diff --git a/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.c b/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.c
index e26a5890ae9..c0e4280b071 100644
--- a/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.c
+++ b/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.c
@@ -145,8 +145,6 @@ radv_amdgpu_request_to_fence(struct radv_amdgpu_ctx *ctx, struct radv_amdgpu_fen
    fence->fence.ip_instance = req->ip_instance;
    fence->fence.ring = req->ring;
    fence->fence.fence = req->seq_no;
-   fence->user_ptr =
-      (volatile uint64_t *)(ctx->fence_map + req->ip_type * MAX_RINGS_PER_TYPE + req->ring);
 }
 
 static void
@@ -1269,19 +1267,9 @@ radv_amdgpu_ctx_create(struct radeon_winsys *_ws, enum radeon_ctx_priority prior
       goto fail_alloc;
    }
 
-   ctx->fence_map = (uint64_t *)ws->base.buffer_map(ctx->fence_bo);
-   if (!ctx->fence_map) {
-      result = VK_ERROR_OUT_OF_DEVICE_MEMORY;
-      goto fail_map;
-   }
-
-   memset(ctx->fence_map, 0, 4096);
-
    *rctx = (struct radeon_winsys_ctx *)ctx;
    return VK_SUCCESS;
 
-fail_map:
-   ws->base.buffer_destroy(&ws->base, ctx->fence_bo);
 fail_alloc:
    amdgpu_cs_ctx_free(ctx->ctx);
 fail_create:
diff --git a/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.h b/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.h
index a901afbf5d8..e66ae3e8e4e 100644
--- a/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.h
+++ b/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.h
@@ -41,7 +41,6 @@ enum { MAX_RINGS_PER_TYPE = 8 };
 
 struct radv_amdgpu_fence {
    struct amdgpu_cs_fence fence;
-   volatile uint64_t *user_ptr;
 };
 
 struct radv_amdgpu_ctx {
@@ -50,7 +49,6 @@ struct radv_amdgpu_ctx {
    struct radv_amdgpu_fence last_submission[AMDGPU_HW_IP_DMA + 1][MAX_RINGS_PER_TYPE];
 
    struct radeon_winsys_bo *fence_bo;
-   uint64_t *fence_map;
 };
 
 static inline struct radv_amdgpu_ctx *