cvs commit: src/sys/netinet sctp.h sctp_asconf.c sctp_asconf.h sctp_bsd_addr.c sctp_constants.h sctp_indata.c sctp_input.c sctp_lock_bsd.h sctp_os_bsd.h sctp_output.c sctp_pcb.c sctp_pcb.h sctp_peeloff.c sctp_sysctl.c sctp_sysctl.h sctp_timer.c sctp_uio.h ...

[John Baldwin]

On Thursday 14 June 2007 10:33:58 pm Randall Stewart wrote:
> Kris Kennaway wrote:
> > On Thu, Jun 14, 2007 at 10:25:15PM -0400, Randall Stewart wrote:
> > 
> >>Kris Kennaway wrote:
> >>
> >>>On Thu, Jun 14, 2007 at 10:59:04PM +0000, Randall Stewart wrote:
> >>>
> >>>
> >>>>rrs         2007-06-14 22:59:04 UTC
> >>>>
> >>>>FreeBSD src repository
> >>>>
> >>>>Modified files:
> >>>>  sys/netinet          sctp.h sctp_asconf.c sctp_asconf.h 
> >>>>                       sctp_bsd_addr.c sctp_constants.h 
> >>>>                       sctp_indata.c sctp_input.c 
> >>>>                       sctp_lock_bsd.h sctp_os_bsd.h 
> >>>>                       sctp_output.c sctp_pcb.c sctp_pcb.h 
> >>>>                       sctp_peeloff.c sctp_sysctl.c 
> >>>>                       sctp_sysctl.h sctp_timer.c sctp_uio.h 
> >>>>                       sctp_usrreq.c sctputil.c sctputil.h 
> >>>>  sys/netinet6         sctp6_usrreq.c 
> >>>>  sys/conf             options 
> >>>>Log:
> >>>>- Fix so ifn's are properly deleted when the ref count goes to 0.
> >>>>- Fix so VRF's will clean themselves up when no references are around.
> >>>>- Allow sctp_ifa to be passed into inpcb_bind, addr_mgmt_ep_sa to bypass
> >>>>  normal validation checks.
> >>>>- turn auto-asconf off for subset bound sockets
> >>>>- Moves all logging to use KTR. This gets rid of most
> >>>>  of the logging #ifdef's with a few exceptions reducing
> >>>>  the number of config options for SCTP.
> >>>
> >>>
> >>>| +#ifndef SCTP_SUBSYS_KTR
> >>>| +#define SCTP_SUBSYS_KTR KTR_GEN
> >>>| +#endif
> >>>
> >>>Brief silence after previous disapproval doesn't equal approval ;-)
> >>>
> >>>What was wrong with the method I suggested, namely using KTR_SUBSYS if
> >>>a SCTP_TRACE option is included in the kernel?  That is the intended
> >>>way that events local to a particular subsystem should be handled.
> >>>
> >>>Kris
> >>
> >>I asked if KTR_GEN was ok.. I can use KTR_SUBSYS.. sure.. but that
> >>means I can't really run witness on my machine as I test.. since
> >>witness is the only one that uses KTR_SUBSYS.. No one else uses
> >>KTR_GEN.. why is it a problem using one.
> > 
> > 
> > WITNESS does not use itq:
> > 
> > #if 0
> > #define KTR_WITNESS     KTR_SUBSYS
> > #else
> > #define KTR_WITNESS     0
> > #endif
> > 
> > KTR_WITNESS is used for debugging of the WITNESS code itself (probably
> > no-one has needed to use it for many years).  It is a completely
> > orthogonal purpose to yours, so they can co-exist happily.
> > 
> > KTR_GEN is for "general events".  SCTP is not a general event, it's
> > localized events belonging to a particular subsystem.
> > 
> > 
> >>So whats wrong with using a unused one?
> > 
> > 
> > If you look at the history there was a big effort to reclaim these
> > fields.  Because space is at such a premium here there needs to be a
> > strong reason for using up a spare field.  SCTP does not strike me as
> > such a reason, particularly since the alternative seems quite
> > practical.
> > 
> > Kris
> Ok.. whatever you say.. I can change it very easily ;-)

Also, I would use KTR_SCTP and use the same #ifdef above so it is 0 or 
KTR_SUBSYS depending on whether or not you want it enabled.

-- 
John Baldwin