Skip to contentSkip to conversation

ACPI / PCI: Fix NULL pointer dereference in acpi_get_pci_dev()

PATCH Applied Rafael J. Wysocki 3 participants 8 messages Oct 5, 2009 – Oct 14, 2009

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ACPI / PCI: Fix NULL pointer dereference in acpi_get_pci_dev()

From: Rafael J. Wysocki <[email protected]>

acpi_get_pci_dev() assumes that every handle it finds in the ACPI CA name space, between given device handle and the PCI root bridge handle, corresponds to a PCI-to-PCI bridge with an existing secondary bus. For this reason, when it finds a struct pci_dev object corresponding to one of them, it doesn't check if its 'subordinate' field is a valid pointer.

However, during resume from a sleep state, as well as at startup, we may get a dock notification, via dock_acpi_notifier registered by dock_init(), for a docking station device that is present in the ACPI tables, but not physically accessible at the moment. In that situation, the device appears to be below a PCI bridge whose 'subordinate' field is NULL. This, in turn, causes acpi_get_pci_dev() to trigger a NULL pointer dereference.

To fix this issue make acpi_get_pci_dev() check if pdev->subordinate is not NULL for every device it finds in the path between the root bridge and the device it's supposed to get to and return NULL if the "target" device is not really accessible.

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14129, which is a regression from 2.6.30.

---
 drivers/acpi/pci_root.c |   12 ++++++++++++
 1 file changed, 12 insertions(+)
drivers/acpi/pci_root.c +12 -0
Index: linux-2.6/drivers/acpi/pci_root.c
===================================================================
@@ -389,6 +389,18 @@ struct pci_dev *acpi_get_pci_dev(acpi_ha
 
 		pbus = pdev->subordinate;
 		pci_dev_put(pdev);
+
+		/*
+		 * During resume from a sleep state we can get a dock
+		 * notification for a device that is present in ACPI tables,
+		 * but not physically accessible at the moment, so tell the
+		 * caller it's not present in that case.
+		 */
+		if (!pbus) {
+			dev_info(&pdev->dev, "Secondary bus not present\n");
+			pdev = NULL;
+			break;
+		}
 	}
 out:
 	list_for_each_entry_safe(node, tmp, &device_list, node)
Reviewed-by on v1 1/1
#1 lore raw

* Rafael J. Wysocki <[email protected]>:

Quoted 3 lines

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14129, which is a regression from 2.6.30.

Thanks for doing this.

One small comment below (not major), but other than that:

Quoted 14 lines
---
 drivers/acpi/pci_root.c |   12 ++++++++++++
 1 file changed, 12 insertions(+)

Index: linux-2.6/drivers/acpi/pci_root.c
===================================================================

drivers/acpi/pci_root.c +0 -0
@@ -389,6 +389,18 @@ struct pci_dev *acpi_get_pci_dev(acpi_ha
 
 		pbus = pdev->subordinate;

drivers/acpi/pci_root.c struct pci_dev *acpi_get_pci_dev(acpi_ha · line 400

Quoted 10 lines
drivers/acpi/pci_root.c +9 -0
 		pci_dev_put(pdev);
+
+		/*
+		 * During resume from a sleep state we can get a dock
+		 * notification for a device that is present in ACPI tables,
+		 * but not physically accessible at the moment, so tell the
+		 * caller it's not present in that case.
+		 */
+		if (!pbus) {
+			dev_info(&pdev->dev, "Secondary bus not present\n");

Should this be dev_dbg() or maybe even strike it completely?

Quoted 7 lines
drivers/acpi/pci_root.c +3 -0
+			pdev = NULL;
+			break;
+		}
 	}
 out:
 	list_for_each_entry_safe(node, tmp, &device_list, node)
Reviewed-by on v1 1/1
#2 lore raw
In reply to Rafael J. Wysocki

* Rafael J. Wysocki <[email protected]>:

Quoted 3 lines

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14129, which is a regression from 2.6.30.

Thanks for doing this.

One small comment below (not major), but other than that:

Quoted 14 lines
---
 drivers/acpi/pci_root.c |   12 ++++++++++++
 1 file changed, 12 insertions(+)

Index: linux-2.6/drivers/acpi/pci_root.c
===================================================================

drivers/acpi/pci_root.c +0 -0
@@ -389,6 +389,18 @@ struct pci_dev *acpi_get_pci_dev(acpi_ha
 
 		pbus = pdev->subordinate;

drivers/acpi/pci_root.c struct pci_dev *acpi_get_pci_dev(acpi_ha · line 400

Quoted 10 lines
drivers/acpi/pci_root.c +9 -0
 		pci_dev_put(pdev);
+
+		/*
+		 * During resume from a sleep state we can get a dock
+		 * notification for a device that is present in ACPI tables,
+		 * but not physically accessible at the moment, so tell the
+		 * caller it's not present in that case.
+		 */
+		if (!pbus) {
+			dev_info(&pdev->dev, "Secondary bus not present\n");

Should this be dev_dbg() or maybe even strike it completely?

Quoted 7 lines
drivers/acpi/pci_root.c +3 -0
+			pdev = NULL;
+			break;
+		}
 	}
 out:
 	list_for_each_entry_safe(node, tmp, &device_list, node)
#3 lore raw
Quoted 24 lines · Alex Chiang

* Rafael J. Wysocki <[email protected]>:

Quoted 3 lines

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14129, which is a regression from 2.6.30.

Thanks for doing this.

One small comment below (not major), but other than that:

Quoted 13 lines
---
 drivers/acpi/pci_root.c |   12 ++++++++++++
 1 file changed, 12 insertions(+)

Index: linux-2.6/drivers/acpi/pci_root.c
===================================================================

drivers/acpi/pci_root.c +0 -0
@@ -389,6 +389,18 @@ struct pci_dev *acpi_get_pci_dev(acpi_ha
 
Quoted 13 lines

drivers/acpi/pci_root.c struct pci_dev *acpi_get_pci_dev(acpi_ha · line 400

Quoted 11 lines
drivers/acpi/pci_root.c +9 -0
 		pbus = pdev->subordinate;
 		pci_dev_put(pdev);
+
+		/*
+		 * During resume from a sleep state we can get a dock
+		 * notification for a device that is present in ACPI tables,
+		 * but not physically accessible at the moment, so tell the
+		 * caller it's not present in that case.
+		 */
+		if (!pbus) {
+			dev_info(&pdev->dev, "Secondary bus not present\n");

Should this be dev_dbg() or maybe even strike it completely?

I thought about that, but I decided not to just in order to get some idea about how many affected systems are out there. If that happens to be a problem for anyone, we can simply remove the message in the future.

Quoted 8 lines

drivers/acpi/pci_root.c struct pci_dev *acpi_get_pci_dev(acpi_ha · line 406

Quoted 8 lines
drivers/acpi/pci_root.c +3 -0
+			pdev = NULL;
+			break;
+		}
 	}
 out:
 	list_for_each_entry_safe(node, tmp, &device_list, node)

Thanks,
Rafael

#5 lore raw

PATCH 1/1 ACPI / PCI: Fix NULL pointer dereference in acpi_get_pci_dev() (rev. 2)

In reply to Rafael J. Wysocki

From: Rafael J. Wysocki <[email protected]>

acpi_get_pci_dev() may be called for a non-PCI device, in which case it should return NULL. However, it assumes that every handle it finds in the ACPI CA name space, between given device handle and the PCI root bridge handle, corresponds to a PCI-to-PCI bridge with an existing secondary bus. For this reason, when it finds a struct pci_dev object corresponding to one of them, it doesn't check if its 'subordinate' field is a valid pointer. This obviously leads to a NULL pointer dereference if acpi_get_pci_dev() is called for a non-PCI device with a PCI parent which is not a bridge.

To fix this issue make acpi_get_pci_dev() check if pdev->subordinate is not NULL for every device it finds on the path between the root bridge and the device it's supposed to get to and return NULL if the "target" device cannot be found.

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14129, which is a regression from 2.6.30.

---

Jesse, Len,

This is a replacement for the original $subject patch (now in the Len's tree as commit 5988eaded02e3cca2702f46efc255143468255bd).

The code was correct, but the comment and the changelog were not. Please use the one below instead.

Thanks,
Rafael

---
 drivers/acpi/pci_root.c |   11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/acpi/pci_root.c +11 -0
Index: linux-2.6/drivers/acpi/pci_root.c
===================================================================
@@ -389,6 +389,17 @@ struct pci_dev *acpi_get_pci_dev(acpi_ha
 
 		pbus = pdev->subordinate;
 		pci_dev_put(pdev);
+
+		/*
+		 * This function may be called for a non-PCI device that has a
+		 * PCI parent (eg. a disk under a PCI SATA controller).  In that
+		 * case pdev->subordinate will be NULL for the parent.
+		 */
+		if (!pbus) {
+			dev_dbg(&pdev->dev, "Not a PCI-to-PCI bridge\n");
+			pdev = NULL;
+			break;
+		}
 	}
 out:
 	list_for_each_entry_safe(node, tmp, &device_list, node)
reviewer
#6 lore raw

Re: [PATCH] ACPI / PCI: Fix NULL pointer dereference in acpi_get_pci_dev() (rev. 2)

Quoted 1 line

Please use the one below instead.

done.

thanks,
Len Brown, Intel Open Source Technology Center

Acked-by on v1 1/1
#7 lore raw

Re: [PATCH] ACPI / PCI: Fix NULL pointer dereference in acpi_get_pci_dev() (rev. 2)

In reply to Rafael J. Wysocki

* Rafael J. Wysocki <[email protected]>:

Quoted 12 lines

From: Rafael J. Wysocki <[email protected]>

acpi_get_pci_dev() may be called for a non-PCI device, in which case it should return NULL. However, it assumes that every handle it finds in the ACPI CA name space, between given device handle and the PCI root bridge handle, corresponds to a PCI-to-PCI bridge with an existing secondary bus. For this reason, when it finds a struct pci_dev object corresponding to one of them, it doesn't check if its 'subordinate' field is a valid pointer. This obviously leads to a NULL pointer dereference if acpi_get_pci_dev() is called for a non-PCI device with a PCI parent which is not a bridge.

Quoted 7 lines

To fix this issue make acpi_get_pci_dev() check if pdev->subordinate is not NULL for every device it finds on the path between the root bridge and the device it's supposed to get to and return NULL if the "target" device cannot be found.

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14129, which is a regression from 2.6.30.

Quoted 14 lines

---

Jesse, Len,

This is a replacement for the original $subject patch (now in the Len's tree as commit 5988eaded02e3cca2702f46efc255143468255bd).

The code was correct, but the comment and the changelog were not. Please use the one below instead.

Thanks,
Rafael