* [PATCH] vfio-pci: Fix the check on pci device type in vfio_pci_probe()
@ 2014-12-08 1:22 Wei Yang
2014-12-22 6:12 ` Wei Yang
0 siblings, 1 reply; 6+ messages in thread
From: Wei Yang @ 2014-12-08 1:22 UTC (permalink / raw)
To: alex.williamson, bhelgaas; +Cc: linux-pci, Wei Yang
Current vfio-pci just supports normal pci device, so vfio_pci_probe() will
return if the pci device is not a normal device. While current code makes a
mistake. PCI_HEADER_TYPE is the offset in configuration space of the device
type, but we use this value to mask the type value.
This patch adds a function to check whether the pci device is a normal type
and use it in vfio_pci_probe().
Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
---
drivers/vfio/pci/vfio_pci.c | 4 +---
include/linux/pci.h | 11 +++++++++++
2 files changed, 12 insertions(+), 3 deletions(-)
diff --git a/drivers/vfio/pci/vfio_pci.c b/drivers/vfio/pci/vfio_pci.c
index 9558da3..f82bf62 100644
--- a/drivers/vfio/pci/vfio_pci.c
+++ b/drivers/vfio/pci/vfio_pci.c
@@ -839,13 +839,11 @@ static const struct vfio_device_ops vfio_pci_ops = {
static int vfio_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
{
- u8 type;
struct vfio_pci_device *vdev;
struct iommu_group *group;
int ret;
- pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
- if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
+ if (!pci_is_normal(pdev))
return -EINVAL;
group = iommu_group_get(&pdev->dev);
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 48d1f98..2027f66 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -519,6 +519,17 @@ static inline bool pci_is_bridge(struct pci_dev *dev)
dev->hdr_type == PCI_HEADER_TYPE_CARDBUS;
}
+/**
+ * pci_is_normal - check if the PCI device is a normal device
+ * @dev: PCI device
+ *
+ * Return true if the PCI device is a normal device
+ */
+static inline bool pci_is_normal(struct pci_dev *dev)
+{
+ return dev->hdr_type == PCI_HEADER_TYPE_NORMAL;
+}
+
static inline struct pci_dev *pci_upstream_bridge(struct pci_dev *dev)
{
dev = pci_physfn(dev);
--
1.7.9.5
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH] vfio-pci: Fix the check on pci device type in vfio_pci_probe()
2014-12-08 1:22 [PATCH] vfio-pci: Fix the check on pci device type in vfio_pci_probe() Wei Yang
@ 2014-12-22 6:12 ` Wei Yang
2014-12-22 23:38 ` Alex Williamson
0 siblings, 1 reply; 6+ messages in thread
From: Wei Yang @ 2014-12-22 6:12 UTC (permalink / raw)
To: alex.williamson; +Cc: bhelgaas, linux-pci
Alex,
Does it look good to you?
On Mon, Dec 08, 2014 at 09:22:18AM +0800, Wei Yang wrote:
>Current vfio-pci just supports normal pci device, so vfio_pci_probe() will
>return if the pci device is not a normal device. While current code makes a
>mistake. PCI_HEADER_TYPE is the offset in configuration space of the device
>type, but we use this value to mask the type value.
>
>This patch adds a function to check whether the pci device is a normal type
>and use it in vfio_pci_probe().
>
>Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
>---
> drivers/vfio/pci/vfio_pci.c | 4 +---
> include/linux/pci.h | 11 +++++++++++
> 2 files changed, 12 insertions(+), 3 deletions(-)
>
>diff --git a/drivers/vfio/pci/vfio_pci.c b/drivers/vfio/pci/vfio_pci.c
>index 9558da3..f82bf62 100644
>--- a/drivers/vfio/pci/vfio_pci.c
>+++ b/drivers/vfio/pci/vfio_pci.c
>@@ -839,13 +839,11 @@ static const struct vfio_device_ops vfio_pci_ops = {
>
> static int vfio_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
> {
>- u8 type;
> struct vfio_pci_device *vdev;
> struct iommu_group *group;
> int ret;
>
>- pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
>- if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
>+ if (!pci_is_normal(pdev))
> return -EINVAL;
>
> group = iommu_group_get(&pdev->dev);
>diff --git a/include/linux/pci.h b/include/linux/pci.h
>index 48d1f98..2027f66 100644
>--- a/include/linux/pci.h
>+++ b/include/linux/pci.h
>@@ -519,6 +519,17 @@ static inline bool pci_is_bridge(struct pci_dev *dev)
> dev->hdr_type == PCI_HEADER_TYPE_CARDBUS;
> }
>
>+/**
>+ * pci_is_normal - check if the PCI device is a normal device
>+ * @dev: PCI device
>+ *
>+ * Return true if the PCI device is a normal device
>+ */
>+static inline bool pci_is_normal(struct pci_dev *dev)
>+{
>+ return dev->hdr_type == PCI_HEADER_TYPE_NORMAL;
>+}
>+
> static inline struct pci_dev *pci_upstream_bridge(struct pci_dev *dev)
> {
> dev = pci_physfn(dev);
>--
>1.7.9.5
--
Richard Yang
Help you, Help me
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] vfio-pci: Fix the check on pci device type in vfio_pci_probe()
2014-12-22 6:12 ` Wei Yang
@ 2014-12-22 23:38 ` Alex Williamson
2014-12-23 1:16 ` Wei Yang
0 siblings, 1 reply; 6+ messages in thread
From: Alex Williamson @ 2014-12-22 23:38 UTC (permalink / raw)
To: Wei Yang; +Cc: bhelgaas, linux-pci
On Mon, 2014-12-22 at 14:12 +0800, Wei Yang wrote:
> Alex,
>
> Does it look good to you?
Well, mixing changes to a subsystem and a driver in a single patch is
surely going to slow down acceptance because neither Bjorn or I know who
is going to grab it. I'd probably want to mark the vfio-pci change for
stable, which would be a lot easier to do if it was just a 4-line
correction to the existing code rather than extending the pci-core
interface for a single, trivial user. Thanks,
Alex
On Mon, Dec 08, 2014 at 09:22:18AM +0800, Wei Yang wrote:
> >Current vfio-pci just supports normal pci device, so vfio_pci_probe() will
> >return if the pci device is not a normal device. While current code makes a
> >mistake. PCI_HEADER_TYPE is the offset in configuration space of the device
> >type, but we use this value to mask the type value.
> >
> >This patch adds a function to check whether the pci device is a normal type
> >and use it in vfio_pci_probe().
> >
> >Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
> >---
> > drivers/vfio/pci/vfio_pci.c | 4 +---
> > include/linux/pci.h | 11 +++++++++++
> > 2 files changed, 12 insertions(+), 3 deletions(-)
> >
> >diff --git a/drivers/vfio/pci/vfio_pci.c b/drivers/vfio/pci/vfio_pci.c
> >index 9558da3..f82bf62 100644
> >--- a/drivers/vfio/pci/vfio_pci.c
> >+++ b/drivers/vfio/pci/vfio_pci.c
> >@@ -839,13 +839,11 @@ static const struct vfio_device_ops vfio_pci_ops = {
> >
> > static int vfio_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
> > {
> >- u8 type;
> > struct vfio_pci_device *vdev;
> > struct iommu_group *group;
> > int ret;
> >
> >- pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
> >- if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
> >+ if (!pci_is_normal(pdev))
> > return -EINVAL;
> >
> > group = iommu_group_get(&pdev->dev);
> >diff --git a/include/linux/pci.h b/include/linux/pci.h
> >index 48d1f98..2027f66 100644
> >--- a/include/linux/pci.h
> >+++ b/include/linux/pci.h
> >@@ -519,6 +519,17 @@ static inline bool pci_is_bridge(struct pci_dev *dev)
> > dev->hdr_type == PCI_HEADER_TYPE_CARDBUS;
> > }
> >
> >+/**
> >+ * pci_is_normal - check if the PCI device is a normal device
> >+ * @dev: PCI device
> >+ *
> >+ * Return true if the PCI device is a normal device
> >+ */
> >+static inline bool pci_is_normal(struct pci_dev *dev)
> >+{
> >+ return dev->hdr_type == PCI_HEADER_TYPE_NORMAL;
> >+}
> >+
> > static inline struct pci_dev *pci_upstream_bridge(struct pci_dev *dev)
> > {
> > dev = pci_physfn(dev);
> >--
> >1.7.9.5
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] vfio-pci: Fix the check on pci device type in vfio_pci_probe()
2014-12-22 23:38 ` Alex Williamson
@ 2014-12-23 1:16 ` Wei Yang
2014-12-23 2:58 ` [PATCH V2] " Wei Yang
0 siblings, 1 reply; 6+ messages in thread
From: Wei Yang @ 2014-12-23 1:16 UTC (permalink / raw)
To: Alex Williamson; +Cc: Wei Yang, bhelgaas, linux-pci
On Mon, Dec 22, 2014 at 04:38:24PM -0700, Alex Williamson wrote:
>On Mon, 2014-12-22 at 14:12 +0800, Wei Yang wrote:
>> Alex,
>>
>> Does it look good to you?
>
>Well, mixing changes to a subsystem and a driver in a single patch is
>surely going to slow down acceptance because neither Bjorn or I know who
>is going to grab it. I'd probably want to mark the vfio-pci change for
>stable, which would be a lot easier to do if it was just a 4-line
>correction to the existing code rather than extending the pci-core
>interface for a single, trivial user. Thanks,
>
>Alex
Hi,
I got your concern.
The reason I introduce a helper to check the pci device type is in my mind we
will try our best to hide those pci detailed information from outside. The
easiest way to fix it is to check the pci_dev->hdr_type field, but this will
export some internal information from pci core to vfio-pci.
I will first make a 4-line change to fix this error. If Bjorn like the idea of
the helper, I will make up another one.
>
>On Mon, Dec 08, 2014 at 09:22:18AM +0800, Wei Yang wrote:
>> >Current vfio-pci just supports normal pci device, so vfio_pci_probe() will
>> >return if the pci device is not a normal device. While current code makes a
>> >mistake. PCI_HEADER_TYPE is the offset in configuration space of the device
>> >type, but we use this value to mask the type value.
>> >
>> >This patch adds a function to check whether the pci device is a normal type
>> >and use it in vfio_pci_probe().
>> >
>> >Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
>> >---
>> > drivers/vfio/pci/vfio_pci.c | 4 +---
>> > include/linux/pci.h | 11 +++++++++++
>> > 2 files changed, 12 insertions(+), 3 deletions(-)
>> >
>> >diff --git a/drivers/vfio/pci/vfio_pci.c b/drivers/vfio/pci/vfio_pci.c
>> >index 9558da3..f82bf62 100644
>> >--- a/drivers/vfio/pci/vfio_pci.c
>> >+++ b/drivers/vfio/pci/vfio_pci.c
>> >@@ -839,13 +839,11 @@ static const struct vfio_device_ops vfio_pci_ops = {
>> >
>> > static int vfio_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
>> > {
>> >- u8 type;
>> > struct vfio_pci_device *vdev;
>> > struct iommu_group *group;
>> > int ret;
>> >
>> >- pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
>> >- if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
>> >+ if (!pci_is_normal(pdev))
>> > return -EINVAL;
>> >
>> > group = iommu_group_get(&pdev->dev);
>> >diff --git a/include/linux/pci.h b/include/linux/pci.h
>> >index 48d1f98..2027f66 100644
>> >--- a/include/linux/pci.h
>> >+++ b/include/linux/pci.h
>> >@@ -519,6 +519,17 @@ static inline bool pci_is_bridge(struct pci_dev *dev)
>> > dev->hdr_type == PCI_HEADER_TYPE_CARDBUS;
>> > }
>> >
>> >+/**
>> >+ * pci_is_normal - check if the PCI device is a normal device
>> >+ * @dev: PCI device
>> >+ *
>> >+ * Return true if the PCI device is a normal device
>> >+ */
>> >+static inline bool pci_is_normal(struct pci_dev *dev)
>> >+{
>> >+ return dev->hdr_type == PCI_HEADER_TYPE_NORMAL;
>> >+}
>> >+
>> > static inline struct pci_dev *pci_upstream_bridge(struct pci_dev *dev)
>> > {
>> > dev = pci_physfn(dev);
>> >--
>> >1.7.9.5
>>
>
>
--
Richard Yang
Help you, Help me
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH V2] vfio-pci: Fix the check on pci device type in vfio_pci_probe()
2014-12-23 1:16 ` Wei Yang
@ 2014-12-23 2:58 ` Wei Yang
2015-01-07 19:00 ` Alex Williamson
0 siblings, 1 reply; 6+ messages in thread
From: Wei Yang @ 2014-12-23 2:58 UTC (permalink / raw)
To: alex.williamson, bhelgaas; +Cc: linux-kernel, linux-pci, Wei Yang
Current vfio-pci just supports normal pci device, so vfio_pci_probe() will
return if the pci device is not a normal device. While current code makes a
mistake. PCI_HEADER_TYPE is the offset in configuration space of the device
type, but we use this value to mask the type value.
This patch fixs this by do the check directly on the pci_dev->hdr_type.
Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
V2:
* Fix the problem by checking pci_dev->hdr_type
---
drivers/vfio/pci/vfio_pci.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/drivers/vfio/pci/vfio_pci.c b/drivers/vfio/pci/vfio_pci.c
index 9558da3..2f8a055 100644
--- a/drivers/vfio/pci/vfio_pci.c
+++ b/drivers/vfio/pci/vfio_pci.c
@@ -839,13 +839,11 @@ static const struct vfio_device_ops vfio_pci_ops = {
static int vfio_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
{
- u8 type;
struct vfio_pci_device *vdev;
struct iommu_group *group;
int ret;
- pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
- if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
+ if (pdev->hdr_type != PCI_HEADER_TYPE_NORMAL)
return -EINVAL;
group = iommu_group_get(&pdev->dev);
--
1.7.9.5
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH V2] vfio-pci: Fix the check on pci device type in vfio_pci_probe()
2014-12-23 2:58 ` [PATCH V2] " Wei Yang
@ 2015-01-07 19:00 ` Alex Williamson
0 siblings, 0 replies; 6+ messages in thread
From: Alex Williamson @ 2015-01-07 19:00 UTC (permalink / raw)
To: Wei Yang; +Cc: bhelgaas, linux-kernel, linux-pci
On Tue, 2014-12-23 at 10:58 +0800, Wei Yang wrote:
> Current vfio-pci just supports normal pci device, so vfio_pci_probe() will
> return if the pci device is not a normal device. While current code makes a
> mistake. PCI_HEADER_TYPE is the offset in configuration space of the device
> type, but we use this value to mask the type value.
>
> This patch fixs this by do the check directly on the pci_dev->hdr_type.
>
> Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
>
> V2:
> * Fix the problem by checking pci_dev->hdr_type
>
> ---
> drivers/vfio/pci/vfio_pci.c | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
Applied to my for-linus branch for v3.19. I've also tagged it for
stable. Thanks,
Alex
> diff --git a/drivers/vfio/pci/vfio_pci.c b/drivers/vfio/pci/vfio_pci.c
> index 9558da3..2f8a055 100644
> --- a/drivers/vfio/pci/vfio_pci.c
> +++ b/drivers/vfio/pci/vfio_pci.c
> @@ -839,13 +839,11 @@ static const struct vfio_device_ops vfio_pci_ops = {
>
> static int vfio_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
> {
> - u8 type;
> struct vfio_pci_device *vdev;
> struct iommu_group *group;
> int ret;
>
> - pci_read_config_byte(pdev, PCI_HEADER_TYPE, &type);
> - if ((type & PCI_HEADER_TYPE) != PCI_HEADER_TYPE_NORMAL)
> + if (pdev->hdr_type != PCI_HEADER_TYPE_NORMAL)
> return -EINVAL;
>
> group = iommu_group_get(&pdev->dev);
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2015-01-07 19:00 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-12-08 1:22 [PATCH] vfio-pci: Fix the check on pci device type in vfio_pci_probe() Wei Yang
2014-12-22 6:12 ` Wei Yang
2014-12-22 23:38 ` Alex Williamson
2014-12-23 1:16 ` Wei Yang
2014-12-23 2:58 ` [PATCH V2] " Wei Yang
2015-01-07 19:00 ` Alex Williamson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).