[PATCH 00/21] replace strtok with strtok_r

[PATCH 00/21] replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Jie Hai (21):
  app/graph: replace strtok with strtok_r
  app/test-bbdev: replace strtok with strtok_r
  app/test-compress-perf: replace strtok with strtok_r
  app/test-crypto-perf: replace strtok with strtok_r
  app/test-dma-perf: replace strtok with strtok_r
  app/test-fib: replace strtok with strtok_r
  app/dpdk-test-flow-perf: replace strtok with strtok_r
  app/test-mldev: replace strtok with strtok_r
  lib/dmadev: replace strtok with strtok_r
  lib/eal: replace strtok with strtok_r
  lib/ethdev: replace strtok with strtok_r
  lib/eventdev: replace strtok with strtok_r
  lib/telemetry: replace strtok with strtok_r
  lib/telemetry: replace strtok with strtok_r
  bus/fslmc: replace strtok with strtok_r
  common/cnxk: replace strtok with strtok_r
  event/cnxk: replace strtok with strtok_r
  net/ark: replace strtok with strtok_r
  raw/cnxk_gpio: replace strtok with strtok_r
  examples/l2fwd-crypto: replace strtok with strtok_r
  examples/vhost: replace strtok with strtok_r

 app/graph/graph.c                             |  5 ++-
 app/graph/utils.c                             | 15 +++++---
 app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
 .../comp_perf_options_parse.c                 | 16 ++++----
 app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
 .../cperf_test_vector_parsing.c               | 10 +++--
 app/test-dma-perf/main.c                      | 13 ++++---
 app/test-fib/main.c                           | 10 ++---
 app/test-flow-perf/main.c                     | 22 ++++++-----
 app/test-mldev/ml_options.c                   | 18 ++++-----
 drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
 drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
 drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
 drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
 drivers/net/ark/ark_pktchkr.c                 | 10 ++---
 drivers/net/ark/ark_pktgen.c                  | 10 ++---
 drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
 examples/l2fwd-crypto/main.c                  |  6 +--
 examples/vhost/main.c                         |  3 +-
 lib/dmadev/rte_dmadev.c                       |  4 +-
 lib/eal/common/eal_common_memory.c            |  8 ++--
 lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
 lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
 lib/eventdev/rte_eventdev.c                   | 18 ++++-----
 lib/security/rte_security.c                   |  3 +-
 lib/telemetry/telemetry.c                     |  5 ++-
 27 files changed, 169 insertions(+), 140 deletions(-)

-- 
2.30.0

[PATCH 01/21] app/graph: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 app/graph/graph.c |  5 +++--
 app/graph/utils.c | 15 +++++++++------
 2 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/app/graph/graph.c b/app/graph/graph.c
index a65723a196da..316950948160 100644
--- a/app/graph/graph.c
+++ b/app/graph/graph.c
@@ -103,9 +103,10 @@ parser_usecases_read(char *usecases)
 {
 	bool valid = false;
 	uint32_t i, j = 0;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(usecases, ",");
+	token = strtok_r(usecases, ",", &sp);
 	while (token != NULL) {
 		for (i = 0; i < RTE_DIM(supported_usecases); i++) {
 			if (strcmp(supported_usecases[i], token) == 0) {
@@ -116,7 +117,7 @@ parser_usecases_read(char *usecases)
 				break;
 			}
 		}
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	return valid;
diff --git a/app/graph/utils.c b/app/graph/utils.c
index c7b6ae83cf1f..ae2ff225c394 100644
--- a/app/graph/utils.c
+++ b/app/graph/utils.c
@@ -101,13 +101,14 @@ int
 parser_ip4_read(uint32_t *value, char *p)
 {
 	uint8_t shift = 24;
+	char *sp = NULL;
 	uint32_t ip = 0;
 	char *token;
 
-	token = strtok(p, ".");
+	token = strtok_r(p, ".", &sp);
 	while (token != NULL) {
 		ip |= (((uint32_t)strtoul(token, NULL, 10)) << shift);
-		token = strtok(NULL, ".");
+		token = strtok_r(NULL, ".", &sp);
 		shift -= 8;
 	}
 
@@ -120,13 +121,14 @@ int
 parser_ip6_read(uint8_t *value, char *p)
 {
 	uint64_t val = 0;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(p, ":");
+	token = strtok_r(p, ":", &sp);
 	while (token != NULL) {
 		hex_string_to_uint64(&val, token);
 		*value = val;
-		token = strtok(NULL, ":");
+		token = strtok_r(NULL, ":", &sp);
 		value++;
 		val = 0;
 	}
@@ -139,13 +141,14 @@ parser_mac_read(uint64_t *value, char *p)
 {
 	uint64_t mac = 0, val = 0;
 	uint8_t shift = 40;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(p, ":");
+	token = strtok_r(p, ":", &sp);
 	while (token != NULL) {
 		hex_string_to_uint64(&val, token);
 		mac |= val << shift;
-		token = strtok(NULL, ":");
+		token = strtok_r(NULL, ":", &sp);
 		shift -= 8;
 		val = 0;
 	}
-- 
2.30.0

[PATCH 02/21] app/test-bbdev: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 app/test-bbdev/test_bbdev_vector.c | 25 +++++++++++++++----------
 1 file changed, 15 insertions(+), 10 deletions(-)

diff --git a/app/test-bbdev/test_bbdev_vector.c b/app/test-bbdev/test_bbdev_vector.c
index c26727cd35c4..5c966e79645e 100644
--- a/app/test-bbdev/test_bbdev_vector.c
+++ b/app/test-bbdev/test_bbdev_vector.c
@@ -63,8 +63,9 @@ parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
 
 	uint32_t *values, *values_resized;
 	char *tok, *error = NULL;
+	char *sp = NULL;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -98,7 +99,7 @@ parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
 
 		*data_length = *data_length + (strlen(tok) - strlen("0x"))/2;
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 
@@ -299,8 +300,9 @@ parse_turbo_flags(char *tokens, uint32_t *op_flags,
 {
 	char *tok = NULL;
 	uint32_t op_flag_value = 0;
+	char *sp = NULL;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -330,7 +332,7 @@ parse_turbo_flags(char *tokens, uint32_t *op_flags,
 
 		*op_flags = *op_flags | op_flag_value;
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 	}
@@ -368,9 +370,10 @@ static int
 parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
 {
 	char *tok = NULL;
+	char *sp = NULL;
 	bool status_ok = false;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -401,7 +404,7 @@ parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
 			return -1;
 		}
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 	}
@@ -894,6 +897,7 @@ parse_fft_params(const char *key_token, char *token,
 	int ret = 0, status = 0, i, shift;
 	uint32_t op_flags = 0;
 	char *tok, *err = NULL;
+	char *sp = NULL;
 
 	struct rte_bbdev_op_fft *fft = &vector->fft;
 
@@ -922,7 +926,7 @@ parse_fft_params(const char *key_token, char *token,
 		fft->output_leading_depadding = (uint32_t) strtoul(token, &err, 0);
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "window_index")) {
-		tok = strtok(token, VALUE_DELIMITER);
+		tok = strtok_r(token, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			return -1;
 		for (i = 0; i < FFT_WIN_SIZE; i++) {
@@ -930,7 +934,7 @@ parse_fft_params(const char *key_token, char *token,
 			fft->window_index[i / 2] |= (uint32_t) strtoul(tok, &err, 0)
 					<< shift;
 			if (i < (FFT_WIN_SIZE - 1)) {
-				tok = strtok(NULL, VALUE_DELIMITER);
+				tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 				if (tok == NULL)
 					return -1;
 			}
@@ -995,6 +999,7 @@ static int
 parse_entry(char *entry, struct test_bbdev_vector *vector)
 {
 	int ret = 0;
+	char *sp = NULL;
 	char *token, *key_token;
 	enum rte_bbdev_op_type op_type = RTE_BBDEV_OP_NONE;
 
@@ -1004,10 +1009,10 @@ parse_entry(char *entry, struct test_bbdev_vector *vector)
 	}
 
 	/* get key */
-	token = strtok(entry, ENTRY_DELIMITER);
+	token = strtok_r(entry, ENTRY_DELIMITER, &sp);
 	key_token = token;
 	/* get values for key */
-	token = strtok(NULL, ENTRY_DELIMITER);
+	token = strtok_r(NULL, ENTRY_DELIMITER, &sp);
 
 	if (key_token == NULL || token == NULL) {
 		printf("Expected 'key = values' but was '%.40s'..\n", entry);
-- 
2.30.0

[PATCH 03/21] app/test-compress-perf: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 app/test-compress-perf/comp_perf_options_parse.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/app/test-compress-perf/comp_perf_options_parse.c b/app/test-compress-perf/comp_perf_options_parse.c
index 6d8c370fc2ea..d997fa667a94 100644
--- a/app/test-compress-perf/comp_perf_options_parse.c
+++ b/app/test-compress-perf/comp_perf_options_parse.c
@@ -177,6 +177,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 {
 	char *token;
 	uint8_t number;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -184,7 +185,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ":");
+	token = strtok_r(copy_arg, ":", &sp);
 
 	/* Parse minimum value */
 	if (token != NULL) {
@@ -197,7 +198,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse increment value */
 	if (token != NULL) {
@@ -211,7 +212,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse maximum value */
 	if (token != NULL) {
@@ -225,7 +226,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	if (strtok(NULL, ":") != NULL)
+	if (strtok_r(NULL, ":", &sp) != NULL)
 		goto err_range;
 
 	free(copy_arg);
@@ -244,6 +245,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 	uint8_t count = 0;
 	uint32_t temp_min;
 	uint32_t temp_max;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -251,7 +253,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ",");
+	token = strtok_r(copy_arg, ",", &sp);
 
 	/* Parse first value */
 	if (token != NULL) {
@@ -266,7 +268,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 	} else
 		goto err_list;
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 
 	while (token != NULL) {
 		if (count == MAX_LIST) {
@@ -288,7 +290,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 		if (number > temp_max)
 			temp_max = number;
 
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	if (min)
-- 
2.30.0

[PATCH 04/21] app/test-crypto-perf: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 app/test-crypto-perf/cperf_options_parsing.c     | 16 +++++++++-------
 app/test-crypto-perf/cperf_test_vector_parsing.c | 10 ++++++----
 2 files changed, 15 insertions(+), 11 deletions(-)

diff --git a/app/test-crypto-perf/cperf_options_parsing.c b/app/test-crypto-perf/cperf_options_parsing.c
index 75afedc7fd6e..77cda0eb8e6f 100644
--- a/app/test-crypto-perf/cperf_options_parsing.c
+++ b/app/test-crypto-perf/cperf_options_parsing.c
@@ -161,6 +161,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 {
 	char *token;
 	uint32_t number;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -168,7 +169,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ":");
+	token = strtok_r(copy_arg, ":", &sp);
 
 	/* Parse minimum value */
 	if (token != NULL) {
@@ -182,7 +183,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse increment value */
 	if (token != NULL) {
@@ -196,7 +197,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse maximum value */
 	if (token != NULL) {
@@ -211,7 +212,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	if (strtok(NULL, ":") != NULL)
+	if (strtok_r(NULL, ":", &sp) != NULL)
 		goto err_range;
 
 	free(copy_arg);
@@ -230,6 +231,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 	uint8_t count = 0;
 	uint32_t temp_min;
 	uint32_t temp_max;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -237,7 +239,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ",");
+	token = strtok_r(copy_arg, ",", &sp);
 
 	/* Parse first value */
 	if (token != NULL) {
@@ -253,7 +255,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 	} else
 		goto err_list;
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 
 	while (token != NULL) {
 		if (count == MAX_LIST) {
@@ -275,7 +277,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 		if (number > temp_max)
 			temp_max = number;
 
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	if (min)
diff --git a/app/test-crypto-perf/cperf_test_vector_parsing.c b/app/test-crypto-perf/cperf_test_vector_parsing.c
index 737d61d4af6b..4d2e6053e745 100644
--- a/app/test-crypto-perf/cperf_test_vector_parsing.c
+++ b/app/test-crypto-perf/cperf_test_vector_parsing.c
@@ -220,8 +220,9 @@ parse_values(char *tokens, uint8_t **data, uint32_t *data_length)
 
 	uint8_t *values, *values_resized;
 	char *tok, *error = NULL;
+	char *sp = NULL;
 
-	tok = strtok(tokens, CPERF_VALUE_DELIMITER);
+	tok = strtok_r(tokens, CPERF_VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -252,7 +253,7 @@ parse_values(char *tokens, uint8_t **data, uint32_t *data_length)
 			return -1;
 		}
 
-		tok = strtok(NULL, CPERF_VALUE_DELIMITER);
+		tok = strtok_r(NULL, CPERF_VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 
@@ -283,6 +284,7 @@ parse_entry(char *entry, struct cperf_test_vector *vector,
 
 	uint8_t *data = NULL;
 	char *token, *key_token;
+	char *sp = NULL;
 
 	if (entry == NULL) {
 		printf("Expected entry value\n");
@@ -290,10 +292,10 @@ parse_entry(char *entry, struct cperf_test_vector *vector,
 	}
 
 	/* get key */
-	token = strtok(entry, CPERF_ENTRY_DELIMITER);
+	token = strtok_r(entry, CPERF_ENTRY_DELIMITER, &sp);
 	key_token = token;
 	/* get values for key */
-	token = strtok(NULL, CPERF_ENTRY_DELIMITER);
+	token = strtok_r(NULL, CPERF_ENTRY_DELIMITER, &sp);
 
 	if (key_token == NULL || token == NULL) {
 		printf("Expected 'key = values' but was '%.40s'..\n", entry);
-- 
2.30.0

[PATCH 05/21] app/test-dma-perf: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 app/test-dma-perf/main.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/app/test-dma-perf/main.c b/app/test-dma-perf/main.c
index e5bccc27da5e..38780939578e 100644
--- a/app/test-dma-perf/main.c
+++ b/app/test-dma-perf/main.c
@@ -164,6 +164,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 	uint16_t len;
 	char *input;
 	struct lcore_dma_map_t *lcore_dma_map;
+	char *sp = NULL;
 
 	if (test_case == NULL || value == NULL)
 		return -1;
@@ -175,7 +176,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 
 	memset(lcore_dma_map, 0, sizeof(struct lcore_dma_map_t));
 
-	char *token = strtok(input, ", ");
+	char *token = strtok_r(input, ", ", &sp);
 	while (token != NULL) {
 		if (lcore_dma_map->cnt >= MAX_LCORE_NB) {
 			free(input);
@@ -185,7 +186,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 		uint16_t lcore_id = atoi(token);
 		lcore_dma_map->lcores[lcore_dma_map->cnt++] = lcore_id;
 
-		token = strtok(NULL, ", ");
+		token = strtok_r(NULL, ", ", &sp);
 	}
 
 	free(input);
@@ -201,6 +202,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 	char *start, *end, *substr;
 	uint16_t lcore_id;
 	int ret = 0;
+	char *sp = NULL;
 
 	if (test_case == NULL || value == NULL)
 		return -1;
@@ -216,7 +218,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 		goto out;
 	}
 
-	substr = strtok(addrs, ",");
+	substr = strtok_r(addrs, ",", &sp);
 	if (substr == NULL) {
 		fprintf(stderr, "No input DMA address\n");
 		ret = -1;
@@ -258,7 +260,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 		strlcpy(lcore_dma_map->dma_names[lcore_dma_map->cnt], ptrs[1],
 				RTE_DEV_NAME_MAX_LEN);
 		lcore_dma_map->cnt++;
-		substr = strtok(NULL, ",");
+		substr = strtok_r(NULL, ",", &sp);
 	} while (substr != NULL);
 
 out:
@@ -486,6 +488,7 @@ main(int argc, char *argv[])
 	char *rst_path_ptr = NULL;
 	char rst_path[PATH_MAX];
 	int new_argc;
+	char *sp = NULL;
 
 	memset(args, 0, sizeof(args));
 
@@ -504,7 +507,7 @@ main(int argc, char *argv[])
 	}
 	if (rst_path_ptr == NULL) {
 		strlcpy(rst_path, cfg_path_ptr, PATH_MAX);
-		char *token = strtok(basename(rst_path), ".");
+		char *token = strtok_r(basename(rst_path), ".", &sp);
 		if (token == NULL) {
 			printf("Config file error.\n");
 			return -1;
-- 
2.30.0

[PATCH 06/21] app/test-fib: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 app/test-fib/main.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/app/test-fib/main.c b/app/test-fib/main.c
index 75a56135f212..8e5d17b13028 100644
--- a/app/test-fib/main.c
+++ b/app/test-fib/main.c
@@ -223,9 +223,9 @@ parse_distrib(uint8_t depth_lim, const uint32_t n)
 	uint32_t nrpd[128 + 1] = {0}; /* number of routes per depth */
 	uint32_t n_routes;
 	uint8_t depth, ratio, ratio_acc = 0;
-	char *in;
+	char *in, *sp = NULL;
 
-	in = strtok(distrib_string, ",");
+	in = strtok_r(distrib_string, ",", &sp);
 
 	/*parse configures routes percentage ratios*/
 	while (in != NULL) {
@@ -265,7 +265,7 @@ parse_distrib(uint8_t depth_lim, const uint32_t n)
 		}
 
 		/*number of configured depths in*/
-		in = strtok(NULL, ",");
+		in = strtok_r(NULL, ",", &sp);
 	}
 
 	if (ratio_acc > 100) {
@@ -542,10 +542,10 @@ parse_lookup(FILE *f, int af)
 	int ret, i = 0;
 	uint8_t *tbl = (uint8_t *)config.lookup_tbl;
 	int step = (af == AF_INET) ? 4 : 16;
-	char *s;
+	char *s, *sp = NULL;
 
 	while (fgets(line, sizeof(line), f) != NULL) {
-		s = strtok(line, " \t\n");
+		s = strtok_r(line, " \t\n", &sp);
 		if (s == NULL)
 			return -EINVAL;
 		ret = inet_pton(af, s, &tbl[i]);
-- 
2.30.0

[PATCH 07/21] app/dpdk-test-flow-perf: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 app/test-flow-perf/main.c | 22 ++++++++++++----------
 1 file changed, 12 insertions(+), 10 deletions(-)

diff --git a/app/test-flow-perf/main.c b/app/test-flow-perf/main.c
index e224ef67983d..36e7c1d72019 100644
--- a/app/test-flow-perf/main.c
+++ b/app/test-flow-perf/main.c
@@ -602,6 +602,7 @@ read_meter_policy(char *prog, char *arg)
 {
 	char *token;
 	size_t i, j, k;
+	char *sp = NULL;
 
 	j = 0;
 	k = 0;
@@ -612,9 +613,9 @@ read_meter_policy(char *prog, char *arg)
 		token = strsep(&arg, ":\0");
 	}
 	j = 0;
-	token = strtok(actions_str[0], ",\0");
+	token = strtok_r(actions_str[0], ",\0", &sp);
 	while (token == NULL && j < RTE_COLORS - 1)
-		token = strtok(actions_str[++j], ",\0");
+		token = strtok_r(actions_str[++j], ",\0", &sp);
 	while (j < RTE_COLORS && token != NULL) {
 		for (i = 0; i < RTE_DIM(flow_options); i++) {
 			if (!strcmp(token, flow_options[i].str)) {
@@ -628,9 +629,9 @@ read_meter_policy(char *prog, char *arg)
 			usage(prog);
 			rte_exit(EXIT_SUCCESS, "Invalid colored actions\n");
 		}
-		token = strtok(NULL, ",\0");
+		token = strtok_r(NULL, ",\0", &sp);
 		while (!token && j < RTE_COLORS - 1) {
-			token = strtok(actions_str[++j], ",\0");
+			token = strtok_r(actions_str[++j], ",\0", &sp);
 			k = 0;
 		}
 	}
@@ -641,6 +642,7 @@ args_parse(int argc, char **argv)
 {
 	uint64_t pm, seed;
 	uint64_t hp_conf;
+	char *sp = NULL;
 	char **argvopt;
 	uint32_t prio;
 	char *token;
@@ -804,7 +806,7 @@ args_parse(int argc, char **argv)
 						RTE_FLOW_ACTION_TYPE_RAW_ENCAP
 					);
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					for (i = 0; i < RTE_DIM(flow_options); i++) {
 						if (strcmp(flow_options[i].str, token) == 0) {
@@ -817,7 +819,7 @@ args_parse(int argc, char **argv)
 							rte_exit(EXIT_FAILURE,
 								"Invalid encap item: %s\n", token);
 					}
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 				printf(" / ");
 			}
@@ -828,7 +830,7 @@ args_parse(int argc, char **argv)
 						RTE_FLOW_ACTION_TYPE_RAW_DECAP
 					);
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					for (i = 0; i < RTE_DIM(flow_options); i++) {
 						if (strcmp(flow_options[i].str, token) == 0) {
@@ -841,7 +843,7 @@ args_parse(int argc, char **argv)
 							rte_exit(EXIT_FAILURE,
 								"Invalid decap item %s\n", token);
 					}
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 				printf(" / ");
 			}
@@ -910,10 +912,10 @@ args_parse(int argc, char **argv)
 				uint16_t port_idx = 0;
 				char *token;
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					dst_ports[port_idx++] = atoi(token);
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 			}
 			if (strcmp(lgopts[opt_idx].name, "rxq") == 0) {
-- 
2.30.0

[PATCH 08/21] app/test-mldev: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 app/test-mldev/ml_options.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/app/test-mldev/ml_options.c b/app/test-mldev/ml_options.c
index 72357c1c393d..3f3dc8f69890 100644
--- a/app/test-mldev/ml_options.c
+++ b/app/test-mldev/ml_options.c
@@ -75,12 +75,12 @@ ml_parse_models(struct ml_options *opt, const char *arg)
 {
 	const char *delim = ",";
 	char models[PATH_MAX];
-	char *token;
+	char *token, *sp = NULL;
 	int ret = 0;
 
 	strlcpy(models, arg, PATH_MAX);
 
-	token = strtok(models, delim);
+	token = strtok_r(models, delim, &sp);
 	while (token != NULL) {
 		strlcpy(opt->filelist[opt->nb_filelist].model, token, PATH_MAX);
 		opt->nb_filelist++;
@@ -90,7 +90,7 @@ ml_parse_models(struct ml_options *opt, const char *arg)
 			ret = -EINVAL;
 			break;
 		}
-		token = strtok(NULL, delim);
+		token = strtok_r(NULL, delim, &sp);
 	}
 
 	if (opt->nb_filelist == 0) {
@@ -106,7 +106,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 {
 	const char *delim = ",";
 	char filelist[PATH_MAX];
-	char *token;
+	char *token, *sp = NULL;
 
 	if (opt->nb_filelist >= ML_TEST_MAX_MODELS) {
 		ml_err("Exceeded filelist count, max = %d\n", ML_TEST_MAX_MODELS);
@@ -116,7 +116,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(filelist, arg, PATH_MAX);
 
 	/* model */
-	token = strtok(filelist, delim);
+	token = strtok_r(filelist, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, model not specified = %s\n", arg);
 		return -EINVAL;
@@ -124,7 +124,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].model, token, PATH_MAX);
 
 	/* input */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, input not specified = %s\n", arg);
 		return -EINVAL;
@@ -132,7 +132,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].input, token, PATH_MAX);
 
 	/* output */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, output not specified = %s\n", arg);
 		return -EINVAL;
@@ -140,14 +140,14 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].output, token, PATH_MAX);
 
 	/* reference - optional */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token != NULL)
 		strlcpy(opt->filelist[opt->nb_filelist].reference, token, PATH_MAX);
 	else
 		memset(opt->filelist[opt->nb_filelist].reference, 0, PATH_MAX);
 
 	/* check for extra tokens */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token != NULL) {
 		ml_err("Invalid filelist. Entries > 4\n.");
 		return -EINVAL;
-- 
2.30.0

[PATCH 09/21] lib/dmadev: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 lib/dmadev/rte_dmadev.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/dmadev/rte_dmadev.c b/lib/dmadev/rte_dmadev.c
index 4e5e420c82a5..f856cc7d0905 100644
--- a/lib/dmadev/rte_dmadev.c
+++ b/lib/dmadev/rte_dmadev.c
@@ -971,7 +971,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	struct rte_dma_info dma_info;
 	struct rte_dma_stats dma_stats;
 	int dev_id, ret, vchan_id;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *vchan_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -990,7 +990,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	if (dma_info.nb_vchans == 1 && *end_param == '\0')
 		vchan_id = 0;
 	else {
-		vchan_param = strtok(end_param, ",");
+		vchan_param = strtok_r(end_param, ",", &sp);
 		if (!vchan_param || strlen(vchan_param) == 0 || !isdigit(*vchan_param))
 			return -EINVAL;
 
-- 
2.30.0

[PATCH 10/21] lib/eal: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 lib/eal/common/eal_common_memory.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/eal/common/eal_common_memory.c b/lib/eal/common/eal_common_memory.c
index d9433db62345..a05eb0442b0b 100644
--- a/lib/eal/common/eal_common_memory.c
+++ b/lib/eal/common/eal_common_memory.c
@@ -1273,22 +1273,22 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	char dlim[2] = ",";
 	char *params_args;
 	size_t count = 0;
-	char *token;
+	char *token, *sp = NULL;
 
 	if (vals == NULL || params == NULL || strlen(params) == 0)
 		return -1;
 
-	/* strtok expects char * and param is const char *. Hence on using
+	/* strtok_r expects char * and param is const char *. Hence on using
 	 * params as "const char *" compiler throws warning.
 	 */
 	params_args = strdup(params);
 	if (params_args == NULL)
 		return -1;
 
-	token = strtok(params_args, dlim);
+	token = strtok_r(params_args, dlim, &sp);
 	while (token && isdigit(*token) && count < n_vals) {
 		vals[count++] = strtoul(token, NULL, 10);
-		token = strtok(NULL, dlim);
+		token = strtok_r(NULL, dlim, &sp);
 	}
 
 	free(params_args);
-- 
2.30.0

Re: [PATCH 10/21] lib/eal: replace strtok with strtok_r

[Stephen Hemminger]

On Mon, 13 Nov 2023 18:45:39 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.

This code is only called in startup and can not be called by
multiple threads.

Re: [PATCH 10/21] lib/eal: replace strtok with strtok_r

[fengchengwen]

Hi Stephen,

On 2023/11/14 0:27, Stephen Hemminger wrote:
> On Mon, 13 Nov 2023 18:45:39 +0800
> Jie Hai <haijie1@huawei.com> wrote:
> 
>> Multiple threads calling the same function may cause condition
>> race issues, which often leads to abnormal behavior and can cause
>> more serious vulnerabilities such as abnormal termination, denial
>> of service, and compromised data integrity.
> 
> This code is only called in startup and can not be called by
> multiple threads.

For the DPDK examples and apps, I think it's OK if it called in startup then may not need replace.

But as for lib and drivers, I think we shouldn't make such an assumption.

At last I also recommend fix the examples and apps, because people may refer it.

Thanks
Chengwen

> 
> .
> 

Re: [PATCH 10/21] lib/eal: replace strtok with strtok_r

[Stephen Hemminger]

On Tue, 14 Nov 2023 09:05:10 +0800
fengchengwen <fengchengwen@huawei.com> wrote:

> Hi Stephen,
> 
> On 2023/11/14 0:27, Stephen Hemminger wrote:
> > On Mon, 13 Nov 2023 18:45:39 +0800
> > Jie Hai <haijie1@huawei.com> wrote:
> >   
> >> Multiple threads calling the same function may cause condition
> >> race issues, which often leads to abnormal behavior and can cause
> >> more serious vulnerabilities such as abnormal termination, denial
> >> of service, and compromised data integrity.  
> > 
> > This code is only called in startup and can not be called by
> > multiple threads.  
> 
> For the DPDK examples and apps, I think it's OK if it called in startup then may not need replace.
> 
> But as for lib and drivers, I think we shouldn't make such an assumption.
> 
> At last I also recommend fix the examples and apps, because people may refer it.

For startup of libs and driver initialization the code is always single threaded.
If not many, many other things would break.

[PATCH 11/21] lib/ethdev: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 lib/ethdev/rte_ethdev_telemetry.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/lib/ethdev/rte_ethdev_telemetry.c b/lib/ethdev/rte_ethdev_telemetry.c
index b01028ce9b60..6810c54109a5 100644
--- a/lib/ethdev/rte_ethdev_telemetry.c
+++ b/lib/ethdev/rte_ethdev_telemetry.c
@@ -477,6 +477,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	const char *qid_param;
 	uint16_t nb_queues;
 	char *end_param;
+	char *sp = NULL;
 	uint64_t qid;
 	int ret;
 
@@ -489,7 +490,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	if (nb_queues == 1 && *end_param == '\0')
 		qid = 0;
 	else {
-		qid_param = strtok(end_param, ",");
+		qid_param = strtok_r(end_param, ",", &sp);
 		if (!qid_param || strlen(qid_param) == 0 || !isdigit(*qid_param))
 			return -EINVAL;
 
@@ -1221,9 +1222,10 @@ static int
 eth_dev_parse_tm_params(char *params, uint32_t *result)
 {
 	const char *splited_param;
+	char *sp = NULL;
 	uint64_t ret;
 
-	splited_param = strtok(params, ",");
+	splited_param = strtok_r(params, ",", &sp);
 	if (!splited_param || strlen(splited_param) == 0 || !isdigit(*splited_param))
 		return -EINVAL;
 
-- 
2.30.0

[PATCH 12/21] lib/eventdev: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 lib/eventdev/rte_event_eth_rx_adapter.c | 38 ++++++++++++-------------
 lib/eventdev/rte_eventdev.c             | 18 ++++++------
 2 files changed, 28 insertions(+), 28 deletions(-)

diff --git a/lib/eventdev/rte_event_eth_rx_adapter.c b/lib/eventdev/rte_event_eth_rx_adapter.c
index 6db03adf0463..4107d600dfb7 100644
--- a/lib/eventdev/rte_event_eth_rx_adapter.c
+++ b/lib/eventdev/rte_event_eth_rx_adapter.c
@@ -3651,7 +3651,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp;
 	struct rte_event_eth_rx_adapter_queue_conf queue_conf;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3661,19 +3661,19 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3684,7 +3684,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3723,7 +3723,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 	struct rte_event_eth_rx_adapter_queue_stats q_stats;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3733,19 +3733,19 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3756,7 +3756,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3794,7 +3794,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3803,19 +3803,19 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3826,7 +3826,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3855,7 +3855,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	uint8_t instance_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3863,14 +3863,14 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3881,7 +3881,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
diff --git a/lib/eventdev/rte_eventdev.c b/lib/eventdev/rte_eventdev.c
index 0ca32d672175..fabba0228672 100644
--- a/lib/eventdev/rte_eventdev.c
+++ b/lib/eventdev/rte_eventdev.c
@@ -1784,7 +1784,7 @@ handle_queue_links(const char *cmd __rte_unused,
 		   struct rte_tel_data *d)
 {
 	int i, ret, port_id = 0;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	uint8_t dev_id;
 	uint8_t queues[RTE_EVENT_MAX_QUEUES_PER_DEV];
 	uint8_t priorities[RTE_EVENT_MAX_QUEUES_PER_DEV];
@@ -1797,12 +1797,12 @@ handle_queue_links(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
 		return -1;
 
 	port_id = strtoul(p_param, &end_param, 10);
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1922,7 +1922,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1932,7 +1932,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_PORT;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1940,7 +1940,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1956,7 +1956,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1966,7 +1966,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_QUEUE;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1974,7 +1974,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
-- 
2.30.0

[PATCH 13/21] lib/telemetry: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 lib/security/rte_security.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/lib/security/rte_security.c b/lib/security/rte_security.c
index b082a290296b..e20d610172ef 100644
--- a/lib/security/rte_security.c
+++ b/lib/security/rte_security.c
@@ -496,13 +496,14 @@ security_handle_cryptodev_crypto_caps(const char *cmd __rte_unused, const char *
 	int dev_id, capa_id;
 	int crypto_caps_n;
 	char *end_param;
+	char *sp = NULL;
 	int rc;
 
 	if (!params || strlen(params) == 0 || !isdigit(*params))
 		return -EINVAL;
 
 	dev_id = strtoul(params, &end_param, 0);
-	capa_param = strtok(end_param, ",");
+	capa_param = strtok_r(end_param, ",", &sp);
 	if (!capa_param || strlen(capa_param) == 0 || !isdigit(*capa_param))
 		return -EINVAL;
 
-- 
2.30.0

[PATCH 14/21] lib/telemetry: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 lib/telemetry/telemetry.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/lib/telemetry/telemetry.c b/lib/telemetry/telemetry.c
index 92982842a860..bb028d9381f7 100644
--- a/lib/telemetry/telemetry.c
+++ b/lib/telemetry/telemetry.c
@@ -371,6 +371,7 @@ static void *
 client_handler(void *sock_id)
 {
 	int s = (int)(uintptr_t)sock_id;
+	char *sp = NULL;
 	char buffer[1024];
 	char info_str[1024];
 	snprintf(info_str, sizeof(info_str),
@@ -385,8 +386,8 @@ client_handler(void *sock_id)
 	int bytes = read(s, buffer, sizeof(buffer) - 1);
 	while (bytes > 0) {
 		buffer[bytes] = 0;
-		const char *cmd = strtok(buffer, ",");
-		const char *param = strtok(NULL, "\0");
+		const char *cmd = strtok_r(buffer, ",", &sp);
+		const char *param = strtok_r(NULL, "\0", &sp);
 		telemetry_cb fn = unknown_command;
 		int i;
 
-- 
2.30.0

[PATCH 15/21] bus/fslmc: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 drivers/bus/fslmc/fslmc_bus.c            | 5 +++--
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 4 ++--
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index 57bfb5111a97..7960ad3979ef 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -131,6 +131,7 @@ scan_one_fslmc_device(char *dev_name)
 {
 	char *dup_dev_name, *t_ptr;
 	struct rte_dpaa2_device *dev = NULL;
+	char *sp = NULL;
 	int ret = -1;
 
 	if (!dev_name)
@@ -168,7 +169,7 @@ scan_one_fslmc_device(char *dev_name)
 	}
 
 	/* Parse the device name and ID */
-	t_ptr = strtok(dup_dev_name, ".");
+	t_ptr = strtok_r(dup_dev_name, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Invalid device found: (%s)", dup_dev_name);
 		ret = -EINVAL;
@@ -199,7 +200,7 @@ scan_one_fslmc_device(char *dev_name)
 	else
 		dev->dev_type = DPAA2_UNKNOWN;
 
-	t_ptr = strtok(NULL, ".");
+	t_ptr = strtok_r(NULL, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Skipping invalid device (%s)", dup_dev_name);
 		ret = 0;
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index 4aec7b2cd8ba..09a1a2b23787 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -129,7 +129,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 	uint32_t cpu_mask = 1;
 	int ret;
 	size_t len = 0;
-	char *temp = NULL, *token = NULL;
+	char *temp = NULL, *token = NULL, *sp = NULL;
 	char string[STRING_LEN], command[COMMAND_LEN];
 	FILE *file;
 
@@ -141,7 +141,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 	}
 	while (getline(&temp, &len, file) != -1) {
 		if ((strstr(temp, string)) != NULL) {
-			token = strtok(temp, ":");
+			token = strtok_r(temp, ":", &sp);
 			break;
 		}
 	}
-- 
2.30.0

RE: [PATCH 15/21] bus/fslmc: replace strtok with strtok_r

[Sachin Saxena]

Acked-by: Sachin Saxena <sachin.saxena@nxp.com>

> -----Original Message-----
> From: Jie Hai <haijie1@huawei.com>
> Sent: 13 November 2023 04:16 PM
> To: dev@dpdk.org; Hemant Agrawal <hemant.agrawal@nxp.com>; Sachin
> Saxena <sachin.saxena@nxp.com>
> Cc: haijie1@huawei.com; lihuisong@huawei.com; fengchengwen@huawei.com
> Subject: [PATCH 15/21] bus/fslmc: replace strtok with strtok_r
> 
> Multiple threads calling the same function may cause condition race issues,
> which often leads to abnormal behavior and can cause more serious
> vulnerabilities such as abnormal termination, denial of service, and
> compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a reentrant function.
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> ---
>  drivers/bus/fslmc/fslmc_bus.c            | 5 +++--
>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 4 ++--
>  2 files changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
> index 57bfb5111a97..7960ad3979ef 100644
> --- a/drivers/bus/fslmc/fslmc_bus.c
> +++ b/drivers/bus/fslmc/fslmc_bus.c
> @@ -131,6 +131,7 @@ scan_one_fslmc_device(char *dev_name)  {
>  	char *dup_dev_name, *t_ptr;
>  	struct rte_dpaa2_device *dev = NULL;
> +	char *sp = NULL;
>  	int ret = -1;
> 
>  	if (!dev_name)
> @@ -168,7 +169,7 @@ scan_one_fslmc_device(char *dev_name)
>  	}
> 
>  	/* Parse the device name and ID */
> -	t_ptr = strtok(dup_dev_name, ".");
> +	t_ptr = strtok_r(dup_dev_name, ".", &sp);
>  	if (!t_ptr) {
>  		DPAA2_BUS_ERR("Invalid device found: (%s)", dup_dev_name);
>  		ret = -EINVAL;
> @@ -199,7 +200,7 @@ scan_one_fslmc_device(char *dev_name)
>  	else
>  		dev->dev_type = DPAA2_UNKNOWN;
> 
> -	t_ptr = strtok(NULL, ".");
> +	t_ptr = strtok_r(NULL, ".", &sp);
>  	if (!t_ptr) {
>  		DPAA2_BUS_ERR("Skipping invalid device (%s)",
> dup_dev_name);
>  		ret = 0;
> diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
> b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
> index 4aec7b2cd8ba..09a1a2b23787 100644
> --- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
> +++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
> @@ -129,7 +129,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t
> dpio_id, int cpu_id)
>  	uint32_t cpu_mask = 1;
>  	int ret;
>  	size_t len = 0;
> -	char *temp = NULL, *token = NULL;
> +	char *temp = NULL, *token = NULL, *sp = NULL;
>  	char string[STRING_LEN], command[COMMAND_LEN];
>  	FILE *file;
> 
> @@ -141,7 +141,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t
> dpio_id, int cpu_id)
>  	}
>  	while (getline(&temp, &len, file) != -1) {
>  		if ((strstr(temp, string)) != NULL) {
> -			token = strtok(temp, ":");
> +			token = strtok_r(temp, ":", &sp);
>  			break;
>  		}
>  	}
> --
> 2.30.0

[PATCH 16/21] common/cnxk: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 drivers/common/cnxk/cnxk_telemetry_nix.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/common/cnxk/cnxk_telemetry_nix.c b/drivers/common/cnxk/cnxk_telemetry_nix.c
index ccae5d7853af..8054c3cf910c 100644
--- a/drivers/common/cnxk/cnxk_telemetry_nix.c
+++ b/drivers/common/cnxk/cnxk_telemetry_nix.c
@@ -761,7 +761,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			   struct plt_tel_data *d)
 {
 	struct nix_tel_node *node;
-	char *name, *param;
+	char *name, *param, *sp = NULL;
 	char buf[1024];
 	int rc = -1;
 
@@ -769,11 +769,11 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 		goto exit;
 
 	plt_strlcpy(buf, params, PCI_PRI_STR_SIZE + 1);
-	name = strtok(buf, ",");
+	name = strtok_r(buf, ",", &sp);
 	if (name == NULL)
 		goto exit;
 
-	param = strtok(NULL, "\0");
+	param = strtok_r(NULL, "\0", &sp);
 
 	node = nix_tel_node_get_by_pcidev_name(name);
 	if (!node)
@@ -782,7 +782,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 	plt_tel_data_start_dict(d);
 
 	if (strstr(cmd, "rq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int rq;
 
 		if (!tok)
@@ -798,7 +798,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_rq(node->rqs[rq], d);
 
 	} else if (strstr(cmd, "cq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int cq;
 
 		if (!tok)
@@ -814,7 +814,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_cq(node->cqs[cq], d);
 
 	} else if (strstr(cmd, "sq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int sq;
 
 		if (!tok)
-- 
2.30.0

[PATCH 17/21] event/cnxk: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 drivers/event/cnxk/cnxk_eventdev.c  | 10 ++++++----
 drivers/event/cnxk/cnxk_tim_evdev.c | 11 ++++++-----
 2 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/drivers/event/cnxk/cnxk_eventdev.c b/drivers/event/cnxk/cnxk_eventdev.c
index 0c61f4c20eec..fe7a86797c25 100644
--- a/drivers/event/cnxk/cnxk_eventdev.c
+++ b/drivers/event/cnxk/cnxk_eventdev.c
@@ -478,7 +478,8 @@ parse_queue_param(char *value, void *opaque)
 	struct cnxk_sso_qos queue_qos = {0};
 	uint16_t *val = (uint16_t *)&queue_qos;
 	struct cnxk_sso_evdev *dev = opaque;
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "-", &sp);
 	struct cnxk_sso_qos *old_ptr;
 
 	if (!strlen(value))
@@ -486,7 +487,7 @@ parse_queue_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		val++;
 	}
 
@@ -514,7 +515,8 @@ parse_stash_param(char *value, void *opaque)
 	struct cnxk_sso_stash queue_stash = {0};
 	struct cnxk_sso_evdev *dev = opaque;
 	struct cnxk_sso_stash *old_ptr;
-	char *tok = strtok(value, "|");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "|", &sp);
 	uint16_t *val;
 
 	if (!strlen(value))
@@ -523,7 +525,7 @@ parse_stash_param(char *value, void *opaque)
 	val = (uint16_t *)&queue_stash;
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "|");
+		tok = strtok_r(NULL, "|", &sp);
 		val++;
 	}
 
diff --git a/drivers/event/cnxk/cnxk_tim_evdev.c b/drivers/event/cnxk/cnxk_tim_evdev.c
index 6d59fdf90983..86ef7dc3d578 100644
--- a/drivers/event/cnxk/cnxk_tim_evdev.c
+++ b/drivers/event/cnxk/cnxk_tim_evdev.c
@@ -420,7 +420,8 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 {
 	struct cnxk_tim_evdev *dev = opaque;
 	struct cnxk_tim_ctl ring_ctl = {0};
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "-", &sp);
 	struct cnxk_tim_ctl *old_ptr;
 	uint16_t *val;
 
@@ -431,7 +432,7 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		val++;
 	}
 
@@ -507,16 +508,16 @@ cnxk_tim_parse_clk_list(const char *value, void *opaque)
 				      ROC_TIM_CLK_SRC_INVALID};
 	struct cnxk_tim_evdev *dev = opaque;
 	char *str = strdup(value);
-	char *tok;
+	char *tok, *sp = NULL;
 	int i = 0;
 
 	if (str == NULL || !strlen(str))
 		goto free;
 
-	tok = strtok(str, "-");
+	tok = strtok_r(str, "-", &sp);
 	while (tok != NULL && src[i] != ROC_TIM_CLK_SRC_INVALID) {
 		dev->ext_clk_freq[src[i]] = strtoull(tok, NULL, 10);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		i++;
 	}
 
-- 
2.30.0

[PATCH 18/21] net/ark: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 drivers/net/ark/ark_pktchkr.c | 10 +++++-----
 drivers/net/ark/ark_pktgen.c  | 10 +++++-----
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ark/ark_pktchkr.c b/drivers/net/ark/ark_pktchkr.c
index e1f336c73c2a..6e8e5339607b 100644
--- a/drivers/net/ark/ark_pktchkr.c
+++ b/drivers/net/ark/ark_pktchkr.c
@@ -359,14 +359,14 @@ set_arg(char *arg, char *val)
 void
 ark_pktchkr_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = "=\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_r(args, toks, &sp);
+	v = strtok_r(NULL, toks, &sp);
 	while (argv && v) {
 		set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_r(NULL, toks, &sp);
+		v = strtok_r(NULL, toks, &sp);
 	}
 }
 
diff --git a/drivers/net/ark/ark_pktgen.c b/drivers/net/ark/ark_pktgen.c
index 69ff7072b2ab..d611406a1b46 100644
--- a/drivers/net/ark/ark_pktgen.c
+++ b/drivers/net/ark/ark_pktgen.c
@@ -340,14 +340,14 @@ pmd_set_arg(char *arg, char *val)
 void
 ark_pktgen_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = " =\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_r(args, toks, &sp);
+	v = strtok_r(NULL, toks, &sp);
 	while (argv && v) {
 		pmd_set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_r(NULL, toks, &sp);
+		v = strtok_r(NULL, toks, &sp);
 	}
 }
 
-- 
2.30.0

[PATCH 19/21] raw/cnxk_gpio: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 drivers/raw/cnxk_gpio/cnxk_gpio.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/raw/cnxk_gpio/cnxk_gpio.c b/drivers/raw/cnxk_gpio/cnxk_gpio.c
index 29c250672646..a499a258f315 100644
--- a/drivers/raw/cnxk_gpio/cnxk_gpio.c
+++ b/drivers/raw/cnxk_gpio/cnxk_gpio.c
@@ -190,7 +190,7 @@ static int
 cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 {
 	int i, ret, val, queue = 0;
-	char *token;
+	char *token, *sp = NULL;
 	int *list;
 
 	list = rte_calloc(NULL, gpiochip->num_gpios, sizeof(*list), 0);
@@ -208,7 +208,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 	allowlist[strlen(allowlist) - 1] = ' ';
 
 	/* quiesce -Wcast-qual */
-	token = strtok((char *)(uintptr_t)allowlist, ",");
+	token = strtok_r((char *)(uintptr_t)allowlist, ",", &sp);
 	do {
 		errno = 0;
 		val = strtol(token, NULL, 10);
@@ -234,7 +234,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 		}
 		if (i == queue)
 			list[queue++] = val;
-	} while ((token = strtok(NULL, ",")));
+	} while ((token = strtok_r(NULL, ",", &sp)));
 
 	free(allowlist);
 	gpiochip->allowlist = list;
-- 
2.30.0

[PATCH 20/21] examples/l2fwd-crypto: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 examples/l2fwd-crypto/main.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/examples/l2fwd-crypto/main.c b/examples/l2fwd-crypto/main.c
index efe7eea2a768..06e296776b95 100644
--- a/examples/l2fwd-crypto/main.c
+++ b/examples/l2fwd-crypto/main.c
@@ -1105,12 +1105,12 @@ static int
 parse_bytes(uint8_t *data, char *input_arg, uint16_t max_size)
 {
 	unsigned byte_count;
-	char *token;
+	char *token, *sp = NULL;
 
 	errno = 0;
-	for (byte_count = 0, token = strtok(input_arg, ":");
+	for (byte_count = 0, token = strtok_r(input_arg, ":", &sp);
 			(byte_count < max_size) && (token != NULL);
-			token = strtok(NULL, ":")) {
+			token = strtok_r(NULL, ":", &sp)) {
 
 		int number = (int)strtol(token, NULL, 16);
 
-- 
2.30.0

[PATCH 21/21] examples/vhost: replace strtok with strtok_r

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant function.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 examples/vhost/main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index ce5c1efddf5c..e8b3a97c48a4 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -246,6 +246,7 @@ open_dma(const char *value)
 	char *ptrs[2];
 	char *start, *end, *substr;
 	int64_t socketid, vring_id;
+	char *sp = NULL;
 
 	struct rte_dma_info info;
 	struct rte_dma_conf dev_config = { .nb_vchans = 1 };
@@ -269,7 +270,7 @@ open_dma(const char *value)
 
 	/* process DMA devices within bracket. */
 	addrs++;
-	substr = strtok(addrs, ";]");
+	substr = strtok_r(addrs, ";]", &sp);
 	if (!substr) {
 		ret = -1;
 		goto out;
-- 
2.30.0

Re: [PATCH 21/21] examples/vhost: replace strtok with strtok_r

[Stephen Hemminger]

On Mon, 13 Nov 2023 18:45:50 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant function.
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>

This is in initialization code, can not be called by multiple threads.

Re: [PATCH 00/21] replace strtok with strtok_r

[Thomas Monjalon]

13/11/2023 11:45, Jie Hai:
> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant function.

You should add a check in checkpatches.sh so we won't add more in future.

Re: [PATCH 00/21] replace strtok with strtok_r

[fengchengwen]

Hi Jie,

Good fix

There are two minor I think need to modify:
1. The [PATCH 13/21] lib/telemetry should be lib/security
2. All commits should add Cc because it's potential bug.

The other LGTM, with above fixed
Series-acked-by: Chengwen Feng <fengchengwen@huawei.com>

Thanks
Chengwen

On 2023/11/13 18:45, Jie Hai wrote:
> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant function.
> 
> Jie Hai (21):
>   app/graph: replace strtok with strtok_r
>   app/test-bbdev: replace strtok with strtok_r
>   app/test-compress-perf: replace strtok with strtok_r
>   app/test-crypto-perf: replace strtok with strtok_r
>   app/test-dma-perf: replace strtok with strtok_r
>   app/test-fib: replace strtok with strtok_r
>   app/dpdk-test-flow-perf: replace strtok with strtok_r
>   app/test-mldev: replace strtok with strtok_r
>   lib/dmadev: replace strtok with strtok_r
>   lib/eal: replace strtok with strtok_r
>   lib/ethdev: replace strtok with strtok_r
>   lib/eventdev: replace strtok with strtok_r
>   lib/telemetry: replace strtok with strtok_r
>   lib/telemetry: replace strtok with strtok_r
>   bus/fslmc: replace strtok with strtok_r
>   common/cnxk: replace strtok with strtok_r
>   event/cnxk: replace strtok with strtok_r
>   net/ark: replace strtok with strtok_r
>   raw/cnxk_gpio: replace strtok with strtok_r
>   examples/l2fwd-crypto: replace strtok with strtok_r
>   examples/vhost: replace strtok with strtok_r
> 
>  app/graph/graph.c                             |  5 ++-
>  app/graph/utils.c                             | 15 +++++---
>  app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
>  .../comp_perf_options_parse.c                 | 16 ++++----
>  app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
>  .../cperf_test_vector_parsing.c               | 10 +++--
>  app/test-dma-perf/main.c                      | 13 ++++---
>  app/test-fib/main.c                           | 10 ++---
>  app/test-flow-perf/main.c                     | 22 ++++++-----
>  app/test-mldev/ml_options.c                   | 18 ++++-----
>  drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
>  drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
>  drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
>  drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
>  drivers/net/ark/ark_pktchkr.c                 | 10 ++---
>  drivers/net/ark/ark_pktgen.c                  | 10 ++---
>  drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
>  examples/l2fwd-crypto/main.c                  |  6 +--
>  examples/vhost/main.c                         |  3 +-
>  lib/dmadev/rte_dmadev.c                       |  4 +-
>  lib/eal/common/eal_common_memory.c            |  8 ++--
>  lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
>  lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
>  lib/eventdev/rte_eventdev.c                   | 18 ++++-----
>  lib/security/rte_security.c                   |  3 +-
>  lib/telemetry/telemetry.c                     |  5 ++-
>  27 files changed, 169 insertions(+), 140 deletions(-)
> 

Re: [PATCH 00/21] replace strtok with strtok_r

[Stephen Hemminger]

On Mon, 13 Nov 2023 18:45:29 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant function.

Lots of churn. And most of these places are never callable from
multiple threads.

It does indicate that some better  arg handling is needed.

Re: [PATCH 00/21] replace strtok with strtok_r

[Tyler Retzlaff]

On Mon, Nov 13, 2023 at 06:45:29PM +0800, Jie Hai wrote:
> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant function.

could you please use strtok_s instead of strtok_r the former is part of
the C11 standard the latter is not.

thanks!

> 
> Jie Hai (21):
>   app/graph: replace strtok with strtok_r
>   app/test-bbdev: replace strtok with strtok_r
>   app/test-compress-perf: replace strtok with strtok_r
>   app/test-crypto-perf: replace strtok with strtok_r
>   app/test-dma-perf: replace strtok with strtok_r
>   app/test-fib: replace strtok with strtok_r
>   app/dpdk-test-flow-perf: replace strtok with strtok_r
>   app/test-mldev: replace strtok with strtok_r
>   lib/dmadev: replace strtok with strtok_r
>   lib/eal: replace strtok with strtok_r
>   lib/ethdev: replace strtok with strtok_r
>   lib/eventdev: replace strtok with strtok_r
>   lib/telemetry: replace strtok with strtok_r
>   lib/telemetry: replace strtok with strtok_r
>   bus/fslmc: replace strtok with strtok_r
>   common/cnxk: replace strtok with strtok_r
>   event/cnxk: replace strtok with strtok_r
>   net/ark: replace strtok with strtok_r
>   raw/cnxk_gpio: replace strtok with strtok_r
>   examples/l2fwd-crypto: replace strtok with strtok_r
>   examples/vhost: replace strtok with strtok_r
> 
>  app/graph/graph.c                             |  5 ++-
>  app/graph/utils.c                             | 15 +++++---
>  app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
>  .../comp_perf_options_parse.c                 | 16 ++++----
>  app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
>  .../cperf_test_vector_parsing.c               | 10 +++--
>  app/test-dma-perf/main.c                      | 13 ++++---
>  app/test-fib/main.c                           | 10 ++---
>  app/test-flow-perf/main.c                     | 22 ++++++-----
>  app/test-mldev/ml_options.c                   | 18 ++++-----
>  drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
>  drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
>  drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
>  drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
>  drivers/net/ark/ark_pktchkr.c                 | 10 ++---
>  drivers/net/ark/ark_pktgen.c                  | 10 ++---
>  drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
>  examples/l2fwd-crypto/main.c                  |  6 +--
>  examples/vhost/main.c                         |  3 +-
>  lib/dmadev/rte_dmadev.c                       |  4 +-
>  lib/eal/common/eal_common_memory.c            |  8 ++--
>  lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
>  lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
>  lib/eventdev/rte_eventdev.c                   | 18 ++++-----
>  lib/security/rte_security.c                   |  3 +-
>  lib/telemetry/telemetry.c                     |  5 ++-
>  27 files changed, 169 insertions(+), 140 deletions(-)
> 
> -- 
> 2.30.0

Re: [PATCH 00/21] replace strtok with strtok_r

[Jie Hai]

On 2023/11/14 1:09, Tyler Retzlaff wrote:
> On Mon, Nov 13, 2023 at 06:45:29PM +0800, Jie Hai wrote:
>> Multiple threads calling the same function may cause condition
>> race issues, which often leads to abnormal behavior and can cause
>> more serious vulnerabilities such as abnormal termination, denial
>> of service, and compromised data integrity.
>>
>> The strtok() is non-reentrant, it is better to replace it with a
>> reentrant function.
> 
> could you please use strtok_s instead of strtok_r the former is part of
> the C11 standard the latter is not.
> 
> thanks!
> 
Hi, Tyler Retzlaff

Thanks for your comment.

For the use of strtok_s, I have consulted some documents, see
	https://en.cppreference.com/w/c/string/byte/strtok
It says
"As with all bounds-checked functions, strtok_s only guaranteed to be
available if __STDC_LIB_EXT1__ is defined by the implementation and if
the user defines __STDC_WANT_LIB_EXT1__ to the integer constant 1 before
including <string.h>.
"

I use strtok_s with "#ifdef __STDC_LIB_EXT1__ ... #endif" around and 
compiled
locally and found that __STDC_LIB_EXT1__ was not defined, so the related
code  was not called. I'm afraid there's a problem with this
modification.

Am I using strtok_s wrong?

Thanks,
Jie Hai
>>
>> Jie Hai (21):
>>    app/graph: replace strtok with strtok_r
>>    app/test-bbdev: replace strtok with strtok_r
>>    app/test-compress-perf: replace strtok with strtok_r
>>    app/test-crypto-perf: replace strtok with strtok_r
>>    app/test-dma-perf: replace strtok with strtok_r
>>    app/test-fib: replace strtok with strtok_r
>>    app/dpdk-test-flow-perf: replace strtok with strtok_r
>>    app/test-mldev: replace strtok with strtok_r
>>    lib/dmadev: replace strtok with strtok_r
>>    lib/eal: replace strtok with strtok_r
>>    lib/ethdev: replace strtok with strtok_r
>>    lib/eventdev: replace strtok with strtok_r
>>    lib/telemetry: replace strtok with strtok_r
>>    lib/telemetry: replace strtok with strtok_r
>>    bus/fslmc: replace strtok with strtok_r
>>    common/cnxk: replace strtok with strtok_r
>>    event/cnxk: replace strtok with strtok_r
>>    net/ark: replace strtok with strtok_r
>>    raw/cnxk_gpio: replace strtok with strtok_r
>>    examples/l2fwd-crypto: replace strtok with strtok_r
>>    examples/vhost: replace strtok with strtok_r
>>
>>   app/graph/graph.c                             |  5 ++-
>>   app/graph/utils.c                             | 15 +++++---
>>   app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
>>   .../comp_perf_options_parse.c                 | 16 ++++----
>>   app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
>>   .../cperf_test_vector_parsing.c               | 10 +++--
>>   app/test-dma-perf/main.c                      | 13 ++++---
>>   app/test-fib/main.c                           | 10 ++---
>>   app/test-flow-perf/main.c                     | 22 ++++++-----
>>   app/test-mldev/ml_options.c                   | 18 ++++-----
>>   drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
>>   drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
>>   drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
>>   drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
>>   drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
>>   drivers/net/ark/ark_pktchkr.c                 | 10 ++---
>>   drivers/net/ark/ark_pktgen.c                  | 10 ++---
>>   drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
>>   examples/l2fwd-crypto/main.c                  |  6 +--
>>   examples/vhost/main.c                         |  3 +-
>>   lib/dmadev/rte_dmadev.c                       |  4 +-
>>   lib/eal/common/eal_common_memory.c            |  8 ++--
>>   lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
>>   lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
>>   lib/eventdev/rte_eventdev.c                   | 18 ++++-----
>>   lib/security/rte_security.c                   |  3 +-
>>   lib/telemetry/telemetry.c                     |  5 ++-
>>   27 files changed, 169 insertions(+), 140 deletions(-)
>>
>> -- 
>> 2.30.0
> .

Re: [PATCH 00/21] replace strtok with strtok_r

[Tyler Retzlaff]

On Tue, Nov 14, 2023 at 08:50:17PM +0800, Jie Hai wrote:
> On 2023/11/14 1:09, Tyler Retzlaff wrote:
> >On Mon, Nov 13, 2023 at 06:45:29PM +0800, Jie Hai wrote:
> >>Multiple threads calling the same function may cause condition
> >>race issues, which often leads to abnormal behavior and can cause
> >>more serious vulnerabilities such as abnormal termination, denial
> >>of service, and compromised data integrity.
> >>
> >>The strtok() is non-reentrant, it is better to replace it with a
> >>reentrant function.
> >
> >could you please use strtok_s instead of strtok_r the former is part of
> >the C11 standard the latter is not.
> >
> >thanks!
> >
> Hi, Tyler Retzlaff
> 
> Thanks for your comment.
> 
> For the use of strtok_s, I have consulted some documents, see
> 	https://en.cppreference.com/w/c/string/byte/strtok
> It says
> "As with all bounds-checked functions, strtok_s only guaranteed to be
> available if __STDC_LIB_EXT1__ is defined by the implementation and if
> the user defines __STDC_WANT_LIB_EXT1__ to the integer constant 1 before
> including <string.h>.
> "
> 
> I use strtok_s with "#ifdef __STDC_LIB_EXT1__ ... #endif" around and
> compiled
> locally and found that __STDC_LIB_EXT1__ was not defined, so the related
> code  was not called. I'm afraid there's a problem with this
> modification.
> 
> Am I using strtok_s wrong?

no i overlooked that it is optional my fault sorry.

since there is no portable strtok_r i'm going to have to agree with others
that only places where you actually need reentrant strtok be converted to
strtok_r.

for windows i think we'll either need to introduce an abstracted strtok_r
name for portability or something in the rte_ namespace (dependent on
what other revieweres would prefer)

thanks!

> 
> Thanks,
> Jie Hai
> >>
> >>Jie Hai (21):
> >>   app/graph: replace strtok with strtok_r
> >>   app/test-bbdev: replace strtok with strtok_r
> >>   app/test-compress-perf: replace strtok with strtok_r
> >>   app/test-crypto-perf: replace strtok with strtok_r
> >>   app/test-dma-perf: replace strtok with strtok_r
> >>   app/test-fib: replace strtok with strtok_r
> >>   app/dpdk-test-flow-perf: replace strtok with strtok_r
> >>   app/test-mldev: replace strtok with strtok_r
> >>   lib/dmadev: replace strtok with strtok_r
> >>   lib/eal: replace strtok with strtok_r
> >>   lib/ethdev: replace strtok with strtok_r
> >>   lib/eventdev: replace strtok with strtok_r
> >>   lib/telemetry: replace strtok with strtok_r
> >>   lib/telemetry: replace strtok with strtok_r
> >>   bus/fslmc: replace strtok with strtok_r
> >>   common/cnxk: replace strtok with strtok_r
> >>   event/cnxk: replace strtok with strtok_r
> >>   net/ark: replace strtok with strtok_r
> >>   raw/cnxk_gpio: replace strtok with strtok_r
> >>   examples/l2fwd-crypto: replace strtok with strtok_r
> >>   examples/vhost: replace strtok with strtok_r
> >>
> >>  app/graph/graph.c                             |  5 ++-
> >>  app/graph/utils.c                             | 15 +++++---
> >>  app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
> >>  .../comp_perf_options_parse.c                 | 16 ++++----
> >>  app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
> >>  .../cperf_test_vector_parsing.c               | 10 +++--
> >>  app/test-dma-perf/main.c                      | 13 ++++---
> >>  app/test-fib/main.c                           | 10 ++---
> >>  app/test-flow-perf/main.c                     | 22 ++++++-----
> >>  app/test-mldev/ml_options.c                   | 18 ++++-----
> >>  drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
> >>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
> >>  drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
> >>  drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
> >>  drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
> >>  drivers/net/ark/ark_pktchkr.c                 | 10 ++---
> >>  drivers/net/ark/ark_pktgen.c                  | 10 ++---
> >>  drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
> >>  examples/l2fwd-crypto/main.c                  |  6 +--
> >>  examples/vhost/main.c                         |  3 +-
> >>  lib/dmadev/rte_dmadev.c                       |  4 +-
> >>  lib/eal/common/eal_common_memory.c            |  8 ++--
> >>  lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
> >>  lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
> >>  lib/eventdev/rte_eventdev.c                   | 18 ++++-----
> >>  lib/security/rte_security.c                   |  3 +-
> >>  lib/telemetry/telemetry.c                     |  5 ++-
> >>  27 files changed, 169 insertions(+), 140 deletions(-)
> >>
> >>-- 
> >>2.30.0
> >.

Re: [PATCH 00/21] replace strtok with strtok_r

[Tyler Retzlaff]

On Tue, Nov 14, 2023 at 09:32:48AM -0800, Tyler Retzlaff wrote:
> On Tue, Nov 14, 2023 at 08:50:17PM +0800, Jie Hai wrote:
> > On 2023/11/14 1:09, Tyler Retzlaff wrote:
> > >On Mon, Nov 13, 2023 at 06:45:29PM +0800, Jie Hai wrote:
> > >>Multiple threads calling the same function may cause condition
> > >>race issues, which often leads to abnormal behavior and can cause
> > >>more serious vulnerabilities such as abnormal termination, denial
> > >>of service, and compromised data integrity.
> > >>
> > >>The strtok() is non-reentrant, it is better to replace it with a
> > >>reentrant function.
> > >
> > >could you please use strtok_s instead of strtok_r the former is part of
> > >the C11 standard the latter is not.
> > >
> > >thanks!
> > >
> > Hi, Tyler Retzlaff
> > 
> > Thanks for your comment.
> > 
> > For the use of strtok_s, I have consulted some documents, see
> > 	https://en.cppreference.com/w/c/string/byte/strtok
> > It says
> > "As with all bounds-checked functions, strtok_s only guaranteed to be
> > available if __STDC_LIB_EXT1__ is defined by the implementation and if
> > the user defines __STDC_WANT_LIB_EXT1__ to the integer constant 1 before
> > including <string.h>.
> > "
> > 
> > I use strtok_s with "#ifdef __STDC_LIB_EXT1__ ... #endif" around and
> > compiled
> > locally and found that __STDC_LIB_EXT1__ was not defined, so the related
> > code  was not called. I'm afraid there's a problem with this
> > modification.
> > 
> > Am I using strtok_s wrong?
> 
> no i overlooked that it is optional my fault sorry.
> 
> since there is no portable strtok_r i'm going to have to agree with others
> that only places where you actually need reentrant strtok be converted to
> strtok_r.
> 
> for windows i think we'll either need to introduce an abstracted strtok_r
> name for portability or something in the rte_ namespace (dependent on
> what other revieweres would prefer)

just as a follow up here maybe it would be optimal if we could use
strtok_s *if* we test that the toolchain makes it available and if not
provide a mapping of strtok_s -> strtok_r.

what do others think?

> 
> thanks!
> 
> > 
> > Thanks,
> > Jie Hai
> > >>
> > >>Jie Hai (21):
> > >>   app/graph: replace strtok with strtok_r
> > >>   app/test-bbdev: replace strtok with strtok_r
> > >>   app/test-compress-perf: replace strtok with strtok_r
> > >>   app/test-crypto-perf: replace strtok with strtok_r
> > >>   app/test-dma-perf: replace strtok with strtok_r
> > >>   app/test-fib: replace strtok with strtok_r
> > >>   app/dpdk-test-flow-perf: replace strtok with strtok_r
> > >>   app/test-mldev: replace strtok with strtok_r
> > >>   lib/dmadev: replace strtok with strtok_r
> > >>   lib/eal: replace strtok with strtok_r
> > >>   lib/ethdev: replace strtok with strtok_r
> > >>   lib/eventdev: replace strtok with strtok_r
> > >>   lib/telemetry: replace strtok with strtok_r
> > >>   lib/telemetry: replace strtok with strtok_r
> > >>   bus/fslmc: replace strtok with strtok_r
> > >>   common/cnxk: replace strtok with strtok_r
> > >>   event/cnxk: replace strtok with strtok_r
> > >>   net/ark: replace strtok with strtok_r
> > >>   raw/cnxk_gpio: replace strtok with strtok_r
> > >>   examples/l2fwd-crypto: replace strtok with strtok_r
> > >>   examples/vhost: replace strtok with strtok_r
> > >>
> > >>  app/graph/graph.c                             |  5 ++-
> > >>  app/graph/utils.c                             | 15 +++++---
> > >>  app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
> > >>  .../comp_perf_options_parse.c                 | 16 ++++----
> > >>  app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
> > >>  .../cperf_test_vector_parsing.c               | 10 +++--
> > >>  app/test-dma-perf/main.c                      | 13 ++++---
> > >>  app/test-fib/main.c                           | 10 ++---
> > >>  app/test-flow-perf/main.c                     | 22 ++++++-----
> > >>  app/test-mldev/ml_options.c                   | 18 ++++-----
> > >>  drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
> > >>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
> > >>  drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
> > >>  drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
> > >>  drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
> > >>  drivers/net/ark/ark_pktchkr.c                 | 10 ++---
> > >>  drivers/net/ark/ark_pktgen.c                  | 10 ++---
> > >>  drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
> > >>  examples/l2fwd-crypto/main.c                  |  6 +--
> > >>  examples/vhost/main.c                         |  3 +-
> > >>  lib/dmadev/rte_dmadev.c                       |  4 +-
> > >>  lib/eal/common/eal_common_memory.c            |  8 ++--
> > >>  lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
> > >>  lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
> > >>  lib/eventdev/rte_eventdev.c                   | 18 ++++-----
> > >>  lib/security/rte_security.c                   |  3 +-
> > >>  lib/telemetry/telemetry.c                     |  5 ++-
> > >>  27 files changed, 169 insertions(+), 140 deletions(-)
> > >>
> > >>-- 
> > >>2.30.0
> > >.

Re: [PATCH 00/21] replace strtok with strtok_r

[Tyler Retzlaff]

On Tue, Nov 14, 2023 at 09:34:33AM -0800, Tyler Retzlaff wrote:
> On Tue, Nov 14, 2023 at 09:32:48AM -0800, Tyler Retzlaff wrote:
> > On Tue, Nov 14, 2023 at 08:50:17PM +0800, Jie Hai wrote:
> > > On 2023/11/14 1:09, Tyler Retzlaff wrote:
> > > >On Mon, Nov 13, 2023 at 06:45:29PM +0800, Jie Hai wrote:
> > > >>Multiple threads calling the same function may cause condition
> > > >>race issues, which often leads to abnormal behavior and can cause
> > > >>more serious vulnerabilities such as abnormal termination, denial
> > > >>of service, and compromised data integrity.
> > > >>
> > > >>The strtok() is non-reentrant, it is better to replace it with a
> > > >>reentrant function.
> > > >
> > > >could you please use strtok_s instead of strtok_r the former is part of
> > > >the C11 standard the latter is not.
> > > >
> > > >thanks!
> > > >
> > > Hi, Tyler Retzlaff
> > > 
> > > Thanks for your comment.
> > > 
> > > For the use of strtok_s, I have consulted some documents, see
> > > 	https://en.cppreference.com/w/c/string/byte/strtok
> > > It says
> > > "As with all bounds-checked functions, strtok_s only guaranteed to be
> > > available if __STDC_LIB_EXT1__ is defined by the implementation and if
> > > the user defines __STDC_WANT_LIB_EXT1__ to the integer constant 1 before
> > > including <string.h>.
> > > "
> > > 
> > > I use strtok_s with "#ifdef __STDC_LIB_EXT1__ ... #endif" around and
> > > compiled
> > > locally and found that __STDC_LIB_EXT1__ was not defined, so the related
> > > code  was not called. I'm afraid there's a problem with this
> > > modification.
> > > 
> > > Am I using strtok_s wrong?
> > 
> > no i overlooked that it is optional my fault sorry.
> > 
> > since there is no portable strtok_r i'm going to have to agree with others
> > that only places where you actually need reentrant strtok be converted to
> > strtok_r.
> > 
> > for windows i think we'll either need to introduce an abstracted strtok_r
> > name for portability or something in the rte_ namespace (dependent on
> > what other revieweres would prefer)
> 
> just as a follow up here maybe it would be optimal if we could use
> strtok_s *if* we test that the toolchain makes it available and if not
> provide a mapping of strtok_s -> strtok_r.

just a final follow up, i can see that we already have a rte_strerror
here to do the replace with reentrant dance. it is probably good to
follow the already established pattern for this and have a rte_strtok.

what do others think?
 
> > 
> > thanks!
> > 
> > > 
> > > Thanks,
> > > Jie Hai
> > > >>
> > > >>Jie Hai (21):
> > > >>   app/graph: replace strtok with strtok_r
> > > >>   app/test-bbdev: replace strtok with strtok_r
> > > >>   app/test-compress-perf: replace strtok with strtok_r
> > > >>   app/test-crypto-perf: replace strtok with strtok_r
> > > >>   app/test-dma-perf: replace strtok with strtok_r
> > > >>   app/test-fib: replace strtok with strtok_r
> > > >>   app/dpdk-test-flow-perf: replace strtok with strtok_r
> > > >>   app/test-mldev: replace strtok with strtok_r
> > > >>   lib/dmadev: replace strtok with strtok_r
> > > >>   lib/eal: replace strtok with strtok_r
> > > >>   lib/ethdev: replace strtok with strtok_r
> > > >>   lib/eventdev: replace strtok with strtok_r
> > > >>   lib/telemetry: replace strtok with strtok_r
> > > >>   lib/telemetry: replace strtok with strtok_r
> > > >>   bus/fslmc: replace strtok with strtok_r
> > > >>   common/cnxk: replace strtok with strtok_r
> > > >>   event/cnxk: replace strtok with strtok_r
> > > >>   net/ark: replace strtok with strtok_r
> > > >>   raw/cnxk_gpio: replace strtok with strtok_r
> > > >>   examples/l2fwd-crypto: replace strtok with strtok_r
> > > >>   examples/vhost: replace strtok with strtok_r
> > > >>
> > > >>  app/graph/graph.c                             |  5 ++-
> > > >>  app/graph/utils.c                             | 15 +++++---
> > > >>  app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
> > > >>  .../comp_perf_options_parse.c                 | 16 ++++----
> > > >>  app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
> > > >>  .../cperf_test_vector_parsing.c               | 10 +++--
> > > >>  app/test-dma-perf/main.c                      | 13 ++++---
> > > >>  app/test-fib/main.c                           | 10 ++---
> > > >>  app/test-flow-perf/main.c                     | 22 ++++++-----
> > > >>  app/test-mldev/ml_options.c                   | 18 ++++-----
> > > >>  drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
> > > >>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
> > > >>  drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
> > > >>  drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
> > > >>  drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
> > > >>  drivers/net/ark/ark_pktchkr.c                 | 10 ++---
> > > >>  drivers/net/ark/ark_pktgen.c                  | 10 ++---
> > > >>  drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
> > > >>  examples/l2fwd-crypto/main.c                  |  6 +--
> > > >>  examples/vhost/main.c                         |  3 +-
> > > >>  lib/dmadev/rte_dmadev.c                       |  4 +-
> > > >>  lib/eal/common/eal_common_memory.c            |  8 ++--
> > > >>  lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
> > > >>  lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
> > > >>  lib/eventdev/rte_eventdev.c                   | 18 ++++-----
> > > >>  lib/security/rte_security.c                   |  3 +-
> > > >>  lib/telemetry/telemetry.c                     |  5 ++-
> > > >>  27 files changed, 169 insertions(+), 140 deletions(-)
> > > >>
> > > >>-- 
> > > >>2.30.0
> > > >.

Re: [PATCH 00/21] replace strtok with strtok_r

[fengchengwen]

On 2023/11/15 1:49, Tyler Retzlaff wrote:
> On Tue, Nov 14, 2023 at 09:34:33AM -0800, Tyler Retzlaff wrote:
>> On Tue, Nov 14, 2023 at 09:32:48AM -0800, Tyler Retzlaff wrote:
>>> On Tue, Nov 14, 2023 at 08:50:17PM +0800, Jie Hai wrote:
>>>> On 2023/11/14 1:09, Tyler Retzlaff wrote:
>>>>> On Mon, Nov 13, 2023 at 06:45:29PM +0800, Jie Hai wrote:
>>>>>> Multiple threads calling the same function may cause condition
>>>>>> race issues, which often leads to abnormal behavior and can cause
>>>>>> more serious vulnerabilities such as abnormal termination, denial
>>>>>> of service, and compromised data integrity.
>>>>>>
>>>>>> The strtok() is non-reentrant, it is better to replace it with a
>>>>>> reentrant function.
>>>>>
>>>>> could you please use strtok_s instead of strtok_r the former is part of
>>>>> the C11 standard the latter is not.
>>>>>
>>>>> thanks!
>>>>>
>>>> Hi, Tyler Retzlaff
>>>>
>>>> Thanks for your comment.
>>>>
>>>> For the use of strtok_s, I have consulted some documents, see
>>>> 	https://en.cppreference.com/w/c/string/byte/strtok
>>>> It says
>>>> "As with all bounds-checked functions, strtok_s only guaranteed to be
>>>> available if __STDC_LIB_EXT1__ is defined by the implementation and if
>>>> the user defines __STDC_WANT_LIB_EXT1__ to the integer constant 1 before
>>>> including <string.h>.
>>>> "
>>>>
>>>> I use strtok_s with "#ifdef __STDC_LIB_EXT1__ ... #endif" around and
>>>> compiled
>>>> locally and found that __STDC_LIB_EXT1__ was not defined, so the related
>>>> code  was not called. I'm afraid there's a problem with this
>>>> modification.
>>>>
>>>> Am I using strtok_s wrong?
>>>
>>> no i overlooked that it is optional my fault sorry.
>>>
>>> since there is no portable strtok_r i'm going to have to agree with others
>>> that only places where you actually need reentrant strtok be converted to
>>> strtok_r.
>>>
>>> for windows i think we'll either need to introduce an abstracted strtok_r
>>> name for portability or something in the rte_ namespace (dependent on
>>> what other revieweres would prefer)
>>
>> just as a follow up here maybe it would be optimal if we could use
>> strtok_s *if* we test that the toolchain makes it available and if not
>> provide a mapping of strtok_s -> strtok_r.
> 
> just a final follow up, i can see that we already have a rte_strerror
> here to do the replace with reentrant dance. it is probably good to
> follow the already established pattern for this and have a rte_strtok.

+1 for have rte_strtok which could cover different platform.

> 
> what do others think?
>  
>>>
>>> thanks!
>>>
>>>>
>>>> Thanks,
>>>> Jie Hai
>>>>>>
>>>>>> Jie Hai (21):
>>>>>>   app/graph: replace strtok with strtok_r
>>>>>>   app/test-bbdev: replace strtok with strtok_r
>>>>>>   app/test-compress-perf: replace strtok with strtok_r
>>>>>>   app/test-crypto-perf: replace strtok with strtok_r
>>>>>>   app/test-dma-perf: replace strtok with strtok_r
>>>>>>   app/test-fib: replace strtok with strtok_r
>>>>>>   app/dpdk-test-flow-perf: replace strtok with strtok_r
>>>>>>   app/test-mldev: replace strtok with strtok_r
>>>>>>   lib/dmadev: replace strtok with strtok_r
>>>>>>   lib/eal: replace strtok with strtok_r
>>>>>>   lib/ethdev: replace strtok with strtok_r
>>>>>>   lib/eventdev: replace strtok with strtok_r
>>>>>>   lib/telemetry: replace strtok with strtok_r
>>>>>>   lib/telemetry: replace strtok with strtok_r
>>>>>>   bus/fslmc: replace strtok with strtok_r
>>>>>>   common/cnxk: replace strtok with strtok_r
>>>>>>   event/cnxk: replace strtok with strtok_r
>>>>>>   net/ark: replace strtok with strtok_r
>>>>>>   raw/cnxk_gpio: replace strtok with strtok_r
>>>>>>   examples/l2fwd-crypto: replace strtok with strtok_r
>>>>>>   examples/vhost: replace strtok with strtok_r
>>>>>>
>>>>>>  app/graph/graph.c                             |  5 ++-
>>>>>>  app/graph/utils.c                             | 15 +++++---
>>>>>>  app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
>>>>>>  .../comp_perf_options_parse.c                 | 16 ++++----
>>>>>>  app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
>>>>>>  .../cperf_test_vector_parsing.c               | 10 +++--
>>>>>>  app/test-dma-perf/main.c                      | 13 ++++---
>>>>>>  app/test-fib/main.c                           | 10 ++---
>>>>>>  app/test-flow-perf/main.c                     | 22 ++++++-----
>>>>>>  app/test-mldev/ml_options.c                   | 18 ++++-----
>>>>>>  drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
>>>>>>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
>>>>>>  drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
>>>>>>  drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
>>>>>>  drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
>>>>>>  drivers/net/ark/ark_pktchkr.c                 | 10 ++---
>>>>>>  drivers/net/ark/ark_pktgen.c                  | 10 ++---
>>>>>>  drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
>>>>>>  examples/l2fwd-crypto/main.c                  |  6 +--
>>>>>>  examples/vhost/main.c                         |  3 +-
>>>>>>  lib/dmadev/rte_dmadev.c                       |  4 +-
>>>>>>  lib/eal/common/eal_common_memory.c            |  8 ++--
>>>>>>  lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
>>>>>>  lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
>>>>>>  lib/eventdev/rte_eventdev.c                   | 18 ++++-----
>>>>>>  lib/security/rte_security.c                   |  3 +-
>>>>>>  lib/telemetry/telemetry.c                     |  5 ++-
>>>>>>  27 files changed, 169 insertions(+), 140 deletions(-)
>>>>>>
>>>>>> -- 
>>>>>> 2.30.0
>>>>> .
> .
> 

RE: [PATCH 00/21] replace strtok with strtok_r

[Morten Brørup]

> From: fengchengwen [mailto:fengchengwen@huawei.com]
> Sent: Wednesday, 15 November 2023 04.03
> 
> On 2023/11/15 1:49, Tyler Retzlaff wrote:
> > On Tue, Nov 14, 2023 at 09:34:33AM -0800, Tyler Retzlaff wrote:
> >> On Tue, Nov 14, 2023 at 09:32:48AM -0800, Tyler Retzlaff wrote:
> >>> On Tue, Nov 14, 2023 at 08:50:17PM +0800, Jie Hai wrote:
> >>>> On 2023/11/14 1:09, Tyler Retzlaff wrote:
> >>>>> On Mon, Nov 13, 2023 at 06:45:29PM +0800, Jie Hai wrote:
> >>>>>> Multiple threads calling the same function may cause condition
> >>>>>> race issues, which often leads to abnormal behavior and can
> cause
> >>>>>> more serious vulnerabilities such as abnormal termination,
> denial
> >>>>>> of service, and compromised data integrity.
> >>>>>>
> >>>>>> The strtok() is non-reentrant, it is better to replace it with a
> >>>>>> reentrant function.
> >>>>>
> >>>>> could you please use strtok_s instead of strtok_r the former is
> part of
> >>>>> the C11 standard the latter is not.
> >>>>>
> >>>>> thanks!
> >>>>>
> >>>> Hi, Tyler Retzlaff
> >>>>
> >>>> Thanks for your comment.
> >>>>
> >>>> For the use of strtok_s, I have consulted some documents, see
> >>>> 	https://en.cppreference.com/w/c/string/byte/strtok
> >>>> It says
> >>>> "As with all bounds-checked functions, strtok_s only guaranteed to
> be
> >>>> available if __STDC_LIB_EXT1__ is defined by the implementation
> and if
> >>>> the user defines __STDC_WANT_LIB_EXT1__ to the integer constant 1
> before
> >>>> including <string.h>.
> >>>> "
> >>>>
> >>>> I use strtok_s with "#ifdef __STDC_LIB_EXT1__ ... #endif" around
> and
> >>>> compiled
> >>>> locally and found that __STDC_LIB_EXT1__ was not defined, so the
> related
> >>>> code  was not called. I'm afraid there's a problem with this
> >>>> modification.
> >>>>
> >>>> Am I using strtok_s wrong?
> >>>
> >>> no i overlooked that it is optional my fault sorry.
> >>>
> >>> since there is no portable strtok_r i'm going to have to agree with
> others
> >>> that only places where you actually need reentrant strtok be
> converted to
> >>> strtok_r.
> >>>
> >>> for windows i think we'll either need to introduce an abstracted
> strtok_r
> >>> name for portability or something in the rte_ namespace (dependent
> on
> >>> what other revieweres would prefer)
> >>
> >> just as a follow up here maybe it would be optimal if we could use
> >> strtok_s *if* we test that the toolchain makes it available and if
> not
> >> provide a mapping of strtok_s -> strtok_r.
> >
> > just a final follow up, i can see that we already have a rte_strerror
> > here to do the replace with reentrant dance. it is probably good to
> > follow the already established pattern for this and have a
> rte_strtok.
> 
> +1 for have rte_strtok which could cover different platform.

+1 to rte_strtok doing the reentrant dance for us.

If we had such an rte_strtok(), we could also generally disallow the use of strtok().

> 
> >
> > what do others think?
> >

Re: [PATCH 00/21] replace strtok with strtok_r

[Stephen Hemminger]

On Wed, 15 Nov 2023 12:27:37 +0100
Morten Brørup <mb@smartsharesystems.com> wrote:

> > > just a final follow up, i can see that we already have a rte_strerror
> > > here to do the replace with reentrant dance. it is probably good to
> > > follow the already established pattern for this and have a  
> > rte_strtok.
> > 
> > +1 for have rte_strtok which could cover different platform.  
> 
> +1 to rte_strtok doing the reentrant dance for us.
> 
> If we had such an rte_strtok(), we could also generally disallow the use of strtok().

Good idea, I like this.
Would be good to have a version on Windows as well.

Re: [PATCH 00/21] replace strtok with strtok_r

[Jie Hai]

On 2023/11/15 23:08, Stephen Hemminger wrote:
> On Wed, 15 Nov 2023 12:27:37 +0100
> Morten Brørup <mb@smartsharesystems.com> wrote:
> 
>>>> just a final follow up, i can see that we already have a rte_strerror
>>>> here to do the replace with reentrant dance. it is probably good to
>>>> follow the already established pattern for this and have a
>>> rte_strtok.
>>>
>>> +1 for have rte_strtok which could cover different platform.
>>
>> +1 to rte_strtok doing the reentrant dance for us.
>>
>> If we had such an rte_strtok(), we could also generally disallow the use of strtok().
> 
> Good idea, I like this.
> Would be good to have a version on Windows as well.
> 
> .
Hi, all maintainers,

Since the map of strtok_r to strtok_s with three patramaters has
been done in "rte_os_shim.h", I just include this file.
And the rte_strtok is defined as below.
My question is whether the
"strtok_s(s, stringlen, delim, save_ptr)" and
"strtok_s(str, delim, saveptr)" are compatible for C11.

And I check the glibc codes here,
https://sourceware.org/git/glibc.git
there is no definition of strtok_s, is strtok_s not in use?

If so, should we delayed processing for the C11 standard strtok_s 
function? That is, delete the "#ifdef __STDC_LIB_EXT1__...
#endif" part, but keep the extension of the four parameters for later 
use. Or keep the following change but never use stringlen as not null
until the function can be used.

What do you think?

diff --git a/lib/eal/include/rte_string_fns.h 
b/lib/eal/include/rte_string_fns.h
index bb43b2cba3eb..6746bfec384b 100644
--- a/lib/eal/include/rte_string_fns.h
+++ b/lib/eal/include/rte_string_fns.h
@@ -15,10 +15,13 @@
  extern "C" {
  #endif

+#define __STDC_WANT_LIB_EXT1__ 1
  #include <stdio.h>
  #include <string.h>

  #include <rte_common.h>
+#include <rte_compat.h>
+#include <rte_os_shim.h>

  /**
   * Takes string "string" parameter and splits it at character "delim"
@@ -115,6 +118,35 @@ rte_strlcat(char *dst, const char *src, size_t size)
  ssize_t
  rte_strscpy(char *dst, const char *src, size_t dsize);

+/*
+ * Divide string s into tokens separated by characters in delim.
+ * Information passed between calls are stored in save_ptr.
+ * The size of the string to be separated is stored in *stringlen.
+ *
+ * @param s
+ *   The string to be separated, it could be NULL.
+ *
+ * @param stringlen
+ *   The pointor to the size of the string to be separated, it could be 
NULL.
+ *
+ * @param delim
+ *   The characters on which the split of the data will be done.
+ *
+ * @param save_ptr
+ *   The internal state of the string separated.
+ */
+static inline char *
+rte_strtok(char *__rte_restrict s, size_t *__rte_restrict stringlen,
+          const char *__rte_restrict delim,
+          char **__rte_restrict save_ptr)
+{
+#ifdef __STDC_LIB_EXT1__
+	if (stringlen != NULL)
+		return strtok_s(s, stringlen, delim, save_ptr);

+#endif
+       (void)stringlen;
+       return strtok_r(s, delim, save_ptr);
+}
+
  #ifdef __cplusplus
  }
  #endif

Thanks,
Jie Hai

Re: [PATCH 00/21] replace strtok with strtok_r

[David Marchand]

On Tue, Nov 21, 2023 at 4:33 AM Jie Hai <haijie1@huawei.com> wrote:
>
> On 2023/11/15 23:08, Stephen Hemminger wrote:
> > On Wed, 15 Nov 2023 12:27:37 +0100
> > Morten Brørup <mb@smartsharesystems.com> wrote:
> >
> >>>> just a final follow up, i can see that we already have a rte_strerror
> >>>> here to do the replace with reentrant dance. it is probably good to
> >>>> follow the already established pattern for this and have a
> >>> rte_strtok.
> >>>
> >>> +1 for have rte_strtok which could cover different platform.
> >>
> >> +1 to rte_strtok doing the reentrant dance for us.
> >>
> >> If we had such an rte_strtok(), we could also generally disallow the use of strtok().
> >
> > Good idea, I like this.
> > Would be good to have a version on Windows as well.
> >
> > .
> Hi, all maintainers,
>
> Since the map of strtok_r to strtok_s with three patramaters has
> been done in "rte_os_shim.h", I just include this file.
> And the rte_strtok is defined as below.
> My question is whether the
> "strtok_s(s, stringlen, delim, save_ptr)" and
> "strtok_s(str, delim, saveptr)" are compatible for C11.
>
> And I check the glibc codes here,
> https://sourceware.org/git/glibc.git
> there is no definition of strtok_s, is strtok_s not in use?
>
> If so, should we delayed processing for the C11 standard strtok_s
> function? That is, delete the "#ifdef __STDC_LIB_EXT1__...
> #endif" part, but keep the extension of the four parameters for later
> use. Or keep the following change but never use stringlen as not null
> until the function can be used.
>
> What do you think?

Regardless of how it is done internally, I would not inline this helper.
This will minimise headaches with checks against toolchain/compiler
support in a DPDK header.


-- 
David Marchand

[PATCH v2 00/22] replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

This patchset replaces strtok with strtok_s in app, example, lib
and drivers. And adds check for use of strtok in checkpatches.sh.

--
v2:
1. fix commit log.
2. add check in checkpatches.sh.
3. replace strtok_r with strtok_s.
4. add Acked-by.
--

Jie Hai (22):
  app/graph: replace strtok with reentrant version
  app/bbdev: replace strtok with reentrant version
  app/compress-perf: replace strtok with reentrant version
  app/crypto-perf: replace strtok with reentrant version
  app/dma-perf: replace strtok with reentrant version
  app/test-fib: replace strtok with reentrant version
  app/flow-perf: replace strtok with reentrant version
  app/test-mldev: replace strtok with reentrant version
  dmadev: replace strtok with reentrant version
  eal: replace strtok with reentrant version
  ethdev: replace strtok with reentrant version
  eventdev: replace strtok with reentrant version
  security: replace strtok with reentrant version
  telemetry: replace strtok with reentrant version
  bus/fslmc: replace strtok with reentrant version
  common/cnxk: replace strtok with reentrant version
  event/cnxk: replace strtok with reentrant version
  net/ark: replace strtok with reentrant version
  raw/cnxk_gpio: replace strtok with reentrant version
  examples/l2fwd-crypto: replace strtok with reentrant version
  examples/vhost: replace strtok with reentrant version
  devtools: check for some reentrant function

 app/graph/graph.c                             |  5 ++-
 app/graph/utils.c                             | 15 +++++---
 app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
 .../comp_perf_options_parse.c                 | 16 ++++----
 app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
 .../cperf_test_vector_parsing.c               | 10 +++--
 app/test-dma-perf/main.c                      | 13 ++++---
 app/test-fib/main.c                           | 10 ++---
 app/test-flow-perf/main.c                     | 22 ++++++-----
 app/test-mldev/ml_options.c                   | 18 ++++-----
 devtools/checkpatches.sh                      |  8 ++++
 drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
 drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
 drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
 drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
 drivers/net/ark/ark_pktchkr.c                 | 10 ++---
 drivers/net/ark/ark_pktgen.c                  | 10 ++---
 drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
 examples/l2fwd-crypto/main.c                  |  6 +--
 examples/vhost/main.c                         |  3 +-
 lib/dmadev/rte_dmadev.c                       |  4 +-
 lib/eal/common/eal_common_memory.c            |  8 ++--
 lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
 lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
 lib/eventdev/rte_eventdev.c                   | 18 ++++-----
 lib/security/rte_security.c                   |  3 +-
 lib/telemetry/telemetry.c                     |  5 ++-
 28 files changed, 177 insertions(+), 140 deletions(-)

-- 
2.30.0

[PATCH v2 01/22] app/graph: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 5c59002a34f3 ("app/graph: add graph commands")
Fixes: 984a315a5804 ("app/graph: add parser utility")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/graph/graph.c |  5 +++--
 app/graph/utils.c | 15 +++++++++------
 2 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/app/graph/graph.c b/app/graph/graph.c
index a65723a196da..ed9405ea4bac 100644
--- a/app/graph/graph.c
+++ b/app/graph/graph.c
@@ -103,9 +103,10 @@ parser_usecases_read(char *usecases)
 {
 	bool valid = false;
 	uint32_t i, j = 0;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(usecases, ",");
+	token = strtok_s(usecases, ",", &sp);
 	while (token != NULL) {
 		for (i = 0; i < RTE_DIM(supported_usecases); i++) {
 			if (strcmp(supported_usecases[i], token) == 0) {
@@ -116,7 +117,7 @@ parser_usecases_read(char *usecases)
 				break;
 			}
 		}
-		token = strtok(NULL, ",");
+		token = strtok_s(NULL, ",", &sp);
 	}
 
 	return valid;
diff --git a/app/graph/utils.c b/app/graph/utils.c
index c7b6ae83cf1f..f7525f0b70af 100644
--- a/app/graph/utils.c
+++ b/app/graph/utils.c
@@ -101,13 +101,14 @@ int
 parser_ip4_read(uint32_t *value, char *p)
 {
 	uint8_t shift = 24;
+	char *sp = NULL;
 	uint32_t ip = 0;
 	char *token;
 
-	token = strtok(p, ".");
+	token = strtok_s(p, ".", &sp);
 	while (token != NULL) {
 		ip |= (((uint32_t)strtoul(token, NULL, 10)) << shift);
-		token = strtok(NULL, ".");
+		token = strtok_s(NULL, ".", &sp);
 		shift -= 8;
 	}
 
@@ -120,13 +121,14 @@ int
 parser_ip6_read(uint8_t *value, char *p)
 {
 	uint64_t val = 0;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(p, ":");
+	token = strtok_s(p, ":", &sp);
 	while (token != NULL) {
 		hex_string_to_uint64(&val, token);
 		*value = val;
-		token = strtok(NULL, ":");
+		token = strtok_s(NULL, ":", &sp);
 		value++;
 		val = 0;
 	}
@@ -139,13 +141,14 @@ parser_mac_read(uint64_t *value, char *p)
 {
 	uint64_t mac = 0, val = 0;
 	uint8_t shift = 40;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(p, ":");
+	token = strtok_s(p, ":", &sp);
 	while (token != NULL) {
 		hex_string_to_uint64(&val, token);
 		mac |= val << shift;
-		token = strtok(NULL, ":");
+		token = strtok_s(NULL, ":", &sp);
 		shift -= 8;
 		val = 0;
 	}
-- 
2.30.0

[PATCH v2 02/22] app/bbdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 0acdb9866756 ("test/bbdev: add FFT operations cases")
Fixes: f714a18885a6 ("app/testbbdev: add test application for bbdev")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-bbdev/test_bbdev_vector.c | 25 +++++++++++++++----------
 1 file changed, 15 insertions(+), 10 deletions(-)

diff --git a/app/test-bbdev/test_bbdev_vector.c b/app/test-bbdev/test_bbdev_vector.c
index c26727cd35c4..00a0f99801b7 100644
--- a/app/test-bbdev/test_bbdev_vector.c
+++ b/app/test-bbdev/test_bbdev_vector.c
@@ -63,8 +63,9 @@ parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
 
 	uint32_t *values, *values_resized;
 	char *tok, *error = NULL;
+	char *sp = NULL;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_s(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -98,7 +99,7 @@ parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
 
 		*data_length = *data_length + (strlen(tok) - strlen("0x"))/2;
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_s(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 
@@ -299,8 +300,9 @@ parse_turbo_flags(char *tokens, uint32_t *op_flags,
 {
 	char *tok = NULL;
 	uint32_t op_flag_value = 0;
+	char *sp = NULL;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_s(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -330,7 +332,7 @@ parse_turbo_flags(char *tokens, uint32_t *op_flags,
 
 		*op_flags = *op_flags | op_flag_value;
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_s(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 	}
@@ -368,9 +370,10 @@ static int
 parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
 {
 	char *tok = NULL;
+	char *sp = NULL;
 	bool status_ok = false;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_s(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -401,7 +404,7 @@ parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
 			return -1;
 		}
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_s(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 	}
@@ -894,6 +897,7 @@ parse_fft_params(const char *key_token, char *token,
 	int ret = 0, status = 0, i, shift;
 	uint32_t op_flags = 0;
 	char *tok, *err = NULL;
+	char *sp = NULL;
 
 	struct rte_bbdev_op_fft *fft = &vector->fft;
 
@@ -922,7 +926,7 @@ parse_fft_params(const char *key_token, char *token,
 		fft->output_leading_depadding = (uint32_t) strtoul(token, &err, 0);
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "window_index")) {
-		tok = strtok(token, VALUE_DELIMITER);
+		tok = strtok_s(token, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			return -1;
 		for (i = 0; i < FFT_WIN_SIZE; i++) {
@@ -930,7 +934,7 @@ parse_fft_params(const char *key_token, char *token,
 			fft->window_index[i / 2] |= (uint32_t) strtoul(tok, &err, 0)
 					<< shift;
 			if (i < (FFT_WIN_SIZE - 1)) {
-				tok = strtok(NULL, VALUE_DELIMITER);
+				tok = strtok_s(NULL, VALUE_DELIMITER, &sp);
 				if (tok == NULL)
 					return -1;
 			}
@@ -995,6 +999,7 @@ static int
 parse_entry(char *entry, struct test_bbdev_vector *vector)
 {
 	int ret = 0;
+	char *sp = NULL;
 	char *token, *key_token;
 	enum rte_bbdev_op_type op_type = RTE_BBDEV_OP_NONE;
 
@@ -1004,10 +1009,10 @@ parse_entry(char *entry, struct test_bbdev_vector *vector)
 	}
 
 	/* get key */
-	token = strtok(entry, ENTRY_DELIMITER);
+	token = strtok_s(entry, ENTRY_DELIMITER, &sp);
 	key_token = token;
 	/* get values for key */
-	token = strtok(NULL, ENTRY_DELIMITER);
+	token = strtok_s(NULL, ENTRY_DELIMITER, &sp);
 
 	if (key_token == NULL || token == NULL) {
 		printf("Expected 'key = values' but was '%.40s'..\n", entry);
-- 
2.30.0

[PATCH v2 03/22] app/compress-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: e0b6287c035d ("app/compress-perf: add parser")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-compress-perf/comp_perf_options_parse.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/app/test-compress-perf/comp_perf_options_parse.c b/app/test-compress-perf/comp_perf_options_parse.c
index 6d8c370fc2ea..a390fa36c56c 100644
--- a/app/test-compress-perf/comp_perf_options_parse.c
+++ b/app/test-compress-perf/comp_perf_options_parse.c
@@ -177,6 +177,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 {
 	char *token;
 	uint8_t number;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -184,7 +185,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ":");
+	token = strtok_s(copy_arg, ":", &sp);
 
 	/* Parse minimum value */
 	if (token != NULL) {
@@ -197,7 +198,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_s(NULL, ":", &sp);
 
 	/* Parse increment value */
 	if (token != NULL) {
@@ -211,7 +212,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_s(NULL, ":", &sp);
 
 	/* Parse maximum value */
 	if (token != NULL) {
@@ -225,7 +226,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	if (strtok(NULL, ":") != NULL)
+	if (strtok_s(NULL, ":", &sp) != NULL)
 		goto err_range;
 
 	free(copy_arg);
@@ -244,6 +245,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 	uint8_t count = 0;
 	uint32_t temp_min;
 	uint32_t temp_max;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -251,7 +253,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ",");
+	token = strtok_s(copy_arg, ",", &sp);
 
 	/* Parse first value */
 	if (token != NULL) {
@@ -266,7 +268,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 	} else
 		goto err_list;
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 
 	while (token != NULL) {
 		if (count == MAX_LIST) {
@@ -288,7 +290,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 		if (number > temp_max)
 			temp_max = number;
 
-		token = strtok(NULL, ",");
+		token = strtok_s(NULL, ",", &sp);
 	}
 
 	if (min)
-- 
2.30.0

[PATCH v2 04/22] app/crypto-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: f6cefe253cc8 ("app/crypto-perf: add range/list of sizes")
Fixes: f8be1786b1b8 ("app/crypto-perf: introduce performance test application")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-crypto-perf/cperf_options_parsing.c     | 16 +++++++++-------
 app/test-crypto-perf/cperf_test_vector_parsing.c | 10 ++++++----
 2 files changed, 15 insertions(+), 11 deletions(-)

diff --git a/app/test-crypto-perf/cperf_options_parsing.c b/app/test-crypto-perf/cperf_options_parsing.c
index 75afedc7fd6e..81fa4a925291 100644
--- a/app/test-crypto-perf/cperf_options_parsing.c
+++ b/app/test-crypto-perf/cperf_options_parsing.c
@@ -161,6 +161,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 {
 	char *token;
 	uint32_t number;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -168,7 +169,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ":");
+	token = strtok_s(copy_arg, ":", &sp);
 
 	/* Parse minimum value */
 	if (token != NULL) {
@@ -182,7 +183,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_s(NULL, ":", &sp);
 
 	/* Parse increment value */
 	if (token != NULL) {
@@ -196,7 +197,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_s(NULL, ":", &sp);
 
 	/* Parse maximum value */
 	if (token != NULL) {
@@ -211,7 +212,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	if (strtok(NULL, ":") != NULL)
+	if (strtok_s(NULL, ":", &sp) != NULL)
 		goto err_range;
 
 	free(copy_arg);
@@ -230,6 +231,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 	uint8_t count = 0;
 	uint32_t temp_min;
 	uint32_t temp_max;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -237,7 +239,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ",");
+	token = strtok_s(copy_arg, ",", &sp);
 
 	/* Parse first value */
 	if (token != NULL) {
@@ -253,7 +255,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 	} else
 		goto err_list;
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 
 	while (token != NULL) {
 		if (count == MAX_LIST) {
@@ -275,7 +277,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 		if (number > temp_max)
 			temp_max = number;
 
-		token = strtok(NULL, ",");
+		token = strtok_s(NULL, ",", &sp);
 	}
 
 	if (min)
diff --git a/app/test-crypto-perf/cperf_test_vector_parsing.c b/app/test-crypto-perf/cperf_test_vector_parsing.c
index 737d61d4af6b..3557d767b55c 100644
--- a/app/test-crypto-perf/cperf_test_vector_parsing.c
+++ b/app/test-crypto-perf/cperf_test_vector_parsing.c
@@ -220,8 +220,9 @@ parse_values(char *tokens, uint8_t **data, uint32_t *data_length)
 
 	uint8_t *values, *values_resized;
 	char *tok, *error = NULL;
+	char *sp = NULL;
 
-	tok = strtok(tokens, CPERF_VALUE_DELIMITER);
+	tok = strtok_s(tokens, CPERF_VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -252,7 +253,7 @@ parse_values(char *tokens, uint8_t **data, uint32_t *data_length)
 			return -1;
 		}
 
-		tok = strtok(NULL, CPERF_VALUE_DELIMITER);
+		tok = strtok_s(NULL, CPERF_VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 
@@ -283,6 +284,7 @@ parse_entry(char *entry, struct cperf_test_vector *vector,
 
 	uint8_t *data = NULL;
 	char *token, *key_token;
+	char *sp = NULL;
 
 	if (entry == NULL) {
 		printf("Expected entry value\n");
@@ -290,10 +292,10 @@ parse_entry(char *entry, struct cperf_test_vector *vector,
 	}
 
 	/* get key */
-	token = strtok(entry, CPERF_ENTRY_DELIMITER);
+	token = strtok_s(entry, CPERF_ENTRY_DELIMITER, &sp);
 	key_token = token;
 	/* get values for key */
-	token = strtok(NULL, CPERF_ENTRY_DELIMITER);
+	token = strtok_s(NULL, CPERF_ENTRY_DELIMITER, &sp);
 
 	if (key_token == NULL || token == NULL) {
 		printf("Expected 'key = values' but was '%.40s'..\n", entry);
-- 
2.30.0

RE: [PATCH v2 04/22] app/crypto-perf: replace strtok with reentrant version

[Power, Ciara]

Hi Jie Hai,

> -----Original Message-----
> From: Jie Hai <haijie1@huawei.com>
> Sent: Tuesday, November 14, 2023 8:41 AM
> To: dev@dpdk.org; Power, Ciara <ciara.power@intel.com>; Sergio Gonzalez
> Monroy <sergio.gonzalez.monroy@intel.com>; De Lara Guarch, Pablo
> <pablo.de.lara.guarch@intel.com>; Piotr Azarewicz
> <piotr.azarewicz@intel.com>; Kobylinski, Michal
> <michal.kobylinski@intel.com>; Mrozowicz, SlawomirX
> <slawomirx.mrozowicz@intel.com>; Kerlin, Marcin <marcin.kerlin@intel.com>;
> Doherty, Declan <declan.doherty@intel.com>
> Cc: haijie1@huawei.com; lihuisong@huawei.com;
> fengchengwen@huawei.com
> Subject: [PATCH v2 04/22] app/crypto-perf: replace strtok with reentrant
> version
> 
> Multiple threads calling the same function may cause condition race issues,
> which often leads to abnormal behavior and can cause more serious
> vulnerabilities such as abnormal termination, denial of service, and
> compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a reentrant version.
> 
> Fixes: f6cefe253cc8 ("app/crypto-perf: add range/list of sizes")
> Fixes: f8be1786b1b8 ("app/crypto-perf: introduce performance test
> application")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>

Acked-by: Ciara Power <ciara.power@intel.com>

[PATCH v2 05/22] app/dma-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 623dc9364dc6 ("app/dma-perf: introduce DMA performance test")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-dma-perf/main.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/app/test-dma-perf/main.c b/app/test-dma-perf/main.c
index e5bccc27da5e..e34f9f8aaa7c 100644
--- a/app/test-dma-perf/main.c
+++ b/app/test-dma-perf/main.c
@@ -164,6 +164,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 	uint16_t len;
 	char *input;
 	struct lcore_dma_map_t *lcore_dma_map;
+	char *sp = NULL;
 
 	if (test_case == NULL || value == NULL)
 		return -1;
@@ -175,7 +176,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 
 	memset(lcore_dma_map, 0, sizeof(struct lcore_dma_map_t));
 
-	char *token = strtok(input, ", ");
+	char *token = strtok_s(input, ", ", &sp);
 	while (token != NULL) {
 		if (lcore_dma_map->cnt >= MAX_LCORE_NB) {
 			free(input);
@@ -185,7 +186,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 		uint16_t lcore_id = atoi(token);
 		lcore_dma_map->lcores[lcore_dma_map->cnt++] = lcore_id;
 
-		token = strtok(NULL, ", ");
+		token = strtok_s(NULL, ", ", &sp);
 	}
 
 	free(input);
@@ -201,6 +202,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 	char *start, *end, *substr;
 	uint16_t lcore_id;
 	int ret = 0;
+	char *sp = NULL;
 
 	if (test_case == NULL || value == NULL)
 		return -1;
@@ -216,7 +218,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 		goto out;
 	}
 
-	substr = strtok(addrs, ",");
+	substr = strtok_s(addrs, ",", &sp);
 	if (substr == NULL) {
 		fprintf(stderr, "No input DMA address\n");
 		ret = -1;
@@ -258,7 +260,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 		strlcpy(lcore_dma_map->dma_names[lcore_dma_map->cnt], ptrs[1],
 				RTE_DEV_NAME_MAX_LEN);
 		lcore_dma_map->cnt++;
-		substr = strtok(NULL, ",");
+		substr = strtok_s(NULL, ",", &sp);
 	} while (substr != NULL);
 
 out:
@@ -486,6 +488,7 @@ main(int argc, char *argv[])
 	char *rst_path_ptr = NULL;
 	char rst_path[PATH_MAX];
 	int new_argc;
+	char *sp = NULL;
 
 	memset(args, 0, sizeof(args));
 
@@ -504,7 +507,7 @@ main(int argc, char *argv[])
 	}
 	if (rst_path_ptr == NULL) {
 		strlcpy(rst_path, cfg_path_ptr, PATH_MAX);
-		char *token = strtok(basename(rst_path), ".");
+		char *token = strtok_s(basename(rst_path), ".", &sp);
 		if (token == NULL) {
 			printf("Config file error.\n");
 			return -1;
-- 
2.30.0

[PATCH v2 06/22] app/test-fib: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 103809d032cd ("app/test-fib: add test application for FIB")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-fib/main.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/app/test-fib/main.c b/app/test-fib/main.c
index 75a56135f212..95f6bd467b49 100644
--- a/app/test-fib/main.c
+++ b/app/test-fib/main.c
@@ -223,9 +223,9 @@ parse_distrib(uint8_t depth_lim, const uint32_t n)
 	uint32_t nrpd[128 + 1] = {0}; /* number of routes per depth */
 	uint32_t n_routes;
 	uint8_t depth, ratio, ratio_acc = 0;
-	char *in;
+	char *in, *sp = NULL;
 
-	in = strtok(distrib_string, ",");
+	in = strtok_s(distrib_string, ",", &sp);
 
 	/*parse configures routes percentage ratios*/
 	while (in != NULL) {
@@ -265,7 +265,7 @@ parse_distrib(uint8_t depth_lim, const uint32_t n)
 		}
 
 		/*number of configured depths in*/
-		in = strtok(NULL, ",");
+		in = strtok_s(NULL, ",", &sp);
 	}
 
 	if (ratio_acc > 100) {
@@ -542,10 +542,10 @@ parse_lookup(FILE *f, int af)
 	int ret, i = 0;
 	uint8_t *tbl = (uint8_t *)config.lookup_tbl;
 	int step = (af == AF_INET) ? 4 : 16;
-	char *s;
+	char *s, *sp = NULL;
 
 	while (fgets(line, sizeof(line), f) != NULL) {
-		s = strtok(line, " \t\n");
+		s = strtok_s(line, " \t\n", &sp);
 		if (s == NULL)
 			return -EINVAL;
 		ret = inet_pton(af, s, &tbl[i]);
-- 
2.30.0

[PATCH v2 07/22] app/flow-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 0c8f1f4ab90e ("app/flow-perf: support raw encap/decap actions")
Fixes: 7f37f0936a19 ("app/flow-perf: support meter policy API")
Fixes: 80a323319745 ("app/flow-perf: add destination ports parameter")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-flow-perf/main.c | 22 ++++++++++++----------
 1 file changed, 12 insertions(+), 10 deletions(-)

diff --git a/app/test-flow-perf/main.c b/app/test-flow-perf/main.c
index e224ef67983d..04ce7e38a539 100644
--- a/app/test-flow-perf/main.c
+++ b/app/test-flow-perf/main.c
@@ -602,6 +602,7 @@ read_meter_policy(char *prog, char *arg)
 {
 	char *token;
 	size_t i, j, k;
+	char *sp = NULL;
 
 	j = 0;
 	k = 0;
@@ -612,9 +613,9 @@ read_meter_policy(char *prog, char *arg)
 		token = strsep(&arg, ":\0");
 	}
 	j = 0;
-	token = strtok(actions_str[0], ",\0");
+	token = strtok_s(actions_str[0], ",\0", &sp);
 	while (token == NULL && j < RTE_COLORS - 1)
-		token = strtok(actions_str[++j], ",\0");
+		token = strtok_s(actions_str[++j], ",\0", &sp);
 	while (j < RTE_COLORS && token != NULL) {
 		for (i = 0; i < RTE_DIM(flow_options); i++) {
 			if (!strcmp(token, flow_options[i].str)) {
@@ -628,9 +629,9 @@ read_meter_policy(char *prog, char *arg)
 			usage(prog);
 			rte_exit(EXIT_SUCCESS, "Invalid colored actions\n");
 		}
-		token = strtok(NULL, ",\0");
+		token = strtok_s(NULL, ",\0", &sp);
 		while (!token && j < RTE_COLORS - 1) {
-			token = strtok(actions_str[++j], ",\0");
+			token = strtok_s(actions_str[++j], ",\0", &sp);
 			k = 0;
 		}
 	}
@@ -641,6 +642,7 @@ args_parse(int argc, char **argv)
 {
 	uint64_t pm, seed;
 	uint64_t hp_conf;
+	char *sp = NULL;
 	char **argvopt;
 	uint32_t prio;
 	char *token;
@@ -804,7 +806,7 @@ args_parse(int argc, char **argv)
 						RTE_FLOW_ACTION_TYPE_RAW_ENCAP
 					);
 
-				token = strtok(optarg, ",");
+				token = strtok_s(optarg, ",", &sp);
 				while (token != NULL) {
 					for (i = 0; i < RTE_DIM(flow_options); i++) {
 						if (strcmp(flow_options[i].str, token) == 0) {
@@ -817,7 +819,7 @@ args_parse(int argc, char **argv)
 							rte_exit(EXIT_FAILURE,
 								"Invalid encap item: %s\n", token);
 					}
-					token = strtok(NULL, ",");
+					token = strtok_s(NULL, ",", &sp);
 				}
 				printf(" / ");
 			}
@@ -828,7 +830,7 @@ args_parse(int argc, char **argv)
 						RTE_FLOW_ACTION_TYPE_RAW_DECAP
 					);
 
-				token = strtok(optarg, ",");
+				token = strtok_s(optarg, ",", &sp);
 				while (token != NULL) {
 					for (i = 0; i < RTE_DIM(flow_options); i++) {
 						if (strcmp(flow_options[i].str, token) == 0) {
@@ -841,7 +843,7 @@ args_parse(int argc, char **argv)
 							rte_exit(EXIT_FAILURE,
 								"Invalid decap item %s\n", token);
 					}
-					token = strtok(NULL, ",");
+					token = strtok_s(NULL, ",", &sp);
 				}
 				printf(" / ");
 			}
@@ -910,10 +912,10 @@ args_parse(int argc, char **argv)
 				uint16_t port_idx = 0;
 				char *token;
 
-				token = strtok(optarg, ",");
+				token = strtok_s(optarg, ",", &sp);
 				while (token != NULL) {
 					dst_ports[port_idx++] = atoi(token);
-					token = strtok(NULL, ",");
+					token = strtok_s(NULL, ",", &sp);
 				}
 			}
 			if (strcmp(lgopts[opt_idx].name, "rxq") == 0) {
-- 
2.30.0

[PATCH v2 08/22] app/test-mldev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-mldev/ml_options.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/app/test-mldev/ml_options.c b/app/test-mldev/ml_options.c
index 72357c1c393d..3e93a8b7356c 100644
--- a/app/test-mldev/ml_options.c
+++ b/app/test-mldev/ml_options.c
@@ -75,12 +75,12 @@ ml_parse_models(struct ml_options *opt, const char *arg)
 {
 	const char *delim = ",";
 	char models[PATH_MAX];
-	char *token;
+	char *token, *sp = NULL;
 	int ret = 0;
 
 	strlcpy(models, arg, PATH_MAX);
 
-	token = strtok(models, delim);
+	token = strtok_s(models, delim, &sp);
 	while (token != NULL) {
 		strlcpy(opt->filelist[opt->nb_filelist].model, token, PATH_MAX);
 		opt->nb_filelist++;
@@ -90,7 +90,7 @@ ml_parse_models(struct ml_options *opt, const char *arg)
 			ret = -EINVAL;
 			break;
 		}
-		token = strtok(NULL, delim);
+		token = strtok_s(NULL, delim, &sp);
 	}
 
 	if (opt->nb_filelist == 0) {
@@ -106,7 +106,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 {
 	const char *delim = ",";
 	char filelist[PATH_MAX];
-	char *token;
+	char *token, *sp = NULL;
 
 	if (opt->nb_filelist >= ML_TEST_MAX_MODELS) {
 		ml_err("Exceeded filelist count, max = %d\n", ML_TEST_MAX_MODELS);
@@ -116,7 +116,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(filelist, arg, PATH_MAX);
 
 	/* model */
-	token = strtok(filelist, delim);
+	token = strtok_s(filelist, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, model not specified = %s\n", arg);
 		return -EINVAL;
@@ -124,7 +124,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].model, token, PATH_MAX);
 
 	/* input */
-	token = strtok(NULL, delim);
+	token = strtok_s(NULL, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, input not specified = %s\n", arg);
 		return -EINVAL;
@@ -132,7 +132,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].input, token, PATH_MAX);
 
 	/* output */
-	token = strtok(NULL, delim);
+	token = strtok_s(NULL, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, output not specified = %s\n", arg);
 		return -EINVAL;
@@ -140,14 +140,14 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].output, token, PATH_MAX);
 
 	/* reference - optional */
-	token = strtok(NULL, delim);
+	token = strtok_s(NULL, delim, &sp);
 	if (token != NULL)
 		strlcpy(opt->filelist[opt->nb_filelist].reference, token, PATH_MAX);
 	else
 		memset(opt->filelist[opt->nb_filelist].reference, 0, PATH_MAX);
 
 	/* check for extra tokens */
-	token = strtok(NULL, delim);
+	token = strtok_s(NULL, delim, &sp);
 	if (token != NULL) {
 		ml_err("Invalid filelist. Entries > 4\n.");
 		return -EINVAL;
-- 
2.30.0

[PATCH v2 09/22] dmadev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 39b5ab60df30 ("dmadev: add telemetry")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/dmadev/rte_dmadev.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/dmadev/rte_dmadev.c b/lib/dmadev/rte_dmadev.c
index 4e5e420c82a5..e9c052304393 100644
--- a/lib/dmadev/rte_dmadev.c
+++ b/lib/dmadev/rte_dmadev.c
@@ -971,7 +971,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	struct rte_dma_info dma_info;
 	struct rte_dma_stats dma_stats;
 	int dev_id, ret, vchan_id;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *vchan_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -990,7 +990,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	if (dma_info.nb_vchans == 1 && *end_param == '\0')
 		vchan_id = 0;
 	else {
-		vchan_param = strtok(end_param, ",");
+		vchan_param = strtok_s(end_param, ",", &sp);
 		if (!vchan_param || strlen(vchan_param) == 0 || !isdigit(*vchan_param))
 			return -EINVAL;
 
-- 
2.30.0

[PATCH v2 10/22] eal: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/eal/common/eal_common_memory.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/eal/common/eal_common_memory.c b/lib/eal/common/eal_common_memory.c
index d9433db62345..327eb0f65289 100644
--- a/lib/eal/common/eal_common_memory.c
+++ b/lib/eal/common/eal_common_memory.c
@@ -1273,22 +1273,22 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	char dlim[2] = ",";
 	char *params_args;
 	size_t count = 0;
-	char *token;
+	char *token, *sp = NULL;
 
 	if (vals == NULL || params == NULL || strlen(params) == 0)
 		return -1;
 
-	/* strtok expects char * and param is const char *. Hence on using
+	/* strtok_s expects char * and param is const char *. Hence on using
 	 * params as "const char *" compiler throws warning.
 	 */
 	params_args = strdup(params);
 	if (params_args == NULL)
 		return -1;
 
-	token = strtok(params_args, dlim);
+	token = strtok_s(params_args, dlim, &sp);
 	while (token && isdigit(*token) && count < n_vals) {
 		vals[count++] = strtoul(token, NULL, 10);
-		token = strtok(NULL, dlim);
+		token = strtok_s(NULL, dlim, &sp);
 	}
 
 	free(params_args);
-- 
2.30.0

[PATCH v2 11/22] ethdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: f38f62650f7b ("ethdev: add Rx queue telemetry query")
Fixes: 9e7533aeb80a ("ethdev: add telemetry command for TM level capabilities")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/ethdev/rte_ethdev_telemetry.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/lib/ethdev/rte_ethdev_telemetry.c b/lib/ethdev/rte_ethdev_telemetry.c
index b01028ce9b60..658f7e54de03 100644
--- a/lib/ethdev/rte_ethdev_telemetry.c
+++ b/lib/ethdev/rte_ethdev_telemetry.c
@@ -477,6 +477,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	const char *qid_param;
 	uint16_t nb_queues;
 	char *end_param;
+	char *sp = NULL;
 	uint64_t qid;
 	int ret;
 
@@ -489,7 +490,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	if (nb_queues == 1 && *end_param == '\0')
 		qid = 0;
 	else {
-		qid_param = strtok(end_param, ",");
+		qid_param = strtok_s(end_param, ",", &sp);
 		if (!qid_param || strlen(qid_param) == 0 || !isdigit(*qid_param))
 			return -EINVAL;
 
@@ -1221,9 +1222,10 @@ static int
 eth_dev_parse_tm_params(char *params, uint32_t *result)
 {
 	const char *splited_param;
+	char *sp = NULL;
 	uint64_t ret;
 
-	splited_param = strtok(params, ",");
+	splited_param = strtok_s(params, ",", &sp);
 	if (!splited_param || strlen(splited_param) == 0 || !isdigit(*splited_param))
 		return -EINVAL;
 
-- 
2.30.0

[PATCH v2 12/22] eventdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/eventdev/rte_event_eth_rx_adapter.c | 38 ++++++++++++-------------
 lib/eventdev/rte_eventdev.c             | 18 ++++++------
 2 files changed, 28 insertions(+), 28 deletions(-)

diff --git a/lib/eventdev/rte_event_eth_rx_adapter.c b/lib/eventdev/rte_event_eth_rx_adapter.c
index 6db03adf0463..91e535e983d3 100644
--- a/lib/eventdev/rte_event_eth_rx_adapter.c
+++ b/lib/eventdev/rte_event_eth_rx_adapter.c
@@ -3651,7 +3651,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp;
 	struct rte_event_eth_rx_adapter_queue_conf queue_conf;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3661,19 +3661,19 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_s(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3684,7 +3684,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_s(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3723,7 +3723,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 	struct rte_event_eth_rx_adapter_queue_stats q_stats;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3733,19 +3733,19 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_s(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3756,7 +3756,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_s(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3794,7 +3794,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3803,19 +3803,19 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_s(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3826,7 +3826,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_s(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3855,7 +3855,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	uint8_t instance_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3863,14 +3863,14 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_s(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_s(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3881,7 +3881,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_s(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
diff --git a/lib/eventdev/rte_eventdev.c b/lib/eventdev/rte_eventdev.c
index 0ca32d672175..9e753c3720c0 100644
--- a/lib/eventdev/rte_eventdev.c
+++ b/lib/eventdev/rte_eventdev.c
@@ -1784,7 +1784,7 @@ handle_queue_links(const char *cmd __rte_unused,
 		   struct rte_tel_data *d)
 {
 	int i, ret, port_id = 0;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	uint8_t dev_id;
 	uint8_t queues[RTE_EVENT_MAX_QUEUES_PER_DEV];
 	uint8_t priorities[RTE_EVENT_MAX_QUEUES_PER_DEV];
@@ -1797,12 +1797,12 @@ handle_queue_links(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_s(end_param, ",", &sp);
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
 		return -1;
 
 	port_id = strtoul(p_param, &end_param, 10);
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_s(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1922,7 +1922,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1932,7 +1932,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_s(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_PORT;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1940,7 +1940,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_s(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1956,7 +1956,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1966,7 +1966,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_s(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_QUEUE;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1974,7 +1974,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_s(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
-- 
2.30.0

[PATCH v2 13/22] security: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 259ca6d1617f ("security: add telemetry endpoint for capabilities")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/security/rte_security.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/lib/security/rte_security.c b/lib/security/rte_security.c
index b082a290296b..51e3a669e6ec 100644
--- a/lib/security/rte_security.c
+++ b/lib/security/rte_security.c
@@ -496,13 +496,14 @@ security_handle_cryptodev_crypto_caps(const char *cmd __rte_unused, const char *
 	int dev_id, capa_id;
 	int crypto_caps_n;
 	char *end_param;
+	char *sp = NULL;
 	int rc;
 
 	if (!params || strlen(params) == 0 || !isdigit(*params))
 		return -EINVAL;
 
 	dev_id = strtoul(params, &end_param, 0);
-	capa_param = strtok(end_param, ",");
+	capa_param = strtok_s(end_param, ",", &sp);
 	if (!capa_param || strlen(capa_param) == 0 || !isdigit(*capa_param))
 		return -EINVAL;
 
-- 
2.30.0

[PATCH v2 14/22] telemetry: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 6dd571fd07c3 ("telemetry: introduce new functionality")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/telemetry/telemetry.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/lib/telemetry/telemetry.c b/lib/telemetry/telemetry.c
index 92982842a860..ef5cc87c34d1 100644
--- a/lib/telemetry/telemetry.c
+++ b/lib/telemetry/telemetry.c
@@ -371,6 +371,7 @@ static void *
 client_handler(void *sock_id)
 {
 	int s = (int)(uintptr_t)sock_id;
+	char *sp = NULL;
 	char buffer[1024];
 	char info_str[1024];
 	snprintf(info_str, sizeof(info_str),
@@ -385,8 +386,8 @@ client_handler(void *sock_id)
 	int bytes = read(s, buffer, sizeof(buffer) - 1);
 	while (bytes > 0) {
 		buffer[bytes] = 0;
-		const char *cmd = strtok(buffer, ",");
-		const char *param = strtok(NULL, "\0");
+		const char *cmd = strtok_s(buffer, ",", &sp);
+		const char *param = strtok_s(NULL, "\0", &sp);
 		telemetry_cb fn = unknown_command;
 		int i;
 
-- 
2.30.0

RE: [PATCH v2 14/22] telemetry: replace strtok with reentrant version

[Power, Ciara]

> -----Original Message-----
> From: Jie Hai <haijie1@huawei.com>
> Sent: Tuesday, November 14, 2023 8:41 AM
> To: dev@dpdk.org; Power, Ciara <ciara.power@intel.com>; Wiles, Keith
> <keith.wiles@intel.com>; Richardson, Bruce <bruce.richardson@intel.com>
> Cc: haijie1@huawei.com; lihuisong@huawei.com;
> fengchengwen@huawei.com
> Subject: [PATCH v2 14/22] telemetry: replace strtok with reentrant version
> 
> Multiple threads calling the same function may cause condition race issues,
> which often leads to abnormal behavior and can cause more serious
> vulnerabilities such as abnormal termination, denial of service, and
> compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a reentrant version.
> 
> Fixes: 6dd571fd07c3 ("telemetry: introduce new functionality")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> ---

Acked-by: Ciara Power <ciara.power@intel.com>

[PATCH v2 15/22] bus/fslmc: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 9ccb76b24c1d ("bus/fslmc: enable portal interrupt handling")
Fixes: 828d51d8fc3e ("bus/fslmc: refactor scan and probe functions")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/bus/fslmc/fslmc_bus.c            | 5 +++--
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 4 ++--
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index 57bfb5111a97..2aa0e45c3d1b 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -131,6 +131,7 @@ scan_one_fslmc_device(char *dev_name)
 {
 	char *dup_dev_name, *t_ptr;
 	struct rte_dpaa2_device *dev = NULL;
+	char *sp = NULL;
 	int ret = -1;
 
 	if (!dev_name)
@@ -168,7 +169,7 @@ scan_one_fslmc_device(char *dev_name)
 	}
 
 	/* Parse the device name and ID */
-	t_ptr = strtok(dup_dev_name, ".");
+	t_ptr = strtok_s(dup_dev_name, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Invalid device found: (%s)", dup_dev_name);
 		ret = -EINVAL;
@@ -199,7 +200,7 @@ scan_one_fslmc_device(char *dev_name)
 	else
 		dev->dev_type = DPAA2_UNKNOWN;
 
-	t_ptr = strtok(NULL, ".");
+	t_ptr = strtok_s(NULL, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Skipping invalid device (%s)", dup_dev_name);
 		ret = 0;
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index 4aec7b2cd8ba..b788d7681866 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -129,7 +129,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 	uint32_t cpu_mask = 1;
 	int ret;
 	size_t len = 0;
-	char *temp = NULL, *token = NULL;
+	char *temp = NULL, *token = NULL, *sp = NULL;
 	char string[STRING_LEN], command[COMMAND_LEN];
 	FILE *file;
 
@@ -141,7 +141,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 	}
 	while (getline(&temp, &len, file) != -1) {
 		if ((strstr(temp, string)) != NULL) {
-			token = strtok(temp, ":");
+			token = strtok_s(temp, ":", &sp);
 			break;
 		}
 	}
-- 
2.30.0

[PATCH v2 16/22] common/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: af75aac78978 ("common/cnxk: support telemetry for NIX")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/common/cnxk/cnxk_telemetry_nix.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/common/cnxk/cnxk_telemetry_nix.c b/drivers/common/cnxk/cnxk_telemetry_nix.c
index ccae5d7853af..eff1517951e9 100644
--- a/drivers/common/cnxk/cnxk_telemetry_nix.c
+++ b/drivers/common/cnxk/cnxk_telemetry_nix.c
@@ -761,7 +761,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			   struct plt_tel_data *d)
 {
 	struct nix_tel_node *node;
-	char *name, *param;
+	char *name, *param, *sp = NULL;
 	char buf[1024];
 	int rc = -1;
 
@@ -769,11 +769,11 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 		goto exit;
 
 	plt_strlcpy(buf, params, PCI_PRI_STR_SIZE + 1);
-	name = strtok(buf, ",");
+	name = strtok_s(buf, ",", &sp);
 	if (name == NULL)
 		goto exit;
 
-	param = strtok(NULL, "\0");
+	param = strtok_s(NULL, "\0", &sp);
 
 	node = nix_tel_node_get_by_pcidev_name(name);
 	if (!node)
@@ -782,7 +782,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 	plt_tel_data_start_dict(d);
 
 	if (strstr(cmd, "rq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_s(param, ",", &sp);
 		int rq;
 
 		if (!tok)
@@ -798,7 +798,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_rq(node->rqs[rq], d);
 
 	} else if (strstr(cmd, "cq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_s(param, ",", &sp);
 		int cq;
 
 		if (!tok)
@@ -814,7 +814,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_cq(node->cqs[cq], d);
 
 	} else if (strstr(cmd, "sq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_s(param, ",", &sp);
 		int sq;
 
 		if (!tok)
-- 
2.30.0

[PATCH v2 17/22] event/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/event/cnxk/cnxk_eventdev.c  | 10 ++++++----
 drivers/event/cnxk/cnxk_tim_evdev.c | 11 ++++++-----
 2 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/drivers/event/cnxk/cnxk_eventdev.c b/drivers/event/cnxk/cnxk_eventdev.c
index 0c61f4c20eec..1067c403ee02 100644
--- a/drivers/event/cnxk/cnxk_eventdev.c
+++ b/drivers/event/cnxk/cnxk_eventdev.c
@@ -478,7 +478,8 @@ parse_queue_param(char *value, void *opaque)
 	struct cnxk_sso_qos queue_qos = {0};
 	uint16_t *val = (uint16_t *)&queue_qos;
 	struct cnxk_sso_evdev *dev = opaque;
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_s(value, "-", &sp);
 	struct cnxk_sso_qos *old_ptr;
 
 	if (!strlen(value))
@@ -486,7 +487,7 @@ parse_queue_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_s(NULL, "-", &sp);
 		val++;
 	}
 
@@ -514,7 +515,8 @@ parse_stash_param(char *value, void *opaque)
 	struct cnxk_sso_stash queue_stash = {0};
 	struct cnxk_sso_evdev *dev = opaque;
 	struct cnxk_sso_stash *old_ptr;
-	char *tok = strtok(value, "|");
+	char *sp = NULL;
+	char *tok = strtok_s(value, "|", &sp);
 	uint16_t *val;
 
 	if (!strlen(value))
@@ -523,7 +525,7 @@ parse_stash_param(char *value, void *opaque)
 	val = (uint16_t *)&queue_stash;
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "|");
+		tok = strtok_s(NULL, "|", &sp);
 		val++;
 	}
 
diff --git a/drivers/event/cnxk/cnxk_tim_evdev.c b/drivers/event/cnxk/cnxk_tim_evdev.c
index 6d59fdf90983..82975ec8c6e7 100644
--- a/drivers/event/cnxk/cnxk_tim_evdev.c
+++ b/drivers/event/cnxk/cnxk_tim_evdev.c
@@ -420,7 +420,8 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 {
 	struct cnxk_tim_evdev *dev = opaque;
 	struct cnxk_tim_ctl ring_ctl = {0};
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_s(value, "-", &sp);
 	struct cnxk_tim_ctl *old_ptr;
 	uint16_t *val;
 
@@ -431,7 +432,7 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_s(NULL, "-", &sp);
 		val++;
 	}
 
@@ -507,16 +508,16 @@ cnxk_tim_parse_clk_list(const char *value, void *opaque)
 				      ROC_TIM_CLK_SRC_INVALID};
 	struct cnxk_tim_evdev *dev = opaque;
 	char *str = strdup(value);
-	char *tok;
+	char *tok, *sp = NULL;
 	int i = 0;
 
 	if (str == NULL || !strlen(str))
 		goto free;
 
-	tok = strtok(str, "-");
+	tok = strtok_s(str, "-", &sp);
 	while (tok != NULL && src[i] != ROC_TIM_CLK_SRC_INVALID) {
 		dev->ext_clk_freq[src[i]] = strtoull(tok, NULL, 10);
-		tok = strtok(NULL, "-");
+		tok = strtok_s(NULL, "-", &sp);
 		i++;
 	}
 
-- 
2.30.0

[PATCH v2 18/22] net/ark: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 9c7188a68d7b ("net/ark: provide API for hardware modules pktchkr and pktgen")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/net/ark/ark_pktchkr.c | 10 +++++-----
 drivers/net/ark/ark_pktgen.c  | 10 +++++-----
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ark/ark_pktchkr.c b/drivers/net/ark/ark_pktchkr.c
index e1f336c73c2a..c1c1033eb75e 100644
--- a/drivers/net/ark/ark_pktchkr.c
+++ b/drivers/net/ark/ark_pktchkr.c
@@ -359,14 +359,14 @@ set_arg(char *arg, char *val)
 void
 ark_pktchkr_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = "=\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_s(args, toks, &sp);
+	v = strtok_s(NULL, toks, &sp);
 	while (argv && v) {
 		set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_s(NULL, toks, &sp);
+		v = strtok_s(NULL, toks, &sp);
 	}
 }
 
diff --git a/drivers/net/ark/ark_pktgen.c b/drivers/net/ark/ark_pktgen.c
index 69ff7072b2ab..4b564099d2d6 100644
--- a/drivers/net/ark/ark_pktgen.c
+++ b/drivers/net/ark/ark_pktgen.c
@@ -340,14 +340,14 @@ pmd_set_arg(char *arg, char *val)
 void
 ark_pktgen_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = " =\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_s(args, toks, &sp);
+	v = strtok_s(NULL, toks, &sp);
 	while (argv && v) {
 		pmd_set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_s(NULL, toks, &sp);
+		v = strtok_s(NULL, toks, &sp);
 	}
 }
 
-- 
2.30.0

[PATCH v2 19/22] raw/cnxk_gpio: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: ecc0dd455e9a ("raw/cnxk_gpio: add option to select subset of GPIOs")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/raw/cnxk_gpio/cnxk_gpio.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/raw/cnxk_gpio/cnxk_gpio.c b/drivers/raw/cnxk_gpio/cnxk_gpio.c
index 29c250672646..0e6ee31d35c9 100644
--- a/drivers/raw/cnxk_gpio/cnxk_gpio.c
+++ b/drivers/raw/cnxk_gpio/cnxk_gpio.c
@@ -190,7 +190,7 @@ static int
 cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 {
 	int i, ret, val, queue = 0;
-	char *token;
+	char *token, *sp = NULL;
 	int *list;
 
 	list = rte_calloc(NULL, gpiochip->num_gpios, sizeof(*list), 0);
@@ -208,7 +208,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 	allowlist[strlen(allowlist) - 1] = ' ';
 
 	/* quiesce -Wcast-qual */
-	token = strtok((char *)(uintptr_t)allowlist, ",");
+	token = strtok_s((char *)(uintptr_t)allowlist, ",", &sp);
 	do {
 		errno = 0;
 		val = strtol(token, NULL, 10);
@@ -234,7 +234,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 		}
 		if (i == queue)
 			list[queue++] = val;
-	} while ((token = strtok(NULL, ",")));
+	} while ((token = strtok_s(NULL, ",", &sp)));
 
 	free(allowlist);
 	gpiochip->allowlist = list;
-- 
2.30.0

[PATCH v2 20/22] examples/l2fwd-crypto: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: ff5d5b01f8f2 ("examples/l2fwd-crypto: support AES-CCM")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 examples/l2fwd-crypto/main.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/examples/l2fwd-crypto/main.c b/examples/l2fwd-crypto/main.c
index efe7eea2a768..7327ff1128df 100644
--- a/examples/l2fwd-crypto/main.c
+++ b/examples/l2fwd-crypto/main.c
@@ -1105,12 +1105,12 @@ static int
 parse_bytes(uint8_t *data, char *input_arg, uint16_t max_size)
 {
 	unsigned byte_count;
-	char *token;
+	char *token, *sp = NULL;
 
 	errno = 0;
-	for (byte_count = 0, token = strtok(input_arg, ":");
+	for (byte_count = 0, token = strtok_s(input_arg, ":", &sp);
 			(byte_count < max_size) && (token != NULL);
-			token = strtok(NULL, ":")) {
+			token = strtok_s(NULL, ":", &sp)) {
 
 		int number = (int)strtol(token, NULL, 16);
 
-- 
2.30.0

[PATCH v2 21/22] examples/vhost: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 3a04ecb21420 ("examples/vhost: add async vhost args parsing")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 examples/vhost/main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index ce5c1efddf5c..3949d83be212 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -246,6 +246,7 @@ open_dma(const char *value)
 	char *ptrs[2];
 	char *start, *end, *substr;
 	int64_t socketid, vring_id;
+	char *sp = NULL;
 
 	struct rte_dma_info info;
 	struct rte_dma_conf dev_config = { .nb_vchans = 1 };
@@ -269,7 +270,7 @@ open_dma(const char *value)
 
 	/* process DMA devices within bracket. */
 	addrs++;
-	substr = strtok(addrs, ";]");
+	substr = strtok_s(addrs, ";]", &sp);
 	if (!substr) {
 		ret = -1;
 		goto out;
-- 
2.30.0

[PATCH v2 22/22] devtools: check for some reentrant function

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

This patch adds check in checkpatches.sh for strtok, which is
non-reentrant.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 devtools/checkpatches.sh | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/devtools/checkpatches.sh b/devtools/checkpatches.sh
index 066449d147e8..0a936219c069 100755
--- a/devtools/checkpatches.sh
+++ b/devtools/checkpatches.sh
@@ -119,6 +119,14 @@ check_forbidden_additions() { # <patch>
 		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
 		"$1" || res=1
 
+	# refrain from using some non-reentrant functions
+	awk -v FOLDERS="lib drivers app examples" \
+		-v EXPRESSIONS="strtok\\\(" \
+		-v RET_ON_FAIL=1 \
+		-v MESSAGE='Using non-reentrant function strtok, prefer strtok_s' \
+		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
+		"$1" || res=1
+
 	# refrain from using some pthread functions
 	awk -v FOLDERS="lib drivers app examples" \
 		-v EXPRESSIONS="pthread_(create|join|detach|set(_?name_np|affinity_np)|attr_set(inheritsched|schedpolicy))\\\(" \
-- 
2.30.0

[PATCH v3 00/22] replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

This patchset replaces strtok with strtok_s in app, example, lib
and drivers. And adds check for use of strtok in checkpatches.sh.

--
v3:
1. fix compile error.
2. use strtok_r instead.
v2:
1. fix commit log.
2. add check in checkpatches.sh.
3. replace strtok_r with strtok_s.
4. add Acked-by.
--

Jie Hai (22):
  app/graph: replace strtok with reentrant version
  app/bbdev: replace strtok with reentrant version
  app/compress-perf: replace strtok with reentrant version
  app/crypto-perf: replace strtok with reentrant version
  app/dma-perf: replace strtok with reentrant version
  app/test-fib: replace strtok with reentrant version
  app/flow-perf: replace strtok with reentrant version
  app/test-mldev: replace strtok with reentrant version
  dmadev: replace strtok with reentrant version
  eal: replace strtok with reentrant version
  ethdev: replace strtok with reentrant version
  eventdev: replace strtok with reentrant version
  security: replace strtok with reentrant version
  telemetry: replace strtok with reentrant version
  bus/fslmc: replace strtok with reentrant version
  common/cnxk: replace strtok with reentrant version
  event/cnxk: replace strtok with reentrant version
  net/ark: replace strtok with reentrant version
  raw/cnxk_gpio: replace strtok with reentrant version
  examples/l2fwd-crypto: replace strtok with reentrant version
  examples/vhost: replace strtok with reentrant version
  devtools: check for some reentrant function

 app/graph/graph.c                             |  5 ++-
 app/graph/utils.c                             | 15 +++++---
 app/test-bbdev/test_bbdev_vector.c            | 25 +++++++-----
 .../comp_perf_options_parse.c                 | 16 ++++----
 app/test-crypto-perf/cperf_options_parsing.c  | 16 ++++----
 .../cperf_test_vector_parsing.c               | 10 +++--
 app/test-dma-perf/main.c                      | 13 ++++---
 app/test-fib/main.c                           | 10 ++---
 app/test-flow-perf/main.c                     | 22 ++++++-----
 app/test-mldev/ml_options.c                   | 18 ++++-----
 devtools/checkpatches.sh                      |  8 ++++
 drivers/bus/fslmc/fslmc_bus.c                 |  5 ++-
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  4 +-
 drivers/common/cnxk/cnxk_telemetry_nix.c      | 12 +++---
 drivers/event/cnxk/cnxk_eventdev.c            | 10 +++--
 drivers/event/cnxk/cnxk_tim_evdev.c           | 11 +++---
 drivers/net/ark/ark_pktchkr.c                 | 10 ++---
 drivers/net/ark/ark_pktgen.c                  | 10 ++---
 drivers/raw/cnxk_gpio/cnxk_gpio.c             |  6 +--
 examples/l2fwd-crypto/main.c                  |  6 +--
 examples/vhost/main.c                         |  3 +-
 lib/dmadev/rte_dmadev.c                       |  4 +-
 lib/eal/common/eal_common_memory.c            |  8 ++--
 lib/ethdev/rte_ethdev_telemetry.c             |  6 ++-
 lib/eventdev/rte_event_eth_rx_adapter.c       | 38 +++++++++----------
 lib/eventdev/rte_eventdev.c                   | 18 ++++-----
 lib/security/rte_security.c                   |  3 +-
 lib/telemetry/telemetry.c                     |  5 ++-
 28 files changed, 177 insertions(+), 140 deletions(-)

-- 
2.30.0

[PATCH v3 01/22] app/graph: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 5c59002a34f3 ("app/graph: add graph commands")
Fixes: 984a315a5804 ("app/graph: add parser utility")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/graph/graph.c |  5 +++--
 app/graph/utils.c | 15 +++++++++------
 2 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/app/graph/graph.c b/app/graph/graph.c
index a65723a196da..316950948160 100644
--- a/app/graph/graph.c
+++ b/app/graph/graph.c
@@ -103,9 +103,10 @@ parser_usecases_read(char *usecases)
 {
 	bool valid = false;
 	uint32_t i, j = 0;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(usecases, ",");
+	token = strtok_r(usecases, ",", &sp);
 	while (token != NULL) {
 		for (i = 0; i < RTE_DIM(supported_usecases); i++) {
 			if (strcmp(supported_usecases[i], token) == 0) {
@@ -116,7 +117,7 @@ parser_usecases_read(char *usecases)
 				break;
 			}
 		}
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	return valid;
diff --git a/app/graph/utils.c b/app/graph/utils.c
index c7b6ae83cf1f..ae2ff225c394 100644
--- a/app/graph/utils.c
+++ b/app/graph/utils.c
@@ -101,13 +101,14 @@ int
 parser_ip4_read(uint32_t *value, char *p)
 {
 	uint8_t shift = 24;
+	char *sp = NULL;
 	uint32_t ip = 0;
 	char *token;
 
-	token = strtok(p, ".");
+	token = strtok_r(p, ".", &sp);
 	while (token != NULL) {
 		ip |= (((uint32_t)strtoul(token, NULL, 10)) << shift);
-		token = strtok(NULL, ".");
+		token = strtok_r(NULL, ".", &sp);
 		shift -= 8;
 	}
 
@@ -120,13 +121,14 @@ int
 parser_ip6_read(uint8_t *value, char *p)
 {
 	uint64_t val = 0;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(p, ":");
+	token = strtok_r(p, ":", &sp);
 	while (token != NULL) {
 		hex_string_to_uint64(&val, token);
 		*value = val;
-		token = strtok(NULL, ":");
+		token = strtok_r(NULL, ":", &sp);
 		value++;
 		val = 0;
 	}
@@ -139,13 +141,14 @@ parser_mac_read(uint64_t *value, char *p)
 {
 	uint64_t mac = 0, val = 0;
 	uint8_t shift = 40;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(p, ":");
+	token = strtok_r(p, ":", &sp);
 	while (token != NULL) {
 		hex_string_to_uint64(&val, token);
 		mac |= val << shift;
-		token = strtok(NULL, ":");
+		token = strtok_r(NULL, ":", &sp);
 		shift -= 8;
 		val = 0;
 	}
-- 
2.30.0

Re: [PATCH v3 01/22] app/graph: replace strtok with reentrant version

[Stephen Hemminger]

On Tue, 14 Nov 2023 18:59:45 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: 5c59002a34f3 ("app/graph: add graph commands")
> Fixes: 984a315a5804 ("app/graph: add parser utility")

NAK
These are only called from graph CLI which must be single threaded

[PATCH v3 02/22] app/bbdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 0acdb9866756 ("test/bbdev: add FFT operations cases")
Fixes: f714a18885a6 ("app/testbbdev: add test application for bbdev")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-bbdev/test_bbdev_vector.c | 25 +++++++++++++++----------
 1 file changed, 15 insertions(+), 10 deletions(-)

diff --git a/app/test-bbdev/test_bbdev_vector.c b/app/test-bbdev/test_bbdev_vector.c
index c26727cd35c4..5c966e79645e 100644
--- a/app/test-bbdev/test_bbdev_vector.c
+++ b/app/test-bbdev/test_bbdev_vector.c
@@ -63,8 +63,9 @@ parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
 
 	uint32_t *values, *values_resized;
 	char *tok, *error = NULL;
+	char *sp = NULL;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -98,7 +99,7 @@ parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
 
 		*data_length = *data_length + (strlen(tok) - strlen("0x"))/2;
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 
@@ -299,8 +300,9 @@ parse_turbo_flags(char *tokens, uint32_t *op_flags,
 {
 	char *tok = NULL;
 	uint32_t op_flag_value = 0;
+	char *sp = NULL;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -330,7 +332,7 @@ parse_turbo_flags(char *tokens, uint32_t *op_flags,
 
 		*op_flags = *op_flags | op_flag_value;
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 	}
@@ -368,9 +370,10 @@ static int
 parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
 {
 	char *tok = NULL;
+	char *sp = NULL;
 	bool status_ok = false;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -401,7 +404,7 @@ parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
 			return -1;
 		}
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 	}
@@ -894,6 +897,7 @@ parse_fft_params(const char *key_token, char *token,
 	int ret = 0, status = 0, i, shift;
 	uint32_t op_flags = 0;
 	char *tok, *err = NULL;
+	char *sp = NULL;
 
 	struct rte_bbdev_op_fft *fft = &vector->fft;
 
@@ -922,7 +926,7 @@ parse_fft_params(const char *key_token, char *token,
 		fft->output_leading_depadding = (uint32_t) strtoul(token, &err, 0);
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "window_index")) {
-		tok = strtok(token, VALUE_DELIMITER);
+		tok = strtok_r(token, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			return -1;
 		for (i = 0; i < FFT_WIN_SIZE; i++) {
@@ -930,7 +934,7 @@ parse_fft_params(const char *key_token, char *token,
 			fft->window_index[i / 2] |= (uint32_t) strtoul(tok, &err, 0)
 					<< shift;
 			if (i < (FFT_WIN_SIZE - 1)) {
-				tok = strtok(NULL, VALUE_DELIMITER);
+				tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 				if (tok == NULL)
 					return -1;
 			}
@@ -995,6 +999,7 @@ static int
 parse_entry(char *entry, struct test_bbdev_vector *vector)
 {
 	int ret = 0;
+	char *sp = NULL;
 	char *token, *key_token;
 	enum rte_bbdev_op_type op_type = RTE_BBDEV_OP_NONE;
 
@@ -1004,10 +1009,10 @@ parse_entry(char *entry, struct test_bbdev_vector *vector)
 	}
 
 	/* get key */
-	token = strtok(entry, ENTRY_DELIMITER);
+	token = strtok_r(entry, ENTRY_DELIMITER, &sp);
 	key_token = token;
 	/* get values for key */
-	token = strtok(NULL, ENTRY_DELIMITER);
+	token = strtok_r(NULL, ENTRY_DELIMITER, &sp);
 
 	if (key_token == NULL || token == NULL) {
 		printf("Expected 'key = values' but was '%.40s'..\n", entry);
-- 
2.30.0

Re: [PATCH v3 02/22] app/bbdev: replace strtok with reentrant version

[Stephen Hemminger]

On Tue, 14 Nov 2023 18:59:46 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: 0acdb9866756 ("test/bbdev: add FFT operations cases")
> Fixes: f714a18885a6 ("app/testbbdev: add test application for bbdev")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>

NAK
Please don't make wide assertions about multi-threading.
All these are only called from test_bbdev_vector_read which is single
threaded.

Please limit patches to places where you are sure the use is multi-threaded.

[PATCH v3 03/22] app/compress-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: e0b6287c035d ("app/compress-perf: add parser")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-compress-perf/comp_perf_options_parse.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/app/test-compress-perf/comp_perf_options_parse.c b/app/test-compress-perf/comp_perf_options_parse.c
index 6d8c370fc2ea..d997fa667a94 100644
--- a/app/test-compress-perf/comp_perf_options_parse.c
+++ b/app/test-compress-perf/comp_perf_options_parse.c
@@ -177,6 +177,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 {
 	char *token;
 	uint8_t number;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -184,7 +185,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ":");
+	token = strtok_r(copy_arg, ":", &sp);
 
 	/* Parse minimum value */
 	if (token != NULL) {
@@ -197,7 +198,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse increment value */
 	if (token != NULL) {
@@ -211,7 +212,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse maximum value */
 	if (token != NULL) {
@@ -225,7 +226,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	if (strtok(NULL, ":") != NULL)
+	if (strtok_r(NULL, ":", &sp) != NULL)
 		goto err_range;
 
 	free(copy_arg);
@@ -244,6 +245,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 	uint8_t count = 0;
 	uint32_t temp_min;
 	uint32_t temp_max;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -251,7 +253,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ",");
+	token = strtok_r(copy_arg, ",", &sp);
 
 	/* Parse first value */
 	if (token != NULL) {
@@ -266,7 +268,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 	} else
 		goto err_list;
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 
 	while (token != NULL) {
 		if (count == MAX_LIST) {
@@ -288,7 +290,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 		if (number > temp_max)
 			temp_max = number;
 
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	if (min)
-- 
2.30.0

[PATCH v3 04/22] app/crypto-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: f6cefe253cc8 ("app/crypto-perf: add range/list of sizes")
Fixes: f8be1786b1b8 ("app/crypto-perf: introduce performance test application")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-crypto-perf/cperf_options_parsing.c     | 16 +++++++++-------
 app/test-crypto-perf/cperf_test_vector_parsing.c | 10 ++++++----
 2 files changed, 15 insertions(+), 11 deletions(-)

diff --git a/app/test-crypto-perf/cperf_options_parsing.c b/app/test-crypto-perf/cperf_options_parsing.c
index 75afedc7fd6e..77cda0eb8e6f 100644
--- a/app/test-crypto-perf/cperf_options_parsing.c
+++ b/app/test-crypto-perf/cperf_options_parsing.c
@@ -161,6 +161,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 {
 	char *token;
 	uint32_t number;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -168,7 +169,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ":");
+	token = strtok_r(copy_arg, ":", &sp);
 
 	/* Parse minimum value */
 	if (token != NULL) {
@@ -182,7 +183,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse increment value */
 	if (token != NULL) {
@@ -196,7 +197,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse maximum value */
 	if (token != NULL) {
@@ -211,7 +212,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	if (strtok(NULL, ":") != NULL)
+	if (strtok_r(NULL, ":", &sp) != NULL)
 		goto err_range;
 
 	free(copy_arg);
@@ -230,6 +231,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 	uint8_t count = 0;
 	uint32_t temp_min;
 	uint32_t temp_max;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -237,7 +239,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ",");
+	token = strtok_r(copy_arg, ",", &sp);
 
 	/* Parse first value */
 	if (token != NULL) {
@@ -253,7 +255,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 	} else
 		goto err_list;
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 
 	while (token != NULL) {
 		if (count == MAX_LIST) {
@@ -275,7 +277,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 		if (number > temp_max)
 			temp_max = number;
 
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	if (min)
diff --git a/app/test-crypto-perf/cperf_test_vector_parsing.c b/app/test-crypto-perf/cperf_test_vector_parsing.c
index 737d61d4af6b..4d2e6053e745 100644
--- a/app/test-crypto-perf/cperf_test_vector_parsing.c
+++ b/app/test-crypto-perf/cperf_test_vector_parsing.c
@@ -220,8 +220,9 @@ parse_values(char *tokens, uint8_t **data, uint32_t *data_length)
 
 	uint8_t *values, *values_resized;
 	char *tok, *error = NULL;
+	char *sp = NULL;
 
-	tok = strtok(tokens, CPERF_VALUE_DELIMITER);
+	tok = strtok_r(tokens, CPERF_VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -252,7 +253,7 @@ parse_values(char *tokens, uint8_t **data, uint32_t *data_length)
 			return -1;
 		}
 
-		tok = strtok(NULL, CPERF_VALUE_DELIMITER);
+		tok = strtok_r(NULL, CPERF_VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 
@@ -283,6 +284,7 @@ parse_entry(char *entry, struct cperf_test_vector *vector,
 
 	uint8_t *data = NULL;
 	char *token, *key_token;
+	char *sp = NULL;
 
 	if (entry == NULL) {
 		printf("Expected entry value\n");
@@ -290,10 +292,10 @@ parse_entry(char *entry, struct cperf_test_vector *vector,
 	}
 
 	/* get key */
-	token = strtok(entry, CPERF_ENTRY_DELIMITER);
+	token = strtok_r(entry, CPERF_ENTRY_DELIMITER, &sp);
 	key_token = token;
 	/* get values for key */
-	token = strtok(NULL, CPERF_ENTRY_DELIMITER);
+	token = strtok_r(NULL, CPERF_ENTRY_DELIMITER, &sp);
 
 	if (key_token == NULL || token == NULL) {
 		printf("Expected 'key = values' but was '%.40s'..\n", entry);
-- 
2.30.0

[PATCH v3 05/22] app/dma-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 623dc9364dc6 ("app/dma-perf: introduce DMA performance test")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-dma-perf/main.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/app/test-dma-perf/main.c b/app/test-dma-perf/main.c
index e5bccc27da5e..38780939578e 100644
--- a/app/test-dma-perf/main.c
+++ b/app/test-dma-perf/main.c
@@ -164,6 +164,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 	uint16_t len;
 	char *input;
 	struct lcore_dma_map_t *lcore_dma_map;
+	char *sp = NULL;
 
 	if (test_case == NULL || value == NULL)
 		return -1;
@@ -175,7 +176,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 
 	memset(lcore_dma_map, 0, sizeof(struct lcore_dma_map_t));
 
-	char *token = strtok(input, ", ");
+	char *token = strtok_r(input, ", ", &sp);
 	while (token != NULL) {
 		if (lcore_dma_map->cnt >= MAX_LCORE_NB) {
 			free(input);
@@ -185,7 +186,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 		uint16_t lcore_id = atoi(token);
 		lcore_dma_map->lcores[lcore_dma_map->cnt++] = lcore_id;
 
-		token = strtok(NULL, ", ");
+		token = strtok_r(NULL, ", ", &sp);
 	}
 
 	free(input);
@@ -201,6 +202,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 	char *start, *end, *substr;
 	uint16_t lcore_id;
 	int ret = 0;
+	char *sp = NULL;
 
 	if (test_case == NULL || value == NULL)
 		return -1;
@@ -216,7 +218,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 		goto out;
 	}
 
-	substr = strtok(addrs, ",");
+	substr = strtok_r(addrs, ",", &sp);
 	if (substr == NULL) {
 		fprintf(stderr, "No input DMA address\n");
 		ret = -1;
@@ -258,7 +260,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 		strlcpy(lcore_dma_map->dma_names[lcore_dma_map->cnt], ptrs[1],
 				RTE_DEV_NAME_MAX_LEN);
 		lcore_dma_map->cnt++;
-		substr = strtok(NULL, ",");
+		substr = strtok_r(NULL, ",", &sp);
 	} while (substr != NULL);
 
 out:
@@ -486,6 +488,7 @@ main(int argc, char *argv[])
 	char *rst_path_ptr = NULL;
 	char rst_path[PATH_MAX];
 	int new_argc;
+	char *sp = NULL;
 
 	memset(args, 0, sizeof(args));
 
@@ -504,7 +507,7 @@ main(int argc, char *argv[])
 	}
 	if (rst_path_ptr == NULL) {
 		strlcpy(rst_path, cfg_path_ptr, PATH_MAX);
-		char *token = strtok(basename(rst_path), ".");
+		char *token = strtok_r(basename(rst_path), ".", &sp);
 		if (token == NULL) {
 			printf("Config file error.\n");
 			return -1;
-- 
2.30.0

[PATCH v3 06/22] app/test-fib: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 103809d032cd ("app/test-fib: add test application for FIB")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-fib/main.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/app/test-fib/main.c b/app/test-fib/main.c
index 75a56135f212..8e5d17b13028 100644
--- a/app/test-fib/main.c
+++ b/app/test-fib/main.c
@@ -223,9 +223,9 @@ parse_distrib(uint8_t depth_lim, const uint32_t n)
 	uint32_t nrpd[128 + 1] = {0}; /* number of routes per depth */
 	uint32_t n_routes;
 	uint8_t depth, ratio, ratio_acc = 0;
-	char *in;
+	char *in, *sp = NULL;
 
-	in = strtok(distrib_string, ",");
+	in = strtok_r(distrib_string, ",", &sp);
 
 	/*parse configures routes percentage ratios*/
 	while (in != NULL) {
@@ -265,7 +265,7 @@ parse_distrib(uint8_t depth_lim, const uint32_t n)
 		}
 
 		/*number of configured depths in*/
-		in = strtok(NULL, ",");
+		in = strtok_r(NULL, ",", &sp);
 	}
 
 	if (ratio_acc > 100) {
@@ -542,10 +542,10 @@ parse_lookup(FILE *f, int af)
 	int ret, i = 0;
 	uint8_t *tbl = (uint8_t *)config.lookup_tbl;
 	int step = (af == AF_INET) ? 4 : 16;
-	char *s;
+	char *s, *sp = NULL;
 
 	while (fgets(line, sizeof(line), f) != NULL) {
-		s = strtok(line, " \t\n");
+		s = strtok_r(line, " \t\n", &sp);
 		if (s == NULL)
 			return -EINVAL;
 		ret = inet_pton(af, s, &tbl[i]);
-- 
2.30.0

[PATCH v3 07/22] app/flow-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 0c8f1f4ab90e ("app/flow-perf: support raw encap/decap actions")
Fixes: 7f37f0936a19 ("app/flow-perf: support meter policy API")
Fixes: 80a323319745 ("app/flow-perf: add destination ports parameter")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-flow-perf/main.c | 22 ++++++++++++----------
 1 file changed, 12 insertions(+), 10 deletions(-)

diff --git a/app/test-flow-perf/main.c b/app/test-flow-perf/main.c
index e224ef67983d..36e7c1d72019 100644
--- a/app/test-flow-perf/main.c
+++ b/app/test-flow-perf/main.c
@@ -602,6 +602,7 @@ read_meter_policy(char *prog, char *arg)
 {
 	char *token;
 	size_t i, j, k;
+	char *sp = NULL;
 
 	j = 0;
 	k = 0;
@@ -612,9 +613,9 @@ read_meter_policy(char *prog, char *arg)
 		token = strsep(&arg, ":\0");
 	}
 	j = 0;
-	token = strtok(actions_str[0], ",\0");
+	token = strtok_r(actions_str[0], ",\0", &sp);
 	while (token == NULL && j < RTE_COLORS - 1)
-		token = strtok(actions_str[++j], ",\0");
+		token = strtok_r(actions_str[++j], ",\0", &sp);
 	while (j < RTE_COLORS && token != NULL) {
 		for (i = 0; i < RTE_DIM(flow_options); i++) {
 			if (!strcmp(token, flow_options[i].str)) {
@@ -628,9 +629,9 @@ read_meter_policy(char *prog, char *arg)
 			usage(prog);
 			rte_exit(EXIT_SUCCESS, "Invalid colored actions\n");
 		}
-		token = strtok(NULL, ",\0");
+		token = strtok_r(NULL, ",\0", &sp);
 		while (!token && j < RTE_COLORS - 1) {
-			token = strtok(actions_str[++j], ",\0");
+			token = strtok_r(actions_str[++j], ",\0", &sp);
 			k = 0;
 		}
 	}
@@ -641,6 +642,7 @@ args_parse(int argc, char **argv)
 {
 	uint64_t pm, seed;
 	uint64_t hp_conf;
+	char *sp = NULL;
 	char **argvopt;
 	uint32_t prio;
 	char *token;
@@ -804,7 +806,7 @@ args_parse(int argc, char **argv)
 						RTE_FLOW_ACTION_TYPE_RAW_ENCAP
 					);
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					for (i = 0; i < RTE_DIM(flow_options); i++) {
 						if (strcmp(flow_options[i].str, token) == 0) {
@@ -817,7 +819,7 @@ args_parse(int argc, char **argv)
 							rte_exit(EXIT_FAILURE,
 								"Invalid encap item: %s\n", token);
 					}
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 				printf(" / ");
 			}
@@ -828,7 +830,7 @@ args_parse(int argc, char **argv)
 						RTE_FLOW_ACTION_TYPE_RAW_DECAP
 					);
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					for (i = 0; i < RTE_DIM(flow_options); i++) {
 						if (strcmp(flow_options[i].str, token) == 0) {
@@ -841,7 +843,7 @@ args_parse(int argc, char **argv)
 							rte_exit(EXIT_FAILURE,
 								"Invalid decap item %s\n", token);
 					}
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 				printf(" / ");
 			}
@@ -910,10 +912,10 @@ args_parse(int argc, char **argv)
 				uint16_t port_idx = 0;
 				char *token;
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					dst_ports[port_idx++] = atoi(token);
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 			}
 			if (strcmp(lgopts[opt_idx].name, "rxq") == 0) {
-- 
2.30.0

[PATCH v3 08/22] app/test-mldev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-mldev/ml_options.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/app/test-mldev/ml_options.c b/app/test-mldev/ml_options.c
index 72357c1c393d..3f3dc8f69890 100644
--- a/app/test-mldev/ml_options.c
+++ b/app/test-mldev/ml_options.c
@@ -75,12 +75,12 @@ ml_parse_models(struct ml_options *opt, const char *arg)
 {
 	const char *delim = ",";
 	char models[PATH_MAX];
-	char *token;
+	char *token, *sp = NULL;
 	int ret = 0;
 
 	strlcpy(models, arg, PATH_MAX);
 
-	token = strtok(models, delim);
+	token = strtok_r(models, delim, &sp);
 	while (token != NULL) {
 		strlcpy(opt->filelist[opt->nb_filelist].model, token, PATH_MAX);
 		opt->nb_filelist++;
@@ -90,7 +90,7 @@ ml_parse_models(struct ml_options *opt, const char *arg)
 			ret = -EINVAL;
 			break;
 		}
-		token = strtok(NULL, delim);
+		token = strtok_r(NULL, delim, &sp);
 	}
 
 	if (opt->nb_filelist == 0) {
@@ -106,7 +106,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 {
 	const char *delim = ",";
 	char filelist[PATH_MAX];
-	char *token;
+	char *token, *sp = NULL;
 
 	if (opt->nb_filelist >= ML_TEST_MAX_MODELS) {
 		ml_err("Exceeded filelist count, max = %d\n", ML_TEST_MAX_MODELS);
@@ -116,7 +116,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(filelist, arg, PATH_MAX);
 
 	/* model */
-	token = strtok(filelist, delim);
+	token = strtok_r(filelist, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, model not specified = %s\n", arg);
 		return -EINVAL;
@@ -124,7 +124,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].model, token, PATH_MAX);
 
 	/* input */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, input not specified = %s\n", arg);
 		return -EINVAL;
@@ -132,7 +132,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].input, token, PATH_MAX);
 
 	/* output */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, output not specified = %s\n", arg);
 		return -EINVAL;
@@ -140,14 +140,14 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].output, token, PATH_MAX);
 
 	/* reference - optional */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token != NULL)
 		strlcpy(opt->filelist[opt->nb_filelist].reference, token, PATH_MAX);
 	else
 		memset(opt->filelist[opt->nb_filelist].reference, 0, PATH_MAX);
 
 	/* check for extra tokens */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token != NULL) {
 		ml_err("Invalid filelist. Entries > 4\n.");
 		return -EINVAL;
-- 
2.30.0

[PATCH v3 09/22] dmadev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 39b5ab60df30 ("dmadev: add telemetry")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/dmadev/rte_dmadev.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/dmadev/rte_dmadev.c b/lib/dmadev/rte_dmadev.c
index 4e5e420c82a5..f856cc7d0905 100644
--- a/lib/dmadev/rte_dmadev.c
+++ b/lib/dmadev/rte_dmadev.c
@@ -971,7 +971,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	struct rte_dma_info dma_info;
 	struct rte_dma_stats dma_stats;
 	int dev_id, ret, vchan_id;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *vchan_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -990,7 +990,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	if (dma_info.nb_vchans == 1 && *end_param == '\0')
 		vchan_id = 0;
 	else {
-		vchan_param = strtok(end_param, ",");
+		vchan_param = strtok_r(end_param, ",", &sp);
 		if (!vchan_param || strlen(vchan_param) == 0 || !isdigit(*vchan_param))
 			return -EINVAL;
 
-- 
2.30.0

[PATCH v3 10/22] eal: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/eal/common/eal_common_memory.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/eal/common/eal_common_memory.c b/lib/eal/common/eal_common_memory.c
index d9433db62345..a05eb0442b0b 100644
--- a/lib/eal/common/eal_common_memory.c
+++ b/lib/eal/common/eal_common_memory.c
@@ -1273,22 +1273,22 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	char dlim[2] = ",";
 	char *params_args;
 	size_t count = 0;
-	char *token;
+	char *token, *sp = NULL;
 
 	if (vals == NULL || params == NULL || strlen(params) == 0)
 		return -1;
 
-	/* strtok expects char * and param is const char *. Hence on using
+	/* strtok_r expects char * and param is const char *. Hence on using
 	 * params as "const char *" compiler throws warning.
 	 */
 	params_args = strdup(params);
 	if (params_args == NULL)
 		return -1;
 
-	token = strtok(params_args, dlim);
+	token = strtok_r(params_args, dlim, &sp);
 	while (token && isdigit(*token) && count < n_vals) {
 		vals[count++] = strtoul(token, NULL, 10);
-		token = strtok(NULL, dlim);
+		token = strtok_r(NULL, dlim, &sp);
 	}
 
 	free(params_args);
-- 
2.30.0

RE: [EXT] [PATCH v3 10/22] eal: replace strtok with reentrant version

[Amit Prakash Shukla]

> -----Original Message-----
> From: Jie Hai <haijie1@huawei.com>
> Sent: Tuesday, November 14, 2023 4:30 PM
> To: dev@dpdk.org; Anatoly Burakov <anatoly.burakov@intel.com>; Amit
> Prakash Shukla <amitprakashs@marvell.com>
> Cc: haijie1@huawei.com; lihuisong@huawei.com;
> fengchengwen@huawei.com
> Subject: [EXT] [PATCH v3 10/22] eal: replace strtok with reentrant version
> 
> External Email
> 
> ----------------------------------------------------------------------
> Multiple threads calling the same function may cause condition race issues,
> which often leads to abnormal behavior and can cause more serious
> vulnerabilities such as abnormal termination, denial of service, and
> compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a reentrant
> version.
> 
> Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> ---
>  lib/eal/common/eal_common_memory.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/lib/eal/common/eal_common_memory.c
> b/lib/eal/common/eal_common_memory.c
> index d9433db62345..a05eb0442b0b 100644
> --- a/lib/eal/common/eal_common_memory.c
> +++ b/lib/eal/common/eal_common_memory.c
> @@ -1273,22 +1273,22 @@ parse_params(const char *params, uint32_t
> *vals, size_t n_vals)
>  	char dlim[2] = ",";
>  	char *params_args;
>  	size_t count = 0;
> -	char *token;
> +	char *token, *sp = NULL;
> 
>  	if (vals == NULL || params == NULL || strlen(params) == 0)
>  		return -1;
> 
> -	/* strtok expects char * and param is const char *. Hence on using
> +	/* strtok_r expects char * and param is const char *. Hence on using
>  	 * params as "const char *" compiler throws warning.
>  	 */
>  	params_args = strdup(params);
>  	if (params_args == NULL)
>  		return -1;
> 
> -	token = strtok(params_args, dlim);
> +	token = strtok_r(params_args, dlim, &sp);
>  	while (token && isdigit(*token) && count < n_vals) {
>  		vals[count++] = strtoul(token, NULL, 10);
> -		token = strtok(NULL, dlim);
> +		token = strtok_r(NULL, dlim, &sp);
>  	}
> 
>  	free(params_args);
> --
> 2.30.0

Acked-by: Amit Prakash Shukla <amitprakashs@marvell.com>

[PATCH v3 11/22] ethdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: f38f62650f7b ("ethdev: add Rx queue telemetry query")
Fixes: 9e7533aeb80a ("ethdev: add telemetry command for TM level capabilities")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/ethdev/rte_ethdev_telemetry.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/lib/ethdev/rte_ethdev_telemetry.c b/lib/ethdev/rte_ethdev_telemetry.c
index b01028ce9b60..6810c54109a5 100644
--- a/lib/ethdev/rte_ethdev_telemetry.c
+++ b/lib/ethdev/rte_ethdev_telemetry.c
@@ -477,6 +477,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	const char *qid_param;
 	uint16_t nb_queues;
 	char *end_param;
+	char *sp = NULL;
 	uint64_t qid;
 	int ret;
 
@@ -489,7 +490,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	if (nb_queues == 1 && *end_param == '\0')
 		qid = 0;
 	else {
-		qid_param = strtok(end_param, ",");
+		qid_param = strtok_r(end_param, ",", &sp);
 		if (!qid_param || strlen(qid_param) == 0 || !isdigit(*qid_param))
 			return -EINVAL;
 
@@ -1221,9 +1222,10 @@ static int
 eth_dev_parse_tm_params(char *params, uint32_t *result)
 {
 	const char *splited_param;
+	char *sp = NULL;
 	uint64_t ret;
 
-	splited_param = strtok(params, ",");
+	splited_param = strtok_r(params, ",", &sp);
 	if (!splited_param || strlen(splited_param) == 0 || !isdigit(*splited_param))
 		return -EINVAL;
 
-- 
2.30.0

Re: [PATCH v3 11/22] ethdev: replace strtok with reentrant version

[Andrew Rybchenko]

On 11/14/23 13:59, Jie Hai wrote:
> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: f38f62650f7b ("ethdev: add Rx queue telemetry query")
> Fixes: 9e7533aeb80a ("ethdev: add telemetry command for TM level capabilities")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>

Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>

[PATCH v3 12/22] eventdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/eventdev/rte_event_eth_rx_adapter.c | 38 ++++++++++++-------------
 lib/eventdev/rte_eventdev.c             | 18 ++++++------
 2 files changed, 28 insertions(+), 28 deletions(-)

diff --git a/lib/eventdev/rte_event_eth_rx_adapter.c b/lib/eventdev/rte_event_eth_rx_adapter.c
index 6db03adf0463..4107d600dfb7 100644
--- a/lib/eventdev/rte_event_eth_rx_adapter.c
+++ b/lib/eventdev/rte_event_eth_rx_adapter.c
@@ -3651,7 +3651,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp;
 	struct rte_event_eth_rx_adapter_queue_conf queue_conf;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3661,19 +3661,19 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3684,7 +3684,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3723,7 +3723,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 	struct rte_event_eth_rx_adapter_queue_stats q_stats;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3733,19 +3733,19 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3756,7 +3756,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3794,7 +3794,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3803,19 +3803,19 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3826,7 +3826,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3855,7 +3855,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	uint8_t instance_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3863,14 +3863,14 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_ETH_VALID_PORTID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3881,7 +3881,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
diff --git a/lib/eventdev/rte_eventdev.c b/lib/eventdev/rte_eventdev.c
index 0ca32d672175..fabba0228672 100644
--- a/lib/eventdev/rte_eventdev.c
+++ b/lib/eventdev/rte_eventdev.c
@@ -1784,7 +1784,7 @@ handle_queue_links(const char *cmd __rte_unused,
 		   struct rte_tel_data *d)
 {
 	int i, ret, port_id = 0;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	uint8_t dev_id;
 	uint8_t queues[RTE_EVENT_MAX_QUEUES_PER_DEV];
 	uint8_t priorities[RTE_EVENT_MAX_QUEUES_PER_DEV];
@@ -1797,12 +1797,12 @@ handle_queue_links(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
 		return -1;
 
 	port_id = strtoul(p_param, &end_param, 10);
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1922,7 +1922,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1932,7 +1932,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_PORT;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1940,7 +1940,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1956,7 +1956,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1966,7 +1966,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_QUEUE;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1974,7 +1974,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
-- 
2.30.0

[PATCH v3 13/22] security: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 259ca6d1617f ("security: add telemetry endpoint for capabilities")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/security/rte_security.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/lib/security/rte_security.c b/lib/security/rte_security.c
index b082a290296b..e20d610172ef 100644
--- a/lib/security/rte_security.c
+++ b/lib/security/rte_security.c
@@ -496,13 +496,14 @@ security_handle_cryptodev_crypto_caps(const char *cmd __rte_unused, const char *
 	int dev_id, capa_id;
 	int crypto_caps_n;
 	char *end_param;
+	char *sp = NULL;
 	int rc;
 
 	if (!params || strlen(params) == 0 || !isdigit(*params))
 		return -EINVAL;
 
 	dev_id = strtoul(params, &end_param, 0);
-	capa_param = strtok(end_param, ",");
+	capa_param = strtok_r(end_param, ",", &sp);
 	if (!capa_param || strlen(capa_param) == 0 || !isdigit(*capa_param))
 		return -EINVAL;
 
-- 
2.30.0

[PATCH v3 14/22] telemetry: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 6dd571fd07c3 ("telemetry: introduce new functionality")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 lib/telemetry/telemetry.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/lib/telemetry/telemetry.c b/lib/telemetry/telemetry.c
index 92982842a860..bb028d9381f7 100644
--- a/lib/telemetry/telemetry.c
+++ b/lib/telemetry/telemetry.c
@@ -371,6 +371,7 @@ static void *
 client_handler(void *sock_id)
 {
 	int s = (int)(uintptr_t)sock_id;
+	char *sp = NULL;
 	char buffer[1024];
 	char info_str[1024];
 	snprintf(info_str, sizeof(info_str),
@@ -385,8 +386,8 @@ client_handler(void *sock_id)
 	int bytes = read(s, buffer, sizeof(buffer) - 1);
 	while (bytes > 0) {
 		buffer[bytes] = 0;
-		const char *cmd = strtok(buffer, ",");
-		const char *param = strtok(NULL, "\0");
+		const char *cmd = strtok_r(buffer, ",", &sp);
+		const char *param = strtok_r(NULL, "\0", &sp);
 		telemetry_cb fn = unknown_command;
 		int i;
 
-- 
2.30.0

[PATCH v3 15/22] bus/fslmc: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 9ccb76b24c1d ("bus/fslmc: enable portal interrupt handling")
Fixes: 828d51d8fc3e ("bus/fslmc: refactor scan and probe functions")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/bus/fslmc/fslmc_bus.c            | 5 +++--
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 4 ++--
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index 57bfb5111a97..7960ad3979ef 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -131,6 +131,7 @@ scan_one_fslmc_device(char *dev_name)
 {
 	char *dup_dev_name, *t_ptr;
 	struct rte_dpaa2_device *dev = NULL;
+	char *sp = NULL;
 	int ret = -1;
 
 	if (!dev_name)
@@ -168,7 +169,7 @@ scan_one_fslmc_device(char *dev_name)
 	}
 
 	/* Parse the device name and ID */
-	t_ptr = strtok(dup_dev_name, ".");
+	t_ptr = strtok_r(dup_dev_name, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Invalid device found: (%s)", dup_dev_name);
 		ret = -EINVAL;
@@ -199,7 +200,7 @@ scan_one_fslmc_device(char *dev_name)
 	else
 		dev->dev_type = DPAA2_UNKNOWN;
 
-	t_ptr = strtok(NULL, ".");
+	t_ptr = strtok_r(NULL, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Skipping invalid device (%s)", dup_dev_name);
 		ret = 0;
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index 4aec7b2cd8ba..09a1a2b23787 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -129,7 +129,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 	uint32_t cpu_mask = 1;
 	int ret;
 	size_t len = 0;
-	char *temp = NULL, *token = NULL;
+	char *temp = NULL, *token = NULL, *sp = NULL;
 	char string[STRING_LEN], command[COMMAND_LEN];
 	FILE *file;
 
@@ -141,7 +141,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 	}
 	while (getline(&temp, &len, file) != -1) {
 		if ((strstr(temp, string)) != NULL) {
-			token = strtok(temp, ":");
+			token = strtok_r(temp, ":", &sp);
 			break;
 		}
 	}
-- 
2.30.0

[PATCH v3 16/22] common/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: af75aac78978 ("common/cnxk: support telemetry for NIX")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/common/cnxk/cnxk_telemetry_nix.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/common/cnxk/cnxk_telemetry_nix.c b/drivers/common/cnxk/cnxk_telemetry_nix.c
index ccae5d7853af..8054c3cf910c 100644
--- a/drivers/common/cnxk/cnxk_telemetry_nix.c
+++ b/drivers/common/cnxk/cnxk_telemetry_nix.c
@@ -761,7 +761,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			   struct plt_tel_data *d)
 {
 	struct nix_tel_node *node;
-	char *name, *param;
+	char *name, *param, *sp = NULL;
 	char buf[1024];
 	int rc = -1;
 
@@ -769,11 +769,11 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 		goto exit;
 
 	plt_strlcpy(buf, params, PCI_PRI_STR_SIZE + 1);
-	name = strtok(buf, ",");
+	name = strtok_r(buf, ",", &sp);
 	if (name == NULL)
 		goto exit;
 
-	param = strtok(NULL, "\0");
+	param = strtok_r(NULL, "\0", &sp);
 
 	node = nix_tel_node_get_by_pcidev_name(name);
 	if (!node)
@@ -782,7 +782,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 	plt_tel_data_start_dict(d);
 
 	if (strstr(cmd, "rq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int rq;
 
 		if (!tok)
@@ -798,7 +798,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_rq(node->rqs[rq], d);
 
 	} else if (strstr(cmd, "cq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int cq;
 
 		if (!tok)
@@ -814,7 +814,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_cq(node->cqs[cq], d);
 
 	} else if (strstr(cmd, "sq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int sq;
 
 		if (!tok)
-- 
2.30.0

[PATCH v3 17/22] event/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/event/cnxk/cnxk_eventdev.c  | 10 ++++++----
 drivers/event/cnxk/cnxk_tim_evdev.c | 11 ++++++-----
 2 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/drivers/event/cnxk/cnxk_eventdev.c b/drivers/event/cnxk/cnxk_eventdev.c
index 0c61f4c20eec..fe7a86797c25 100644
--- a/drivers/event/cnxk/cnxk_eventdev.c
+++ b/drivers/event/cnxk/cnxk_eventdev.c
@@ -478,7 +478,8 @@ parse_queue_param(char *value, void *opaque)
 	struct cnxk_sso_qos queue_qos = {0};
 	uint16_t *val = (uint16_t *)&queue_qos;
 	struct cnxk_sso_evdev *dev = opaque;
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "-", &sp);
 	struct cnxk_sso_qos *old_ptr;
 
 	if (!strlen(value))
@@ -486,7 +487,7 @@ parse_queue_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		val++;
 	}
 
@@ -514,7 +515,8 @@ parse_stash_param(char *value, void *opaque)
 	struct cnxk_sso_stash queue_stash = {0};
 	struct cnxk_sso_evdev *dev = opaque;
 	struct cnxk_sso_stash *old_ptr;
-	char *tok = strtok(value, "|");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "|", &sp);
 	uint16_t *val;
 
 	if (!strlen(value))
@@ -523,7 +525,7 @@ parse_stash_param(char *value, void *opaque)
 	val = (uint16_t *)&queue_stash;
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "|");
+		tok = strtok_r(NULL, "|", &sp);
 		val++;
 	}
 
diff --git a/drivers/event/cnxk/cnxk_tim_evdev.c b/drivers/event/cnxk/cnxk_tim_evdev.c
index 6d59fdf90983..86ef7dc3d578 100644
--- a/drivers/event/cnxk/cnxk_tim_evdev.c
+++ b/drivers/event/cnxk/cnxk_tim_evdev.c
@@ -420,7 +420,8 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 {
 	struct cnxk_tim_evdev *dev = opaque;
 	struct cnxk_tim_ctl ring_ctl = {0};
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "-", &sp);
 	struct cnxk_tim_ctl *old_ptr;
 	uint16_t *val;
 
@@ -431,7 +432,7 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		val++;
 	}
 
@@ -507,16 +508,16 @@ cnxk_tim_parse_clk_list(const char *value, void *opaque)
 				      ROC_TIM_CLK_SRC_INVALID};
 	struct cnxk_tim_evdev *dev = opaque;
 	char *str = strdup(value);
-	char *tok;
+	char *tok, *sp = NULL;
 	int i = 0;
 
 	if (str == NULL || !strlen(str))
 		goto free;
 
-	tok = strtok(str, "-");
+	tok = strtok_r(str, "-", &sp);
 	while (tok != NULL && src[i] != ROC_TIM_CLK_SRC_INVALID) {
 		dev->ext_clk_freq[src[i]] = strtoull(tok, NULL, 10);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		i++;
 	}
 
-- 
2.30.0

[PATCH v3 18/22] net/ark: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 9c7188a68d7b ("net/ark: provide API for hardware modules pktchkr and pktgen")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/net/ark/ark_pktchkr.c | 10 +++++-----
 drivers/net/ark/ark_pktgen.c  | 10 +++++-----
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ark/ark_pktchkr.c b/drivers/net/ark/ark_pktchkr.c
index e1f336c73c2a..6e8e5339607b 100644
--- a/drivers/net/ark/ark_pktchkr.c
+++ b/drivers/net/ark/ark_pktchkr.c
@@ -359,14 +359,14 @@ set_arg(char *arg, char *val)
 void
 ark_pktchkr_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = "=\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_r(args, toks, &sp);
+	v = strtok_r(NULL, toks, &sp);
 	while (argv && v) {
 		set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_r(NULL, toks, &sp);
+		v = strtok_r(NULL, toks, &sp);
 	}
 }
 
diff --git a/drivers/net/ark/ark_pktgen.c b/drivers/net/ark/ark_pktgen.c
index 69ff7072b2ab..d611406a1b46 100644
--- a/drivers/net/ark/ark_pktgen.c
+++ b/drivers/net/ark/ark_pktgen.c
@@ -340,14 +340,14 @@ pmd_set_arg(char *arg, char *val)
 void
 ark_pktgen_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = " =\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_r(args, toks, &sp);
+	v = strtok_r(NULL, toks, &sp);
 	while (argv && v) {
 		pmd_set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_r(NULL, toks, &sp);
+		v = strtok_r(NULL, toks, &sp);
 	}
 }
 
-- 
2.30.0

[PATCH v3 19/22] raw/cnxk_gpio: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: ecc0dd455e9a ("raw/cnxk_gpio: add option to select subset of GPIOs")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/raw/cnxk_gpio/cnxk_gpio.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/raw/cnxk_gpio/cnxk_gpio.c b/drivers/raw/cnxk_gpio/cnxk_gpio.c
index 29c250672646..a499a258f315 100644
--- a/drivers/raw/cnxk_gpio/cnxk_gpio.c
+++ b/drivers/raw/cnxk_gpio/cnxk_gpio.c
@@ -190,7 +190,7 @@ static int
 cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 {
 	int i, ret, val, queue = 0;
-	char *token;
+	char *token, *sp = NULL;
 	int *list;
 
 	list = rte_calloc(NULL, gpiochip->num_gpios, sizeof(*list), 0);
@@ -208,7 +208,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 	allowlist[strlen(allowlist) - 1] = ' ';
 
 	/* quiesce -Wcast-qual */
-	token = strtok((char *)(uintptr_t)allowlist, ",");
+	token = strtok_r((char *)(uintptr_t)allowlist, ",", &sp);
 	do {
 		errno = 0;
 		val = strtol(token, NULL, 10);
@@ -234,7 +234,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 		}
 		if (i == queue)
 			list[queue++] = val;
-	} while ((token = strtok(NULL, ",")));
+	} while ((token = strtok_r(NULL, ",", &sp)));
 
 	free(allowlist);
 	gpiochip->allowlist = list;
-- 
2.30.0

[PATCH v3 20/22] examples/l2fwd-crypto: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: ff5d5b01f8f2 ("examples/l2fwd-crypto: support AES-CCM")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 examples/l2fwd-crypto/main.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/examples/l2fwd-crypto/main.c b/examples/l2fwd-crypto/main.c
index efe7eea2a768..06e296776b95 100644
--- a/examples/l2fwd-crypto/main.c
+++ b/examples/l2fwd-crypto/main.c
@@ -1105,12 +1105,12 @@ static int
 parse_bytes(uint8_t *data, char *input_arg, uint16_t max_size)
 {
 	unsigned byte_count;
-	char *token;
+	char *token, *sp = NULL;
 
 	errno = 0;
-	for (byte_count = 0, token = strtok(input_arg, ":");
+	for (byte_count = 0, token = strtok_r(input_arg, ":", &sp);
 			(byte_count < max_size) && (token != NULL);
-			token = strtok(NULL, ":")) {
+			token = strtok_r(NULL, ":", &sp)) {
 
 		int number = (int)strtol(token, NULL, 16);
 
-- 
2.30.0

[PATCH v3 21/22] examples/vhost: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 3a04ecb21420 ("examples/vhost: add async vhost args parsing")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 examples/vhost/main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index ce5c1efddf5c..e8b3a97c48a4 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -246,6 +246,7 @@ open_dma(const char *value)
 	char *ptrs[2];
 	char *start, *end, *substr;
 	int64_t socketid, vring_id;
+	char *sp = NULL;
 
 	struct rte_dma_info info;
 	struct rte_dma_conf dev_config = { .nb_vchans = 1 };
@@ -269,7 +270,7 @@ open_dma(const char *value)
 
 	/* process DMA devices within bracket. */
 	addrs++;
-	substr = strtok(addrs, ";]");
+	substr = strtok_r(addrs, ";]", &sp);
 	if (!substr) {
 		ret = -1;
 		goto out;
-- 
2.30.0

[PATCH v3 22/22] devtools: check for some reentrant function

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

This patch adds check in checkpatches.sh for strtok, which is
non-reentrant.

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 devtools/checkpatches.sh | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/devtools/checkpatches.sh b/devtools/checkpatches.sh
index 066449d147e8..e831fc723dfb 100755
--- a/devtools/checkpatches.sh
+++ b/devtools/checkpatches.sh
@@ -119,6 +119,14 @@ check_forbidden_additions() { # <patch>
 		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
 		"$1" || res=1
 
+	# refrain from using some non-reentrant functions
+	awk -v FOLDERS="lib drivers app examples" \
+		-v EXPRESSIONS="strtok\\\(" \
+		-v RET_ON_FAIL=1 \
+		-v MESSAGE='Using non-reentrant function strtok, prefer strtok_r' \
+		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
+		"$1" || res=1
+
 	# refrain from using some pthread functions
 	awk -v FOLDERS="lib drivers app examples" \
 		-v EXPRESSIONS="pthread_(create|join|detach|set(_?name_np|affinity_np)|attr_set(inheritsched|schedpolicy))\\\(" \
-- 
2.30.0

RE: [PATCH v3 00/22] replace strtok with reentrant version

[Morten Brørup]

> From: Jie Hai [mailto:haijie1@huawei.com]
> Sent: Tuesday, 14 November 2023 12.00
> 
> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> This patchset replaces strtok with strtok_s in app, example, lib

Typo: strtok_s -> strtok_r

> and drivers. And adds check for use of strtok in checkpatches.sh.
> 
> --
> v3:
> 1. fix compile error.
> 2. use strtok_r instead.
> v2:
> 1. fix commit log.
> 2. add check in checkpatches.sh.
> 3. replace strtok_r with strtok_s.
> 4. add Acked-by.
> --

+1 to doing this, also in functions that are supposed to be single threaded; because it allows the addition to checkpatches.

Acked-by: Morten Brørup <mb@smartsharesystems.com>

[PATCH v4 00/13] replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

This patchset replaces strtok with strtok_r in app, example, lib
and drivers. And adds check for use of strtok in checkpatches.sh.

--
v4:
1. fix mispellings.
2. add Acked-bys and Reviewd-bys.
3. remove some patch and add new. 
v3:
1. fix compile error.
2. use strtok_r instead.
v2:
1. fix commit log.
2. add check in checkpatches.sh.
3. replace strtok_r with strtok_s.
4. add Acked-by.
--

Jie Hai (13):
  dmadev: replace strtok with reentrant version
  eal: replace strtok with reentrant version
  ethdev: replace strtok with reentrant version
  eventdev: replace strtok with reentrant version
  security: replace strtok with reentrant version
  telemetry: replace strtok with reentrant version
  bus/fslmc: replace strtok with reentrant version
  common/cnxk: replace strtok with reentrant version
  event/cnxk: replace strtok with reentrant version
  net/ark: replace strtok with reentrant version
  raw/cnxk_gpio: replace strtok with reentrant version
  net/cnxk: replace strtok with reentrant version
  devtools: check for some reentrant function

 devtools/checkpatches.sh                     |  8 +++++
 drivers/bus/fslmc/fslmc_bus.c                |  5 +--
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c     |  4 +--
 drivers/common/cnxk/cnxk_telemetry_nix.c     | 12 +++----
 drivers/event/cnxk/cnxk_eventdev.c           | 10 +++---
 drivers/event/cnxk/cnxk_tim_evdev.c          | 11 +++---
 drivers/net/ark/ark_pktchkr.c                | 10 +++---
 drivers/net/ark/ark_pktgen.c                 | 10 +++---
 drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c |  5 +--
 drivers/raw/cnxk_gpio/cnxk_gpio.c            |  6 ++--
 lib/dmadev/rte_dmadev.c                      |  4 +--
 lib/eal/common/eal_common_memory.c           |  8 ++---
 lib/ethdev/rte_ethdev_telemetry.c            |  9 +++--
 lib/eventdev/rte_event_eth_rx_adapter.c      | 38 ++++++++++----------
 lib/eventdev/rte_eventdev.c                  | 18 +++++-----
 lib/security/rte_security.c                  |  3 +-
 lib/telemetry/telemetry.c                    |  5 +--
 17 files changed, 92 insertions(+), 74 deletions(-)

-- 
2.22.0

[PATCH v4 01/13] dmadev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 39b5ab60df30 ("dmadev: add telemetry")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/dmadev/rte_dmadev.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/dmadev/rte_dmadev.c b/lib/dmadev/rte_dmadev.c
index 8bb7824aa129..5c5ac78ae9c4 100644
--- a/lib/dmadev/rte_dmadev.c
+++ b/lib/dmadev/rte_dmadev.c
@@ -1016,7 +1016,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	struct rte_dma_info dma_info;
 	struct rte_dma_stats dma_stats;
 	int dev_id, ret, vchan_id;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *vchan_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1035,7 +1035,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	if (dma_info.nb_vchans == 1 && *end_param == '\0')
 		vchan_id = 0;
 	else {
-		vchan_param = strtok(end_param, ",");
+		vchan_param = strtok_r(end_param, ",", &sp);
 		if (!vchan_param || strlen(vchan_param) == 0 || !isdigit(*vchan_param))
 			return -EINVAL;
 
-- 
2.22.0

Re: [PATCH v4 01/13] dmadev: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:39 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: 39b5ab60df30 ("dmadev: add telemetry")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> Acked-by: Morten Brørup <mb@smartsharesystems.com>

No reason to bother stable with this one. That function is only called
from RTE_INIT() which is a constructor.

[PATCH v4 02/13] eal: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Amit Prakash Shukla <amitprakashs@marvell.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/eal/common/eal_common_memory.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/eal/common/eal_common_memory.c b/lib/eal/common/eal_common_memory.c
index a185e0b580c7..ace63313ea75 100644
--- a/lib/eal/common/eal_common_memory.c
+++ b/lib/eal/common/eal_common_memory.c
@@ -1275,22 +1275,22 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	char dlim[2] = ",";
 	char *params_args;
 	size_t count = 0;
-	char *token;
+	char *token, *sp = NULL;
 
 	if (vals == NULL || params == NULL || strlen(params) == 0)
 		return -1;
 
-	/* strtok expects char * and param is const char *. Hence on using
+	/* strtok_r expects char * and param is const char *. Hence on using
 	 * params as "const char *" compiler throws warning.
 	 */
 	params_args = strdup(params);
 	if (params_args == NULL)
 		return -1;
 
-	token = strtok(params_args, dlim);
+	token = strtok_r(params_args, dlim, &sp);
 	while (token && isdigit(*token) && count < n_vals) {
 		vals[count++] = strtoul(token, NULL, 10);
-		token = strtok(NULL, dlim);
+		token = strtok_r(NULL, dlim, &sp);
 	}
 
 	free(params_args);
-- 
2.22.0

Re: [PATCH v4 02/13] eal: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:40 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> Acked-by: Amit Prakash Shukla <amitprakashs@marvell.com>
> Acked-by: Morten Brørup <mb@smartsharesystems.com>

This doesn't need to go to stable. parse_params is always single threaded.

Re: [PATCH v4 02/13] eal: replace strtok with reentrant version

[fengchengwen]

On 2024/10/27 11:00, Stephen Hemminger wrote:
> On Sat, 26 Oct 2024 18:14:40 +0800
> Jie Hai <haijie1@huawei.com> wrote:
> 
>> Multiple threads calling the same function may cause condition
>> race issues, which often leads to abnormal behavior and can cause
>> more serious vulnerabilities such as abnormal termination, denial
>> of service, and compromised data integrity.
>>
>> The strtok() is non-reentrant, it is better to replace it with a
>> reentrant version.
>>
>> Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")
>> Cc: stable@dpdk.org
>>
>> Signed-off-by: Jie Hai <haijie1@huawei.com>
>> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
>> Acked-by: Amit Prakash Shukla <amitprakashs@marvell.com>
>> Acked-by: Morten Brørup <mb@smartsharesystems.com>
> 
> This doesn't need to go to stable. parse_params is always single threaded.

I recommend replacing all, based on:
1\ almost at no cost.
2\ reduce analysis costs, if don't we have to analyze the callers of strtok when you encounter it.

Re: [PATCH v4 02/13] eal: replace strtok with reentrant version

[Stephen Hemminger]

On Mon, 28 Oct 2024 21:04:08 +0800
fengchengwen <fengchengwen@huawei.com> wrote:

> On 2024/10/27 11:00, Stephen Hemminger wrote:
> > On Sat, 26 Oct 2024 18:14:40 +0800
> > Jie Hai <haijie1@huawei.com> wrote:
> >   
> >> Multiple threads calling the same function may cause condition
> >> race issues, which often leads to abnormal behavior and can cause
> >> more serious vulnerabilities such as abnormal termination, denial
> >> of service, and compromised data integrity.
> >>
> >> The strtok() is non-reentrant, it is better to replace it with a
> >> reentrant version.
> >>
> >> Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")
> >> Cc: stable@dpdk.org
> >>
> >> Signed-off-by: Jie Hai <haijie1@huawei.com>
> >> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> >> Acked-by: Amit Prakash Shukla <amitprakashs@marvell.com>
> >> Acked-by: Morten Brørup <mb@smartsharesystems.com>  
> > 
> > This doesn't need to go to stable. parse_params is always single threaded.  
> 
> I recommend replacing all, based on:
> 1\ almost at no cost.
> 2\ reduce analysis costs, if don't we have to analyze the callers of strtok when you encounter it.
> 

Yes but. The replacement should not go to stable.
One of the rules of stable is that changes should be minimized, and fixes should
not be accepted for things that can not ever happen with current code.

Re: [PATCH v4 02/13] eal: replace strtok with reentrant version

[fengchengwen]

On 2024/10/28 23:31, Stephen Hemminger wrote:
> On Mon, 28 Oct 2024 21:04:08 +0800
> fengchengwen <fengchengwen@huawei.com> wrote:
> 
>> On 2024/10/27 11:00, Stephen Hemminger wrote:
>>> On Sat, 26 Oct 2024 18:14:40 +0800
>>> Jie Hai <haijie1@huawei.com> wrote:
>>>   
>>>> Multiple threads calling the same function may cause condition
>>>> race issues, which often leads to abnormal behavior and can cause
>>>> more serious vulnerabilities such as abnormal termination, denial
>>>> of service, and compromised data integrity.
>>>>
>>>> The strtok() is non-reentrant, it is better to replace it with a
>>>> reentrant version.
>>>>
>>>> Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")
>>>> Cc: stable@dpdk.org
>>>>
>>>> Signed-off-by: Jie Hai <haijie1@huawei.com>
>>>> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
>>>> Acked-by: Amit Prakash Shukla <amitprakashs@marvell.com>
>>>> Acked-by: Morten Brørup <mb@smartsharesystems.com>  
>>>
>>> This doesn't need to go to stable. parse_params is always single threaded.  
>>
>> I recommend replacing all, based on:
>> 1\ almost at no cost.
>> 2\ reduce analysis costs, if don't we have to analyze the callers of strtok when you encounter it.
>>
> 
> Yes but. The replacement should not go to stable.
> One of the rules of stable is that changes should be minimized, and fixes should
> not be accepted for things that can not ever happen with current code.

Hope more opinion from TB members.

Re: [PATCH v4 02/13] eal: replace strtok with reentrant version

[Stephen Hemminger]

On Tue, 29 Oct 2024 08:56:20 +0800
fengchengwen <fengchengwen@huawei.com> wrote:

> >>>
> >>> This doesn't need to go to stable. parse_params is always single threaded.    
> >>
> >> I recommend replacing all, based on:
> >> 1\ almost at no cost.
> >> 2\ reduce analysis costs, if don't we have to analyze the callers of strtok when you encounter it.
> >>  
> > 
> > Yes but. The replacement should not go to stable.
> > One of the rules of stable is that changes should be minimized, and fixes should
> > not be accepted for things that can not ever happen with current code.  
> 
> Hope more opinion from TB members.

My assumption is that DPDK operates under similar rules as the Linux kernel.
Linux kernel rules are:

Rules on what kind of patches are accepted, and which ones are not, into the “-stable” tree:
    • It or an equivalent fix must already exist in Linux mainline (upstream).
    • It must be obviously correct and tested.
    • It cannot be bigger than 100 lines, with context.
    • It must follow the Documentation/process/submitting-patches.rst rules.
    • It must either fix a real bug that bothers people or just add a device ID. To elaborate on the former:
        ◦ It fixes a problem like an oops, a hang, data corruption, a real security issue, a hardware quirk, a build error (but not for things marked CONFIG_BROKEN), or some “oh, that’s not good” issue.
        ◦ Serious issues as reported by a user of a distribution kernel may also be considered if they fix a notable performance or interactivity issue. As these fixes are not as obvious and have a higher risk of a subtle regression they should only be submitted by a distribution kernel maintainer and include an addendum linking to a bugzilla entry if it exists and additional information on the user-visible impact.
        ◦ No “This could be a problem...” type of things like a “theoretical race condition”, unless an explanation of how the bug can be exploited is also provided.
        ◦ No “trivial” fixes without benefit for users (spelling changes, whitespace cleanups, etc).

Re: [PATCH v4 02/13] eal: replace strtok with reentrant version

[Jie Hai]

Hi, Stephen Hemminger,

Thanks for your reviews.
I'll replace all strtok with strtok_r, and Cc to stable only the 
necessary ones.

Best regards,
Jie Hai

On 2024/10/29 10:51, Stephen Hemminger wrote:
> On Tue, 29 Oct 2024 08:56:20 +0800
> fengchengwen <fengchengwen@huawei.com> wrote:
> 
>>>>>
>>>>> This doesn't need to go to stable. parse_params is always single threaded.
>>>>
>>>> I recommend replacing all, based on:
>>>> 1\ almost at no cost.
>>>> 2\ reduce analysis costs, if don't we have to analyze the callers of strtok when you encounter it.
>>>>   
>>>
>>> Yes but. The replacement should not go to stable.
>>> One of the rules of stable is that changes should be minimized, and fixes should
>>> not be accepted for things that can not ever happen with current code.
>>
>> Hope more opinion from TB members.
> 
> My assumption is that DPDK operates under similar rules as the Linux kernel.
> Linux kernel rules are:
> 
> Rules on what kind of patches are accepted, and which ones are not, into the “-stable” tree:
>      • It or an equivalent fix must already exist in Linux mainline (upstream).
>      • It must be obviously correct and tested.
>      • It cannot be bigger than 100 lines, with context.
>      • It must follow the Documentation/process/submitting-patches.rst rules.
>      • It must either fix a real bug that bothers people or just add a device ID. To elaborate on the former:
>          ◦ It fixes a problem like an oops, a hang, data corruption, a real security issue, a hardware quirk, a build error (but not for things marked CONFIG_BROKEN), or some “oh, that’s not good” issue.
>          ◦ Serious issues as reported by a user of a distribution kernel may also be considered if they fix a notable performance or interactivity issue. As these fixes are not as obvious and have a higher risk of a subtle regression they should only be submitted by a distribution kernel maintainer and include an addendum linking to a bugzilla entry if it exists and additional information on the user-visible impact.
>          ◦ No “This could be a problem...” type of things like a “theoretical race condition”, unless an explanation of how the bug can be exploited is also provided.
>          ◦ No “trivial” fixes without benefit for users (spelling changes, whitespace cleanups, etc).
> .

[PATCH v4 03/13] ethdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: f38f62650f7b ("ethdev: add Rx queue telemetry query")
Fixes: 9e7533aeb80a ("ethdev: add telemetry command for TM level capabilities")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/ethdev/rte_ethdev_telemetry.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/lib/ethdev/rte_ethdev_telemetry.c b/lib/ethdev/rte_ethdev_telemetry.c
index 5e6c4172d3be..0362b1b46d4a 100644
--- a/lib/ethdev/rte_ethdev_telemetry.c
+++ b/lib/ethdev/rte_ethdev_telemetry.c
@@ -477,6 +477,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	const char *qid_param;
 	uint16_t nb_queues;
 	char *end_param;
+	char *sp = NULL;
 	uint64_t qid;
 	int ret;
 
@@ -489,7 +490,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	if (nb_queues == 1 && *end_param == '\0')
 		qid = 0;
 	else {
-		qid_param = strtok(end_param, ",");
+		qid_param = strtok_r(end_param, ",", &sp);
 		if (!qid_param || strlen(qid_param) == 0 || !isdigit(*qid_param))
 			return -EINVAL;
 
@@ -1221,9 +1222,10 @@ static int
 eth_dev_parse_tm_params(char *params, uint32_t *result)
 {
 	const char *splited_param;
+	char *sp = NULL;
 	uint64_t ret;
 
-	splited_param = strtok(params, ",");
+	splited_param = strtok_r(params, ",", &sp);
 	if (!splited_param || strlen(splited_param) == 0 || !isdigit(*splited_param))
 		return -EINVAL;
 
@@ -1510,13 +1512,14 @@ eth_dev_handle_port_regs(const char *cmd __rte_unused,
 {
 	char *filter, *end_param;
 	uint16_t port_id;
+	char *sp = NULL;
 	int ret;
 
 	ret = eth_dev_parse_port_params(params, &port_id, &end_param, true);
 	if (ret != 0)
 		return ret;
 
-	filter = strtok(end_param, ",");
+	filter = strtok_r(end_param, ",", &sp);
 	if (filter != NULL && strlen(filter) == 0)
 		filter = NULL;
 
-- 
2.22.0

Re: [PATCH v4 03/13] ethdev: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:41 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: f38f62650f7b ("ethdev: add Rx queue telemetry query")
> Fixes: 9e7533aeb80a ("ethdev: add telemetry command for TM level capabilities")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
> Acked-by: Morten Brørup <mb@smartsharesystems.com>

Only called from RTE_INIT() so naturally single threaded.
No need for this one in stable.

[PATCH v4 04/13] eventdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/eventdev/rte_event_eth_rx_adapter.c | 38 ++++++++++++-------------
 lib/eventdev/rte_eventdev.c             | 18 ++++++------
 2 files changed, 28 insertions(+), 28 deletions(-)

diff --git a/lib/eventdev/rte_event_eth_rx_adapter.c b/lib/eventdev/rte_event_eth_rx_adapter.c
index 3ee20d95f372..da864a8e1593 100644
--- a/lib/eventdev/rte_event_eth_rx_adapter.c
+++ b/lib/eventdev/rte_event_eth_rx_adapter.c
@@ -3651,7 +3651,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp;
 	struct rte_event_eth_rx_adapter_queue_conf queue_conf;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3661,19 +3661,19 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_PORTID_VALID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3684,7 +3684,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3723,7 +3723,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 	struct rte_event_eth_rx_adapter_queue_stats q_stats;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3733,19 +3733,19 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_PORTID_VALID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3756,7 +3756,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3794,7 +3794,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3803,19 +3803,19 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_PORTID_VALID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3826,7 +3826,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3855,7 +3855,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	uint8_t instance_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3863,14 +3863,14 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_PORTID_VALID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3881,7 +3881,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
diff --git a/lib/eventdev/rte_eventdev.c b/lib/eventdev/rte_eventdev.c
index ca295c87c47d..d3dacffdfab2 100644
--- a/lib/eventdev/rte_eventdev.c
+++ b/lib/eventdev/rte_eventdev.c
@@ -1787,7 +1787,7 @@ handle_queue_links(const char *cmd __rte_unused,
 		   struct rte_tel_data *d)
 {
 	int i, ret, port_id = 0;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	uint8_t dev_id;
 	uint8_t queues[RTE_EVENT_MAX_QUEUES_PER_DEV];
 	uint8_t priorities[RTE_EVENT_MAX_QUEUES_PER_DEV];
@@ -1800,12 +1800,12 @@ handle_queue_links(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
 		return -1;
 
 	port_id = strtoul(p_param, &end_param, 10);
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1925,7 +1925,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1935,7 +1935,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_PORT;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1943,7 +1943,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1959,7 +1959,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1969,7 +1969,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_QUEUE;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1977,7 +1977,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
-- 
2.22.0

Re: [PATCH v4 04/13] eventdev: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:42 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> Acked-by: Morten Brørup <mb@smartsharesystems.com>

Yet another function only called from RTE_INIT()

[PATCH v4 05/13] security: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 259ca6d1617f ("security: add telemetry endpoint for capabilities")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/security/rte_security.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/lib/security/rte_security.c b/lib/security/rte_security.c
index e5c862f5f570..c34b09bedd6a 100644
--- a/lib/security/rte_security.c
+++ b/lib/security/rte_security.c
@@ -497,13 +497,14 @@ security_handle_cryptodev_crypto_caps(const char *cmd __rte_unused, const char *
 	int dev_id, capa_id;
 	int crypto_caps_n;
 	char *end_param;
+	char *sp = NULL;
 	int rc;
 
 	if (!params || strlen(params) == 0 || !isdigit(*params))
 		return -EINVAL;
 
 	dev_id = strtoul(params, &end_param, 0);
-	capa_param = strtok(end_param, ",");
+	capa_param = strtok_r(end_param, ",", &sp);
 	if (!capa_param || strlen(capa_param) == 0 || !isdigit(*capa_param))
 		return -EINVAL;
 
-- 
2.22.0

Re: [PATCH v4 05/13] security: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:43 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: 259ca6d1617f ("security: add telemetry endpoint for capabilities")
> Cc: stable@dpdk.org

Another patch that is not needed in stable.

[PATCH v4 06/13] telemetry: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 6dd571fd07c3 ("telemetry: introduce new functionality")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Ciara Power <ciara.power@intel.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/telemetry/telemetry.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/lib/telemetry/telemetry.c b/lib/telemetry/telemetry.c
index 31a2c91c0657..000b0f9f80df 100644
--- a/lib/telemetry/telemetry.c
+++ b/lib/telemetry/telemetry.c
@@ -398,6 +398,7 @@ static void *
 client_handler(void *sock_id)
 {
 	int s = (int)(uintptr_t)sock_id;
+	char *sp = NULL;
 	char buffer[1024];
 	char info_str[1024];
 	snprintf(info_str, sizeof(info_str),
@@ -412,8 +413,8 @@ client_handler(void *sock_id)
 	int bytes = read(s, buffer, sizeof(buffer) - 1);
 	while (bytes > 0) {
 		buffer[bytes] = 0;
-		const char *cmd = strtok(buffer, ",");
-		const char *param = strtok(NULL, "\0");
+		const char *cmd = strtok_r(buffer, ",", &sp);
+		const char *param = strtok_r(NULL, "\0", &sp);
 		struct cmd_callback cb = {.fn = unknown_command};
 		int i;
 
-- 
2.22.0

Re: [PATCH v4 06/13] telemetry: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:44 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: 6dd571fd07c3 ("telemetry: introduce new functionality")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> Acked-by: Ciara Power <ciara.power@intel.com>
> Acked-by: Morten Brørup <mb@smartsharesystems.com>
> ---
>  lib/telemetry/telemetry.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)

This one could be hit while application is running.

Acked-by: Stephen Hemminger <stephen@networkplumber.org>

[PATCH v4 07/13] bus/fslmc: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 9ccb76b24c1d ("bus/fslmc: enable portal interrupt handling")
Fixes: 828d51d8fc3e ("bus/fslmc: refactor scan and probe functions")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Sachin Saxena <sachin.saxena@nxp.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/bus/fslmc/fslmc_bus.c            | 5 +++--
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 4 ++--
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index 097d6dca08b5..42b6b270a419 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -132,6 +132,7 @@ scan_one_fslmc_device(char *dev_name)
 {
 	char *dup_dev_name, *t_ptr;
 	struct rte_dpaa2_device *dev = NULL;
+	char *sp = NULL;
 	int ret = -1;
 
 	if (!dev_name)
@@ -169,7 +170,7 @@ scan_one_fslmc_device(char *dev_name)
 	}
 
 	/* Parse the device name and ID */
-	t_ptr = strtok(dup_dev_name, ".");
+	t_ptr = strtok_r(dup_dev_name, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Invalid device found: (%s)", dup_dev_name);
 		ret = -EINVAL;
@@ -200,7 +201,7 @@ scan_one_fslmc_device(char *dev_name)
 	else
 		dev->dev_type = DPAA2_UNKNOWN;
 
-	t_ptr = strtok(NULL, ".");
+	t_ptr = strtok_r(NULL, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Skipping invalid device (%s)", dup_dev_name);
 		ret = 0;
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index d8a98326d9de..204662ad9eaf 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -128,7 +128,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 #define AFFINITY_LEN	128
 	uint32_t cpu_mask = 1;
 	size_t len = 0;
-	char *temp = NULL, *token = NULL;
+	char *temp = NULL, *token = NULL, *sp = NULL;
 	char string[STRING_LEN];
 	char smp_affinity[AFFINITY_LEN];
 	FILE *file;
@@ -141,7 +141,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 	}
 	while (getline(&temp, &len, file) != -1) {
 		if ((strstr(temp, string)) != NULL) {
-			token = strtok(temp, ":");
+			token = strtok_r(temp, ":", &sp);
 			break;
 		}
 	}
-- 
2.22.0

Re: [PATCH v4 07/13] bus/fslmc: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:45 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: 9ccb76b24c1d ("bus/fslmc: enable portal interrupt handling")
> Fixes: 828d51d8fc3e ("bus/fslmc: refactor scan and probe functions")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> Acked-by: Sachin Saxena <sachin.saxena@nxp.com>
> Acked-by: Morten Brørup <mb@smartsharesystems.com>
> ---
>  drivers/bus/fslmc/fslmc_bus.c            | 5 +++--
>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 4 ++--
>  2 files changed, 5 insertions(+), 4 deletions(-)

Only called during bus scan so single threaded.
Not needed for stable.

[PATCH v4 08/13] common/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: af75aac78978 ("common/cnxk: support telemetry for NIX")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/common/cnxk/cnxk_telemetry_nix.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/common/cnxk/cnxk_telemetry_nix.c b/drivers/common/cnxk/cnxk_telemetry_nix.c
index abeefafe1e19..00b8ab9e5c18 100644
--- a/drivers/common/cnxk/cnxk_telemetry_nix.c
+++ b/drivers/common/cnxk/cnxk_telemetry_nix.c
@@ -1015,7 +1015,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			   struct plt_tel_data *d)
 {
 	struct nix_tel_node *node;
-	char *name, *param;
+	char *name, *param, *sp = NULL;
 	char buf[1024];
 	int rc = -1;
 
@@ -1023,11 +1023,11 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 		goto exit;
 
 	plt_strlcpy(buf, params, PCI_PRI_STR_SIZE + 1);
-	name = strtok(buf, ",");
+	name = strtok_r(buf, ",", &sp);
 	if (name == NULL)
 		goto exit;
 
-	param = strtok(NULL, "\0");
+	param = strtok_r(NULL, "\0", &sp);
 
 	node = nix_tel_node_get_by_pcidev_name(name);
 	if (!node)
@@ -1036,7 +1036,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 	plt_tel_data_start_dict(d);
 
 	if (strstr(cmd, "rq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int rq;
 
 		if (!tok)
@@ -1052,7 +1052,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_rq(node->rqs[rq], d);
 
 	} else if (strstr(cmd, "cq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int cq;
 
 		if (!tok)
@@ -1068,7 +1068,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_cq(node->cqs[cq], d);
 
 	} else if (strstr(cmd, "sq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int sq;
 
 		if (!tok)
-- 
2.22.0

Re: [PATCH v4 08/13] common/cnxk: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:46 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: af75aac78978 ("common/cnxk: support telemetry for NIX")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> Acked-by: Morten Brørup <mb@smartsharesystems.com>
> ---

These info functions can be called when running.

Acked-by: Stephen Hemminger <stephen@networkplumber.org>

[PATCH v4 09/13] event/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 8a3d58c189fd ("event/cnxk: add option to control timer adapters")
Fixes: 8bdbae66b299 ("event/cnxk: add external clock support for timer")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/event/cnxk/cnxk_eventdev.c  | 10 ++++++----
 drivers/event/cnxk/cnxk_tim_evdev.c | 11 ++++++-----
 2 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/drivers/event/cnxk/cnxk_eventdev.c b/drivers/event/cnxk/cnxk_eventdev.c
index 84a55511a328..feb23e0b2177 100644
--- a/drivers/event/cnxk/cnxk_eventdev.c
+++ b/drivers/event/cnxk/cnxk_eventdev.c
@@ -482,7 +482,8 @@ parse_queue_param(char *value, void *opaque)
 	struct cnxk_sso_qos queue_qos = {0};
 	uint16_t *val = (uint16_t *)&queue_qos;
 	struct cnxk_sso_evdev *dev = opaque;
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "-", &sp);
 	struct cnxk_sso_qos *old_ptr;
 
 	if (!strlen(value))
@@ -490,7 +491,7 @@ parse_queue_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		val++;
 	}
 
@@ -518,7 +519,8 @@ parse_stash_param(char *value, void *opaque)
 	struct cnxk_sso_stash queue_stash = {0};
 	struct cnxk_sso_evdev *dev = opaque;
 	struct cnxk_sso_stash *old_ptr;
-	char *tok = strtok(value, "|");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "|", &sp);
 	uint16_t *val;
 
 	if (!strlen(value))
@@ -527,7 +529,7 @@ parse_stash_param(char *value, void *opaque)
 	val = (uint16_t *)&queue_stash;
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "|");
+		tok = strtok_r(NULL, "|", &sp);
 		val++;
 	}
 
diff --git a/drivers/event/cnxk/cnxk_tim_evdev.c b/drivers/event/cnxk/cnxk_tim_evdev.c
index bba70646fa16..63f868bdb33d 100644
--- a/drivers/event/cnxk/cnxk_tim_evdev.c
+++ b/drivers/event/cnxk/cnxk_tim_evdev.c
@@ -420,7 +420,8 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 {
 	struct cnxk_tim_evdev *dev = opaque;
 	struct cnxk_tim_ctl ring_ctl = {0};
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "-", &sp);
 	struct cnxk_tim_ctl *old_ptr;
 	uint16_t *val;
 
@@ -431,7 +432,7 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		val++;
 	}
 
@@ -507,16 +508,16 @@ cnxk_tim_parse_clk_list(const char *value, void *opaque)
 				      ROC_TIM_CLK_SRC_INVALID};
 	struct cnxk_tim_evdev *dev = opaque;
 	char *str = strdup(value);
-	char *tok;
+	char *tok, *sp = NULL;
 	int i = 0;
 
 	if (str == NULL || !strlen(str))
 		goto free;
 
-	tok = strtok(str, "-");
+	tok = strtok_r(str, "-", &sp);
 	while (tok != NULL && src[i] != ROC_TIM_CLK_SRC_INVALID) {
 		dev->ext_clk_freq[src[i]] = strtoull(tok, NULL, 10);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		i++;
 	}
 
-- 
2.22.0

Re: [PATCH v4 09/13] event/cnxk: replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:47 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> The strtok() is non-reentrant, it is better to replace it with a
> reentrant version.
> 
> Fixes: 8a3d58c189fd ("event/cnxk: add option to control timer adapters")
> Fixes: 8bdbae66b299 ("event/cnxk: add external clock support for timer")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jie Hai <haijie1@huawei.com>
> Acked-by: Chengwen Feng <fengchengwen@huawei.com>
> Acked-by: Morten Brørup <mb@smartsharesystems.com>

Another set of parse routines only called during startup.
Patch is unnecessary for stable.

[PATCH v4 10/13] net/ark: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 9c7188a68d7b ("net/ark: provide API for hardware modules pktchkr and pktgen")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/net/ark/ark_pktchkr.c | 10 +++++-----
 drivers/net/ark/ark_pktgen.c  | 10 +++++-----
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ark/ark_pktchkr.c b/drivers/net/ark/ark_pktchkr.c
index e1f336c73c2a..6e8e5339607b 100644
--- a/drivers/net/ark/ark_pktchkr.c
+++ b/drivers/net/ark/ark_pktchkr.c
@@ -359,14 +359,14 @@ set_arg(char *arg, char *val)
 void
 ark_pktchkr_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = "=\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_r(args, toks, &sp);
+	v = strtok_r(NULL, toks, &sp);
 	while (argv && v) {
 		set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_r(NULL, toks, &sp);
+		v = strtok_r(NULL, toks, &sp);
 	}
 }
 
diff --git a/drivers/net/ark/ark_pktgen.c b/drivers/net/ark/ark_pktgen.c
index 69ff7072b2ab..d611406a1b46 100644
--- a/drivers/net/ark/ark_pktgen.c
+++ b/drivers/net/ark/ark_pktgen.c
@@ -340,14 +340,14 @@ pmd_set_arg(char *arg, char *val)
 void
 ark_pktgen_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = " =\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_r(args, toks, &sp);
+	v = strtok_r(NULL, toks, &sp);
 	while (argv && v) {
 		pmd_set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_r(NULL, toks, &sp);
+		v = strtok_r(NULL, toks, &sp);
 	}
 }
 
-- 
2.22.0

[PATCH v4 11/13] raw/cnxk_gpio: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: ecc0dd455e9a ("raw/cnxk_gpio: add option to select subset of GPIOs")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/raw/cnxk_gpio/cnxk_gpio.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/raw/cnxk_gpio/cnxk_gpio.c b/drivers/raw/cnxk_gpio/cnxk_gpio.c
index 329ac28a2736..bad70ecb1985 100644
--- a/drivers/raw/cnxk_gpio/cnxk_gpio.c
+++ b/drivers/raw/cnxk_gpio/cnxk_gpio.c
@@ -192,7 +192,7 @@ static int
 cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 {
 	int i, ret, val, queue = 0;
-	char *token;
+	char *token, *sp = NULL;
 	int *list;
 
 	list = rte_calloc(NULL, gpiochip->num_gpios, sizeof(*list), 0);
@@ -210,7 +210,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 	allowlist[strlen(allowlist) - 1] = ' ';
 
 	/* quiesce -Wcast-qual */
-	token = strtok((char *)(uintptr_t)allowlist, ",");
+	token = strtok_r((char *)(uintptr_t)allowlist, ",", &sp);
 	do {
 		errno = 0;
 		val = strtol(token, NULL, 10);
@@ -236,7 +236,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 		}
 		if (i == queue)
 			list[queue++] = val;
-	} while ((token = strtok(NULL, ",")));
+	} while ((token = strtok_r(NULL, ",", &sp)));
 
 	free(allowlist);
 	gpiochip->allowlist = list;
-- 
2.22.0

[PATCH v4 12/13] net/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: c8f91985331c ("raw/cnxk_gpio: replace strtok with reentrant version")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c b/drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c
index 86c2453c0983..fa1d7c063536 100644
--- a/drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c
+++ b/drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c
@@ -214,6 +214,7 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	char dlim[2] = ",";
 	char *params_args;
 	size_t count = 0;
+	char *sp = NULL;
 	char *token;
 
 	if (vals == NULL || params == NULL || strlen(params) == 0)
@@ -226,10 +227,10 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	if (params_args == NULL)
 		return -1;
 
-	token = strtok(params_args, dlim);
+	token = strtok_r(params_args, dlim, &sp);
 	while (token && isdigit(*token) && count < n_vals) {
 		vals[count++] = strtoul(token, NULL, 10);
-		token = strtok(NULL, dlim);
+		token = strtok_r(NULL, dlim, &sp);
 	}
 
 	free(params_args);
-- 
2.22.0

[PATCH v4 13/13] devtools: check for some reentrant function

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

This patch adds check in checkpatches.sh for strtok, which is
non-reentrant.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 devtools/checkpatches.sh | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/devtools/checkpatches.sh b/devtools/checkpatches.sh
index d860f190457e..8c3d32d8e552 100755
--- a/devtools/checkpatches.sh
+++ b/devtools/checkpatches.sh
@@ -145,6 +145,14 @@ check_forbidden_additions() { # <patch>
 		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
 		"$1" || res=1
 
+	# refrain from using some non-reentrant functions
+	awk -v FOLDERS="lib drivers app examples" \
+		-v EXPRESSIONS="strtok\\\(" \
+		-v RET_ON_FAIL=1 \
+		-v MESSAGE='Using non-reentrant function strtok, prefer strtok_r' \
+		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
+		"$1" || res=1
+
 	# refrain from using some pthread functions
 	awk -v FOLDERS="lib drivers app examples" \
 		-v EXPRESSIONS="pthread_(create|join|detach|set(_?name_np|affinity_np)|attr_set(inheritsched|schedpolicy))\\\(" \
-- 
2.22.0

Re: [PATCH v4 13/13] devtools: check for some reentrant function

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:51 +0800
Jie Hai <haijie1@huawei.com> wrote:

> diff --git a/devtools/checkpatches.sh b/devtools/checkpatches.sh
> index d860f190457e..8c3d32d8e552 100755
> --- a/devtools/checkpatches.sh
> +++ b/devtools/checkpatches.sh
> @@ -145,6 +145,14 @@ check_forbidden_additions() { # <patch>
>  		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
>  		"$1" || res=1
>  
> +	# refrain from using some non-reentrant functions
> +	awk -v FOLDERS="lib drivers app examples" \
> +		-v EXPRESSIONS="strtok\\\(" \
> +		-v RET_ON_FAIL=1 \
> +		-v MESSAGE='Using non-reentrant function strtok, prefer strtok_r' \
> +		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
> +		"$1" || res=1
> +
>  	# refrain from using some pthread functions
>  	awk -v FOLDERS="lib drivers app examples" \
>  		-v EXPRESSIONS="pthread_(create|join|detach|set(_?name_np|affinity_np)|attr_set(inheritsched|schedpolicy))\\\(" \

Ok, but checkpatch keeps getting slower and slower with all this copy baset.
Could someone make it into a python script without so much copy/paste?

Re: [PATCH v4 00/13] replace strtok with reentrant version

[Stephen Hemminger]

On Sat, 26 Oct 2024 18:14:38 +0800
Jie Hai <haijie1@huawei.com> wrote:

> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
> 
> This patchset replaces strtok with strtok_r in app, example, lib
> and drivers. And adds check for use of strtok in checkpatches.sh.
> 
> --
> v4:
> 1. fix mispellings.
> 2. add Acked-bys and Reviewd-bys.
> 3. remove some patch and add new. 
> v3:
> 1. fix compile error.
> 2. use strtok_r instead.
> v2:
> 1. fix commit log.
> 2. add check in checkpatches.sh.
> 3. replace strtok_r with strtok_s.
> 4. add Acked-by.
> --
> 
> Jie Hai (13):
>   dmadev: replace strtok with reentrant version
>   eal: replace strtok with reentrant version
>   ethdev: replace strtok with reentrant version
>   eventdev: replace strtok with reentrant version
>   security: replace strtok with reentrant version
>   telemetry: replace strtok with reentrant version
>   bus/fslmc: replace strtok with reentrant version
>   common/cnxk: replace strtok with reentrant version
>   event/cnxk: replace strtok with reentrant version
>   net/ark: replace strtok with reentrant version
>   raw/cnxk_gpio: replace strtok with reentrant version
>   net/cnxk: replace strtok with reentrant version
>   devtools: check for some reentrant function
> 
>  devtools/checkpatches.sh                     |  8 +++++
>  drivers/bus/fslmc/fslmc_bus.c                |  5 +--
>  drivers/bus/fslmc/portal/dpaa2_hw_dpio.c     |  4 +--
>  drivers/common/cnxk/cnxk_telemetry_nix.c     | 12 +++----
>  drivers/event/cnxk/cnxk_eventdev.c           | 10 +++---
>  drivers/event/cnxk/cnxk_tim_evdev.c          | 11 +++---
>  drivers/net/ark/ark_pktchkr.c                | 10 +++---
>  drivers/net/ark/ark_pktgen.c                 | 10 +++---
>  drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c |  5 +--
>  drivers/raw/cnxk_gpio/cnxk_gpio.c            |  6 ++--
>  lib/dmadev/rte_dmadev.c                      |  4 +--
>  lib/eal/common/eal_common_memory.c           |  8 ++---
>  lib/ethdev/rte_ethdev_telemetry.c            |  9 +++--
>  lib/eventdev/rte_event_eth_rx_adapter.c      | 38 ++++++++++----------
>  lib/eventdev/rte_eventdev.c                  | 18 +++++-----
>  lib/security/rte_security.c                  |  3 +-
>  lib/telemetry/telemetry.c                    |  5 +--
>  17 files changed, 92 insertions(+), 74 deletions(-)
> 

This should go into main (for 24.11) but not be backported to stable.

Series-Acked-by: Stephen Hemminger <stephen@networkplumber.org>

Re: [PATCH v4 00/13] replace strtok with reentrant version

[David Marchand]

Hello,

On Sat, Oct 26, 2024 at 12:26 PM Jie Hai <haijie1@huawei.com> wrote:
>
> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
>
> This patchset replaces strtok with strtok_r in app, example, lib
> and drivers. And adds check for use of strtok in checkpatches.sh.

- The current v4 series breaks compilation on Windows with clang.
http://mails.dpdk.org/archives/test-report/2024-November/819978.html

The reason is that some include of rte_os_shim.h are missing.

==== 20 line log output for Windows Server 2022 (dpdk_win_llvm_compile): ====
vchan_param = strtok_r(end_param, ",", &sp);
^
../lib/dmadev/rte_dmadev.c:1038:17: note: did you mean 'strtok_s'?
C:\Program Files (x86)\Windows
Kits\10\Include\10.0.22621.0\ucrt\string.h:68:29: note: 'strtok_s'
declared here
_ACRTIMP char*  __cdecl strtok_s(
^
../lib/dmadev/rte_dmadev.c:1038:15: error: incompatible integer to
pointer conversion assigning to 'const char *' from 'int'
[-Wint-conversion]
vchan_param = strtok_r(end_param, ",", &sp);
^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2 errors generated.
[10/836] Linking static target lib/librte_gpudev.a
[11/836] Compiling C object
lib/librte_mldev.a.p/mldev_mldev_utils_scalar_bfloat16.c.obj
[12/836] Compiling C object lib/librte_gro.a.p/gro_gro_udp4.c.obj
[13/836] Compiling C object lib/librte_mldev.a.p/mldev_mldev_utils_scalar.c.obj
[14/836] Compiling C object lib/librte_gro.a.p/gro_rte_gro.c.obj
[15/836] Compiling C object lib/librte_gro.a.p/gro_gro_tcp6.c.obj
[16/836] Compiling C object lib/librte_gro.a.p/gro_gro_tcp4.c.obj
[17/836] Compiling C object lib/librte_gro.a.p/gro_gro_vxlan_tcp4.c.obj
[18/836] Compiling C object
lib/librte_cryptodev.a.p/cryptodev_rte_cryptodev.c.obj
ninja: build stopped: subcommand failed.


- grep shows that there are a number of missed places in drivers and
app, which contradicts the commitlog.
I don't like leaving some places with strtok while checkpatches.sh
warn on the rest.
And I think others were expecting too that the whole tree is fixed
after this series.

Re-reading the thread, were the changes on app/ dropped, following
comments from Stephen?
I understand his request was to remove the backport request (iow no
Cc: stable@dpdk.org in the commitlogs), not to drop the changes.


-- 
David Marchand

Re: [PATCH v4 00/13] replace strtok with reentrant version

[Jie Hai]

Hi, David Marchand ,

Thank you very much for your reminder and comments. I will correct it in 
the next version.

Best Regards,
Jie Hai

On 2024/11/7 4:11, David Marchand wrote:
> Hello,
> 
> On Sat, Oct 26, 2024 at 12:26 PM Jie Hai <haijie1@huawei.com> wrote:
>>
>> Multiple threads calling the same function may cause condition
>> race issues, which often leads to abnormal behavior and can cause
>> more serious vulnerabilities such as abnormal termination, denial
>> of service, and compromised data integrity.
>>
>> This patchset replaces strtok with strtok_r in app, example, lib
>> and drivers. And adds check for use of strtok in checkpatches.sh.
> 
> - The current v4 series breaks compilation on Windows with clang.
> http://mails.dpdk.org/archives/test-report/2024-November/819978.html
> 
> The reason is that some include of rte_os_shim.h are missing.
> 
> ==== 20 line log output for Windows Server 2022 (dpdk_win_llvm_compile): ====
> vchan_param = strtok_r(end_param, ",", &sp);
> ^
> ../lib/dmadev/rte_dmadev.c:1038:17: note: did you mean 'strtok_s'?
> C:\Program Files (x86)\Windows
> Kits\10\Include\10.0.22621.0\ucrt\string.h:68:29: note: 'strtok_s'
> declared here
> _ACRTIMP char*  __cdecl strtok_s(
> ^
> ../lib/dmadev/rte_dmadev.c:1038:15: error: incompatible integer to
> pointer conversion assigning to 'const char *' from 'int'
> [-Wint-conversion]
> vchan_param = strtok_r(end_param, ",", &sp);
> ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 2 errors generated.
> [10/836] Linking static target lib/librte_gpudev.a
> [11/836] Compiling C object
> lib/librte_mldev.a.p/mldev_mldev_utils_scalar_bfloat16.c.obj
> [12/836] Compiling C object lib/librte_gro.a.p/gro_gro_udp4.c.obj
> [13/836] Compiling C object lib/librte_mldev.a.p/mldev_mldev_utils_scalar.c.obj
> [14/836] Compiling C object lib/librte_gro.a.p/gro_rte_gro.c.obj
> [15/836] Compiling C object lib/librte_gro.a.p/gro_gro_tcp6.c.obj
> [16/836] Compiling C object lib/librte_gro.a.p/gro_gro_tcp4.c.obj
> [17/836] Compiling C object lib/librte_gro.a.p/gro_gro_vxlan_tcp4.c.obj
> [18/836] Compiling C object
> lib/librte_cryptodev.a.p/cryptodev_rte_cryptodev.c.obj
> ninja: build stopped: subcommand failed.
> 
> 
> - grep shows that there are a number of missed places in drivers and
> app, which contradicts the commitlog.
> I don't like leaving some places with strtok while checkpatches.sh
> warn on the rest.
> And I think others were expecting too that the whole tree is fixed
> after this series.
> 
> Re-reading the thread, were the changes on app/ dropped, following
> comments from Stephen?
> I understand his request was to remove the backport request (iow no
> Cc: stable@dpdk.org in the commitlogs), not to drop the changes.
> 
> 

[PATCH v5 00/25]

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

This patchset replaces strtok with strtok_r in app, example, lib
and drivers. And adds check for use of strtok in checkpatches.sh.

--
v5:
1. remove CC stable for some patch.
2. replace strtok for all files.

v4:
1. fix mispellings.
2. add Acked-bys and Reviewd-bys.
3. remove some patch and add new. 
v3:
1. fix compile error.
2. use strtok_r instead.
v2:
1. fix commit log.
2. add check in checkpatches.sh.
3. replace strtok_r with strtok_s.
4. add Acked-by.
--

Jie Hai (25):
  app/graph: replace strtok with reentrant version
  app/bbdev: replace strtok with reentrant version
  app/compress-perf: replace strtok with reentrant version
  app/crypto-perf: replace strtok with reentrant version
  app/dma-perf: replace strtok with reentrant version
  app/flow-perf: replace strtok with reentrant version
  app/test-mldev: replace strtok with reentrant version
  app/test-fib: replace strtok with reentrant version
  dmadev: replace strtok with reentrant version
  eal: replace strtok with reentrant version
  ethdev: replace strtok with reentrant version
  eventdev: replace strtok with reentrant version
  security: replace strtok with reentrant version
  telemetry: replace strtok with reentrant version
  bus/fslmc: replace strtok with reentrant version
  common/cnxk: replace strtok with reentrant version
  event/cnxk: replace strtok with reentrant version
  net/ark: replace strtok with reentrant version
  raw/cnxk_gpio: replace strtok with reentrant version
  net/cnxk: replace strtok with reentrant version
  common/qat: replace strtok with reentrant version
  net/mlx5: replace strtok with reentrant version
  examples/l2fwd-crypto: replace strtok with reentrant version
  examples/vhost: replace strtok with reentrant version
  devtools: check for some reentrant function

 app/graph/graph.c                             |  6 ++-
 app/graph/utils.c                             | 16 ++++---
 app/test-bbdev/test_bbdev_vector.c            | 42 +++++++++++--------
 .../comp_perf_options_parse.c                 | 17 ++++----
 app/test-crypto-perf/cperf_options_parsing.c  | 17 ++++----
 .../cperf_test_vector_parsing.c               | 11 +++--
 app/test-dma-perf/main.c                      | 14 ++++---
 app/test-fib/main.c                           | 11 ++---
 app/test-flow-perf/main.c                     | 23 +++++-----
 app/test-mldev/ml_options.c                   | 19 +++++----
 devtools/checkpatches.sh                      |  8 ++++
 drivers/bus/fslmc/fslmc_bus.c                 |  6 ++-
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c      |  5 ++-
 drivers/common/cnxk/cnxk_telemetry_nix.c      | 13 +++---
 drivers/common/qat/qat_device.c               |  6 ++-
 drivers/event/cnxk/cnxk_eventdev.c            | 11 +++--
 drivers/event/cnxk/cnxk_tim_evdev.c           | 12 +++---
 drivers/net/ark/ark_pktchkr.c                 | 11 ++---
 drivers/net/ark/ark_pktgen.c                  | 11 ++---
 drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c  |  6 ++-
 drivers/net/mlx5/mlx5_testpmd.c               |  5 ++-
 drivers/raw/cnxk_gpio/cnxk_gpio.c             |  7 ++--
 examples/l2fwd-crypto/main.c                  |  7 ++--
 examples/vhost/main.c                         |  4 +-
 lib/dmadev/rte_dmadev.c                       |  5 ++-
 lib/eal/common/eal_common_memory.c            |  8 ++--
 lib/ethdev/rte_ethdev_telemetry.c             | 10 +++--
 lib/eventdev/rte_event_eth_rx_adapter.c       | 39 ++++++++---------
 lib/eventdev/rte_eventdev.c                   | 18 ++++----
 lib/security/rte_security.c                   |  4 +-
 lib/telemetry/telemetry.c                     |  6 ++-
 31 files changed, 223 insertions(+), 155 deletions(-)

-- 
2.22.0

[PATCH v5 01/25] app/graph: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 5c59002a34f3 ("app/graph: add graph commands")
Fixes: 984a315a5804 ("app/graph: add parser utility")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/graph/graph.c |  6 ++++--
 app/graph/utils.c | 16 ++++++++++------
 2 files changed, 14 insertions(+), 8 deletions(-)

diff --git a/app/graph/graph.c b/app/graph/graph.c
index 17717738d7d9..6849b8da5da4 100644
--- a/app/graph/graph.c
+++ b/app/graph/graph.c
@@ -13,6 +13,7 @@
 #include <rte_ethdev.h>
 #include <rte_graph_worker.h>
 #include <rte_log.h>
+#include <rte_os_shim.h>
 
 #include "graph_priv.h"
 #include "module_api.h"
@@ -103,9 +104,10 @@ parser_usecases_read(char *usecases)
 {
 	bool valid = false;
 	uint32_t i, j = 0;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(usecases, ",");
+	token = strtok_r(usecases, ",", &sp);
 	while (token != NULL) {
 		for (i = 0; i < RTE_DIM(supported_usecases); i++) {
 			if (strcmp(supported_usecases[i], token) == 0) {
@@ -116,7 +118,7 @@ parser_usecases_read(char *usecases)
 				break;
 			}
 		}
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	return valid;
diff --git a/app/graph/utils.c b/app/graph/utils.c
index 3e8099ea8821..6c53948aa277 100644
--- a/app/graph/utils.c
+++ b/app/graph/utils.c
@@ -10,6 +10,7 @@
 
 #include <rte_common.h>
 #include <rte_string_fns.h>
+#include <rte_os_shim.h>
 
 #include "module_api.h"
 
@@ -94,13 +95,14 @@ int
 parser_ip4_read(uint32_t *value, char *p)
 {
 	uint8_t shift = 24;
+	char *sp = NULL;
 	uint32_t ip = 0;
 	char *token;
 
-	token = strtok(p, ".");
+	token = strtok_r(p, ".", &sp);
 	while (token != NULL) {
 		ip |= (((uint32_t)strtoul(token, NULL, 10)) << shift);
-		token = strtok(NULL, ".");
+		token = strtok_r(NULL, ".", &sp);
 		shift -= 8;
 	}
 
@@ -113,13 +115,14 @@ int
 parser_ip6_read(uint8_t *value, char *p)
 {
 	uint64_t val = 0;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(p, ":");
+	token = strtok_r(p, ":", &sp);
 	while (token != NULL) {
 		hex_string_to_uint64(&val, token);
 		*value = val;
-		token = strtok(NULL, ":");
+		token = strtok_r(NULL, ":", &sp);
 		value++;
 		val = 0;
 	}
@@ -132,13 +135,14 @@ parser_mac_read(uint64_t *value, char *p)
 {
 	uint64_t mac = 0, val = 0;
 	uint8_t shift = 40;
+	char *sp = NULL;
 	char *token;
 
-	token = strtok(p, ":");
+	token = strtok_r(p, ":", &sp);
 	while (token != NULL) {
 		hex_string_to_uint64(&val, token);
 		mac |= val << shift;
-		token = strtok(NULL, ":");
+		token = strtok_r(NULL, ":", &sp);
 		shift -= 8;
 		val = 0;
 	}
-- 
2.22.0

[PATCH v5 02/25] app/bbdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 0acdb9866756 ("test/bbdev: add FFT operations cases")
Fixes: f714a18885a6 ("app/testbbdev: add test application for bbdev")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-bbdev/test_bbdev_vector.c | 42 +++++++++++++++++-------------
 1 file changed, 24 insertions(+), 18 deletions(-)

diff --git a/app/test-bbdev/test_bbdev_vector.c b/app/test-bbdev/test_bbdev_vector.c
index 8b32850982c3..fbbdde70bdfc 100644
--- a/app/test-bbdev/test_bbdev_vector.c
+++ b/app/test-bbdev/test_bbdev_vector.c
@@ -10,6 +10,7 @@
 #include <stdlib.h>
 #include <stdbool.h>
 #include <rte_malloc.h>
+#include <rte_os_shim.h>
 
 #include "test_bbdev_vector.h"
 
@@ -63,8 +64,9 @@ parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
 
 	uint32_t *values, *values_resized;
 	char *tok, *error = NULL;
+	char *sp = NULL;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -98,7 +100,7 @@ parse_values(char *tokens, uint32_t **data, uint32_t *data_length)
 
 		*data_length = *data_length + (strlen(tok) - strlen("0x"))/2;
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 
@@ -324,8 +326,9 @@ parse_turbo_flags(char *tokens, uint32_t *op_flags,
 {
 	char *tok = NULL;
 	uint32_t op_flag_value = 0;
+	char *sp = NULL;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -359,7 +362,7 @@ parse_turbo_flags(char *tokens, uint32_t *op_flags,
 
 		*op_flags = *op_flags | op_flag_value;
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 	}
@@ -399,9 +402,10 @@ static int
 parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
 {
 	char *tok = NULL;
+	char *sp = NULL;
 	bool status_ok = false;
 
-	tok = strtok(tokens, VALUE_DELIMITER);
+	tok = strtok_r(tokens, VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -432,7 +436,7 @@ parse_expected_status(char *tokens, int *status, enum rte_bbdev_op_type op_type)
 			return -1;
 		}
 
-		tok = strtok(NULL, VALUE_DELIMITER);
+		tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 	}
@@ -932,6 +936,7 @@ parse_fft_params(const char *key_token, char *token,
 	int ret = 0, status = 0, i, shift;
 	uint32_t op_flags = 0;
 	char *tok, *err = NULL;
+	char *sp = NULL;
 
 	struct rte_bbdev_op_fft *fft = &vector->fft;
 
@@ -964,7 +969,7 @@ parse_fft_params(const char *key_token, char *token,
 		fft->output_leading_depadding = (uint32_t) strtoul(token, &err, 0);
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "window_index")) {
-		tok = strtok(token, VALUE_DELIMITER);
+		tok = strtok_r(token, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			return -1;
 		for (i = 0; i < FFT_WIN_SIZE; i++) {
@@ -972,7 +977,7 @@ parse_fft_params(const char *key_token, char *token,
 			fft->window_index[i / 2] |= (uint32_t) strtoul(tok, &err, 0)
 					<< shift;
 			if (i < (FFT_WIN_SIZE - 1)) {
-				tok = strtok(NULL, VALUE_DELIMITER);
+				tok = strtok_r(NULL, VALUE_DELIMITER, &sp);
 				if (tok == NULL)
 					return -1;
 			}
@@ -1016,53 +1021,53 @@ parse_fft_params(const char *key_token, char *token,
 		fft->output_depadded_size = (uint32_t) strtoul(token, &err, 0);
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "cs_theta_0")) {
-		tok = strtok(token, VALUE_DELIMITER);
+		tok = strtok_r(token, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			return -1;
 		for (i = 0; i < FFT_WIN_SIZE; i++) {
 			fft->cs_theta_0[i] = (uint32_t) strtoul(tok, &err, 0);
 			if (i < (FFT_WIN_SIZE - 1)) {
-				tok = strtok(NULL, VALUE_DELIMITER);
+				tok = strtok_r(token, VALUE_DELIMITER, &sp);
 				if (tok == NULL)
 					return -1;
 			}
 		}
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "cs_theta_d")) {
-		tok = strtok(token, VALUE_DELIMITER);
+		tok = strtok_r(token, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			return -1;
 		for (i = 0; i < FFT_WIN_SIZE; i++) {
 			fft->cs_theta_d[i] = (uint32_t) strtoul(tok, &err, 0);
 			if (i < (FFT_WIN_SIZE - 1)) {
-				tok = strtok(NULL, VALUE_DELIMITER);
+				tok = strtok_r(token, VALUE_DELIMITER, &sp);
 				if (tok == NULL)
 					return -1;
 			}
 		}
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "time_offset")) {
-		tok = strtok(token, VALUE_DELIMITER);
+		tok = strtok_r(token, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			return -1;
 		for (i = 0; i < FFT_WIN_SIZE; i++) {
 			fft->time_offset[i] = (uint32_t) strtoul(tok, &err, 0);
 			if (i < (FFT_WIN_SIZE - 1)) {
-				tok = strtok(NULL, VALUE_DELIMITER);
+				tok = strtok_r(token, VALUE_DELIMITER, &sp);
 				if (tok == NULL)
 					return -1;
 			}
 		}
 		ret = ((err == NULL) || (*err != '\0')) ? -1 : 0;
 	} else if (!strcmp(key_token, "fft_window_width")) {
-		tok = strtok(token, VALUE_DELIMITER);
+		tok = strtok_r(token, VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			return -1;
 		for (i = 0; i < FFT_WIN_SIZE; i++) {
 			if (i == 0)
 				vector->fft_window_width_vec = (uint32_t) strtoul(tok, &err, 0);
 			if (i < (FFT_WIN_SIZE - 1)) {
-				tok = strtok(NULL, VALUE_DELIMITER);
+				tok = strtok_r(token, VALUE_DELIMITER, &sp);
 				if (tok == NULL)
 					return -1;
 			}
@@ -1163,6 +1168,7 @@ static int
 parse_entry(char *entry, struct test_bbdev_vector *vector)
 {
 	int ret = 0;
+	char *sp = NULL;
 	char *token, *key_token;
 	enum rte_bbdev_op_type op_type = RTE_BBDEV_OP_NONE;
 
@@ -1172,10 +1178,10 @@ parse_entry(char *entry, struct test_bbdev_vector *vector)
 	}
 
 	/* get key */
-	token = strtok(entry, ENTRY_DELIMITER);
+	token = strtok_r(entry, ENTRY_DELIMITER, &sp);
 	key_token = token;
 	/* get values for key */
-	token = strtok(NULL, ENTRY_DELIMITER);
+	token = strtok_r(NULL, ENTRY_DELIMITER, &sp);
 
 	if (key_token == NULL || token == NULL) {
 		printf("Expected 'key = values' but was '%.40s'..\n", entry);
-- 
2.22.0

[PATCH v5 03/25] app/compress-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: e0b6287c035d ("app/compress-perf: add parser")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 .../comp_perf_options_parse.c                   | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/app/test-compress-perf/comp_perf_options_parse.c b/app/test-compress-perf/comp_perf_options_parse.c
index 6d8c370fc2ea..6f3f907162e6 100644
--- a/app/test-compress-perf/comp_perf_options_parse.c
+++ b/app/test-compress-perf/comp_perf_options_parse.c
@@ -12,6 +12,7 @@
 
 #include <rte_string_fns.h>
 #include <rte_comp.h>
+#include <rte_os_shim.h>
 
 #include "comp_perf_options.h"
 
@@ -177,6 +178,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 {
 	char *token;
 	uint8_t number;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -184,7 +186,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ":");
+	token = strtok_r(copy_arg, ":", &sp);
 
 	/* Parse minimum value */
 	if (token != NULL) {
@@ -197,7 +199,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse increment value */
 	if (token != NULL) {
@@ -211,7 +213,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse maximum value */
 	if (token != NULL) {
@@ -225,7 +227,7 @@ parse_range(const char *arg, uint8_t *min, uint8_t *max, uint8_t *inc)
 	} else
 		goto err_range;
 
-	if (strtok(NULL, ":") != NULL)
+	if (strtok_r(NULL, ":", &sp) != NULL)
 		goto err_range;
 
 	free(copy_arg);
@@ -244,6 +246,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 	uint8_t count = 0;
 	uint32_t temp_min;
 	uint32_t temp_max;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -251,7 +254,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ",");
+	token = strtok_r(copy_arg, ",", &sp);
 
 	/* Parse first value */
 	if (token != NULL) {
@@ -266,7 +269,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 	} else
 		goto err_list;
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 
 	while (token != NULL) {
 		if (count == MAX_LIST) {
@@ -288,7 +291,7 @@ parse_list(const char *arg, uint8_t *list, uint8_t *min, uint8_t *max)
 		if (number > temp_max)
 			temp_max = number;
 
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	if (min)
-- 
2.22.0

[PATCH v5 04/25] app/crypto-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: f6cefe253cc8 ("app/crypto-perf: add range/list of sizes")
Fixes: f8be1786b1b8 ("app/crypto-perf: introduce performance test application")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Ciara Power <ciara.power@intel.com>
---
 app/test-crypto-perf/cperf_options_parsing.c    | 17 ++++++++++-------
 .../cperf_test_vector_parsing.c                 | 11 +++++++----
 2 files changed, 17 insertions(+), 11 deletions(-)

diff --git a/app/test-crypto-perf/cperf_options_parsing.c b/app/test-crypto-perf/cperf_options_parsing.c
index 8abee2d68826..07130e2c27f7 100644
--- a/app/test-crypto-perf/cperf_options_parsing.c
+++ b/app/test-crypto-perf/cperf_options_parsing.c
@@ -9,6 +9,7 @@
 #include <rte_cryptodev.h>
 #include <rte_malloc.h>
 #include <rte_ether.h>
+#include <rte_os_shim.h>
 
 #include "cperf_options.h"
 #include "cperf_test_common.h"
@@ -166,6 +167,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 {
 	char *token;
 	uint32_t number;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -173,7 +175,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ":");
+	token = strtok_r(copy_arg, ":", &sp);
 
 	/* Parse minimum value */
 	if (token != NULL) {
@@ -187,7 +189,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse increment value */
 	if (token != NULL) {
@@ -201,7 +203,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	token = strtok(NULL, ":");
+	token = strtok_r(NULL, ":", &sp);
 
 	/* Parse maximum value */
 	if (token != NULL) {
@@ -216,7 +218,7 @@ parse_range(const char *arg, uint32_t *min, uint32_t *max, uint32_t *inc)
 	} else
 		goto err_range;
 
-	if (strtok(NULL, ":") != NULL)
+	if (strtok_r(NULL, ":", &sp) != NULL)
 		goto err_range;
 
 	free(copy_arg);
@@ -235,6 +237,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 	uint8_t count = 0;
 	uint32_t temp_min;
 	uint32_t temp_max;
+	char *sp = NULL;
 
 	char *copy_arg = strdup(arg);
 
@@ -242,7 +245,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 		return -1;
 
 	errno = 0;
-	token = strtok(copy_arg, ",");
+	token = strtok_r(copy_arg, ",", &sp);
 
 	/* Parse first value */
 	if (token != NULL) {
@@ -258,7 +261,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 	} else
 		goto err_list;
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 
 	while (token != NULL) {
 		if (count == MAX_LIST) {
@@ -280,7 +283,7 @@ parse_list(const char *arg, uint32_t *list, uint32_t *min, uint32_t *max)
 		if (number > temp_max)
 			temp_max = number;
 
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 
 	if (min)
diff --git a/app/test-crypto-perf/cperf_test_vector_parsing.c b/app/test-crypto-perf/cperf_test_vector_parsing.c
index 737d61d4af6b..b1cc1ef3c91d 100644
--- a/app/test-crypto-perf/cperf_test_vector_parsing.c
+++ b/app/test-crypto-perf/cperf_test_vector_parsing.c
@@ -9,6 +9,7 @@
 #include <stdlib.h>
 
 #include <rte_malloc.h>
+#include <rte_os_shim.h>
 
 #include "cperf_options.h"
 #include "cperf_test_vectors.h"
@@ -220,8 +221,9 @@ parse_values(char *tokens, uint8_t **data, uint32_t *data_length)
 
 	uint8_t *values, *values_resized;
 	char *tok, *error = NULL;
+	char *sp = NULL;
 
-	tok = strtok(tokens, CPERF_VALUE_DELIMITER);
+	tok = strtok_r(tokens, CPERF_VALUE_DELIMITER, &sp);
 	if (tok == NULL)
 		return -1;
 
@@ -252,7 +254,7 @@ parse_values(char *tokens, uint8_t **data, uint32_t *data_length)
 			return -1;
 		}
 
-		tok = strtok(NULL, CPERF_VALUE_DELIMITER);
+		tok = strtok_r(NULL, CPERF_VALUE_DELIMITER, &sp);
 		if (tok == NULL)
 			break;
 
@@ -283,6 +285,7 @@ parse_entry(char *entry, struct cperf_test_vector *vector,
 
 	uint8_t *data = NULL;
 	char *token, *key_token;
+	char *sp = NULL;
 
 	if (entry == NULL) {
 		printf("Expected entry value\n");
@@ -290,10 +293,10 @@ parse_entry(char *entry, struct cperf_test_vector *vector,
 	}
 
 	/* get key */
-	token = strtok(entry, CPERF_ENTRY_DELIMITER);
+	token = strtok_r(entry, CPERF_ENTRY_DELIMITER, &sp);
 	key_token = token;
 	/* get values for key */
-	token = strtok(NULL, CPERF_ENTRY_DELIMITER);
+	token = strtok_r(NULL, CPERF_ENTRY_DELIMITER, &sp);
 
 	if (key_token == NULL || token == NULL) {
 		printf("Expected 'key = values' but was '%.40s'..\n", entry);
-- 
2.22.0

[PATCH v5 05/25] app/dma-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 623dc9364dc6 ("app/dma-perf: introduce DMA performance test")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-dma-perf/main.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/app/test-dma-perf/main.c b/app/test-dma-perf/main.c
index 18219918ccab..25d8397ebe5c 100644
--- a/app/test-dma-perf/main.c
+++ b/app/test-dma-perf/main.c
@@ -18,6 +18,7 @@
 #include <rte_lcore.h>
 #include <rte_dmadev.h>
 #include <rte_kvargs.h>
+#include <rte_os_shim.h>
 
 #include "main.h"
 
@@ -183,6 +184,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 	uint16_t len;
 	char *input;
 	struct lcore_dma_map_t *lcore_dma_map;
+	char *sp = NULL;
 
 	if (test_case == NULL || value == NULL)
 		return -1;
@@ -194,7 +196,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 
 	memset(lcore_dma_map, 0, sizeof(struct lcore_dma_map_t));
 
-	char *token = strtok(input, ", ");
+	char *token = strtok_r(input, ", ", &sp);
 	while (token != NULL) {
 		if (lcore_dma_map->cnt >= MAX_WORKER_NB) {
 			free(input);
@@ -204,7 +206,7 @@ parse_lcore(struct test_configure *test_case, const char *value)
 		uint16_t lcore_id = atoi(token);
 		lcore_dma_map->lcores[lcore_dma_map->cnt++] = lcore_id;
 
-		token = strtok(NULL, ", ");
+		token = strtok_r(NULL, ", ", &sp);
 	}
 
 	free(input);
@@ -220,6 +222,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 	char *start, *end, *substr;
 	uint16_t lcore_id;
 	int ret = 0;
+	char *sp = NULL;
 
 	if (test_case == NULL || value == NULL)
 		return -1;
@@ -237,7 +240,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 		goto out;
 	}
 
-	substr = strtok(addrs, ",");
+	substr = strtok_r(addrs, ",", &sp);
 	if (substr == NULL) {
 		fprintf(stderr, "No input DMA address\n");
 		ret = -1;
@@ -279,7 +282,7 @@ parse_lcore_dma(struct test_configure *test_case, const char *value)
 		strlcpy(lcore_dma_map->dma_names[lcore_dma_map->cnt], ptrs[1],
 				RTE_DEV_NAME_MAX_LEN);
 		lcore_dma_map->cnt++;
-		substr = strtok(NULL, ",");
+		substr = strtok_r(NULL, ",", &sp);
 	} while (substr != NULL);
 
 out:
@@ -625,6 +628,7 @@ main(int argc, char *argv[])
 	char *rst_path_ptr = NULL;
 	char rst_path[PATH_MAX];
 	int new_argc;
+	char *sp = NULL;
 
 	memset(args, 0, sizeof(args));
 
@@ -643,7 +647,7 @@ main(int argc, char *argv[])
 	}
 	if (rst_path_ptr == NULL) {
 		strlcpy(rst_path, cfg_path_ptr, PATH_MAX);
-		char *token = strtok(basename(rst_path), ".");
+		char *token = strtok_r(basename(rst_path), ".", &sp);
 		if (token == NULL) {
 			printf("Config file error.\n");
 			return -1;
-- 
2.22.0

[PATCH v5 06/25] app/flow-perf: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 0c8f1f4ab90e ("app/flow-perf: support raw encap/decap actions")
Fixes: 7f37f0936a19 ("app/flow-perf: support meter policy API")
Fixes: 80a323319745 ("app/flow-perf: add destination ports parameter")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-flow-perf/main.c | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/app/test-flow-perf/main.c b/app/test-flow-perf/main.c
index 07ddfe0e46df..7e68146c16d7 100644
--- a/app/test-flow-perf/main.c
+++ b/app/test-flow-perf/main.c
@@ -36,6 +36,7 @@
 #include <rte_ethdev.h>
 #include <rte_flow.h>
 #include <rte_mtr.h>
+#include <rte_os_shim.h>
 
 #include "config.h"
 #include "actions_gen.h"
@@ -602,6 +603,7 @@ read_meter_policy(char *prog, char *arg)
 {
 	char *token;
 	size_t i, j, k;
+	char *sp = NULL;
 
 	j = 0;
 	k = 0;
@@ -612,9 +614,9 @@ read_meter_policy(char *prog, char *arg)
 		token = strsep(&arg, ":\0");
 	}
 	j = 0;
-	token = strtok(actions_str[0], ",\0");
+	token = strtok_r(actions_str[0], ",\0", &sp);
 	while (token == NULL && j < RTE_COLORS - 1)
-		token = strtok(actions_str[++j], ",\0");
+		token = strtok_r(actions_str[++j], ",\0", &sp);
 	while (j < RTE_COLORS && token != NULL) {
 		for (i = 0; i < RTE_DIM(flow_options); i++) {
 			if (!strcmp(token, flow_options[i].str)) {
@@ -628,9 +630,9 @@ read_meter_policy(char *prog, char *arg)
 			usage(prog);
 			rte_exit(EXIT_SUCCESS, "Invalid colored actions\n");
 		}
-		token = strtok(NULL, ",\0");
+		token = strtok_r(NULL, ",\0", &sp);
 		while (!token && j < RTE_COLORS - 1) {
-			token = strtok(actions_str[++j], ",\0");
+			token = strtok_r(actions_str[++j], ",\0", &sp);
 			k = 0;
 		}
 	}
@@ -641,6 +643,7 @@ args_parse(int argc, char **argv)
 {
 	uint64_t pm, seed;
 	uint64_t hp_conf;
+	char *sp = NULL;
 	char **argvopt;
 	uint32_t prio;
 	char *token;
@@ -804,7 +807,7 @@ args_parse(int argc, char **argv)
 						RTE_FLOW_ACTION_TYPE_RAW_ENCAP
 					);
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					for (i = 0; i < RTE_DIM(flow_options); i++) {
 						if (strcmp(flow_options[i].str, token) == 0) {
@@ -817,7 +820,7 @@ args_parse(int argc, char **argv)
 							rte_exit(EXIT_FAILURE,
 								"Invalid encap item: %s\n", token);
 					}
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 				printf(" / ");
 			}
@@ -828,7 +831,7 @@ args_parse(int argc, char **argv)
 						RTE_FLOW_ACTION_TYPE_RAW_DECAP
 					);
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					for (i = 0; i < RTE_DIM(flow_options); i++) {
 						if (strcmp(flow_options[i].str, token) == 0) {
@@ -841,7 +844,7 @@ args_parse(int argc, char **argv)
 							rte_exit(EXIT_FAILURE,
 								"Invalid decap item %s\n", token);
 					}
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 				printf(" / ");
 			}
@@ -910,10 +913,10 @@ args_parse(int argc, char **argv)
 				uint16_t port_idx = 0;
 				char *token;
 
-				token = strtok(optarg, ",");
+				token = strtok_r(optarg, ",", &sp);
 				while (token != NULL) {
 					dst_ports[port_idx++] = atoi(token);
-					token = strtok(NULL, ",");
+					token = strtok_r(NULL, ",", &sp);
 				}
 			}
 			if (strcmp(lgopts[opt_idx].name, "rxq") == 0) {
-- 
2.22.0

[PATCH v5 07/25] app/test-mldev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: bbd272edcb14 ("app/mldev: add ordered inferences")
Fixes: 28a4a819c850 ("app/mldev: improve checks for invalid options")
Fixes: da6793390596 ("app/mldev: support inference validation")
Fixes: f6661e6d9a3a ("app/mldev: validate model operations")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-mldev/ml_options.c | 19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)

diff --git a/app/test-mldev/ml_options.c b/app/test-mldev/ml_options.c
index 320f6325ae67..1033444de0e1 100644
--- a/app/test-mldev/ml_options.c
+++ b/app/test-mldev/ml_options.c
@@ -9,6 +9,7 @@
 #include <rte_memory.h>
 #include <rte_mldev.h>
 #include <rte_string_fns.h>
+#include <rte_os_shim.h>
 
 #include "ml_common.h"
 #include "ml_test.h"
@@ -76,12 +77,12 @@ ml_parse_models(struct ml_options *opt, const char *arg)
 {
 	const char *delim = ",";
 	char models[PATH_MAX];
-	char *token;
+	char *token, *sp = NULL;
 	int ret = 0;
 
 	strlcpy(models, arg, PATH_MAX);
 
-	token = strtok(models, delim);
+	token = strtok_r(models, delim, &sp);
 	while (token != NULL) {
 		if (opt->nb_filelist >= ML_TEST_MAX_MODELS) {
 			ml_err("Exceeded model count, max = %d\n", ML_TEST_MAX_MODELS);
@@ -92,7 +93,7 @@ ml_parse_models(struct ml_options *opt, const char *arg)
 		strlcpy(opt->filelist[opt->nb_filelist].model, token, PATH_MAX);
 		opt->nb_filelist++;
 
-		token = strtok(NULL, delim);
+		token = strtok_r(NULL, delim, &sp);
 	}
 
 	if (opt->nb_filelist == 0) {
@@ -108,7 +109,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 {
 	const char *delim = ",";
 	char filelist[PATH_MAX];
-	char *token;
+	char *token, *sp = NULL;
 
 	if (opt->nb_filelist >= ML_TEST_MAX_MODELS) {
 		ml_err("Exceeded filelist count, max = %d\n", ML_TEST_MAX_MODELS);
@@ -118,7 +119,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(filelist, arg, PATH_MAX);
 
 	/* model */
-	token = strtok(filelist, delim);
+	token = strtok_r(filelist, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, model not specified = %s\n", arg);
 		return -EINVAL;
@@ -126,7 +127,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].model, token, PATH_MAX);
 
 	/* input */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, input not specified = %s\n", arg);
 		return -EINVAL;
@@ -134,7 +135,7 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].input, token, PATH_MAX);
 
 	/* output */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token == NULL) {
 		ml_err("Invalid filelist, output not specified = %s\n", arg);
 		return -EINVAL;
@@ -142,14 +143,14 @@ ml_parse_filelist(struct ml_options *opt, const char *arg)
 	strlcpy(opt->filelist[opt->nb_filelist].output, token, PATH_MAX);
 
 	/* reference - optional */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token != NULL)
 		strlcpy(opt->filelist[opt->nb_filelist].reference, token, PATH_MAX);
 	else
 		memset(opt->filelist[opt->nb_filelist].reference, 0, PATH_MAX);
 
 	/* check for extra tokens */
-	token = strtok(NULL, delim);
+	token = strtok_r(NULL, delim, &sp);
 	if (token != NULL) {
 		ml_err("Invalid filelist. Entries > 4\n.");
 		return -EINVAL;
-- 
2.22.0

[PATCH v5 08/25] app/test-fib: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 103809d032cd ("app/test-fib: add test application for FIB")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 app/test-fib/main.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/app/test-fib/main.c b/app/test-fib/main.c
index 6479f48cdf6c..11810a1a80a8 100644
--- a/app/test-fib/main.c
+++ b/app/test-fib/main.c
@@ -17,6 +17,7 @@
 #include <rte_lpm6.h>
 #include <rte_fib.h>
 #include <rte_fib6.h>
+#include <rte_os_shim.h>
 
 #define	PRINT_USAGE_START	"%s [EAL options] --\n"
 
@@ -204,9 +205,9 @@ parse_distrib(uint8_t depth_lim, const uint32_t n)
 	uint32_t nrpd[128 + 1] = {0}; /* number of routes per depth */
 	uint32_t n_routes;
 	uint8_t depth, ratio, ratio_acc = 0;
-	char *in;
+	char *in, *sp = NULL;
 
-	in = strtok(distrib_string, ",");
+	in = strtok_r(distrib_string, ",", &sp);
 
 	/*parse configures routes percentage ratios*/
 	while (in != NULL) {
@@ -246,7 +247,7 @@ parse_distrib(uint8_t depth_lim, const uint32_t n)
 		}
 
 		/*number of configured depths in*/
-		in = strtok(NULL, ",");
+		in = strtok_r(NULL, ",", &sp);
 	}
 
 	if (ratio_acc > 100) {
@@ -522,10 +523,10 @@ parse_lookup(FILE *f, int af)
 	int ret, i = 0;
 	uint8_t *tbl = (uint8_t *)config.lookup_tbl;
 	int step = (af == AF_INET) ? 4 : 16;
-	char *s;
+	char *s, *sp = NULL;
 
 	while (fgets(line, sizeof(line), f) != NULL) {
-		s = strtok(line, " \t\n");
+		s = strtok_r(line, " \t\n", &sp);
 		if (s == NULL)
 			return -EINVAL;
 		ret = inet_pton(af, s, &tbl[i]);
-- 
2.22.0

[PATCH v5 09/25] dmadev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 39b5ab60df30 ("dmadev: add telemetry")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/dmadev/rte_dmadev.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/lib/dmadev/rte_dmadev.c b/lib/dmadev/rte_dmadev.c
index 8bb7824aa129..eaf649e7d982 100644
--- a/lib/dmadev/rte_dmadev.c
+++ b/lib/dmadev/rte_dmadev.c
@@ -14,6 +14,7 @@
 #include <rte_memzone.h>
 #include <rte_string_fns.h>
 #include <rte_telemetry.h>
+#include <rte_os_shim.h>
 
 #include "rte_dmadev.h"
 #include "rte_dmadev_pmd.h"
@@ -1016,7 +1017,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	struct rte_dma_info dma_info;
 	struct rte_dma_stats dma_stats;
 	int dev_id, ret, vchan_id;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *vchan_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1035,7 +1036,7 @@ dmadev_handle_dev_stats(const char *cmd __rte_unused,
 	if (dma_info.nb_vchans == 1 && *end_param == '\0')
 		vchan_id = 0;
 	else {
-		vchan_param = strtok(end_param, ",");
+		vchan_param = strtok_r(end_param, ",", &sp);
 		if (!vchan_param || strlen(vchan_param) == 0 || !isdigit(*vchan_param))
 			return -EINVAL;
 
-- 
2.22.0

[PATCH v5 10/25] eal: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 2054f31a1fcd ("mem: add memseg info in telemetry")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Amit Prakash Shukla <amitprakashs@marvell.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/eal/common/eal_common_memory.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/eal/common/eal_common_memory.c b/lib/eal/common/eal_common_memory.c
index a185e0b580c7..ace63313ea75 100644
--- a/lib/eal/common/eal_common_memory.c
+++ b/lib/eal/common/eal_common_memory.c
@@ -1275,22 +1275,22 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	char dlim[2] = ",";
 	char *params_args;
 	size_t count = 0;
-	char *token;
+	char *token, *sp = NULL;
 
 	if (vals == NULL || params == NULL || strlen(params) == 0)
 		return -1;
 
-	/* strtok expects char * and param is const char *. Hence on using
+	/* strtok_r expects char * and param is const char *. Hence on using
 	 * params as "const char *" compiler throws warning.
 	 */
 	params_args = strdup(params);
 	if (params_args == NULL)
 		return -1;
 
-	token = strtok(params_args, dlim);
+	token = strtok_r(params_args, dlim, &sp);
 	while (token && isdigit(*token) && count < n_vals) {
 		vals[count++] = strtoul(token, NULL, 10);
-		token = strtok(NULL, dlim);
+		token = strtok_r(NULL, dlim, &sp);
 	}
 
 	free(params_args);
-- 
2.22.0

[PATCH v5 11/25] ethdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: f38f62650f7b ("ethdev: add Rx queue telemetry query")
Fixes: 9e7533aeb80a ("ethdev: add telemetry command for TM level capabilities")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/ethdev/rte_ethdev_telemetry.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/lib/ethdev/rte_ethdev_telemetry.c b/lib/ethdev/rte_ethdev_telemetry.c
index 5e6c4172d3be..e589032dd368 100644
--- a/lib/ethdev/rte_ethdev_telemetry.c
+++ b/lib/ethdev/rte_ethdev_telemetry.c
@@ -7,6 +7,7 @@
 
 #include <rte_kvargs.h>
 #include <rte_telemetry.h>
+#include <rte_os_shim.h>
 
 #include "rte_ethdev.h"
 #include "ethdev_driver.h"
@@ -477,6 +478,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	const char *qid_param;
 	uint16_t nb_queues;
 	char *end_param;
+	char *sp = NULL;
 	uint64_t qid;
 	int ret;
 
@@ -489,7 +491,7 @@ ethdev_parse_queue_params(const char *params, bool is_rx,
 	if (nb_queues == 1 && *end_param == '\0')
 		qid = 0;
 	else {
-		qid_param = strtok(end_param, ",");
+		qid_param = strtok_r(end_param, ",", &sp);
 		if (!qid_param || strlen(qid_param) == 0 || !isdigit(*qid_param))
 			return -EINVAL;
 
@@ -1221,9 +1223,10 @@ static int
 eth_dev_parse_tm_params(char *params, uint32_t *result)
 {
 	const char *splited_param;
+	char *sp = NULL;
 	uint64_t ret;
 
-	splited_param = strtok(params, ",");
+	splited_param = strtok_r(params, ",", &sp);
 	if (!splited_param || strlen(splited_param) == 0 || !isdigit(*splited_param))
 		return -EINVAL;
 
@@ -1510,13 +1513,14 @@ eth_dev_handle_port_regs(const char *cmd __rte_unused,
 {
 	char *filter, *end_param;
 	uint16_t port_id;
+	char *sp = NULL;
 	int ret;
 
 	ret = eth_dev_parse_port_params(params, &port_id, &end_param, true);
 	if (ret != 0)
 		return ret;
 
-	filter = strtok(end_param, ",");
+	filter = strtok_r(end_param, ",", &sp);
 	if (filter != NULL && strlen(filter) == 0)
 		filter = NULL;
 
-- 
2.22.0

[PATCH v5 12/25] eventdev: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/eventdev/rte_event_eth_rx_adapter.c | 39 +++++++++++++------------
 lib/eventdev/rte_eventdev.c             | 18 ++++++------
 2 files changed, 29 insertions(+), 28 deletions(-)

diff --git a/lib/eventdev/rte_event_eth_rx_adapter.c b/lib/eventdev/rte_event_eth_rx_adapter.c
index 3ee20d95f372..96af9c0c8b65 100644
--- a/lib/eventdev/rte_event_eth_rx_adapter.c
+++ b/lib/eventdev/rte_event_eth_rx_adapter.c
@@ -23,6 +23,7 @@
 #include <rte_interrupts.h>
 #include <rte_mbuf_dyn.h>
 #include <rte_telemetry.h>
+#include <rte_os_shim.h>
 
 #include "rte_eventdev.h"
 #include "eventdev_pmd.h"
@@ -3651,7 +3652,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp;
 	struct rte_event_eth_rx_adapter_queue_conf queue_conf;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3661,19 +3662,19 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_PORTID_VALID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3684,7 +3685,7 @@ handle_rxa_get_queue_conf(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3723,7 +3724,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 	struct rte_event_eth_rx_adapter_queue_stats q_stats;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -3733,19 +3734,19 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_PORTID_VALID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3756,7 +3757,7 @@ handle_rxa_get_queue_stats(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3794,7 +3795,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	uint8_t rx_adapter_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3803,19 +3804,19 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 	rx_adapter_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_ID_VALID_OR_GOTO_ERR_RET(rx_adapter_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_PORTID_VALID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3826,7 +3827,7 @@ handle_rxa_queue_stats_reset(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
@@ -3855,7 +3856,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	uint8_t instance_id;
 	uint16_t rx_queue_id;
 	int eth_dev_id, ret = -1;
-	char *token, *l_params;
+	char *token, *l_params, *sp = NULL;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
 		return -1;
@@ -3863,14 +3864,14 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 	l_params = strdup(params);
 	if (l_params == NULL)
 		return -ENOMEM;
-	token = strtok(l_params, ",");
+	token = strtok_r(l_params, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get device ID from parameter string */
 	eth_dev_id = strtoul(token, NULL, 10);
 	RTE_EVENT_ETH_RX_ADAPTER_PORTID_VALID_OR_GOTO_ERR_RET(eth_dev_id, -EINVAL);
 
-	token = strtok(NULL, ",");
+	token = strtok_r(NULL, ",", &sp);
 	RTE_EVENT_ETH_RX_ADAPTER_TOKEN_VALID_OR_GOTO_ERR_RET(token, -1);
 
 	/* Get Rx queue ID from parameter string */
@@ -3881,7 +3882,7 @@ handle_rxa_instance_get(const char *cmd __rte_unused,
 		goto error;
 	}
 
-	token = strtok(NULL, "\0");
+	token = strtok_r(NULL, "\0", &sp);
 	if (token != NULL)
 		RTE_EDEV_LOG_ERR("Extra parameters passed to eventdev"
 				 " telemetry command, ignoring");
diff --git a/lib/eventdev/rte_eventdev.c b/lib/eventdev/rte_eventdev.c
index ca295c87c47d..d3dacffdfab2 100644
--- a/lib/eventdev/rte_eventdev.c
+++ b/lib/eventdev/rte_eventdev.c
@@ -1787,7 +1787,7 @@ handle_queue_links(const char *cmd __rte_unused,
 		   struct rte_tel_data *d)
 {
 	int i, ret, port_id = 0;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	uint8_t dev_id;
 	uint8_t queues[RTE_EVENT_MAX_QUEUES_PER_DEV];
 	uint8_t priorities[RTE_EVENT_MAX_QUEUES_PER_DEV];
@@ -1800,12 +1800,12 @@ handle_queue_links(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
 		return -1;
 
 	port_id = strtoul(p_param, &end_param, 10);
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1925,7 +1925,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1935,7 +1935,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_PORT;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1943,7 +1943,7 @@ handle_port_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
@@ -1959,7 +1959,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	int dev_id;
 	int port_queue_id = 0;
 	enum rte_event_dev_xstats_mode mode;
-	char *end_param;
+	char *end_param, *sp = NULL;
 	const char *p_param;
 
 	if (params == NULL || strlen(params) == 0 || !isdigit(*params))
@@ -1969,7 +1969,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 	dev_id = strtoul(params, &end_param, 10);
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 
-	p_param = strtok(end_param, ",");
+	p_param = strtok_r(end_param, ",", &sp);
 	mode = RTE_EVENT_DEV_XSTATS_QUEUE;
 
 	if (p_param == NULL || strlen(p_param) == 0 || !isdigit(*p_param))
@@ -1977,7 +1977,7 @@ handle_queue_xstats(const char *cmd __rte_unused,
 
 	port_queue_id = strtoul(p_param, &end_param, 10);
 
-	p_param = strtok(NULL, "\0");
+	p_param = strtok_r(NULL, "\0", &sp);
 	if (p_param != NULL)
 		RTE_EDEV_LOG_DEBUG(
 			"Extra parameters passed to eventdev telemetry command, ignoring");
-- 
2.22.0

[PATCH v5 13/25] security: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 259ca6d1617f ("security: add telemetry endpoint for capabilities")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/security/rte_security.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lib/security/rte_security.c b/lib/security/rte_security.c
index e5c862f5f570..669888ef0807 100644
--- a/lib/security/rte_security.c
+++ b/lib/security/rte_security.c
@@ -11,6 +11,7 @@
 #include <rte_cryptodev.h>
 #include <dev_driver.h>
 #include <rte_telemetry.h>
+#include <rte_os_shim.h>
 #include "rte_security.h"
 #include "rte_security_driver.h"
 
@@ -497,13 +498,14 @@ security_handle_cryptodev_crypto_caps(const char *cmd __rte_unused, const char *
 	int dev_id, capa_id;
 	int crypto_caps_n;
 	char *end_param;
+	char *sp = NULL;
 	int rc;
 
 	if (!params || strlen(params) == 0 || !isdigit(*params))
 		return -EINVAL;
 
 	dev_id = strtoul(params, &end_param, 0);
-	capa_param = strtok(end_param, ",");
+	capa_param = strtok_r(end_param, ",", &sp);
 	if (!capa_param || strlen(capa_param) == 0 || !isdigit(*capa_param))
 		return -EINVAL;
 
-- 
2.22.0

[PATCH v5 14/25] telemetry: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 6dd571fd07c3 ("telemetry: introduce new functionality")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Ciara Power <ciara.power@intel.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
---
 lib/telemetry/telemetry.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/lib/telemetry/telemetry.c b/lib/telemetry/telemetry.c
index 31a2c91c0657..cd1eddc5609c 100644
--- a/lib/telemetry/telemetry.c
+++ b/lib/telemetry/telemetry.c
@@ -19,6 +19,7 @@
 #include <rte_common.h>
 #include <rte_spinlock.h>
 #include <rte_log.h>
+#include <rte_os_shim.h>
 
 #include "rte_telemetry.h"
 #include "telemetry_json.h"
@@ -398,6 +399,7 @@ static void *
 client_handler(void *sock_id)
 {
 	int s = (int)(uintptr_t)sock_id;
+	char *sp = NULL;
 	char buffer[1024];
 	char info_str[1024];
 	snprintf(info_str, sizeof(info_str),
@@ -412,8 +414,8 @@ client_handler(void *sock_id)
 	int bytes = read(s, buffer, sizeof(buffer) - 1);
 	while (bytes > 0) {
 		buffer[bytes] = 0;
-		const char *cmd = strtok(buffer, ",");
-		const char *param = strtok(NULL, "\0");
+		const char *cmd = strtok_r(buffer, ",", &sp);
+		const char *param = strtok_r(NULL, "\0", &sp);
 		struct cmd_callback cb = {.fn = unknown_command};
 		int i;
 
-- 
2.22.0

[PATCH v5 15/25] bus/fslmc: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 9ccb76b24c1d ("bus/fslmc: enable portal interrupt handling")
Fixes: 828d51d8fc3e ("bus/fslmc: refactor scan and probe functions")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Sachin Saxena <sachin.saxena@nxp.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/bus/fslmc/fslmc_bus.c            | 6 ++++--
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 5 +++--
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index 097d6dca08b5..fe7eb26940cb 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -16,6 +16,7 @@
 #include <rte_memcpy.h>
 #include <ethdev_driver.h>
 #include <rte_mbuf_dyn.h>
+#include <rte_os_shim.h>
 
 #include "private.h"
 #include <fslmc_vfio.h>
@@ -132,6 +133,7 @@ scan_one_fslmc_device(char *dev_name)
 {
 	char *dup_dev_name, *t_ptr;
 	struct rte_dpaa2_device *dev = NULL;
+	char *sp = NULL;
 	int ret = -1;
 
 	if (!dev_name)
@@ -169,7 +171,7 @@ scan_one_fslmc_device(char *dev_name)
 	}
 
 	/* Parse the device name and ID */
-	t_ptr = strtok(dup_dev_name, ".");
+	t_ptr = strtok_r(dup_dev_name, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Invalid device found: (%s)", dup_dev_name);
 		ret = -EINVAL;
@@ -200,7 +202,7 @@ scan_one_fslmc_device(char *dev_name)
 	else
 		dev->dev_type = DPAA2_UNKNOWN;
 
-	t_ptr = strtok(NULL, ".");
+	t_ptr = strtok_r(NULL, ".", &sp);
 	if (!t_ptr) {
 		DPAA2_BUS_ERR("Skipping invalid device (%s)", dup_dev_name);
 		ret = 0;
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index d8a98326d9de..8b863af010da 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -31,6 +31,7 @@
 #include <rte_cycles.h>
 #include <rte_kvargs.h>
 #include <dev_driver.h>
+#include <rte_os_shim.h>
 
 #include <fslmc_logs.h>
 #include <bus_fslmc_driver.h>
@@ -128,7 +129,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 #define AFFINITY_LEN	128
 	uint32_t cpu_mask = 1;
 	size_t len = 0;
-	char *temp = NULL, *token = NULL;
+	char *temp = NULL, *token = NULL, *sp = NULL;
 	char string[STRING_LEN];
 	char smp_affinity[AFFINITY_LEN];
 	FILE *file;
@@ -141,7 +142,7 @@ dpaa2_affine_dpio_intr_to_respective_core(int32_t dpio_id, int cpu_id)
 	}
 	while (getline(&temp, &len, file) != -1) {
 		if ((strstr(temp, string)) != NULL) {
-			token = strtok(temp, ":");
+			token = strtok_r(temp, ":", &sp);
 			break;
 		}
 	}
-- 
2.22.0

[PATCH v5 16/25] common/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: af75aac78978 ("common/cnxk: support telemetry for NIX")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
---
 drivers/common/cnxk/cnxk_telemetry_nix.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/common/cnxk/cnxk_telemetry_nix.c b/drivers/common/cnxk/cnxk_telemetry_nix.c
index abeefafe1e19..a0b587c97c63 100644
--- a/drivers/common/cnxk/cnxk_telemetry_nix.c
+++ b/drivers/common/cnxk/cnxk_telemetry_nix.c
@@ -3,6 +3,7 @@
  */
 
 #include <ctype.h>
+#include <rte_os_shim.h>
 #include "cnxk_telemetry.h"
 #include "roc_api.h"
 #include "roc_priv.h"
@@ -1015,7 +1016,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			   struct plt_tel_data *d)
 {
 	struct nix_tel_node *node;
-	char *name, *param;
+	char *name, *param, *sp = NULL;
 	char buf[1024];
 	int rc = -1;
 
@@ -1023,11 +1024,11 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 		goto exit;
 
 	plt_strlcpy(buf, params, PCI_PRI_STR_SIZE + 1);
-	name = strtok(buf, ",");
+	name = strtok_r(buf, ",", &sp);
 	if (name == NULL)
 		goto exit;
 
-	param = strtok(NULL, "\0");
+	param = strtok_r(NULL, "\0", &sp);
 
 	node = nix_tel_node_get_by_pcidev_name(name);
 	if (!node)
@@ -1036,7 +1037,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 	plt_tel_data_start_dict(d);
 
 	if (strstr(cmd, "rq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int rq;
 
 		if (!tok)
@@ -1052,7 +1053,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_rq(node->rqs[rq], d);
 
 	} else if (strstr(cmd, "cq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int cq;
 
 		if (!tok)
@@ -1068,7 +1069,7 @@ cnxk_nix_tel_handle_info_x(const char *cmd, const char *params,
 			rc = cnxk_tel_nix_cq(node->cqs[cq], d);
 
 	} else if (strstr(cmd, "sq")) {
-		char *tok = strtok(param, ",");
+		char *tok = strtok_r(param, ",", &sp);
 		int sq;
 
 		if (!tok)
-- 
2.22.0

[PATCH v5 17/25] event/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 8a3d58c189fd ("event/cnxk: add option to control timer adapters")
Fixes: 8bdbae66b299 ("event/cnxk: add external clock support for timer")

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/event/cnxk/cnxk_eventdev.c  | 11 +++++++----
 drivers/event/cnxk/cnxk_tim_evdev.c | 12 +++++++-----
 2 files changed, 14 insertions(+), 9 deletions(-)

diff --git a/drivers/event/cnxk/cnxk_eventdev.c b/drivers/event/cnxk/cnxk_eventdev.c
index 84a55511a328..facbe7fa0928 100644
--- a/drivers/event/cnxk/cnxk_eventdev.c
+++ b/drivers/event/cnxk/cnxk_eventdev.c
@@ -2,6 +2,7 @@
  * Copyright(C) 2021 Marvell.
  */
 
+#include <rte_os_shim.h>
 #include "roc_npa.h"
 
 #include "cnxk_eventdev.h"
@@ -482,7 +483,8 @@ parse_queue_param(char *value, void *opaque)
 	struct cnxk_sso_qos queue_qos = {0};
 	uint16_t *val = (uint16_t *)&queue_qos;
 	struct cnxk_sso_evdev *dev = opaque;
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "-", &sp);
 	struct cnxk_sso_qos *old_ptr;
 
 	if (!strlen(value))
@@ -490,7 +492,7 @@ parse_queue_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		val++;
 	}
 
@@ -518,7 +520,8 @@ parse_stash_param(char *value, void *opaque)
 	struct cnxk_sso_stash queue_stash = {0};
 	struct cnxk_sso_evdev *dev = opaque;
 	struct cnxk_sso_stash *old_ptr;
-	char *tok = strtok(value, "|");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "|", &sp);
 	uint16_t *val;
 
 	if (!strlen(value))
@@ -527,7 +530,7 @@ parse_stash_param(char *value, void *opaque)
 	val = (uint16_t *)&queue_stash;
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "|");
+		tok = strtok_r(NULL, "|", &sp);
 		val++;
 	}
 
diff --git a/drivers/event/cnxk/cnxk_tim_evdev.c b/drivers/event/cnxk/cnxk_tim_evdev.c
index bba70646fa16..82d4a1878dd5 100644
--- a/drivers/event/cnxk/cnxk_tim_evdev.c
+++ b/drivers/event/cnxk/cnxk_tim_evdev.c
@@ -3,6 +3,7 @@
  */
 
 #include <math.h>
+#include <rte_os_shim.h>
 
 #include "roc_npa.h"
 
@@ -420,7 +421,8 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 {
 	struct cnxk_tim_evdev *dev = opaque;
 	struct cnxk_tim_ctl ring_ctl = {0};
-	char *tok = strtok(value, "-");
+	char *sp = NULL;
+	char *tok = strtok_r(value, "-", &sp);
 	struct cnxk_tim_ctl *old_ptr;
 	uint16_t *val;
 
@@ -431,7 +433,7 @@ cnxk_tim_parse_ring_param(char *value, void *opaque)
 
 	while (tok != NULL) {
 		*val = atoi(tok);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		val++;
 	}
 
@@ -507,16 +509,16 @@ cnxk_tim_parse_clk_list(const char *value, void *opaque)
 				      ROC_TIM_CLK_SRC_INVALID};
 	struct cnxk_tim_evdev *dev = opaque;
 	char *str = strdup(value);
-	char *tok;
+	char *tok, *sp = NULL;
 	int i = 0;
 
 	if (str == NULL || !strlen(str))
 		goto free;
 
-	tok = strtok(str, "-");
+	tok = strtok_r(str, "-", &sp);
 	while (tok != NULL && src[i] != ROC_TIM_CLK_SRC_INVALID) {
 		dev->ext_clk_freq[src[i]] = strtoull(tok, NULL, 10);
-		tok = strtok(NULL, "-");
+		tok = strtok_r(NULL, "-", &sp);
 		i++;
 	}
 
-- 
2.22.0

[PATCH v5 18/25] net/ark: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 9c7188a68d7b ("net/ark: provide API for hardware modules pktchkr and pktgen")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/net/ark/ark_pktchkr.c | 11 ++++++-----
 drivers/net/ark/ark_pktgen.c  | 11 ++++++-----
 2 files changed, 12 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ark/ark_pktchkr.c b/drivers/net/ark/ark_pktchkr.c
index e1f336c73c2a..2bb3dd7b5f36 100644
--- a/drivers/net/ark/ark_pktchkr.c
+++ b/drivers/net/ark/ark_pktchkr.c
@@ -7,6 +7,7 @@
 
 #include <rte_string_fns.h>
 #include <rte_malloc.h>
+#include <rte_os_shim.h>
 
 #include "ark_pktchkr.h"
 #include "ark_logs.h"
@@ -359,14 +360,14 @@ set_arg(char *arg, char *val)
 void
 ark_pktchkr_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = "=\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_r(args, toks, &sp);
+	v = strtok_r(NULL, toks, &sp);
 	while (argv && v) {
 		set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_r(NULL, toks, &sp);
+		v = strtok_r(NULL, toks, &sp);
 	}
 }
 
diff --git a/drivers/net/ark/ark_pktgen.c b/drivers/net/ark/ark_pktgen.c
index 69ff7072b2ab..4765b8f0992a 100644
--- a/drivers/net/ark/ark_pktgen.c
+++ b/drivers/net/ark/ark_pktgen.c
@@ -8,6 +8,7 @@
 #include <rte_string_fns.h>
 #include <rte_malloc.h>
 #include <rte_thread.h>
+#include <rte_os_shim.h>
 
 #include "ark_pktgen.h"
 #include "ark_logs.h"
@@ -340,14 +341,14 @@ pmd_set_arg(char *arg, char *val)
 void
 ark_pktgen_parse(char *args)
 {
-	char *argv, *v;
+	char *argv, *v, *sp = NULL;
 	const char toks[] = " =\n\t\v\f \r";
-	argv = strtok(args, toks);
-	v = strtok(NULL, toks);
+	argv = strtok_r(args, toks, &sp);
+	v = strtok_r(NULL, toks, &sp);
 	while (argv && v) {
 		pmd_set_arg(argv, v);
-		argv = strtok(NULL, toks);
-		v = strtok(NULL, toks);
+		argv = strtok_r(NULL, toks, &sp);
+		v = strtok_r(NULL, toks, &sp);
 	}
 }
 
-- 
2.22.0

[PATCH v5 19/25] raw/cnxk_gpio: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: ecc0dd455e9a ("raw/cnxk_gpio: add option to select subset of GPIOs")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 drivers/raw/cnxk_gpio/cnxk_gpio.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/raw/cnxk_gpio/cnxk_gpio.c b/drivers/raw/cnxk_gpio/cnxk_gpio.c
index 329ac28a2736..e6408db824de 100644
--- a/drivers/raw/cnxk_gpio/cnxk_gpio.c
+++ b/drivers/raw/cnxk_gpio/cnxk_gpio.c
@@ -11,6 +11,7 @@
 #include <rte_kvargs.h>
 #include <rte_lcore.h>
 #include <rte_rawdev_pmd.h>
+#include <rte_os_shim.h>
 
 #include <roc_api.h>
 
@@ -192,7 +193,7 @@ static int
 cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 {
 	int i, ret, val, queue = 0;
-	char *token;
+	char *token, *sp = NULL;
 	int *list;
 
 	list = rte_calloc(NULL, gpiochip->num_gpios, sizeof(*list), 0);
@@ -210,7 +211,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 	allowlist[strlen(allowlist) - 1] = ' ';
 
 	/* quiesce -Wcast-qual */
-	token = strtok((char *)(uintptr_t)allowlist, ",");
+	token = strtok_r((char *)(uintptr_t)allowlist, ",", &sp);
 	do {
 		errno = 0;
 		val = strtol(token, NULL, 10);
@@ -236,7 +237,7 @@ cnxk_gpio_parse_allowlist(struct cnxk_gpiochip *gpiochip, char *allowlist)
 		}
 		if (i == queue)
 			list[queue++] = val;
-	} while ((token = strtok(NULL, ",")));
+	} while ((token = strtok_r(NULL, ",", &sp)));
 
 	free(allowlist);
 	gpiochip->allowlist = list;
-- 
2.22.0

[PATCH v5 20/25] net/cnxk: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: c8f91985331c ("raw/cnxk_gpio: replace strtok with reentrant version")
Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
---
 drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c b/drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c
index 86c2453c0983..0b766be11a17 100644
--- a/drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c
+++ b/drivers/net/cnxk/cnxk_ethdev_sec_telemetry.c
@@ -5,6 +5,7 @@
 #include <ctype.h>
 
 #include <rte_telemetry.h>
+#include <rte_os_shim.h>
 
 #include <roc_api.h>
 
@@ -214,6 +215,7 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	char dlim[2] = ",";
 	char *params_args;
 	size_t count = 0;
+	char *sp = NULL;
 	char *token;
 
 	if (vals == NULL || params == NULL || strlen(params) == 0)
@@ -226,10 +228,10 @@ parse_params(const char *params, uint32_t *vals, size_t n_vals)
 	if (params_args == NULL)
 		return -1;
 
-	token = strtok(params_args, dlim);
+	token = strtok_r(params_args, dlim, &sp);
 	while (token && isdigit(*token) && count < n_vals) {
 		vals[count++] = strtoul(token, NULL, 10);
-		token = strtok(NULL, dlim);
+		token = strtok_r(NULL, dlim, &sp);
 	}
 
 	free(params_args);
-- 
2.22.0

[PATCH v5 21/25] common/qat: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 99ab2806687b ("common/qat: isolate parser arguments configuration")

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 drivers/common/qat/qat_device.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/common/qat/qat_device.c b/drivers/common/qat/qat_device.c
index bca88fd9bded..99153775d883 100644
--- a/drivers/common/qat/qat_device.c
+++ b/drivers/common/qat/qat_device.c
@@ -4,6 +4,7 @@
 
 #include <rte_string_fns.h>
 #include <rte_devargs.h>
+#include <rte_os_shim.h>
 #include <ctype.h>
 
 #include "qat_device.h"
@@ -222,6 +223,7 @@ qat_dev_parse_command_line(struct qat_pci_device *qat_dev,
 {
 	int len = 0;
 	char *token = NULL;
+	char *sp = NULL;
 
 	if (!devargs)
 		return 0;
@@ -236,14 +238,14 @@ qat_dev_parse_command_line(struct qat_pci_device *qat_dev,
 		return -1;
 	}
 	strcpy(qat_dev->command_line, devargs->drv_str);
-	token = strtok(qat_dev->command_line, ",");
+	token = strtok_r(qat_dev->command_line, ",", &sp);
 	while (token != NULL) {
 		if (!cmdline_validate(token)) {
 			QAT_LOG(ERR, "Incorrect command line argument: %s",
 				token);
 			return -1;
 		}
-		token = strtok(NULL, ",");
+		token = strtok_r(NULL, ",", &sp);
 	}
 	/* Copy once againe the entire string, strtok already altered the contents */
 	strcpy(qat_dev->command_line, devargs->drv_str);
-- 
2.22.0

[PATCH v5 22/25] net/mlx5: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 0683c002f7f5 ("net/mlx5: add testpmd commands for GENEVE TLV parser")

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 drivers/net/mlx5/mlx5_testpmd.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_testpmd.c b/drivers/net/mlx5/mlx5_testpmd.c
index 1bb5a89559fe..7bd220fafa46 100644
--- a/drivers/net/mlx5/mlx5_testpmd.c
+++ b/drivers/net/mlx5/mlx5_testpmd.c
@@ -353,6 +353,7 @@ mlx5_test_parse_geneve_option_data(const char *buff, uint8_t data_len,
 				   rte_be32_t **match_data_mask)
 {
 	rte_be32_t *data;
+	char *sp = NULL;
 	char *buff2;
 	char *token;
 	uint8_t i = 0;
@@ -377,7 +378,7 @@ mlx5_test_parse_geneve_option_data(const char *buff, uint8_t data_len,
 		return -ENOMEM;
 	}
 
-	token = strtok(buff2, SPACE_DELIMITER);
+	token = strtok_r(buff2, SPACE_DELIMITER, &sp);
 	while (token != NULL) {
 		if (i == data_len) {
 			TESTPMD_LOG(ERR,
@@ -393,7 +394,7 @@ mlx5_test_parse_geneve_option_data(const char *buff, uint8_t data_len,
 		else
 			data[i] = 0x0;
 
-		token = strtok(NULL, SPACE_DELIMITER);
+		token = strtok_r(NULL, SPACE_DELIMITER, &sp);
 		i++;
 	}
 
-- 
2.22.0

[PATCH v5 23/25] examples/l2fwd-crypto: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 1df9c0109f4c ("examples/l2fwd-crypto: parse key parameters")

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 examples/l2fwd-crypto/main.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/examples/l2fwd-crypto/main.c b/examples/l2fwd-crypto/main.c
index a441312f5524..7128bd2e72cf 100644
--- a/examples/l2fwd-crypto/main.c
+++ b/examples/l2fwd-crypto/main.c
@@ -43,6 +43,7 @@
 #include <rte_prefetch.h>
 #include <rte_random.h>
 #include <rte_hexdump.h>
+#include <rte_os_shim.h>
 #ifdef RTE_CRYPTO_SCHEDULER
 #include <rte_cryptodev_scheduler.h>
 #endif
@@ -1105,12 +1106,12 @@ static int
 parse_bytes(uint8_t *data, char *input_arg, uint16_t max_size)
 {
 	unsigned byte_count;
-	char *token;
+	char *token, *sp = NULL;
 
 	errno = 0;
-	for (byte_count = 0, token = strtok(input_arg, ":");
+	for (byte_count = 0, token = strtok_r(input_arg, ":", &sp);
 			(byte_count < max_size) && (token != NULL);
-			token = strtok(NULL, ":")) {
+			token = strtok_r(NULL, ":", &sp)) {
 
 		int number = (int)strtol(token, NULL, 16);
 
-- 
2.22.0

[PATCH v5 24/25] examples/vhost: replace strtok with reentrant version

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

The strtok() is non-reentrant, it is better to replace it with a
reentrant version.

Fixes: 53d3f4778c1d ("vhost: integrate dmadev in asynchronous data-path")

Signed-off-by: Jie Hai <haijie1@huawei.com>
---
 examples/vhost/main.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index 4391d88c3d15..0fbb11b1d4f4 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -29,6 +29,7 @@
 #include <rte_dmadev.h>
 #include <rte_vhost_async.h>
 #include <rte_thread.h>
+#include <rte_os_shim.h>
 
 #include "main.h"
 
@@ -259,6 +260,7 @@ open_dma(const char *value)
 	uint16_t i = 0;
 	char *dma_arg[RTE_MAX_VHOST_DEVICE];
 	int args_nr;
+	char *sp = NULL;
 
 	if (input == NULL)
 		return -1;
@@ -272,7 +274,7 @@ open_dma(const char *value)
 
 	/* process DMA devices within bracket. */
 	addrs++;
-	substr = strtok(addrs, ";]");
+	substr = strtok_r(addrs, ";]", &sp);
 	if (!substr) {
 		ret = -1;
 		goto out;
-- 
2.22.0

[PATCH v5 25/25] devtools: check for some reentrant function

[Jie Hai]

Multiple threads calling the same function may cause condition
race issues, which often leads to abnormal behavior and can cause
more serious vulnerabilities such as abnormal termination, denial
of service, and compromised data integrity.

This patch adds check in checkpatches.sh for strtok, which is
non-reentrant.

Cc: stable@dpdk.org

Signed-off-by: Jie Hai <haijie1@huawei.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 devtools/checkpatches.sh | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/devtools/checkpatches.sh b/devtools/checkpatches.sh
index 4a8591be225e..80fe41604427 100755
--- a/devtools/checkpatches.sh
+++ b/devtools/checkpatches.sh
@@ -145,6 +145,14 @@ check_forbidden_additions() { # <patch>
 		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
 		"$1" || res=1
 
+	# refrain from using some non-reentrant functions
+	awk -v FOLDERS="lib drivers app examples" \
+		-v EXPRESSIONS="strtok\\\(" \
+		-v RET_ON_FAIL=1 \
+		-v MESSAGE='Using non-reentrant function strtok, prefer strtok_r' \
+		-f $(dirname $(readlink -f $0))/check-forbidden-tokens.awk \
+		"$1" || res=1
+
 	# refrain from using some pthread functions
 	awk -v FOLDERS="lib drivers app examples" \
 		-v EXPRESSIONS="pthread_(create|join|detach|set(_?name_np|affinity_np)|attr_set(inheritsched|schedpolicy))\\\(" \
-- 
2.22.0

Re: [PATCH v5 00/25]

[David Marchand]

On Fri, Nov 8, 2024 at 12:15 PM Jie Hai <haijie1@huawei.com> wrote:
>
> Multiple threads calling the same function may cause condition
> race issues, which often leads to abnormal behavior and can cause
> more serious vulnerabilities such as abnormal termination, denial
> of service, and compromised data integrity.
>
> This patchset replaces strtok with strtok_r in app, example, lib
> and drivers. And adds check for use of strtok in checkpatches.sh.
>
> --
> v5:
> 1. remove CC stable for some patch.
> 2. replace strtok for all files.

In case you did not notice, including rte_os_shim.h in examples breaks
compilation.
Too late for taking in rc2, I may consider in rc3 if the series is
fixed by then.


Thanks.

-- 
David Marchand