[PATCH v1] PCI: PM: Power up all devices during runtime resume

[PATCH v1] PCI: PM: Power up all devices during runtime resume

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Currently, endpoint devices may not be powered up entirely during
runtime resume that follows a D3hot -> D0 transition of the parent
bridge.

Namely, even if the power state of an endpoint device, as indicated
by its PCI_PM_CTRL register, is D0 after powering up its parent
bridge, it may be still necessary to bring its ACPI companion into
D0 and that should be done before accessing it.  However, the current
code assumes that reading the PCI_PM_CTRL register is sufficient to
establish the endpoint device's power state, which may lead to
problems.

Address that by forcing a power-up of all PCI devices, including the
platform firmware part of it, during runtime resume.

Link: https://lore.kernel.org/linux-pm/11967527.O9o76ZdvQC@kreacher
Fixes: 5775b843a619 ("PCI: Restore config space on runtime resume despite being unbound")
Reported-by: Abhishek Sahu <abhsahu@nvidia.com>
Tested-by: Abhishek Sahu <abhsahu@nvidia.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---
 drivers/pci/pci-driver.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: linux-pm/drivers/pci/pci-driver.c
===================================================================
--- linux-pm.orig/drivers/pci/pci-driver.c
+++ linux-pm/drivers/pci/pci-driver.c
@@ -1312,7 +1312,7 @@ static int pci_pm_runtime_resume(struct
 	 * to a driver because although we left it in D0, it may have gone to
 	 * D3cold when the bridge above it runtime suspended.
 	 */
-	pci_restore_standard_config(pci_dev);
+	pci_pm_default_resume_early(pci_dev);
 
 	if (!pci_dev->driver)
 		return 0;

Re: [PATCH v1] PCI: PM: Power up all devices during runtime resume

[Mika Westerberg]

On Wed, Apr 06, 2022 at 09:00:52PM +0200, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Currently, endpoint devices may not be powered up entirely during
> runtime resume that follows a D3hot -> D0 transition of the parent
> bridge.
> 
> Namely, even if the power state of an endpoint device, as indicated
> by its PCI_PM_CTRL register, is D0 after powering up its parent
> bridge, it may be still necessary to bring its ACPI companion into
> D0 and that should be done before accessing it.  However, the current
> code assumes that reading the PCI_PM_CTRL register is sufficient to
> establish the endpoint device's power state, which may lead to
> problems.
> 
> Address that by forcing a power-up of all PCI devices, including the
> platform firmware part of it, during runtime resume.
> 
> Link: https://lore.kernel.org/linux-pm/11967527.O9o76ZdvQC@kreacher
> Fixes: 5775b843a619 ("PCI: Restore config space on runtime resume despite being unbound")
> Reported-by: Abhishek Sahu <abhsahu@nvidia.com>
> Tested-by: Abhishek Sahu <abhsahu@nvidia.com>
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>

Re: [PATCH v1] PCI: PM: Power up all devices during runtime resume

[Bjorn Helgaas]

On Wed, Apr 06, 2022 at 09:00:52PM +0200, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Currently, endpoint devices may not be powered up entirely during
> runtime resume that follows a D3hot -> D0 transition of the parent
> bridge.
> 
> Namely, even if the power state of an endpoint device, as indicated
> by its PCI_PM_CTRL register, is D0 after powering up its parent
> bridge, it may be still necessary to bring its ACPI companion into
> D0 and that should be done before accessing it.  However, the current
> code assumes that reading the PCI_PM_CTRL register is sufficient to
> establish the endpoint device's power state, which may lead to
> problems.
> 
> Address that by forcing a power-up of all PCI devices, including the
> platform firmware part of it, during runtime resume.
> 
> Link: https://lore.kernel.org/linux-pm/11967527.O9o76ZdvQC@kreacher
> Fixes: 5775b843a619 ("PCI: Restore config space on runtime resume despite being unbound")
> Reported-by: Abhishek Sahu <abhsahu@nvidia.com>
> Tested-by: Abhishek Sahu <abhsahu@nvidia.com>
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Applied with Mika's reviewed-by to pci/pm for v5.19, thanks!

> ---
>  drivers/pci/pci-driver.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> Index: linux-pm/drivers/pci/pci-driver.c
> ===================================================================
> --- linux-pm.orig/drivers/pci/pci-driver.c
> +++ linux-pm/drivers/pci/pci-driver.c
> @@ -1312,7 +1312,7 @@ static int pci_pm_runtime_resume(struct
>  	 * to a driver because although we left it in D0, it may have gone to
>  	 * D3cold when the bridge above it runtime suspended.
>  	 */
> -	pci_restore_standard_config(pci_dev);
> +	pci_pm_default_resume_early(pci_dev);
>  
>  	if (!pci_dev->driver)
>  		return 0;
> 
> 
> 

Re: [PATCH v1] PCI: PM: Power up all devices during runtime resume

[Rafael J. Wysocki]

On Thu, Apr 7, 2022 at 5:49 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
>
> On Wed, Apr 06, 2022 at 09:00:52PM +0200, Rafael J. Wysocki wrote:
> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> >
> > Currently, endpoint devices may not be powered up entirely during
> > runtime resume that follows a D3hot -> D0 transition of the parent
> > bridge.
> >
> > Namely, even if the power state of an endpoint device, as indicated
> > by its PCI_PM_CTRL register, is D0 after powering up its parent
> > bridge, it may be still necessary to bring its ACPI companion into
> > D0 and that should be done before accessing it.  However, the current
> > code assumes that reading the PCI_PM_CTRL register is sufficient to
> > establish the endpoint device's power state, which may lead to
> > problems.
> >
> > Address that by forcing a power-up of all PCI devices, including the
> > platform firmware part of it, during runtime resume.
> >
> > Link: https://lore.kernel.org/linux-pm/11967527.O9o76ZdvQC@kreacher
> > Fixes: 5775b843a619 ("PCI: Restore config space on runtime resume despite being unbound")
> > Reported-by: Abhishek Sahu <abhsahu@nvidia.com>
> > Tested-by: Abhishek Sahu <abhsahu@nvidia.com>
> > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>
> Applied with Mika's reviewed-by to pci/pm for v5.19, thanks!

This doesn't compile when CONFIG_PM_SLEEP in uset, sorry about this.

I'll send a v2, but if you prefer an incremental fix, please let me know.

>
> > ---
> >  drivers/pci/pci-driver.c |    2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > Index: linux-pm/drivers/pci/pci-driver.c
> > ===================================================================
> > --- linux-pm.orig/drivers/pci/pci-driver.c
> > +++ linux-pm/drivers/pci/pci-driver.c
> > @@ -1312,7 +1312,7 @@ static int pci_pm_runtime_resume(struct
> >        * to a driver because although we left it in D0, it may have gone to
> >        * D3cold when the bridge above it runtime suspended.
> >        */
> > -     pci_restore_standard_config(pci_dev);
> > +     pci_pm_default_resume_early(pci_dev);
> >
> >       if (!pci_dev->driver)
> >               return 0;
> >
> >
> >

[PATCH v2] PCI: PM: Power up all devices during runtime resume

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Subject: [PATCH] PCI: PM: Power up all devices during runtime resume

Currently, endpoint devices may not be powered up entirely during
runtime resume that follows a D3hot -> D0 transition of the parent
bridge.

Namely, even if the power state of an endpoint device, as indicated
by its PCI_PM_CTRL register, is D0 after powering up its parent
bridge, it may be still necessary to bring its ACPI companion into
D0 and that should be done before accessing it.  However, the current
code assumes that reading the PCI_PM_CTRL register is sufficient to
establish the endpoint device's power state, which may lead to
problems.

Address that by forcing a power-up of all PCI devices, including the
platform firmware part of it, during runtime resume.

Link: https://lore.kernel.org/linux-pm/11967527.O9o76ZdvQC@kreacher
Fixes: 5775b843a619 ("PCI: Restore config space on runtime resume despite being unbound")
Reported-by: Abhishek Sahu <abhsahu@nvidia.com>
Tested-by: Abhishek Sahu <abhsahu@nvidia.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---

v1 -> v2:
   * Move pci_pm_default_resume_early() away from #ifdef CONFIG_PM_SLEEP.
   * Add R-by from Mika.

---
 drivers/pci/pci-driver.c |   10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

Index: linux-pm/drivers/pci/pci-driver.c
===================================================================
--- linux-pm.orig/drivers/pci/pci-driver.c
+++ linux-pm/drivers/pci/pci-driver.c
@@ -551,10 +551,6 @@ static void pci_pm_default_resume(struct
 	pci_enable_wake(pci_dev, PCI_D0, false);
 }
 
-#endif
-
-#ifdef CONFIG_PM_SLEEP
-
 static void pci_pm_default_resume_early(struct pci_dev *pci_dev)
 {
 	pci_power_up(pci_dev);
@@ -563,6 +559,10 @@ static void pci_pm_default_resume_early(
 	pci_pme_restore(pci_dev);
 }
 
+#endif
+
+#ifdef CONFIG_PM_SLEEP
+
 /*
  * Default "suspend" method for devices that have no driver provided suspend,
  * or not even a driver at all (second part).
@@ -1312,7 +1312,7 @@ static int pci_pm_runtime_resume(struct
 	 * to a driver because although we left it in D0, it may have gone to
 	 * D3cold when the bridge above it runtime suspended.
 	 */
-	pci_restore_standard_config(pci_dev);
+	pci_pm_default_resume_early(pci_dev);
 
 	if (!pci_dev->driver)
 		return 0;

Re: [PATCH v2] PCI: PM: Power up all devices during runtime resume

[Bjorn Helgaas]

On Fri, Apr 08, 2022 at 08:29:01PM +0200, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> Subject: [PATCH] PCI: PM: Power up all devices during runtime resume
> 
> Currently, endpoint devices may not be powered up entirely during
> runtime resume that follows a D3hot -> D0 transition of the parent
> bridge.
> 
> Namely, even if the power state of an endpoint device, as indicated
> by its PCI_PM_CTRL register, is D0 after powering up its parent
> bridge, it may be still necessary to bring its ACPI companion into
> D0 and that should be done before accessing it.  However, the current
> code assumes that reading the PCI_PM_CTRL register is sufficient to
> establish the endpoint device's power state, which may lead to
> problems.
> 
> Address that by forcing a power-up of all PCI devices, including the
> platform firmware part of it, during runtime resume.
> 
> Link: https://lore.kernel.org/linux-pm/11967527.O9o76ZdvQC@kreacher
> Fixes: 5775b843a619 ("PCI: Restore config space on runtime resume despite being unbound")
> Reported-by: Abhishek Sahu <abhsahu@nvidia.com>
> Tested-by: Abhishek Sahu <abhsahu@nvidia.com>
> Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

I replaced the v1 patch with this one on pci/pm, thanks!

> ---
> 
> v1 -> v2:
>    * Move pci_pm_default_resume_early() away from #ifdef CONFIG_PM_SLEEP.
>    * Add R-by from Mika.
> 
> ---
>  drivers/pci/pci-driver.c |   10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> Index: linux-pm/drivers/pci/pci-driver.c
> ===================================================================
> --- linux-pm.orig/drivers/pci/pci-driver.c
> +++ linux-pm/drivers/pci/pci-driver.c
> @@ -551,10 +551,6 @@ static void pci_pm_default_resume(struct
>  	pci_enable_wake(pci_dev, PCI_D0, false);
>  }
>  
> -#endif
> -
> -#ifdef CONFIG_PM_SLEEP
> -
>  static void pci_pm_default_resume_early(struct pci_dev *pci_dev)
>  {
>  	pci_power_up(pci_dev);
> @@ -563,6 +559,10 @@ static void pci_pm_default_resume_early(
>  	pci_pme_restore(pci_dev);
>  }
>  
> +#endif
> +
> +#ifdef CONFIG_PM_SLEEP
> +
>  /*
>   * Default "suspend" method for devices that have no driver provided suspend,
>   * or not even a driver at all (second part).
> @@ -1312,7 +1312,7 @@ static int pci_pm_runtime_resume(struct
>  	 * to a driver because although we left it in D0, it may have gone to
>  	 * D3cold when the bridge above it runtime suspended.
>  	 */
> -	pci_restore_standard_config(pci_dev);
> +	pci_pm_default_resume_early(pci_dev);
>  
>  	if (!pci_dev->driver)
>  		return 0;
> 
> 
>