[PATCH] pci: allow multiple calls to pcim_enable_device()

[PATCH] pci: allow multiple calls to pcim_enable_device()

[Tejun Heo]

There's no reason not to allow multiple calls to pcim_enable_device().
Calls after the first one can simply be noop.  All PCI resources will
be released when the initial pcim_enable_device() resource is
released.

This allows more flexibility to managed PCI users.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Greg Kroah-Hartman <gregkh@suse.de>
---
This will be used to fix some libata issues.  After Greg's ACK, would
it be okay to push this through libata-dev#upstream?

Thanks.

 drivers/pci/pci.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index 71d561f..7d4ce90 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -823,7 +823,8 @@ int pcim_enable_device(struct pci_dev *pdev)
 	dr = get_pci_dr(pdev);
 	if (unlikely(!dr))
 		return -ENOMEM;
-	WARN_ON(!!dr->enabled);
+	if (dr->enabled)
+		return 0;
 
 	rc = pci_enable_device(pdev);
 	if (!rc) {
-- 
1.5.2.4

Re: [PATCH] pci: allow multiple calls to pcim_enable_device()

[Greg KH]

On Wed, Jan 30, 2008 at 06:20:04PM +0900, Tejun Heo wrote:
> There's no reason not to allow multiple calls to pcim_enable_device().
> Calls after the first one can simply be noop.  All PCI resources will
> be released when the initial pcim_enable_device() resource is
> released.
> 
> This allows more flexibility to managed PCI users.
> 
> Signed-off-by: Tejun Heo <htejun@gmail.com>
> Cc: Greg Kroah-Hartman <gregkh@suse.de>
> ---
> This will be used to fix some libata issues.  After Greg's ACK, would
> it be okay to push this through libata-dev#upstream?

That's fine with me, please add:
	Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

To the patch if you want to send it through the ata tree.

thanks,

greg k-h

Re: [PATCH] pci: allow multiple calls to pcim_enable_device()

[Jeff Garzik]

Tejun Heo wrote:
> There's no reason not to allow multiple calls to pcim_enable_device().
> Calls after the first one can simply be noop.  All PCI resources will
> be released when the initial pcim_enable_device() resource is
> released.
> 
> This allows more flexibility to managed PCI users.
> 
> Signed-off-by: Tejun Heo <htejun@gmail.com>
> Cc: Greg Kroah-Hartman <gregkh@suse.de>
> ---
> This will be used to fix some libata issues.  After Greg's ACK, would
> it be okay to push this through libata-dev#upstream?
> 
> Thanks.
> 
>  drivers/pci/pci.c |    3 ++-
>  1 files changed, 2 insertions(+), 1 deletions(-)

applied, w/ greg's sign-off