[PATCH] pci: Ensure we re-enable devices on resume

[PATCH] pci: Ensure we re-enable devices on resume

[Matthew Garrett]

If the firmware puts a device back into D0 state at resume time, we'll
update its state in resume_noirq and thus skip the platform resume code.
Calling that code twice should be safe and we ought to avoid getting to
that point anyway, so remove the check and also allow the platform pci
code to be called for D0.

Fixes USB not being powered after resume on recent Lenovo machines.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
---
 drivers/pci/pci.c |    6 +-----
 1 files changed, 1 insertions(+), 5 deletions(-)

diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index 5ea587e..3749912 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -679,7 +679,7 @@ static void __pci_start_power_transition(struct pci_dev *dev, pci_power_t state)
  */
 int __pci_complete_power_transition(struct pci_dev *dev, pci_power_t state)
 {
-	return state > PCI_D0 ?
+	return state >= PCI_D0 ?
 			pci_platform_power_transition(dev, state) : -EINVAL;
 }
 EXPORT_SYMBOL_GPL(__pci_complete_power_transition);
@@ -716,10 +716,6 @@ int pci_set_power_state(struct pci_dev *dev, pci_power_t state)
 		 */
 		return 0;
 
-	/* Check if we're already there */
-	if (dev->current_state == state)
-		return 0;
-
 	__pci_start_power_transition(dev, state);
 
 	/* This device is quirked not to be put into D3, so
-- 
1.7.0.1

Re: [PATCH] pci: Ensure we re-enable devices on resume

[Rafael J. Wysocki]

On Thursday 22 April 2010, Matthew Garrett wrote:
> If the firmware puts a device back into D0 state at resume time, we'll
> update its state in resume_noirq and thus skip the platform resume code.
> Calling that code twice should be safe and we ought to avoid getting to
> that point anyway, so remove the check and also allow the platform pci
> code to be called for D0.
> 
> Fixes USB not being powered after resume on recent Lenovo machines.

OK ...

> Signed-off-by: Matthew Garrett <mjg@redhat.com>
> ---
>  drivers/pci/pci.c |    6 +-----
>  1 files changed, 1 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 5ea587e..3749912 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -679,7 +679,7 @@ static void __pci_start_power_transition(struct pci_dev *dev, pci_power_t state)
>   */
>  int __pci_complete_power_transition(struct pci_dev *dev, pci_power_t state)
>  {
> -	return state > PCI_D0 ?
> +	return state >= PCI_D0 ?
>  			pci_platform_power_transition(dev, state) : -EINVAL;
>  }
>  EXPORT_SYMBOL_GPL(__pci_complete_power_transition);
> @@ -716,10 +716,6 @@ int pci_set_power_state(struct pci_dev *dev, pci_power_t state)
>  		 */
>  		return 0;
>  
> -	/* Check if we're already there */
> -	if (dev->current_state == state)
> -		return 0;
> -

If my memory serves me right, this check is actually there for a reason,
although I don't remember _what_ exactly the reason is right now.  Sigh.

[/me puts a note for himself to _always_ add specific comments into the code
 in such cases.]

>  	__pci_start_power_transition(dev, state);
>  
>  	/* This device is quirked not to be put into D3, so
> 

Well, I'd say let's try that.  If it breaks things, we'll know relatively quickly. ;-)

Acked-by: Rafael J. Wysocki <rjw@sisk.pl>

Re: [PATCH] pci: Ensure we re-enable devices on resume

[Alex Chiang]

Looks good to me. Thanks for doing this, Matthew.

Acked-by: Alex Chiang <achiang@canonical.com>

/ac

* Matthew Garrett <mjg@redhat.com>:
> If the firmware puts a device back into D0 state at resume time, we'll
> update its state in resume_noirq and thus skip the platform resume code.
> Calling that code twice should be safe and we ought to avoid getting to
> that point anyway, so remove the check and also allow the platform pci
> code to be called for D0.
> 
> Fixes USB not being powered after resume on recent Lenovo machines.
> 
> Signed-off-by: Matthew Garrett <mjg@redhat.com>
> ---
>  drivers/pci/pci.c |    6 +-----
>  1 files changed, 1 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 5ea587e..3749912 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -679,7 +679,7 @@ static void __pci_start_power_transition(struct pci_dev *dev, pci_power_t state)
>   */
>  int __pci_complete_power_transition(struct pci_dev *dev, pci_power_t state)
>  {
> -	return state > PCI_D0 ?
> +	return state >= PCI_D0 ?
>  			pci_platform_power_transition(dev, state) : -EINVAL;
>  }
>  EXPORT_SYMBOL_GPL(__pci_complete_power_transition);
> @@ -716,10 +716,6 @@ int pci_set_power_state(struct pci_dev *dev, pci_power_t state)
>  		 */
>  		return 0;
>  
> -	/* Check if we're already there */
> -	if (dev->current_state == state)
> -		return 0;
> -
>  	__pci_start_power_transition(dev, state);
>  
>  	/* This device is quirked not to be put into D3, so
> -- 
> 1.7.0.1
> 

Re: [PATCH] pci: Ensure we re-enable devices on resume

[Jesse Barnes]

On Thu, 22 Apr 2010 09:30:51 -0400
Matthew Garrett <mjg@redhat.com> wrote:

> If the firmware puts a device back into D0 state at resume time, we'll
> update its state in resume_noirq and thus skip the platform resume code.
> Calling that code twice should be safe and we ought to avoid getting to
> that point anyway, so remove the check and also allow the platform pci
> code to be called for D0.
> 
> Fixes USB not being powered after resume on recent Lenovo machines.
> 
> Signed-off-by: Matthew Garrett <mjg@redhat.com>
> ---

Applied to for-linus, thanks.

-- 
Jesse Barnes, Intel Open Source Technology Center

Re: [PATCH] pci: Ensure we re-enable devices on resume

[Alex Chiang]

* Jesse Barnes <jbarnes@virtuousgeek.org>:
> On Thu, 22 Apr 2010 09:30:51 -0400
> Matthew Garrett <mjg@redhat.com> wrote:
> 
> > If the firmware puts a device back into D0 state at resume time, we'll
> > update its state in resume_noirq and thus skip the platform resume code.
> > Calling that code twice should be safe and we ought to avoid getting to
> > that point anyway, so remove the check and also allow the platform pci
> > code to be called for D0.
> > 
> > Fixes USB not being powered after resume on recent Lenovo machines.
> > 
> > Signed-off-by: Matthew Garrett <mjg@redhat.com>
> 
> Applied to for-linus, thanks.

Thanks Jesse. 

Stable, can you please pick this up too?

The original commit that introduced the thinko was 4a865905.

achiang@aspen:~/kernels/linux-2.6$ git describe --contains 4a865905
v2.6.30-rc1~609^2~4

which looks like it was introduced in 2.6.30, but I'd be happy with
just hitting v2.6.32, since this is what just about every distro
has settled on these days.

Thanks,
/ac

Re: [stable] [PATCH] pci: Ensure we re-enable devices on resume

[Greg KH]

On Thu, Apr 22, 2010 at 05:46:53PM -0600, Alex Chiang wrote:
> * Jesse Barnes <jbarnes@virtuousgeek.org>:
> > On Thu, 22 Apr 2010 09:30:51 -0400
> > Matthew Garrett <mjg@redhat.com> wrote:
> > 
> > > If the firmware puts a device back into D0 state at resume time, we'll
> > > update its state in resume_noirq and thus skip the platform resume code.
> > > Calling that code twice should be safe and we ought to avoid getting to
> > > that point anyway, so remove the check and also allow the platform pci
> > > code to be called for D0.
> > > 
> > > Fixes USB not being powered after resume on recent Lenovo machines.
> > > 
> > > Signed-off-by: Matthew Garrett <mjg@redhat.com>
> > 
> > Applied to for-linus, thanks.
> 
> Thanks Jesse. 
> 
> Stable, can you please pick this up too?

When it goes into Linus's tree, yes.  Care to let me know what the git
commit id of it is when that happens?

thanks,

greg k-h

Re: [stable] [PATCH] pci: Ensure we re-enable devices on resume

[Alex Chiang]

* Greg KH <greg@kroah.com>:
> On Thu, Apr 22, 2010 at 05:46:53PM -0600, Alex Chiang wrote:
> > * Jesse Barnes <jbarnes@virtuousgeek.org>:
> > > Matthew Garrett <mjg@redhat.com> wrote:
> > > 
> > > > If the firmware puts a device back into D0 state at
> > > > resume time, we'll update its state in resume_noirq and
> > > > thus skip the platform resume code. Calling that code
> > > > twice should be safe and we ought to avoid getting to
> > > > that point anyway, so remove the check and also allow the
> > > > platform pci code to be called for D0.
> > > > 
> > > > Fixes USB not being powered after resume on recent Lenovo
> > > > machines.
> > > > 
> > > 
> > > Applied to for-linus, thanks.
> > 
> > Stable, can you please pick this up too?
> 
> When it goes into Linus's tree, yes.  Care to let me know what
> the git commit id of it is when that happens?

commit cc2893b6af5265baa1d68b17b136cffca9e40cfa

Thanks!
/ac

Re: [stable] [PATCH] pci: Ensure we re-enable devices on resume

[Greg KH]

On Sat, Apr 24, 2010 at 10:21:49PM -0600, Alex Chiang wrote:
> * Greg KH <greg@kroah.com>:
> > On Thu, Apr 22, 2010 at 05:46:53PM -0600, Alex Chiang wrote:
> > > * Jesse Barnes <jbarnes@virtuousgeek.org>:
> > > > Matthew Garrett <mjg@redhat.com> wrote:
> > > > 
> > > > > If the firmware puts a device back into D0 state at
> > > > > resume time, we'll update its state in resume_noirq and
> > > > > thus skip the platform resume code. Calling that code
> > > > > twice should be safe and we ought to avoid getting to
> > > > > that point anyway, so remove the check and also allow the
> > > > > platform pci code to be called for D0.
> > > > > 
> > > > > Fixes USB not being powered after resume on recent Lenovo
> > > > > machines.
> > > > > 
> > > > 
> > > > Applied to for-linus, thanks.
> > > 
> > > Stable, can you please pick this up too?
> > 
> > When it goes into Linus's tree, yes.  Care to let me know what
> > the git commit id of it is when that happens?
> 
> commit cc2893b6af5265baa1d68b17b136cffca9e40cfa

Now queued up.

thanks,

greg k-h