Chronological Thread 

< Chronological >      < Thread >

---

Thanks for your reply, Michael!

An array index reaches the same goal, but I am not really happy about this:

1) Is it one array for every instance? Then SCTP hase the problem to identfy
the instance.

2) Is it one array for all instances? Then the ULP has a problem with its
array size...

3) The maximum number of associations must be fixed, or the array must be
expandable. (is this easy to realize?)

4) A check for a valid pointer seems easy to me:
- e.g. the pointer itself is stored in the assoc data and this can be
checked.
- Or (that's my favorite!) in the API-functions both pointers are handed
over, so everybody can check it and the caller identifies himself (may be
helpful for debugging).

Viele Grüße
Wolfgang Schrüfer

**************************************
Wolfgang Schrüfer SIEMENS AG
ICN WN CC EKJ 14 TEL: 722-61646
mailto:Wolfgang.Schruefer AT icn.siemens.de
**************************************

> -----Ursprüngliche Nachricht-----
> Von: Tuexen Michael
> [SMTP:Michael.Tuexen AT icn.siemens.de]
> Gesendet am: Mittwoch, 14. Februar 2001 13:02
> An:
> 'discussion AT sctp.de'
> Betreff: RE: [SCTP-Discussion] SCTP-API: Pointer instead of
> AssociationID
>
> Seems reasonable for me, but if the ULP uses a wrong pointer
> you probabily modify some important part of your memory ...
> What about having an array of the assoc data (or at least an
> array of pointers to these data) and use an index. Than you
> can do at least some array index bounds checking.
> I think that is a good thing for an API.
> Happy SCTPing
> Michael
>
> Michael Tüxen Tel.: +49 89 722 47210
> Siemens AG Fax: +49 89 722 48212
> ICN WN CS SE 51 E-mail:
> Michael.Tuexen AT icn.siemens.de
>
> > -----Original Message-----
> > From: Schruefer Wolfgang
> > [SMTP:Wolfgang.Schruefer AT icn.siemens.de]
> > Sent: Wednesday, February 14, 2001 11:12 AM
> > To:
> > 'discussion AT sctp.de'
> > Cc: Raisch Juergen
> > Subject: [SCTP-Discussion] SCTP-API: Pointer instead of AssociationID
> >
> > Dear all,
> >
> > we have a future requirement to handle 12000! associations at one machine.
> >
> > So I thought about the search algorithm to get the data related to a
> > single association, which is necessary for the most function calls
> > between ULP and SCTP.
> >
> > But at the interface SCTP <> ULP any searching could be avoided if there
> > is a possibility to define an association via address.
> >
> >
> > Idea:
> > The function sctp_associate returns the address of the
> > SCTP-association-data-structure.
> > A new parameter announces the ULP-association-data-structure to sctp,
> > e.g.:
> >
> > void *sctp_associate ( ...
> > ..., void *ULP_association)
> >
> > For all API-functions the associationID can be replaced by the pointer to
> > the association of the called layer.
> >
> >
> > Would this be a way to go for?
> >
> > Viele Grüße
> > Wolfgang Schrüfer
> >
> > **************************************
> > Wolfgang Schrüfer SIEMENS AG
> > ICN WN CC EKJ 14 TEL: 722-61646
> > mailto:Wolfgang.Schruefer AT icn.siemens.de
> > **************************************
> >
> > _______________________________________________
> > discussion mailing list
> > discussion AT sctp.de
> > http://www.sctp.de/mailman/listinfo/discussion
> _______________________________________________
> discussion mailing list
> discussion AT sctp.de
> http://www.sctp.de/mailman/listinfo/discussion

---

• AW: [SCTP-Discussion] SCTP-API: Pointer instead of AssociationID, Schruefer Wolfgang, 02/14/2001

---

Archive powered by MHonArc 2.6.19+.