CODEWITHSUNDEEP
clinuxposixsemaphore
Ruslan R. Laishev
Ruslan R. Laishev

Reputation: 143

A SIGSEGV signal on calling sem_open()

I use follows piece of code to check how to use semaphores :

char    sema_name [NAME_MAX];
sem_t   *sema_hnd = NULL;
const mode_t semprot = 0777; /*S_IRWXU | S_IRWXG ;*/

int cid = 79;


int main()
{

    sprintf(sema_name, "/StarLet-TV-CID=%d", cid);

    if ( !(sema_hnd = sem_open(sema_name, O_CREAT, semprot, cid)) )
        printf("errno=%d", errno);


    printf("sema_hnd=%#x", sema_hnd);
}

Getting SIGSEGV in the sem_open(), so what I do wrong?

root@SysMan-Ubuntu:/dev/shm# uname -a
Linux SysMan-Ubuntu 4.15.0-34-generic #37-Ubuntu SMP Mon Aug 27 15:21:48 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

It's looks like that trouble in the shm-directory.c module:

__shm_directory (size_t *len)
{
  /* Determine where the shmfs is mounted.  */
  __libc_once (once, where_is_shmfs);

  /* If we don't know the mount points there is nothing we can do.  Ever.  */
  if (__glibc_unlikely (mountpoint.dir == NULL))
    {
      __set_errno (ENOSYS);
      return NULL;
    }

  *len = mountpoint.dirlen;
  return mountpoint.dir;
}

At this line "if (__glibc_unlikely (mountpoint.dir == NULL))"

Take a look to : 

root@SysMan-Ubuntu:/home/sysman# cd /dev/shm
    root@SysMan-Ubuntu:/dev/shm# ll
    total 0
    drwxrwxrwt  2 root root   40 окт 13 19:17 ./
    drwxr-xr-x 18 root root 3860 окт 13 19:18 ../
    root@SysMan-Ubuntu:/dev/shm# ls >zz.log
    root@SysMan-Ubuntu:/dev/shm# ll
    total 4
    drwxrwxrwt  2 root root   60 окт 13 19:19 ./
    drwxr-xr-x 18 root root 3860 окт 13 19:18 ../
    -rw-r--r--  1 root root    7 окт 13 19:19 zz.log
    root@SysMan-Ubuntu:/dev/shm# cat zz.log
    zz.log
    root@SysMan-Ubuntu:/dev/shm#


root@SysMan-Ubuntu:/dev/shm# df
Filesystem     1K-blocks      Used Available Use% Mounted on
udev              990112         0    990112   0% /dev
tmpfs             204076     21896    182180  11% /run
/dev/sda1       48250196  34281344  11865744  75% /
tmpfs            1020376         4   1020372   1% /dev/shm
tmpfs               5120         0      5120   0% /run/lock
tmpfs            1020376         0   1020376   0% /sys/fs/cgroup
Downloads      976759804 802679056 174080748  83% /media/sf_Downloads
Works          976759804 802679056 174080748  83% /media/sf_Works
tmpfs             204072        12    204060   1% /run/user/122
tmpfs             204072        20    204052   1% /run/user/1000
/dev/sr0           56618     56618         0 100% /media/sysman/VBox_GAs_5.2.18
tmpfs             204072         0    204072   0% /run/user/0
root@SysMan-Ubuntu:/dev/shm#


root@SysMan-Ubuntu:/dev/shm# mount | grep shm
tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev)
root@SysMan-Ubuntu:/dev/shm#

Moreover, the first run of the program was successful and the semaphore with the name 'StarLet ..." was created. After second run the code is finished with SIGSEGV at sem_open(). Reboot was not help.

Upvotes: 2

Views: 264

Answers (1)

alk
alk

Reputation: 70893

Your system seems to be missing the filesystem for named semphores:

From man 7 sem_overview:

Accessing named semaphores via the filesystem

On Linux, named semaphores are created in a virtual filesystem, normally mounted under /dev/shm, with names of the form sem.somename. (This is the reason that semaphore names are limited to NAME_MAX-4 rather than NAME_MAX characters.)

Use the mount command to check if shm is mounted as type tmpfs.

The list returned by mount should contain an entry like:

tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev)

And if it isn't mounted, mounting can be done by:

mount -t tmpfs tmpfs /dev/shm -o nosuid,nodev

Upvotes: 2

Related Questions