[DI] Move scripts not used by the CI

[mpg]

[Moved and re-organized from #12 and #16.]

Note: the scope (as a first step) is now only to move to the framework, not to adapt to tf-psa-crypto. Some comments and size estimates are still from the time where the scope included adapting to tf-psa-crypto.

The lists below are temporary and might still evolve.

Table legend:
I: Currently identical between 3.6 and development?
S: Work size for the move and potentially changes needed for TF-PSA-Crypto

Moving to the framework - used by tf-psa-crypto

name	I	S	Issue/PR	Comment
ecc-heap.sh	y	S		depends on benchmark.c, adapt to tf_psa_crypto_config.h
gen_ctr_drbg.pl	y	XS
gen_gcm_decrypt.pl	y	-
gen_gcm_encrypt.pl	y	-
gen_pkcs1_v21_sign_verify.pl	y	-
generate-afl-tests.sh	y	-
prepare_release.sh	y	?		do we keep it?
test_config_script.py	y	XS		changes for crypto as part of work in crypto_config.py
tmp_ignore_makefiles.sh	y	S		do we keep it? if yes, adapt to TF-PSA-Crypto tree
bump_version.sh	n	S		adapt to TF-PSA-Crypto
code_size_compare.py	n	M		adapt to mbedtls_config.h split, adapt to tf_psa_crypto_config.h

Moving to the framework - not used by tf-psa-crypto

name	I	S	Issue/PR	Comment
audit-validity-dates.py	y	XS
footprint.sh	y			do we keep it? broken, redundant with code_size_compare.py?
generate_server9_bad_saltlen.py	y	-
set_psa_test_dependencies.py	y	-

Not moving to the framework

name	I	S	Issue/PR	Comment
ecp_comb_table.py	y	-		Utility to add a new curve in ecp_curves.c, tied to ecp_curves.c code / Manuel has plan to complete rewrite this, see Mbed-TLS/mbedtls#7623
memory.sh + massif_max.pl	n			do we keep it? if yes, adapt to mbedtls_config.h split

[gilles-peskine-arm]

set_psa_test_dependencies.py is an obsolete script that was written and used during the days of migrating test dependencies test in data files. We haven't had a need for it in years and we won't. So we can remove it in both 3.6 and development.

It would be good to remove set_psa_test_dependencies.py for other reasons as well: it refers to legacy identifiers that we're removing (MBEDTLS_AES_C etc.), and we shouldn't waste time trying to figure out what to do about it — we should do nothing about it, and it would be beneficial even just for this script not to show up in git grep.