[2.6 patch] CONFIGFS_FS must depend on SYSFS

[2.6 patch] CONFIGFS_FS must depend on SYSFS

[Adrian Bunk]

This patch fixes the following compile error with CONFIG_SYSFS=n:

<--  snip  -->

...
  LD      .tmp_vmlinux1
fs/built-in.o: In function `configfs_init':mount.c:(.init.text+0x3d5d): undefined reference to `kernel_subsys'
make: *** [.tmp_vmlinux1] Error 1

<--  snip  -->

Signed-off-by: Adrian Bunk <bunk@stusta.de>

---

 fs/Kconfig |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

--- linux-2.6.16-mm1-full/fs/Kconfig.old	2006-03-26 03:02:17.000000000 +0200
+++ linux-2.6.16-mm1-full/fs/Kconfig	2006-03-26 03:02:34.000000000 +0200
@@ -328,7 +328,7 @@
 
 config OCFS2_FS
 	tristate "OCFS2 file system support (EXPERIMENTAL)"
-	depends on NET && EXPERIMENTAL
+	depends on NET && SYSFS && EXPERIMENTAL
 	select CONFIGFS_FS
 	select JBD
 	select CRC32
@@ -863,7 +863,7 @@
 
 config CONFIGFS_FS
 	tristate "Userspace-driven configuration filesystem (EXPERIMENTAL)"
-	depends on EXPERIMENTAL
+	depends on SYSFS && EXPERIMENTAL
 	help
 	  configfs is a ram-based filesystem that provides the converse
 	  of sysfs's functionality. Where sysfs is a filesystem-based

Re: [2.6 patch] CONFIGFS_FS must depend on SYSFS

[Joel Becker]

On Sun, Mar 26, 2006 at 02:25:52PM +0200, Adrian Bunk wrote:
> This patch fixes the following compile error with CONFIG_SYSFS=n:
> 
> <--  snip  -->
> 
> ...
>   LD      .tmp_vmlinux1
> fs/built-in.o: In function `configfs_init':mount.c:(.init.text+0x3d5d): undefined reference to `kernel_subsys'
> make: *** [.tmp_vmlinux1] Error 1

	Hmm.  We only rely on SYSFS because of the policy decree of
default mounting at /sys/kernel/config (just like GregKH was asked to
put debugfs at /sys/kernel/debug).  While we create the mount point, we
don't rely on it.  You can mount configfs anywhere you like.
	So, what do you think of making that part of the code check for
CONFIG_SYSFS?  Then the module can be built without SYSFS instead of
depending on it.  I'm open to either idea.  Certainly, we don't leave
the compile error.

Joel

-- 

Life's Little Instruction Book #313

	"Never underestimate the power of love."

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker@oracle.com
Phone: (650) 506-8127

Re: [2.6 patch] CONFIGFS_FS must depend on SYSFS

[Adrian Bunk]

On Sun, Mar 26, 2006 at 07:06:22PM -0800, Joel Becker wrote:
> On Sun, Mar 26, 2006 at 02:25:52PM +0200, Adrian Bunk wrote:
> > This patch fixes the following compile error with CONFIG_SYSFS=n:
> > 
> > <--  snip  -->
> > 
> > ...
> >   LD      .tmp_vmlinux1
> > fs/built-in.o: In function `configfs_init':mount.c:(.init.text+0x3d5d): undefined reference to `kernel_subsys'
> > make: *** [.tmp_vmlinux1] Error 1
> 
> 	Hmm.  We only rely on SYSFS because of the policy decree of
> default mounting at /sys/kernel/config (just like GregKH was asked to
> put debugfs at /sys/kernel/debug).  While we create the mount point, we
> don't rely on it.  You can mount configfs anywhere you like.
> 	So, what do you think of making that part of the code check for
> CONFIG_SYSFS?  Then the module can be built without SYSFS instead of
> depending on it.  I'm open to either idea.  Certainly, we don't leave
> the compile error.

First of all sorry for the late answer.

My intention is to get the compile error fixed, and I don't care that 
much how it gets fixed.

OTOH, there's the question whether it matters at all - is the 
intersection of the people who are in such space-limited environments 
that they are setting CONFIG_EMBEDDED=y and then CONFIG_SYSFS=n and
the people using OCFS2 actually non-empty?

> Joel

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: [2.6 patch] CONFIGFS_FS must depend on SYSFS

[Joel Becker]

On Mon, Apr 10, 2006 at 08:35:22PM +0200, Adrian Bunk wrote:
> OTOH, there's the question whether it matters at all - is the 
> intersection of the people who are in such space-limited environments 
> that they are setting CONFIG_EMBEDDED=y and then CONFIG_SYSFS=n and
> the people using OCFS2 actually non-empty?

	We decided that (configfs + !sysfs == 0), and we just went with
the direct dependancy.  It's in my tree, going to Linus whenever it gets
there.

Joel

-- 

"I don't even butter my bread; I consider that cooking."
         - Katherine Cebrian

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker@oracle.com
Phone: (650) 506-8127