Re: [PATCH v2 02/11] PCI: Add the pci_get_class_masked() helper

2023-08-11 Thread Ilpo Järvinen
On Thu, 10 Aug 2023, suijingfeng wrote:

> Hi,
> 
> 
> On 2023/8/9 22:01, Ilpo Järvinen wrote:
> > On Wed, 9 Aug 2023, Sui Jingfeng wrote:
> > 
> > > From: Sui Jingfeng 
> > > 
> > > Because there is no good way to get the mask member used to searching for
> > > devices that conform to a specific PCI class code, an application needs to
> > > process all PCI display devices can achieve its goal as follows:
> > This is mixing old and new way in a single sentence (which is confusing)?
> 
> 
> Thanks for reviewing, but I can't understand this sentence.
> Are you telling me that my description have grammar problem or something else?

I think it's a bit of both.

> I means that before apply this patch, we don't have a function can be used
> to get all PCI(e) devices in a system by matching against its the PCI base
> class code only,
> while keep the Sub-Class code and the Programming Interface ignored.
> By supply a mask as argument, such thing become possible.

This explanation you put into this reply is much easier to follow and 
understand. I recommend you'd use it to replace the unclear fragment 
above. So something along the lines of:

  There is no function that can be used to get all PCI(e) devices in a 
  system by matching against its the PCI base class code only, while keep 
  the Sub-Class code and the Programming Interface ignored.

  Add pci_get_class_masked() to allow supplying a mask for the get.

  [After this you can put the explanining code block+its intro if you 
   want]

-- 
 i.

> If an application want to process all PCI display devices in the system,
> then it can achieve its goal by calling pci_get_class_masked() function.
> 
> 
> > > pdev = NULL;
> > > do {
> > >   pdev = pci_get_class_masked(PCI_BASE_CLASS_DISPLAY << 16, 0xFF,
> > > pdev);
> > >   if (pdev)
> > >   do_something_for_pci_display_device(pdev);
> > > } while (pdev);
> > > 
> > > While previously, we just can not ignore Sub-Class code and the
> > > Programming
> > cannot
> > 
> > > Interface byte when do the searching.
> > doing the search.
> 
> 
> OK, will be fixed at the next version.
> 
> 
> > > Signed-off-by: Sui Jingfeng 
> > > ---
> > >   drivers/pci/search.c | 30 ++
> > >   include/linux/pci.h  |  7 +++
> > >   2 files changed, 37 insertions(+)
> > > 
> > > diff --git a/drivers/pci/search.c b/drivers/pci/search.c
> > > index b4c138a6ec02..f1c15aea868b 100644
> > > --- a/drivers/pci/search.c
> > > +++ b/drivers/pci/search.c
> > > @@ -334,6 +334,36 @@ struct pci_dev *pci_get_device(unsigned int vendor,
> > > unsigned int device,
> > >   }
> > >   EXPORT_SYMBOL(pci_get_device);
> > >   +/**
> > > + * pci_get_class_masked - begin or continue searching for a PCI device by
> > > class and mask
> > > + * @class: search for a PCI device with this class designation
> > > + * @from: Previous PCI device found in search, or %NULL for new search.
> > > + *
> > > + * Iterates through the list of known PCI devices.  If a PCI device is
> > No double spaces in kernel comments. Perhaps your editor might be adding
> > them on reflow (might be configurable to not do that).
> > 
> > > + * found with a matching @class, the reference count to the device is
> > > + * incremented and a pointer to its device structure is returned.
> > > + * Otherwise, %NULL is returned.
> > > + * A new search is initiated by passing %NULL as the @from argument.
> > > + * Otherwise if @from is not %NULL, searches continue from next device
> > > + * on the global list.  The reference count for @from is always
> > > decremented
> > > + * if it is not %NULL.
> > Use kerneldoc's Return: section for describing return value.
> > 
> > > + */
> > > +struct pci_dev *pci_get_class_masked(unsigned int class, unsigned int
> > > mask,
> > > +  struct pci_dev *from)
> > > +{
> > > + struct pci_device_id id = {
> > > + .vendor = PCI_ANY_ID,
> > > + .device = PCI_ANY_ID,
> > > + .subvendor = PCI_ANY_ID,
> > > + .subdevice = PCI_ANY_ID,
> > > + .class_mask = mask,
> > > + .class = class,
> > > + };
> > > +
> > > + return pci_get_dev_by_id(, from);
> > > +}
> > > +EXPORT_SYMBOL(pci_get_class_masked);
> > > +
> > >   /**
> > >* pci_get_class - begin or continue searching for a PCI device by class
> > >* @class: search for a PCI device with this class designation
> > > diff --git a/include/linux/pci.h b/include/linux/pci.h
> > > index 0ff7500772e6..b20e7ba844bf 100644
> > > --- a/include/linux/pci.h
> > > +++ b/include/linux/pci.h
> > > @@ -1180,6 +1180,9 @@ struct pci_dev *pci_get_slot(struct pci_bus *bus,
> > > unsigned int devfn);
> > >   struct pci_dev *pci_get_domain_bus_and_slot(int domain, unsigned int
> > > bus,
> > >   unsigned int devfn);
> > >   struct pci_dev *pci_get_class(unsigned int class, struct pci_dev *from);
> > > +struct pci_dev *pci_get_class_masked(unsigned int class, unsigned int
> > > 

Re: [PATCH v2 02/11] PCI: Add the pci_get_class_masked() helper

2023-08-10 Thread suijingfeng

Hi,


On 2023/8/9 22:01, Ilpo Järvinen wrote:

On Wed, 9 Aug 2023, Sui Jingfeng wrote:


From: Sui Jingfeng 

Because there is no good way to get the mask member used to searching for
devices that conform to a specific PCI class code, an application needs to
process all PCI display devices can achieve its goal as follows:

This is mixing old and new way in a single sentence (which is confusing)?



Thanks for reviewing, but I can't understand this sentence.
Are you telling me that my description have grammar problem or something else?


I means that before apply this patch, we don't have a function can be used
to get all PCI(e) devices in a system by matching against its the PCI base 
class code only,
while keep the Sub-Class code and the Programming Interface ignored.
By supply a mask as argument, such thing become possible.

If an application want to process all PCI display devices in the system,
then it can achieve its goal by calling pci_get_class_masked() function.



pdev = NULL;
do {
pdev = pci_get_class_masked(PCI_BASE_CLASS_DISPLAY << 16, 0xFF, 
pdev);
if (pdev)
do_something_for_pci_display_device(pdev);
} while (pdev);

While previously, we just can not ignore Sub-Class code and the Programming

cannot


Interface byte when do the searching.

doing the search.



OK, will be fixed at the next version.



Signed-off-by: Sui Jingfeng 
---
  drivers/pci/search.c | 30 ++
  include/linux/pci.h  |  7 +++
  2 files changed, 37 insertions(+)

diff --git a/drivers/pci/search.c b/drivers/pci/search.c
index b4c138a6ec02..f1c15aea868b 100644
--- a/drivers/pci/search.c
+++ b/drivers/pci/search.c
@@ -334,6 +334,36 @@ struct pci_dev *pci_get_device(unsigned int vendor, 
unsigned int device,
  }
  EXPORT_SYMBOL(pci_get_device);
  
+/**

+ * pci_get_class_masked - begin or continue searching for a PCI device by 
class and mask
+ * @class: search for a PCI device with this class designation
+ * @from: Previous PCI device found in search, or %NULL for new search.
+ *
+ * Iterates through the list of known PCI devices.  If a PCI device is

No double spaces in kernel comments. Perhaps your editor might be adding
them on reflow (might be configurable to not do that).


+ * found with a matching @class, the reference count to the device is
+ * incremented and a pointer to its device structure is returned.
+ * Otherwise, %NULL is returned.
+ * A new search is initiated by passing %NULL as the @from argument.
+ * Otherwise if @from is not %NULL, searches continue from next device
+ * on the global list.  The reference count for @from is always decremented
+ * if it is not %NULL.

Use kerneldoc's Return: section for describing return value.


+ */
+struct pci_dev *pci_get_class_masked(unsigned int class, unsigned int mask,
+struct pci_dev *from)
+{
+   struct pci_device_id id = {
+   .vendor = PCI_ANY_ID,
+   .device = PCI_ANY_ID,
+   .subvendor = PCI_ANY_ID,
+   .subdevice = PCI_ANY_ID,
+   .class_mask = mask,
+   .class = class,
+   };
+
+   return pci_get_dev_by_id(, from);
+}
+EXPORT_SYMBOL(pci_get_class_masked);
+
  /**
   * pci_get_class - begin or continue searching for a PCI device by class
   * @class: search for a PCI device with this class designation
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 0ff7500772e6..b20e7ba844bf 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -1180,6 +1180,9 @@ struct pci_dev *pci_get_slot(struct pci_bus *bus, 
unsigned int devfn);
  struct pci_dev *pci_get_domain_bus_and_slot(int domain, unsigned int bus,
unsigned int devfn);
  struct pci_dev *pci_get_class(unsigned int class, struct pci_dev *from);
+struct pci_dev *pci_get_class_masked(unsigned int class, unsigned int mask,
+struct pci_dev *from);
+
  int pci_dev_present(const struct pci_device_id *ids);
  
  int pci_bus_read_config_byte(struct pci_bus *bus, unsigned int devfn,

@@ -1895,6 +1898,10 @@ static inline struct pci_dev *pci_get_class(unsigned int 
class,
struct pci_dev *from)
  { return NULL; }
  
+static inline struct pci_dev *pci_get_class_masked(unsigned int class,

+  unsigned int mask,
+  struct pci_dev *from)
+{ return NULL; }
  
  static inline int pci_dev_present(const struct pci_device_id *ids)

  { return 0; }





Re: [PATCH v2 02/11] PCI: Add the pci_get_class_masked() helper

2023-08-10 Thread Ilpo Järvinen
On Wed, 9 Aug 2023, Sui Jingfeng wrote:

> From: Sui Jingfeng 
> 
> Because there is no good way to get the mask member used to searching for
> devices that conform to a specific PCI class code, an application needs to
> process all PCI display devices can achieve its goal as follows:

This is mixing old and new way in a single sentence (which is confusing)?
 
> pdev = NULL;
> do {
>   pdev = pci_get_class_masked(PCI_BASE_CLASS_DISPLAY << 16, 0xFF, 
> pdev);
>   if (pdev)
>   do_something_for_pci_display_device(pdev);
> } while (pdev);
> 
> While previously, we just can not ignore Sub-Class code and the Programming

cannot

> Interface byte when do the searching.

doing the search.

> 
> Signed-off-by: Sui Jingfeng 
> ---
>  drivers/pci/search.c | 30 ++
>  include/linux/pci.h  |  7 +++
>  2 files changed, 37 insertions(+)
> 
> diff --git a/drivers/pci/search.c b/drivers/pci/search.c
> index b4c138a6ec02..f1c15aea868b 100644
> --- a/drivers/pci/search.c
> +++ b/drivers/pci/search.c
> @@ -334,6 +334,36 @@ struct pci_dev *pci_get_device(unsigned int vendor, 
> unsigned int device,
>  }
>  EXPORT_SYMBOL(pci_get_device);
>  
> +/**
> + * pci_get_class_masked - begin or continue searching for a PCI device by 
> class and mask
> + * @class: search for a PCI device with this class designation
> + * @from: Previous PCI device found in search, or %NULL for new search.
> + *
> + * Iterates through the list of known PCI devices.  If a PCI device is

No double spaces in kernel comments. Perhaps your editor might be adding 
them on reflow (might be configurable to not do that).

> + * found with a matching @class, the reference count to the device is
> + * incremented and a pointer to its device structure is returned.
> + * Otherwise, %NULL is returned.
> + * A new search is initiated by passing %NULL as the @from argument.
> + * Otherwise if @from is not %NULL, searches continue from next device
> + * on the global list.  The reference count for @from is always decremented
> + * if it is not %NULL.

Use kerneldoc's Return: section for describing return value.

> + */
> +struct pci_dev *pci_get_class_masked(unsigned int class, unsigned int mask,
> +  struct pci_dev *from)
> +{
> + struct pci_device_id id = {
> + .vendor = PCI_ANY_ID,
> + .device = PCI_ANY_ID,
> + .subvendor = PCI_ANY_ID,
> + .subdevice = PCI_ANY_ID,
> + .class_mask = mask,
> + .class = class,
> + };
> +
> + return pci_get_dev_by_id(, from);
> +}
> +EXPORT_SYMBOL(pci_get_class_masked);
> +
>  /**
>   * pci_get_class - begin or continue searching for a PCI device by class
>   * @class: search for a PCI device with this class designation
> diff --git a/include/linux/pci.h b/include/linux/pci.h
> index 0ff7500772e6..b20e7ba844bf 100644
> --- a/include/linux/pci.h
> +++ b/include/linux/pci.h
> @@ -1180,6 +1180,9 @@ struct pci_dev *pci_get_slot(struct pci_bus *bus, 
> unsigned int devfn);
>  struct pci_dev *pci_get_domain_bus_and_slot(int domain, unsigned int bus,
>   unsigned int devfn);
>  struct pci_dev *pci_get_class(unsigned int class, struct pci_dev *from);
> +struct pci_dev *pci_get_class_masked(unsigned int class, unsigned int mask,
> +  struct pci_dev *from);
> +
>  int pci_dev_present(const struct pci_device_id *ids);
>  
>  int pci_bus_read_config_byte(struct pci_bus *bus, unsigned int devfn,
> @@ -1895,6 +1898,10 @@ static inline struct pci_dev *pci_get_class(unsigned 
> int class,
>   struct pci_dev *from)
>  { return NULL; }
>  
> +static inline struct pci_dev *pci_get_class_masked(unsigned int class,
> +unsigned int mask,
> +struct pci_dev *from)
> +{ return NULL; }
>  
>  static inline int pci_dev_present(const struct pci_device_id *ids)
>  { return 0; }
> 

-- 
 i.



[PATCH v2 02/11] PCI: Add the pci_get_class_masked() helper

2023-08-08 Thread Sui Jingfeng
From: Sui Jingfeng 

Because there is no good way to get the mask member used to searching for
devices that conform to a specific PCI class code, an application needs to
process all PCI display devices can achieve its goal as follows:

pdev = NULL;
do {
pdev = pci_get_class_masked(PCI_BASE_CLASS_DISPLAY << 16, 0xFF, 
pdev);
if (pdev)
do_something_for_pci_display_device(pdev);
} while (pdev);

While previously, we just can not ignore Sub-Class code and the Programming
Interface byte when do the searching.

Signed-off-by: Sui Jingfeng 
---
 drivers/pci/search.c | 30 ++
 include/linux/pci.h  |  7 +++
 2 files changed, 37 insertions(+)

diff --git a/drivers/pci/search.c b/drivers/pci/search.c
index b4c138a6ec02..f1c15aea868b 100644
--- a/drivers/pci/search.c
+++ b/drivers/pci/search.c
@@ -334,6 +334,36 @@ struct pci_dev *pci_get_device(unsigned int vendor, 
unsigned int device,
 }
 EXPORT_SYMBOL(pci_get_device);
 
+/**
+ * pci_get_class_masked - begin or continue searching for a PCI device by 
class and mask
+ * @class: search for a PCI device with this class designation
+ * @from: Previous PCI device found in search, or %NULL for new search.
+ *
+ * Iterates through the list of known PCI devices.  If a PCI device is
+ * found with a matching @class, the reference count to the device is
+ * incremented and a pointer to its device structure is returned.
+ * Otherwise, %NULL is returned.
+ * A new search is initiated by passing %NULL as the @from argument.
+ * Otherwise if @from is not %NULL, searches continue from next device
+ * on the global list.  The reference count for @from is always decremented
+ * if it is not %NULL.
+ */
+struct pci_dev *pci_get_class_masked(unsigned int class, unsigned int mask,
+struct pci_dev *from)
+{
+   struct pci_device_id id = {
+   .vendor = PCI_ANY_ID,
+   .device = PCI_ANY_ID,
+   .subvendor = PCI_ANY_ID,
+   .subdevice = PCI_ANY_ID,
+   .class_mask = mask,
+   .class = class,
+   };
+
+   return pci_get_dev_by_id(, from);
+}
+EXPORT_SYMBOL(pci_get_class_masked);
+
 /**
  * pci_get_class - begin or continue searching for a PCI device by class
  * @class: search for a PCI device with this class designation
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 0ff7500772e6..b20e7ba844bf 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -1180,6 +1180,9 @@ struct pci_dev *pci_get_slot(struct pci_bus *bus, 
unsigned int devfn);
 struct pci_dev *pci_get_domain_bus_and_slot(int domain, unsigned int bus,
unsigned int devfn);
 struct pci_dev *pci_get_class(unsigned int class, struct pci_dev *from);
+struct pci_dev *pci_get_class_masked(unsigned int class, unsigned int mask,
+struct pci_dev *from);
+
 int pci_dev_present(const struct pci_device_id *ids);
 
 int pci_bus_read_config_byte(struct pci_bus *bus, unsigned int devfn,
@@ -1895,6 +1898,10 @@ static inline struct pci_dev *pci_get_class(unsigned int 
class,
struct pci_dev *from)
 { return NULL; }
 
+static inline struct pci_dev *pci_get_class_masked(unsigned int class,
+  unsigned int mask,
+  struct pci_dev *from)
+{ return NULL; }
 
 static inline int pci_dev_present(const struct pci_device_id *ids)
 { return 0; }
-- 
2.34.1