Memfault Firmware SDK 1.24.0 (Build 13888)

Author: Memfault Inc

CHANGELOG.md

@@ -6,7 +6,96 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to
 [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [1.23.1] - 2024-04-17
+## [1.24.0] - 2025-05-06
+
+### 📈 Added
+
+- General:
+
+  - Add a new API, `memfault_log_get_unsent_count()`, which returns the log
+    count and size in bytes of unsent logs in the log buffer. This can be used
+    inside `memfault_log_handle_saved_callback()` for example to drain the
+    packetized logs when a certain watermark is reached.
+
+- ESP-IDF:
+
+  - Add a Kconfig option, `CONFIG_MEMFAULT_TIME_SINCE_BOOT_CUSTOM`, which when
+    set `=n`, enables using a custom implementation of
+    `memfault_platform_get_time_since_boot_ms()`.
+
+  - Add 2 new metrics for tracking raw network bytes rx / tx. These metrics
+    track network IO traffic on the default netif:
+
+    - `network_rx_bytes`
+    - `network_tx_bytes`
+
+    These metrics are enabled by default, and can be disabled with
+    `CONFIG_MEMFAULT_METRICS_NETWORK_IO=n`
+
+### 🛠️ Changed
+
+- FreeRTOS (including ESP-IDF):
+
+  - Rename the thread stack usage measurement variable included in coredumps
+    when `MEMFAULT_COREDUMP_COMPUTE_THREAD_STACK_USAGE` is enabled from
+    `high_watermark` to `stack_unused`. This change is to make the
+    implementation more readable. The Memfault backend is updated to process
+    both the old and new formats.
+
+  - FreeRTOS-detected stack overflows (via `vApplicationStackOverflowHook`) will
+    now be correctly tagged as `Stack Overflow` for the trace reason in
+    Memfault, instead of `Assert`.
+
+### 🐛 Fixed
+
+- Zephyr:
+
+  - Fix a null dereference when calling
+    `memfault_coredump_storage_compute_size_required()` (eg the shell command
+    `mflt coredump_size`).
+
+- General:
+
+  - For the emlib WDOG port implementation,
+    [`ports/emlib/wdog_software_watchdog.c`](ports/emlib/wdog_software_watchdog.c),
+    enable the WDOG when in EM1 mode for series2 chips.
+
+  - Added support for `MEMFAULT_REBOOT_REASON_CLEAR` in the
+    [`ports/nrf5_sdk/resetreas_reboot_tracking.c`](ports/nrf5_sdk/resetreas_reboot_tracking.c)
+    implementation, by default enabled (like other ports). This permits opting
+    out of the auto-clearing of the `NRF_POWER->RESETREAS` register, in case the
+    user needs it after the function runs.
+
+- FreeRTOS (including ESP-IDF):
+
+  - Fix incorrect computation of per-thread stack usage metrics
+    (`MEMFAULT_FREERTOS_COLLECT_THREAD_METRICS`). Before this fix, the returned
+    values had 2 errors:
+
+    1. the value is the _unused_ stack space, not the used stack space
+    2. on platforms where `sizeof(StackType_t)` is not 1 byte, the numerator
+       when computing percentage is incorrectly scaled down by
+       `sizeof(StackType_t)`, resulting in significant under-reporting of the
+       stack usage percentage.
+
+    Users can apply device and software version filtering on dashboards to
+    filter out reports from devices that are running an old version of the SDK.
+
+- ESP-IDF:
+
+  - Fix a potential issue that would cause the wrong implementation of
+    `memfault_platform_time_get_current()` to be included in the final link,
+    when `CONFIG_MEMFAULT_SYSTEM_TIME=y` is enabled (default).
+
+### 🛠️ Changed
+
+- ESP-IDF:
+
+  - `CONFIG_MEMFAULT_ENABLE_REBOOT_DIAG_DUMP` now defaults to `y` instead of
+    `n`, to print out the reboot reason on boot. Disable it with
+    `CONFIG_MEMFAULT_ENABLE_REBOOT_DIAG_DUMP=n`.
+
+## [1.23.1] - 2025-04-17
 
 ### 🐛 Fixed
 
@@ -15,7 +104,7 @@ and this project adheres to
   [CircleCI's v2 container runtime](https://discuss.circleci.com/t/docker-executor-infrastructure-upgrade/52282),
   which uses `cgroupv2` instead of `cgroupv1`.
 
-## [1.23.0] - 2024-04-16
+## [1.23.0] - 2025-04-16
 
 ### 📈 Added
 

VERSION

@@ -1,3 +1,3 @@
-BUILD ID: 13641
-GIT COMMIT: 5f0dd98704
-VERSION: 1.23.1
+BUILD ID: 13888
+GIT COMMIT: 58d5865404
+VERSION: 1.24.0

components/core/src/memfault_event_storage.c

@@ -34,9 +34,11 @@
 //  memfault/core/platform/event.h
 //
 
+#if MEMFAULT_EVENT_STORAGE_STUB_SYSTEM_TIME
 MEMFAULT_WEAK bool memfault_platform_time_get_current(MEMFAULT_UNUSED sMemfaultCurrentTime *time) {
   return false;
 }
+#endif
 
 MEMFAULT_WEAK void memfault_lock(void) { }
 

components/core/src/memfault_log_data_source.c

@@ -37,12 +37,15 @@ static bool prv_log_is_sent(uint8_t hdr) {
 
 typedef struct {
   size_t num_logs;
+  size_t bytes;
 } sMfltLogCountingCtx;
 
 static bool prv_log_iterate_counting_callback(sMfltLogIterator *iter) {
   sMfltLogCountingCtx *const ctx = (sMfltLogCountingCtx *)(iter->user_ctx);
   if (!prv_log_is_sent(iter->entry.hdr)) {
     ++ctx->num_logs;
+    // tally the size of the log message
+    ctx->bytes += iter->entry.len + sizeof(iter->entry.hdr);
   }
   return true;
 }
@@ -303,4 +306,15 @@ size_t memfault_log_data_source_count_unsent_logs(void) {
   return ctx.num_logs;
 }
 
+sMfltLogUnsentCount memfault_log_get_unsent_count(void) {
+  sMfltLogCountingCtx ctx = { 0 };
+  sMfltLogIterator iter = { .user_ctx = &ctx };
+  memfault_log_iterate(prv_log_iterate_counting_callback, &iter);
+
+  return (sMfltLogUnsentCount){
+    .num_logs = ctx.num_logs,
+    .bytes = ctx.bytes,
+  };
+}
+
 #endif /* MEMFAULT_LOG_DATA_SOURCE_ENABLED */

components/include/memfault/core/log.h

@@ -241,6 +241,14 @@ uint32_t memfault_log_get_dropped_count(void);
 //! Return the count of lines that have been written to the logging buffer.
 uint32_t memfault_log_get_recorded_count(void);
 
+typedef struct {
+  size_t num_logs;
+  size_t bytes;
+} sMfltLogUnsentCount;
+
+//! Return the number of unsent logs and their total size
+sMfltLogUnsentCount memfault_log_get_unsent_count(void);
+
 #ifdef __cplusplus
 }
 #endif

components/include/memfault/default_config.h

@@ -206,6 +206,14 @@ extern "C" {
 
 #endif /* MEMFAULT_EVENT_STORAGE_READ_BATCHING_ENABLED */
 
+//! Include a stub, weakly defined implementation of
+//! memfault_platform_time_get_current(). Most build systems will be able to
+//! override the symbol with a custom implementation at link time, but some will
+//! need to exclude the weak definition entirely.
+#ifndef MEMFAULT_EVENT_STORAGE_STUB_SYSTEM_TIME
+  #define MEMFAULT_EVENT_STORAGE_STUB_SYSTEM_TIME 1
+#endif
+
 //! The max size of a chunk. Should be a size suitable to write to transport
 //! data is being dumped over.
 #ifndef MEMFAULT_DATA_EXPORT_CHUNK_MAX_LEN

components/include/memfault/version.h

@@ -20,8 +20,8 @@ typedef struct {
 } sMfltSdkVersion;
 
 #define MEMFAULT_SDK_VERSION \
-  { .major = 1, .minor = 23, .patch = 1 }
-#define MEMFAULT_SDK_VERSION_STR "1.23.1"
+  { .major = 1, .minor = 24, .patch = 0 }
+#define MEMFAULT_SDK_VERSION_STR "1.24.0"
 
 #ifdef __cplusplus
 }

components/panics/src/memfault_fault_handling_arm.c

@@ -51,11 +51,17 @@ size_t memfault_coredump_storage_compute_size_required(void) {
     .trace_reason = kMfltRebootReason_UnknownError,
   };
 
+  const sMfltRegState exception_regs = {
+    // spoof the EXC_RETURN value as "PSP active", to avoid checking PSP
+    // register when computing size. it's unused in bare metal applications.
+    .exc_return = 1 << 2,
+  };
+
   sCoredumpCrashInfo info = {
     // we'll just pass the current stack pointer, value shouldn't matter
     .stack_address = (void *)&core_regs,
     .trace_reason = save_info.trace_reason,
-    .exception_reg_state = NULL,
+    .exception_reg_state = &exception_regs,
   };
   save_info.regions = memfault_platform_coredump_get_regions(&info, &save_info.num_regions);
 

examples/esp32/apps/memfault_demo_app/main/main.c

@@ -410,6 +410,13 @@ static void prv_load_memfault_settings_from_nvs(void) {
       (s_mflt_device_url_len > 1) ? s_mflt_device_url : NULL;
   }
 }
+
+  #if defined(CONFIG_MEMFAULT_TIME_SINCE_BOOT_CUSTOM)
+uint64_t memfault_platform_get_time_since_boot_ms(void) {
+  const int64_t time_since_boot_us = esp_timer_get_time();
+  return (uint64_t)(time_since_boot_us / 1000) /* us per ms */;
+}
+  #endif
 #endif  // defined(CONFIG_MEMFAULT)
 
 // This task started by cpu_start.c::start_cpu0_default().

examples/esp32/apps/memfault_demo_app/sdkconfig.defaults

@@ -53,9 +53,6 @@ CONFIG_MEMFAULT_HTTP_PERIODIC_UPLOAD=y
 # Upload logs. Not recommended for production.
 CONFIG_MEMFAULT_HTTP_PERIODIC_UPLOAD_LOGS=y
 
-# Pretty-print reboot reason
-CONFIG_MEMFAULT_ENABLE_REBOOT_DIAG_DUMP=y
-
 # Support older ESP32-C3 variants
 CONFIG_ESP32C3_REV_MIN_2=y
 

examples/freertos/Makefile

@@ -49,15 +49,24 @@ SIZE ?= $(shell $(CC) -print-prog-name=size)
 READELF ?= $(shell $(CC) -print-prog-name=readelf)
 
 # Check if FreeRTOS location set, add rule to download if not
+FREERTOS_DIR ?=
 ifeq ($(FREERTOS_DIR),)
 FREERTOS_DIR = FreeRTOS-Kernel
 FREERTOS_VER = V11.2.0
+FREERTOS_SHA = 0adc196d4bd52a2d91102b525b0aafc1e14a2386
 
-$(FREERTOS_DIR)/.clonedone:
-	$(info FREERTOS_DIR does not exist, downloading local copy)
+# Verify the FreeRTOS directory is present and matches expected SHA
+FREERTOS_STALE = \
+    $(shell \
+        if ! (echo "$(FREERTOS_SHA)" | diff -q $(FREERTOS_DIR)/.git/HEAD - > /dev/null 2>&1); then \
+            echo FREERTOS_STALE; \
+        fi \
+    )
+.PHONY: FREERTOS_STALE
+$(FREERTOS_DIR)/.git/HEAD: $(FREERTOS_STALE)
+	$(info FREERTOS_DIR needs updating, deleting stale copy (if present) and fetching from git)
+	rm -rf $(FREERTOS_DIR)
 	git -c advice.detachedHead=false clone --branch $(FREERTOS_VER) --depth 1 https://github.com/FreeRTOS/FreeRTOS-Kernel.git $(FREERTOS_DIR)
-	# sad check for clone completion, :(
-	touch $@
 
 .DEFAULT_GOAL :=
 endif
@@ -174,7 +183,7 @@ endif
 all: $(ELF)
 
 # Require clone to complete as prereq for sources
-$(SRCS): $(FREERTOS_DIR)/.clonedone
+$(SRCS): $(FREERTOS_DIR)/.git/HEAD
 
 # Store computed cflags in a file; it's a prerequisite for all objects. Use a
 # shell hack to check if the current cflags are different from the stored file,

examples/freertos/src/FreeRTOSConfig.h

@@ -87,7 +87,7 @@ extern void vAssertCalled(const char *file, int line);
  * to exclude the API function. */
 
 #if MEMFAULT_TEST_USE_PORT_TEMPLATE != 1
-  #define configCHECK_FOR_STACK_OVERFLOW 1
+  #define configCHECK_FOR_STACK_OVERFLOW 2
 #else
   #define configCHECK_FOR_STACK_OVERFLOW 0
 #endif

examples/freertos/src/console.c

@@ -168,6 +168,38 @@ static int prv_assert_with_reason(int argc, char *argv[]) {
   return 0;
 }
 
+static int prv_get_log_count(int argc, char *argv[]) {
+  (void)argc, (void)argv;
+
+  sMfltLogUnsentCount log_count = memfault_log_get_unsent_count();
+  printf("Number of logs: %u\nSize: %u bytes\n", (unsigned int)log_count.num_logs,
+         (unsigned int)log_count.bytes);
+
+  return 0;
+}
+
+static int prv_stack_overflow(int argc, char *argv[]) {
+  (void)argc, (void)argv;
+  TaskHandle_t task_handle = xTaskGetCurrentTaskHandle();
+  MEMFAULT_LOG_INFO("Triggering stack overflow test on task %p / %s", (void *)task_handle,
+                    pcTaskGetName(task_handle));
+  // get the current address- it should be ~ the address of the above variable,
+  // offset it by 4 words to hopefully prevent clobbering our current variables.
+  // we'll write from this address, so the stack watermark will be updated.
+  volatile uint32_t *stack_start = ((volatile uint32_t *)&task_handle) - 4;
+  volatile uint32_t *stack_current = stack_start;
+
+  // starting at the stack start address, write decrementing values. this should crash!
+  for (; stack_current > (stack_start - CONSOLE_INPUT_STACK_SIZE); stack_current--) {
+    *stack_current = 0xDEADBEEF;
+    // sleep a bit
+    vTaskDelay((10) / portTICK_PERIOD_MS);
+  }
+
+  MEMFAULT_LOG_ERROR("Stack overflow test failed, stack overflow was not triggered!");
+  return 0;
+}
+
 static const sMemfaultShellCommand s_freertos_example_shell_extension_list[] = {
   {
     .command = "freertos_vassert",
@@ -208,6 +240,16 @@ static const sMemfaultShellCommand s_freertos_example_shell_extension_list[] = {
     .command = "assert_with_reason",
     .handler = prv_assert_with_reason,
     .help = "Execute an assert with a custom reason code",
+  },
+  {
+    .command = "log_count",
+    .handler = prv_get_log_count,
+    .help = "Get the number and size of unsent logs",
+  },
+  {
+    .command = "stack_overflow",
+    .handler = prv_stack_overflow,
+    .help = "Trigger a stack overflow",
   }
 };
 #endif

examples/freertos/src/memfault/memfault_platform_port.c

@@ -85,6 +85,12 @@ void memfault_platform_log_raw(const char *fmt, ...) {
   va_end(args);
 }
 
+//! Route the 'export' command to output via printf, so we don't drop messages
+//! from logging in a big burst.
+void memfault_data_export_base64_encoded_chunk(const char *base64_chunk) {
+  printf("%s\n", base64_chunk);
+}
+
 bool memfault_platform_time_get_current(sMemfaultCurrentTime *time_output) {
 #if MEMFAULT_EXAMPLE_USE_REAL_TIME
   // Get time from time.h

ports/emlib/wdog_software_watchdog.c

@@ -62,6 +62,9 @@ static void prv_build_configuration(WDOG_Init_TypeDef *cfg, uint32_t persel) {
     .enable = true,
     // freeze watchdog when a debugger halts the system
     .debugRun = false,
+#if defined(_WDOG_CFG_EM1RUN_MASK)
+    .em1Run = true,
+#endif
     .em2Run = true,
     .em3Run = true,
     .em4Block = false,