0.81.1 of the i4b code - namely support of the I4B_VR_REQ
ioctl via the i4brbchX device.
Ppp controls the phone number, but idle timers and
SYNC/RAW decisions are still made by isdnd (in isdnd.rc).
This involves a new datalink state machine phase. The
``wait for carrier'' phase happens after dialing but
before logging in. The whole dial state should really
be abstracted so that each device type can deal with it
in its own way (thinking about PPPoE) - but that'll have
to wait.
The ``set cd'' symantics remain the same for tty devices,
but we now delay until we either get CD or timeout waiting
(at which time we drop the link if we require CD).
For i4b devices we always insist on carrier.
Thanks to hm@ for his help, and especially for pointing out
that I *don't* need to re-implement isdnd (that was a huge
waste of time !) :-]
o Show more information about missing MP fragments in ``show mp''.
o Do away with mbuf_Log(). It was showing mbuf stats twice on
receipt of LCP/CCP/IPCP packets.... ???!!?
o Pre-allocate a bit extra when creating LQR packets to avoid having
to allocate another mbuf in mbuf_Prepend().
the layering.
We now ``stack'' layers as soon as we open the device (when we figure
out what we're dealing with). A static set of `dispatch' routines are
also declared for dealing with incoming packets after they've been
`pulled' up through the stacked layers.
Physical devices are now assigned handlers based on the device type
when they're opened. For the moment there are three device types;
ttys, execs and tcps.
o Increment version number to 2.2
o Make an entry in [uw]tmp for non-tty -direct invocations (after
pap/chap authentication).
o Make throughput counters quad_t's
o Account for the absolute number of mbuf malloc()s and free()s in
``show mem''.
o ``show modem'' becomes ``show physical''.
when we've simply missed a packet.
When our Predictor1 CRC is wrong (implying we've dropped
a packet), don't send a ResetReq(). Instead, send another
CCP ConfigReq(). *shrug* My tests show this as being far
worse than the ResetReq as we may have further Nak/Rejs etc
and we're basically resetting both our incoming and outgoing
compression dictionaries, but rfc1978 says the ConfigReq is
correct, so we'd better go along...
(see the new ``set callback'' and ``set cbcp'' commands)
o Add a ``cbcp'' log level and mbuf type.
o Don't dump core when \T is given in ``set login'' or
``set hangup''.
o Allow ``*'' and blanks as placeholders in ppp.secret and
allow a fifth field for specifying auth/cbcp dialback
parameters.
o Remove a few extraneous #includes
o Define the default number of REQs (restart counter) in defs.h
rather than hardcoding ``5'' all over the place.
o Fix a few man page inconsistencies.
o Bring the static ``ttystate'' into struct prompt so that
the tilde context is per prompt and not global.
o Comment the remaining static variables so that it's
clear why they're static.
o Add some XXX comments suggesting that our interface list
and our hostname should be re-generated after a signal
(say SIGUSR1) so that a machine with PCCARDs has a chance.
into the ST_STOPPED state.
o Allow an optional ccp|lcp argument to `down'. The default is
still lcp (as before). You can now call down with no context
in multilink mode, in which case it'll down the multilink ccp
or the entire bundle (*very* rude).
o Allow an optional `!' after `close ccp' (close ccp!) to tell
ccp to stay in the CLOSED state after the terminate ACK. The
default is now to re-enter STOPPED so that the peer can bring
the layer back up if desired.
o Always handle proto-compressed packets, even if we've agreed
(in LCP) that the peer will not send us 1 byte protocols.
If the peer violates the LCP agreement, log it to the HDLC
log.
o Fix some comments.
bundle (non-negotiated vars) or to their respective IPCP,
LCP or CCP.
o Enable rolling throughput statistics by default.
o Remove the `display' command. These values now appear in
`show bundle', `show ipcp', `show ccp' and `show lcp'.
o Initialise auth name & key at bundle create time (oops).
o Rename pppd-deflate (the id-24 hack) to deflate24.
o Don't send both a REJ and a NAK to an IPCP or LCP REQ.
Favour the REJ (already done at the CCP level).
o Recurse in datalink_UpdateSet() when we change state, otherwise
we end up setting no descriptors and getting jammed in the
imminent select() instead of doing the dial/login/hangup.
o Display our CHAP encryption method despite being built with DES.
o Display VJ as not negotiated in ``show ipcp'' when necessary.
o Move Var*Version into command.c
o Remove struct pppVars (and there was much rejoicing) !
o Forward-decl some structs in .h files to avoid include
ordering requirements and remove a few more redundant
#includes.
Increment OutPackets for any packet - not just LQRs
MFC:
o Fix a few comment typos.
o Fix ``set timeout'' usage message and documentation.
o Change ifOutPackets, ifOutOctets and ifOutLQRs to `u_int32_t's
so that they wrap correctly.
o Put the LQR in network byte order using the correct struct size
(sizeof u_int32_t, not sizeof u_long).
o Wrap LQR ECHO counters correctly.
o Don't increment OutLQR count if the last LQR hasn't been replied
to.
o Initialise last received LQR in StartLqm.
o Don't start the LQR timer if we're `disabled' and `accepted'.
o Generate LQR responses when both sides are using a timer and
we're not going to send our next LQR before the peers max timeout.
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.
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.
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).
(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!
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.
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 *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).
