minix/lib/libasyn/asyn_special.c

105 lines
3.1 KiB
C
Raw Normal View History

This patch switches the MINIX3 ethernet driver stack from a port-based model to an instance-based model. Each ethernet driver instance is now responsible for exactly one network interface card. The port field in /etc/inet.conf now acts as an instance field instead. This patch also updates the data link protocol. This update: - eliminates the concept of ports entirely; - eliminates DL_GETNAME entirely; - standardizes on using m_source for IPC and DL_ENDPT for safecopies; - removes error codes from TASK/STAT replies, as they were unused; - removes a number of other old or unused fields; - names and renames a few other fields. All ethernet drivers have been changed to: - conform to the new protocol, and exactly that; - take on an instance number based on a given "instance" argument; - skip that number of PCI devices in probe iterations; - use config tables and environment variables based on that number; - no longer be limited to a predefined maximum of cards in any way; - get rid of any leftover non-safecopy support and other ancient junk; - have a correct banner protocol figure, or none at all. Other changes: * Inet.conf is now taken to be line-based, and supports #-comments. No existing installations are expected to be affected by this. * A new, select-based asynchio library replaces the old one. Kindly contributed by Kees J. Bot. * Inet now supports use of select() on IP devices. Combined, the last two changes together speed up dhcpd considerably in the presence of multiple interfaces. * A small bug has been fixed in nonamed.
2010-05-18 00:22:53 +02:00
/* asyn_special(), asyn_result() Author: Kees J. Bot
* 8 Jul 1997
*/
#include "asyn.h"
#include <signal.h>
/* Saved signal mask between asyn_special() and asyn_result(). */
static sigset_t mask;
int asyn_special(asynchio_t *asyn, int fd, int op)
/* Wait for an operation. This is an odd one out compared to asyn_read()
* and asyn_write(). It does not do an operation itself, but together with
* asyn_result() it is a set of brackets around a system call xxx that has
* no asyn_xxx() for itself. It can be used to build an asyn_accept() or
* asyn_connect() for instance. (Minix-vmd has asyn_ioctl() instead,
* which is used for any other event like TCP/IP listen/connect. BSD has
* a myriad of calls that can't be given an asyn_xxx() counterpart each.)
* Asyn_special() returns -1 for "forget it", 0 for "try it", and 1 for
* "very first call, maybe you should try it once, maybe not". Errno is
* set to EAGAIN if the result is -1 or 1. After trying the system call
* make sure errno equals EAGAIN if the call is still in progress and call
* asyn_result with the result of the system call. Asyn_result() must be
* called if asyn_special() returns 0 or 1.
*
* Example use:
*
* int asyn_accept(asynchio_t *asyn, int s, struct sockaddr *addr, int *addrlen)
* {
* int r;
* if ((r= asyn_special(asyn, fd, SEL_READ)) < 0) return -1;
* r= r == 0 ? accept(fd, addr, addrlen) : -1;
* return asyn_result(asyn, fd, SEL_READ, r);
* }
*
* int asyn_connect(asynchio_t *asyn, int s, struct sockaddr *name, int namelen)
* {
* int r;
* if ((r= asyn_special(asyn, fd, SEL_WRITE)) < 0) return -1;
* if (r == 1 && (r= connect(fd, name, namelen)) < 0) {
* if (errno == EINPROGRESS) errno= EAGAIN;
* }
* return asyn_result(asyn, fd, SEL_WRITE, r);
* }
*/
{
asynfd_t *afd;
int seen;
asyn->asyn_more++;
if ((unsigned) fd >= FD_SETSIZE) { errno= EBADF; return -1; }
afd= &asyn->asyn_afd[fd];
/* If this is the first async call on this filedescriptor then
* remember its file flags.
*/
if (!(seen= afd->afd_seen)) {
if ((afd->afd_flags= fcntl(fd, F_GETFL)) < 0) return -1;
afd->afd_seen= 1;
}
/* Try to read if I/O is pending. */
if (!seen || afd->afd_state[op] == PENDING) {
sigemptyset(&mask);
if (sigprocmask(SIG_SETMASK, &mask, &mask) < 0) return -1;
(void) fcntl(fd, F_SETFL, afd->afd_flags | O_NONBLOCK);
/* Let the caller try the system call. */
errno= EAGAIN;
return seen ? 0 : 1;
}
/* Record this read as "waiting". */
afd->afd_state[op]= WAITING;
FD_SET(fd, &asyn->asyn_fdset[op]);
errno= EAGAIN;
asyn->asyn_more--;
return -1;
}
int asyn_result(asynchio_t *asyn, int fd, int op, int result)
/* The caller has tried the system call with the given result. Finish up. */
{
int err;
asynfd_t *afd= &asyn->asyn_afd[fd];
err= errno;
(void) fcntl(fd, F_SETFL, afd->afd_flags);
(void) sigprocmask(SIG_SETMASK, &mask, nil);
errno= err;
if (result != -1 || errno != EAGAIN) {
afd->afd_state[op]= IDLE;
return result;
}
/* Record this operation as "waiting". */
afd->afd_state[op]= WAITING;
FD_SET(fd, &asyn->asyn_fdset[op]);
errno= EAGAIN;
asyn->asyn_more--;
return -1;
}