Commit Graph

39 Commits

Author SHA1 Message Date
Alin Jerpelea 67d02a45eb net: migrate to SPDX identifier
Most tools used for compliance and SBOM generation use SPDX identifiers
This change brings us a step closer to an easy SBOM generation.

Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
2024-09-12 01:08:11 +08:00
Xiang Xiao eddd90de78 poll: pollsetup should notify only one fd passd by caller
since it's redundant to iterate the whole fds array in setup

Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
2023-11-21 09:07:17 +01:00
Zhe Weng 075eb6a6d2 net/udp: Change conn->readahead to I/O buffer chain
When using IOB queue to store readahead data, we use one IOB for each
UDP packet. Then if the packets are very small, like 10Bytes per packet,
we'll use ~1600 IOBs just for 16KB recv buffer size, which is wasteful
and dangerous. So change conn->readahead to a single IOB chain like TCP.

Benefits:
- Using memory and IOBs more efficiently (small packets are common in
  UDP)

Side effects:
- UDP recv buffer size may count the overhead
- A little bit drop in performance (<1%, more seek & copy)

Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-07-18 10:51:45 +08:00
wangbowen6 344c8be049 poll: add poll_notify() api and call it in all drivers
Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com>
2022-09-26 12:06:32 +08:00
chao.an 162fcd10ca net: cleanup pvconn reference to avoid confuse
More reference:
https://github.com/apache/incubator-nuttx/pull/5252
https://github.com/apache/incubator-nuttx/pull/5434

Signed-off-by: chao.an <anchao@xiaomi.com>
2022-08-26 20:58:11 +08:00
chao.an 8ae8c10954 net/poll: fix race condition if connect free before poll teardown
Net poll teardown is not protected by net lock, if the conn is released
before teardown, the assertion failure will be triggered during free dev
callback, this patch will add the net lock around net poll teardown to
fix race condition

nuttx/libs/libc/assert/lib_assert.c:36
nuttx/net/devif/devif_callback.c:85
nuttx/net/tcp/tcp_netpoll.c:405
nuttx/fs/vfs/fs_poll.c:244
nuttx/fs/vfs/fs_poll.c:500

Signed-off-by: chao.an <anchao@xiaomi.com>
2022-07-09 19:11:42 +08:00
chao.an eac13a113d net/udp: remove psock hook to avoid invalid reference
Signed-off-by: chao.an <anchao@xiaomi.com>
2022-02-10 15:04:33 -03:00
Alin Jerpelea 37d5c1b0d9 net: Author Gregory Nutt: update licenses to Apache
Gregory Nutt has submitted the SGA and we can migrate the licenses
 to Apache.

Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
2021-02-20 00:38:18 -08:00
Xiang Xiao 63e3054ced Don't need monitor IOB buffer empty event for POLLOUT implementation
It's enough to check the buffer available in the net event handler

Change-Id: I2d7c7a03675cf6eff6ffb42a81b7c7245253e92c
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
2020-05-13 06:50:07 -06:00
Xiang Xiao 346336bb9e Make the read ahead buffer unselectable
Here is the email loop talk about why it is better to remove the option:
https://groups.google.com/forum/#!topic/nuttx/AaNkS7oU6R0

Change-Id: Ib66c037752149ad4b2787ef447f966c77aa12aad
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
2020-01-11 08:24:49 -06:00
Xiang Xiao 02b652ce1c Minor fix (#23)
* fix ieee802154/ieee802154_input.c:179:7: error: too few arguments to function 'iob_free'
       iob_free(container->ic_iob);
       ^~~~~~~~

ieee802154/ieee802154_input.c:180:7: error: too many arguments to function 'ieee802154_container_free'
       ieee802154_container_free(container, IOBUSER_NET_SOCK_IEEE802154);
       ^~~~~~~~~~~~~~~~~~~~~~~~~

* fix udp/udp_netpoll.c:327:10: warning: 'ret' may be used uninitialized in this function [-Wmaybe-uninitialized]

* fix local/local_netpoll.c:154:15: warning: implicit declaration of function 'nxsem_post'; did you mean 'sem_post'? [-Wimplicit-function-declaration]
               nxsem_post(fds->sem);
               ^~~~~~~~~~
               sem_post
2020-01-02 09:35:50 -06:00
Xiang Xiao 90c52e6f8f Squashed commit of the following:
Author: Gregory Nutt <gnutt@nuttx.org>

    Run all .h and .c files modified in last PR through nxstyle.

Author: Xiang Xiao <xiaoxiang@xiaomi.com>

    Net cleanup (#17)

    * Fix the semaphore usage issue found in tcp/udp

    1. The count semaphore need disable priority inheritance
    2. Loop again if net_lockedwait return -EINTR
    3. Call nxsem_trywait to avoid the race condition
    4. Call nxsem_post instead of sem_post

    * Put the work notifier into free list to avoid the heap fragment in the long run.  Since the allocation strategy is encapsulated internally, we can even refine the implementation later.

    * Network stack shouldn't allocate memory in the poll implementation to avoid the heap fragment in the long run, other modification include:

    1. Select MM_IOB automatically since ICMP[v6] socket can't work without the read ahead buffer
    2. Remove the net lock since xxx_callback_free already do the same thing
    3. TCP/UDP poll should work even the read ahead buffer isn't enabled at all

    * Add NET_ prefix for UDP_NOTIFIER and TCP_NOTIFIER option to align with other UDP/TCP option convention

    * Remove the unused _SF_[IDLE|ACCEPT|SEND|RECV|MASK] flags since there are code to set/clear these flags, but nobody check them.
2019-12-31 09:26:14 -06:00
Xiang Xiao 1905e01fda net/: ICMP/ICMPv6/TCP/UDP poll shouldn't set POLLHUP and POLLOUT at the same time the standard require that only report POLLHUP:
https://pubs.opengroup.org/onlinepubs/009695399/functions/poll.html.
2019-11-25 09:59:50 -06:00
Xiang Xiao af9c67ab58 net/tcp/tcp_netpoll.c and net/udp/udp_netpoll.c: In [tcp|udp]_iob_work fix the wrong condition logic (& vs &&). 2019-11-25 09:32:24 -06:00
Xiang Xiao 98fc60eb75 [tcp|udp]_poll_eventhandler check psock_[tcp|udp]_cansend before report POLLOUT. Change the unbuffered psock_[tcp|udp]_cansend return OK to unify the code logic and remove the unnecessary [tcp|udp]_poll_txnotify call. 2019-11-24 10:30:48 -06:00
Xiang Xiao 893fc6e1cf net/udp/udp_netpoll.c: Report POLLHUP and POLLERR in 'revents' regardless the requested 'events' set. Per Opengroup.org, these bits must be ignored in the 'events' set. 2019-11-24 10:18:08 -06:00
Gregory Nutt 1346f29151 net/: Fix alignment and spacing problems found by tools/nxstyle. 2019-07-02 18:02:23 -06:00
Gregory Nutt 88130a996a net/: Fixes for more coding standard issues detected by nxstyle. 2018-10-29 18:00:30 -06:00
Gregory Nutt 6e60af91e4 net/udp/udp_netpoll.c: Port TCP poll changes to UDP. 2018-09-12 10:10:54 -06:00
Gregory Nutt a680553f35 net/tcp/tcp_netpoll.c: Add logic to receive notifications when IOBs are freed (Needs CONFIG_NET_TCP_WRITE_BUFFERS and CONFIG_IOB_NOTIFIER). At present, does nothing because the logic in in psock_tcp_cansend() does not check for the availability of IOBs. That will change. 2018-09-12 06:40:18 -06:00
Gregory Nutt 884ee6e43e net/tcp and net/udp: In the POLLOUT poll logic, request an immediate Tx poll from the network device bound to the socket. This obviously cannot work if there is not single device bound to the network device. 2018-09-11 10:31:11 -06:00
Gregory Nutt fb8cf9373c net/udp: Resolve race condition in connection-less UDP sockets with read-ahead buffering.
In connection-mode UDP sockets, a remote address is retained in the UDP connection structure.  This determines both there send() will send the packets and which packets recv() will accept.

This same mechanism is used for connection-less UDP sendto:  A temporary remote address is written into the connection structure to support the sendto() operation.  That address persists until the next recvfrom() when it is reset to accept any address.

When UDP read-ahead buffering is enabled, however, that means that the old, invalid remote address can be left in the connection structure for some time.  This can cause read-ahead buffer to fail, dropping UDP packets.

Shortening the time between when he remote address is reset (i.e., immediately after the sendto() completes) is not a solution, that does not eliminate the race condition; in only makes it smaller.

With this change, a flag was added to the connection structure to indicate if the UDP socket is in connection mode or if it is connection-less.  This change effects only UDP receive operations:  The remote address in the UDP connection is always ignored if the UDP socket is not in connection-mode.

No for connection-mode sockets, that remote address behaves as before.  But for connection-less sockets, it is only used by sendto().
2018-05-13 09:57:34 -06:00
Gregory Nutt 091e3f732e net/udp: Fix several issues using connected UDP sockets: sendto() should not accept and address. Normally send() should be used instead. recvfrom() is no longer necessary. recvfrom() and poll() should not reset to accept any addresses but should, instead, only accept inputs and events from the connected remore peer. On of these issues was noted by 李桂丁 <liguiding@pinecone.net>. Changes were made as summarized here: http://www.masterraghu.com/subjects/np/introduction/unix_network_programming_v1.3/ch08lev1sec11.html 2018-04-25 08:43:38 -06:00
Gregory Nutt b54ffe858a Standardization of some function headers. 2018-03-13 09:52:27 -06:00
Gregory Nutt 82cb799bb6 net/udp: Fix memory leak with UDP + write buffer is closed. Also update TODO and comments and refresh a configuration. 2018-01-23 11:54:03 -06:00
Gregory Nutt cccc86da09 This change adds support for semi-standard IPPROTO_ICMP AF_INET datagram sockets. This replaces the old ad hoc, nonstandard way of implementing ping with a more standard, socket interface.
Squashed commit of the following:

    net/icmp:  Finishes off icmp_recvfrom().
    net/icmp:  Add readahead support for IPPROTO_ICMP sockets.
    net/icmp:  Add poll() support for IPPROTO_ICMP sockets.
    net/icmp:  Add a connection structure for IPPROTO_ICMP sockets.
    net/icmp:  Implements sendto for the IPPROTO_ICMP socket.
    net/icmp:  Move icmp_sendto() and icmp_recvfrom() to separate files.  They are likely to be complex (when they are implemented).
    net/icmp:  Hook IPPROTO_ICMP sockets into network.  Fix some naming collisions.  Still missing basic ICMP send/receive logic.
    configs: apps/system/ping current need poll() enabled.
    configs: All defconfig files that use to enable low-level support must now enabled CONFIG_SYSTEM_PING.
    net/icmp:  Adds basic build framework to support IPPROTO_ICMP sockets.
2017-10-23 08:45:12 -06:00
Gregory Nutt cf5cba953d There was a possible recursion that could eventually overflow the stack. The error occurred when closing the socket with inet_close() while a socket callback was still queued. When the socket callback was executed by devif_conn_event(), this resulted in a call to psock_send_eventhandler() with TCP_CLOSE flag set which then called tcp_lost_connection(). tcp_shutdown_monitor() then called tcp_callback() again, which again called psock_send_eventhandler(), and so on.... Noted by Pascal Speck. Solution is also similar to a solution proposed by Pascal Speck. 2017-10-19 11:55:51 -06:00
Gregory Nutt 42a0796615 Squashed commit of the following:
sched/semaphore:  Add nxsem_post() which is identical to sem_post() except that it never modifies the errno variable.  Changed all references to sem_post in the OS to nxsem_post().

    sched/semaphore:  Add nxsem_destroy() which is identical to sem_destroy() except that it never modifies the errno variable.  Changed all references to sem_destroy() in the OS to nxsem_destroy().

    libc/semaphore and sched/semaphore:  Add nxsem_getprotocol() and nxsem_setprotocola which are identical to sem_getprotocol() and set_setprotocol() except that they never modifies the errno variable.  Changed all references to sem_setprotocol in the OS to nxsem_setprotocol().  sem_getprotocol() was not used in the OS
2017-10-03 15:35:24 -06:00
Gregory Nutt 8ffb103adb networking: IGMP: Remove special support for interrupt level processing (there is none) and fix some timer cancellation logic. In many files, correct comments. There is no interrupt level processing in the networking layer. 2017-09-02 10:27:03 -06:00
Gregory Nutt 0b2a4eb4bd Networking: A little more wording changes related to interrupts vs. events 2017-08-29 15:08:38 -06:00
Gregory Nutt 04ad162540 Networking: Clean up some naming that has bothered me for a long time... There are no interrupts and no interrupt handlers in the network. There are events and event handler (there used to to be interrupt logic in there years ago but that is long, long gone). 2017-08-29 14:08:04 -06:00
Gregory Nutt d5207efb5a Be consistent... Use Name: consistent in function headers vs Function: 2017-04-21 16:33:14 -06:00
Gregory Nutt 7467329a98 Eliminate CONFIG_NO_NOINTS. Lots of files changed -> lots of testing needed. 2016-12-03 16:28:19 -06:00
Gregory Nutt 43eb04bb8f Without lowsyslog() *llinfo() is not useful. Eliminate and replace with *info(). 2016-06-20 11:59:15 -06:00
Gregory Nutt 1cdc746726 Rename CONFIG_DEBUG to CONFIG_DEBUG_FEATURES 2016-06-11 14:14:08 -06:00
Gregory Nutt fc3540cffe Replace all occurrences of vdbg with vinfo 2016-06-11 11:59:51 -06:00
Gregory Nutt edbb5d0840 networking: Correct a typo introduced with recent networking changes 2015-06-02 15:04:02 -06:00
Gregory Nutt e672bcebfd UDP Networking: Add support for device event notification for UDP transfers. 2015-05-29 10:21:06 -06:00
Gregory Nutt de91d34a19 Networking: Separate out UDP poll logic from socket/net_poll.c into a new udp/udp_netpoll.c; Create a skeleton local/local_netpoll.c for future poll support on Unix domain sockets. 2015-01-30 07:09:25 -06:00