Vorago demo by dgarske · Pull Request #683 · wolfSSL/wolfBoot

dgarske · 2026-02-10T16:54:48Z

Fix for tools/scripts/va416x0/build_test.sh portability (MacOS)
Added option to support wolfCrypt test/benchmark in test-app
Add some checking if partition size is too large

Copilot

Pull request overview

This PR enhances the Vorago VA416x0 platform support by improving script portability for macOS, adding wolfCrypt test/benchmark capabilities to the test application, and implementing partition size validation.

Changes:

Fixed JLinkExe detection for macOS compatibility in the build script
Added wolfCrypt test and benchmark support to test-app with necessary infrastructure
Implemented partition size checking to prevent oversized images

Reviewed changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tools/test.mk	Updated ML_DSA size limit for test
tools/scripts/va416x0/build_test.sh	Added macOS JLinkExe detection and partition size validation
tools/keytools/sign.py	Added partition size overflow checking
tools/keytools/sign.c	Added partition size overflow checking
test-app/wolfcrypt_support.c	New file providing time and RNG functions for wolfCrypt
test-app/syscalls.c	New file providing newlib syscall stubs for embedded systems
test-app/app_va416x0.c	Added wolfCrypt test/benchmark integration
test-app/Makefile	Added build support for wolfCrypt test/benchmark
test-app/ARM-va416x0.ld	Increased heap and stack sizes
src/update_flash.c	Fixed MAX_UPDATE_SIZE calculation to include header size
include/user_settings.h	Added wolfCrypt test/benchmark configuration options
config/examples/vorago_va416x0.config	Updated partition layout and optimization settings
Makefile	Added WOLFBOOT_PARTITION_SIZE to sign tool environment

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

Copilot · 2026-02-24T22:26:42Z

tools/scripts/va416x0/build_test.sh

+    JLINK_PATH=$(find /Applications/SEGGER -name "JLinkExe" 2>/dev/null | head -n1)
+    if [ -n "$JLINK_PATH" ] && [ -x "$JLINK_PATH" ]; then
+        JLINK="$JLINK_PATH"
+    else
+        echo "Error: JLinkExe not found. Please install SEGGER J-Link software."


The error message on line 28 doesn't account for the case where /Applications/SEGGER doesn't exist. The find command will fail silently but the error message will still say 'JLinkExe not found' which could be confusing. Consider adding a check for the directory's existence first or improving the error message to indicate both possible failure modes.

Suggested change

JLINK_PATH=$(find /Applications/SEGGER -name "JLinkExe" 2>/dev/null | head -n1)

if [ -n "$JLINK_PATH" ] && [ -x "$JLINK_PATH" ]; then

JLINK="$JLINK_PATH"

else

echo "Error: JLinkExe not found. Please install SEGGER J-Link software."

if [ -d /Applications/SEGGER ]; then

JLINK_PATH=$(find /Applications/SEGGER -name "JLinkExe" 2>/dev/null | head -n1)

if [ -n "$JLINK_PATH" ] && [ -x "$JLINK_PATH" ]; then

JLINK="$JLINK_PATH"

else

echo "Error: JLinkExe not found in PATH or under /Applications/SEGGER. Please install SEGGER J-Link software."

exit 1

fi

else

echo "Error: JLinkExe not found in PATH and /Applications/SEGGER does not exist. Please install SEGGER J-Link software."

Copilot · 2026-02-24T22:26:42Z

test-app/syscalls.c

+    prev_heap_end = heap_end;
+
+    /* Calculate heap limit */
+    heap_limit = &_end + _Min_Heap_Size;


Pointer arithmetic error: _Min_Heap_Size is an extern char (treated as pointer), not an integer value. This should cast _Min_Heap_Size to an integer type (e.g., (int)&_Min_Heap_Size) to correctly calculate the heap limit.

Copilot · 2026-02-24T22:26:42Z

test-app/syscalls.c

+
+void _exit(int status)
+{
+    while(1);


Missing space between 'while' and '(1)'. Should be 'while (1);' for consistency with standard C formatting conventions.

Suggested change

while(1);

while (1);

Copilot · 2026-02-24T22:26:42Z

src/update_flash.c

 #endif

-/* Reserve space for two sectors in case of NVM_FLASH_WRITEONCE, for redundancy */
+/* Max firmware size: partition must hold header + fw + trailer sector(s) */


The comment on line 647 is now misleading. It says 'partition must hold header + fw + trailer sector(s)' but the calculation subtracts both IMAGE_HEADER_SIZE and WOLFBOOT_SECTOR_SIZE from WOLFBOOT_PARTITION_SIZE, meaning MAX_UPDATE_SIZE represents only the firmware size, not including the header. Update the comment to clarify that MAX_UPDATE_SIZE is the maximum firmware size (excluding header).

Suggested change

/* Max firmware size: partition must hold header + fw + trailer sector(s) */

/* Max firmware size (excluding header); partition must also hold header +

* trailer sector(s) */

Copilot

Pull request overview

Copilot reviewed 14 out of 14 changed files in this pull request and generated 8 comments.

Comments suppressed due to low confidence (1)

tools/keytools/sign.py:97

Inside make_header(), assignments to WOLFBOOT_PARTITION_SIZE / WOLFBOOT_SECTOR_SIZE create locals unless declared global, so the later size check will still see the module-level values (currently 0) and never run. Declare these as global in the function (or refactor to pass them through) so the partition-size guard actually takes effect.

def make_header(image_file, fw_version, extra_fields=[]):

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

Copilot · 2026-02-25T22:27:47Z

tools/scripts/va416x0/build_test.sh

+if command -v JLinkExe &> /dev/null; then
+    JLINK="JLinkExe"


&> /dev/null is not POSIX-sh compatible (it’s a bash-ism). If this script is run under /bin/sh (common on macOS/Linux), this can fail. Prefer >/dev/null 2>&1, or explicitly set the shebang to bash if bash is required.

Copilot · 2026-02-25T22:27:48Z

tools/keytools/sign.py

+        if "WOLFBOOT_PARTITION_SIZE" in l and "ADDRESS" not in l:
+            val=l.split('=')[1].rstrip('\n')
+            WOLFBOOT_PARTITION_SIZE = int(val,0)
+        if "WOLFBOOT_SECTOR_SIZE" in l:
+            val=l.split('=')[1].rstrip('\n')
+            WOLFBOOT_SECTOR_SIZE = int(val,0)


Inside make_header(), assignments to WOLFBOOT_PARTITION_SIZE / WOLFBOOT_SECTOR_SIZE create locals unless declared global, so the later size check will still see the module-level values (currently 0) and never run. Declare these as global in the function (or refactor to pass them through) so the partition-size guard actually takes effect.

Copilot · 2026-02-25T22:27:48Z

test-app/syscalls.c

+extern char _end; /* Defined by linker */
+extern char _Min_Heap_Size; /* Defined by linker (if available) */


_Min_Heap_Size is a linker symbol value, but it’s being used as if it were a char variable (&_end + _Min_Heap_Size). This will either fail to compile (invalid operands) or read a byte from an address that is not intended to be dereferenced. Use the symbol’s address/value correctly (e.g., cast (uintptr_t)&_Min_Heap_Size as the size), or define/export an explicit heap limit symbol in the linker script and compare against that.

Copilot · 2026-02-25T22:27:48Z

test-app/syscalls.c

+    prev_heap_end = heap_end;
+
+    /* Calculate heap limit */
+    heap_limit = &_end + _Min_Heap_Size;


_Min_Heap_Size is a linker symbol value, but it’s being used as if it were a char variable (&_end + _Min_Heap_Size). This will either fail to compile (invalid operands) or read a byte from an address that is not intended to be dereferenced. Use the symbol’s address/value correctly (e.g., cast (uintptr_t)&_Min_Heap_Size as the size), or define/export an explicit heap limit symbol in the linker script and compare against that.

Copilot · 2026-02-25T22:27:48Z

test-app/Makefile

+  APP_OBJS+=syscalls.o
+


When WOLFCRYPT_SUPPORT=1, you compile/link your own syscall stubs (syscalls.o) while also linking with --specs=nosys.specs (which typically pulls in libnosys syscall stubs). This commonly causes multiple-definition link errors for _write, _sbrk, etc. Consider making --specs=nosys.specs conditional (disable it when syscalls.o is linked), or drop syscalls.o and rely on nosys (with any required overrides).

Suggested change

APP_OBJS+=syscalls.o

Copilot · 2026-02-25T22:27:49Z

test-app/Makefile

    APP_OBJS+=../src/keystore.o
  endif
+  # Reduce size: newlib-nano and section GC
+  LDFLAGS+=--specs=nano.specs --specs=nosys.specs


When WOLFCRYPT_SUPPORT=1, you compile/link your own syscall stubs (syscalls.o) while also linking with --specs=nosys.specs (which typically pulls in libnosys syscall stubs). This commonly causes multiple-definition link errors for _write, _sbrk, etc. Consider making --specs=nosys.specs conditional (disable it when syscalls.o is linked), or drop syscalls.o and rely on nosys (with any required overrides).

Suggested change

LDFLAGS+=--specs=nano.specs --specs=nosys.specs

LDFLAGS+=--specs=nano.specs

ifneq ($(WOLFCRYPT_SUPPORT),1)

# Only pull in nosys stubs when not providing our own syscalls

LDFLAGS+=--specs=nosys.specs

endif

Copilot · 2026-02-25T22:27:49Z

tools/keytools/sign.c

+            uint32_t partition_sz = (uint32_t)strtol(env_psize, NULL, 0);
+            uint32_t sector_sz = env_ssize ?
+                (uint32_t)strtol(env_ssize, NULL, 0) : 0;
+            uint32_t total_img_sz = CMD.header_sz + image_sz;
+            /* Only subtract sector for trailer when sector < partition.
+             * When sector >= partition (e.g. update_ram targets), the
+             * entire partition is available for the image. */
+            uint32_t max_img_sz = (sector_sz < partition_sz) ?
+                (partition_sz - sector_sz) : partition_sz;
+            if (total_img_sz > max_img_sz) {
+                printf("Error: Image size %u (header %u + firmware %u) "
+                    "exceeds max %u (partition %u - sector %u)\n",
+                    total_img_sz, CMD.header_sz, image_sz,
+                    max_img_sz, partition_sz, sector_sz);
+                goto failure;
+            }


The strtol(..., NULL, 0) conversions ignore parse errors and allow negative values, which then get cast to uint32_t and can produce very large sizes. This can make the size check behave incorrectly. Use an endptr and validate that the full string was parsed, and reject non-positive / out-of-range values (fail early with a clear error).

Suggested change

uint32_t partition_sz = (uint32_t)strtol(env_psize, NULL, 0);

uint32_t sector_sz = env_ssize ?

(uint32_t)strtol(env_ssize, NULL, 0) : 0;

uint32_t total_img_sz = CMD.header_sz + image_sz;

/* Only subtract sector for trailer when sector < partition.

* When sector >= partition (e.g. update_ram targets), the

* entire partition is available for the image. */

uint32_t max_img_sz = (sector_sz < partition_sz) ?

(partition_sz - sector_sz) : partition_sz;

if (total_img_sz > max_img_sz) {

printf("Error: Image size %u (header %u + firmware %u) "

"exceeds max %u (partition %u - sector %u)\n",

total_img_sz, CMD.header_sz, image_sz,

max_img_sz, partition_sz, sector_sz);

goto failure;

}

char *endptr;

unsigned long tmp_ul;

uint32_t partition_sz;

uint32_t sector_sz = 0;

errno = 0;

tmp_ul = strtoul(env_psize, &endptr, 0);

if (endptr == env_psize || *endptr != '\0' || errno == ERANGE ||

tmp_ul == 0 || tmp_ul > UINT32_MAX) {

printf("Error: Invalid WOLFBOOT_PARTITION_SIZE value '%s'\n",

env_psize);

goto failure;

}

partition_sz = (uint32_t)tmp_ul;

if (env_ssize) {

errno = 0;

tmp_ul = strtoul(env_ssize, &endptr, 0);

if (endptr == env_ssize || *endptr != '\0' || errno == ERANGE ||

tmp_ul == 0 || tmp_ul > UINT32_MAX) {

printf("Error: Invalid WOLFBOOT_SECTOR_SIZE value '%s'\n",

env_ssize);

goto failure;

}

sector_sz = (uint32_t)tmp_ul;

}

{

uint32_t total_img_sz = CMD.header_sz + image_sz;

/* Only subtract sector for trailer when sector < partition.

* When sector >= partition (e.g. update_ram targets), the

* entire partition is available for the image. */

uint32_t max_img_sz = (sector_sz < partition_sz) ?

(partition_sz - sector_sz) : partition_sz;

if (total_img_sz > max_img_sz) {

printf("Error: Image size %u (header %u + firmware %u) "

"exceeds max %u (partition %u - sector %u)\n",

total_img_sz, CMD.header_sz, image_sz,

max_img_sz, partition_sz, sector_sz);

goto failure;

}

}

Copilot · 2026-02-25T22:27:49Z

include/user_settings.h

+    #if defined(WOLFCRYPT_TEST) || defined(WOLFCRYPT_BENCHMARK)
+        /* Use custom RNG for tests (saves ~7KB vs HASHDRBG) */
+        #define WC_NO_HASHDRBG
+        #define CUSTOM_RAND_GENERATE_SEED my_rng_seed_gen
+        #define CUSTOM_RAND_GENERATE_BLOCK my_rng_seed_gen
+        extern int my_rng_seed_gen(unsigned char* output, unsigned int sz);
+    #else


This wires a deterministic, non-cryptographic RNG into wolfCrypt whenever WOLFCRYPT_TEST/WOLFCRYPT_BENCHMARK is enabled. That’s fine for tests, but it’s a sharp edge if these flags are accidentally enabled outside test builds. Consider adding a stronger compile-time safeguard (e.g., require an explicit ALLOW_INSECURE_TEST_RNG macro, or emit a preprocessor error unless a dedicated test configuration is selected) to prevent shipping builds with this RNG.

dgarske self-assigned this Feb 10, 2026

dgarske force-pushed the vorago_demo branch 3 times, most recently from b1c0e3b to a2d630f Compare February 10, 2026 20:40

dgarske marked this pull request as ready for review February 11, 2026 01:28

dgarske requested a review from danielinux February 11, 2026 01:28

dgarske assigned danielinux and unassigned dgarske Feb 11, 2026

danielinux previously approved these changes Feb 24, 2026

View reviewed changes

danielinux assigned dgarske and unassigned danielinux Feb 24, 2026

dgarske added 3 commits February 24, 2026 14:20

Fixes for portability

8ce6ff2

Add option to enable wolfCrypt test/benchmark to test-app

b6c22aa

Add some checking if partition size is too large

5910923

Copilot AI review requested due to automatic review settings February 24, 2026 22:21

dgarske dismissed danielinux’s stale review via 5910923 February 24, 2026 22:21

dgarske force-pushed the vorago_demo branch from a2d630f to 5910923 Compare February 24, 2026 22:21

dgarske assigned danielinux and wolfSSL-Bot and unassigned dgarske Feb 24, 2026

dgarske requested review from Copilot and removed request for Copilot February 24, 2026 22:21

dgarske self-assigned this Feb 24, 2026

Copilot AI reviewed Feb 24, 2026

View reviewed changes

Update submodule

9af0051

Copilot started reviewing on behalf of dgarske February 24, 2026 22:49 View session

Copilot started reviewing on behalf of dgarske February 24, 2026 23:03 View session

Further improvements to support benchmark

a220564

dgarske unassigned danielinux and wolfSSL-Bot Feb 25, 2026

Peer review fixes

1e96682

Copilot AI review requested due to automatic review settings February 25, 2026 22:21

Copilot AI reviewed Feb 25, 2026

View reviewed changes

	/* Max firmware size: partition must hold header + fw + trailer sector(s) */
	/* Max firmware size (excluding header); partition must also hold header +
	* trailer sector(s) */

		extern char _end; /* Defined by linker */
		extern char _Min_Heap_Size; /* Defined by linker (if available) */

-  LDFLAGS+=--specs=nano.specs --specs=nosys.specs
+  LDFLAGS+=--specs=nano.specs
+  ifneq ($(WOLFCRYPT_SUPPORT),1)
+    # Only pull in nosys stubs when not providing our own syscalls
+    LDFLAGS+=--specs=nosys.specs
+  endif

Conversation

dgarske commented Feb 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Feb 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 25, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

dgarske commented Feb 10, 2026 •

edited

Loading