linux-kernel - Re: [RFC 8/8] x86_64: Support for new UV apic

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20080330210356.GA13383@sgi.com>
Date:	Sun, 30 Mar 2008 16:03:56 -0500
From:	Jack Steiner <steiner@....com>
To:	Yinghai Lu <yhlu.kernel@...il.com>
Cc:	Ingo Molnar <mingo@...e.hu>, Andi Kleen <andi@...stfloor.org>,
	tglx@...utronix.de, linux-mm@...ck.org,
	linux-kernel@...r.kernel.org
Subject: Re: [RFC 8/8] x86_64: Support for new UV apic

On Sun, Mar 30, 2008 at 01:23:12PM -0700, Yinghai Lu wrote:
> On Wed, Mar 26, 2008 at 12:38 AM, Ingo Molnar <mingo@...e.hu> wrote:
> >
> >  * Jack Steiner <steiner@....com> wrote:
> >
> >  > > > -        obj-y                            += genapic_64.o genapic_flat_64.o
> >  > > > +        obj-y                            += genapic_64.o genapic_flat_64.o genx2apic_uv_x.o
> >  > >
> >  > > Definitely should be a CONFIG
> >  >
> >  > Not sure that I understand why. The overhead of UV is minimal & we
> >  > want UV enabled in all distro kernels. OTOH, small embedded systems
> >  > probably want to eliminate every last bit of unneeded code.
> >  >
> >  > Might make sense to have a config option. Thoughts????
> >
> >  i wouldnt mind having UV enabled by default (it can be a config option
> >  but default-enabled on generic kernels so all distros will pick this hw
> >  support up), but we definitely need the genapic unification before we
> >  can add more features.
> 
> config option would be reasonable.
> for x86_64
> subarch already have X86_PC, X86_VSMP.
> we have X86_UVSMP

If there was a significant differece between UV and generic kernels
(or hardware), then I would agree. However, the only significant
difference is the APIC model on large systems. Small systems are
exactly compatible.

The problem with subarch is that we want 1 binary kernel to support
both generic hardware AND uv hardware. This restriction is desirable
for the distros and software vendors. Otherwise, additional kernel
images would have to be built, released, & certified.

--- jack

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/