[PATCH v5 00/17] Introduce PMF Smart PC Solution Builder Feature

[PATCH v5 00/17] Introduce PMF Smart PC Solution Builder Feature

[Shyam Sundar S K]

Smart PC Solutions Builder allows for OEM to define a large number of
custom system states to dynamically switch to. The system states are
referred to as policies, and multiple policies can be loaded onto the
system at any given time, however only one policy can be active at a
given time.

Policy is a combination of PMF input and output capabilities. The inputs
are the incoming information from the other kernel subsystems like LID
state, Sensor info, GPU info etc and the actions are the updating the 
power limits of SMU etc.

The policy binary is signed and encrypted by a special key from AMD. This
policy binary shall have the inputs and outputs which the OEMs can build
for the platform customization that can enhance the user experience and
system behavior.

This series adds the initial support for Smart PC solution to PMF driver.

Note that, on platforms where CnQF and Smart PC is advertised, Smart PC
shall have higher precedence and same applies for Auto Mode.

v4->v5:
---------
 - Remove PMF-GPU interface from amdgpu driver and add DRM/backlight
   changes within PMF
 - Add module_softdep for AMDGPU
 - remove error checks for debugfs_create_file()
 - Add "Reviewed-by:" tags
 - Add kerneldoc for kernel-wide headers
 - Add checks for acpi_backlight_native
 - Add early return for SFH path
 - other cosmetic changes
 
v3->v4:
---------
- Split v3 9/16 into 2 patches, that addresses using generic fn names
- Add softdep [Ilpo] instead of request_module()
- return proper ACPI status [Mario]
- Update comments in code [Mario]
- Remove missed double _ remarks
- handle battery status branches [Ilpo]
- Address KASAN problems 

v2->v3:
---------
- Remove pci_get_device() for getting gpu handle
- add .suspend handler for pmf driver
- remove unwanted type caste
- Align comments, spaces etc.
- add wrapper for print_hex_dump_debug()
- Remove lkp tags in commit-msg
- Add macros for magic numbers
- use right format specifiers for printing
- propagate error codes back to the caller
- remove unwanted comments


v1->v2:
---------
- Remove __func__ macros
- Remove manual function names inside prints
- Handle tee_shm_get_va() failure
- Remove double _
- Add meaningful prints
- pass amd_pmf_set_dram_addr() failure errors
- Add more information to commit messages
- use right format specifiers
- use devm_ioremap() instead of ioremap()
- address unsigned long vs u32 problems
- Fix lkp reported issues
- Add amd_pmf_remove_pb() to remove the debugfs files created(if any).
- Make amd_pmf_open_pb() as static.
- Add cooling device APIs for controlling amdgpu backlight
- handle amd_pmf_apply_policies() failures
- Split v1 14/15 into 2 patches further
- use linux/units.h for better handling
- add "depends on" AMD_SFH_HID for interaction with SFH
- other cosmetic remarks

Basavaraj Natikar (3):
  HID: amd_sfh: rename float_to_int() to amd_sfh_float_to_int()
  platform/x86/amd/pmf: Add PMF-AMDSFH interface for HPD
  platform/x86/amd/pmf: Add PMF-AMDSFH interface for ALS

Shyam Sundar S K (14):
  platform/x86/amd/pmf: Add PMF TEE interface
  platform/x86/amd/pmf: Add support for PMF-TA interaction
  platform/x86/amd/pmf: Change return type of amd_pmf_set_dram_addr()
  platform/x86/amd/pmf: Add support for PMF Policy Binary
  platform/x86/amd/pmf: change amd_pmf_init_features() call sequence
  platform/x86/amd/pmf: Add support to get inputs from other subsystems
  platform/x86/amd/pmf: Add support update p3t limit
  platform/x86/amd/pmf: Add support to update system state
  platform/x86/amd/pmf: Make source_as_str() as non-static
  platform/x86/amd/pmf: Add facility to dump TA inputs
  platform/x86/amd/pmf: Add capability to sideload of policy binary
  platform/x86/amd/pmf: dump policy binary data
  platform/x86/amd/pmf: Add PMF-AMDGPU get interface
  platform/x86/amd/pmf: Add PMF-AMDGPU set interface

 Documentation/admin-guide/index.rst           |   1 +
 Documentation/admin-guide/pmf.rst             |  24 +
 drivers/hid/amd-sfh-hid/amd_sfh_common.h      |   6 +
 drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_desc.c |  28 +-
 drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c |  17 +
 .../amd-sfh-hid/sfh1_1/amd_sfh_interface.c    |  47 ++
 .../amd-sfh-hid/sfh1_1/amd_sfh_interface.h    |   1 +
 drivers/platform/x86/amd/pmf/Kconfig          |   2 +
 drivers/platform/x86/amd/pmf/Makefile         |   3 +-
 drivers/platform/x86/amd/pmf/acpi.c           |  37 ++
 drivers/platform/x86/amd/pmf/core.c           |  64 +-
 drivers/platform/x86/amd/pmf/pmf.h            | 223 +++++++
 drivers/platform/x86/amd/pmf/spc.c            | 227 +++++++
 drivers/platform/x86/amd/pmf/sps.c            |   5 +-
 drivers/platform/x86/amd/pmf/tee-if.c         | 564 ++++++++++++++++++
 include/linux/amd-pmf-io.h                    |  50 ++
 16 files changed, 1269 insertions(+), 30 deletions(-)
 create mode 100644 Documentation/admin-guide/pmf.rst
 create mode 100644 drivers/platform/x86/amd/pmf/spc.c
 create mode 100644 drivers/platform/x86/amd/pmf/tee-if.c
 create mode 100644 include/linux/amd-pmf-io.h

-- 
2.25.1

[PATCH v5 01/17] platform/x86/amd/pmf: Add PMF TEE interface

[Shyam Sundar S K]

AMD PMF driver loads the PMF TA (Trusted Application) into the AMD
ASP's (AMD Security Processor) TEE (Trusted Execution Environment).

PMF Trusted Application is a secured firmware placed under
/lib/firmware/amdtee gets loaded only when the TEE environment is
initialized. Add the initial code path to build these pipes.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/Kconfig  |   1 +
 drivers/platform/x86/amd/pmf/Makefile |   3 +-
 drivers/platform/x86/amd/pmf/core.c   |  10 ++-
 drivers/platform/x86/amd/pmf/pmf.h    |  16 ++++
 drivers/platform/x86/amd/pmf/tee-if.c | 105 ++++++++++++++++++++++++++
 5 files changed, 130 insertions(+), 5 deletions(-)
 create mode 100644 drivers/platform/x86/amd/pmf/tee-if.c

diff --git a/drivers/platform/x86/amd/pmf/Kconfig b/drivers/platform/x86/amd/pmf/Kconfig
index 3064bc8ea167..32a029e8db80 100644
--- a/drivers/platform/x86/amd/pmf/Kconfig
+++ b/drivers/platform/x86/amd/pmf/Kconfig
@@ -9,6 +9,7 @@ config AMD_PMF
 	depends on POWER_SUPPLY
 	depends on AMD_NB
 	select ACPI_PLATFORM_PROFILE
+	depends on TEE
 	help
 	  This driver provides support for the AMD Platform Management Framework.
 	  The goal is to enhance end user experience by making AMD PCs smarter,
diff --git a/drivers/platform/x86/amd/pmf/Makefile b/drivers/platform/x86/amd/pmf/Makefile
index fdededf54392..d2746ee7369f 100644
--- a/drivers/platform/x86/amd/pmf/Makefile
+++ b/drivers/platform/x86/amd/pmf/Makefile
@@ -6,4 +6,5 @@
 
 obj-$(CONFIG_AMD_PMF) += amd-pmf.o
 amd-pmf-objs := core.o acpi.o sps.o \
-		auto-mode.o cnqf.o
+		auto-mode.o cnqf.o \
+		tee-if.o
diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c
index 78ed3ee22555..ec92d1cc0dac 100644
--- a/drivers/platform/x86/amd/pmf/core.c
+++ b/drivers/platform/x86/amd/pmf/core.c
@@ -309,13 +309,13 @@ static void amd_pmf_init_features(struct amd_pmf_dev *dev)
 		dev_dbg(dev->dev, "SPS enabled and Platform Profiles registered\n");
 	}
 
-	/* Enable Auto Mode */
-	if (is_apmf_func_supported(dev, APMF_FUNC_AUTO_MODE)) {
+	if (!amd_pmf_init_smart_pc(dev)) {
+		dev_dbg(dev->dev, "Smart PC Solution Enabled\n");
+	} else if (is_apmf_func_supported(dev, APMF_FUNC_AUTO_MODE)) {
 		amd_pmf_init_auto_mode(dev);
 		dev_dbg(dev->dev, "Auto Mode Init done\n");
 	} else if (is_apmf_func_supported(dev, APMF_FUNC_DYN_SLIDER_AC) ||
 			  is_apmf_func_supported(dev, APMF_FUNC_DYN_SLIDER_DC)) {
-		/* Enable Cool n Quiet Framework (CnQF) */
 		ret = amd_pmf_init_cnqf(dev);
 		if (ret)
 			dev_warn(dev->dev, "CnQF Init failed\n");
@@ -330,7 +330,9 @@ static void amd_pmf_deinit_features(struct amd_pmf_dev *dev)
 		amd_pmf_deinit_sps(dev);
 	}
 
-	if (is_apmf_func_supported(dev, APMF_FUNC_AUTO_MODE)) {
+	if (!dev->smart_pc_enabled) {
+		amd_pmf_deinit_smart_pc(dev);
+	} else if (is_apmf_func_supported(dev, APMF_FUNC_AUTO_MODE)) {
 		amd_pmf_deinit_auto_mode(dev);
 	} else if (is_apmf_func_supported(dev, APMF_FUNC_DYN_SLIDER_AC) ||
 			  is_apmf_func_supported(dev, APMF_FUNC_DYN_SLIDER_DC)) {
diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index deba88e6e4c8..bd40458937ba 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -179,6 +179,12 @@ struct amd_pmf_dev {
 	bool cnqf_enabled;
 	bool cnqf_supported;
 	struct notifier_block pwr_src_notifier;
+	/* Smart PC solution builder */
+	struct tee_context *tee_ctx;
+	struct tee_shm *fw_shm_pool;
+	u32 session_id;
+	void *shbuf;
+	bool smart_pc_enabled;
 };
 
 struct apmf_sps_prop_granular {
@@ -389,6 +395,13 @@ struct apmf_dyn_slider_output {
 	struct apmf_cnqf_power_set ps[APMF_CNQF_MAX];
 } __packed;
 
+struct ta_pmf_shared_memory {
+	int command_id;
+	int resp_id;
+	u32 pmf_result;
+	u32 if_version;
+};
+
 /* Core Layer */
 int apmf_acpi_init(struct amd_pmf_dev *pmf_dev);
 void apmf_acpi_deinit(struct amd_pmf_dev *pmf_dev);
@@ -433,4 +446,7 @@ void amd_pmf_deinit_cnqf(struct amd_pmf_dev *dev);
 int amd_pmf_trans_cnqf(struct amd_pmf_dev *dev, int socket_power, ktime_t time_lapsed_ms);
 extern const struct attribute_group cnqf_feature_attribute_group;
 
+/* Smart PC builder Layer */
+int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev);
+void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev);
 #endif /* PMF_H */
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
new file mode 100644
index 000000000000..6ec8c3726624
--- /dev/null
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -0,0 +1,105 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * AMD Platform Management Framework Driver - TEE Interface
+ *
+ * Copyright (c) 2023, Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Author: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
+ */
+
+#include <linux/tee_drv.h>
+#include <linux/uuid.h>
+#include "pmf.h"
+
+#define MAX_TEE_PARAM	4
+static const uuid_t amd_pmf_ta_uuid = UUID_INIT(0x6fd93b77, 0x3fb8, 0x524d,
+						0xb1, 0x2d, 0xc5, 0x29, 0xb1, 0x3d, 0x85, 0x43);
+
+static int amd_pmf_amdtee_ta_match(struct tee_ioctl_version_data *ver, const void *data)
+{
+	return ver->impl_id == TEE_IMPL_ID_AMDTEE;
+}
+
+static int amd_pmf_ta_open_session(struct tee_context *ctx, u32 *id)
+{
+	struct tee_ioctl_open_session_arg sess_arg = {};
+	int rc;
+
+	export_uuid(sess_arg.uuid, &amd_pmf_ta_uuid);
+	sess_arg.clnt_login = TEE_IOCTL_LOGIN_PUBLIC;
+	sess_arg.num_params = 0;
+
+	rc = tee_client_open_session(ctx, &sess_arg, NULL);
+	if (rc < 0 || sess_arg.ret != 0) {
+		pr_err("Failed to open TEE session err:%#x, rc:%d\n", sess_arg.ret, rc);
+		return rc;
+	}
+
+	*id = sess_arg.session;
+
+	return rc;
+}
+
+static int amd_pmf_tee_init(struct amd_pmf_dev *dev)
+{
+	u32 size;
+	int ret;
+
+	dev->tee_ctx = tee_client_open_context(NULL, amd_pmf_amdtee_ta_match, NULL, NULL);
+	if (IS_ERR(dev->tee_ctx)) {
+		dev_err(dev->dev, "Failed to open TEE context\n");
+		return PTR_ERR(dev->tee_ctx);
+	}
+
+	ret = amd_pmf_ta_open_session(dev->tee_ctx, &dev->session_id);
+	if (ret) {
+		dev_err(dev->dev, "Failed to open TA session (%d)\n", ret);
+		ret = -EINVAL;
+		goto out_ctx;
+	}
+
+	size = sizeof(struct ta_pmf_shared_memory);
+	dev->fw_shm_pool = tee_shm_alloc_kernel_buf(dev->tee_ctx, size);
+	if (IS_ERR(dev->fw_shm_pool)) {
+		dev_err(dev->dev, "Failed to alloc TEE shared memory\n");
+		ret = PTR_ERR(dev->fw_shm_pool);
+		goto out_sess;
+	}
+
+	dev->shbuf = tee_shm_get_va(dev->fw_shm_pool, 0);
+	if (IS_ERR(dev->shbuf)) {
+		dev_err(dev->dev, "Failed to get TEE virtual address\n");
+		ret = PTR_ERR(dev->shbuf);
+		goto out_shm;
+	}
+	dev_dbg(dev->dev, "TEE init done\n");
+
+	return 0;
+
+out_shm:
+	tee_shm_free(dev->fw_shm_pool);
+out_sess:
+	tee_client_close_session(dev->tee_ctx, dev->session_id);
+out_ctx:
+	tee_client_close_context(dev->tee_ctx);
+
+	return ret;
+}
+
+static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
+{
+	tee_shm_free(dev->fw_shm_pool);
+	tee_client_close_session(dev->tee_ctx, dev->session_id);
+	tee_client_close_context(dev->tee_ctx);
+}
+
+int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
+{
+	return amd_pmf_tee_init(dev);
+}
+
+void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev)
+{
+	amd_pmf_tee_deinit(dev);
+}
-- 
2.25.1

[PATCH v5 02/17] platform/x86/amd/pmf: Add support for PMF-TA interaction

[Shyam Sundar S K]

PMF TA (Trusted Application) loads via the TEE environment into the
AMD ASP.

PMF-TA supports two commands:
1) Init: Initialize the TA with the PMF Smart PC policy binary and
start the policy engine. A policy is a combination of inputs and
outputs, where;
 - the inputs are the changing dynamics of the system like the user
   behaviour, system heuristics etc.
 - the outputs, which are the actions to be set on the system which
   lead to better power management and enhanced user experience.

PMF driver acts as a central manager in this case to supply the
inputs required to the TA (either by getting the information from
the other kernel subsystems or from userland)

2) Enact: Enact the output actions from the TA. The action could be
applying a new thermal limit to boost/throttle the power limits or
change system behavior.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/pmf.h    | 10 +++
 drivers/platform/x86/amd/pmf/tee-if.c | 97 ++++++++++++++++++++++++++-
 2 files changed, 106 insertions(+), 1 deletion(-)

diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index bd40458937ba..a24e34e42032 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -59,6 +59,9 @@
 #define ARG_NONE 0
 #define AVG_SAMPLE_SIZE 3
 
+/* TA macros */
+#define PMF_TA_IF_VERSION_MAJOR				1
+
 /* AMD PMF BIOS interfaces */
 struct apmf_verify_interface {
 	u16 size;
@@ -184,6 +187,7 @@ struct amd_pmf_dev {
 	struct tee_shm *fw_shm_pool;
 	u32 session_id;
 	void *shbuf;
+	struct delayed_work pb_work;
 	bool smart_pc_enabled;
 };
 
@@ -395,6 +399,12 @@ struct apmf_dyn_slider_output {
 	struct apmf_cnqf_power_set ps[APMF_CNQF_MAX];
 } __packed;
 
+/* Command ids for TA communication */
+enum ta_pmf_command {
+	TA_PMF_COMMAND_POLICY_BUILDER_INITIALIZE,
+	TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES,
+};
+
 struct ta_pmf_shared_memory {
 	int command_id;
 	int resp_id;
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index 6ec8c3726624..4036f435f1e2 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -13,9 +13,96 @@
 #include "pmf.h"
 
 #define MAX_TEE_PARAM	4
+
+/* Policy binary actions sampling frequency (in ms) */
+static int pb_actions_ms = MSEC_PER_SEC;
+#ifdef CONFIG_AMD_PMF_DEBUG
+module_param(pb_actions_ms, int, 0644);
+MODULE_PARM_DESC(pb_actions_ms, "Policy binary actions sampling frequency (default = 1000ms)");
+#endif
+
 static const uuid_t amd_pmf_ta_uuid = UUID_INIT(0x6fd93b77, 0x3fb8, 0x524d,
 						0xb1, 0x2d, 0xc5, 0x29, 0xb1, 0x3d, 0x85, 0x43);
 
+static void amd_pmf_prepare_args(struct amd_pmf_dev *dev, int cmd,
+				 struct tee_ioctl_invoke_arg *arg,
+				 struct tee_param *param)
+{
+	memset(arg, 0, sizeof(*arg));
+	memset(param, 0, MAX_TEE_PARAM * sizeof(*param));
+
+	arg->func = cmd;
+	arg->session = dev->session_id;
+	arg->num_params = MAX_TEE_PARAM;
+
+	/* Fill invoke cmd params */
+	param[0].u.memref.size = sizeof(struct ta_pmf_shared_memory);
+	param[0].attr = TEE_IOCTL_PARAM_ATTR_TYPE_MEMREF_INOUT;
+	param[0].u.memref.shm = dev->fw_shm_pool;
+	param[0].u.memref.shm_offs = 0;
+}
+
+static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
+{
+	struct ta_pmf_shared_memory *ta_sm = NULL;
+	struct tee_param param[MAX_TEE_PARAM];
+	struct tee_ioctl_invoke_arg arg;
+	int ret = 0;
+
+	if (!dev->tee_ctx)
+		return -ENODEV;
+
+	ta_sm = dev->shbuf;
+	memset(ta_sm, 0, sizeof(*ta_sm));
+	ta_sm->command_id = TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES;
+	ta_sm->if_version = PMF_TA_IF_VERSION_MAJOR;
+
+	amd_pmf_prepare_args(dev, TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES, &arg, param);
+
+	ret = tee_client_invoke_func(dev->tee_ctx, &arg, param);
+	if (ret < 0 || arg.ret != 0) {
+		dev_err(dev->dev, "TEE enact cmd failed. err: %x, ret:%d\n", arg.ret, ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int amd_pmf_invoke_cmd_init(struct amd_pmf_dev *dev)
+{
+	struct ta_pmf_shared_memory *ta_sm = NULL;
+	struct tee_param param[MAX_TEE_PARAM];
+	struct tee_ioctl_invoke_arg arg;
+	int ret = 0;
+
+	if (!dev->tee_ctx) {
+		dev_err(dev->dev, "Failed to get TEE context\n");
+		return -ENODEV;
+	}
+
+	ta_sm = dev->shbuf;
+	ta_sm->command_id = TA_PMF_COMMAND_POLICY_BUILDER_INITIALIZE;
+	ta_sm->if_version = PMF_TA_IF_VERSION_MAJOR;
+
+	amd_pmf_prepare_args(dev, TA_PMF_COMMAND_POLICY_BUILDER_INITIALIZE, &arg, param);
+
+	ret = tee_client_invoke_func(dev->tee_ctx, &arg, param);
+	if (ret < 0 || arg.ret != 0) {
+		dev_err(dev->dev, "Failed to invoke TEE init cmd. err: %x, ret:%d\n", arg.ret, ret);
+		return ret;
+	}
+
+	return ta_sm->pmf_result;
+}
+
+static void amd_pmf_invoke_cmd(struct work_struct *work)
+{
+	struct amd_pmf_dev *dev = container_of(work, struct amd_pmf_dev, pb_work.work);
+
+	amd_pmf_invoke_cmd_enact(dev);
+	schedule_delayed_work(&dev->pb_work, msecs_to_jiffies(pb_actions_ms));
+}
+
 static int amd_pmf_amdtee_ta_match(struct tee_ioctl_version_data *ver, const void *data)
 {
 	return ver->impl_id == TEE_IMPL_ID_AMDTEE;
@@ -96,10 +183,18 @@ static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
 
 int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
 {
-	return amd_pmf_tee_init(dev);
+	int ret;
+
+	ret = amd_pmf_tee_init(dev);
+	if (ret)
+		return ret;
+
+	INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
+	return 0;
 }
 
 void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev)
 {
+	cancel_delayed_work_sync(&dev->pb_work);
 	amd_pmf_tee_deinit(dev);
 }
-- 
2.25.1

[PATCH v5 03/17] platform/x86/amd/pmf: Change return type of amd_pmf_set_dram_addr()

[Shyam Sundar S K]

In the current code, the metrics table information was required only
for auto-mode or CnQF at a given time. Hence keeping the return type
of amd_pmf_set_dram_addr() as static made sense.

But with the addition of Smart PC builder feature, the metrics table
information has to be shared by the Smart PC also and this feature
resides outside of core.c.

To make amd_pmf_set_dram_addr() visible outside of core.c make it
as a non-static function and move the allocation of memory for
metrics table from amd_pmf_init_metrics_table() to amd_pmf_set_dram_addr()
as amd_pmf_set_dram_addr() is the common function to set the DRAM
address.

Add a suspend handler that can free up the allocated memory for getting
the metrics table information.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/core.c | 42 ++++++++++++++++++++++-------
 drivers/platform/x86/amd/pmf/pmf.h  |  1 +
 2 files changed, 34 insertions(+), 9 deletions(-)

diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c
index ec92d1cc0dac..f50b7ec9a625 100644
--- a/drivers/platform/x86/amd/pmf/core.c
+++ b/drivers/platform/x86/amd/pmf/core.c
@@ -251,29 +251,35 @@ static const struct pci_device_id pmf_pci_ids[] = {
 	{ }
 };
 
-static void amd_pmf_set_dram_addr(struct amd_pmf_dev *dev)
+int amd_pmf_set_dram_addr(struct amd_pmf_dev *dev)
 {
 	u64 phys_addr;
 	u32 hi, low;
 
+	/* Get Metrics Table Address */
+	dev->buf = kzalloc(sizeof(dev->m_table), GFP_KERNEL);
+	if (!dev->buf)
+		return -ENOMEM;
+
 	phys_addr = virt_to_phys(dev->buf);
 	hi = phys_addr >> 32;
 	low = phys_addr & GENMASK(31, 0);
 
 	amd_pmf_send_cmd(dev, SET_DRAM_ADDR_HIGH, 0, hi, NULL);
 	amd_pmf_send_cmd(dev, SET_DRAM_ADDR_LOW, 0, low, NULL);
+
+	return 0;
 }
 
 int amd_pmf_init_metrics_table(struct amd_pmf_dev *dev)
 {
-	/* Get Metrics Table Address */
-	dev->buf = kzalloc(sizeof(dev->m_table), GFP_KERNEL);
-	if (!dev->buf)
-		return -ENOMEM;
+	int ret;
 
 	INIT_DELAYED_WORK(&dev->work_buffer, amd_pmf_get_metrics);
 
-	amd_pmf_set_dram_addr(dev);
+	ret = amd_pmf_set_dram_addr(dev);
+	if (ret)
+		return ret;
 
 	/*
 	 * Start collecting the metrics data after a small delay
@@ -284,17 +290,35 @@ int amd_pmf_init_metrics_table(struct amd_pmf_dev *dev)
 	return 0;
 }
 
+static int amd_pmf_suspend_handler(struct device *dev)
+{
+	struct amd_pmf_dev *pdev = dev_get_drvdata(dev);
+
+	/*
+	 * Free the buffer allocated for storing the metrics table
+	 * information, as will have to allocate it freshly after
+	 * resume.
+	 */
+	kfree(pdev->buf);
+
+	return 0;
+}
+
 static int amd_pmf_resume_handler(struct device *dev)
 {
 	struct amd_pmf_dev *pdev = dev_get_drvdata(dev);
+	int ret;
 
-	if (pdev->buf)
-		amd_pmf_set_dram_addr(pdev);
+	if (pdev->buf) {
+		ret = amd_pmf_set_dram_addr(pdev);
+		if (ret)
+			return ret;
+	}
 
 	return 0;
 }
 
-static DEFINE_SIMPLE_DEV_PM_OPS(amd_pmf_pm, NULL, amd_pmf_resume_handler);
+static DEFINE_SIMPLE_DEV_PM_OPS(amd_pmf_pm, amd_pmf_suspend_handler, amd_pmf_resume_handler);
 
 static void amd_pmf_init_features(struct amd_pmf_dev *dev)
 {
diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index a24e34e42032..6a0e4c446dd3 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -421,6 +421,7 @@ int amd_pmf_init_metrics_table(struct amd_pmf_dev *dev);
 int amd_pmf_get_power_source(void);
 int apmf_install_handler(struct amd_pmf_dev *pmf_dev);
 int apmf_os_power_slider_update(struct amd_pmf_dev *dev, u8 flag);
+int amd_pmf_set_dram_addr(struct amd_pmf_dev *dev);
 
 /* SPS Layer */
 int amd_pmf_get_pprof_modes(struct amd_pmf_dev *pmf);
-- 
2.25.1

[PATCH v5 04/17] platform/x86/amd/pmf: Add support for PMF Policy Binary

[Shyam Sundar S K]

PMF Policy binary is a encrypted and signed binary that will be part
of the BIOS. PMF driver via the ACPI interface checks the existence
of Smart PC bit. If the advertised bit is found, PMF driver walks
the acpi namespace to find out the policy binary size and the address
which has to be passed to the TA during the TA init sequence.

The policy binary is comprised of inputs (or the events) and outputs
(or the actions). With the PMF ecosystem, OEMs generate the policy
binary (or could be multiple binaries) that contains a supported set
of inputs and outputs which could be specifically carved out for each
usage segment (or for each user also) that could influence the system
behavior either by enriching the user experience or/and boost/throttle
power limits.

Once the TA init command succeeds, the PMF driver sends the changing
events in the current environment to the TA for a constant sampling
frequency time (the event here could be a lid close or open) and
if the policy binary has corresponding action built within it, the
TA sends the action for it in the subsequent enact command.

If the inputs sent to the TA has no output defined in the policy
binary generated by OEMs, there will be no action to be performed
by the PMF driver.

Example policies:

1) if slider is performance ; set the SPL to 40W
Here PMF driver registers with the platform profile interface and
when the slider position is changed, PMF driver lets the TA know
about this. TA sends back an action to update the Sustained
Power Limit (SPL). PMF driver updates this limit via the PMFW mailbox.

2) if user_away ; then lock the system
Here PMF driver hooks to the AMD SFH driver to know the user presence
and send the inputs to TA and if the condition is met, the TA sends
the action of locking the system. PMF driver generates a uevent and
based on the udev rule in the userland the system gets locked with
systemctl.

The intent here is to provide the OEM's to make a policy to lock the
system when the user is away ; but the userland can make a choice to
ignore it.

and so on.

The OEMs will have an utility to create numerous such policies and
the policies shall be reviewed by AMD before signing and encrypting
them. Policies are shared between operating systems to have seemless user
experience.

Since all this action has to happen via the "amdtee" driver, currently
there is no caller for it in the kernel which can load the amdtee driver.
Without amdtee driver loading onto the system the "tee" calls shall fail
from the PMF driver. Hence an explicit MODULE_SOFTDEP has been added
to address this.

Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/Kconfig  |   2 +-
 drivers/platform/x86/amd/pmf/acpi.c   |  37 +++++++
 drivers/platform/x86/amd/pmf/core.c   |   9 ++
 drivers/platform/x86/amd/pmf/pmf.h    | 141 ++++++++++++++++++++++++
 drivers/platform/x86/amd/pmf/tee-if.c | 147 +++++++++++++++++++++++++-
 5 files changed, 333 insertions(+), 3 deletions(-)

diff --git a/drivers/platform/x86/amd/pmf/Kconfig b/drivers/platform/x86/amd/pmf/Kconfig
index 32a029e8db80..f246252bddd8 100644
--- a/drivers/platform/x86/amd/pmf/Kconfig
+++ b/drivers/platform/x86/amd/pmf/Kconfig
@@ -9,7 +9,7 @@ config AMD_PMF
 	depends on POWER_SUPPLY
 	depends on AMD_NB
 	select ACPI_PLATFORM_PROFILE
-	depends on TEE
+	depends on TEE && AMDTEE
 	help
 	  This driver provides support for the AMD Platform Management Framework.
 	  The goal is to enhance end user experience by making AMD PCs smarter,
diff --git a/drivers/platform/x86/amd/pmf/acpi.c b/drivers/platform/x86/amd/pmf/acpi.c
index 3fc5e4547d9f..4ec7957eb707 100644
--- a/drivers/platform/x86/amd/pmf/acpi.c
+++ b/drivers/platform/x86/amd/pmf/acpi.c
@@ -286,6 +286,43 @@ int apmf_install_handler(struct amd_pmf_dev *pmf_dev)
 	return 0;
 }
 
+static acpi_status apmf_walk_resources(struct acpi_resource *res, void *data)
+{
+	struct amd_pmf_dev *dev = data;
+
+	switch (res->type) {
+	case ACPI_RESOURCE_TYPE_ADDRESS64:
+		dev->policy_addr = res->data.address64.address.minimum;
+		dev->policy_sz = res->data.address64.address.address_length;
+		break;
+	case ACPI_RESOURCE_TYPE_FIXED_MEMORY32:
+		dev->policy_addr = res->data.fixed_memory32.address;
+		dev->policy_sz = res->data.fixed_memory32.address_length;
+		break;
+	}
+
+	if (!dev->policy_addr || dev->policy_sz > POLICY_BUF_MAX_SZ || dev->policy_sz == 0) {
+		pr_err("Incorrect Policy params, possibly a SBIOS bug\n");
+		return AE_ERROR;
+	}
+
+	return AE_OK;
+}
+
+int apmf_check_smart_pc(struct amd_pmf_dev *pmf_dev)
+{
+	acpi_handle ahandle = ACPI_HANDLE(pmf_dev->dev);
+	acpi_status status;
+
+	status = acpi_walk_resources(ahandle, METHOD_NAME__CRS, apmf_walk_resources, pmf_dev);
+	if (ACPI_FAILURE(status)) {
+		dev_err(pmf_dev->dev, "acpi_walk_resources failed :%d\n", status);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 void apmf_acpi_deinit(struct amd_pmf_dev *pmf_dev)
 {
 	acpi_handle ahandle = ACPI_HANDLE(pmf_dev->dev);
diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c
index f50b7ec9a625..c8f6ec4c2e2c 100644
--- a/drivers/platform/x86/amd/pmf/core.c
+++ b/drivers/platform/x86/amd/pmf/core.c
@@ -395,6 +395,14 @@ static int amd_pmf_probe(struct platform_device *pdev)
 		return -ENOMEM;
 
 	dev->dev = &pdev->dev;
+	err = apmf_check_smart_pc(dev);
+	if (err)
+		/*
+		 * Lets not return from here if Smart PC bit is not advertised in
+		 * the BIOS. This way, there will be some amount of power savings
+		 * to the user with static slider (if enabled).
+		 */
+		pr_err("PMF Smart PC not advertised in BIOS!:%d\n", err);
 
 	rdev = pci_get_domain_bus_and_slot(0, 0, PCI_DEVFN(0, 0));
 	if (!rdev || !pci_match_id(pmf_pci_ids, rdev)) {
@@ -474,3 +482,4 @@ module_platform_driver(amd_pmf_driver);
 
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("AMD Platform Management Framework Driver");
+MODULE_SOFTDEP("pre: amdtee");
diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index 6a0e4c446dd3..092be501d4d3 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -14,6 +14,11 @@
 #include <linux/acpi.h>
 #include <linux/platform_profile.h>
 
+#define POLICY_BUF_MAX_SZ		0x4b000
+#define POLICY_SIGN_COOKIE		0x31535024
+#define POLICY_COOKIE_OFFSET		0x10
+#define POLICY_COOKIE_LEN		0x14
+
 /* APMF Functions */
 #define APMF_FUNC_VERIFY_INTERFACE			0
 #define APMF_FUNC_GET_SYS_PARAMS			1
@@ -59,8 +64,21 @@
 #define ARG_NONE 0
 #define AVG_SAMPLE_SIZE 3
 
+/* Policy Actions */
+#define PMF_POLICY_SPL						2
+#define PMF_POLICY_SPPT						3
+#define PMF_POLICY_FPPT						4
+#define PMF_POLICY_SPPT_APU_ONLY				5
+#define PMF_POLICY_STT_MIN					6
+#define PMF_POLICY_STT_SKINTEMP_APU				7
+#define PMF_POLICY_STT_SKINTEMP_HS2				8
+
 /* TA macros */
 #define PMF_TA_IF_VERSION_MAJOR				1
+#define TA_PMF_ACTION_MAX					32
+#define TA_PMF_UNDO_MAX						8
+#define TA_OUTPUT_RESERVED_MEM				906
+#define MAX_OPERATION_PARAMS					4
 
 /* AMD PMF BIOS interfaces */
 struct apmf_verify_interface {
@@ -183,11 +201,16 @@ struct amd_pmf_dev {
 	bool cnqf_supported;
 	struct notifier_block pwr_src_notifier;
 	/* Smart PC solution builder */
+	unsigned char *policy_buf;
+	u32 policy_sz;
 	struct tee_context *tee_ctx;
 	struct tee_shm *fw_shm_pool;
 	u32 session_id;
 	void *shbuf;
 	struct delayed_work pb_work;
+	struct pmf_action_table *prev_data;
+	u64 policy_addr;
+	void *policy_base;
 	bool smart_pc_enabled;
 };
 
@@ -399,17 +422,134 @@ struct apmf_dyn_slider_output {
 	struct apmf_cnqf_power_set ps[APMF_CNQF_MAX];
 } __packed;
 
+enum smart_pc_status {
+	PMF_SMART_PC_ENABLED,
+	PMF_SMART_PC_DISABLED,
+};
+
+/* Smart PC - TA internals */
+enum ta_slider {
+	TA_BEST_BATTERY,
+	TA_BETTER_BATTERY,
+	TA_BETTER_PERFORMANCE,
+	TA_BEST_PERFORMANCE,
+	TA_MAX,
+};
+
 /* Command ids for TA communication */
 enum ta_pmf_command {
 	TA_PMF_COMMAND_POLICY_BUILDER_INITIALIZE,
 	TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES,
 };
 
+enum ta_pmf_error_type {
+	TA_PMF_TYPE_SUCCESS,
+	TA_PMF_ERROR_TYPE_GENERIC,
+	TA_PMF_ERROR_TYPE_CRYPTO,
+	TA_PMF_ERROR_TYPE_CRYPTO_VALIDATE,
+	TA_PMF_ERROR_TYPE_CRYPTO_VERIFY_OEM,
+	TA_PMF_ERROR_TYPE_POLICY_BUILDER,
+	TA_PMF_ERROR_TYPE_PB_CONVERT,
+	TA_PMF_ERROR_TYPE_PB_SETUP,
+	TA_PMF_ERROR_TYPE_PB_ENACT,
+	TA_PMF_ERROR_TYPE_ASD_GET_DEVICE_INFO,
+	TA_PMF_ERROR_TYPE_ASD_GET_DEVICE_PCIE_INFO,
+	TA_PMF_ERROR_TYPE_SYS_DRV_FW_VALIDATION,
+	TA_PMF_ERROR_TYPE_MAX,
+};
+
+struct pmf_action_table {
+	u32 spl;		/* in mW */
+	u32 sppt;		/* in mW */
+	u32 sppt_apuonly;	/* in mW */
+	u32 fppt;		/* in mW */
+	u32 stt_minlimit;	/* in mW */
+	u32 stt_skintemp_apu;	/* in C */
+	u32 stt_skintemp_hs2;	/* in C */
+};
+
+/* Input conditions */
+struct ta_pmf_condition_info {
+	u32 power_source;
+	u32 bat_percentage;
+	u32 power_slider;
+	u32 lid_state;
+	bool user_present;
+	u32 rsvd1[2];
+	u32 monitor_count;
+	u32 rsvd2[2];
+	u32 bat_design;
+	u32 full_charge_capacity;
+	int drain_rate;
+	bool user_engaged;
+	u32 device_state;
+	u32 socket_power;
+	u32 skin_temperature;
+	u32 rsvd3[5];
+	u32 ambient_light;
+	u32 length;
+	u32 avg_c0residency;
+	u32 max_c0residency;
+	u32 s0i3_entry;
+	u32 gfx_busy;
+	u32 rsvd4[7];
+	bool camera_state;
+	u32 workload_type;
+	u32 display_type;
+	u32 display_state;
+	u32 rsvd5[150];
+};
+
+struct ta_pmf_load_policy_table {
+	u32 table_size;
+	u8 table[POLICY_BUF_MAX_SZ];
+};
+
+/* TA initialization params */
+struct ta_pmf_init_table {
+	u32 frequency; /* SMU sampling frequency */
+	bool validate;
+	bool sku_check;
+	bool metadata_macrocheck;
+	struct ta_pmf_load_policy_table policies_table;
+};
+
+/* Everything the TA needs to Enact Policies */
+struct ta_pmf_enact_table {
+	struct ta_pmf_condition_info ev_info;
+	u32 name;
+};
+
+struct ta_pmf_action {
+	u32 action_index;
+	u32 value;
+};
+
+/* Output actions from TA */
+struct ta_pmf_enact_result {
+	u32 actions_count;
+	struct ta_pmf_action actions_list[TA_PMF_ACTION_MAX];
+	u32 undo_count;
+	struct ta_pmf_action undo_list[TA_PMF_UNDO_MAX];
+};
+
+union ta_pmf_input {
+	struct ta_pmf_enact_table enact_table;
+	struct ta_pmf_init_table init_table;
+};
+
+union ta_pmf_output {
+	struct ta_pmf_enact_result policy_apply_table;
+	u32 rsvd[TA_OUTPUT_RESERVED_MEM];
+};
+
 struct ta_pmf_shared_memory {
 	int command_id;
 	int resp_id;
 	u32 pmf_result;
 	u32 if_version;
+	union ta_pmf_output pmf_output;
+	union ta_pmf_input pmf_input;
 };
 
 /* Core Layer */
@@ -460,4 +600,5 @@ extern const struct attribute_group cnqf_feature_attribute_group;
 /* Smart PC builder Layer */
 int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev);
 void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev);
+int apmf_check_smart_pc(struct amd_pmf_dev *pmf_dev);
 #endif /* PMF_H */
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index 4036f435f1e2..468f3797a848 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -42,9 +42,77 @@ static void amd_pmf_prepare_args(struct amd_pmf_dev *dev, int cmd,
 	param[0].u.memref.shm_offs = 0;
 }
 
+static void amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_result *out)
+{
+	u32 val;
+	int idx;
+
+	for (idx = 0; idx < out->actions_count; idx++) {
+		val = out->actions_list[idx].value;
+		switch (out->actions_list[idx].action_index) {
+		case PMF_POLICY_SPL:
+			if (dev->prev_data->spl != val) {
+				amd_pmf_send_cmd(dev, SET_SPL, false, val, NULL);
+				dev_dbg(dev->dev, "update SPL: %u\n", val);
+				dev->prev_data->spl = val;
+			}
+			break;
+
+		case PMF_POLICY_SPPT:
+			if (dev->prev_data->sppt != val) {
+				amd_pmf_send_cmd(dev, SET_SPPT, false, val, NULL);
+				dev_dbg(dev->dev, "update SPPT: %u\n", val);
+				dev->prev_data->sppt = val;
+			}
+			break;
+
+		case PMF_POLICY_FPPT:
+			if (dev->prev_data->fppt != val) {
+				amd_pmf_send_cmd(dev, SET_FPPT, false, val, NULL);
+				dev_dbg(dev->dev, "update FPPT: %u\n", val);
+				dev->prev_data->fppt = val;
+			}
+			break;
+
+		case PMF_POLICY_SPPT_APU_ONLY:
+			if (dev->prev_data->sppt_apuonly != val) {
+				amd_pmf_send_cmd(dev, SET_SPPT_APU_ONLY, false, val, NULL);
+				dev_dbg(dev->dev, "update SPPT_APU_ONLY: %u\n", val);
+				dev->prev_data->sppt_apuonly = val;
+			}
+			break;
+
+		case PMF_POLICY_STT_MIN:
+			if (dev->prev_data->stt_minlimit != val) {
+				amd_pmf_send_cmd(dev, SET_STT_MIN_LIMIT, false, val, NULL);
+				dev_dbg(dev->dev, "update STT_MIN: %u\n", val);
+				dev->prev_data->stt_minlimit = val;
+			}
+			break;
+
+		case PMF_POLICY_STT_SKINTEMP_APU:
+			if (dev->prev_data->stt_skintemp_apu != val) {
+				amd_pmf_send_cmd(dev, SET_STT_LIMIT_APU, false, val, NULL);
+				dev_dbg(dev->dev, "update STT_SKINTEMP_APU: %u\n", val);
+				dev->prev_data->stt_skintemp_apu = val;
+			}
+			break;
+
+		case PMF_POLICY_STT_SKINTEMP_HS2:
+			if (dev->prev_data->stt_skintemp_hs2 != val) {
+				amd_pmf_send_cmd(dev, SET_STT_LIMIT_HS2, false, val, NULL);
+				dev_dbg(dev->dev, "update STT_SKINTEMP_HS2: %u\n", val);
+				dev->prev_data->stt_skintemp_hs2 = val;
+			}
+			break;
+		}
+	}
+}
+
 static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
 {
 	struct ta_pmf_shared_memory *ta_sm = NULL;
+	struct ta_pmf_enact_result *out = NULL;
 	struct tee_param param[MAX_TEE_PARAM];
 	struct tee_ioctl_invoke_arg arg;
 	int ret = 0;
@@ -52,7 +120,10 @@ static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
 	if (!dev->tee_ctx)
 		return -ENODEV;
 
+	memset(dev->shbuf, 0, dev->policy_sz);
 	ta_sm = dev->shbuf;
+	out = &ta_sm->pmf_output.policy_apply_table;
+
 	memset(ta_sm, 0, sizeof(*ta_sm));
 	ta_sm->command_id = TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES;
 	ta_sm->if_version = PMF_TA_IF_VERSION_MAJOR;
@@ -65,6 +136,12 @@ static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
 		return ret;
 	}
 
+	if (ta_sm->pmf_result == TA_PMF_TYPE_SUCCESS && out->actions_count) {
+		dev_dbg(dev->dev, "action count:%u result:%x\n", out->actions_count,
+			ta_sm->pmf_result);
+		amd_pmf_apply_policies(dev, out);
+	}
+
 	return 0;
 }
 
@@ -72,6 +149,7 @@ static int amd_pmf_invoke_cmd_init(struct amd_pmf_dev *dev)
 {
 	struct ta_pmf_shared_memory *ta_sm = NULL;
 	struct tee_param param[MAX_TEE_PARAM];
+	struct ta_pmf_init_table *in = NULL;
 	struct tee_ioctl_invoke_arg arg;
 	int ret = 0;
 
@@ -80,10 +158,21 @@ static int amd_pmf_invoke_cmd_init(struct amd_pmf_dev *dev)
 		return -ENODEV;
 	}
 
+	dev_dbg(dev->dev, "Policy Binary size: %u bytes\n", dev->policy_sz);
+	memset(dev->shbuf, 0, dev->policy_sz);
 	ta_sm = dev->shbuf;
+	in = &ta_sm->pmf_input.init_table;
+
 	ta_sm->command_id = TA_PMF_COMMAND_POLICY_BUILDER_INITIALIZE;
 	ta_sm->if_version = PMF_TA_IF_VERSION_MAJOR;
 
+	in->metadata_macrocheck = false;
+	in->sku_check = false;
+	in->validate = true;
+	in->frequency = pb_actions_ms;
+	in->policies_table.table_size = dev->policy_sz;
+
+	memcpy(in->policies_table.table, dev->policy_buf, dev->policy_sz);
 	amd_pmf_prepare_args(dev, TA_PMF_COMMAND_POLICY_BUILDER_INITIALIZE, &arg, param);
 
 	ret = tee_client_invoke_func(dev->tee_ctx, &arg, param);
@@ -103,6 +192,52 @@ static void amd_pmf_invoke_cmd(struct work_struct *work)
 	schedule_delayed_work(&dev->pb_work, msecs_to_jiffies(pb_actions_ms));
 }
 
+static int amd_pmf_start_policy_engine(struct amd_pmf_dev *dev)
+{
+	u32 cookie, length;
+	int res;
+
+	cookie = readl(dev->policy_buf + POLICY_COOKIE_OFFSET);
+	length = readl(dev->policy_buf + POLICY_COOKIE_LEN);
+
+	if (cookie != POLICY_SIGN_COOKIE || !length)
+		return -EINVAL;
+
+	/* Update the actual length */
+	dev->policy_sz = length + 512;
+	res = amd_pmf_invoke_cmd_init(dev);
+	if (res == TA_PMF_TYPE_SUCCESS) {
+		/* Now its safe to announce that smart pc is enabled */
+		dev->smart_pc_enabled = PMF_SMART_PC_ENABLED;
+		/*
+		 * Start collecting the data from TA FW after a small delay
+		 * or else, we might end up getting stale values.
+		 */
+		schedule_delayed_work(&dev->pb_work, msecs_to_jiffies(pb_actions_ms * 3));
+	} else {
+		dev_err(dev->dev, "ta invoke cmd init failed err: %x\n", res);
+		dev->smart_pc_enabled = PMF_SMART_PC_DISABLED;
+		return res;
+	}
+
+	return 0;
+}
+
+static int amd_pmf_get_bios_buffer(struct amd_pmf_dev *dev)
+{
+	dev->policy_buf = kzalloc(dev->policy_sz, GFP_KERNEL);
+	if (!dev->policy_buf)
+		return -ENOMEM;
+
+	dev->policy_base = devm_ioremap(dev->dev, dev->policy_addr, dev->policy_sz);
+	if (!dev->policy_base)
+		return -ENOMEM;
+
+	memcpy(dev->policy_buf, dev->policy_base, dev->policy_sz);
+
+	return amd_pmf_start_policy_engine(dev);
+}
+
 static int amd_pmf_amdtee_ta_match(struct tee_ioctl_version_data *ver, const void *data)
 {
 	return ver->impl_id == TEE_IMPL_ID_AMDTEE;
@@ -146,7 +281,7 @@ static int amd_pmf_tee_init(struct amd_pmf_dev *dev)
 		goto out_ctx;
 	}
 
-	size = sizeof(struct ta_pmf_shared_memory);
+	size = sizeof(struct ta_pmf_shared_memory) + dev->policy_sz;
 	dev->fw_shm_pool = tee_shm_alloc_kernel_buf(dev->tee_ctx, size);
 	if (IS_ERR(dev->fw_shm_pool)) {
 		dev_err(dev->dev, "Failed to alloc TEE shared memory\n");
@@ -190,11 +325,19 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
 		return ret;
 
 	INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
-	return 0;
+	amd_pmf_set_dram_addr(dev);
+	amd_pmf_get_bios_buffer(dev);
+	dev->prev_data = kzalloc(sizeof(*dev->prev_data), GFP_KERNEL);
+	if (!dev->prev_data)
+		return -ENOMEM;
+
+	return dev->smart_pc_enabled;
 }
 
 void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev)
 {
+	kfree(dev->prev_data);
+	kfree(dev->policy_buf);
 	cancel_delayed_work_sync(&dev->pb_work);
 	amd_pmf_tee_deinit(dev);
 }
-- 
2.25.1

[PATCH v5 05/17] platform/x86/amd/pmf: change amd_pmf_init_features() call sequence

[Shyam Sundar S K]

To sideload pmf policy binaries, the Smart PC Solution Builder provides a
debugfs file called "update_policy"; that gets created under a new debugfs
directory called "pb" and this new directory has to be associated with
existing parent directory for PMF driver called "amd_pmf".

In the current code structure, amd_pmf_dbgfs_register() is called after
amd_pmf_init_features(). This will not help when the Smart PC builder
feature has to be assoicated to the parent directory.

Hence change the order of amd_pmf_dbgfs_register() and call it before
amd_pmf_init_features() so that when the Smart PC init happens, it has the
parent debugfs directory to get itself hooked.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c
index c8f6ec4c2e2c..4b8156033fa6 100644
--- a/drivers/platform/x86/amd/pmf/core.c
+++ b/drivers/platform/x86/amd/pmf/core.c
@@ -442,9 +442,9 @@ static int amd_pmf_probe(struct platform_device *pdev)
 
 	apmf_acpi_init(dev);
 	platform_set_drvdata(pdev, dev);
+	amd_pmf_dbgfs_register(dev);
 	amd_pmf_init_features(dev);
 	apmf_install_handler(dev);
-	amd_pmf_dbgfs_register(dev);
 
 	dev_info(dev->dev, "registered PMF device successfully\n");
 
-- 
2.25.1

[PATCH v5 06/17] platform/x86/amd/pmf: Add support to get inputs from other subsystems

[Shyam Sundar S K]

PMF driver sends changing inputs from each subystem to TA for evaluating
the conditions in the policy binary.

Add initial support of plumbing in the PMF driver for Smart PC to get
information from other subsystems in the kernel.

Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/Makefile |   2 +-
 drivers/platform/x86/amd/pmf/pmf.h    |  18 ++++
 drivers/platform/x86/amd/pmf/spc.c    | 120 ++++++++++++++++++++++++++
 drivers/platform/x86/amd/pmf/tee-if.c |   3 +
 4 files changed, 142 insertions(+), 1 deletion(-)
 create mode 100644 drivers/platform/x86/amd/pmf/spc.c

diff --git a/drivers/platform/x86/amd/pmf/Makefile b/drivers/platform/x86/amd/pmf/Makefile
index d2746ee7369f..6b26e48ce8ad 100644
--- a/drivers/platform/x86/amd/pmf/Makefile
+++ b/drivers/platform/x86/amd/pmf/Makefile
@@ -7,4 +7,4 @@
 obj-$(CONFIG_AMD_PMF) += amd-pmf.o
 amd-pmf-objs := core.o acpi.o sps.o \
 		auto-mode.o cnqf.o \
-		tee-if.o
+		tee-if.o spc.o
diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index 092be501d4d3..a4a73b845c09 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -150,6 +150,21 @@ struct smu_pmf_metrics {
 	u16 infra_gfx_maxfreq; /* in MHz */
 	u16 skin_temp; /* in centi-Celsius */
 	u16 device_state;
+	u16 curtemp; /* in centi-Celsius */
+	u16 filter_alpha_value;
+	u16 avg_gfx_clkfrequency;
+	u16 avg_fclk_frequency;
+	u16 avg_gfx_activity;
+	u16 avg_socclk_frequency;
+	u16 avg_vclk_frequency;
+	u16 avg_vcn_activity;
+	u16 avg_dram_reads;
+	u16 avg_dram_writes;
+	u16 avg_socket_power;
+	u16 avg_core_power[2];
+	u16 avg_core_c0residency[16];
+	u16 spare1;
+	u32 metrics_counter;
 } __packed;
 
 enum amd_stt_skin_temp {
@@ -601,4 +616,7 @@ extern const struct attribute_group cnqf_feature_attribute_group;
 int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev);
 void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev);
 int apmf_check_smart_pc(struct amd_pmf_dev *pmf_dev);
+
+/* Smart PC - TA interfaces */
+void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in);
 #endif /* PMF_H */
diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
new file mode 100644
index 000000000000..bd5061fdfdbe
--- /dev/null
+++ b/drivers/platform/x86/amd/pmf/spc.c
@@ -0,0 +1,120 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * AMD Platform Management Framework Driver - Smart PC Capabilities
+ *
+ * Copyright (c) 2023, Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Authors: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
+ *          Patil Rajesh Reddy <Patil.Reddy@amd.com>
+ */
+
+#include <acpi/button.h>
+#include <linux/power_supply.h>
+#include <linux/units.h>
+#include "pmf.h"
+
+static void amd_pmf_get_smu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
+{
+	u16 max, avg = 0;
+	int i;
+
+	memset(dev->buf, 0, sizeof(dev->m_table));
+	amd_pmf_send_cmd(dev, SET_TRANSFER_TABLE, 0, 7, NULL);
+	memcpy(&dev->m_table, dev->buf, sizeof(dev->m_table));
+
+	in->ev_info.socket_power = dev->m_table.apu_power + dev->m_table.dgpu_power;
+	in->ev_info.skin_temperature = dev->m_table.skin_temp;
+
+	/* Get the avg and max C0 residency of all the cores */
+	max = dev->m_table.avg_core_c0residency[0];
+	for (i = 0; i < ARRAY_SIZE(dev->m_table.avg_core_c0residency); i++) {
+		avg += dev->m_table.avg_core_c0residency[i];
+		if (dev->m_table.avg_core_c0residency[i] > max)
+			max = dev->m_table.avg_core_c0residency[i];
+	}
+
+	avg = DIV_ROUND_CLOSEST(avg, ARRAY_SIZE(dev->m_table.avg_core_c0residency));
+	in->ev_info.avg_c0residency = avg;
+	in->ev_info.max_c0residency = max;
+	in->ev_info.gfx_busy = dev->m_table.avg_gfx_activity;
+}
+
+static const char * const pmf_battery_supply_name[] = {
+	"BATT",
+	"BAT0",
+};
+
+static int amd_pmf_get_battery_prop(enum power_supply_property prop)
+{
+	union power_supply_propval value;
+	struct power_supply *psy;
+	int i, ret;
+
+	for (i = 0; i < ARRAY_SIZE(pmf_battery_supply_name); i++) {
+		psy = power_supply_get_by_name(pmf_battery_supply_name[i]);
+		if (!psy)
+			continue;
+
+		ret = power_supply_get_property(psy, prop, &value);
+		if (ret) {
+			power_supply_put(psy);
+			return ret;
+		}
+	}
+
+	return value.intval;
+}
+
+static int amd_pmf_get_battery_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
+{
+	int val;
+
+	val = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_PRESENT);
+	if (val >= 0 && val != 1)
+		return -ENODEV;
+
+	in->ev_info.bat_percentage = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_CAPACITY);
+	/* all values in mWh metrics */
+	in->ev_info.bat_design = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_ENERGY_FULL_DESIGN) /
+		MILLIWATT_PER_WATT;
+	in->ev_info.full_charge_capacity = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_ENERGY_FULL) /
+		MILLIWATT_PER_WATT;
+	in->ev_info.drain_rate = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_POWER_NOW) /
+		MILLIWATT_PER_WATT;
+
+	return 0;
+}
+
+static int amd_pmf_get_slider_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
+{
+	int val;
+
+	switch (dev->current_profile) {
+	case PLATFORM_PROFILE_PERFORMANCE:
+		val = TA_BEST_PERFORMANCE;
+		break;
+	case PLATFORM_PROFILE_BALANCED:
+		val = TA_BETTER_PERFORMANCE;
+		break;
+	case PLATFORM_PROFILE_LOW_POWER:
+		val = TA_BEST_BATTERY;
+		break;
+	default:
+		dev_err(dev->dev, "Unknown Platform Profile.\n");
+		return -EOPNOTSUPP;
+	}
+	in->ev_info.power_slider = val;
+
+	return 0;
+}
+
+void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
+{
+	/* TA side lid open is 1 and close is 0, hence the ! here */
+	in->ev_info.lid_state = !acpi_lid_open();
+	in->ev_info.power_source = amd_pmf_get_power_source();
+	amd_pmf_get_smu_info(dev, in);
+	amd_pmf_get_battery_info(dev, in);
+	amd_pmf_get_slider_info(dev, in);
+}
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index 468f3797a848..e3f17852d532 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -113,6 +113,7 @@ static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
 {
 	struct ta_pmf_shared_memory *ta_sm = NULL;
 	struct ta_pmf_enact_result *out = NULL;
+	struct ta_pmf_enact_table *in = NULL;
 	struct tee_param param[MAX_TEE_PARAM];
 	struct tee_ioctl_invoke_arg arg;
 	int ret = 0;
@@ -123,11 +124,13 @@ static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
 	memset(dev->shbuf, 0, dev->policy_sz);
 	ta_sm = dev->shbuf;
 	out = &ta_sm->pmf_output.policy_apply_table;
+	in = &ta_sm->pmf_input.enact_table;
 
 	memset(ta_sm, 0, sizeof(*ta_sm));
 	ta_sm->command_id = TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES;
 	ta_sm->if_version = PMF_TA_IF_VERSION_MAJOR;
 
+	amd_pmf_populate_ta_inputs(dev, in);
 	amd_pmf_prepare_args(dev, TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES, &arg, param);
 
 	ret = tee_client_invoke_func(dev->tee_ctx, &arg, param);
-- 
2.25.1

Re: [PATCH v5 06/17] platform/x86/amd/pmf: Add support to get inputs from other subsystems

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 1418 bytes --]

On Fri, 17 Nov 2023, Shyam Sundar S K wrote:

> PMF driver sends changing inputs from each subystem to TA for evaluating
> the conditions in the policy binary.
> 
> Add initial support of plumbing in the PMF driver for Smart PC to get
> information from other subsystems in the kernel.
> 
> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> ---

> +static int amd_pmf_get_battery_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
> +{
> +	int val;
> +
> +	val = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_PRESENT);
> +	if (val >= 0 && val != 1)
> +		return -ENODEV;

This no longer handles errors and was not what I suggested. With 
splitting the check into two I meant this:

	val = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_PRESENT);
	if (val < 0)
		return val;
	if (val != 1)
		return -NODEV;

After fixing this, please add:

Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>

> +	in->ev_info.bat_percentage = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_CAPACITY);
> +	/* all values in mWh metrics */
> +	in->ev_info.bat_design = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_ENERGY_FULL_DESIGN) /
> +		MILLIWATT_PER_WATT;
> +	in->ev_info.full_charge_capacity = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_ENERGY_FULL) /
> +		MILLIWATT_PER_WATT;
> +	in->ev_info.drain_rate = amd_pmf_get_battery_prop(POWER_SUPPLY_PROP_POWER_NOW) /
> +		MILLIWATT_PER_WATT;
> +
> +	return 0;
> +}

-- 
 i.

[PATCH v5 07/17] platform/x86/amd/pmf: Add support update p3t limit

[Shyam Sundar S K]

P3T (Peak Package Power Limit) is a metric within the SMU controller
that can influence the power limits. Add support from the driver
to update P3T limits accordingly.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/pmf.h    | 3 +++
 drivers/platform/x86/amd/pmf/tee-if.c | 8 ++++++++
 2 files changed, 11 insertions(+)

diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index a4a73b845c09..781989c7dddd 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -49,6 +49,7 @@
 #define GET_STT_MIN_LIMIT	0x1F
 #define GET_STT_LIMIT_APU	0x20
 #define GET_STT_LIMIT_HS2	0x21
+#define SET_P3T				0x23 /* P3T: Peak Package Power Limit */
 
 /* OS slider update notification */
 #define DC_BEST_PERF		0
@@ -72,6 +73,7 @@
 #define PMF_POLICY_STT_MIN					6
 #define PMF_POLICY_STT_SKINTEMP_APU				7
 #define PMF_POLICY_STT_SKINTEMP_HS2				8
+#define PMF_POLICY_P3T						38
 
 /* TA macros */
 #define PMF_TA_IF_VERSION_MAJOR				1
@@ -481,6 +483,7 @@ struct pmf_action_table {
 	u32 stt_minlimit;	/* in mW */
 	u32 stt_skintemp_apu;	/* in C */
 	u32 stt_skintemp_hs2;	/* in C */
+	u32 p3t_limit;		/* in mW */
 };
 
 /* Input conditions */
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index e3f17852d532..9861c2702251 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -105,6 +105,14 @@ static void amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_
 				dev->prev_data->stt_skintemp_hs2 = val;
 			}
 			break;
+
+		case PMF_POLICY_P3T:
+			if (dev->prev_data->p3t_limit != val) {
+				amd_pmf_send_cmd(dev, SET_P3T, false, val, NULL);
+				dev_dbg(dev->dev, "update P3T: %u\n", val);
+				dev->prev_data->p3t_limit = val;
+			}
+			break;
 		}
 	}
 }
-- 
2.25.1

[PATCH v5 08/17] platform/x86/amd/pmf: Add support to update system state

[Shyam Sundar S K]

PMF driver based on the output actions from the TA can request to update
the system states like entering s0i3, lock screen etc. by generating
an uevent. Based on the udev rules set in the userspace the event id
matching the uevent shall get updated accordingly using the systemctl.

Sample udev rules under Documentation/admin-guide/pmf.rst.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 Documentation/admin-guide/index.rst   |  1 +
 Documentation/admin-guide/pmf.rst     | 24 +++++++++++++++++++
 drivers/platform/x86/amd/pmf/pmf.h    |  9 +++++++
 drivers/platform/x86/amd/pmf/tee-if.c | 34 +++++++++++++++++++++++++++
 4 files changed, 68 insertions(+)
 create mode 100644 Documentation/admin-guide/pmf.rst

diff --git a/Documentation/admin-guide/index.rst b/Documentation/admin-guide/index.rst
index 43ea35613dfc..fb40a1f6f79e 100644
--- a/Documentation/admin-guide/index.rst
+++ b/Documentation/admin-guide/index.rst
@@ -119,6 +119,7 @@ configure specific aspects of kernel behavior to your liking.
    parport
    perf-security
    pm/index
+   pmf
    pnp
    rapidio
    ras
diff --git a/Documentation/admin-guide/pmf.rst b/Documentation/admin-guide/pmf.rst
new file mode 100644
index 000000000000..9ee729ffc19b
--- /dev/null
+++ b/Documentation/admin-guide/pmf.rst
@@ -0,0 +1,24 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+Set udev rules for PMF Smart PC Builder
+---------------------------------------
+
+AMD PMF(Platform Management Framework) Smart PC Solution builder has to set the system states
+like S0i3, Screen lock, hibernate etc, based on the output actions provided by the PMF
+TA (Trusted Application).
+
+In order for this to work the PMF driver generates a uevent for userspace to react to. Below are
+sample udev rules that can facilitate this experience when a machine has PMF Smart PC solution builder
+enabled.
+
+Please add the following line(s) to
+``/etc/udev/rules.d/99-local.rules``::
+
+        DRIVERS=="amd-pmf", ACTION=="change", ENV{EVENT_ID}=="0", RUN+="/usr/bin/systemctl suspend"
+        DRIVERS=="amd-pmf", ACTION=="change", ENV{EVENT_ID}=="1", RUN+="/usr/bin/systemctl hibernate"
+        DRIVERS=="amd-pmf", ACTION=="change", ENV{EVENT_ID}=="2", RUN+="/bin/loginctl lock-sessions"
+
+EVENT_ID values:
+0= Put the system to S0i3/S2Idle
+1= Put the system to hibernate
+2= Lock the screen
diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index 781989c7dddd..fb9ce2593236 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -73,6 +73,7 @@
 #define PMF_POLICY_STT_MIN					6
 #define PMF_POLICY_STT_SKINTEMP_APU				7
 #define PMF_POLICY_STT_SKINTEMP_HS2				8
+#define PMF_POLICY_SYSTEM_STATE					9
 #define PMF_POLICY_P3T						38
 
 /* TA macros */
@@ -445,6 +446,13 @@ enum smart_pc_status {
 };
 
 /* Smart PC - TA internals */
+enum system_state {
+	SYSTEM_STATE_S0i3,
+	SYSTEM_STATE_S4,
+	SYSTEM_STATE_SCREEN_LOCK,
+	SYSTEM_STATE_MAX,
+};
+
 enum ta_slider {
 	TA_BEST_BATTERY,
 	TA_BETTER_BATTERY,
@@ -476,6 +484,7 @@ enum ta_pmf_error_type {
 };
 
 struct pmf_action_table {
+	enum system_state system_state;
 	u32 spl;		/* in mW */
 	u32 sppt;		/* in mW */
 	u32 sppt_apuonly;	/* in mW */
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index 9861c2702251..4c2475a28d55 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -24,6 +24,20 @@ MODULE_PARM_DESC(pb_actions_ms, "Policy binary actions sampling frequency (defau
 static const uuid_t amd_pmf_ta_uuid = UUID_INIT(0x6fd93b77, 0x3fb8, 0x524d,
 						0xb1, 0x2d, 0xc5, 0x29, 0xb1, 0x3d, 0x85, 0x43);
 
+static const char *amd_pmf_uevent_as_str(unsigned int state)
+{
+	switch (state) {
+	case SYSTEM_STATE_S0i3:
+		return "S0i3";
+	case SYSTEM_STATE_S4:
+		return "S4";
+	case SYSTEM_STATE_SCREEN_LOCK:
+		return "SCREEN_LOCK";
+	default:
+		return "Unknown Smart PC event";
+	}
+}
+
 static void amd_pmf_prepare_args(struct amd_pmf_dev *dev, int cmd,
 				 struct tee_ioctl_invoke_arg *arg,
 				 struct tee_param *param)
@@ -42,6 +56,20 @@ static void amd_pmf_prepare_args(struct amd_pmf_dev *dev, int cmd,
 	param[0].u.memref.shm_offs = 0;
 }
 
+static int amd_pmf_update_uevents(struct amd_pmf_dev *dev, u16 event)
+{
+	char *envp[2] = {};
+
+	envp[0] = kasprintf(GFP_KERNEL, "EVENT_ID=%d", event);
+	if (!envp[0])
+		return -EINVAL;
+
+	kobject_uevent_env(&dev->dev->kobj, KOBJ_CHANGE, envp);
+
+	kfree(envp[0]);
+	return 0;
+}
+
 static void amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_result *out)
 {
 	u32 val;
@@ -113,6 +141,12 @@ static void amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_
 				dev->prev_data->p3t_limit = val;
 			}
 			break;
+
+		case PMF_POLICY_SYSTEM_STATE:
+			amd_pmf_update_uevents(dev, val);
+			dev_dbg(dev->dev, "update SYSTEM_STATE: %s\n",
+				amd_pmf_uevent_as_str(val));
+			break;
 		}
 	}
 }
-- 
2.25.1

[PATCH v5 09/17] platform/x86/amd/pmf: Make source_as_str() as non-static

[Shyam Sundar S K]

Add amd_pmf prefix to source_as_str() function, so that the function name
does not look generic. As this is a helper function make it as non-static
so that it can be reused across multiple PMF features.

Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/pmf.h | 1 +
 drivers/platform/x86/amd/pmf/sps.c | 5 +++--
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index fb9ce2593236..216a9f795436 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -600,6 +600,7 @@ int apmf_get_static_slider_granular(struct amd_pmf_dev *pdev,
 				    struct apmf_static_slider_granular_output *output);
 bool is_pprof_balanced(struct amd_pmf_dev *pmf);
 int amd_pmf_power_slider_update_event(struct amd_pmf_dev *dev);
+const char *amd_pmf_source_as_str(unsigned int state);
 
 
 int apmf_update_fan_idx(struct amd_pmf_dev *pdev, bool manual, u32 idx);
diff --git a/drivers/platform/x86/amd/pmf/sps.c b/drivers/platform/x86/amd/pmf/sps.c
index a70e67749be3..33e23e25c8b1 100644
--- a/drivers/platform/x86/amd/pmf/sps.c
+++ b/drivers/platform/x86/amd/pmf/sps.c
@@ -27,7 +27,7 @@ static const char *slider_as_str(unsigned int state)
 	}
 }
 
-static const char *source_as_str(unsigned int state)
+const char *amd_pmf_source_as_str(unsigned int state)
 {
 	switch (state) {
 	case POWER_SOURCE_AC:
@@ -47,7 +47,8 @@ static void amd_pmf_dump_sps_defaults(struct amd_pmf_static_slider_granular *dat
 
 	for (i = 0; i < POWER_SOURCE_MAX; i++) {
 		for (j = 0; j < POWER_MODE_MAX; j++) {
-			pr_debug("--- Source:%s Mode:%s ---\n", source_as_str(i), slider_as_str(j));
+			pr_debug("--- Source:%s Mode:%s ---\n", amd_pmf_source_as_str(i),
+				 slider_as_str(j));
 			pr_debug("SPL: %u mW\n", data->prop[i][j].spl);
 			pr_debug("SPPT: %u mW\n", data->prop[i][j].sppt);
 			pr_debug("SPPT_ApuOnly: %u mW\n", data->prop[i][j].sppt_apu_only);
-- 
2.25.1

Re: [PATCH v5 09/17] platform/x86/amd/pmf: Make source_as_str() as non-static

[Mario Limonciello]

On 11/17/2023 02:07, Shyam Sundar S K wrote:
> Add amd_pmf prefix to source_as_str() function, so that the function name
> does not look generic. As this is a helper function make it as non-static
> so that it can be reused across multiple PMF features.
> 
> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>

> ---
>   drivers/platform/x86/amd/pmf/pmf.h | 1 +
>   drivers/platform/x86/amd/pmf/sps.c | 5 +++--
>   2 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
> index fb9ce2593236..216a9f795436 100644
> --- a/drivers/platform/x86/amd/pmf/pmf.h
> +++ b/drivers/platform/x86/amd/pmf/pmf.h
> @@ -600,6 +600,7 @@ int apmf_get_static_slider_granular(struct amd_pmf_dev *pdev,
>   				    struct apmf_static_slider_granular_output *output);
>   bool is_pprof_balanced(struct amd_pmf_dev *pmf);
>   int amd_pmf_power_slider_update_event(struct amd_pmf_dev *dev);
> +const char *amd_pmf_source_as_str(unsigned int state);
>   
>   
>   int apmf_update_fan_idx(struct amd_pmf_dev *pdev, bool manual, u32 idx);
> diff --git a/drivers/platform/x86/amd/pmf/sps.c b/drivers/platform/x86/amd/pmf/sps.c
> index a70e67749be3..33e23e25c8b1 100644
> --- a/drivers/platform/x86/amd/pmf/sps.c
> +++ b/drivers/platform/x86/amd/pmf/sps.c
> @@ -27,7 +27,7 @@ static const char *slider_as_str(unsigned int state)
>   	}
>   }
>   
> -static const char *source_as_str(unsigned int state)
> +const char *amd_pmf_source_as_str(unsigned int state)
>   {
>   	switch (state) {
>   	case POWER_SOURCE_AC:
> @@ -47,7 +47,8 @@ static void amd_pmf_dump_sps_defaults(struct amd_pmf_static_slider_granular *dat
>   
>   	for (i = 0; i < POWER_SOURCE_MAX; i++) {
>   		for (j = 0; j < POWER_MODE_MAX; j++) {
> -			pr_debug("--- Source:%s Mode:%s ---\n", source_as_str(i), slider_as_str(j));
> +			pr_debug("--- Source:%s Mode:%s ---\n", amd_pmf_source_as_str(i),
> +				 slider_as_str(j));
>   			pr_debug("SPL: %u mW\n", data->prop[i][j].spl);
>   			pr_debug("SPPT: %u mW\n", data->prop[i][j].sppt);
>   			pr_debug("SPPT_ApuOnly: %u mW\n", data->prop[i][j].sppt_apu_only);

[PATCH v5 10/17] platform/x86/amd/pmf: Add facility to dump TA inputs

[Shyam Sundar S K]

PMF driver sends constant inputs to TA which its gets via the other
subsystems in the kernel. To debug certain TA issues knowing what inputs
being sent to TA becomes critical. Add debug facility to the driver which
can isolate Smart PC and TA related issues.

Also, make source_as_str() as non-static function as this helper is
required outside of sps.c file.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/pmf.h    |  3 +++
 drivers/platform/x86/amd/pmf/spc.c    | 37 +++++++++++++++++++++++++++
 drivers/platform/x86/amd/pmf/tee-if.c |  1 +
 3 files changed, 41 insertions(+)

diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index 216a9f795436..593930519039 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -602,6 +602,7 @@ bool is_pprof_balanced(struct amd_pmf_dev *pmf);
 int amd_pmf_power_slider_update_event(struct amd_pmf_dev *dev);
 const char *amd_pmf_source_as_str(unsigned int state);
 
+const char *amd_pmf_source_as_str(unsigned int state);
 
 int apmf_update_fan_idx(struct amd_pmf_dev *pdev, bool manual, u32 idx);
 int amd_pmf_set_sps_power_limits(struct amd_pmf_dev *pmf);
@@ -632,4 +633,6 @@ int apmf_check_smart_pc(struct amd_pmf_dev *pmf_dev);
 
 /* Smart PC - TA interfaces */
 void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in);
+void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in);
+
 #endif /* PMF_H */
diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
index bd5061fdfdbe..959146fd483f 100644
--- a/drivers/platform/x86/amd/pmf/spc.c
+++ b/drivers/platform/x86/amd/pmf/spc.c
@@ -14,6 +14,43 @@
 #include <linux/units.h>
 #include "pmf.h"
 
+#ifdef CONFIG_AMD_PMF_DEBUG
+static const char *ta_slider_as_str(unsigned int state)
+{
+	switch (state) {
+	case TA_BEST_PERFORMANCE:
+		return "PERFORMANCE";
+	case TA_BETTER_PERFORMANCE:
+		return "BALANCED";
+	case TA_BEST_BATTERY:
+		return "POWER_SAVER";
+	default:
+		return "Unknown TA Slider State";
+	}
+}
+
+void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
+{
+	dev_dbg(dev->dev, "==== TA inputs START ====\n");
+	dev_dbg(dev->dev, "Slider State: %s\n", ta_slider_as_str(in->ev_info.power_slider));
+	dev_dbg(dev->dev, "Power Source: %s\n", amd_pmf_source_as_str(in->ev_info.power_source));
+	dev_dbg(dev->dev, "Battery Percentage: %u\n", in->ev_info.bat_percentage);
+	dev_dbg(dev->dev, "Designed Battery Capacity: %u\n", in->ev_info.bat_design);
+	dev_dbg(dev->dev, "Fully Charged Capacity: %u\n", in->ev_info.full_charge_capacity);
+	dev_dbg(dev->dev, "Drain Rate: %d\n", in->ev_info.drain_rate);
+	dev_dbg(dev->dev, "Socket Power: %u\n", in->ev_info.socket_power);
+	dev_dbg(dev->dev, "Skin Temperature: %u\n", in->ev_info.skin_temperature);
+	dev_dbg(dev->dev, "Avg C0 Residency: %u\n", in->ev_info.avg_c0residency);
+	dev_dbg(dev->dev, "Max C0 Residency: %u\n", in->ev_info.max_c0residency);
+	dev_dbg(dev->dev, "GFX Busy: %u\n", in->ev_info.gfx_busy);
+	dev_dbg(dev->dev, "Connected Display Count: %u\n", in->ev_info.monitor_count);
+	dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ? "close" : "open");
+	dev_dbg(dev->dev, "==== TA inputs END ====\n");
+}
+#else
+void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in) {}
+#endif
+
 static void amd_pmf_get_smu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
 {
 	u16 max, avg = 0;
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index 4c2475a28d55..5f10e5c6335e 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -182,6 +182,7 @@ static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
 	}
 
 	if (ta_sm->pmf_result == TA_PMF_TYPE_SUCCESS && out->actions_count) {
+		amd_pmf_dump_ta_inputs(dev, in);
 		dev_dbg(dev->dev, "action count:%u result:%x\n", out->actions_count,
 			ta_sm->pmf_result);
 		amd_pmf_apply_policies(dev, out);
-- 
2.25.1

Re: [PATCH v5 10/17] platform/x86/amd/pmf: Add facility to dump TA inputs

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 620 bytes --]

On Fri, 17 Nov 2023, Shyam Sundar S K wrote:

> PMF driver sends constant inputs to TA which its gets via the other
> subsystems in the kernel. To debug certain TA issues knowing what inputs
> being sent to TA becomes critical. Add debug facility to the driver which
> can isolate Smart PC and TA related issues.
> 
> Also, make source_as_str() as non-static function as this helper is
> required outside of sps.c file.
> 
> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> ---

Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>

-- 
 i.

[PATCH v5 11/17] platform/x86/amd/pmf: Add capability to sideload of policy binary

[Shyam Sundar S K]

A policy binary is OS agnostic, and the same policies are expected to work
across the OSes.  At times it becomes difficult to debug when the policies
inside the policy binaries starts to misbehave. Add a way to sideload such
policies independently to debug them via a debugfs entry.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/pmf.h    |  1 +
 drivers/platform/x86/amd/pmf/tee-if.c | 54 +++++++++++++++++++++++++++
 2 files changed, 55 insertions(+)

diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index 593930519039..8712299ad52b 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -219,6 +219,7 @@ struct amd_pmf_dev {
 	bool cnqf_supported;
 	struct notifier_block pwr_src_notifier;
 	/* Smart PC solution builder */
+	struct dentry *esbin;
 	unsigned char *policy_buf;
 	u32 policy_sz;
 	struct tee_context *tee_ctx;
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index 5f10e5c6335e..f73663c629fe 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -8,6 +8,7 @@
  * Author: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
  */
 
+#include <linux/debugfs.h>
 #include <linux/tee_drv.h>
 #include <linux/uuid.h>
 #include "pmf.h"
@@ -16,9 +17,14 @@
 
 /* Policy binary actions sampling frequency (in ms) */
 static int pb_actions_ms = MSEC_PER_SEC;
+/* Sideload policy binaries to debug policy failures */
+static bool pb_side_load;
+
 #ifdef CONFIG_AMD_PMF_DEBUG
 module_param(pb_actions_ms, int, 0644);
 MODULE_PARM_DESC(pb_actions_ms, "Policy binary actions sampling frequency (default = 1000ms)");
+module_param(pb_side_load, bool, 0444);
+MODULE_PARM_DESC(pb_side_load, "Sideload policy binaries debug policy failures");
 #endif
 
 static const uuid_t amd_pmf_ta_uuid = UUID_INIT(0x6fd93b77, 0x3fb8, 0x524d,
@@ -269,6 +275,48 @@ static int amd_pmf_start_policy_engine(struct amd_pmf_dev *dev)
 	return 0;
 }
 
+#ifdef CONFIG_AMD_PMF_DEBUG
+static ssize_t amd_pmf_get_pb_data(struct file *filp, const char __user *buf,
+				   size_t length, loff_t *pos)
+{
+	struct amd_pmf_dev *dev = filp->private_data;
+	int ret;
+
+	/* Policy binary size cannot exceed POLICY_BUF_MAX_SZ */
+	if (length > POLICY_BUF_MAX_SZ || length == 0)
+		return -EINVAL;
+
+	dev->policy_sz = length;
+	if (copy_from_user(dev->policy_buf, buf, dev->policy_sz))
+		return -EFAULT;
+
+	ret = amd_pmf_start_policy_engine(dev);
+	if (ret)
+		return -EINVAL;
+
+	return length;
+}
+
+static const struct file_operations pb_fops = {
+	.write = amd_pmf_get_pb_data,
+	.open = simple_open,
+};
+
+static void amd_pmf_open_pb(struct amd_pmf_dev *dev, struct dentry *debugfs_root)
+{
+	dev->esbin = debugfs_create_dir("pb", debugfs_root);
+	debugfs_create_file("update_policy", 0644, dev->esbin, dev, &pb_fops);
+}
+
+static void amd_pmf_remove_pb(struct amd_pmf_dev *dev)
+{
+	debugfs_remove_recursive(dev->esbin);
+}
+#else
+static void amd_pmf_open_pb(struct amd_pmf_dev *dev, struct dentry *debugfs_root) {}
+static void amd_pmf_remove_pb(struct amd_pmf_dev *dev) {}
+#endif
+
 static int amd_pmf_get_bios_buffer(struct amd_pmf_dev *dev)
 {
 	dev->policy_buf = kzalloc(dev->policy_sz, GFP_KERNEL);
@@ -281,6 +329,9 @@ static int amd_pmf_get_bios_buffer(struct amd_pmf_dev *dev)
 
 	memcpy(dev->policy_buf, dev->policy_base, dev->policy_sz);
 
+	if (pb_side_load)
+		amd_pmf_open_pb(dev, dev->dbgfs_dir);
+
 	return amd_pmf_start_policy_engine(dev);
 }
 
@@ -382,6 +433,9 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
 
 void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev)
 {
+	if (pb_side_load)
+		amd_pmf_remove_pb(dev);
+
 	kfree(dev->prev_data);
 	kfree(dev->policy_buf);
 	cancel_delayed_work_sync(&dev->pb_work);
-- 
2.25.1

Re: [PATCH v5 11/17] platform/x86/amd/pmf: Add capability to sideload of policy binary

[Ilpo Järvinen]

On Fri, 17 Nov 2023, Shyam Sundar S K wrote:

> A policy binary is OS agnostic, and the same policies are expected to work
> across the OSes.  At times it becomes difficult to debug when the policies
> inside the policy binaries starts to misbehave. Add a way to sideload such
> policies independently to debug them via a debugfs entry.
> 
> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> ---

> diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
> index 5f10e5c6335e..f73663c629fe 100644
> --- a/drivers/platform/x86/amd/pmf/tee-if.c
> +++ b/drivers/platform/x86/amd/pmf/tee-if.c

> +#ifdef CONFIG_AMD_PMF_DEBUG
> +static ssize_t amd_pmf_get_pb_data(struct file *filp, const char __user *buf,
> +				   size_t length, loff_t *pos)
> +{
> +	struct amd_pmf_dev *dev = filp->private_data;
> +	int ret;
> +
> +	/* Policy binary size cannot exceed POLICY_BUF_MAX_SZ */
> +	if (length > POLICY_BUF_MAX_SZ || length == 0)
> +		return -EINVAL;
> +
> +	dev->policy_sz = length;
> +	if (copy_from_user(dev->policy_buf, buf, dev->policy_sz))
> +		return -EFAULT;
> +
> +	ret = amd_pmf_start_policy_engine(dev);

Is this call safe against concurrent invocations from two racing writes?

Other than that, this change looked fine.

> +	if (ret)
> +		return -EINVAL;
> +
> +	return length;
> +}


-- 
 i.

Re: [PATCH v5 11/17] platform/x86/amd/pmf: Add capability to sideload of policy binary

[Shyam Sundar S K]

Hi Ilpo,

On 11/21/2023 5:47 PM, Ilpo Järvinen wrote:
> On Fri, 17 Nov 2023, Shyam Sundar S K wrote:
> 
>> A policy binary is OS agnostic, and the same policies are expected to work
>> across the OSes.  At times it becomes difficult to debug when the policies
>> inside the policy binaries starts to misbehave. Add a way to sideload such
>> policies independently to debug them via a debugfs entry.
>>
>> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
>> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
>> ---
> 
>> diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
>> index 5f10e5c6335e..f73663c629fe 100644
>> --- a/drivers/platform/x86/amd/pmf/tee-if.c
>> +++ b/drivers/platform/x86/amd/pmf/tee-if.c
> 
>> +#ifdef CONFIG_AMD_PMF_DEBUG
>> +static ssize_t amd_pmf_get_pb_data(struct file *filp, const char __user *buf,
>> +				   size_t length, loff_t *pos)
>> +{
>> +	struct amd_pmf_dev *dev = filp->private_data;
>> +	int ret;
>> +
>> +	/* Policy binary size cannot exceed POLICY_BUF_MAX_SZ */
>> +	if (length > POLICY_BUF_MAX_SZ || length == 0)
>> +		return -EINVAL;
>> +
>> +	dev->policy_sz = length;
>> +	if (copy_from_user(dev->policy_buf, buf, dev->policy_sz))
>> +		return -EFAULT;
>> +
>> +	ret = amd_pmf_start_policy_engine(dev);
> 
> Is this call safe against concurrent invocations from two racing writes?
> 
> Other than that, this change looked fine.


This path gets enabled only when CONFIG_AMD_PMF_DEBUG option is
enabled. Also when enabled, did not observe anything really unusual
(like races). So I have retained the same code in v6.

Kindly take a look.

Thanks,
Shyam

> 
>> +	if (ret)
>> +		return -EINVAL;
>> +
>> +	return length;
>> +}
> 
> 

[PATCH v5 12/17] platform/x86/amd/pmf: dump policy binary data

[Shyam Sundar S K]

Sometimes policy binary retrieved from the BIOS maybe incorrect that can
end up in failing to enable the Smart PC solution feature.

Use print_hex_dump_debug() to dump the policy binary in hex, so that we
debug the issues related to the binary even before sending that to TA.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/tee-if.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index f73663c629fe..81b1bd356e83 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -276,6 +276,12 @@ static int amd_pmf_start_policy_engine(struct amd_pmf_dev *dev)
 }
 
 #ifdef CONFIG_AMD_PMF_DEBUG
+static void amd_pmf_hex_dump_pb(struct amd_pmf_dev *dev)
+{
+	print_hex_dump_debug("(pb):  ", DUMP_PREFIX_OFFSET, 16, 1, dev->policy_buf,
+			     dev->policy_sz, false);
+}
+
 static ssize_t amd_pmf_get_pb_data(struct file *filp, const char __user *buf,
 				   size_t length, loff_t *pos)
 {
@@ -290,6 +296,7 @@ static ssize_t amd_pmf_get_pb_data(struct file *filp, const char __user *buf,
 	if (copy_from_user(dev->policy_buf, buf, dev->policy_sz))
 		return -EFAULT;
 
+	amd_pmf_hex_dump_pb(dev);
 	ret = amd_pmf_start_policy_engine(dev);
 	if (ret)
 		return -EINVAL;
@@ -315,6 +322,7 @@ static void amd_pmf_remove_pb(struct amd_pmf_dev *dev)
 #else
 static void amd_pmf_open_pb(struct amd_pmf_dev *dev, struct dentry *debugfs_root) {}
 static void amd_pmf_remove_pb(struct amd_pmf_dev *dev) {}
+static void amd_pmf_hex_dump_pb(struct amd_pmf_dev *dev) {}
 #endif
 
 static int amd_pmf_get_bios_buffer(struct amd_pmf_dev *dev)
@@ -329,6 +337,7 @@ static int amd_pmf_get_bios_buffer(struct amd_pmf_dev *dev)
 
 	memcpy(dev->policy_buf, dev->policy_base, dev->policy_sz);
 
+	amd_pmf_hex_dump_pb(dev);
 	if (pb_side_load)
 		amd_pmf_open_pb(dev, dev->dbgfs_dir);
 
-- 
2.25.1

[PATCH v5 13/17] platform/x86/amd/pmf: Add PMF-AMDGPU get interface

[Shyam Sundar S K]

In order to provide GPU inputs to TA for the Smart PC solution to work, we
need to have interface between the PMF driver and the AMDGPU driver.

Add the initial code path for get interface from AMDGPU.

Co-developed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/core.c   |  3 +-
 drivers/platform/x86/amd/pmf/pmf.h    | 18 +++++++
 drivers/platform/x86/amd/pmf/spc.c    | 41 ++++++++++++++
 drivers/platform/x86/amd/pmf/tee-if.c | 77 +++++++++++++++++++++++++++
 4 files changed, 138 insertions(+), 1 deletion(-)

diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c
index 4b8156033fa6..9b14a997cd48 100644
--- a/drivers/platform/x86/amd/pmf/core.c
+++ b/drivers/platform/x86/amd/pmf/core.c
@@ -411,6 +411,7 @@ static int amd_pmf_probe(struct platform_device *pdev)
 	}
 
 	dev->cpu_id = rdev->device;
+	dev->root = rdev;
 
 	err = amd_smn_read(0, AMD_PMF_BASE_ADDR_LO, &val);
 	if (err) {
@@ -482,4 +483,4 @@ module_platform_driver(amd_pmf_driver);
 
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("AMD Platform Management Framework Driver");
-MODULE_SOFTDEP("pre: amdtee");
+MODULE_SOFTDEP("pre: amdtee amdgpu");
diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index 8712299ad52b..525308519fa3 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -11,7 +11,11 @@
 #ifndef PMF_H
 #define PMF_H
 
+#include <acpi/video.h>
+#include <drm/drm_connector.h>
 #include <linux/acpi.h>
+#include <linux/backlight.h>
+#include <linux/pci.h>
 #include <linux/platform_profile.h>
 
 #define POLICY_BUF_MAX_SZ		0x4b000
@@ -83,6 +87,8 @@
 #define TA_OUTPUT_RESERVED_MEM				906
 #define MAX_OPERATION_PARAMS					4
 
+#define MAX_SUPPORTED_DISPLAY		4
+
 /* AMD PMF BIOS interfaces */
 struct apmf_verify_interface {
 	u16 size;
@@ -194,6 +200,15 @@ enum power_modes {
 	POWER_MODE_MAX,
 };
 
+struct amd_gpu_pmf_data {
+	struct pci_dev *gpu_dev;
+	struct backlight_device *raw_bd;
+	struct thermal_cooling_device *cooling_dev;
+	enum drm_connector_status con_status[MAX_SUPPORTED_DISPLAY];
+	int display_count;
+	int connector_type[MAX_SUPPORTED_DISPLAY];
+};
+
 struct amd_pmf_dev {
 	void __iomem *regbase;
 	void __iomem *smu_virt_addr;
@@ -228,9 +243,12 @@ struct amd_pmf_dev {
 	void *shbuf;
 	struct delayed_work pb_work;
 	struct pmf_action_table *prev_data;
+	struct amd_gpu_pmf_data gfx_data;
 	u64 policy_addr;
 	void *policy_base;
 	bool smart_pc_enabled;
+	struct pci_dev *root;
+	struct drm_device *drm_dev;
 };
 
 struct apmf_sps_prop_granular {
diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
index 959146fd483f..47ec563088b8 100644
--- a/drivers/platform/x86/amd/pmf/spc.c
+++ b/drivers/platform/x86/amd/pmf/spc.c
@@ -44,6 +44,10 @@ void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *
 	dev_dbg(dev->dev, "Max C0 Residency: %u\n", in->ev_info.max_c0residency);
 	dev_dbg(dev->dev, "GFX Busy: %u\n", in->ev_info.gfx_busy);
 	dev_dbg(dev->dev, "Connected Display Count: %u\n", in->ev_info.monitor_count);
+	dev_dbg(dev->dev, "Primary Display Type: %s\n",
+		drm_get_connector_type_name(in->ev_info.display_type));
+	dev_dbg(dev->dev, "Primary Display State: %s\n", in->ev_info.display_state ?
+			"Connected" : "disconnected/unknown");
 	dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ? "close" : "open");
 	dev_dbg(dev->dev, "==== TA inputs END ====\n");
 }
@@ -146,6 +150,41 @@ static int amd_pmf_get_slider_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_
 	return 0;
 }
 
+static int amd_pmf_get_gfx_data(struct amd_pmf_dev *pmf)
+{
+	struct drm_connector_list_iter iter;
+	struct drm_connector *connector;
+	int i = 0;
+
+	/* Reset the count to zero */
+	pmf->gfx_data.display_count = 0;
+
+	drm_connector_list_iter_begin(pmf->drm_dev, &iter);
+	drm_for_each_connector_iter(connector, &iter) {
+		if (connector->status == connector_status_connected)
+			pmf->gfx_data.display_count++;
+		if (connector->status != pmf->gfx_data.con_status[i])
+			pmf->gfx_data.con_status[i] = connector->status;
+		if (connector->connector_type != pmf->gfx_data.connector_type[i])
+			pmf->gfx_data.connector_type[i] = connector->connector_type;
+
+		i++;
+		if (i >= MAX_SUPPORTED_DISPLAY)
+			break;
+	}
+	drm_connector_list_iter_end(&iter);
+
+	return 0;
+}
+
+static void amd_pmf_get_gpu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
+{
+	amd_pmf_get_gfx_data(dev);
+	in->ev_info.monitor_count = dev->gfx_data.display_count;
+	in->ev_info.display_type = dev->gfx_data.connector_type[0];
+	in->ev_info.display_state = dev->gfx_data.con_status[0];
+}
+
 void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
 {
 	/* TA side lid open is 1 and close is 0, hence the ! here */
@@ -154,4 +193,6 @@ void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_tab
 	amd_pmf_get_smu_info(dev, in);
 	amd_pmf_get_battery_info(dev, in);
 	amd_pmf_get_slider_info(dev, in);
+	if (dev->drm_dev)
+		amd_pmf_get_gpu_info(dev, in);
 }
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index 81b1bd356e83..82ee2b1c627f 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -10,6 +10,7 @@
 
 #include <linux/debugfs.h>
 #include <linux/tee_drv.h>
+#include <linux/thermal.h>
 #include <linux/uuid.h>
 #include "pmf.h"
 
@@ -422,6 +423,60 @@ static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
 	tee_client_close_context(dev->tee_ctx);
 }
 
+static int amd_pmf_gpu_get_cur_state(struct thermal_cooling_device *cooling_dev,
+				     unsigned long *state)
+{
+	struct backlight_device *bd;
+
+	if (acpi_video_get_backlight_type() != acpi_backlight_native)
+		return -ENODEV;
+
+	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
+	if (!bd)
+		return -ENODEV;
+
+	*state = backlight_get_brightness(bd);
+
+	return 0;
+}
+
+static int amd_pmf_gpu_get_max_state(struct thermal_cooling_device *cooling_dev,
+				     unsigned long *state)
+{
+	struct backlight_device *bd;
+
+	if (acpi_video_get_backlight_type() != acpi_backlight_native)
+		return -ENODEV;
+
+	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
+	if (!bd)
+		return -ENODEV;
+
+	if (backlight_is_blank(bd))
+		*state = 0;
+	else
+		*state = bd->props.max_brightness;
+
+	return 0;
+}
+
+static const struct thermal_cooling_device_ops bd_cooling_ops = {
+	.get_max_state = amd_pmf_gpu_get_max_state,
+	.get_cur_state = amd_pmf_gpu_get_cur_state,
+};
+
+static int amd_pmf_get_gpu_handle(struct pci_dev *pdev, void *data)
+{
+	struct amd_pmf_dev *dev = data;
+
+	if (pdev->vendor == PCI_VENDOR_ID_ATI && pdev->devfn == 0) {
+		dev->gfx_data.gpu_dev = pdev;
+		return 1; /* Stop walking */
+	}
+
+	return 0; /* Continue walking */
+}
+
 int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
 {
 	int ret;
@@ -433,10 +488,30 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
 	INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
 	amd_pmf_set_dram_addr(dev);
 	amd_pmf_get_bios_buffer(dev);
+
 	dev->prev_data = kzalloc(sizeof(*dev->prev_data), GFP_KERNEL);
 	if (!dev->prev_data)
 		return -ENOMEM;
 
+	pci_walk_bus(dev->root->bus, amd_pmf_get_gpu_handle, dev);
+	if (dev->gfx_data.gpu_dev) {
+		dev->drm_dev = pci_get_drvdata(dev->gfx_data.gpu_dev);
+		if (!dev->drm_dev)
+			return -EINVAL;
+
+		if (acpi_video_get_backlight_type() != acpi_backlight_native)
+			return -ENODEV;
+
+		dev->gfx_data.raw_bd = backlight_device_get_by_type(BACKLIGHT_RAW);
+		if (!dev->gfx_data.raw_bd)
+			return -ENODEV;
+
+		dev->gfx_data.cooling_dev = thermal_cooling_device_register("pmf_gpu_bd",
+									    NULL, &bd_cooling_ops);
+		if (IS_ERR(dev->gfx_data.cooling_dev))
+			return -ENODEV;
+	}
+
 	return dev->smart_pc_enabled;
 }
 
@@ -448,5 +523,7 @@ void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev)
 	kfree(dev->prev_data);
 	kfree(dev->policy_buf);
 	cancel_delayed_work_sync(&dev->pb_work);
+	if (dev->gfx_data.cooling_dev)
+		thermal_cooling_device_unregister(dev->gfx_data.cooling_dev);
 	amd_pmf_tee_deinit(dev);
 }
-- 
2.25.1

Re: [PATCH v5 13/17] platform/x86/amd/pmf: Add PMF-AMDGPU get interface

[Hans de Goede]

Hi Shyam,

On 11/17/23 09:07, Shyam Sundar S K wrote:
> In order to provide GPU inputs to TA for the Smart PC solution to work, we
> need to have interface between the PMF driver and the AMDGPU driver.
> 
> Add the initial code path for get interface from AMDGPU.
> 
> Co-developed-by: Mario Limonciello <mario.limonciello@amd.com>
> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> ---
>  drivers/platform/x86/amd/pmf/core.c   |  3 +-
>  drivers/platform/x86/amd/pmf/pmf.h    | 18 +++++++
>  drivers/platform/x86/amd/pmf/spc.c    | 41 ++++++++++++++
>  drivers/platform/x86/amd/pmf/tee-if.c | 77 +++++++++++++++++++++++++++
>  4 files changed, 138 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c
> index 4b8156033fa6..9b14a997cd48 100644
> --- a/drivers/platform/x86/amd/pmf/core.c
> +++ b/drivers/platform/x86/amd/pmf/core.c
> @@ -411,6 +411,7 @@ static int amd_pmf_probe(struct platform_device *pdev)
>  	}
>  
>  	dev->cpu_id = rdev->device;
> +	dev->root = rdev;
>  
>  	err = amd_smn_read(0, AMD_PMF_BASE_ADDR_LO, &val);
>  	if (err) {
> @@ -482,4 +483,4 @@ module_platform_driver(amd_pmf_driver);
>  
>  MODULE_LICENSE("GPL");
>  MODULE_DESCRIPTION("AMD Platform Management Framework Driver");
> -MODULE_SOFTDEP("pre: amdtee");
> +MODULE_SOFTDEP("pre: amdtee amdgpu");
> diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
> index 8712299ad52b..525308519fa3 100644
> --- a/drivers/platform/x86/amd/pmf/pmf.h
> +++ b/drivers/platform/x86/amd/pmf/pmf.h
> @@ -11,7 +11,11 @@
>  #ifndef PMF_H
>  #define PMF_H
>  
> +#include <acpi/video.h>
> +#include <drm/drm_connector.h>
>  #include <linux/acpi.h>
> +#include <linux/backlight.h>
> +#include <linux/pci.h>
>  #include <linux/platform_profile.h>
>  
>  #define POLICY_BUF_MAX_SZ		0x4b000
> @@ -83,6 +87,8 @@
>  #define TA_OUTPUT_RESERVED_MEM				906
>  #define MAX_OPERATION_PARAMS					4
>  
> +#define MAX_SUPPORTED_DISPLAY		4
> +
>  /* AMD PMF BIOS interfaces */
>  struct apmf_verify_interface {
>  	u16 size;
> @@ -194,6 +200,15 @@ enum power_modes {
>  	POWER_MODE_MAX,
>  };
>  
> +struct amd_gpu_pmf_data {
> +	struct pci_dev *gpu_dev;
> +	struct backlight_device *raw_bd;
> +	struct thermal_cooling_device *cooling_dev;
> +	enum drm_connector_status con_status[MAX_SUPPORTED_DISPLAY];
> +	int display_count;
> +	int connector_type[MAX_SUPPORTED_DISPLAY];
> +};
> +
>  struct amd_pmf_dev {
>  	void __iomem *regbase;
>  	void __iomem *smu_virt_addr;
> @@ -228,9 +243,12 @@ struct amd_pmf_dev {
>  	void *shbuf;
>  	struct delayed_work pb_work;
>  	struct pmf_action_table *prev_data;
> +	struct amd_gpu_pmf_data gfx_data;
>  	u64 policy_addr;
>  	void *policy_base;
>  	bool smart_pc_enabled;
> +	struct pci_dev *root;
> +	struct drm_device *drm_dev;
>  };
>  
>  struct apmf_sps_prop_granular {
> diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
> index 959146fd483f..47ec563088b8 100644
> --- a/drivers/platform/x86/amd/pmf/spc.c
> +++ b/drivers/platform/x86/amd/pmf/spc.c
> @@ -44,6 +44,10 @@ void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *
>  	dev_dbg(dev->dev, "Max C0 Residency: %u\n", in->ev_info.max_c0residency);
>  	dev_dbg(dev->dev, "GFX Busy: %u\n", in->ev_info.gfx_busy);
>  	dev_dbg(dev->dev, "Connected Display Count: %u\n", in->ev_info.monitor_count);
> +	dev_dbg(dev->dev, "Primary Display Type: %s\n",
> +		drm_get_connector_type_name(in->ev_info.display_type));
> +	dev_dbg(dev->dev, "Primary Display State: %s\n", in->ev_info.display_state ?
> +			"Connected" : "disconnected/unknown");
>  	dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ? "close" : "open");
>  	dev_dbg(dev->dev, "==== TA inputs END ====\n");
>  }
> @@ -146,6 +150,41 @@ static int amd_pmf_get_slider_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_
>  	return 0;
>  }
>  
> +static int amd_pmf_get_gfx_data(struct amd_pmf_dev *pmf)
> +{
> +	struct drm_connector_list_iter iter;
> +	struct drm_connector *connector;
> +	int i = 0;
> +
> +	/* Reset the count to zero */
> +	pmf->gfx_data.display_count = 0;
> +
> +	drm_connector_list_iter_begin(pmf->drm_dev, &iter);
> +	drm_for_each_connector_iter(connector, &iter) {
> +		if (connector->status == connector_status_connected)
> +			pmf->gfx_data.display_count++;
> +		if (connector->status != pmf->gfx_data.con_status[i])
> +			pmf->gfx_data.con_status[i] = connector->status;
> +		if (connector->connector_type != pmf->gfx_data.connector_type[i])
> +			pmf->gfx_data.connector_type[i] = connector->connector_type;
> +
> +		i++;
> +		if (i >= MAX_SUPPORTED_DISPLAY)
> +			break;
> +	}
> +	drm_connector_list_iter_end(&iter);
> +
> +	return 0;
> +}
> +
> +static void amd_pmf_get_gpu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
> +{
> +	amd_pmf_get_gfx_data(dev);
> +	in->ev_info.monitor_count = dev->gfx_data.display_count;
> +	in->ev_info.display_type = dev->gfx_data.connector_type[0];
> +	in->ev_info.display_state = dev->gfx_data.con_status[0];
> +}
> +
>  void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
>  {
>  	/* TA side lid open is 1 and close is 0, hence the ! here */
> @@ -154,4 +193,6 @@ void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_tab
>  	amd_pmf_get_smu_info(dev, in);
>  	amd_pmf_get_battery_info(dev, in);
>  	amd_pmf_get_slider_info(dev, in);
> +	if (dev->drm_dev)
> +		amd_pmf_get_gpu_info(dev, in);
>  }
> diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
> index 81b1bd356e83..82ee2b1c627f 100644
> --- a/drivers/platform/x86/amd/pmf/tee-if.c
> +++ b/drivers/platform/x86/amd/pmf/tee-if.c
> @@ -10,6 +10,7 @@
>  
>  #include <linux/debugfs.h>
>  #include <linux/tee_drv.h>
> +#include <linux/thermal.h>
>  #include <linux/uuid.h>
>  #include "pmf.h"
>  
> @@ -422,6 +423,60 @@ static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
>  	tee_client_close_context(dev->tee_ctx);
>  }
>  
> +static int amd_pmf_gpu_get_cur_state(struct thermal_cooling_device *cooling_dev,
> +				     unsigned long *state)
> +{
> +	struct backlight_device *bd;
> +
> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
> +		return -ENODEV;
> +
> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
> +	if (!bd)
> +		return -ENODEV;
> +
> +	*state = backlight_get_brightness(bd);
> +
> +	return 0;
> +}
> +
> +static int amd_pmf_gpu_get_max_state(struct thermal_cooling_device *cooling_dev,
> +				     unsigned long *state)
> +{
> +	struct backlight_device *bd;
> +
> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
> +		return -ENODEV;
> +
> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
> +	if (!bd)
> +		return -ENODEV;
> +
> +	if (backlight_is_blank(bd))
> +		*state = 0;
> +	else
> +		*state = bd->props.max_brightness;
> +
> +	return 0;
> +}
> +
> +static const struct thermal_cooling_device_ops bd_cooling_ops = {
> +	.get_max_state = amd_pmf_gpu_get_max_state,
> +	.get_cur_state = amd_pmf_gpu_get_cur_state,
> +};

This is still the wrong thing to do. The new PMF code MUST only
be a *consumer* of the thermal_cooling_device API.

The thermal-cooling device for the backlight itself MUST be
registered by the amdgpu driver.

I believe that the correct fix for this is to add a new flag to
the backlight_properties struct;
And then make backlight_device_register() register
a thermal_cooling_device for the backlight when this new flag is set.

This way we get a nice generic way to use backlight class devices
as thermal cooling devices and you can make the amdgpu driver
register the thermal cooling device by simply adding the new flag
to the backlight_properties struct used in the amdgpu driver.

> +static int amd_pmf_get_gpu_handle(struct pci_dev *pdev, void *data)
> +{
> +	struct amd_pmf_dev *dev = data;
> +
> +	if (pdev->vendor == PCI_VENDOR_ID_ATI && pdev->devfn == 0) {
> +		dev->gfx_data.gpu_dev = pdev;
> +		return 1; /* Stop walking */
> +	}
> +
> +	return 0; /* Continue walking */
> +}
> +
>  int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>  {
>  	int ret;
> @@ -433,10 +488,30 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>  	INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
>  	amd_pmf_set_dram_addr(dev);
>  	amd_pmf_get_bios_buffer(dev);
> +
>  	dev->prev_data = kzalloc(sizeof(*dev->prev_data), GFP_KERNEL);
>  	if (!dev->prev_data)
>  		return -ENOMEM;
>  
> +	pci_walk_bus(dev->root->bus, amd_pmf_get_gpu_handle, dev);
> +	if (dev->gfx_data.gpu_dev) {
> +		dev->drm_dev = pci_get_drvdata(dev->gfx_data.gpu_dev);
> +		if (!dev->drm_dev)
> +			return -EINVAL;

You cannot just call pci_get_drvdata() on a device for which you
are not the driver. You have no idea of the lifetime of this device,
the driver may unbind and release the memory pci_get_drvdata()
points to, leaving this code with a dangling pointer which will
crash the kernel the first time you try to use it.

Also since you are not the owner you MUST not assume any specific
type for this memory, you cannot be sure this actually is of
the type drm_device. Basically you MUST not touch another
driver's drvdata *at all*.

The proper way to fix this would be to add some API to the DRM
subsystem to query the information you are looking for form
the DRM subsystem.

Poking directly inside other driver's internals is NOT acceptable,
NACK for this patch.

Regards,

Hans


> +
> +		if (acpi_video_get_backlight_type() != acpi_backlight_native)
> +			return -ENODEV;
> +
> +		dev->gfx_data.raw_bd = backlight_device_get_by_type(BACKLIGHT_RAW);
> +		if (!dev->gfx_data.raw_bd)
> +			return -ENODEV;
> +
> +		dev->gfx_data.cooling_dev = thermal_cooling_device_register("pmf_gpu_bd",
> +									    NULL, &bd_cooling_ops);
> +		if (IS_ERR(dev->gfx_data.cooling_dev))
> +			return -ENODEV;
> +	}
> +
>  	return dev->smart_pc_enabled;
>  }
>  
> @@ -448,5 +523,7 @@ void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev)
>  	kfree(dev->prev_data);
>  	kfree(dev->policy_buf);
>  	cancel_delayed_work_sync(&dev->pb_work);
> +	if (dev->gfx_data.cooling_dev)
> +		thermal_cooling_device_unregister(dev->gfx_data.cooling_dev);
>  	amd_pmf_tee_deinit(dev);
>  }

Re: [PATCH v5 13/17] platform/x86/amd/pmf: Add PMF-AMDGPU get interface

[Shyam Sundar S K]

Adding AMDGPU folks (Alex and Christian) - I had to drop them as the
changes from amdgpu were moved to PMF driver.

Hi Hans,


On 11/17/2023 4:18 PM, Hans de Goede wrote:
> Hi Shyam,
> 
> On 11/17/23 09:07, Shyam Sundar S K wrote:
>> In order to provide GPU inputs to TA for the Smart PC solution to work, we
>> need to have interface between the PMF driver and the AMDGPU driver.
>>
>> Add the initial code path for get interface from AMDGPU.
>>
>> Co-developed-by: Mario Limonciello <mario.limonciello@amd.com>
>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
>> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
>> ---
>>  drivers/platform/x86/amd/pmf/core.c   |  3 +-
>>  drivers/platform/x86/amd/pmf/pmf.h    | 18 +++++++
>>  drivers/platform/x86/amd/pmf/spc.c    | 41 ++++++++++++++
>>  drivers/platform/x86/amd/pmf/tee-if.c | 77 +++++++++++++++++++++++++++
>>  4 files changed, 138 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c
>> index 4b8156033fa6..9b14a997cd48 100644
>> --- a/drivers/platform/x86/amd/pmf/core.c
>> +++ b/drivers/platform/x86/amd/pmf/core.c
>> @@ -411,6 +411,7 @@ static int amd_pmf_probe(struct platform_device *pdev)
>>  	}
>>  
>>  	dev->cpu_id = rdev->device;
>> +	dev->root = rdev;
>>  
>>  	err = amd_smn_read(0, AMD_PMF_BASE_ADDR_LO, &val);
>>  	if (err) {
>> @@ -482,4 +483,4 @@ module_platform_driver(amd_pmf_driver);
>>  
>>  MODULE_LICENSE("GPL");
>>  MODULE_DESCRIPTION("AMD Platform Management Framework Driver");
>> -MODULE_SOFTDEP("pre: amdtee");
>> +MODULE_SOFTDEP("pre: amdtee amdgpu");
>> diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
>> index 8712299ad52b..525308519fa3 100644
>> --- a/drivers/platform/x86/amd/pmf/pmf.h
>> +++ b/drivers/platform/x86/amd/pmf/pmf.h
>> @@ -11,7 +11,11 @@
>>  #ifndef PMF_H
>>  #define PMF_H
>>  
>> +#include <acpi/video.h>
>> +#include <drm/drm_connector.h>
>>  #include <linux/acpi.h>
>> +#include <linux/backlight.h>
>> +#include <linux/pci.h>
>>  #include <linux/platform_profile.h>
>>  
>>  #define POLICY_BUF_MAX_SZ		0x4b000
>> @@ -83,6 +87,8 @@
>>  #define TA_OUTPUT_RESERVED_MEM				906
>>  #define MAX_OPERATION_PARAMS					4
>>  
>> +#define MAX_SUPPORTED_DISPLAY		4
>> +
>>  /* AMD PMF BIOS interfaces */
>>  struct apmf_verify_interface {
>>  	u16 size;
>> @@ -194,6 +200,15 @@ enum power_modes {
>>  	POWER_MODE_MAX,
>>  };
>>  
>> +struct amd_gpu_pmf_data {
>> +	struct pci_dev *gpu_dev;
>> +	struct backlight_device *raw_bd;
>> +	struct thermal_cooling_device *cooling_dev;
>> +	enum drm_connector_status con_status[MAX_SUPPORTED_DISPLAY];
>> +	int display_count;
>> +	int connector_type[MAX_SUPPORTED_DISPLAY];
>> +};
>> +
>>  struct amd_pmf_dev {
>>  	void __iomem *regbase;
>>  	void __iomem *smu_virt_addr;
>> @@ -228,9 +243,12 @@ struct amd_pmf_dev {
>>  	void *shbuf;
>>  	struct delayed_work pb_work;
>>  	struct pmf_action_table *prev_data;
>> +	struct amd_gpu_pmf_data gfx_data;
>>  	u64 policy_addr;
>>  	void *policy_base;
>>  	bool smart_pc_enabled;
>> +	struct pci_dev *root;
>> +	struct drm_device *drm_dev;
>>  };
>>  
>>  struct apmf_sps_prop_granular {
>> diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
>> index 959146fd483f..47ec563088b8 100644
>> --- a/drivers/platform/x86/amd/pmf/spc.c
>> +++ b/drivers/platform/x86/amd/pmf/spc.c
>> @@ -44,6 +44,10 @@ void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *
>>  	dev_dbg(dev->dev, "Max C0 Residency: %u\n", in->ev_info.max_c0residency);
>>  	dev_dbg(dev->dev, "GFX Busy: %u\n", in->ev_info.gfx_busy);
>>  	dev_dbg(dev->dev, "Connected Display Count: %u\n", in->ev_info.monitor_count);
>> +	dev_dbg(dev->dev, "Primary Display Type: %s\n",
>> +		drm_get_connector_type_name(in->ev_info.display_type));
>> +	dev_dbg(dev->dev, "Primary Display State: %s\n", in->ev_info.display_state ?
>> +			"Connected" : "disconnected/unknown");
>>  	dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ? "close" : "open");
>>  	dev_dbg(dev->dev, "==== TA inputs END ====\n");
>>  }
>> @@ -146,6 +150,41 @@ static int amd_pmf_get_slider_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_
>>  	return 0;
>>  }
>>  
>> +static int amd_pmf_get_gfx_data(struct amd_pmf_dev *pmf)
>> +{
>> +	struct drm_connector_list_iter iter;
>> +	struct drm_connector *connector;
>> +	int i = 0;
>> +
>> +	/* Reset the count to zero */
>> +	pmf->gfx_data.display_count = 0;
>> +
>> +	drm_connector_list_iter_begin(pmf->drm_dev, &iter);
>> +	drm_for_each_connector_iter(connector, &iter) {
>> +		if (connector->status == connector_status_connected)
>> +			pmf->gfx_data.display_count++;
>> +		if (connector->status != pmf->gfx_data.con_status[i])
>> +			pmf->gfx_data.con_status[i] = connector->status;
>> +		if (connector->connector_type != pmf->gfx_data.connector_type[i])
>> +			pmf->gfx_data.connector_type[i] = connector->connector_type;
>> +
>> +		i++;
>> +		if (i >= MAX_SUPPORTED_DISPLAY)
>> +			break;
>> +	}
>> +	drm_connector_list_iter_end(&iter);
>> +
>> +	return 0;
>> +}
>> +
>> +static void amd_pmf_get_gpu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
>> +{
>> +	amd_pmf_get_gfx_data(dev);
>> +	in->ev_info.monitor_count = dev->gfx_data.display_count;
>> +	in->ev_info.display_type = dev->gfx_data.connector_type[0];
>> +	in->ev_info.display_state = dev->gfx_data.con_status[0];
>> +}
>> +
>>  void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
>>  {
>>  	/* TA side lid open is 1 and close is 0, hence the ! here */
>> @@ -154,4 +193,6 @@ void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_tab
>>  	amd_pmf_get_smu_info(dev, in);
>>  	amd_pmf_get_battery_info(dev, in);
>>  	amd_pmf_get_slider_info(dev, in);
>> +	if (dev->drm_dev)
>> +		amd_pmf_get_gpu_info(dev, in);
>>  }
>> diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
>> index 81b1bd356e83..82ee2b1c627f 100644
>> --- a/drivers/platform/x86/amd/pmf/tee-if.c
>> +++ b/drivers/platform/x86/amd/pmf/tee-if.c
>> @@ -10,6 +10,7 @@
>>  
>>  #include <linux/debugfs.h>
>>  #include <linux/tee_drv.h>
>> +#include <linux/thermal.h>
>>  #include <linux/uuid.h>
>>  #include "pmf.h"
>>  
>> @@ -422,6 +423,60 @@ static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
>>  	tee_client_close_context(dev->tee_ctx);
>>  }
>>  
>> +static int amd_pmf_gpu_get_cur_state(struct thermal_cooling_device *cooling_dev,
>> +				     unsigned long *state)
>> +{
>> +	struct backlight_device *bd;
>> +
>> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
>> +		return -ENODEV;
>> +
>> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>> +	if (!bd)
>> +		return -ENODEV;
>> +
>> +	*state = backlight_get_brightness(bd);
>> +
>> +	return 0;
>> +}
>> +
>> +static int amd_pmf_gpu_get_max_state(struct thermal_cooling_device *cooling_dev,
>> +				     unsigned long *state)
>> +{
>> +	struct backlight_device *bd;
>> +
>> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
>> +		return -ENODEV;
>> +
>> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>> +	if (!bd)
>> +		return -ENODEV;
>> +
>> +	if (backlight_is_blank(bd))
>> +		*state = 0;
>> +	else
>> +		*state = bd->props.max_brightness;
>> +
>> +	return 0;
>> +}
>> +
>> +static const struct thermal_cooling_device_ops bd_cooling_ops = {
>> +	.get_max_state = amd_pmf_gpu_get_max_state,
>> +	.get_cur_state = amd_pmf_gpu_get_cur_state,
>> +};
> 
> This is still the wrong thing to do. The new PMF code MUST only
> be a *consumer* of the thermal_cooling_device API.
> 
> The thermal-cooling device for the backlight itself MUST be
> registered by the amdgpu driver.
> 
> I believe that the correct fix for this is to add a new flag to
> the backlight_properties struct;
> And then make backlight_device_register() register
> a thermal_cooling_device for the backlight when this new flag is set.
> 
> This way we get a nice generic way to use backlight class devices
> as thermal cooling devices and you can make the amdgpu driver
> register the thermal cooling device by simply adding the new flag
> to the backlight_properties struct used in the amdgpu driver.

Agreed. I am also of the same opinion.

> 
>> +static int amd_pmf_get_gpu_handle(struct pci_dev *pdev, void *data)
>> +{
>> +	struct amd_pmf_dev *dev = data;
>> +
>> +	if (pdev->vendor == PCI_VENDOR_ID_ATI && pdev->devfn == 0) {
>> +		dev->gfx_data.gpu_dev = pdev;
>> +		return 1; /* Stop walking */
>> +	}
>> +
>> +	return 0; /* Continue walking */
>> +}
>> +
>>  int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>>  {
>>  	int ret;
>> @@ -433,10 +488,30 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>>  	INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
>>  	amd_pmf_set_dram_addr(dev);
>>  	amd_pmf_get_bios_buffer(dev);
>> +
>>  	dev->prev_data = kzalloc(sizeof(*dev->prev_data), GFP_KERNEL);
>>  	if (!dev->prev_data)
>>  		return -ENOMEM;
>>  
>> +	pci_walk_bus(dev->root->bus, amd_pmf_get_gpu_handle, dev);
>> +	if (dev->gfx_data.gpu_dev) {
>> +		dev->drm_dev = pci_get_drvdata(dev->gfx_data.gpu_dev);
>> +		if (!dev->drm_dev)
>> +			return -EINVAL;
> 
> You cannot just call pci_get_drvdata() on a device for which you
> are not the driver. You have no idea of the lifetime of this device,
> the driver may unbind and release the memory pci_get_drvdata()
> points to, leaving this code with a dangling pointer which will
> crash the kernel the first time you try to use it.
> 
> Also since you are not the owner you MUST not assume any specific
> type for this memory, you cannot be sure this actually is of
> the type drm_device. Basically you MUST not touch another
> driver's drvdata *at all*.
> 
> The proper way to fix this would be to add some API to the DRM
> subsystem to query the information you are looking for form
> the DRM subsystem.
> 
> Poking directly inside other driver's internals is NOT acceptable,
> NACK for this patch.
> 

I am inline with your remarks, but I could find a way where the
thermal driver registration, handling the backlight without having the
changes in the amdgpu driver very tricky.

Like you said, I am also of the same opinion that PMF driver should
call the DRM/GPU subsystem APIs (like it does with other subsystems).

But Christian had concerns on adding all of these into the GPU driver.
So I had to roll back these into the PMF driver, and hence you see a
pci_get_drvdata() call.

I can add the thermal device registration into the amdgpu driver and
then call the DRM APIs from the PMF driver.

Christian, do you have any feedback here please?

For the sake of simplicity, I can drop patches 13/17 and 14/17 from
the series and send them separately later.

Thanks,
Shyam

> Regards,
> 
> Hans
> 
> 
>> +
>> +		if (acpi_video_get_backlight_type() != acpi_backlight_native)
>> +			return -ENODEV;
>> +
>> +		dev->gfx_data.raw_bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>> +		if (!dev->gfx_data.raw_bd)
>> +			return -ENODEV;
>> +
>> +		dev->gfx_data.cooling_dev = thermal_cooling_device_register("pmf_gpu_bd",
>> +									    NULL, &bd_cooling_ops);
>> +		if (IS_ERR(dev->gfx_data.cooling_dev))
>> +			return -ENODEV;
>> +	}
>> +
>>  	return dev->smart_pc_enabled;
>>  }
>>  
>> @@ -448,5 +523,7 @@ void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev)
>>  	kfree(dev->prev_data);
>>  	kfree(dev->policy_buf);
>>  	cancel_delayed_work_sync(&dev->pb_work);
>> +	if (dev->gfx_data.cooling_dev)
>> +		thermal_cooling_device_unregister(dev->gfx_data.cooling_dev);
>>  	amd_pmf_tee_deinit(dev);
>>  }
> 
> 

Re: [PATCH v5 13/17] platform/x86/amd/pmf: Add PMF-AMDGPU get interface

[Hans de Goede]

Hi,

On 11/17/23 12:08, Shyam Sundar S K wrote:
> Adding AMDGPU folks (Alex and Christian) - I had to drop them as the
> changes from amdgpu were moved to PMF driver.
> 
> Hi Hans,
> 
> 
> On 11/17/2023 4:18 PM, Hans de Goede wrote:
>> Hi Shyam,
>>
>> On 11/17/23 09:07, Shyam Sundar S K wrote:
>>> In order to provide GPU inputs to TA for the Smart PC solution to work, we
>>> need to have interface between the PMF driver and the AMDGPU driver.
>>>
>>> Add the initial code path for get interface from AMDGPU.
>>>
>>> Co-developed-by: Mario Limonciello <mario.limonciello@amd.com>
>>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
>>> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>

<snip>

>>> diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
>>> index 81b1bd356e83..82ee2b1c627f 100644
>>> --- a/drivers/platform/x86/amd/pmf/tee-if.c
>>> +++ b/drivers/platform/x86/amd/pmf/tee-if.c
>>> @@ -10,6 +10,7 @@
>>>  
>>>  #include <linux/debugfs.h>
>>>  #include <linux/tee_drv.h>
>>> +#include <linux/thermal.h>
>>>  #include <linux/uuid.h>
>>>  #include "pmf.h"
>>>  
>>> @@ -422,6 +423,60 @@ static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
>>>  	tee_client_close_context(dev->tee_ctx);
>>>  }
>>>  
>>> +static int amd_pmf_gpu_get_cur_state(struct thermal_cooling_device *cooling_dev,
>>> +				     unsigned long *state)
>>> +{
>>> +	struct backlight_device *bd;
>>> +
>>> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
>>> +		return -ENODEV;
>>> +
>>> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>>> +	if (!bd)
>>> +		return -ENODEV;
>>> +
>>> +	*state = backlight_get_brightness(bd);
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static int amd_pmf_gpu_get_max_state(struct thermal_cooling_device *cooling_dev,
>>> +				     unsigned long *state)
>>> +{
>>> +	struct backlight_device *bd;
>>> +
>>> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
>>> +		return -ENODEV;
>>> +
>>> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>>> +	if (!bd)
>>> +		return -ENODEV;
>>> +
>>> +	if (backlight_is_blank(bd))
>>> +		*state = 0;
>>> +	else
>>> +		*state = bd->props.max_brightness;
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static const struct thermal_cooling_device_ops bd_cooling_ops = {
>>> +	.get_max_state = amd_pmf_gpu_get_max_state,
>>> +	.get_cur_state = amd_pmf_gpu_get_cur_state,
>>> +};
>>
>> This is still the wrong thing to do. The new PMF code MUST only
>> be a *consumer* of the thermal_cooling_device API.
>>
>> The thermal-cooling device for the backlight itself MUST be
>> registered by the amdgpu driver.
>>
>> I believe that the correct fix for this is to add a new flag to
>> the backlight_properties struct;
>> And then make backlight_device_register() register
>> a thermal_cooling_device for the backlight when this new flag is set.
>>
>> This way we get a nice generic way to use backlight class devices
>> as thermal cooling devices and you can make the amdgpu driver
>> register the thermal cooling device by simply adding the new flag
>> to the backlight_properties struct used in the amdgpu driver.
> 
> Agreed. I am also of the same opinion.

So the change to the AMDGPU driver here would just be setting
this one new flag in the backlight_properties struct.

Alex, Christian, would that be acceptable to you ?


>>> +static int amd_pmf_get_gpu_handle(struct pci_dev *pdev, void *data)
>>> +{
>>> +	struct amd_pmf_dev *dev = data;
>>> +
>>> +	if (pdev->vendor == PCI_VENDOR_ID_ATI && pdev->devfn == 0) {
>>> +		dev->gfx_data.gpu_dev = pdev;
>>> +		return 1; /* Stop walking */
>>> +	}
>>> +
>>> +	return 0; /* Continue walking */
>>> +}
>>> +
>>>  int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>>>  {
>>>  	int ret;
>>> @@ -433,10 +488,30 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>>>  	INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
>>>  	amd_pmf_set_dram_addr(dev);
>>>  	amd_pmf_get_bios_buffer(dev);
>>> +
>>>  	dev->prev_data = kzalloc(sizeof(*dev->prev_data), GFP_KERNEL);
>>>  	if (!dev->prev_data)
>>>  		return -ENOMEM;
>>>  
>>> +	pci_walk_bus(dev->root->bus, amd_pmf_get_gpu_handle, dev);
>>> +	if (dev->gfx_data.gpu_dev) {
>>> +		dev->drm_dev = pci_get_drvdata(dev->gfx_data.gpu_dev);
>>> +		if (!dev->drm_dev)
>>> +			return -EINVAL;
>>
>> You cannot just call pci_get_drvdata() on a device for which you
>> are not the driver. You have no idea of the lifetime of this device,
>> the driver may unbind and release the memory pci_get_drvdata()
>> points to, leaving this code with a dangling pointer which will
>> crash the kernel the first time you try to use it.
>>
>> Also since you are not the owner you MUST not assume any specific
>> type for this memory, you cannot be sure this actually is of
>> the type drm_device. Basically you MUST not touch another
>> driver's drvdata *at all*.
>>
>> The proper way to fix this would be to add some API to the DRM
>> subsystem to query the information you are looking for form
>> the DRM subsystem.
>>
>> Poking directly inside other driver's internals is NOT acceptable,
>> NACK for this patch.
>>
> 
> I am inline with your remarks, but I could find a way where the
> thermal driver registration, handling the backlight without having the
> changes in the amdgpu driver very tricky.

As mentioned above I think there should be generic thermal cooling
device support added to drivers/video/backlight/backlight.c, then
the amdgpu code just needs to pass a flag when registering
the backlight to enable this.

> Like you said, I am also of the same opinion that PMF driver should
> call the DRM/GPU subsystem APIs (like it does with other subsystems).
> 
> But Christian had concerns on adding all of these into the GPU driver.
> So I had to roll back these into the PMF driver, and hence you see a
> pci_get_drvdata() call.

Ok, so I can see how this AMD PMF code is all kinda special
and how the DRM folks don't want to have to add APIs just for
that. But IMHO calling pci_get_drvdata() on an unowned
device is completely unacceptable.

At a minimum we need life-cycle management for the drm_device
which the PMF code is using, something like:

struct drm_device *drm_device_find(const void *data,
   int (*match)(struct drm_device *dev, const void *data));

which works similar to class_find_device() and returns
a reference to the drm_device for which match has returned 0
(which also stops it from looping over devices).

Combined with a:

void drm_device_put(struct drm_device *dev);

for when the PMF code is done with the device.

This way the PMF code can safely get a reference to drm_device
and release it when it is done. Rather then just getting
some random pointer which may or not actually be a drm_device
and the lifetime of which is not guaranteed in anyway.

E.g. if the PMF driver loads before amdgpu then
pci_get_drvdata() will just return NULL.

And as mentioned if the amdgpu driver gets unbound after
the PMF code has called  pci_get_drvdata() the PMF driver
now has a dangling pointer.

So IMHO adding: drm_device_find() + drm_device_put()
to the DRM core are minimum which is necessary here.

If the PMF code then ends up doing things like
drm_for_each_connector_iter() on the gotten drm_device
referemce so be it. But we must make sure we have
a properly lifecycle managed reference first and
pci_get_drvdata() does not give us that.

> For the sake of simplicity, I can drop patches 13/17 and 14/17 from
> the series and send them separately later.

Yes I think that dropping the GPU related patches for
now would be best.

Regards,

Hans

Re: [PATCH v5 13/17] platform/x86/amd/pmf: Add PMF-AMDGPU get interface

[Shyam Sundar S K]

Hi Hans,

On 11/17/2023 5:01 PM, Hans de Goede wrote:
> Hi,
> 
> On 11/17/23 12:08, Shyam Sundar S K wrote:
>> Adding AMDGPU folks (Alex and Christian) - I had to drop them as the
>> changes from amdgpu were moved to PMF driver.
>>
>> Hi Hans,
>>
>>
>> On 11/17/2023 4:18 PM, Hans de Goede wrote:
>>> Hi Shyam,
>>>
>>> On 11/17/23 09:07, Shyam Sundar S K wrote:
>>>> In order to provide GPU inputs to TA for the Smart PC solution to work, we
>>>> need to have interface between the PMF driver and the AMDGPU driver.
>>>>
>>>> Add the initial code path for get interface from AMDGPU.
>>>>
>>>> Co-developed-by: Mario Limonciello <mario.limonciello@amd.com>
>>>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
>>>> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> 
> <snip>
> 
>>>> diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
>>>> index 81b1bd356e83..82ee2b1c627f 100644
>>>> --- a/drivers/platform/x86/amd/pmf/tee-if.c
>>>> +++ b/drivers/platform/x86/amd/pmf/tee-if.c
>>>> @@ -10,6 +10,7 @@
>>>>  
>>>>  #include <linux/debugfs.h>
>>>>  #include <linux/tee_drv.h>
>>>> +#include <linux/thermal.h>
>>>>  #include <linux/uuid.h>
>>>>  #include "pmf.h"
>>>>  
>>>> @@ -422,6 +423,60 @@ static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
>>>>  	tee_client_close_context(dev->tee_ctx);
>>>>  }
>>>>  
>>>> +static int amd_pmf_gpu_get_cur_state(struct thermal_cooling_device *cooling_dev,
>>>> +				     unsigned long *state)
>>>> +{
>>>> +	struct backlight_device *bd;
>>>> +
>>>> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
>>>> +		return -ENODEV;
>>>> +
>>>> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>>>> +	if (!bd)
>>>> +		return -ENODEV;
>>>> +
>>>> +	*state = backlight_get_brightness(bd);
>>>> +
>>>> +	return 0;
>>>> +}
>>>> +
>>>> +static int amd_pmf_gpu_get_max_state(struct thermal_cooling_device *cooling_dev,
>>>> +				     unsigned long *state)
>>>> +{
>>>> +	struct backlight_device *bd;
>>>> +
>>>> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
>>>> +		return -ENODEV;
>>>> +
>>>> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>>>> +	if (!bd)
>>>> +		return -ENODEV;
>>>> +
>>>> +	if (backlight_is_blank(bd))
>>>> +		*state = 0;
>>>> +	else
>>>> +		*state = bd->props.max_brightness;
>>>> +
>>>> +	return 0;
>>>> +}
>>>> +
>>>> +static const struct thermal_cooling_device_ops bd_cooling_ops = {
>>>> +	.get_max_state = amd_pmf_gpu_get_max_state,
>>>> +	.get_cur_state = amd_pmf_gpu_get_cur_state,
>>>> +};
>>>
>>> This is still the wrong thing to do. The new PMF code MUST only
>>> be a *consumer* of the thermal_cooling_device API.
>>>
>>> The thermal-cooling device for the backlight itself MUST be
>>> registered by the amdgpu driver.
>>>
>>> I believe that the correct fix for this is to add a new flag to
>>> the backlight_properties struct;
>>> And then make backlight_device_register() register
>>> a thermal_cooling_device for the backlight when this new flag is set.
>>>
>>> This way we get a nice generic way to use backlight class devices
>>> as thermal cooling devices and you can make the amdgpu driver
>>> register the thermal cooling device by simply adding the new flag
>>> to the backlight_properties struct used in the amdgpu driver.
>>
>> Agreed. I am also of the same opinion.
> 
> So the change to the AMDGPU driver here would just be setting
> this one new flag in the backlight_properties struct.
> 
> Alex, Christian, would that be acceptable to you ?
> 
> 
>>>> +static int amd_pmf_get_gpu_handle(struct pci_dev *pdev, void *data)
>>>> +{
>>>> +	struct amd_pmf_dev *dev = data;
>>>> +
>>>> +	if (pdev->vendor == PCI_VENDOR_ID_ATI && pdev->devfn == 0) {
>>>> +		dev->gfx_data.gpu_dev = pdev;
>>>> +		return 1; /* Stop walking */
>>>> +	}
>>>> +
>>>> +	return 0; /* Continue walking */
>>>> +}
>>>> +
>>>>  int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>>>>  {
>>>>  	int ret;
>>>> @@ -433,10 +488,30 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>>>>  	INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
>>>>  	amd_pmf_set_dram_addr(dev);
>>>>  	amd_pmf_get_bios_buffer(dev);
>>>> +
>>>>  	dev->prev_data = kzalloc(sizeof(*dev->prev_data), GFP_KERNEL);
>>>>  	if (!dev->prev_data)
>>>>  		return -ENOMEM;
>>>>  
>>>> +	pci_walk_bus(dev->root->bus, amd_pmf_get_gpu_handle, dev);
>>>> +	if (dev->gfx_data.gpu_dev) {
>>>> +		dev->drm_dev = pci_get_drvdata(dev->gfx_data.gpu_dev);
>>>> +		if (!dev->drm_dev)
>>>> +			return -EINVAL;
>>>
>>> You cannot just call pci_get_drvdata() on a device for which you
>>> are not the driver. You have no idea of the lifetime of this device,
>>> the driver may unbind and release the memory pci_get_drvdata()
>>> points to, leaving this code with a dangling pointer which will
>>> crash the kernel the first time you try to use it.
>>>
>>> Also since you are not the owner you MUST not assume any specific
>>> type for this memory, you cannot be sure this actually is of
>>> the type drm_device. Basically you MUST not touch another
>>> driver's drvdata *at all*.
>>>
>>> The proper way to fix this would be to add some API to the DRM
>>> subsystem to query the information you are looking for form
>>> the DRM subsystem.
>>>
>>> Poking directly inside other driver's internals is NOT acceptable,
>>> NACK for this patch.
>>>
>>
>> I am inline with your remarks, but I could find a way where the
>> thermal driver registration, handling the backlight without having the
>> changes in the amdgpu driver very tricky.
> 
> As mentioned above I think there should be generic thermal cooling
> device support added to drivers/video/backlight/backlight.c, then
> the amdgpu code just needs to pass a flag when registering
> the backlight to enable this.
> 
>> Like you said, I am also of the same opinion that PMF driver should
>> call the DRM/GPU subsystem APIs (like it does with other subsystems).
>>
>> But Christian had concerns on adding all of these into the GPU driver.
>> So I had to roll back these into the PMF driver, and hence you see a
>> pci_get_drvdata() call.
> 
> Ok, so I can see how this AMD PMF code is all kinda special
> and how the DRM folks don't want to have to add APIs just for
> that. But IMHO calling pci_get_drvdata() on an unowned
> device is completely unacceptable.
> 
> At a minimum we need life-cycle management for the drm_device
> which the PMF code is using, something like:
> 
> struct drm_device *drm_device_find(const void *data,
>    int (*match)(struct drm_device *dev, const void *data));
> 
> which works similar to class_find_device() and returns
> a reference to the drm_device for which match has returned 0
> (which also stops it from looping over devices).
> 
> Combined with a:
> 
> void drm_device_put(struct drm_device *dev);
> 
> for when the PMF code is done with the device.
> 
> This way the PMF code can safely get a reference to drm_device
> and release it when it is done. Rather then just getting
> some random pointer which may or not actually be a drm_device
> and the lifetime of which is not guaranteed in anyway.
> 
> E.g. if the PMF driver loads before amdgpu then
> pci_get_drvdata() will just return NULL.
> 
> And as mentioned if the amdgpu driver gets unbound after
> the PMF code has called  pci_get_drvdata() the PMF driver
> now has a dangling pointer.
> 
> So IMHO adding: drm_device_find() + drm_device_put()
> to the DRM core are minimum which is necessary here.
> 
> If the PMF code then ends up doing things like
> drm_for_each_connector_iter() on the gotten drm_device
> referemce so be it. But we must make sure we have
> a properly lifecycle managed reference first and
> pci_get_drvdata() does not give us that.
> 

Ok I will work on your feedback.

>> For the sake of simplicity, I can drop patches 13/17 and 14/17 from
>> the series and send them separately later.
> 
> Yes I think that dropping the GPU related patches for
> now would be best.
> 

Thank you. Let me wait for feedback from others before I drop the GPU
patches.

Thanks,
Shyam

> Regards,
> 
> Hans
> 
> 
> 
> 

Re: [PATCH v5 13/17] platform/x86/amd/pmf: Add PMF-AMDGPU get interface

[Mario Limonciello]

On 11/17/2023 02:07, Shyam Sundar S K wrote:
> In order to provide GPU inputs to TA for the Smart PC solution to work, we
> need to have interface between the PMF driver and the AMDGPU driver.
> 
> Add the initial code path for get interface from AMDGPU.

Make sure you update the commit message when you resumbit v6.

> 
> Co-developed-by: Mario Limonciello <mario.limonciello@amd.com>
> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> ---
>   drivers/platform/x86/amd/pmf/core.c   |  3 +-
>   drivers/platform/x86/amd/pmf/pmf.h    | 18 +++++++
>   drivers/platform/x86/amd/pmf/spc.c    | 41 ++++++++++++++
>   drivers/platform/x86/amd/pmf/tee-if.c | 77 +++++++++++++++++++++++++++
>   4 files changed, 138 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c
> index 4b8156033fa6..9b14a997cd48 100644
> --- a/drivers/platform/x86/amd/pmf/core.c
> +++ b/drivers/platform/x86/amd/pmf/core.c
> @@ -411,6 +411,7 @@ static int amd_pmf_probe(struct platform_device *pdev)
>   	}
>   
>   	dev->cpu_id = rdev->device;
> +	dev->root = rdev;
>   
>   	err = amd_smn_read(0, AMD_PMF_BASE_ADDR_LO, &val);
>   	if (err) {
> @@ -482,4 +483,4 @@ module_platform_driver(amd_pmf_driver);
>   
>   MODULE_LICENSE("GPL");
>   MODULE_DESCRIPTION("AMD Platform Management Framework Driver");
> -MODULE_SOFTDEP("pre: amdtee");
> +MODULE_SOFTDEP("pre: amdtee amdgpu");
> diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
> index 8712299ad52b..525308519fa3 100644
> --- a/drivers/platform/x86/amd/pmf/pmf.h
> +++ b/drivers/platform/x86/amd/pmf/pmf.h
> @@ -11,7 +11,11 @@
>   #ifndef PMF_H
>   #define PMF_H
>   
> +#include <acpi/video.h>
> +#include <drm/drm_connector.h>
>   #include <linux/acpi.h>
> +#include <linux/backlight.h>
> +#include <linux/pci.h>
>   #include <linux/platform_profile.h>
>   
>   #define POLICY_BUF_MAX_SZ		0x4b000
> @@ -83,6 +87,8 @@
>   #define TA_OUTPUT_RESERVED_MEM				906
>   #define MAX_OPERATION_PARAMS					4
>   
> +#define MAX_SUPPORTED_DISPLAY		4
> +
>   /* AMD PMF BIOS interfaces */
>   struct apmf_verify_interface {
>   	u16 size;
> @@ -194,6 +200,15 @@ enum power_modes {
>   	POWER_MODE_MAX,
>   };
>   
> +struct amd_gpu_pmf_data {
> +	struct pci_dev *gpu_dev;
> +	struct backlight_device *raw_bd;
> +	struct thermal_cooling_device *cooling_dev;
> +	enum drm_connector_status con_status[MAX_SUPPORTED_DISPLAY];
> +	int display_count;
> +	int connector_type[MAX_SUPPORTED_DISPLAY];
> +};
> +
>   struct amd_pmf_dev {
>   	void __iomem *regbase;
>   	void __iomem *smu_virt_addr;
> @@ -228,9 +243,12 @@ struct amd_pmf_dev {
>   	void *shbuf;
>   	struct delayed_work pb_work;
>   	struct pmf_action_table *prev_data;
> +	struct amd_gpu_pmf_data gfx_data;
>   	u64 policy_addr;
>   	void *policy_base;
>   	bool smart_pc_enabled;
> +	struct pci_dev *root;
> +	struct drm_device *drm_dev;
>   };
>   
>   struct apmf_sps_prop_granular {
> diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
> index 959146fd483f..47ec563088b8 100644
> --- a/drivers/platform/x86/amd/pmf/spc.c
> +++ b/drivers/platform/x86/amd/pmf/spc.c
> @@ -44,6 +44,10 @@ void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *
>   	dev_dbg(dev->dev, "Max C0 Residency: %u\n", in->ev_info.max_c0residency);
>   	dev_dbg(dev->dev, "GFX Busy: %u\n", in->ev_info.gfx_busy);
>   	dev_dbg(dev->dev, "Connected Display Count: %u\n", in->ev_info.monitor_count);
> +	dev_dbg(dev->dev, "Primary Display Type: %s\n",
> +		drm_get_connector_type_name(in->ev_info.display_type));
> +	dev_dbg(dev->dev, "Primary Display State: %s\n", in->ev_info.display_state ?
> +			"Connected" : "disconnected/unknown");
>   	dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ? "close" : "open");
>   	dev_dbg(dev->dev, "==== TA inputs END ====\n");
>   }
> @@ -146,6 +150,41 @@ static int amd_pmf_get_slider_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_
>   	return 0;
>   }
>   
> +static int amd_pmf_get_gfx_data(struct amd_pmf_dev *pmf)
> +{
> +	struct drm_connector_list_iter iter;
> +	struct drm_connector *connector;
> +	int i = 0;
> +
> +	/* Reset the count to zero */
> +	pmf->gfx_data.display_count = 0;
> +
> +	drm_connector_list_iter_begin(pmf->drm_dev, &iter);
> +	drm_for_each_connector_iter(connector, &iter) {
> +		if (connector->status == connector_status_connected)
> +			pmf->gfx_data.display_count++;
> +		if (connector->status != pmf->gfx_data.con_status[i])
> +			pmf->gfx_data.con_status[i] = connector->status;
> +		if (connector->connector_type != pmf->gfx_data.connector_type[i])
> +			pmf->gfx_data.connector_type[i] = connector->connector_type;
> +
> +		i++;
> +		if (i >= MAX_SUPPORTED_DISPLAY)
> +			break;
> +	}
> +	drm_connector_list_iter_end(&iter);
> +
> +	return 0;
> +}
> +
> +static void amd_pmf_get_gpu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
> +{
> +	amd_pmf_get_gfx_data(dev);
> +	in->ev_info.monitor_count = dev->gfx_data.display_count;
> +	in->ev_info.display_type = dev->gfx_data.connector_type[0];
> +	in->ev_info.display_state = dev->gfx_data.con_status[0];

Can you elaborate on future expansion areas for the TA as it pertains to 
this info?

Do you think it's going to need information from more than just the 
first display?
Is monitor count boolean or does it actually care about the count > 1?
Is it *really* looking at whether the eDP is active (this is what would 
affect power most signifiantly IIUC)?

If this isn't an area that is likely to expand much in the future and is 
really "just about eDP", I wonder if the better answer is a DRM helper 
that is something like:

bool drm_edp_connected(void);

If it is likely to expand, it could be a set of multiple helpers.

> +}
> +
>   void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
>   {
>   	/* TA side lid open is 1 and close is 0, hence the ! here */
> @@ -154,4 +193,6 @@ void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_tab
>   	amd_pmf_get_smu_info(dev, in);
>   	amd_pmf_get_battery_info(dev, in);
>   	amd_pmf_get_slider_info(dev, in);
> +	if (dev->drm_dev)
> +		amd_pmf_get_gpu_info(dev, in);
>   }
> diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
> index 81b1bd356e83..82ee2b1c627f 100644
> --- a/drivers/platform/x86/amd/pmf/tee-if.c
> +++ b/drivers/platform/x86/amd/pmf/tee-if.c
> @@ -10,6 +10,7 @@
>   
>   #include <linux/debugfs.h>
>   #include <linux/tee_drv.h>
> +#include <linux/thermal.h>
>   #include <linux/uuid.h>
>   #include "pmf.h"
>   
> @@ -422,6 +423,60 @@ static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
>   	tee_client_close_context(dev->tee_ctx);
>   }
>   
> +static int amd_pmf_gpu_get_cur_state(struct thermal_cooling_device *cooling_dev,
> +				     unsigned long *state)
> +{
> +	struct backlight_device *bd;
> +
> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
> +		return -ENODEV;
> +
> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
> +	if (!bd)
> +		return -ENODEV;
> +
> +	*state = backlight_get_brightness(bd);
> +
> +	return 0;
> +}
> +
> +static int amd_pmf_gpu_get_max_state(struct thermal_cooling_device *cooling_dev,
> +				     unsigned long *state)
> +{
> +	struct backlight_device *bd;
> +
> +	if (acpi_video_get_backlight_type() != acpi_backlight_native)
> +		return -ENODEV;
> +
> +	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
> +	if (!bd)
> +		return -ENODEV;
> +
> +	if (backlight_is_blank(bd))
> +		*state = 0;
> +	else
> +		*state = bd->props.max_brightness;
> +
> +	return 0;
> +}
> +
> +static const struct thermal_cooling_device_ops bd_cooling_ops = {
> +	.get_max_state = amd_pmf_gpu_get_max_state,
> +	.get_cur_state = amd_pmf_gpu_get_cur_state,
> +};
> +
> +static int amd_pmf_get_gpu_handle(struct pci_dev *pdev, void *data)
> +{
> +	struct amd_pmf_dev *dev = data;
> +
> +	if (pdev->vendor == PCI_VENDOR_ID_ATI && pdev->devfn == 0) {
> +		dev->gfx_data.gpu_dev = pdev;
> +		return 1; /* Stop walking */
> +	}
> +
> +	return 0; /* Continue walking */
> +}
> +
>   int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>   {
>   	int ret;
> @@ -433,10 +488,30 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>   	INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
>   	amd_pmf_set_dram_addr(dev);
>   	amd_pmf_get_bios_buffer(dev);
> +
>   	dev->prev_data = kzalloc(sizeof(*dev->prev_data), GFP_KERNEL);
>   	if (!dev->prev_data)
>   		return -ENOMEM;
>   
> +	pci_walk_bus(dev->root->bus, amd_pmf_get_gpu_handle, dev);
> +	if (dev->gfx_data.gpu_dev) {
> +		dev->drm_dev = pci_get_drvdata(dev->gfx_data.gpu_dev);

I did see Hans response; but I want to mention that specifically a good 
litmus test whether what you're doing to get the data from GPU device 
side is safe would be to ssh into the machine, stop the GUI and then 
unbind the PCI device from amdgpu driver and make sure nothing explodes.

If it does explode, take PMF out of the picture and see if it was caused 
by what you did or an existing problem.


> +		if (!dev->drm_dev)
> +			return -EINVAL;
> +
> +		if (acpi_video_get_backlight_type() != acpi_backlight_native)
> +			return -ENODEV;
> +
> +		dev->gfx_data.raw_bd = backlight_device_get_by_type(BACKLIGHT_RAW);
> +		if (!dev->gfx_data.raw_bd)
> +			return -ENODEV;
> +
> +		dev->gfx_data.cooling_dev = thermal_cooling_device_register("pmf_gpu_bd",
> +									    NULL, &bd_cooling_ops);
> +		if (IS_ERR(dev->gfx_data.cooling_dev))
> +			return -ENODEV;
> +	}
> +
>   	return dev->smart_pc_enabled;
>   }
>   
> @@ -448,5 +523,7 @@ void amd_pmf_deinit_smart_pc(struct amd_pmf_dev *dev)
>   	kfree(dev->prev_data);
>   	kfree(dev->policy_buf);
>   	cancel_delayed_work_sync(&dev->pb_work);
> +	if (dev->gfx_data.cooling_dev)
> +		thermal_cooling_device_unregister(dev->gfx_data.cooling_dev);
>   	amd_pmf_tee_deinit(dev);
>   }

Re: [PATCH v5 13/17] platform/x86/amd/pmf: Add PMF-AMDGPU get interface

[Shyam Sundar S K]

On 11/17/2023 11:43 PM, Mario Limonciello wrote:
> On 11/17/2023 02:07, Shyam Sundar S K wrote:
>> In order to provide GPU inputs to TA for the Smart PC solution to
>> work, we
>> need to have interface between the PMF driver and the AMDGPU driver.
>>
>> Add the initial code path for get interface from AMDGPU.
> 
> Make sure you update the commit message when you resumbit v6.
> 
>>
>> Co-developed-by: Mario Limonciello <mario.limonciello@amd.com>
>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
>> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
>> ---
>>   drivers/platform/x86/amd/pmf/core.c   |  3 +-
>>   drivers/platform/x86/amd/pmf/pmf.h    | 18 +++++++
>>   drivers/platform/x86/amd/pmf/spc.c    | 41 ++++++++++++++
>>   drivers/platform/x86/amd/pmf/tee-if.c | 77
>> +++++++++++++++++++++++++++
>>   4 files changed, 138 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/platform/x86/amd/pmf/core.c
>> b/drivers/platform/x86/amd/pmf/core.c
>> index 4b8156033fa6..9b14a997cd48 100644
>> --- a/drivers/platform/x86/amd/pmf/core.c
>> +++ b/drivers/platform/x86/amd/pmf/core.c
>> @@ -411,6 +411,7 @@ static int amd_pmf_probe(struct platform_device
>> *pdev)
>>       }
>>         dev->cpu_id = rdev->device;
>> +    dev->root = rdev;
>>         err = amd_smn_read(0, AMD_PMF_BASE_ADDR_LO, &val);
>>       if (err) {
>> @@ -482,4 +483,4 @@ module_platform_driver(amd_pmf_driver);
>>     MODULE_LICENSE("GPL");
>>   MODULE_DESCRIPTION("AMD Platform Management Framework Driver");
>> -MODULE_SOFTDEP("pre: amdtee");
>> +MODULE_SOFTDEP("pre: amdtee amdgpu");
>> diff --git a/drivers/platform/x86/amd/pmf/pmf.h
>> b/drivers/platform/x86/amd/pmf/pmf.h
>> index 8712299ad52b..525308519fa3 100644
>> --- a/drivers/platform/x86/amd/pmf/pmf.h
>> +++ b/drivers/platform/x86/amd/pmf/pmf.h
>> @@ -11,7 +11,11 @@
>>   #ifndef PMF_H
>>   #define PMF_H
>>   +#include <acpi/video.h>
>> +#include <drm/drm_connector.h>
>>   #include <linux/acpi.h>
>> +#include <linux/backlight.h>
>> +#include <linux/pci.h>
>>   #include <linux/platform_profile.h>
>>     #define POLICY_BUF_MAX_SZ        0x4b000
>> @@ -83,6 +87,8 @@
>>   #define TA_OUTPUT_RESERVED_MEM                906
>>   #define MAX_OPERATION_PARAMS                    4
>>   +#define MAX_SUPPORTED_DISPLAY        4
>> +
>>   /* AMD PMF BIOS interfaces */
>>   struct apmf_verify_interface {
>>       u16 size;
>> @@ -194,6 +200,15 @@ enum power_modes {
>>       POWER_MODE_MAX,
>>   };
>>   +struct amd_gpu_pmf_data {
>> +    struct pci_dev *gpu_dev;
>> +    struct backlight_device *raw_bd;
>> +    struct thermal_cooling_device *cooling_dev;
>> +    enum drm_connector_status con_status[MAX_SUPPORTED_DISPLAY];
>> +    int display_count;
>> +    int connector_type[MAX_SUPPORTED_DISPLAY];
>> +};
>> +
>>   struct amd_pmf_dev {
>>       void __iomem *regbase;
>>       void __iomem *smu_virt_addr;
>> @@ -228,9 +243,12 @@ struct amd_pmf_dev {
>>       void *shbuf;
>>       struct delayed_work pb_work;
>>       struct pmf_action_table *prev_data;
>> +    struct amd_gpu_pmf_data gfx_data;
>>       u64 policy_addr;
>>       void *policy_base;
>>       bool smart_pc_enabled;
>> +    struct pci_dev *root;
>> +    struct drm_device *drm_dev;
>>   };
>>     struct apmf_sps_prop_granular {
>> diff --git a/drivers/platform/x86/amd/pmf/spc.c
>> b/drivers/platform/x86/amd/pmf/spc.c
>> index 959146fd483f..47ec563088b8 100644
>> --- a/drivers/platform/x86/amd/pmf/spc.c
>> +++ b/drivers/platform/x86/amd/pmf/spc.c
>> @@ -44,6 +44,10 @@ void amd_pmf_dump_ta_inputs(struct amd_pmf_dev
>> *dev, struct ta_pmf_enact_table *
>>       dev_dbg(dev->dev, "Max C0 Residency: %u\n",
>> in->ev_info.max_c0residency);
>>       dev_dbg(dev->dev, "GFX Busy: %u\n", in->ev_info.gfx_busy);
>>       dev_dbg(dev->dev, "Connected Display Count: %u\n",
>> in->ev_info.monitor_count);
>> +    dev_dbg(dev->dev, "Primary Display Type: %s\n",
>> +        drm_get_connector_type_name(in->ev_info.display_type));
>> +    dev_dbg(dev->dev, "Primary Display State: %s\n",
>> in->ev_info.display_state ?
>> +            "Connected" : "disconnected/unknown");
>>       dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ?
>> "close" : "open");
>>       dev_dbg(dev->dev, "==== TA inputs END ====\n");
>>   }
>> @@ -146,6 +150,41 @@ static int amd_pmf_get_slider_info(struct
>> amd_pmf_dev *dev, struct ta_pmf_enact_
>>       return 0;
>>   }
>>   +static int amd_pmf_get_gfx_data(struct amd_pmf_dev *pmf)
>> +{
>> +    struct drm_connector_list_iter iter;
>> +    struct drm_connector *connector;
>> +    int i = 0;
>> +
>> +    /* Reset the count to zero */
>> +    pmf->gfx_data.display_count = 0;
>> +
>> +    drm_connector_list_iter_begin(pmf->drm_dev, &iter);
>> +    drm_for_each_connector_iter(connector, &iter) {
>> +        if (connector->status == connector_status_connected)
>> +            pmf->gfx_data.display_count++;
>> +        if (connector->status != pmf->gfx_data.con_status[i])
>> +            pmf->gfx_data.con_status[i] = connector->status;
>> +        if (connector->connector_type !=
>> pmf->gfx_data.connector_type[i])
>> +            pmf->gfx_data.connector_type[i] =
>> connector->connector_type;
>> +
>> +        i++;
>> +        if (i >= MAX_SUPPORTED_DISPLAY)
>> +            break;
>> +    }
>> +    drm_connector_list_iter_end(&iter);
>> +
>> +    return 0;
>> +}
>> +
>> +static void amd_pmf_get_gpu_info(struct amd_pmf_dev *dev, struct
>> ta_pmf_enact_table *in)
>> +{
>> +    amd_pmf_get_gfx_data(dev);
>> +    in->ev_info.monitor_count = dev->gfx_data.display_count;
>> +    in->ev_info.display_type = dev->gfx_data.connector_type[0];
>> +    in->ev_info.display_state = dev->gfx_data.con_status[0];
> 
> Can you elaborate on future expansion areas for the TA as it pertains
> to this info?
> 

For now, the planned items are:
- GPU Engine Running Time
- GPU Engine Utilization Percent
- GPU Workload type
- GPU Power

more items are still being worked out from the TA. So definately there
is a need to build the pipe to get information from amdgpu.

> Do you think it's going to need information from more than just the
> first display?

Yes. We are looking at if there is an extended display also. Future,
we can include to see if that behind the dock etc.

But to make it simple, today TA only looks if there is a first display.

> Is monitor count boolean or does it actually care about the count > 1?
> Is it *really* looking at whether the eDP is active (this is what
> would affect power most signifiantly IIUC)?
> 

Not Boolean.

> If this isn't an area that is likely to expand much in the future and
> is really "just about eDP", I wonder if the better answer is a DRM
> helper that is something like:
> 
> bool drm_edp_connected(void);
> 
> If it is likely to expand, it could be a set of multiple helpers.
> 

Ok. Thanks. I will see how it gets shaped up.

>> +}
>> +
>>   void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct
>> ta_pmf_enact_table *in)
>>   {
>>       /* TA side lid open is 1 and close is 0, hence the ! here */
>> @@ -154,4 +193,6 @@ void amd_pmf_populate_ta_inputs(struct
>> amd_pmf_dev *dev, struct ta_pmf_enact_tab
>>       amd_pmf_get_smu_info(dev, in);
>>       amd_pmf_get_battery_info(dev, in);
>>       amd_pmf_get_slider_info(dev, in);
>> +    if (dev->drm_dev)
>> +        amd_pmf_get_gpu_info(dev, in);
>>   }
>> diff --git a/drivers/platform/x86/amd/pmf/tee-if.c
>> b/drivers/platform/x86/amd/pmf/tee-if.c
>> index 81b1bd356e83..82ee2b1c627f 100644
>> --- a/drivers/platform/x86/amd/pmf/tee-if.c
>> +++ b/drivers/platform/x86/amd/pmf/tee-if.c
>> @@ -10,6 +10,7 @@
>>     #include <linux/debugfs.h>
>>   #include <linux/tee_drv.h>
>> +#include <linux/thermal.h>
>>   #include <linux/uuid.h>
>>   #include "pmf.h"
>>   @@ -422,6 +423,60 @@ static void amd_pmf_tee_deinit(struct
>> amd_pmf_dev *dev)
>>       tee_client_close_context(dev->tee_ctx);
>>   }
>>   +static int amd_pmf_gpu_get_cur_state(struct
>> thermal_cooling_device *cooling_dev,
>> +                     unsigned long *state)
>> +{
>> +    struct backlight_device *bd;
>> +
>> +    if (acpi_video_get_backlight_type() != acpi_backlight_native)
>> +        return -ENODEV;
>> +
>> +    bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>> +    if (!bd)
>> +        return -ENODEV;
>> +
>> +    *state = backlight_get_brightness(bd);
>> +
>> +    return 0;
>> +}
>> +
>> +static int amd_pmf_gpu_get_max_state(struct thermal_cooling_device
>> *cooling_dev,
>> +                     unsigned long *state)
>> +{
>> +    struct backlight_device *bd;
>> +
>> +    if (acpi_video_get_backlight_type() != acpi_backlight_native)
>> +        return -ENODEV;
>> +
>> +    bd = backlight_device_get_by_type(BACKLIGHT_RAW);
>> +    if (!bd)
>> +        return -ENODEV;
>> +
>> +    if (backlight_is_blank(bd))
>> +        *state = 0;
>> +    else
>> +        *state = bd->props.max_brightness;
>> +
>> +    return 0;
>> +}
>> +
>> +static const struct thermal_cooling_device_ops bd_cooling_ops = {
>> +    .get_max_state = amd_pmf_gpu_get_max_state,
>> +    .get_cur_state = amd_pmf_gpu_get_cur_state,
>> +};
>> +
>> +static int amd_pmf_get_gpu_handle(struct pci_dev *pdev, void *data)
>> +{
>> +    struct amd_pmf_dev *dev = data;
>> +
>> +    if (pdev->vendor == PCI_VENDOR_ID_ATI && pdev->devfn == 0) {
>> +        dev->gfx_data.gpu_dev = pdev;
>> +        return 1; /* Stop walking */
>> +    }
>> +
>> +    return 0; /* Continue walking */
>> +}
>> +
>>   int amd_pmf_init_smart_pc(struct amd_pmf_dev *dev)
>>   {
>>       int ret;
>> @@ -433,10 +488,30 @@ int amd_pmf_init_smart_pc(struct amd_pmf_dev
>> *dev)
>>       INIT_DELAYED_WORK(&dev->pb_work, amd_pmf_invoke_cmd);
>>       amd_pmf_set_dram_addr(dev);
>>       amd_pmf_get_bios_buffer(dev);
>> +
>>       dev->prev_data = kzalloc(sizeof(*dev->prev_data), GFP_KERNEL);
>>       if (!dev->prev_data)
>>           return -ENOMEM;
>>   +    pci_walk_bus(dev->root->bus, amd_pmf_get_gpu_handle, dev);
>> +    if (dev->gfx_data.gpu_dev) {
>> +        dev->drm_dev = pci_get_drvdata(dev->gfx_data.gpu_dev);
> 
> I did see Hans response; but I want to mention that specifically a
> good litmus test whether what you're doing to get the data from GPU
> device side is safe would be to ssh into the machine, stop the GUI and
> then unbind the PCI device from amdgpu driver and make sure nothing
> explodes.
> 
> If it does explode, take PMF out of the picture and see if it was
> caused by what you did or an existing problem.
> 

Ok. I will take a look.

Thanks,
Shyam

> 
>> +        if (!dev->drm_dev)
>> +            return -EINVAL;
>> +
>> +        if (acpi_video_get_backlight_type() != acpi_backlight_native)
>> +            return -ENODEV;
>> +
>> +        dev->gfx_data.raw_bd =
>> backlight_device_get_by_type(BACKLIGHT_RAW);
>> +        if (!dev->gfx_data.raw_bd)
>> +            return -ENODEV;
>> +
>> +        dev->gfx_data.cooling_dev =
>> thermal_cooling_device_register("pmf_gpu_bd",
>> +                                        NULL, &bd_cooling_ops);
>> +        if (IS_ERR(dev->gfx_data.cooling_dev))
>> +            return -ENODEV;
>> +    }
>> +
>>       return dev->smart_pc_enabled;
>>   }
>>   @@ -448,5 +523,7 @@ void amd_pmf_deinit_smart_pc(struct
>> amd_pmf_dev *dev)
>>       kfree(dev->prev_data);
>>       kfree(dev->policy_buf);
>>       cancel_delayed_work_sync(&dev->pb_work);
>> +    if (dev->gfx_data.cooling_dev)
>> +        thermal_cooling_device_unregister(dev->gfx_data.cooling_dev);
>>       amd_pmf_tee_deinit(dev);
>>   }
> 

[PATCH v5 14/17] platform/x86/amd/pmf: Add PMF-AMDGPU set interface

[Shyam Sundar S K]

For the Smart PC Solution to fully work, it has to enact to the actions
coming from TA. Add the initial code path for set interface to AMDGPU.

Change amd_pmf_apply_policies() return type, so that it can return
errors when the call to retrieve information from amdgpu fails.

Co-developed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
---
 drivers/platform/x86/amd/pmf/pmf.h    |  2 ++
 drivers/platform/x86/amd/pmf/tee-if.c | 39 +++++++++++++++++++++++++--
 2 files changed, 39 insertions(+), 2 deletions(-)

diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h
index 525308519fa3..6e680637a631 100644
--- a/drivers/platform/x86/amd/pmf/pmf.h
+++ b/drivers/platform/x86/amd/pmf/pmf.h
@@ -78,6 +78,7 @@
 #define PMF_POLICY_STT_SKINTEMP_APU				7
 #define PMF_POLICY_STT_SKINTEMP_HS2				8
 #define PMF_POLICY_SYSTEM_STATE					9
+#define PMF_POLICY_DISPLAY_BRIGHTNESS				12
 #define PMF_POLICY_P3T						38
 
 /* TA macros */
@@ -503,6 +504,7 @@ enum ta_pmf_error_type {
 };
 
 struct pmf_action_table {
+	unsigned long display_brightness;
 	enum system_state system_state;
 	u32 spl;		/* in mW */
 	u32 sppt;		/* in mW */
diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c
index 82ee2b1c627f..94a2afd84b56 100644
--- a/drivers/platform/x86/amd/pmf/tee-if.c
+++ b/drivers/platform/x86/amd/pmf/tee-if.c
@@ -77,8 +77,10 @@ static int amd_pmf_update_uevents(struct amd_pmf_dev *dev, u16 event)
 	return 0;
 }
 
-static void amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_result *out)
+static int amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_result *out)
 {
+	struct thermal_cooling_device *cdev = dev->gfx_data.cooling_dev;
+	unsigned long state;
 	u32 val;
 	int idx;
 
@@ -154,8 +156,21 @@ static void amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_
 			dev_dbg(dev->dev, "update SYSTEM_STATE: %s\n",
 				amd_pmf_uevent_as_str(val));
 			break;
+
+		case PMF_POLICY_DISPLAY_BRIGHTNESS:
+			if (!dev->drm_dev)
+				return -ENODEV;
+
+			cdev->ops->get_cur_state(cdev, &state);
+			if (state != val) {
+				cdev->ops->set_cur_state(cdev, val);
+				dev_dbg(dev->dev, "update DISPLAY_BRIGHTNESS: %u\n", val);
+			}
+			break;
 		}
 	}
+
+	return 0;
 }
 
 static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
@@ -192,7 +207,9 @@ static int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
 		amd_pmf_dump_ta_inputs(dev, in);
 		dev_dbg(dev->dev, "action count:%u result:%x\n", out->actions_count,
 			ta_sm->pmf_result);
-		amd_pmf_apply_policies(dev, out);
+		ret = amd_pmf_apply_policies(dev, out);
+		if (ret)
+			return ret;
 	}
 
 	return 0;
@@ -423,6 +440,23 @@ static void amd_pmf_tee_deinit(struct amd_pmf_dev *dev)
 	tee_client_close_context(dev->tee_ctx);
 }
 
+static int amd_pmf_gpu_set_cur_state(struct thermal_cooling_device *cooling_dev,
+				     unsigned long state)
+{
+	struct backlight_device *bd;
+
+	if (acpi_video_get_backlight_type() != acpi_backlight_native)
+		return -ENODEV;
+
+	bd = backlight_device_get_by_type(BACKLIGHT_RAW);
+	if (!bd)
+		return -ENODEV;
+
+	backlight_device_set_brightness(bd, state);
+
+	return 0;
+}
+
 static int amd_pmf_gpu_get_cur_state(struct thermal_cooling_device *cooling_dev,
 				     unsigned long *state)
 {
@@ -463,6 +497,7 @@ static int amd_pmf_gpu_get_max_state(struct thermal_cooling_device *cooling_dev,
 static const struct thermal_cooling_device_ops bd_cooling_ops = {
 	.get_max_state = amd_pmf_gpu_get_max_state,
 	.get_cur_state = amd_pmf_gpu_get_cur_state,
+	.set_cur_state = amd_pmf_gpu_set_cur_state,
 };
 
 static int amd_pmf_get_gpu_handle(struct pci_dev *pdev, void *data)
-- 
2.25.1

[PATCH v5 15/17] HID: amd_sfh: rename float_to_int() to amd_sfh_float_to_int()

[Shyam Sundar S K]

From: Basavaraj Natikar <Basavaraj.Natikar@amd.com>

Current amd_sfh driver has float_to_int() to convert units from
float to int. This is fine until this function gets called outside of
the current scope of file.

Add a prefix "amd_sfh" to float_to_int() so that function represents
the driver name. This function will be called by multiple callers in the
next patch.

Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Co-developed-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
---
 drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_desc.c | 28 ++++++++++---------
 .../amd-sfh-hid/sfh1_1/amd_sfh_interface.h    |  1 +
 2 files changed, 16 insertions(+), 13 deletions(-)

diff --git a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_desc.c b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_desc.c
index 8a037de08e92..33fbdde8aff0 100644
--- a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_desc.c
+++ b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_desc.c
@@ -132,7 +132,7 @@ static void get_common_inputs(struct common_input_property *common, int report_i
 	common->event_type = HID_USAGE_SENSOR_EVENT_DATA_UPDATED_ENUM;
 }
 
-static int float_to_int(u32 flt32_val)
+int amd_sfh_float_to_int(u32 flt32_val)
 {
 	int fraction, shift, mantissa, sign, exp, zeropre;
 
@@ -201,9 +201,9 @@ static u8 get_input_rep(u8 current_index, int sensor_idx, int report_id,
 			     OFFSET_SENSOR_DATA_DEFAULT;
 		memcpy_fromio(&accel_data, sensoraddr, sizeof(struct sfh_accel_data));
 		get_common_inputs(&acc_input.common_property, report_id);
-		acc_input.in_accel_x_value = float_to_int(accel_data.acceldata.x) / 100;
-		acc_input.in_accel_y_value = float_to_int(accel_data.acceldata.y) / 100;
-		acc_input.in_accel_z_value = float_to_int(accel_data.acceldata.z) / 100;
+		acc_input.in_accel_x_value = amd_sfh_float_to_int(accel_data.acceldata.x) / 100;
+		acc_input.in_accel_y_value = amd_sfh_float_to_int(accel_data.acceldata.y) / 100;
+		acc_input.in_accel_z_value = amd_sfh_float_to_int(accel_data.acceldata.z) / 100;
 		memcpy(input_report, &acc_input, sizeof(acc_input));
 		report_size = sizeof(acc_input);
 		break;
@@ -212,9 +212,9 @@ static u8 get_input_rep(u8 current_index, int sensor_idx, int report_id,
 			     OFFSET_SENSOR_DATA_DEFAULT;
 		memcpy_fromio(&gyro_data, sensoraddr, sizeof(struct sfh_gyro_data));
 		get_common_inputs(&gyro_input.common_property, report_id);
-		gyro_input.in_angel_x_value = float_to_int(gyro_data.gyrodata.x) / 1000;
-		gyro_input.in_angel_y_value = float_to_int(gyro_data.gyrodata.y) / 1000;
-		gyro_input.in_angel_z_value = float_to_int(gyro_data.gyrodata.z) / 1000;
+		gyro_input.in_angel_x_value = amd_sfh_float_to_int(gyro_data.gyrodata.x) / 1000;
+		gyro_input.in_angel_y_value = amd_sfh_float_to_int(gyro_data.gyrodata.y) / 1000;
+		gyro_input.in_angel_z_value = amd_sfh_float_to_int(gyro_data.gyrodata.z) / 1000;
 		memcpy(input_report, &gyro_input, sizeof(gyro_input));
 		report_size = sizeof(gyro_input);
 		break;
@@ -223,9 +223,9 @@ static u8 get_input_rep(u8 current_index, int sensor_idx, int report_id,
 			     OFFSET_SENSOR_DATA_DEFAULT;
 		memcpy_fromio(&mag_data, sensoraddr, sizeof(struct sfh_mag_data));
 		get_common_inputs(&magno_input.common_property, report_id);
-		magno_input.in_magno_x = float_to_int(mag_data.magdata.x) / 100;
-		magno_input.in_magno_y = float_to_int(mag_data.magdata.y) / 100;
-		magno_input.in_magno_z = float_to_int(mag_data.magdata.z) / 100;
+		magno_input.in_magno_x = amd_sfh_float_to_int(mag_data.magdata.x) / 100;
+		magno_input.in_magno_y = amd_sfh_float_to_int(mag_data.magdata.y) / 100;
+		magno_input.in_magno_z = amd_sfh_float_to_int(mag_data.magdata.z) / 100;
 		magno_input.in_magno_accuracy = mag_data.accuracy / 100;
 		memcpy(input_report, &magno_input, sizeof(magno_input));
 		report_size = sizeof(magno_input);
@@ -235,13 +235,15 @@ static u8 get_input_rep(u8 current_index, int sensor_idx, int report_id,
 			     OFFSET_SENSOR_DATA_DEFAULT;
 		memcpy_fromio(&als_data, sensoraddr, sizeof(struct sfh_als_data));
 		get_common_inputs(&als_input.common_property, report_id);
-		als_input.illuminance_value = float_to_int(als_data.lux);
+		als_input.illuminance_value = amd_sfh_float_to_int(als_data.lux);
 
 		memcpy_fromio(&binfo, mp2->vsbase, sizeof(struct sfh_base_info));
 		if (binfo.sbase.s_prop[ALS_IDX].sf.feat & 0x2) {
 			als_input.light_color_temp = als_data.light_color_temp;
-			als_input.chromaticity_x_value = float_to_int(als_data.chromaticity_x);
-			als_input.chromaticity_y_value = float_to_int(als_data.chromaticity_y);
+			als_input.chromaticity_x_value =
+				amd_sfh_float_to_int(als_data.chromaticity_x);
+			als_input.chromaticity_y_value =
+				amd_sfh_float_to_int(als_data.chromaticity_y);
 		}
 
 		report_size = sizeof(als_input);
diff --git a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.h b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.h
index 656c3e95ef8c..75267b0fec70 100644
--- a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.h
+++ b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.h
@@ -166,4 +166,5 @@ struct hpd_status {
 
 void sfh_interface_init(struct amd_mp2_dev *mp2);
 void amd_sfh1_1_set_desc_ops(struct amd_mp2_ops *mp2_ops);
+int amd_sfh_float_to_int(u32 flt32_val);
 #endif
-- 
2.25.1

[PATCH v5 16/17] platform/x86/amd/pmf: Add PMF-AMDSFH interface for HPD

[Shyam Sundar S K]

From: Basavaraj Natikar <Basavaraj.Natikar@amd.com>

AMDSFH has information about the User presence information via the Human
Presence Detection (HPD) sensor which is part of the AMD sensor fusion hub.
Add PMF and AMDSFH interface to get this information.

Co-developed-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
---
 drivers/hid/amd-sfh-hid/amd_sfh_common.h      |  5 ++
 drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c | 11 +++++
 .../amd-sfh-hid/sfh1_1/amd_sfh_interface.c    | 28 +++++++++++
 drivers/platform/x86/amd/pmf/Kconfig          |  1 +
 drivers/platform/x86/amd/pmf/spc.c            | 22 +++++++++
 include/linux/amd-pmf-io.h                    | 46 +++++++++++++++++++
 6 files changed, 113 insertions(+)
 create mode 100644 include/linux/amd-pmf-io.h

diff --git a/drivers/hid/amd-sfh-hid/amd_sfh_common.h b/drivers/hid/amd-sfh-hid/amd_sfh_common.h
index 2643bb14fee2..cd57037bf217 100644
--- a/drivers/hid/amd-sfh-hid/amd_sfh_common.h
+++ b/drivers/hid/amd-sfh-hid/amd_sfh_common.h
@@ -37,6 +37,10 @@ struct amd_mp2_sensor_info {
 	dma_addr_t dma_address;
 };
 
+struct sfh_dev_status {
+	bool is_hpd_present;
+};
+
 struct amd_mp2_dev {
 	struct pci_dev *pdev;
 	struct amdtp_cl_data *cl_data;
@@ -47,6 +51,7 @@ struct amd_mp2_dev {
 	struct amd_input_data in_data;
 	/* mp2 active control status */
 	u32 mp2_acs;
+	struct sfh_dev_status dev_en;
 };
 
 struct amd_mp2_ops {
diff --git a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
index e9c6413af24a..3dc652d41d7d 100644
--- a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
+++ b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
@@ -73,6 +73,12 @@ static int amd_sfh_hid_client_deinit(struct amd_mp2_dev *privdata)
 	int i, status;
 
 	for (i = 0; i < cl_data->num_hid_devices; i++) {
+		switch (cl_data->sensor_idx[i]) {
+		case HPD_IDX:
+			privdata->dev_en.is_hpd_present = false;
+			break;
+		}
+
 		if (cl_data->sensor_sts[i] == SENSOR_ENABLED) {
 			privdata->mp2_ops->stop(privdata, cl_data->sensor_idx[i]);
 			status = amd_sfh_wait_for_response
@@ -178,6 +184,11 @@ static int amd_sfh1_1_hid_client_init(struct amd_mp2_dev *privdata)
 			rc = amdtp_hid_probe(i, cl_data);
 			if (rc)
 				goto cleanup;
+			switch (cl_data->sensor_idx[i]) {
+			case HPD_IDX:
+				privdata->dev_en.is_hpd_present = true;
+				break;
+			}
 		}
 		dev_dbg(dev, "sid 0x%x (%s) status 0x%x\n",
 			cl_data->sensor_idx[i], get_sensor_name(cl_data->sensor_idx[i]),
diff --git a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
index 4f81ef2d4f56..c0f31efe4927 100644
--- a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
+++ b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
@@ -7,11 +7,14 @@
  *
  * Author: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
  */
+#include <linux/amd-pmf-io.h>
 #include <linux/io-64-nonatomic-lo-hi.h>
 #include <linux/iopoll.h>
 
 #include "amd_sfh_interface.h"
 
+static struct amd_mp2_dev *emp2;
+
 static int amd_sfh_wait_response(struct amd_mp2_dev *mp2, u8 sid, u32 cmd_id)
 {
 	struct sfh_cmd_response cmd_resp;
@@ -76,4 +79,29 @@ static struct amd_mp2_ops amd_sfh_ops = {
 void sfh_interface_init(struct amd_mp2_dev *mp2)
 {
 	mp2->mp2_ops = &amd_sfh_ops;
+	emp2 = mp2;
+}
+
+static int amd_sfh_hpd_info(u8 *user_present)
+{
+	struct hpd_status hpdstatus;
+
+	if (!emp2 || !emp2->dev_en.is_hpd_present)
+		return -ENODEV;
+
+	hpdstatus.val = readl(emp2->mmio + AMD_C2P_MSG(4));
+	*user_present = hpdstatus.shpd.presence;
+	return 0;
+}
+
+int amd_get_sfh_info(struct amd_sfh_info *sfh_info, enum sfh_message_type op)
+{
+	if (sfh_info) {
+		switch (op) {
+		case MT_HPD:
+			return amd_sfh_hpd_info(&sfh_info->user_present);
+		}
+	}
+	return -EINVAL;
 }
+EXPORT_SYMBOL_GPL(amd_get_sfh_info);
diff --git a/drivers/platform/x86/amd/pmf/Kconfig b/drivers/platform/x86/amd/pmf/Kconfig
index f246252bddd8..f4fa8bd8bda8 100644
--- a/drivers/platform/x86/amd/pmf/Kconfig
+++ b/drivers/platform/x86/amd/pmf/Kconfig
@@ -10,6 +10,7 @@ config AMD_PMF
 	depends on AMD_NB
 	select ACPI_PLATFORM_PROFILE
 	depends on TEE && AMDTEE
+	depends on AMD_SFH_HID
 	help
 	  This driver provides support for the AMD Platform Management Framework.
 	  The goal is to enhance end user experience by making AMD PCs smarter,
diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
index 47ec563088b8..7f63973ab40e 100644
--- a/drivers/platform/x86/amd/pmf/spc.c
+++ b/drivers/platform/x86/amd/pmf/spc.c
@@ -10,6 +10,7 @@
  */
 
 #include <acpi/button.h>
+#include <linux/amd-pmf-io.h>
 #include <linux/power_supply.h>
 #include <linux/units.h>
 #include "pmf.h"
@@ -49,6 +50,7 @@ void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *
 	dev_dbg(dev->dev, "Primary Display State: %s\n", in->ev_info.display_state ?
 			"Connected" : "disconnected/unknown");
 	dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ? "close" : "open");
+	dev_dbg(dev->dev, "User Presence: %s\n", in->ev_info.user_present ? "Present" : "Away");
 	dev_dbg(dev->dev, "==== TA inputs END ====\n");
 }
 #else
@@ -185,6 +187,25 @@ static void amd_pmf_get_gpu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_ta
 	in->ev_info.display_state = dev->gfx_data.con_status[0];
 }
 
+static void amd_pmf_get_sensor_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
+{
+	struct amd_sfh_info sfh_info;
+
+	/* get HPD data */
+	amd_get_sfh_info(&sfh_info, MT_HPD);
+	switch (sfh_info.user_present) {
+	case SFH_NOT_DETECTED:
+		in->ev_info.user_present = 0xff; /* assume no sensors connected */
+		break;
+	case SFH_USER_PRESENT:
+		in->ev_info.user_present = 1;
+		break;
+	case SFH_USER_AWAY:
+		in->ev_info.user_present = 0;
+		break;
+	}
+}
+
 void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
 {
 	/* TA side lid open is 1 and close is 0, hence the ! here */
@@ -195,4 +216,5 @@ void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_tab
 	amd_pmf_get_slider_info(dev, in);
 	if (dev->drm_dev)
 		amd_pmf_get_gpu_info(dev, in);
+	amd_pmf_get_sensor_info(dev, in);
 }
diff --git a/include/linux/amd-pmf-io.h b/include/linux/amd-pmf-io.h
new file mode 100644
index 000000000000..5b6d29d36922
--- /dev/null
+++ b/include/linux/amd-pmf-io.h
@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * AMD Platform Management Framework Interface
+ *
+ * Copyright (c) 2023, Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Authors: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
+ *          Basavaraj Natikar <Basavaraj.Natikar@amd.com>
+ */
+
+#ifndef AMD_PMF_IO_H
+#define AMD_PMF_IO_H
+
+#include <linux/types.h>
+
+/**
+ * enum sfh_message_type - Query the SFH message type
+ * @MT_HPD: Message ID to know the Human presence info from MP2 FW
+ */
+enum sfh_message_type {
+	MT_HPD,
+};
+
+/**
+ * enum sfh_hpd_info - Query the Human presence information
+ * @SFH_NOT_DETECTED: Check the HPD connection information from MP2 FW
+ * @SFH_USER_PRESENT: Check if the user is present from HPD sensor
+ * @SFH_USER_AWAY: Check if the user is away from HPD sensor
+ */
+enum sfh_hpd_info {
+	SFH_NOT_DETECTED,
+	SFH_USER_PRESENT,
+	SFH_USER_AWAY,
+};
+
+/**
+ * struct amd_sfh_info - get HPD sensor info from MP2 FW
+ * @user_present: Populates the user presence information
+ */
+struct amd_sfh_info {
+	u8 user_present;
+};
+
+int amd_get_sfh_info(struct amd_sfh_info *sfh_info, enum sfh_message_type op);
+#endif
-- 
2.25.1

Re: [PATCH v5 16/17] platform/x86/amd/pmf: Add PMF-AMDSFH interface for HPD

[Mario Limonciello]

On 11/17/2023 02:07, Shyam Sundar S K wrote:
> From: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
> 
> AMDSFH has information about the User presence information via the Human
> Presence Detection (HPD) sensor which is part of the AMD sensor fusion hub.
> Add PMF and AMDSFH interface to get this information.
> 
> Co-developed-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
> ---
>   drivers/hid/amd-sfh-hid/amd_sfh_common.h      |  5 ++
>   drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c | 11 +++++
>   .../amd-sfh-hid/sfh1_1/amd_sfh_interface.c    | 28 +++++++++++
>   drivers/platform/x86/amd/pmf/Kconfig          |  1 +
>   drivers/platform/x86/amd/pmf/spc.c            | 22 +++++++++
>   include/linux/amd-pmf-io.h                    | 46 +++++++++++++++++++
>   6 files changed, 113 insertions(+)
>   create mode 100644 include/linux/amd-pmf-io.h
> 
> diff --git a/drivers/hid/amd-sfh-hid/amd_sfh_common.h b/drivers/hid/amd-sfh-hid/amd_sfh_common.h
> index 2643bb14fee2..cd57037bf217 100644
> --- a/drivers/hid/amd-sfh-hid/amd_sfh_common.h
> +++ b/drivers/hid/amd-sfh-hid/amd_sfh_common.h
> @@ -37,6 +37,10 @@ struct amd_mp2_sensor_info {
>   	dma_addr_t dma_address;
>   };
>   
> +struct sfh_dev_status {
> +	bool is_hpd_present;
> +};
> +
>   struct amd_mp2_dev {
>   	struct pci_dev *pdev;
>   	struct amdtp_cl_data *cl_data;
> @@ -47,6 +51,7 @@ struct amd_mp2_dev {
>   	struct amd_input_data in_data;
>   	/* mp2 active control status */
>   	u32 mp2_acs;
> +	struct sfh_dev_status dev_en;
>   };
>   
>   struct amd_mp2_ops {
> diff --git a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
> index e9c6413af24a..3dc652d41d7d 100644
> --- a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
> +++ b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
> @@ -73,6 +73,12 @@ static int amd_sfh_hid_client_deinit(struct amd_mp2_dev *privdata)
>   	int i, status;
>   
>   	for (i = 0; i < cl_data->num_hid_devices; i++) {
> +		switch (cl_data->sensor_idx[i]) {
> +		case HPD_IDX:
> +			privdata->dev_en.is_hpd_present = false;
> +			break;
> +		}
> +
>   		if (cl_data->sensor_sts[i] == SENSOR_ENABLED) {
>   			privdata->mp2_ops->stop(privdata, cl_data->sensor_idx[i]);
>   			status = amd_sfh_wait_for_response
> @@ -178,6 +184,11 @@ static int amd_sfh1_1_hid_client_init(struct amd_mp2_dev *privdata)
>   			rc = amdtp_hid_probe(i, cl_data);
>   			if (rc)
>   				goto cleanup;
> +			switch (cl_data->sensor_idx[i]) {
> +			case HPD_IDX:
> +				privdata->dev_en.is_hpd_present = true;
> +				break;
> +			}
>   		}
>   		dev_dbg(dev, "sid 0x%x (%s) status 0x%x\n",
>   			cl_data->sensor_idx[i], get_sensor_name(cl_data->sensor_idx[i]),
> diff --git a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
> index 4f81ef2d4f56..c0f31efe4927 100644
> --- a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
> +++ b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
> @@ -7,11 +7,14 @@
>    *
>    * Author: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
>    */
> +#include <linux/amd-pmf-io.h>
>   #include <linux/io-64-nonatomic-lo-hi.h>
>   #include <linux/iopoll.h>
>   
>   #include "amd_sfh_interface.h"
>   
> +static struct amd_mp2_dev *emp2;
> +
>   static int amd_sfh_wait_response(struct amd_mp2_dev *mp2, u8 sid, u32 cmd_id)
>   {
>   	struct sfh_cmd_response cmd_resp;
> @@ -76,4 +79,29 @@ static struct amd_mp2_ops amd_sfh_ops = {
>   void sfh_interface_init(struct amd_mp2_dev *mp2)
>   {
>   	mp2->mp2_ops = &amd_sfh_ops;
> +	emp2 = mp2;

How about when SFH is destroyed?  I think you want to explicitly set 
emp2 to NULL;

> +}
> +
> +static int amd_sfh_hpd_info(u8 *user_present)
> +{
> +	struct hpd_status hpdstatus;
> +
> +	if (!emp2 || !emp2->dev_en.is_hpd_present)
> +		return -ENODEV;
> +
> +	hpdstatus.val = readl(emp2->mmio + AMD_C2P_MSG(4));
> +	*user_present = hpdstatus.shpd.presence;
> +	return 0;
> +}
> +
> +int amd_get_sfh_info(struct amd_sfh_info *sfh_info, enum sfh_message_type op)
> +{
> +	if (sfh_info) {
> +		switch (op) {
> +		case MT_HPD:
> +			return amd_sfh_hpd_info(&sfh_info->user_present);
> +		}
> +	}
> +	return -EINVAL;
>   }
> +EXPORT_SYMBOL_GPL(amd_get_sfh_info);
> diff --git a/drivers/platform/x86/amd/pmf/Kconfig b/drivers/platform/x86/amd/pmf/Kconfig
> index f246252bddd8..f4fa8bd8bda8 100644
> --- a/drivers/platform/x86/amd/pmf/Kconfig
> +++ b/drivers/platform/x86/amd/pmf/Kconfig
> @@ -10,6 +10,7 @@ config AMD_PMF
>   	depends on AMD_NB
>   	select ACPI_PLATFORM_PROFILE
>   	depends on TEE && AMDTEE
> +	depends on AMD_SFH_HID
>   	help
>   	  This driver provides support for the AMD Platform Management Framework.
>   	  The goal is to enhance end user experience by making AMD PCs smarter,
> diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
> index 47ec563088b8..7f63973ab40e 100644
> --- a/drivers/platform/x86/amd/pmf/spc.c
> +++ b/drivers/platform/x86/amd/pmf/spc.c
> @@ -10,6 +10,7 @@
>    */
>   
>   #include <acpi/button.h>
> +#include <linux/amd-pmf-io.h>
>   #include <linux/power_supply.h>
>   #include <linux/units.h>
>   #include "pmf.h"
> @@ -49,6 +50,7 @@ void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *
>   	dev_dbg(dev->dev, "Primary Display State: %s\n", in->ev_info.display_state ?
>   			"Connected" : "disconnected/unknown");
>   	dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ? "close" : "open");
> +	dev_dbg(dev->dev, "User Presence: %s\n", in->ev_info.user_present ? "Present" : "Away");
>   	dev_dbg(dev->dev, "==== TA inputs END ====\n");
>   }
>   #else
> @@ -185,6 +187,25 @@ static void amd_pmf_get_gpu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_ta
>   	in->ev_info.display_state = dev->gfx_data.con_status[0];
>   }
>   
> +static void amd_pmf_get_sensor_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
> +{
> +	struct amd_sfh_info sfh_info;
> +
> +	/* get HPD data */
> +	amd_get_sfh_info(&sfh_info, MT_HPD);
> +	switch (sfh_info.user_present) {
> +	case SFH_NOT_DETECTED:
> +		in->ev_info.user_present = 0xff; /* assume no sensors connected */
> +		break;
> +	case SFH_USER_PRESENT:
> +		in->ev_info.user_present = 1;
> +		break;
> +	case SFH_USER_AWAY:
> +		in->ev_info.user_present = 0;
> +		break;
> +	}
> +}
> +
>   void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
>   {
>   	/* TA side lid open is 1 and close is 0, hence the ! here */
> @@ -195,4 +216,5 @@ void amd_pmf_populate_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_tab
>   	amd_pmf_get_slider_info(dev, in);
>   	if (dev->drm_dev)
>   		amd_pmf_get_gpu_info(dev, in);
> +	amd_pmf_get_sensor_info(dev, in);
>   }
> diff --git a/include/linux/amd-pmf-io.h b/include/linux/amd-pmf-io.h
> new file mode 100644
> index 000000000000..5b6d29d36922
> --- /dev/null
> +++ b/include/linux/amd-pmf-io.h
> @@ -0,0 +1,46 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/*
> + * AMD Platform Management Framework Interface
> + *
> + * Copyright (c) 2023, Advanced Micro Devices, Inc.
> + * All Rights Reserved.
> + *
> + * Authors: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> + *          Basavaraj Natikar <Basavaraj.Natikar@amd.com>
> + */
> +
> +#ifndef AMD_PMF_IO_H
> +#define AMD_PMF_IO_H
> +
> +#include <linux/types.h>
> +
> +/**
> + * enum sfh_message_type - Query the SFH message type
> + * @MT_HPD: Message ID to know the Human presence info from MP2 FW
> + */
> +enum sfh_message_type {
> +	MT_HPD,
> +};
> +
> +/**
> + * enum sfh_hpd_info - Query the Human presence information
> + * @SFH_NOT_DETECTED: Check the HPD connection information from MP2 FW
> + * @SFH_USER_PRESENT: Check if the user is present from HPD sensor
> + * @SFH_USER_AWAY: Check if the user is away from HPD sensor
> + */
> +enum sfh_hpd_info {
> +	SFH_NOT_DETECTED,
> +	SFH_USER_PRESENT,
> +	SFH_USER_AWAY,
> +};
> +
> +/**
> + * struct amd_sfh_info - get HPD sensor info from MP2 FW
> + * @user_present: Populates the user presence information
> + */
> +struct amd_sfh_info {
> +	u8 user_present;
> +};
> +
> +int amd_get_sfh_info(struct amd_sfh_info *sfh_info, enum sfh_message_type op);
> +#endif

[PATCH v5 17/17] platform/x86/amd/pmf: Add PMF-AMDSFH interface for ALS

[Shyam Sundar S K]

From: Basavaraj Natikar <Basavaraj.Natikar@amd.com>

AMDSFH has information about the Ambient light via the Ambient
Light Sensor (ALS) which is part of the AMD sensor fusion hub.
Add PMF and AMDSFH interface to get this information.

Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Co-developed-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
---
 drivers/hid/amd-sfh-hid/amd_sfh_common.h      |  1 +
 drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c |  6 ++++++
 .../amd-sfh-hid/sfh1_1/amd_sfh_interface.c    | 19 +++++++++++++++++++
 drivers/platform/x86/amd/pmf/spc.c            |  9 ++++++++-
 include/linux/amd-pmf-io.h                    |  4 ++++
 5 files changed, 38 insertions(+), 1 deletion(-)

diff --git a/drivers/hid/amd-sfh-hid/amd_sfh_common.h b/drivers/hid/amd-sfh-hid/amd_sfh_common.h
index cd57037bf217..a1950bc6e6ce 100644
--- a/drivers/hid/amd-sfh-hid/amd_sfh_common.h
+++ b/drivers/hid/amd-sfh-hid/amd_sfh_common.h
@@ -39,6 +39,7 @@ struct amd_mp2_sensor_info {
 
 struct sfh_dev_status {
 	bool is_hpd_present;
+	bool is_als_present;
 };
 
 struct amd_mp2_dev {
diff --git a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
index 3dc652d41d7d..f2890d329459 100644
--- a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
+++ b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c
@@ -77,6 +77,9 @@ static int amd_sfh_hid_client_deinit(struct amd_mp2_dev *privdata)
 		case HPD_IDX:
 			privdata->dev_en.is_hpd_present = false;
 			break;
+		case ALS_IDX:
+			privdata->dev_en.is_als_present = false;
+			break;
 		}
 
 		if (cl_data->sensor_sts[i] == SENSOR_ENABLED) {
@@ -188,6 +191,9 @@ static int amd_sfh1_1_hid_client_init(struct amd_mp2_dev *privdata)
 			case HPD_IDX:
 				privdata->dev_en.is_hpd_present = true;
 				break;
+			case ALS_IDX:
+				privdata->dev_en.is_als_present = true;
+				break;
 			}
 		}
 		dev_dbg(dev, "sid 0x%x (%s) status 0x%x\n",
diff --git a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
index c0f31efe4927..0ae087d5bc44 100644
--- a/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
+++ b/drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_interface.c
@@ -94,12 +94,31 @@ static int amd_sfh_hpd_info(u8 *user_present)
 	return 0;
 }
 
+static int amd_sfh_als_info(u32 *ambient_light)
+{
+	struct sfh_als_data als_data;
+	void __iomem *sensoraddr;
+
+	if (!emp2 || !emp2->dev_en.is_als_present)
+		return -ENODEV;
+
+	sensoraddr = emp2->vsbase +
+		(ALS_IDX * SENSOR_DATA_MEM_SIZE_DEFAULT) +
+		OFFSET_SENSOR_DATA_DEFAULT;
+	memcpy_fromio(&als_data, sensoraddr, sizeof(struct sfh_als_data));
+	*ambient_light = amd_sfh_float_to_int(als_data.lux);
+
+	return 0;
+}
+
 int amd_get_sfh_info(struct amd_sfh_info *sfh_info, enum sfh_message_type op)
 {
 	if (sfh_info) {
 		switch (op) {
 		case MT_HPD:
 			return amd_sfh_hpd_info(&sfh_info->user_present);
+		case MT_ALS:
+			return amd_sfh_als_info(&sfh_info->ambient_light);
 		}
 	}
 	return -EINVAL;
diff --git a/drivers/platform/x86/amd/pmf/spc.c b/drivers/platform/x86/amd/pmf/spc.c
index 7f63973ab40e..07f6bd238aa2 100644
--- a/drivers/platform/x86/amd/pmf/spc.c
+++ b/drivers/platform/x86/amd/pmf/spc.c
@@ -51,6 +51,7 @@ void amd_pmf_dump_ta_inputs(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *
 			"Connected" : "disconnected/unknown");
 	dev_dbg(dev->dev, "LID State: %s\n", in->ev_info.lid_state ? "close" : "open");
 	dev_dbg(dev->dev, "User Presence: %s\n", in->ev_info.user_present ? "Present" : "Away");
+	dev_dbg(dev->dev, "Ambient Light: %d\n", in->ev_info.ambient_light);
 	dev_dbg(dev->dev, "==== TA inputs END ====\n");
 }
 #else
@@ -190,8 +191,14 @@ static void amd_pmf_get_gpu_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_ta
 static void amd_pmf_get_sensor_info(struct amd_pmf_dev *dev, struct ta_pmf_enact_table *in)
 {
 	struct amd_sfh_info sfh_info;
+	int ret;
 
-	/* get HPD data */
+	/* Get ALS data */
+	ret = amd_get_sfh_info(&sfh_info, MT_ALS);
+	if (!ret)
+		in->ev_info.ambient_light = sfh_info.ambient_light;
+
+	/* Get HPD data */
 	amd_get_sfh_info(&sfh_info, MT_HPD);
 	switch (sfh_info.user_present) {
 	case SFH_NOT_DETECTED:
diff --git a/include/linux/amd-pmf-io.h b/include/linux/amd-pmf-io.h
index 5b6d29d36922..b4f818205216 100644
--- a/include/linux/amd-pmf-io.h
+++ b/include/linux/amd-pmf-io.h
@@ -17,9 +17,11 @@
 /**
  * enum sfh_message_type - Query the SFH message type
  * @MT_HPD: Message ID to know the Human presence info from MP2 FW
+ * @MT_ALS: Message ID to know the Ambient light info from MP2 FW
  */
 enum sfh_message_type {
 	MT_HPD,
+	MT_ALS,
 };
 
 /**
@@ -36,9 +38,11 @@ enum sfh_hpd_info {
 
 /**
  * struct amd_sfh_info - get HPD sensor info from MP2 FW
+ * @ambient_light: Populates the ambient light information
  * @user_present: Populates the user presence information
  */
 struct amd_sfh_info {
+	u32 ambient_light;
 	u8 user_present;
 };
 
-- 
2.25.1