Rebase

[AravindanNC]

No description provided.

[Copilot]

Pull request overview

Copilot reviewed 233 out of 332 changed files in this pull request and generated 7 comments.

Comments suppressed due to low confidence (2)

recipes-kernel/linux-libc-headers/files_516/0001-idle_metrics_v2.patch:1

• This patch adds non-upstream PRCTL constants and large custom structs directly into include/uapi/linux/prctl.h. In particular, hardcoding NUM_CPUS 4 in a UAPI header is not ABI-safe across targets and will break consumers on systems with different CPU counts. Consider moving these definitions into a separate UAPI header, using a flexible array/size field instead of a fixed CPU count, and ensuring PR_* numbers do not collide with upstream allocations.

--- a/include/uapi/linux/prctl.h	2025-12-16 15:50:41.761927991 +0000

recipes-connectivity/socat/files/socat_restrict_localhost.patch:1

• IPv6 loopback literals are typically passed as ::1 (without brackets); using \"[::1]\" will likely reject valid loopback usage. Also, calling exit(1) inside an address-resolution helper is a hard process abort that bypasses normal error propagation. Accept ::1 (and optionally localhost if desired) and return an error through the existing error-handling path instead of printing and exiting.

Date: Apr 27 16:25:26 2018

---

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

[Copilot]

pipefail is not POSIX and will fail under common /bin/sh implementations (e.g., dash). Either switch the shebang to /bin/bash (and ensure bash is in RDEPENDS) or remove/replace the pipefail usage.

Suggested change

	set -o pipefail
	cd tests || exit 1
	# Quick functional tests only - skip long-running stress tests
	cd tests || exit 1
	# Quick functional tests only - skip long-running stress tests
	# Quick functional tests only - skip long-running stress tests

[Copilot]

The conditional on line 22 is invalid ([ ! -name \"*.sh\" ] is not a valid test expression), and the runner invokes every test via bash, which will break ELF binaries picked up by -perm -u+x. Update the logic to (1) correctly detect .sh scripts vs binaries and (2) execute binaries directly instead of forcing bash.

[Copilot]

The script prints FAIL but never returns a non-zero exit status when a test fails, so ptest will report success even with failures. Track a failure flag/counter and exit 1 if any test fails.

[Copilot]

FAIL_COUNT is never incremented, so the summary will always report 0 failures even for genuine failures. Either increment FAIL_COUNT for mismatches (and optionally only SKIP for known/expected missing-feature cases), or remove FAIL_COUNT entirely to avoid misleading output. Also consider exiting non-zero when failures occur so ptest can fail correctly.

[Copilot]

FAIL_COUNT is never incremented, so the summary will always report 0 failures even for genuine failures. Either increment FAIL_COUNT for mismatches (and optionally only SKIP for known/expected missing-feature cases), or remove FAIL_COUNT entirely to avoid misleading output. Also consider exiting non-zero when failures occur so ptest can fail correctly.

[Copilot]

The declared project version (1.18.0), the shared library VERSION (1.0.0), and the pkg-config version (18.1.0) are inconsistent. Align these values or derive them from a single source of truth to avoid producing mismatched artifacts (.so metadata vs pkg-config version).

[Copilot]

The declared project version (1.18.0), the shared library VERSION (1.0.0), and the pkg-config version (18.1.0) are inconsistent. Align these values or derive them from a single source of truth to avoid producing mismatched artifacts (.so metadata vs pkg-config version).

Suggested change

	VERSION 1.0.0 SOVERSION 1)
	VERSION ${PROJECT_VERSION} SOVERSION ${PROJECT_VERSION_MAJOR})

[Copilot]

Pull request overview

Copilot reviewed 233 out of 332 changed files in this pull request and generated 7 comments.

---

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

[Copilot]

This hardcodes prefix/libdir/include paths (and forces lib even on multilib systems like lib64), which can install pkg-config metadata and libraries into the wrong locations. Use CMake/GNUInstallDirs variables (CMAKE_INSTALL_PREFIX, CMAKE_INSTALL_LIBDIR, CMAKE_INSTALL_INCLUDEDIR) and generate the .pc file from those values to respect the Yocto toolchain configuration.

[Copilot]

set -o pipefail is not POSIX and will fail under many /bin/sh implementations (e.g., BusyBox ash), causing the ptest to abort immediately. Either switch the shebang to /bin/bash (and ensure bash is in RDEPENDS) or remove pipefail and handle pipeline errors explicitly.

[Copilot]

This script has unquoted expansions and legacy backticks, which can break if the path contains whitespace or special characters. Use $(...) and quote $0, $workdir, and the cd target to make the runner robust.

Suggested change

	workdir=$(dirname `realpath $0`)
	cd ${workdir}/testing
	workdir=$(dirname "$(realpath "$0")")
	cd "${workdir}/testing"

[Copilot]

This patch introduces new prctl() numbers, hardcodes NUM_CPUS, and adds typedefs/structs directly into a UAPI header (include/uapi/linux/prctl.h). UAPI headers are part of the stable ABI surface: hardcoding CPU counts and adding new PR_* numbers without ensuring they are aligned with upstream/kernel-side definitions risks ABI conflicts and user-space breakage. A safer approach is to (a) keep these definitions in a separate, project-specific UAPI header, or (b) ensure the kernel implementation and UAPI definitions are upstream-aligned and do not reuse/overlap existing PR_* values.

Suggested change

	@@ -272,4 +272,73 @@
	# define PR_SCHED_CORE_SCOPE_THREAD_GROUP 1
	# define PR_SCHED_CORE_SCOPE_PROCESS_GROUP 2
	+#define PR_SET_FTRACE_MARKER 99
	+
	+/* sets max size of string buffer */
	+#define FTRACE_BUF_SIZE 1024
	+
	+
	+#define NUM_CPUS 4
	+
	+/* define structure(s) used for logging idle metrics */
	+typedef struct {
	+ __u64 metricStartTime;
	+ __u64 metricEndTime;
	+ __u64 idle_time[NUM_CPUS];
	+} IDLE_METRICS;
	+
	+typedef struct {
	+ __u64 sum_idle_time;
	+ __u32 max_idle_time;
	+ __u32 idle_count;
	+} PER_CPU_IDLE_METRIC;
	+
	+/* define structure(s) used for logging idle metrics */
	+typedef struct {
	+ __u64 metricStartTime;
	+ __u64 metricEndTime;
	+ PER_CPU_IDLE_METRIC idle[NUM_CPUS];
	+ __u64 sum_all_cpus_running_time;
	+ __u64 sum_run_time_lt_1ms;
	+ __u64 sum_run_time_gt_1ms;
	+ __u64 sum_run_time_gt_5ms;
	+ __u64 sum_run_time_gt_10ms;
	+ __u64 sum_run_time_gt_20ms;
	+ __u64 sum_run_time_gt_30ms;
	+ __u64 sum_run_time_gt_40ms;
	+ __u64 sum_run_time_gt_50ms;
	+ __u64 sum_run_time_gt_75ms;
	+ __u64 sum_run_time_gt_100ms;
	+ __u32 max_run_time_lt_1ms;
	+ __u32 max_run_time_gt_1ms;
	+ __u32 max_run_time_gt_5ms;
	+ __u32 max_run_time_gt_10ms;
	+ __u32 max_run_time_gt_20ms;
	+ __u32 max_run_time_gt_30ms;
	+ __u32 max_run_time_gt_40ms;
	+ __u32 max_run_time_gt_50ms;
	+ __u32 max_run_time_gt_75ms;
	+ __u32 max_run_time_gt_100ms;
	+ __u32 max_all_cpus_running_time;
	+ __u32 count;
	+ __u32 run_time_gt_1ms;
	+ __u32 run_time_gt_5ms;
	+ __u32 run_time_gt_10ms;
	+ __u32 run_time_gt_20ms;
	+ __u32 run_time_gt_30ms;
	+ __u32 run_time_gt_40ms;
	+ __u32 run_time_gt_50ms;
	+ __u32 run_time_gt_75ms;
	+ __u32 run_time_gt_100ms;
	+} IDLE_METRICS_V2;
	+
	+#define PR_GET_IDLE_METRICS 104
	+#define PR_GET_IDLE_METRICS_VERSION 105
	+
	+typedef enum {
	+ IDLE_METRICS_VERSION_V1 = 0,
	+ IDLE_METRICS_VERSION_V2,
	+ IDLE_METRICS_VERSION_NUM_VERSIONS
	+} IDLE_METRICS_VERSION;
	+
	@@ -272,4 +272,4 @@
	# define PR_SCHED_CORE_SCOPE_THREAD_GROUP 1
	# define PR_SCHED_CORE_SCOPE_PROCESS_GROUP 2

[Copilot]

This patch introduces new prctl() numbers, hardcodes NUM_CPUS, and adds typedefs/structs directly into a UAPI header (include/uapi/linux/prctl.h). UAPI headers are part of the stable ABI surface: hardcoding CPU counts and adding new PR_* numbers without ensuring they are aligned with upstream/kernel-side definitions risks ABI conflicts and user-space breakage. A safer approach is to (a) keep these definitions in a separate, project-specific UAPI header, or (b) ensure the kernel implementation and UAPI definitions are upstream-aligned and do not reuse/overlap existing PR_* values.

Suggested change

	@@ -272,4 +272,73 @@
	# define PR_SCHED_CORE_SCOPE_THREAD_GROUP 1
	# define PR_SCHED_CORE_SCOPE_PROCESS_GROUP 2
	+#define PR_SET_FTRACE_MARKER 99
	+
	+/* sets max size of string buffer */
	+#define FTRACE_BUF_SIZE 1024
	+
	+
	+#define NUM_CPUS 4
	+
	+/* define structure(s) used for logging idle metrics */
	+typedef struct {
	+ __u64 metricStartTime;
	+ __u64 metricEndTime;
	+ __u64 idle_time[NUM_CPUS];
	+} IDLE_METRICS;
	+
	+typedef struct {
	+ __u64 sum_idle_time;
	+ __u32 max_idle_time;
	+ __u32 idle_count;
	+} PER_CPU_IDLE_METRIC;
	+
	+/* define structure(s) used for logging idle metrics */
	+typedef struct {
	+ __u64 metricStartTime;
	+ __u64 metricEndTime;
	+ PER_CPU_IDLE_METRIC idle[NUM_CPUS];
	+ __u64 sum_all_cpus_running_time;
	+ __u64 sum_run_time_lt_1ms;
	+ __u64 sum_run_time_gt_1ms;
	+ __u64 sum_run_time_gt_5ms;
	+ __u64 sum_run_time_gt_10ms;
	+ __u64 sum_run_time_gt_20ms;
	+ __u64 sum_run_time_gt_30ms;
	+ __u64 sum_run_time_gt_40ms;
	+ __u64 sum_run_time_gt_50ms;
	+ __u64 sum_run_time_gt_75ms;
	+ __u64 sum_run_time_gt_100ms;
	+ __u32 max_run_time_lt_1ms;
	+ __u32 max_run_time_gt_1ms;
	+ __u32 max_run_time_gt_5ms;
	+ __u32 max_run_time_gt_10ms;
	+ __u32 max_run_time_gt_20ms;
	+ __u32 max_run_time_gt_30ms;
	+ __u32 max_run_time_gt_40ms;
	+ __u32 max_run_time_gt_50ms;
	+ __u32 max_run_time_gt_75ms;
	+ __u32 max_run_time_gt_100ms;
	+ __u32 max_all_cpus_running_time;
	+ __u32 count;
	+ __u32 run_time_gt_1ms;
	+ __u32 run_time_gt_5ms;
	+ __u32 run_time_gt_10ms;
	+ __u32 run_time_gt_20ms;
	+ __u32 run_time_gt_30ms;
	+ __u32 run_time_gt_40ms;
	+ __u32 run_time_gt_50ms;
	+ __u32 run_time_gt_75ms;
	+ __u32 run_time_gt_100ms;
	+} IDLE_METRICS_V2;
	+
	+#define PR_GET_IDLE_METRICS 104
	+#define PR_GET_IDLE_METRICS_VERSION 105
	+
	+typedef enum {
	+ IDLE_METRICS_VERSION_V1 = 0,
	+ IDLE_METRICS_VERSION_V2,
	+ IDLE_METRICS_VERSION_NUM_VERSIONS
	+} IDLE_METRICS_VERSION;
	+
	@@ -272,4 +272,4 @@
	# define PR_SCHED_CORE_SCOPE_THREAD_GROUP 1
	# define PR_SCHED_CORE_SCOPE_PROCESS_GROUP 2

[Copilot]

This patch introduces new prctl() numbers, hardcodes NUM_CPUS, and adds typedefs/structs directly into a UAPI header (include/uapi/linux/prctl.h). UAPI headers are part of the stable ABI surface: hardcoding CPU counts and adding new PR_* numbers without ensuring they are aligned with upstream/kernel-side definitions risks ABI conflicts and user-space breakage. A safer approach is to (a) keep these definitions in a separate, project-specific UAPI header, or (b) ensure the kernel implementation and UAPI definitions are upstream-aligned and do not reuse/overlap existing PR_* values.

Suggested change

	@@ -272,4 +272,73 @@
	# define PR_SCHED_CORE_SCOPE_THREAD_GROUP 1
	# define PR_SCHED_CORE_SCOPE_PROCESS_GROUP 2
	+#define PR_SET_FTRACE_MARKER 99
	+
	+/* sets max size of string buffer */
	+#define FTRACE_BUF_SIZE 1024
	+
	+
	+#define NUM_CPUS 4
	+
	+/* define structure(s) used for logging idle metrics */
	+typedef struct {
	+ __u64 metricStartTime;
	+ __u64 metricEndTime;
	+ __u64 idle_time[NUM_CPUS];
	+} IDLE_METRICS;
	+
	+typedef struct {
	+ __u64 sum_idle_time;
	+ __u32 max_idle_time;
	+ __u32 idle_count;
	+} PER_CPU_IDLE_METRIC;
	+
	+/* define structure(s) used for logging idle metrics */
	+typedef struct {
	+ __u64 metricStartTime;
	+ __u64 metricEndTime;
	+ PER_CPU_IDLE_METRIC idle[NUM_CPUS];
	+ __u64 sum_all_cpus_running_time;
	+ __u64 sum_run_time_lt_1ms;
	+ __u64 sum_run_time_gt_1ms;
	+ __u64 sum_run_time_gt_5ms;
	+ __u64 sum_run_time_gt_10ms;
	+ __u64 sum_run_time_gt_20ms;
	+ __u64 sum_run_time_gt_30ms;
	+ __u64 sum_run_time_gt_40ms;
	+ __u64 sum_run_time_gt_50ms;
	+ __u64 sum_run_time_gt_75ms;
	+ __u64 sum_run_time_gt_100ms;
	+ __u32 max_run_time_lt_1ms;
	+ __u32 max_run_time_gt_1ms;
	+ __u32 max_run_time_gt_5ms;
	+ __u32 max_run_time_gt_10ms;
	+ __u32 max_run_time_gt_20ms;
	+ __u32 max_run_time_gt_30ms;
	+ __u32 max_run_time_gt_40ms;
	+ __u32 max_run_time_gt_50ms;
	+ __u32 max_run_time_gt_75ms;
	+ __u32 max_run_time_gt_100ms;
	+ __u32 max_all_cpus_running_time;
	+ __u32 count;
	+ __u32 run_time_gt_1ms;
	+ __u32 run_time_gt_5ms;
	+ __u32 run_time_gt_10ms;
	+ __u32 run_time_gt_20ms;
	+ __u32 run_time_gt_30ms;
	+ __u32 run_time_gt_40ms;
	+ __u32 run_time_gt_50ms;
	+ __u32 run_time_gt_75ms;
	+ __u32 run_time_gt_100ms;
	+} IDLE_METRICS_V2;
	+
	+#define PR_GET_IDLE_METRICS 104
	+#define PR_GET_IDLE_METRICS_VERSION 105
	+
	+typedef enum {
	+ IDLE_METRICS_VERSION_V1 = 0,
	+ IDLE_METRICS_VERSION_V2,
	+ IDLE_METRICS_VERSION_NUM_VERSIONS
	+} IDLE_METRICS_VERSION;
	+
	@@ -272,4 +272,4 @@
	# define PR_SCHED_CORE_SCOPE_THREAD_GROUP 1
	# define PR_SCHED_CORE_SCOPE_PROCESS_GROUP 2

[Copilot]

FAIL_COUNT is never incremented and any unexpected return code is treated as SKIP, which can mask genuine regressions (e.g., a test that should pass but fails for functional reasons). Consider treating mismatches as FAIL by default, and only SKIP when you can positively detect an environmental reason (missing capability/module) or when the expected rc indicates skip semantics.

Suggested change

	# Treat as SKIP - likely missing kernel modules (ip6_tables, nat, mangle, etc.)
	echo "SKIP: $test_name (expected $rc_spec got $rc_got)"
	SKIP_COUNT=$((SKIP_COUNT + 1))
	# Treat unexpected return codes as FAIL by default.
	# Only explicitly detected environmental conditions should be reported as SKIP.
	echo "FAIL: $test_name (expected $rc_spec got $rc_got)"
	FAIL_COUNT=$((FAIL_COUNT + 1))