qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Chen, Tiejun" <tiejun.chen@intel.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: pbonzini@redhat.com, xen-devel@lists.xensource.com,
	qemu-devel@nongnu.org, stefano.stabellini@eu.citrix.com
Subject: Re: [Qemu-devel] [v4][PATCH 3/5] I440FX_PCI_DEVICE: add pci_type to index
Date: Mon, 11 Aug 2014 10:50:48 +0800	[thread overview]
Message-ID: <53E82F88.5020209@intel.com> (raw)
In-Reply-To: <20140810202759.GA5199@redhat.com>

On 2014/8/11 4:27, Michael S. Tsirkin wrote:
> On Thu, Aug 07, 2014 at 09:40:49AM +0800, Chen, Tiejun wrote:
>>
>>
>> On 2014/8/7 5:07, Michael S. Tsirkin wrote:
>>> On Wed, Aug 06, 2014 at 06:17:02PM +0800, Chen, Tiejun wrote:
>>>> On 2014/8/6 17:45, Michael S. Tsirkin wrote:
>>>>> On Wed, Aug 06, 2014 at 02:50:33PM +0800, Tiejun Chen wrote:
>>>>>> We need to use this index to reuse this macro later
>>>>>>
>>>>>> Signed-off-by: Tiejun Chen <tiejun.chen@intel.com>
>>>>>
>>>>> Which index?
>>>>> Most users don't need to change.
>>>>> Just open-code OBJECT_CHECK where necessary, or add
>>>>> a new wrapper.
>>>>
>>>> Okay so what about this?
>>>>
>>>>      hw:pci-host:piix: define I440FX_PCI_DEVICE_FROM_TYPE
>>>>
>>>>      We need to introduce I440FX_PCI_DEVICE_FROM_TYPE to get
>>>>      object with type, then we can reuse i440fx_init() simply.
>>>>
>>>>      Signed-off-by: Tiejun Chen <tiejun.chen@intel.com>
>>>>
>>>> diff --git a/hw/pci-host/piix.c b/hw/pci-host/piix.c
>>>> index 0cd82b8..8c74653 100644
>>>> --- a/hw/pci-host/piix.c
>>>> +++ b/hw/pci-host/piix.c
>>>> @@ -93,6 +93,9 @@ typedef struct PIIX3State {
>>>>   #define I440FX_PCI_DEVICE(obj) \
>>>>       OBJECT_CHECK(PCII440FXState, (obj), TYPE_I440FX_PCI_DEVICE)
>>>>
>>>> +#define I440FX_PCI_DEVICE_FROM_TYPE(obj, type) \
>>>> +    OBJECT_CHECK(PCII440FXState, (obj), type)
>>>
>>> This is just wrong. If you are casting to PCII440FXState,
>>
>> Why? We will have two different QOM typenames of PCII440FXStates.
>>
>>> there is no reason not to use TYPE_I440FX_PCI_DEVICE.
>>
>> As you know we already have this original,
>>
>> static const TypeInfo i440fx_info = {
>>      .name          = TYPE_I440FX_PCI_DEVICE,
>>      .parent        = TYPE_PCI_DEVICE,
>>      .instance_size = sizeof(PCII440FXState),
>>      .class_init    = i440fx_class_init,
>> };
>>
>> and in patch #4, we will register that new host bridge to IGD passthrough:
>>
>> static const TypeInfo xen_igd_passthrough_i440fx_info = {
>>      .name          = TYPE_XEN_IGD_PASSTHROUGH_I440FX_PCI_DEVICE,
>>      .parent        = TYPE_PCI_DEVICE,
>>      .instance_size = sizeof(PCII440FXState),
>>      .class_init    = xen_igd_passthrough_i440fx_class_init,
>> };
>
> My idea is to inherit TYPE_I440FX_PCI_DEVICE instead.

So here, this mean xen_igd_passthrough_i440fx_info's parent should be 
TYPE_I440FX_PCI_DEVICE, right?

> Then you can reuse regular piix code which casts to
> TYPE_I440FX_PCI_DEVICE because
> TYPE_XEN_IGD_PASSTHROUGH_I440FX_PCI_DEVICE would
> be a subclass of TYPE_I440FX_PCI_DEVICE.
>

So since then, we can reuse i440fx_initfn.

As a summary, what we should do is like the following,

xen:hw:pci-host:piix: create host bridge to passthrough

Implement a pci host bridge specific to passthrough. Actually
this just inherits the standard one.

This is based on http://patchwork.ozlabs.org/patch/363810/.

Signed-off-by: Tiejun Chen <tiejun.chen@intel.com>
---
  hw/pci-host/piix.c   | 28 ++++++++++++++++++++++++++++
  include/hw/i386/pc.h |  2 ++
  2 files changed, 30 insertions(+)

diff --git a/hw/pci-host/piix.c b/hw/pci-host/piix.c
index 0cd82b8..2ccd9ee 100644
--- a/hw/pci-host/piix.c
+++ b/hw/pci-host/piix.c
@@ -703,6 +703,33 @@ static const TypeInfo i440fx_info = {
      .class_init    = i440fx_class_init,
  };

+static void xen_igd_passthrough_i440fx_class_init(ObjectClass *klass, 
void *data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+    PCIDeviceClass *k = PCI_DEVICE_CLASS(klass);
+
+    k->init = i440fx_initfn;
+    k->vendor_id = PCI_VENDOR_ID_INTEL;
+    k->device_id = PCI_DEVICE_ID_INTEL_82441;
+    k->revision = 0x02;
+    k->class_id = PCI_CLASS_BRIDGE_HOST;
+    dc->desc = "IGD PT XEN Host bridge";
+    dc->vmsd = &vmstate_i440fx;
+    /*
+     * PCI-facing part of the host bridge, not usable without the
+     * host-facing part, which can't be device_add'ed, yet.
+     */
+    dc->cannot_instantiate_with_device_add_yet = true;
+    dc->hotpluggable   = false;
+}
+
+static const TypeInfo xen_igd_passthrough_i440fx_info = {
+    .name          = TYPE_XEN_IGD_PASSTHROUGH_I440FX_PCI_DEVICE,
+    .parent        = TYPE_I440FX_PCI_DEVICE,
+    .instance_size = sizeof(PCII440FXState),
+    .class_init    = xen_igd_passthrough_i440fx_class_init,
+};
+
  static const char *i440fx_pcihost_root_bus_path(PCIHostState *host_bridge,
                                                  PCIBus *rootbus)
  {
@@ -744,6 +771,7 @@ static const TypeInfo i440fx_pcihost_info = {
  static void i440fx_register_types(void)
  {
      type_register_static(&i440fx_info);
+    type_register_static(&xen_igd_passthrough_i440fx_info);
      type_register_static(&piix3_info);
      type_register_static(&piix3_xen_info);
      type_register_static(&i440fx_pcihost_info);
diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
index 11fb72f..de34aa6 100644
--- a/include/hw/i386/pc.h
+++ b/include/hw/i386/pc.h
@@ -238,6 +238,8 @@ typedef struct PCII440FXState PCII440FXState;
  #define TYPE_I440FX_PCI_HOST_BRIDGE "i440FX-pcihost"
  #define TYPE_I440FX_PCI_DEVICE "i440FX"

+#define TYPE_XEN_IGD_PASSTHROUGH_I440FX_PCI_DEVICE 
"xen-igd-passthrough-i440FX"
+
  PCIBus *i440fx_init(const char *host_type, const char *pci_type,
                      PCII440FXState **pi440fx_state, int *piix_devfn,
                      ISABus **isa_bus, qemu_irq *pic,
-- 
1.9.1

Thanks
Tiejun

  reply	other threads:[~2014-08-11  2:51 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-08-06  6:50 [Qemu-devel] [v4][PATCH 0/5] xen: introduce new machine for IGD passthrough Tiejun Chen
2014-08-06  6:50 ` [Qemu-devel] [v4][PATCH 1/5] i440fx: make types configurable at run-time Tiejun Chen
2014-08-06  6:50 ` [Qemu-devel] [v4][PATCH 2/5] pc_init1: pass parameters just with types Tiejun Chen
2014-08-06  6:50 ` [Qemu-devel] [v4][PATCH 3/5] I440FX_PCI_DEVICE: add pci_type to index Tiejun Chen
2014-08-06  9:45   ` Michael S. Tsirkin
2014-08-06 10:17     ` Chen, Tiejun
2014-08-06 21:07       ` Michael S. Tsirkin
2014-08-07  1:40         ` Chen, Tiejun
2014-08-10 20:27           ` Michael S. Tsirkin
2014-08-11  2:50             ` Chen, Tiejun [this message]
2014-08-12  8:54               ` Michael S. Tsirkin
2014-08-12  9:25                 ` Chen, Tiejun
2014-08-12  9:39                   ` Chen, Tiejun
2014-08-06  6:50 ` [Qemu-devel] [v4][PATCH 4/5] xen:hw:pci-host:piix: create host bridge to passthrough Tiejun Chen
2014-08-06  9:42   ` Michael S. Tsirkin
2014-08-06  9:47     ` Chen, Tiejun
2014-08-06 21:04       ` Michael S. Tsirkin
2014-08-07  1:44         ` Chen, Tiejun
2014-08-06  6:50 ` [Qemu-devel] [v4][PATCH 5/5] xen:hw:i386:pc_piix: introduce new machine for IGD passthrough Tiejun Chen

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=53E82F88.5020209@intel.com \
    --to=tiejun.chen@intel.com \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=stefano.stabellini@eu.citrix.com \
    --cc=xen-devel@lists.xensource.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).