cmp: Check the status of stdout.

[des]

Summary

POSIX requires us to print an error message and exit non-zero if
writing to stdout fails. This can only happen if sflag is unset.

MFC after: 3 days
Sponsored by: Klara, Inc.

[markj]

There are two many separate exit points, so install an atexit handler before calling anything that may print to stdout.

But almost all of them are exiting with a non-zero status. Don't you just need to handle those that exit with status 0?

[kevans]

In D47020#1071814, @markj wrote:

There are two many separate exit points, so install an atexit handler before calling anything that may print to stdout.

But almost all of them are exiting with a non-zero status. Don't you just need to handle those that exit with status 0?

Strictly speaking, yes- other errors should take precedence over a flush error, even if we could have encountered the flush error first but didn't bother error checking all of the writes to stdout.

[des]

DIFF_EXIT is non-zero but not an error. Unless -s was specified, it is preceded by a message to stdout, and POSIX requires us to not only exit non-zero but also print an error message to stderr if writing to stdout fails. I suppose this would be clearer if I rewrote cmp to return non-error results back to main so there was only a single non-error exit point, and then try to fflush(stdout) iff !sflag, rather than use atexit().

[des]

avoid atexit

[des]

more thorough testing

[markj]

Thank you, I think this is nicer.

[markj]

usr.bin/cmp/tests/cmp_test2.sh

144

If I run this test case in a loop with some load in the background, e.g., from a build, there is an occasional failure because the result is 1 instead of 2.

I wonder if the test is racy because the subshell is able to write to the pipe buffer before the other end exits, triggering EPIPE?

usr.bin/cmp/cmp.c

 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 .\" SUCH DAMAGE.
 .\"
 .\"     @(#)mountd.8       8.4 (Berkeley) 4/28/95
 .\" $FreeBSD$
 .\"
-.Dd October 11, 2020
+.Dd March 18, 2023
 .Dt MOUNTD 8
 .Os
 .Sh NAME
 .Nm mountd
 .Nd service remote
 .Tn NFS
 mount requests
 .Sh SYNOPSIS

 option is used.
 .El
 .Pp
 When
 .Nm
 is started,
 it loads the export host addresses and options into the kernel
 using the
-.Xr mount 2
+.Xr nmount 2
 system call.
 After changing the exports file,
 a hangup signal should be sent to the
 .Nm
 daemon
 to get it to reload the export information.
 After sending the SIGHUP
 (kill \-s HUP `cat /var/run/mountd.pid`),
 check the syslog output to see if
 .Nm
 logged any parsing
 errors in the exports file.
 .Pp
+If multiple instances of
+.Nm
+are being run, either in multiple jails or both within
+and outside of a jail, care must be taken to export
+any given file system in only one of the instances.
+Note that the
+.Va allow.nfsd
+jail parameter is required to allow
+.Nm
+to run in a jail.
+See
+.Xr jail 8
+for more information.
+.Pp
 If
 .Nm
 detects that the running kernel does not include
 .Tn NFS
 support, it will attempt to load a loadable kernel module containing
 .Tn NFS
 code, using
 .Xr kldload 2 .
 If this fails, or no
 .Tn NFS
 KLD was available,
 .Nm
 exits with an error.
+When run in a jail, the
+.Xr kldload 2
+must be done outside the jail, typically by adding
+.Dq nfsd
+to
+.Va kld_list
+in the
+.Xr rc.conf 5
+file on the jail host.
 .Sh FILES
 .Bl -tag -width /var/run/mountd.pid -compact
 .It Pa /etc/exports
 the list of exported file systems
 .It Pa /var/run/mountd.pid
 the pid of the currently running mountd
 .It Pa /var/db/mountdtab
 the current list of remote mounted file systems
 .El
 .Sh SEE ALSO
 .Xr nfsstat 1 ,
 .Xr kldload 2 ,
 .Xr nfsv4 4 ,
 .Xr exports 5 ,
+.Xr rc.conf 5 ,
+.Xr jail 8 ,
 .Xr nfsd 8 ,
 .Xr rpcbind 8 ,
 .Xr showmount 8
 .Sh HISTORY
 The
 .Nm
 utility first appeared in
 .Bx 4.4 .

usr.bin/cmp/extern.h

 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 .\" SUCH DAMAGE.
 .\"
 .\"     @(#)mountd.8       8.4 (Berkeley) 4/28/95
 .\" $FreeBSD$
 .\"
-.Dd October 11, 2020
+.Dd March 18, 2023
 .Dt MOUNTD 8
 .Os
 .Sh NAME
 .Nm mountd
 .Nd service remote
 .Tn NFS
 mount requests
 .Sh SYNOPSIS

 option is used.
 .El
 .Pp
 When
 .Nm
 is started,
 it loads the export host addresses and options into the kernel
 using the
-.Xr mount 2
+.Xr nmount 2
 system call.
 After changing the exports file,
 a hangup signal should be sent to the
 .Nm
 daemon
 to get it to reload the export information.
 After sending the SIGHUP
 (kill \-s HUP `cat /var/run/mountd.pid`),
 check the syslog output to see if
 .Nm
 logged any parsing
 errors in the exports file.
 .Pp
+If multiple instances of
+.Nm
+are being run, either in multiple jails or both within
+and outside of a jail, care must be taken to export
+any given file system in only one of the instances.
+Note that the
+.Va allow.nfsd
+jail parameter is required to allow
+.Nm
+to run in a jail.
+See
+.Xr jail 8
+for more information.
+.Pp
 If
 .Nm
 detects that the running kernel does not include
 .Tn NFS
 support, it will attempt to load a loadable kernel module containing
 .Tn NFS
 code, using
 .Xr kldload 2 .
 If this fails, or no
 .Tn NFS
 KLD was available,
 .Nm
 exits with an error.
+When run in a jail, the
+.Xr kldload 2
+must be done outside the jail, typically by adding
+.Dq nfsd
+to
+.Va kld_list
+in the
+.Xr rc.conf 5
+file on the jail host.
 .Sh FILES
 .Bl -tag -width /var/run/mountd.pid -compact
 .It Pa /etc/exports
 the list of exported file systems
 .It Pa /var/run/mountd.pid
 the pid of the currently running mountd
 .It Pa /var/db/mountdtab
 the current list of remote mounted file systems
 .El
 .Sh SEE ALSO
 .Xr nfsstat 1 ,
 .Xr kldload 2 ,
 .Xr nfsv4 4 ,
 .Xr exports 5 ,
+.Xr rc.conf 5 ,
+.Xr jail 8 ,
 .Xr nfsd 8 ,
 .Xr rpcbind 8 ,
 .Xr showmount 8
 .Sh HISTORY
 The
 .Nm
 utility first appeared in
 .Bx 4.4 .

usr.bin/cmp/link.c

 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 .\" SUCH DAMAGE.
 .\"
 .\"     @(#)mountd.8       8.4 (Berkeley) 4/28/95
 .\" $FreeBSD$
 .\"
-.Dd October 11, 2020
+.Dd March 18, 2023
 .Dt MOUNTD 8
 .Os
 .Sh NAME
 .Nm mountd
 .Nd service remote
 .Tn NFS
 mount requests
 .Sh SYNOPSIS

 option is used.
 .El
 .Pp
 When
 .Nm
 is started,
 it loads the export host addresses and options into the kernel
 using the
-.Xr mount 2
+.Xr nmount 2
 system call.
 After changing the exports file,
 a hangup signal should be sent to the
 .Nm
 daemon
 to get it to reload the export information.
 After sending the SIGHUP
 (kill \-s HUP `cat /var/run/mountd.pid`),
 check the syslog output to see if
 .Nm
 logged any parsing
 errors in the exports file.
 .Pp
+If multiple instances of
+.Nm
+are being run, either in multiple jails or both within
+and outside of a jail, care must be taken to export
+any given file system in only one of the instances.
+Note that the
+.Va allow.nfsd
+jail parameter is required to allow
+.Nm
+to run in a jail.
+See
+.Xr jail 8
+for more information.
+.Pp
 If
 .Nm
 detects that the running kernel does not include
 .Tn NFS
 support, it will attempt to load a loadable kernel module containing
 .Tn NFS
 code, using
 .Xr kldload 2 .
 If this fails, or no
 .Tn NFS
 KLD was available,
 .Nm
 exits with an error.
+When run in a jail, the
+.Xr kldload 2
+must be done outside the jail, typically by adding
+.Dq nfsd
+to
+.Va kld_list
+in the
+.Xr rc.conf 5
+file on the jail host.
 .Sh FILES
 .Bl -tag -width /var/run/mountd.pid -compact
 .It Pa /etc/exports
 the list of exported file systems
 .It Pa /var/run/mountd.pid
 the pid of the currently running mountd
 .It Pa /var/db/mountdtab
 the current list of remote mounted file systems
 .El
 .Sh SEE ALSO
 .Xr nfsstat 1 ,
 .Xr kldload 2 ,
 .Xr nfsv4 4 ,
 .Xr exports 5 ,
+.Xr rc.conf 5 ,
+.Xr jail 8 ,
 .Xr nfsd 8 ,
 .Xr rpcbind 8 ,
 .Xr showmount 8
 .Sh HISTORY
 The
 .Nm
 utility first appeared in
 .Bx 4.4 .

usr.bin/cmp/misc.c

 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 .\" SUCH DAMAGE.
 .\"
 .\"     @(#)mountd.8       8.4 (Berkeley) 4/28/95
 .\" $FreeBSD$
 .\"
-.Dd October 11, 2020
+.Dd March 18, 2023
 .Dt MOUNTD 8
 .Os
 .Sh NAME
 .Nm mountd
 .Nd service remote
 .Tn NFS
 mount requests
 .Sh SYNOPSIS

 option is used.
 .El
 .Pp
 When
 .Nm
 is started,
 it loads the export host addresses and options into the kernel
 using the
-.Xr mount 2
+.Xr nmount 2
 system call.
 After changing the exports file,
 a hangup signal should be sent to the
 .Nm
 daemon
 to get it to reload the export information.
 After sending the SIGHUP
 (kill \-s HUP `cat /var/run/mountd.pid`),
 check the syslog output to see if
 .Nm
 logged any parsing
 errors in the exports file.
 .Pp
+If multiple instances of
+.Nm
+are being run, either in multiple jails or both within
+and outside of a jail, care must be taken to export
+any given file system in only one of the instances.
+Note that the
+.Va allow.nfsd
+jail parameter is required to allow
+.Nm
+to run in a jail.
+See
+.Xr jail 8
+for more information.
+.Pp
 If
 .Nm
 detects that the running kernel does not include
 .Tn NFS
 support, it will attempt to load a loadable kernel module containing
 .Tn NFS
 code, using
 .Xr kldload 2 .
 If this fails, or no
 .Tn NFS
 KLD was available,
 .Nm
 exits with an error.
+When run in a jail, the
+.Xr kldload 2
+must be done outside the jail, typically by adding
+.Dq nfsd
+to
+.Va kld_list
+in the
+.Xr rc.conf 5
+file on the jail host.
 .Sh FILES
 .Bl -tag -width /var/run/mountd.pid -compact
 .It Pa /etc/exports
 the list of exported file systems
 .It Pa /var/run/mountd.pid
 the pid of the currently running mountd
 .It Pa /var/db/mountdtab
 the current list of remote mounted file systems
 .El
 .Sh SEE ALSO
 .Xr nfsstat 1 ,
 .Xr kldload 2 ,
 .Xr nfsv4 4 ,
 .Xr exports 5 ,
+.Xr rc.conf 5 ,
+.Xr jail 8 ,
 .Xr nfsd 8 ,
 .Xr rpcbind 8 ,
 .Xr showmount 8
 .Sh HISTORY
 The
 .Nm
 utility first appeared in
 .Bx 4.4 .