Re: New Xen boot infrastructure proposal

From: Keir Fraser <keir.xen@gmail.com>
To: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>,
	Jan Beulich <JBeulich@suse.com>
Cc: Daniel Kiper <daniel.kiper@oracle.com>,
	stefano.stabellini@eu.citrix.com, keir@xen.org,
	ian.campbell@citrix.com, xen-devel <xen-devel@lists.xen.org>
Subject: Re: New Xen boot infrastructure proposal
Date: Wed, 22 May 2013 17:56:15 +0100	[thread overview]
Message-ID: <CDC2B73F.5361F%keir.xen@gmail.com> (raw)
In-Reply-To: <20130522164745.GD9712@phenom.dumpdata.com>

On 22/05/2013 17:47, "Konrad Rzeszutek Wilk" <konrad.wilk@oracle.com> wrote:

> On Wed, May 22, 2013 at 04:16:30PM +0100, Jan Beulich wrote:
>>>>> On 22.05.13 at 17:01, Daniel Kiper <daniel.kiper@oracle.com> wrote:
>>> If we stick to current MBI I am not able to pass (in sensible way),
>>> from preloader to __start_xen(), e.g. ACPI and EFI stuff from multiboot2
>>> protocol.
>> 
>> Why? You get handed a list (almost like an array) of items, and you'd
>> pass the base address instead of the base address of the multiboot
>> structure that we pass right now, together with an indicator which
>> of the two it is. Then __start_xen() has to adopt its behavior to this.
>> Not a big deal afaict.
> 
> Won't you have to do a bunch of 'if (multibootv1) { use_this_offset } else
> if (multibootv2) { use this other offset }' in the code to support
> both formats?
> 
> If we just have a mesh of both of them we only have to do this sort
> of copying only once and can just use the struct that encompasses
> v2, v1, and whatever else we need (say pointer to RSDT).

Yes, having all the x86 (say) preloaders marshal into one single
x86-specific format makes sense. That's the sort of patch I would support.

 -- Keir

> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel