.0 ,get_snr () /10.0 ); + +for (i =0 ; i < n; ++i) { + m = inputs[i].data; + int8_t *llrs = rte_pktmbuf_mtod_offset(m, int8_t *, 0); +for (j =0 ; j < range; ++j) +gen_turbo_llr (llrs, j,N0 , llr_max); + } +} + static void copy_reference_ldpc_dec_op(struct rte_bbdev_dec_op **ops, unsigned int n, unsigned int start_idx,
@@ -2306,6 +2351,30 @@ validate_ldpc_bler(struct rte_bbdev_dec_op **ops, const uint16_t n)return errors; } +/* Check Number of code blocks errors */ +static int +validate_turbo_bler (struct rte_bbdev_dec_op **ops,const uint16_t n) +{ +unsigned int i; +struct op_data_entries *hard_data_orig = &test_vector.entries[DATA_HARD_OUTPUT ]; +struct rte_bbdev_op_turbo_dec *ops_td; +struct rte_bbdev_op_data *hard_output; +int errors =0 ; +struct rte_mbuf *m; + +for (i =0 ; i < n; ++i) { + ops_td = &ops[i]->turbo_dec; + hard_output = &ops_td->hard_output; + m = hard_output->data; + if (memcmp(rte_pktmbuf_mtod_offset(m, uint32_t *, 0), + hard_data_orig->segments[0 ].addr, + hard_data_orig->segments[0 ].length)) + errors++; + } +return errors; +} + +static int validate_ldpc_dec_op (struct rte_bbdev_dec_op **ops,const uint16_t n,struct rte_bbdev_dec_op *ref_op,const int vector_mask)
@@ -3738,6 +3807,114 @@ bler_pmd_lcore_ldpc_dec(void *arg)return TEST_SUCCESS ; } + +static int +bler_pmd_lcore_turbo_dec (void *arg) +{ +struct thread_params *tp = arg; +uint16_t enq, deq; +uint64_t total_time =0 , start_time; +const uint16_t queue_id = tp->queue_id; +const uint16_t burst_sz = tp->op_params->burst_sz; +const uint16_t num_ops = tp->op_params->num_to_process; +struct rte_bbdev_dec_op *ops_enq[num_ops]; +struct rte_bbdev_dec_op *ops_deq[num_ops]; +struct rte_bbdev_dec_op *ref_op = tp->op_params->ref_dec_op; +struct test_buffers *bufs =NULL ; +int i, j, ret; +struct rte_bbdev_info info; +uint16_t num_to_enq; + +TEST_ASSERT_SUCCESS ((burst_sz >MAX_BURST ), +"BURST_SIZE should be <= %u" ,MAX_BURST ); + +rte_bbdev_info_get (tp->dev_id, &info); + +TEST_ASSERT_SUCCESS ((num_ops > info.drv.queue_size_lim), +"NUM_OPS cannot exceed %u for this device" , + info.drv.queue_size_lim); + + bufs = &tp->op_params->q_bufs[GET_SOCKET (info.socket_id)][queue_id]; + +rte_wait_until_equal_16 (&tp->op_params->sync,SYNC_START , __ATOMIC_RELAXED); + + ret =rte_bbdev_dec_op_alloc_bulk (tp->op_params->mp, ops_enq, num_ops); +TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_ops); + +/* For BLER tests we need to enable early termination */ +if (!check_bit (ref_op->turbo_dec.op_flags, +RTE_BBDEV_TURBO_EARLY_TERMINATION )) + ref_op->turbo_dec.op_flags += +RTE_BBDEV_TURBO_EARLY_TERMINATION ; + ref_op->turbo_dec.iter_max =get_iter_max (); + ref_op->turbo_dec.iter_count = ref_op->turbo_dec.iter_max; + +if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) +copy_reference_dec_op (ops_enq, num_ops,0 , bufs->inputs, + bufs->hard_outputs, bufs->soft_outputs, + ref_op); +generate_turbo_llr_input (num_ops, bufs->inputs, ref_op); + +/* Set counter to validate the ordering */ +for (j =0 ; j < num_ops; ++j) + ops_enq[j]->opaque_data = (void *)(uintptr_t )j; + +for (i =0 ; i <1 ; ++i) {/* Could add more iterations */ +for (j =0 ; j < num_ops; ++j) { +mbuf_reset ( + ops_enq[j]->turbo_dec.hard_output.data); + } + + start_time =rte_rdtsc_precise (); + +for (enq =0 , deq =0 ; enq < num_ops;) { + num_to_enq = burst_sz; + +if (unlikely (num_ops - enq < num_to_enq)) + num_to_enq = num_ops - enq; + + enq +=rte_bbdev_enqueue_dec_ops (tp->dev_id, + queue_id, &ops_enq[enq], num_to_enq); + + deq +=rte_bbdev_dequeue_dec_ops (tp->dev_id, + queue_id, &ops_deq[deq], enq - deq); + } + +/* dequeue the remaining */ +while (deq < enq) { + deq +=rte_bbdev_dequeue_dec_ops (tp->dev_id, + queue_id, &ops_deq[deq], enq - deq); + } + + total_time +=rte_rdtsc_precise () - start_time; + } + + tp->iter_count =0 ; + tp->iter_average =0 ; +/* get the max of iter_count for all dequeued ops */ +for (i =0 ; i < num_ops; ++i) { + tp->iter_count =RTE_MAX (ops_enq[i]->turbo_dec.iter_count, + tp->iter_count); + tp->iter_average += (double ) ops_enq[i]->turbo_dec.iter_count; + } + + tp->iter_average /= num_ops; + tp->bler = (double )validate_turbo_bler (ops_deq, num_ops) / num_ops; + +rte_bbdev_dec_op_free_bulk (ops_enq, num_ops); + +double tb_len_bits =calc_dec_TB_size (ref_op); + tp->ops_per_sec = ((double )num_ops *1 ) / + ((double )total_time / (double )rte_get_tsc_hz ()); + tp->mbps = (((double )(num_ops *1 * tb_len_bits)) / +1000000.0 ) / ((double )total_time / + (double )rte_get_tsc_hz ()); +printf ("TBS %.0f Time %.0f\n" , tb_len_bits,1000000.0 * + ((double )total_time / (double )rte_get_tsc_hz ())); + +return TEST_SUCCESS ; +} +static int throughput_pmd_lcore_ldpc_dec (void *arg) {
@@ -4195,7 +4372,7 @@ print_dec_bler(struct thread_params *t_params, unsigned int used_cores) total_bler /= used_cores; total_iter /= used_cores; -printf ("SNR %.2f BLER %.1f %% - Iterations %.1f %d - Tp %.1f Mbps %s\n" , +printf ("SNR %.2f BLER %.1f %% - Iterations %.1f %d - Tp %.3f Mbps %s\n" , snr, total_bler *100 , total_iter,get_iter_max (), total_mbps,get_vector_filename ()); }
@@ -4247,6 +4424,10 @@ bler_test(struct active_device *ad, && !check_bit(test_vector.ldpc_dec.op_flags, RTE_BBDEV_LDPC_LLR_COMPRESSION)) bler_function = bler_pmd_lcore_ldpc_dec; + else if ((test_vector.op_type == RTE_BBDEV_OP_TURBO_DEC) && + !check_bit(test_vector.turbo_dec.op_flags, + RTE_BBDEV_TURBO_SOFT_OUTPUT)) + bler_function = bler_pmd_lcore_turbo_dec; else return TEST_SKIPPED;
Signature · 1 line
2.37.1
Re: [PATCH v2 08/16] test/bbdev: add support for BLER for 4G
Quoted 166 lines · Hernan Vargas
New feature to add BLER support for 4G in test-bbdev.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 183 ++++++++++++++++++++++++++++++- 1 file changed, 182 insertions(+), 1 deletion(-)app/test-bbdev/test_bbdev_perf.c +130
-0@@ -1757,6 +1757,30 @@ gen_qm2_llr(int8_t *llrs, uint32_t j, double N0, double llr_max) llrs[j] = (int8_t ) b; } +/* Simple LLR generation assuming AWGN and QPSK */ +static void +gen_turbo_llr (int8_t *llrs,uint32_t j,double N0 ,double llr_max) +{ +double b, b1, n; +double coeff =2.0 *sqrt (N0 ); + +/* Ignore in vectors null LLRs not to be saturated */ +if (llrs[j] ==0 ) +return ; + +/* Note don't change sign here */ + n =randn (j %2 ); + b1 = ((llrs[j] >0 ?2.0 :-2.0 ) + + coeff * n) /N0 ; + b = b1 * (1 <<4 ); + b =round (b); +if (b > llr_max) + b = llr_max; +if (b < -llr_max) + b = -llr_max; + llrs[j] = (int8_t ) b; +} +/* Generate LLR for a given SNR */ static void generate_llr_input(uint16_t n, struct rte_bbdev_op_data *inputs,@@ -1792,6 +1816,27 @@ generate_llr_input(uint16_t n, struct rte_bbdev_op_data *inputs, } } +/* Generate LLR for turbo decoder for a given SNR */ +static void +generate_turbo_llr_input (uint16_t n,struct rte_bbdev_op_data *inputs, +struct rte_bbdev_dec_op *ref_op) +{ +struct rte_mbuf *m; +uint32_t i, j, range; +double N0 , llr_max; + + llr_max =127 ; + range = ref_op->turbo_dec.input.length; +N0 =1.0 /pow (10.0 ,get_snr () /10.0 ); + +for (i =0 ; i < n; ++i) { + m = inputs[i].data; + int8_t *llrs = rte_pktmbuf_mtod_offset(m, int8_t *, 0); +for (j =0 ; j < range; ++j) +gen_turbo_llr (llrs, j,N0 , llr_max); + } +} + static void copy_reference_ldpc_dec_op(struct rte_bbdev_dec_op **ops, unsigned int n, unsigned int start_idx,@@ -2306,6 +2351,30 @@ validate_ldpc_bler(struct rte_bbdev_dec_op **ops, const uint16_t n)return errors; } +/* Check Number of code blocks errors */ +static int +validate_turbo_bler (struct rte_bbdev_dec_op **ops,const uint16_t n) +{ +unsigned int i; +struct op_data_entries *hard_data_orig = &test_vector.entries[DATA_HARD_OUTPUT ]; +struct rte_bbdev_op_turbo_dec *ops_td; +struct rte_bbdev_op_data *hard_output; +int errors =0 ; +struct rte_mbuf *m; + +for (i =0 ; i < n; ++i) { + ops_td = &ops[i]->turbo_dec; + hard_output = &ops_td->hard_output; + m = hard_output->data; + if (memcmp(rte_pktmbuf_mtod_offset(m, uint32_t *, 0), + hard_data_orig->segments[0 ].addr, + hard_data_orig->segments[0 ].length)) + errors++; + } +return errors; +} + +static int validate_ldpc_dec_op (struct rte_bbdev_dec_op **ops,const uint16_t n,struct rte_bbdev_dec_op *ref_op,const int vector_mask)@@ -3738,6 +3807,114 @@ bler_pmd_lcore_ldpc_dec(void *arg)return TEST_SUCCESS ; } + +static int +bler_pmd_lcore_turbo_dec (void *arg) +{ +struct thread_params *tp = arg; +uint16_t enq, deq; +uint64_t total_time =0 , start_time; +const uint16_t queue_id = tp->queue_id; +const uint16_t burst_sz = tp->op_params->burst_sz; +const uint16_t num_ops = tp->op_params->num_to_process; +struct rte_bbdev_dec_op *ops_enq[num_ops]; +struct rte_bbdev_dec_op *ops_deq[num_ops]; +struct rte_bbdev_dec_op *ref_op = tp->op_params->ref_dec_op; +struct test_buffers *bufs =NULL ; +int i, j, ret; +struct rte_bbdev_info info; +uint16_t num_to_enq; + +TEST_ASSERT_SUCCESS ((burst_sz >MAX_BURST ), +"BURST_SIZE should be <= %u" ,MAX_BURST ); + +rte_bbdev_info_get (tp->dev_id, &info); + +TEST_ASSERT_SUCCESS ((num_ops > info.drv.queue_size_lim), +"NUM_OPS cannot exceed %u for this device" , + info.drv.queue_size_lim); + + bufs = &tp->op_params->q_bufs[GET_SOCKET (info.socket_id)][queue_id]; + +rte_wait_until_equal_16 (&tp->op_params->sync,SYNC_START , __ATOMIC_RELAXED); + + ret =rte_bbdev_dec_op_alloc_bulk (tp->op_params->mp, ops_enq, num_ops); +TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_ops); + +/* For BLER tests we need to enable early termination */ +if (!check_bit (ref_op->turbo_dec.op_flags, +RTE_BBDEV_TURBO_EARLY_TERMINATION )) + ref_op->turbo_dec.op_flags += +RTE_BBDEV_TURBO_EARLY_TERMINATION ; + ref_op->turbo_dec.iter_max =get_iter_max (); + ref_op->turbo_dec.iter_count = ref_op->turbo_dec.iter_max; + +if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) +copy_reference_dec_op (ops_enq, num_ops,0 , bufs->inputs, + bufs->hard_outputs, bufs->soft_outputs, + ref_op); +generate_turbo_llr_input (num_ops, bufs->inputs, ref_op); + +/* Set counter to validate the ordering */ +for (j =0 ; j < num_ops; ++j) + ops_enq[j]->opaque_data = (void *)(uintptr_t )j; + +for (i =0 ; i <1 ; ++i) {/* Could add more iterations */ +for (j =0 ; j < num_ops; ++j) { +mbuf_reset ( + ops_enq[j]->turbo_dec.hard_output.data); + } + + start_time =rte_rdtsc_precise (); + +for (enq =0 , deq =0 ; enq < num_ops;) {
app/test-bbdev/test_bbdev_perf.c bler_pmd_lcore_ldpc_dec(void *arg) · line 3880
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +10
-0+ num_to_enq = burst_sz; + +if (unlikely (num_ops - enq < num_to_enq)) + num_to_enq = num_ops - enq; + + enq +=rte_bbdev_enqueue_dec_ops (tp->dev_id, + queue_id, &ops_enq[enq], num_to_enq); + + deq +=rte_bbdev_dequeue_dec_ops (tp->dev_id, + queue_id, &ops_deq[deq], enq - deq);
Please check my reply on v1.
Quoted 60 lines
app/test-bbdev/test_bbdev_perf.c +42
-1+ } + +/* dequeue the remaining */ +while (deq < enq) { + deq +=rte_bbdev_dequeue_dec_ops (tp->dev_id, + queue_id, &ops_deq[deq], enq - deq); + } + + total_time +=rte_rdtsc_precise () - start_time; + } + + tp->iter_count =0 ; + tp->iter_average =0 ; +/* get the max of iter_count for all dequeued ops */ +for (i =0 ; i < num_ops; ++i) { + tp->iter_count =RTE_MAX (ops_enq[i]->turbo_dec.iter_count, + tp->iter_count); + tp->iter_average += (double ) ops_enq[i]->turbo_dec.iter_count; + } + + tp->iter_average /= num_ops; + tp->bler = (double )validate_turbo_bler (ops_deq, num_ops) / num_ops; + +rte_bbdev_dec_op_free_bulk (ops_enq, num_ops); + +double tb_len_bits =calc_dec_TB_size (ref_op); + tp->ops_per_sec = ((double )num_ops *1 ) / + ((double )total_time / (double )rte_get_tsc_hz ()); + tp->mbps = (((double )(num_ops *1 * tb_len_bits)) / +1000000.0 ) / ((double )total_time / + (double )rte_get_tsc_hz ()); +printf ("TBS %.0f Time %.0f\n" , tb_len_bits,1000000.0 * + ((double )total_time / (double )rte_get_tsc_hz ())); + +return TEST_SUCCESS ; +} +static int throughput_pmd_lcore_ldpc_dec (void *arg) {@@ -4195,7 +4372,7 @@ print_dec_bler(struct thread_params *t_params, unsigned int used_cores) total_bler /= used_cores; total_iter /= used_cores; -printf ("SNR %.2f BLER %.1f %% - Iterations %.1f %d - Tp %.1f Mbps %s\n" , +printf ("SNR %.2f BLER %.1f %% - Iterations %.1f %d - Tp %.3f Mbps %s\n" , snr, total_bler *100 , total_iter,get_iter_max (), total_mbps,get_vector_filename ()); }@@ -4247,6 +4424,10 @@ bler_test(struct active_device *ad, && !check_bit(test_vector.ldpc_dec.op_flags, RTE_BBDEV_LDPC_LLR_COMPRESSION)) bler_function = bler_pmd_lcore_ldpc_dec; + else if ((test_vector.op_type == RTE_BBDEV_OP_TURBO_DEC) && + !check_bit(test_vector.turbo_dec.op_flags, + RTE_BBDEV_TURBO_SOFT_OUTPUT)) + bler_function = bler_pmd_lcore_turbo_dec; else return TEST_SKIPPED;
PATCH v2 9/16 test/bbdev: extend support for large TB
In reply to Hernan VargasAdd support for large TB when it cannot fit into a true mbuf.
--- app/test-bbdev/test_bbdev_perf.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-)
app/test-bbdev/test_bbdev_perf.c +12 -9
@@ -1073,8 +1073,6 @@ init_op_data_objs(struct rte_bbdev_op_data *bufs,* Special case when DPDK mbuf cannot handle * the required input size */ -printf ("Warning: Larger input size than DPDK mbuf %d\n" , - seg->length); large_input = true; } bufs[i].data = m_head;
@@ -2031,6 +2029,7 @@ validate_op_chain(struct rte_bbdev_op_data *op,struct rte_mbuf *m = op->data;uint8_t nb_dst_segments = orig_op->nb_segments;uint32_t total_data_size =0 ; +bool ignore_mbuf = false;/* ignore mbuf limitations */ TEST_ASSERT(nb_dst_segments == m->nb_segs, "Number of segments differ in original (%u) and filled (%u) op",
@@ -2043,21 +2042,25 @@ 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; -TEST_ASSERT (orig_op->segments[i].length == data_len, -"Length of segment differ in original (%u) and filled (%u) op" , - orig_op->segments[i].length, data_len); +if (orig_op->segments[i].length >RTE_BBDEV_LDPC_E_MAX_MBUF ) + ignore_mbuf = true; +if (!ignore_mbuf) +TEST_ASSERT (orig_op->segments[i].length == data_len, +"Length of segment differ in original (%u) and filled (%u) op" , + orig_op->segments[i].length, data_len);TEST_ASSERT_BUFFERS_ARE_EQUAL (orig_op->segments[i].addr, rte_pktmbuf_mtod_offset(m, uint32_t *, offset), - data_len, + orig_op->segments[i].length,"Output buffers (CB=%u) are not equal" , i); m = m->next; }/* Validate total mbuf pkt length */ uint32_t pkt_len =rte_pktmbuf_pkt_len (op->data) - op->offset; -TEST_ASSERT (total_data_size == pkt_len, -"Length of data differ in original (%u) and filled (%u) op" , - total_data_size, pkt_len); +if (!ignore_mbuf) +TEST_ASSERT (total_data_size == pkt_len, +"Length of data differ in original (%u) and filled (%u) op" , + total_data_size, pkt_len);return TEST_SUCCESS ; }
Signature · 1 line
2.37.1
Re: [PATCH v2 09/16] test/bbdev: extend support for large TB
Quoted 51 lines · Hernan Vargas
Add support for large TB when it cannot fit into a true mbuf.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-)app/test-bbdev/test_bbdev_perf.c +8
-6@@ -1073,8 +1073,6 @@ init_op_data_objs(struct rte_bbdev_op_data *bufs,* Special case when DPDK mbuf cannot handle * the required input size */ -printf ("Warning: Larger input size than DPDK mbuf %d\n" , - seg->length); large_input = true; } bufs[i].data = m_head;@@ -2031,6 +2029,7 @@ validate_op_chain(struct rte_bbdev_op_data *op,struct rte_mbuf *m = op->data;uint8_t nb_dst_segments = orig_op->nb_segments;uint32_t total_data_size =0 ; +bool ignore_mbuf = false;/* ignore mbuf limitations */ TEST_ASSERT(nb_dst_segments == m->nb_segs, "Number of segments differ in original (%u) and filled (%u) op",@@ -2043,21 +2042,25 @@ 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; -TEST_ASSERT (orig_op->segments[i].length == data_len, -"Length of segment differ in original (%u) and filled (%u) op" , - orig_op->segments[i].length, data_len); +if (orig_op->segments[i].length >RTE_BBDEV_LDPC_E_MAX_MBUF ) + ignore_mbuf = true; +if (!ignore_mbuf) +TEST_ASSERT (orig_op->segments[i].length == data_len, +"Length of segment differ in original (%u) and filled (%u) op" , + orig_op->segments[i].length, data_len);TEST_ASSERT_BUFFERS_ARE_EQUAL (orig_op->segments[i].addr, rte_pktmbuf_mtod_offset(m, uint32_t *, offset), - data_len, + orig_op->segments[i].length,"Output buffers (CB=%u) are not equal" , i); m = m->next; }/* Validate total mbuf pkt length */ uint32_t pkt_len =rte_pktmbuf_pkt_len (op->data) - op->offset;
app/test-bbdev/test_bbdev_perf.c validate_op_chain(struct rte_bbdev_op_data *op, · line 2066
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-3-TEST_ASSERT (total_data_size == pkt_len, -"Length of data differ in original (%u) and filled (%u) op" , - total_data_size, pkt_len); +if (!ignore_mbuf) +TEST_ASSERT (total_data_size == pkt_len, +"Length of data differ in original (%u) and filled (%u) op" , + total_data_size, pkt_len);return TEST_SUCCESS ; }
Thanks,
Maxime
PATCH v2 10/16 test/bbdev: adjustment for soft output
In reply to Hernan VargasExtend test-bbdev for soft output check, notably due to the logic in test-bbdev to enable termination changing.
--- app/test-bbdev/test_bbdev_perf.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-)
app/test-bbdev/test_bbdev_perf.c +13 -13
@@ -1349,6 +1349,7 @@ fill_queue_buffers(struct test_op_params *op_params,RTE_BBDEV_LDPC_LLR_COMPRESSION ;bool harq_comp = op_params->ref_dec_op->ldpc_dec.op_flags &RTE_BBDEV_LDPC_HARQ_6BIT_COMPRESSION ; + ldpc_llr_decimals = capabilities->cap.ldpc_dec.llr_decimals; ldpc_llr_size = capabilities->cap.ldpc_dec.llr_size; ldpc_cap_flags = capabilities->cap.ldpc_dec.capability_flags;
@@ -2424,7 +2425,7 @@ validate_ldpc_dec_op(struct rte_bbdev_dec_op **ops, const uint16_t n, i);if (ref_op->ldpc_dec.op_flags &RTE_BBDEV_LDPC_SOFT_OUT_ENABLE ) -TEST_ASSERT_SUCCESS (validate_op_chain (soft_output, +TEST_ASSERT_SUCCESS (validate_op_so_chain (soft_output, soft_data_orig),"Soft output buffers (CB=%u) are not equal" , i);
@@ -2494,7 +2495,6 @@ validate_ldpc_enc_op(struct rte_bbdev_enc_op **ops, const uint16_t n,return TEST_SUCCESS ; } -static inline int validate_op_fft_chain (struct rte_bbdev_op_data *op,struct op_data_entries *orig_op) {
@@ -3181,11 +3181,11 @@ throughput_intr_lcore_ldpc_dec(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) {if (!loopback) -rte_pktmbuf_reset ( - ops[i]->ldpc_dec.hard_output.data); +rte_pktmbuf_reset (ops[i]->ldpc_dec.hard_output.data);if (hc_out || loopback) -mbuf_reset ( - ops[i]->ldpc_dec.harq_combined_output.data); +mbuf_reset (ops[i]->ldpc_dec.harq_combined_output.data); +if (ops[i]->ldpc_dec.soft_output.data !=NULL ) +rte_pktmbuf_reset (ops[i]->ldpc_dec.soft_output.data); } tp->start_time =rte_rdtsc_precise ();
@@ -3280,7 +3280,6 @@ throughput_intr_lcore_dec(void *arg)rte_pktmbuf_reset (ops[i]->turbo_dec.soft_output.data); } - tp->start_time =rte_rdtsc_precise ();for (enqueued =0 ; enqueued < num_to_process;) { num_to_enq = burst_sz;
@@ -3742,10 +3741,11 @@ bler_pmd_lcore_ldpc_dec(void *arg)for (i =0 ; i <1 ; ++i) {/* Could add more iterations */ for (j =0 ; j < num_ops; ++j) {if (!loopback) -mbuf_reset ( - ops_enq[j]->ldpc_dec.hard_output.data); +mbuf_reset (ops_enq[j]->ldpc_dec.hard_output.data);if (hc_out || loopback)mbuf_reset (ops_enq[j]->ldpc_dec.harq_combined_output.data); +if (ops_enq[j]->ldpc_dec.soft_output.data !=NULL ) +mbuf_reset (ops_enq[j]->ldpc_dec.soft_output.data); }if (extDdr) preload_harq_ddr(tp->dev_id, queue_id, ops_enq,
@@ -3977,11 +3977,11 @@ throughput_pmd_lcore_ldpc_dec(void *arg)for (i =0 ; i <TEST_REPETITIONS ; ++i) {for (j =0 ; j < num_ops; ++j) {if (!loopback) -mbuf_reset ( - ops_enq[j]->ldpc_dec.hard_output.data); +mbuf_reset (ops_enq[j]->ldpc_dec.hard_output.data);if (hc_out || loopback) -mbuf_reset ( - ops_enq[j]->ldpc_dec.harq_combined_output.data); +mbuf_reset (ops_enq[j]->ldpc_dec.harq_combined_output.data); +if (ops_enq[j]->ldpc_dec.soft_output.data !=NULL ) +mbuf_reset (ops_enq[j]->ldpc_dec.soft_output.data); }if (extDdr) preload_harq_ddr(tp->dev_id, queue_id, ops_enq,
Signature · 1 line
2.37.1
Re: [PATCH v2 10/16] test/bbdev: adjustment for soft output
Quoted 81 lines · Hernan Vargas
Extend test-bbdev for soft output check, notably due to the logic in test-bbdev to enable termination changing.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-)app/test-bbdev/test_bbdev_perf.c +9
-11@@ -1349,6 +1349,7 @@ fill_queue_buffers(struct test_op_params *op_params,RTE_BBDEV_LDPC_LLR_COMPRESSION ;bool harq_comp = op_params->ref_dec_op->ldpc_dec.op_flags &RTE_BBDEV_LDPC_HARQ_6BIT_COMPRESSION ; + ldpc_llr_decimals = capabilities->cap.ldpc_dec.llr_decimals; ldpc_llr_size = capabilities->cap.ldpc_dec.llr_size; ldpc_cap_flags = capabilities->cap.ldpc_dec.capability_flags;@@ -2424,7 +2425,7 @@ validate_ldpc_dec_op(struct rte_bbdev_dec_op **ops, const uint16_t n, i);if (ref_op->ldpc_dec.op_flags &RTE_BBDEV_LDPC_SOFT_OUT_ENABLE ) -TEST_ASSERT_SUCCESS (validate_op_chain (soft_output, +TEST_ASSERT_SUCCESS (validate_op_so_chain (soft_output, soft_data_orig),"Soft output buffers (CB=%u) are not equal" , i);@@ -2494,7 +2495,6 @@ validate_ldpc_enc_op(struct rte_bbdev_enc_op **ops, const uint16_t n,return TEST_SUCCESS ; } -static inline int validate_op_fft_chain (struct rte_bbdev_op_data *op,struct op_data_entries *orig_op) {@@ -3181,11 +3181,11 @@ throughput_intr_lcore_ldpc_dec(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) {if (!loopback) -rte_pktmbuf_reset ( - ops[i]->ldpc_dec.hard_output.data); +rte_pktmbuf_reset (ops[i]->ldpc_dec.hard_output.data);if (hc_out || loopback) -mbuf_reset ( - ops[i]->ldpc_dec.harq_combined_output.data); +mbuf_reset (ops[i]->ldpc_dec.harq_combined_output.data); +if (ops[i]->ldpc_dec.soft_output.data !=NULL ) +rte_pktmbuf_reset (ops[i]->ldpc_dec.soft_output.data); } tp->start_time =rte_rdtsc_precise ();@@ -3280,7 +3280,6 @@ throughput_intr_lcore_dec(void *arg)rte_pktmbuf_reset (ops[i]->turbo_dec.soft_output.data); } - tp->start_time =rte_rdtsc_precise ();for (enqueued =0 ; enqueued < num_to_process;) { num_to_enq = burst_sz;@@ -3742,10 +3741,11 @@ bler_pmd_lcore_ldpc_dec(void *arg)for (i =0 ; i <1 ; ++i) {/* Could add more iterations */ for (j =0 ; j < num_ops; ++j) {if (!loopback) -mbuf_reset ( - ops_enq[j]->ldpc_dec.hard_output.data); +mbuf_reset (ops_enq[j]->ldpc_dec.hard_output.data);if (hc_out || loopback)mbuf_reset (ops_enq[j]->ldpc_dec.harq_combined_output.data); +if (ops_enq[j]->ldpc_dec.soft_output.data !=NULL ) +mbuf_reset (ops_enq[j]->ldpc_dec.soft_output.data); }if (extDdr) preload_harq_ddr(tp->dev_id, queue_id, ops_enq,@@ -3977,11 +3977,11 @@ throughput_pmd_lcore_ldpc_dec(void *arg)for (i =0 ; i <TEST_REPETITIONS ; ++i) {for (j =0 ; j < num_ops; ++j) {if (!loopback) -mbuf_reset ( - ops_enq[j]->ldpc_dec.hard_output.data);
app/test-bbdev/test_bbdev_perf.c throughput_pmd_lcore_ldpc_dec(void *arg) · line 3987
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +4
-2+mbuf_reset (ops_enq[j]->ldpc_dec.hard_output.data);if (hc_out || loopback) -mbuf_reset ( - ops_enq[j]->ldpc_dec.harq_combined_output.data); +mbuf_reset (ops_enq[j]->ldpc_dec.harq_combined_output.data); +if (ops_enq[j]->ldpc_dec.soft_output.data !=NULL ) +mbuf_reset (ops_enq[j]->ldpc_dec.soft_output.data); }if (extDdr) preload_harq_ddr(tp->dev_id, queue_id, ops_enq,
Thanks,
Maxime
PATCH v2 11/16 test/bbdev: expose warning counters
In reply to Hernan VargasPrint warnings reported on queues for offload test. No functional impact.
--- app/test-bbdev/test_bbdev_perf.c | 3 +++ 1 file changed, 3 insertions(+)
app/test-bbdev/test_bbdev_perf.c +3 -0
@@ -5673,6 +5673,9 @@ offload_cost_test(struct active_device *ad,struct rte_bbdev_stats stats = {0 };get_bbdev_queue_stats (ad->dev_id, queue_id, &stats); +if (stats.enqueue_warn_count >0 ) +printf ("Warning reported on the queue : %10" PRIu64"\n" , + stats.enqueue_warn_count);if (op_type !=RTE_BBDEV_OP_LDPC_DEC ) { TEST_ASSERT_SUCCESS(stats.enqueued_count != num_to_process, "Mismatch in enqueue count %10"PRIu64" %d",
Signature · 1 line
2.37.1
Re: [PATCH v2 11/16] test/bbdev: expose warning counters
Quoted 13 lines · Hernan Vargas
Print warnings reported on queues for offload test. No functional impact.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 3 +++ 1 file changed, 3 insertions(+)app/test-bbdev/test_bbdev_perf.c +0
-0
app/test-bbdev/test_bbdev_perf.c offload_cost_test(struct active_device *ad, · line 5681
Quoted 9 lines
app/test-bbdev/test_bbdev_perf.c +3
-0@@ -5673,6 +5673,9 @@ offload_cost_test(struct active_device *ad,struct rte_bbdev_stats stats = {0 };get_bbdev_queue_stats (ad->dev_id, queue_id, &stats); +if (stats.enqueue_warn_count >0 ) +printf ("Warning reported on the queue : %10" PRIu64"\n" , + stats.enqueue_warn_count);if (op_type !=RTE_BBDEV_OP_LDPC_DEC ) { TEST_ASSERT_SUCCESS(stats.enqueued_count != num_to_process, "Mismatch in enqueue count %10"PRIu64" %d",
Reporting from v1:
Thanks,
Maxime
PATCH v2 12/16 test/bbdev: remove check for invalid opaque data
In reply to Hernan VargasAssert also if the opaque data is invalid.
--- app/test-bbdev/test_bbdev_perf.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-)
app/test-bbdev/test_bbdev_perf.c +3 -5
@@ -79,7 +79,6 @@#define SYNC_WAIT 0#define SYNC_START 1 -#define INVALID_OPAQUE -1#define INVALID_QUEUE_ID -1/* Increment for next code block in external HARQ memory */
@@ -1999,10 +1998,9 @@ check_enc_status_and_ordering(struct rte_bbdev_enc_op *op,"op_status (%d) != expected_status (%d)" , op->status, expected_status); -if (op->opaque_data != (void *)(uintptr_t )INVALID_OPAQUE ) -TEST_ASSERT ((void *)(uintptr_t )order_idx == op->opaque_data, -"Ordering error, expected %p, got %p" , - (void *)(uintptr_t )order_idx, op->opaque_data); +TEST_ASSERT ((void *)(uintptr_t )order_idx == op->opaque_data, +"Ordering error, expected %p, got %p" , + (void *)(uintptr_t )order_idx, op->opaque_data);return TEST_SUCCESS ; }
Signature · 1 line
2.37.1
Re: [PATCH v2 12/16] test/bbdev: remove check for invalid opaque data
Quoted 1 line · Hernan Vargas
Assert also if the opaque data is invalid.
New line needed before Fixes tag.
Quoted 24 lines
Fixes 335c11fd276 ("app/bbdev: support HARQ validation")Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-)app/test-bbdev/test_bbdev_perf.c +0
-1@@ -79,7 +79,6 @@#define SYNC_WAIT 0#define SYNC_START 1 -#define INVALID_OPAQUE -1#define INVALID_QUEUE_ID -1/* Increment for next code block in external HARQ memory */ @@ -1999,10 +1998,9 @@ check_enc_status_and_ordering(struct rte_bbdev_enc_op *op,"op_status (%d) != expected_status (%d)" , op->status, expected_status);
app/test-bbdev/test_bbdev_perf.c check_enc_status_and_ordering(struct rte_bbdev_enc_op *op, · line 2006
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +3
-4-if (op->opaque_data != (void *)(uintptr_t )INVALID_OPAQUE ) -TEST_ASSERT ((void *)(uintptr_t )order_idx == op->opaque_data, -"Ordering error, expected %p, got %p" , - (void *)(uintptr_t )order_idx, op->opaque_data); +TEST_ASSERT ((void *)(uintptr_t )order_idx == op->opaque_data, +"Ordering error, expected %p, got %p" , + (void *)(uintptr_t )order_idx, op->opaque_data);return TEST_SUCCESS ; }
With typo fixed:
Thanks,
Maxime
PATCH v2 13/16 test/bbdev: remove iteration count check
In reply to Hernan VargasTo make the test compatible with devices that do not support early termination, the iteration count assert can be removed.
--- app/test-bbdev/test_bbdev_perf.c | 24 ++++++++---------------- 1 file changed, 8 insertions(+), 16 deletions(-)
app/test-bbdev/test_bbdev_perf.c +8 -16
@@ -2288,7 +2288,7 @@ validate_op_so_chain(struct rte_bbdev_op_data *op,static int validate_dec_op (struct rte_bbdev_dec_op **ops,const uint16_t n, -struct rte_bbdev_dec_op *ref_op,const int vector_mask) +struct rte_bbdev_dec_op *ref_op) {unsigned int i;int ret;
@@ -2299,17 +2299,12 @@ validate_dec_op(struct rte_bbdev_dec_op **ops, const uint16_t n,struct rte_bbdev_op_turbo_dec *ops_td;struct rte_bbdev_op_data *hard_output;struct rte_bbdev_op_data *soft_output; -struct rte_bbdev_op_turbo_dec *ref_td = &ref_op->turbo_dec;for (i =0 ; i < n; ++i) { ops_td = &ops[i]->turbo_dec; hard_output = &ops_td->hard_output; soft_output = &ops_td->soft_output; -if (vector_mask &TEST_BBDEV_VF_EXPECTED_ITER_COUNT ) -TEST_ASSERT (ops_td->iter_count <= ref_td->iter_count, -"Returned iter_count (%d) > expected iter_count (%d)" , - ops_td->iter_count, ref_td->iter_count); ret =check_dec_status_and_ordering (ops[i], i, ref_op->status);TEST_ASSERT_SUCCESS (ret,"Checking status and ordering for decoder failed" );
@@ -3058,8 +3053,7 @@ dequeue_event_callback(uint16_t dev_id,if (test_vector.op_type ==RTE_BBDEV_OP_TURBO_DEC ) {struct rte_bbdev_dec_op *ref_op = tp->op_params->ref_dec_op; - ret =validate_dec_op (tp->dec_ops, num_ops, ref_op, - tp->op_params->vector_mask); + ret =validate_dec_op (tp->dec_ops, num_ops, ref_op);/* get the max of iter_count for all dequeued ops */ for (i =0 ; i < num_ops; ++i) tp->iter_count = RTE_MAX(
@@ -3660,8 +3654,7 @@ throughput_pmd_lcore_dec(void *arg) }if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { - ret =validate_dec_op (ops_deq, num_ops, ref_op, - tp->op_params->vector_mask); + ret =validate_dec_op (ops_deq, num_ops, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }
@@ -4649,7 +4642,7 @@ throughput_test(struct active_device *ad,static int latency_test_dec (struct rte_mempool *mempool,struct test_buffers *bufs,struct rte_bbdev_dec_op *ref_op, -int vector_mask,uint16_t dev_id,uint16_t queue_id, +uint16_t dev_id,uint16_t queue_id,const uint16_t num_to_process,uint16_t burst_sz,uint64_t *total_time,uint64_t *min_time,uint64_t *max_time,bool disable_et) {
@@ -4709,8 +4702,7 @@ latency_test_dec(struct rte_mempool *mempool, *total_time += last_time;if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { - ret =validate_dec_op (ops_deq, burst_sz, ref_op, - vector_mask); + ret =validate_dec_op (ops_deq, burst_sz, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }
@@ -5065,9 +5057,9 @@ validation_latency_test(struct active_device *ad,if (op_type ==RTE_BBDEV_OP_TURBO_DEC ) iter =latency_test_dec (op_params->mp, bufs, - op_params->ref_dec_op, op_params->vector_mask, - ad->dev_id, queue_id, num_to_process, - burst_sz, &total_time, &min_time, &max_time, latency_flag); + op_params->ref_dec_op, ad->dev_id, queue_id, + num_to_process, burst_sz, &total_time, + &min_time, &max_time, latency_flag);else if (op_type ==RTE_BBDEV_OP_LDPC_ENC ) iter = latency_test_ldpc_enc(op_params->mp, bufs, op_params->ref_enc_op, ad->dev_id, queue_id,
Signature · 1 line
2.37.1
Re: [PATCH v2 13/16] test/bbdev: remove iteration count check
Quoted 81 lines · Hernan Vargas
To make the test compatible with devices that do not support early termination, the iteration count assert can be removed.
Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 24 ++++++++---------------- 1 file changed, 8 insertions(+), 16 deletions(-)app/test-bbdev/test_bbdev_perf.c +5
-13@@ -2288,7 +2288,7 @@ validate_op_so_chain(struct rte_bbdev_op_data *op,static int validate_dec_op (struct rte_bbdev_dec_op **ops,const uint16_t n, -struct rte_bbdev_dec_op *ref_op,const int vector_mask) +struct rte_bbdev_dec_op *ref_op) {unsigned int i;int ret;@@ -2299,17 +2299,12 @@ validate_dec_op(struct rte_bbdev_dec_op **ops, const uint16_t n,struct rte_bbdev_op_turbo_dec *ops_td;struct rte_bbdev_op_data *hard_output;struct rte_bbdev_op_data *soft_output; -struct rte_bbdev_op_turbo_dec *ref_td = &ref_op->turbo_dec;for (i =0 ; i < n; ++i) { ops_td = &ops[i]->turbo_dec; hard_output = &ops_td->hard_output; soft_output = &ops_td->soft_output; -if (vector_mask &TEST_BBDEV_VF_EXPECTED_ITER_COUNT ) -TEST_ASSERT (ops_td->iter_count <= ref_td->iter_count, -"Returned iter_count (%d) > expected iter_count (%d)" , - ops_td->iter_count, ref_td->iter_count); ret =check_dec_status_and_ordering (ops[i], i, ref_op->status);TEST_ASSERT_SUCCESS (ret,"Checking status and ordering for decoder failed" );@@ -3058,8 +3053,7 @@ dequeue_event_callback(uint16_t dev_id,if (test_vector.op_type ==RTE_BBDEV_OP_TURBO_DEC ) {struct rte_bbdev_dec_op *ref_op = tp->op_params->ref_dec_op; - ret =validate_dec_op (tp->dec_ops, num_ops, ref_op, - tp->op_params->vector_mask); + ret =validate_dec_op (tp->dec_ops, num_ops, ref_op);/* get the max of iter_count for all dequeued ops */ for (i =0 ; i < num_ops; ++i) tp->iter_count = RTE_MAX(@@ -3660,8 +3654,7 @@ throughput_pmd_lcore_dec(void *arg) }if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { - ret =validate_dec_op (ops_deq, num_ops, ref_op, - tp->op_params->vector_mask); + ret =validate_dec_op (ops_deq, num_ops, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }@@ -4649,7 +4642,7 @@ throughput_test(struct active_device *ad,static int latency_test_dec (struct rte_mempool *mempool,struct test_buffers *bufs,struct rte_bbdev_dec_op *ref_op, -int vector_mask,uint16_t dev_id,uint16_t queue_id, +uint16_t dev_id,uint16_t queue_id,const uint16_t num_to_process,uint16_t burst_sz,uint64_t *total_time,uint64_t *min_time,uint64_t *max_time,bool disable_et) {@@ -4709,8 +4702,7 @@ latency_test_dec(struct rte_mempool *mempool, *total_time += last_time;if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) { - ret =validate_dec_op (ops_deq, burst_sz, ref_op, - vector_mask); + ret =validate_dec_op (ops_deq, burst_sz, ref_op);TEST_ASSERT_SUCCESS (ret,"Validation failed!" ); }@@ -5065,9 +5057,9 @@ validation_latency_test(struct active_device *ad,if (op_type ==RTE_BBDEV_OP_TURBO_DEC )
app/test-bbdev/test_bbdev_perf.c validation_latency_test(struct active_device *ad, · line 5065
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +3
-3iter =latency_test_dec (op_params->mp, bufs, - op_params->ref_dec_op, op_params->vector_mask, - ad->dev_id, queue_id, num_to_process, - burst_sz, &total_time, &min_time, &max_time, latency_flag); + op_params->ref_dec_op, ad->dev_id, queue_id, + num_to_process, burst_sz, &total_time, + &min_time, &max_time, latency_flag); else if (op_type == RTE_BBDEV_OP_LDPC_ENC) iter = latency_test_ldpc_enc(op_params->mp, bufs, op_params->ref_enc_op, ad->dev_id, queue_id,
Thanks,
Maxime
PATCH v2 14/16 test/bbdev: use mbuf reset function
In reply to Hernan VargasUse mbuf_reset function for code consistency.
No functional impact.
--- app/test-bbdev/test_bbdev_perf.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
app/test-bbdev/test_bbdev_perf.c +7 -7
@@ -3173,11 +3173,11 @@ throughput_intr_lcore_ldpc_dec(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) {if (!loopback) -rte_pktmbuf_reset (ops[i]->ldpc_dec.hard_output.data); +mbuf_reset (ops[i]->ldpc_dec.hard_output.data);if (hc_out || loopback)mbuf_reset (ops[i]->ldpc_dec.harq_combined_output.data);if (ops[i]->ldpc_dec.soft_output.data !=NULL ) -rte_pktmbuf_reset (ops[i]->ldpc_dec.soft_output.data); +mbuf_reset (ops[i]->ldpc_dec.soft_output.data); } tp->start_time =rte_rdtsc_precise ();
@@ -3267,9 +3267,9 @@ throughput_intr_lcore_dec(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) { -rte_pktmbuf_reset (ops[i]->turbo_dec.hard_output.data); +mbuf_reset (ops[i]->turbo_dec.hard_output.data);if (ops[i]->turbo_dec.soft_output.data !=NULL ) -rte_pktmbuf_reset (ops[i]->turbo_dec.soft_output.data); +mbuf_reset (ops[i]->turbo_dec.soft_output.data); } tp->start_time =rte_rdtsc_precise ();
@@ -3356,7 +3356,7 @@ throughput_intr_lcore_enc(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) -rte_pktmbuf_reset (ops[i]->turbo_enc.output.data); +mbuf_reset (ops[i]->turbo_enc.output.data); tp->start_time =rte_rdtsc_precise ();for (enqueued =0 ; enqueued < num_to_process;) {
@@ -3444,7 +3444,7 @@ throughput_intr_lcore_ldpc_enc(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) -rte_pktmbuf_reset (ops[i]->turbo_enc.output.data); +mbuf_reset (ops[i]->turbo_enc.output.data); tp->start_time =rte_rdtsc_precise ();for (enqueued =0 ; enqueued < num_to_process;) {
@@ -3532,7 +3532,7 @@ throughput_intr_lcore_fft(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) -rte_pktmbuf_reset (ops[i]->fft.base_output.data); +mbuf_reset (ops[i]->fft.base_output.data); tp->start_time =rte_rdtsc_precise ();for (enqueued =0 ; enqueued < num_to_process;) {
Signature · 1 line
2.37.1
Re: [PATCH v2 14/16] test/bbdev: use mbuf reset function
Quoted 57 lines · Hernan Vargas
Use
mbuf_resetfunction for code consistency.
No functional impact.Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)app/test-bbdev/test_bbdev_perf.c +6
-6@@ -3173,11 +3173,11 @@ throughput_intr_lcore_ldpc_dec(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) {if (!loopback) -rte_pktmbuf_reset (ops[i]->ldpc_dec.hard_output.data); +mbuf_reset (ops[i]->ldpc_dec.hard_output.data);if (hc_out || loopback)mbuf_reset (ops[i]->ldpc_dec.harq_combined_output.data);if (ops[i]->ldpc_dec.soft_output.data !=NULL ) -rte_pktmbuf_reset (ops[i]->ldpc_dec.soft_output.data); +mbuf_reset (ops[i]->ldpc_dec.soft_output.data); } tp->start_time =rte_rdtsc_precise ();@@ -3267,9 +3267,9 @@ throughput_intr_lcore_dec(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) { -rte_pktmbuf_reset (ops[i]->turbo_dec.hard_output.data); +mbuf_reset (ops[i]->turbo_dec.hard_output.data);if (ops[i]->turbo_dec.soft_output.data !=NULL ) -rte_pktmbuf_reset (ops[i]->turbo_dec.soft_output.data); +mbuf_reset (ops[i]->turbo_dec.soft_output.data); } tp->start_time =rte_rdtsc_precise ();@@ -3356,7 +3356,7 @@ throughput_intr_lcore_enc(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) -rte_pktmbuf_reset (ops[i]->turbo_enc.output.data); +mbuf_reset (ops[i]->turbo_enc.output.data); tp->start_time =rte_rdtsc_precise ();for (enqueued =0 ; enqueued < num_to_process;) {@@ -3444,7 +3444,7 @@ throughput_intr_lcore_ldpc_enc(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) -rte_pktmbuf_reset (ops[i]->turbo_enc.output.data); +mbuf_reset (ops[i]->turbo_enc.output.data); tp->start_time =rte_rdtsc_precise ();for (enqueued =0 ; enqueued < num_to_process;) {
app/test-bbdev/test_bbdev_perf.c throughput_intr_lcore_fft(void *arg) · line 3538
Quoted 8 lines
app/test-bbdev/test_bbdev_perf.c +1
-1@@ -3532,7 +3532,7 @@ throughput_intr_lcore_fft(void *arg)for (j =0 ; j <TEST_REPETITIONS ; ++j) {for (i =0 ; i < num_to_process; ++i) -rte_pktmbuf_reset (ops[i]->fft.base_output.data); +mbuf_reset (ops[i]->fft.base_output.data); tp->start_time =rte_rdtsc_precise ();for (enqueued =0 ; enqueued < num_to_process;) {
Thanks,
Maxime
PATCH v2 15/16 test/bbdev: remove max iteration from vectors
In reply to Hernan VargasRemove iter_max from test vectors as this value is passed as an argument to the test.
No functional impact.
--- app/test-bbdev/test_bbdev_perf.c | 4 ++++ app/test-bbdev/test_bbdev_vector.c | 14 -------------- app/test-bbdev/test_bbdev_vector.h | 1 - app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_0.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_1.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_2.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v11835.data | 3 --- .../test_vectors/ldpc_dec_v2342_drop.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v7813.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v8480.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v8568.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v9503.data | 3 --- .../turbo_dec_c1_k40_r0_e17280_sbd_negllr.data | 3 --- ...k6144_r0_e10376_crc24b_sbd_negllr_high_snr.data | 3 --- ..._k6144_r0_e10376_crc24b_sbd_negllr_low_snr.data | 3 --- .../turbo_dec_c1_k6144_r0_e34560_posllr.data | 3 --- .../turbo_dec_c1_k6144_r0_e34560_sbd_negllr.data | 3 --- .../turbo_dec_c1_k6144_r0_e34560_sbd_posllr.data | 3 --- ...bo_dec_c2_k3136_r0_e4920_sbd_negllr_crc24b.data | 3 --- 19 files changed, 4 insertions(+), 63 deletions(-)
app/test-bbdev/test_bbdev_perf.c +4 -0
@@ -3161,6 +3161,8 @@ throughput_intr_lcore_ldpc_dec(void *arg) num_to_process);TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_to_process); + ref_op->ldpc_dec.iter_max =get_iter_max (); +if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_ldpc_dec_op(ops, num_to_process, 0, bufs->inputs, bufs->hard_outputs, bufs->soft_outputs,
@@ -5238,6 +5240,7 @@ offload_latency_test_dec(struct rte_mempool *mempool, struct test_buffers *bufs, ret =rte_bbdev_dec_op_alloc_bulk (mempool, ops_enq, burst_sz);TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , burst_sz); + ref_op->turbo_dec.iter_max =get_iter_max ();if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_dec_op(ops_enq, burst_sz, dequeued, bufs->inputs,
@@ -5324,6 +5327,7 @@ offload_latency_test_ldpc_dec(struct rte_mempool *mempool, ret =rte_bbdev_dec_op_alloc_bulk (mempool, ops_enq, burst_sz);TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , burst_sz); + ref_op->ldpc_dec.iter_max =get_iter_max ();if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_ldpc_dec_op(ops_enq, burst_sz, dequeued, bufs->inputs,
app/test-bbdev/test_bbdev_vector.c +0 -14
@@ -532,10 +532,6 @@ parse_decoder_params(const char *key_token, char *token, vector->mask |=TEST_BBDEV_VF_RV_INDEX ; turbo_dec->rv_index = (uint8_t )strtoul (token, &err,0 ); ret = ((err ==NULL ) || (*err !='\0' )) ?-1 :0 ; - }else if (!strcmp (key_token,"iter_max" )) { - vector->mask |=TEST_BBDEV_VF_ITER_MAX ; - turbo_dec->iter_max = (uint8_t )strtoul (token, &err,0 ); - ret = ((err ==NULL ) || (*err !='\0' )) ?-1 :0 ; }else if (!strcmp (key_token,"iter_min" )) { vector->mask |=TEST_BBDEV_VF_ITER_MIN ; turbo_dec->iter_min = (uint8_t )strtoul (token, &err,0 );
@@ -862,10 +858,6 @@ parse_ldpc_decoder_params(const char *key_token, char *token, vector->mask |=TEST_BBDEV_VF_EXPECTED_ITER_COUNT ; ldpc_dec->iter_count = (uint8_t )strtoul (token, &err,0 ); ret = ((err ==NULL ) || (*err !='\0' )) ?-1 :0 ; - }else if (!strcmp (key_token,"iter_max" )) { - vector->mask |=TEST_BBDEV_VF_ITER_MAX ; - ldpc_dec->iter_max = (uint8_t )strtoul (token, &err,0 ); - ret = ((err ==NULL ) || (*err !='\0' )) ?-1 :0 ; }else if (!strcmp (key_token,"code_block_mode" )) { vector->mask |=TEST_BBDEV_VF_CODE_BLOCK_MODE ; ldpc_dec->code_block_mode = (uint8_t )strtoul (token, &err,0 );
@@ -1260,9 +1252,6 @@ check_decoder(struct test_bbdev_vector *vector)if (!(mask &TEST_BBDEV_VF_ITER_MIN ))printf ("WARNING: iter_min was not specified in vector file and will be set to 0\n" ); -if (!(mask &TEST_BBDEV_VF_ITER_MAX )) -printf ( -"WARNING: iter_max was not specified in vector file and will be set to 0\n" );if (!(mask &TEST_BBDEV_VF_EXPECTED_ITER_COUNT ))printf ("WARNING: expected_iter_count was not specified in vector file and iter_count will not be validated\n" );
@@ -1335,9 +1324,6 @@ check_ldpc_decoder(struct test_bbdev_vector *vector)if (!(mask &TEST_BBDEV_VF_RV_INDEX ))printf ("INFO: rv_index was not specified in vector file and will be set to 0\n" ); -if (!(mask &TEST_BBDEV_VF_ITER_MAX )) -printf ( -"WARNING: iter_max was not specified in vector file and will be set to 0\n" );if (!(mask &TEST_BBDEV_VF_EXPECTED_ITER_COUNT ))printf ("WARNING: expected_iter_count was not specified in vector file and iter_count will not be validated\n" );
app/test-bbdev/test_bbdev_vector.h +0 -1
@@ -19,7 +19,6 @@ enum {TEST_BBDEV_VF_C = (1ULL <<7 ),TEST_BBDEV_VF_CAB = (1ULL <<8 ),TEST_BBDEV_VF_RV_INDEX = (1ULL <<9 ), -TEST_BBDEV_VF_ITER_MAX = (1ULL <<10 ),TEST_BBDEV_VF_ITER_MIN = (1ULL <<11 ),TEST_BBDEV_VF_EXPECTED_ITER_COUNT = (1ULL <<12 ), TEST_BBDEV_VF_EXT_SCALE = (1ULL << 13),
app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_0.data +0 -3
@@ -340,9 +340,6 @@ rv_index= code_block_mode= 1 -iter_max= -20 - expected_iter_count= 20
app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_1.data +0 -3
@@ -671,9 +671,6 @@ rv_index= code_block_mode= 1 -iter_max= -20 - expected_iter_count= 4
app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_2.data +0 -3
@@ -889,9 +889,6 @@ rv_index= code_block_mode= 1 -iter_max= -20 - expected_iter_count= 3
app/test-bbdev/test_vectors/ldpc_dec_v11835.data +0 -3
@@ -36,9 +36,6 @@ rv_index= code_block_mode= 1 -iter_max= -20 - expected_iter_count= 4
app/test-bbdev/test_vectors/ldpc_dec_v2342_drop.data +0 -3
@@ -732,9 +732,6 @@ rv_index = code_block_mode = 1 -iter_max = -20 - expected_iter_count = 3
app/test-bbdev/test_vectors/ldpc_dec_v7813.data +0 -3
@@ -35,9 +35,6 @@ rv_index = code_block_mode = 1 -iter_max = -8 - expected_iter_count = 6
app/test-bbdev/test_vectors/ldpc_dec_v8480.data +0 -3
@@ -61,9 +61,6 @@ rv_index = code_block_mode = 1 -iter_max = -8 - expected_iter_count = 3
app/test-bbdev/test_vectors/ldpc_dec_v8568.data +0 -3
@@ -242,9 +242,6 @@ rv_index = code_block_mode = 1 -iter_max = -20 - expected_iter_count = 6
app/test-bbdev/test_vectors/ldpc_dec_v9503.data +0 -3
@@ -1202,9 +1202,6 @@ rv_index = code_block_mode = 1 -iter_max = -20 - expected_iter_count = 3
app/test-bbdev/test_vectors/turbo_dec_c1_k40_r0_e17280_sbd_negllr.data +0 -3
@@ -31,9 +31,6 @@ k = rv_index = 1 -iter_max = -8 - iter_min = 4
app/test-bbdev/test_vectors/turbo_dec_c1_k6144_r0_e10376_crc24b_sbd_negllr_high_snr.data +0 -3
@@ -617,9 +617,6 @@ k = e = 10376 -iter_max = -8 - iter_min = 4
app/test-bbdev/test_vectors/turbo_dec_c1_k6144_r0_e10376_crc24b_sbd_negllr_low_snr.data +0 -3
@@ -617,9 +617,6 @@ k = e = 10376 -iter_max = -8 - iter_min = 4
app/test-bbdev/test_vectors/turbo_dec_c1_k6144_r0_e34560_posllr.data +0 -3
@@ -620,9 +620,6 @@ k = rv_index = 0 -iter_max = -8 - iter_min = 4
app/test-bbdev/test_vectors/turbo_dec_c1_k6144_r0_e34560_sbd_negllr.data +0 -3
@@ -1198,9 +1198,6 @@ k = rv_index = 0 -iter_max = -8 - iter_min = 4
app/test-bbdev/test_vectors/turbo_dec_c1_k6144_r0_e34560_sbd_posllr.data +0 -3
@@ -1199,9 +1199,6 @@ k = rv_index = 0 -iter_max = -8 - iter_min = 4
app/test-bbdev/test_vectors/turbo_dec_c2_k3136_r0_e4920_sbd_negllr_crc24b.data +0 -3
@@ -651,9 +651,6 @@ k_pos = rv_index = 0 -iter_max = -8 - iter_min = 4
Signature · 1 line
2.37.1
Re: [PATCH v2 15/16] test/bbdev: remove max iteration from vectors
Quoted 18 lines · Hernan Vargas
Remove
iter_maxfrom test vectors as this value is passed as an argument to the test.
No functional impact.Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 4 ++++ app/test-bbdev/test_bbdev_vector.c | 14 -------------- app/test-bbdev/test_bbdev_vector.h | 1 - app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_0.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_1.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_HARQ_1_2.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v11835.data | 3 --- .../test_vectors/ldpc_dec_v2342_drop.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v7813.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v8480.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v8568.data | 3 --- app/test-bbdev/test_vectors/ldpc_dec_v9503.data | 3 ---
Quoted 9 lines
.../turbo_dec_c1_k40_r0_e17280_sbd_negllr.data | 3 --- ...k6144_r0_e10376_crc24b_sbd_negllr_high_snr.data | 3 --- ..._k6144_r0_e10376_crc24b_sbd_negllr_low_snr.data | 3 --- .../turbo_dec_c1_k6144_r0_e34560_posllr.data | 3 --- .../turbo_dec_c1_k6144_r0_e34560_sbd_negllr.data | 3 --- .../turbo_dec_c1_k6144_r0_e34560_sbd_posllr.data | 3 --- ...bo_dec_c2_k3136_r0_e4920_sbd_negllr_crc24b.data | 3 --- 19 files changed, 4 insertions(+), 63 deletions(-)
Thanks,
Maxime
PATCH v2 16/16 test/bbdev: remove iter count from bler test
In reply to Hernan Vargasiter_count doesn't need to be set equal to iter_max for bler tests. This is only needed in throughput and latency tests because early termination is disabled for those tests.
--- app/test-bbdev/test_bbdev_perf.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-)
app/test-bbdev/test_bbdev_perf.c +6 -10
@@ -3714,12 +3714,10 @@ bler_pmd_lcore_ldpc_dec(void *arg)TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_ops);/* For BLER tests we need to enable early termination */ -if (!check_bit (ref_op->ldpc_dec.op_flags, -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) - ref_op->ldpc_dec.op_flags += -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; +if (!check_bit (ref_op->ldpc_dec.op_flags,RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) + ref_op->ldpc_dec.op_flags +=RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; + ref_op->ldpc_dec.iter_max =get_iter_max (); - ref_op->ldpc_dec.iter_count = ref_op->ldpc_dec.iter_max;if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_ldpc_dec_op(ops_enq, num_ops, 0, bufs->inputs,
@@ -3838,12 +3836,10 @@ bler_pmd_lcore_turbo_dec(void *arg)TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_ops);/* For BLER tests we need to enable early termination */ -if (!check_bit (ref_op->turbo_dec.op_flags, -RTE_BBDEV_TURBO_EARLY_TERMINATION )) - ref_op->turbo_dec.op_flags += -RTE_BBDEV_TURBO_EARLY_TERMINATION ; +if (!check_bit (ref_op->turbo_dec.op_flags,RTE_BBDEV_TURBO_EARLY_TERMINATION )) + ref_op->turbo_dec.op_flags +=RTE_BBDEV_TURBO_EARLY_TERMINATION ; + ref_op->turbo_dec.iter_max =get_iter_max (); - ref_op->turbo_dec.iter_count = ref_op->turbo_dec.iter_max;if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_dec_op(ops_enq, num_ops, 0, bufs->inputs,
Signature · 1 line
2.37.1
Re: [PATCH v2 16/16] test/bbdev: remove iter count from bler test
Quoted 35 lines · Hernan Vargas
iter_countdoesn't need to be set equal toiter_maxfor bler tests. This is only needed in throughput and latency tests because early termination is disabled for those tests.Signed-off-by Hernan Vargas <[email protected]>--- app/test-bbdev/test_bbdev_perf.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-)app/test-bbdev/test_bbdev_perf.c +3
-7@@ -3714,12 +3714,10 @@ bler_pmd_lcore_ldpc_dec(void *arg)TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_ops);/* For BLER tests we need to enable early termination */ -if (!check_bit (ref_op->ldpc_dec.op_flags, -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) - ref_op->ldpc_dec.op_flags += -RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; +if (!check_bit (ref_op->ldpc_dec.op_flags,RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE )) + ref_op->ldpc_dec.op_flags +=RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE ; + ref_op->ldpc_dec.iter_max =get_iter_max (); - ref_op->ldpc_dec.iter_count = ref_op->ldpc_dec.iter_max;if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_ldpc_dec_op(ops_enq, num_ops, 0, bufs->inputs,@@ -3838,12 +3836,10 @@ bler_pmd_lcore_turbo_dec(void *arg)TEST_ASSERT_SUCCESS (ret,"Allocation failed for %d ops" , num_ops);/* For BLER tests we need to enable early termination */ -if (!check_bit (ref_op->turbo_dec.op_flags, -RTE_BBDEV_TURBO_EARLY_TERMINATION ))
app/test-bbdev/test_bbdev_perf.c bler_pmd_lcore_turbo_dec(void *arg) · line 3845
Quoted 10 lines
app/test-bbdev/test_bbdev_perf.c +3
-3- ref_op->turbo_dec.op_flags += -RTE_BBDEV_TURBO_EARLY_TERMINATION ; +if (!check_bit (ref_op->turbo_dec.op_flags,RTE_BBDEV_TURBO_EARLY_TERMINATION )) + ref_op->turbo_dec.op_flags +=RTE_BBDEV_TURBO_EARLY_TERMINATION ; + ref_op->turbo_dec.iter_max =get_iter_max (); - ref_op->turbo_dec.iter_count = ref_op->turbo_dec.iter_max;if (test_vector.op_type !=RTE_BBDEV_OP_NONE ) copy_reference_dec_op(ops_enq, num_ops, 0, bufs->inputs,
Thanks,
Maxime
Hi Hernan,
In order to make reviewers and maintainers task easier, please apply the R-by and A-by given on previous revisions for patches that haven't changed.
You can do it automatically when initiating a new revision by applying the previous one from patchwork, which takes care of adding them.
Thanks,
Maxime
Quoted 42 lines · Hernan Vargas
v2: Made changes requested during review. Added 3 commits.
v1: Upstreaming bbdev-test changes for 23.03.Hernan Vargas (16):
test/bbdev: fix seg fault for non supported HARQ len test/bbdev: extend HARQ tolerance test/bbdev: refactor TB throughput report test/bbdev: add timeout for latency tests test/bbdev: enable early termination for validation test/bbdev: report device status in test-bbdev test/bbdev: test start/stop bbdev API test/bbdev: add support for BLER for 4G test/bbdev: extend support for large TB test/bbdev: adjustment for soft output test/bbdev: expose warning counters test/bbdev: remove check for invalid opaque data test/bbdev: remove iteration count check test/bbdev: use mbuf reset function test/bbdev: remove max iteration from vectors test/bbdev: remove iter count from bler testapp/test-bbdev/test_bbdev_perf.c | 400 +++++++++++++----- app/test-bbdev/test_bbdev_vector.c | 14 - app/test-bbdev/test_bbdev_vector.h | 1 - .../test_vectors/ldpc_dec_HARQ_1_0.data | 3 - .../test_vectors/ldpc_dec_HARQ_1_1.data | 3 - .../test_vectors/ldpc_dec_HARQ_1_2.data | 3 - .../test_vectors/ldpc_dec_v11835.data | 3 - .../test_vectors/ldpc_dec_v2342_drop.data | 3 - .../test_vectors/ldpc_dec_v7813.data | 3 - .../test_vectors/ldpc_dec_v8480.data | 3 - .../test_vectors/ldpc_dec_v8568.data | 3 - .../test_vectors/ldpc_dec_v9503.data | 3 - ...turbo_dec_c1_k40_r0_e17280_sbd_negllr.data | 3 - ..._r0_e10376_crc24b_sbd_negllr_high_snr.data | 3 - ...4_r0_e10376_crc24b_sbd_negllr_low_snr.data | 3 - .../turbo_dec_c1_k6144_r0_e34560_posllr.data | 3 - ...rbo_dec_c1_k6144_r0_e34560_sbd_negllr.data | 3 - ...rbo_dec_c1_k6144_r0_e34560_sbd_posllr.data | 3 - ...c_c2_k3136_r0_e4920_sbd_negllr_crc24b.data | 3 - 19 files changed, 301 insertions(+), 162 deletions(-)