Skip to contentSkip to conversation

fix multi maps that have an implied '/' entry

PATCH In review Jeff Moyer 3 participants 4 messages Apr 29, 2005 – May 1, 2005 Kernel automounter

lore
Share & tools
lore

Jeff Moyer

author
#0 lore raw

PATCH 1/1 fix multi maps that have an implied '/' entry

This is apparently a legal format for a multi-mount map:

g3b     segfault.boston.redhat.com:/export/automount/export5/testing/test \
        /s1     segfault.boston.redhat.com:/export/automount/export2 \
        /s2     segfault.boston.redhat.com:/export/automount/export3

The first entry is an implied root level entry, or '/'. Solaris handles this case just fine, and the automounter test sweet uses this as one of its tests. The attached patch causes this map format to be recognized. I think the patch could probably re-use the memory allocatd for "path" by the dequote function. That's up to you, Ian, if you want to change that to be a little prettier.

-Jeff

modules/parse_sun.c renamed from modules/parse_sun.c.orig +12 -1
@@ -886,7 +886,18 @@ int parse_mount(const char *root, const 
 				return 1;
 			}
 
-			p += l;
+			if (*path != '/') {
+				free(path);
+				path = strdup("/");
+				if (!path) {
+					error(MODPREFIX "out of memory");
+					free(myoptions);
+					free(options);
+					multi_free_list(head);
+					return 1;
+				}
+			} else
+				p += l;
 			p = skipspace(p);
 
 			/* Local options are appended to per-map options */
#1 lore raw
Quoted 6 lines · Jeff Moyer

This is apparently a legal format for a multi-mount map:

g3b     segfault.boston.redhat.com:/export/automount/export5/testing/test \
       /s1     segfault.boston.redhat.com:/export/automount/export2 \
       /s2     segfault.boston.redhat.com:/export/automount/export3
Quoted 6 lines

The first entry is an implied root level entry, or '/'. Solaris handles this case just fine, and the automounter test sweet uses this as one of its tests. The attached patch causes this map format to be recognized. I think the patch could probably re-use the memory allocatd for "path" by the dequote function. That's up to you, Ian, if you want to change that to be a little prettier.

I've already placed a patch on kernel.org for that.
Check out autofs-4.1.4-multi-parse-fix.patch and let me know what you think.

It also addresses the multi-mount recognition problem you mentioned a short while ago.

Ian

#2 lore raw

Patches for 4.1.4 (was: fix multi maps that have an implied '/' entry)

Quoted 3 lines · On Sat, Apr 30, 2005 at 11:12:56PM +0800, [email protected] wrote:

I've already placed a patch on kernel.org for that.
Check out autofs-4.1.4-multi-parse-fix.patch and let me know what you think.

I'm rapidly losing overview over what's going in and not here, with all the traffic :-) I'm planning an upload of autofs 4.1.4 to Debian unstable RSN; what patches do I want with this? Vanilla 4.1.4 + multi-parse-fix + non-replicated ping?

BTW: Ian, have you looked at
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=305872 ? I'm not really sure what the & is supposed to mean :-)

/* Steinar */

Signature · 1 line
Homepage: http://www.sesse.net/
#3 lore raw

Re: Patches for 4.1.4 (was: fix multi maps that have an implied '/' entry)

Quoted 9 lines · Steinar H. Gunderson
Quoted 3 lines · On Sat, Apr 30, 2005 at 11:12:56PM +0800, [email protected] wrote:

I've already placed a patch on kernel.org for that.
Check out autofs-4.1.4-multi-parse-fix.patch and let me know what you think.

I'm rapidly losing overview over what's going in and not here, with all the traffic :-) I'm planning an upload of autofs 4.1.4 to Debian unstable RSN; what patches do I want with this? Vanilla 4.1.4 + multi-parse-fix + non-replicated ping?

Sorry, I don't mean to be confusing.

This is the way that I do things normally.
Any smallish bug fixes that I think are really needed I post on kernel.org with the tarball.

I would recommend you include both of those patches.

Quoted 4 lines

BTW: Ian, have you looked at
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=305872 ? I'm not really sure what the & is supposed to mean :-)

Interesting but not nearly enough information to resolve the issue. The problem that is reported shouldn't be happening.

Debug logs needed.

The & is a placeholder for the key and the * is the wildcard key.

In

* &:/var/local/bulk

if this map is mounted on /amnt then

ls /amnt/somehost

resolves to somehost:/var/local.bulk

Ian