wolfCrypt SRAM PUF: configurable error correction and footprint

[dgarske]

Summary

Makes the wolfCrypt SRAM PUF configurable instead of fixed at BCH(127,64,t=10) with 16 codewords. The default build is unchanged and byte-for-byte compatible with prior releases (identical identity, derived key, and helper data).

Features

• WC_PUF_BCH_T selects the BCH error-correction profile: t=7, t=10 (default), t=13, t=15. Higher t corrects more bit flips per 127-bit codeword; unsupported values are a compile-time error.
• WC_PUF_NUM_CODEWORDS (default 16) trades SRAM footprint and helper-data size against derived-key entropy.
• wc_PufGetParams() and WC_PUF_PROFILE_ID let an application persist the profile with its helper data and detect an enroll/reconstruct build mismatch.
• Profile selection via ./configure --enable-puf=small|balanced|strong and CMake -DWOLFSSL_PUF_PROFILE=...; other values (and codeword count) via CPPFLAGS.

Testing

Profile-parametrized unit tests, plus on-silicon regression across STM32 Cortex-M0+/M4/M7/M33 boards (see the wolfssl-examples-stm32 companion PR).

[Copilot]

Pull request overview

This PR makes the wolfCrypt SRAM PUF BCH(127,k,t) fuzzy-extractor configurable (t-profile and number of codewords) while keeping the default build byte-for-byte compatible with prior releases. It also adds a way for applications to query/persist the selected profile to detect enrollment/reconstruction mismatches.

Changes:

• Add configurable PUF parameters (WC_PUF_BCH_T, WC_PUF_NUM_CODEWORDS), derived sizing macros, and a WC_PUF_PROFILE_ID/wc_PufGetParams() API for mismatch detection.
• Generalize BCH encode/decode implementation to work across supported (t,k) profiles and adjust stable/helper data packing accordingly.
• Update unit tests, build-system toggles (Autoconf/CMake), and Doxygen docs for profile selection and sizing.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
wolfssl/wolfcrypt/puf.h	Adds configuration knobs, derived sizing macros, profile fingerprint, and new public wc_PufGetParams() declaration.
wolfcrypt/src/puf.c	Implements generalized BCH encode/decode for selectable profiles and adds wc_PufGetParams() implementation.
wolfcrypt/test/test.c	Updates PUF unit test to be profile-parameterized (dynamic SRAM fill, noise injection, new failure-mode expectations).
scripts/puf_bch_genpoly.py	Adds a generator script for BCH generator polynomials/KAT vectors used by puf.c.
doc/dox_comments/header_files/puf.h	Updates public documentation/examples to reflect configurable sizing and the new params API.
configure.ac	Adds `--enable-puf=small
CMakeLists.txt	Adds WOLFSSL_PUF_PROFILE cache option and maps to WC_PUF_BCH_T definitions when PUF is enabled.

Suppressed comments (1)

wolfcrypt/test/test.c:24126

• The Test 4b temporary tooNoisySram buffer is also allocated on the stack and scales with WC_PUF_RAW_BYTES. This should follow the same small-stack pattern as the other PUF test buffers.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

MemBrowse Memory Report

No memory changes detected for:

• gcc-arm-cortex-m0plus
• gcc-arm-cortex-m3
• gcc-arm-cortex-m4
• gcc-arm-cortex-m4-baremetal
• gcc-arm-cortex-m4-crypto-only
• gcc-arm-cortex-m4-dtls13
• gcc-arm-cortex-m4-min-ecc
• gcc-arm-cortex-m4-openssl-compat
• gcc-arm-cortex-m4-pkcs7
• gcc-arm-cortex-m4-pq
• gcc-arm-cortex-m4-rsa-only
• gcc-arm-cortex-m4-sp-math
• gcc-arm-cortex-m4-tls12
• gcc-arm-cortex-m4-tls13
• gcc-arm-cortex-m7
• gcc-arm-cortex-m7-pq
• gcc-arm-cortex-m7-tls13
• linuxkm-pie
• linuxkm-standard
• stm32-sim-stm32h753

[dgarske]

Jenkins retest this please

[Copilot]

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated no new comments.

Suppressed comments (2)

wolfssl/wolfcrypt/puf.h:43

• The build-flag documentation here is out of sync with the actual build-system support added in this PR: configure.ac/README/.github/workflows/CMakeLists all accept a “strongest” profile (t=15), but this header only documents small|balanced|strong. This can mislead users into thinking t=15 isn’t supported via configure/CMake.

    Build: ./configure --enable-puf[=small|balanced|strong] (auto-enables
    HKDF). CMake: -DWOLFSSL_PUF=yes with
    -DWOLFSSL_PUF_PROFILE=small|balanced|strong.

CMakeLists.txt:2323

• CMake currently has no supported way to set WC_PUF_NUM_CODEWORDS such that BOTH (1) the library is compiled with the override and (2) the generated wolfssl/options.h records it for separately-built applications. The comment says it can be set via CFLAGS, but CMake’s options.h generation only reflects macros present in WOLFSSL_DEFINITIONS (see the foreach(DEF IN LISTS WOLFSSL_DEFINITIONS) block later in this file), so a CFLAGS-only override risks an ABI/config mismatch and thus a silently wrong key/profile ID at the application level.

    # PUF BCH error-correction profile (only surfaced when PUF is enabled).
    # Other t values and WC_PUF_NUM_CODEWORDS can still be set via CFLAGS.
    add_option("WOLFSSL_PUF_PROFILE"
        "PUF BCH profile: balanced (t=10), small (t=7), strong (t=13), strongest (t=15)"
        "balanced" "balanced;small;strong;strongest")

[dgarske]

Jenkins retest this please