close(): Difference between revisions

From vegard.wiki
Jump to navigation Jump to search
Content added Content deleted
(new page)
 
(fix line wrapping)
 
Line 3: Line 3:
From the [https://github.com/mkerrisk/man-pages/blob/5373f62f1e4352e665c24dfe49b7e3fe03721cab/man2/close.2#L134 current Linux man page]:
From the [https://github.com/mkerrisk/man-pages/blob/5373f62f1e4352e665c24dfe49b7e3fe03721cab/man2/close.2#L134 current Linux man page]:
<blockquote>
<blockquote>
A careful programmer will check the return value of <tt>close()</tt>, since it is quite possi‐
A careful programmer will check the return value of <tt>close()</tt>, since it is quite possible
ble that errors on a previous <tt>write(2)</tt> operation are reported only on the final
that errors on a previous <tt>write(2)</tt> operation are reported only on the final
<tt>close()</tt> that releases the open file description. Failing to check the return value
<tt>close()</tt> that releases the open file description. Failing to check the return value
when closing a file may lead to silent loss of data. This can especially be observed
when closing a file may lead to silent loss of data. This can especially be observed
Line 11: Line 11:
Note, however, that a failure return should be used only for diagnostic purposes
Note, however, that a failure return should be used only for diagnostic purposes
(i.e., a warning to the application that there may still be I/O pending or there may
(i.e., a warning to the application that there may still be I/O pending or there may
have been failed I/O) or remedial purposes (e.g., writing the file once more or cre‐
have been failed I/O) or remedial purposes (e.g., writing the file once more or creating
ating a backup).
a backup).


Retrying the <tt>close()</tt> after a failure return is the wrong thing to do, since this may
Retrying the <tt>close()</tt> after a failure return is the wrong thing to do, since this may
cause a reused file descriptor from another thread to be closed. This can occur
cause a reused file descriptor from another thread to be closed. This can occur
because the Linux kernel always releases the file descriptor early in the close oper‐
because the Linux kernel always releases the file descriptor early in the close operation,
ation, freeing it for reuse; the steps that may return an error, such as flushing
freeing it for reuse; the steps that may return an error, such as flushing
data to the filesystem or device, occur only later in the close operation.
data to the filesystem or device, occur only later in the close operation.


Many other implementations similarly always close the file descriptor (except in the
Many other implementations similarly always close the file descriptor (except in the
case of <tt>EBADF</tt>, meaning that the file descriptor was invalid) even if they subse‐
case of <tt>EBADF</tt>, meaning that the file descriptor was invalid) even if they subsequently
quently report an error on return from <tt>close()</tt>. POSIX.1 is currently silent on this
report an error on return from <tt>close()</tt>. POSIX.1 is currently silent on this
point, but there are plans to mandate this behavior in the next major release of the
point, but there are plans to mandate this behavior in the next major release of the
standard
standard

Latest revision as of 06:48, 18 December 2019

close() is notoriously difficult to get right. You are supposed to check the return value, but not really handle it (other than printing a diagnostic). If you are wrapping a file descriptor using RAII, then you may not be able to throw an exception or return an error if you are calling close() from the destructor.

From the current Linux man page:

A careful programmer will check the return value of close(), since it is quite possible that errors on a previous write(2) operation are reported only on the final close() that releases the open file description. Failing to check the return value when closing a file may lead to silent loss of data. This can especially be observed with NFS and with disk quota.

Note, however, that a failure return should be used only for diagnostic purposes (i.e., a warning to the application that there may still be I/O pending or there may have been failed I/O) or remedial purposes (e.g., writing the file once more or creating a backup).

Retrying the close() after a failure return is the wrong thing to do, since this may cause a reused file descriptor from another thread to be closed. This can occur because the Linux kernel always releases the file descriptor early in the close operation, freeing it for reuse; the steps that may return an error, such as flushing data to the filesystem or device, occur only later in the close operation.

Many other implementations similarly always close the file descriptor (except in the case of EBADF, meaning that the file descriptor was invalid) even if they subsequently report an error on return from close(). POSIX.1 is currently silent on this point, but there are plans to mandate this behavior in the next major release of the standard

A careful programmer who wants to know about I/O errors may precede close() with a call to fsync(2).

Thus, best practice would seem to be:

  • Call fsync() at some point after writing data, but before close(); this should report any I/O errors
  • Only use close() to actually destroy the file descriptor when the file descriptor is no longer in use
  • On Linux, the file descriptor is always closed regardless of the return value (and should therefore not be retried); on other systems, EINTR may signal that the file descriptor was not closed