Re: [LTP] [PATCH] send*: fix errno of invalid flags

[Wanlong Gao <gaowanlong@cn.fujitsu.com>]

On 11/29/2013 05:10 PM, Jan Stancek wrote:
> 
> 
> ----- Original Message -----
>> From: "Wanlong Gao" <gaowanlong@cn.fujitsu.com>
>> To: "LTP" <ltp-list@lists.sourceforge.net>
>> Sent: Thursday, 28 November, 2013 10:13:08 AM
>> Subject: [LTP] [PATCH] send*: fix errno of invalid flags
>>
>> After following kernel commit, the tcp_fastopen
>> sysctl is enabled default, so the errno will
>> no be ENOTSUP any more.
>>
>> Actually, the flags of send* syscall is "unsigned int",
>> then our test case passed "-1" to it will not be treated
>> as an invalid flags.
>>
>> commit 0d41cca490c274352211efac50e9598d39a9dc80
>> Author: Yuchung Cheng <ycheng@google.com>
>> Date:   Thu Oct 31 09:19:32 2013 -0700
>>
>>     tcp: enable sockets to use MSG_FASTOPEN by default
>>
>>     Applications have started to use Fast Open (e.g., Chrome browser has
>>     such an optional flag) and the feature has gone through several
>>     generations of kernels since 3.7 with many real network tests. It's
>>     time to enable this flag by default for applications to test more
>>     conveniently and extensively.
>>
>>     Signed-off-by: Yuchung Cheng <ycheng@google.com>
>>     Signed-off-by: Neal Cardwell <ncardwell@google.com>
>>     Acked-by: Eric Dumazet <edumazet@google.com>
>>     Signed-off-by: David S. Miller <davem@davemloft.net>
>>
>> Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
>> ---
>>  testcases/kernel/syscalls/send/send01.c       | 6 ++++--
>>  testcases/kernel/syscalls/sendmsg/sendmsg01.c | 6 +++++-
>>  testcases/kernel/syscalls/sendto/sendto01.c   | 5 ++++-
>>  3 files changed, 13 insertions(+), 4 deletions(-)
> 
> Hi,
> 
> my guess is that intention of testcase "invalid flags set"
> was to get EOPNOTSUPP out of it. But this never worked for older kernels.
> And on new kernels this is still likely going to depend on "sysctl_tcp_fastopen".


Agreed, this is what I mean. Could you please make a patch to fix send* like this?


Thanks,
Wanlong Gao


> 
> I'm running kernel < 3.6 and if I apply patch for send01.c
> it fails with:
>   send01      6  TFAIL  :  call succeeded unexpectedly
> 
> sendto(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 1024, MSG_OOB|MSG_DONTROUTE|MSG_PEEK|MSG_CTRUNC|MSG_PROXY|MSG_EOR|MSG_WAITALL|MSG_TRUNC|MSG_ERRQUEUE|MSG_DONTWAIT|MSG_CONFIRM|MSG_FIN|MSG_SYN|MSG_RST|MSG_NOSIGNAL|MSG_MORE|MSG_CMSG_CLOEXEC|0xbfff0000, NULL, 0) = 1024
> 
> I think we should get rid of that -1 in flags and use concrete flags we know
> to be invalid. I recall suggesting SOCK_DGRAM + MSG_OOB last time this test needed update:
> 
> diff --git a/testcases/kernel/syscalls/send/send01.c b/testcases/kernel/syscalls/send/send01.c
> index d228422..6ebe131 100644
> --- a/testcases/kernel/syscalls/send/send01.c
> +++ b/testcases/kernel/syscalls/send/send01.c
> @@ -72,7 +72,6 @@ static void setup(void);
>  static void setup0(void);
>  static void setup1(void);
>  static void setup2(void);
> -static void setup3(void);
>  static void cleanup0(void);
>  static void cleanup1(void);
>  
> @@ -143,14 +142,14 @@ static struct test_case_t tdat[] = {
>  #ifndef UCLINUX
>         /* Skip since uClinux does not implement memory protection */
>         {.domain = PF_INET,
> -        .type = SOCK_STREAM,
> +        .type = SOCK_DGRAM,
>          .proto = 0,
> -        .buf = (void *)-1,
> +        .buf = buf,
>          .buflen = sizeof(buf),
> -        .flags = -1,
> +        .flags = MSG_OOB,
>          .retval = -1,
> -        .experrno = EFAULT,
> -        .setup = setup3,
> +        .experrno = EOPNOTSUPP,
> +        .setup = setup1,
>          .cleanup = cleanup1,
>          .desc = "invalid flags set"}
>  #endif
> @@ -352,11 +351,3 @@ static void setup2(void)
>                 tst_brkm(TBROK | TERRNO, cleanup, "socket setup failed connect "
>                          "test %d", testno);
>  }
> -
> -static void setup3(void)
> -{
> -       setup1();
> -
> -       if (tst_kvercmp(3, 6, 0) >= 0)
> -               tdat[testno].experrno = ENOTSUP;
> -}
> 
> Regards,
> Jan
> 
> 
>>
>> diff --git a/testcases/kernel/syscalls/send/send01.c
>> b/testcases/kernel/syscalls/send/send01.c
>> index d228422..0d91e92 100644
>> --- a/testcases/kernel/syscalls/send/send01.c
>> +++ b/testcases/kernel/syscalls/send/send01.c
>> @@ -145,7 +145,7 @@ static struct test_case_t tdat[] = {
>>  	{.domain = PF_INET,
>>  	 .type = SOCK_STREAM,
>>  	 .proto = 0,
>> -	 .buf = (void *)-1,
>> +	 .buf = buf,
>>  	 .buflen = sizeof(buf),
>>  	 .flags = -1,
>>  	 .retval = -1,
>> @@ -357,6 +357,8 @@ static void setup3(void)
>>  {
>>  	setup1();
>>  
>> -	if (tst_kvercmp(3, 6, 0) >= 0)
>> +	if (tst_kvercmp(3, 12, 0) > 0)
>> +		tdat[testno].experrno = EINVAL;
>> +	else if (tst_kvercmp(3, 6, 0) >= 0)
>>  		tdat[testno].experrno = ENOTSUP;
>>  }
>> diff --git a/testcases/kernel/syscalls/sendmsg/sendmsg01.c
>> b/testcases/kernel/syscalls/sendmsg/sendmsg01.c
>> index 90fbb5d..890834c 100644
>> --- a/testcases/kernel/syscalls/sendmsg/sendmsg01.c
>> +++ b/testcases/kernel/syscalls/sendmsg/sendmsg01.c
>> @@ -730,8 +730,12 @@ static void setup7(void)
>>  {
>>  	setup1();
>>  
>> -	if (tst_kvercmp(3, 6, 0) >= 0)
>> +	if (tst_kvercmp(3, 12, 0) > 0) {
>>  		tdat[testno].retval = -1;
>> +		tdat[testno].experrno = EISCONN;
>> +	} else if (tst_kvercmp(3, 6, 0) >= 0) {
>> +		tdat[testno].retval = -1;
>> +	}
>>  }
>>  
>>  static void setup8(void)
>> diff --git a/testcases/kernel/syscalls/sendto/sendto01.c
>> b/testcases/kernel/syscalls/sendto/sendto01.c
>> index a571afd..e7c8805 100644
>> --- a/testcases/kernel/syscalls/sendto/sendto01.c
>> +++ b/testcases/kernel/syscalls/sendto/sendto01.c
>> @@ -435,7 +435,10 @@ static void setup4(void)
>>  {
>>  	setup1();
>>  
>> -	if (tst_kvercmp(3, 6, 0) >= 0) {
>> +	if (tst_kvercmp(3, 12, 0) > 0) {
>> +		tdat[testno].retval = -1;
>> +		tdat[testno].experrno = EISCONN;
>> +	} else if (tst_kvercmp(3, 6, 0) >= 0) {
>>  		tdat[testno].retval = -1;
>>  		tdat[testno].experrno = ENOTSUP;
>>  	}
>> --
>> 1.8.5.rc3
>>
>>
>> ------------------------------------------------------------------------------
>> Rapidly troubleshoot problems before they affect your business. Most IT
>> organizations don't have a clear picture of how application performance
>> affects their revenue. With AppDynamics, you get 100% visibility into your
>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics
>> Pro!
>> http://pubads.g.doubleclick.net/gampad/clk?id=84349351&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Ltp-list mailing list
>> Ltp-list@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/ltp-list
>>
> 


------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349351&iu=/4140/ostg.clktrk
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list