d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
18e6466b0f
freebsd-nq/sys/netatalk/aarp.c

692 lines
18 KiB
C
Raw Normal View History

/* -> /*- for license, minor formatting changes, insert COPYRIGHT into files
2005-01-07 02:35:34 +00:00
/*-
Convert the aa_ifaddr timeout to a callout, and run the aarprobe callout MPSAFE. Acquire the aarptab_mtx to make sure that the callout and msleep in the ioctl thread don't race. MFC after: 1 week
2005-02-22 14:20:29 +00:00
* Copyright (c) 2004-2005 Robert N. M. Watson
Inline umich license from COPYRIGHT to make it clear what license the umich copyright is asserting. Clarify that the copyright I'm asserting is the standard Berkeley license. Remove Giant assertions from AARP and DDP input routines.
2004-08-10 03:23:05 +00:00
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
/* -> /*- for license, minor formatting changes, insert COPYRIGHT into files
2005-01-07 02:35:34 +00:00
/*-
Inline umich license from COPYRIGHT to make it clear what license the umich copyright is asserting. Clarify that the copyright I'm asserting is the standard Berkeley license. Remove Giant assertions from AARP and DDP input routines.
2004-08-10 03:23:05 +00:00
* Copyright (c) 1990,1991,1994 Regents of The University of Michigan.
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
* All Rights Reserved.
Slight cleanup of aarp code Submitted by: Ian Dowse <iedowse@maths.tcd.ie> PR: kern/17583 This PR is still open as Ther is still ongoing investigation.
2000-03-29 17:12:09 +00:00
*
Inline umich license from COPYRIGHT to make it clear what license the umich copyright is asserting. Clarify that the copyright I'm asserting is the standard Berkeley license. Remove Giant assertions from AARP and DDP input routines.
2004-08-10 03:23:05 +00:00
* Permission to use, copy, modify, and distribute this software and
* its documentation for any purpose and without fee is hereby granted,
* provided that the above copyright notice appears in all copies and
* that both that copyright notice and this permission notice appear
* in supporting documentation, and that the name of The University
* of Michigan not be used in advertising or publicity pertaining to
* distribution of the software without specific, written prior
* permission. This software is supplied as is without expressed or
* implied warranties of any kind.
*
* This product includes software developed by the University of
* California, Berkeley and its contributors.
*
* Research Systems Unix Group
* The University of Michigan
* c/o Wesley Craig
* 535 W. William Street
* Ann Arbor, Michigan
* +1-313-764-2278
* netatalk@umich.edu
*
Slight cleanup of aarp code Submitted by: Ian Dowse <iedowse@maths.tcd.ie> PR: kern/17583 This PR is still open as Ther is still ongoing investigation.
2000-03-29 17:12:09 +00:00
* $FreeBSD$
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
*/
Make the debug options new-style. This also zaps a DPT option from lint; it wasn't referenced from anywhere.
1998-01-31 07:23:16 +00:00
#include "opt_atalk.h"
Introduce experimental support for MAC in the AppleTalk/EtherTalk stack. Label link layer mbufs as they are created for transmission, check mbufs before delivering them to sockets, label mbufs as they are created from sockets, and preserve mbuf labels if mbufs are copied. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-15 18:58:44 +00:00
#include "opt_mac.h"
Make the debug options new-style. This also zaps a DPT option from lint; it wasn't referenced from anywhere.
1998-01-31 07:23:16 +00:00
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
#include <sys/param.h>
#include <sys/systm.h>
Introduce experimental support for MAC in the AppleTalk/EtherTalk stack. Label link layer mbufs as they are created for transmission, check mbufs before delivering them to sockets, label mbufs as they are created from sockets, and preserve mbuf labels if mbufs are copied. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-15 18:58:44 +00:00
#include <sys/mac.h>
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
#include <sys/mbuf.h>
#include <sys/kernel.h>
Removed unused #includes.
1997-09-07 08:30:24 +00:00
#include <sys/socket.h>
#include <sys/syslog.h>
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
#include <net/if.h>
Removed unused #includes.
1997-09-07 08:30:24 +00:00
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
#include <netinet/in.h>
#undef s_net
#include <netinet/if_ether.h>
#include <netatalk/at.h>
#include <netatalk/at_var.h>
#include <netatalk/aarp.h>
#include <netatalk/phase2.h>
#include <netatalk/at_extern.h>
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
static void aarptfree(struct aarptab *aat);
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
static void at_aarpinput(struct ifnet *ifp, struct mbuf *m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
#define AARPTAB_BSIZ 9
#define AARPTAB_NB 19
#define AARPTAB_SIZE (AARPTAB_BSIZ * AARPTAB_NB)
Staticize.
1998-02-09 06:11:36 +00:00
static struct aarptab aarptab[AARPTAB_SIZE];
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Convert the aa_ifaddr timeout to a callout, and run the aarprobe callout MPSAFE. Acquire the aarptab_mtx to make sure that the callout and msleep in the ioctl thread don't race. MFC after: 1 week
2005-02-22 14:20:29 +00:00
struct mtx aarptab_mtx;
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
MTX_SYSINIT(aarptab_mtx, &aarptab_mtx, "aarptab_mtx", MTX_DEF);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
#define AARPTAB_HASH(a) \
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
((((a).s_net << 8) + (a).s_node) % AARPTAB_NB)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Compare pointers with NULL rather than 0, or treating them as boolans in if statements. at_rmx gets a $FreeBSD$ out of the deal also (this code appears to be unused).
2004-03-22 03:57:01 +00:00
#define AARPTAB_LOOK(aat, addr) { \
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
int n; \
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_LOCK_ASSERT(); \
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
aat = &aarptab[ AARPTAB_HASH(addr) * AARPTAB_BSIZ ]; \
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
for (n = 0; n < AARPTAB_BSIZ; n++, aat++) \
if (aat->aat_ataddr.s_net == (addr).s_net && \
aat->aat_ataddr.s_node == (addr).s_node) \
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
break; \
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (n >= AARPTAB_BSIZ) \
Compare pointers with NULL rather than 0, or treating them as boolans in if statements. at_rmx gets a $FreeBSD$ out of the deal also (this code appears to be unused).
2004-03-22 03:57:01 +00:00
aat = NULL; \
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
#define AARPT_AGE (60 * 1)
#define AARPT_KILLC 20
#define AARPT_KILLI 3
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
# if !defined(__FreeBSD__)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
extern u_char etherbroadcastaddr[6];
Fix warning; Extra tokens after #endif
2002-05-24 06:11:07 +00:00
# endif /* __FreeBSD__ */
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Const-poison atmulticastaddr, which should be read but not modified. While there, remove (caddr_t) casting of ethernet addresses, which among other things discards the qualifier. This makes it clear that atmulticastaddr does not require synchronization.
2004-03-13 05:27:17 +00:00
static const u_char atmulticastaddr[ 6 ] = {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
0x09, 0x00, 0x07, 0xff, 0xff, 0xff,
};
u_char at_org_code[ 3 ] = {
0x08, 0x00, 0x07,
};
Const-poison atmulticastaddr, which should be read but not modified. While there, remove (caddr_t) casting of ethernet addresses, which among other things discards the qualifier. This makes it clear that atmulticastaddr does not require synchronization.
2004-03-13 05:27:17 +00:00
const u_char aarp_org_code[ 3 ] = {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
0x00, 0x00, 0x00,
};
Update for new callout interface.
1997-09-21 22:02:25 +00:00
static struct callout_handle aarptimer_ch =
CALLOUT_HANDLE_INITIALIZER(&aarptimer_ch);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
static void
aarptimer(void *ignored)
{
struct aarptab *aat;
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
int i;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarptimer_ch = timeout(aarptimer, (caddr_t)0, AARPT_AGE * hz);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
aat = aarptab;
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_LOCK();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
for (i = 0; i < AARPTAB_SIZE; i++, aat++) {
if (aat->aat_flags == 0 || (aat->aat_flags & ATF_PERM))
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
continue;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (++aat->aat_timer < ((aat->aat_flags & ATF_COM) ?
AARPT_KILLC : AARPT_KILLI))
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
continue;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarptfree(aat);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_UNLOCK();
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Massively COMMENT at_control.c and fix some bugs.. also fix a bug in aarp.c that didn't take netranges into account. default routes now work with appletalk, which is a poor-man's way of being able to access netranges if you only have one network :) Hopefully the full netranges fix will happen soon.
1996-08-29 23:16:34 +00:00
/*
* search through the network addresses to find one that includes
* the given network.. remember to take netranges into
* consideration.
*/
Fix various problems with netatalk kernel support. Some of these changes are a bit rough and will become more polished later. the changes to if_ethersubr should largely be moved to within the appletalk code, but that will happen later. A few of these were related to network-byteorder problems, and more were related to loopback failures.
1997-10-29 00:30:52 +00:00
struct at_ifaddr *
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
at_ifawithnet(struct sockaddr_at *sat)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
{
Fix various problems with netatalk kernel support. Some of these changes are a bit rough and will become more polished later. the changes to if_ethersubr should largely be moved to within the appletalk code, but that will happen later. A few of these were related to network-byteorder problems, and more were related to loopback failures.
1997-10-29 00:30:52 +00:00
struct at_ifaddr *aa;
Massively COMMENT at_control.c and fix some bugs.. also fix a bug in aarp.c that didn't take netranges into account. default routes now work with appletalk, which is a poor-man's way of being able to access netranges if you only have one network :) Hopefully the full netranges fix will happen soon.
1996-08-29 23:16:34 +00:00
struct sockaddr_at *sat2;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Rename 'at_ifaddr' list to 'at_ifaddr_list' so that the variable is more easily mechanically distinguished from 'struct at_ifaddr'.
2004-03-22 04:50:36 +00:00
for (aa = at_ifaddr_list; aa != NULL; aa = aa->aa_next) {
Fix various problems with netatalk kernel support. Some of these changes are a bit rough and will become more polished later. the changes to if_ethersubr should largely be moved to within the appletalk code, but that will happen later. A few of these were related to network-byteorder problems, and more were related to loopback failures.
1997-10-29 00:30:52 +00:00
sat2 = &(aa->aa_addr);
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (sat2->sat_addr.s_net == sat->sat_addr.s_net) {
Fix various problems with netatalk kernel support. Some of these changes are a bit rough and will become more polished later. the changes to if_ethersubr should largely be moved to within the appletalk code, but that will happen later. A few of these were related to network-byteorder problems, and more were related to loopback failures.
1997-10-29 00:30:52 +00:00
break;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if((aa->aa_flags & AFA_PHASE2)
Fix various problems with netatalk kernel support. Some of these changes are a bit rough and will become more polished later. the changes to if_ethersubr should largely be moved to within the appletalk code, but that will happen later. A few of these were related to network-byteorder problems, and more were related to loopback failures.
1997-10-29 00:30:52 +00:00
&& (ntohs(aa->aa_firstnet) <= ntohs(sat->sat_addr.s_net))
&& (ntohs(aa->aa_lastnet) >= ntohs(sat->sat_addr.s_net))) {
break;
}
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
return (aa);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
static void
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
aarpwhohas(struct ifnet *ifp, struct sockaddr_at *sat)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
{
struct mbuf *m;
struct ether_header *eh;
struct ether_aarp *ea;
struct at_ifaddr *aa;
struct llc *llc;
struct sockaddr sa;
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_UNLOCK_ASSERT();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if ((m = m_gethdr(M_DONTWAIT, MT_DATA)) == NULL) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Introduce experimental support for MAC in the AppleTalk/EtherTalk stack. Label link layer mbufs as they are created for transmission, check mbufs before delivering them to sockets, label mbufs as they are created from sockets, and preserve mbuf labels if mbufs are copied. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-15 18:58:44 +00:00
#ifdef MAC
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
mac_create_mbuf_linklayer(ifp, m);
Introduce experimental support for MAC in the AppleTalk/EtherTalk stack. Label link layer mbufs as they are created for transmission, check mbufs before delivering them to sockets, label mbufs as they are created from sockets, and preserve mbuf labels if mbufs are copied. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-15 18:58:44 +00:00
#endif
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
m->m_len = sizeof(*ea);
m->m_pkthdr.len = sizeof(*ea);
MH_ALIGN(m, sizeof(*ea));
ea = mtod(m, struct ether_aarp *);
bzero((caddr_t)ea, sizeof(*ea));
ea->aarp_hrd = htons(AARPHRD_ETHER);
ea->aarp_pro = htons(ETHERTYPE_AT);
ea->aarp_hln = sizeof(ea->aarp_sha);
ea->aarp_pln = sizeof(ea->aarp_spu);
ea->aarp_op = htons(AARPOP_REQUEST);
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
bcopy(IFP2AC(ifp)->ac_enaddr, (caddr_t)ea->aarp_sha,
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
sizeof(ea->aarp_sha));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
/*
* We need to check whether the output ethernet type should
* be phase 1 or 2. We have the interface that we'll be sending
* the aarp out. We need to find an AppleTalk network on that
* interface with the same address as we're looking for. If the
* net is phase 2, generate an 802.2 and SNAP header.
*/
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if ((aa = at_ifawithnet(sat)) == NULL) {
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
eh = (struct ether_header *)sa.sa_data;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aa->aa_flags & AFA_PHASE2) {
bcopy(atmulticastaddr, eh->ether_dhost, sizeof(eh->ether_dhost));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
eh->ether_type = htons(sizeof(struct llc) + sizeof(struct ether_aarp));
When generating a phase II ARP lookup from aarpwhohas(), use a non-sleeping mbuf allocation. MFC after: 1 week
2005-02-22 14:37:22 +00:00
M_PREPEND(m, sizeof(struct llc), M_DONTWAIT);
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (m == NULL) {
More checking of M_PREPEND() return values: pass through link-layer Appletalk code. These changes are untested since I don't have an Appletalk environment.
2003-08-29 19:27:09 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
llc = mtod(m, struct llc *);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
llc->llc_dsap = llc->llc_ssap = LLC_SNAP_LSAP;
llc->llc_control = LLC_UI;
Const-poison atmulticastaddr, which should be read but not modified. While there, remove (caddr_t) casting of ethernet addresses, which among other things discards the qualifier. This makes it clear that atmulticastaddr does not require synchronization.
2004-03-13 05:27:17 +00:00
bcopy(aarp_org_code, llc->llc_org_code, sizeof(aarp_org_code));
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
llc->llc_ether_type = htons(ETHERTYPE_AARP);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
bcopy(&AA_SAT(aa)->sat_addr.s_net, ea->aarp_spnet,
sizeof(ea->aarp_spnet));
bcopy(&sat->sat_addr.s_net, ea->aarp_tpnet,
sizeof(ea->aarp_tpnet));
ea->aarp_spnode = AA_SAT(aa)->sat_addr.s_node;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
ea->aarp_tpnode = sat->sat_addr.s_node;
} else {
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
bcopy(ifp->if_broadcastaddr, (caddr_t)eh->ether_dhost,
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
sizeof(eh->ether_dhost));
eh->ether_type = htons(ETHERTYPE_AARP);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
ea->aarp_spa = AA_SAT(aa)->sat_addr.s_node;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
ea->aarp_tpa = sat->sat_addr.s_node;
}
#ifdef NETATALKDEBUG
printf("aarp: sending request for %u.%u\n",
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
ntohs(AA_SAT(aa)->sat_addr.s_net),
AA_SAT(aa)->sat_addr.s_node);
Make the debug options new-style. This also zaps a DPT option from lint; it wasn't referenced from anywhere.
1998-01-31 07:23:16 +00:00
#endif /* NETATALKDEBUG */
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
sa.sa_len = sizeof(struct sockaddr);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
sa.sa_family = AF_UNSPEC;
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
ifp->if_output(ifp, m, &sa, NULL /* route */);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
int
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
aarpresolve(ifp, m, destsat, desten)
struct ifnet *ifp;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
struct mbuf *m;
struct sockaddr_at *destsat;
u_char *desten;
{
struct at_ifaddr *aa;
struct aarptab *aat;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (at_broadcast(destsat)) {
Fix various problems with netatalk kernel support. Some of these changes are a bit rough and will become more polished later. the changes to if_ethersubr should largely be moved to within the appletalk code, but that will happen later. A few of these were related to network-byteorder problems, and more were related to loopback failures.
1997-10-29 00:30:52 +00:00
m->m_flags |= M_BCAST;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if ((aa = at_ifawithnet(destsat)) == NULL) {
m_freem(m);
return (0);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aa->aa_flags & AFA_PHASE2) {
Const-poison atmulticastaddr, which should be read but not modified. While there, remove (caddr_t) casting of ethernet addresses, which among other things discards the qualifier. This makes it clear that atmulticastaddr does not require synchronization.
2004-03-13 05:27:17 +00:00
bcopy(atmulticastaddr, (caddr_t)desten, sizeof(atmulticastaddr));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
} else {
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
bcopy(ifp->if_broadcastaddr, (caddr_t)desten,
sizeof(ifp->if_addrlen));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
return (1);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_LOCK();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
AARPTAB_LOOK(aat, destsat->sat_addr);
Compare pointers with NULL rather than 0, or treating them as boolans in if statements. at_rmx gets a $FreeBSD$ out of the deal also (this code appears to be unused).
2004-03-22 03:57:01 +00:00
if (aat == NULL) { /* No entry */
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aat = aarptnew(&destsat->sat_addr);
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
if (aat == NULL) { /* we should fail more gracefully! */
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
panic("aarpresolve: no free entry");
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
goto done;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
/* found an entry */
aat->aat_timer = 0;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aat->aat_flags & ATF_COM) { /* entry is COMplete */
bcopy((caddr_t)aat->aat_enaddr, (caddr_t)desten,
sizeof(aat->aat_enaddr));
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_UNLOCK();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
return (1);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
/* entry has not completed */
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aat->aat_hold) {
m_freem(aat->aat_hold);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
done:
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
aat->aat_hold = m;
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_UNLOCK();
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
aarpwhohas(ifp, destsat);
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
return (0);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
void
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarpintr(m)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
struct mbuf *m;
{
struct arphdr *ar;
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
struct ifnet *ifp;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
ifp = m->m_pkthdr.rcvif;
if (ifp->if_flags & IFF_NOARP)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
goto out;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (m->m_len < sizeof(struct arphdr)) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
goto out;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
ar = mtod(m, struct arphdr *);
if (ntohs(ar->ar_hrd) != AARPHRD_ETHER) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
goto out;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (m->m_len < sizeof(struct arphdr) + 2 * ar->ar_hln +
2 * ar->ar_pln) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
goto out;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
switch(ntohs(ar->ar_pro)) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
case ETHERTYPE_AT :
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
at_aarpinput(ifp, m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
default:
break;
}
out:
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
static void
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
at_aarpinput(struct ifnet *ifp, struct mbuf *m)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
{
struct ether_aarp *ea;
struct at_ifaddr *aa;
struct aarptab *aat;
struct ether_header *eh;
struct llc *llc;
struct sockaddr_at sat;
struct sockaddr sa;
struct at_addr spa, tpa, ma;
Submitted by: archie@whistle.com appletalk cleanups
1996-07-23 01:18:47 +00:00
int op;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
u_short net;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
ea = mtod(m, struct ether_aarp *);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
/* Check to see if from my hardware address */
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
if (!bcmp((caddr_t)ea->aarp_sha, IFP2AC(ifp)->ac_enaddr,
sizeof(IFP2AC(ifp)->ac_enaddr))) {
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
op = ntohs(ea->aarp_op);
bcopy(ea->aarp_tpnet, &net, sizeof(net));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (net != 0) { /* should be ATADDR_ANYNET? */
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
sat.sat_len = sizeof(struct sockaddr_at);
sat.sat_family = AF_APPLETALK;
sat.sat_addr.s_net = net;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if ((aa = at_ifawithnet(&sat)) == NULL) {
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
bcopy(ea->aarp_spnet, &spa.s_net, sizeof(spa.s_net));
bcopy(ea->aarp_tpnet, &tpa.s_net, sizeof(tpa.s_net));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
} else {
/*
* Since we don't know the net, we just look for the first
* phase 1 address on the interface.
*/
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
for (aa = (struct at_ifaddr *)TAILQ_FIRST(&ifp->if_addrhead); aa;
Convert the interface address and IP interface address structures to TAILQs. Fix places which referenced these for no good reason that I can see (the references remain, but were fixed to compile again; they are still questionable).
1996-12-13 21:29:07 +00:00
aa = (struct at_ifaddr *)aa->aa_ifa.ifa_link.tqe_next) {
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (AA_SAT(aa)->sat_family == AF_APPLETALK &&
(aa->aa_flags & AFA_PHASE2) == 0) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
break;
}
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aa == NULL) {
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
tpa.s_net = spa.s_net = AA_SAT(aa)->sat_addr.s_net;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
spa.s_node = ea->aarp_spnode;
tpa.s_node = ea->aarp_tpnode;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
ma.s_net = AA_SAT(aa)->sat_addr.s_net;
ma.s_node = AA_SAT(aa)->sat_addr.s_node;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
/*
* This looks like it's from us.
*/
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (spa.s_net == ma.s_net && spa.s_node == ma.s_node) {
if (aa->aa_flags & AFA_PROBING) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
/*
* We're probing, someone either responded to our probe, or
* probed for the same address we'd like to use. Change the
* address we're probing for.
*/
Convert the aa_ifaddr timeout to a callout, and run the aarprobe callout MPSAFE. Acquire the aarptab_mtx to make sure that the callout and msleep in the ioctl thread don't race. MFC after: 1 week
2005-02-22 14:20:29 +00:00
callout_stop(&aa->aa_callout);
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
wakeup(aa);
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
} else if (op != AARPOP_PROBE) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
/*
* This is not a probe, and we're not probing. This means
* that someone's saying they have the same source address
* as the one we're using. Get upset...
*/
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
log(LOG_ERR,
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
"aarp: duplicate AT address!! %x:%x:%x:%x:%x:%x\n",
ea->aarp_sha[ 0 ], ea->aarp_sha[ 1 ], ea->aarp_sha[ 2 ],
ea->aarp_sha[ 3 ], ea->aarp_sha[ 4 ], ea->aarp_sha[ 5 ]);
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
}
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_LOCK();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
AARPTAB_LOOK(aat, spa);
Compare pointers with NULL rather than 0, or treating them as boolans in if statements. at_rmx gets a $FreeBSD$ out of the deal also (this code appears to be unused).
2004-03-22 03:57:01 +00:00
if (aat != NULL) {
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (op == AARPOP_PROBE) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
/*
* Someone's probing for spa, dealocate the one we've got,
* so that if the prober keeps the address, we'll be able
* to arp for him.
*/
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarptfree(aat);
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_UNLOCK();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
bcopy((caddr_t)ea->aarp_sha, (caddr_t)aat->aat_enaddr,
sizeof(ea->aarp_sha));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
aat->aat_flags |= ATF_COM;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aat->aat_hold) {
Slight cleanup of aarp code Submitted by: Ian Dowse <iedowse@maths.tcd.ie> PR: kern/17583 This PR is still open as Ther is still ongoing investigation.
2000-03-29 17:12:09 +00:00
struct mbuf *mhold = aat->aat_hold;
aat->aat_hold = NULL;
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_UNLOCK();
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
sat.sat_len = sizeof(struct sockaddr_at);
sat.sat_family = AF_APPLETALK;
sat.sat_addr = spa;
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
(*ifp->if_output)(ifp, mhold,
Submitted by: archie@whistle.com appletalk cleanups
1996-07-23 01:18:47 +00:00
(struct sockaddr *)&sat, NULL); /* XXX */
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
} else
AARPTAB_UNLOCK();
Slight cleanup of aarp code Submitted by: Ian Dowse <iedowse@maths.tcd.ie> PR: kern/17583 This PR is still open as Ther is still ongoing investigation.
2000-03-29 17:12:09 +00:00
} else if ((tpa.s_net == ma.s_net)
&& (tpa.s_node == ma.s_node)
&& (op != AARPOP_PROBE)
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
&& ((aat = aarptnew(&spa)) != NULL)) {
bcopy((caddr_t)ea->aarp_sha, (caddr_t)aat->aat_enaddr,
sizeof(ea->aarp_sha));
aat->aat_flags |= ATF_COM;
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_UNLOCK();
} else
AARPTAB_UNLOCK();
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
/*
* Don't respond to responses, and never respond if we're
* still probing.
*/
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (tpa.s_net != ma.s_net || tpa.s_node != ma.s_node ||
op == AARPOP_RESPONSE || (aa->aa_flags & AFA_PROBING)) {
m_freem(m);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
bcopy((caddr_t)ea->aarp_sha, (caddr_t)ea->aarp_tha,
sizeof(ea->aarp_sha));
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
bcopy(IFP2AC(ifp)->ac_enaddr, (caddr_t)ea->aarp_sha,
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
sizeof(ea->aarp_sha));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
/* XXX */
eh = (struct ether_header *)sa.sa_data;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
bcopy((caddr_t)ea->aarp_tha, (caddr_t)eh->ether_dhost,
sizeof(eh->ether_dhost));
if (aa->aa_flags & AFA_PHASE2) {
eh->ether_type = htons(sizeof(struct llc) +
sizeof(struct ether_aarp));
M_PREPEND(m, sizeof(struct llc), M_DONTWAIT);
if (m == NULL) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
llc = mtod(m, struct llc *);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
llc->llc_dsap = llc->llc_ssap = LLC_SNAP_LSAP;
llc->llc_control = LLC_UI;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
bcopy(aarp_org_code, llc->llc_org_code, sizeof(aarp_org_code));
llc->llc_ether_type = htons(ETHERTYPE_AARP);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
bcopy(ea->aarp_spnet, ea->aarp_tpnet, sizeof(ea->aarp_tpnet));
bcopy(&ma.s_net, ea->aarp_spnet, sizeof(ea->aarp_spnet));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
} else {
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
eh->ether_type = htons(ETHERTYPE_AARP);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
ea->aarp_tpnode = ea->aarp_spnode;
ea->aarp_spnode = ma.s_node;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
ea->aarp_op = htons(AARPOP_RESPONSE);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
sa.sa_len = sizeof(struct sockaddr);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
sa.sa_family = AF_UNSPEC;
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
(*ifp->if_output)(ifp, m, &sa, NULL); /* XXX */
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
static void
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarptfree(struct aarptab *aat)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
{
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_LOCK_ASSERT();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aat->aat_hold)
m_freem(aat->aat_hold);
Slight cleanup of aarp code Submitted by: Ian Dowse <iedowse@maths.tcd.ie> PR: kern/17583 This PR is still open as Ther is still ongoing investigation.
2000-03-29 17:12:09 +00:00
aat->aat_hold = NULL;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
aat->aat_timer = aat->aat_flags = 0;
aat->aat_ataddr.s_net = 0;
aat->aat_ataddr.s_node = 0;
}
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
struct aarptab *
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarptnew(addr)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
struct at_addr *addr;
{
int n;
int oldest = -1;
struct aarptab *aat, *aato = NULL;
static int first = 1;
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_LOCK_ASSERT();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (first) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
first = 0;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarptimer_ch = timeout(aarptimer, (caddr_t)0, hz);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aat = &aarptab[ AARPTAB_HASH(*addr) * AARPTAB_BSIZ ];
for (n = 0; n < AARPTAB_BSIZ; n++, aat++) {
if (aat->aat_flags == 0)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
goto out;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aat->aat_flags & ATF_PERM)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
continue;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if ((int) aat->aat_timer > oldest) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
oldest = aat->aat_timer;
aato = aat;
}
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aato == NULL)
return (NULL);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
aat = aato;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarptfree(aat);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
out:
aat->aat_ataddr = *addr;
aat->aat_flags = ATF_INUSE;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
return (aat);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
void
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
aarpprobe(void *arg)
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
{
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
struct ifnet *ifp = arg;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
struct mbuf *m;
struct ether_header *eh;
struct ether_aarp *ea;
struct at_ifaddr *aa;
struct llc *llc;
struct sockaddr sa;
/*
* We need to check whether the output ethernet type should
* be phase 1 or 2. We have the interface that we'll be sending
* the aarp out. We need to find an AppleTalk network on that
* interface with the same address as we're looking for. If the
* net is phase 2, generate an 802.2 and SNAP header.
*/
Convert the aa_ifaddr timeout to a callout, and run the aarprobe callout MPSAFE. Acquire the aarptab_mtx to make sure that the callout and msleep in the ioctl thread don't race. MFC after: 1 week
2005-02-22 14:20:29 +00:00
AARPTAB_LOCK();
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
for (aa = (struct at_ifaddr *)TAILQ_FIRST(&ifp->if_addrhead); aa;
Convert the interface address and IP interface address structures to TAILQs. Fix places which referenced these for no good reason that I can see (the references remain, but were fixed to compile again; they are still questionable).
1996-12-13 21:29:07 +00:00
aa = (struct at_ifaddr *)aa->aa_ifa.ifa_link.tqe_next) {
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (AA_SAT(aa)->sat_family == AF_APPLETALK &&
(aa->aa_flags & AFA_PROBING)) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
break;
}
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aa == NULL) { /* serious error XXX */
Convert the aa_ifaddr timeout to a callout, and run the aarprobe callout MPSAFE. Acquire the aarptab_mtx to make sure that the callout and msleep in the ioctl thread don't race. MFC after: 1 week
2005-02-22 14:20:29 +00:00
AARPTAB_UNLOCK();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
printf("aarpprobe why did this happen?!\n");
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aa->aa_probcnt <= 0) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
aa->aa_flags &= ~AFA_PROBING;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
wakeup(aa);
Convert the aa_ifaddr timeout to a callout, and run the aarprobe callout MPSAFE. Acquire the aarptab_mtx to make sure that the callout and msleep in the ioctl thread don't race. MFC after: 1 week
2005-02-22 14:20:29 +00:00
AARPTAB_UNLOCK();
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
} else {
Convert the aa_ifaddr timeout to a callout, and run the aarprobe callout MPSAFE. Acquire the aarptab_mtx to make sure that the callout and msleep in the ioctl thread don't race. MFC after: 1 week
2005-02-22 14:20:29 +00:00
callout_reset(&aa->aa_callout, hz / 5, aarpprobe, ifp);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Convert the aa_ifaddr timeout to a callout, and run the aarprobe callout MPSAFE. Acquire the aarptab_mtx to make sure that the callout and msleep in the ioctl thread don't race. MFC after: 1 week
2005-02-22 14:20:29 +00:00
AARPTAB_UNLOCK();
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if ((m = m_gethdr(M_DONTWAIT, MT_DATA)) == NULL) {
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
return;
}
Introduce experimental support for MAC in the AppleTalk/EtherTalk stack. Label link layer mbufs as they are created for transmission, check mbufs before delivering them to sockets, label mbufs as they are created from sockets, and preserve mbuf labels if mbufs are copied. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-15 18:58:44 +00:00
#ifdef MAC
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
mac_create_mbuf_linklayer(ifp, m);
Introduce experimental support for MAC in the AppleTalk/EtherTalk stack. Label link layer mbufs as they are created for transmission, check mbufs before delivering them to sockets, label mbufs as they are created from sockets, and preserve mbuf labels if mbufs are copied. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-15 18:58:44 +00:00
#endif
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
m->m_len = sizeof(*ea);
m->m_pkthdr.len = sizeof(*ea);
MH_ALIGN(m, sizeof(*ea));
ea = mtod(m, struct ether_aarp *);
bzero((caddr_t)ea, sizeof(*ea));
ea->aarp_hrd = htons(AARPHRD_ETHER);
ea->aarp_pro = htons(ETHERTYPE_AT);
ea->aarp_hln = sizeof(ea->aarp_sha);
ea->aarp_pln = sizeof(ea->aarp_spu);
ea->aarp_op = htons(AARPOP_PROBE);
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
bcopy(IFP2AC(ifp)->ac_enaddr, (caddr_t)ea->aarp_sha,
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
sizeof(ea->aarp_sha));
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
eh = (struct ether_header *)sa.sa_data;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
if (aa->aa_flags & AFA_PHASE2) {
bcopy(atmulticastaddr, eh->ether_dhost, sizeof(eh->ether_dhost));
eh->ether_type = htons(sizeof(struct llc) +
sizeof(struct ether_aarp));
M_PREPEND(m, sizeof(struct llc), M_TRYWAIT);
if (m == NULL) {
More checking of M_PREPEND() return values: pass through link-layer Appletalk code. These changes are untested since I don't have an Appletalk environment.
2003-08-29 19:27:09 +00:00
return;
}
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
llc = mtod(m, struct llc *);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
llc->llc_dsap = llc->llc_ssap = LLC_SNAP_LSAP;
llc->llc_control = LLC_UI;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
bcopy(aarp_org_code, llc->llc_org_code, sizeof(aarp_org_code));
llc->llc_ether_type = htons(ETHERTYPE_AARP);
bcopy(&AA_SAT(aa)->sat_addr.s_net, ea->aarp_spnet,
sizeof(ea->aarp_spnet));
bcopy(&AA_SAT(aa)->sat_addr.s_net, ea->aarp_tpnet,
sizeof(ea->aarp_tpnet));
ea->aarp_spnode = ea->aarp_tpnode = AA_SAT(aa)->sat_addr.s_node;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
} else {
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
bcopy(ifp->if_broadcastaddr, (caddr_t)eh->ether_dhost,
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
sizeof(eh->ether_dhost));
eh->ether_type = htons(ETHERTYPE_AARP);
ea->aarp_spa = ea->aarp_tpa = AA_SAT(aa)->sat_addr.s_node;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
#ifdef NETATALKDEBUG
printf("aarp: sending probe for %u.%u\n",
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
ntohs(AA_SAT(aa)->sat_addr.s_net),
AA_SAT(aa)->sat_addr.s_node);
Make the debug options new-style. This also zaps a DPT option from lint; it wasn't referenced from anywhere.
1998-01-31 07:23:16 +00:00
#endif /* NETATALKDEBUG */
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
sa.sa_len = sizeof(struct sockaddr);
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
sa.sa_family = AF_UNSPEC;
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
(*ifp->if_output)(ifp, m, &sa, NULL); /* XXX */
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
aa->aa_probcnt--;
}
void
aarp_clean(void)
{
struct aarptab *aat;
int i;
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
untimeout(aarptimer, 0, aarptimer_ch);
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_LOCK();
Since I have my hands all over netatalk adding locking and restructuring it, cinch the file's style closer to style(9) with regard to parenthesis: s/( /(/g s/ )/)/g s/return(/return (/g s/return 0/return (0)/ s/return 1/return (1)/
2004-03-22 03:24:10 +00:00
for (i = 0, aat = aarptab; i < AARPTAB_SIZE; i++, aat++) {
if (aat->aat_hold) {
m_freem(aat->aat_hold);
Slight cleanup of aarp code Submitted by: Ian Dowse <iedowse@maths.tcd.ie> PR: kern/17583 This PR is still open as Ther is still ongoing investigation.
2000-03-29 17:12:09 +00:00
aat->aat_hold = NULL;
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
}
Lock down the netatalk AARP code, which is responsible for appletalk address discovery and caching (similar to inet ARP). Use a single global mutex, aarptab_mtx, to protect the table. Remove spl/spx. Tested by: Bob Bishop <rb@gid.co.uk>
2004-04-09 01:40:12 +00:00
AARPTAB_UNLOCK();
Obtained from: netatalk distribution netatalk@itd.umich.edu Kernel Appletalk protocol support both CAP and netatalk can make use of this.. still needs some owrk but it seemd the right tiime to commit it so other can experiment.
1996-05-24 01:35:45 +00:00
}
Reference in New Issue Copy Permalink