[PATCH i-g-t] tests/amdgpu: fix secure bounce test failure on APUs

[vitaly prosyak]

The changed looks good to me.

Reviewed-by: Vitaly Prosyak <vitaly.prosyak at amd.com>

On 2024-08-14 22:01, Tim Huang wrote:
> Some APUs have address tweak secure mechanisms that encrypt
> destination data preventing it from matching the original data.
>
> Rework the test case to ensure compatibility with these APUs.
>
> Signed-off-by: Tim Huang <tim.huang at amd.com>
> ---
>  tests/amdgpu/amd_security.c | 28 ++++++----------------------
>  1 file changed, 6 insertions(+), 22 deletions(-)
>
> diff --git a/tests/amdgpu/amd_security.c b/tests/amdgpu/amd_security.c
> index ec6883d3d..024cadc05 100644
> --- a/tests/amdgpu/amd_security.c
> +++ b/tests/amdgpu/amd_security.c
> @@ -147,8 +147,6 @@ amdgpu_bo_move(amdgpu_device_handle device, int fd,
>   */
>  static const uint8_t secure_pattern[] = { 0x5A, 0xFE, 0x05, 0xEC };
>  
> -static const uint8_t secure_pattern2[] = { 0xDE, 0xAD, 0xBE, 0xEF };
> -
>  static void
>  amdgpu_secure_bounce(amdgpu_device_handle device_handle, int fd,
>  		struct drm_amdgpu_info_hw_ip  *sdma_info,
> @@ -209,27 +207,13 @@ amdgpu_secure_bounce(amdgpu_device_handle device_handle, int fd,
>  	amdgpu_bo_lcopy(device_handle, ring_context, ip_block, SECURE_BUFFER_SIZE,
>  			secure == true ? 1 : 0);
>  
> -	/* Verify the contents of Bob. */
> -	for (pp = (__typeof__(pp))ring_context->bo2_cpu;
> -	     pp < (__typeof__(pp)) ring_context->bo2_cpu + SECURE_BUFFER_SIZE;
> -	     pp += sizeof(secure_pattern)) {
> -		r = memcmp(pp, secure_pattern, sizeof(secure_pattern));
> -		if (r) {
> -			// test failure
> -			igt_assert(false);
> -			break;
> -		}
> -	}
> -	/* Fill Bob with a pattern2 */
> -	for (pp = (__typeof__(pp))ring_context->bo2_cpu;
> -		pp < (__typeof__(pp)) ring_context->bo2_cpu + SECURE_BUFFER_SIZE;
> -		pp += sizeof(secure_pattern2))
> -		memcpy(pp, secure_pattern2, sizeof(secure_pattern2));
> -
>  	/* Move Bob to the GTT domain. */
>  	amdgpu_bo_move(device_handle, fd, ring_context, ip_block,
>  			AMDGPU_GEM_DOMAIN_GTT, 0);
>  
> +	/* Clean Alice first before do the copy from bob. */
> +	memset((void *)ring_context->bo_cpu, 0, SECURE_BUFFER_SIZE);
> +
>  	/* sDMA TMZ copy from Bob to Alice.
>  	 * bo is a source ,bo2 is destination
>  	 */
> @@ -250,11 +234,11 @@ amdgpu_secure_bounce(amdgpu_device_handle device_handle, int fd,
>  	amdgpu_bo_lcopy(device_handle, ring_context, ip_block, SECURE_BUFFER_SIZE,
>  			secure == true ? 1 : 0);
>  
> -	/* Verify the content of Alice if it matches to pattern2*/
> +	/* Verify the content of Alice if it matches to pattern */
>  	for (pp = (__typeof__(pp))ring_context->bo_cpu;
>  	     pp < (__typeof__(pp)) ring_context->bo_cpu + SECURE_BUFFER_SIZE;
> -	     pp += sizeof(secure_pattern2)) {
> -		r = memcmp(pp, secure_pattern2, sizeof(secure_pattern2));
> +	     pp += sizeof(secure_pattern)) {
> +		r = memcmp(pp, secure_pattern, sizeof(secure_pattern));
>  		if (r) {
>  			// test failure
>  			igt_assert(false);