[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180720134026.6saxekgxxgltv3hg@kshutemo-mobl1>
Date: Fri, 20 Jul 2018 16:40:26 +0300
From: "Kirill A. Shutemov" <kirill@...temov.name>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: Dave Hansen <dave.hansen@...el.com>,
"Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
Ingo Molnar <mingo@...hat.com>, x86@...nel.org,
"H. Peter Anvin" <hpa@...or.com>,
Tom Lendacky <thomas.lendacky@....com>,
Kai Huang <kai.huang@...ux.intel.com>,
Jacob Pan <jacob.jun.pan@...ux.intel.com>,
linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: [PATCHv5 08/19] x86/mm: Introduce variables to store number,
shift and mask of KeyIDs
On Fri, Jul 20, 2018 at 03:17:54PM +0200, Thomas Gleixner wrote:
> On Fri, 20 Jul 2018, Kirill A. Shutemov wrote:
> > On Thu, Jul 19, 2018 at 03:40:41PM +0200, Thomas Gleixner wrote:
> > > > > I still don't see how that's supposed to work.
> > > > >
> > > > > When the inconsistent CPU is brought up _AFTER_ MKTME is enabled, then how
> > > > > does clearing the variables help? It does not magically make all the other
> > > > > stuff go away.
> > > >
> > > > We don't actually enable MKTME in kernel. BIOS does. Kernel makes choose
> > > > to use it or not. Current design targeted to be used by userspace.
> > > > So until init we don't have any other stuff to go away. We can just
> > > > pretend that MKTME was never there.
> > >
> > > Hotplug is not guaranteed to happen _BEFORE_ init. Think about physical
> > > hotplug.
> >
> > Ouch. I didn't think about this. :/
> >
> > In this case I don't see how to handle the situation properly.
> > Is it okay to WARN() && pray()?
>
> Not really. First of all, you want to do the initial checking on the boot
> CPU and then when secondary CPUs are brought up, verify that they have
> matching parameters. If they do not, then we should just shut them down
> right away before they can touch anything which is TME related and mark
> them as 'don't online again'. That needs some extra logic in the hotplug
> code, but I already have played with that for different reasons. Stick a
> fat comment into that 'not matching' code path for now and I'll give you
> the magic for preventing full bringup after polishing it a bit.
Got it. Thanks!
--
Kirill A. Shutemov
Powered by blists - more mailing lists