diff mbox

[v2,1/2] pci: Call pcibios_sriov_enable() before IOV BARs are enabled

Message ID 1475192870-7763-1-git-send-email-gwshan@linux.vnet.ibm.com (mailing list archive)
State Superseded
Headers show

Commit Message

Gavin Shan Sept. 29, 2016, 11:47 p.m. UTC
In current implementation, pcibios_sriov_enable() is used by PPC
PowerNV platform only. In PowerNV specific pcibios_sriov_enable(),
PF's IOV BARs might be updated (shifted) by pci_update_resource().
It means the IOV BARs aren't ready for decoding incoming memory
address until pcibios_sriov_enable() returns.

This calls pcibios_sriov_enable() earlier before the IOV BARs are
enabled. As the result, the IOV BARs have been configured correctly
when they are enabled.

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Tested-by: Carol Soto <clsoto@us.ibm.com>
---
 drivers/pci/iov.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

Comments

Gavin Shan Oct. 11, 2016, 5:33 a.m. UTC | #1
On Fri, Sep 30, 2016 at 09:47:49AM +1000, Gavin Shan wrote:
>In current implementation, pcibios_sriov_enable() is used by PPC
>PowerNV platform only. In PowerNV specific pcibios_sriov_enable(),
>PF's IOV BARs might be updated (shifted) by pci_update_resource().
>It means the IOV BARs aren't ready for decoding incoming memory
>address until pcibios_sriov_enable() returns.
>
>This calls pcibios_sriov_enable() earlier before the IOV BARs are
>enabled. As the result, the IOV BARs have been configured correctly
>when they are enabled.
>
>Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
>Tested-by: Carol Soto <clsoto@us.ibm.com>

Bjorn, is there any chance to put those two patches to 4.9? :)

Thanks,
Gavin

>---
> drivers/pci/iov.c | 14 +++++++-------
> 1 file changed, 7 insertions(+), 7 deletions(-)
>
>diff --git a/drivers/pci/iov.c b/drivers/pci/iov.c
>index 2194b44..f1343f0 100644
>--- a/drivers/pci/iov.c
>+++ b/drivers/pci/iov.c
>@@ -303,13 +303,6 @@ static int sriov_enable(struct pci_dev *dev, int nr_virtfn)
> 			return rc;
> 	}
>
>-	pci_iov_set_numvfs(dev, nr_virtfn);
>-	iov->ctrl |= PCI_SRIOV_CTRL_VFE | PCI_SRIOV_CTRL_MSE;
>-	pci_cfg_access_lock(dev);
>-	pci_write_config_word(dev, iov->pos + PCI_SRIOV_CTRL, iov->ctrl);
>-	msleep(100);
>-	pci_cfg_access_unlock(dev);
>-
> 	iov->initial_VFs = initial;
> 	if (nr_virtfn < initial)
> 		initial = nr_virtfn;
>@@ -320,6 +313,13 @@ static int sriov_enable(struct pci_dev *dev, int nr_virtfn)
> 		goto err_pcibios;
> 	}
>
>+	pci_iov_set_numvfs(dev, nr_virtfn);
>+	iov->ctrl |= PCI_SRIOV_CTRL_VFE | PCI_SRIOV_CTRL_MSE;
>+	pci_cfg_access_lock(dev);
>+	pci_write_config_word(dev, iov->pos + PCI_SRIOV_CTRL, iov->ctrl);
>+	msleep(100);
>+	pci_cfg_access_unlock(dev);
>+
> 	for (i = 0; i < initial; i++) {
> 		rc = pci_iov_add_virtfn(dev, i, 0);
> 		if (rc)
>-- 
>2.1.0
>
Bjorn Helgaas Oct. 21, 2016, 8:23 p.m. UTC | #2
On Fri, Sep 30, 2016 at 09:47:49AM +1000, Gavin Shan wrote:
> In current implementation, pcibios_sriov_enable() is used by PPC
> PowerNV platform only. In PowerNV specific pcibios_sriov_enable(),
> PF's IOV BARs might be updated (shifted) by pci_update_resource().
> It means the IOV BARs aren't ready for decoding incoming memory
> address until pcibios_sriov_enable() returns.
> 
> This calls pcibios_sriov_enable() earlier before the IOV BARs are
> enabled. As the result, the IOV BARs have been configured correctly
> when they are enabled.
> 
> Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
> Tested-by: Carol Soto <clsoto@us.ibm.com>
> ---
>  drivers/pci/iov.c | 14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/pci/iov.c b/drivers/pci/iov.c
> index 2194b44..f1343f0 100644
> --- a/drivers/pci/iov.c
> +++ b/drivers/pci/iov.c
> @@ -303,13 +303,6 @@ static int sriov_enable(struct pci_dev *dev, int nr_virtfn)
>  			return rc;
>  	}
>  
> -	pci_iov_set_numvfs(dev, nr_virtfn);
> -	iov->ctrl |= PCI_SRIOV_CTRL_VFE | PCI_SRIOV_CTRL_MSE;
> -	pci_cfg_access_lock(dev);
> -	pci_write_config_word(dev, iov->pos + PCI_SRIOV_CTRL, iov->ctrl);
> -	msleep(100);
> -	pci_cfg_access_unlock(dev);
> -
>  	iov->initial_VFs = initial;
>  	if (nr_virtfn < initial)
>  		initial = nr_virtfn;
> @@ -320,6 +313,13 @@ static int sriov_enable(struct pci_dev *dev, int nr_virtfn)
>  		goto err_pcibios;
>  	}
>  
> +	pci_iov_set_numvfs(dev, nr_virtfn);
> +	iov->ctrl |= PCI_SRIOV_CTRL_VFE | PCI_SRIOV_CTRL_MSE;
> +	pci_cfg_access_lock(dev);
> +	pci_write_config_word(dev, iov->pos + PCI_SRIOV_CTRL, iov->ctrl);
> +	msleep(100);
> +	pci_cfg_access_unlock(dev);

This one looks fine to me.

>  	for (i = 0; i < initial; i++) {
>  		rc = pci_iov_add_virtfn(dev, i, 0);
>  		if (rc)
> -- 
> 2.1.0
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/pci/iov.c b/drivers/pci/iov.c
index 2194b44..f1343f0 100644
--- a/drivers/pci/iov.c
+++ b/drivers/pci/iov.c
@@ -303,13 +303,6 @@  static int sriov_enable(struct pci_dev *dev, int nr_virtfn)
 			return rc;
 	}
 
-	pci_iov_set_numvfs(dev, nr_virtfn);
-	iov->ctrl |= PCI_SRIOV_CTRL_VFE | PCI_SRIOV_CTRL_MSE;
-	pci_cfg_access_lock(dev);
-	pci_write_config_word(dev, iov->pos + PCI_SRIOV_CTRL, iov->ctrl);
-	msleep(100);
-	pci_cfg_access_unlock(dev);
-
 	iov->initial_VFs = initial;
 	if (nr_virtfn < initial)
 		initial = nr_virtfn;
@@ -320,6 +313,13 @@  static int sriov_enable(struct pci_dev *dev, int nr_virtfn)
 		goto err_pcibios;
 	}
 
+	pci_iov_set_numvfs(dev, nr_virtfn);
+	iov->ctrl |= PCI_SRIOV_CTRL_VFE | PCI_SRIOV_CTRL_MSE;
+	pci_cfg_access_lock(dev);
+	pci_write_config_word(dev, iov->pos + PCI_SRIOV_CTRL, iov->ctrl);
+	msleep(100);
+	pci_cfg_access_unlock(dev);
+
 	for (i = 0; i < initial; i++) {
 		rc = pci_iov_add_virtfn(dev, i, 0);
 		if (rc)