Q. DEBUG_FS and SYSFS config (Re: [RFC Aufs2 #3 2/2] split 'xino' entry under sysfs)

[hooanon05@yahoo.co.jp]

Greg KH:
> No one disables sysfs that I know of.  Heck, my phone enables sysfs...
> 
> > Addition to set limit, there is one more reason to adopt seq_file.
> > Because the printed string is a path, it may contain unprintable
> > characters. seq_file has a good interface seq_path() which supports
> > escaping such characters.
> 
> That is true.  Ok, let's see the final result and I'll be glad to look
> at it.

Hi Greg,

Currently I am testing locally about new aufs entries under debugfs, and
I found a strange issue about configuration.

First, configure "normaly".
General setup
  [ ] Configure standard kernel features (for small systems)  --->

$ egrep -i '(sysfs|debug_fs)' .config
CONFIG_SYSFS_DEPRECATED=y
CONFIG_SYSFS_DEPRECATED_V2=y
CONFIG_SYSFS=y
CONFIG_DEBUG_FS=y
# CONFIG_FAULT_INJECTION_DEBUG_FS is not set

Next, enable EMBEDED, and then disable SYSFS.
General setup
  [*] Configure standard kernel features (for small systems)  --->
File systems
  Pseudo filesystems  --->
    [ ] sysfs file system support

$ egrep -i '(sysfs|debug_fs)' .config
# CONFIG_SYSFS is not set
CONFIG_DEBUG_FS=y

SYSFS is disabled expectedly, but DEBUG_FS is still enabled,
while "depends on SYSFS" is specified for DEBUG_FS.

Is this expected behaviour?
In other words, can I assume "when DEBUG_FS is enabled, SYSFS must be
enabled too" safely?


J. R. Okajima