| Message ID | 1396886819-2637-13-git-send-email-rui.zhang@intel.com (mailing list archive) |
|---|---|
| State | Changes Requested, archived |
| Headers | show |
On Tuesday, April 08, 2014 12:06:59 AM Zhang Rui wrote: > For some devices with scan handler attached, their children devices > are enumerated by the scan handler, indirectly. This isn't the case really. They are enumerated by bus controller drivers for the buses they are on. > In this case, we do not want to enumerate the children devices in > acpi scan code explicitly. > > Thus a new flag .handle_children is introduced in this patch. > > For scan handlers with this flag set, we will do default enumeration neither > for the attached devices nor for the children of the attached devices. I'm not sure if that is the right approach. I would prefer that to be handled in a more fine-graind manner, like a flag per device ID or something similar? > Signed-off-by: Zhang Rui <rui.zhang@intel.com> > --- > drivers/acpi/acpi_lpss.c | 2 ++ > drivers/acpi/scan.c | 28 ++++++++++++++++++++++++++-- > include/acpi/acpi_bus.h | 4 +++- > 3 files changed, 31 insertions(+), 3 deletions(-) > > diff --git a/drivers/acpi/acpi_lpss.c b/drivers/acpi/acpi_lpss.c > index 965428f..da0a3d4 100644 > --- a/drivers/acpi/acpi_lpss.c > +++ b/drivers/acpi/acpi_lpss.c > @@ -521,6 +521,7 @@ static struct acpi_scan_handler lpss_handler = { > .attach = acpi_lpss_create_device, > .bind = acpi_lpss_bind, > .unbind = acpi_lpss_unbind, > + .handle_children = true, > }; > > #endif /* CONFIG_X86_INTEL_LPSS */ > @@ -534,6 +535,7 @@ static int acpi_lpss_dummy_attach(struct acpi_device *adev, > static struct acpi_scan_handler lpss_dummy_handler = { > .ids = acpi_lpss_device_ids, > .attach = acpi_lpss_dummy_attach, > + .handle_children = true, > }; > > void __init acpi_lpss_init(void) > diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c > index 44c4668..4ea867e 100644 > --- a/drivers/acpi/scan.c > +++ b/drivers/acpi/scan.c > @@ -2073,6 +2073,31 @@ static acpi_status acpi_bus_check_add(acpi_handle handle, u32 lvl_not_used, > return AE_OK; > } > > +static void acpi_do_default_enumeration(struct acpi_device *device) > +{ > + /* > + * Do not do enumeration for device object that > + * its parent doesn't want to > + */ > + if (device->parent && device->parent->flags.no_child_enumeration) { > + device->flags.no_child_enumeration = 1; > + return; > + } > + > + /* Do not do enumeration for device object with scan handler attached */ > + if (device->handler) { > + if (device->handler->handle_children) > + device->flags.no_child_enumeration = 1; > + return; > + } > + > + /* Do not do enumeration for device object w/o platform_id */ > + if (!device->pnp.type.platform_id) > + return; > + > + acpi_create_platform_device(device, NULL); > +} > + > static int acpi_scan_attach_handler(struct acpi_device *device) > { > struct acpi_hardware_id *hwid; > @@ -2095,8 +2120,7 @@ static int acpi_scan_attach_handler(struct acpi_device *device) > } > } > > - if (device->pnp.type.platform_id && !device->handler) > - acpi_create_platform_device(device, NULL); > + acpi_do_default_enumeration(device); > > return ret; > } > diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h > index ec92ad3..4724fe5 100644 > --- a/include/acpi/acpi_bus.h > +++ b/include/acpi/acpi_bus.h > @@ -137,6 +137,7 @@ struct acpi_scan_handler { > void (*bind)(struct device *phys_dev); > void (*unbind)(struct device *phys_dev); > struct acpi_hotplug_profile hotplug; > + bool handle_children; > }; > > /* > @@ -207,7 +208,8 @@ struct acpi_device_flags { > u32 no_hotplug:1; > u32 hotplug_notify:1; > u32 is_dock_station:1; > - u32 reserved:22; > + u32 no_child_enumeration:1; > + u32 reserved:21; > }; > > /* File System */ >
On Mon, 2014-04-28 at 00:26 +0200, Rafael J. Wysocki wrote: > On Tuesday, April 08, 2014 12:06:59 AM Zhang Rui wrote: > > For some devices with scan handler attached, their children devices > > are enumerated by the scan handler, indirectly. > > This isn't the case really. They are enumerated by bus controller drivers > for the buses they are on. > that's what I mean by saying "indirectly". :) > > In this case, we do not want to enumerate the children devices in > > acpi scan code explicitly. > > > > Thus a new flag .handle_children is introduced in this patch. > > > > For scan handlers with this flag set, we will do default enumeration neither > > for the attached devices nor for the children of the attached devices. > > I'm not sure if that is the right approach. I would prefer that to be > handled in a more fine-graind manner, like a flag per device ID or something > similar? > hmmm, how about this, first, keep the device->flags.no_child_enumeration flag introduced in this patch second, set the flag explicitly, for specified devices, in the scan handler .attach() function. thanks, rui > > Signed-off-by: Zhang Rui <rui.zhang@intel.com> > > --- > > drivers/acpi/acpi_lpss.c | 2 ++ > > drivers/acpi/scan.c | 28 ++++++++++++++++++++++++++-- > > include/acpi/acpi_bus.h | 4 +++- > > 3 files changed, 31 insertions(+), 3 deletions(-) > > > > diff --git a/drivers/acpi/acpi_lpss.c b/drivers/acpi/acpi_lpss.c > > index 965428f..da0a3d4 100644 > > --- a/drivers/acpi/acpi_lpss.c > > +++ b/drivers/acpi/acpi_lpss.c > > @@ -521,6 +521,7 @@ static struct acpi_scan_handler lpss_handler = { > > .attach = acpi_lpss_create_device, > > .bind = acpi_lpss_bind, > > .unbind = acpi_lpss_unbind, > > + .handle_children = true, > > }; > > > > #endif /* CONFIG_X86_INTEL_LPSS */ > > @@ -534,6 +535,7 @@ static int acpi_lpss_dummy_attach(struct acpi_device *adev, > > static struct acpi_scan_handler lpss_dummy_handler = { > > .ids = acpi_lpss_device_ids, > > .attach = acpi_lpss_dummy_attach, > > + .handle_children = true, > > }; > > > > void __init acpi_lpss_init(void) > > diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c > > index 44c4668..4ea867e 100644 > > --- a/drivers/acpi/scan.c > > +++ b/drivers/acpi/scan.c > > @@ -2073,6 +2073,31 @@ static acpi_status acpi_bus_check_add(acpi_handle handle, u32 lvl_not_used, > > return AE_OK; > > } > > > > +static void acpi_do_default_enumeration(struct acpi_device *device) > > +{ > > + /* > > + * Do not do enumeration for device object that > > + * its parent doesn't want to > > + */ > > + if (device->parent && device->parent->flags.no_child_enumeration) { > > + device->flags.no_child_enumeration = 1; > > + return; > > + } > > + > > + /* Do not do enumeration for device object with scan handler attached */ > > + if (device->handler) { > > + if (device->handler->handle_children) > > + device->flags.no_child_enumeration = 1; > > + return; > > + } > > + > > + /* Do not do enumeration for device object w/o platform_id */ > > + if (!device->pnp.type.platform_id) > > + return; > > + > > + acpi_create_platform_device(device, NULL); > > +} > > + > > static int acpi_scan_attach_handler(struct acpi_device *device) > > { > > struct acpi_hardware_id *hwid; > > @@ -2095,8 +2120,7 @@ static int acpi_scan_attach_handler(struct acpi_device *device) > > } > > } > > > > - if (device->pnp.type.platform_id && !device->handler) > > - acpi_create_platform_device(device, NULL); > > + acpi_do_default_enumeration(device); > > > > return ret; > > } > > diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h > > index ec92ad3..4724fe5 100644 > > --- a/include/acpi/acpi_bus.h > > +++ b/include/acpi/acpi_bus.h > > @@ -137,6 +137,7 @@ struct acpi_scan_handler { > > void (*bind)(struct device *phys_dev); > > void (*unbind)(struct device *phys_dev); > > struct acpi_hotplug_profile hotplug; > > + bool handle_children; > > }; > > > > /* > > @@ -207,7 +208,8 @@ struct acpi_device_flags { > > u32 no_hotplug:1; > > u32 hotplug_notify:1; > > u32 is_dock_station:1; > > - u32 reserved:22; > > + u32 no_child_enumeration:1; > > + u32 reserved:21; > > }; > > > > /* File System */ > > > -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 4/28/2014 4:07 AM, Zhang Rui wrote: > On Mon, 2014-04-28 at 00:26 +0200, Rafael J. Wysocki wrote: >> On Tuesday, April 08, 2014 12:06:59 AM Zhang Rui wrote: >>> For some devices with scan handler attached, their children devices >>> are enumerated by the scan handler, indirectly. >> This isn't the case really. They are enumerated by bus controller drivers >> for the buses they are on. >> > that's what I mean by saying "indirectly". :) > >>> In this case, we do not want to enumerate the children devices in >>> acpi scan code explicitly. >>> >>> Thus a new flag .handle_children is introduced in this patch. >>> >>> For scan handlers with this flag set, we will do default enumeration neither >>> for the attached devices nor for the children of the attached devices. >> I'm not sure if that is the right approach. I would prefer that to be >> handled in a more fine-graind manner, like a flag per device ID or something >> similar? >> > hmmm, how about this, > first, keep the device->flags.no_child_enumeration flag introduced in > this patch > second, set the flag explicitly, for specified devices, in the scan > handler .attach() function. But then it could simply clear the platform_id flag for them, couldn't it? Rafael -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/acpi/acpi_lpss.c b/drivers/acpi/acpi_lpss.c index 965428f..da0a3d4 100644 --- a/drivers/acpi/acpi_lpss.c +++ b/drivers/acpi/acpi_lpss.c @@ -521,6 +521,7 @@ static struct acpi_scan_handler lpss_handler = { .attach = acpi_lpss_create_device, .bind = acpi_lpss_bind, .unbind = acpi_lpss_unbind, + .handle_children = true, }; #endif /* CONFIG_X86_INTEL_LPSS */ @@ -534,6 +535,7 @@ static int acpi_lpss_dummy_attach(struct acpi_device *adev, static struct acpi_scan_handler lpss_dummy_handler = { .ids = acpi_lpss_device_ids, .attach = acpi_lpss_dummy_attach, + .handle_children = true, }; void __init acpi_lpss_init(void) diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c index 44c4668..4ea867e 100644 --- a/drivers/acpi/scan.c +++ b/drivers/acpi/scan.c @@ -2073,6 +2073,31 @@ static acpi_status acpi_bus_check_add(acpi_handle handle, u32 lvl_not_used, return AE_OK; } +static void acpi_do_default_enumeration(struct acpi_device *device) +{ + /* + * Do not do enumeration for device object that + * its parent doesn't want to + */ + if (device->parent && device->parent->flags.no_child_enumeration) { + device->flags.no_child_enumeration = 1; + return; + } + + /* Do not do enumeration for device object with scan handler attached */ + if (device->handler) { + if (device->handler->handle_children) + device->flags.no_child_enumeration = 1; + return; + } + + /* Do not do enumeration for device object w/o platform_id */ + if (!device->pnp.type.platform_id) + return; + + acpi_create_platform_device(device, NULL); +} + static int acpi_scan_attach_handler(struct acpi_device *device) { struct acpi_hardware_id *hwid; @@ -2095,8 +2120,7 @@ static int acpi_scan_attach_handler(struct acpi_device *device) } } - if (device->pnp.type.platform_id && !device->handler) - acpi_create_platform_device(device, NULL); + acpi_do_default_enumeration(device); return ret; } diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h index ec92ad3..4724fe5 100644 --- a/include/acpi/acpi_bus.h +++ b/include/acpi/acpi_bus.h @@ -137,6 +137,7 @@ struct acpi_scan_handler { void (*bind)(struct device *phys_dev); void (*unbind)(struct device *phys_dev); struct acpi_hotplug_profile hotplug; + bool handle_children; }; /* @@ -207,7 +208,8 @@ struct acpi_device_flags { u32 no_hotplug:1; u32 hotplug_notify:1; u32 is_dock_station:1; - u32 reserved:22; + u32 no_child_enumeration:1; + u32 reserved:21; }; /* File System */
For some devices with scan handler attached, their children devices are enumerated by the scan handler, indirectly. In this case, we do not want to enumerate the children devices in acpi scan code explicitly. Thus a new flag .handle_children is introduced in this patch. For scan handlers with this flag set, we will do default enumeration neither for the attached devices nor for the children of the attached devices. Signed-off-by: Zhang Rui <rui.zhang@intel.com> --- drivers/acpi/acpi_lpss.c | 2 ++ drivers/acpi/scan.c | 28 ++++++++++++++++++++++++++-- include/acpi/acpi_bus.h | 4 +++- 3 files changed, 31 insertions(+), 3 deletions(-)