Struct bundle will have its own struct ccp in the future
too.
o The ``set stopped'' command now requires context and doesn't
work on the IPCP FSM.
o Check if it's time to break out of our top level loop before
doing a select - otherwise, we'll select forever :-(
o Remove `struct link'::ccp (a temporary hack). It turns out
that IpStartOutput() calls link_Output() and link_Output()
incorrectly calls StartOutput() (really modem_StartOutput)
requiring the ccp knowledge so that it can call
IpStartOutput()... The end result is that the whole IP
output queue gets dumped into the modem output queue
and a pile of physical writes are done prematurely. This
makes the (original) code in main() actually work in that
it would not bother selecting() on the tun descriptor when
our modem queue length was 20 or greater. Instead, we now
make that decision based on the overall queue length.
This will need improvement later.
packets when we haven't agreed a protocol).
o Move the complication of passing incoming data to the
PROTO_COMPD input or dictionary setup routine into ccp.c
o Move our LCP report timer into struct hdlc - it's really
a hdlc timer (fcs errors etc).
o Make `show hdlc' require context and make the output more
friendly.
o Remove all non-const globals from hdlc.c
o Output peer-rejected protocols by name - not just ones
that we reject.
is available, but LCP hasn't yet been started. We get to this
state in ``term'' mode.
Remove PacketMode(). LCP startup and shutdown is now controlled
by the datalink.
Add ``show links'' command.
Make ``close'' capable of running with and without a context.
Make ``down'' require a context.
Make ``set parity'' and ``set rtscts'' use the correct context.
datalink_Up() can now be told to skip the dial/login/hangup
scripts and can be told whether to enter packet mode when
entering the DATALINK_OPENED state.
This is a type of physical link that can chat and talk
LCP & CCP. A bundle contains a list of these (only one
in the list for the moment).
The datalink is a type of descriptor, and dials, enters
LCP (& does CCP), kicks the bundle when its FSMs do
something interesting and does the hangup chat script
on the way down. It also handles redials and reconnects.
There are lots of loose ends, and probably lots of bugs,
but the data structures are getting there !
Allow for NULL fd_sets in descriptor_UpdateSet()
Reimplement the entire chat module, creating
`struct chat' - a `type' of struct descriptor.
Remove CARRIER logging.
CONNECT logging now only logs "CONNECT" lines. CHAT logging
masks it with an entire log of the conversation.
Modem dialing is now asynchronous, including pauses
and timeouts :-)
The hooks in DoLoop() in main.c are *very* messy ! I'll have
to rewrite DoLoop fairly soon, so I don't care too much for the
moment. This code is pretty raw.
may result in a our modem closing after it's made its way into
the fd_set, resulting in a program exit (with select(): bad file
descriptor) rather than a dropped link.
This will ultimately be a member of a list of descriptors and
their handler functions on which we need to select() in the
main loop.
o Make struct physical into a `sort' of struct descriptor.
o Don't remove routes and DOWN the interface when we're
closed in auto mode.
o Initialize the FSMs in bundle_Create.
o Initialize ipcp::if_mine & ipcp::if_peer only once (so
that we don't forget that we've SIOCAIFADDR'd the interface).
o Do a SIOCDIFADDR on the specific address to avoid hurting any
other (still non-existent) NCPs.
o Fix some error/diagnostic messages.
o The FSM layering is now more sane.
o Move a lot of the NCP stuff into our ipcpstate rather than having it
in the bundle, including control of the configured IP addresses. We
don't need hacks like the global `linkup' variable any more as the
FSM decides when our ppp.link* files get run. This is going to eventually
be configurable based on FSM events anyway.
o Fix a few inconsistencies when both sides require authentication.
o We now have "Ppp..." and "PPp" prompts, reflecting authentication
and network phase. We don't print loads of spurious prompts as we
change phases any more.
o Our phase is part of the bundle now.
o Fix a bug where the FSM wasn't calling LayerFinish.
o Close the FSM down correctly with a signal rather than slamming it
down as if the line was dropped (the undocumented ``down'' command
is still available though).
o Remove the forgotten `tunno' variable and fix references to it.
This structure contains the asynchronous state of the physical
link.
Unfortunately, just about every .h file is included in every .c
file now. Fixing this can be one of the last jobs.
This structure will eventually contain a list of NCPs (currently
only IPCP is supported) and a list of physical `struct link's.
It will also derive from a struct link itself.
Make ModemTimeout() static - it's way to dangerous to be called
from outside !
Bump version to 1.9. Our first MP release should be 2.0.
IPCP, CCP and LCP are now just derived FSMs.
Comment each of the FSM implementations so that we can
tell what's going on.
Revise the state transitions so that CCP and IPCP actually
send terminate REQs when appropriate.
The OS & IPCP layers are still like spagetti (next job).
o Use INADDR_ANY and INADDR_BROADCAST instead of 0 and 0xffffffff
for IP addresses & masks.
o Move struct compreq into ipcp.c (it's none of anyone elses
business).
o Rename MAX_STATES to MAX_VJ_STATES.
o Take note of the return value of the compression layers
init routine, and if it fails, bring the whole layer
cleanly back down again.
o Only use srandomdev() if ``__FreeBSD__ > 2'' - this makes
these sources buildable on 2.2 (where I'm testing).
Don't mention the authors name at startup. He's already credited
in the man page. Instead, make the message consistent with the
one given to the diagnostic port (and fix the grammar when entering
`term' mode).
Don't credit the zlib author in the man page as ppp isn't linked
directly with zlib (it's shared).
Mention when the OpenBSD port was first made available.
The delay defaults to 1 sec (as it always has) unless we've done
a ~p in interactive mode or we've actually detected a HDLC frame.
This is now cleanly implemented (via async timers) so that it is
possible for LCP to come up despite the delay if an LCP REQ is
received.
This will hopefully solve situations with slow servers or slirp
scenarios (where ECHO is left on the port for a second or so before
the peer enters packet mode).
Also, ~p in interactive mode no longer changes the value of the default
openmode delay and -dedicated mode enters packet mode in the right state
according to the value of openmode.
When building a release, RELEASE_CRUNCH is defined for a `make' of
the objects required by the crunch of each program. The object list
is still obtained in the same way, so you must make sure that all
objects are built (empty if necessary) by this make. ppp/Makefile
provides an example.
Reviewed by: jkh
(I completely mis-read the rfc last time 'round!)
This means:
o Better CCP/WARN Reset diagnostics.
o After we've sent a REQ and before we've received an ACK, we drop
incoming compressed data and send another REQ.
o Before sending an ACK, re-sequence all pending PRI_NORMAL data in
the modem queue so that pending packets won't get to the peer
*after* the ResetAck.
o Send ACKs with the `identifier' from the REQ frame.
o After we've received a correct ACK, duplicate ACKs are ok (and will
reset our history).
o Incorrect ACKs (not matching the last REQ) are moaned about and dropped.
Also,
o Calculate the correct FCS after compressing a packet. DEFLATE
*may* produce an mbuf with more than a single link in the chain,
but HdlcOutput didn't know how to calculate the FCS :-(
o Make `struct fsm'::reqid a u_char, not an int.
This fix will prevent us from sending id `255' 2,000,000,000 times
before wrapping to `0' for another 2,000,000,000 sends :-/
o Bump the version number a little.
The end result: DEFLATE now works over an unreliable link layer.
I can txfr a 1.5Mb kernel over a (rather bad) null-modem
cable at an average of 21679 bytes per second using rcp.
Repeat after me: Don't test compression using a loopback ppp/tcp setup as
we never lose packets and therefore never have to reset!
o Allow a forth argument in ppp.secret, specifying a new
label. This gives control over which section of
ppp.link{up,down} is used based on the authenticated user.
o Support random address ranges in ppp.secret (not just in ppp.conf).
o Add a AUTHENTICATING INCOMING CONNECTIONS section to the man page.
o Add a bit more about DEFLATE in the man page.
o Fix the incorrect "you must specify a password in interactive
mode" bit of the manual.
o Space things in the man page consistently.
o Be more precice about where you can use MYADDR, HISADDR and INTERFACE
in the "add" command documentation.
not in -auto mode isn't a good idea, and that the
add should be done in ppp.linkup instead.
Change "add 0 0 HISADDR" to "add default HISADDR". It's
more intuitive.
interactive mode.
Use `netfd' in fcntl() and tc[gs]etattr() calls rather than
the hard coded descriptor 0.
Use _FILENO constants from unistd.h
This un-breaks things after my recent `close(0)' in interactive
mode.
Close STDIN_FILENO, and open _PATH_TTY O_RDONLY as `netfd'. This
has the effect of allowing `show route' to output more than about
a page of data (on FreeBSD, not OpenBSD....). I have no idea why,
except that it was a direct consequence of the tcsetattr() in
TtyCommandMode(). My previous fix (closing descriptor 0) `fixed'
this because all calls to tcsetattr() failed :-(
RTM_CHANGE if the RTM_ADD fails with an EEXIST.
Allow "delete! dst" (note the ``!'') to silently
fail if the RTM_DELETE fails with an ESRCH.
Also, make the ESRCH and EEXIST error conditions
more understandable to the casual observer.
interrupted with a SIGALRM. In fact, select() sets the
passed time to zero, making the previous implementation
terminate always after 1/10th of a second !
Also, deal with someone changing the clock while we're
sleeping (and restart the whole sleep).
Dangers pointed out by: Theo de Raadt <deraadt@cvs.openbsd.org>
When CCP is originally negotiated, the only thing we can be
sure about is that we've started adding data to the inflate
dictionary either before or at the same time as the peer. This
is ok, 'cos DEFLATE is a `sliding window' compressor.
Show the IP range (if specified) in "show ipcp".
Close unused descriptors 0 and 2 in interactive mode.
Pass (size_t *) rather than (int *) to sysctl().
strcpy(a, b); /* a and b are the same size */
with
strncpy(a, b, sizeof(a));
a[sizeof(a)-1] = '\0';
Making the code `correct at a glance'.
Suggested by: Theo de Raadt <deraadt@cvs.openbsd.org>
Disallow connections from port numbers <1024 as low port numbers
can only really mean trouble (ftp bounces etc.).
Discussed at length with: Theo de Raadt <deraadt@cvs.openbsd.org>
fd_set, make sure that all descriptors >2 are closed
when we start - otherwise we're asking for a dump in
FD_SET().
Problem pointed out by: Theo de Raadt <deraadt@cvs.openbsd.org>
when command logging is switched on.
o Display ******** for the authkey for "show auth"
o Document how \P should be used, and document the other chat escapes
while I'm there.
o Make sure the full command is displayed when a compound command
fails - ie, "set novar rubbish" should say "set novar: Invalid command"
rather than "novar: Invalid command"
Problem pointed out by: Theo de Raadt <deraadt@cvs.openbsd.org> (among others)
line is > LINE_LEN (512 bytes), we scribble (*blush*).
Hinted at by: Theo de Raadt <deraadt@cvs.openbsd.org>
Change sprintf(buf, "fixedstring") to strcpy(buf, "fixedstring").
all RTF_WASCLONED routes, and the second removes the
others. This avoids the situation where we've added an
RTF_CLONING route (such as ``default''), created some
clones, then deleted the CLONING route before the
WASCLONED route(s). Without the two passes, we get
errno (not rtm_errno) set to ESRCH when deleting the
WASCLONED route, despite the deletion succeeding !
Also:
Enhance the route operation failure diagnostics.
Make portability #ifdefs a bit more generic.
Validate the peers suggested IP by attempting to make a routing table
entry.
Give up IPCP negotiation if the peer NAKs us with an unusable IP.
Always SIOCDIFADDR then SIOCAIFADDR when configuring the tun device.
Using SIOCSIFDSTADDR allows duplicate dst addresses (which we don't
want)!!!
Allow up to 200 interface names (was 50) (now that ppp can play server
properly).
Up the version number (1.5 -> 1.6).
Cosmetic:
Log unexpected CCP packets in the CCP log rather than the ERROR log.
Log unexpected Config Reqs in the appropriate LCP/IPCP/CCP log rather
than the ERROR log.
Log failed route additions and deletions with WARN, not TCPIP.
Log the option id and length for unrecognised IPCP options.
Change some .Sq to .Ar in the man page.
We must call inflate again in case there's any pending output
despite our input buffer being empty. If the output buffer
is in fact already flushed, inflate() returns Z_BUF_ERROR.
There isn't really an error !
I did the deflate re-org).
Make PAP & CHAP negotiation prettier in the log file.
If both PAP & CHAP are `enabled' and the peer NAKs CHAP
suggesting PAP, be friendly and REQ PAP the next time.
This is in line with the rfc.
Delete AF_LINK routes as well as AF_INET.
Allow the word `default' as the arg to `delete' or in place of the
first two args (dest & netmask) to `add'.
Accept INTERFACE as the third arg to `add'.
You can now say `add default interface' to create a default route
through the tun interface. It's reported that subsequent bind()s
will bind to a broadcast address and not to the address currently
assigned to the tun device - this is the first step towards
supporting that first connection that was around from before the
dynamic IP negotiation....
(I *really* meant to do this *before* committing the
deflate changes in the first place - oops).
Pppd is horribly broken in this respect - refer to the
ppp man page for details. Ppp *WON'T* negotiate deflate
with pppd by default - you must ``enable'' and ``accept''
``pppd-deflate'' in your config.
While I'm in there, update the cftypes in ccp.c so that
we recognise some more protocols (we don't actually do
anything with them - just send a REJ).
Remove extraneous decls.
Add ``const'' to several places.
Allow ``make NOALIAS=1'' to remove IP aliasing.
Merge with OpenBSD - only the Makefiles vary.
We can now survive a compile with
-Wall -Wbad-function-cast -Wcast-align -Wcast-qual
-Winline -Wmissing-declarations -Wmissing-prototypes
-Wnested-externs -Wpointer-arith -Wredundant-decls
-Wshadow -Wstrict-prototypes -Wwrite-strings -Wchar-subscripts
(although the Makefile just contains -Wall).
avoids the situation where we specify label ``x'' on the command
line, and label ``x'' has a ``load y'' command embedded in it.
When the line comes up, we want to use ``x'' from ppp.linkup,
not ``y''.
allowing for a possible header on the front of all packets.
In OpenBSD, there's a structure containing the address
family here.
If we're building under OpenBSD, set up the ``flags'' part
of struct tuninfo (not there under FreeBSD) so that we config
the interface as POINTOPOINT.
Prefix prototypes with ``extern'' in os.c for consistency.
These changes are cosmetic under FreeBSD, but allow ppp to
build & work under OpenBSD (bar the srandomdev() stuff,
the inclusing of <net/if_var.h> and some Makefile symantecs).
Print out inteface names rather than numbers, and use the
same "find out the names" code in SetIfAddr(). This means
that the nasty ioctl(,SIOCGIFCONF,)/realloc loop is now
buried :-)
o Add "allow" command:
"allow users a b c" gives access to users a, b and c.
"allow modes auto" gives those users access to auto mode only.
"allow users *" and "allow modes *" are accepted.
No users and all modes are allowed by default.
UID 0 can do anything.
o Set the current label with the "load" and "dial" commands
so that the call to ppp.linkdown makes sense.
o Up the verison number.
o Don't OR MODE_AUTO for -background and -ddial.
o Don't OR MODE_INTER when we get a diagnostic connection.
o Allow up to 40 args per line (was 20).
o "set ifaddr" only changes the interface in AUTO mode (with other
modes, it happens after IPCP negotiation).
o Sort command descriptions in the man page.
o Support -dedicated mode where we just talk ppp forever (no login etc).
Stay as the invoking uid as much as possible.
Execution as a normal user is still forbidden for now,
so these changes are pretty ineffective.
The next commit will implement the modifications suggested
on -hackers a number of days ago.
ppp.secret to get a key value based on the peers name,
then send the value of AuthName with the encrypted result.
This is *way* too confusing.
Instead, always use AuthName and AuthKey as the documentation
says.
o Report modem connect time properly
o Report bytes in/out over physical media
o Fix phases (TERMINATE is *higher than* DEAD)
o Do a LayerFinish from LcpDown
o Bring down IPCP & CCP when we enter PHASE_TERMINATE
o Give a new prompt when we go to PHASE_DEAD
o Stop the modem timer properly when idle
o Treat sig 15 like an exiting carrier loss
o Log (DEBUG) offline & online transitions
o Add missing $Id$s
o Move extern decls from .c -> .h files
o Staticize
o Remove #includes from .h files
o style(9)ify includes
o bcopy -> memcpy
bzero -> memset
bcmp -> memcmp
index -> strchr
rindex -> strrchr
o Move timeout.h -> timer.h (making it consistent w/ timer.c)
o Add -Wmissing-prototypes
Without this, in -auto mode, we stay in ST_STOPPING
and never check our dial filters to see if it's time
to bring the line up again.
This may make "set stopped" redundant.
Support VJ slot id compression.
Previously, ppp would negotiate a max slot between 2 & 15
(if asked), and would agree to slot id compression (if asked).
It would then proceed to use 16 slots and no compression
anyway. The result was a rather unusable connection.
o LcpLayerDown() no longer does a NewPhase(PHASE_TERMINATE).
Instead, it's done in LcpLayerFinish(). LayerFinish() gets
called by the FSM after the LCP FSM goes through the Stopping
and Stopped states.
o -direct and -background mode exit at PHASE_TERMINATE, not
PHASE_DEAD.
The result is that LCP, CCP & IPCP are brought down cleanly on both
sides of the link (not just our side). Killing ppp rather than just
closing it still makes it get out after the LCP SendTerminateReq().
I'll have a look at that soon. We're probably not actually sending
the REQ :-(
than /usr/include/des.h before building with MSChap.
support. Also allow -DNOCRYPT (as well as -DNOSECURE)
as an override
sbin/init example pointed out by: bde
negotiation. Instead, incrementally pause after
receiving LCPs with the same magic. We can now
suffer a server that waits more than 1 second before
responding. Pauses greater than a second get
hopelessly confusing as when the server eventually
starts, it sees a flood of Config Requests followed
by config NAKs and changes of magic. This causes the
server to change its magic over and over.....
file get created. We don't create lock files over non-tty
connections, but we *do* create lock files in -direct mode.
This leaves us capable of adding utmp/wtmp support for
successful pap & chap logins (coming soon).
config request. This stops us from squirting stuff
down a line that still has ECHO turned on because the
peer hasn't had a chance to start yet.
Lead to the cause by: Greg Lehey <grog@lemis.com>
close(1);
close(2);
x = open(ctermid(NULL), O_RDWR|O_NONBLOCK);
close(0)
on a tty causes select() to return an exception for descriptor x !
This is the case in RELENG_2_2, but not in 2.2.2. I'm not sure why.
Instead of doing the x=open() and close(0), we just do x=0 now.
Problem pointed out by: Greg Lehey <grog@lemis.com>
Tomi Vainio <tomppa@fidata.fi>
isatty(1) ! Keep 0 open for this till the modem's
been set up by either dup()ing 0 or by opening
ctermid(NULL) (if isatty(0)).
Discussed problem with: Tomi Vainio <tomppa@fidata.fi>
Made it finally dawn on me: Angelo Turetta <ATuretta@stylo.it>
o Output the correct device for "show modem"
while in -direct mode.
o Cosmetic: Moan a bit more when we can't open
the [modem] device.
o Call OpenModem() in a more "natural" way.
o Add some LogDEBUG in OpenModem().
receive the ResetAck, NOT when we send the
ResetReq (as per the rfc).
o SILENTLY ignore CCPs that arrive *before*
the network phase (as per the rfc).
o Check that we've actually negotiated PRED1
before sending PRED1 output.
This bug has been around for a *VERY* long
time ! We shouldn't need to explicitly disable
PRED1 now :-)
mode. We don't want to be forced to type a password
here :-(
Pointed out by: mouth@ibm.net (John Kelly)
While I'm there, don't allow a "set server" in
interactive mode.
Insist that uid == 0 for client ppp
Disallow client sockets if no password is specified
Don't exit on failure to open client socket for listening
Allow specification of null local password
Use reasonable size (smaller) ``vector''s in auth.c
Fix "passwd ..." usage message
Insist on "all" as arg to "quit" (if any)
Drop client socket connection before Cleanup() when "quit all"
