patch 'test/debug: fix crash with mlx5 devices' has been queued to stable release 22.11.11

[luca.boccassi at gmail.com]

Hi,

FYI, your patch has been queued to stable release 22.11.11

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 10/29/25. So please
shout if anyone has objections.

Also note that after the patch there's a diff of the upstream commit vs the
patch applied to the branch. This will indicate if there was any rebasing
needed to apply to the stable branch. If there were code changes for rebasing
(ie: not only metadata diffs), please double check that the rebase was
correctly done.

Queued patches are on a temporary branch at:
https://github.com/bluca/dpdk-stable

This queued commit can be viewed at:
https://github.com/bluca/dpdk-stable/commit/ec185479ebb5c833bca7da93731fed2760b57ba7

Thanks.

Luca Boccassi

---
>From ec185479ebb5c833bca7da93731fed2760b57ba7 Mon Sep 17 00:00:00 2001
From: David Marchand <david.marchand at redhat.com>
Date: Thu, 2 Oct 2025 17:36:50 +0200
Subject: [PATCH] test/debug: fix crash with mlx5 devices

[ upstream commit 2b403dd8fb37d0ba13723e44ffc7ee2c2795f838 ]

Running rte_exit() in a forked process means that shared memory will be
released by the child process before the parent process does the same.
This issue has been seen recently when some GHA virtual machine (with
some mlx5 devices) runs the debug_autotest unit test.

Instead, run rte_panic() and rte_exit() from a new DPDK process spawned
like for other recursive unit tests.

Bugzilla ID: 1796
Fixes: af75078fece3 ("first public release")

Signed-off-by: David Marchand <david.marchand at redhat.com>
Acked-by: Bruce Richardson <bruce.richardson at intel.com>
Acked-by: Dariusz Sosnowski <dsosnowski at nvidia.com>
---
 app/test/process.h    |  2 +-
 app/test/test.c       |  2 +
 app/test/test.h       |  2 +
 app/test/test_debug.c | 92 ++++++++++++++++++++++++++++++-------------
 4 files changed, 69 insertions(+), 29 deletions(-)

diff --git a/app/test/process.h b/app/test/process.h
index e8e7e5ab60..610d657c2e 100644
--- a/app/test/process.h
+++ b/app/test/process.h
@@ -203,7 +203,7 @@ process_dup(const char *const argv[], int numargs, const char *env_value)
  * tests attempting to use this function on FreeBSD.
  */
 #ifdef RTE_EXEC_ENV_LINUX
-static char *
+static inline char *
 get_current_prefix(char *prefix, int size)
 {
 	char path[PATH_MAX] = {0};
diff --git a/app/test/test.c b/app/test/test.c
index 5cf9f51c28..02cdf44fc8 100644
--- a/app/test/test.c
+++ b/app/test/test.c
@@ -82,6 +82,8 @@ do_recursive_call(void)
 			{ "test_memory_flags", no_action },
 			{ "test_file_prefix", no_action },
 			{ "test_no_huge_flag", no_action },
+			{ "test_panic", test_panic },
+			{ "test_exit", test_exit },
 #ifdef RTE_LIB_TIMER
 #ifndef RTE_EXEC_ENV_WINDOWS
 			{ "timer_secondary_spawn_wait", test_timer_secondary },
diff --git a/app/test/test.h b/app/test/test.h
index 6a4fa0b1d7..4142c22c1d 100644
--- a/app/test/test.h
+++ b/app/test/test.h
@@ -173,7 +173,9 @@ extern const char *prgname;
 int commands_init(void);
 int command_valid(const char *cmd);
 
+int test_exit(void);
 int test_mp_secondary(void);
+int test_panic(void);
 int test_timer_secondary(void);
 
 int test_set_rxtx_conf(cmdline_fixed_string_t mode);
diff --git a/app/test/test_debug.c b/app/test/test_debug.c
index 2704f5b927..f016e2f8f7 100644
--- a/app/test/test_debug.c
+++ b/app/test/test_debug.c
@@ -8,6 +8,18 @@
 #include <stdint.h>
 
 #ifdef RTE_EXEC_ENV_WINDOWS
+int
+test_panic(void)
+{
+	printf("debug not supported on Windows, skipping test\n");
+	return TEST_SKIPPED;
+}
+int
+test_exit(void)
+{
+	printf("debug not supported on Windows, skipping test\n");
+	return TEST_SKIPPED;
+}
 static int
 test_debug(void)
 {
@@ -25,34 +37,31 @@ test_debug(void)
 #include <rte_debug.h>
 #include <rte_common.h>
 #include <rte_eal.h>
-#include <rte_service_component.h>
+#include <rte_lcore.h>
+
+#include "process.h"
 
 /*
  * Debug test
  * ==========
  */
 
-/* use fork() to test rte_panic() */
-static int
+static const char *test_args[7];
+
+int
 test_panic(void)
 {
-	int pid;
 	int status;
 
-	pid = fork();
-
-	if (pid == 0) {
+	if (getenv(RECURSIVE_ENV_VAR) != NULL) {
 		struct rlimit rl;
 
 		/* No need to generate a coredump when panicking. */
 		rl.rlim_cur = rl.rlim_max = 0;
 		setrlimit(RLIMIT_CORE, &rl);
 		rte_panic("Test Debug\n");
-	} else if (pid < 0) {
-		printf("Fork Failed\n");
-		return -1;
 	}
-	wait(&status);
+	status = process_dup(test_args, RTE_DIM(test_args), "test_panic");
 	if(status == 0){
 		printf("Child process terminated normally!\n");
 		return -1;
@@ -62,27 +71,16 @@ test_panic(void)
 	return 0;
 }
 
-/* use fork() to test rte_exit() */
 static int
 test_exit_val(int exit_val)
 {
-	int pid;
+	char buf[5];
 	int status;
 
-	/* manually cleanup EAL memory, as the fork() below would otherwise
-	 * cause the same hugepages to be free()-ed multiple times.
-	 */
-	rte_service_finalize();
-
-	pid = fork();
-
-	if (pid == 0)
-		rte_exit(exit_val, __func__);
-	else if (pid < 0){
-		printf("Fork Failed\n");
-		return -1;
-	}
-	wait(&status);
+	sprintf(buf, "%d", exit_val);
+	if (setenv("TEST_DEBUG_EXIT_VAL", buf, 1) == -1)
+		rte_panic("Failed to set exit value in env\n");
+	status = process_dup(test_args, RTE_DIM(test_args), "test_exit");
 	printf("Child process status: %d\n", status);
 	if(!WIFEXITED(status) || WEXITSTATUS(status) != (uint8_t)exit_val){
 		printf("Child process terminated with incorrect status (expected = %d)!\n",
@@ -92,11 +90,22 @@ test_exit_val(int exit_val)
 	return 0;
 }
 
-static int
+int
 test_exit(void)
 {
 	int test_vals[] = { 0, 1, 2, 255, -1 };
 	unsigned i;
+
+	if (getenv(RECURSIVE_ENV_VAR) != NULL) {
+		int exit_val;
+
+		if (!getenv("TEST_DEBUG_EXIT_VAL"))
+			rte_panic("No exit value set in env\n");
+
+		exit_val = strtol(getenv("TEST_DEBUG_EXIT_VAL"), NULL, 0);
+		rte_exit(exit_val, __func__);
+	}
+
 	for (i = 0; i < RTE_DIM(test_vals); i++) {
 		if (test_exit_val(test_vals[i]) < 0)
 			return -1;
@@ -128,6 +137,33 @@ test_usage(void)
 static int
 test_debug(void)
 {
+#ifdef RTE_EXEC_ENV_FREEBSD
+	/* BSD target doesn't support prefixes at this point, and we also need to
+	 * run another primary process here.
+	 */
+	const char * prefix = "--no-shconf";
+#else
+	const char * prefix = "--file-prefix=debug";
+#endif
+	char core[10];
+
+	sprintf(core, "%d", rte_get_main_lcore());
+
+	test_args[0] = prgname;
+	test_args[1] = prefix;
+	test_args[2] = "-l";
+	test_args[3] = core;
+
+	if (rte_eal_has_hugepages()) {
+		test_args[4] = "";
+		test_args[5] = "";
+		test_args[6] = "";
+	} else {
+		test_args[4] = "--no-huge";
+		test_args[5] = "-m";
+		test_args[6] = "2048";
+	}
+
 	rte_dump_stack();
 	if (test_panic() < 0)
 		return -1;
-- 
2.47.3

---
  Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- -	2025-10-27 15:54:36.247535374 +0000
+++ 0038-test-debug-fix-crash-with-mlx5-devices.patch	2025-10-27 15:54:34.807949850 +0000
@@ -1 +1 @@
-From 2b403dd8fb37d0ba13723e44ffc7ee2c2795f838 Mon Sep 17 00:00:00 2001
+From ec185479ebb5c833bca7da93731fed2760b57ba7 Mon Sep 17 00:00:00 2001
@@ -5,0 +6,2 @@
+[ upstream commit 2b403dd8fb37d0ba13723e44ffc7ee2c2795f838 ]
+
@@ -16 +17,0 @@
-Cc: stable at dpdk.org
@@ -29 +30 @@
-index 9fb2bf481c..8e11d0b059 100644
+index e8e7e5ab60..610d657c2e 100644
@@ -42 +43 @@
-index fd653cbbfd..8a4598baee 100644
+index 5cf9f51c28..02cdf44fc8 100644
@@ -45 +46 @@
-@@ -80,6 +80,8 @@ do_recursive_call(void)
+@@ -82,6 +82,8 @@ do_recursive_call(void)
@@ -55 +56 @@
-index ebc4864bf8..c6d7d23313 100644
+index 6a4fa0b1d7..4142c22c1d 100644
@@ -58 +59 @@
-@@ -174,7 +174,9 @@ extern const char *prgname;
+@@ -173,7 +173,9 @@ extern const char *prgname;
@@ -69 +70 @@
-index 8ad6d40fcb..fe5dd5b02d 100644
+index 2704f5b927..f016e2f8f7 100644