LDP
/
man-pages
mirror of https://github.com/mkerrisk/man-pages
0
1
Fork 0
Code Releases Activity

close.2: Further clarify that close() should not be retried after an error 

See Linus ancient comments re EINTR in
    https://lkml.org/lkml/headers/2005/9/10/129
    Date	Sat, 10 Sep 2005 12:00:01 -0700 (PDT)
    From	Linus Torvalds <>
    Subject	Re: [patch 7/7] uml: retry host close() on EINTR

The FreeBSD 11.0 close() man page says similar:

    In case of any error except EBADF, the supplied file
    descriptor is deallocated and therefore is no longer valid.

For AIX:
http://publib16.boulder.ibm.com/doc_link/en_US/a_doc_lib/libs/basetrf1/close.htm

    If the FileDescriptor parameter refers to a device and the
    close subroutine actually results in a device close, and the
    device close routine returns an error, the error is returned
    to the application. However, the FileDescriptor parameter is
    considered closed and it may not be used in any subsequent
    calls.

See also:
http://austingroupbugs.net/view.php?id=529
and in particular:
http://austingroupbugs.net/view.php?id=529#c1200

Reported-by: Daniel Wagner <wagi@monom.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
This commit is contained in:
Michael Kerrisk 2016-12-05 14:23:20 +01:00
parent 2f3db2a58f
commit 2904e19a2b
1 changed files with 45 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
man2/close.2
View File

@ -81,6 +81,10 @@ call was interrupted by a signal; see
.TP
.B EIO
An I/O error occurred.
.PP
See NOTES for a discussion of why
.BR close ()
should not be retried after an error.
.SH CONFORMING TO
POSIX.1-2001, POSIX.1-2008, SVr4, 4.3BSD.
.\" SVr4 documents an additional ENOLINK error condition.
@ -98,12 +102,51 @@ Not checking the return value when closing the file may lead to
silent loss of data.
This can especially be observed with NFS
and with disk quota.
Note that the return value should be used only for diagnostics.
Note, however, that the return value should be used only for diagnostics.
.\" The FreeBSD 11.0 close(2) man page says similar:
.\" In case of any error except EBADF, the supplied file descriptor is
.\" deallocated and therefore is no longer valid.
Retrying the
.BR close ()
after an error is the wrong thing to do,
.\" The file descriptor is released early in close();
.\" close() ==> __close_fd():
.\" __put_unused_fd() ==> __clear_open_fd()
.\" return filp_close(file, files);
.\"
.\" The errors are returned bt filp_close() after the FD has been
.\" cleared for re-use.
since this may cause a reused file descriptor
from another thread to be closed.
(The kernel releases the file descriptor early in the close
operation, freeing it for reuse;
the steps that may return an error,
.\" filp_close()
such as flushing data to disk, occur only later in the close operation.)
In particular,
.BR close ()
should not be retried after an
.B EINTR
since this may cause a reused file descriptor from another thread to be closed.
error.
(POSIX.1 says:
"If
.BR close ()
is interrupted by a signal that is to be caught, it shall return \-1 with
.I errno
set to
.B EINTR
and the state of fildes is
.IR unspecified .")
.\" FIXME . for later review when Issue 8 is one day released...
.\" POSIX proposes further changes for EINTR
.\" http://austingroupbugs.net/tag_view_page.php?tag_id=8
.\" http://austingroupbugs.net/view.php?id=529
Careful users who want to know about I/O errors may precede
.BR close ()
with a call to
.BR fsync (2).
.PP
A successful close does not guarantee that the data has been successfully
saved to disk, as the kernel uses the buffer cache to defer writes.