Mirrors_Framework/nuttx
1
0
Fork 0
mirror of https://github.com/apache/nuttx.git synced 2025-12-21 03:15:18 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

Signals were not a good choice of IPC to implement the poll function for several reasons: In order to handle the asynchrnous poll-related event, a substantial amount of state information is needed. Signals are only capable of passing minimal amounts of data. There are also complexities with performing kernel space signal handlers in kernel space code that is better to avoid. So, instead of signals, the equivalent logic was converted to run via a callback that executes on the high-priority work queue.

Browse Source 
Squashed commit of the following:

    Fix up some final compile isses.

    net/netdev:  Convert the network down notification logic to use the new wqueue-based notification factility.

    net/udp:  Convert the UDP readahead notification logic to use the new wqueue-based notification factility.

    net/tcp:  Convert the TCP readahead notification logic to use the new wqueue-based notification factility.

    mm/iob:  Convert the IOB notification logic to use the new wqueue-based notification factility.

    sched/wqueue:  Signals are not good IPCs to support the target poll functionality for several reasons including the amount of data that can be passed with a signal and in the fact that in protected and kernel modes, user threads executing signal handlers in protected, kernel memory is problematic.  Instead, convert the same logic to perform the notifications via function callback on the high priority work queue.
This commit is contained in:
Gregory Nutt
2018-09-09 15:01:44 -06:00
parent 20814acad2
commit 9d3148406c
23 changed files with 570 additions and 448 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
net/tcp/tcp.h
View File

@@ -49,6 +49,10 @@
#include <nuttx/mm/iob.h>
#include <nuttx/net/ip.h>
#ifdef CONFIG_TCP_READAHEAD_NOTIFIER
# include <nuttx/wqueue.h>
#endif
#if defined(CONFIG_NET_TCP) && !defined(CONFIG_NET_TCP_NO_STACK)
/****************************************************************************
@@ -1546,21 +1550,16 @@ int tcp_pollteardown(FAR struct socket *psock, FAR struct pollfd *fds);
* Name: tcp_notifier_setup
*
* Description:
* Set up to notify the specified PID with the provided signal number.
*
* NOTE: To avoid race conditions, the caller should set the sigprocmask
* to block signal delivery. The signal will be delivered once the
* signal is removed from the sigprocmask.
*
* NOTE: If sigwaitinfo() or sigtimedwait() are used to catch the signal
* then then TCP connection structure pointer may be recovered in the
* sival_ptr value of the struct siginfo instance.
* Set up to perform a callback to the worker function when an TCP data
* is added to the read-ahead buffer. The worker function will execute
* on the high priority worker thread.
*
* Input Parameters:
* pid - The PID to be notified. If a zero value is provided, then the
* PID of the calling thread will be used.
* signo - The signal number to use with the notification.
* worker - The worker function to execute on the high priority work
* queue when data is available in the TCP readahead buffer.
* conn - The TCP connection where read-ahead data is needed.
* arg - A user-defined argument that will be available to the worker
* function when it runs.
*
* Returned Value:
* > 0 - The signal notification is in place. The returned value is a
@@ -1575,7 +1574,8 @@ int tcp_pollteardown(FAR struct socket *psock, FAR struct pollfd *fds);
****************************************************************************/
#ifdef CONFIG_TCP_READAHEAD_NOTIFIER
int tcp_notifier_setup(int pid, int signo, FAR struct tcp_conn_s *conn);
int tcp_notifier_setup(worker_t worker, FAR struct tcp_conn_s *conn,
FAR void *arg);
#endif
/****************************************************************************
@@ -1608,14 +1608,10 @@ int tcp_notifier_teardown(int key);
* Read-ahead data has been buffered. Signal all threads waiting for
* read-ahead data to become available.
*
* When the read-ahead data becomes available, *all* of the waiters in
* this thread will be signaled. If there are multiple waiters then only
* the highest priority thread will get the data. Lower priority threads
* will need to call tcp_notifier_setup() once again.
*
* NOTE: If sigwaitinfo() or sigtimedwait() are used to catch the signal
* then then TCP connection structure pointer may be obtained in the
* sival_ptr value of the struct siginfo instance.
* When read-ahead data becomes available, *all* of the workers waiting
* for read-ahead data will be executed. If there are multiple workers
* waiting for read-ahead data then only the first to execute will get the
* data. Others will need to call tcp_notifier_setup() once again.
*
* Input Parameters:
* conn - The TCP connection where read-ahead data was just buffered.