Re: Change in functionality of futex() system call.

[Eric Dumazet <eric.dumazet@gmail.com>]

Le lundi 06 juin 2011 à 10:53 -0700, Darren Hart a écrit :
> 

> If I understand the problem correctly, RO private mapping really doesn't
> make any sense and we should probably explicitly not support it, while
> special casing the RO shared mapping in support of David's scenario.
> 

We supported them in 2.6.18 kernels, apparently. This might sounds
stupid but who knows ?

> > 
> > In David Oliver case, this is needed : He wants to catch a change in a
> > file/memory region written by another process.
> 
> But with shared mapping and shared futexes. He just needs the ability to
> FUTEX_WAIT on a RO mapping. Or is that what you were saying?
> 

I am saying that in David Oliver case, he sure uses a MAP_SHARED ro
mapping.

Now, what if other software uses a MAP_PRIVATE ro mapping ?

It was working in previous kernels as well.

We can say its stupid, but IMHO its not.

In other words, this program should work, if process never touches
(writes) into first page.

This program on previous kernels gave :
rc=-1 errno=11
(allowing to wait for a value change and a futex_WAKE)

With new kernel :
rc=-1 errno=14  [ no sleep allowed ]

#include <errno.h>
#include <fcntl.h>
#include <stdint.h>
typedef uint32_t u32;   // for futex.h
#include <linux/futex.h>
#include <sys/mman.h>
#include <sys/syscall.h>
#include <unistd.h>


int main() {
 	int fd, *futex, rc;

	fd = open("/tmp/futex_test", O_RDWR|O_CREAT, 0644);
	write(fd, "\1\1\1\1", 4);
	futex = (int *)mmap(0, sizeof(int), PROT_READ, MAP_PRIVATE, fd, 0);
	rc = syscall(SYS_futex, futex, FUTEX_WAIT, 42, 0, 0, 0);
	printf("rc=%d errno=%d\n", rc, errno);
}