Skip to contentSkip to conversation

Kernel Patch to support nested bind mounts

DISCUSSION Lukas Kolbe 2 participants 3 messages May 20, 2008 Kernel automounter

lore
Share & tools
lore
#0 lore raw

Hi Ian, list,

this is just a ping to remind you of this little patch. It still applies to the latest kernel source, and it still fixes our problems we have with nested mount maps such as (taken from the map for /vol/):

eclipse-3.2 \

  /bin          -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/bin/linx86 \
  /include      -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/include \
  /lib          -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/lib/linx86 \
  /man          -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/man \
  /share        -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/share

eclipse \

  /bin          -rw,nosuid          :/vol/eclipse-3.2/bin \
  /include      -rw,nosuid          :/vol/eclipse-3.2/include \
  /lib          -rw,nosuid          :/vol/eclipse-3.2/lib \
  /man          -rw,nosuid          :/vol/eclipse-3.2/man \
  /share        -rw,nosuid          :/vol/eclipse-3.2/share \

and:

jdk             -rw,nosuid          :/vol/jdk-1.6
jdk-1.6         -rw,nosuid          :/vol/jdk-1.6.6
jdk-1.6.6       -rw,nosuid          fileserver:/export/stud/vol/jdk-1.6.6/linx86

We're running this since quite a while, and have not encountered any serious problems since then. Home's and volumes get unmounted as they should, and (even more important) accessing /vol/jdk or /vol/eclipse/bin now actually works (it did _not_ without the patch).

I'm just tired of constantly rebuilding distribution kerneles :)

Signature · 35 lines
Lukas


patch:

diff -ru linux-2.6-2.6.23-a/fs/autofs4/root.c linux-2.6-2.6.23/fs/autofs4/root.c
--- linux-2.6-2.6.23-a/fs/autofs4/root.c        2007-11-30 16:20:33.000000000 +0100
+++ linux-2.6-2.6.23/fs/autofs4/root.c  2007-11-30 16:22:24.000000000 +0100
@@ -19,6 +19,9 @@
 #include <linux/time.h>
 #include "autofs_i.h"

+#define DIRECT_TRIGGER_FLAGS (LOOKUP_CONTINUE|LOOKUP_DIRECTORY|LOOKUP_ACCESS)
+
+
 static int autofs4_dir_symlink(struct inode *,struct dentry *,const char *);
 static int autofs4_dir_unlink(struct inode *,struct dentry *);
 static int autofs4_dir_rmdir(struct inode *,struct dentry *);
@@ -291,7 +294,7 @@
                        return status;
                }
        /* Trigger mount for path component or follow link */
-       } else if (flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY) ||
+       } else if (flags & DIRECT_TRIGGER_FLAGS ||
                        current->link_count) {
                DPRINTK("waiting for mount name=%.*s",
                        dentry->d_name.len, dentry->d_name.name);
@@ -335,7 +338,7 @@
                nd->flags);

        /* If it's our master or we shouldn't trigger a mount we're done */
-       lookup_type = nd->flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY);
+       lookup_type = nd->flags & DIRECT_TRIGGER_FLAGS;
        if (oz_mode || !lookup_type)
                goto done;
maintainer
#1 lore raw
Quoted 28 lines · Lukas Kolbe

Hi Ian, list,

this is just a ping to remind you of this little patch. It still applies to the latest kernel source, and it still fixes our problems we have with nested mount maps such as (taken from the map for /vol/):

eclipse-3.2 \

  /bin          -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/bin/linx86 \
  /include      -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/include \
  /lib          -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/lib/linx86 \
  /man          -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/man \
  /share        -rw,nosuid          fileserver:/export/stud/vol/eclipse-3.2/share

eclipse \

  /bin          -rw,nosuid          :/vol/eclipse-3.2/bin \
  /include      -rw,nosuid          :/vol/eclipse-3.2/include \
  /lib          -rw,nosuid          :/vol/eclipse-3.2/lib \
  /man          -rw,nosuid          :/vol/eclipse-3.2/man \
  /share        -rw,nosuid          :/vol/eclipse-3.2/share \

and:

jdk             -rw,nosuid          :/vol/jdk-1.6
jdk-1.6         -rw,nosuid          :/vol/jdk-1.6.6
jdk-1.6.6       -rw,nosuid          fileserver:/export/stud/vol/jdk-1.6.6/linx86
Quoted 6 lines

We're running this since quite a while, and have not encountered any serious problems since then. Home's and volumes get unmounted as they should, and (even more important) accessing /vol/jdk or /vol/eclipse/bin now actually works (it did _not_ without the patch).

I'm just tired of constantly rebuilding distribution kerneles :)

I thought the email replies to the kernel submission I made also went to the autofs list. I didn't notice if that wasn't the case.

Anyway, Al Viro vetoed the patch with little justification except to say that there are VFS changes coming that the patch won't fit well with, but also didn't elaborate those changes when asked.

So all we can do is wait and see.

Ian

#2 lore raw

Hi Ian,

Quoted 10 lines
Quoted 1 line

I'm just tired of constantly rebuilding distribution kerneles :)

I thought the email replies to the kernel submission I made also went to the autofs list. I didn't notice if that wasn't the case.

Anyway, Al Viro vetoed the patch with little justification except to say that there are VFS changes coming that the patch won't fit well with, but also didn't elaborate those changes when asked.

So all we can do is wait and see.

Thanks, didn't notice that. I'll look in the archives and recompile everything for the next security fix then :)

Quoted 1 line

Ian

Signature · 2 lines
kind regards,
Lukas