Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 3CC5ECCFA04 for ; Tue, 4 Nov 2025 14:29:55 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id D4B2910E26E; Tue, 4 Nov 2025 14:29:54 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=intel.com header.i=@intel.com header.b="cnJnJwqy"; dkim-atps=neutral Received: from mgamail.intel.com (mgamail.intel.com [192.198.163.11]) by gabe.freedesktop.org (Postfix) with ESMTPS id CE4E810E60F for ; Tue, 4 Nov 2025 14:29:52 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1762266593; x=1793802593; h=from:to:subject:date:message-id:in-reply-to:references: mime-version:content-transfer-encoding; bh=JgKIbic5Yo4xndx96aNIvMKx8F6inKmaNZphkV1TLtI=; b=cnJnJwqyhLg8b4+1MtQNh+mlDtu2kOnSACzNjeH5cHY0KnuhDxJmFzy4 enwTivFgZj+sgSHvf4t+6QkbiKu+qV5TcMVahepGPS1ClvmMAfT1WJ38Y vnab8xLUdQyL21adwOkvRF5Y3Tnpr2tA2i+M20QqTCuHZod2nyZ4xm1aB le1xwMCMdp0coOQ7GHPfJscXff5T5XCUf2gVge17Fd015QdxkbQZ30Sjd xcGkw2AKhEAcSZ21rEgU7SfIk232MiLqqBXJJgyk9pYyK9GPu56INyz99 hg4/I7XfyVdySDA254YONSDH1VrwGPbT8uS9LshrzOrLUtzYbJcfIOKRP Q==; X-CSE-ConnectionGUID: Djhdf7nYTSqa+lNcytLEKw== X-CSE-MsgGUID: e0fkAoF5QS2BpFPJWDtEFw== X-IronPort-AV: E=McAfee;i="6800,10657,11602"; a="74966174" X-IronPort-AV: E=Sophos;i="6.19,279,1754982000"; d="scan'208";a="74966174" Received: from orviesa005.jf.intel.com ([10.64.159.145]) by fmvoesa105.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 04 Nov 2025 06:29:53 -0800 X-CSE-ConnectionGUID: iTPAK2zvQdm09OjOKBvE1A== X-CSE-MsgGUID: jk1BAtyPS9OGkC77j5Gygw== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.19,279,1754982000"; d="scan'208";a="192339862" Received: from dut2943pvc.iind.intel.com ([10.190.215.39]) by orviesa005.jf.intel.com with ESMTP; 04 Nov 2025 06:29:51 -0800 From: nishit.sharma@intel.com To: igt-dev@lists.freedesktop.org, thomas.hellstrom@intel.com, nishit.sharma@intel.com Subject: [PATCH i-g-t v1 2/7] tests/intel/xe_multi-gpusvm.c: Add SVM multi-GPU atomic operations Date: Tue, 4 Nov 2025 14:29:39 +0000 Message-ID: <20251104142947.51744-3-nishit.sharma@intel.com> X-Mailer: git-send-email 2.43.0 In-Reply-To: <20251104142947.51744-1-nishit.sharma@intel.com> References: <20251104142947.51744-1-nishit.sharma@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-BeenThere: igt-dev@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Development mailing list for IGT GPU Tools List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: igt-dev-bounces@lists.freedesktop.org Sender: "igt-dev" From: Nishit Sharma This test performs atomic increment operation on a shared SVM buffer from both GPUs and the CPU in a multi-GPU environment. It uses madvise and prefetch to control buffer placement and verifies correctness and ordering of atomic updates across agents. Signed-off-by: Nishit Sharma --- tests/intel/xe_multi_gpusvm.c | 186 ++++++++++++++++++++++++++++++++++ 1 file changed, 186 insertions(+) diff --git a/tests/intel/xe_multi_gpusvm.c b/tests/intel/xe_multi_gpusvm.c index a88b46323..71bf01ba8 100644 --- a/tests/intel/xe_multi_gpusvm.c +++ b/tests/intel/xe_multi_gpusvm.c @@ -31,6 +31,11 @@ * region both remotely and locally and copies to it. Reads back to * system memory and checks the result. * + * SUBTEST: atomic-inc-gpu-op + * Description: + * This test does atomic operation in multi-gpu by executing atomic + * operation on GPU1 and then atomic operation on GPU2 using same + * adress */ #define MAX_XE_REGIONS 8 @@ -40,6 +45,7 @@ #define BIND_SYNC_VAL 0x686868 #define EXEC_SYNC_VAL 0x676767 #define COPY_SIZE SZ_64M +#define ATOMIC_OP_VAL 56 struct xe_svm_gpu_info { bool supports_faults; @@ -49,8 +55,45 @@ struct xe_svm_gpu_info { int fd; }; +struct test_exec_data { + uint32_t batch[32]; + uint64_t pad; + uint64_t vm_sync; + uint64_t exec_sync; + uint32_t data; + uint32_t expected_data; + uint64_t batch_addr; +}; + static void open_pagemaps(int fd, struct xe_svm_gpu_info *info); +static void +atomic_batch_init(int fd, uint32_t vm, uint64_t src_addr, + uint32_t *bo, uint64_t *addr) +{ + uint32_t batch_bo_size = BATCH_SIZE(fd); + uint32_t batch_bo; + uint64_t batch_addr; + void *batch; + uint32_t *cmd; + int i = 0; + + batch_bo = xe_bo_create(fd, vm, batch_bo_size, vram_if_possible(fd, 0), 0); + batch = xe_bo_map(fd, batch_bo, batch_bo_size); + cmd = (uint32_t *)batch; + + cmd[i++] = MI_ATOMIC | MI_ATOMIC_INC; + cmd[i++] = src_addr; + cmd[i++] = src_addr >> 32; + cmd[i++] = MI_BATCH_BUFFER_END; + + batch_addr = to_user_pointer(batch); + /* Punch a gap in the SVM map where we map the batch_bo */ + xe_vm_bind_lr_sync(fd, vm, batch_bo, 0, batch_addr, batch_bo_size, 0); + *bo = batch_bo; + *addr = batch_addr; +} + static void batch_init(int fd, uint32_t vm, uint64_t src_addr, uint64_t dst_addr, uint64_t copy_size, uint32_t *bo, uint64_t *addr) @@ -222,6 +265,144 @@ gpu_mem_access(struct xe_svm_gpu_info *src_gpu, copy_src_dst(src_gpu, dst_gpu, eci); } +static void +atomic_inc_op(struct xe_svm_gpu_info *gpu0, + struct xe_svm_gpu_info *gpu1, + struct drm_xe_engine_class_instance *eci, + bool prefetch_req) +{ + uint64_t addr; + uint32_t vm[2]; + uint32_t exec_queue[2]; + uint32_t batch_bo; + struct test_exec_data *data; + uint64_t batch_addr; + struct drm_xe_sync sync = {}; + volatile uint64_t *sync_addr; + volatile uint32_t *shared_val; + + vm[0] = xe_vm_create(gpu0->fd, DRM_XE_VM_CREATE_FLAG_LR_MODE | + DRM_XE_VM_CREATE_FLAG_FAULT_MODE, 0); + exec_queue[0] = xe_exec_queue_create(gpu0->fd, vm[0], eci, 0); + xe_vm_bind_lr_sync(gpu0->fd, vm[0], 0, 0, 0, 1ull << gpu0->va_bits, + DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR); + + vm[1] = xe_vm_create(gpu1->fd, DRM_XE_VM_CREATE_FLAG_LR_MODE | + DRM_XE_VM_CREATE_FLAG_FAULT_MODE, 0); + exec_queue[1] = xe_exec_queue_create(gpu1->fd, vm[1], eci, 0); + xe_vm_bind_lr_sync(gpu1->fd, vm[1], 0, 0, 0, 1ull << gpu1->va_bits, + DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR); + + data = aligned_alloc(SZ_2M, SZ_4K); + igt_assert(data); + data[0].vm_sync = 0; + addr = to_user_pointer(data); + + shared_val = (volatile uint32_t *)addr; + *shared_val = ATOMIC_OP_VAL - 1; + + atomic_batch_init(gpu0->fd, vm[0], addr, &batch_bo, &batch_addr); + + /* Place destination in an optionally remote location to test */ + xe_vm_madvise(gpu0->fd, vm[0], addr, SZ_4K, 0, + DRM_XE_MEM_RANGE_ATTR_PREFERRED_LOC, + gpu0->fd, 0, gpu0->vram_regions[0]); + + sync_addr = malloc(sizeof(*sync_addr)); + igt_assert(!!sync_addr); + sync.flags = DRM_XE_SYNC_FLAG_SIGNAL; + sync.type = DRM_XE_SYNC_TYPE_USER_FENCE; + sync.addr = to_user_pointer((uint64_t *)sync_addr); + sync.timeline_value = BIND_SYNC_VAL; + *sync_addr = 0; + + if (prefetch_req) { + xe_vm_prefetch_async(gpu0->fd, vm[0], 0, 0, addr, + SZ_4K, &sync, 1, + DRM_XE_CONSULT_MEM_ADVISE_PREF_LOC); + if (*sync_addr != BIND_SYNC_VAL) + xe_wait_ufence(gpu0->fd, (uint64_t *)sync_addr, BIND_SYNC_VAL, + exec_queue[0], NSEC_PER_SEC * 10); + } + free((void *)sync_addr); + + sync_addr = (void *)((char *)batch_addr + SZ_4K); + sync.addr = to_user_pointer((uint64_t *)sync_addr); + sync.timeline_value = EXEC_SYNC_VAL; + *sync_addr = 0; + + /* Executing ATOMIC_INC on GPU0. */ + xe_exec_sync(gpu0->fd, exec_queue[0], batch_addr, &sync, 1); + if (*sync_addr != EXEC_SYNC_VAL) + xe_wait_ufence(gpu0->fd, (uint64_t *)sync_addr, EXEC_SYNC_VAL, exec_queue[0], + NSEC_PER_SEC * 10); + + igt_assert_eq(*shared_val, ATOMIC_OP_VAL); + + atomic_batch_init(gpu1->fd, vm[1], addr, &batch_bo, &batch_addr); + + /* Place destination in an optionally remote location to test */ + xe_vm_madvise(gpu1->fd, vm[1], addr, SZ_4K, 0, + DRM_XE_MEM_RANGE_ATTR_PREFERRED_LOC, + gpu1->fd, 0, gpu1->vram_regions[0]); + + sync_addr = malloc(sizeof(*sync_addr)); + igt_assert(!!sync_addr); + sync.flags = DRM_XE_SYNC_FLAG_SIGNAL; + sync.type = DRM_XE_SYNC_TYPE_USER_FENCE; + sync.addr = to_user_pointer((uint64_t *)sync_addr); + sync.timeline_value = BIND_SYNC_VAL; + *sync_addr = 0; + + if (prefetch_req) { + xe_vm_prefetch_async(gpu1->fd, vm[1], 0, 0, addr, + SZ_4K, &sync, 1, + DRM_XE_CONSULT_MEM_ADVISE_PREF_LOC); + if (*sync_addr != BIND_SYNC_VAL) + xe_wait_ufence(gpu1->fd, (uint64_t *)sync_addr, BIND_SYNC_VAL, + exec_queue[1], NSEC_PER_SEC * 10); + } + free((void *)sync_addr); + + sync_addr = (void *)((char *)batch_addr + SZ_4K); + sync.addr = to_user_pointer((uint64_t *)sync_addr); + sync.timeline_value = EXEC_SYNC_VAL; + *sync_addr = 0; + + /* Execute ATOMIC_INC on GPU1 */ + xe_exec_sync(gpu1->fd, exec_queue[1], batch_addr, &sync, 1); + if (*sync_addr != EXEC_SYNC_VAL) + xe_wait_ufence(gpu1->fd, (uint64_t *)sync_addr, EXEC_SYNC_VAL, exec_queue[1], + NSEC_PER_SEC * 10); + + igt_assert_eq(*shared_val, ATOMIC_OP_VAL + 1); + + munmap((void *)batch_addr, BATCH_SIZE(gpu0->fd)); + batch_fini(gpu0->fd, vm[0], batch_bo, batch_addr); + batch_fini(gpu1->fd, vm[1], batch_bo, batch_addr); + free(data); + + xe_vm_unbind_lr_sync(gpu0->fd, vm[0], 0, 0, 1ull << gpu0->va_bits); + xe_exec_queue_destroy(gpu0->fd, exec_queue[0]); + xe_vm_destroy(gpu0->fd, vm[0]); + + xe_vm_unbind_lr_sync(gpu1->fd, vm[1], 0, 0, 1ull << gpu1->va_bits); + xe_exec_queue_destroy(gpu1->fd, exec_queue[1]); + xe_vm_destroy(gpu1->fd, vm[1]); +} + +static void +gpu_atomic_inc(struct xe_svm_gpu_info *src_gpu, + struct xe_svm_gpu_info *dst_gpu, + struct drm_xe_engine_class_instance *eci, + bool prefetch_req) +{ + igt_assert(src_gpu); + igt_assert(dst_gpu); + + atomic_inc_op(src_gpu, dst_gpu, eci, prefetch_req); +} + igt_main { struct xe_svm_gpu_info gpus[MAX_XE_GPUS]; @@ -256,6 +437,11 @@ igt_main igt_subtest("cross-gpu-mem-access") gpu_mem_access(&gpus[0], &gpus[1], &eci); + igt_subtest("atomic-inc-gpu-op") { + gpu_atomic_inc(&gpus[0], &gpus[1], &eci, 1); + gpu_atomic_inc(&gpus[0], &gpus[1], &eci, 0); + } + igt_fixture { int cnt; -- 2.48.1