[PATCH 1/2] powerpc/pseries: Fix corrupted pdn list

[PATCH 1/2] powerpc/pseries: Fix corrupted pdn list

[Gavin Shan]

Commit cca87d30 ("powerpc/pci: Refactor pci_dn") introduced pdn
list for SRIOV VFs. It means the pdn is be put into the child list
of its parent pdn when the pdn is created. When doing PCI hot
unplugging on pSeries, the PCI device node as well as its pdn are
released through procfs entry "powerpc/ofdt". Some one else grabs
the memory chunk of the pdn and update it accordingly. At the same
time, the pdn is still tracked in the child list of parent pdn. It
leads to corrupted child list in the parent pdn.

This fixes above issue by removing the pdn from the child list of
its parent pdn when the device node is detached from the system.
Note the pdn is free'd when the device node is released if the
device node is dynamic one. Otherwise, the device node as well
as the pdn won't be released.

Fixes: cca87d30 ("powerpc/pci: Refactor pci_dn")
Reported-by: Santwana Samantray <santwana.samantray@in.ibm.com>
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
---
 arch/powerpc/platforms/pseries/setup.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index df6a704..e6e8b24 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -268,6 +268,11 @@ static int pci_dn_reconfig_notifier(struct notifier_block *nb, unsigned long act
 			eeh_dev_init(PCI_DN(np), pci->phb);
 		}
 		break;
+	case OF_RECONFIG_DETACH_NODE:
+		pci = PCI_DN(np);
+		if (pci)
+			list_del(&pci->list);
+		break;
 	default:
 		err = NOTIFY_DONE;
 		break;
-- 
2.1.0

[PATCH 2/2] powerpc/pseries: Cleanup on pci_dn_reconfig_notifier()

[Gavin Shan]

This applies cleanup on pci_dn_reconfig_notifier(), no functional
changes:

   * Rename variable "pci" to "pdn" to indicate its purpose clearly.
   * The parent node can be released at any time. So it should be
     hold with of_get_parent() before accessing it.
   * The device node doesn't have to have parent node in theory.
     More check on this.

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
---
 arch/powerpc/platforms/pseries/setup.c | 24 +++++++++++++-----------
 1 file changed, 13 insertions(+), 11 deletions(-)

diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index e6e8b24..39a74fa 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -254,24 +254,26 @@ static void __init pseries_discover_pic(void)
 static int pci_dn_reconfig_notifier(struct notifier_block *nb, unsigned long action, void *data)
 {
 	struct of_reconfig_data *rd = data;
-	struct device_node *np = rd->dn;
-	struct pci_dn *pci = NULL;
+	struct device_node *parent, *np = rd->dn;
+	struct pci_dn *pdn;
 	int err = NOTIFY_OK;
 
 	switch (action) {
 	case OF_RECONFIG_ATTACH_NODE:
-		pci = np->parent->data;
-		if (pci) {
-			update_dn_pci_info(np, pci->phb);
-
-			/* Create EEH device for the OF node */
-			eeh_dev_init(PCI_DN(np), pci->phb);
+		parent = of_get_parent(np);
+		pdn = parent ? PCI_DN(parent) : NULL;
+		if (pdn) {
+			/* Create pdn and EEH device */
+			update_dn_pci_info(np, pdn->phb);
+			eeh_dev_init(PCI_DN(np), pdn->phb);
 		}
+
+		of_node_put(parent);
 		break;
 	case OF_RECONFIG_DETACH_NODE:
-		pci = PCI_DN(np);
-		if (pci)
-			list_del(&pci->list);
+		pdn = PCI_DN(np);
+		if (pdn)
+			list_del(&pdn->list);
 		break;
 	default:
 		err = NOTIFY_DONE;
-- 
2.1.0

Re: [1/2] powerpc/pseries: Fix corrupted pdn list

[Michael Ellerman]

On Thu, 2015-27-08 at 04:12:36 UTC, Gavin Shan wrote:
> Commit cca87d30 ("powerpc/pci: Refactor pci_dn") introduced pdn
> list for SRIOV VFs. It means the pdn is be put into the child list
> of its parent pdn when the pdn is created. When doing PCI hot
> unplugging on pSeries, the PCI device node as well as its pdn are
> released through procfs entry "powerpc/ofdt". Some one else grabs
> the memory chunk of the pdn and update it accordingly. At the same
> time, the pdn is still tracked in the child list of parent pdn. It
> leads to corrupted child list in the parent pdn.
> 
> This fixes above issue by removing the pdn from the child list of
> its parent pdn when the device node is detached from the system.
> Note the pdn is free'd when the device node is released if the
> device node is dynamic one. Otherwise, the device node as well
> as the pdn won't be released.
> 
> Fixes: cca87d30 ("powerpc/pci: Refactor pci_dn")

That went into 4.1, so this should presumably go to stable?

cheers

Re: [1/2] powerpc/pseries: Fix corrupted pdn list

[Gavin Shan]

On Fri, Aug 28, 2015 at 10:55:35AM +1000, Michael Ellerman wrote:
>On Thu, 2015-27-08 at 04:12:36 UTC, Gavin Shan wrote:
>> Commit cca87d30 ("powerpc/pci: Refactor pci_dn") introduced pdn
>> list for SRIOV VFs. It means the pdn is be put into the child list
>> of its parent pdn when the pdn is created. When doing PCI hot
>> unplugging on pSeries, the PCI device node as well as its pdn are
>> released through procfs entry "powerpc/ofdt". Some one else grabs
>> the memory chunk of the pdn and update it accordingly. At the same
>> time, the pdn is still tracked in the child list of parent pdn. It
>> leads to corrupted child list in the parent pdn.
>> 
>> This fixes above issue by removing the pdn from the child list of
>> its parent pdn when the device node is detached from the system.
>> Note the pdn is free'd when the device node is released if the
>> device node is dynamic one. Otherwise, the device node as well
>> as the pdn won't be released.
>> 
>> Fixes: cca87d30 ("powerpc/pci: Refactor pci_dn")
>
>That went into 4.1, so this should presumably go to stable?
>

Yeah, This should be put into 4.1 stable.

Thanks,
Gavin

>cheers
>

Re: [1/2] powerpc/pseries: Fix corrupted pdn list

[Michael Ellerman]

On Fri, 2015-08-28 at 11:09 +1000, Gavin Shan wrote:
> On Fri, Aug 28, 2015 at 10:55:35AM +1000, Michael Ellerman wrote:
> >On Thu, 2015-27-08 at 04:12:36 UTC, Gavin Shan wrote:
> >> Commit cca87d30 ("powerpc/pci: Refactor pci_dn") introduced pdn
> >> list for SRIOV VFs. It means the pdn is be put into the child list
> >> of its parent pdn when the pdn is created. When doing PCI hot
> >> unplugging on pSeries, the PCI device node as well as its pdn are
> >> released through procfs entry "powerpc/ofdt". Some one else grabs
> >> the memory chunk of the pdn and update it accordingly. At the same
> >> time, the pdn is still tracked in the child list of parent pdn. It
> >> leads to corrupted child list in the parent pdn.
> >> 
> >> This fixes above issue by removing the pdn from the child list of
> >> its parent pdn when the device node is detached from the system.
> >> Note the pdn is free'd when the device node is released if the
> >> device node is dynamic one. Otherwise, the device node as well
> >> as the pdn won't be released.
> >> 
> >> Fixes: cca87d30 ("powerpc/pci: Refactor pci_dn")
> >
> >That went into 4.1, so this should presumably go to stable?
> 
> Yeah, This should be put into 4.1 stable.

OK thanks.

I've added:

  Cc: stable@vger.kernel.org # 4.1

cheers

Re: [1/2] powerpc/pseries: Fix corrupted pdn list

[Gavin Shan]

On Fri, Aug 28, 2015 at 11:32:40AM +1000, Michael Ellerman wrote:
>On Fri, 2015-08-28 at 11:09 +1000, Gavin Shan wrote:
>> On Fri, Aug 28, 2015 at 10:55:35AM +1000, Michael Ellerman wrote:
>> >On Thu, 2015-27-08 at 04:12:36 UTC, Gavin Shan wrote:
>> >> Commit cca87d30 ("powerpc/pci: Refactor pci_dn") introduced pdn
>> >> list for SRIOV VFs. It means the pdn is be put into the child list
>> >> of its parent pdn when the pdn is created. When doing PCI hot
>> >> unplugging on pSeries, the PCI device node as well as its pdn are
>> >> released through procfs entry "powerpc/ofdt". Some one else grabs
>> >> the memory chunk of the pdn and update it accordingly. At the same
>> >> time, the pdn is still tracked in the child list of parent pdn. It
>> >> leads to corrupted child list in the parent pdn.
>> >> 
>> >> This fixes above issue by removing the pdn from the child list of
>> >> its parent pdn when the device node is detached from the system.
>> >> Note the pdn is free'd when the device node is released if the
>> >> device node is dynamic one. Otherwise, the device node as well
>> >> as the pdn won't be released.
>> >> 
>> >> Fixes: cca87d30 ("powerpc/pci: Refactor pci_dn")
>> >
>> >That went into 4.1, so this should presumably go to stable?
>> 
>> Yeah, This should be put into 4.1 stable.
>
>OK thanks.
>
>I've added:
>
>  Cc: stable@vger.kernel.org # 4.1
>

Thanks, Michael :-)

>cheers
>
>

Re: [1/2] powerpc/pseries: Fix corrupted pdn list

[Michael Ellerman]

On Thu, 2015-27-08 at 04:12:36 UTC, Gavin Shan wrote:
> Commit cca87d30 ("powerpc/pci: Refactor pci_dn") introduced pdn
> list for SRIOV VFs. It means the pdn is be put into the child list
> of its parent pdn when the pdn is created. When doing PCI hot
> unplugging on pSeries, the PCI device node as well as its pdn are
> released through procfs entry "powerpc/ofdt". Some one else grabs
> the memory chunk of the pdn and update it accordingly. At the same
> time, the pdn is still tracked in the child list of parent pdn. It
> leads to corrupted child list in the parent pdn.
> 
> This fixes above issue by removing the pdn from the child list of
> its parent pdn when the device node is detached from the system.
> Note the pdn is free'd when the device node is released if the
> device node is dynamic one. Otherwise, the device node as well
> as the pdn won't be released.
> 
> Fixes: cca87d30 ("powerpc/pci: Refactor pci_dn")
> Reported-by: Santwana Samantray <santwana.samantray@in.ibm.com>
> Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>

Applied to powerpc next, thanks.

https://git.kernel.org/powerpc/c/590c7567a2895f939525ead5

cheers

Re: [2/2] powerpc/pseries: Cleanup on pci_dn_reconfig_notifier()

[Michael Ellerman]

On Thu, 2015-27-08 at 04:12:37 UTC, Gavin Shan wrote:
> This applies cleanup on pci_dn_reconfig_notifier(), no functional
> changes:
> 
>    * Rename variable "pci" to "pdn" to indicate its purpose clearly.
>    * The parent node can be released at any time. So it should be
>      hold with of_get_parent() before accessing it.
>    * The device node doesn't have to have parent node in theory.
>      More check on this.
> 
> Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>

Applied to powerpc next, thanks.

https://git.kernel.org/powerpc/c/ea0f8acf4d44727f7d3a3807

cheers