- most of the kernel code will not care about the actual encoding of
scope zone IDs and won't touch "s6_addr16[1]" directly.
- similarly, most of the kernel code will not care about link-local
scoped addresses as a special case.
- scope boundary check will be stricter. For example, the current
*BSD code allows a packet with src=::1 and dst=(some global IPv6
address) to be sent outside of the node, if the application do:
s = socket(AF_INET6);
bind(s, "::1");
sendto(s, some_global_IPv6_addr);
This is clearly wrong, since ::1 is only meaningful within a single
node, but the current implementation of the *BSD kernel cannot
reject this attempt.
- and, while there, don't try to remove the ff02::/32 interface route
entry in in6_ifdetach() as it's already gone.
This also includes some level of support for the standard source
address selection algorithm defined in RFC3484, which will be
completed on in the future.
From the KAME project via JINMEI Tatuya.
Approved by core@.
could modify the struct mbuf and calling functions (udp_input() and
udp4_realinput()) would have used a garbled local copy of the pointer.
The fix is not perfect. udp4_espinudp() should use m_pulldown()...
pass it to in_pcbbind() so that can allocate a low numbered port
if setsockopt() has been used to set IP_PORTRANGE to IP_PORTRANGE_LOW.
While there, fail in_pcbconnect() if the in_pcbbind() fails - rather
than sending the request out from a port of zero.
This has been largely broken since the socket option was added in 1998.
the decapsulator dispatch changes in 2001. Problems found and fixed
by Christine Jones of BBN. Specifically:
Check for a packet's protocol to be ENCAP_PROTO, not AF_INET.
Remove one-back cache for last vif, because vif_encapcheck is called
for each vif, rather than being expected to find the appropriate vif.
The cache usage caused packets to be input on the wrong vif and hence
usually dropped.
In vif_encapcheck, verify the local source as well. While mrouted
endeavors not to create multiple tunnels with a peer, a packet
arriving with the wrong local address is still wrong and should not be
accepted. (This is a correctness nit, not a security issue.) Order
checks to fail quickly for packets being checked to see if they match
a vif other than the one they belong on (essentially, check peer
source address in outer header first).
Claim 69 bits of match (32 each from outer src/dst and 5 from checking
that inner dst is within 224/5). This should result in the vif having
a higher priority for multicast packets compared to a parallel gif(4)
tunnel, and that both seems appropriate if both are configured and
seems to match the semantics expected by the decapsulator dispatch
machinery.
(These changes were made in 2.99.15 and about a dozen nodes are
running them with many vifs. ip_mroute.c has not changed
significantly since then (February 2005) and the changes applied
cleanly to current and compile cleanly.)
http://www.gont.com.ar/drafts/icmp-attacks-against-tcp.html
1. Don't act on ICMP-need-frag immediately if adhoc checks on the
advertised MTU fail. The MTU update is delayed until a TCP retransmit
happens.
2. Ignore ICMP Source Quench messages meant for TCP connections.
From OpenBSD.
store a struct ifnet *, and define it for udp/tcp/rawip for INET and
INET6. When deleting a struct ifnet, invoke PRU_PURGEIF on all
protocols marked with PR_PURGEIF. Closes PR kern/29580 (mine).
The __UNCONST macro is now used only where necessary and the RW macros
are gone. Most of the changes here are consumers of the
sysctl_createv(9) interface that now takes a pair of const pointers
which used not to be.
"const struct mbuf *" to "struct mbuf *". Without this change the
actual implementation cannot even use m_copydata() on the mbuf chain
which is broken.
chgsbsize() as mentioned in the PR can be called from an interrupt context
via tcp_close(). Avoid calling uid_find() in chgsbsize().
- Instead of storing so_uid in struct socketvar, store *so_uidinfo
- Add a simple lock to struct uidinfo.
- for ipv4, defer decision to ip layer as h/w checksum offloading does
so that it can check the actual interface the packet is going to.
- for ipv6, disable it.
(maybe will be revisited when it implements h/w checksum offloading.)
ok'ed by Jason Thorpe.
- introduce t_segqlen, the number of segments in segq/timeq.
the name is from freebsd.
- rather than maintaining a copy of sack blocks (rcv_sack_block[]),
build it directly from the segment list when needed.
- Use the correct IP header length variable for other-than-first packets.
- Remove redundant setting of the original IP header length in the first
packet's csum_data. (It's already set before ip_fragment() is called
in 1.147.)
net.local.stream.pcblist
net.local.dgram.pcblist
net.inet.tcp.pcblist
net.inet.udp.pcblist
net.inet.raw.pcblist
net.inet6.tcp6.pcblist
net.inet6.udp6.pcblist
net.inet6.raw6.pcblist
which allow retrieval of the pcbs in use for those protocols. The
struct involved is 32/64 bit clean and incorporates parts of struct
inpcb, struct unpcb, a bit of struct tcpcb, and two socket addresses.
on gnats-bugs in PR kern/29544.
Tested with an NFS client using default rwsize on an NFS server
with wm(4) interface configured IP4CSUM,TCP4CSUM,UDP4CSUM.
Prior revision required the server to have checksum offload disabled.
it in the TCP stack to test which of the REXMT or PERSIST timer is in use.
This fixes a race condition that could cause "panic: tcp_output REXMT". See
tech-net for details.
http://www.sigusr1.org/~kurahone/tcp-sack-netbsd-02152005.diff.gz
Fixes in that patch for pre-existing TCP pcb initializations were already
committed to NetBSD-current, so are not included in this commit.
The SACK patch has been observed to correctly negotiate and respond,
to SACKs in wide-area traffic.
There are two indepenently-observed, as-yet-unresolved anomalies:
First, seeing unexplained delays between in fast retransmission
(potentially explainable by an 0.2sec RTT between adjacent
ethernet/wifi NICs); and second, peculiar and unepxlained TCP
retransmits observed over an ath0 card.
After discussion with several interested developers, I'm committing
this now, as-is, for more eyes to use and look over. Current hypothesis
is that the anomalies above may in fact be due to link/level (hardware,
driver, HAL, firmware) abberations in the test setup, affecting both
Kentaro's wired-Ethernet NIC and in my two (different) WiFi NICs.
headers and LKM.
Add MKPF; if set to no, don't build and install the pf(4) programs,
headers, LKM and spamd.
Both options default to yes, so nothing changed in the default build.
Reviewed by lukem.
checksum is always in the L4 header by the time we get to this point. It
was occasionally not there due to a bug in tcp_respond, which has since
been fixed.
So, instead just stash the length of the L3 header in the high 16 bits of
csum_data.
(from an offset to the end of the packet), the pseudo-header checksum must be
calculated by software. So, provide it in the TCP/UDP header when
M_CSUM_NO_PSEUDOHDR is set in the interface's if_csum_flags_tx.
The start offset, the end of the IP header, is also provided in the high 16
bits of pkthdr.csum_data. Such that the driver need not examine the packet
at all.
XXX At the request of Jonathan Stone, note that sharing of if_csum_flags_tx &
pkthdr.csum_flags for checksum quirks should be re-evaluated.
Hans Rosenfeld (rosenfeld at grumpf.hope-2000.org)
This change makes it possible to add gif interfaces to bridges, which
will then send and receive IP protocol 97 packets. Packets are Ethernet
frames with an EtherIP header prepended.
1) dupseg_fix_=true from NS: do not count a segment with completely duplicate
data as a duplicate ack. This can occur due to duplicate packets in the
network, or due to fast retransmit from the other side.
2) dupack_reset_=false from NS: do not reset the duplicate ack counter or exit
fast recovery if we happen to get data or a window update along with a
duplicate ack.
3) In the "very old ack" case that itojun added, send an ACK before dropping
the segment, to try to update the other side's send sequence number.
4) Check the ssthresh crossover point with >= rather than >. Otherwise we
start to do "exponential" growth immediately following recovery, where we
should be doing "linear". This is what NS does.
and it could detect even older time stamps. (Really, to be 100% correct, there
should be a timer that clears these out -- but it probably doesn't matter in
the real world.)
* t_partialacks<0 means we are not in fast recovery.
* t_partialacks==0 means we are in fast recovery, but we have not received
any partial acks yet.
* t_partialacks>0 means we are in fast recovery, and we have received
partial acks.
This is used to implement 2 changes in RFC 3782:
* We keep the notion that we are in fast recovery separate from t_dupacks, so
it is not reset due to out-of-order acks. (This affects both the Reno and
NewReno cases.)
* We only reset the retransmit timer on the first partial ack -- preventing us
from possibly taking one RTO per segment once fast recovery is initiated.
As before, it is hard to measure any difference between Reno and NewReno in the
real-world cases that I've tested.
1) If an echoed RFC 1323 time stamp appears to be later than the current time,
ignore it and fall back to old-style RTT calculation. This prevents ending
up with a negative RTT and panicking later.
2) Fix NewReno. This involves a few changes:
a) Implement the send_high variable in RFC 2582. Our implementation is
subtly different; it is one *past* the last sequence number transmitted
rather than being equal to it. This simplifies some logic and makes
the code smaller. Additional logic was required to prevent sequence
number wraparound problems; this is not mentioned in RFC 2582.
b) Make sure we reset t_dupacks on new acks, but *not* on a partial ack.
All of the new ack code is pushed out into tcp_newreno(). (Later this
will probably be a pluggable function.) Thus t_dupacks keeps track of
whether we're in fast recovery all the time, with Reno or NewReno, which
keeps some logic simpler.
c) We do not need to update snd_recover when we're not in fast recovery.
See tech-net for an explanation of this.
d) In the gratuitous fast retransmit prevention case, do not send a packet.
RFC 2582 specifically says that we should "do nothing".
e) Do not inflate the congestion window on a partial ack. (This is done by
testing t_dupacks to see whether we're still in fast recovery.)
This brings the performance of NewReno back up to the same as Reno in a few
random test cases (e.g. transferring peer-to-peer over my wireless network).
I have not concocted a good test case for the behavior specific to NewReno.
- Allow rn_init() to be called multiple times, but do nothing except the
first call.
- Include opt_inet.h so that #ifdef INET works.
- Call rn_init() from encap_init() explicitly rather than depending on the
order of initialization.
received packet so that the checksum is not performed twice. Also,
tcp_respond() does not fill-in the m_pkthdr.csum_data, so a h/w checksum may
have the wrong offset.
OK from Jason Thorpe.
Connect over tcp on the loopback is broken:
4729 amq 0.000007 CALL connect(4,0x804f2a0,0x1c)
4729 amq 75.007420 RET connect -1 errno 60 Connection timed out
and marked as out of window - trying to do the add will result in a failure
and the packet being blocked, incorrectly.
Committed By: darrenr
Tested By: smb
("no domain for AF 0") on if_detach.
- SIOCAIFADDR, SIOCSIFADDR: free an address on error.
- SIOCSIFNETMASK, SIOCSIFDSTADDR: reject operations for an interface which
has no AF_INET addresses.
partly from OpenBSD and FreeBSD.
reviewed by Christos Zoulas on tech-net@.
the interface route and various internal state. Also, it should use an ifreq,
not an if_aliasreq. Addresses PR 9604. (Nothing in our source tree uses
SIOCSIFNETMASK, though. Perhaps it should be deprecated.)
1.) Make sure that "pass" is always initialized.
2.) Make sure the code doesn't use a stale mbuf pointer after fr_makefrip()
has been called. This fixes PR kern/25868.
Analyzed and reviewed by Steve Woodford.
support IPv6 if KAME IPSEC (RFC is not explicit about how we make data stream
for checksum with IPv6, but i'm pretty sure using normal pseudo-header is the
right thing).
XXX
current TCP MD5 signature code has giant flaw:
it does not validate signature on input (can't believe it! what is the point?)
(MD5 signatures for TCP, as used with BGP). Credit for original
FreeBSD code goes to Bruce M. Simpson, with FreeBSD sponsorship
credited to sentex.net. Shortening of the setsockopt() name
attributed to Vincent Jardin.
This commit is a minimal, working version of the FreeBSD code, as
MFC'ed to FreeBSD-4. It has received minimal testing with a ttcp
modified to set the TCP-MD5 option; BMS's additions to tcpdump-current
(tcpdump -M) confirm that the MD5 signatures are correct. Committed
as-is for further testing between a NetBSD BGP speaker (e.g., quagga)
and industry-standard BGP speakers (e.g., Cisco, Juniper).
NOTE: This version has two potential flaws. First, I do see any code
that verifies recieved TCP-MD5 signatures. Second, the TCP-MD5
options are internally padded and assumed to be 32-bit aligned. A more
space-efficient scheme is to pack all TCP options densely (and
possibly unaligned) into the TCP header ; then do one final padding to
a 4-byte boundary. Pre-existing comments note that accounting for
TCP-option space when we add SACK is yet to be done. For now, I'm
punting on that; we can solve it properly, in a way that will handle
SACK blocks, as a separate exercise.
In case a pullup to NetBSD-2 is requested, this adds sys/netipsec/xform_tcp.c
,and modifies:
sys/net/pfkeyv2.h,v 1.15
sys/netinet/files.netinet,v 1.5
sys/netinet/ip.h,v 1.25
sys/netinet/tcp.h,v 1.15
sys/netinet/tcp_input.c,v 1.200
sys/netinet/tcp_output.c,v 1.109
sys/netinet/tcp_subr.c,v 1.165
sys/netinet/tcp_usrreq.c,v 1.89
sys/netinet/tcp_var.h,v 1.109
sys/netipsec/files.netipsec,v 1.3
sys/netipsec/ipsec.c,v 1.11
sys/netipsec/ipsec.h,v 1.7
sys/netipsec/key.c,v 1.11
share/man/man4/tcp.4,v 1.16
lib/libipsec/pfkey.c,v 1.20
lib/libipsec/pfkey_dump.c,v 1.17
lib/libipsec/policy_token.l,v 1.8
sbin/setkey/parse.y,v 1.14
sbin/setkey/setkey.8,v 1.27
sbin/setkey/token.l,v 1.15
Note that the preceding two revisions to tcp.4 will be
required to cleanly apply this diff.
to pool_init. Untouched pools are ones that either in arch-specific
code, or aren't initialiased during initial system startup.
Convert struct session, ucred and lockf to pools.
closer to normal behaviour for the current century.
New Reno is now on by default (which is really the only reasonable
choice, since we don't do SACK); instead of an initial window of 1
for non-local nets, we now use Sally Floyd's magic 4K rule.
christos