* Query
@ 2010-01-11 14:12 RahulSrivastava 71616
2010-01-11 17:05 ` Query Vlad Yasevich
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: RahulSrivastava 71616 @ 2010-01-11 14:12 UTC (permalink / raw)
To: linux-sctp
Hi All,
I have a query. When using Linux SCTP one-to-many style sockets sendmsg succedes even though message is not delivered at the peer. When I ask the socket to wait for notification it gave send failed in notification.
Is this possible for sendmsg to fail for even for one-to-many style socket (for peer errors)?
Regards,
Rahul
******************************************************************************************
This email and its attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained here in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this email in error, please notify the sender by phone or email immediately and delete it!
*****************************************************************************************
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Query
2010-01-11 14:12 Query RahulSrivastava 71616
@ 2010-01-11 17:05 ` Vlad Yasevich
2010-01-12 4:56 ` Query RahulSrivastava 71616
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Vlad Yasevich @ 2010-01-11 17:05 UTC (permalink / raw)
To: linux-sctp
RahulSrivastava 71616 wrote:
> Hi All,
> I have a query. When using Linux SCTP one-to-many style sockets sendmsg succedes even though message is not delivered at the peer. When I ask the socket to wait for notification it gave send failed in notification.
> Is this possible for sendmsg to fail for even for one-to-many style socket (for peer errors)?
>
Yes. sendmsg() will return success when the data has successfully been accepted
by the kernel. However, if peer has problems, you may get a SEND_FAILED
notification.
This is not linux SCTP specific. TCP can behave the same way (minus the
notification).
-vlad
> Regards,
> Rahul
>
>
>
> ******************************************************************************************
> This email and its attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained here in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this email in error, please notify the sender by phone or email immediately and delete it!
> *****************************************************************************************
> --
> To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Query
2010-01-11 14:12 Query RahulSrivastava 71616
2010-01-11 17:05 ` Query Vlad Yasevich
@ 2010-01-12 4:56 ` RahulSrivastava 71616
2010-01-12 20:52 ` Query Vlad Yasevich
2011-04-05 6:30 ` Query RahulSrivastava 71616
3 siblings, 0 replies; 5+ messages in thread
From: RahulSrivastava 71616 @ 2010-01-12 4:56 UTC (permalink / raw)
To: linux-sctp
Thanks!
If sendmsg succedes like this then how its possible to maintain ordering of messages:
Consider below order:
send1 ----> send2
send1 was not recieved but send2 was successfully recived by peer on same association. Application gets a notification of send1 (indicating send failed).
Regards,
Rahul
******************************************************************************************
This email and its attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained here in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this email in error, please notify the sender by phone or email immediately and delete it!
*****************************************************************************************
----- Original Message -----
From: Vlad Yasevich <vladislav.yasevich@hp.com>
Date: Monday, January 11, 2010 10:35 pm
Subject: Re: Query
To: RahulSrivastava 71616 <rahuls@huawei.com>
Cc: linux-sctp@vger.kernel.org
>
>
> RahulSrivastava 71616 wrote:
> > Hi All,
> > I have a query. When using Linux SCTP one-to-many style
> sockets sendmsg succedes even though message is not delivered at
> the peer. When I ask the socket to wait for notification it gave
> send failed in notification.
> > Is this possible for sendmsg to fail for even for one-to-many
> style socket (for peer errors)?
> >
>
> Yes. sendmsg() will return success when the data has successfully
> been accepted
> by the kernel. However, if peer has problems, you may get a
> SEND_FAILEDnotification.
>
> This is not linux SCTP specific. TCP can behave the same way
> (minus the
> notification).
>
> -vlad
>
> > Regards,
> > Rahul
> >
> >
> >
> >
> ******************************************************************************************> This email and its attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained here in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this email in error, please notify the sender by phone or email immediately and delete it!
> >
> *****************************************************************************************> --
> > To unsubscribe from this list: send the line "unsubscribe linux-
> sctp" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
> >
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Query
2010-01-11 14:12 Query RahulSrivastava 71616
2010-01-11 17:05 ` Query Vlad Yasevich
2010-01-12 4:56 ` Query RahulSrivastava 71616
@ 2010-01-12 20:52 ` Vlad Yasevich
2011-04-05 6:30 ` Query RahulSrivastava 71616
3 siblings, 0 replies; 5+ messages in thread
From: Vlad Yasevich @ 2010-01-12 20:52 UTC (permalink / raw)
To: linux-sctp
RahulSrivastava 71616 wrote:
> Thanks!
> If sendmsg succedes like this then how its possible to maintain ordering of messages:
> Consider below order:
> send1 ----> send2
>
> send1 was not recieved but send2 was successfully recived by peer on same association. Application gets a notification of send1 (indicating send failed).
>
No, that will not happen on the same association unless partial reliability is
used.
In the normal scenario (full reliability), if send1 fails, then the association
is destroyed. The SEND_FAILED event is typically generated in the case of
excessive retransmissions that cause the association to terminate. There are
a few other times it can be generated, but all those times have to do with
setting a time_to_live on the message.
So, if you are using implicit connect (i.e used the sendto/sendmsg() call
to establish an association), then the steps are these:
1) queue the message to the kernel queue.
2) start association procedure.
a) if association is successful, send the message.
b) if association fails, notify the user to association and send failure.
If if you already have an association, the protocol will attempt to deliver
all messages in order. If the message has already been queued to the kernel,
and the association terminates due to retransmissions, you will get an
association termination notification and a send failed notification.
Any subsequent send calls by the applications will attempt to establish a
new association.
-vlad
>
>
> Regards,
> Rahul
>
>
> ******************************************************************************************
> This email and its attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained here in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this email in error, please notify the sender by phone or email immediately and delete it!
> *****************************************************************************************
>
> ----- Original Message -----
> From: Vlad Yasevich <vladislav.yasevich@hp.com>
> Date: Monday, January 11, 2010 10:35 pm
> Subject: Re: Query
> To: RahulSrivastava 71616 <rahuls@huawei.com>
> Cc: linux-sctp@vger.kernel.org
>
>>
>> RahulSrivastava 71616 wrote:
>>> Hi All,
>>> I have a query. When using Linux SCTP one-to-many style
>> sockets sendmsg succedes even though message is not delivered at
>> the peer. When I ask the socket to wait for notification it gave
>> send failed in notification.
>>> Is this possible for sendmsg to fail for even for one-to-many
>> style socket (for peer errors)?
>> Yes. sendmsg() will return success when the data has successfully
>> been accepted
>> by the kernel. However, if peer has problems, you may get a
>> SEND_FAILEDnotification.
>>
>> This is not linux SCTP specific. TCP can behave the same way
>> (minus the
>> notification).
>>
>> -vlad
>>
>>> Regards,
>>> Rahul
>>>
>>>
>>>
>>>
>> ******************************************************************************************> This email and its attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained here in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this email in error, please notify the sender by phone or email immediately and delete it!
>>>
>> *****************************************************************************************> --
>>> To unsubscribe from this list: send the line "unsubscribe linux-
>> sctp" in
>>> the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>>>
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Query
2010-01-11 14:12 Query RahulSrivastava 71616
` (2 preceding siblings ...)
2010-01-12 20:52 ` Query Vlad Yasevich
@ 2011-04-05 6:30 ` RahulSrivastava 71616
3 siblings, 0 replies; 5+ messages in thread
From: RahulSrivastava 71616 @ 2011-04-05 6:30 UTC (permalink / raw)
To: linux-sctp
Hi All,
I am running a test application and I am finding lksctp performance is at least 25% of UDP on Linux.
Specifically my profiler tells sctp_sendmsg is at least 5 times slower than UDP sendto.
Can anybody provide pointers on this?
Regards,
Rahul
******************************************************************************************
This email and its attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained here in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this email in error, please notify the sender by phone or email immediately and delete it!
*****************************************************************************************
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2011-04-05 6:30 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-01-11 14:12 Query RahulSrivastava 71616
2010-01-11 17:05 ` Query Vlad Yasevich
2010-01-12 4:56 ` Query RahulSrivastava 71616
2010-01-12 20:52 ` Query Vlad Yasevich
2011-04-05 6:30 ` Query RahulSrivastava 71616
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).