diff mbox series

sctp: bugfix for utils/sctp/func_tests/test_1_to_1_events.c

Message ID 20240329154715.7130-1-yangfeng59949@163.com
State Accepted
Headers show
Series sctp: bugfix for utils/sctp/func_tests/test_1_to_1_events.c | expand

Commit Message

yangfeng59949 March 29, 2024, 3:47 p.m. UTC 
From: yangfeng <yangfeng@kylinos.cn>

event not initialized to 0 causes "Got a notification,
expecting a datamsg" issues

Signed-off-by: yangfeng <yangfeng@kylinos.cn>
---
 utils/sctp/func_tests/test_1_to_1_events.c | 1 +
 1 file changed, 1 insertion(+)

--
2.25.1

Comments

Petr Vorel April 4, 2024, 1:53 p.m. UTC | #1 
Hi,

> event not initialized to 0 causes "Got a notification,
> expecting a datamsg" issues

Out of curiosity which architecture, kernel version, toolchain version and
distro do you see the issue? I cannot reproduce it on any of systems we test.

...
> +++ b/utils/sctp/func_tests/test_1_to_1_events.c
> @@ -96,6 +96,7 @@ main(void)
>  	/* Create the client socket.  */
>  	clt_sk = test_socket(AF_INET, SOCK_STREAM, IPPROTO_SCTP);

> +	memset(&event, 0, sizeof(struct sctp_event_subscribe));

Obviously correct.

Reviewed-by: Petr Vorel <pvorel@suse.cz>

Note, nowadays it'd be enough to initialize as:

struct sctp_event_subscribe event = {0};

But because we use the old style with the rest of the functions, memset() is OK.

Kind regards,
Petr
yangfeng59949 April 5, 2024, 6:53 a.m. UTC | #2 
Hi,<br/>Thanks for getting back to me.<br/><br/>gcc (GCC) 7.3.0<br/>Linux localhost.localdomain 4.19.90-89.6.v2401.ky10.aarch64 #1 SMP Thu Mar 14 20:06:10 CST 2024 aarch64 aarch64 aarch64 GNU/Linux<br/>Manufacturers ID：                   HiSilicon<br/>BIOS Vendor ID:                        QEMU<br/>Model name：                          Kunpeng-920<br/>BIOS Model name:                    virt-4.1<br/><br/>Kind regards,<br/>yangfeng
At 2024-04-04 21:53:23, "Petr Vorel" <pvorel@suse.cz> wrote:
>Hi,
>
>> event not initialized to 0 causes "Got a notification,
>> expecting a datamsg" issues
>
>Out of curiosity which architecture, kernel version, toolchain version and
>distro do you see the issue? I cannot reproduce it on any of systems we test.
>
>...
>> +++ b/utils/sctp/func_tests/test_1_to_1_events.c
>> @@ -96,6 +96,7 @@ main(void)
>>  	/* Create the client socket.  */
>>  	clt_sk = test_socket(AF_INET, SOCK_STREAM, IPPROTO_SCTP);
>
>> +	memset(&event, 0, sizeof(struct sctp_event_subscribe));
>
>Obviously correct.
>
>Reviewed-by: Petr Vorel <pvorel@suse.cz>
>
>Note, nowadays it'd be enough to initialize as:
>
>struct sctp_event_subscribe event = {0};
>
>But because we use the old style with the rest of the functions, memset() is OK.
>
>Kind regards,
>Petr
Petr Vorel April 5, 2024, 7:34 a.m. UTC | #3 
Hi,

> Hi,<br/>Thanks for getting back to me.<br/><br/>gcc (GCC) 7.3.0<br/>Linux localhost.localdomain 4.19.90-89.6.v2401.ky10.aarch64 #1 SMP Thu Mar 14 20:06:10 CST 2024 aarch64 aarch64 aarch64 GNU/Linux<br/>Manufacturers ID：                   HiSilicon<br/>BIOS Vendor ID:                        QEMU<br/>Model name：                          Kunpeng-920<br/>BIOS Model name:                    virt-4.1<br/><br/>Kind regards,<br/>yangfeng

Thanks for info, merged.

Kind regards,
Petr
diff mbox series

Patch

diff --git a/utils/sctp/func_tests/test_1_to_1_events.c b/utils/sctp/func_tests/test_1_to_1_events.c
index 447845ff3..889d2ff33 100644
--- a/utils/sctp/func_tests/test_1_to_1_events.c
+++ b/utils/sctp/func_tests/test_1_to_1_events.c
@@ -96,6 +96,7 @@  main(void)
 	/* Create the client socket.  */
 	clt_sk = test_socket(AF_INET, SOCK_STREAM, IPPROTO_SCTP);

+	memset(&event, 0, sizeof(struct sctp_event_subscribe));
 	event.sctp_data_io_event = 1;
 	event.sctp_association_event = 1;
 	event.sctp_shutdown_event = 1;