SE050 and CryptoCell are supported with this.
WOLFSSL_NO_RSA_SW to disable the RSA software implementation.
WOLFSSL_NO_ECC_SW to disable the ECC software implementation.
- fsp.ld Removed
- README additional description to place the RTT control block at a fixed
- renesas_fspsm_aes.c : WOLFSSL_MAX_32BIT use, fixes indent
- .gitignore Added fsp.ld in RA6M4 projects
Replace USE_FAST_MATH/TFM_TIMING_RESISTANT with SP math, sized for the
fixed key sizes this board uses (RSA 2048, ECC P-256, Curve25519,
FFDHE 2048): WOLFSSL_SP_MATH + WOLFSSL_HAVE_SP_RSA/DH/ECC, with
WOLFSSL_SP_ARM_CORTEX_M_ASM for the Cortex-M33 assembly path and
WOLFSSL_SP_SMALL/WOLFSSL_SP_NO_MALLOC/WOLFSSL_SP_NO_DYN_STACK/
WOLFSSL_SP_NO_3072 to keep it small and heap/VLA-free for this
embedded target.
Pass 1 only matched the exact "/*#define X*/" form, but the checked-in
wolfssl_demo.h spells BENCHMARK's line with spaces ("/* #define
BENCHMARK */"). That left -Mode bench unable to enable BENCHMARK, and
any other mode's pass 2 matched "#define BENCHMARK" inside the spaced
comment and produced an invalid nested comment ("/* /*#define
BENCHMARK*/ */"), breaking the header. Use a whitespace-tolerant regex
instead of a literal match.
build.bat: %ERRORLEVEL% inside a parenthesized if-block is expanded at
parse time, before the PowerShell call runs, so it silently ignored a
failed set_demo_mode.ps1. Use "if errorlevel 1", matching the same fix
already applied to the clean step below.
debug_run.bat: QC_DELAY_MS was read into a variable and shown in
status messages, but the generated J-Link command file never actually
emitted a "sleep" command between "g" and "qc" -- the probe
disconnected immediately instead of waiting. Emit "sleep !QC_DELAY_MS!"
in both the restart and reload command files.
user_settings.h: enable GCM_TABLE_4BIT (faster software GHASH) and
default to TLS 1.3 instead of requiring a manual edit to try it.
test_main.c: guser_PKCbInfo was only declared in the #else branch of
TLS_MULTITHREAD_TEST, but sce_test()'s CRYPT_TEST/BENCHMARK branches
use it unconditionally -- building with TLS_MULTITHREAD_TEST and
CRYPT_TEST/BENCHMARK together left it undeclared. Declare it
unconditionally; guser_PKCbInfo_taskA/taskB stay under
TLS_MULTITHREAD_TEST since only the TLS_CLIENT branch's per-task
connections use those.
Rewrite the SEGGER_RTT setup steps in README.md to add the
.txt.rtt_block section to the persistent script/fsp.ld instead of the
auto-generated fsp_gen.ld, which gets overwritten every time the FSP
Configurator regenerates project content. Note that build.bat now
prints the RTT control block address after a build.
Add script/fsp.ld for both the wolfssl and test projects (the
project's linker config looks for fsp.ld under ${ProjName}/script,
which didn't exist there before); test/script/fsp.ld additionally
defines the .txt.rtt_block section.
Add an Appendix section documenting build.bat and debug_run.bat usage.
The portable aesgcm_misaligned_test() (wolfcrypt/test/test.c) sets a
plaintext key, so it never reaches wc_fspsm_AesGcmEncrypt()/Decrypt()
-- wc_fspsm_AesCipher() only dispatches to the SCE hardware path once
keyflgs_crypt.bits.aes*_installedkey_set is set by SCE_KeyGeneration().
Add sce_aesgcm_misaligned_test(), following the existing
sce_aesgcm128_test()/sce_aesgcm256_test() wrapped-key pattern, to
actually exercise the hardware with misaligned AAD/data at every byte
offset within a block, for both AES-128 and AES-256. Verified passing
on RA6M4 hardware.
ca_cert_der_sign and ca_ecc_cert_der_sign are regenerated output of the
SCE-signing tool (tools/example_keys/generate_SignedCA.*) against the
current CA certs, so they no longer match the checked-in bytes.
Also drop the stray `static` on sizeof_ca_ecc_cert_der_sign so its
linkage matches sizeof_ca_cert_der_sign right above it; nothing in
this tree references either symbol from another translation unit yet,
but the RSA/ECC pair should behave the same way once something does.
The network stack (PHY link, ARP for the gateway) isn't necessarily
ready the instant a connection task starts, so a connect attempt made
too early fails with ETIMEDOUT(-116) or ENOTCONN(-128). The
non-multithreaded caller in test_main.c already retries around its own
wolfSSL_TLS_client_do() call for exactly this; TLS_MULTITHREAD_TEST's
per-task connections had no equivalent, so they'd fail outright instead
of retrying. Close and recreate the socket each attempt, since its
state after a failed connect isn't reusable.
FreeRTOS's TaskFunction_t is "void (*)(void*)", but wolfSSL_TLS_client_do()
was declared/defined to always return int, and the non-multithread caller
in test_main.c uses that return code. Under TLS_MULTITHREAD_TEST, make it
return void instead (it already ends itself via vTaskDelete(NULL) rather
than returning a value there), matching TaskFunction_t exactly; keep the
int-returning definition for the non-multithread, direct-call path.
wc_GetTime()/time() resolves to ARM semihosting (_gettimeofday, from
librdimon, via --specs=rdimon.specs) on this board, which has no RTC.
Semihosting blocks forever unless an active debugger is servicing the
call -- fine under an e2studio GUI debug session, but hangs a target
flashed and run standalone (e.g. via debug_run.bat). asn_test()'s
wc_GetTime() call in CRYPT_TEST, and peer certificate date validation
in TLS_CLIENT, both hit this.
Register a wc_SetTimeCb() callback that derives a plausible "now" from
the build date/time (__DATE__/__TIME__), the same approach already
used in IDE/Renesas/e2studio/RX65N/GR-ROSE/common/wolfssl_dummy.c, so
it stays roughly current as the project keeps getting rebuilt instead
of a hand-maintained literal. Replace with a real RTC-backed callback
if wall-clock time is ever needed (e.g. real certificate expiry
checks).
- build.bat / debug_run.bat: build and flash+run the RA6M4 wolfssl/test
e2studio projects from cmd.exe/PowerShell, mirroring the RX72N
EnvisionKit Simple demo's existing batch tooling. debug_run.bat drives
a real SEGGER J-Link via JLink Commander (unlike RX72N's RFP-based
workaround) and prints the _SEGGER_RTT control block address after a
successful build to save manually locating it for RTT Viewer.
- set_demo_mode.ps1: switch common/wolfssl_demo.h between CRYPT_TEST /
BENCHMARK / TLS_CLIENT, matching the RX72N demo's mode switcher.
- tools/example_keys/generate_SignedCA.bat: Windows wrapper for
generate_SignedCA.sh via Git for Windows' bash, with fixed RSA/ECC
arguments.
- tools/README.md: document the above and correct a stale script name.
- include.am: list the new files for source distribution.
Inside the "if defined FORCE_WOLFSSL_REBUILD (...)" block, %ERRORLEVEL%
was expanded once when the block was parsed -- before "%MAKE%" clean had
even run -- so it always read the value from before the block (0),
letting a failed clean go undetected and the build continue from a
partially cleaned tree. Verified by reproducing with a stand-in "make"
that fails on "clean": the old check silently proceeded to build "all"
regardless; "if errorlevel 1" (tested at the current point, not
text-substituted at parse time) plus a literal exit code catch it
correctly.
wc_tsip_AesGcmEncrypt() allowed sz == 0 and authInSz == 0 in its own
argument validation, but then unconditionally allocated plainBuf (sized
sz) and aadBuf (sized authInSz) via XMALLOC and treated a NULL result as
an allocation failure. XMALLOC(0, ...) is implementation-defined and may
legitimately return NULL, so a valid empty payload or empty AAD could
fail depending on allocator behavior rather than on the actual inputs.
wc_tsip_AesGcmDecrypt() has the same problem for authInSz == 0 (its
sz == 0 case is already rejected by validation, by design, so it never
reaches a zero-size allocation there).
Skip allocating (and later copying into) plainBuf/aadBuf when their
size is 0, and only require them non-NULL in that case, so the
zero-length case no longer depends on what the platform's allocator
returns for a zero-byte request.
Add tsip_aesgcm_zerolen_test() to wolfssl_tsip_unit_test.c covering
empty payload with non-empty AAD (encrypt-only, since TSIP decrypt
rejects sz == 0 by design), non-empty payload with empty AAD
(full encrypt/decrypt round trip), and both empty together.
Generated makefiles call BusyBox sed/rm directly, and edits to shared
headers like user_settings.h don't trigger incremental rebuilds, so
build.bat needed both a fix and a fast path to force-rebuild just the
wolfSSL-dependent sources.
It was missing from the project's linked resources, so it never got
compiled; wolfssl.rcpc (the e2studio project record) picks up the same
addition plus a couple of unrelated toolchain/build-option updates it
already carried.
