Create individual crypto callbacks for hkdf extract and hkdf expand by twcook86 · Pull Request #10598 · wolfSSL/wolfssl

twcook86 · 2026-06-04T16:41:57Z

Description

Fixes zd#21919

Testing

built-in test.c expanded

Checklist

added tests
updated/added doxygen
updated appropriate READMEs
Updated manual and documentation

Copilot

Pull request overview

This PR refines the crypto-callback (WOLF_CRYPTO_CB) KDF interface by splitting HKDF into distinct callback operations for HKDF-Extract and HKDF-Expand, and updates library internals and tests accordingly.

Changes:

Introduces new KDF callback types and wc_CryptoInfo payloads for HKDF extract/expand.
Adds new internal crypto-callback dispatch helpers (wc_CryptoCb_Hkdf_Extract / wc_CryptoCb_Hkdf_Expand) and wires them into HKDF code paths.
Expands HKDF tests to exercise extract+expand flows in addition to one-shot HKDF.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
wolfssl/wolfcrypt/types.h	Adds new `wc_KdfType` values for HKDF extract/expand.
wolfssl/wolfcrypt/cryptocb.h	Extends `wc_CryptoInfo` KDF union with HKDF extract/expand structs and declares new callback helpers.
wolfcrypt/src/hmac.c	Routes HKDF Extract/Expand through crypto callbacks when available.
wolfcrypt/src/cryptocb.c	Implements new crypto-callback helper functions and updates debug strings for KDF type.
wolfcrypt/test/test.c	Adds HKDF extract/expand test coverage and updates test crypto-device callback handling.

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

    WC_KDF_TYPE_NONE = 0,
    WC_KDF_TYPE_HKDF = 1,
-    WC_KDF_TYPE_TWOSTEP_CMAC = 2 /* NIST SP 800-56C two-step cmac kdf. */
-    /* Future: WC_KDF_TYPE_PBKDF2 = 3, WC_KDF_TYPE_SCRYPT = 4, etc. */
+    WC_KDF_TYPE_HKDF_EXTRACT = 2,
+    WC_KDF_TYPE_HKDF_EXPAND = 3,
+    WC_KDF_TYPE_TWOSTEP_CMAC = 4 /* NIST SP 800-56C two-step cmac kdf. */
+    /* Future: WC_KDF_TYPE_PBKDF2 = 5, WC_KDF_TYPE_SCRYPT = 6, etc. */


+#ifdef WOLF_CRYPTO_CB
+        /* Try crypto callback first for complete operation */
+        if (devId != INVALID_DEVID) {
+             ret = wc_CryptoCb_Hkdf_Expand(type, inKey, inKeySz, info, infoSz,
+                                           out, outSz, devId);


bigbrett

Would prefer you append-only to public structs, but its fine as is since it is only used by CryptoCbs and those aren't ABI restricted

bigbrett · 2026-06-04T17:08:57Z

+    WC_KDF_TYPE_HKDF_EXTRACT = 2,
+    WC_KDF_TYPE_HKDF_EXPAND = 3,


In general, I would try to not modify existing public enum values once they are declared, just in case they are used by customers.

In this case I think it is fine as I believe it is ONLY used internally. But something to keep in mind going forward

wolfSSL-Fenrir-bot

Fenrir Automated Review — PR #10598

Scan targets checked: wolfcrypt-bugs, wolfcrypt-port-bugs, wolfcrypt-rs-bugs, wolfcrypt-src, wolfssl-bugs, wolfssl-src

Findings: 3
3 finding(s) posted as inline comments (see file-level comments below)

This review was generated automatically by Fenrir. Findings are non-blocking.

wolfSSL-Fenrir-bot · 2026-06-04T17:30:39Z

-    /* Future: WC_KDF_TYPE_PBKDF2 = 3, WC_KDF_TYPE_SCRYPT = 4, etc. */
+    WC_KDF_TYPE_HKDF_EXTRACT = 2,
+    WC_KDF_TYPE_HKDF_EXPAND = 3,
+    WC_KDF_TYPE_TWOSTEP_CMAC = 4 /* NIST SP 800-56C two-step cmac kdf. */


🟠 [Medium] Existing KDF enum value is renumbered · API contract violations

WC_KDF_TYPE_TWOSTEP_CMAC changes from value 2 to 4, so existing crypto callback implementations that dispatch on the public enum value stop recognizing two-step CMAC KDF requests.

Fix: Keep WC_KDF_TYPE_TWOSTEP_CMAC at 2 and assign the new HKDF split values after existing enum constants.

wolfSSL-Fenrir-bot · 2026-06-04T17:30:39Z

+#ifdef WOLF_CRYPTO_CB
+        /* Try crypto callback first for complete operation */
+        if (devId != INVALID_DEVID) {
+             ret = wc_CryptoCb_Hkdf_Expand(type, inKey, inKeySz, info, infoSz,


🟠 [Medium] HKDF expand callback bypasses argument validation · Incorrect error handling

wc_HKDF_Expand_ex() dispatches to the new expand callback before the existing out == NULL and 255-block length checks, so invalid calls can bypass BAD_FUNC_ARG or reach a callback with invalid buffers.

Fix: Move the callback dispatch after wc_HmacSizeByType() and the out/length validation.

wolfSSL-Fenrir-bot · 2026-06-04T17:30:40Z

-    /* Future: WC_KDF_TYPE_PBKDF2 = 3, WC_KDF_TYPE_SCRYPT = 4, etc. */
+    WC_KDF_TYPE_HKDF_EXTRACT = 2,
+    WC_KDF_TYPE_HKDF_EXPAND = 3,
+    WC_KDF_TYPE_TWOSTEP_CMAC = 4 /* NIST SP 800-56C two-step cmac kdf. */


🔵 [Low] Public CMAC KDF callback type value changed · API contract violations

WC_KDF_TYPE_TWOSTEP_CMAC changed from value 2 to 4, so existing crypto callbacks built against the public KDF enum no longer recognize CMAC KDF requests sent by wc_CryptoCb_Kdf_TwostepCmac.

Fix: Keep WC_KDF_TYPE_TWOSTEP_CMAC = 2 and assign the new HKDF split values after it.

github-actions · 2026-06-04T17:35:18Z

billphipps

Please move the check for cryptocb into the hdkf functions, not in the side that is using it. Check on unit testing the interface as well. Looks like a great start!

billphipps · 2026-06-04T17:31:55Z

+#ifdef WOLF_CRYPTO_CB
+        /* Try crypto callback first */
+        if (devId != INVALID_DEVID) {
+             ret = wc_CryptoCb_Hkdf_Extract(type, salt, saltSz, inKey, inKeySz,
+                                            out, devId);
+            if (ret != WC_NO_ERR_TRACE(CRYPTOCB_UNAVAILABLE))
+                return ret;
+        }
+#endif
+


This should be moved to the wc_Hkdf_Extract public function so ANY potential invocation will trigger this check/offload. This means that if someone invokes the one-shot, it will try to offload the one-shot. If the one-shot is not offloaded, then the each of the extract/expand will be subsequently attempted to be offloaded since the SW version of the one-shot invokes the Expand/Extract directly. This allows a customer to invoke the one-shot separately since it may optimize better.

billphipps · 2026-06-04T17:33:04Z

+int wc_CryptoCb_Hkdf_Extract(int hashType, const byte* salt, word32 saltSz,
+                        const byte* inKey, word32 inKeySz, byte* out, int devId)


Are all of these parameters necessary? Please confirm. Fill in NULL/0 as necessary in the shared cryptoInfo structure below.

billphipps · 2026-06-04T17:33:27Z

+int wc_CryptoCb_Hkdf_Expand(int hashType, const byte* inKey, word32 inKeySz,
+                     const byte* info, word32 infoSz, byte* out, word32 outSz,
+                     int devId)


Samme comment about if all of these are necessary

billphipps · 2026-06-04T18:46:15Z

Consider looking at the test/swdev series of tests to support testing of the cryptocb plumbing.

Create individual crypto callbacks for hkdf extract and hkdf expand

2e9fe0e

twcook86 requested a review from billphipps June 4, 2026 16:45

twcook86 assigned billphipps and wolfSSL-Bot and unassigned billphipps Jun 4, 2026

bigbrett requested review from Copilot and wolfSSL-Fenrir-bot June 4, 2026 16:56

Copilot started reviewing on behalf of bigbrett June 4, 2026 16:56 View session

Copilot AI reviewed Jun 4, 2026

View reviewed changes

bigbrett approved these changes Jun 4, 2026

View reviewed changes

wolfSSL-Fenrir-bot reviewed Jun 4, 2026

View reviewed changes

billphipps requested changes Jun 4, 2026

View reviewed changes

		int wc_CryptoCb_Hkdf_Extract(int hashType, const byte* salt, word32 saltSz,
		const byte* inKey, word32 inKeySz, byte* out, int devId)

Conversation

twcook86 commented Jun 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Testing

Checklist

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

bigbrett left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

wolfSSL-Fenrir-bot left a comment

Choose a reason for hiding this comment

Fenrir Automated Review — PR #10598

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Jun 4, 2026

MemBrowse Memory Report

Uh oh!

billphipps left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

twcook86 commented Jun 4, 2026 •

edited

Loading