ACPI: platform: Fix SMB0001 enumeration on Kontron devices
v4 cover ACPI: platform: Fix SMB0001 enumeration on Kontron devices
After switching i2c-scmi driver to be a plaform one, it stopped being enumerated on number of Kontron platforms, because it's listed in the forbidden_id_list.
To resolve the situation, add a flag and check if there are resources privided for the forbiden device. If it is the case, the device must be skipped.
Changelog v4:
- dropped applied patch
- added kernel doc for new helper (Rafael)
- rewritten the commit message for the reason of a new helper (Rafael)
- added a couple of patches for ACPI scan.c that can be applied later on
Changelog v3:
- provided completely rewritten solution (Rafael)
- due to above, added two new patches
- due to above, dropped tags from patch 3
Andy Shevchenko (5):
ACPI: bus: Introduce acpi_match_acpi_device() helper
ACPI: platform: Ignore SMB0001 only when it has resources
ACPI: platform: Move SMB0001 HID to the header and reuse
ACPI: scan: Use the acpi_match_acpi_device() helper
ACPI: scan: Provide symbol declarations
drivers/acpi/acpi_platform.c | 30 +++++++++++++++++++++++++++--- drivers/acpi/bus.c | 25 +++++++++++++++++++++---- drivers/acpi/scan.c | 22 +++++++++++++--------- drivers/i2c/busses/i2c-scmi.c | 3 --- include/acpi/acpi_drivers.h | 2 ++ include/linux/acpi.h | 9 +++++++++ 6 files changed, 72 insertions(+), 19 deletions(-)
Signature · 1 line
2.40.0.1.gaa8946217a0b
PATCH v4 1/5 ACPI: bus: Introduce acpi_match_acpi_device() helper
Match the ACPI device against a given list of ACPI IDs. This may be helpful in different cases.
--- drivers/acpi/bus.c | 25 +++++++++++++++++++++---- include/linux/acpi.h | 9 +++++++++ 2 files changed, 30 insertions(+), 4 deletions(-)
drivers/acpi/bus.c +21 -4
@@ -850,6 +850,26 @@ static bool __acpi_match_device(const struct acpi_device *device,return true; } +/** +* acpi_match_acpi_device - Match an ACPI device against a given list of ACPI IDs +* @ids: Array of struct acpi_device_id objects to match against. +* @adev: The ACPI device pointer to match. +* +* Match the ACPI device @adev against a given list of ACPI IDs @ids. +* +* Return: +* a pointer to the first matching ACPI ID on success or %NULL on failure. +*/ +const struct acpi_device_id *acpi_match_acpi_device (const struct acpi_device_id *ids, +const struct acpi_device *adev) +{ +const struct acpi_device_id *id =NULL ; + +__acpi_match_device (adev, ids,NULL , &id,NULL ); +return id; +} +EXPORT_SYMBOL_GPL (acpi_match_acpi_device); +/** * acpi_match_device - Match a struct device against a given list of ACPI IDs * @ids: Array of struct acpi_device_id object to match against.
@@ -864,10 +884,7 @@ static bool __acpi_match_device(const struct acpi_device *device,const struct acpi_device_id *acpi_match_device (const struct acpi_device_id *ids,const struct device *dev) { -const struct acpi_device_id *id =NULL ; - -__acpi_match_device (acpi_companion_match (dev), ids,NULL , &id,NULL ); -return id; +return acpi_match_acpi_device (ids,acpi_companion_match (dev)); }EXPORT_SYMBOL_GPL (acpi_match_device);
include/linux/acpi.h +9 -0
@@ -707,6 +707,9 @@ extern int acpi_nvs_register(__u64 start, __u64 size);extern int acpi_nvs_for_each_region (int (*func)(__u64 ,__u64 ,void *),void *data); +const struct acpi_device_id *acpi_match_acpi_device (const struct acpi_device_id *ids, +const struct acpi_device *adev); +const struct acpi_device_id *acpi_match_device (const struct acpi_device_id *ids,const struct device *dev);
@@ -922,6 +925,12 @@ static inline int acpi_nvs_for_each_region(int (*func)(__u64, __u64, void *),struct acpi_device_id ; +static inline const struct acpi_device_id *acpi_match_acpi_device ( +const struct acpi_device_id *ids,const struct acpi_device *adev) +{ +return NULL ; +} +static inline const struct acpi_device_id *acpi_match_device (const struct acpi_device_id *ids,const struct device *dev) {
Signature · 1 line
2.40.0.1.gaa8946217a0b
PATCH v4 2/5 ACPI: platform: Ignore SMB0001 only when it has resources
In reply to Andy ShevchenkoAfter switching i2c-scmi driver to be a platform one, it stopped being enumerated on number of Kontron platforms, because it's listed in the forbidden_id_list.
To resolve the situation, add a flag to driver data to allow devices with no resources in _CRS to be enumerated via platform bus.
--- drivers/acpi/acpi_platform.c | 30 +++++++++++++++++++++++++++--- 1 file changed, 27 insertions(+), 3 deletions(-)
drivers/acpi/acpi_platform.c +27 -3
@@ -9,6 +9,7 @@*/ #include <linux/acpi.h> +#include <linux/bits.h> #include <linux/device.h> #include <linux/err.h> #include <linux/kernel.h>
@@ -19,13 +20,16 @@#include "internal.h" +/* Exclude devices that have no _CRS resources provided */ +#define ACPI_ALLOW_WO_RESOURCES BIT(0) +static const struct acpi_device_id forbidden_id_list[] = { {"ACPI0009" ,0 },/* IOxAPIC */ {"ACPI000A" ,0 },/* IOAPIC */ {"PNP0000" ,0 },/* PIC */ {"PNP0100" ,0 },/* Timer */ {"PNP0200" ,0 },/* AT DMA Controller */ - {"SMB0001" ,0 },/* ACPI SMBUS virtual device */ + {"SMB0001" ,ACPI_ALLOW_WO_RESOURCES },/* ACPI SMBUS virtual device */ { } };
@@ -83,6 +87,15 @@ static void acpi_platform_fill_resource(struct acpi_device *adev, dest->parent =pci_find_resource (to_pci_dev (parent), dest); } +static unsigned int acpi_platform_resource_count (struct acpi_resource *ares,void *data) +{ +bool *has_resources = data; + + *has_resources = true; + +return AE_CTRL_TERMINATE ; +} +/** * acpi_create_platform_device - Create platform device for ACPI device node * @adev: ACPI device node to create a platform device for.
@@ -100,17 +113,28 @@ struct platform_device *acpi_create_platform_device(struct acpi_device *adev,struct acpi_device *parent =acpi_dev_parent (adev);struct platform_device *pdev =NULL ;struct platform_device_info pdevinfo; +const struct acpi_device_id *match;struct resource_entry *rentry;struct list_head resource_list;struct resource *resources =NULL ; +bool has_resources;int count;/* If the ACPI node already has a physical device attached, skip it. */ if (adev->physical_node_count)return NULL ; -if (!acpi_match_device_ids (adev, forbidden_id_list)) -return ERR_PTR (-EINVAL ); + match =acpi_match_acpi_device (forbidden_id_list, adev); +if (match) { +if (match->driver_data &ACPI_ALLOW_WO_RESOURCES ) { +acpi_walk_resources (adev->handle,METHOD_NAME__CRS , + acpi_platform_resource_count, &has_resources); +if (has_resources) +return ERR_PTR (-EINVAL ); + }else { +return ERR_PTR (-EINVAL ); + } + }INIT_LIST_HEAD (&resource_list); count =acpi_dev_get_resources (adev, &resource_list,NULL ,NULL );
Signature · 1 line
2.40.0.1.gaa8946217a0b
Rafael J. Wysocki
maintainerRe: [PATCH v4 2/5] ACPI: platform: Ignore SMB0001 only when it has resources
Quoted 85 lines · Andy Shevchenko
After switching i2c-scmi driver to be a platform one, it stopped being enumerated on number of Kontron platforms, because it's listed in the
forbidden_id_list.To resolve the situation, add a flag to driver data to allow devices with no resources in
_CRSto be enumerated via platform bus.Fixes 03d4287add6e ("i2c: scmi: Convert to be a platform driver")Signed-off-by Andy Shevchenko <[email protected]>Reviewed-by Andi Shyti <[email protected]>--- drivers/acpi/acpi_platform.c | 30 +++++++++++++++++++++++++++--- 1 file changed, 27 insertions(+), 3 deletions(-)drivers/acpi/acpi_platform.c +21
-3@@ -9,6 +9,7 @@*/ #include <linux/acpi.h> +#include <linux/bits.h> #include <linux/device.h> #include <linux/err.h> #include <linux/kernel.h> @@ -19,13 +20,16 @@#include "internal.h" +/* Exclude devices that have no _CRS resources provided */ +#define ACPI_ALLOW_WO_RESOURCES BIT(0) +static const struct acpi_device_id forbidden_id_list[] = { {"ACPI0009" ,0 },/* IOxAPIC */ {"ACPI000A" ,0 },/* IOAPIC */ {"PNP0000" ,0 },/* PIC */ {"PNP0100" ,0 },/* Timer */ {"PNP0200" ,0 },/* AT DMA Controller */ - {"SMB0001" ,0 },/* ACPI SMBUS virtual device */ + {"SMB0001" ,ACPI_ALLOW_WO_RESOURCES },/* ACPI SMBUS virtual device */ { } };@@ -83,6 +87,15 @@ static void acpi_platform_fill_resource(struct acpi_device *adev, dest->parent =pci_find_resource (to_pci_dev (parent), dest); } +static unsigned int acpi_platform_resource_count (struct acpi_resource *ares,void *data) +{ +bool *has_resources = data; + + *has_resources = true; + +return AE_CTRL_TERMINATE ; +} +/** * acpi_create_platform_device - Create platform device for ACPI device node * @adev: ACPI device node to create a platform device for. @@ -100,17 +113,28 @@ struct platform_device *acpi_create_platform_device(struct acpi_device *adev,struct acpi_device *parent =acpi_dev_parent (adev);struct platform_device *pdev =NULL ;struct platform_device_info pdevinfo; +const struct acpi_device_id *match;struct resource_entry *rentry;struct list_head resource_list;struct resource *resources =NULL ; +bool has_resources;int count;/* If the ACPI node already has a physical device attached, skip it. */ if (adev->physical_node_count)return NULL ; -if (!acpi_match_device_ids (adev, forbidden_id_list)) -return ERR_PTR (-EINVAL ); + match =acpi_match_acpi_device (forbidden_id_list, adev); +if (match) { +if (match->driver_data &ACPI_ALLOW_WO_RESOURCES ) { +acpi_walk_resources (adev->handle,METHOD_NAME__CRS , + acpi_platform_resource_count, &has_resources);
drivers/acpi/acpi_platform.c struct platform_device *acpi_create_platform_device(struct acpi_device *adev,
Quoted 10 lines
drivers/acpi/acpi_platform.c +6
-1+if (has_resources) +return ERR_PTR (-EINVAL ); + }else { +return ERR_PTR (-EINVAL ); + } + }INIT_LIST_HEAD (&resource_list); count =acpi_dev_get_resources (adev, &resource_list,NULL ,NULL ); --
Applied as 6.5-rc material along with the rest of the series (except for the last patch that has been superseded), but I moved the has_resources definition to the block where it is used and added a 'false' initial value to it. as the code might not work as expected without initializing it.
Please double check the bleeding-edge branch of linux-pm.
Thanks!
Re: [PATCH v4 2/5] ACPI: platform: Ignore SMB0001 only when it has resources
Quoted 2 lines · Rafael J. Wysocki
On Mon, Jul 3, 2023 at 2:14 PM Andy Shevchenko
<[email protected]> wrote:
...
Quoted 7 lines
Applied as 6.5-rc material along with the rest of the series (except for the last patch that has been superseded), but I moved the
has_resourcesdefinition to the block where it is used and added a 'false' initial value to it. as the code might not work as expected without initializing it.Please double check the bleeding-edge branch of linux-pm.
Now it's in Linux Next and everything looks fine to me, thank you!
Signature · 2 lines
With Best Regards, Andy Shevchenko
PATCH v4 3/5 ACPI: platform: Move SMB0001 HID to the header and reuse
In reply to Andy ShevchenkoThere are at least two places in the kernel that are using the SMB0001 HID. Make it to be available via acpi_drivers.h header file. While at it, replace hard coded one with a definition.
--- drivers/acpi/acpi_platform.c | 2 +- drivers/i2c/busses/i2c-scmi.c | 3 --- include/acpi/acpi_drivers.h | 2 ++ 3 files changed, 3 insertions(+), 4 deletions(-)
drivers/acpi/acpi_platform.c +1 -1
@@ -29,7 +29,7 @@ static const struct acpi_device_id forbidden_id_list[] = { {"PNP0000", 0}, /* PIC */ {"PNP0100", 0}, /* Timer */ {"PNP0200", 0}, /* AT DMA Controller */ - {"SMB0001", ACPI_ALLOW_WO_RESOURCES}, /* ACPI SMBUS virtual device */ + {ACPI_SMBUS_MS_HID, ACPI_ALLOW_WO_RESOURCES}, /* ACPI SMBUS virtual device */ { } };
drivers/i2c/busses/i2c-scmi.c +0 -3
@@ -13,9 +13,6 @@#include <linux/i2c.h> #include <linux/acpi.h> -/* SMBUS HID definition as supported by Microsoft Windows */ -#define ACPI_SMBUS_MS_HID "SMB0001" -struct smbus_methods_t {char *mt_info;char *mt_sbr;
include/acpi/acpi_drivers.h +2 -0
@@ -27,6 +27,8 @@#define ACPI_BAY_HID "LNXIOBAY"#define ACPI_DOCK_HID "LNXDOCK"#define ACPI_ECDT_HID "LNXEC" +/* SMBUS HID definition as supported by Microsoft Windows */ +#define ACPI_SMBUS_MS_HID "SMB0001"/* Quirk for broken IBM BIOSes */ #define ACPI_SMBUS_IBM_HID "SMBUSIBM"
Signature · 1 line
2.40.0.1.gaa8946217a0b
PATCH v4 4/5 ACPI: scan: Use the acpi_match_acpi_device() helper
In reply to Andy ShevchenkoInstead of doing two pass parsing of the table, replace acpi_match_device_ids() with acpi_match_acpi_device().
--- drivers/acpi/scan.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-)
drivers/acpi/scan.c +12 -9
@@ -928,26 +928,29 @@ static int acpi_bus_extract_wakeup_device_power_package(struct acpi_device *dev)return err; } +/* Do not use a button for S5 wakeup */ +#define ACPI_AVOID_WAKE_FROM_S5 BIT(0) +static bool acpi_wakeup_gpe_init (struct acpi_device *device) {static const struct acpi_device_id button_device_ids[] = { - {"PNP0C0C" ,0 },/* Power button */ - {"PNP0C0D" ,0 },/* Lid */ - {"PNP0C0E" ,0 },/* Sleep button */ + {"PNP0C0C" ,0 },/* Power button */ + {"PNP0C0D" ,ACPI_AVOID_WAKE_FROM_S5 },/* Lid */ + {"PNP0C0E" ,ACPI_AVOID_WAKE_FROM_S5 },/* Sleep button */ {"" ,0 }, };struct acpi_device_wakeup *wakeup = &device->wakeup; +const struct acpi_device_id *match;acpi_status status; wakeup->flags.notifier_present =0 ;/* Power button, Lid switch always enable wakeup */ -if (!acpi_match_device_ids (device, button_device_ids)) { -if (!acpi_match_device_ids (device, &button_device_ids[1 ])) { -/* Do not use Lid/sleep button for S5 wakeup */ -if (wakeup->sleep_state ==ACPI_STATE_S5 ) - wakeup->sleep_state =ACPI_STATE_S4 ; - } + match =acpi_match_acpi_device (button_device_ids, device); +if (match) { +if ((match->driver_data &ACPI_AVOID_WAKE_FROM_S5 ) && + wakeup->sleep_state ==ACPI_STATE_S5 ) + wakeup->sleep_state =ACPI_STATE_S4 ;acpi_mark_gpe_for_wake (wakeup->gpe_device, wakeup->gpe_number);device_set_wakeup_capable (&device->dev, true);return true;
Signature · 1 line
2.40.0.1.gaa8946217a0b
PATCH v4 5/5 ACPI: scan: Provide symbol declarations
In reply to Andy ShevchenkoCompiler is not happy about a couple of symbols that, it thinks, are not declared:
warning: symbol 'acpi_device_lock' was not declared. Should it be static?
warning: symbol 'acpi_wakeup_device_list' was not declared. Should it be static?
Include "sleep.h" to have them explicitly declared and make the compiler happy.
--- drivers/acpi/scan.c | 1 + 1 file changed, 1 insertion(+)
drivers/acpi/scan.c +1 -0
@@ -23,6 +23,7 @@#include <linux/dma-direct.h> #include "internal.h" +#include "sleep.h" #define ACPI_BUS_CLASS "system_bus"#define ACPI_BUS_HID "LNXSYBUS"
Signature · 1 line
2.40.0.1.gaa8946217a0b
Rafael J. Wysocki
maintainerRe: [PATCH v4 5/5] ACPI: scan: Provide symbol declarations
Quoted 11 lines · Andy Shevchenko
Compiler is not happy about a couple of symbols that, it thinks, are not declared:
warning: symbol '
acpi_device_lock' was not declared. Should it be static?
warning: symbol 'acpi_wakeup_device_list' was not declared. Should it be static?Include "
sleep.h" to have them explicitly declared and make the compiler happy.Signed-off-by Andy Shevchenko <[email protected]>
I've just applied a patch from Ben Dooks making the exact same change.
Quoted 17 lines
--- drivers/acpi/scan.c | 1 + 1 file changed, 1 insertion(+)drivers/acpi/scan.c +1
-1@@ -23,6 +23,7 @@#include <linux/dma-direct.h> #include "internal.h" +#include "sleep.h" #define ACPI_BUS_CLASS "system_bus"#define ACPI_BUS_HID "LNXSYBUS" --
Quoted 7 lines · Andy Shevchenko
After switching i2c-scmi driver to be a plaform one, it stopped being enumerated on number of Kontron platforms, because it's listed in the
forbidden_id_list.To resolve the situation, add a flag and check if there are resources privided for the forbiden device. If it is the case, the device must be skipped.
...
Quoted 5 lines
Changelog v4:
- dropped applied patch
- added kernel doc for new helper (Rafael)
- rewritten the commit message for the reason of a new helper (Rafael)
- added a couple of patches for ACPIscan.cthat can be applied later on
- converted to use boolean instead of integer (Rafael)
- added tag (Andi)
Quoted 4 lines
Changelog v3:
- provided completely rewritten solution (Rafael)
- due to above, added two new patches
- due to above, dropped tags from patch 3
Signature · 2 lines
With Best Regards, Andy Shevchenko