Hi Len,
I hope this isn't too late for the .31 merge window. If so, no big deal;
this patchset isn't urgent, it's just a nice cleanup that was inspired
from investigating the recent bug in acpi_pci_bind() (fixed by dacd254).
This patch series eliminates static boot-time binding of ACPI
and PCI devices, and introduces an API to perform this lookup
during runtime.
This change has the following advantages:
- eliminates struct acpi_device vs struct pci_dev lifetime issues
- lays groundwork for eliminating .bind/.unbind from acpi_device_ops
- lays more groundwork for eliminating .start from acpi_device_ops
and thus simplifying ACPI drivers
- whacks out a lot of code
This patchset is based on lenb/test (66c74fa1d4), and has been boot
tested on ia64 and x86. I also performed physical PCI hotplug tests
using acpiphp on ia64. I do not have an acpiphp-capable x86 platform.
I didn't test the changes in the video driver either, as I don't have
the hardware.
v1 -> v2
- rearrange series into a more logical order
- much simpler acpi_is_root_bridge() implementation
- no longer export acpi_pci_find_root()
- no longer leak memory in acpi_get_pci_dev()
- no longer leak references in acpi_pci_unbind/acpi_pci_bind
- convert video driver to use acpi_get_pci_dev()
- kill off acpi_get_physical_pci_device()
- incorporate Bjorn's other comments. :)
---
Alex Chiang (11):
ACPI: kill acpi_get_physical_pci_device()
ACPI: video: convert to acpi_get_pci_dev
ACPI: kill acpi_get_pci_id
PCI Hotplug: acpiphp: convert to acpi_get_pci_dev
ACPI: acpi_pci_unbind should clean up properly after acpi_pci_bind
ACPI: simplify acpi_pci_irq_del_prt() API
ACPI: simplify acpi_pci_irq_add_prt() API
ACPI: eviscerate pci_bind.c
ACPI: Introduce acpi_get_pci_dev()
ACPI: Introduce acpi_is_root_bridge()
ACPI: make acpi_pci_bind() static
drivers/acpi/glue.c | 40 -----
drivers/acpi/pci_bind.c | 315 +++++-------------------------------
drivers/acpi/pci_irq.c | 17 +-
drivers/acpi/pci_root.c | 112 ++++++++++++-
drivers/acpi/video.c | 6 -
drivers/acpi/video_detect.c | 9 +
drivers/pci/hotplug/acpi_pcihp.c | 40 -----
drivers/pci/hotplug/acpiphp_glue.c | 27 +--
include/acpi/acpi_bus.h | 2
include/acpi/acpi_drivers.h | 10 -
include/linux/pci_hotplug.h | 1
11 files changed, 179 insertions(+), 400 deletions(-)
acpi_pci_root_add() explicitly assigns device->ops.bind, and later
calls acpi_pci_bind_root(), which also does the same thing.
We don't need to repeat ourselves; removing the explicit assignment
allows us to make acpi_pci_bind() static.
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/pci_bind.c | 3 ++-
drivers/acpi/pci_root.c | 2 --
include/acpi/acpi_drivers.h | 1 -
3 files changed, 2 insertions(+), 4 deletions(-)
diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
index bc46de3..236765c 100644
--- a/drivers/acpi/pci_bind.c
+++ b/drivers/acpi/pci_bind.c
@@ -44,6 +44,7 @@ struct acpi_pci_data {
struct pci_dev *dev;
};
+static int acpi_pci_bind(struct acpi_device *device);
static int acpi_pci_unbind(struct acpi_device *device);
static void acpi_pci_data_handler(acpi_handle handle, u32 function,
@@ -108,7 +109,7 @@ acpi_status acpi_get_pci_id(acpi_handle handle, struct acpi_pci_id *id)
EXPORT_SYMBOL(acpi_get_pci_id);
-int acpi_pci_bind(struct acpi_device *device)
+static int acpi_pci_bind(struct acpi_device *device)
{
int result = 0;
acpi_status status;
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
index 196f97d..ca8dba3 100644
--- a/drivers/acpi/pci_root.c
+++ b/drivers/acpi/pci_root.c
@@ -386,8 +386,6 @@ static int __devinit acpi_pci_root_add(struct acpi_device *device)
strcpy(acpi_device_class(device), ACPI_PCI_ROOT_CLASS);
device->driver_data = root;
- device->ops.bind = acpi_pci_bind;
-
/*
* All supported architectures that use ACPI have support for
* PCI domains, so we indicate this in _OSC support capabilities.
diff --git a/include/acpi/acpi_drivers.h b/include/acpi/acpi_drivers.h
index 5e8ed3a..bbe9207 100644
--- a/include/acpi/acpi_drivers.h
+++ b/include/acpi/acpi_drivers.h
@@ -98,7 +98,6 @@ void acpi_pci_irq_del_prt(int segment, int bus);
struct pci_bus;
acpi_status acpi_get_pci_id(acpi_handle handle, struct acpi_pci_id *id);
-int acpi_pci_bind(struct acpi_device *device);
int acpi_pci_bind_root(struct acpi_device *device, struct acpi_pci_id *id,
struct pci_bus *bus);
Returns whether an ACPI CA node is a PCI root bridge or not.
This API is generically useful, and shouldn't just be a hotplug function.
The implementation becomes much simpler as well.
Cc: Bjorn Helgaas <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/pci_root.c | 24 ++++++++++++++++++++++
drivers/pci/hotplug/acpi_pcihp.c | 40 ++----------------------------------
drivers/pci/hotplug/acpiphp_glue.c | 2 +-
include/acpi/acpi_bus.h | 1 +
include/linux/pci_hotplug.h | 1 -
5 files changed, 28 insertions(+), 40 deletions(-)
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
index ca8dba3..888cb9f 100644
--- a/drivers/acpi/pci_root.c
+++ b/drivers/acpi/pci_root.c
@@ -142,6 +142,30 @@ acpi_handle acpi_get_pci_rootbridge_handle(unsigned int seg, unsigned int bus)
EXPORT_SYMBOL_GPL(acpi_get_pci_rootbridge_handle);
+/**
+ * acpi_is_root_bridge - determine whether an ACPI CA node is a PCI root bridge
+ * @handle - the ACPI CA node in question.
+ *
+ * Note: we could make this API take a struct acpi_device * instead, but
+ * for now, it's more convenient to operate on an acpi_handle.
+ */
+int acpi_is_root_bridge(acpi_handle handle)
+{
+ int ret;
+ struct acpi_device *device;
+
+ ret = acpi_bus_get_device(handle, &device);
+ if (ret)
+ return 0;
+
+ ret = acpi_match_device_ids(device, root_device_ids);
+ if (ret)
+ return 0;
+ else
+ return 1;
+}
+EXPORT_SYMBOL_GPL(acpi_is_root_bridge);
+
static acpi_status
get_root_bridge_busnr_callback(struct acpi_resource *resource, void *data)
{
diff --git a/drivers/pci/hotplug/acpi_pcihp.c b/drivers/pci/hotplug/acpi_pcihp.c
index fbc63d5..eb15958 100644
--- a/drivers/pci/hotplug/acpi_pcihp.c
+++ b/drivers/pci/hotplug/acpi_pcihp.c
@@ -354,7 +354,7 @@ acpi_status acpi_get_hp_params_from_firmware(struct pci_bus *bus,
status = acpi_run_hpp(handle, hpp);
if (ACPI_SUCCESS(status))
break;
- if (acpi_root_bridge(handle))
+ if (acpi_is_root_bridge(handle))
break;
status = acpi_get_parent(handle, &phandle);
if (ACPI_FAILURE(status))
@@ -428,7 +428,7 @@ int acpi_get_hp_hw_control_from_firmware(struct pci_dev *pdev, u32 flags)
status = acpi_run_oshp(handle);
if (ACPI_SUCCESS(status))
goto got_one;
- if (acpi_root_bridge(handle))
+ if (acpi_is_root_bridge(handle))
break;
chandle = handle;
status = acpi_get_parent(chandle, &handle);
@@ -449,42 +449,6 @@ got_one:
}
EXPORT_SYMBOL(acpi_get_hp_hw_control_from_firmware);
-/* acpi_root_bridge - check to see if this acpi object is a root bridge
- *
- * @handle - the acpi object in question.
- */
-int acpi_root_bridge(acpi_handle handle)
-{
- acpi_status status;
- struct acpi_device_info *info;
- struct acpi_buffer buffer = {ACPI_ALLOCATE_BUFFER, NULL};
- int i;
-
- status = acpi_get_object_info(handle, &buffer);
- if (ACPI_SUCCESS(status)) {
- info = buffer.pointer;
- if ((info->valid & ACPI_VALID_HID) &&
- !strcmp(PCI_ROOT_HID_STRING,
- info->hardware_id.value)) {
- kfree(buffer.pointer);
- return 1;
- }
- if (info->valid & ACPI_VALID_CID) {
- for (i=0; i < info->compatibility_id.count; i++) {
- if (!strcmp(PCI_ROOT_HID_STRING,
- info->compatibility_id.id[i].value)) {
- kfree(buffer.pointer);
- return 1;
- }
- }
- }
- kfree(buffer.pointer);
- }
- return 0;
-}
-EXPORT_SYMBOL_GPL(acpi_root_bridge);
-
-
static int is_ejectable(acpi_handle handle)
{
acpi_status status;
diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c
index 3a6064b..fc6636e 100644
--- a/drivers/pci/hotplug/acpiphp_glue.c
+++ b/drivers/pci/hotplug/acpiphp_glue.c
@@ -1631,7 +1631,7 @@ find_root_bridges(acpi_handle handle, u32 lvl, void *context, void **rv)
{
int *count = (int *)context;
- if (acpi_root_bridge(handle)) {
+ if (acpi_is_root_bridge(handle)) {
acpi_install_notify_handler(handle, ACPI_SYSTEM_NOTIFY,
handle_hotplug_event_bridge, NULL);
(*count)++;
diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h
index 0f50167..494088b 100644
--- a/include/acpi/acpi_bus.h
+++ b/include/acpi/acpi_bus.h
@@ -372,6 +372,7 @@ struct device *acpi_get_physical_pci_device(acpi_handle);
/* helper */
acpi_handle acpi_get_child(acpi_handle, acpi_integer);
+int acpi_is_root_bridge(acpi_handle);
acpi_handle acpi_get_pci_rootbridge_handle(unsigned int, unsigned int);
#define DEVICE_ACPI_HANDLE(dev) ((acpi_handle)((dev)->archdata.acpi_handle))
diff --git a/include/linux/pci_hotplug.h b/include/linux/pci_hotplug.h
index 2099874..a3576ef 100644
--- a/include/linux/pci_hotplug.h
+++ b/include/linux/pci_hotplug.h
@@ -226,7 +226,6 @@ struct hotplug_params {
extern acpi_status acpi_get_hp_params_from_firmware(struct pci_bus *bus,
struct hotplug_params *hpp);
int acpi_get_hp_hw_control_from_firmware(struct pci_dev *dev, u32 flags);
-int acpi_root_bridge(acpi_handle handle);
int acpi_pci_check_ejectable(struct pci_bus *pbus, acpi_handle handle);
int acpi_pci_detect_ejectable(struct pci_bus *pbus);
#endif
Convert an ACPI CA handle to a struct pci_dev.
Performing this lookup dynamically allows us to get rid of the
ACPI-PCI binding code, which:
- eliminates struct acpi_device vs struct pci_dev lifetime issues
- lays more groundwork for eliminating .start from acpi_device_ops
and thus simplifying ACPI drivers
- whacks out a lot of code
This change lays the groundwork for eliminating much of pci_bind.c.
Although pci_root.c may not be the most logical place for this
change, putting it here saves us from having to export acpi_pci_find_root.
Cc: Bjorn Helgaas <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/pci_root.c | 81 +++++++++++++++++++++++++++++++++++++++++++
include/acpi/acpi_drivers.h | 1 +
2 files changed, 82 insertions(+), 0 deletions(-)
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
index 888cb9f..e509991 100644
--- a/drivers/acpi/pci_root.c
+++ b/drivers/acpi/pci_root.c
@@ -329,6 +329,87 @@ static struct acpi_pci_root *acpi_pci_find_root(acpi_handle handle)
return NULL;
}
+struct acpi_handle_node {
+ struct list_head node;
+ acpi_handle handle;
+};
+
+/**
+ * acpi_get_pci_dev - convert ACPI CA handle to struct pci_dev
+ * @handle: the handle in question
+ *
+ * Given an ACPI CA handle, the desired PCI device is located in the
+ * list of PCI devices.
+ *
+ * If the device is found, its reference count is increased and this
+ * function returns a pointer to its data structure. The caller must
+ * decrement the reference count by calling pci_dev_put().
+ * If no device is found, %NULL is returned.
+ */
+struct pci_dev *acpi_get_pci_dev(acpi_handle handle)
+{
+ int dev, fn;
+ unsigned long long adr;
+ acpi_status status;
+ acpi_handle phandle;
+ struct pci_bus *pbus;
+ struct pci_dev *pdev = NULL;
+ struct acpi_handle_node *node, *tmp;
+ struct acpi_pci_root *root;
+ LIST_HEAD(device_list);
+
+ /*
+ * Walk up the ACPI CA namespace until we reach a PCI root bridge.
+ */
+ phandle = handle;
+ while (!acpi_is_root_bridge(phandle)) {
+ node = kzalloc(sizeof(struct acpi_handle_node), GFP_KERNEL);
+ if (!node)
+ goto out;
+
+ INIT_LIST_HEAD(&node->node);
+ node->handle = phandle;
+ list_add(&node->node, &device_list);
+
+ status = acpi_get_parent(phandle, &phandle);
+ if (ACPI_FAILURE(status))
+ goto out;
+ }
+
+ root = acpi_pci_find_root(phandle);
+ if (!root)
+ goto out;
+
+ pbus = root->bus;
+
+ /*
+ * Now, walk back down the PCI device tree until we return to our
+ * original handle. Assumes that everything between the PCI root
+ * bridge and the device we're looking for must be a P2P bridge.
+ */
+ list_for_each_entry(node, &device_list, node) {
+ acpi_handle hnd = node->handle;
+ status = acpi_evaluate_integer(hnd, "_ADR", NULL, &adr);
+ if (ACPI_FAILURE(status))
+ goto out;
+ dev = (adr >> 16) & 0xffff;
+ fn = adr & 0xffff;
+
+ pdev = pci_get_slot(pbus, PCI_DEVFN(dev, fn));
+ if (hnd == handle)
+ break;
+
+ pbus = pdev->subordinate;
+ pci_dev_put(pdev);
+ }
+out:
+ list_for_each_entry_safe(node, tmp, &device_list, node)
+ kfree(node);
+
+ return pdev;
+}
+EXPORT_SYMBOL_GPL(acpi_get_pci_dev);
+
/**
* acpi_pci_osc_control_set - commit requested control to Firmware
* @handle: acpi_handle for the target ACPI object
diff --git a/include/acpi/acpi_drivers.h b/include/acpi/acpi_drivers.h
index bbe9207..1ef529b 100644
--- a/include/acpi/acpi_drivers.h
+++ b/include/acpi/acpi_drivers.h
@@ -97,6 +97,7 @@ void acpi_pci_irq_del_prt(int segment, int bus);
struct pci_bus;
+struct pci_dev *acpi_get_pci_dev(acpi_handle);
acpi_status acpi_get_pci_id(acpi_handle handle, struct acpi_pci_id *id);
int acpi_pci_bind_root(struct acpi_device *device, struct acpi_pci_id *id,
struct pci_bus *bus);
Now that we can dynamically convert an ACPI CA handle to a
struct pci_dev at runtime, there's no need to statically bind
them during boot.
acpi_pci_bind/unbind are vastly simplified, and are only used
to evaluate _PRT methods on P2P bridges and non-bridge children.
This patch also changes the time-space tradeoff ever so slightly.
Looking up the ACPI-PCI binding is never in the performance path, and by
eliminating this caching, we save 24 bytes for each _ADR device in the
ACPI namespace.
This patch lays further groundwork to eventually eliminate
the acpi_driver_ops.bind callback.
Cc: Bjorn Helgaas <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/pci_bind.c | 251 +++++++------------------------------------
drivers/acpi/pci_root.c | 2
include/acpi/acpi_drivers.h | 3 -
3 files changed, 42 insertions(+), 214 deletions(-)
diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
index 236765c..703d2a3 100644
--- a/drivers/acpi/pci_bind.c
+++ b/drivers/acpi/pci_bind.c
@@ -24,12 +24,7 @@
*/
#include <linux/kernel.h>
-#include <linux/module.h>
-#include <linux/init.h>
#include <linux/types.h>
-#include <linux/proc_fs.h>
-#include <linux/spinlock.h>
-#include <linux/pm.h>
#include <linux/pci.h>
#include <linux/acpi.h>
#include <acpi/acpi_bus.h>
@@ -109,240 +104,74 @@ acpi_status acpi_get_pci_id(acpi_handle handle, struct acpi_pci_id *id)
EXPORT_SYMBOL(acpi_get_pci_id);
-static int acpi_pci_bind(struct acpi_device *device)
+static int acpi_pci_unbind(struct acpi_device *device)
{
- int result = 0;
- acpi_status status;
- struct acpi_pci_data *data;
- struct acpi_pci_data *pdata;
- struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
- acpi_handle handle;
-
- if (!device || !device->parent)
- return -EINVAL;
-
- data = kzalloc(sizeof(struct acpi_pci_data), GFP_KERNEL);
- if (!data)
- return -ENOMEM;
-
- status = acpi_get_name(device->handle, ACPI_FULL_PATHNAME, &buffer);
- if (ACPI_FAILURE(status)) {
- kfree(data);
- return -ENODEV;
- }
-
- ACPI_DEBUG_PRINT((ACPI_DB_INFO, "Binding PCI device [%s]...\n",
- (char *)buffer.pointer));
+ struct pci_dev *dev;
- /*
- * Segment & Bus
- * -------------
- * These are obtained via the parent device's ACPI-PCI context.
- */
- status = acpi_get_data(device->parent->handle, acpi_pci_data_handler,
- (void **)&pdata);
- if (ACPI_FAILURE(status) || !pdata || !pdata->bus) {
- ACPI_EXCEPTION((AE_INFO, status,
- "Invalid ACPI-PCI context for parent device %s",
- acpi_device_bid(device->parent)));
- result = -ENODEV;
- goto end;
- }
- data->id.segment = pdata->id.segment;
- data->id.bus = pdata->bus->number;
+ dev = acpi_get_pci_dev(device->handle);
+ if (!dev)
+ return 0;
- /*
- * Device & Function
- * -----------------
- * These are simply obtained from the device's _ADR method. Note
- * that a value of zero is valid.
- */
- data->id.device = device->pnp.bus_address >> 16;
- data->id.function = device->pnp.bus_address & 0xFFFF;
+ if (dev->subordinate)
+ acpi_pci_irq_del_prt(pci_domain_nr(dev->bus),
+ dev->subordinate->number);
- ACPI_DEBUG_PRINT((ACPI_DB_INFO, "...to %04x:%02x:%02x.%d\n",
- data->id.segment, data->id.bus, data->id.device,
- data->id.function));
+ pci_dev_put(dev);
+ return 0;
+}
- /*
- * TBD: Support slot devices (e.g. function=0xFFFF).
- */
+static int acpi_pci_bind(struct acpi_device *device)
+{
+ acpi_status status;
+ acpi_handle handle;
+ unsigned char bus;
+ struct pci_dev *dev;
- /*
- * Locate PCI Device
- * -----------------
- * Locate matching device in PCI namespace. If it doesn't exist
- * this typically means that the device isn't currently inserted
- * (e.g. docking station, port replicator, etc.).
- */
- data->dev = pci_get_slot(pdata->bus,
- PCI_DEVFN(data->id.device, data->id.function));
- if (!data->dev) {
- ACPI_DEBUG_PRINT((ACPI_DB_INFO,
- "Device %04x:%02x:%02x.%d not present in PCI namespace\n",
- data->id.segment, data->id.bus,
- data->id.device, data->id.function));
- result = -ENODEV;
- goto end;
- }
- if (!data->dev->bus) {
- printk(KERN_ERR PREFIX
- "Device %04x:%02x:%02x.%d has invalid 'bus' field\n",
- data->id.segment, data->id.bus,
- data->id.device, data->id.function);
- result = -ENODEV;
- goto end;
- }
+ dev = acpi_get_pci_dev(device->handle);
+ if (!dev)
+ return 0;
/*
- * PCI Bridge?
- * -----------
- * If so, set the 'bus' field and install the 'bind' function to
- * facilitate callbacks for all of its children.
+ * Install the 'bind' function to facilitate callbacks for
+ * children of the P2P bridge.
*/
- if (data->dev->subordinate) {
+ if (dev->subordinate) {
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"Device %04x:%02x:%02x.%d is a PCI bridge\n",
- data->id.segment, data->id.bus,
- data->id.device, data->id.function));
- data->bus = data->dev->subordinate;
+ pci_domain_nr(dev->bus), dev->bus->number,
+ PCI_SLOT(dev->devfn), PCI_FUNC(dev->devfn)));
device->ops.bind = acpi_pci_bind;
device->ops.unbind = acpi_pci_unbind;
}
/*
- * Attach ACPI-PCI Context
- * -----------------------
- * Thus binding the ACPI and PCI devices.
- */
- status = acpi_attach_data(device->handle, acpi_pci_data_handler, data);
- if (ACPI_FAILURE(status)) {
- ACPI_EXCEPTION((AE_INFO, status,
- "Unable to attach ACPI-PCI context to device %s",
- acpi_device_bid(device)));
- result = -ENODEV;
- goto end;
- }
-
- /*
- * PCI Routing Table
- * -----------------
- * Evaluate and parse _PRT, if exists. This code is independent of
- * PCI bridges (above) to allow parsing of _PRT objects within the
- * scope of non-bridge devices. Note that _PRTs within the scope of
- * a PCI bridge assume the bridge's subordinate bus number.
+ * Evaluate and parse _PRT, if exists. This code allows parsing of
+ * _PRT objects within the scope of non-bridge devices. Note that
+ * _PRTs within the scope of a PCI bridge assume the bridge's
+ * subordinate bus number.
*
* TBD: Can _PRTs exist within the scope of non-bridge PCI devices?
*/
status = acpi_get_handle(device->handle, METHOD_NAME__PRT, &handle);
- if (ACPI_SUCCESS(status)) {
- if (data->bus) /* PCI-PCI bridge */
- acpi_pci_irq_add_prt(device->handle, data->id.segment,
- data->bus->number);
- else /* non-bridge PCI device */
- acpi_pci_irq_add_prt(device->handle, data->id.segment,
- data->id.bus);
- }
-
- end:
- kfree(buffer.pointer);
- if (result) {
- pci_dev_put(data->dev);
- kfree(data);
- }
- return result;
-}
-
-static int acpi_pci_unbind(struct acpi_device *device)
-{
- int result = 0;
- acpi_status status;
- struct acpi_pci_data *data;
- struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
-
-
- if (!device || !device->parent)
- return -EINVAL;
-
- status = acpi_get_name(device->handle, ACPI_FULL_PATHNAME, &buffer);
if (ACPI_FAILURE(status))
- return -ENODEV;
-
- ACPI_DEBUG_PRINT((ACPI_DB_INFO, "Unbinding PCI device [%s]...\n",
- (char *) buffer.pointer));
- kfree(buffer.pointer);
+ goto out;
- status =
- acpi_get_data(device->handle, acpi_pci_data_handler,
- (void **)&data);
- if (ACPI_FAILURE(status)) {
- result = -ENODEV;
- goto end;
- }
+ if (dev->subordinate)
+ bus = dev->subordinate->number;
+ else
+ bus = dev->bus->number;
- status = acpi_detach_data(device->handle, acpi_pci_data_handler);
- if (ACPI_FAILURE(status)) {
- ACPI_EXCEPTION((AE_INFO, status,
- "Unable to detach data from device %s",
- acpi_device_bid(device)));
- result = -ENODEV;
- goto end;
- }
- if (data->dev->subordinate) {
- acpi_pci_irq_del_prt(data->id.segment, data->bus->number);
- }
- pci_dev_put(data->dev);
- kfree(data);
+ acpi_pci_irq_add_prt(device->handle, pci_domain_nr(dev->bus), bus);
- end:
- return result;
+out:
+ pci_dev_put(dev);
+ return 0;
}
-int
-acpi_pci_bind_root(struct acpi_device *device,
- struct acpi_pci_id *id, struct pci_bus *bus)
+int acpi_pci_bind_root(struct acpi_device *device)
{
- int result = 0;
- acpi_status status;
- struct acpi_pci_data *data = NULL;
- struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
-
- if (!device || !id || !bus) {
- return -EINVAL;
- }
-
- data = kzalloc(sizeof(struct acpi_pci_data), GFP_KERNEL);
- if (!data)
- return -ENOMEM;
-
- data->id = *id;
- data->bus = bus;
device->ops.bind = acpi_pci_bind;
device->ops.unbind = acpi_pci_unbind;
- status = acpi_get_name(device->handle, ACPI_FULL_PATHNAME, &buffer);
- if (ACPI_FAILURE(status)) {
- kfree (data);
- return -ENODEV;
- }
-
- ACPI_DEBUG_PRINT((ACPI_DB_INFO, "Binding PCI root bridge [%s] to "
- "%04x:%02x\n", (char *)buffer.pointer,
- id->segment, id->bus));
-
- status = acpi_attach_data(device->handle, acpi_pci_data_handler, data);
- if (ACPI_FAILURE(status)) {
- ACPI_EXCEPTION((AE_INFO, status,
- "Unable to attach ACPI-PCI context to device %s",
- (char *)buffer.pointer));
- result = -ENODEV;
- goto end;
- }
-
- end:
- kfree(buffer.pointer);
- if (result != 0)
- kfree(data);
-
- return result;
+ return 0;
}
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
index e509991..f23fcc5 100644
--- a/drivers/acpi/pci_root.c
+++ b/drivers/acpi/pci_root.c
@@ -603,7 +603,7 @@ static int __devinit acpi_pci_root_add(struct acpi_device *device)
* -----------------------
* Thus binding the ACPI and PCI devices.
*/
- result = acpi_pci_bind_root(device, &root->id, root->bus);
+ result = acpi_pci_bind_root(device);
if (result)
goto end;
diff --git a/include/acpi/acpi_drivers.h b/include/acpi/acpi_drivers.h
index 1ef529b..01d9ce4 100644
--- a/include/acpi/acpi_drivers.h
+++ b/include/acpi/acpi_drivers.h
@@ -99,8 +99,7 @@ struct pci_bus;
struct pci_dev *acpi_get_pci_dev(acpi_handle);
acpi_status acpi_get_pci_id(acpi_handle handle, struct acpi_pci_id *id);
-int acpi_pci_bind_root(struct acpi_device *device, struct acpi_pci_id *id,
- struct pci_bus *bus);
+int acpi_pci_bind_root(struct acpi_device *device);
/* Arch-defined function to add a bus to the system */
Now that acpi_get_pci_dev is available, let's use it instead of
acpi_get_pci_id.
Cc: Jesse Barnes <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/pci/hotplug/acpiphp_glue.c | 25 +++++++------------------
1 files changed, 7 insertions(+), 18 deletions(-)
diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c
index fc6636e..0cb0f83 100644
--- a/drivers/pci/hotplug/acpiphp_glue.c
+++ b/drivers/pci/hotplug/acpiphp_glue.c
@@ -678,18 +678,9 @@ static void remove_bridge(acpi_handle handle)
static struct pci_dev * get_apic_pci_info(acpi_handle handle)
{
- struct acpi_pci_id id;
- struct pci_bus *bus;
struct pci_dev *dev;
- if (ACPI_FAILURE(acpi_get_pci_id(handle, &id)))
- return NULL;
-
- bus = pci_find_bus(id.segment, id.bus);
- if (!bus)
- return NULL;
-
- dev = pci_get_slot(bus, PCI_DEVFN(id.device, id.function));
+ dev = acpi_get_pci_dev(handle);
if (!dev)
return NULL;
@@ -1396,19 +1387,16 @@ static void acpiphp_sanitize_bus(struct pci_bus *bus)
/* Program resources in newly inserted bridge */
static int acpiphp_configure_bridge (acpi_handle handle)
{
- struct acpi_pci_id pci_id;
+ struct pci_dev *dev;
struct pci_bus *bus;
- if (ACPI_FAILURE(acpi_get_pci_id(handle, &pci_id))) {
+ dev = acpi_get_pci_dev(handle);
+ if (!dev) {
err("cannot get PCI domain and bus number for bridge\n");
return -EINVAL;
}
- bus = pci_find_bus(pci_id.segment, pci_id.bus);
- if (!bus) {
- err("cannot find bus %d:%d\n",
- pci_id.segment, pci_id.bus);
- return -EINVAL;
- }
+
+ bus = dev->bus;
pci_bus_size_bridges(bus);
pci_bus_assign_resources(bus);
@@ -1416,6 +1404,7 @@ static int acpiphp_configure_bridge (acpi_handle handle)
acpiphp_set_hpp_values(handle, bus);
pci_enable_bridges(bus);
acpiphp_configure_ioapics(handle);
+ pci_dev_put(dev);
return 0;
}
In acpi_pci_bind, we add _PRT information for non-bridge devices, but
we never delete that information in acpi_pci_unbind.
We also set device->ops.bind and device->ops.unbind, but never clear
them out.
Let's make acpi_pci_unbind clean up what we did in acpi_pci_bind.
Cc: Bjorn Helgaas <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/pci_bind.c | 11 +++++++++--
1 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
index 62cb383..5fa1c5e 100644
--- a/drivers/acpi/pci_bind.c
+++ b/drivers/acpi/pci_bind.c
@@ -106,14 +106,21 @@ EXPORT_SYMBOL(acpi_get_pci_id);
static int acpi_pci_unbind(struct acpi_device *device)
{
+ struct pci_bus *bus;
struct pci_dev *dev;
dev = acpi_get_pci_dev(device->handle);
if (!dev)
return 0;
- if (dev->subordinate)
- acpi_pci_irq_del_prt(dev->subordinate);
+ if (dev->subordinate) {
+ bus = dev->subordinate;
+ device->ops.bind = NULL;
+ device->ops.unbind = NULL;
+ } else
+ bus = dev->bus;
+
+ acpi_pci_irq_del_prt(bus);
pci_dev_put(dev);
return 0;
A PCI domain cannot change as you descend down subordinate buses, which
makes the 'segment' argument to acpi_pci_irq_add_prt() useless.
Change the interface to take a struct pci_bus *, from whence we can derive
the bus number and segment. Reducing the number of arguments makes life
simpler for callers.
Cc: Bjorn Helgaas <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/pci_bind.c | 8 ++++----
drivers/acpi/pci_irq.c | 10 +++++-----
drivers/acpi/pci_root.c | 3 +--
include/acpi/acpi_drivers.h | 2 +-
4 files changed, 11 insertions(+), 12 deletions(-)
diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
index 703d2a3..6eb58ef 100644
--- a/drivers/acpi/pci_bind.c
+++ b/drivers/acpi/pci_bind.c
@@ -124,7 +124,7 @@ static int acpi_pci_bind(struct acpi_device *device)
{
acpi_status status;
acpi_handle handle;
- unsigned char bus;
+ struct pci_bus *bus;
struct pci_dev *dev;
dev = acpi_get_pci_dev(device->handle);
@@ -157,11 +157,11 @@ static int acpi_pci_bind(struct acpi_device *device)
goto out;
if (dev->subordinate)
- bus = dev->subordinate->number;
+ bus = dev->subordinate;
else
- bus = dev->bus->number;
+ bus = dev->bus;
- acpi_pci_irq_add_prt(device->handle, pci_domain_nr(dev->bus), bus);
+ acpi_pci_irq_add_prt(device->handle, bus);
out:
pci_dev_put(dev);
diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c
index 51b9f82..3ed944c 100644
--- a/drivers/acpi/pci_irq.c
+++ b/drivers/acpi/pci_irq.c
@@ -182,7 +182,7 @@ static void do_prt_fixups(struct acpi_prt_entry *entry,
}
}
-static int acpi_pci_irq_add_entry(acpi_handle handle, int segment, int bus,
+static int acpi_pci_irq_add_entry(acpi_handle handle, struct pci_bus *bus,
struct acpi_pci_routing_table *prt)
{
struct acpi_prt_entry *entry;
@@ -196,8 +196,8 @@ static int acpi_pci_irq_add_entry(acpi_handle handle, int segment, int bus,
* 1=INTA, 2=INTB. We use the PCI encoding throughout, so convert
* it here.
*/
- entry->id.segment = segment;
- entry->id.bus = bus;
+ entry->id.segment = pci_domain_nr(bus);
+ entry->id.bus = bus->number;
entry->id.device = (prt->address >> 16) & 0xFFFF;
entry->pin = prt->pin + 1;
@@ -242,7 +242,7 @@ static int acpi_pci_irq_add_entry(acpi_handle handle, int segment, int bus,
return 0;
}
-int acpi_pci_irq_add_prt(acpi_handle handle, int segment, int bus)
+int acpi_pci_irq_add_prt(acpi_handle handle, struct pci_bus *bus)
{
acpi_status status;
struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
@@ -271,7 +271,7 @@ int acpi_pci_irq_add_prt(acpi_handle handle, int segment, int bus)
entry = buffer.pointer;
while (entry && (entry->length > 0)) {
- acpi_pci_irq_add_entry(handle, segment, bus, entry);
+ acpi_pci_irq_add_entry(handle, bus, entry);
entry = (struct acpi_pci_routing_table *)
((unsigned long)entry + entry->length);
}
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
index f23fcc5..f341b07 100644
--- a/drivers/acpi/pci_root.c
+++ b/drivers/acpi/pci_root.c
@@ -614,8 +614,7 @@ static int __devinit acpi_pci_root_add(struct acpi_device *device)
*/
status = acpi_get_handle(device->handle, METHOD_NAME__PRT, &handle);
if (ACPI_SUCCESS(status))
- result = acpi_pci_irq_add_prt(device->handle, root->id.segment,
- root->id.bus);
+ result = acpi_pci_irq_add_prt(device->handle, root->bus);
/*
* Scan and bind all _ADR-Based Devices
diff --git a/include/acpi/acpi_drivers.h b/include/acpi/acpi_drivers.h
index 01d9ce4..702b12e 100644
--- a/include/acpi/acpi_drivers.h
+++ b/include/acpi/acpi_drivers.h
@@ -90,7 +90,7 @@ int acpi_pci_link_free_irq(acpi_handle handle);
/* ACPI PCI Interrupt Routing (pci_irq.c) */
-int acpi_pci_irq_add_prt(acpi_handle handle, int segment, int bus);
+int acpi_pci_irq_add_prt(acpi_handle handle, struct pci_bus *bus);
void acpi_pci_irq_del_prt(int segment, int bus);
/* ACPI PCI Device Binding (pci_bind.c) */
There is no need to pass a segment/bus tuple to this API, as the callsite
always has a struct pci_bus. We can derive segment/bus from the
struct pci_bus, so let's take this opportunit to simplify the API and
make life easier for the callers.
Cc: Bjorn Helgaas <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/pci_bind.c | 3 +--
drivers/acpi/pci_irq.c | 7 ++++---
include/acpi/acpi_drivers.h | 2 +-
3 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
index 6eb58ef..62cb383 100644
--- a/drivers/acpi/pci_bind.c
+++ b/drivers/acpi/pci_bind.c
@@ -113,8 +113,7 @@ static int acpi_pci_unbind(struct acpi_device *device)
return 0;
if (dev->subordinate)
- acpi_pci_irq_del_prt(pci_domain_nr(dev->bus),
- dev->subordinate->number);
+ acpi_pci_irq_del_prt(dev->subordinate);
pci_dev_put(dev);
return 0;
diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c
index 3ed944c..ef9509e 100644
--- a/drivers/acpi/pci_irq.c
+++ b/drivers/acpi/pci_irq.c
@@ -280,16 +280,17 @@ int acpi_pci_irq_add_prt(acpi_handle handle, struct pci_bus *bus)
return 0;
}
-void acpi_pci_irq_del_prt(int segment, int bus)
+void acpi_pci_irq_del_prt(struct pci_bus *bus)
{
struct acpi_prt_entry *entry, *tmp;
printk(KERN_DEBUG
"ACPI: Delete PCI Interrupt Routing Table for %04x:%02x\n",
- segment, bus);
+ pci_domain_nr(bus), bus->number);
spin_lock(&acpi_prt_lock);
list_for_each_entry_safe(entry, tmp, &acpi_prt_list, list) {
- if (segment == entry->id.segment && bus == entry->id.bus) {
+ if (pci_domain_nr(bus) == entry->id.segment
+ && bus->number == entry->id.bus) {
list_del(&entry->list);
kfree(entry);
}
diff --git a/include/acpi/acpi_drivers.h b/include/acpi/acpi_drivers.h
index 702b12e..12e99c3 100644
--- a/include/acpi/acpi_drivers.h
+++ b/include/acpi/acpi_drivers.h
@@ -91,7 +91,7 @@ int acpi_pci_link_free_irq(acpi_handle handle);
/* ACPI PCI Interrupt Routing (pci_irq.c) */
int acpi_pci_irq_add_prt(acpi_handle handle, struct pci_bus *bus);
-void acpi_pci_irq_del_prt(int segment, int bus);
+void acpi_pci_irq_del_prt(struct pci_bus *bus);
/* ACPI PCI Device Binding (pci_bind.c) */
acpi_get_pci_dev() is better, and all callers have been converted, so
eliminate acpi_get_pci_id().
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/pci_bind.c | 71 -------------------------------------------
include/acpi/acpi_drivers.h | 1 -
2 files changed, 0 insertions(+), 72 deletions(-)
diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
index 5fa1c5e..ffe0046 100644
--- a/drivers/acpi/pci_bind.c
+++ b/drivers/acpi/pci_bind.c
@@ -33,77 +33,6 @@
#define _COMPONENT ACPI_PCI_COMPONENT
ACPI_MODULE_NAME("pci_bind");
-struct acpi_pci_data {
- struct acpi_pci_id id;
- struct pci_bus *bus;
- struct pci_dev *dev;
-};
-
-static int acpi_pci_bind(struct acpi_device *device);
-static int acpi_pci_unbind(struct acpi_device *device);
-
-static void acpi_pci_data_handler(acpi_handle handle, u32 function,
- void *context)
-{
-
- /* TBD: Anything we need to do here? */
-
- return;
-}
-
-/**
- * acpi_get_pci_id
- * ------------------
- * This function is used by the ACPI Interpreter (a.k.a. Core Subsystem)
- * to resolve PCI information for ACPI-PCI devices defined in the namespace.
- * This typically occurs when resolving PCI operation region information.
- */
-acpi_status acpi_get_pci_id(acpi_handle handle, struct acpi_pci_id *id)
-{
- int result = 0;
- acpi_status status = AE_OK;
- struct acpi_device *device = NULL;
- struct acpi_pci_data *data = NULL;
-
-
- if (!id)
- return AE_BAD_PARAMETER;
-
- result = acpi_bus_get_device(handle, &device);
- if (result) {
- printk(KERN_ERR PREFIX
- "Invalid ACPI Bus context for device %s\n",
- acpi_device_bid(device));
- return AE_NOT_EXIST;
- }
-
- status = acpi_get_data(handle, acpi_pci_data_handler, (void **)&data);
- if (ACPI_FAILURE(status) || !data) {
- ACPI_EXCEPTION((AE_INFO, status,
- "Invalid ACPI-PCI context for device %s",
- acpi_device_bid(device)));
- return status;
- }
-
- *id = data->id;
-
- /*
- id->segment = data->id.segment;
- id->bus = data->id.bus;
- id->device = data->id.device;
- id->function = data->id.function;
- */
-
- ACPI_DEBUG_PRINT((ACPI_DB_INFO,
- "Device %s has PCI address %04x:%02x:%02x.%d\n",
- acpi_device_bid(device), id->segment, id->bus,
- id->device, id->function));
-
- return AE_OK;
-}
-
-EXPORT_SYMBOL(acpi_get_pci_id);
-
static int acpi_pci_unbind(struct acpi_device *device)
{
struct pci_bus *bus;
diff --git a/include/acpi/acpi_drivers.h b/include/acpi/acpi_drivers.h
index 12e99c3..f4906f6 100644
--- a/include/acpi/acpi_drivers.h
+++ b/include/acpi/acpi_drivers.h
@@ -98,7 +98,6 @@ void acpi_pci_irq_del_prt(struct pci_bus *bus);
struct pci_bus;
struct pci_dev *acpi_get_pci_dev(acpi_handle);
-acpi_status acpi_get_pci_id(acpi_handle handle, struct acpi_pci_id *id);
int acpi_pci_bind_root(struct acpi_device *device);
/* Arch-defined function to add a bus to the system */
Now that acpi_get_pci_dev is available, let's use it instead of
acpi_get_physical_pci_device()
Cc: Thomas Renninger <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/video.c | 6 +++---
drivers/acpi/video_detect.c | 9 +++++----
2 files changed, 8 insertions(+), 7 deletions(-)
diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
index 1bdfb37..5adbf93 100644
--- a/drivers/acpi/video.c
+++ b/drivers/acpi/video.c
@@ -1054,15 +1054,15 @@ static void acpi_video_bus_find_cap(struct acpi_video_bus *video)
static int acpi_video_bus_check(struct acpi_video_bus *video)
{
acpi_status status = -ENOENT;
- struct device *dev;
+ struct pci_dev *dev;
if (!video)
return -EINVAL;
- dev = acpi_get_physical_pci_device(video->device->handle);
+ dev = acpi_get_pci_dev(video->device->handle);
if (!dev)
return -ENODEV;
- put_device(dev);
+ pci_dev_put(dev);
/* Since there is no HID, CID and so on for VGA driver, we have
* to check well known required nodes.
diff --git a/drivers/acpi/video_detect.c b/drivers/acpi/video_detect.c
index 0973727..7cd2b63 100644
--- a/drivers/acpi/video_detect.c
+++ b/drivers/acpi/video_detect.c
@@ -10,7 +10,7 @@
* assinged
*
* After PCI devices are glued with ACPI devices
- * acpi_get_physical_pci_device() can be called to identify ACPI graphics
+ * acpi_get_pci_dev() can be called to identify ACPI graphics
* devices for which a real graphics card is plugged in
*
* Now acpi_video_get_capabilities() can be called to check which
@@ -36,6 +36,7 @@
#include <linux/acpi.h>
#include <linux/dmi.h>
+#include <linux/pci.h>
ACPI_MODULE_NAME("video");
#define _COMPONENT ACPI_VIDEO_COMPONENT
@@ -109,7 +110,7 @@ static acpi_status
find_video(acpi_handle handle, u32 lvl, void *context, void **rv)
{
long *cap = context;
- struct device *dev;
+ struct pci_dev *dev;
struct acpi_device *acpi_dev;
const struct acpi_device_id video_ids[] = {
@@ -120,10 +121,10 @@ find_video(acpi_handle handle, u32 lvl, void *context, void **rv)
return AE_OK;
if (!acpi_match_device_ids(acpi_dev, video_ids)) {
- dev = acpi_get_physical_pci_device(handle);
+ dev = acpi_get_pci_dev(handle);
if (!dev)
return AE_OK;
- put_device(dev);
+ pci_dev_put(dev);
*cap |= acpi_is_video_device(acpi_dev);
}
return AE_OK;
acpi_get_pci_dev() is (hopefully) better, and all callers have been
converted, so let's get rid of this duplicated functionality.
Cc: Thomas Renninger <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
drivers/acpi/glue.c | 40 ----------------------------------------
include/acpi/acpi_bus.h | 1 -
2 files changed, 0 insertions(+), 41 deletions(-)
diff --git a/drivers/acpi/glue.c b/drivers/acpi/glue.c
index 8bd2c2a..a8a5c29 100644
--- a/drivers/acpi/glue.c
+++ b/drivers/acpi/glue.c
@@ -140,46 +140,6 @@ struct device *acpi_get_physical_device(acpi_handle handle)
EXPORT_SYMBOL(acpi_get_physical_device);
-/* ToDo: When a PCI bridge is found, return the PCI device behind the bridge
- * This should work in general, but did not on a Lenovo T61 for the
- * graphics card. But this must be fixed when the PCI device is
- * bound and the kernel device struct is attached to the acpi device
- * Note: A success call will increase reference count by one
- * Do call put_device(dev) on the returned device then
- */
-struct device *acpi_get_physical_pci_device(acpi_handle handle)
-{
- struct device *dev;
- long long device_id;
- acpi_status status;
-
- status =
- acpi_evaluate_integer(handle, "_ADR", NULL, &device_id);
-
- if (ACPI_FAILURE(status))
- return NULL;
-
- /* We need to attempt to determine whether the _ADR refers to a
- PCI device or not. There's no terribly good way to do this,
- so the best we can hope for is to assume that there'll never
- be a device in the host bridge */
- if (device_id >= 0x10000) {
- /* It looks like a PCI device. Does it exist? */
- dev = acpi_get_physical_device(handle);
- } else {
- /* It doesn't look like a PCI device. Does its parent
- exist? */
- acpi_handle phandle;
- if (acpi_get_parent(handle, &phandle))
- return NULL;
- dev = acpi_get_physical_device(phandle);
- }
- if (!dev)
- return NULL;
- return dev;
-}
-EXPORT_SYMBOL(acpi_get_physical_pci_device);
-
static int acpi_bind_one(struct device *dev, acpi_handle handle)
{
struct acpi_device *acpi_dev;
diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h
index 494088b..c65e4ce 100644
--- a/include/acpi/acpi_bus.h
+++ b/include/acpi/acpi_bus.h
@@ -368,7 +368,6 @@ struct acpi_bus_type {
int register_acpi_bus_type(struct acpi_bus_type *);
int unregister_acpi_bus_type(struct acpi_bus_type *);
struct device *acpi_get_physical_device(acpi_handle);
-struct device *acpi_get_physical_pci_device(acpi_handle);
/* helper */
acpi_handle acpi_get_child(acpi_handle, acpi_integer);
Alex Chiang wrote:
> In acpi_pci_bind, we add _PRT information for non-bridge devices, but
> we never delete that information in acpi_pci_unbind.
>
> We also set device->ops.bind and device->ops.unbind, but never clear
> them out.
>
> Let's make acpi_pci_unbind clean up what we did in acpi_pci_bind.
>
> Cc: Bjorn Helgaas <[email protected]>
> Signed-off-by: Alex Chiang <[email protected]>
> ---
>
> drivers/acpi/pci_bind.c | 11 +++++++++--
> 1 files changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
> index 62cb383..5fa1c5e 100644
> --- a/drivers/acpi/pci_bind.c
> +++ b/drivers/acpi/pci_bind.c
> @@ -106,14 +106,21 @@ EXPORT_SYMBOL(acpi_get_pci_id);
>
> static int acpi_pci_unbind(struct acpi_device *device)
> {
> + struct pci_bus *bus;
> struct pci_dev *dev;
>
> dev = acpi_get_pci_dev(device->handle);
> if (!dev)
> return 0;
>
> - if (dev->subordinate)
> - acpi_pci_irq_del_prt(dev->subordinate);
> + if (dev->subordinate) {
> + bus = dev->subordinate;
> + device->ops.bind = NULL;
> + device->ops.unbind = NULL;
> + } else
> + bus = dev->bus;
> +
> + acpi_pci_irq_del_prt(bus);
>
I have a concern about this change.
The acpi_pci_irq_del_prt() against dev->bus removes not only
the _PRT entries for PCI function corresponding to specified
acpi_device, but also other _PRT entries for working PCI
devices/functions on the same bus. As a result, interrupt
initialization for those PCI functions would no longer work
properly after that.
So I think we should not call acpi_pci_irq_del_prt() against
dev->bus.
Thanks,
Kenji Kaneshige
* Kenji Kaneshige <[email protected]>:
>>
>
> I have a concern about this change.
>
> The acpi_pci_irq_del_prt() against dev->bus removes not only
> the _PRT entries for PCI function corresponding to specified
> acpi_device, but also other _PRT entries for working PCI
> devices/functions on the same bus. As a result, interrupt
> initialization for those PCI functions would no longer work
> properly after that.
>
> So I think we should not call acpi_pci_irq_del_prt() against
> dev->bus.
Thanks for the review. I agree with you.
Here is a respun version of this patch.
From: Alex Chiang <[email protected]>
ACPI: acpi_pci_unbind should clean up properly after acpi_pci_bind
In acpi_pci_bind, we set device->ops.bind and device->ops.unbind, but
never clear them out.
Cc: Bjorn Helgaas <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
index 62cb383..c9cc650 100644
--- a/drivers/acpi/pci_bind.c
+++ b/drivers/acpi/pci_bind.c
@@ -109,11 +109,13 @@ static int acpi_pci_unbind(struct acpi_device *device)
struct pci_dev *dev;
dev = acpi_get_pci_dev(device->handle);
- if (!dev)
+ if (!dev || !dev->subordinate)
return 0;
- if (dev->subordinate)
- acpi_pci_irq_del_prt(dev->subordinate);
+ acpi_pci_irq_del_prt(dev->subordinate);
+
+ device->ops.bind = NULL;
+ device->ops.unbind = NULL;
pci_dev_put(dev);
return 0;
On Thursday 04 June 2009 05:35:21 pm Alex Chiang wrote:
> * Kenji Kaneshige <[email protected]>:
> >
> > I have a concern about this change.
> >
> > The acpi_pci_irq_del_prt() against dev->bus removes not only
> > the _PRT entries for PCI function corresponding to specified
> > acpi_device, but also other _PRT entries for working PCI
> > devices/functions on the same bus. As a result, interrupt
> > initialization for those PCI functions would no longer work
> > properly after that.
> >
> > So I think we should not call acpi_pci_irq_del_prt() against
> > dev->bus.
>
> Thanks for the review. I agree with you.
I agree that this respun version makes things more the way they were,
so in that sense, it should do no harm. But I still have the niggling
concern that .bind() adds _PRT info for non-bridges, and there's no
corresponding removal. There should be some path that makes this
more symmetric.
> Here is a respun version of this patch.
>
> From: Alex Chiang <[email protected]>
>
> ACPI: acpi_pci_unbind should clean up properly after acpi_pci_bind
>
> In acpi_pci_bind, we set device->ops.bind and device->ops.unbind, but
> never clear them out.
>
> Cc: Bjorn Helgaas <[email protected]>
> Signed-off-by: Alex Chiang <[email protected]>
> ---
> diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
> index 62cb383..c9cc650 100644
> --- a/drivers/acpi/pci_bind.c
> +++ b/drivers/acpi/pci_bind.c
> @@ -109,11 +109,13 @@ static int acpi_pci_unbind(struct acpi_device *device)
> struct pci_dev *dev;
>
> dev = acpi_get_pci_dev(device->handle);
> - if (!dev)
> + if (!dev || !dev->subordinate)
> return 0;
>
> - if (dev->subordinate)
> - acpi_pci_irq_del_prt(dev->subordinate);
> + acpi_pci_irq_del_prt(dev->subordinate);
> +
> + device->ops.bind = NULL;
> + device->ops.unbind = NULL;
>
> pci_dev_put(dev);
> return 0;
>
* Bjorn Helgaas <[email protected]>:
> On Thursday 04 June 2009 05:35:21 pm Alex Chiang wrote:
> > * Kenji Kaneshige <[email protected]>:
> > >
> > > I have a concern about this change.
> > >
> > > The acpi_pci_irq_del_prt() against dev->bus removes not only
> > > the _PRT entries for PCI function corresponding to specified
> > > acpi_device, but also other _PRT entries for working PCI
> > > devices/functions on the same bus. As a result, interrupt
> > > initialization for those PCI functions would no longer work
> > > properly after that.
> > >
> > > So I think we should not call acpi_pci_irq_del_prt() against
> > > dev->bus.
> >
> > Thanks for the review. I agree with you.
>
> I agree that this respun version makes things more the way they were,
> so in that sense, it should do no harm. But I still have the niggling
> concern that .bind() adds _PRT info for non-bridges, and there's no
> corresponding removal. There should be some path that makes this
> more symmetric.
The comment in acpi_pci_bind doesn't seem to know if you can even
have a _PRT for non-bridges. The spec (3.0b) says that _PRT is
required for all root bridges, but doesn't mention anything about
non-bridge devices.
My gut feeling is the way to cure the symmetry is to remove the
path in .bind() that adds _PRT for non-bridges, but I'm a little
hesitant to do that without knowing for sure.
I think for this patch, the Hippocratic approach is the right
one for now.
Thanks.
/ac
> > Here is a respun version of this patch.
> >
> > From: Alex Chiang <[email protected]>
> >
> > ACPI: acpi_pci_unbind should clean up properly after acpi_pci_bind
> >
> > In acpi_pci_bind, we set device->ops.bind and device->ops.unbind, but
> > never clear them out.
> >
> > Cc: Bjorn Helgaas <[email protected]>
> > Signed-off-by: Alex Chiang <[email protected]>
> > ---
> > diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
> > index 62cb383..c9cc650 100644
> > --- a/drivers/acpi/pci_bind.c
> > +++ b/drivers/acpi/pci_bind.c
> > @@ -109,11 +109,13 @@ static int acpi_pci_unbind(struct acpi_device *device)
> > struct pci_dev *dev;
> >
> > dev = acpi_get_pci_dev(device->handle);
> > - if (!dev)
> > + if (!dev || !dev->subordinate)
> > return 0;
> >
> > - if (dev->subordinate)
> > - acpi_pci_irq_del_prt(dev->subordinate);
> > + acpi_pci_irq_del_prt(dev->subordinate);
> > +
> > + device->ops.bind = NULL;
> > + device->ops.unbind = NULL;
> >
> > pci_dev_put(dev);
> > return 0;
> >
>
>
Alex Chiang wrote:
> * Kenji Kaneshige <[email protected]>:
>>>
>> I have a concern about this change.
>>
>> The acpi_pci_irq_del_prt() against dev->bus removes not only
>> the _PRT entries for PCI function corresponding to specified
>> acpi_device, but also other _PRT entries for working PCI
>> devices/functions on the same bus. As a result, interrupt
>> initialization for those PCI functions would no longer work
>> properly after that.
>>
>> So I think we should not call acpi_pci_irq_del_prt() against
>> dev->bus.
>
> Thanks for the review. I agree with you.
>
> Here is a respun version of this patch.
>
> From: Alex Chiang <[email protected]>
>
> ACPI: acpi_pci_unbind should clean up properly after acpi_pci_bind
>
> In acpi_pci_bind, we set device->ops.bind and device->ops.unbind, but
> never clear them out.
>
> Cc: Bjorn Helgaas <[email protected]>
> Signed-off-by: Alex Chiang <[email protected]>
> ---
> diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
> index 62cb383..c9cc650 100644
> --- a/drivers/acpi/pci_bind.c
> +++ b/drivers/acpi/pci_bind.c
> @@ -109,11 +109,13 @@ static int acpi_pci_unbind(struct acpi_device *device)
> struct pci_dev *dev;
>
> dev = acpi_get_pci_dev(device->handle);
> - if (!dev)
> + if (!dev || !dev->subordinate)
> return 0;
This would leak the pci_dev's refcount
if dev != NULL && dev->subordinate == NULL.
Thanks,
Kenji Kaneshige
>
> - if (dev->subordinate)
> - acpi_pci_irq_del_prt(dev->subordinate);
> + acpi_pci_irq_del_prt(dev->subordinate);
> +
> + device->ops.bind = NULL;
> + device->ops.unbind = NULL;
>
> pci_dev_put(dev);
> return 0;
> --
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
>
* Kenji Kaneshige <[email protected]>:
> Alex Chiang wrote:
>> diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
>> index 62cb383..c9cc650 100644
>> --- a/drivers/acpi/pci_bind.c
>> +++ b/drivers/acpi/pci_bind.c
>> @@ -109,11 +109,13 @@ static int acpi_pci_unbind(struct acpi_device *device)
>> struct pci_dev *dev;
>> dev = acpi_get_pci_dev(device->handle);
>> - if (!dev)
>> + if (!dev || !dev->subordinate)
>> return 0;
>
> This would leak the pci_dev's refcount
> if dev != NULL && dev->subordinate == NULL.
Yes, of course you are correct. Thanks for correcting my
sloppiness.
Here is the correct fix.
From: Alex Chiang <[email protected]>
ACPI: acpi_pci_unbind should clean up properly after acpi_pci_bind
In acpi_pci_bind, we set device->ops.bind and device->ops.unbind, but
never clear them out.
Cc: Bjorn Helgaas <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
---
diff --git a/drivers/acpi/pci_bind.c b/drivers/acpi/pci_bind.c
index 62cb383..af75784 100644
--- a/drivers/acpi/pci_bind.c
+++ b/drivers/acpi/pci_bind.c
@@ -109,12 +109,15 @@ static int acpi_pci_unbind(struct acpi_device *device)
struct pci_dev *dev;
dev = acpi_get_pci_dev(device->handle);
- if (!dev)
- return 0;
+ if (!dev || !dev->subordinate)
+ goto out:
- if (dev->subordinate)
- acpi_pci_irq_del_prt(dev->subordinate);
+ acpi_pci_irq_del_prt(dev->subordinate);
+ device->ops.bind = NULL;
+ device->ops.unbind = NULL;
+
+out:
pci_dev_put(dev);
return 0;
}
* Bjorn Helgaas <[email protected]>:
> On Thursday 04 June 2009 05:35:21 pm Alex Chiang wrote:
> > * Kenji Kaneshige <[email protected]>:
> > >
> > > I have a concern about this change.
> > >
> > > The acpi_pci_irq_del_prt() against dev->bus removes not only
> > > the _PRT entries for PCI function corresponding to specified
> > > acpi_device, but also other _PRT entries for working PCI
> > > devices/functions on the same bus. As a result, interrupt
> > > initialization for those PCI functions would no longer work
> > > properly after that.
> > >
> > > So I think we should not call acpi_pci_irq_del_prt() against
> > > dev->bus.
> >
> > Thanks for the review. I agree with you.
>
> I agree that this respun version makes things more the way they were,
> so in that sense, it should do no harm. But I still have the niggling
> concern that .bind() adds _PRT info for non-bridges, and there's no
> corresponding removal. There should be some path that makes this
> more symmetric.
Hm, in another forum, you suggested that dynamic PRT lookups
might be a solution, which I kinda like.
So, the plan that I would prefer is:
a) get this patchset in [and we 'do no harm' here so
_hopefully_ aren't introducing regressions]
b) work on dynamic PRT lookups in a future patchset.
Thanks.
/ac
On Wed, 03 Jun 2009 23:58:57 -0600
Alex Chiang <[email protected]> wrote:
> Now that acpi_get_pci_dev is available, let's use it instead of
> acpi_get_pci_id.
>
> Cc: Jesse Barnes <[email protected]>
> Signed-off-by: Alex Chiang <[email protected]>
> ---
>
Applied to linux-next, thanks.
--
Jesse Barnes, Intel Open Source Technology Center
On Thu, 11 Jun 2009 14:48:45 -0700
Jesse Barnes <[email protected]> wrote:
> On Wed, 03 Jun 2009 23:58:57 -0600
> Alex Chiang <[email protected]> wrote:
>
> > Now that acpi_get_pci_dev is available, let's use it instead of
> > acpi_get_pci_id.
> >
> > Cc: Jesse Barnes <[email protected]>
> > Signed-off-by: Alex Chiang <[email protected]>
> > ---
> >
>
> Applied to linux-next, thanks.
Oops, since this one depends on others in the series it's probably best
if they just go in through Len's tree. This one has my
Acked-by: Jesse Barnes <[email protected]>
too.
--
Jesse Barnes, Intel Open Source Technology Center