[PATCH] portdrv: Use pci_enable_msix_range() instead of pci_enable_msix()

[PATCH] portdrv: Use pci_enable_msix_range() instead of pci_enable_msix()

[Alexander Gordeev]

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
---
 drivers/pci/pcie/portdrv_core.c |   17 ++++++++++-------
 1 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c
index 986f8ea..4e46eca 100644
--- a/drivers/pci/pcie/portdrv_core.c
+++ b/drivers/pci/pcie/portdrv_core.c
@@ -99,8 +99,9 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
 	for (i = 0; i < nr_entries; i++)
 		msix_entries[i].entry = i;
 
-	status = pci_enable_msix(dev, msix_entries, nr_entries);
-	if (status)
+	status = pci_enable_msix_range(dev,
+				       msix_entries, nr_entries, nr_entries);
+	if (status < 0)
 		goto Exit;
 
 	for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
@@ -159,21 +160,23 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
 		idx[PCIE_PORT_SERVICE_AER_SHIFT] = i;
 	}
 
+	status = 0;
+
 	/*
 	 * If nvec is equal to the allocated number of entries, we can just use
 	 * what we have.  Otherwise, the port has some extra entries not for the
 	 * services we know and we need to work around that.
 	 */
-	if (nvec == nr_entries) {
-		status = 0;
-	} else {
+	if (nvec != nr_entries) {
 		/* Drop the temporary MSI-X setup */
 		pci_disable_msix(dev);
 
 		/* Now allocate the MSI-X vectors for real */
-		status = pci_enable_msix(dev, msix_entries, nvec);
-		if (status)
+		nvec = pci_enable_msix_range(dev, msix_entries, nvec, nvec);
+		if (nvec < 0) {
+			status = nvec;
 			goto Exit;
+		}
 	}
 
 	for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
-- 
1.7.7.6

Re: [PATCH] portdrv: Use pci_enable_msix_range() instead of pci_enable_msix()

[Bjorn Helgaas]

On Wed, Feb 19, 2014 at 3:15 AM, Alexander Gordeev <agordeev@redhat.com> wrote:
> As result of deprecation of MSI-X/MSI enablement functions
> pci_enable_msix() and pci_enable_msi_block() all drivers
> using these two interfaces need to be updated to use the
> new pci_enable_msi_range() and pci_enable_msix_range()
> interfaces.
>
> Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
> Cc: Bjorn Helgaas <bhelgaas@google.com>
> Cc: linux-pci@vger.kernel.org
> ---
>  drivers/pci/pcie/portdrv_core.c |   17 ++++++++++-------
>  1 files changed, 10 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c
> index 986f8ea..4e46eca 100644
> --- a/drivers/pci/pcie/portdrv_core.c
> +++ b/drivers/pci/pcie/portdrv_core.c
> @@ -99,8 +99,9 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
>         for (i = 0; i < nr_entries; i++)
>                 msix_entries[i].entry = i;
>
> -       status = pci_enable_msix(dev, msix_entries, nr_entries);
> -       if (status)
> +       status = pci_enable_msix_range(dev,
> +                                      msix_entries, nr_entries, nr_entries);

This is a place we could use pci_enable_msix_exact(), isn't it?  (Also below.)

> +       if (status < 0)
>                 goto Exit;
>
>         for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
> @@ -159,21 +160,23 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
>                 idx[PCIE_PORT_SERVICE_AER_SHIFT] = i;
>         }
>
> +       status = 0;
> +
>         /*
>          * If nvec is equal to the allocated number of entries, we can just use
>          * what we have.  Otherwise, the port has some extra entries not for the
>          * services we know and we need to work around that.
>          */
> -       if (nvec == nr_entries) {
> -               status = 0;
> -       } else {
> +       if (nvec != nr_entries) {
>                 /* Drop the temporary MSI-X setup */
>                 pci_disable_msix(dev);
>
>                 /* Now allocate the MSI-X vectors for real */
> -               status = pci_enable_msix(dev, msix_entries, nvec);
> -               if (status)
> +               nvec = pci_enable_msix_range(dev, msix_entries, nvec, nvec);
> +               if (nvec < 0) {
> +                       status = nvec;
>                         goto Exit;
> +               }
>         }
>
>         for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
> --
> 1.7.7.6
>

Re: [PATCH] portdrv: Use pci_enable_msix_range() instead of pci_enable_msix()

[Alexander Gordeev]

On Wed, Feb 19, 2014 at 03:09:41PM -0700, Bjorn Helgaas wrote:
> On Wed, Feb 19, 2014 at 3:15 AM, Alexander Gordeev <agordeev@redhat.com> wrote:
> > As result of deprecation of MSI-X/MSI enablement functions
> > pci_enable_msix() and pci_enable_msi_block() all drivers
> > using these two interfaces need to be updated to use the
> > new pci_enable_msi_range() and pci_enable_msix_range()
> > interfaces.
> >
> > Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
> > Cc: Bjorn Helgaas <bhelgaas@google.com>
> > Cc: linux-pci@vger.kernel.org
> > ---
> >  drivers/pci/pcie/portdrv_core.c |   17 ++++++++++-------
> >  1 files changed, 10 insertions(+), 7 deletions(-)
> >
> > diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c
> > index 986f8ea..4e46eca 100644
> > --- a/drivers/pci/pcie/portdrv_core.c
> > +++ b/drivers/pci/pcie/portdrv_core.c
> > @@ -99,8 +99,9 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
> >         for (i = 0; i < nr_entries; i++)
> >                 msix_entries[i].entry = i;
> >
> > -       status = pci_enable_msix(dev, msix_entries, nr_entries);
> > -       if (status)
> > +       status = pci_enable_msix_range(dev,
> > +                                      msix_entries, nr_entries, nr_entries);
> 
> This is a place we could use pci_enable_msix_exact(), isn't it?  (Also below.)

Yes, absolutely.

I posted it (and other updates) as is, since I understood
pci_enable_msix_exact() is not going to make it into 3.14. Was I wrong here?

I think more time in testing with either pci_enable_msix_range() or
pci_enable_msix_exact() is good. And a follow-up switch from
pci_enable_msix_range() to pci_enable_msix_exact() should be easy eigher.

What do you think is the best course of action now? 

> > +       if (status < 0)
> >                 goto Exit;
> >
> >         for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
> > @@ -159,21 +160,23 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
> >                 idx[PCIE_PORT_SERVICE_AER_SHIFT] = i;
> >         }
> >
> > +       status = 0;
> > +
> >         /*
> >          * If nvec is equal to the allocated number of entries, we can just use
> >          * what we have.  Otherwise, the port has some extra entries not for the
> >          * services we know and we need to work around that.
> >          */
> > -       if (nvec == nr_entries) {
> > -               status = 0;
> > -       } else {
> > +       if (nvec != nr_entries) {
> >                 /* Drop the temporary MSI-X setup */
> >                 pci_disable_msix(dev);
> >
> >                 /* Now allocate the MSI-X vectors for real */
> > -               status = pci_enable_msix(dev, msix_entries, nvec);
> > -               if (status)
> > +               nvec = pci_enable_msix_range(dev, msix_entries, nvec, nvec);
> > +               if (nvec < 0) {
> > +                       status = nvec;
> >                         goto Exit;
> > +               }
> >         }
> >
> >         for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
> > --
> > 1.7.7.6
> >
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

-- 
Regards,
Alexander Gordeev
agordeev@redhat.com

Re: [PATCH] portdrv: Use pci_enable_msix_range() instead of pci_enable_msix()

[Bjorn Helgaas]

On Thu, Feb 20, 2014 at 5:44 AM, Alexander Gordeev <agordeev@redhat.com> wrote:
> On Wed, Feb 19, 2014 at 03:09:41PM -0700, Bjorn Helgaas wrote:
>> On Wed, Feb 19, 2014 at 3:15 AM, Alexander Gordeev <agordeev@redhat.com> wrote:
>> > As result of deprecation of MSI-X/MSI enablement functions
>> > pci_enable_msix() and pci_enable_msi_block() all drivers
>> > using these two interfaces need to be updated to use the
>> > new pci_enable_msi_range() and pci_enable_msix_range()
>> > interfaces.
>> >
>> > Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
>> > Cc: Bjorn Helgaas <bhelgaas@google.com>
>> > Cc: linux-pci@vger.kernel.org
>> > ---
>> >  drivers/pci/pcie/portdrv_core.c |   17 ++++++++++-------
>> >  1 files changed, 10 insertions(+), 7 deletions(-)
>> >
>> > diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c
>> > index 986f8ea..4e46eca 100644
>> > --- a/drivers/pci/pcie/portdrv_core.c
>> > +++ b/drivers/pci/pcie/portdrv_core.c
>> > @@ -99,8 +99,9 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
>> >         for (i = 0; i < nr_entries; i++)
>> >                 msix_entries[i].entry = i;
>> >
>> > -       status = pci_enable_msix(dev, msix_entries, nr_entries);
>> > -       if (status)
>> > +       status = pci_enable_msix_range(dev,
>> > +                                      msix_entries, nr_entries, nr_entries);
>>
>> This is a place we could use pci_enable_msix_exact(), isn't it?  (Also below.)
>
> Yes, absolutely.
>
> I posted it (and other updates) as is, since I understood
> pci_enable_msix_exact() is not going to make it into 3.14. Was I wrong here?

Ooh, sorry, I do plan to merge pci_enable_msix_exact() and some of the
other infrastructure patches for v3.14.  They've been in my for-linus
branch for a week or so, and I plan to ask Linus to pull them today.
Sorry I didn't make that more clear.

> I think more time in testing with either pci_enable_msix_range() or
> pci_enable_msix_exact() is good. And a follow-up switch from
> pci_enable_msix_range() to pci_enable_msix_exact() should be easy eigher.
>
> What do you think is the best course of action now?

You should be able to post patches that use pci_enable_msix_exact();
that's been in -next for a while now.  Maybe include a cover note
saying what they depend on.

Bjorn

>> > +       if (status < 0)
>> >                 goto Exit;
>> >
>> >         for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
>> > @@ -159,21 +160,23 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
>> >                 idx[PCIE_PORT_SERVICE_AER_SHIFT] = i;
>> >         }
>> >
>> > +       status = 0;
>> > +
>> >         /*
>> >          * If nvec is equal to the allocated number of entries, we can just use
>> >          * what we have.  Otherwise, the port has some extra entries not for the
>> >          * services we know and we need to work around that.
>> >          */
>> > -       if (nvec == nr_entries) {
>> > -               status = 0;
>> > -       } else {
>> > +       if (nvec != nr_entries) {
>> >                 /* Drop the temporary MSI-X setup */
>> >                 pci_disable_msix(dev);
>> >
>> >                 /* Now allocate the MSI-X vectors for real */
>> > -               status = pci_enable_msix(dev, msix_entries, nvec);
>> > -               if (status)
>> > +               nvec = pci_enable_msix_range(dev, msix_entries, nvec, nvec);
>> > +               if (nvec < 0) {
>> > +                       status = nvec;
>> >                         goto Exit;
>> > +               }
>> >         }
>> >
>> >         for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
>> > --
>> > 1.7.7.6
>> >
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
>
> --
> Regards,
> Alexander Gordeev
> agordeev@redhat.com

[PATCH v2] portdrv: Use pci_enable_msix_exact() instead of pci_enable_msix()

[Alexander Gordeev]

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range()  or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
---
 drivers/pci/pcie/portdrv_core.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c
index 986f8ea..0b1efb2 100644
--- a/drivers/pci/pcie/portdrv_core.c
+++ b/drivers/pci/pcie/portdrv_core.c
@@ -99,7 +99,7 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
 	for (i = 0; i < nr_entries; i++)
 		msix_entries[i].entry = i;
 
-	status = pci_enable_msix(dev, msix_entries, nr_entries);
+	status = pci_enable_msix_exact(dev, msix_entries, nr_entries);
 	if (status)
 		goto Exit;
 
@@ -171,7 +171,7 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
 		pci_disable_msix(dev);
 
 		/* Now allocate the MSI-X vectors for real */
-		status = pci_enable_msix(dev, msix_entries, nvec);
+		status = pci_enable_msix_exact(dev, msix_entries, nvec);
 		if (status)
 			goto Exit;
 	}
-- 
1.7.7.6

-- 
Regards,
Alexander Gordeev
agordeev@redhat.com