moving sctp to a separate directory ? (Re: svn commit: r201523 - head/sys/netinet)

[Michael Tuexen]

On Jan 4, 2010, at 8:00 PM, Luigi Rizzo wrote:

> taking a random commit to this tree -- would you guys consider moving
> the sctp sources to its own directory, e.g. netinet/sctp/ or something
> that suits better ?
Why do you think that a place different from netinet/ does suit
better for the sctp_*.[ch] files than netinet/ ? This is at least the
place where TCP and UDP live...

Best regards
Michael
> 
> It would help to browse through the directory:
> 
> 	NAME		FILES		LINES
> 
> 	netinet/	156		153k
> 	sctp*		 36		 66k
> 	tcp*		 24		 16k
> 	ipfw/		  9		  9k
> 
> 
> I understand that some of the headers should remain in netinet/
> but surely the code is not bound to a specific place...
> 
> cheers
> luigi
> 
> On Mon, Jan 04, 2010 at 06:25:38PM +0000, Michael Tuexen wrote:
>> Author: tuexen
>> Date: Mon Jan  4 18:25:38 2010
>> New Revision: 201523
>> URL: http://svn.freebsd.org/changeset/base/201523
>> 
>> Log:
>>  Correct usage of parenthesis.
>> 
>>  PR:	kern/142066
>>  Approved by: rrs (mentor)
>>  Obtained from: Henning Petersen, Bruce Cran.
>>  MFC after: 2 weeks
>> 
>> Modified:
>>  head/sys/netinet/sctp_pcb.c
>> 
>> Modified: head/sys/netinet/sctp_pcb.c
>> ==============================================================================
>> --- head/sys/netinet/sctp_pcb.c	Mon Jan  4 18:21:27 2010	(r201522)
>> +++ head/sys/netinet/sctp_pcb.c	Mon Jan  4 18:25:38 2010	(r201523)
>> @@ -5528,7 +5528,7 @@ sctp_pcb_init()
>> 
>> 	/* Init the TIMEWAIT list */
>> 	for (i = 0; i < SCTP_STACK_VTAG_HASH_SIZE; i++) {
>> -		LIST_INIT(&SCTP_BASE_INFO(vtag_timewait[i]));
>> +		LIST_INIT(&SCTP_BASE_INFO(vtag_timewait)[i]);
>> 	}
>> 
>> #if defined(SCTP_USE_THREAD_BASED_ITERATOR)
>> @@ -6385,7 +6385,7 @@ sctp_is_vtag_good(struct sctp_inpcb *inp
>> 	}
>> skip_vtag_check:
>> 
>> -	chain = &SCTP_BASE_INFO(vtag_timewait[(tag % SCTP_STACK_VTAG_HASH_SIZE))];
>> +	chain = &SCTP_BASE_INFO(vtag_timewait)[(tag % SCTP_STACK_VTAG_HASH_SIZE)];
>> 	/* Now what about timed wait ? */
>> 	if (!LIST_EMPTY(chain)) {
>> 		/*
>