Re: [PATCH security-next v4 23/32] selinux: Remove boot parameter

[John Johansen]

On 10/02/2018 01:29 PM, Kees Cook wrote:
> On Tue, Oct 2, 2018 at 12:47 PM, John Johansen
> <john.johansen@xxxxxxxxxxxxx> wrote:
>> On 10/02/2018 12:17 PM, Kees Cook wrote:
>>> I could define CONFIG_LSM_ENABLE as being "additive" to
>>> SECURITY_APPARMOR_BOOTPARAM_VALUE and
>>> SECURITY_SELINUX_BOOTPARAM_VALUE?
>>
>> Oh sure lets deal with my complaint about too many ways to configure
>> this beast by adding yet another config option :P
> 
> This is what v3 already does: SEC...BOOTPARAM_VALUE trumps ...LSM_ENABLE.
> 
sure but I sent in a patch to kill SECURITY_APPARMOR_BOOTPARAM_VALUE
because I really dislike the extra levels of config and getting rid
of the  SEC..BOOTPARAM_VALUE seems to be the easy way to fix it

Now if only we can convince Paul and Stephen :)

>> seriously though, please no. That just adds another layer of confusion
>> even if it is only being foisted on the distro/builder
> 
> You've already sent a patch removing
> SECURITY_APPARMOR_BOOTPARAM_VALUE. If SELinux is fine to do that too,
> then I think we'll be sorted out. I'll just need to make "lsm.enable="
> be an explicit list. (Do you have a problem with "lsm.disable=..." ?)
> 

why yes, glad you asked

If lsm.enabled is an explicit list lsm.disabled isn't required its a
convenience option that can introduce confusion and conflicts. If
both lsm.enabled and lsm.disabled are being used at the same time.

I realize that some times the convenience of specifying

  lsm.disable=$LSM

is easier than specifying an entire list of what should be enabled
when you just want to disable a single LSM.

I don't think the convenience is worth the potential confusion, but
I don't feel strongly about it and realize others feel the other
way.


tldr: I can live with it, but don't like it if you are asking :)