X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;ds=sidebyside;f=app%2Ftest-gpudev%2Fmain.c;h=f065e6cd81a0aa20591933f51ba0395331f21e05;hb=9b8cae4d991ee4fb0faf92d8994879d6776df181;hp=0b44a0990c53e0b50aa9fd38fa93e10ef3101e0c;hpb=e818c4e2bf5010944fc5b4ef9a52b37df6e47c1a;p=dpdk.git diff --git a/app/test-gpudev/main.c b/app/test-gpudev/main.c index 0b44a0990c..f065e6cd81 100644 --- a/app/test-gpudev/main.c +++ b/app/test-gpudev/main.c @@ -68,48 +68,68 @@ alloc_gpu_memory(uint16_t gpu_id) void *ptr_1 = NULL; void *ptr_2 = NULL; size_t buf_bytes = 1024; + unsigned int align = 4096; int ret; - printf("\n=======> TEST: Allocate GPU memory\n"); + printf("\n=======> TEST: Allocate GPU memory\n\n"); - /* Alloc memory on GPU 0 */ - ptr_1 = rte_gpu_mem_alloc(gpu_id, buf_bytes); + /* Alloc memory on GPU 0 without any specific alignment */ + ptr_1 = rte_gpu_mem_alloc(gpu_id, buf_bytes, 0); if (ptr_1 == NULL) { fprintf(stderr, "rte_gpu_mem_alloc GPU memory returned error\n"); - return -1; + goto error; } - printf("GPU memory allocated at 0x%p %zdB\n", ptr_1, buf_bytes); + printf("GPU memory allocated at 0x%p size is %zd bytes\n", + ptr_1, buf_bytes); - ptr_2 = rte_gpu_mem_alloc(gpu_id, buf_bytes); + /* Alloc memory on GPU 0 with 4kB alignment */ + ptr_2 = rte_gpu_mem_alloc(gpu_id, buf_bytes, align); if (ptr_2 == NULL) { fprintf(stderr, "rte_gpu_mem_alloc GPU memory returned error\n"); - return -1; + goto error; + } + printf("GPU memory allocated at 0x%p size is %zd bytes\n", + ptr_2, buf_bytes); + + if (((uintptr_t)ptr_2) % align) { + fprintf(stderr, "Memory address 0x%p is not aligned to %u\n", ptr_2, align); + goto error; } - printf("GPU memory allocated at 0x%p %zdB\n", ptr_2, buf_bytes); ret = rte_gpu_mem_free(gpu_id, (uint8_t *)(ptr_1)+0x700); if (ret < 0) { - printf("GPU memory 0x%p + 0x700 NOT freed because of memory address not recognized by driver\n", ptr_1); + printf("GPU memory 0x%p NOT freed: GPU driver didn't find this memory address internally.\n", + (uint8_t *)(ptr_1)+0x700); } else { - fprintf(stderr, "rte_gpu_mem_free erroneously freed GPU memory 0x%p + 0x700\n", ptr_1); - return -1; + fprintf(stderr, "ERROR: rte_gpu_mem_free freed GPU memory 0x%p\n", + (uint8_t *)(ptr_1)+0x700); + goto error; } ret = rte_gpu_mem_free(gpu_id, ptr_2); if (ret < 0) { fprintf(stderr, "rte_gpu_mem_free returned error %d\n", ret); - return -1; + goto error; } printf("GPU memory 0x%p freed\n", ptr_2); ret = rte_gpu_mem_free(gpu_id, ptr_1); if (ret < 0) { fprintf(stderr, "rte_gpu_mem_free returned error %d\n", ret); - return -1; + goto error; } printf("GPU memory 0x%p freed\n", ptr_1); + printf("\n=======> TEST: PASSED\n"); return 0; + +error: + + rte_gpu_mem_free(gpu_id, ptr_1); + rte_gpu_mem_free(gpu_id, ptr_2); + + printf("\n=======> TEST: FAILED\n"); + return -1; } static int @@ -119,39 +139,286 @@ register_cpu_memory(uint16_t gpu_id) size_t buf_bytes = 1024; int ret; - printf("\n=======> TEST: Register CPU memory\n"); + printf("\n=======> TEST: Register CPU memory\n\n"); /* Alloc memory on CPU visible from GPU 0 */ ptr = rte_zmalloc(NULL, buf_bytes, 0); if (ptr == NULL) { fprintf(stderr, "Failed to allocate CPU memory.\n"); - return -1; + goto error; } ret = rte_gpu_mem_register(gpu_id, buf_bytes, ptr); if (ret < 0) { fprintf(stderr, "rte_gpu_mem_register CPU memory returned error %d\n", ret); - return -1; + goto error; } printf("CPU memory registered at 0x%p %zdB\n", ptr, buf_bytes); ret = rte_gpu_mem_unregister(gpu_id, (uint8_t *)(ptr)+0x700); if (ret < 0) { - printf("CPU memory 0x%p + 0x700 NOT unregistered because of memory address not recognized by driver\n", ptr); + printf("CPU memory 0x%p NOT unregistered: GPU driver didn't find this memory address internally\n", + (uint8_t *)(ptr)+0x700); } else { - fprintf(stderr, "rte_gpu_mem_free erroneously freed GPU memory 0x%p + 0x700\n", ptr); - return -1; + fprintf(stderr, "ERROR: rte_gpu_mem_unregister unregistered GPU memory 0x%p\n", + (uint8_t *)(ptr)+0x700); + goto error; } - printf("CPU memory 0x%p unregistered\n", ptr); ret = rte_gpu_mem_unregister(gpu_id, ptr); if (ret < 0) { fprintf(stderr, "rte_gpu_mem_unregister returned error %d\n", ret); - return -1; + goto error; } printf("CPU memory 0x%p unregistered\n", ptr); + rte_free(ptr); + + printf("\n=======> TEST: PASSED\n"); + return 0; + +error: + + rte_gpu_mem_unregister(gpu_id, ptr); + rte_free(ptr); + printf("\n=======> TEST: FAILED\n"); + return -1; +} + +static int +gpu_mem_cpu_map(uint16_t gpu_id) +{ + void *ptr_gpu = NULL; + void *ptr_cpu = NULL; + size_t buf_bytes = 1024; + unsigned int align = 4096; + int ret; + + printf("\n=======> TEST: Map GPU memory for CPU visibility\n\n"); + + /* Alloc memory on GPU 0 with 4kB alignment */ + ptr_gpu = rte_gpu_mem_alloc(gpu_id, buf_bytes, align); + if (ptr_gpu == NULL) { + fprintf(stderr, "rte_gpu_mem_alloc GPU memory returned error\n"); + goto error; + } + printf("GPU memory allocated at 0x%p size is %zd bytes\n", + ptr_gpu, buf_bytes); + + ptr_cpu = rte_gpu_mem_cpu_map(gpu_id, buf_bytes, ptr_gpu); + if (ptr_cpu == NULL) { + fprintf(stderr, "rte_gpu_mem_cpu_map returned error\n"); + goto error; + } + printf("GPU memory CPU mapped at 0x%p\n", ptr_cpu); + + ((uint8_t *)ptr_cpu)[0] = 0x4; + ((uint8_t *)ptr_cpu)[1] = 0x5; + ((uint8_t *)ptr_cpu)[2] = 0x6; + + printf("GPU memory first 3 bytes set from CPU: %x %x %x\n", + ((uint8_t *)ptr_cpu)[0], + ((uint8_t *)ptr_cpu)[1], + ((uint8_t *)ptr_cpu)[2]); + + ret = rte_gpu_mem_cpu_unmap(gpu_id, ptr_gpu); + if (ret < 0) { + fprintf(stderr, "rte_gpu_mem_cpu_unmap returned error %d\n", ret); + goto error; + } + printf("GPU memory CPU unmapped, 0x%p not valid anymore\n", ptr_cpu); + + ret = rte_gpu_mem_free(gpu_id, ptr_gpu); + if (ret < 0) { + fprintf(stderr, "rte_gpu_mem_free returned error %d\n", ret); + goto error; + } + printf("GPU memory 0x%p freed\n", ptr_gpu); + + printf("\n=======> TEST: PASSED\n"); + return 0; + +error: + + rte_gpu_mem_cpu_unmap(gpu_id, ptr_gpu); + rte_gpu_mem_free(gpu_id, ptr_gpu); + + printf("\n=======> TEST: FAILED\n"); + return -1; +} + +static int +create_update_comm_flag(uint16_t gpu_id) +{ + struct rte_gpu_comm_flag devflag; + int ret = 0; + uint32_t set_val; + uint32_t get_val; + + printf("\n=======> TEST: Communication flag\n\n"); + + ret = rte_gpu_comm_create_flag(gpu_id, &devflag, RTE_GPU_COMM_FLAG_CPU); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_create_flag returned error %d\n", ret); + goto error; + } + + set_val = 25; + ret = rte_gpu_comm_set_flag(&devflag, set_val); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_set_flag returned error %d\n", ret); + goto error; + } + + ret = rte_gpu_comm_get_flag_value(&devflag, &get_val); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_get_flag_value returned error %d\n", ret); + goto error; + } + + printf("Communication flag value at 0x%p was set to %d and current value is %d\n", + devflag.ptr, set_val, get_val); + + set_val = 38; + ret = rte_gpu_comm_set_flag(&devflag, set_val); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_set_flag returned error %d\n", ret); + goto error; + } + + ret = rte_gpu_comm_get_flag_value(&devflag, &get_val); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_get_flag_value returned error %d\n", ret); + goto error; + } + + printf("Communication flag value at 0x%p was set to %d and current value is %d\n", + devflag.ptr, set_val, get_val); + + ret = rte_gpu_comm_destroy_flag(&devflag); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_destroy_flags returned error %d\n", ret); + goto error; + } + + printf("\n=======> TEST: PASSED\n"); return 0; + +error: + + rte_gpu_comm_destroy_flag(&devflag); + printf("\n=======> TEST: FAILED\n"); + return -1; +} + +static int +simulate_gpu_task(struct rte_gpu_comm_list *comm_list_item, int num_pkts) +{ + int idx; + + if (comm_list_item == NULL) + return -1; + + for (idx = 0; idx < num_pkts; idx++) { + /** + * consume(comm_list_item->pkt_list[idx].addr); + */ + } + /* + * A real GPU workload function can't directly call rte_gpu_comm_set_status + * because it's a CPU-only function. + * A real GPU workload should implement the content + * of rte_gpu_comm_set_status() in GPU specific code. + */ + rte_gpu_comm_set_status(comm_list_item, RTE_GPU_COMM_LIST_DONE); + + return 0; +} + +static int +create_update_comm_list(uint16_t gpu_id) +{ + int ret = 0; + int i = 0; + struct rte_gpu_comm_list *comm_list = NULL; + uint32_t num_comm_items = 1024; + struct rte_mbuf *mbufs[10]; + + printf("\n=======> TEST: Communication list\n\n"); + + comm_list = rte_gpu_comm_create_list(gpu_id, num_comm_items); + if (comm_list == NULL) { + fprintf(stderr, "rte_gpu_comm_create_list returned error %d\n", ret); + goto error; + } + + /** + * Simulate DPDK receive functions like rte_eth_rx_burst() + */ + for (i = 0; i < 10; i++) { + mbufs[i] = rte_zmalloc(NULL, sizeof(struct rte_mbuf), 0); + if (mbufs[i] == NULL) { + fprintf(stderr, "Failed to allocate fake mbufs in CPU memory.\n"); + goto error; + } + + memset(mbufs[i], 0, sizeof(struct rte_mbuf)); + } + + /** + * Populate just the first item of the list + */ + ret = rte_gpu_comm_populate_list_pkts(&(comm_list[0]), mbufs, 10); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_populate_list_pkts returned error %d\n", ret); + goto error; + } + + ret = rte_gpu_comm_cleanup_list(&(comm_list[0])); + if (ret == 0) { + fprintf(stderr, "rte_gpu_comm_cleanup_list erroneously cleaned the list even if packets have not been consumed yet\n"); + goto error; + } + printf("Communication list not cleaned because packets have not been consumed yet.\n"); + + /** + * Simulate a GPU tasks going through the packet list to consume + * mbufs packets and release them + */ + printf("Consuming packets...\n"); + simulate_gpu_task(&(comm_list[0]), 10); + + /** + * Packets have been consumed, now the communication item + * and the related mbufs can be all released + */ + ret = rte_gpu_comm_cleanup_list(&(comm_list[0])); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_cleanup_list returned error %d\n", ret); + goto error; + } + + printf("Communication list cleaned because packets have been consumed now.\n"); + + ret = rte_gpu_comm_destroy_list(comm_list, num_comm_items); + if (ret < 0) { + fprintf(stderr, "rte_gpu_comm_destroy_list returned error %d\n", ret); + goto error; + } + + for (i = 0; i < 10; i++) + rte_free(mbufs[i]); + + printf("\n=======> TEST: PASSED\n"); + return 0; + +error: + + rte_gpu_comm_destroy_list(comm_list, num_comm_items); + for (i = 0; i < 10; i++) + rte_free(mbufs[i]); + printf("\n=======> TEST: FAILED\n"); + return -1; } int @@ -203,10 +470,16 @@ main(int argc, char **argv) */ alloc_gpu_memory(gpu_id); register_cpu_memory(gpu_id); + gpu_mem_cpu_map(gpu_id); + + /** + * Communication items test + */ + create_update_comm_flag(gpu_id); + create_update_comm_list(gpu_id); /* clean up the EAL */ rte_eal_cleanup(); - printf("Bye...\n"); return EXIT_SUCCESS; }