Skip to contentSkip to conversation

PCI/PM: Skip resuming to D0 if disconnected

PATCH Superseded Mario Limonciello 4 participants 6 messages Jun 23, 2025 – Jun 26, 2025 PCI

v4 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

PATCH v4 1/1 PCI/PM: Skip resuming to D0 if disconnected

From: Mario Limonciello <[email protected]>

When a USB4 dock is unplugged the PCIe bridge it's connected to will issue a "Link Down" and "Card not detected event". The PCI core will treat this as a surprise hotplug event and unconfigure all downstream devices. This involves setting the device error state to pci_channel_io_perm_failure which pci_dev_is_disconnected() will check.

It doesn't make sense to runtime resume disconnected devices to D0 and report the (expected) error, so bail early.

---
v4:
* no info message
v3:
* Adjust text and subject
* Add an info message instead
v2:
* Use pci_dev_is_disconnected()
v1: https://lore.kernel.org/linux-usb/[email protected]/T/#mf95c947990d016fbfccfd11afe60b8ae08aafa0b

---
 drivers/pci/pci.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/pci/pci.c +5 -0
index 9e42090fb1089..160a9a482c732 100644
@@ -1374,6 +1374,11 @@ int pci_power_up(struct pci_dev *dev)
 		return -EIO;
 	}
 
+	if (pci_dev_is_disconnected(dev)) {
+		dev->current_state = PCI_D3cold;
+		return -EIO;
+	}
+
 	pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);
 	if (PCI_POSSIBLE_ERROR(pmcsr)) {
 		pci_err(dev, "Unable to change power state from %s to D0, device inaccessible\n",
Signature · 1 line
2.43.0
Reviewed-by on v4 1/1
#1 lore raw

[cc += Rafael, Mika]

Quoted 8 lines · Mario Limonciello

From: Mario Limonciello <[email protected]>

When a USB4 dock is unplugged the PCIe bridge it's connected to will issue a "Link Down" and "Card not detected event". The PCI core will treat this as a surprise hotplug event and unconfigure all downstream devices. This involves setting the device error state to pci_channel_io_perm_failure which pci_dev_is_disconnected() will check.

Quoted 5 lines

It doesn't make sense to runtime resume disconnected devices to D0 and report the (expected) error, so bail early.

Quoted 31 lines

---
v4:
* no info message
v3:
* Adjust text and subject
* Add an info message instead
v2:
* Use pci_dev_is_disconnected()
v1: https://lore.kernel.org/linux-usb/[email protected]/T/#mf95c947990d016fbfccfd11afe60b8ae08aafa0b

---
 drivers/pci/pci.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/pci/pci.c +5 -0
index 9e42090fb1089..160a9a482c732 100644
@@ -1374,6 +1374,11 @@ int pci_power_up(struct pci_dev *dev)
 		return -EIO;
 	}
 
+	if (pci_dev_is_disconnected(dev)) {
+		dev->current_state = PCI_D3cold;
+		return -EIO;
+	}
+
 	pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);
 	if (PCI_POSSIBLE_ERROR(pmcsr)) {
 		pci_err(dev, "Unable to change power state from %s to D0, device inaccessible\n",
Signature · 1 line
2.43.0
#2 lore raw
Quoted 41 lines · Lukas Wunner

[cc += Rafael, Mika]

Quoted 8 lines · Mario Limonciello

From: Mario Limonciello <[email protected]>

When a USB4 dock is unplugged the PCIe bridge it's connected to will issue a "Link Down" and "Card not detected event". The PCI core will treat this as a surprise hotplug event and unconfigure all downstream devices. This involves setting the device error state to pci_channel_io_perm_failure which pci_dev_is_disconnected() will check.

Quoted 5 lines

It doesn't make sense to runtime resume disconnected devices to D0 and report the (expected) error, so bail early.

Quoted 21 lines

---
v4:
* no info message
v3:
* Adjust text and subject
* Add an info message instead
v2:
* Use pci_dev_is_disconnected()
v1: https://lore.kernel.org/linux-usb/[email protected]/T/#mf95c947990d016fbfccfd11afe60b8ae08aafa0b

---
 drivers/pci/pci.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/pci/pci.c +0 -0
index 9e42090fb1089..160a9a482c732 100644
@@ -1374,6 +1374,11 @@ int pci_power_up(struct pci_dev *dev)
              return -EIO;
      }
 
Quoted 2 lines

drivers/pci/pci.c int pci_power_up(struct pci_dev *dev) · line 1378

Quoted 2 lines
drivers/pci/pci.c +2 -0
+     if (pci_dev_is_disconnected(dev)) {
+             dev->current_state = PCI_D3cold;

Why not PCI_UNKNOWN?

Quoted 8 lines
Quoted 8 lines
drivers/pci/pci.c +3 -1
+             return -EIO;
+     }
+
      pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);
      if (PCI_POSSIBLE_ERROR(pmcsr)) {
              pci_err(dev, "Unable to change power state from %s to D0, device inaccessible\n",
--

2.43.0

#3 lore raw
Quoted 42 lines · Rafael J. Wysocki
Quoted 41 lines · Lukas Wunner

[cc += Rafael, Mika]

Quoted 8 lines · Mario Limonciello

From: Mario Limonciello <[email protected]>

When a USB4 dock is unplugged the PCIe bridge it's connected to will issue a "Link Down" and "Card not detected event". The PCI core will treat this as a surprise hotplug event and unconfigure all downstream devices. This involves setting the device error state to pci_channel_io_perm_failure which pci_dev_is_disconnected() will check.

Quoted 5 lines

It doesn't make sense to runtime resume disconnected devices to D0 and report the (expected) error, so bail early.

Quoted 21 lines

---
v4:
* no info message
v3:
* Adjust text and subject
* Add an info message instead
v2:
* Use pci_dev_is_disconnected()
v1: https://lore.kernel.org/linux-usb/[email protected]/T/#mf95c947990d016fbfccfd11afe60b8ae08aafa0b

---
  drivers/pci/pci.c | 5 +++++
  1 file changed, 5 insertions(+)
drivers/pci/pci.c +0 -0
index 9e42090fb1089..160a9a482c732 100644
@@ -1374,6 +1374,11 @@ int pci_power_up(struct pci_dev *dev)
               return -EIO;
       }
 
Quoted 4 lines
Quoted 2 lines

drivers/pci/pci.c int pci_power_up(struct pci_dev *dev) · line 1378

Quoted 2 lines
drivers/pci/pci.c +2 -0
+     if (pci_dev_is_disconnected(dev)) {
+             dev->current_state = PCI_D3cold;

Why not PCI_UNKNOWN?

It was following what other situations of failure did:
* existing error in pci_power_up()
* error in pci_update_current_state()
* error in pci_set_low_power_state()

I view all of these cases as unrecoverable failures.

So perhaps if changing this one to PCI_UNKNOWN those three should those also be PCI_UNKNOWN?

Bjorn, Lukas, thoughts?

Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
drivers/pci/pci.c +3 -1
+             return -EIO;
+     }
+
       pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);
       if (PCI_POSSIBLE_ERROR(pmcsr)) {
               pci_err(dev, "Unable to change power state from %s to D0, device inaccessible\n",
--

2.43.0

Reviewed-by
#4 lore raw
Quoted 55 lines · Mario Limonciello
Quoted 44 lines · Rafael J. Wysocki
Quoted 43 lines · Lukas Wunner

[cc += Rafael, Mika]

Quoted 9 lines · Mario Limonciello

From: Mario Limonciello <[email protected]>

When a USB4 dock is unplugged the PCIe bridge it's connected to will issue a "Link Down" and "Card not detected event". The PCI core will treat this as a surprise hotplug event and unconfigure all downstream devices. This involves setting the device error state to pci_channel_io_perm_failure which pci_dev_is_disconnected() will check.

Quoted 5 lines

It doesn't make sense to runtime resume disconnected devices to D0 and report the (expected) error, so bail early.

Quoted 22 lines

---
v4:
* no info message
v3:
* Adjust text and subject
* Add an info message instead
v2:
* Use pci_dev_is_disconnected()
v1:
https://lore.kernel.org/linux-usb/[email protected]/T/#mf95c947990d016fbfccfd11afe60b8ae08aafa0b

---
  drivers/pci/pci.c | 5 +++++
  1 file changed, 5 insertions(+)
drivers/pci/pci.c +0 -0
index 9e42090fb1089..160a9a482c732 100644
@@ -1374,6 +1374,11 @@ int pci_power_up(struct pci_dev *dev)
               return -EIO;
       }
 
Quoted 4 lines
Quoted 2 lines

drivers/pci/pci.c int pci_power_up(struct pci_dev *dev) · line 1378

Quoted 2 lines
drivers/pci/pci.c +2 -0
+     if (pci_dev_is_disconnected(dev)) {
+             dev->current_state = PCI_D3cold;

Why not PCI_UNKNOWN?

It was following what other situations of failure did:
* existing error in pci_power_up()
* error in pci_update_current_state()
* error in pci_set_low_power_state()

Quoted 15 lines

I view all of these cases as unrecoverable failures.

So perhaps if changing this one to PCI_UNKNOWN those three should those also be PCI_UNKNOWN?

Bjorn, Lukas, thoughts?

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
drivers/pci/pci.c +3 -0
+             return -EIO;
+     }
+
       pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);
       if (PCI_POSSIBLE_ERROR(pmcsr)) {
               pci_err(dev, "Unable to change power state from %s to D0,

device inaccessible\n",

(I've no opinion on PCI_UNKNOWN or not.)

Signature · 1 line
 i.
#5 lore raw
In reply to Mario Limonciello
Quoted 54 lines · Mario Limonciello
Quoted 42 lines · Rafael J. Wysocki
Quoted 41 lines · Lukas Wunner

[cc += Rafael, Mika]

Quoted 8 lines · Mario Limonciello

From: Mario Limonciello <[email protected]>

When a USB4 dock is unplugged the PCIe bridge it's connected to will issue a "Link Down" and "Card not detected event". The PCI core will treat this as a surprise hotplug event and unconfigure all downstream devices. This involves setting the device error state to pci_channel_io_perm_failure which pci_dev_is_disconnected() will check.

Quoted 5 lines

It doesn't make sense to runtime resume disconnected devices to D0 and report the (expected) error, so bail early.

Quoted 21 lines

---
v4:
* no info message
v3:
* Adjust text and subject
* Add an info message instead
v2:
* Use pci_dev_is_disconnected()
v1: https://lore.kernel.org/linux-usb/[email protected]/T/#mf95c947990d016fbfccfd11afe60b8ae08aafa0b

---
  drivers/pci/pci.c | 5 +++++
  1 file changed, 5 insertions(+)
drivers/pci/pci.c +0 -0
index 9e42090fb1089..160a9a482c732 100644
@@ -1374,6 +1374,11 @@ int pci_power_up(struct pci_dev *dev)
               return -EIO;
       }
 
Quoted 4 lines
Quoted 2 lines

drivers/pci/pci.c int pci_power_up(struct pci_dev *dev) · line 1378

Quoted 2 lines
drivers/pci/pci.c +2 -0
+     if (pci_dev_is_disconnected(dev)) {
+             dev->current_state = PCI_D3cold;

Why not PCI_UNKNOWN?

It was following what other situations of failure did:
* existing error in pci_power_up()
* error in pci_update_current_state()
* error in pci_set_low_power_state()

Quoted 6 lines

I view all of these cases as unrecoverable failures.

So perhaps if changing this one to PCI_UNKNOWN those three should those also be PCI_UNKNOWN?

Bjorn, Lukas, thoughts?

Note that pci_device_remove() sets power_state to PCI_UNKNOWN, but only if the old value is PCI_D0, so it would be good to make all of that consistent at one point.

Anyway, feel free to add

to this patch.

Thanks!