Follow-up of the #44565 bug fix, renamed the misnamed
PPP_INPROC_MULTITHREADED to PPP_INPROC_IRQ_SAFE because it is
IRQ safe but not thread safe.
Updated PPP documentation which now clearly state when and how
this feature can be used.
lwip/src/netif/ppp/pppol2tp.c: In function ‘pppol2tp_udp_send’:
lwip/src/netif/ppp/pppol2tp.c:1229:3: warning: ‘err’ may be used uninitialized in this function [-Wmaybe-uninitialized]
return err;
^
Renamed pppos_drop() → pppos_input_drop()
Renamed pppos_free_current_input_packet() → pppos_input_free_current_packet()
Moved pppos_output_last() after pppos_output_append()
Moved pppos_input_free_current_packet() before pppos_input_drop()
We actually allocated a pbuf chain only to iterate later the linked list
calling sio_write() for each pbuf, improved by calling sio_write() when
buffer is full and by recycling the pbuf, therefore only using one pbuf
for PPPoS output path.
Reworked pppos_write() and pppos_netif_output() to share more common
code into pppos_output_append() and pppos_output_last().
If PPP_INPROC_MULTITHREADED is true, then user does not what to use
the TCPIP API. Disabling the TCPIP API helps the user to understand
that PPP_INPROC_MULTITHREADED must not be used if he wish to use
the TCPIP API.
!NO_SYS users may now use as well the TCPIP API for PPPoS input data,
this way they can disable PPP_INPROC_MULTITHREADED and run pppos_input()
inside the lwIP thread, which fixes, at least for them, all the
threading issues related to PPP_INPROC_MULTITHREADED.
If PPP_INPROC_MULTITHREADED is not enabled, we can free unfinished
RX pbuf from the pppos_disconnect() function because pppos_input()
is running in the same context. Thanks to the pppos->open flags we
now only need to free remaining pbuf in the disconnect function
if PPP_INPROC_MULTITHREADED is not enabled.
Don't process input data if PPPoS is closed, it helps using
pppos_input() from a different context to prevent pppos_input() to
modify PPPoS RX machine state on a closed PPPoS session. It also
prevents allocating pbuf (which are going to be tossed out by PPP core)
and parsing serial input on a closed session.
It only mitigates the fact that this function is actually NOT thread
safe in absolutely all cases, it does not fix it but it helps for a low
cost.
For example user application should never call pppos_input() while
pppos_connect() or pppos_listen() is currently running because both of
them are freeing any input pbuf left over from the last session before
resetting the PPPoS state, they really have to to prevent pbuf leaks.
We cannot fix that easily because we don't have spinlock with an
irqsave/irqrestore helper for IRQ contexts. Mutex cannot be used in
interrupt contexts (or again, with an IRQ mutex helper).
We are going to improve the documentation on this point.
Reduced static buffer to the strict necessary (26 bytes), removed
call to eui64_ntoa(), merged eui64_ntoa() into llv6_ntoa().
272 bytes (code + static buffers) to 140 bytes on x86_64.
Improved eui64_ntoa() as well, we don't need it anymore but I'd
like to keep all eui64_* helpers functions.
Set IPv6 link local invalid before clearing it. Don't wait sif6up() to set
the link local valid, netif_ip6_addr_set_state() is only an accessor function
so it does not start any discovery whatsoever.
PPPoS was actually not thread safe, pppos_input() can be called from
lwIP user port at any time, whatever the PPP state is. It might even be
called during pppos_connect() and pppos_listen(), this is quite unlikely
the port do that but nothing prevent the user to since we document
pppos_input() as being thread safe.
Added a mutex if PPP_INPROC_MULTITHREADED is set and ensure pppos_input()
is safe in regard to other pppos_* functions.
Now setting DNS addresses in sdns() using dns_setserver() instead
of keeping a local copy. This is probably the wanted PPP behavior,
we will see if users want this conditional.
This was confusing, recent lwIP changes fixed the meaning as well as
how it is used everywhere, making the administrative status a user-only
controlled flag. Now that it's clear, updated PPP to follow lwIP
core change.
Using netif_set_link_{up,down} instead of netif_set_{up,down} when PPP
reaches/leaves running state. PPP interface is now set to administratively
UP when created with link state down.
lwIP core might send packet when calling netif_set_up(), don't drop packets
anymore in ppp_netif_output_ip4() and ppp_netif_output_ip6() because flags
are cleared.
Only used for debug, reworked in a more clever way. It was actually broken
by design: setting an interface integer into a string without formatting
was quite a dumb idea.
Don't disconnect when we receive a PADT, we let the LCP Echo/Reply find
the fact that PPP session is down. Asking the PPP stack to end the
session require strict checking about the PPP phase to prevent endless
disconnection loops.
Luckily it previously does nothing because PADT frames are rarely sent
with a hunique tag and we only set the sc pointer if we receive a
hunique tag.
Makes it clear we are initiating the PPP session with ppp_connect
(i.e. acting as a PPP client) so there is no confusion possible
between ppp_connect and ppp_listen.
Prepare for PPP_SERVER support, we need to move auth configuration before
lcp_allowoptions is copied into lcp_gotoptions. Restore unused
auth_reset() function using pppd original source code.
We don't need ask_for_local boolean, this is only useful for setup which
can determine the local IP address from the system hostname, which is
probably meaningless for embedded devices (and probably any devices).
It was actually only set by ip_check_options() which is commented out in
lwIP because we don't parse a config file nor check PPP configuration
(user is responsible about writing a configuration which is logical ;-).
Furthermore ask_for_local boolean never set actually had the wrong
default for PPP server setups.
PPP is now pointerful for a while, we don't need anymore accessor functions
for the unique PPP local and static control block. Replaced
ppp_set_netif_statuscallback() and ppp_set_netif_linkcallback() functions to
defines.
Removed pppapi_do_ppp_set_netif_statuscallback() and
pppapi_do_ppp_set_netif_linkcallback(), they were useless because
netif_set_status_callback() and netif_set_link_callback() can be
safely called while PPP status is in dead (= non open) state
and even before the PPP session is actually created at all.
this changes the callback signatures of the ip_output and the
{udp,raw}_recv functions.
changes were made by going through all header files, searching for
occurrences of ip_addr_t, qualifying them as const and if required
modifying the corresponding c files, looking for other uses of ip_addr_t
that would be required.
the following header files were not treated as i'm not using them and
wouldn't see them compiled: api.h api_msg.h dhcp.h dns.h igmp.h
netifapi.h pppapi.h snmp.h snmp_msg.h snmp_structs.h ppp.h pppol2tp.h
test/*
no modifications were done on ip6_addr_t.
pppos_create() can be called whether the modem is ready to process the
PPP session since pppos_create() does not start the PPP session anymore,
moved the advise from pppos_create() to ppp_open().
ppp_sighup() hard change the PPP FSM phase, it ends up with a FSM
mismatch if PPP is currently connecting or disconnecting.
Only do "sighup" on the stable running phase, fallback to the close
method. Handle special DEAD and HOLDOFF states as well.
We can have an IPv6 only PPP interface, checking if6_up instead of if_up fixes
IPv6 only setup.
ppp_netif_output() which were only used for common code between
ppp_netif_output_ip4() and ppp_netif_output_ip6() is not necessary
anymore, removed, reducing call stack by one.
This is now totally useless, it was used for the PPP core code but it is not
used anymore, remove then the user ability to set the PPP error code through
PPPCTLS_ERRCODE.
Removed useless calls to ppp_ioctl(pcb, PPPCTLS_ERRCODE, …), we now assign
pcb->err_code directly instead. ppp_ioctl() is not linked anymore if user
application don't use it.
Added macro PPP_AUTH_SUPPORT, if none of auth protocols are enabled
(PAP, CHAP, EAP) we reduce PPP memory usage by compiling out all
struct fields and source code used for authentication.
We only need to keep track of existing PPPoS interfaces if PPPoS
is not the only enabled protocol.
PPP CORE does not have callbacks pointers for all PPPoS callbacks
which should actually be required for PPPoS (VJ config, asyncmap, ...),
there is too much callbacks to create and PPPoS must be kept light,
especially for users who are only using PPPoS.
But there is a drawback, PPP CORE does not know which
lower protocols it is talking to thanks to the abstraction,
therefore if PPPoS is enabled as well as PPPoE or PPPoL2TP there
might be situation where PPP CORE calls pppos_ config functions
on interfaces which are NOT PPPoS one. This is very unlikely to
happen because protocols not supported by PPPoE or PPPoL2TP are
disabled at LCP/IPCP negotiation but we are better safe than sorry.
So we check if passed PPP pointer to PPPoS configuration functions
is a PPPoS interface by checking against a linked list of existing
PPPoS interfaces.
lwip/src/netif/ppp/ppp.c: In function ‘ppp_input’:
lwip/src/netif/ppp/ppp.c:769:5: warning: ISO C90 forbids mixed declarations and code [-Wpedantic]
const char *pname = protocol_name(protocol);
^
Reordered functions in the order they are declared in headers.
Removed useless ppp_link_down() function.
Merged ppp_stop() and ppp_close().
Merged ppp_hup() and ppp_sighup().
xmit_accm was meant to be a user configurable asyncmap, it was actually
broken since the introduction of ppp_new by the way we now reset the PPP
initial state, looks like no one until now is needing it anymore. If
necessary we will reintroduce this feature later properly instead of a
dirty and ugly hack into the PPP code.
We don't have callbacks which can be set or cleared for PPPoS
configuration, there is too much callbacks to create and PPPoS must be
kept light, therefore PPPoS functions can be called when PPP core
configure a PPPoE or PPPoL2TP interface, this is very unlikely to
happens because protocols not supported by PPPoE or PPPoL2TP are
disabled at LCP/IPCP negotiation but being safe is still better.
Check if passed PPP pointer to PPPoS configuration functions is a PPPoS
interface by using a linked list of exiting PPPoS interfaces.
Moved ppp_write_over_ethernet() and ppp_netif_output_over_ethernet() to pppoe.c
Moved ppp_write_over_l2tp() and ppp_netif_output_over_l2tp() to pppol2tp.c
Added necessary PPP core functions for PPPoE and PPPoL2TP status
notificaton (ppp_link_failed and ppp_link_end), removed callback,
low level protocol are now calling PPP core "link" functions.
First step of a rework of how low level protocols are using the
PPP core. Low level protocols are now going to use the core instead
of core using the low level protocols.
Final goal: separate PPP core code from low level protocols.
Re-order ppp.[ch] functions in the order functions should be called from
user application. Moved create functions, which actually return a PPP
control block before functions needing a PPP control block.
The only benefit of ppp_delete() call was about having a persistent
netif interface. netif was moved out of PPP pcb so we don't need
ppp_delete() anymore, second step in simplifying the weird
new/open/free/delete PPP API.
The only benefit of ppp_new() call was about having a persistent netif
interface. netif was moved out of PPP pcb so we don't need ppp_new()
anymore, first step in simplifying the weird new/open/free/delete PPP
API.
Users might want to share a netif control block between an Ethernet
interface and a PPPoS interface (I want actually) in case PPP is just
used as redundancy if Ethernet is down (eg. PPPoS GPRS fail over).
Moved netif out of PPP control block in a similar way it is currently
done for Ethernet interfaces. Furthermore, this is a first step on
removing the "new/create/free/delete" API which is awful but currently
necessary to handle fail over from PPPoX to another PPPoX (eg. from PPoE
on xDSL to PPPoS on GPRS fail over) without free()ing the netif which
might be used on udp_sendto() or L2TP VPN links.
The current code is buggy regarding handling of link state when using
both IPCP and IPv6CP: if IPv6CP has been set up and if during IPCP
negociation, ipcp_up() fails, it will incorrectly take the interface
down. The simple solution here is to change the platform code to do the
same as on Solaris: separate IPv6CP up/down state handling with sif6up()
and sif6down(), so that we really know when the interface is allowed to
go down.
(Based from pppd commit b04d2dc6df5c6b5650fea44250d58757ee3dac4a)
The style u_int32_t is not used anywhere else in the project, and is not
supported by the C standard, now using lwIP u32_t type. It was introduced
in 25e398a.
pppd: Accept IPCP ConfAck packets containing MS-WINS options
Since last week I'm seeing IPCP negotiations going like this (and
eventually failing) when connecting to my ISP:
Jul 11 20:03:25 * pppd[4833]: sent [IPCP ConfReq id=0x2 <addr 0.0.0.0> <ms-dns1
0.0.0.0> <ms-dns2 0.0.0.0>]
Jul 11 20:03:26 * pppd[4833]: sent [IPCP ConfReq id=0x2 <addr 0.0.0.0> <ms-dns1
0.0.0.0> <ms-dns2 0.0.0.0>]
Jul 11 20:03:26 * pppd[4833]: rcvd [IPCP ConfNak id=0x2 <addr 10.167.246.198>
<ms-dns1 213.162.69.1> <ms-dns2 213.162.69.169> <ms-wins 124.6.168.55> <ms-wins
17.17.17.17>]
Jul 11 20:03:26 * pppd[4833]: sent [IPCP ConfReq id=0x3 <addr 10.167.246.198>
<ms-dns1 213.162.69.1> <ms-dns2 213.162.69.169> <ms-wins 124.6.168.55> <ms-wins
17.17.17.17>]
Jul 11 20:03:26 * pppd[4833]: rcvd [IPCP ConfAck id=0x3 <addr 10.167.246.198>
<ms-dns1 213.162.69.1> <ms-dns2 213.162.69.169> <ms-wins 124.6.168.55> <ms-wins
17.17.17.17>]
Jul 11 20:03:27 * pppd[4833]: sent [IPCP ConfReq id=0x3 <addr 10.167.246.198>
<ms-dns1 213.162.69.1> <ms-dns2 213.162.69.169> <ms-wins 124.6.168.55> <ms-wins
17.17.17.17>]
...
with the last two lines repeating until the IPCP error limit is
reached. As you can see, the peer added two extra fields in the
ConfNak reply. This is allowed, and indeed the following sent
ConfReq packet reflects this. However, when the ConfAck packet
is received, pppd discards it as invalid, because of the ms-wins
fields.
This fixes it.
pppd: Take out unused %r conversion completely
This just removes some code surrounded by #if 0/#endif, which Fedora
apparently feels the need to patch...
PPP notify phase support, using compile-time PPP_NOTIFY_PHASE macro.
This can be used for example to set a LED pattern depending on the
current phase of the PPP session.
Callback example:
static void ppp_notify_phase_cb(ppp_pcb *pcb, u8_t phase, void *ctx) {
switch(phase) {
case PPP_PHASE_DEAD: /* Kept off */
case PPP_PHASE_MASTER:
/* LED Off */
break;
case PPP_PHASE_INITIALIZE: /* Session opened */
/* LED FastBlink */
break;
case PPP_PHASE_RUNNING: /* Session running */
/* LED On */
break;
default:
/* LED SlowBlink */
}
}
Removed one unecessary allocated PBUF per PPPoS RX packet if PPP_INPROC_MULTITHREADED is set by adding the necessary data for
pppos_input_callback() in front of the first pbuf instead of allocating a new buffer.
pbuf_type PPP is using for LCP, PAP, CHAP, EAP, IPCP and IP6CP packets.
Memory allocated must be single buffered for PPP to works, it requires pbuf
that are not going to be chained when allocated. This requires setting
PBUF_POOL_BUFSIZE to at least 512 bytes, which is quite huge for small systems.
Setting PPP_USE_PBUF_RAM to 1 makes PPP use memory from heap where continuous
buffers are required, allowing you to use a smaller PBUF_POOL_BUFSIZE.
I consider to remove the PPP_INPROC_OWNTHREAD crap in ppp-new,
as said in bugs #37278 and #37353.
1. It requires the ppp_input_thread() function to be modified to match
user system, like some did by adding the vTaskDelete(NULL); FreeRTOS
call at the end of the function, for example.
This is a tiny-tiny fonction that should be, in my opinion, on the user
port, like the Ethernet input thread we see in many Ethernet port.
2. It is actually not that thread safe.
2.1. pcb->phase IS modified by the lwIP core thread so it should at
least be set to volatile, otherwise the pcb->phase copy may live
indefinitely in CPU register. It works because of the sio_read()
function call which without doubt flush pcb->phase copy from CPU
register. I dont want to set ppp_pcb struct to volatile for obvious
performance reasons.
2.2. This function assume PCB still exists whatever is happening, which
is not the case after you called ppp_delete() function outside of this
thread. If sio_read() is blocking waiting data and pcb destroyed, it is
going to read a deallocated pcb which luckily should still have
pcb->phase set to 0 (=PHASE_DEAD) due to preallocated "control block"
structures of lwIP. Even with sio_read_abort(), there might be timings
issue due to a lack of a synchronization mechanism.
3. I dislike the sys_msleep(1), it means that systems should have at
least a 11 chr buffer at 115200/10 byte/s, and bigger with higher serial
speed, for example with 3G/HSDPA modems accessed through SPI, at 20
Mbits/s this is a ~2000 bytes buffer required to keep incoming data
during this sleep, I don't see why we require systems to do so,
sio_read() should obviously be a blocking call. I cannot easily
remove this sleep because some systems might have wrongfully used this
call as a CPU idle feature with a non blocking sio_read() call.
Free the control block, clean everything except the PPP PCB itself
and the netif, it allows you to change the underlying PPP protocol
(eg. from PPPoE to PPPoS to switch from DSL to GPRS) without losing
your PPP and netif handlers.
Created new ppp_over_X_create() functions which only prepare the PPP session without starting it
Removed ppp_reopen() and all of its sub ppp_over_X_reopen()
Removed PPPoL2TP reconnect() function, merged to connect()
Added ppp_open() able to start or restart any session
LCP is stealing a bit from fsm->flags struct member for LCP delayed up feature.
Bit stealed used to be the 9th bit (0x100) but fsm->flags was reduced to u8_t to save memory,
we are now stealing the 8th bit (0x80).
these are the compiler warnings I get with the head of ppp-new. All
of them are trivial, [...] (I'm using IAR EWARM 6.4).
ppp.c
Warning[Pe550]: variable "c" was set but never used
lwip\src\netif\ppp\ppp.c 1012
Warning[Pe111]: statement is unreachable
lwip\src\netif\ppp\ppp.c 1132
Warning[Pe111]: statement is unreachable
lwip\src\netif\ppp\ppp.c 1377
Warning[Pe111]: statement is unreachable
lwip\src\netif\ppp\ppp.c 1412
utils.c
Warning[Pe186]: pointless comparison of unsigned integer with zero
lwip\src\netif\ppp\utils.c 210
The word "class" is reserved in IAR's EWARM compiler since it looks like c++.
This causes a failure to compile in the lcp code.
Arguably it is a bug in the compiler, but it is easy to work around with a
name change in the lcp.[ch] code. I fixed it by changing "class" to "class_".
PPP stack does not handle chained pbuf, but PPP negociation packets
are at most ~40 bytes long, so we are only checking if the payload
can fit into the allocated pbuf (p->tot_len == p->len).
The lwIP PPP support is based from pppd 2.4.5 (http://ppp.samba.org) with
huge changes to match code size and memory requirements for embedded devices.
Anyway, pppd has a mature codebase for years and the average commit count
is getting low on their Git repositories, meaning that we can follow what
is happening on their side and merge what is relevant for lwIP.
So, here is the pppd follow up, so that we don't get away too far from pppd.
