Re: [RFC] epoll

[Davide Libenzi]

On Thu, 22 Sep 2005, Vadim Lobanov wrote:

> On Thu, 22 Sep 2005, Davide Libenzi wrote:
>
> > This born from the initial epoll implementation, that was using the size
> > parameter to size the hash. Since now epoll uses rbtrees, the parameter is
> > no more used. Changing the API to remove it, and break userspace, was/is
> > not a good idea.
>
> Any thoughts on allowing a size of '0'?

We would need to change man pages for that, and this IMHO w/out a strong 
reason of doing so.



> > Yes, it is nicer if epoll behaves like poll. (*)
>
> No patch needed from me here, then?

No need thx. I'll be sending the one-liner to the list right now.



> > > 3. Wakeup
> > > As determined by testing with userland code, the sys_tgkill() and
> > > sys_tkill() functions currently will NOT wake up a sleeping
> > > epoll_wait(). Effectively, this means that epoll_wait() is NOT a pthread
> > > cancellation point. There are two potential issues with this:
> > > - epoll_wait() meets the unofficial(?) definition of a "system call that
> > > may block".
> > > - epoll_wait() behaves differently from poll() and friends.
> >
> > The epoll_wait() wait loop is the standard one that even poll() uses (prep
> > wait, make interruptible, test signals, sched timeo). So if poll() is woke
> > up, so should epoll_wait(). A minimal code snippet that proves poll()
> > behing woke up, and epoll_wait() not, would help.
>
> Certainly. :-) See end of email for sample program.

I'm afraid you need to bug the glibc guys, since I think they wrap 
sys_poll(). Try the test program below, when defining _X_, that makes it 
call sys_poll() directly. It will have the same epoll_wait() behaviour.




- Davide





#include <sys/epoll.h>
#include <sys/poll.h>
#include <pthread.h>
#include <stdio.h>
#include <linux/unistd.h>
#include <errno.h>
#include <unistd.h>


#define __NR_xpoll __NR_poll

#define __sys_xpoll(ufds, nfds, timeout) _syscall3(int, xpoll, struct pollfd *, ufds, \
						   unsigned int, nfds, long, timeout)

__sys_xpoll(ufds, nfds, timeout);


void * run (int * fd) {
	struct epoll_event result;

	printf("Wait forever while polling.\n");
#if defined(_E_)
	epoll_wait(*fd, &result, 1, -1);
#elif defined(_X_)
	xpoll(NULL, 0, -1);
#else
	poll(NULL, 0, -1);
#endif
	printf("Uhoh! Something is borked!\n");

	return NULL;
}

int main (void) {
	int events;
	pthread_t thread;

#if defined(_E_)
	events = epoll_create(1);
#endif
	pthread_create(&thread, NULL, (void * (*) (void *))run, &events);
	getchar();
	printf("Try to kill the thread.\n");
	pthread_cancel(thread);
	pthread_join(thread, NULL);
	printf("Success.\n");
#if defined(_E_)
	close(events);
#endif

	return 0;
}

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/