Commit Graph

4734 Commits

Author SHA1 Message Date
Dirk Ziegelmeier
a420d2530f arch.h: Add hint for struct packing #defines on GCC/clang 2016-12-08 11:13:51 +01:00
Dirk Ziegelmeier
89cb7b7aa1 Try to remove #include <stdlib.h> from many files. Does not seem necessary any more and might cause problems when porting lwIP. 2016-12-08 11:05:01 +01:00
goldsimon
9c3bbcf4e6 removed old disabled code 2016-12-08 10:24:56 +01:00
Dirk Ziegelmeier
f446194c8a pppos.c: Use arch.h to get size_t instead of including stddef.h 2016-12-07 22:29:24 +01:00
Dirk Ziegelmeier
1dd563a0ae arch.h: Implement possibility to provide own header for size_t, default is stddef.h 2016-12-07 22:27:57 +01:00
Dirk Ziegelmeier
aae0fc4908 Use new IPv6 init macro in netif.c 2016-12-07 22:19:37 +01:00
Dirk Ziegelmeier
5bb83301dd Add macro to init an IPv6 address with U32 in host byte order 2016-12-07 22:19:16 +01:00
Dirk Ziegelmeier
0ca82df062 Cleanup byte order handling a bit.
- Create LWIP_MAKEU32(a,b,c,d) to create an U32 value from bytes
- Use PP_HTONL() in some macros to emphasize network byte order conversion
2016-12-07 20:18:58 +01:00
goldsimon
ff3656f4f5 Added improved macros for ip address initialization: IPADDR4_INIT_BYTES(), IPADDR6_INIT_HOST() and IP4_ADDR_MAKEU32() 2016-12-07 13:06:07 +01:00
goldsimon
b31b0c8148 remove bogus LWIP_MAKE_U16() define and use PP_NTOHS() in ip4.c instead 2016-12-07 12:44:57 +01:00
goldsimon
c87855423c DNS: added compile-time check for some defines to fit into an u8_t (bug #49658) 2016-12-07 09:09:45 +01:00
Dirk Ziegelmeier
e00a131160 Fix bug #49778: sntp_stop does not cancel all timers
Patch by Ari Suutari
2016-12-06 20:29:12 +01:00
Dirk Ziegelmeier
cb29a49a64 Update CHANGELOG 2016-12-06 11:39:03 +01:00
Axel Lin
795acf020e lwiperf: Simplify #if LWIPERF_CHECK_RX_DATA guard
The variable i is equal to q->len after exit the for loop.
Check the received data should not change the logic of update packet_idx.
So let's simplify the code a bit.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
2016-12-06 11:36:25 +01:00
Dirk Ziegelmeier
cc25c2634b Cleanup: move struct ip6_addr_packed and ip6_addr_p_t type to prot/ip6.h - these types are used in prot/ip6.h and prot/mld6.h 2016-12-06 09:36:36 +01:00
Dirk Ziegelmeier
1687721600 Fix compile when IPv4 is disabled 2016-12-05 22:01:58 +01:00
Dirk Ziegelmeier
a6bc422729 Move declaration of struct ip4_addr_packed and ip4_addr_p_t to prot/ip4.h
The types are used in structs declared in ip4.h.
2016-12-05 21:53:43 +01:00
Axel Lin
92183bb354 icmp: Increment mib2.icmpintimeexcds counter if got ICMP_TE
Increment mib2.icmpintimeexcds rather than mib2.icmpindestunreachs if got ICMP_TE.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
2016-12-05 12:30:09 +01:00
Sylvain Rochet
bcaf2f08aa PPP, PPPoS: fix memory leak when disconnecting if there are remaining input bytes
Art says:
  pppos_input() can call ppp_input() which can call pppos_disconnect() to
  disconnect the interface.  However, it will continue to read in
  characters and allocate a pbuf from the PBUF_POOL and keep it in
  pppos->in_head and in_tail. When a re-connect happens and pppos_connect()
  is called, this pppos->in_head and in_tail are zeroed, hence a memory
  leak. (This happens with PPP_INPROC_IRQ_SAFE not defined.)

  A fix would be inside pppos_input() to break out of the loop inputting
  characters after calling ppp_input() if pppos->open == 0.  Note that
  the loop is not even entered if pppos->open == 0.

          ppp_input(ppp, inp);
          if(pppos->open == 0) //get out if they disconnected
            break;

Fix it in a similar way which doesn't add new code by moving the
existing pppos->open check inside the byte loop.
2016-12-03 16:12:51 +01:00
Dirk Ziegelmeier
4bbed75cc4 Minor code layout cleanup in err.c and errno.h 2016-12-01 08:55:01 +01:00
Joel Cunningham
7d0aeaf539 Doc: correct minor mis-spelling in sys_arch.txt
This corrects a minor mis-spelling where "by" was mis-spelled as "ny"
2016-11-30 10:28:38 -06:00
Joel Cunningham
f28e63b2a3 Add netifapi macros for set link up/down
This commit extends the netifapi macros to support netif_set_link_up
and netif_set_link_down
2016-11-30 08:49:23 -06:00
Joel Cunningham
7f48289fcd Increment ip.drop when dropping due to NULL netif
This commit increments the ip.drop statistic when an IP packet is
dropped due to no matching netif found and forwarding is disabled

This adds parity to the other places where mib2.ipinaddrerrors and
mib2.ipindiscards are incremented which also increment ip.drop
2016-11-30 08:49:22 -06:00
Axel Lin
12e35c4c12 mdns: Fix assertion message in mdns_resp_add_service_txtitem()
So we know which function emits the assertion.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
2016-11-30 11:36:05 +01:00
Dirk Ziegelmeier
182d7c138a Add #include <stddef.h> to a central place (arch.h) instead of #including it in several other files throughout lwip since size_t is needed in many places
See http://lwip.100.n7.nabble.com/Issue-in-arch-h-for-lwIP-2-0-0-td27948.html
2016-11-30 07:43:59 +01:00
Dirk Ziegelmeier
47fd67a35c Remove TODO comments and one check from sockets.c indicating IPV6_V6ONLY socket option handling does not work 2016-11-28 15:56:59 +01:00
goldsimon
b934c3f471 fixed bug #49726: setsockopt() set TCP_NODELAY TCP_KEEPALIVE ... with a listen state TCP will crash 2016-11-28 15:50:59 +01:00
goldsimon
2a882b6387 minor: fixed indent 2016-11-28 12:54:17 +01:00
goldsimon
ca9342c549 fixed bug #49725 (send-timeout: netwonn_write() can return ERR_OK without all bytes being written) 2016-11-28 12:51:45 +01:00
goldsimon
6f1304e03e patch by Ambroz Bizjak: fixed bug #49717 (window size in received SYN and SYN-ACK assumed scaled) 2016-11-28 10:27:21 +01:00
Axel Lin
cac3dc8a46 netif: Trivial indent fix
Signed-off-by: Axel Lin <axel.lin@ingics.com>
2016-11-28 10:24:23 +01:00
Dirk Ziegelmeier
aea872431c Fix naming of some inet_addr_* macros in inet.h
From inet4_addr_* to inet_addr_* - inet_addr is a "known word", don't change it
2016-11-25 22:13:12 +01:00
Axel Lin
1d4cbe768d netif: Add proper lock protect for accessing netif->loop_first
All the reset part of the code accessing netif->loop_first has lock protection,
the only missing part is "while (netif->loop_first != NULL)".
Fix it by adding lock protect around the while loop.

Also convert the code to use while{} loop instead of do .. while{} loop,
then we can avoid NULL test for in pointer in each loop and reduce a level of indent.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
2016-11-25 13:02:23 +01:00
goldsimon
0e07ed4b13 fixed bug #49676 (Possible endless loop when parsing dhcp options) & added unit test for that 2016-11-25 10:03:43 +01:00
Richard Sailer
2ed755764e doxygen/generate.sh: Add shebang line
This file had the x bit set.
But executing it produced an error, since it was
missing the "#!/bin/sh" line.

This patch adds the "#!/bin/sh" line and makes generate.sh directly executable.
2016-11-25 09:12:40 +01:00
goldsimon
f419231dc3 fixed typo in CHANGELOG 2016-11-24 11:31:46 +01:00
Ambroz Bizjak
8ba7363d11 Optimize passing contiguous nocopy buffers to tcp_write
While TCP_OVERSIZE works only when tcp_write() is used with
TCP_WRITE_FLAG_COPY, this new code achieves
similar benefits for the use case that the caller manages their own
send buffers and passes successive chunks of those to tcp_write()
without TCP_WRITE_FLAG_COPY.

In particular, if a buffer is passed to
tcp_write() that is adjacent in memory to the previously passed
buffer, it will be combined into the previous ROM pbuf reference
whenever possible, thus extending that ROM pbuf rather than allocating
a new ROM pbuf.

For the aforementioned use case, the advantages of this code are
twofold:
1) fewer ROM pbufs need to be allocated to send the same data, and,
2) the MAC layer gets outgoing TCP packets with shorter pbuf chains.

Original patch by Ambroz Bizjak <ambrop7@gmail.com>
Edited by David van Moolenbroek <david@minix3.org>
Signed-off-by: goldsimon <goldsimon@gmx.de>
2016-11-24 11:27:34 +01:00
goldsimon
4c8620e03b Added important post-2.0.0 changes to CHANGELOG to keep track of changes for future 2.0.1 release 2016-11-24 11:21:00 +01:00
goldsimon
eb1de78ce1 prepare CHANGELOG for post-2.0.0 2016-11-24 11:12:22 +01:00
Joel Cunningham
5030fa81a0 bug #49684, api_msg: treat non-blocking ERR_MEM as ERR_WOULDBLOCK
This corrects a case in lwip_netconn_do_writemore() where if a
non-blocking socket receives ERR_MEM in a call to tcp_write(), it would
return ERR_MEM, which would result in ENOMEM coming out of the socket
layer

This case can be gracefully handled by returning ERR_WOULDBLOCK since the
socket is already marked as no longer writable and sent_tcp/poll_tcp will
mark the socket as writable again based on available buffer space

This is very similiar to how ERR_MEM is resolved for blocking sockets
2016-11-23 15:03:43 -06:00
Dirk Ziegelmeier
0f87cb92b8 Add note about UDP multicast behavior fix to UPGRADING document 2016-11-23 13:10:16 +01:00
Dirk Ziegelmeier
09547832ba Fix bug #49662: UDP layer should filter incoming multicast datagrams against the bound IP address
Change lwIP UDP API to match socket behavior. Multicast traffic is now only received on a UDP PCB (and therefore on a UDP socket/netconn) when the PCB is bound to IP_ADDR_ANY.
2016-11-23 12:46:35 +01:00
Sylvain Rochet
8c3c96baf7 PPP, L2TP: fix PPPOL2TP_AUTH_SUPPORT == 0 support
Fix compiler warnings on unused parameters and a function signature
mismatch in PPPAPI.
2016-11-22 22:13:24 +01:00
sg
03a9aac157 dns_enqueue(): minor readability improvement: add local variable "age" to store result of subtraction 2016-11-22 21:34:12 +01:00
David van Moolenbroek
68ec20fffc ipv4/ipv6: restrict loopback-destined traffic
Generally speaking, packets with a loopback destination address -
127.0.0.1 for IPv4 and ::1 for IPv6 - should not be accepted on
non-loopback interfaces.  For IPv4, this is implied by RFC 1122
Sec. 3.2.1.3.  For IPv6, it is mandated by RFC 4291 Sec. 2.5.3.
Failure to perform this filtering may have security implications, as
applications that bind sockets to loopback addresses may not expect
that nodes on the local external network be able to produce traffic
that will arrive at such sockets.

With this patch, lwIP drops packets that are sent to a loopback
address but do not originate from the interface that has the loopback
address assigned to it.  This approach works regardless of whether it
is lwIP or the system using it that implements a loopback netif.  The
only exception that must be made is for configurations that enable
netif packet loopback but disable the lwIP loopback netif: in that
case, loopback packets are routed across non-loopback netifs and would
thus be lost by the new filter as well.

For IPv6, loopback-destined packets are also no longer forwarded; the
IPv4 forwarding code already had a check for that.

As a small performance improvement, the IPv6 link-local/loopback
address check is now performed only once per packet rather than
repeatedly for every candidate netif.
2016-11-22 20:51:36 +01:00
Dirk Ziegelmeier
4076b12ee9 Revert "Apply patch #9165: Allowing udp src port to be 0 in cases when we don't care about outgoing port"
This reverts commit 31b0237c50.
2016-11-22 14:35:45 +01:00
David van Moolenbroek
0034abfa45 Always check whether netif_default is NULL
In general, netif_default may be NULL, and various places in the code
already check for this case before attempting to dereference the
netif_default pointer.  Some places do not perform this check though,
and may cause null pointer dereferences if netif_default is not set.
This patch adds NULL checks to those places as well.
2016-11-21 10:16:28 +01:00
Dirk Ziegelmeier
31b0237c50 Apply patch #9165: Allowing udp src port to be 0 in cases when we don't care about outgoing port 2016-11-18 08:13:15 +01:00
Dirk Ziegelmeier
9366c0eaab I decided to keep the "complexity" of handling IPv6 mapped IPv4 addresses out of netconn API.
Only socket API understands this address type now.
2016-11-17 12:41:00 +01:00
Dirk Ziegelmeier
5d5eeca008 Once more: Try fix compile with clang 2016-11-17 09:23:57 +01:00