[dpdk-stable] patch 'metrics: add check for invalid key' has been queued to LTS release 17.11.4

[dpdk-stable] patch 'metrics: add check for invalid key' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From f4f7cd84a0527f12005b59dfaca10340727a2977 Mon Sep 17 00:00:00 2001
From: Remy Horton <remy.horton@intel.com>
Date: Fri, 6 Jul 2018 14:21:19 +0100
Subject: [PATCH] metrics: add check for invalid key

[ upstream commit 50d2459fdde1916c2ebce1a3751ab2fa59c0766f ]

This patchset adds a check to rte_metrics_update_values()
that prevents the updating of metrics when presented with
an invalid metric key. Previously, doing the latter could
result in a crash.

Fixes: 349950ddb9c5 ("metrics: add information metrics library")

Signed-off-by: Remy Horton <remy.horton@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
 lib/librte_metrics/rte_metrics.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/lib/librte_metrics/rte_metrics.c b/lib/librte_metrics/rte_metrics.c
index d9404001a..26c65b173 100644
--- a/lib/librte_metrics/rte_metrics.c
+++ b/lib/librte_metrics/rte_metrics.c
@@ -190,6 +190,11 @@ rte_metrics_update_values(int port_id,
 	stats = memzone->addr;
 
 	rte_spinlock_lock(&stats->lock);
+
+	if (key >= stats->cnt_stats) {
+		rte_spinlock_unlock(&stats->lock);
+		return -EINVAL;
+	}
 	idx_metric = key;
 	cnt_setsize = 1;
 	while (idx_metric < stats->cnt_stats) {
-- 
2.11.0

[dpdk-stable] patch 'metrics: do not fail silently when uninitialised' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From 2191a73c5584a610d097f34fb8b419c666255488 Mon Sep 17 00:00:00 2001
From: Remy Horton <remy.horton@intel.com>
Date: Mon, 2 Jul 2018 15:55:47 +0100
Subject: [PATCH] metrics: do not fail silently when uninitialised

[ upstream commit a84bdf6a101f2ab3c3859a40f366962d78c12868 ]

If rte_metrics_init() had not been called and hence the internal
metric storage is not allocated, rte_metrics_get_values() and
rte_metrics_get_name() would silently fail by returning zero
(i.e. no metrics registered). This patch changes the result of
this scenario to an explicit fail by returning -EIO.

Fixes: 349950ddb9c5 ("metrics: add information metrics library")

Signed-off-by: Remy Horton <remy.horton@intel.com>
---
 lib/librte_metrics/rte_metrics.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/lib/librte_metrics/rte_metrics.c b/lib/librte_metrics/rte_metrics.c
index 26c65b173..3ddf9caa4 100644
--- a/lib/librte_metrics/rte_metrics.c
+++ b/lib/librte_metrics/rte_metrics.c
@@ -236,9 +236,8 @@ rte_metrics_get_names(struct rte_metric_name *names,
 	int return_value;
 
 	memzone = rte_memzone_lookup(RTE_METRICS_MEMZONE_NAME);
-	/* If not allocated, fail silently */
 	if (memzone == NULL)
-		return 0;
+		return -EIO;
 
 	stats = memzone->addr;
 	rte_spinlock_lock(&stats->lock);
@@ -274,9 +273,9 @@ rte_metrics_get_values(int port_id,
 		return -EINVAL;
 
 	memzone = rte_memzone_lookup(RTE_METRICS_MEMZONE_NAME);
-	/* If not allocated, fail silently */
 	if (memzone == NULL)
-		return 0;
+		return -EIO;
+
 	stats = memzone->addr;
 	rte_spinlock_lock(&stats->lock);
 
-- 
2.11.0

[dpdk-stable] patch 'metrics: disallow null as metric name' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From 9b7a4b363682184c5b7e186c3301b5cf98fcb48d Mon Sep 17 00:00:00 2001
From: Remy Horton <remy.horton@intel.com>
Date: Mon, 2 Jul 2018 14:46:08 +0100
Subject: [PATCH] metrics: disallow null as metric name

[ upstream commit 219731302a0aa5b8edb30a6bc1cdea64e7824a98 ]

This patch adds a sanity check so that names passed into
rte_metrics_reg_names() and the wrapper rte_metrics_reg_name()
cannot be NULL.

Fixes: 349950ddb9c5 ("metrics: add information metrics library")

Signed-off-by: Remy Horton <remy.horton@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
 lib/librte_metrics/rte_metrics.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/lib/librte_metrics/rte_metrics.c b/lib/librte_metrics/rte_metrics.c
index 3ddf9caa4..dc6344673 100644
--- a/lib/librte_metrics/rte_metrics.c
+++ b/lib/librte_metrics/rte_metrics.c
@@ -124,6 +124,9 @@ rte_metrics_reg_names(const char * const *names, uint16_t cnt_names)
 	/* Some sanity checks */
 	if (cnt_names < 1 || names == NULL)
 		return -EINVAL;
+	for (idx_name = 0; idx_name < cnt_names; idx_name++)
+		if (names[idx_name] == NULL)
+			return -EINVAL;
 
 	memzone = rte_memzone_lookup(RTE_METRICS_MEMZONE_NAME);
 	if (memzone == NULL)
-- 
2.11.0

[dpdk-stable] patch 'test: fix EAL flags autotest on FreeBSD' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From b5eb3351a6e41a2f69a687b3de473187b789bb02 Mon Sep 17 00:00:00 2001
From: Anatoly Burakov <anatoly.burakov@intel.com>
Date: Mon, 16 Jul 2018 17:34:04 +0100
Subject: [PATCH] test: fix EAL flags autotest on FreeBSD

[ upstream commit b5d878e6db56c76ffe5e8c69b6c19faac4e97d80 ]

FreeBSD does not support running multiple primary processes
concurrently, because all DPDK instances will allocate memory
from the same place (memory provided by contigmem driver).
While it is technically possible to launch a DPDK process
using no-shconf switch, it will actually corrupt main process'
for the above reason.

Fix EAL flags autotest to not run primary processes unless
both no-shconf and no-huge are specified.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Tested-by: Changqing Wu <changqingx.wu@intel.com>
---
 test/test/test_eal_flags.c | 33 +++++++++++++++++----------------
 1 file changed, 17 insertions(+), 16 deletions(-)

diff --git a/test/test/test_eal_flags.c b/test/test/test_eal_flags.c
index 473ea113d..e82bb963a 100644
--- a/test/test/test_eal_flags.c
+++ b/test/test/test_eal_flags.c
@@ -404,17 +404,17 @@ test_invalid_vdev_flag(void)
 #endif
 
 	/* Test with invalid vdev option */
-	const char *vdevinval[] = {prgname, prefix, "-n", "1",
+	const char *vdevinval[] = {prgname, prefix, no_huge, "-n", "1",
 				"-c", "1", vdev, "eth_dummy"};
 
 	/* Test with valid vdev option */
-	const char *vdevval1[] = {prgname, prefix, "-n", "1",
+	const char *vdevval1[] = {prgname, prefix, no_huge, "-n", "1",
 	"-c", "1", vdev, "net_ring0"};
 
-	const char *vdevval2[] = {prgname, prefix, "-n", "1",
+	const char *vdevval2[] = {prgname, prefix, no_huge, "-n", "1",
 	"-c", "1", vdev, "net_ring0,args=test"};
 
-	const char *vdevval3[] = {prgname, prefix, "-n", "1",
+	const char *vdevval3[] = {prgname, prefix, no_huge, "-n", "1",
 	"-c", "1", vdev, "net_ring0,nodeaction=r1:0:CREATE"};
 
 	if (launch_proc(vdevinval) == 0) {
@@ -877,13 +877,10 @@ test_misc_flags(void)
 	const char *argv4[] = {prgname, prefix, mp_flag, "-c", "1", "--syslog"};
 	/* With invalid --syslog */
 	const char *argv5[] = {prgname, prefix, mp_flag, "-c", "1", "--syslog", "error"};
-	/* With no-sh-conf */
+	/* With no-sh-conf, also use no-huge to ensure this test runs on BSD */
 	const char *argv6[] = {prgname, "-c", "1", "-n", "2", "-m", DEFAULT_MEM_SIZE,
-			no_shconf, nosh_prefix };
+			no_shconf, nosh_prefix, no_huge};
 
-#ifdef RTE_EXEC_ENV_BSDAPP
-	return 0;
-#endif
 	/* With --huge-dir */
 	const char *argv7[] = {prgname, "-c", "1", "-n", "2", "-m", DEFAULT_MEM_SIZE,
 			"--file-prefix=hugedir", "--huge-dir", hugepath};
@@ -917,6 +914,7 @@ test_misc_flags(void)
 	const char *argv15[] = {prgname, "--file-prefix=intr",
 			"-c", "1", "-n", "2", "--vfio-intr=invalid"};
 
+	/* run all tests also applicable to FreeBSD first */
 
 	if (launch_proc(argv0) == 0) {
 		printf("Error - process ran ok with invalid flag\n");
@@ -930,6 +928,16 @@ test_misc_flags(void)
 		printf("Error - process did not run ok with -v flag\n");
 		return -1;
 	}
+	if (launch_proc(argv6) != 0) {
+		printf("Error - process did not run ok with --no-shconf flag\n");
+		return -1;
+	}
+
+#ifdef RTE_EXEC_ENV_BSDAPP
+	/* no more tests to be done on FreeBSD */
+	return 0;
+#endif
+
 	if (launch_proc(argv3) != 0) {
 		printf("Error - process did not run ok with --syslog flag\n");
 		return -1;
@@ -942,13 +950,6 @@ test_misc_flags(void)
 		printf("Error - process run ok with invalid --syslog flag\n");
 		return -1;
 	}
-	if (launch_proc(argv6) != 0) {
-		printf("Error - process did not run ok with --no-shconf flag\n");
-		return -1;
-	}
-#ifdef RTE_EXEC_ENV_BSDAPP
-	return 0;
-#endif
 	if (launch_proc(argv7) != 0) {
 		printf("Error - process did not run ok with --huge-dir flag\n");
 		return -1;
-- 
2.11.0

[dpdk-stable] patch 'test/flow_classify: fix return types' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From d67b45ebbb287517f4528ac116d5b0d25005ab66 Mon Sep 17 00:00:00 2001
From: Reshma Pattan <reshma.pattan@intel.com>
Date: Tue, 17 Jul 2018 17:44:45 +0100
Subject: [PATCH] test/flow_classify: fix return types

[ upstream commit 4f872116c59a1342dd7ab7165adf05a624480b70 ]

UTs should return either TEST_SUCCESS or TEST_FAILED only.
They should not return 0, -1 and any other value.

Fixes: 9c9befea4f ("test: add flow classify unit tests")

Signed-off-by: Reshma Pattan <reshma.pattan@intel.com>
Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>
---
 test/test/test_flow_classify.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/test/test/test_flow_classify.c b/test/test/test_flow_classify.c
index 9f331cd8e..a8a1cad41 100644
--- a/test/test/test_flow_classify.c
+++ b/test/test/test_flow_classify.c
@@ -641,32 +641,32 @@ test_flow_classify(void)
 		printf("Line %i: f_create has failed!\n", __LINE__);
 		rte_flow_classifier_free(cls->cls);
 		rte_free(cls);
-		return -1;
+		return TEST_FAILED;
 	}
 	printf("Created table_acl for for IPv4 five tuple packets\n");
 
 	ret = init_mbufpool();
 	if (ret) {
 		printf("Line %i: init_mbufpool has failed!\n", __LINE__);
-		return -1;
+		return TEST_FAILED;
 	}
 
 	if (test_invalid_parameters() < 0)
-		return -1;
+		return TEST_FAILED;
 	if (test_valid_parameters() < 0)
-		return -1;
+		return TEST_FAILED;
 	if (test_invalid_patterns() < 0)
-		return -1;
+		return TEST_FAILED;
 	if (test_invalid_actions() < 0)
-		return -1;
+		return TEST_FAILED;
 	if (test_query_udp() < 0)
-		return -1;
+		return TEST_FAILED;
 	if (test_query_tcp() < 0)
-		return -1;
+		return TEST_FAILED;
 	if (test_query_sctp() < 0)
-		return -1;
+		return TEST_FAILED;
 
-	return 0;
+	return TEST_SUCCESS;
 }
 
 REGISTER_TEST_COMMAND(flow_classify_autotest, test_flow_classify);
-- 
2.11.0

[dpdk-stable] patch 'test/hash: fix multiwriter with non consecutive cores' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From 1716698055dafe82310c285cf0c49fbfb94dc8bc Mon Sep 17 00:00:00 2001
From: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Date: Tue, 17 Jul 2018 14:31:50 +0100
Subject: [PATCH] test/hash: fix multiwriter with non consecutive cores

[ upstream commit 0c992822901464d440534d337838fd08acf233d4 ]

When non consecutive cores are passed into the test application,
the distribution of the keys that each thread needs to insert
is not correct, since it assumes that there are no cores skipped
between the master core and the worker core.

Fixes: be856325cba3 ("hash: add scalable multi-writer insertion with Intel TSX")

Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---
 test/test/test_hash_multiwriter.c | 41 ++++++++++++++++++++++++++++++++++-----
 1 file changed, 36 insertions(+), 5 deletions(-)

diff --git a/test/test/test_hash_multiwriter.c b/test/test/test_hash_multiwriter.c
index 4dcbd9d56..951c3e743 100644
--- a/test/test/test_hash_multiwriter.c
+++ b/test/test/test_hash_multiwriter.c
@@ -76,18 +76,29 @@ static rte_atomic64_t ginsertions;
 static int use_htm;
 
 static int
-test_hash_multiwriter_worker(__attribute__((unused)) void *arg)
+test_hash_multiwriter_worker(void *arg)
 {
 	uint64_t i, offset;
+	uint16_t pos_core;
 	uint32_t lcore_id = rte_lcore_id();
 	uint64_t begin, cycles;
+	uint16_t *enabled_core_ids = (uint16_t *)arg;
 
-	offset = (lcore_id - rte_get_master_lcore())
-		* tbl_multiwriter_test_params.nb_tsx_insertion;
+	for (pos_core = 0; pos_core < rte_lcore_count(); pos_core++) {
+		if (enabled_core_ids[pos_core] == lcore_id)
+			break;
+	}
+
+	/*
+	 * Calculate offset for entries based on the position of the
+	 * logical core, from the master core (not counting not enabled cores)
+	 */
+	offset = pos_core * tbl_multiwriter_test_params.nb_tsx_insertion;
 
 	printf("Core #%d inserting %d: %'"PRId64" - %'"PRId64"\n",
 	       lcore_id, tbl_multiwriter_test_params.nb_tsx_insertion,
-	       offset, offset + tbl_multiwriter_test_params.nb_tsx_insertion);
+	       offset,
+	       offset + tbl_multiwriter_test_params.nb_tsx_insertion - 1);
 
 	begin = rte_rdtsc_precise();
 
@@ -116,6 +127,8 @@ test_hash_multiwriter(void)
 {
 	unsigned int i, rounded_nb_total_tsx_insertion;
 	static unsigned calledCount = 1;
+	uint16_t enabled_core_ids[RTE_MAX_LCORE];
+	uint16_t core_id;
 
 	uint32_t *keys;
 	uint32_t *found;
@@ -186,9 +199,27 @@ test_hash_multiwriter(void)
 	rte_atomic64_init(&ginsertions);
 	rte_atomic64_clear(&ginsertions);
 
+	/* Get list of enabled cores */
+	i = 0;
+	for (core_id = 0; core_id < RTE_MAX_LCORE; core_id++) {
+		if (i == rte_lcore_count())
+			break;
+
+		if (rte_lcore_is_enabled(core_id)) {
+			enabled_core_ids[i] = core_id;
+			i++;
+		}
+	}
+
+	if (i != rte_lcore_count()) {
+		printf("Number of enabled cores in list is different from "
+				"number given by rte_lcore_count()\n");
+		goto err3;
+	}
+
 	/* Fire all threads. */
 	rte_eal_mp_remote_launch(test_hash_multiwriter_worker,
-				 NULL, CALL_MASTER);
+				 enabled_core_ids, CALL_MASTER);
 	rte_eal_mp_wait_lcore();
 
 	while (rte_hash_iterate(handle, &next_key, &next_data, &iter) >= 0) {
-- 
2.11.0

[dpdk-stable] patch 'test/hash: fix potential memory leak' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From 13621ecc7ba85128300cee9599a37d5a58400442 Mon Sep 17 00:00:00 2001
From: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Date: Tue, 17 Jul 2018 14:31:51 +0100
Subject: [PATCH] test/hash: fix potential memory leak

[ upstream commit a4492ab49ec3b829745c9060a1719354e0a572b7 ]

In the multiwriter test, if "found" array allocation failed,
the memory of "keys" array, which was successfully allocated
could not be freed, since by this time, tbl_multiwriter_test_params.keys
was not set to this array, which is the pointer freed when finishing
the test or when a failure happens.

To solve this, tbl_multiwriter_test_params.keys is set to the "keys"
address, just after allocating and filling the array.

Fixes: be856325cba3 ("hash: add scalable multi-writer insertion with Intel TSX")

Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---
 test/test/test_hash_multiwriter.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/test/test/test_hash_multiwriter.c b/test/test/test_hash_multiwriter.c
index 951c3e743..0e9ac7162 100644
--- a/test/test/test_hash_multiwriter.c
+++ b/test/test/test_hash_multiwriter.c
@@ -181,16 +181,17 @@ test_hash_multiwriter(void)
 		goto err1;
 	}
 
+	for (i = 0; i < nb_entries; i++)
+		keys[i] = i;
+
+	tbl_multiwriter_test_params.keys = keys;
+
 	found = rte_zmalloc(NULL, sizeof(uint32_t) * nb_entries, 0);
 	if (found == NULL) {
 		printf("RTE_ZMALLOC failed\n");
 		goto err2;
 	}
 
-	for (i = 0; i < nb_entries; i++)
-		keys[i] = i;
-
-	tbl_multiwriter_test_params.keys = keys;
 	tbl_multiwriter_test_params.found = found;
 
 	rte_atomic64_init(&gcycles);
-- 
2.11.0

[dpdk-stable] patch 'doc: fix bonding command in testpmd' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From b92b2f5760e77b4537141e6b3d2eb0f271274b19 Mon Sep 17 00:00:00 2001
From: Shreyansh Jain <shreyansh.jain@nxp.com>
Date: Wed, 25 Jul 2018 17:01:32 +0530
Subject: [PATCH] doc: fix bonding command in testpmd

[ upstream commit 0c00abfd85448449d73acc09cc6fe445dba37b50 ]

Fixes: ac718398f477 ("doc: testpmd application user guide")

Signed-off-by: Shreyansh Jain <shreyansh.jain@nxp.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
---
 doc/guides/testpmd_app_ug/testpmd_funcs.rst | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/doc/guides/testpmd_app_ug/testpmd_funcs.rst b/doc/guides/testpmd_app_ug/testpmd_funcs.rst
index 9789139af..c9ce85c90 100644
--- a/doc/guides/testpmd_app_ug/testpmd_funcs.rst
+++ b/doc/guides/testpmd_app_ug/testpmd_funcs.rst
@@ -1865,7 +1865,7 @@ Create a new bonding device::
 
 For example, to create a bonded device in mode 1 on socket 0::
 
-   testpmd> create bonded 1 0
+   testpmd> create bonded device 1 0
    created new bonded device (port X)
 
 add bonding slave
-- 
2.11.0

[dpdk-stable] patch 'net/mlx5: add missing sanity checks for Tx completion queue' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From bcbe141318c63055bf3f87df0965ae84d01c0d94 Mon Sep 17 00:00:00 2001
From: Yongseok Koh <yskoh@mellanox.com>
Date: Tue, 26 Dec 2017 19:55:45 -0800
Subject: [PATCH] net/mlx5: add missing sanity checks for Tx completion queue

[ upstream commit 2eefbec531c7abf4c9c668f7941b285858c06733 ]

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5_rxtx.c          | 19 ++++++++++++++++++-
 drivers/net/mlx5/mlx5_rxtx.h          |  2 ++
 drivers/net/mlx5/mlx5_rxtx_vec_neon.h |  8 ++++++++
 drivers/net/mlx5/mlx5_rxtx_vec_sse.h  |  8 ++++++++
 drivers/net/mlx5/mlx5_txq.c           |  2 ++
 5 files changed, 38 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 2e003aeac..38f036c4b 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -401,6 +401,8 @@ mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	/* Start processing. */
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
+	/* A CQE slot must always be available. */
+	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	if (unlikely(!max_wqe))
 		return 0;
@@ -745,6 +747,9 @@ next_wqe:
 		/* Save elts_head in unused "immediate" field of WQE. */
 		last_wqe->ctrl3 = txq->elts_head;
 		txq->elts_comp = 0;
+#ifndef NDEBUG
+		++txq->cq_pi;
+#endif
 	} else {
 		txq->elts_comp = comp;
 	}
@@ -863,6 +868,8 @@ mlx5_tx_burst_mpw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	/* Start processing. */
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
+	/* A CQE slot must always be available. */
+	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	if (unlikely(!max_wqe))
 		return 0;
@@ -956,6 +963,9 @@ mlx5_tx_burst_mpw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 		/* Save elts_head in unused "immediate" field of WQE. */
 		wqe->ctrl[3] = elts_head;
 		txq->elts_comp = 0;
+#ifndef NDEBUG
+		++txq->cq_pi;
+#endif
 	} else {
 		txq->elts_comp = comp;
 	}
@@ -1087,6 +1097,8 @@ mlx5_tx_burst_mpw_inline(void *dpdk_txq, struct rte_mbuf **pkts,
 	/* Start processing. */
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
+	/* A CQE slot must always be available. */
+	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	do {
 		struct rte_mbuf *buf = *(pkts++);
 		uintptr_t addr;
@@ -1248,6 +1260,9 @@ mlx5_tx_burst_mpw_inline(void *dpdk_txq, struct rte_mbuf **pkts,
 		/* Save elts_head in unused "immediate" field of WQE. */
 		wqe->ctrl[3] = elts_head;
 		txq->elts_comp = 0;
+#ifndef NDEBUG
+		++txq->cq_pi;
+#endif
 	} else {
 		txq->elts_comp = comp;
 	}
@@ -1590,7 +1605,9 @@ mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 		wqe->ctrl[3] = elts_head;
 		txq->elts_comp = 0;
 		txq->mpw_comp = txq->wqe_ci;
-		txq->cq_pi++;
+#ifndef NDEBUG
+		++txq->cq_pi;
+#endif
 	} else {
 		txq->elts_comp += j;
 	}
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index 29019f792..dac3b39ff 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -185,7 +185,9 @@ struct mlx5_txq_data {
 	uint16_t elts_comp; /* Counter since last completion request. */
 	uint16_t mpw_comp; /* WQ index since last completion request. */
 	uint16_t cq_ci; /* Consumer index for completion queue. */
+#ifndef NDEBUG
 	uint16_t cq_pi; /* Producer index for completion queue. */
+#endif
 	uint16_t wqe_ci; /* Consumer index for work queue. */
 	uint16_t wqe_pi; /* Producer index for work queue. */
 	uint16_t elts_n:4; /* (*elts)[] length (in log2). */
diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
index 1604d0430..32d1e6de9 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
+++ b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
@@ -135,6 +135,8 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 
 	assert(elts_n > pkts_n);
 	mlx5_tx_complete(txq);
+	/* A CQE slot must always be available. */
+	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	if (unlikely(!pkts_n))
 		return 0;
 	for (n = 0; n < pkts_n; ++n) {
@@ -205,7 +207,9 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 		wqe->ctrl[2] = rte_cpu_to_be_32(8);
 		wqe->ctrl[3] = txq->elts_head;
 		txq->elts_comp = 0;
+#ifndef NDEBUG
 		++txq->cq_pi;
+#endif
 	}
 #ifdef MLX5_PMD_SOFT_COUNTERS
 	txq->stats.opackets += n;
@@ -269,6 +273,8 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	assert(elts_n > pkts_n);
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
+	/* A CQE slot must always be available. */
+	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	pkts_n = RTE_MIN((unsigned int)RTE_MIN(pkts_n, max_wqe), max_elts);
 	if (unlikely(!pkts_n))
@@ -306,7 +312,9 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	} else {
 		/* Request a completion. */
 		txq->elts_comp = 0;
+#ifndef NDEBUG
 		++txq->cq_pi;
+#endif
 		comp_req = 8;
 	}
 	/* Fill CTRL in the header. */
diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
index 01a93a4b4..1a56fca9f 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
+++ b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
@@ -135,6 +135,8 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 
 	assert(elts_n > pkts_n);
 	mlx5_tx_complete(txq);
+	/* A CQE slot must always be available. */
+	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	if (unlikely(!pkts_n))
 		return 0;
 	for (n = 0; n < pkts_n; ++n) {
@@ -206,7 +208,9 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 		wqe->ctrl[2] = rte_cpu_to_be_32(8);
 		wqe->ctrl[3] = txq->elts_head;
 		txq->elts_comp = 0;
+#ifndef NDEBUG
 		++txq->cq_pi;
+#endif
 	}
 #ifdef MLX5_PMD_SOFT_COUNTERS
 	txq->stats.opackets += n;
@@ -268,6 +272,8 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	assert(elts_n > pkts_n);
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
+	/* A CQE slot must always be available. */
+	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	pkts_n = RTE_MIN((unsigned int)RTE_MIN(pkts_n, max_wqe), max_elts);
 	assert(pkts_n <= MLX5_DSEG_MAX - nb_dword_in_hdr);
@@ -307,7 +313,9 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	} else {
 		/* Request a completion. */
 		txq->elts_comp = 0;
+#ifndef NDEBUG
 		++txq->cq_pi;
+#endif
 		comp_req = 8;
 	}
 	/* Fill CTRL in the header. */
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index a5c6b5851..760ac92d4 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -467,7 +467,9 @@ mlx5_txq_ibv_new(struct rte_eth_dev *dev, uint16_t idx)
 		(volatile struct mlx5_cqe (*)[])
 		(uintptr_t)cq_info.buf;
 	txq_data->cq_ci = 0;
+#ifndef NDEBUG
 	txq_data->cq_pi = 0;
+#endif
 	txq_data->wqe_ci = 0;
 	txq_data->wqe_pi = 0;
 	txq_ibv->qp = tmpl.qp;
-- 
2.11.0

[dpdk-stable] patch 'net/mlx5: fix assert for Tx completion queue count' has been queued to LTS release 17.11.4

[Yongseok Koh]

Hi,

FYI, your patch has been queued to LTS release 17.11.4

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 07/28/18. So please
shout if anyone has objections.

Thanks.

Yongseok

---
>From 3bf985f340ab3ebe9318331cbd42ed05e08781ab Mon Sep 17 00:00:00 2001
From: Yongseok Koh <yskoh@mellanox.com>
Date: Mon, 23 Jul 2018 11:00:10 -0700
Subject: [PATCH] net/mlx5: fix assert for Tx completion queue count

[ upstream commit c618e7e82bfed1da91f1dbabdd06e0a39fdb7b78 ]

There should be at least one Tx CQE remained if Tx WQ and txq->elts[] have
available slots to send a packet because the size of Tx CQ is exactly
calculated from the size of other resources. As it is guaranteed, it is
checked by an assertion.

max_elts is checked after the assertion for Tx CQ. If no slot is available
in txq->elts[], the assertion would be wrong.

Fixes: 2eefbec531c7 ("net/mlx5: add missing sanity checks for Tx completion queue")
Fixes: 6ce84bd88919 ("net/mlx5: add enhanced multi-packet send for ConnectX-5")

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Xueming Li <xuemingl@mellanox.com>
---
 drivers/net/mlx5/mlx5_rxtx.c          | 28 ++++++++--------------------
 drivers/net/mlx5/mlx5_rxtx_vec_neon.h | 14 ++++----------
 drivers/net/mlx5/mlx5_rxtx_vec_sse.h  | 14 ++++----------
 3 files changed, 16 insertions(+), 40 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 38f036c4b..1bbce3b75 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -401,8 +401,6 @@ mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	/* Start processing. */
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
-	/* A CQE slot must always be available. */
-	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	if (unlikely(!max_wqe))
 		return 0;
@@ -742,14 +740,13 @@ next_wqe:
 	/* Check whether completion threshold has been reached. */
 	comp = txq->elts_comp + i + j + k;
 	if (comp >= MLX5_TX_COMP_THRESH) {
+		/* A CQE slot must always be available. */
+		assert((1u << txq->cqe_n) - (txq->cq_pi++ - txq->cq_ci));
 		/* Request completion on last WQE. */
 		last_wqe->ctrl2 = rte_cpu_to_be_32(8);
 		/* Save elts_head in unused "immediate" field of WQE. */
 		last_wqe->ctrl3 = txq->elts_head;
 		txq->elts_comp = 0;
-#ifndef NDEBUG
-		++txq->cq_pi;
-#endif
 	} else {
 		txq->elts_comp = comp;
 	}
@@ -868,8 +865,6 @@ mlx5_tx_burst_mpw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	/* Start processing. */
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
-	/* A CQE slot must always be available. */
-	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	if (unlikely(!max_wqe))
 		return 0;
@@ -958,14 +953,13 @@ mlx5_tx_burst_mpw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	if (comp >= MLX5_TX_COMP_THRESH) {
 		volatile struct mlx5_wqe *wqe = mpw.wqe;
 
+		/* A CQE slot must always be available. */
+		assert((1u << txq->cqe_n) - (txq->cq_pi++ - txq->cq_ci));
 		/* Request completion on last WQE. */
 		wqe->ctrl[2] = rte_cpu_to_be_32(8);
 		/* Save elts_head in unused "immediate" field of WQE. */
 		wqe->ctrl[3] = elts_head;
 		txq->elts_comp = 0;
-#ifndef NDEBUG
-		++txq->cq_pi;
-#endif
 	} else {
 		txq->elts_comp = comp;
 	}
@@ -1097,8 +1091,6 @@ mlx5_tx_burst_mpw_inline(void *dpdk_txq, struct rte_mbuf **pkts,
 	/* Start processing. */
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
-	/* A CQE slot must always be available. */
-	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	do {
 		struct rte_mbuf *buf = *(pkts++);
 		uintptr_t addr;
@@ -1255,14 +1247,13 @@ mlx5_tx_burst_mpw_inline(void *dpdk_txq, struct rte_mbuf **pkts,
 	if (comp >= MLX5_TX_COMP_THRESH) {
 		volatile struct mlx5_wqe *wqe = mpw.wqe;
 
+		/* A CQE slot must always be available. */
+		assert((1u << txq->cqe_n) - (txq->cq_pi++ - txq->cq_ci));
 		/* Request completion on last WQE. */
 		wqe->ctrl[2] = rte_cpu_to_be_32(8);
 		/* Save elts_head in unused "immediate" field of WQE. */
 		wqe->ctrl[3] = elts_head;
 		txq->elts_comp = 0;
-#ifndef NDEBUG
-		++txq->cq_pi;
-#endif
 	} else {
 		txq->elts_comp = comp;
 	}
@@ -1385,8 +1376,6 @@ mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	/* Start processing. */
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
-	/* A CQE slot must always be available. */
-	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	if (unlikely(!max_wqe))
 		return 0;
@@ -1599,15 +1588,14 @@ mlx5_tx_burst_empw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 			 (1 << txq->wqe_n) / MLX5_TX_COMP_THRESH_INLINE_DIV) {
 		volatile struct mlx5_wqe *wqe = mpw.wqe;
 
+		/* A CQE slot must always be available. */
+		assert((1u << txq->cqe_n) - (txq->cq_pi++ - txq->cq_ci));
 		/* Request completion on last WQE. */
 		wqe->ctrl[2] = rte_cpu_to_be_32(8);
 		/* Save elts_head in unused "immediate" field of WQE. */
 		wqe->ctrl[3] = elts_head;
 		txq->elts_comp = 0;
 		txq->mpw_comp = txq->wqe_ci;
-#ifndef NDEBUG
-		++txq->cq_pi;
-#endif
 	} else {
 		txq->elts_comp += j;
 	}
diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
index 32d1e6de9..e748615ee 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
+++ b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h
@@ -135,8 +135,6 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 
 	assert(elts_n > pkts_n);
 	mlx5_tx_complete(txq);
-	/* A CQE slot must always be available. */
-	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	if (unlikely(!pkts_n))
 		return 0;
 	for (n = 0; n < pkts_n; ++n) {
@@ -204,12 +202,11 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 	txq->elts_comp += (uint16_t)(elts_head - txq->elts_head);
 	txq->elts_head = elts_head;
 	if (txq->elts_comp >= MLX5_TX_COMP_THRESH) {
+		/* A CQE slot must always be available. */
+		assert((1u << txq->cqe_n) - (txq->cq_pi++ - txq->cq_ci));
 		wqe->ctrl[2] = rte_cpu_to_be_32(8);
 		wqe->ctrl[3] = txq->elts_head;
 		txq->elts_comp = 0;
-#ifndef NDEBUG
-		++txq->cq_pi;
-#endif
 	}
 #ifdef MLX5_PMD_SOFT_COUNTERS
 	txq->stats.opackets += n;
@@ -273,8 +270,6 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	assert(elts_n > pkts_n);
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
-	/* A CQE slot must always be available. */
-	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	pkts_n = RTE_MIN((unsigned int)RTE_MIN(pkts_n, max_wqe), max_elts);
 	if (unlikely(!pkts_n))
@@ -310,11 +305,10 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	if (txq->elts_comp + pkts_n < MLX5_TX_COMP_THRESH) {
 		txq->elts_comp += pkts_n;
 	} else {
+		/* A CQE slot must always be available. */
+		assert((1u << txq->cqe_n) - (txq->cq_pi++ - txq->cq_ci));
 		/* Request a completion. */
 		txq->elts_comp = 0;
-#ifndef NDEBUG
-		++txq->cq_pi;
-#endif
 		comp_req = 8;
 	}
 	/* Fill CTRL in the header. */
diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
index 1a56fca9f..7e8c9b885 100644
--- a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
+++ b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
@@ -135,8 +135,6 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 
 	assert(elts_n > pkts_n);
 	mlx5_tx_complete(txq);
-	/* A CQE slot must always be available. */
-	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	if (unlikely(!pkts_n))
 		return 0;
 	for (n = 0; n < pkts_n; ++n) {
@@ -205,12 +203,11 @@ txq_scatter_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts,
 	txq->elts_comp += (uint16_t)(elts_head - txq->elts_head);
 	txq->elts_head = elts_head;
 	if (txq->elts_comp >= MLX5_TX_COMP_THRESH) {
+		/* A CQE slot must always be available. */
+		assert((1u << txq->cqe_n) - (txq->cq_pi++ - txq->cq_ci));
 		wqe->ctrl[2] = rte_cpu_to_be_32(8);
 		wqe->ctrl[3] = txq->elts_head;
 		txq->elts_comp = 0;
-#ifndef NDEBUG
-		++txq->cq_pi;
-#endif
 	}
 #ifdef MLX5_PMD_SOFT_COUNTERS
 	txq->stats.opackets += n;
@@ -272,8 +269,6 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	assert(elts_n > pkts_n);
 	mlx5_tx_complete(txq);
 	max_elts = (elts_n - (elts_head - txq->elts_tail));
-	/* A CQE slot must always be available. */
-	assert((1u << txq->cqe_n) - (txq->cq_pi - txq->cq_ci));
 	max_wqe = (1u << txq->wqe_n) - (txq->wqe_ci - txq->wqe_pi);
 	pkts_n = RTE_MIN((unsigned int)RTE_MIN(pkts_n, max_wqe), max_elts);
 	assert(pkts_n <= MLX5_DSEG_MAX - nb_dword_in_hdr);
@@ -311,11 +306,10 @@ txq_burst_v(struct mlx5_txq_data *txq, struct rte_mbuf **pkts, uint16_t pkts_n,
 	if (txq->elts_comp + pkts_n < MLX5_TX_COMP_THRESH) {
 		txq->elts_comp += pkts_n;
 	} else {
+		/* A CQE slot must always be available. */
+		assert((1u << txq->cqe_n) - (txq->cq_pi++ - txq->cq_ci));
 		/* Request a completion. */
 		txq->elts_comp = 0;
-#ifndef NDEBUG
-		++txq->cq_pi;
-#endif
 		comp_req = 8;
 	}
 	/* Fill CTRL in the header. */
-- 
2.11.0