Skip to contentSkip to conversation

PCI/VGA: Look at all PCI display devices in VGA arbiter

RFC PATCH In review Mario Limonciello 2 participants 4 messages Jun 9, 2025 – Jun 11, 2025 PCI

lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 PCI/VGA: Look at all PCI display devices in VGA arbiter

From: Mario Limonciello <[email protected]>

On an A+N mobile system the APU is not being selected by some desktop environments for any rendering tasks. This is because the neither GPU is being treated as "boot_vga" but that is what some environments use to select a GPU [1].

The VGA arbiter driver only looks at devices that report as PCI display VGA class devices. Neither GPU on the system is a display VGA class device:

c5:00.0 3D controller: NVIDIA Corporation Device 2db9 (rev a1)
c6:00.0 Display controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 150e (rev d1)

So neither device gets the boot_vga sysfs file. The vga_is_boot_device() function already has some handling for integrated GPUs by looking at the ACPI HID entries, so if all PCI display class devices are looked at it actually can detect properly with this device ordering. However if there is a different ordering it could flag the wrong device.

Modify the VGA arbiter code and matching sysfs file entries to examine all PCI display class devices. After every device is added to the arbiter list make a pass on all devices and explicitly check whether one is integrated.

This will cause all GPUs to gain a boot_vga file, but the correct device (APU in this case) will now show 1 and the incorrect device shows 0. Userspace then picks the right device as well.

---
 drivers/pci/pci-sysfs.c |  2 +-
 drivers/pci/vgaarb.c    | 53 ++++++++++++++++++++++++++---------------
 include/linux/pci.h     | 15 ++++++++++++
 3 files changed, 50 insertions(+), 20 deletions(-)
drivers/pci/pci-sysfs.c +1 -1
index 268c69daa4d57..c314ee1b3f9ac 100644
@@ -1707,7 +1707,7 @@ static umode_t pci_dev_attrs_are_visible(struct kobject *kobj,
 	struct device *dev = kobj_to_dev(kobj);
 	struct pci_dev *pdev = to_pci_dev(dev);
 
-	if (a == &dev_attr_boot_vga.attr && pci_is_vga(pdev))
+	if (a == &dev_attr_boot_vga.attr && pci_is_display(pdev))
 		return a->mode;
 
 	return 0;
drivers/pci/vgaarb.c +34 -19

95 lines not on this page · raw message

include/linux/pci.h +15 -0
index 05e68f35f3923..e77754e43c629 100644
@@ -744,6 +744,21 @@ static inline bool pci_is_vga(struct pci_dev *pdev)
 	return false;
 }
 
+/**
+ * pci_is_display - Check if a PCI device is a display controller
+ * @pdev: Pointer to the PCI device structure
+ *
+ * This function determines whether the given PCI device corresponds
+ * to a display controller. Display controllers are typically used
+ * for graphical output and are identified based on their class code.
+ *
+ * Return: true if the PCI device is a display controller, false otherwise.
+ */
+static inline bool pci_is_display(struct pci_dev *pdev)
+{
+	return (pdev->class >> 16) == PCI_BASE_CLASS_DISPLAY;
+}
+
 #define for_each_pci_bridge(dev, bus)				\
 	list_for_each_entry(dev, &bus->devices, bus_list)	\
 		if (!pci_is_bridge(dev)) {} else
Signature · 1 line
2.43.0
#1 lore raw

Hi

Am 09.06.25 um 04:24 schrieb Mario Limonciello:

Quoted 7 lines

From: Mario Limonciello <[email protected]>

On an A+N mobile system the APU is not being selected by some desktop environments for any rendering tasks. This is because the neither GPU is being treated as "boot_vga" but that is what some environments use to select a GPU [1].

Quoted 6 lines

The VGA arbiter driver only looks at devices that report as PCI display VGA class devices. Neither GPU on the system is a display VGA class device:

c5:00.0 3D controller: NVIDIA Corporation Device 2db9 (rev a1)
c6:00.0 Display controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 150e (rev d1)

My understanding of vgaarb is that it manages concurrent usage of the fixed VGA I/O ports. So are these actually VGA devices? I'm concerned about vgaarb handling devices that aren't VGA and possible side effects of that.

As a side note, there's also video_is_primary_device(), which we use for fbcon and which also uses vga_default_device() on x86. [1] This helper should likely return the same value as sysfs' boot_vga attribute.

[1] https://elixir.bootlin.com/linux/v6.15.1/C/ident/video_is_primary_device

Quoted 41 lines

So neither device gets the boot_vga sysfs file. The vga_is_boot_device() function already has some handling for integrated GPUs by looking at the ACPI HID entries, so if all PCI display class devices are looked at it actually can detect properly with this device ordering. However if there is a different ordering it could flag the wrong device.

Modify the VGA arbiter code and matching sysfs file entries to examine all PCI display class devices. After every device is added to the arbiter list make a pass on all devices and explicitly check whether one is integrated.

This will cause all GPUs to gain a boot_vga file, but the correct device (APU in this case) will now show 1 and the incorrect device shows 0. Userspace then picks the right device as well.

---
  drivers/pci/pci-sysfs.c |  2 +-
  drivers/pci/vgaarb.c    | 53 ++++++++++++++++++++++++++---------------
  include/linux/pci.h     | 15 ++++++++++++
  3 files changed, 50 insertions(+), 20 deletions(-)
drivers/pci/pci-sysfs.c +1 -1
index 268c69daa4d57..c314ee1b3f9ac 100644
@@ -1707,7 +1707,7 @@ static umode_t pci_dev_attrs_are_visible(struct kobject *kobj,
  	struct device *dev = kobj_to_dev(kobj);
  	struct pci_dev *pdev = to_pci_dev(dev);
 
-	if (a == &dev_attr_boot_vga.attr && pci_is_vga(pdev))
+	if (a == &dev_attr_boot_vga.attr && pci_is_display(pdev))
  		return a->mode;
 
  	return 0;
drivers/pci/vgaarb.c +0 -0
index 78748e8d2dbae..8281144747487 100644

drivers/pci/vgaarb.c void vga_set_default_device(struct pci_dev *pdev) · line 142

Quoted 5 lines
drivers/pci/vgaarb.c +1 -1
@@ -139,7 +139,7 @@ void vga_set_default_device(struct pci_dev *pdev)
  {
  	if (vga_default == pdev)
  		return;
-
+	vgaarb_info(&pdev->dev, "selecting as VGA boot device\n");

vgaarb_dbg() please.

drivers/pci/vgaarb.c static bool vga_is_boot_device(struct vga_device *vgadev) · line 681

Quoted 13 lines
drivers/pci/vgaarb.c +3 -3
  	pci_dev_put(vga_default);
  	vga_default = pci_dev_get(pdev);
  }
@@ -676,9 +676,9 @@ static bool vga_is_boot_device(struct vga_device *vgadev)
  	}
 
  	/*
-	 * Vgadev has neither IO nor MEM enabled.  If we haven't found any
-	 * other VGA devices, it is the best candidate so far.
-	 */
+	* Vgadev has neither IO nor MEM enabled.  If we haven't found any
+	* other VGA devices, it is the best candidate so far.
+	*/

This should be a separate patch.

Best regards
Thomas

Quoted 109 lines
drivers/pci/vgaarb.c +30 -15
  	if (!boot_vga)
  		return true;
 
@@ -751,6 +751,31 @@ static void vga_arbiter_check_bridge_sharing(struct vga_device *vgadev)
  		vgaarb_info(&vgadev->pdev->dev, "no bridge control possible\n");
  }
 
+static
+void vga_arbiter_select_default_device(void)
+{
+	struct pci_dev *candidate = vga_default_device();
+	struct vga_device *vgadev;
+
+	list_for_each_entry(vgadev, &vga_list, list) {
+		if (vga_is_boot_device(vgadev)) {
+			/* check if one is an integrated GPU, use that if so */
+			if (candidate) {
+				if (vga_arb_integrated_gpu(&candidate->dev))
+					break;
+				if (vga_arb_integrated_gpu(&vgadev->pdev->dev)) {
+					candidate = vgadev->pdev;
+					break;
+				}
+			} else
+				candidate = vgadev->pdev;
+		}
+	}
+
+	if (candidate)
+		vga_set_default_device(candidate);
+}
+
  /*
   * Currently, we assume that the "initial" setup of the system is not sane,
   * that is, we come up with conflicting devices and let the arbiter's
@@ -816,13 +841,6 @@ static bool vga_arbiter_add_pci_device(struct pci_dev *pdev)
  		bus = bus->parent;
  	}
 
-	if (vga_is_boot_device(vgadev)) {
-		vgaarb_info(&pdev->dev, "setting as boot VGA device%s\n",
-			    vga_default_device() ?
-			    " (overriding previous)" : "");
-		vga_set_default_device(pdev);
-	}
-
  	vga_arbiter_check_bridge_sharing(vgadev);
 
  	/* Add to the list */
@@ -833,6 +851,7 @@ static bool vga_arbiter_add_pci_device(struct pci_dev *pdev)
  		vga_iostate_to_str(vgadev->owns),
  		vga_iostate_to_str(vgadev->locks));
 
+	vga_arbiter_select_default_device();
  	spin_unlock_irqrestore(&vga_lock, flags);
  	return true;
  fail:
@@ -1499,8 +1518,8 @@ static int pci_notify(struct notifier_block *nb, unsigned long action,
 
  	vgaarb_dbg(dev, "%s\n", __func__);
 
-	/* Only deal with VGA class devices */
-	if (!pci_is_vga(pdev))
+	/* Only deal with display devices */
+	if (!pci_is_display(pdev))
  		return 0;
 
  	/*
@@ -1548,12 +1567,8 @@ static int __init vga_arb_device_init(void)
 
  	/* Add all VGA class PCI devices by default */
  	pdev = NULL;
-	while ((pdev =
-		pci_get_subsys(PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID,
-			       PCI_ANY_ID, pdev)) != NULL) {
-		if (pci_is_vga(pdev))
-			vga_arbiter_add_pci_device(pdev);
-	}
+	while ((pdev = pci_get_base_class(PCI_BASE_CLASS_DISPLAY, pdev)))
+		vga_arbiter_add_pci_device(pdev);
 
  	pr_info("loaded\n");
  	return rc;
include/linux/pci.h +15 -0
index 05e68f35f3923..e77754e43c629 100644
@@ -744,6 +744,21 @@ static inline bool pci_is_vga(struct pci_dev *pdev)
  	return false;
  }
 
+/**
+ * pci_is_display - Check if a PCI device is a display controller
+ * @pdev: Pointer to the PCI device structure
+ *
+ * This function determines whether the given PCI device corresponds
+ * to a display controller. Display controllers are typically used
+ * for graphical output and are identified based on their class code.
+ *
+ * Return: true if the PCI device is a display controller, false otherwise.
+ */
+static inline bool pci_is_display(struct pci_dev *pdev)
+{
+	return (pdev->class >> 16) == PCI_BASE_CLASS_DISPLAY;
+}
+
  #define for_each_pci_bridge(dev, bus)				\
  	list_for_each_entry(dev, &bus->devices, bus_list)	\
  		if (!pci_is_bridge(dev)) {} else
Signature · 7 lines
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nuernberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman
HRB 36809 (AG Nuernberg)
#2 lore raw
Quoted 10 lines · Thomas Zimmermann

Hi

Am 09.06.25 um 04:24 schrieb Mario Limonciello:

Quoted 7 lines

From: Mario Limonciello <[email protected]>

On an A+N mobile system the APU is not being selected by some desktop environments for any rendering tasks. This is because the neither GPU is being treated as "boot_vga" but that is what some environments use to select a GPU [1].

Quoted 8 lines
Quoted 6 lines

The VGA arbiter driver only looks at devices that report as PCI display VGA class devices. Neither GPU on the system is a display VGA class device:

c5:00.0 3D controller: NVIDIA Corporation Device 2db9 (rev a1)
c6:00.0 Display controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 150e (rev d1)

My understanding of vgaarb is that it manages concurrent usage of the fixed VGA I/O ports. So are these actually VGA devices? I'm concerned about vgaarb handling devices that aren't VGA and possible side effects of that.

No; neither is a VGA device. There was a suggestion to do this from userspace in libpciaccess [1] but Dave Airlie suggested it would be better to adjust the "meaning" of boot_vga, which is essentially what this RFC does.

[1] https://gitlab.freedesktop.org/xorg/lib/libpciaccess/-/merge_requests/37

Quoted 176 lines

As a side note, there's also video_is_primary_device(), which we use for fbcon and which also uses vga_default_device() on x86. [1] This helper should likely return the same value as sysfs' boot_vga attribute.

[1] https://elixir.bootlin.com/linux/v6.15.1/C/ident/ video_is_primary_device

Quoted 40 lines

So neither device gets the boot_vga sysfs file. The vga_is_boot_device() function already has some handling for integrated GPUs by looking at the ACPI HID entries, so if all PCI display class devices are looked at it actually can detect properly with this device ordering.  However if there is a different ordering it could flag the wrong device.

Modify the VGA arbiter code and matching sysfs file entries to examine all PCI display class devices. After every device is added to the arbiter list make a pass on all devices and explicitly check whether one is integrated.

This will cause all GPUs to gain a boot_vga file, but the correct device (APU in this case) will now show 1 and the incorrect device shows 0. Userspace then picks the right device as well.

Link https://github.com/robherring/libpciaccess/commit/ b2838fb61c3542f107014b285cbda097acae1e12 [1]

---
  drivers/pci/pci-sysfs.c |  2 +-
  drivers/pci/vgaarb.c    | 53 ++++++++++++++++++++++++++--------------- include/linux/pci.h     | 15 ++++++++++++

  3 files changed, 50 insertions(+), 20 deletions(-)
drivers/pci/pci-sysfs.c +1 -1
index 268c69daa4d57..c314ee1b3f9ac 100644
@@ -1707,7 +1707,7 @@ static umode_t pci_dev_attrs_are_visible(struct 
 kobject *kobj,
       struct device *dev = kobj_to_dev(kobj);
       struct pci_dev *pdev = to_pci_dev(dev);
-    if (a == &dev_attr_boot_vga.attr && pci_is_vga(pdev))
+    if (a == &dev_attr_boot_vga.attr && pci_is_display(pdev))
           return a->mode;
       return 0;
drivers/pci/vgaarb.c +0 -0
index 78748e8d2dbae..8281144747487 100644

drivers/pci/vgaarb.c void vga_set_default_device(struct pci_dev *pdev) · line 142

Quoted 5 lines
drivers/pci/vgaarb.c +1 -1
@@ -139,7 +139,7 @@ void vga_set_default_device(struct pci_dev *pdev)
   {
       if (vga_default == pdev)
           return;
-
+    vgaarb_info(&pdev->dev, "selecting as VGA boot device\n");

vgaarb_dbg() please.

drivers/pci/vgaarb.c static bool vga_is_boot_device(struct vga_device · line 681

Quoted 13 lines
drivers/pci/vgaarb.c +3 -3
       pci_dev_put(vga_default);
       vga_default = pci_dev_get(pdev);
   }
@@ -676,9 +676,9 @@ static bool vga_is_boot_device(struct vga_device 
 *vgadev)
       }
       /*
-     * Vgadev has neither IO nor MEM enabled.  If we haven't found any
-     * other VGA devices, it is the best candidate so far.
-     */
+    * Vgadev has neither IO nor MEM enabled.  If we haven't found any
+    * other VGA devices, it is the best candidate so far.
+    */

This should be a separate patch.

Best regards
Thomas

Quoted 103 lines
drivers/pci/vgaarb.c +26 -7
       if (!boot_vga)
           return true;
@@ -751,6 +751,31 @@ static void 
 vga_arbiter_check_bridge_sharing(struct vga_device *vgadev)
           vgaarb_info(&vgadev->pdev->dev, "no bridge control possible\n");
   }
+static
+void vga_arbiter_select_default_device(void)
+{
+    struct pci_dev *candidate = vga_default_device();
+    struct vga_device *vgadev;
+
+    list_for_each_entry(vgadev, &vga_list, list) {
+        if (vga_is_boot_device(vgadev)) {
+            /* check if one is an integrated GPU, use that if so */
+            if (candidate) {
+                if (vga_arb_integrated_gpu(&candidate->dev))
+                    break;
+                if (vga_arb_integrated_gpu(&vgadev->pdev->dev)) {
+                    candidate = vgadev->pdev;
+                    break;
+                }
+            } else
+                candidate = vgadev->pdev;
+        }
+    }
+
+    if (candidate)
+        vga_set_default_device(candidate);
+}
+
   /*
    * Currently, we assume that the "initial" setup of the system is not sane,
    * that is, we come up with conflicting devices and let the arbiter's
@@ -816,13 +841,6 @@ static bool vga_arbiter_add_pci_device(struct 
 pci_dev *pdev)
           bus = bus->parent;
       }
-    if (vga_is_boot_device(vgadev)) {
-        vgaarb_info(&pdev->dev, "setting as boot VGA device%s\n",
-                vga_default_device() ?
-                " (overriding previous)" : "");
-        vga_set_default_device(pdev);
-    }
-
       vga_arbiter_check_bridge_sharing(vgadev);
       /* Add to the list */
@@ -833,6 +851,7 @@ static bool vga_arbiter_add_pci_device(struct 
 pci_dev *pdev)
           vga_iostate_to_str(vgadev->owns),
           vga_iostate_to_str(vgadev->locks));
+    vga_arbiter_select_default_device();
       spin_unlock_irqrestore(&vga_lock, flags);
       return true;

  fail:

drivers/pci/vgaarb.c +4 -8
@@ -1499,8 +1518,8 @@ static int pci_notify(struct notifier_block *nb, 
 unsigned long action,
       vgaarb_dbg(dev, "%s\n", __func__);
-    /* Only deal with VGA class devices */
-    if (!pci_is_vga(pdev))
+    /* Only deal with display devices */
+    if (!pci_is_display(pdev))
           return 0;
       /*
@@ -1548,12 +1567,8 @@ static int __init vga_arb_device_init(void)
       /* Add all VGA class PCI devices by default */
       pdev = NULL;
-    while ((pdev =
-        pci_get_subsys(PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID,
-                   PCI_ANY_ID, pdev)) != NULL) {
-        if (pci_is_vga(pdev))
-            vga_arbiter_add_pci_device(pdev);
-    }
+    while ((pdev = pci_get_base_class(PCI_BASE_CLASS_DISPLAY, pdev)))
+        vga_arbiter_add_pci_device(pdev);
       pr_info("loaded\n");
       return rc;
include/linux/pci.h +9 -0
index 05e68f35f3923..e77754e43c629 100644
@@ -744,6 +744,21 @@ static inline bool pci_is_vga(struct pci_dev *pdev)
       return false;
   }
+/**
+ * pci_is_display - Check if a PCI device is a display controller
+ * @pdev: Pointer to the PCI device structure
+ *
+ * This function determines whether the given PCI device corresponds
+ * to a display controller. Display controllers are typically used
+ * for graphical output and are identified based on their class code.
+ *
+ * Return: true if the PCI device is a display controller, false 

otherwise.
+ */

include/linux/pci.h +5 -0
+static inline bool pci_is_display(struct pci_dev *pdev)
+{
+    return (pdev->class >> 16) == PCI_BASE_CLASS_DISPLAY;
+}
+

  #define for_each_pci_bridge(dev, bus)                \
      list_for_each_entry(dev, &bus->devices, bus_list)    \
          if (!pci_is_bridge(dev)) {} else

#3 lore raw

Hi

Am 10.06.25 um 15:56 schrieb Mario Limonciello:

Quoted 15 lines · Thomas Zimmermann
Quoted 14 lines

Hi

Am 09.06.25 um 04:24 schrieb Mario Limonciello:

Quoted 11 lines

From: Mario Limonciello <[email protected]>

On an A+N mobile system the APU is not being selected by some desktop environments for any rendering tasks. This is because the neither GPU is being treated as "boot_vga" but that is what some environments use to select a GPU [1].

The VGA arbiter driver only looks at devices that report as PCI display VGA class devices. Neither GPU on the system is a display VGA class device:

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

c5:00.0 3D controller: NVIDIA Corporation Device 2db9 (rev a1)
c6:00.0 Display controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 150e (rev d1)

My understanding of vgaarb is that it manages concurrent usage of the fixed VGA I/O ports. So are these actually VGA devices? I'm concerned about vgaarb handling devices that aren't VGA and possible side effects of that.

No; neither is a VGA device.  There was a suggestion to do this from userspace in libpciaccess [1] but Dave Airlie suggested it would be better to adjust the "meaning" of boot_vga, which is essentially what this RFC does.

[1] https://gitlab.freedesktop.org/xorg/lib/libpciaccess/-/merge_requests/37

I would have done this in userspace as well, but *shrug*.

I still think that the vgaarb is too complicated and needs an overhaul, but that's for another series.

Best regards
Thomas

Quoted 180 lines
Quoted 179 lines

As a side note, there's also video_is_primary_device(), which we use for fbcon and which also uses vga_default_device() on x86. [1] This helper should likely return the same value as sysfs' boot_vga attribute.

[1] https://elixir.bootlin.com/linux/v6.15.1/C/ident/ video_is_primary_device

Quoted 41 lines

So neither device gets the boot_vga sysfs file. The vga_is_boot_device() function already has some handling for integrated GPUs by looking at the ACPI HID entries, so if all PCI display class devices are looked at it actually can detect properly with this device ordering. However if there is a different ordering it could flag the wrong device.

Modify the VGA arbiter code and matching sysfs file entries to examine all PCI display class devices. After every device is added to the arbiter list make a pass on all devices and explicitly check whether one is integrated.

This will cause all GPUs to gain a boot_vga file, but the correct device (APU in this case) will now show 1 and the incorrect device shows 0. Userspace then picks the right device as well.

Link https://github.com/robherring/libpciaccess/commit/ b2838fb61c3542f107014b285cbda097acae1e12 [1]

---
  drivers/pci/pci-sysfs.c |  2 +-
  drivers/pci/vgaarb.c    | 53
++++++++++++++++++++++++++---------------
  include/linux/pci.h     | 15 ++++++++++++

  3 files changed, 50 insertions(+), 20 deletions(-)
drivers/pci/pci-sysfs.c +1 -1
index 268c69daa4d57..c314ee1b3f9ac 100644
@@ -1707,7 +1707,7 @@ static umode_t 
 pci_dev_attrs_are_visible(struct kobject *kobj,
       struct device *dev = kobj_to_dev(kobj);
       struct pci_dev *pdev = to_pci_dev(dev);
-    if (a == &dev_attr_boot_vga.attr && pci_is_vga(pdev))
+    if (a == &dev_attr_boot_vga.attr && pci_is_display(pdev))
           return a->mode;
       return 0;
drivers/pci/vgaarb.c +0 -0
index 78748e8d2dbae..8281144747487 100644

drivers/pci/vgaarb.c void vga_set_default_device(struct pci_dev *pdev) · line 142

Quoted 5 lines
drivers/pci/vgaarb.c +1 -1
@@ -139,7 +139,7 @@ void vga_set_default_device(struct pci_dev *pdev)
   {
       if (vga_default == pdev)
           return;
-
+    vgaarb_info(&pdev->dev, "selecting as VGA boot device\n");

vgaarb_dbg() please.

drivers/pci/vgaarb.c static bool vga_is_boot_device(struct vga_device · line 681

Quoted 13 lines
drivers/pci/vgaarb.c +3 -3
       pci_dev_put(vga_default);
       vga_default = pci_dev_get(pdev);
   }
@@ -676,9 +676,9 @@ static bool vga_is_boot_device(struct vga_device 
 *vgadev)
       }
       /*
-     * Vgadev has neither IO nor MEM enabled.  If we haven't found any
-     * other VGA devices, it is the best candidate so far.
-     */
+    * Vgadev has neither IO nor MEM enabled.  If we haven't found any
+    * other VGA devices, it is the best candidate so far.
+    */

This should be a separate patch.

Best regards
Thomas

Quoted 104 lines
drivers/pci/vgaarb.c +26 -7
       if (!boot_vga)
           return true;
@@ -751,6 +751,31 @@ static void 
 vga_arbiter_check_bridge_sharing(struct vga_device *vgadev)
           vgaarb_info(&vgadev->pdev->dev, "no bridge control possible\n");
   }
+static
+void vga_arbiter_select_default_device(void)
+{
+    struct pci_dev *candidate = vga_default_device();
+    struct vga_device *vgadev;
+
+    list_for_each_entry(vgadev, &vga_list, list) {
+        if (vga_is_boot_device(vgadev)) {
+            /* check if one is an integrated GPU, use that if so */
+            if (candidate) {
+                if (vga_arb_integrated_gpu(&candidate->dev))
+                    break;
+                if (vga_arb_integrated_gpu(&vgadev->pdev->dev)) {
+                    candidate = vgadev->pdev;
+                    break;
+                }
+            } else
+                candidate = vgadev->pdev;
+        }
+    }
+
+    if (candidate)
+        vga_set_default_device(candidate);
+}
+
   /*
    * Currently, we assume that the "initial" setup of the system is not sane,
    * that is, we come up with conflicting devices and let the arbiter's
@@ -816,13 +841,6 @@ static bool vga_arbiter_add_pci_device(struct 
 pci_dev *pdev)
           bus = bus->parent;
       }
-    if (vga_is_boot_device(vgadev)) {
-        vgaarb_info(&pdev->dev, "setting as boot VGA device%s\n",
-                vga_default_device() ?
-                " (overriding previous)" : "");
-        vga_set_default_device(pdev);
-    }
-
       vga_arbiter_check_bridge_sharing(vgadev);
       /* Add to the list */
@@ -833,6 +851,7 @@ static bool vga_arbiter_add_pci_device(struct 
 pci_dev *pdev)
           vga_iostate_to_str(vgadev->owns),
           vga_iostate_to_str(vgadev->locks));
+    vga_arbiter_select_default_device();
       spin_unlock_irqrestore(&vga_lock, flags);
       return true;

  fail:

drivers/pci/vgaarb.c +4 -8
@@ -1499,8 +1518,8 @@ static int pci_notify(struct notifier_block 
 *nb, unsigned long action,
       vgaarb_dbg(dev, "%s\n", __func__);
-    /* Only deal with VGA class devices */
-    if (!pci_is_vga(pdev))
+    /* Only deal with display devices */
+    if (!pci_is_display(pdev))
           return 0;
       /*
@@ -1548,12 +1567,8 @@ static int __init vga_arb_device_init(void)
       /* Add all VGA class PCI devices by default */
       pdev = NULL;
-    while ((pdev =
-        pci_get_subsys(PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID,
-                   PCI_ANY_ID, pdev)) != NULL) {
-        if (pci_is_vga(pdev))
-            vga_arbiter_add_pci_device(pdev);
-    }
+    while ((pdev = pci_get_base_class(PCI_BASE_CLASS_DISPLAY, pdev)))
+        vga_arbiter_add_pci_device(pdev);
       pr_info("loaded\n");
       return rc;
include/linux/pci.h +9 -0
index 05e68f35f3923..e77754e43c629 100644
@@ -744,6 +744,21 @@ static inline bool pci_is_vga(struct pci_dev 
 *pdev)
       return false;
   }
+/**
+ * pci_is_display - Check if a PCI device is a display controller
+ * @pdev: Pointer to the PCI device structure
+ *
+ * This function determines whether the given PCI device corresponds
+ * to a display controller. Display controllers are typically used
+ * for graphical output and are identified based on their class code.
+ *
+ * Return: true if the PCI device is a display controller, false 

otherwise.
+ */

include/linux/pci.h +5 -0
+static inline bool pci_is_display(struct pci_dev *pdev)
+{
+    return (pdev->class >> 16) == PCI_BASE_CLASS_DISPLAY;
+}
+

  #define for_each_pci_bridge(dev, bus)                \
      list_for_each_entry(dev, &bus->devices, bus_list)    \
          if (!pci_is_bridge(dev)) {} else

Signature · 7 lines
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nuernberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman
HRB 36809 (AG Nuernberg)