Marking a DRHD as controlling an IGD isn't very sensible without
checking that at the very least it's a graphics device that lives at
0000:00:02.0. Re-use the reading of the class-code to control both the
clearing of "gfx_only" and the setting of "igd_drhd_address".

Signed-off-by: Jan Beulich <jbeulich@suse.com>
---
This way we also read the class code when we don't really need it, but
I think that's tolerable, and the simpler logic is preferable. The more
that with Marek's proposed patch, adjusted as suggested, we'd read
vendor and device IDs unconditionally for all devices (enumerated for a
DRHD) anyway.

--- a/xen/drivers/passthrough/vtd/dmar.c
+++ b/xen/drivers/passthrough/vtd/dmar.c
@@ -391,15 +391,12 @@ static int __init acpi_parse_dev_scope(
 
             if ( drhd )
             {
-                if ( (seg == 0) && (bus == 0) && (path->dev == 2) &&
-                     (path->fn == 0) )
-                    igd_drhd_address = drhd->address;
-
-                if ( gfx_only &&
-                     pci_conf_read8(PCI_SBDF(seg, bus, path->dev, path->fn),
+                if ( pci_conf_read8(PCI_SBDF(seg, bus, path->dev, path->fn),
                                     PCI_CLASS_DEVICE + 1) != 0x03
                                     /* PCI_BASE_CLASS_DISPLAY */ )
                     gfx_only = false;
+                else if ( !seg && !bus && path->dev == 2 && !path->fn )
+                    igd_drhd_address = drhd->address;
             }
 
             break;

> From: Jan Beulich <jbeulich@suse.com>
> Sent: Tuesday, February 28, 2023 5:52 PM
> 
> Marking a DRHD as controlling an IGD isn't very sensible without
> checking that at the very least it's a graphics device that lives at
> 0000:00:02.0. Re-use the reading of the class-code to control both the
> clearing of "gfx_only" and the setting of "igd_drhd_address".
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Reviewed-by: Kevin Tian <kevin.tian@intel.com>

On 28.02.2023 10:51, Jan Beulich wrote:
> Marking a DRHD as controlling an IGD isn't very sensible without
> checking that at the very least it's a graphics device that lives at
> 0000:00:02.0. Re-use the reading of the class-code to control both the
> clearing of "gfx_only" and the setting of "igd_drhd_address".
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Kevin?

Thanks, Jan

> ---
> This way we also read the class code when we don't really need it, but
> I think that's tolerable, and the simpler logic is preferable. The more
> that with Marek's proposed patch, adjusted as suggested, we'd read
> vendor and device IDs unconditionally for all devices (enumerated for a
> DRHD) anyway.
> 
> --- a/xen/drivers/passthrough/vtd/dmar.c
> +++ b/xen/drivers/passthrough/vtd/dmar.c
> @@ -391,15 +391,12 @@ static int __init acpi_parse_dev_scope(
>  
>              if ( drhd )
>              {
> -                if ( (seg == 0) && (bus == 0) && (path->dev == 2) &&
> -                     (path->fn == 0) )
> -                    igd_drhd_address = drhd->address;
> -
> -                if ( gfx_only &&
> -                     pci_conf_read8(PCI_SBDF(seg, bus, path->dev, path->fn),
> +                if ( pci_conf_read8(PCI_SBDF(seg, bus, path->dev, path->fn),
>                                      PCI_CLASS_DEVICE + 1) != 0x03
>                                      /* PCI_BASE_CLASS_DISPLAY */ )
>                      gfx_only = false;
> +                else if ( !seg && !bus && path->dev == 2 && !path->fn )
> +                    igd_drhd_address = drhd->address;
>              }
>  
>              break;