[openssl/openssl] c3832d: Fix a no-ecx failure in test_tls13hrr

[Matt Caswell]

  Branch: refs/heads/master
  Home:   https://github.com/openssl/openssl
  Commit: c3832d79db0c8dc08935d86200d427f945deb92e
      https://github.com/openssl/openssl/commit/c3832d79db0c8dc08935d86200d427f945deb92e
  Author: Matt Caswell <matt at openssl.org>
  Date:   2023-07-06 (Thu, 06 Jul 2023)

  Changed paths:
    M test/recipes/70-test_tls13hrr.t

  Log Message:
  -----------
  Fix a no-ecx failure in test_tls13hrr

In the case of no-ecx test 3 in test_tls13hrr was failing because it was
setting the server side support groups to on P-256 in order to induce an
HRR. However with no-ecx the client insteads issues an initial key share
using P-256 anyway and so an HRR is not used. We swap to use P-384 instead.

Reviewed-by: Hugo Landau <hlandau at openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove at arm.com>
(Merged from https://github.com/openssl/openssl/pull/21372)


  Commit: 47ef3b9fc03c92f6a387927f1b8b0fe1014eb646
      https://github.com/openssl/openssl/commit/47ef3b9fc03c92f6a387927f1b8b0fe1014eb646
  Author: Matt Caswell <matt at openssl.org>
  Date:   2023-07-06 (Thu, 06 Jul 2023)

  Changed paths:
    M test/quicapitest.c

  Log Message:
  -----------
  Don't run the ssl trace test if no-ecx

no-ecx causes SSL_trace to give different output. The test compares
the output to a reference sample - so we disable it in the case of no-ecx.

Reviewed-by: Hugo Landau <hlandau at openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove at arm.com>
(Merged from https://github.com/openssl/openssl/pull/21372)


Compare: https://github.com/openssl/openssl/compare/f9fcc7c727f0...47ef3b9fc03c