Skip to contentSkip to conversation

Generate device tree node for pci devices

PATCH Superseded Lizhi Hou 2 participants 16 messages Dec 15, 2022 – Jan 4, 2023

Subsystems (2)

v5 of 13 Superseded Latest is v13 →

lore
Share & tools
lore

Lizhi Hou

author
#0 lore raw

v5 cover Generate device tree node for pci devices

This patch series introduces OF overlay support for PCI devices which primarily addresses two use cases. First, it provides a data driven method to describe hardware peripherals that are present in a PCI endpoint and hence can be accessed by the PCI host. Second, it allows reuse of a OF compatible driver -- often used in SoC platforms -- in a PCI host based system.

There are 2 series devices rely on this patch:

1) Xilinx Alveo Accelerator cards (FPGA based device)
2) Microchip LAN9662 Ethernet Controller

     Please see: https://lore.kernel.org/lkml/[email protected]/

Normally, the PCI core discovers PCI devices and their BARs using the PCI enumeration process. However, the process does not provide a way to discover the hardware peripherals that are present in a PCI device, and which can be accessed through the PCI BARs. Also, the enumeration process does not provide a way to associate MSI-X vectors of a PCI device with the hardware peripherals that are present in the device. PCI device drivers often use header files to describe the hardware peripherals and their resources as there is no standard data driven way to do so. This patch series proposes to use flattened device tree blob to describe the peripherals in a data driven way. Based on previous discussion, using device tree overlay is the best way to unflatten the blob and populate platform devices. To use device tree overlay, there are three obvious problems that need to be resolved.

First, we need to create a base tree for non-DT system such as x86_64. A patch series has been submitted for this:
https://lore.kernel.org/lkml/[email protected]/
https://lore.kernel.org/lkml/[email protected]/

Second, a device tree node corresponding to the PCI endpoint is required for overlaying the flattened device tree blob for that PCI endpoint. Because PCI is a self-discoverable bus, a device tree node is usually not created for PCI devices. This series adds support to generate a device tree node for a PCI device which advertises itself using PCI quirks infrastructure.

Third, we need to generate device tree nodes for PCI bridges since a child PCI endpoint may choose to have a device tree node created.

This patch series is made up of three patches.

The first patch is adding OF interface to create or destroy OF node dynamically.

The second patch introduces a kernel option, CONFIG_DYNAMIC_PCI_OF_NODEX. When the option is turned on, the kernel will generate device tree nodes for all PCI bridges unconditionally. The patch also shows how to use the PCI quirks infrastructure, DECLARE_PCI_FIXUP_FINAL to generate a device tree node for a device. Specifically, the patch generates a device tree node for Xilinx Alveo U50 PCIe accelerator device. The generated device tree nodes do not have any property.

The third patch adds basic properties ('reg', 'compatible' and 'device_type') to the dynamically generated device tree nodes. More properties can be added in the future.

Here is the example of device tree nodes generated within the ARM64 QEMU.

# lspci -t    
-[0000:00]-+-00.0
           +-01.0-[01]--
           +-01.1-[02]----00.0
           +-01.2-[03]----00.0
           +-01.3-[04]----00.0
           +-01.4-[05]----00.0
           +-01.5-[06]--
           +-01.6-[07]--
           +-01.7-[08]--
           +-02.0-[09-0b]----00.0-[0a-0b]----00.0-[0b]--+-00.0
           |                                            \-00.1
           +-02.1-[0c]--
           \-03.0-[0d-0e]----00.0-[0e]----01.0
# tree /sys/firmware/devicetree/base/pcie\@10000000
/sys/firmware/devicetree/base/pcie@10000000
|-- #address-cells
|-- #interrupt-cells
|-- #size-cells
|-- bus-range
|-- compatible
|-- device_type
|-- dma-coherent
|-- interrupt-map
|-- interrupt-map-mask
|-- linux,pci-domain
|-- msi-parent
|-- name
|-- pci@1,0
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@1,1
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@1,2
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@1,3
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@1,4
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@1,5
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@1,6
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@1,7
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@2,0
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- pci@0,0
|   |   |-- #address_cells
|   |   |-- #size_cells
|   |   |-- compatible
|   |   |-- device_type
|   |   |-- pci@0,0
|   |   |   |-- #address_cells
|   |   |   |-- #size_cells
|   |   |   |-- compatible
|   |   |   |-- dev@0,0
|   |   |   |   |-- compatible
|   |   |   |   `-- reg
|   |   |   |-- dev@0,1
|   |   |   |   |-- compatible
|   |   |   |   `-- reg
|   |   |   |-- device_type
|   |   |   |-- ranges
|   |   |   `-- reg
|   |   |-- ranges
|   |   `-- reg
|   |-- ranges
|   `-- reg
|-- pci@2,1
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- ranges
|   `-- reg
|-- pci@3,0
|   |-- #address_cells
|   |-- #size_cells
|   |-- compatible
|   |-- device_type
|   |-- pci@0,0
|   |   |-- #address_cells
|   |   |-- #size_cells
|   |   |-- compatible
|   |   |-- device_type
|   |   |-- ranges
|   |   `-- reg
|   |-- ranges
|   `-- reg
|-- ranges
`-- reg

Changes since RFC v4:
- Fixed code review comments

Changes since RFC v3:
- Split the Xilinx Alveo U50 PCI quirk to a separate patch
- Minor changes in commit description and code comment

Changes since RFC v2:
- Merged patch 3 with patch 2
- Added OF interfaces of_changeset_add_prop_* and use them to create properties.
- Added '#address-cells', '#size-cells' and 'ranges' properties.

Changes since RFC v1:
- Added one patch to create basic properties.
- To move DT related code out of PCI subsystem, replaced of_node_alloc() with of_create_node()/of_destroy_node()

Lizhi Hou (3):
of: dynamic: Add interfaces for creating device node dynamically
PCI: Create device tree node for selected devices
PCI: Add PCI quirks to generate device tree node for Xilinx Alveo U50

 drivers/of/dynamic.c        | 197 ++++++++++++++++++++++++++++++++
 drivers/pci/Kconfig         |  12 ++
 drivers/pci/Makefile        |   1 +
 drivers/pci/bus.c           |   2 +
 drivers/pci/msi/irqdomain.c |   6 +-
 drivers/pci/of.c            |  71 ++++++++++++
 drivers/pci/of_property.c   | 222 ++++++++++++++++++++++++++++++++++++
 drivers/pci/pci-driver.c    |   3 +-
 drivers/pci/pci.h           |  19 +++
 drivers/pci/quirks.c        |  11 ++
 drivers/pci/remove.c        |   1 +
 include/linux/of.h          |  24 ++++
 12 files changed, 566 insertions(+), 3 deletions(-)
 create mode 100644 drivers/pci/of_property.c
Signature · 1 line
2.17.1
author
#1 lore raw

PATCH v5 1/3 of: dynamic: Add interfaces for creating device node dynamically

of_create_node() creates device node dynamically. The parent device node and full name are required for creating the node. It optionally creates an OF changeset and attaches the newly created node to the changeset. The device node pointer and the changeset pointer can be used to add properties to the device node and apply the node to the base tree.

of_destroy_node() frees the device node created by of_create_node(). If an OF changeset was also created for this node, it will destroy the changeset before freeing the device node.

Expand of_changeset APIs to handle specific types of properties.

    of_changeset_add_prop_string()
    of_changeset_add_prop_string_array()
    of_changeset_add_prop_u32_array()
---
 drivers/of/dynamic.c | 197 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/of.h   |  24 ++++++
 2 files changed, 221 insertions(+)
drivers/of/dynamic.c +197 -0
index cd3821a6444f..067d996a9f79 100644
@@ -461,6 +461,71 @@ struct device_node *__of_node_dup(const struct device_node *np,
 	return NULL;
 }
 
+/**
+ * of_create_node - Dynamically create a device node
+ *
+ * @parent: Pointer to parent device node
+ * @full_name: Node full name
+ * @cset: Pointer to returning changeset
+ *
+ * Return: Pointer to the created device node or NULL in case of an error.
+ */
+struct device_node *of_create_node(struct device_node *parent,
+				   const char *full_name,
+				   struct of_changeset **cset)
+{
+	struct of_changeset *ocs;
+	struct device_node *np;
+	int ret;
+
+	np = __of_node_dup(NULL, full_name);
+	if (!np)
+		return NULL;
+	np->parent = parent;
+
+	if (!cset)
+		return np;
+
+	ocs = kmalloc(sizeof(*cset), GFP_KERNEL);
+	if (!ocs) {
+		of_node_put(np);
+		return NULL;
+	}
+
+	of_changeset_init(ocs);
+	ret = of_changeset_attach_node(ocs, np);
+	if (ret) {
+		of_changeset_destroy(ocs);
+		of_node_put(np);
+		kfree(ocs);
+		return NULL;
+	}
+
+	np->data = ocs;
+	*cset = ocs;
+
+	return np;
+}
+EXPORT_SYMBOL(of_create_node);
+
+/**
+ * of_destroy_node - Destroy a dynamically created device node
+ *
+ * @np: Pointer to dynamically created device node
+ *
+ */
+void of_destroy_node(struct device_node *np)
+{
+	struct of_changeset *ocs;
+
+	if (np->data) {
+		ocs = (struct of_changeset *)np->data;
+		of_changeset_destroy(ocs);
+	}
+	of_node_put(np);
+}
+EXPORT_SYMBOL(of_destroy_node);
+
 static void __of_changeset_entry_destroy(struct of_changeset_entry *ce)
 {
 	if (ce->action == OF_RECONFIG_ATTACH_NODE &&
@@ -934,3 +999,135 @@ int of_changeset_action(struct of_changeset *ocs, unsigned long action,
 	return 0;
 }
 EXPORT_SYMBOL_GPL(of_changeset_action);
+
+static int of_changeset_add_prop_helper(struct of_changeset *ocs,
+					struct device_node *np,
+					const struct property *pp)
+{
+	struct property *new_pp;
+	int ret;
+
+	new_pp = __of_prop_dup(pp, GFP_KERNEL);
+	if (!new_pp)
+		return -ENOMEM;
+
+	ret = of_changeset_add_property(ocs, np, new_pp);
+	if (ret) {
+		kfree(new_pp->name);
+		kfree(new_pp->value);
+		kfree(new_pp);
+	}
+
+	return ret;
+}
+
+/**
+ * of_changeset_add_prop_string - Add a string property to a changeset
+ *
+ * @ocs:	changeset pointer
+ * @np:		device node pointer
+ * @prop_name:	name of the property to be added
+ * @str:	pointer to null terminated string
+ *
+ * Create a string property and add it to a changeset.
+ *
+ * Return: 0 on success, a negative error value in case of an error.
+ */
+int of_changeset_add_prop_string(struct of_changeset *ocs,
+				 struct device_node *np,
+				 const char *prop_name, const char *str)
+{
+	struct property prop;
+
+	prop.name = (char *)prop_name;
+	prop.length = strlen(str) + 1;
+	prop.value = (void *)str;
+
+	return of_changeset_add_prop_helper(ocs, np, &prop);
+}
+EXPORT_SYMBOL_GPL(of_changeset_add_prop_string);
+
+/**
+ * of_changeset_add_prop_string_array - Add a string list property to
+ * a changeset
+ *
+ * @ocs:	changeset pointer
+ * @np:		device node pointer
+ * @prop_name:	name of the property to be added
+ * @str_array:	pointer to an array of null terminated strings
+ * @sz:		number of string array elements
+ *
+ * Create a string list property and add it to a changeset.
+ *
+ * Return: 0 on success, a negative error value in case of an error.
+ */
+int of_changeset_add_prop_string_array(struct of_changeset *ocs,
+				       struct device_node *np,
+				       const char *prop_name,
+				       const char **str_array, size_t sz)
+{
+	struct property prop;
+	int i, ret;
+	char *vp;
+
+	prop.name = (char *)prop_name;
+
+	prop.length = 0;
+	for (i = 0; i < sz; i++)
+		prop.length += strlen(str_array[i]) + 1;
+
+	prop.value = kmalloc(prop.length, GFP_KERNEL);
+	if (!prop.value)
+		return -ENOMEM;
+
+	vp = prop.value;
+	for (i = 0; i < sz; i++) {
+		vp += snprintf(vp, (char *)prop.value + prop.length - vp, "%s",
+			       str_array[i]) + 1;
+	}
+	ret = of_changeset_add_prop_helper(ocs, np, &prop);
+	kfree(prop.value);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(of_changeset_add_prop_string_array);
+
+/**
+ * of_changeset_add_prop_u32_array - Add a property of 32 bit integers
+ * property to a changeset
+ *
+ * @ocs:	changeset pointer
+ * @np:		device node pointer
+ * @prop_name:	name of the property to be added
+ * @array:	pointer to an array of 32 bit integers
+ * @sz:		number of array elements
+ *
+ * Create a property of 32 bit integers and add it to a changeset.
+ *
+ * Return: 0 on success, a negative error value in case of an error.
+ */
+int of_changeset_add_prop_u32_array(struct of_changeset *ocs,
+				    struct device_node *np,
+				    const char *prop_name,
+				    const u32 *array, size_t sz)
+{
+	struct property prop;
+	__be32 *val;
+	int i, ret;
+
+	val = kcalloc(sz, sizeof(__be32), GFP_KERNEL);
+	if (!val)
+		return -ENOMEM;
+
+	for (i = 0; i < sz; i++)
+		val[i] = cpu_to_be32(array[i]);
+	prop.name = (char *)prop_name;
+	prop.length = sizeof(u32) * sz;
+	prop.value = (void *)val;
+
+	ret = of_changeset_add_prop_helper(ocs, np, &prop);
+	kfree(val);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(of_changeset_add_prop_u32_array);
include/linux/of.h +24 -0
index 766d002bddb9..ba31036f0876 100644
@@ -1505,6 +1505,30 @@ static inline int of_changeset_update_property(struct of_changeset *ocs,
 {
 	return of_changeset_action(ocs, OF_RECONFIG_UPDATE_PROPERTY, np, prop);
 }
+
+struct device_node *of_create_node(struct device_node *parent,
+				   const char *full_name,
+				   struct of_changeset **cset);
+void of_destroy_node(struct device_node *np);
+int of_changeset_add_prop_string(struct of_changeset *ocs,
+				 struct device_node *np,
+				 const char *prop_name, const char *str);
+int of_changeset_add_prop_string_array(struct of_changeset *ocs,
+				       struct device_node *np,
+				       const char *prop_name,
+				       const char **str_array, size_t sz);
+int of_changeset_add_prop_u32_array(struct of_changeset *ocs,
+				    struct device_node *np,
+				    const char *prop_name,
+				    const u32 *array, size_t sz);
+static inline int of_changeset_add_prop_u32(struct of_changeset *ocs,
+					    struct device_node *np,
+					    const char *prop_name,
+					    const u32 val)
+{
+	return of_changeset_add_prop_u32_array(ocs, np, prop_name, &val, 1);
+}
+
 #else /* CONFIG_OF_DYNAMIC */
 static inline int of_reconfig_notifier_register(struct notifier_block *nb)
 {
Signature · 1 line
2.17.1
#2 lore raw

Re: [PATCH V5 1/3] of: dynamic: Add interfaces for creating device node dynamically

Quoted 48 lines · Lizhi Hou

of_create_node() creates device node dynamically. The parent device node and full name are required for creating the node. It optionally creates an OF changeset and attaches the newly created node to the changeset. The device node pointer and the changeset pointer can be used to add properties to the device node and apply the node to the base tree.

of_destroy_node() frees the device node created by of_create_node(). If an OF changeset was also created for this node, it will destroy the changeset before freeing the device node.

Expand of_changeset APIs to handle specific types of properties.

    of_changeset_add_prop_string()
    of_changeset_add_prop_string_array()
    of_changeset_add_prop_u32_array()
---
 drivers/of/dynamic.c | 197 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/of.h   |  24 ++++++
 2 files changed, 221 insertions(+)
drivers/of/dynamic.c +16 -0
index cd3821a6444f..067d996a9f79 100644
@@ -461,6 +461,71 @@ struct device_node *__of_node_dup(const struct device_node *np,
 	return NULL;
 }
 
+/**
+ * of_create_node - Dynamically create a device node
+ *
+ * @parent: Pointer to parent device node
+ * @full_name: Node full name
+ * @cset: Pointer to returning changeset
+ *
+ * Return: Pointer to the created device node or NULL in case of an error.
+ */
+struct device_node *of_create_node(struct device_node *parent,
+				   const char *full_name,
+				   struct of_changeset **cset)
+{
+	struct of_changeset *ocs;
+	struct device_node *np;
+	int ret;

drivers/of/dynamic.c struct device_node *__of_node_dup(const struct device_node *np, · line 489

Quoted 10 lines
drivers/of/dynamic.c +10 -0
+
+	np = __of_node_dup(NULL, full_name);
+	if (!np)
+		return NULL;
+	np->parent = parent;
+
+	if (!cset)
+		return np;
+
+	ocs = kmalloc(sizeof(*cset), GFP_KERNEL);

I started to test this series and this sizeof(*cset) is probably wrong, it should be sizeof(*ocs) or it will yield the size of a struct of_changeset pointer and not the struct of_changeset itself.

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
author
#3 lore raw

Re: [PATCH V5 1/3] of: dynamic: Add interfaces for creating device node dynamically

Quoted 56 lines · Clément Léger
Quoted 53 lines · Lizhi Hou

of_create_node() creates device node dynamically. The parent device node and full name are required for creating the node. It optionally creates an OF changeset and attaches the newly created node to the changeset. The device node pointer and the changeset pointer can be used to add properties to the device node and apply the node to the base tree.

of_destroy_node() frees the device node created by of_create_node(). If an OF changeset was also created for this node, it will destroy the changeset before freeing the device node.

Expand of_changeset APIs to handle specific types of properties.

     of_changeset_add_prop_string()
     of_changeset_add_prop_string_array()
     of_changeset_add_prop_u32_array()
---
  drivers/of/dynamic.c | 197 +++++++++++++++++++++++++++++++++++++++++++
  include/linux/of.h   |  24 ++++++
  2 files changed, 221 insertions(+)
drivers/of/dynamic.c +21 -0
index cd3821a6444f..067d996a9f79 100644
@@ -461,6 +461,71 @@ struct device_node *__of_node_dup(const struct device_node *np,
  	return NULL;
  }
 
+/**
+ * of_create_node - Dynamically create a device node
+ *
+ * @parent: Pointer to parent device node
+ * @full_name: Node full name
+ * @cset: Pointer to returning changeset
+ *
+ * Return: Pointer to the created device node or NULL in case of an error.
+ */
+struct device_node *of_create_node(struct device_node *parent,
+				   const char *full_name,
+				   struct of_changeset **cset)
+{
+	struct of_changeset *ocs;
+	struct device_node *np;
+	int ret;
+
+	np = __of_node_dup(NULL, full_name);
+	if (!np)
+		return NULL;
+	np->parent = parent;
Quoted 8 lines

drivers/of/dynamic.c struct device_node *__of_node_dup(const struct device_node *np, · line 489

Quoted 5 lines
drivers/of/dynamic.c +5 -0
+
+	if (!cset)
+		return np;
+
+	ocs = kmalloc(sizeof(*cset), GFP_KERNEL);

I started to test this series and this sizeof(*cset) is probably wrong, it should be sizeof(*ocs) or it will yield the size of a struct of_changeset pointer and not the struct of_changeset itself.

I will fix this. Thanks.

Lizhi

Quoted 1 line
author
#4 lore raw

PATCH v5 2/3 PCI: Create device tree node for selected devices

In reply to Lizhi Hou

The PCI endpoint device such as Xilinx Alveo PCI card maps the register spaces from multiple hardware peripherals to its PCI BAR. Normally, the PCI core discovers devices and BARs using the PCI enumeration process. There is no infrastructure to discover the hardware peripherals that are present in a PCI device, and which can be accessed through the PCI BARs.

For Alveo PCI card, the card firmware provides a flattened device tree to describe the hardware peripherals on its BARs. The Alveo card driver can load this flattened device tree and leverage device tree framework to generate platform devices for the hardware peripherals eventually.

Apparently, the device tree framework requires a device tree node for the PCI device. Thus, it can generate the device tree nodes for hardware peripherals underneath. Because PCI is self discoverable bus, there might not be a device tree node created for PCI devices. This patch is to add support to generate device tree node for PCI devices.

Added a kernel option. When the option is turned on, the kernel will generate device tree nodes for PCI bridges unconditionally.

Initially, the basic properties are added for the dynamically generated device tree nodes.

---
 drivers/pci/Kconfig         |  12 ++
 drivers/pci/Makefile        |   1 +
 drivers/pci/bus.c           |   2 +
 drivers/pci/msi/irqdomain.c |   6 +-
 drivers/pci/of.c            |  71 ++++++++++++
 drivers/pci/of_property.c   | 222 ++++++++++++++++++++++++++++++++++++
 drivers/pci/pci-driver.c    |   3 +-
 drivers/pci/pci.h           |  19 +++
 drivers/pci/remove.c        |   1 +
 9 files changed, 334 insertions(+), 3 deletions(-)
 create mode 100644 drivers/pci/of_property.c
drivers/pci/Kconfig +12 -0
index 55c028af4bd9..1b2347aa1e5b 100644
@@ -198,6 +198,18 @@ config PCI_HYPERV
 	  The PCI device frontend driver allows the kernel to import arbitrary
 	  PCI devices from a PCI backend to support PCI driver domains.
 
+config PCI_DYNAMIC_OF_NODES
+	bool "Create Devicetree nodes for PCI devices"
+	depends on OF
+	select OF_DYNAMIC
+	help
+	  This option enables support for generating device tree nodes for some
+	  PCI devices. Thus, the driver of this kind can load and overlay
+	  flattened device tree for its downstream devices.
+
+	  Once this option is selected, the device tree nodes will be generated
+	  for all PCI bridges.
+
 choice
 	prompt "PCI Express hierarchy optimization setting"
 	default PCIE_BUS_DEFAULT
drivers/pci/Makefile +1 -0
index 2680e4c92f0a..cc8b4e01e29d 100644
@@ -32,6 +32,7 @@ obj-$(CONFIG_PCI_P2PDMA)	+= p2pdma.o
 obj-$(CONFIG_XEN_PCIDEV_FRONTEND) += xen-pcifront.o
 obj-$(CONFIG_VGA_ARB)		+= vgaarb.o
 obj-$(CONFIG_PCI_DOE)		+= doe.o
+obj-$(CONFIG_PCI_DYNAMIC_OF_NODES) += of_property.o
 
 # Endpoint library must be initialized before its users
 obj-$(CONFIG_PCI_ENDPOINT)	+= endpoint/
drivers/pci/bus.c +2 -0
index 3cef835b375f..8507cc32b61d 100644
@@ -316,6 +316,8 @@ void pci_bus_add_device(struct pci_dev *dev)
 	 */
 	pcibios_bus_add_device(dev);
 	pci_fixup_device(pci_fixup_final, dev);
+	if (pci_is_bridge(dev))
+		of_pci_make_dev_node(dev);
 	pci_create_sysfs_dev_files(dev);
 	pci_proc_attach_device(dev);
 	pci_bridge_d3_update(dev);
drivers/pci/msi/irqdomain.c +4 -2
index e9cf318e6670..eeaf44169bfd 100644
@@ -230,8 +230,10 @@ u32 pci_msi_domain_get_msi_rid(struct irq_domain *domain, struct pci_dev *pdev)
 	pci_for_each_dma_alias(pdev, get_msi_id_cb, &rid);
 
 	of_node = irq_domain_get_of_node(domain);
-	rid = of_node ? of_msi_map_id(&pdev->dev, of_node, rid) :
-			iort_msi_map_id(&pdev->dev, rid);
+	if (of_node && !of_node_check_flag(of_node, OF_DYNAMIC))
+		rid = of_msi_map_id(&pdev->dev, of_node, rid);
+	else
+		rid = iort_msi_map_id(&pdev->dev, rid);
 
 	return rid;
 }
drivers/pci/of.c +71 -0

83 lines not on this page · raw message

drivers/pci/of_property.c new file +222 -0

222 lines not on this page · raw message

drivers/pci/pci-driver.c +2 -1
index 49238ddd39ee..1540c4c9a770 100644
@@ -1628,7 +1628,8 @@ static int pci_dma_configure(struct device *dev)
 	bridge = pci_get_host_bridge_device(to_pci_dev(dev));
 
 	if (IS_ENABLED(CONFIG_OF) && bridge->parent &&
-	    bridge->parent->of_node) {
+	    bridge->parent->of_node &&
+	    !of_node_check_flag(bridge->parent->of_node, OF_DYNAMIC)) {
 		ret = of_dma_configure(dev, bridge->parent->of_node, true);
 	} else if (has_acpi_companion(bridge)) {
 		struct acpi_device *adev = to_acpi_device_node(bridge->fwnode);
drivers/pci/pci.h +19 -0
index 785f31086313..bd81dc4ca04f 100644
@@ -678,6 +678,25 @@ static inline int devm_of_pci_bridge_init(struct device *dev, struct pci_host_br
 
 #endif /* CONFIG_OF */
 
+struct of_changeset;
+
+#ifdef CONFIG_PCI_DYNAMIC_OF_NODES
+void of_pci_make_dev_node(struct pci_dev *pdev);
+void of_pci_remove_node(struct pci_dev *pdev);
+int of_pci_add_properties(struct pci_dev *pdev, struct of_changeset *ocs,
+			  struct device_node *np);
+#else
+static inline void
+of_pci_make_dev_node(struct pci_dev *pdev)
+{
+}
+
+static inline void
+of_pci_remove_node(struct pci_dev *pdev)
+{
+}
+#endif /* CONFIG_PCI_DYNAMIC_OF_NODES */
+
 #ifdef CONFIG_PCIEAER
 void pci_no_aer(void);
 void pci_aer_init(struct pci_dev *dev);
drivers/pci/remove.c +1 -0
index 4c54c75050dc..0eaa9d9a3609 100644
@@ -23,6 +23,7 @@ static void pci_stop_dev(struct pci_dev *dev)
 		device_release_driver(&dev->dev);
 		pci_proc_detach_device(dev);
 		pci_remove_sysfs_dev_files(dev);
+		of_pci_remove_node(dev);
 
 		pci_dev_assign_added(dev, false);
 	}
Signature · 1 line
2.17.1
#5 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

Quoted 14 lines · Lizhi Hou
drivers/pci/of_property.c new file +14 -0
+};
+
+static int of_pci_prop_device_type(struct pci_dev *pdev,
+				   struct of_changeset *ocs,
+				   struct device_node *np)
+{
+	return of_changeset_add_prop_string(ocs, np, "device_type", "pci");
+}
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",

drivers/pci/of_property.c line 73

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+					 OF_PCI_ADDRESS_CELLS);
+}
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}

Hi Lizhi,

For all these functions, the "pdev" parameter is actually unused.

[snip]

drivers/pci/of_property.c line 188

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+
+static int of_pci_prop_compatible(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	const char *compat_strs[PROP_COMPAT_NUM] = { 0 };
+	int i, ret;
+
+	compat_strs[PROP_COMPAT_PCI_VVVV_DDDD] =
+		kasprintf(GFP_KERNEL, "pci%x,%x", pdev->vendor, pdev->device);

Maybe it should be better to use "pci%04x,%04x" to keep the existing naming.

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
author
#6 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

Quoted 29 lines · Clément Léger
Quoted 14 lines · Lizhi Hou
drivers/pci/of_property.c new file +14 -0
+};
+
+static int of_pci_prop_device_type(struct pci_dev *pdev,
+				   struct of_changeset *ocs,
+				   struct device_node *np)
+{
+	return of_changeset_add_prop_string(ocs, np, "device_type", "pci");
+}
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",

drivers/pci/of_property.c line 73

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+					 OF_PCI_ADDRESS_CELLS);
+}
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}

Hi Lizhi,

Quoted 3 lines

For all these functions, the "pdev" parameter is actually unused.

[snip]

Ok. I will remove unused pdev.

Quoted 13 lines

drivers/pci/of_property.c line 188

Quoted 11 lines
drivers/pci/of_property.c new file +10 -0
+
+static int of_pci_prop_compatible(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	const char *compat_strs[PROP_COMPAT_NUM] = { 0 };
+	int i, ret;
+
+	compat_strs[PROP_COMPAT_PCI_VVVV_DDDD] =
+		kasprintf(GFP_KERNEL, "pci%x,%x", pdev->vendor, pdev->device);

Maybe it should be better to use "pci%04x,%04x" to keep the existing naming.

Based on https://www.devicetree.org/open-firmware/bindings/pci/pci2_1.pdf, "pci%x,%x" should be used?

"name" Based on the PCI Class Code register, pick a name from Table 1. If none apply, generate a name of the form pciVVVV,DDDD as described below under "compatible".

VVVV, DDDD, SSSS, ssss and RR are lower-case ASCII hexadecimal numbers without leading zeroes.

Thanks,

Lizhi

Quoted 1 line
#7 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

Le 2023-01-03 19:44, Lizhi Hou a écrit :

Quoted 52 lines · Clément Léger
Quoted 29 lines
Quoted 14 lines · Lizhi Hou
drivers/pci/of_property.c new file +14 -0
+};
+
+static int of_pci_prop_device_type(struct pci_dev *pdev,
+				   struct of_changeset *ocs,
+				   struct device_node *np)
+{
+	return of_changeset_add_prop_string(ocs, np, "device_type", "pci");
+}
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",

drivers/pci/of_property.c line 73

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+					 OF_PCI_ADDRESS_CELLS);
+}
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}

Hi Lizhi,

Quoted 3 lines

For all these functions, the "pdev" parameter is actually unused.

[snip]

Ok. I will remove unused pdev.

Quoted 13 lines

drivers/pci/of_property.c line 188

Quoted 11 lines
drivers/pci/of_property.c new file +10 -0
+
+static int of_pci_prop_compatible(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	const char *compat_strs[PROP_COMPAT_NUM] = { 0 };
+	int i, ret;
+
+	compat_strs[PROP_COMPAT_PCI_VVVV_DDDD] =
+		kasprintf(GFP_KERNEL, "pci%x,%x", pdev->vendor, pdev->device);

Maybe it should be better to use "pci%04x,%04x" to keep the existing naming.

Based on
https://www.devicetree.org/open-firmware/bindings/pci/pci2_1.pdf,
"pci%x,%x" should be used?

Quoted 6 lines

"name" Based on the PCI Class Code register, pick a name from Table 1. If none apply, generate a name of the
form pciVVVV,DDDD as described below under "compatible".

VVVV, DDDD, SSSS, ssss and RR are lower-case ASCII hexadecimal numbers without leading zeroes.

You might be right then ! I just looked at the already existing device-tree which adds leading zeroes. Someone like Rob or Frank might be able to answer on that.

Clément

Quoted 7 lines

Thanks,

Lizhi

Quoted 1 line
#8 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

In reply to Lizhi Hou
Quoted 269 lines · Lizhi Hou

The PCI endpoint device such as Xilinx Alveo PCI card maps the register spaces from multiple hardware peripherals to its PCI BAR. Normally, the PCI core discovers devices and BARs using the PCI enumeration process. There is no infrastructure to discover the hardware peripherals that are present in a PCI device, and which can be accessed through the PCI BARs.

For Alveo PCI card, the card firmware provides a flattened device tree to describe the hardware peripherals on its BARs. The Alveo card driver can load this flattened device tree and leverage device tree framework to generate platform devices for the hardware peripherals eventually.

Apparently, the device tree framework requires a device tree node for the PCI device. Thus, it can generate the device tree nodes for hardware peripherals underneath. Because PCI is self discoverable bus, there might not be a device tree node created for PCI devices. This patch is to add support to generate device tree node for PCI devices.

Added a kernel option. When the option is turned on, the kernel will generate device tree nodes for PCI bridges unconditionally.

Initially, the basic properties are added for the dynamically generated device tree nodes.

---
 drivers/pci/Kconfig         |  12 ++
 drivers/pci/Makefile        |   1 +
 drivers/pci/bus.c           |   2 +
 drivers/pci/msi/irqdomain.c |   6 +-
 drivers/pci/of.c            |  71 ++++++++++++
 drivers/pci/of_property.c   | 222 ++++++++++++++++++++++++++++++++++++
 drivers/pci/pci-driver.c    |   3 +-
 drivers/pci/pci.h           |  19 +++
 drivers/pci/remove.c        |   1 +
 9 files changed, 334 insertions(+), 3 deletions(-)
 create mode 100644 drivers/pci/of_property.c
drivers/pci/Kconfig +12 -0
index 55c028af4bd9..1b2347aa1e5b 100644
@@ -198,6 +198,18 @@ config PCI_HYPERV
 	  The PCI device frontend driver allows the kernel to import arbitrary
 	  PCI devices from a PCI backend to support PCI driver domains.
 
+config PCI_DYNAMIC_OF_NODES
+	bool "Create Devicetree nodes for PCI devices"
+	depends on OF
+	select OF_DYNAMIC
+	help
+	  This option enables support for generating device tree nodes for some
+	  PCI devices. Thus, the driver of this kind can load and overlay
+	  flattened device tree for its downstream devices.
+
+	  Once this option is selected, the device tree nodes will be generated
+	  for all PCI bridges.
+
 choice
 	prompt "PCI Express hierarchy optimization setting"
 	default PCIE_BUS_DEFAULT
drivers/pci/Makefile +1 -0
index 2680e4c92f0a..cc8b4e01e29d 100644
@@ -32,6 +32,7 @@ obj-$(CONFIG_PCI_P2PDMA)	+= p2pdma.o
 obj-$(CONFIG_XEN_PCIDEV_FRONTEND) += xen-pcifront.o
 obj-$(CONFIG_VGA_ARB)		+= vgaarb.o
 obj-$(CONFIG_PCI_DOE)		+= doe.o
+obj-$(CONFIG_PCI_DYNAMIC_OF_NODES) += of_property.o
 
 # Endpoint library must be initialized before its users
 obj-$(CONFIG_PCI_ENDPOINT)	+= endpoint/
drivers/pci/bus.c +2 -0
index 3cef835b375f..8507cc32b61d 100644
@@ -316,6 +316,8 @@ void pci_bus_add_device(struct pci_dev *dev)
 	 */
 	pcibios_bus_add_device(dev);
 	pci_fixup_device(pci_fixup_final, dev);
+	if (pci_is_bridge(dev))
+		of_pci_make_dev_node(dev);
 	pci_create_sysfs_dev_files(dev);
 	pci_proc_attach_device(dev);
 	pci_bridge_d3_update(dev);
drivers/pci/msi/irqdomain.c +4 -2
index e9cf318e6670..eeaf44169bfd 100644
@@ -230,8 +230,10 @@ u32 pci_msi_domain_get_msi_rid(struct irq_domain *domain, struct pci_dev *pdev)
 	pci_for_each_dma_alias(pdev, get_msi_id_cb, &rid);
 
 	of_node = irq_domain_get_of_node(domain);
-	rid = of_node ? of_msi_map_id(&pdev->dev, of_node, rid) :
-			iort_msi_map_id(&pdev->dev, rid);
+	if (of_node && !of_node_check_flag(of_node, OF_DYNAMIC))
+		rid = of_msi_map_id(&pdev->dev, of_node, rid);
+	else
+		rid = iort_msi_map_id(&pdev->dev, rid);
 
 	return rid;
 }
drivers/pci/of.c +71 -0
index 196834ed44fe..cb34a73ac8a3 100644
@@ -469,6 +469,8 @@ static int of_irq_parse_pci(const struct pci_dev *pdev, struct of_phandle_args *
 		} else {
 			/* We found a P2P bridge, check if it has a node */
 			ppnode = pci_device_to_OF_node(ppdev);
+			if (ppnode && of_node_check_flag(ppnode, OF_DYNAMIC))
+				ppnode = NULL;
 		}
 
 		/*
@@ -599,6 +601,75 @@ int devm_of_pci_bridge_init(struct device *dev, struct pci_host_bridge *bridge)
 	return pci_parse_request_of_pci_ranges(dev, bridge);
 }
 
+#if IS_ENABLED(CONFIG_PCI_DYNAMIC_OF_NODES)
+
+void of_pci_remove_node(struct pci_dev *pdev)
+{
+	struct device_node *np;
+
+	np = pci_device_to_OF_node(pdev);
+	if (!np || !of_node_check_flag(np, OF_DYNAMIC))
+		return;
+	pdev->dev.of_node = NULL;
+
+	of_destroy_node(np);
+}
+
+void of_pci_make_dev_node(struct pci_dev *pdev)
+{
+	struct device_node *ppnode, *np = NULL;
+	const char *pci_type = "dev";
+	struct of_changeset *cset;
+	const char *name;
+	int ret;
+
+	/*
+	 * If there is already a device tree node linked to this device,
+	 * return immediately.
+	 */
+	if (pci_device_to_OF_node(pdev))
+		return;
+
+	/* Check if there is device tree node for parent device */
+	if (!pdev->bus->self)
+		ppnode = pdev->bus->dev.of_node;
+	else
+		ppnode = pdev->bus->self->dev.of_node;
+	if (!ppnode)
+		return;
+
+	if (pci_is_bridge(pdev))
+		pci_type = "pci";
+
+	name = kasprintf(GFP_KERNEL, "%s@%x,%x", pci_type,
+			 PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn));
+	if (!name)
+		goto failed;
+
+	np = of_create_node(ppnode, name, &cset);
+	if (!np)
+		goto failed;
+
+	ret = of_pci_add_properties(pdev, cset, np);
+	if (ret)
+		goto failed;
+
+	ret = of_changeset_apply(cset);
+	if (ret)
+		goto failed;
+
+	pdev->dev.of_node = np;
+	kfree(name);
+
+	return;
+
+failed:
+	if (np)
+		of_destroy_node(np);
+	kfree(name);
+}
+#endif
+
 #endif /* CONFIG_PCI */
 
 /**
drivers/pci/of_property.c new file +69 -0
new file mode 100644
index 000000000000..05c8ca05a71b
@@ -0,0 +1,222 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2022, Advanced Micro Devices, Inc.
+ */
+
+#include <linux/pci.h>
+#include <linux/of.h>
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <asm/unaligned.h>
+#include "pci.h"
+
+#define OF_PCI_ADDRESS_CELLS		3
+#define OF_PCI_SIZE_CELLS		2
+
+struct of_pci_addr_pair {
+	u32		phys_addr[OF_PCI_ADDRESS_CELLS];
+	u32		size[OF_PCI_SIZE_CELLS];
+};
+
+struct of_pci_range {
+	u32		child_addr[OF_PCI_ADDRESS_CELLS];
+	u32		parent_addr[OF_PCI_ADDRESS_CELLS];
+	u32		size[OF_PCI_SIZE_CELLS];
+};
+
+#define OF_PCI_ADDR_SPACE_CONFIG	0x0
+#define OF_PCI_ADDR_SPACE_IO		0x1
+#define OF_PCI_ADDR_SPACE_MEM32		0x2
+#define OF_PCI_ADDR_SPACE_MEM64		0x3
+
+#define OF_PCI_ADDR_FIELD_NONRELOC	BIT(31)
+#define OF_PCI_ADDR_FIELD_SS		GENMASK(25, 24)
+#define OF_PCI_ADDR_FIELD_PREFETCH	BIT(30)
+#define OF_PCI_ADDR_FIELD_BUS		GENMASK(23, 16)
+#define OF_PCI_ADDR_FIELD_DEV		GENMASK(15, 11)
+#define OF_PCI_ADDR_FIELD_FUNC		GENMASK(10, 8)
+#define OF_PCI_ADDR_FIELD_REG		GENMASK(7, 0)
+
+#define OF_PCI_ADDR_HI			GENMASK_ULL(63, 32)
+#define OF_PCI_ADDR_LO			GENMASK_ULL(31, 0)
+#define OF_PCI_SIZE_HI			GENMASK_ULL(63, 32)
+#define OF_PCI_SIZE_LO			GENMASK_ULL(31, 0)
+
+enum of_pci_prop_compatible {
+	PROP_COMPAT_PCI_VVVV_DDDD,
+	PROP_COMPAT_PCICLASS_CCSSPP,
+	PROP_COMPAT_PCICLASS_CCSS,
+	PROP_COMPAT_NUM,
+};
+
+static int of_pci_prop_device_type(struct pci_dev *pdev,
+				   struct of_changeset *ocs,
+				   struct device_node *np)
+{
+	return of_changeset_add_prop_string(ocs, np, "device_type", "pci");
+}
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",
+					 OF_PCI_ADDRESS_CELLS);
+}
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)

drivers/pci/of_property.c line 79

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}
+
+static void of_pci_set_address(u32 *prop, u64 addr, u32 flags)
+{
+	prop[0] = flags;
+	put_unaligned(addr, &prop[1]);
+}

Here, the put_unaligned() call will assume you want to store a 32bits value since prop is pointing to a u32 array. This can't work.

Moreover, when storing a 64bits address in device-tree properties, they are stored starting with their 32 MSB in the first cell and the 32 LSB in the next cells.

This should probably be something like this:

static void of_pci_set_address(u32 *prop, u64 addr, u32 flags)

{
	prop[0] = flags;
	prop[1] = addr >> 32;
	prop[2] = addr;
}

[snip]

Quoted 15 lines
drivers/pci/of_property.c new file +15 -0
+
+static int of_pci_prop_ranges(struct pci_dev *pdev, struct of_changeset *ocs,
+			      struct device_node *np)
+{
+	struct of_pci_range rp[PCI_BRIDGE_RESOURCE_NUM];
+	struct resource *res;
+	int i = 0, j, ret;
+	u64 val64;
+	u32 flags;
+
+	res = &pdev->resource[PCI_BRIDGE_RESOURCES];
+	for (j = 0; j < PCI_BRIDGE_RESOURCE_NUM; j++) {
+		if (!resource_size(&res[j]))
+			continue;
+

drivers/pci/of_property.c line 126

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+		flags = OF_PCI_ADDR_FIELD_NONRELOC;
+		if (of_pci_get_addr_flags(&res[j], &flags))
+			continue;
+
+		val64 = res[j].start;
+		of_pci_set_address(rp[i].parent_addr, val64, flags);
+		of_pci_set_address(rp[i].child_addr, val64, flags);
+
+		val64 = resource_size(&res[j]);
+		put_unaligned(val64, rp[i].size);

Same problem here, the size is meant to be a 64 bits values but will assume you want to store it using a 32bits pointer.

Quoted 33 lines
drivers/pci/of_property.c new file +33 -0
+
+		i++;
+	}
+
+	ret = of_changeset_add_prop_u32_array(ocs, np, "ranges", (u32 *)rp,
+					      i * sizeof(*rp) / sizeof(u32));
+
+	return ret;
+}
+
+static int of_pci_prop_reg(struct pci_dev *pdev, struct of_changeset *ocs,
+			   struct device_node *np)
+{
+	struct of_pci_addr_pair *reg;
+	int i = 1, resno, ret = 0;
+	u32 reg_val, base_addr;
+	resource_size_t sz;
+
+	reg = kzalloc(sizeof(*reg) * (PCI_STD_NUM_BARS + 1), GFP_KERNEL);
+	if (!reg)
+		return -ENOMEM;
+
+	reg_val = FIELD_PREP(OF_PCI_ADDR_FIELD_SS, OF_PCI_ADDR_SPACE_CONFIG) |
+		FIELD_PREP(OF_PCI_ADDR_FIELD_BUS, pdev->bus->number) |
+		FIELD_PREP(OF_PCI_ADDR_FIELD_DEV, PCI_SLOT(pdev->devfn)) |
+		FIELD_PREP(OF_PCI_ADDR_FIELD_FUNC, PCI_FUNC(pdev->devfn));
+	of_pci_set_address(reg[0].phys_addr, 0, reg_val);
+
+	base_addr = PCI_BASE_ADDRESS_0;
+	for (resno = PCI_STD_RESOURCES; resno <= PCI_STD_RESOURCE_END;
+	     resno++, base_addr += 4) {
+		sz = pci_resource_len(pdev, resno);
+		if (!sz)

drivers/pci/of_property.c line 169

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+			continue;
+
+		ret = of_pci_get_addr_flags(&pdev->resource[resno], &reg_val);
+		if (!ret)
+			continue;
+
+		reg_val &= ~OF_PCI_ADDR_FIELD_REG;
+		reg_val |= FIELD_PREP(OF_PCI_ADDR_FIELD_REG, base_addr);
+		of_pci_set_address(reg[i].phys_addr, 0, reg_val);
+		put_unaligned((u64)sz, reg[i].size);

And same here. Something like this should probably be used (untested):

static void of_pci_set_size(u32 *prop, u64 size)

{
	prop[0] = size >> 32;
	prop[1] = size;
}
Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
author
#9 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

Quoted 291 lines · Clément Léger
Quoted 269 lines · Lizhi Hou

The PCI endpoint device such as Xilinx Alveo PCI card maps the register spaces from multiple hardware peripherals to its PCI BAR. Normally, the PCI core discovers devices and BARs using the PCI enumeration process. There is no infrastructure to discover the hardware peripherals that are present in a PCI device, and which can be accessed through the PCI BARs.

For Alveo PCI card, the card firmware provides a flattened device tree to describe the hardware peripherals on its BARs. The Alveo card driver can load this flattened device tree and leverage device tree framework to generate platform devices for the hardware peripherals eventually.

Apparently, the device tree framework requires a device tree node for the PCI device. Thus, it can generate the device tree nodes for hardware peripherals underneath. Because PCI is self discoverable bus, there might not be a device tree node created for PCI devices. This patch is to add support to generate device tree node for PCI devices.

Added a kernel option. When the option is turned on, the kernel will generate device tree nodes for PCI bridges unconditionally.

Initially, the basic properties are added for the dynamically generated device tree nodes.

---
  drivers/pci/Kconfig         |  12 ++
  drivers/pci/Makefile        |   1 +
  drivers/pci/bus.c           |   2 +
  drivers/pci/msi/irqdomain.c |   6 +-
  drivers/pci/of.c            |  71 ++++++++++++
  drivers/pci/of_property.c   | 222 ++++++++++++++++++++++++++++++++++++
  drivers/pci/pci-driver.c    |   3 +-
  drivers/pci/pci.h           |  19 +++
  drivers/pci/remove.c        |   1 +
  9 files changed, 334 insertions(+), 3 deletions(-)

create mode 100644 drivers/pci/of_property.c

drivers/pci/Kconfig +12 -0
index 55c028af4bd9..1b2347aa1e5b 100644
@@ -198,6 +198,18 @@ config PCI_HYPERV
  	  The PCI device frontend driver allows the kernel to import arbitrary
  	  PCI devices from a PCI backend to support PCI driver domains.
 
+config PCI_DYNAMIC_OF_NODES
+	bool "Create Devicetree nodes for PCI devices"
+	depends on OF
+	select OF_DYNAMIC
+	help
+	  This option enables support for generating device tree nodes for some
+	  PCI devices. Thus, the driver of this kind can load and overlay
+	  flattened device tree for its downstream devices.
+
+	  Once this option is selected, the device tree nodes will be generated
+	  for all PCI bridges.
+
  choice
  	prompt "PCI Express hierarchy optimization setting"
  	default PCIE_BUS_DEFAULT
drivers/pci/Makefile +1 -0
index 2680e4c92f0a..cc8b4e01e29d 100644
@@ -32,6 +32,7 @@ obj-$(CONFIG_PCI_P2PDMA)	+= p2pdma.o
  obj-$(CONFIG_XEN_PCIDEV_FRONTEND) += xen-pcifront.o
  obj-$(CONFIG_VGA_ARB)		+= vgaarb.o
  obj-$(CONFIG_PCI_DOE)		+= doe.o
+obj-$(CONFIG_PCI_DYNAMIC_OF_NODES) += of_property.o
 
  # Endpoint library must be initialized before its users
  obj-$(CONFIG_PCI_ENDPOINT)	+= endpoint/
drivers/pci/bus.c +2 -0
index 3cef835b375f..8507cc32b61d 100644
@@ -316,6 +316,8 @@ void pci_bus_add_device(struct pci_dev *dev)
  	 */
  	pcibios_bus_add_device(dev);
  	pci_fixup_device(pci_fixup_final, dev);
+	if (pci_is_bridge(dev))
+		of_pci_make_dev_node(dev);
  	pci_create_sysfs_dev_files(dev);
  	pci_proc_attach_device(dev);
  	pci_bridge_d3_update(dev);
drivers/pci/msi/irqdomain.c +4 -2
index e9cf318e6670..eeaf44169bfd 100644
@@ -230,8 +230,10 @@ u32 pci_msi_domain_get_msi_rid(struct irq_domain *domain, struct pci_dev *pdev)
  	pci_for_each_dma_alias(pdev, get_msi_id_cb, &rid);
 
  	of_node = irq_domain_get_of_node(domain);
-	rid = of_node ? of_msi_map_id(&pdev->dev, of_node, rid) :
-			iort_msi_map_id(&pdev->dev, rid);
+	if (of_node && !of_node_check_flag(of_node, OF_DYNAMIC))
+		rid = of_msi_map_id(&pdev->dev, of_node, rid);
+	else
+		rid = iort_msi_map_id(&pdev->dev, rid);
 
  	return rid;
  }
drivers/pci/of.c +71 -0
index 196834ed44fe..cb34a73ac8a3 100644
@@ -469,6 +469,8 @@ static int of_irq_parse_pci(const struct pci_dev *pdev, struct of_phandle_args *
  		} else {
  			/* We found a P2P bridge, check if it has a node */
  			ppnode = pci_device_to_OF_node(ppdev);
+			if (ppnode && of_node_check_flag(ppnode, OF_DYNAMIC))
+				ppnode = NULL;
  		}
 
  		/*
@@ -599,6 +601,75 @@ int devm_of_pci_bridge_init(struct device *dev, struct pci_host_bridge *bridge)
  	return pci_parse_request_of_pci_ranges(dev, bridge);
  }
 
+#if IS_ENABLED(CONFIG_PCI_DYNAMIC_OF_NODES)
+
+void of_pci_remove_node(struct pci_dev *pdev)
+{
+	struct device_node *np;
+
+	np = pci_device_to_OF_node(pdev);
+	if (!np || !of_node_check_flag(np, OF_DYNAMIC))
+		return;
+	pdev->dev.of_node = NULL;
+
+	of_destroy_node(np);
+}
+
+void of_pci_make_dev_node(struct pci_dev *pdev)
+{
+	struct device_node *ppnode, *np = NULL;
+	const char *pci_type = "dev";
+	struct of_changeset *cset;
+	const char *name;
+	int ret;
+
+	/*
+	 * If there is already a device tree node linked to this device,
+	 * return immediately.
+	 */
+	if (pci_device_to_OF_node(pdev))
+		return;
+
+	/* Check if there is device tree node for parent device */
+	if (!pdev->bus->self)
+		ppnode = pdev->bus->dev.of_node;
+	else
+		ppnode = pdev->bus->self->dev.of_node;
+	if (!ppnode)
+		return;
+
+	if (pci_is_bridge(pdev))
+		pci_type = "pci";
+
+	name = kasprintf(GFP_KERNEL, "%s@%x,%x", pci_type,
+			 PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn));
+	if (!name)
+		goto failed;
+
+	np = of_create_node(ppnode, name, &cset);
+	if (!np)
+		goto failed;
+
+	ret = of_pci_add_properties(pdev, cset, np);
+	if (ret)
+		goto failed;
+
+	ret = of_changeset_apply(cset);
+	if (ret)
+		goto failed;
+
+	pdev->dev.of_node = np;
+	kfree(name);
+
+	return;
+
+failed:
+	if (np)
+		of_destroy_node(np);
+	kfree(name);
+}
+#endif
+
  #endif /* CONFIG_PCI */
 
  /**
drivers/pci/of_property.c new file +69 -0
new file mode 100644
index 000000000000..05c8ca05a71b
@@ -0,0 +1,222 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2022, Advanced Micro Devices, Inc.
+ */
+
+#include <linux/pci.h>
+#include <linux/of.h>
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <asm/unaligned.h>
+#include "pci.h"
+
+#define OF_PCI_ADDRESS_CELLS		3
+#define OF_PCI_SIZE_CELLS		2
+
+struct of_pci_addr_pair {
+	u32		phys_addr[OF_PCI_ADDRESS_CELLS];
+	u32		size[OF_PCI_SIZE_CELLS];
+};
+
+struct of_pci_range {
+	u32		child_addr[OF_PCI_ADDRESS_CELLS];
+	u32		parent_addr[OF_PCI_ADDRESS_CELLS];
+	u32		size[OF_PCI_SIZE_CELLS];
+};
+
+#define OF_PCI_ADDR_SPACE_CONFIG	0x0
+#define OF_PCI_ADDR_SPACE_IO		0x1
+#define OF_PCI_ADDR_SPACE_MEM32		0x2
+#define OF_PCI_ADDR_SPACE_MEM64		0x3
+
+#define OF_PCI_ADDR_FIELD_NONRELOC	BIT(31)
+#define OF_PCI_ADDR_FIELD_SS		GENMASK(25, 24)
+#define OF_PCI_ADDR_FIELD_PREFETCH	BIT(30)
+#define OF_PCI_ADDR_FIELD_BUS		GENMASK(23, 16)
+#define OF_PCI_ADDR_FIELD_DEV		GENMASK(15, 11)
+#define OF_PCI_ADDR_FIELD_FUNC		GENMASK(10, 8)
+#define OF_PCI_ADDR_FIELD_REG		GENMASK(7, 0)
+
+#define OF_PCI_ADDR_HI			GENMASK_ULL(63, 32)
+#define OF_PCI_ADDR_LO			GENMASK_ULL(31, 0)
+#define OF_PCI_SIZE_HI			GENMASK_ULL(63, 32)
+#define OF_PCI_SIZE_LO			GENMASK_ULL(31, 0)
+
+enum of_pci_prop_compatible {
+	PROP_COMPAT_PCI_VVVV_DDDD,
+	PROP_COMPAT_PCICLASS_CCSSPP,
+	PROP_COMPAT_PCICLASS_CCSS,
+	PROP_COMPAT_NUM,
+};
+
+static int of_pci_prop_device_type(struct pci_dev *pdev,
+				   struct of_changeset *ocs,
+				   struct device_node *np)
+{
+	return of_changeset_add_prop_string(ocs, np, "device_type", "pci");
+}
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",
+					 OF_PCI_ADDRESS_CELLS);
+}
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)

drivers/pci/of_property.c line 79

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}
+
+static void of_pci_set_address(u32 *prop, u64 addr, u32 flags)
+{
+	prop[0] = flags;
+	put_unaligned(addr, &prop[1]);
+}

Here, the put_unaligned() call will assume you want to store a 32bits value since prop is pointing to a u32 array. This can't work.

Moreover, when storing a 64bits address in device-tree properties, they are stored starting with their 32 MSB in the first cell and the 32 LSB in the next cells.

This should probably be something like this:

Quoted 8 lines

static void of_pci_set_address(u32 *prop, u64 addr, u32 flags)

{
	prop[0] = flags;
	prop[1] = addr >> 32;
	prop[2] = addr;
}

[snip]

Thanks for pointing this out. And I should add (__be64 *) as below.

 put_unaligned(addr, (__be64 *)&prop[1]);

Quoted 20 lines
Quoted 20 lines
drivers/pci/of_property.c new file +19 -0
+
+static int of_pci_prop_ranges(struct pci_dev *pdev, struct of_changeset *ocs,
+			      struct device_node *np)
+{
+	struct of_pci_range rp[PCI_BRIDGE_RESOURCE_NUM];
+	struct resource *res;
+	int i = 0, j, ret;
+	u64 val64;
+	u32 flags;
+
+	res = &pdev->resource[PCI_BRIDGE_RESOURCES];
+	for (j = 0; j < PCI_BRIDGE_RESOURCE_NUM; j++) {
+		if (!resource_size(&res[j]))
+			continue;
+
+		flags = OF_PCI_ADDR_FIELD_NONRELOC;
+		if (of_pci_get_addr_flags(&res[j], &flags))
+			continue;
+
Quoted 8 lines

drivers/pci/of_property.c line 126

Quoted 6 lines
drivers/pci/of_property.c new file +6 -0
+		val64 = res[j].start;
+		of_pci_set_address(rp[i].parent_addr, val64, flags);
+		of_pci_set_address(rp[i].child_addr, val64, flags);
+
+		val64 = resource_size(&res[j]);
+		put_unaligned(val64, rp[i].size);

Same problem here, the size is meant to be a 64 bits values but will assume you want to store it using a 32bits pointer.

Will change to: put_unaligned(val64, (__be64 *)rp[i].size);

Quoted 37 lines
Quoted 37 lines
drivers/pci/of_property.c new file +36 -0
+
+		i++;
+	}
+
+	ret = of_changeset_add_prop_u32_array(ocs, np, "ranges", (u32 *)rp,
+					      i * sizeof(*rp) / sizeof(u32));
+
+	return ret;
+}
+
+static int of_pci_prop_reg(struct pci_dev *pdev, struct of_changeset *ocs,
+			   struct device_node *np)
+{
+	struct of_pci_addr_pair *reg;
+	int i = 1, resno, ret = 0;
+	u32 reg_val, base_addr;
+	resource_size_t sz;
+
+	reg = kzalloc(sizeof(*reg) * (PCI_STD_NUM_BARS + 1), GFP_KERNEL);
+	if (!reg)
+		return -ENOMEM;
+
+	reg_val = FIELD_PREP(OF_PCI_ADDR_FIELD_SS, OF_PCI_ADDR_SPACE_CONFIG) |
+		FIELD_PREP(OF_PCI_ADDR_FIELD_BUS, pdev->bus->number) |
+		FIELD_PREP(OF_PCI_ADDR_FIELD_DEV, PCI_SLOT(pdev->devfn)) |
+		FIELD_PREP(OF_PCI_ADDR_FIELD_FUNC, PCI_FUNC(pdev->devfn));
+	of_pci_set_address(reg[0].phys_addr, 0, reg_val);
+
+	base_addr = PCI_BASE_ADDRESS_0;
+	for (resno = PCI_STD_RESOURCES; resno <= PCI_STD_RESOURCE_END;
+	     resno++, base_addr += 4) {
+		sz = pci_resource_len(pdev, resno);
+		if (!sz)
+			continue;
+
+		ret = of_pci_get_addr_flags(&pdev->resource[resno], &reg_val);
Quoted 8 lines

drivers/pci/of_property.c line 169

Quoted 7 lines
drivers/pci/of_property.c new file +7 -0
+		if (!ret)
+			continue;
+
+		reg_val &= ~OF_PCI_ADDR_FIELD_REG;
+		reg_val |= FIELD_PREP(OF_PCI_ADDR_FIELD_REG, base_addr);
+		of_pci_set_address(reg[i].phys_addr, 0, reg_val);
+		put_unaligned((u64)sz, reg[i].size);

And same here. Something like this should probably be used (untested):

will change to: put_unaligned((u64)sz, (__be64 *)reg[i].size);

Thanks,

Lizhi

Quoted 8 lines

static void of_pci_set_size(u32 *prop, u64 size)

{
	prop[0] = size >> 32;
	prop[1] = size;
}
author
#10 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

Quoted 300 lines · Lizhi Hou
Quoted 298 lines · Clément Léger
Quoted 269 lines · Lizhi Hou

The PCI endpoint device such as Xilinx Alveo PCI card maps the register spaces from multiple hardware peripherals to its PCI BAR. Normally, the PCI core discovers devices and BARs using the PCI enumeration process. There is no infrastructure to discover the hardware peripherals that are present in a PCI device, and which can be accessed through the PCI BARs.

For Alveo PCI card, the card firmware provides a flattened device tree to describe the hardware peripherals on its BARs. The Alveo card driver can load this flattened device tree and leverage device tree framework to generate platform devices for the hardware peripherals eventually.

Apparently, the device tree framework requires a device tree node for the PCI device. Thus, it can generate the device tree nodes for hardware peripherals underneath. Because PCI is self discoverable bus, there might not be a device tree node created for PCI devices. This patch is to add support to generate device tree node for PCI devices.

Added a kernel option. When the option is turned on, the kernel will generate device tree nodes for PCI bridges unconditionally.

Initially, the basic properties are added for the dynamically generated device tree nodes.

---
  drivers/pci/Kconfig         |  12 ++
  drivers/pci/Makefile        |   1 +
  drivers/pci/bus.c           |   2 +
  drivers/pci/msi/irqdomain.c |   6 +-
  drivers/pci/of.c            |  71 ++++++++++++
  drivers/pci/of_property.c   | 222
++++++++++++++++++++++++++++++++++++
  drivers/pci/pci-driver.c    |   3 +-
  drivers/pci/pci.h           |  19 +++
  drivers/pci/remove.c        |   1 +

  9 files changed, 334 insertions(+), 3 deletions(-)

  create mode 100644 drivers/pci/of_property.c

drivers/pci/Kconfig +0 -0
index 55c028af4bd9..1b2347aa1e5b 100644
@@ -198,6 +198,18 @@ config PCI_HYPERV

        The PCI device frontend driver allows the kernel to import arbitrary PCI devices from a PCI backend to support PCI driver domains. +config PCI_DYNAMIC_OF_NODES
+    bool "Create Devicetree nodes for PCI devices"
+    depends on OF
+    select OF_DYNAMIC
+    help
+      This option enables support for generating device tree nodes for some
+      PCI devices. Thus, the driver of this kind can load and overlay +      flattened device tree for its downstream devices.
+
+      Once this option is selected, the device tree nodes will be generated
+      for all PCI bridges.
+
  choice
      prompt "PCI Express hierarchy optimization setting" default PCIE_BUS_DEFAULT

drivers/pci/Makefile +0 -0
index 2680e4c92f0a..cc8b4e01e29d 100644
@@ -32,6 +32,7 @@ obj-$(CONFIG_PCI_P2PDMA)    += p2pdma.o

  obj-$(CONFIG_XEN_PCIDEV_FRONTEND) += xen-pcifront.o obj-$(CONFIG_VGA_ARB)        += vgaarb.o obj-$(CONFIG_PCI_DOE)        += doe.o +obj-$(CONFIG_PCI_DYNAMIC_OF_NODES) += of_property.o # Endpoint library must be initialized before its users obj-$(CONFIG_PCI_ENDPOINT)    += endpoint/

drivers/pci/bus.c +2 -0
index 3cef835b375f..8507cc32b61d 100644
@@ -316,6 +316,8 @@ void pci_bus_add_device(struct pci_dev *dev)
        */
       pcibios_bus_add_device(dev);
       pci_fixup_device(pci_fixup_final, dev);
+    if (pci_is_bridge(dev))
+        of_pci_make_dev_node(dev);
       pci_create_sysfs_dev_files(dev);
       pci_proc_attach_device(dev);
       pci_bridge_d3_update(dev);
drivers/pci/msi/irqdomain.c +4 -2
index e9cf318e6670..eeaf44169bfd 100644
@@ -230,8 +230,10 @@ u32 pci_msi_domain_get_msi_rid(struct 
 irq_domain *domain, struct pci_dev *pdev)
       pci_for_each_dma_alias(pdev, get_msi_id_cb, &rid);
         of_node = irq_domain_get_of_node(domain);
-    rid = of_node ? of_msi_map_id(&pdev->dev, of_node, rid) :
-            iort_msi_map_id(&pdev->dev, rid);
+    if (of_node && !of_node_check_flag(of_node, OF_DYNAMIC))
+        rid = of_msi_map_id(&pdev->dev, of_node, rid);
+    else
+        rid = iort_msi_map_id(&pdev->dev, rid);
         return rid;
   }
drivers/pci/of.c +70 -0
index 196834ed44fe..cb34a73ac8a3 100644
@@ -469,6 +469,8 @@ static int of_irq_parse_pci(const struct pci_dev 
 *pdev, struct of_phandle_args *
           } else {
               /* We found a P2P bridge, check if it has a node */
               ppnode = pci_device_to_OF_node(ppdev);
+            if (ppnode && of_node_check_flag(ppnode, OF_DYNAMIC))
+                ppnode = NULL;
           }
             /*
@@ -599,6 +601,75 @@ int devm_of_pci_bridge_init(struct device *dev, 
 struct pci_host_bridge *bridge)
       return pci_parse_request_of_pci_ranges(dev, bridge);
   }
   +#if IS_ENABLED(CONFIG_PCI_DYNAMIC_OF_NODES)
+
+void of_pci_remove_node(struct pci_dev *pdev)
+{
+    struct device_node *np;
+
+    np = pci_device_to_OF_node(pdev);
+    if (!np || !of_node_check_flag(np, OF_DYNAMIC))
+        return;
+    pdev->dev.of_node = NULL;
+
+    of_destroy_node(np);
+}
+
+void of_pci_make_dev_node(struct pci_dev *pdev)
+{
+    struct device_node *ppnode, *np = NULL;
+    const char *pci_type = "dev";
+    struct of_changeset *cset;
+    const char *name;
+    int ret;
+
+    /*
+     * If there is already a device tree node linked to this device,
+     * return immediately.
+     */
+    if (pci_device_to_OF_node(pdev))
+        return;
+
+    /* Check if there is device tree node for parent device */
+    if (!pdev->bus->self)
+        ppnode = pdev->bus->dev.of_node;
+    else
+        ppnode = pdev->bus->self->dev.of_node;
+    if (!ppnode)
+        return;
+
+    if (pci_is_bridge(pdev))
+        pci_type = "pci";
+
+    name = kasprintf(GFP_KERNEL, "%s@%x,%x", pci_type,
+             PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn));
+    if (!name)
+        goto failed;
+
+    np = of_create_node(ppnode, name, &cset);
+    if (!np)
+        goto failed;
+
+    ret = of_pci_add_properties(pdev, cset, np);
+    if (ret)
+        goto failed;
+
+    ret = of_changeset_apply(cset);
+    if (ret)
+        goto failed;
+
+    pdev->dev.of_node = np;
+    kfree(name);
+
+    return;
+
+failed:
+    if (np)
+        of_destroy_node(np);
+    kfree(name);
+}
+#endif
+
   #endif /* CONFIG_PCI */
     /**
drivers/pci/of_property.c new file +56 -0
new file mode 100644
index 000000000000..05c8ca05a71b
@@ -0,0 +1,222 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2022, Advanced Micro Devices, Inc.
+ */
+
+#include <linux/pci.h>
+#include <linux/of.h>
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <asm/unaligned.h>
+#include "pci.h"
+
+#define OF_PCI_ADDRESS_CELLS        3
+#define OF_PCI_SIZE_CELLS        2
+
+struct of_pci_addr_pair {
+    u32        phys_addr[OF_PCI_ADDRESS_CELLS];
+    u32        size[OF_PCI_SIZE_CELLS];
+};
+
+struct of_pci_range {
+    u32        child_addr[OF_PCI_ADDRESS_CELLS];
+    u32        parent_addr[OF_PCI_ADDRESS_CELLS];
+    u32        size[OF_PCI_SIZE_CELLS];
+};
+
+#define OF_PCI_ADDR_SPACE_CONFIG    0x0
+#define OF_PCI_ADDR_SPACE_IO        0x1
+#define OF_PCI_ADDR_SPACE_MEM32        0x2
+#define OF_PCI_ADDR_SPACE_MEM64        0x3
+
+#define OF_PCI_ADDR_FIELD_NONRELOC    BIT(31)
+#define OF_PCI_ADDR_FIELD_SS        GENMASK(25, 24)
+#define OF_PCI_ADDR_FIELD_PREFETCH    BIT(30)
+#define OF_PCI_ADDR_FIELD_BUS        GENMASK(23, 16)
+#define OF_PCI_ADDR_FIELD_DEV        GENMASK(15, 11)
+#define OF_PCI_ADDR_FIELD_FUNC        GENMASK(10, 8)
+#define OF_PCI_ADDR_FIELD_REG        GENMASK(7, 0)
+
+#define OF_PCI_ADDR_HI            GENMASK_ULL(63, 32)
+#define OF_PCI_ADDR_LO            GENMASK_ULL(31, 0)
+#define OF_PCI_SIZE_HI            GENMASK_ULL(63, 32)
+#define OF_PCI_SIZE_LO            GENMASK_ULL(31, 0)
+
+enum of_pci_prop_compatible {
+    PROP_COMPAT_PCI_VVVV_DDDD,
+    PROP_COMPAT_PCICLASS_CCSSPP,
+    PROP_COMPAT_PCICLASS_CCSS,
+    PROP_COMPAT_NUM,
+};
+
+static int of_pci_prop_device_type(struct pci_dev *pdev,
+                   struct of_changeset *ocs,
+                   struct device_node *np)
+{
+    return of_changeset_add_prop_string(ocs, np, "device_type", 

"pci");

drivers/pci/of_property.c new file +13 -0
+}
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+                     struct of_changeset *ocs,
+                     struct device_node *np)
+{
+    return of_changeset_add_prop_u32(ocs, np, "#address_cells",
+                     OF_PCI_ADDRESS_CELLS);
+}
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+                  struct of_changeset *ocs,
+                  struct device_node *np)

drivers/pci/of_property.c line 79

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+{
+    return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+                     OF_PCI_SIZE_CELLS);
+}
+
+static void of_pci_set_address(u32 *prop, u64 addr, u32 flags)
+{
+    prop[0] = flags;
+    put_unaligned(addr, &prop[1]);
+}

Here, the put_unaligned() call will assume you want to store a 32bits value since prop is pointing to a u32 array. This can't work.

Moreover, when storing a 64bits address in device-tree properties, they are stored starting with their 32 MSB in the first cell and the 32 LSB in the next cells.

This should probably be something like this:

static void of_pci_set_address(u32 *prop, u64 addr, u32 flags) {
    prop[0] = flags;
    prop[1] = addr >> 32;
    prop[2] = addr;
}

Quoted 5 lines
Quoted 1 line

[snip]

Thanks for pointing this out. And I should add (__be64 *) as below.

 put_unaligned(addr, (__be64 *)&prop[1]);

I looked into it a bit more. You are correct. put_unaligned() does not work here. I re-write it as below:

        prop[1] = FIELD_GET(OF_PCI_ADDR_HI, addr); prop[2] = FIELD_GET(OF_PCI_ADDR_LO, addr);

Similar changes will be made for "size"

Thanks,

Lizhi

Quoted 95 lines
Quoted 22 lines
Quoted 22 lines
drivers/pci/of_property.c new file +2 -0
+
+static int of_pci_prop_ranges(struct pci_dev *pdev, struct 

of_changeset *ocs,
+                  struct device_node *np)

drivers/pci/of_property.c new file +16 -0
+{
+    struct of_pci_range rp[PCI_BRIDGE_RESOURCE_NUM];
+    struct resource *res;
+    int i = 0, j, ret;
+    u64 val64;
+    u32 flags;
+
+    res = &pdev->resource[PCI_BRIDGE_RESOURCES];
+    for (j = 0; j < PCI_BRIDGE_RESOURCE_NUM; j++) {
+        if (!resource_size(&res[j]))
+            continue;
+
+        flags = OF_PCI_ADDR_FIELD_NONRELOC;
+        if (of_pci_get_addr_flags(&res[j], &flags))
+            continue;
+
Quoted 8 lines
Quoted 6 lines

+        val64 = res[j].start;
+        of_pci_set_address(rp[i].parent_addr, val64, flags); +        of_pci_set_address(rp[i].child_addr, val64, flags); +
+        val64 = resource_size(&res[j]);
+        put_unaligned(val64, rp[i].size);

Same problem here, the size is meant to be a 64 bits values but will assume you want to store it using a 32bits pointer.

Will change to: put_unaligned(val64, (__be64 *)rp[i].size);

Quoted 40 lines
Quoted 40 lines
drivers/pci/of_property.c new file +5 -0
+
+        i++;
+    }
+
+    ret = of_changeset_add_prop_u32_array(ocs, np, "ranges", (u32 

*)rp,
+                          i * sizeof(*rp) / sizeof(u32));

drivers/pci/of_property.c new file +5 -0
+
+    return ret;
+}
+
+static int of_pci_prop_reg(struct pci_dev *pdev, struct 

of_changeset *ocs,
+               struct device_node *np)

drivers/pci/of_property.c new file +11 -0
+{
+    struct of_pci_addr_pair *reg;
+    int i = 1, resno, ret = 0;
+    u32 reg_val, base_addr;
+    resource_size_t sz;
+
+    reg = kzalloc(sizeof(*reg) * (PCI_STD_NUM_BARS + 1), GFP_KERNEL);
+    if (!reg)
+        return -ENOMEM;
+
+    reg_val = FIELD_PREP(OF_PCI_ADDR_FIELD_SS, 

OF_PCI_ADDR_SPACE_CONFIG) |
+        FIELD_PREP(OF_PCI_ADDR_FIELD_BUS, pdev->bus->number) | +        FIELD_PREP(OF_PCI_ADDR_FIELD_DEV, PCI_SLOT(pdev->devfn)) | +        FIELD_PREP(OF_PCI_ADDR_FIELD_FUNC, PCI_FUNC(pdev->devfn)); +    of_pci_set_address(reg[0].phys_addr, 0, reg_val);

drivers/pci/of_property.c new file +9 -0
+
+    base_addr = PCI_BASE_ADDRESS_0;
+    for (resno = PCI_STD_RESOURCES; resno <= PCI_STD_RESOURCE_END;
+         resno++, base_addr += 4) {
+        sz = pci_resource_len(pdev, resno);
+        if (!sz)
+            continue;
+
+        ret = of_pci_get_addr_flags(&pdev->resource[resno], &reg_val);
Quoted 8 lines
Quoted 7 lines

+        if (!ret)
+            continue;
+
+        reg_val &= ~OF_PCI_ADDR_FIELD_REG;
+        reg_val |= FIELD_PREP(OF_PCI_ADDR_FIELD_REG, base_addr); +        of_pci_set_address(reg[i].phys_addr, 0, reg_val); +        put_unaligned((u64)sz, reg[i].size);

And same here. Something like this should probably be used (untested):

will change to: put_unaligned((u64)sz, (__be64 *)reg[i].size);

Thanks,

Lizhi

Quoted 8 lines

static void of_pci_set_size(u32 *prop, u64 size)
{
    prop[0] = size >> 32;
    prop[1] = size;
}

#11 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

In reply to Lizhi Hou

drivers/pci/of_property.c line 65

Quoted 8 lines · Lizhi Hou
drivers/pci/of_property.c new file +8 -0
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",
+					 OF_PCI_ADDRESS_CELLS);
+}

Hi Lizhi,

This should be #address-cells and not #address_cells.

drivers/pci/of_property.c line 73

Quoted 8 lines
drivers/pci/of_property.c new file +8 -0
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}

Ditto here, this should be #size-cells and not #size_cells. Was this really tested ?

Thanks,

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
author
#12 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

Quoted 15 lines · Clément Léger

drivers/pci/of_property.c line 65

Quoted 8 lines · Lizhi Hou
drivers/pci/of_property.c new file +8 -0
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",
+					 OF_PCI_ADDRESS_CELLS);
+}

Hi Lizhi,

This should be #address-cells and not #address_cells.

Quoted 10 lines

drivers/pci/of_property.c line 73

Quoted 8 lines
drivers/pci/of_property.c new file +8 -0
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}

Ditto here, this should be #size-cells and not #size_cells. Was this really tested ?

I will fix this.  I verified the properties are generated and did not notice the '_' error.

Thanks,

Lizhi

Quoted 3 lines

Thanks,

#13 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

Le 2023-01-04 01:31, Lizhi Hou a écrit :

Quoted 27 lines · Clément Léger
Quoted 15 lines

drivers/pci/of_property.c line 65

Quoted 8 lines · Lizhi Hou
drivers/pci/of_property.c new file +8 -0
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",
+					 OF_PCI_ADDRESS_CELLS);
+}

Hi Lizhi,

This should be #address-cells and not #address_cells.

Quoted 10 lines

drivers/pci/of_property.c line 73

Quoted 8 lines
drivers/pci/of_property.c new file +8 -0
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}

Ditto here, this should be #size-cells and not #size_cells. Was this really tested ?

Quoted 2 lines

I will fix this.  I verified the properties are generated and did not notice the '_' error.

Hi Lizhi,

The most important thing to test is if a driver can correctly retrieve resource of the device using a node added below the created PCI device node. This is resolved using the whole hierarchy ranges, address-cells and size-cells properties.

If not correctly set, this will fail (and it fails).

Clément

Quoted 9 lines

Thanks,

Lizhi

Quoted 3 lines

Thanks,

#14 lore raw

Re: [PATCH V5 2/3] PCI: Create device tree node for selected devices

In reply to Lizhi Hou

Le Thu, 15 Dec 2022 09:30:45 -0800,
Lizhi Hou <[email protected]> a écrit :

Further comments

Quoted 39 lines
drivers/pci/of_property.c new file +33 -0
new file mode 100644
index 000000000000..05c8ca05a71b
@@ -0,0 +1,222 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2022, Advanced Micro Devices, Inc.
+ */
+
+#include <linux/pci.h>
+#include <linux/of.h>
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <asm/unaligned.h>
+#include "pci.h"
+
+#define OF_PCI_ADDRESS_CELLS		3
+#define OF_PCI_SIZE_CELLS		2
+
+struct of_pci_addr_pair {
+	u32		phys_addr[OF_PCI_ADDRESS_CELLS];
+	u32		size[OF_PCI_SIZE_CELLS];
+};
+
+struct of_pci_range {
+	u32		child_addr[OF_PCI_ADDRESS_CELLS];
+	u32		parent_addr[OF_PCI_ADDRESS_CELLS];
+	u32		size[OF_PCI_SIZE_CELLS];
+};
+
+#define OF_PCI_ADDR_SPACE_CONFIG	0x0
+#define OF_PCI_ADDR_SPACE_IO		0x1
+#define OF_PCI_ADDR_SPACE_MEM32		0x2
+#define OF_PCI_ADDR_SPACE_MEM64		0x3
+
+#define OF_PCI_ADDR_FIELD_NONRELOC	BIT(31)
+#define OF_PCI_ADDR_FIELD_SS		GENMASK(25, 24)

drivers/pci/of_property.c line 43

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+#define OF_PCI_ADDR_FIELD_PREFETCH	BIT(30)
+#define OF_PCI_ADDR_FIELD_BUS		GENMASK(23, 16)
+#define OF_PCI_ADDR_FIELD_DEV		GENMASK(15, 11)
+#define OF_PCI_ADDR_FIELD_FUNC		GENMASK(10, 8)
+#define OF_PCI_ADDR_FIELD_REG		GENMASK(7, 0)
+
+#define OF_PCI_ADDR_HI			GENMASK_ULL(63, 32)
+#define OF_PCI_ADDR_LO			GENMASK_ULL(31, 0)
+#define OF_PCI_SIZE_HI			GENMASK_ULL(63, 32)
+#define OF_PCI_SIZE_LO			GENMASK_ULL(31, 0)

This is unused.

Quoted 45 lines
drivers/pci/of_property.c new file +45 -0
+
+enum of_pci_prop_compatible {
+	PROP_COMPAT_PCI_VVVV_DDDD,
+	PROP_COMPAT_PCICLASS_CCSSPP,
+	PROP_COMPAT_PCICLASS_CCSS,
+	PROP_COMPAT_NUM,
+};
+
+static int of_pci_prop_device_type(struct pci_dev *pdev,
+				   struct of_changeset *ocs,
+				   struct device_node *np)
+{
+	return of_changeset_add_prop_string(ocs, np, "device_type", "pci");
+}
+
+static int of_pci_prop_address_cells(struct pci_dev *pdev,
+				     struct of_changeset *ocs,
+				     struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#address_cells",
+					 OF_PCI_ADDRESS_CELLS);
+}
+
+static int of_pci_prop_size_cells(struct pci_dev *pdev,
+				  struct of_changeset *ocs,
+				  struct device_node *np)
+{
+	return of_changeset_add_prop_u32(ocs, np, "#size_cells",
+					 OF_PCI_SIZE_CELLS);
+}
+
+static void of_pci_set_address(u32 *prop, u64 addr, u32 flags)
+{
+	prop[0] = flags;
+	put_unaligned(addr, &prop[1]);
+}
+
+static int of_pci_get_addr_flags(struct resource *res, u32 *flags)
+{
+	u32 ss;
+
+	if (res->flags & IORESOURCE_IO)
+		ss = OF_PCI_ADDR_SPACE_IO;
+	else if (res->flags & IORESOURCE_MEM_64)
+		ss = OF_PCI_ADDR_SPACE_MEM64;

drivers/pci/of_property.c line 98

Quoted 10 lines
drivers/pci/of_property.c new file +10 -0
+	else if (res->flags & IORESOURCE_MEM)
+		ss = OF_PCI_ADDR_SPACE_MEM32;
+	else
+		return -EINVAL;
+
+	*flags &= ~(OF_PCI_ADDR_FIELD_SS | OF_PCI_ADDR_FIELD_PREFETCH);
+	if (res->flags & IORESOURCE_PREFETCH)
+		*flags |= OF_PCI_ADDR_FIELD_PREFETCH;
+
+	*flags |= ss;

This seems wrong, should be:

*flags |= FIELD_PREP(OF_PCI_ADDR_FIELD_SS, ss);

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
author
#15 lore raw

PATCH v5 3/3 PCI: Add PCI quirks to generate device tree node for Xilinx Alveo U50

In reply to Lizhi Hou

The Xilinx Alveo U50 PCI card exposes multiple hardware peripherals on its PCI BAR. The card firmware provides a flattened device tree to describe the hardware peripherals on its BARs. This allows U50 driver to load the flattened device tree and generate the device tree node for hardware peripherals underneath.

To generate device tree node for U50 card, added PCI quirks to call of_pci_make_dev_node() for U50.

---
 drivers/pci/quirks.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/pci/quirks.c +11 -0
index 4944798e75b5..5d76932f59ec 100644
@@ -5956,3 +5956,14 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x56b1, aspm_l1_acceptable_latency
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x56c0, aspm_l1_acceptable_latency);
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x56c1, aspm_l1_acceptable_latency);
 #endif
+
+/*
+ * For PCI device which have multiple downstream devices, its driver may use
+ * a flattened device tree to describe the downstream devices.
+ * To overlay the flattened device tree, the PCI device and all its ancestor
+ * devices need to have device tree nodes on system base device tree. Thus,
+ * before driver probing, it might need to add a device tree node as the final
+ * fixup.
+ */
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_XILINX, 0x5020, of_pci_make_dev_node);
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_XILINX, 0x5021, of_pci_make_dev_node);
Signature · 1 line
2.17.1