svn commit: r219779 - head/sys/netinet

[Bjoern A. Zeeb]

Author: bz
Date: Sat Mar 19 19:08:54 2011
New Revision: 219779
URL: http://svn.freebsd.org/changeset/base/219779

Log:
Properly check for an IPv4 socket after r219579.

In some cases as udp6_connect() without an earlier bind(2) to an
address, v4-mapped scokets allowed and a non mapped destination
address, we can end up here with both v4 and v6 indicated:
inp_vflag = (INP_IPV4|INP_IPV6|INP_IPV6PROTO)

In that case however laddrp is NULL as the IPv6 path does not
pass in a copy currently.

Reported by: Pawel Worach (pawel.worach gmail.com)
Tested by: Pawel Worach (pawel.worach gmail.com)
MFC after: 6 days
X-MFC with: r219579

Modified:
head/sys/netinet/in_pcb.c

Modified: head/sys/netinet/in_pcb.c
==============================================================================
--- head/sys/netinet/in_pcb.c Sat Mar 19 18:26:54 2011 (r219778)
+++ head/sys/netinet/in_pcb.c Sat Mar 19 19:08:54 2011 (r219779)
@@ -387,7 +387,7 @@ in_pcb_lport(struct inpcb *inp, struct i
#ifdef INET
/* Make the compiler happy. */
laddr.s_addr = 0;
- if ((inp->inp_vflag & INP_IPV4) != 0) {
+ if ((inp->inp_vflag & (INP_IPV4|INP_IPV6)) == INP_IPV4) {
KASSERT(laddrp != NULL, ("%s: laddrp NULL for v4 inp %p",
__func__, inp));
laddr = *laddrp;
@@ -423,7 +423,7 @@ in_pcb_lport(struct inpcb *inp, struct i
} while (tmpinp != NULL);

#ifdef INET
- if ((inp->inp_vflag & INP_IPV4) != 0)
+ if ((inp->inp_vflag & (INP_IPV4|INP_IPV6)) == INP_IPV4)
laddrp->s_addr = laddr.s_addr;
#endif
*lportp = lport;

[Bjoern A. Zeeb]

Post by Bjoern A. Zeeb
Author: bz
Date: Sat Mar 19 19:08:54 2011
New Revision: 219779
URL: http://svn.freebsd.org/changeset/base/219779
Properly check for an IPv4 socket after r219579.
In some cases as udp6_connect() without an earlier bind(2) to an
address, v4-mapped scokets allowed and a non mapped destination
inp_vflag = (INP_IPV4|INP_IPV6|INP_IPV6PROTO)
In that case however laddrp is NULL as the IPv6 path does not
pass in a copy currently.

For your information and future reference:

UDP and TCP come in with flags set differently for the 6 cases. The
only conistent one is IPv4 only. v4-mapped v6 is the trouble maker.

So while this one fixed the panic with plain IPv6 and v4-mapped
enabled, it actually broke udp on v4-mapped as further investigation
has shown. I am working on regression tests as well as harmonizing
the initial calls setting the flags and documenting things (and
obviously fixing that one case as well;)

/bz

Post by Bjoern A. Zeeb
Reported by: Pawel Worach (pawel.worach gmail.com)
Tested by: Pawel Worach (pawel.worach gmail.com)
MFC after: 6 days
X-MFC with: r219579
head/sys/netinet/in_pcb.c
Modified: head/sys/netinet/in_pcb.c
==============================================================================
--- head/sys/netinet/in_pcb.c Sat Mar 19 18:26:54 2011 (r219778)
+++ head/sys/netinet/in_pcb.c Sat Mar 19 19:08:54 2011 (r219779)
@@ -387,7 +387,7 @@ in_pcb_lport(struct inpcb *inp, struct i
#ifdef INET
/* Make the compiler happy. */
laddr.s_addr = 0;
- if ((inp->inp_vflag & INP_IPV4) != 0) {
+ if ((inp->inp_vflag & (INP_IPV4|INP_IPV6)) == INP_IPV4) {
KASSERT(laddrp != NULL, ("%s: laddrp NULL for v4 inp %p",
__func__, inp));
laddr = *laddrp;
@@ -423,7 +423,7 @@ in_pcb_lport(struct inpcb *inp, struct i
} while (tmpinp != NULL);
#ifdef INET
- if ((inp->inp_vflag & INP_IPV4) != 0)
+ if ((inp->inp_vflag & (INP_IPV4|INP_IPV6)) == INP_IPV4)
laddrp->s_addr = laddr.s_addr;
#endif
*lportp = lport;