5:53 a.m.
Subject: Re: de-modularising for the win!
Message-ID: <20080920105355.GA6663(a)nostromo.devel.redhat.com>
References: <20080918191355.GA9683(a)nostromo.devel.redhat.com>
<48D2D94E.2020902(a)redhat.com>
<20080918231301.GA4734(a)nostromo.devel.redhat.com>
<1221790735.15361.152.camel(a)aglarond.local>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1221790735.15361.152.camel(a)aglarond.local>
User-Agent: Mutt/1.5.18 (2008-05-17)
Jeremy Katz (katzj(a)redhat.com) said:
On Thu, 2008-09-18 at 16:13 -0700, Bill Nottingham wrote:
> - killing the initrd for that general 90% case can be a big win
Ermm, the general 90% (or some large-ish generalizing percentage) are
set up to use LVM. Which then requires an initrd.
Yes, but ... LVM is overkill, in general. (That's another discussion.)
On further consideration, though, the biggest issue with kicking out
the initrd is getting the policy lodaed.
Bill
12:13 p.m.
On Sat, Sep 20, 2008 at 06:53:55AM -0400, Bill Nottingham wrote:
Subject: Re: de-modularising for the win!
Message-ID: <20080920105355.GA6663(a)nostromo.devel.redhat.com>
References: <20080918191355.GA9683(a)nostromo.devel.redhat.com>
<48D2D94E.2020902(a)redhat.com>
<20080918231301.GA4734(a)nostromo.devel.redhat.com>
<1221790735.15361.152.camel(a)aglarond.local>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1221790735.15361.152.camel(a)aglarond.local>
User-Agent: Mutt/1.5.18 (2008-05-17)
Jeremy Katz (katzj(a)redhat.com) said:
> On Thu, 2008-09-18 at 16:13 -0700, Bill Nottingham wrote:
> > - killing the initrd for that general 90% case can be a big win
>
> Ermm, the general 90% (or some large-ish generalizing percentage) are
> set up to use LVM. Which then requires an initrd.
Yes, but ... LVM is overkill, in general. (That's another discussion.)
On further consideration, though, the biggest issue with kicking out
the initrd is getting the policy lodaed.
But, won't we all be using btrfs instead of lvm in F11? ;-)
12:22 p.m.
On Sat, 20 Sep 2008 13:13:38 -0400
Kyle McMartin <kyle(a)redhat.com> wrote:
> On further consideration, though, the biggest issue with kicking
out
> the initrd is getting the policy lodaed.
/me wonders about the option of having selinux using
request_firmware() to get its policy
--
Arjan van de Ven Intel Open Source Technology Centre
For development, discussion and tips for power savings,
visit http://www.lesswatts.org
10:01 p.m.
Kyle McMartin wrote:
On Sat, Sep 20, 2008 at 06:53:55AM -0400, Bill Nottingham wrote:
> Subject: Re: de-modularising for the win!
> Message-ID: <20080920105355.GA6663(a)nostromo.devel.redhat.com>
> References: <20080918191355.GA9683(a)nostromo.devel.redhat.com>
<48D2D94E.2020902(a)redhat.com>
<20080918231301.GA4734(a)nostromo.devel.redhat.com>
<1221790735.15361.152.camel(a)aglarond.local>
> MIME-Version: 1.0
> Content-Type: text/plain; charset=us-ascii
> Content-Disposition: inline
> In-Reply-To: <1221790735.15361.152.camel(a)aglarond.local>
> User-Agent: Mutt/1.5.18 (2008-05-17)
>
> Jeremy Katz (katzj(a)redhat.com) said:
>> On Thu, 2008-09-18 at 16:13 -0700, Bill Nottingham wrote:
>>> - killing the initrd for that general 90% case can be a big win
>> Ermm, the general 90% (or some large-ish generalizing percentage) are
>> set up to use LVM. Which then requires an initrd.
> Yes, but ... LVM is overkill, in general. (That's another discussion.)
>
> On further consideration, though, the biggest issue with kicking out
> the initrd is getting the policy lodaed.
>
But, won't we all be using btrfs instead of lvm in F11? ;-)
Aw crap, you mean F10 was my only fleeting chance to make ext4 the
default? :)
-Eric
10:19 a.m.
On Saturday 20 September 2008 1:22:31 pm Arjan van de Ven wrote:
On Sat, 20 Sep 2008 13:13:38 -0400
Kyle McMartin <kyle(a)redhat.com> wrote:
> > On further consideration, though, the biggest issue with kicking
> > out the initrd is getting the policy lodaed.
/me wonders about the option of having selinux using
request_firmware() to get its policy
I've started ignoring most of the firmware loading threads a while ago
so I've kinda lost most of the plot there ... a few questions:
* Is the firmware request asynchronous?
* Is there currently a way, or at least nothing in the way that would
prevent a mechanism from bring created, to pause the boot process until
the policy/firmware is loaded? I ask because we would want to make
sure that SELinux policy was loaded before any services are started.
Regardless, we would still need to keep the current policy loading
mechanism in place since we can't break userspace. You should post
this idea on the SELinux list to get some further thought on this ...
--
paul moore
linux @ hp
5688
days inactive
5690
days old
kernel@lists.fedoraproject.org
4 comments
5 participants
participants (5)
-
Arjan van de Ven -
Bill Nottingham -
Eric Sandeen -
Kyle McMartin -
Paul Moore