Max Fillinger

@davidhorstmann-arm Thank you for the feedback! I could have used fewer magic numbers. I think I addressed all the comments on the code. I will add some negative tests later.

@davidhorstmann-arm I added some tests, mostly negative tests but I also added tests to check that the client and server export the same keys. One of them is currently failing:...

I fixed the problem with `mbedtls_ssl_tls13_hkdf_expand_label()`. Also, I just realized that the HKDF-Expand algorithm produces at most `255 * hash_size` bytes of output, so I changed the input validation in...

@davidhorstmann-arm I tried to fix the CI failures, please run the tests again.

I don't know what these errors mean: On **Windows-2013-Release-Win32-cmake** (and others): "A volatile slot has not been closed properly" On **all_u16-build_arm_linux_gnueabi_gcc_arm5vte** and others: ``` [2024-10-30T13:33:58.436Z] suites/test_suite_ssl.function: In function 'test_ssl_tls_exporter_rejects_bad_parameters': [2024-10-30T13:33:58.436Z]...

I think I fixed the second point in my previous comment. I used the macro `MD_OR_USE_PSA_INIT()`, which may jump to the `exit` label, before declaring and initializing the variables. I...

> A volatile slot is a key created through the PSA API, excluding persistent keys. This error occurs when a test creates a key but does not free it before...

I reproduced the error by enabling `MBEDTLS_USE_PSA_CRYPTO` and fixed it. That uncovered a new error: In the test `TLS 1.2 Keying Material Exporter: Consistent results, large keys`, I try to...

@gilles-peskine-arm said [elsewhere](https://github.com/Mbed-TLS/mbedtls/issues/9744#issuecomment-2450251375) that the number of commits in this PR is maybe too large to review. If this is purely about the number of commits and not the number...

I fixed the last four failing tests. The issue was that when running `depends.py curves` and `depends.py hashes` my new tests were segfaulting in `mbedtls_test_ssl_endpoint_free()`. I looked at other tests...