Re: [PATCH] x86: Find offset for crashkernel reservation automatically

From: Bernhard Walle <bwalle@suse.de>
To: Vivek Goyal <vgoyal@redhat.com>
Cc: x86@kernel.org, kexec@lists.infradead.org,
	linux-kernel@vger.kernel.org,
	"Eric W. Biederman" <ebiederm@xmission.com>
Subject: Re: [PATCH] x86: Find offset for crashkernel reservation automatically
Date: Fri, 27 Jun 2008 16:22:23 +0200	[thread overview]
Message-ID: <20080627162223.2df5a8b9@halley.suse.de> (raw)
In-Reply-To: <20080627141910.GD5801@redhat.com>

* Vivek Goyal [2008-06-27 10:19]:
> On Fri, Jun 27, 2008 at 04:06:56PM +0200, Bernhard Walle wrote:
> > * Vivek Goyal [2008-06-27 09:42]:
> > > 
> > > Thinking more about. Let me step back. I think it is not good idea to
> > > take this kernel take decision about the capability of kernel being
> > > loaded. There is no way we can find out now that if a kernel is capable
> > > of running from this memory location or not. This is highly variable. So,
> > > please ignore above comment.
> > 
> > Maybe below 4G makes sense. Because you need some 32 bit memory for DMA.
> > That's mostly an architecture limitation, so that could make sense to
> > check here.
> 
> Even if you need some 32bit memory in the lower regions, kexec can handle
> it with backup segment mechanism (currently 640K). We can always increse
> the backup region size.
> 
> So I would think that it is best to leave it without any checking and 
> then let kexec-tools handle it.

Ah, that's true. Only on x86, right? (That would be an alternative for
ia64, too ...)

But in general policy should go in userspace (if possible), so I agree
with you that kexec-tools can handle that.

Bernhard
-- 
Bernhard Walle, SUSE LINUX Products GmbH, Architecture Development

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec