Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 26 Jan 2006 08:14:06 -0700
From:      Scott Long <scottl@samsco.org>
To:        Ceri Davies <ceri@submonkey.net>
Cc:        Murray Stokely <murray@FreeBSD.org>, cvs-doc@FreeBSD.org, cvs-all@FreeBSD.org, doc-committers@FreeBSD.org
Subject:   Re: cvs commit: www/en/releases/6.1R todo.sgml
Message-ID:  <43D8E73E.4050103@samsco.org>
In-Reply-To: <20060126101706.GJ36965@submonkey.net>
References:  <200601260957.k0Q9vCUn054132@repoman.freebsd.org> <20060126101706.GJ36965@submonkey.net>

next in thread | previous in thread | raw e-mail | index | archive | help
Ceri Davies wrote:
> On Thu, Jan 26, 2006 at 09:57:12AM +0000, Murray Stokely wrote:
> 
>>murray      2006-01-26 09:57:12 UTC
>>
>>  FreeBSD doc repository
>>
>>  Modified files:
>>    en/releases/6.1R     todo.sgml 
>>  Log:
>>  Add kbdmux and sysinstall smp kernel install items from the ideas page
>>  to the 6.1 Desired Features list.
> 
> 
> I think it's a little late to mess with sysinstall to that extent
> for 6.1.  Sounds like the kind of thing that could sit in -CURRENT for
> months, but hardly anyone would actually be using it.  It seems that the
> main problem with sysinstall is that hardly any of our developers use
> it.
> 
> On to the question: how often does an SMP kernel fail to boot where a
> UP one might work?  I remember that this used to be a problem, but if
> it's still "too often", can we have just the bits that probe for an
> mptable (or however we determine that there is more that one processor)
> in the UP kernel without suffering that instability?
> 
> What I'm basically asking is how much of the SMP code is really required
> just to detect MP hardware?
> 
> Ceri

As in my other post, I don't think it will ever be safe to make i386
default to an SMP kernel (not counting the performance arguments which
are also relevant).

Anyways, as far as this TODO task goes, the first step is to simply
update the release makefiles to generate both a UP and SMP kernel and
have them both get installed, one as /boot/kernel and the other as
/boot/kernel.SMP.  That overcomes the first hurdle for lowering the SMP
bar for users, as then it would be quite easy to add a bootloader menu
item to select one kernel or the other.

The next step would be add an mptable parser to sysinstall (not as hard
as it sounds, probably about 100 lines of code at most) and have
sysinstall either automatically rename /boot/kernel.SMP to /boot/kernel
as needed, or give the user the choice on renaming.  The only problem
here is that most MPTables don't recognize hyperthreading.  So either we
live with that, or we take extra steps to parse ACPI tables.  Since we
do have ACPI on by default, this also isn't as hard as it sounds.

I don't want SMP enabled by default on i386, not even for the
installation kernel.  It will cause a regression in the amount of
systems that we support, and it will impact performance.  Neither of
these are desirable.  Making detection and installation more automated
is a very good thing, though, and that's why this is a TODO item.  There
is still time to work on this before the 6.1 release, and we can always
back it out if it doesn't work out at first.  So yes, I encourage people
to work on this.

Scott



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?43D8E73E.4050103>