diff mbox

[dpdk-dev,v2] test: Fix memory corruption issues which fails the link_bonding test.

Message ID 1499685226-8659-1-git-send-email-herbert.guan@arm.com (mailing list archive)
State Accepted, archived
Delegated to: Ferruh Yigit
Headers

Checks

Context Check Description
ci/Intel-compilation fail Compilation issues
ci/checkpatch success coding style OK

Commit Message

Herbert Guan July 10, 2017, 11:13 a.m. UTC 
  Patch V2: fix build warnings by deleting unused variables. 

There were double-free problems in some test cases, which will cause
a duplicated mbuf will be added into mempool.  After double-free, 
some new allocated mbuf will hold a same address and thus cause the 
memory corruption.

Another minor issue is that in some test cases, allocated mbuf will
not be released after test case exits.  Hopefully these leaked mbuf
will be released by the next test case in its setup phase when
stopping the virtual pmd ports, while this do is a memory leak of
the exited test case.

To fix above 2 issues, this patch will do:
1) Release virtual pmd ports' tx queue in the clean up function
   remove_slaves_and_stop_bonded_device() of each test cases.
2) Do not release allocated mbufs for test bursts.  These mbufs
   will be released in remove_slaves_and_stop_bonded_device() when
   test case exits.

Signed-off-by: Herbert Guan <herbert.guan@arm.com>
---
 test/test/test_link_bonding.c | 64 ++++++-------------------------------------
 1 file changed, 9 insertions(+), 55 deletions(-)

Comments

Doherty, Declan July 12, 2017, 10:23 a.m. UTC | #1 
On 10/07/17 12:13, Herbert Guan wrote:
> Patch V2: fix build warnings by deleting unused variables.
> 
> There were double-free problems in some test cases, which will cause
> a duplicated mbuf will be added into mempool.  After double-free,
> some new allocated mbuf will hold a same address and thus cause the
> memory corruption.
> 
> Another minor issue is that in some test cases, allocated mbuf will
> not be released after test case exits.  Hopefully these leaked mbuf
> will be released by the next test case in its setup phase when
> stopping the virtual pmd ports, while this do is a memory leak of
> the exited test case.
> 
> To fix above 2 issues, this patch will do:
> 1) Release virtual pmd ports' tx queue in the clean up function
>     remove_slaves_and_stop_bonded_device() of each test cases.
> 2) Do not release allocated mbufs for test bursts.  These mbufs
>     will be released in remove_slaves_and_stop_bonded_device() when
>     test case exits.
> 
> Signed-off-by: Herbert Guan <herbert.guan@arm.com>
> ---
...
> 

Acked-by: Declan Doherty <declan.doherty@intel.com>
Ferruh Yigit July 18, 2017, 4:13 p.m. UTC | #2 
On 7/10/2017 12:13 PM, Herbert Guan wrote:
> Patch V2: fix build warnings by deleting unused variables. 
> 
> There were double-free problems in some test cases, which will cause
> a duplicated mbuf will be added into mempool.  After double-free, 
> some new allocated mbuf will hold a same address and thus cause the 
> memory corruption.
> 
> Another minor issue is that in some test cases, allocated mbuf will
> not be released after test case exits.  Hopefully these leaked mbuf
> will be released by the next test case in its setup phase when
> stopping the virtual pmd ports, while this do is a memory leak of
> the exited test case.
> 
> To fix above 2 issues, this patch will do:
> 1) Release virtual pmd ports' tx queue in the clean up function
>    remove_slaves_and_stop_bonded_device() of each test cases.
> 2) Do not release allocated mbufs for test bursts.  These mbufs
>    will be released in remove_slaves_and_stop_bonded_device() when
>    test case exits.
> 
> Signed-off-by: Herbert Guan <herbert.guan@arm.com>

Updated unit test causing segfault, before and after patch, am I doing
something wrong?

I will postpone this patch until unit test can be verified.

Thanks,
ferruh
Herbert Guan July 19, 2017, 1:54 a.m. UTC | #3 
It's caused by some recent merges.  Applying this patch to 17.05 is okay.  I've identified the place of segfault, but there're still some other problems to this test case.

Thanks,
Herbert

-----Original Message-----
From: Ferruh Yigit [mailto:ferruh.yigit@intel.com]

Sent: Wednesday, July 19, 2017 0:14
To: Herbert Guan <Herbert.Guan@arm.com>; dev@dpdk.org; declan.doherty@intel.com; jianbo.liu@linaro.org
Subject: Re: [dpdk-dev] [PATCH v2] test: Fix memory corruption issues which fails the link_bonding test.

On 7/10/2017 12:13 PM, Herbert Guan wrote:
> Patch V2: fix build warnings by deleting unused variables.

>

> There were double-free problems in some test cases, which will cause a

> duplicated mbuf will be added into mempool.  After double-free, some

> new allocated mbuf will hold a same address and thus cause the memory

> corruption.

>

> Another minor issue is that in some test cases, allocated mbuf will

> not be released after test case exits.  Hopefully these leaked mbuf

> will be released by the next test case in its setup phase when

> stopping the virtual pmd ports, while this do is a memory leak of the

> exited test case.

>

> To fix above 2 issues, this patch will do:

> 1) Release virtual pmd ports' tx queue in the clean up function

>    remove_slaves_and_stop_bonded_device() of each test cases.

> 2) Do not release allocated mbufs for test bursts.  These mbufs

>    will be released in remove_slaves_and_stop_bonded_device() when

>    test case exits.

>

> Signed-off-by: Herbert Guan <herbert.guan@arm.com>


Updated unit test causing segfault, before and after patch, am I doing something wrong?

I will postpone this patch until unit test can be verified.

Thanks,
ferruh

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
Ferruh Yigit July 31, 2017, 3:25 p.m. UTC | #4 
On 7/12/2017 11:23 AM, Declan Doherty wrote:
> On 10/07/17 12:13, Herbert Guan wrote:
>> Patch V2: fix build warnings by deleting unused variables.
>>
>> There were double-free problems in some test cases, which will cause
>> a duplicated mbuf will be added into mempool.  After double-free,
>> some new allocated mbuf will hold a same address and thus cause the
>> memory corruption.
>>
>> Another minor issue is that in some test cases, allocated mbuf will
>> not be released after test case exits.  Hopefully these leaked mbuf
>> will be released by the next test case in its setup phase when
>> stopping the virtual pmd ports, while this do is a memory leak of
>> the exited test case.
>>
>> To fix above 2 issues, this patch will do:
>> 1) Release virtual pmd ports' tx queue in the clean up function
>>     remove_slaves_and_stop_bonded_device() of each test cases.
>> 2) Do not release allocated mbufs for test bursts.  These mbufs
>>     will be released in remove_slaves_and_stop_bonded_device() when
>>     test case exits.
>>
>> Signed-off-by: Herbert Guan <herbert.guan@arm.com>
>> ---
> ...
>>
> 
> Acked-by: Declan Doherty <declan.doherty@intel.com>

Applied to dpdk-next-net/master, thanks.
diff mbox

Patch

diff --git a/test/test/test_link_bonding.c b/test/test/test_link_bonding.c
index aa2a1a2..140d864 100644
--- a/test/test/test_link_bonding.c
+++ b/test/test/test_link_bonding.c
@@ -221,6 +221,10 @@  struct rte_fdir_conf fdir_conf = {
 
 };
 
+static void free_virtualpmd_tx_queue(void);
+
+
+
 static int
 configure_ethdev(uint8_t port_id, uint8_t start, uint8_t en_isr)
 {
@@ -684,6 +688,7 @@  struct rte_fdir_conf fdir_conf = {
 remove_slaves_and_stop_bonded_device(void)
 {
 	/* Clean up and remove slaves from bonded device */
+	free_virtualpmd_tx_queue();
 	while (test_params->bonded_slave_count > 0)
 		TEST_ASSERT_SUCCESS(test_remove_slave_from_bonded_device(),
 				"test_remove_slave_from_bonded_device failed");
@@ -1621,9 +1626,6 @@  struct rte_fdir_conf fdir_conf = {
 
 	/* free mbufs */
 	for (i = 0; i < MAX_PKT_BURST; i++) {
-		if (gen_pkt_burst[i] != NULL)
-			rte_pktmbuf_free(gen_pkt_burst[i]);
-
 		if (rx_pkt_burst[i] != NULL)
 			rte_pktmbuf_free(rx_pkt_burst[i]);
 	}
@@ -1970,12 +1972,6 @@  struct rte_fdir_conf fdir_conf = {
 	for (i = 0; i < MAX_PKT_BURST; i++) {
 		if (rx_pkt_burst[i] != NULL)
 			rte_pktmbuf_free(rx_pkt_burst[i]);
-
-		if (gen_pkt_burst[1][i] != NULL)
-			rte_pktmbuf_free(gen_pkt_burst[1][i]);
-
-		if (gen_pkt_burst[3][i] != NULL)
-			rte_pktmbuf_free(gen_pkt_burst[1][i]);
 	}
 
 	/* Clean up and remove slaves from bonded device */
@@ -2414,7 +2410,7 @@  struct rte_fdir_conf fdir_conf = {
 
 	uint8_t slaves[RTE_MAX_ETHPORTS];
 
-	int i, j, burst_size, slave_count, primary_port;
+	int i, burst_size, slave_count, primary_port;
 
 	burst_size = 21;
 
@@ -2547,16 +2543,6 @@  struct rte_fdir_conf fdir_conf = {
 			"(%d) port_stats.opackets not as expected",
 			test_params->slave_port_ids[3]);
 
-	/* free mbufs */
-	for (i = 0; i < TEST_ACTIVE_BACKUP_RX_BURST_SLAVE_COUNT; i++) {
-		for (j = 0; j < MAX_PKT_BURST; j++) {
-			if (pkt_burst[i][j] != NULL) {
-				rte_pktmbuf_free(pkt_burst[i][j]);
-				pkt_burst[i][j] = NULL;
-			}
-		}
-	}
-
 	/* Clean up and remove slaves from bonded device */
 	return remove_slaves_and_stop_bonded_device();
 }
@@ -3318,7 +3304,7 @@  struct rte_fdir_conf fdir_conf = {
 
 	uint8_t slaves[RTE_MAX_ETHPORTS];
 
-	int i, j, burst_size, slave_count;
+	int i, burst_size, slave_count;
 
 	memset(pkt_burst, 0, sizeof(pkt_burst));
 
@@ -3456,16 +3442,6 @@  struct rte_fdir_conf fdir_conf = {
 			test_params->bonded_port_id, (int)port_stats.ipackets,
 			burst_size * 3);
 
-	/* free mbufs allocate for rx testing */
-	for (i = 0; i < TEST_BALANCE_RX_BURST_SLAVE_COUNT; i++) {
-		for (j = 0; j < MAX_PKT_BURST; j++) {
-			if (pkt_burst[i][j] != NULL) {
-				rte_pktmbuf_free(pkt_burst[i][j]);
-				pkt_burst[i][j] = NULL;
-			}
-		}
-	}
-
 	/* Clean up and remove slaves from bonded device */
 	return remove_slaves_and_stop_bonded_device();
 }
@@ -3887,7 +3863,7 @@  struct rte_fdir_conf fdir_conf = {
 
 	uint8_t slaves[RTE_MAX_ETHPORTS];
 
-	int i, j, burst_size, slave_count;
+	int i, burst_size, slave_count;
 
 	memset(pkt_burst, 0, sizeof(pkt_burst));
 
@@ -3984,16 +3960,6 @@  struct rte_fdir_conf fdir_conf = {
 			"(%d) port_stats.ipackets not as expected\n",
 			test_params->bonded_port_id);
 
-	/* free mbufs allocate for rx testing */
-	for (i = 0; i < BROADCAST_LINK_STATUS_NUM_OF_SLAVES; i++) {
-		for (j = 0; j < MAX_PKT_BURST; j++) {
-			if (pkt_burst[i][j] != NULL) {
-				rte_pktmbuf_free(pkt_burst[i][j]);
-				pkt_burst[i][j] = NULL;
-			}
-		}
-	}
-
 	/* Clean up and remove slaves from bonded device */
 	return remove_slaves_and_stop_bonded_device();
 }
@@ -4409,7 +4375,7 @@  struct rte_fdir_conf fdir_conf = {
 
 	uint8_t slaves[RTE_MAX_ETHPORTS];
 
-	int i, j, burst_size, slave_count, primary_port;
+	int i, burst_size, slave_count, primary_port;
 
 	burst_size = 21;
 
@@ -4527,18 +4493,6 @@  struct rte_fdir_conf fdir_conf = {
 			"(%d) port_stats.ipackets not as expected\n",
 			test_params->bonded_port_id);
 
-	/* free mbufs */
-
-	for (i = 0; i < TEST_ADAPTIVE_TRANSMIT_LOAD_BALANCING_RX_BURST_SLAVE_COUNT; i++) {
-		for (j = 0; j < MAX_PKT_BURST; j++) {
-			if (pkt_burst[i][j] != NULL) {
-				rte_pktmbuf_free(pkt_burst[i][j]);
-				pkt_burst[i][j] = NULL;
-			}
-		}
-	}
-
-
 	/* Clean up and remove slaves from bonded device */
 	return remove_slaves_and_stop_bonded_device();
 }