to either enqueue or free their mbuf chains, but tcp_usr_send() was
dropping them on the floor if the tcpcb/inpcb has been torn down in the
middle of a send/write attempt. This has been responsible for a wide
variety of mbuf leak patterns, ranging from slow gradual leakage to rather
rapid exhaustion. This has been a problem since before 2.2 was branched
and appears to have been fixed in rev 1.16 and lost in 1.23/1.28.
Thanks to Jayanth Vijayaraghavan <jayanth@yahoo-inc.com> for checking
(extensively) into this on a live production 2.2.x system and that it
was the actual cause of the leak and looks like it fixes it. The machine
in question was loosing (from memory) about 150 mbufs per hour under
load and a change similar to this stopped it. (Don't blame Jayanth
for this patch though)
An alternative approach to this would be to recheck SS_CANTSENDMORE etc
inside the splnet() right before calling pru_send() after all the potential
sleeps, interrupts and delays have happened. However, this would mean
exposing knowledge of the tcp stack's reset handling and removal of the
pcb to the generic code. There are other things that call pru_send()
directly though.
Problem originally noted by: John Plevyak <jplevyak@inktomi.com>
The cdevsw_add() function now finds the major number(s) in the
struct cdevsw passed to it. cdevsw_add_generic() is no longer
needed, cdevsw_add() does the same thing.
cdevsw_add() will print an message if the d_maj field looks bogus.
Remove nblkdev and nchrdev variables. Most places they were used
bogusly. Instead check a dev_t for validity by seeing if devsw()
or bdevsw() returns NULL.
Move bdevsw() and devsw() functions to kern/kern_conf.c
Bump __FreeBSD_version to 400006
This commit removes:
72 bogus makedev() calls
26 bogus SYSINIT functions
if_xe.c bogusly accessed cdevsw[], author/maintainer please fix.
I4b and vinum not changed. Patches emailed to authors. LINT
probably broken until they catch up.
Reformat and initialize correctly all "struct cdevsw".
Initialize the d_maj and d_bmaj fields.
The d_reset field was not removed, although it is never used.
I used a program to do most of this, so all the files now use the
same consistent format. Please keep it that way.
Vinum and i4b not modified, patches emailed to respective authors.
(default 1) disables PMTUD globally. Although PMTUD can be disabled in
the standard case by locking the MTU on a static route (including the
default route), this method doesn't work in the face of dynamic routing
protocols like gated.
+ add a missing call to dn_rule_delete() when flushing firewall
rules, thus preventing possible panics due to dangling pointers
(this was already done for single rule deletes).
+ improve "usage" output in ipfw(8)
+ add a few checks to ipfw pipe parameters and make it a bit more
tolerant of common mistakes (such as specifying kbit instead of Kbit)
PR: kern/10889
Submitted by: Ruslan Ermilov
routines. The descriptor contains parameters which could be used
within those routines (eg. ip_output() ).
On passing, add IPPROTO_PGM entry to netinet/in.h
+ plug an mbuf leak when dummynet used with bridging
+ make prototype of dummynet_io consistent with usage
+ code cleanup so that now bandwidth regulation is precise to the
bit/s and not to (8*HZ) bit/s as before.
This is a seriously beefed up chroot kind of thing. The process
is jailed along the same lines as a chroot does it, but with
additional tough restrictions imposed on what the superuser can do.
For all I know, it is safe to hand over the root bit inside a
prison to the customer living in that prison, this is what
it was developed for in fact: "real virtual servers".
Each prison has an ip number associated with it, which all IP
communications will be coerced to use and each prison has its own
hostname.
Needless to say, you need more RAM this way, but the advantage is
that each customer can run their own particular version of apache
and not stomp on the toes of their neighbors.
It generally does what one would expect, but setting up a jail
still takes a little knowledge.
A few notes:
I have no scripts for setting up a jail, don't ask me for them.
The IP number should be an alias on one of the interfaces.
mount a /proc in each jail, it will make ps more useable.
/proc/<pid>/status tells the hostname of the prison for
jailed processes.
Quotas are only sensible if you have a mountpoint per prison.
There are no privisions for stopping resource-hogging.
Some "#ifdef INET" and similar may be missing (send patches!)
If somebody wants to take it from here and develop it into
more of a "virtual machine" they should be most welcome!
Tools, comments, patches & documentation most welcome.
Have fun...
Sponsored by: http://www.rndassociates.com/
Run for almost a year by: http://www.servetheweb.com/
1:
s/suser/suser_xxx/
2:
Add new function: suser(struct proc *), prototyped in <sys/proc.h>.
3:
s/suser_xxx(\([a-zA-Z0-9_]*\)->p_ucred, \&\1->p_acflag)/suser(\1)/
The remaining suser_xxx() calls will be scrutinized and dealt with
later.
There may be some unneeded #include <sys/cred.h>, but they are left
as an exercise for Bruce.
More changes to the suser() API will come along with the "jail" code.
- unifdef -DCOMPAT_IPFW (this was on by default already)
- remove traces of in-kernel ip_nat package, it was never committed.
- Make IPFW and DUMMYNET initialize themselves rather than depend on
compiled-in hooks in ip_init(). This means they initialize the same
way both in-kernel and as kld modules. (IPFW initializes now :-)
the IP header (this would not work for bridged packets).
This has been fixed long ago in the 2.2 branch.
Problem noticed by: a few people
Fix suggested by: Remy Nonnenmacher
also rely less on other modules clearing static values, and clear them
in a few cases we missed before.
Submitted by: Matthew Reimer <mreimer@vpop.net>
Move the Olicom token ring driver to the officially sanctionned location of
/sys/contrib. Also fix some brokenness in the generic token ring support.
Be warned that if_dl.h has been changed and SOME programs might
like recompilation.
- Transparent proxying support added.
- PPTP redirecting support added based on patches
contributed by Dru Nelson <dnelson@redwoodsoft.com>.
Submitted by: Charles Mott <cmott@srv.net>
their ttl). This can be used - in combination with the proper ipfw
incantations - to make a firewall or router invisible to traceroute
and other exploration tools.
This behaviour is controlled by a sysctl variable (net.inet.ip.stealth)
and hidden behind a kernel option (IPSTEALTH).
Reviewed by: eivind, bde
This is for various Olicom cards. An IBM driver is following.
This patch also adds support to tcpdump to decode packets on tokenring.
Congratulations to the proud father.. (below)
Submitted by: Larry Lile <lile@stdio.com>
This makes it possible to change the sysctl tree at runtime.
* Change KLD to find and register any sysctl nodes contained in the loaded
file and to unregister them when the file is unloaded.
Reviewed by: Archie Cobbs <archie@whistle.com>,
Peter Wemm <peter@netplex.com.au> (well they looked at it anyway)
convince myself that nothing will break if we permit IP input while
interface addresses are unconfigured. (At worst, they will hit some
ULP's PCB scan and fail if nobody is listening.) So, remove the restriction
that addresses must be configured before packets can be input. Assume
that any unicast packet we receive while unconfigured is potentially ours.
Divert was not feeding clean data to ifa_ifwithaddr() so it was
giving bad results.
Submitted by: kseel <kseel@utcorp.com>, Ruslan Ermilov <ru@ucb.crimea.ua>
This was missed in the 4.4-Lite2 merge.
Noticed by: Mohan Parthasarathy <Mohan.Parthasarathy@eng.Sun.COM> and
jayanth@loc201.tandem.com (vijayaraghavan_jayanth)
on the tcp-impl mailing list.
flag means that there is more data to be put into the socket buffer.
Use it in TCP to reduce the interaction between mbuf sizes and the
Nagle algorithm.
Based on: "Justin C. Walker" <justin@apple.com>'s description of Apple's
fix for this problem.
state.
Note: this requires a recompilation of netstat (but netstat has been
broken since rev 1.52 of ip_mroute.c anyway)
Obtained from: Significantly based on Steve McCanne's
<mccanne@cs.berkeley.edu> work for BSD/OS
arplookup() to try again. This gets rid of at least one user's
"arpresolve: can't allocate llinfo" errors, and arplookup() gives
better error messages to help track down the problem if there really
is a problem with the routing table.
problems reported recently (the rtentry pointer in the dummynet
queue was not initialized in all cases, resulting in spurious
rt_refcnt decreases in the lucky cases, and memory trashing in
other cases.
have all fields in network order, whereas ipfw expects some to be
in host order. This resulted in some incorrect matching, e.g. some
packets being identified as fragments, or bandwidth not being
correctly enforced.
NOTE: this only affects bridge+ipfw, normal ipfw usage was already
correct).
Reported-By: Dave Alden and others.
Add bounds checking to netbios NS packet resolving code. This should
prevent natd from crashing on badly formed netbios packets (as might be
heard when the machine is sitting on a cable modem or certain DSL
networks), and also closes potential security holes that might have
exploited the lack of bounds checking in the previous version of the
code.
If timer calculation results in degenerate value (0), force it to 1
to avoid divide-by-zero panic later on in calls to IGMP_RANDOM_DELAY().
I considered simply adding 1 to the timer calculation, but was unsure
if the calculation was part of the IGMP standard or not so did not want
to mess with it for all cases.
for possible buffer overflow problems. Replaced most sprintf()'s
with snprintf(); for others cases, added terminating NUL bytes where
appropriate, replaced constants like "16" with sizeof(), etc.
These changes include several bug fixes, but most changes are for
maintainability's sake. Any instance where it wasn't "immediately
obvious" that a buffer overflow could not occur was made safer.
Reviewed by: Bruce Evans <bde@zeta.org.au>
Reviewed by: Matthew Dillon <dillon@apollo.backplane.com>
Reviewed by: Mike Spengler <mks@networkcs.com>
option not defined the sysctl int value is set to -1 and read-only.
#ifdef KERNEL's added appropriately to wall off visibility of kernel
routines from user code.
Add ICMP_BANDLIM option and 'net.inet.icmp.icmplim' sysctl. If option
is specified in kernel config, icmplim defaults to 100 pps. Setting it
to 0 will disable the feature. This feature limits ICMP error responses
for packets sent to bad tcp or udp ports, which does a lot to help the
machine handle network D.O.S. attacks.
The kernel will report packet rates that exceed the limit at a rate of
one kernel printf per second. There is one issue in regards to the
'tail end' of an attack... the kernel will not output the last report
until some unrelated and valid icmp error packet is return at some
point after the attack is over. This is a minor reporting issue only.
when a TCP "stealth" scan is directed at a *BSD box by ensuring the window
is 0 for all RST packets generated through tcp_respond()
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com>
Obtained from: Bugtraq (from: Darren Reed <avalon@COOMBS.ANU.EDU.AU>)
This is the bulk of the support for doing kld modules. Two linker_sets
were replaced by SYSINIT()'s. VFS's and exec handlers are self registered.
kld is now a superset of lkm. I have converted most of them, they will
follow as a seperate commit as samples.
This all still works as a static a.out kernel using LKM's.
- Don't bother checking for conflicting sockets if we're binding to a
multicast address.
- Don't return an error if we're binding to INADDR_ANY, the conflicting
socket is bound to INADDR_ANY, and the conflicting socket has SO_REUSEPORT
set.
PR: kern/7713
addresses by default.
Add a knob "icmp_bmcastecho" to "rc.network" to allow this
behaviour to be controlled from "rc.conf".
Document the controlling sysctl variable "net.inet.icmp.bmcastecho"
in sysctl(3).
Reviewed by: dg, jkh
Reminded on -hackers by: Steinar Haug <sthaug@nethelp.no>
4.1.4. Experimental Protocol
A system should not implement an experimental protocol unless it
is participating in the experiment and has coordinated its use of
the protocol with the developer of the protocol.
Pointed out by: Steinar Haug <sthaug@nethelp.no>
another specialized mbuf type in the process. Also clean up some
of the cruft surrounding IPFW, multicast routing, RSVP, and other
ill-explored corners.
several new features are added:
- support vc/vp shaping
- support pvc shadow interface
code cleanup:
- remove WMAYBE related code. ENI WMAYBE DMA doen't work.
- remove updating if_lastchange for every packet.
- BPF related code is moved to midway.c as it should be.
(bpfwrite should work if atm_pseudohdr and LLC/SNAP are
prepended.)
- BPF link type is changed to DLT_ATM_RFC1483.
BPF now understands only LLC/SNAP!! (because bpf can't
handle variable link header length.)
It is recommended to use LLC/SNAP instead of NULL
encapsulation for various reasons. (BPF, IPv6,
interoperability, etc.)
the code has been used for months in ALTQ and KAME IPv6.
OKed by phk long time ago.
`len = min(so->so_snd.sb_cc, win) - off;'. min() has type u_int
and `off' has type int, so when min() is 0 and `off' is 1, the RHS
overflows to 0U - 1 = UINT_MAX. `len' has type long, so when
sizeof(long) == sizeof(int), the LHS normally overflows to to the
correct value of -1, but when sizeof(long) > sizeof(int), the LHS
is UINT_MAX.
Fixed some u_long's that should have been fixed-sized types.
mismatches exposed by this (the prototype for tcp_respond() didn't
match the function definition lexically, and still depends on a
gcc feature to match if ints have more than 32 bits).
Any packet that can be matched by a ipfw rule can be redirected
transparently to another port or machine. Redirection to another port
mostly makes sense with tcp, where a session can be set up
between a proxy and an unsuspecting client. Redirection to another machine
requires that the other machine also be expecting to receive the forwarded
packets, as their headers will not have been modified.
/sbin/ipfw must be recompiled!!!
Reviewed by: Peter Wemm <peter@freebsd.org>
Submitted by: Chrisy Luke <chrisy@flix.net>
Remove lots'o'hacks.
looutput is now static.
Other callers who want to use loopback to allow shortcutting
should call the special entrypoint for this, if_simloop(), which is
specifically designed for this purpose. Using looutput for this purpose
was problematic, particularly with bpf and trying to keep track
of whether one should be using the charateristics of the loopback interface
or the interface (e.g. if_ethersubr.c) that was requesting the loopback.
There was a whole class of errors due to this mis-use each of which had
hacks to cover them up.
Consists largly of hack removal :-)
or unsigned int (this doesn't change the struct layout, size or
alignment in any of the files changed in this commit, at least for
gcc on i386's. Using bitfields of type u_char may affect size and
alignment but not packing)).
ifdef tangle. The previous commit to ip_fil.h didn't change the
one that actually applies to the current FreeBSD kernel, of course.
Fixed.
Fixed style bugs in previous commit to ip_fil.h.
FreeBSD/alpha. The most significant item is to change the command
argument to ioctl functions from int to u_long. This change brings us
inline with various other BSD versions. Driver writers may like to
use (__FreeBSD_version == 300003) to detect this change.
The prototype FreeBSD/alpha machdep will follow in a couple of days
time.
so that the new behaviour is now default.
Solves the "infinite loop in diversion" problem when more than one diversion
is active.
Man page changes follow.
The new code is in -stable as the NON default option.
net.inet.ip.icmp.bmcastecho = 0 by removing the extra check for the
address being a multicast address. The test now relies on the link
layer flags that indicate it was received via multicast. The previous
logic was broken and replied to ICMP echo/timestamp broadcasts even
when the sysctl option disallowed them.
Reviewed by: wollman
Prior to this change, Accidental recursion protection was done by
the diverted daemon feeding back the divert port number it got
the packet on, as the port number on a sendto(). IPFW knew not to
redivert a packet to this port (again). Processing of the ruleset
started at the beginning again, skipping that divert port.
The new semantic (which is how we should have done it the first time)
is that the port number in the sendto() is the rule number AFTER which
processing should restart, and on a recvfrom(), the port number is the
rule number which caused the diversion. This is much more flexible,
and also more intuitive. If the user uses the same sockaddr received
when resending, processing resumes at the rule number following that
that caused the diversion. The user can however select to resume rule
processing at any rule. (0 is restart at the beginning)
To enable the new code use
option IPFW_DIVERT_RESTART
This should become the default as soon as people have looked at it a bit
passed to the user process for incoming packets. When the sockaddr_in
is passed back to the divert socket later, use thi sas the primary
interface lookup and only revert to the IP address when the name fails.
This solves a long standing bug with divert sockets:
When two interfaces had the same address (P2P for example) the interface
"assigned" to the reinjected packet was sometimes incorect.
Probably we should define a "sockaddr_div" to officially hold this
extended information in teh same manner as sockaddr_dl.
of the TCP payload. See RFC1122 section 4.2.2.6 . This allows
Path MTU discovery to be used along with IP options.
PR: problem discovered by Kevin Lahey <kml@nas.nasa.gov>
so it must be adjusted (minus 1) before using it to do the length check.
I'm not sure who to give the credit to, but the bug was reported by
Jennifer Dawn Myers <jdm@enteract.com>, who also supplied a patch. It
was also fixed in OpenBSD previously by andreas.gunnarsson@emw.ericsson.se,
and of course I did the homework to verify that the fix was correct per
the specification.
PR: 6738
NetBSD, ported to FreeBSD by Pierre Beyssac <pb@fasterix.freenix.org> and
minorly tweaked by me.
This is a standard part of FreeBSD, but must be enabled with:
"sysctl -w net.inet.ip.fastforwarding=1" ...and of course forwarding must
also be enabled. This should probably be modified to use the zone
allocator for speed and space efficiency. The current algorithm also
appears to lose if the number of active paths exceeds IPFLOW_MAX (256),
in which case it wastes lots of time trying to figure out which cache
entry to drop.
Define a parameter which indicates the maximum number of sockets in a
system, and use this to size the zone allocators used for sockets and
for certain PCBs.
Convert PF_LOCAL PCB structures to be type-stable and add a version number.
Define an external format for infomation about socket structures and use
it in several places.
Define a mechanism to get all PF_LOCAL and PF_INET PCB lists through
sysctl(3) without blocking network interrupts for an unreasonable
length of time. This probably still has some bugs and/or race
conditions, but it seems to work well enough on my machines.
It is now possible for `netstat' to get almost all of its information
via the sysctl(3) interface rather than reading kmem (changes to follow).
---------
Make callers of namei() responsible for releasing references or locks
instead of having the underlying filesystems do it. This eliminates
redundancy in all terminal filesystems and makes it possible for stacked
transport layers such as umapfs or nullfs to operate correctly.
Quality testing was done with testvn, and lat_fs from the lmbench suite.
Some NFS client testing courtesy of Patrik Kudo.
vop_mknod and vop_symlink still release the returned vpp. vop_rename
still releases 4 vnode arguments before it returns. These remaining cases
will be corrected in the next set of patches.
---------
Submitted by: Michael Hancock <michaelh@cet.co.jp>
is believed to have been broken with the Brakmo/Peterson srtt
calculation changes. The result of this bug is that TCP connections
could time out extremely quickly (in 12 seconds).
Also backed out jdp's partial fix for this problem in rev 1.17 of
tcp_timer.c as it is obsoleted by this commit.
Bug was pointed out by Kevin Lehey <kml@roller.nas.nasa.gov>.
PR: 6068
ftp://ftp.isi.edu/in-notes/iana/assignments/port-numbers
port numbers are divided into three ranges:
0 - 1023 Well Known Ports
1024 - 49151 Registered Ports
49152 - 65535 Dynamic and/or Private Ports
This patch changes the "local port range" from 40000-44999
to the range shown above (plus fix the comment in in_pcb.c).
WARNING: This may have an impact on firewall configurations!
PR: 5402
Reviewed by: phk
Submitted by: Stephen J. Roznowski <sjr@home.net>
"time" wasn't a atomic variable, so splfoo() protection were needed
around any access to it, unless you just wanted the seconds part.
Most uses of time.tv_sec now uses the new variable time_second instead.
gettime() changed to getmicrotime(0.
Remove a couple of unneeded splfoo() protections, the new getmicrotime()
is atomic, (until Bruce sets a breakpoint in it).
A couple of places needed random data, so use read_random() instead
of mucking about with time which isn't random.
Add a new nfs_curusec() function.
Mark a couple of bogosities involving the now disappeard time variable.
Update ffs_update() to avoid the weird "== &time" checks, by fixing the
one remaining call that passwd &time as args.
Change profiling in ncr.c to use ticks instead of time. Resolution is
the same.
Add new function "tvtohz()" to avoid the bogus "splfoo(), add time, call
hzto() which subtracts time" sequences.
Reviewed by: bde
