Masanobu Koike wrote:

quoted

On Friday, March 02, 2018 12:43 AM, Casey Schaufler wrote:

quoted

On 2/28/2018 11:38 PM, Masanobu Koike wrote:

quoted

@@ -264,6 +266,9 @@ choice
 	config DEFAULT_SECURITY_APPARMOR
 		bool "AppArmor" if SECURITY_APPARMOR=y

+	config DEFAULT_SECURITY_WHITEEGRET
+		bool "WhiteEgret" if SECURITY_WHITEEGRET=y
+

I don't see this module using any security blobs. Is there
a reason you're not making this a minor (like yama) module
instead of a major (like AppArmor) module?

Thank you for your suggestion.
We are now developing WhiteEgret on the environment
it works certainly.

??? What Casey suggested is effectively

----------

--- a/security/whiteegret/init.c
+++ b/security/whiteegret/init.c

@@ -48,9 +48,6 @@ static int __init we_init(void)
 {
 	int rc;

-	if (!security_module_enable("whiteegret"))
-		return 0;
-
 	security_add_hooks(we_hooks, ARRAY_SIZE(we_hooks),

"whiteegret");

 	rc = we_specific_init();
----------

, isn't it? Unlike Yama, adding whiteegret_add_hooks() to security_init()
is not useful, for security_init() is called too early to create securityfs
entries for WhiteEgret.

Current version uses security= parameter as a switch for enabling/disabling
WhiteEgret, doesn't it? If WhiteEgret does not use security= as a switch,
is some other switch (e.g. __setup()) expected?