Re: [Qemu-devel] [RFC][PATCH 2/2] xen:i386:pc_piix: create isa bridge specific to IGD passthrough

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Nov 17, 2014 at 07:18:17PM +0800, Chen, Tiejun wrote:
> On 2014/11/17 18:13, Michael S. Tsirkin wrote:
> >On Mon, Nov 17, 2014 at 05:42:12PM +0800, Chen, Tiejun wrote:
> >>On 2014/11/17 17:25, Michael S. Tsirkin wrote:
> >>>On Mon, Nov 17, 2014 at 04:48:32PM +0800, Chen, Tiejun wrote:
> >>>>On 2014/11/17 14:10, Michael S. Tsirkin wrote:
> >>>>>On Mon, Nov 17, 2014 at 10:47:56AM +0800, Chen, Tiejun wrote:
> >>>>>>On 2014/11/5 22:09, Michael S. Tsirkin wrote:
> >>>>>>>On Wed, Nov 05, 2014 at 03:22:59PM +0800, Tiejun Chen wrote:
> >>>>>>>>Currently IGD drivers always need to access PCH by 1f.0, and
> >>>>>>>>PCH vendor/device id is used to identify the card.
> >>>>>>>>
> >>>>>>>>Signed-off-by: Tiejun Chen <tiejun.chen@intel.com>
> >>>>>>>>---
> >>
> >>[snip]
> >>
> >>>Cleaner:
> >>>	 if (!pci_dev) {
> >>>		fprintf
> >>>		return;
> >>>	}
> >>>          pci_config_set_device_id(pci_dev->config, pch_id);
> >>
> >>I will address all comments and thanks.
> >>
> >>>
> >>>>+    }
> >>>>+}
> >>>>+
> >>>>  /* init */
> >>>>
> >>>>  static int xen_pt_initfn(PCIDevice *d)
> >>>>@@ -682,6 +770,9 @@ static int xen_pt_initfn(PCIDevice *d)
> >>>>          return -1;
> >>>>      }
> >>>>
> >>>>+    /* Register ISA bridge for passthrough GFX. */
> >>>>+    xen_igd_passthrough_isa_bridge_create(s, &s->real_device);
> >>>>+
> >>>>      /* reinitialize each config register to be emulated */
> >>>>      if (xen_pt_config_init(s)) {
> >>>>          XEN_PT_ERR(d, "PCI Config space initialisation failed.\n");
> >>>>
> >>>>Note I will introduce a inline function in another patch,
> >>>>
> >>>>+static inline int is_vga_passthrough(XenHostPCIDevice *dev)
> >>>>+{
> >>>>+    return (xen_has_gfx_passthru && (dev->vendor_id == PCI_VENDOR_ID_INTEL)
> >>>>+            && ((dev->class_code >> 0x8) == PCI_CLASS_DISPLAY_VGA));
> >>>>+}
> >>>>
> >>>>Thanks
> >>>>Tiejun
> >>>
> >>>Why bother with all these conditions?
> >>>Won't it be enough to check dev->vendor_id == PCI_VENDOR_ID_INTEL?
> >>>
> >>
> >>If this is just used for IGD, its always fine without checking vendor_id.
> >
> >You need to match device ID to *know* it's IGD.
> >
> >>So after remove that check, I guess I need to rename that as
> >>is_igd_vga_passthrough() to make sense.
> >>
> >>Thanks
> >>Tiejun
> >
> >There is no need to check class code or xen_has_gfx_passthru flag.
> >Device ID + vendor ID identifies each device uniquely.
> >
> 
> Yeah.
> 
> Here I assume vendor ID is always PCI_VENDOR_ID_INTEL so looks you means I
> also need to check that table to do something like,
> 
> is_igd_vga_passthugh(dev)
> {	
> 	int i;
> 	int num = ARRAY_SIZE(xen_igd_combo_id_infos);
> 	for (i = 0; i < num; i++) {
> 		if (dev->device_id == xen_igd_combo_id_infos[i].gpu_device_id)
> 			return 1;
> 	return 0;
> }
> 
> Then we can simplify the subsequent codes based on this, right?
> 
> Thanks
> Tiejun

Yea.
Basically let's try to treat this simply as a device quirk,
and see where this gets us.

-- 
MST