On Mon, 2008-06-16 at 12:56 -0700, Stephen Biggs wrote:
> Hi…
>
> As part of my compilation and execution efforts, I have accumulated a
> small set of minor fixes to some of the automounter daemon files.
> Presented for your approval. These apply to the git HEAD and the
> previous set of patches that Ian had provided to me. The biggest fix
> is the one for redhat'ish systems in rc.autofs.in that fixes the
> daemon "stop" function.
>
> I am CC'ing this directly to Ian just in case there is a problem with
> posting to the autofs list. Sorry for any noise.
>
> ----
>
> daemon/automount.c: Remove compile warning of unused variable
> Makefile: Remove "modification time in the future" messages generated
> by parallelism in some versions of make
> modules/Makefile: Avoid creating soft link to non-existent library if
> LDAP not specified in configure
> samples/rc.autofs.in: Specify full path for invocation of daemon,
> making pkill function more exact,
> Reverse sense of test of return of pidof to
> correctly kill daemon on stop function.
>
> ---
>
>
> Makefile | 2 ++
> daemon/automount.c | 1 -
> modules/Makefile | 2 ++
> samples/rc.autofs.in | 4 ++--
> 4 files changed, 6 insertions(+), 3 deletions(-)
>
>
> diff -urd autofs-5.0.3-git-ian_kent_patches-vanilla/daemon/automount.c
> autofs-5.0.3-git-ian_kent_patches/daemon/automount.c
>
> --- autofs-5.0.3-git-ian_kent_patches-vanilla/daemon/automount.c
> 2008-05-19 17:02:17.083219000 -0700
> +++ autofs-5.0.3-git-ian_kent_patches/daemon/automount.c
> 2008-05-19 13:55:10.608330000 -0700
> @@ -1554,7 +1554,6 @@
>
> while (ap->state != ST_SHUTDOWN) {
> if (handle_packet(ap)) {
> - struct ioctl_ops *ops = get_ioctl_ops();
> int ret, cur_state;
>
> /*
Yes, already spotted this, it's already done.
> diff -urd autofs-5.0.3-git-ian_kent_patches-vanilla/Makefile
> autofs-5.0.3-git-ian_kent_patches/Makefile
> --- autofs-5.0.3-git-ian_kent_patches-vanilla/Makefile 2008-05-19
> 11:59:15.333785000 -0700
> +++ autofs-5.0.3-git-ian_kent_patches/Makefile 2008-05-19
> 18:28:33.916067000 -0700
> @@ -39,6 +39,8 @@
> sed -e "s/(\.autofs-[0-9.]\+)/(.autofs-`cat .version`)/" <
> configure.in > configure.in.tmp
> mv -f configure.in.tmp configure.in
> rm -f configure
> +# get rid of "modification time in the future" messages
> + @sleep 1
> $(MAKE) configure
How does this help if you're clock is more than one second out?
Why is the time on you're NFS machines not synchronized?
>
> TODAY := $(shell date +'%Y%m%d')
> diff -urd autofs-5.0.3-git-ian_kent_patches-vanilla/modules/Makefile
> autofs-5.0.3-git-ian_kent_patches/modules/Makefile
> --- autofs-5.0.3-git-ian_kent_patches-vanilla/modules/Makefile
> 2008-05-19 11:59:15.528977000 -0700
> +++ autofs-5.0.3-git-ian_kent_patches/modules/Makefile 2008-05-19
> 13:52:40.466366000 -0700
> @@ -63,7 +63,9 @@
> -rm -f $(INSTALLROOT)$(autofslibdir)/mount_smbfs.so
Why?
Why not add a test for existence instead of just removing it?
> ln -fs lookup_file.so
> $(INSTALLROOT)$(autofslibdir)/lookup_files.so
> ln -fs lookup_yp.so
> $(INSTALLROOT)$(autofslibdir)/lookup_nis.so
> +ifeq ($(LDAP), 1)
> ln -fs lookup_ldap.so
> $(INSTALLROOT)$(autofslibdir)/lookup_ldaps.so
> +endif
Yeah, maybe?
> ln -fs mount_nfs.so
> $(INSTALLROOT)$(autofslibdir)/mount_nfs4.so
> ifeq ($(EXT2FS), 1)
> ifeq ($(EXT3FS), 1)
> diff -urd
> autofs-5.0.3-git-ian_kent_patches-vanilla/samples/rc.autofs.in
> autofs-5.0.3-git-ian_kent_patches/samples/rc.autofs.in
>
> --- autofs-5.0.3-git-ian_kent_patches-vanilla/samples/rc.autofs.in
> 2008-05-19 17:02:17.071234000 -0700
> +++ autofs-5.0.3-git-ian_kent_patches/samples/rc.autofs.in
> 2008-05-19 18:15:39.881408000 -0700
> @@ -56,7 +56,7 @@
> fi
> fi
>
> - $prog $OPTIONS
> + $DAEMON $OPTIONS
> RETVAL=$?
> if [ $RETVAL -eq 0 ] ; then
> echo "done."
> @@ -75,7 +75,7 @@
> [ $RETVAL = 0 -a -z "`pidof $DAEMON`" ] || sleep 3
> count=`expr $count + 1`
> done
> - if [ -n "`pidof $DAEMON`" ] ; then
> + if [ -z "`pidof $DAEMON`" ] ; then
> echo "done."
> else
> echo "failed."
>
Both these have come up before.
In fact I think I changed the if test, so now do we change it back?
I'm not sure why I didn't change the daemon path before but maybe there
was a reason for it. I'll need to think about that.
Perhaps someone else can remember.
________________________________________________________________________
Ian
_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs
