Re: [LTP] [PATCH]utils/sctp: bugfix for testlib/sctputil.h

Re: [LTP] [PATCH]utils/sctp: bugfix for testlib/sctputil.h

[Petr Vorel]

Hi Li,

> The socket is created and bound immediately without waiting for the handshake after close(sk).
> "bind(): errno=EADDRINUSE(98): Address already in use" may be reported.
> Use SO_REUSEPORT to allow multiple sockets to be bound to the same port.

How this can happen? Running tests too quickly one after the other?
Or what is different on your SUT? I'm not sure if it's good idea to always test
with SO_REUSEADDR and SO_REUSEPORT.

> Signed-off-by: Min Li <limin154@huawei.com>
> ---
>  utils/sctp/testlib/sctputil.h | 9 +++++++++
>  1 file changed, 9 insertions(+)

> diff --git a/utils/sctp/testlib/sctputil.h b/utils/sctp/testlib/sctputil.h
> index 176d623f0..b51a3f9b0 100644
> --- a/utils/sctp/testlib/sctputil.h
> +++ b/utils/sctp/testlib/sctputil.h
> @@ -133,6 +133,15 @@ extern int TST_CNT;
>  static inline int test_socket(int domain, int type, int protocol)
>  {
>  	int sk = socket(domain, type, protocol);
> +	int true_const=1;
What is this unused variable for?

> +
> +	if(setsockopt(sk, SOL_SOCKET, SO_REUSEADDR, &true_const, sizeof(int))){
> +		printf("set addr err\n");
Besides missing spaces in if (if (..) { ), we have SAFE_SETSOCKOPT().
Or tst_brkm(TWARN, ...) could be used (we don't use printf directly in LTP).

Kind regards,
Petr

> +	}
> +
> +	if(setsockopt(sk, SOL_SOCKET, SO_REUSEPORT, &true_const, sizeof(int))){
> +		printf("set port err\n");
> +	}

>  	if (sk == -1) {
>  		if (errno == EAFNOSUPPORT)

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

[LTP] [PATCH]utils/sctp: bugfix for testlib/sctputil.h

[Li Min via ltp]

The socket is created and bound immediately without waiting for the handshake after close(sk).
"bind(): errno=EADDRINUSE(98): Address already in use" may be reported.
Use SO_REUSEPORT to allow multiple sockets to be bound to the same port.

Signed-off-by: Min Li <limin154@huawei.com>
---
 utils/sctp/testlib/sctputil.h | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/utils/sctp/testlib/sctputil.h b/utils/sctp/testlib/sctputil.h
index 176d623f0..b51a3f9b0 100644
--- a/utils/sctp/testlib/sctputil.h
+++ b/utils/sctp/testlib/sctputil.h
@@ -133,6 +133,15 @@ extern int TST_CNT;
 static inline int test_socket(int domain, int type, int protocol)
 {
 	int sk = socket(domain, type, protocol);
+	int true_const=1;
+
+	if(setsockopt(sk, SOL_SOCKET, SO_REUSEADDR, &true_const, sizeof(int))){
+		printf("set addr err\n");
+	}
+
+	if(setsockopt(sk, SOL_SOCKET, SO_REUSEPORT, &true_const, sizeof(int))){
+		printf("set port err\n");
+	}

 	if (sk == -1) {
 		if (errno == EAFNOSUPPORT)
--
2.33.0


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH]utils/sctp: bugfix for testlib/sctputil.h

[Richard Palethorpe]

Hello,

Petr Vorel <pvorel@suse.cz> writes:

> Hi Li,
>
>> The socket is created and bound immediately without waiting for the handshake after close(sk).
>> "bind(): errno=EADDRINUSE(98): Address already in use" may be reported.
>> Use SO_REUSEPORT to allow multiple sockets to be bound to the same port.
>
> How this can happen? Running tests too quickly one after the other?
> Or what is different on your SUT? I'm not sure if it's good idea to always test
> with SO_REUSEADDR and SO_REUSEPORT.
>
>> Signed-off-by: Min Li <limin154@huawei.com>
>> ---
>>  utils/sctp/testlib/sctputil.h | 9 +++++++++
>>  1 file changed, 9 insertions(+)
>
>> diff --git a/utils/sctp/testlib/sctputil.h b/utils/sctp/testlib/sctputil.h
>> index 176d623f0..b51a3f9b0 100644
>> --- a/utils/sctp/testlib/sctputil.h
>> +++ b/utils/sctp/testlib/sctputil.h
>> @@ -133,6 +133,15 @@ extern int TST_CNT;
>>  static inline int test_socket(int domain, int type, int protocol)
>>  {
>>  	int sk = socket(domain, type, protocol);
>> +	int true_const=1;
> What is this unused variable for?

It's used as the socket option value.

>
>> +
>> +	if(setsockopt(sk, SOL_SOCKET, SO_REUSEADDR, &true_const, sizeof(int))){
>> +		printf("set addr err\n");
> Besides missing spaces in if (if (..) { ), we have SAFE_SETSOCKOPT().
> Or tst_brkm(TWARN, ...) could be used (we don't use printf directly in
> LTP).

+1 for SAFE_SETSOCKOPT

>
> Kind regards,
> Petr
>
>> +	}
>> +
>> +	if(setsockopt(sk, SOL_SOCKET, SO_REUSEPORT, &true_const, sizeof(int))){
>> +		printf("set port err\n");
>> +	}

We don't need to set it twice.

Setting patch as changes requested.

>
>>  	if (sk == -1) {
>>  		if (errno == EAFNOSUPPORT)


-- 
Thank you,
Richard.

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp