From 359feb0d2f2c409a215e969485d9b0311215eccc Mon Sep 17 00:00:00 2001 From: Tom Cosgrove Date: Thu, 15 Sep 2022 14:52:34 +0100 Subject: [PATCH] Better wording for the reason why we use an input MPI for a scalar value Signed-off-by: Tom Cosgrove --- tests/suites/test_suite_mpi.function | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/tests/suites/test_suite_mpi.function b/tests/suites/test_suite_mpi.function index 0a31f9576b..71912b1769 100644 --- a/tests/suites/test_suite_mpi.function +++ b/tests/suites/test_suite_mpi.function @@ -1929,7 +1929,9 @@ void mpi_core_mla( char * input_A, char * input_B, char * input_S, { /* We are testing A += B * s; A, B are MPIs, s is a scalar. * - * However, we encode s as an MPI in the .data file for ease of handling. + * However, we encode s as an MPI in the .data file as the test framework + * currently only supports `int`-typed scalars, and that doesn't cover the + * full range of `mbedtls_mpi_uint`. * * We also have the different results for sizeof(mbedtls_mpi_uint) == 4 or 8. */