qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Michael S. Tsirkin" <mst@redhat.com>
To: "Sean O. Stalley" <sean.stalley@intel.com>
Cc: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH 1/1] Add support for PCI Enhanced Allocation "BARs"
Date: Thu, 14 May 2015 08:16:50 +0200	[thread overview]
Message-ID: <20150514081513-mutt-send-email-mst@redhat.com> (raw)
In-Reply-To: <20150513172942.GA2700@sean.stalley.intel.com>

On Wed, May 13, 2015 at 10:29:42AM -0700, Sean O. Stalley wrote:
> On Wed, May 13, 2015 at 08:09:46AM +0200, Michael S. Tsirkin wrote:
> > On Tue, May 12, 2015 at 02:23:07PM -0700, Sean O. Stalley wrote:
> > > On Tue, May 12, 2015 at 11:33:49AM +0200, Michael S. Tsirkin wrote:
> > > > On Mon, May 11, 2015 at 01:08:05PM -0700, Sean O. Stalley wrote:
> > > > > On Mon, May 11, 2015 at 09:26:08PM +0200, Michael S. Tsirkin wrote:
> > > > > > On Mon, May 11, 2015 at 11:56:44AM -0700, Sean O. Stalley wrote:
> > > > > > > PCI Enhanced Allocation is a new method of allocating MMIO & IO
> > > > > > > resources for PCI devices & bridges. It can be used instead
> > > > > > > of the traditional PCI method of using BARs.
> > > > > > > 
> > > > > > > EA entries are hardware-initialized to a fixed address.
> > > > > > > Unlike BARs, regions described by EA are cannot be moved.
> > > > > > > Because of this, only devices which are perminately connected to
> > > > > > > the PCI bus can use EA. A removeable PCI card must not use EA.
> > > > > > > 
> > > > > > > This patchset enables any existing QEMU PCI model to use EA in leiu of
> > > > > > > BARs. It adds EA options to the PCI device paramaters.
> > > > > > > 
> > > > > > > The Enhanced Allocation ECN is publicly available here:
> > > > > > > https://www.pcisig.com/specifications/conventional/ECN_Enhanced_Allocation_23_Oct_2014_Final.pdf
> > > > > > > 
> > > > > > > Signed-off-by: Sean O. Stalley <sean.stalley@intel.com>
> > > > > > 
> > > > > > 
> > > > > > I will review this, thanks.  Will you also be sending a seabios patch to
> > > > > > support these registers?  When you do, please Cc me.
> > > > > 
> > > > > Thanks Michael,
> > > > > 
> > > > > I wasn't planning on sending a seabios patch.
> > > > 
> > > > Won't this mean that bios might allocate PCI BARs conflicting with
> > > > EA entries, causing guest to crash?
> > > 
> > > The short answer is yes, having EA hardware without an EA-aware bios could cause the guest to crash.
> > > 
> > > To make matters worse, EA may cause old OSes may crash as well.
> > > If the OS supports hot-swap, and tries to remap the BARs, it could have the same problem.
> > 
> > Should _CRS in ACPI be updated to exclude the EA regions?
> > If we do this, this problem will go away, won't it?
> > 
> 
> I think that would work, assuming the OS uses the _CRS.
> (https://lwn.net/Articles/373551/)

Well that's Linux, so just send patches to teach it about EA.

> I wasn't quite sure how EA should affect ACPI.
> This is why I pushed out the QEMU hardware changes before the EDKII changes.

OK, so now we have an idea, I'm just hoping it's the correct one :)

> > > 
> > > This patch allows the user to add EA devices to an existing machine,
> > > which is something that would never happen with real EA hardware.
> > > 
> > > EA devices will only be present on new platforms.
> > > Because of this, I thought it made more sense to add EA support to EDKII than to seabios.
> > > Will adding EA support to seabios be necessary as well?
> > 
> > If the problem with old OSes is fixable, then I think it's a good idea.
> > 
> > > > > I do have some EDKII/OVMF patches I hope to send out soon. I will Cc you on those.
> > > > > 
> > > > > -Sean
> > > > > 
> > > > > > 
> > > > > > 
> > > > > > > ---
> > > > > > >  hw/pci/Makefile.objs      |   2 +-
> > > > > > >  hw/pci/pci.c              |  96 ++++++++++++++++------
> > > > > > >  hw/pci/pci_ea.c           | 203 ++++++++++++++++++++++++++++++++++++++++++++++
> > > > > > >  include/hw/pci/pci.h      |   7 ++
> > > > > > >  include/hw/pci/pci_ea.h   |  39 +++++++++
> > > > > > >  include/hw/pci/pci_regs.h |   4 +
> > > > > > >  include/qemu/typedefs.h   |   1 +
> > > > > > >  7 files changed, 328 insertions(+), 24 deletions(-)
> > > > > > >  create mode 100644 hw/pci/pci_ea.c
> > > > > > >  create mode 100644 include/hw/pci/pci_ea.h

  reply	other threads:[~2015-05-14  6:17 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-11 18:56 [Qemu-devel] [PATCH 1/1] Add support for PCI Enhanced Allocation "BARs" Sean O. Stalley
2015-05-11 19:26 ` Michael S. Tsirkin
2015-05-11 20:08   ` Sean O. Stalley
2015-05-12  9:33     ` Michael S. Tsirkin
2015-05-12 21:23       ` Sean O. Stalley
2015-05-13  6:09         ` Michael S. Tsirkin
2015-05-13 17:29           ` Sean O. Stalley
2015-05-14  6:16             ` Michael S. Tsirkin [this message]
2015-05-31 18:13             ` Michael S. Tsirkin
2015-06-01 16:37               ` Sean O. Stalley
2015-06-19 21:48                 ` Stalley, Sean
2015-06-22 21:15 ` Michael S. Tsirkin
2015-06-23 17:46   ` Stalley, Sean

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=20150514081513-mutt-send-email-mst@redhat.com \
    --to=mst@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=sean.stalley@intel.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).