deal with shortages of the VM maps where the backing pages are mapped
(usually kmem_map). Try to deal with this:
* Group all information about the backend allocator for a pool in a
separate structure. The pool references this structure, rather than
the individual fields.
* Change the pool_init() API accordingly, and adjust all callers.
* Link all pools using the same backend allocator on a list.
* The backend allocator is responsible for waiting for physical memory
to become available, but will still fail if it cannot callocate KVA
space for the pages. If this happens, carefully drain all pools using
the same backend allocator, so that some KVA space can be freed.
* Change pool_reclaim() to indicate if it actually succeeded in freeing
some pages, and use that information to make draining easier and more
efficient.
* Get rid of PR_URGENT. There was only one use of it, and it could be
dealt with by the caller.
From art@openbsd.org.
Add capabilities bits that indicate an interface can only perform
in-bound TCPv4 or UDPv4 checksums. There is at least one Gig-E chip
for which this is true (Level One LXT-1001), and this is also the
case for the Intel i82559 10/100 Ethernet chips.
all open TCP connections in tcp_slowtimo() (which is called 2x
per second). It's fairly rare for TCP timers to actually fire,
so saving this list traversal is good, especially if you want
to scale to thousands of open connections.
Instead of incrementing t_idle and t_rtt in tcp_slowtimo(), we now
take a timstamp (via tcp_now) and use subtraction to compute the
delta when we actually need it (using unsigned arithmetic so that
tcp_now wrapping is handled correctly).
Based on similar changes in FreeBSD.
network interfaces. This works by pre-computing the pseudo-header
checksum and caching it, delaying the actual checksum to ip_output()
if the hardware cannot perform the sum for us. In-bound checksums
can either be fully-checked by hardware, or summed up for final
verification by software. This method was modeled after how this
is done in FreeBSD, although the code is significantly different in
most places.
We don't delay checksums for IPv6/TCP, but we do take advantage of the
cached pseudo-header checksum.
Note: hardware-assisted checksumming defaults to "off". It is
enabled with ifconfig(8). See the manual page for details.
Implement hardware-assisted checksumming on the DP83820 Gigabit Ethernet,
3c90xB/3c90xC 10/100 Ethernet, and Alteon Tigon/Tigon2 Gigabit Ethernet.
ISS attacks (which we already fend off quite well).
1. First-cut implementation of RFC1948, Steve Bellovin's cryptographic
hash method of generating TCP ISS values. Note, this code is experimental
and disabled by default (experimental enough that I don't export the
variable via sysctl yet, either). There are a couple of issues I'd
like to discuss with Steve, so this code should only be used by people
who really know what they're doing.
2. Per a recent thread on Bugtraq, it's possible to determine a system's
uptime by snooping the RFC1323 TCP timestamp options sent by a host; in
4.4BSD, timestamps are created by incrementing the tcp_now variable
at 2 Hz; there's even a company out there that uses this to determine
web server uptime. According to Newsham's paper "The Problem With
Random Increments", while NetBSD's TCP ISS generation method is much
better than the "random increment" method used by FreeBSD and OpenBSD,
it is still theoretically possible to mount an attack against NetBSD's
method if the attacker knows how many times the tcp_iss_seq variable
has been incremented. By not leaking uptime information, we can make
that much harder to determine. So, we avoid the leak by giving each
TCP connection a timebase of 0.
- let ipfilter look at wire-format packet only (not the decapsulated ones),
so that VPN setting can work with NAT/ipfilter settings.
sync with kame.
TODO: use header history for stricter inbound validation
ip_output(). This flag, if set, causes ip_output() to set
DF in the IP header if the MTU in the route is not locked.
This allows a bunch of redundant code, which I was never
really all that happy about adding in the first place, to
be eliminated.
Inspired by a similar change made by provos@openbsd.org when
he integrated NetBSD's Path MTU Discovery code into OpenBSD.
basis. default: 100pps
set default value for net.inet.tcp.rstratelimit to 0 (disabled),
NOTE: it does not work right for smaller-than-1/hz interval. maybe we should
nuke it, or make it impossible to set smaller-than-1/hz value.
unspecified address (::) to mean "unbounded" or "unconnected",
and can be confused by packets from outside.
use of :: as source is not documented well in IPv6 specification.
not sure if it presents a real threat. the worst case scenario is a DoS
against TCP listening socket:
- outsider transmit TCP SYN with :: as IPv6 source
- receiving side creates TCP control block with:
local address = my addres
remote address = :: (meaning "unconnected")
state = SYN_RCVD
note that SYN ACK will not be sent due to ip6_output() filter.
this stays until it timeouts.
- the TCP control block prevents listening TCP control block from
being contacted (DoS).
udp6/raw6 socket may have similar problem, but as they are connectionless,
it may too much to filter it out.
- add protection mechanism against ND cache corruption due to bad NUD hints.
- more stats
- icmp6 pps limitation. TOOD: should implement ppsratecheck(9).
between protocol handlers.
ipsec socket pointers, ipsec decryption/auth information, tunnel
decapsulation information are in my mind - there can be several other usage.
at this moment, we use this for ipsec socket pointer passing. this will
avoid reuse of m->m_pkthdr.rcvif in ipsec code.
due to the change, MHLEN will be decreased by sizeof(void *) - for example,
for i386, MHLEN was 100 bytes, but is now 96 bytes.
we may want to increase MSIZE from 128 to 256 for some of our architectures.
take caution if you use it for keeping some data item for long period
of time - use extra caution on M_PREPEND() or m_adj(), as they may result
in loss of m->m_pkthdr.aux pointer (and mbuf leak).
this will bump kernel version.
(as discussed in tech-net, tested in kame tree)
- Filter out multicast destinations explicitly for every incoming packet,
not just SYNs. Previously, non-SYN multicast destination would be
filtered out as a side effect of PCB lookup. Remove now redundant
similar checks in the dropwithreset case and in syn_cache_add().
- Defer the TCP checksum until we know that we want to process the
packet (i.e. have a non-CLOSED connection or a listen socket).
- interop issues in ipcomp is fixed
- padding type (after ESP) is configurable
- key database memory management (need more fixes)
- policy specification is revisited
XXX m->m_pkthdr.rcvif is still overloaded - hope to fix it soon
due to massive changes in KAME side.
- IPv6 output goes through nd6_output
- faith can capture IPv4 packets as well - you can run IPv4-to-IPv6 translator
using heavily modified DNS servers
- per-interface statistics (required for IPv6 MIB)
- interface autoconfig is revisited
- udp input handling has a big change for mapped address support.
- introduce in4_cksum() for non-overwriting checksumming
- introduce m_pulldown()
- neighbor discovery cleanups/improvements
- netinet/in.h strictly conforms to RFC2553 (no extra defs visible to userland)
- IFA_STATS is fixed a bit (not tested)
- and more more more.
TODO:
- cleanup os-independency #ifdef
- avoid rcvif dual use (for IPsec) to help ifdetach
(sorry for jumbo commit, I can't separate this any more...)
MSS advertisement must always be:
max(if mtu) - ip hdr siz - tcp hdr siz
We violated this in the previous code so it was fixed.
tcp_mss_to_advertise() now takes af (af on wire) as its argument,
to compute right ip hdr siz.
tcp_segsize() will take care of IPsec header size.
One thing I'm not really sure is how to handle IPsec header size in
*rxsegsizep (inbound segment size estimation).
The current code subtracts possible *outbound* IPsec size from *rxsegsizep,
hoping that the peer is using the same IPsec policy as me.
It may not be applicable, could TCP gulu please comment...
New Reno fast recovery code was being executed even when New Reno was
disabled, resulting in an unfortunate interaction with the traditional
fast recovery code, the end resulting being that the very condition
that would trigger the traditional fast recovery mechanism caused fast
recovery to be disabled!
Problem reported by Ted Lemon, and some analytical help from Charles Hannum.
Stale syn cache entries are useless because none of them will be used
if there is no listening socket, as tcp_input looks up listening socket by
in_pcblookup*() before looking into syn cache.
This fixes race condition due to dangling socket pointer from syn cache
entries to listening socket (this was introduced when ipsec is merged in).
This should preserve currently implemented behavior (but not 4.4BSD
behavior prior to syn cache).
Tested in KAME repository before commit, but we'd better run some
regression tests.
- Make sure that snd_recover is always at least snd_una. If we don't do
this, there can be confusion when sequence numbers wrap around on a
large loss-free data transfer.
- When doing a New Reno retransmit, snd_una hasn't been updated yet,
and the socket's send buffer has not yet dropped off ACK'd data, so
don't muddle with snd_una, so that tcp_output() gets the correct data
offset.
- When doing a New Reno retransmit, make sure the congestion window is
open one segment beyond the ACK'd data, so that we can actually perform
the retransmit.
Partially derived from, although more complete than, similar changes in
OpenBSD, which in turn originated from Tom Henderson <tomh@cs.berkeley.edu>.
when ip header and tcp header are not adjacent to each other
(i.e. when ip6 options are attached).
To test this, try
telnet @::1@::1 port
toward a port without responding server. Prior to the fix, the kernel will
generate broken RST packet.
(Sorry for a big commit, I can't separate this into several pieces...)
Pls check sys/netinet6/TODO and sys/netinet6/IMPLEMENTATION for details.
- sys/kern: do not assume single mbuf, accept chained mbuf on passing
data from userland to kernel (or other way round).
- "midway" ATM card: ATM PVC pseudo device support, like those done in ALTQ
package (ftp://ftp.csl.sony.co.jp/pub/kjc/).
- sys/netinet/tcp*: IPv4/v6 dual stack tcp support.
- sys/netinet/{ip6,icmp6}.h, sys/net/pfkeyv2.h: IETF document assumes those
file to be there so we patch it up.
- sys/netinet: IPsec additions are here and there.
- sys/netinet6/*: most of IPv6 code sits here.
- sys/netkey: IPsec key management code
- dev/pci/pcidevs: regen
In my understanding no code here is subject to export control so it
should be safe.
where one side can think a connection exists, where the other side thinks
the connection was never established.
The original problem was first reported by Ty Sarna in PR #5909. The
original fix I made to the code didn't cover all cases. The problem this
fix addresses was reported by Christoph Badura via private e-mail.
Many thanks to Bill Sommerfeld for helping me to test this code, and
for finding a subtle bug.
- Don't use tcp_respond(), instead create the tcp/ip header from scratch,
and send it ourself.
- Reuse the mbuf that carried the SYN, or allocate one if that is not
available.
- Cache the route we look up to do the Path MTU Discovery check, and
transfer the reference to that route to the inpcb when the connection
completes.
* Macro'ize a small, but often repeated code fragment.
syn_cache_unreach() should remove the entry, or just continue on.
Algorithm is to only remove the entry if we've had more than one unreach
error and have retransmitted 3 or more times. This prevents the following
scenario, as noted in PR #5909 (PR from Ty Sarna, scenario from
Charles Hannum):
* Host A sends a SYN.
* Host A retransmits the SYN.
* Host B gets the first SYN and sends a SYN-ACK.
* Host B gets the second SYN and sends a SYN-ACK.
* One of the SYN-ACK bounces with an
ICMP unreachable, causing the `SYN cache' entry to be
removed with no notification.
* Host A receives the other SYN-ACK, sends an ACK, and goes to
ESTABLISHED state.
Should fix PR #5909.
- Don't use home-grown queue manipulation. Use <sys/queue.h> instead. The
data structures are a little larger, but we are otherwise wasting the
memory chunk anyway (we're already a 64-byte malloc bucket).
- Fix a bug in the cache-is-full case: if the oldest element removed from
the first non-empty bucket was the only element in the bucket, the
bucket wouldn't be removed from the bucket cache, causing queue corruption
later.
- Optimize the syn cache timers by using PRT timers rather than home-grown
decrement-and-propagate timers.
This code is now a fair bit smaller, and significantly easier to read
and understand.
rule was to update the timestamp if the sequence numbers are in range. New
rule adds a check that the timestamp is advancing, thus preventing our notion
of the most recent timestamp from incorrectly moving backwards.
TCP connections by using the MTU of the interface. Also added
a knob, mss_ifmtu, to force all connections to use the MTU of
the interface to calculate the advertised MSS.
to ACK immediately any packet that arrived with PSH set. This breaks
delayed ACKs in a few specific common cases that delayed ACKs were
supposed to help, and ends up not making much (if any) difference in
the case where where this ACK-on-PSH change was supposed to help.
Per discussion with several members of the TCPIMPL and TCPSAT IETF
working groups.
code, as clarified in the TCPIMPL WG meeting at IETF #41: If the SYN
(active open) or SYN,ACK (passive open) was retransmitted, the initial
congestion window for the first slow start of that connection must be
one segment.
RTO estimation changes. Under some circumstances it would return a value
of 0, while the old Van Jacobson RTO code would return a minimum of 3.
This would result in 12 retransmissions, each 1 second apart.
This takes care of those instances, and ensures that t_rttmin is
used everywhere as a lower bound.
case. Sending an RST to ourselves is a little silly, considering that
we'll just attempt to remove a non-existent compressed state entry and
then drop the packet anyway.
socket:
- If we received a SYN,ACK, send an RST.
- If we received a SYN, and the connection attempt appears to come from
itself, send an RST, since it cannot possibly be valid.
- Don't overload t_maxseg. Previous behavior was to set it to the min
of the peer's advertised MSS, our advertised MSS, and tcp_mssdflt
(for non-local networks). This breaks PMTU discovery running on
either host. Instead, remember the MSS we advertise, and use it
as appropriate (in silly window avoidance).
- Per last bullet, split tcp_mss() into several functions for handling
MSS (ours and peer's), and performing various tasks when a connection
becomes ESTABLISHED.
- Introduce a new function, tcp_segsize(), which computes the max size
for every segment transmitted in tcp_output(). This will eventually
be used to hook in PMTU discovery.
which happen to have a TCB in TIME_WAIT, where an mbuf which had been
advanced past the IP+TCP headers and TCP options would be reused as if
it had not been advanced. Problem found by Juergen Hannken-Illjes, who
also suggested a work-around on which this fix is based.
fixed in FreeBSD by John Polstra:
Fix a bug (apparently very old) that can cause a TCP connection to
be dropped when it has an unusual traffic pattern. For full details
as well as a test case that demonstrates the failure, see the
referenced PR (FreeBSD's kern/3998).
Under certain circumstances involving the persist state, it is
possible for the receive side's tp->rcv_nxt to advance beyond its
tp->rcv_adv. This causes (tp->rcv_adv - tp->rcv_nxt) to become
negative. However, in the code affected by this fix, that difference
was interpreted as an unsigned number by max(). Since it was
negative, it was taken as a huge unsigned number. The effect was
to cause the receiver to believe that its receive window had negative
size, thereby rejecting all received segments including ACKs. As
the test case shows, this led to fruitless retransmissions and
eventually to a dropped connection. Even connections using the
loopback interface could be dropped. The fix substitutes the signed
imax() for the unsigned max() function.
Bill informs me that his research indicates this bug appeared in Reno.
(like the alpha). Biggest problem: IP headers were overlayed with
structure which included pointers, and which therefore didn't overlay
properly on 64-bit machines. Solution: instead of threading pointers
through IP header overlays, add a "queue element" structure to do
the threading, and point it at the ip headers.
* Convert several data structures to use queue.h.
* Split in_pcbnotify() into two parts; one for notifying a specific PCB, and
one for notifying all PCBs for a particular foreign address.
* Don't add the extra 1/8 of the mss when ramping up the congestion window.
* Scale the RTT values slightly to adjust for rounding errors.
* Set the lower bound of the RTO to RTT+2.
I've found a problem with the TCP delayed ack algorithm. If the writer's
buffer becomes full before sending an entire window, the writer will stop
and the ack will be delayed and the transmission will be stalled pending
a timeout on (and transmission of) the delayed ack.
As an experiment, I've applied the following patch to my (NetBSD) kernel,
and it alleviates the problem.
The worst case for this change is that the writer sets the PSH bit on
every outgoing packet, in which case delayed ack is effectively disabled.
This is not an issue of correctness, however, and since most vendors use
the PSH bit a bit more intelligently, it doesn't seem like a serious
problem.
christos