Commit Graph

4970 Commits

Author SHA1 Message Date
Dirk Ziegelmeier
27c835aa56 Remove netif_invoke_ext_callback() from doxygen docs, it is intended for internal use only 2017-02-27 09:36:33 +01:00
Dirk Ziegelmeier
e4b356f08c Add netif extended callback to doxygen docs 2017-02-26 09:56:16 +01:00
Dirk Ziegelmeier
01f9a04e4a Add pbuf_get_contiguous() to doxygen docs 2017-02-26 09:44:16 +01:00
Dirk Ziegelmeier
9ab2eefc37 Fix documentation in sockets.c 2017-02-26 09:42:06 +01:00
Dirk Ziegelmeier
08931b33c1 Fix documentation on pbuf_get_contiguous() 2017-02-25 17:03:14 +01:00
Joel Cunningham
36b9a45c3b netconn_write_partly cleanups
This commit changes netconn_write_partly to use msg.w.offset to set
bytes_written for both blocking and non-blocking connections

This is correct because msg.w.offset is the canonical output from
the do_write call and in the case that not all bytes were written,
(a bug?) returning the full size to the caller is dangerous

Lastly, this commit adds an assert for the blocking case to sanity
check that all the bytes we requested were written.  This will help
catch bugs in do_write
2017-02-24 15:50:28 -06:00
Joel Cunningham
36fa1a97d4 lwip_netconn_do_writemore() cleanups
This commit makes a couple of cleanups discussed in patch #8882:
  1) msg.w.offset should not be set to 0 in the error case.  It is
     only valid when err == ERR_OK
  2) Remove out-of-date comment which indicated the entire write had
     completed (not true for non-blocking write)

This also updates the documentation on offset to include that offset
is only valid when err == ERR_OK
2017-02-24 15:42:07 -06:00
goldsimon
0da9cf70ea Added pbuf_get_contiguous() to get data in one piece (either zero copy from pbuf or memcpied into a supplied buffer) 2017-02-24 21:36:49 +01:00
goldsimon
a38e937dd6 fixed close race conditions in lwip_select (for LWIP_NETCONN_FULLDUPLEX) 2017-02-24 21:29:29 +01:00
goldsimon
f4d13d52d3 fixed get_socket, fixed LWIP_SOCKET_OFFSET != 0 2017-02-24 21:23:53 +01:00
goldsimon
8bb43e7388 Added a few missing calls to done_socket() 2017-02-24 21:11:20 +01:00
goldsimon
6786c9f143 Start working on bug #44032: added sock->fd_used that is != 0 when at least one thread is using a socket 2017-02-24 21:00:01 +01:00
goldsimon
9afe10e23d make get_socket use tryget_socket, added tryget_socket_unconn in preparation of properly fixing bug #44032 2017-02-24 20:49:18 +01:00
goldsimon
f1f6050ad7 minor: comment, whitespace 2017-02-24 19:50:53 +01:00
goldsimon
d73e225519 lwip_select: check for correct 'maxfdp1' parameter range 2017-02-24 19:50:14 +01:00
goldsimon
3107d4a0fa fixed that select ignored invalid/not open sockets in the fd_sets (bug #50392) 2017-02-24 19:48:59 +01:00
goldsimon
01cc06bcb8 netif_add: ensure netif->num is unique even if more than 255 netifs have been added/removed since system startup (e.g. by ppp) 2017-02-23 20:41:54 +01:00
goldsimon
76763c9bcd pbuf_ref: assert-check for 'ref' overflow 2017-02-23 20:16:51 +01:00
Dirk Ziegelmeier
e94e2da3ec opt.h: Set LWIP_PBUF_REF_T back to u8_t, s8_t has no advantage any more after reverting last commit 2017-02-23 18:10:00 +01:00
Dirk Ziegelmeier
2cf3bbddd4 pbuf_free: Use correct type LWIP_PBUF_REF_T for local variable 2017-02-23 18:08:26 +01:00
Dirk Ziegelmeier
0f6d8ccd90 Revert "Improve pbuf refcount underflow check by checking the local variable on the stack that was assigned in a protected region"
This reverts commit 62c44138da.

Didn't notice due to local changes the ASSERTION was already inside the locks...
2017-02-23 18:04:30 +01:00
Dirk Ziegelmeier
62c44138da Improve pbuf refcount underflow check by checking the local variable on the stack that was assigned in a protected region
The old code was vulnerable to race conditions since it checked ref to be >0 without locks
2017-02-23 16:55:14 +01:00
Dirk Ziegelmeier
0a7734cf64 Preparation for task #14369: Define ways to work with a netif having multiple ports
Let ethernet_input() fill in pbuf's if_idx (if not already set by driver or an L2 bridge)
2017-02-23 16:36:40 +01:00
Dirk Ziegelmeier
0eeb10d4bb Preparation for task #14369: Define ways to work with a netif having multiple ports
Add if_idx member to struct pbuf
2017-02-23 16:34:28 +01:00
Dirk Ziegelmeier
2e1b12d09f Make pbuf refcount type an LWIP option and #define it to s8_t (to be able to check for underflow!) 2017-02-23 15:47:44 +01:00
goldsimon
5c55978f10 Revert "Added LWIP_SOCKERR_T to adapt the size of sock->err to what is needed."
Let's stay with the standard and accept the 4-bytes-per-socket memory increase...

This reverts commit 5d6c9ce77b.
2017-02-20 13:29:33 +01:00
goldsimon
5d6c9ce77b Added LWIP_SOCKERR_T to adapt the size of sock->err to what is needed. 2017-02-20 12:55:18 +01:00
goldsimon
d1049511cc post-fixups for patch #9262: fix SO_ERROR, fix !LWIP_TCPIP_CORE_LOCKING 2017-02-20 12:43:45 +01:00
Joan Lledó
703a22fae5 Support for 32bit errno codes 2017-02-20 12:37:29 +01:00
sg
ff3fe1f489 Fixed bug #50231 (nd6: several option size related issues in Router Advertisement message processing) 2017-02-17 21:59:08 +01:00
Joel Cunningham
4c76fd500c Move write_offset from struct netconn to struct api_msg
This moves the write_offset variable from struct netconn to struct api_msg

This optimizes the storage by only having the space claimed when it is
needed (during a netconn_write_partly() call) and not throughout the
lifetime of the netconn

This also reduces code space/execution by not having to separately manage
clearing/checking write_offset from the current_msg pointer

Lastly, we also save execution by using msg.w.offset as the output
rather than marshaling the result to msg.w.len. Previously, len was used
as input length of dataptr and output for the write operation.
netconn_write_partly() also has access to msg.w.offset, so we can use
that
2017-02-17 13:26:16 -06:00
goldsimon
deaa6e9406 Improved DNS_LOCAL_HOSTLIST interface (bug #50325) 2017-02-17 10:08:49 +01:00
goldsimon
19d63e6aa0 Fix portable initialization of non-dynamic DNS_LOCAL_HOSTLIST 2017-02-17 09:25:31 +01:00
sg
6dca664217 LWIP_NETCONN_FULLDUPLEX: fixed shutdown during write (bug #50274) 2017-02-16 22:09:02 +01:00
Dirk Ziegelmeier
d7b51b5906 MDNS: Fix accessing wrong union member in netif ext callback 2017-02-15 14:38:57 +01:00
Dirk Ziegelmeier
3c9e05a6d9 Minor: Fix documentation in netif.h 2017-02-15 14:38:34 +01:00
Dirk Ziegelmeier
498913b982 Third draft of task #14283: Add more detailed status callback
Implement Joel's idea of a union
2017-02-15 10:14:06 +01:00
Dirk Ziegelmeier
35e148e26d MDNS: Use new netif callback, now does not need mdns_resp_netif_settings_changed() calls from user any more when LWIP_NETIF_EXT_STATUS_CALLBACK is enabled 2017-02-14 21:03:49 +01:00
Dirk Ziegelmeier
454927151d Continue working on ext netif callback
Docs, better callback points (e.g. before netif is set down)
2017-02-14 21:02:38 +01:00
Dirk Ziegelmeier
d46d8bcda2 Add some asserts to netif callback functions 2017-02-14 20:35:48 +01:00
Dirk Ziegelmeier
4434762a08 Some minor documentation fixes 2017-02-14 20:28:43 +01:00
Dirk Ziegelmeier
bbedb35bf3 Fix name clash between typedef and struct definition.
Interesting, all three compilers in Erik Ekman's travis-ci don't complain. But my gcc at home does...
2017-02-14 20:17:27 +01:00
Dirk Ziegelmeier
f4f204b850 Implement a better way to initialize all fields in ip address than using memset() (better fix than my last commit) 2017-02-14 14:20:39 +01:00
Dirk Ziegelmeier
feba38a1f4 Fix unportable struct initializer I introcduced while working on task #14283: Add more detailed status callback 2017-02-14 13:54:14 +01:00
Dirk Ziegelmeier
1b2c8974b8 Forgot to adapt #define in my last commit 2017-02-14 11:00:34 +01:00
Dirk Ziegelmeier
8025b85694 Second draft of task #14283: Add more detailed status callback 2017-02-14 10:59:28 +01:00
Joel Cunningham
b2beb42c14 Remove forward declaration of struct lwip_sock
The forward declaration was added in commit
e2c2afbbe0 for use in struct
lwip_setgetsockopt_data, but became dead after commit
5d2e93e5f0 where the sock pointer was
removed from lwip_setgetsockopt_data
2017-02-13 17:22:52 -06:00
Dirk Ziegelmeier
7dd21e93c4 Fix my last commits (unused variable warning)
my gcc at home did not show this warning...
2017-02-13 22:27:59 +01:00
Dirk Ziegelmeier
3c1713406e Work on task #14283: Add more detailed status callback
Cleanups and IPv6 address status
2017-02-13 21:52:55 +01:00
Dirk Ziegelmeier
58c48af58f Start working on task #14283: Add more detailed status callback
IPv6 stuff is not implemented
2017-02-13 21:40:08 +01:00