All of lore.kernel.org
 help / color / mirror / Atom feed
From: George Dunlap <george.dunlap@citrix.com>
To: Ian Jackson <Ian.Jackson@eu.citrix.com>,
	George Dunlap <George.Dunlap@eu.citrix.com>
Cc: Wei Liu <wei.liu2@citrix.com>,
	Ian Campbell <ian.campbell@citrix.com>,
	Stefano Stabellini <stefano.stabellini@eu.citrix.com>,
	"xen-devel@lists.xen.org" <xen-devel@lists.xen.org>,
	Anthony PERARD <anthony.perard@citrix.com>,
	Boris Ostrovsky <boris.ostrovsky@oracle.com>
Subject: Re: [PATCH] libxl: No emulated disk driver for xvdX disk
Date: Tue, 27 Oct 2015 14:34:12 +0000	[thread overview]
Message-ID: <562F8B64.8090003@citrix.com> (raw)
In-Reply-To: <22063.31205.869213.960160@mariner.uk.xensource.com>

On 27/10/15 13:19, Ian Jackson wrote:
> George Dunlap writes ("Re: [Xen-devel] [PATCH] libxl: No emulated disk driver for xvdX disk"):
>> On Tue, Oct 27, 2015 at 11:11 AM, Stefano Stabellini
>> <stefano.stabellini@eu.citrix.com> wrote:
>>> The guest can access xvdX disks via the PV block protocol, or via
>>> UEFI firmare using EFI BootServices.
>>
>> But it introduces a sort of regression in some setups -- namely,
>> before you could use exactly the same disk and config file for either
>> a PV or HVM guest, simply adding builder=hvm (either to the file or
>> the command line) to switch between the two.
> 
> The purpose of the `vdev' in the xl disk specification is to say how
> the disk should be presented to the guest.
> 
> Did you look at docs/misc/vbd-interface.txt ?
> 
> If you want something that is accessible via the BIOS etc. in HVM
> guests, why not use vdev=hda rather than vdev=xvda ?

If I then take the came config file and remove 'builder="hvm"', will the
resulting thing boot as a PV guest?

I understand the idea of saying, "Well, 'vdev=xvda' means a PV disk; you
haven't asked for an emulated disk, so you shouldn't expect to get one."
 It's all very nice and neat and logical.  But:

1. It's been this way for years, and at least many developers' testing
systems have come to rely on this behavior.  I'd be surprised if this
didn't break a lot of users' configurations on upgrade

2. Being able to boot the exact same config file either PV or HVM is a
feature that makes sense to have

If both SeaBIOS and OVMF had PV drivers I don't think it would really be
an issue; but they don't.

The real issue I guess is that there is "automatic" behavior here that
we don't have a flag to control: PV disks get emulated counterparts
automatically, and emulated disks get PV counterparts automatically.
Would it make sense to have a per-device flag to specify this?
mirror=(true|false|auto), where "auto" is the default?

If we can detect whether we're booting with SeaBIOS or OVMF, we can even
have "auto" default to "false' for OVMF, and "true" for SeaBIOS (until
such time as we have PV drivers in SeaBIOS).

 -George

      parent reply	other threads:[~2015-10-27 14:34 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-10-14 11:05 [PATCH] libxl: No emulated disk driver for xvdX disk Anthony PERARD
2015-10-16 14:01 ` Ian Jackson
2015-10-22 15:54   ` Ian Campbell
2015-10-26 21:53   ` Boris Ostrovsky
2015-10-26 21:53   ` Boris Ostrovsky
2015-10-27 11:11     ` Stefano Stabellini
2015-10-27 11:31       ` George Dunlap
2015-10-27 13:19         ` Ian Jackson
2015-10-27 13:25           ` Boris Ostrovsky
2015-10-27 14:26             ` Ian Jackson
2015-10-27 14:35               ` Boris Ostrovsky
2015-10-27 14:43                 ` Ian Jackson
2015-10-27 14:52                   ` George Dunlap
2015-10-27 14:34           ` George Dunlap [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=562F8B64.8090003@citrix.com \
    --to=george.dunlap@citrix.com \
    --cc=George.Dunlap@eu.citrix.com \
    --cc=Ian.Jackson@eu.citrix.com \
    --cc=anthony.perard@citrix.com \
    --cc=boris.ostrovsky@oracle.com \
    --cc=ian.campbell@citrix.com \
    --cc=stefano.stabellini@eu.citrix.com \
    --cc=wei.liu2@citrix.com \
    --cc=xen-devel@lists.xen.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.