open系统调用的O_CREAT和O_EXCL

     O_CREAT
              If the file does not exist, it will be created.  The owner (user ID) of the file is set to the effective user ID of the process.  The group ownership
              (group  ID) is set either to the effective group ID of the process or to the group ID of the parent directory (depending on filesystem type and mount
              options, and the mode of the parent directory; see the mount options bsdgroups and sysvgroups described in mount(8)).


              mode specifies the mode to use in case a new file is created.  This argument must be supplied when O_CREAT or O_TMPFILE is  specified  in  flags;  if
              neither  O_CREAT  nor  O_TMPFILE  is specified, then mode is ignored.  The effective mode is modified by the process's umask in the usual way: in the
              absence of a default ACL, the mode of the created file is (mode & ~umask).  Note that this mode applies only to future accesses of the newly  created
              file; the open() call that creates a read-only file may well return a read/write file descriptor.


              The following symbolic constants are provided for mode:


              S_IRWXU  00700 user (file owner) has read, write, and execute permission


              S_IRUSR  00400 user has read permission


              S_IWUSR  00200 user has write permission


              S_IXUSR  00100 user has execute permission


              S_IRWXG  00070 group has read, write, and execute permission


              S_IRGRP  00040 group has read permission


              S_IWGRP  00020 group has write permission


              S_IXGRP  00010 group has execute permission


              S_IRWXO  00007 others have read, write, and execute permission


              S_IROTH  00004 others have read permission


              S_IWOTH  00002 others have write permission


              S_IXOTH  00001 others have execute permission


              According to POSIX, the effect when other bits are set in mode is unspecified.  On Linux, the following bits are also honored in mode:


              S_ISUID  0004000 set-user-ID bit


              S_ISGID  0002000 set-group-ID bit (see stat(2))

              S_ISVTX  0001000 sticky bit (see stat(2))

       O_EXCL Ensure that this call creates the file: if this flag is specified in conjunction with O_CREAT, and pathname already exists, then open() will fail.


              When these two flags are specified, symbolic links are not followed: if pathname is a symbolic link, then open() fails regardless of where  the  sym‐
              bolic link points to.


              In  general,  the  behavior of O_EXCL is undefined if it is used without O_CREAT.  There is one exception: on Linux 2.6 and later, O_EXCL can be used
              without O_CREAT if pathname refers to a block device.  If the block device is in use by the system (e.g.,  mounted),  open()  fails  with  the  error
              EBUSY.


              On  NFS,  O_EXCL  is supported only when using NFSv3 or later on kernel 2.6 or later.  In NFS environments where O_EXCL support is not provided, pro‐
              grams that rely on it for performing locking tasks will contain a race condition.  Portable programs that want to perform atomic file locking using a
              lockfile,  and  need  to  avoid reliance on NFS support for O_EXCL, can create a unique file on the same filesystem (e.g., incorporating hostname and
              PID), and use link(2) to make a link to the lockfile.  If link(2) returns 0, the lock is successful.  Otherwise, use stat(2) on the  unique  file  to
              check if its link count has increased to 2, in which case the lock is also successful.