Removed mutiple copies of keys by pre-setting MPPE keys during MSCHAP
negotiation.
Improved MPPE init so we don't need to pass a buffer formatted in a
special way to MPPE, this is necessary for pppd to talk to the kernel,
we don't need that here.
ccp_test() is not only used to test kernel support, but also to set MPPE keys,
we will change that further, but for now, re-add the necessary ccp_test()
We don't need to test our available compression protocols that much, we know
what we are going to be able to use in embedded. Even PPPd don't need that.
Checking only in ccp_resetci() is enough, this way we know our cilen size
before ccp_addci() is called. PPPd is designed that *_addci() can use less
bytes than it previously asked in *_cilen(), we prefer to avoid that here.
We actually don't need at all ccp_test() calls, we know what we are
able to support at compile time, we don't need to ask the kernel like
PPPd does. This is just the first step in ccp_test() removal.
Obviously, it requires some wiring to know if there is new activity
on a not-yet established PPP interface with the default route
already set.
I don't think any lwIP user will ever need that, all should know
when to bring the link up and down.
Removed all the used Linux ioctl(), sys_linux.c is entirely disabled.
Replaced unecessary functions to set up a PPP link to null fonctions,
however all stuff necessary to shutdown a PPP link should be done.
Some utils functions related to I/O files are disabled too.
ECP and CCP support only set PPP attributes into the PPP kernel support,
they are now compile-time options and are disabled by default and
obviously not supported (yet? :p)
OpenBSD DES require more than 70 kB of static memory.
The GNU libcrypt DES, for information, require more than 131 kB,
so this is probably the case with all non memory optimised DES.
PolarSSL only required 132 bytes of stack with some kB of .rodata
precomputed tables :-)
I personally don't need MS CHAP v1 or MS CHAP v2, and that was not
supported in the previous PPP port, so there is no regression,
I feel comfortable about removing those hard to port stuff.
If someone want to do the MS CHAP port, he first have to find or
do a small memory footprint DES implementation.
PPPoE works, PPPoS code is not ported at all.
I am using the RP-PPPoE server to do my tests using the following
configuration:
$ cat /etc/ppp/pppoe-server-options
debug
login
lcp-echo-interval 10
lcp-echo-failure 10
ms-dns 192.168.4.130
ms-dns 192.168.4.231
netmask 255.255.255.0
defaultroute
noipdefault
usepeerdns
$ cat /etc/ppp/allip
192.168.4.1-200
$ pppoe-server -C isp -L 192.168.4.254 -p /etc/ppp/allip -I tap0
Plus the usual auth-lines in /etc/ppp/pap-secrets and
/etc/ppp/chap-secrets .
And the unix port minimal "echo" project slightly modified to use
the "tcpip" API, so with threads, which I am going to commit with
NO_SYS as a -Dmacro.
It still use some of the linux'ism, such as syslog() and crypt(),
I do not want to drop the syslog() supports at the moment, this is
pretty useful to debug, and we may just convert the way the syslog() is
done to provide a trace feature to our PPP users, as a compile-time
option.
