arm/arm64: Unify PSCI client support

We make use of the PSCI function IDs, but don't explicitly include the header which defines them. Relying on transitive header includes is fragile and will be broken as headers are refactored.

This patch includes the relevant header file directly so as to avoid future breakage.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Christoffer Dall <[email protected]>

Cc Marc Zyngier <[email protected]>

---
  arch/arm/kvm/psci.c | 2 ++
  1 file changed, 2 insertions(+)

arch/arm/kvm/psci.c +0 -0

index 02fa8ef..7e9398c 100644

The PSCI MIGRATE_INFO_UP_CPU call returns a physical ID, which we will need to map back to a Linux logical ID.

Implement a reusable get_logical_index to map from a physical ID to a logical ID.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

---
 arch/arm64/include/asm/smp_plat.h | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

arch/arm64/include/asm/smp_plat.h +6 -0

index 8dcd61e..7abf757 100644
@@ -19,6 +19,8 @@
 #ifndef __ASM_SMP_PLAT_H
 #define __ASM_SMP_PLAT_H
 
+#include <linux/cpumask.h>
+
 #include <asm/types.h>
 
 struct mpidr_hash {

@@ -39,6 +41,20 @@ static inline u32 mpidr_hash_size(void)
  */
 extern u64 __cpu_logical_map[NR_CPUS];
 #define cpu_logical_map(cpu)    __cpu_logical_map[cpu]
+/*
+ * Retrieve logical cpu index corresponding to a given MPIDR.Aff*
+ *  - mpidr: MPIDR.Aff* bits to be used for the look-up
+ *

Quoted 40 lines · Mark Rutland

The PSCI MIGRATE_INFO_UP_CPU call returns a physical ID, which we will need to map back to a Linux logical ID.

Implement a reusable get_logical_index to map from a physical ID to a logical ID.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

---
 arch/arm64/include/asm/smp_plat.h | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

arch/arm64/include/asm/smp_plat.h +11 -0

index 8dcd61e..7abf757 100644
@@ -19,6 +19,8 @@
 #ifndef __ASM_SMP_PLAT_H
 #define __ASM_SMP_PLAT_H
 
+#include <linux/cpumask.h>
+
 #include <asm/types.h>
 
 struct mpidr_hash {

@@ -39,6 +41,20 @@ static inline u32 mpidr_hash_size(void)
  */
 extern u64 __cpu_logical_map[NR_CPUS];
 #define cpu_logical_map(cpu)    __cpu_logical_map[cpu]
+/*
+ * Retrieve logical cpu index corresponding to a given MPIDR.Aff*
+ *  - mpidr: MPIDR.Aff* bits to be used for the look-up
+ *
+ * Returns the cpu logical index or -EINVAL on look-up error
+ */
+static inline int get_logical_index(u64 mpidr)
+{
+	int cpu;

arch/arm64/include/asm/smp_plat.h static inline u32 mpidr_hash_size(void) · line 57

Quoted 5 lines

arch/arm64/include/asm/smp_plat.h +5 -0

+	for (cpu = 0; cpu < nr_cpu_ids; cpu++)
+		if (cpu_logical_map(cpu) == mpidr)
+			return cpu;
+	return -EINVAL;
+}

Do we guarantee that mpidr here only contains the affinity bits (i.e. no masking)?

Quoted 41 lines · Catalin Marinas

Quoted 40 lines · Mark Rutland

The PSCI MIGRATE_INFO_UP_CPU call returns a physical ID, which we will need to map back to a Linux logical ID.

Implement a reusable get_logical_index to map from a physical ID to a logical ID.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

---
  arch/arm64/include/asm/smp_plat.h | 16 ++++++++++++++++
  1 file changed, 16 insertions(+)

arch/arm64/include/asm/smp_plat.h +11 -0

index 8dcd61e..7abf757 100644
@@ -19,6 +19,8 @@
  #ifndef __ASM_SMP_PLAT_H
  #define __ASM_SMP_PLAT_H
 
+#include <linux/cpumask.h>
+
  #include <asm/types.h>
 
  struct mpidr_hash {

@@ -39,6 +41,20 @@ static inline u32 mpidr_hash_size(void)
   */
  extern u64 __cpu_logical_map[NR_CPUS];
  #define cpu_logical_map(cpu)    __cpu_logical_map[cpu]
+/*
+ * Retrieve logical cpu index corresponding to a given MPIDR.Aff*
+ *  - mpidr: MPIDR.Aff* bits to be used for the look-up
+ *
+ * Returns the cpu logical index or -EINVAL on look-up error
+ */
+static inline int get_logical_index(u64 mpidr)
+{
+	int cpu;

Quoted 8 lines

arch/arm64/include/asm/smp_plat.h static inline u32 mpidr_hash_size(void) · line 57

Quoted 5 lines

arch/arm64/include/asm/smp_plat.h +5 -0

+	for (cpu = 0; cpu < nr_cpu_ids; cpu++)
+		if (cpu_logical_map(cpu) == mpidr)
+			return cpu;
+	return -EINVAL;
+}

Do we guarantee that mpidr here only contains the affinity bits (i.e. no masking)?

In the DT and ACPI probe paths we reject IDs with bits other than the aff bits set.

Quoted 41 lines · Catalin Marinas

Quoted 40 lines · Mark Rutland

The PSCI MIGRATE_INFO_UP_CPU call returns a physical ID, which we will need to map back to a Linux logical ID.

Implement a reusable get_logical_index to map from a physical ID to a logical ID.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

---
 arch/arm64/include/asm/smp_plat.h | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

arch/arm64/include/asm/smp_plat.h +11 -0

index 8dcd61e..7abf757 100644
@@ -19,6 +19,8 @@
 #ifndef __ASM_SMP_PLAT_H
 #define __ASM_SMP_PLAT_H
 
+#include <linux/cpumask.h>
+
 #include <asm/types.h>
 
 struct mpidr_hash {

@@ -39,6 +41,20 @@ static inline u32 mpidr_hash_size(void)
  */
 extern u64 __cpu_logical_map[NR_CPUS];
 #define cpu_logical_map(cpu)    __cpu_logical_map[cpu]
+/*
+ * Retrieve logical cpu index corresponding to a given MPIDR.Aff*
+ *  - mpidr: MPIDR.Aff* bits to be used for the look-up
+ *
+ * Returns the cpu logical index or -EINVAL on look-up error
+ */
+static inline int get_logical_index(u64 mpidr)
+{
+	int cpu;

Quoted 8 lines

arch/arm64/include/asm/smp_plat.h static inline u32 mpidr_hash_size(void) · line 57

Quoted 5 lines

arch/arm64/include/asm/smp_plat.h +5 -0

+	for (cpu = 0; cpu < nr_cpu_ids; cpu++)
+		if (cpu_logical_map(cpu) == mpidr)
+			return cpu;
+	return -EINVAL;
+}

Do we guarantee that mpidr here only contains the affinity bits (i.e. no masking)?

In the DT and ACPI probe paths we reject IDs with bits other than the aff bits set.

I don't currently have a sanity check in the code handling MIGRATE_INFO_UP_CPU, but I should add one.

Would you like a warning here too?

Quoted 21 lines

Quoted 9 lines

Quoted 9 lines

arch/arm64/include/asm/smp_plat.h +9 -0

+/*
+ * Retrieve logical cpu index corresponding to a given MPIDR.Aff*
+ *  - mpidr: MPIDR.Aff* bits to be used for the look-up
+ *
+ * Returns the cpu logical index or -EINVAL on look-up error
+ */
+static inline int get_logical_index(u64 mpidr)
+{
+	int cpu;

Quoted 8 lines

arch/arm64/include/asm/smp_plat.h static inline u32 mpidr_hash_size(void) · line 57

Quoted 5 lines

arch/arm64/include/asm/smp_plat.h +5 -0

+	for (cpu = 0; cpu < nr_cpu_ids; cpu++)
+		if (cpu_logical_map(cpu) == mpidr)
+			return cpu;
+	return -EINVAL;
+}

Do we guarantee that mpidr here only contains the affinity bits (i.e. no masking)?

In the DT and ACPI probe paths we reject IDs with bits other than the aff bits set.

Quoted 4 lines

I don't currently have a sanity check in the code handling MIGRATE_INFO_UP_CPU, but I should add one.

Would you like a warning here too?

Or maybe just the warning if the code paths to this function already mask out the non-affinity bits.

cpu_kill currently returns one for success and zero for failure, which is unlike all the other cpu_operations, which return zero for success and an error code upon failure. This difference is unnecessarily confusing.

Make cpu_kill consistent with the other cpu_operations.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

cpu_kill currently returns one for success and zero for failure, which is unlike all the other cpu_operations, which return zero for success and an error code upon failure. This difference is unnecessarily confusing.

Make cpu_kill consistent with the other cpu_operations.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

---
  arch/arm64/kernel/psci.c |  7 +++----
  arch/arm64/kernel/smp.c  | 10 +++++++---
  2 files changed, 10 insertions(+), 7 deletions(-)

arch/arm64/kernel/psci.c +3 -4

index ea18cb5..10fa25e 100644
@@ -498,7 +498,7 @@ static int cpu_psci_cpu_kill(unsigned int cpu)
  	int err, i;
 
  	if (!psci_ops.affinity_info)
-		return 1;
+		return 0;
  	/*
  	 * cpu_kill could race with cpu_die and we can
  	 * potentially end up declaring this cpu undead

@@ -509,7 +509,7 @@ static int cpu_psci_cpu_kill(unsigned int cpu)
  		err = psci_ops.affinity_info(cpu_logical_map(cpu), 0);
  		if (err == PSCI_0_2_AFFINITY_LEVEL_OFF) {
  			pr_info("CPU%d killed.\n", cpu);
-			return 1;
+			return 0;
  		}
 
  		msleep(10);

@@ -518,8 +518,7 @@ static int cpu_psci_cpu_kill(unsigned int cpu)
 
  	pr_warn("CPU%d may not have shut down cleanly (AFFINITY_INFO reports %d)\n",
  			cpu, err);
-	/* Make op_cpu_kill() fail. */
-	return 0;
+	return -ETIMEDOUT;
  }
  #endif
  #endif

arch/arm64/kernel/smp.c +3 -1

index 2cb0081..3799df2 100644
@@ -248,7 +248,7 @@ static int op_cpu_kill(unsigned int cpu)
  	 * time and hope that it's dead, so let's skip the wait and just hope.
  	 */
  	if (!cpu_ops[cpu]->cpu_kill)
-		return 1;
+		return 0;
 
  	return cpu_ops[cpu]->cpu_kill(cpu);
  }

@@ -261,6 +261,8 @@ static DECLARE_COMPLETION(cpu_died);
   */
  void __cpu_die(unsigned int cpu)
  {
+	int err;
+
  	if (!wait_for_completion_timeout(&cpu_died, msecs_to_jiffies(5000))) {
  		pr_crit("CPU%u: cpu didn't die\n", cpu);
  		return;

PSCI 0.1 did not define canonical IDs for CPU_ON, CPU_OFF, CPU_SUSPEND, or MIGRATE, and so these need to be provided when using firmware compliant to PSCI 0.1.

However, functions introduced in 0.2 or later have canonical IDs, and these cannot be provided via DT. There's no need to indirect the IDs via a table; they can be used directly at callsites (and already are for SYSTEM_OFF and SYSTEM_RESET).

This patch removes the unnecessary function ID indirection for AFFINITY_INFO and MIGRATE_INFO_TYPE.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

PSCI 0.1 did not define canonical IDs for CPU_ON, CPU_OFF, CPU_SUSPEND, or MIGRATE, and so these need to be provided when using firmware compliant to PSCI 0.1.

However, functions introduced in 0.2 or later have canonical IDs, and these cannot be provided via DT. There's no need to indirect the IDs via a table; they can be used directly at callsites (and already are for SYSTEM_OFF and SYSTEM_RESET).

This patch removes the unnecessary function ID indirection for AFFINITY_INFO and MIGRATE_INFO_TYPE.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

---
  arch/arm64/kernel/psci.c | 20 +++-----------------
  1 file changed, 3 insertions(+), 17 deletions(-)

arch/arm64/kernel/psci.c +3 -14

index 10fa25e..66b5b75 100644
@@ -67,8 +67,6 @@ enum psci_function {
  	PSCI_FN_CPU_ON,
  	PSCI_FN_CPU_OFF,
  	PSCI_FN_MIGRATE,
-	PSCI_FN_AFFINITY_INFO,
-	PSCI_FN_MIGRATE_INFO_TYPE,
  	PSCI_FN_MAX,
  };
 

@@ -168,22 +166,13 @@ static int psci_migrate(unsigned long cpuid)
  static int psci_affinity_info(unsigned long target_affinity,
  		unsigned long lowest_affinity_level)
  {
-	int err;
-	u32 fn;
-
-	fn = psci_function_id[PSCI_FN_AFFINITY_INFO];
-	err = invoke_psci_fn(fn, target_affinity, lowest_affinity_level, 0);
-	return err;
+	return invoke_psci_fn(PSCI_0_2_FN64_AFFINITY_INFO, target_affinity,
+			      lowest_affinity_level, 0);
  }
 
  static int psci_migrate_info_type(void)
  {
-	int err;
-	u32 fn;
-
-	fn = psci_function_id[PSCI_FN_MIGRATE_INFO_TYPE];
-	err = invoke_psci_fn(fn, 0, 0, 0);
-	return err;
+	return invoke_psci_fn(PSCI_0_2_FN_MIGRATE_INFO_TYPE, 0, 0, 0);
  }
 
  static int __maybe_unused cpu_psci_cpu_init_idle(struct device_node *cpu_node,

@@ -290,11 +279,8 @@ static void __init psci_0_2_set_functions(void)
  	psci_function_id[PSCI_FN_MIGRATE] = PSCI_0_2_FN64_MIGRATE;

PSCI_VERSION and MIGRATE_INFO_TYPE_UP_CPU return unsigned values, with the latter returning a 64-bit value. However, the PSCI invocation functions have prototypes returning int.

This patch upgrades the invocation functions to return unsigned long, with a new typedef to keep things legible. As PSCI_VERSION cannot return a negative value, the erroneous check against PSCI_RET_NOT_SUPPORTED is also removed. The unrelated psci_initcall_t typedef is moved closer to its first user, to avoid confusion with the invocation functions.

In preparation for sharing the code with ARM, unsigned long is used in preference of u64. In the SMC32 calling convention, the relevant fields will be 32 bits wide.

Signed-off-by Mark Rutland <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Will Deacon <[email protected]>

PSCI_VERSION and MIGRATE_INFO_TYPE_UP_CPU return unsigned values, with the latter returning a 64-bit value. However, the PSCI invocation functions have prototypes returning int.

This patch upgrades the invocation functions to return unsigned long, with a new typedef to keep things legible. As PSCI_VERSION cannot return a negative value, the erroneous check against PSCI_RET_NOT_SUPPORTED is also removed. The unrelated psci_initcall_t typedef is moved closer to its first user, to avoid confusion with the invocation functions.

In preparation for sharing the code with ARM, unsigned long is used in preference of u64. In the SMC32 calling convention, the relevant fields will be 32 bits wide.

Signed-off-by Mark Rutland <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Will Deacon <[email protected]>

---
  arch/arm64/kernel/psci.c | 47 ++++++++++++++++++-----------------------------
  1 file changed, 18 insertions(+), 29 deletions(-)

arch/arm64/kernel/psci.c +7 -28

index 66b5b75..9c4de4b 100644
@@ -56,11 +56,11 @@ struct psci_operations {
 
  static struct psci_operations psci_ops;
 
-static int (*invoke_psci_fn)(u64, u64, u64, u64);
-typedef int (*psci_initcall_t)(const struct device_node *);
-
-asmlinkage int __invoke_psci_fn_hvc(u64, u64, u64, u64);
-asmlinkage int __invoke_psci_fn_smc(u64, u64, u64, u64);
+typedef unsigned long (psci_fn)(unsigned long, unsigned long,
+				unsigned long, unsigned long);
+asmlinkage psci_fn __invoke_psci_fn_hvc;
+asmlinkage psci_fn __invoke_psci_fn_smc;
+static psci_fn *invoke_psci_fn;
 
  enum psci_function {
  	PSCI_FN_CPU_SUSPEND,

@@ -112,12 +112,9 @@ static void psci_power_state_unpack(u32 power_state,
  			PSCI_0_2_POWER_STATE_AFFL_SHIFT;
  }
 
-static int psci_get_version(void)
+static u32 psci_get_version(void)
  {
-	int err;
-
-	err = invoke_psci_fn(PSCI_0_2_FN_PSCI_VERSION, 0, 0, 0);
-	return err;
+	return invoke_psci_fn(PSCI_0_2_FN_PSCI_VERSION, 0, 0, 0);
  }
 
  static int psci_cpu_suspend(struct psci_power_state state,

@@ -293,25 +290,15 @@ static void __init psci_0_2_set_functions(void)
   */
  static int __init psci_probe(void)
  {
-	int ver = psci_get_version();
-
-	if (ver == PSCI_RET_NOT_SUPPORTED) {
-		/*
-		 * PSCI versions >=0.2 mandates implementation of
-		 * PSCI_VERSION.
-		 */
-		pr_err("PSCI firmware does not comply with the v0.2 spec.\n");
-		return -EOPNOTSUPP;
-	} else {
-		pr_info("PSCIv%d.%d detected in firmware.\n",
-				PSCI_VERSION_MAJOR(ver),
-				PSCI_VERSION_MINOR(ver));
-
-		if (PSCI_VERSION_MAJOR(ver) == 0 &&
-				PSCI_VERSION_MINOR(ver) < 2) {
-			pr_err("Conflicting PSCI version detected.\n");
-			return -EINVAL;

Software resident in the secure world (a "Trusted OS") may cause CPU_OFF calls for the CPU it is resident on to be denied. Such a denial would be fatal for the kernel, and so we must detect when this can happen before the point of no return.

This patch implements Trusted OS detection for PSCI 0.2+ systems, using MIGRATE_INFO_TYPE and MIGRATE_INFO_UP_CPU. When a trusted OS is detected as resident on a particular CPU, attempts to hot unplug that CPU will be denied early, before they can prove fatal.

Trusted OS migration is not implemented by this patch. Implementation of migratable UP trusted OSs seems unlikely, and the right policy for migration is unclear (and will likely differ across implementations). As such, it is likely that migration will require cooperation with Trusted OS drivers.

PSCI implementations prior to 0.1 do not provide the facility to detect the presence of a Trusted OS, nor the CPU any such OS is resident on, so without additional information it is not possible to handle Trusted OSs with PSCI 0.1.

Signed-off-by Mark Rutland <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Will Deacon <[email protected]>

A PSCI 1.0 implementation may choose to use the new extended StateID format, the presence of which may be queried via the PSCI_FEATURES call. The layout of this new StateID format is incompatible with the existing format, and so to handle both we must abstract attempts to parse the fields.

In preparation for PSCI 1.0 support, this patch introduces psci_power_state_loses_context and psci_power_state_is_valid functions to query information from a PSCI power state, which is no longer decomposed (and hence the pack/unpack functions are removed). As it is no longer decomposed, it is now passed round as an opaque u32 token.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

---
 arch/arm64/kernel/psci.c | 89 ++++++++++++++++++++----------------------------
 1 file changed, 37 insertions(+), 52 deletions(-)

arch/arm64/kernel/psci.c +0 -0

index 3ef677e..4ef9472 100644
@@ -37,11 +37,19 @@
 #define PSCI_POWER_STATE_TYPE_STANDBY		0
 #define PSCI_POWER_STATE_TYPE_POWER_DOWN	1

Quoted 28 lines · Mark Rutland

A PSCI 1.0 implementation may choose to use the new extended StateID format, the presence of which may be queried via the PSCI_FEATURES call. The layout of this new StateID format is incompatible with the existing format, and so to handle both we must abstract attempts to parse the fields.

In preparation for PSCI 1.0 support, this patch introduces psci_power_state_loses_context and psci_power_state_is_valid functions to query information from a PSCI power state, which is no longer decomposed (and hence the pack/unpack functions are removed). As it is no longer decomposed, it is now passed round as an opaque u32 token.

Signed-off-by Mark Rutland <[email protected]>

Acked-by Lorenzo Pieralisi <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Will Deacon <[email protected]>

---
 arch/arm64/kernel/psci.c | 89 ++++++++++++++++++++----------------------------
 1 file changed, 37 insertions(+), 52 deletions(-)

arch/arm64/kernel/psci.c +0 -0

index 3ef677e..4ef9472 100644
@@ -37,11 +37,19 @@
 #define PSCI_POWER_STATE_TYPE_STANDBY		0
 #define PSCI_POWER_STATE_TYPE_POWER_DOWN	1
 

arch/arm64/kernel/psci.c line 43

Quoted 9 lines

arch/arm64/kernel/psci.c +4 -5

-struct psci_power_state {
-	u16	id;
-	u8	type;
-	u8	affinity_level;
-};
+static bool psci_power_state_loses_context(u32 state)
+{
+	return !!(state & PSCI_0_2_POWER_STATE_TYPE_MASK);
+}

Nitpick: I don't think you need the !! here since the default conversion is to bool anyway.

Acked-by Catalin Marinas <[email protected]>

The 32-bit ARM port doesn't have ACPI headers, and conditionally including them is going to look horrendous. In preparation for sharing the PSCI invocation code with 32-bit, move the acpi_psci_* function declarations and definitions such that the PSCI client code need not include ACPI headers.

While it would seem like we could simply hide the ACPI includes in psci.h, the ACPI headers have hilarious circular dependencies which make this infeasible without reorganising most of ACPICA. So rather than doing that, move the acpi_psci_* prototypes into psci.h.

The psci_acpi_init function is made dependent on CONFIG_ACPI (with a stub implementation in asm/psci.h) such that it need not be built for 32-bit ARM or kernels without ACPI support. The currently missing __init annotations are added to the prototypes in the header.

Signed-off-by Mark Rutland <[email protected]>

Reviewed-by Ashwin Chaugule <[email protected]>

Cc Al Stone <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Hanjun Guo <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Will Deacon <[email protected]>

The 32-bit ARM port doesn't have ACPI headers, and conditionally including them is going to look horrendous. In preparation for sharing the PSCI invocation code with 32-bit, move the acpi_psci_* function declarations and definitions such that the PSCI client code need not include ACPI headers.

While it would seem like we could simply hide the ACPI includes in psci.h, the ACPI headers have hilarious circular dependencies which make this infeasible without reorganising most of ACPICA. So rather than doing that, move the acpi_psci_* prototypes into psci.h.

The psci_acpi_init function is made dependent on CONFIG_ACPI (with a stub implementation in asm/psci.h) such that it need not be built for 32-bit ARM or kernels without ACPI support. The currently missing __init annotations are added to the prototypes in the header.

Signed-off-by Mark Rutland <[email protected]>

Reviewed-by Ashwin Chaugule <[email protected]>

Cc Al Stone <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Hanjun Guo <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Will Deacon <[email protected]>

---
  arch/arm64/include/asm/acpi.h | 14 --------------
  arch/arm64/include/asm/psci.h | 11 +++++++++--
  arch/arm64/kernel/acpi.c      | 11 +++++++++++
  arch/arm64/kernel/psci.c      |  4 ++--
  4 files changed, 22 insertions(+), 18 deletions(-)

arch/arm64/include/asm/acpi.h +0 -14

index 59c05d8..b0fecad 100644
@@ -39,18 +39,6 @@ extern int acpi_disabled;
  extern int acpi_noirq;
  extern int acpi_pci_disabled;
 
-/* 1 to indicate PSCI 0.2+ is implemented */
-static inline bool acpi_psci_present(void)
-{
-	return acpi_gbl_FADT.arm_boot_flags & ACPI_FADT_PSCI_COMPLIANT;
-}
-
-/* 1 to indicate HVC must be used instead of SMC as the PSCI conduit */
-static inline bool acpi_psci_use_hvc(void)
-{
-	return acpi_gbl_FADT.arm_boot_flags & ACPI_FADT_PSCI_USE_HVC;
-}
-
  static inline void disable_acpi(void)
  {
  	acpi_disabled = 1;

@@ -88,8 +76,6 @@ static inline void arch_fix_phys_package_id(int num, u32 slot) { }
  void __init acpi_init_cpus(void);
 
  #else
-static inline bool acpi_psci_present(void) { return false; }
-static inline bool acpi_psci_use_hvc(void) { return false; }
  static inline void acpi_init_cpus(void) { }
  #endif /* CONFIG_ACPI */
 

arch/arm64/include/asm/psci.h +0 -0

index 2454bc5..b5d4c1d 100644
@@ -14,7 +14,14 @@
  #ifndef __ASM_PSCI_H

Quoted 8 lines

arch/arm64/include/asm/psci.h +0 -0

index 2454bc5..b5d4c1d 100644
@@ -14,7 +14,14 @@
  #ifndef __ASM_PSCI_H
  #define __ASM_PSCI_H
 

arch/arm64/include/asm/psci.h line 22

Quoted 8 lines

arch/arm64/include/asm/psci.h +6 -2

-int psci_dt_init(void);
-int psci_acpi_init(void);
+int __init psci_dt_init(void);
+
+#ifdef CONFIG_ACPI
+int __init psci_acpi_init(void);
+bool __init acpi_psci_present(void);
+bool __init acpi_psci_use_hvc(void);

really some nit, do we need __init for function declaration?

other than that,

Acked-by Hanjun Guo <[email protected]>

To enable sharing with arm, move the core PSCI framework code to drivers/firmware. This results in a minor gain in lines of code, but this will quickly be amortised by the removal of code currently duplicated in arch/arm.

Signed-off-by Mark Rutland <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Will Deacon <[email protected]>

---
 arch/arm64/Kconfig            |   1 +
 arch/arm64/include/asm/psci.h |  27 ----
 arch/arm64/kernel/psci.c      | 354 +---------------------------------------
 arch/arm64/kernel/setup.c     |   2 +-
 drivers/firmware/Kconfig      |   3 +
 drivers/firmware/Makefile     |   1 +
 drivers/firmware/psci.c       | 363 ++++++++++++++++++++++++++++++++++++++++++
 include/linux/psci.h          |  51 ++++++
 8 files changed, 422 insertions(+), 380 deletions(-)
 delete mode 100644 arch/arm64/include/asm/psci.h
 create mode 100644 drivers/firmware/psci.c
 create mode 100644 include/linux/psci.h

To enable sharing with arm, move the core PSCI framework code to drivers/firmware. This results in a minor gain in lines of code, but this will quickly be amortised by the removal of code currently duplicated in arch/arm.

Signed-off-by Mark Rutland <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Will Deacon <[email protected]>

---
  arch/arm64/Kconfig            |   1 +
  arch/arm64/include/asm/psci.h |  27 ----
  arch/arm64/kernel/psci.c      | 354 +---------------------------------------
  arch/arm64/kernel/setup.c     |   2 +-
  drivers/firmware/Kconfig      |   3 +
  drivers/firmware/Makefile     |   1 +
  drivers/firmware/psci.c       | 363 ++++++++++++++++++++++++++++++++++++++++++
  include/linux/psci.h          |  51 ++++++
  8 files changed, 422 insertions(+), 380 deletions(-)

delete mode 100644 arch/arm64/include/asm/psci.h create mode 100644 drivers/firmware/psci.c create mode 100644 include/linux/psci.h

A 32-bit OS cannot make calls with SMC64 IDs, while a 64-bit OS must invoke some PSCI functions with SMC64 IDs.

This patch introduces and makes use of a new macro to choose the appropriate IDs based on the register width of the OS, which will allow 32-bit callers to use the PSCI client code.

Signed-off-by Mark Rutland <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

---
 drivers/firmware/psci.c | 25 +++++++++++++++++++------
 1 file changed, 19 insertions(+), 6 deletions(-)

drivers/firmware/psci.c +12 -0

index 792f4a7..e071ada 100644
@@ -27,6 +27,18 @@
 #include <asm/smp_plat.h>
 
 /*
+ * While a 64-bit OS can make calls with SMC32 calling conventions, for some
+ * calls it is necessary to use SMC64 to pass or return 64-bit values. For such
+ * calls PSCI_0_2_FN_NATIVE(x) will choose the appropriate (native-width)
+ * function ID.
+ */
+#ifdef CONFIG_64BIT
+#define PSCI_0_2_FN_NATIVE(name)	PSCI_0_2_FN_##name
+#else
+#define PSCI_0_2_FN_NATIVE(name)	PSCI_0_2_FN64_##name
+#endif
+
+/*
  * The CPU any Trusted OS is resident on. The trusted OS may reject CPU_OFF
  * calls to its resident CPU, so we must avoid issuing those. We never migrate
  * a Trusted OS even if it claims to be capable of migration -- doing so will

Quoted 38 lines · Mark Rutland

A 32-bit OS cannot make calls with SMC64 IDs, while a 64-bit OS must invoke some PSCI functions with SMC64 IDs.

This patch introduces and makes use of a new macro to choose the appropriate IDs based on the register width of the OS, which will allow 32-bit callers to use the PSCI client code.

Signed-off-by Mark Rutland <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

---
 drivers/firmware/psci.c | 25 +++++++++++++++++++------
 1 file changed, 19 insertions(+), 6 deletions(-)

drivers/firmware/psci.c +12 -0

index 792f4a7..e071ada 100644
@@ -27,6 +27,18 @@
 #include <asm/smp_plat.h>
 
 /*
+ * While a 64-bit OS can make calls with SMC32 calling conventions, for some
+ * calls it is necessary to use SMC64 to pass or return 64-bit values. For such
+ * calls PSCI_0_2_FN_NATIVE(x) will choose the appropriate (native-width)
+ * function ID.
+ */
+#ifdef CONFIG_64BIT
+#define PSCI_0_2_FN_NATIVE(name)	PSCI_0_2_FN_##name
+#else
+#define PSCI_0_2_FN_NATIVE(name)	PSCI_0_2_FN64_##name
+#endif
+
+/*
  * The CPU any Trusted OS is resident on. The trusted OS may reject CPU_OFF
  * calls to its resident CPU, so we must avoid issuing those. We never migrate
  * a Trusted OS even if it claims to be capable of migration -- doing so will

@@ -121,8 +133,8 @@ static int psci_migrate(unsigned long cpuid)
 static int psci_affinity_info(unsigned long target_affinity,

drivers/firmware/psci.c static int psci_migrate(unsigned long cpuid) · line 137

Quoted 6 lines

drivers/firmware/psci.c +2 -2

 		unsigned long lowest_affinity_level)
 {
-	return invoke_psci_fn(PSCI_0_2_FN64_AFFINITY_INFO, target_affinity,
-			      lowest_affinity_level, 0);
+	return invoke_psci_fn(PSCI_0_2_FN_NATIVE(AFFINITY_INFO),
+			      target_affinity, lowest_affinity_level, 0);

That's wrong ;).

Now that the common PSCI client code has been factored out to drivers/firmware, and made safe for 32-bit use, move the 32-bit ARM code over to it. This results in a moderate reduction of duplicated lines, and will prevent further duplication as the PSCI client code is updated for PSCI 1.0 and beyond.

The two legacy platform users of the PSCI invocation code are updated to account for interface changes. In both cases the power state parameter is changed to an opaque u32 token in preparation for PSCI 1.0 power state changes.

Signed-off-by Mark Rutland <[email protected]>

Cc Ashwin Chaugule <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Rob Herring <[email protected]>

Cc Will Deacon <[email protected]>

---
 arch/arm/Kconfig                  |   1 +
 arch/arm/include/asm/psci.h       |  23 ---
 arch/arm/kernel/Makefile          |   2 +-
 arch/arm/kernel/psci.c            | 299 --------------------------------------
 arch/arm/kernel/psci_smp.c        |  29 +++-
 arch/arm/kernel/setup.c           |   3 +-
 arch/arm/mach-highbank/highbank.c |   2 +-
 arch/arm/mach-highbank/pm.c       |   8 +-
 drivers/cpuidle/cpuidle-calxeda.c |   7 +-
 9 files changed, 32 insertions(+), 342 deletions(-)
 delete mode 100644 arch/arm/kernel/psci.c

Add myself and Lorenzo as maintainers of the PSCI client code.

Signed-off-by Mark Rutland <[email protected]>

Cc Catalin Marinas <[email protected]>

Cc Lorenzo Pieralisi <[email protected]>

Cc Will Deacon <[email protected]>

---
 MAINTAINERS | 9 +++++++++
 1 file changed, 9 insertions(+)

MAINTAINERS +0 -0

index 590304b..1694148 100644
@@ -7770,6 +7770,15 @@ S:	Maintained
 F:	include/linux/power_supply.h

This series unifies the 32-bit and 64-bit PSCI client code, moving the bulk of the FW invocation and probing out to a common locaiton in drivers/firmware. As part of this, the remaining edge cases for PSCI 0.2 (Trusted OSs rejecting CPU_OFF) are accounted for, maknig both 32-bit adn 64-bit clients more robust.

This results in a reasonable saving in terms of lines of code, and will allow for PSCI 1.0 support to be unified form the beginning, avoiding further duplication.

The series is based on v4.1-rc3.

Since v1 [1]:
* Fix build when PSCI isn't selected
* Don't indirect migrate_info_up_cpu
* Fix ver to u32 in psci_get_version

Thanks,
Mark.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2015-May/341770.html

Mark Rutland (12):
arm/arm64: kvm: add missing PSCI include
arm64: smp_plat: add get_logical_index
arm64: smp: consistently use error codes
arm64: psci: remove unnecessary id indirection
arm64: psci: support unsigned return values
arm64: psci: account for Trusted OS instances
arm64: psci: kill psci_power_state
arm64: psci: remove ACPI coupling
arm64: psci: factor invocation code to drivers
drivers: psci: support native SMC{32,64} calls
ARM: migrate to common PSCI client code
MAINTAINERS: add PSCI entry

Quoted 21 lines · Mark Rutland

This series unifies the 32-bit and 64-bit PSCI client code, moving the bulk of the FW invocation and probing out to a common locaiton in drivers/firmware. As part of this, the remaining edge cases for PSCI 0.2 (Trusted OSs rejecting CPU_OFF) are accounted for, maknig both 32-bit adn 64-bit clients more robust.

This results in a reasonable saving in terms of lines of code, and will allow for PSCI 1.0 support to be unified form the beginning, avoiding further duplication.

The series is based on v4.1-rc3.

Since v1 [1]:
* Fix build when PSCI isn't selected
* Don't indirect migrate_info_up_cpu
* Fix ver to u32 in psci_get_version

Thanks,
Mark.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2015-May/341770.html

Quoted 13 lines

Mark Rutland (12):
arm/arm64: kvm: add missing PSCI include
arm64: smp_plat: add get_logical_index
arm64: smp: consistently use error codes
arm64: psci: remove unnecessary id indirection
arm64: psci: support unsigned return values
arm64: psci: account for Trusted OS instances
arm64: psci: kill psci_power_state
arm64: psci: remove ACPI coupling
arm64: psci: factor invocation code to drivers
drivers: psci: support native SMC{32,64} calls
ARM: migrate to common PSCI client code
MAINTAINERS: add PSCI entry

So, apart from some comments I had, the series looks fine.

How do you plan to merge this? I'm happy for it to go through the arm-soc tree (hopefully there are no conflicts with some of Lorenzo's patches for the unification of ACPI/DT CPU initialisation).

Quoted 38 lines · Catalin Marinas

Quoted 21 lines · Mark Rutland

This series unifies the 32-bit and 64-bit PSCI client code, moving the bulk of the FW invocation and probing out to a common locaiton in drivers/firmware. As part of this, the remaining edge cases for PSCI 0.2 (Trusted OSs rejecting CPU_OFF) are accounted for, maknig both 32-bit adn 64-bit clients more robust.

This results in a reasonable saving in terms of lines of code, and will allow for PSCI 1.0 support to be unified form the beginning, avoiding further duplication.

The series is based on v4.1-rc3.

Since v1 [1]:
* Fix build when PSCI isn't selected
* Don't indirect migrate_info_up_cpu
* Fix ver to u32 in psci_get_version

Thanks,
Mark.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2015-May/341770.html

Quoted 13 lines

Mark Rutland (12):
arm/arm64: kvm: add missing PSCI include
arm64: smp_plat: add get_logical_index
arm64: smp: consistently use error codes
arm64: psci: remove unnecessary id indirection
arm64: psci: support unsigned return values
arm64: psci: account for Trusted OS instances
arm64: psci: kill psci_power_state
arm64: psci: remove ACPI coupling
arm64: psci: factor invocation code to drivers
drivers: psci: support native SMC{32,64} calls
ARM: migrate to common PSCI client code
MAINTAINERS: add PSCI entry

So, apart from some comments I had, the series looks fine.

Quoted 3 lines

How do you plan to merge this? I'm happy for it to go through the arm-soc tree (hopefully there are no conflicts with some of Lorenzo's patches for the unification of ACPI/DT CPU initialisation).

Unfortunately I think there are (minor, owing to PSCI headers reshuffling, but to be sorted out), let me know please the preferred way to merge the two sets.

Quoted 43 lines · Lorenzo Pieralisi

Quoted 41 lines · Catalin Marinas

Quoted 21 lines · Mark Rutland

This series unifies the 32-bit and 64-bit PSCI client code, moving the bulk of the FW invocation and probing out to a common locaiton in drivers/firmware. As part of this, the remaining edge cases for PSCI 0.2 (Trusted OSs rejecting CPU_OFF) are accounted for, maknig both 32-bit adn 64-bit clients more robust.

This results in a reasonable saving in terms of lines of code, and will allow for PSCI 1.0 support to be unified form the beginning, avoiding further duplication.

The series is based on v4.1-rc3.

Since v1 [1]:
* Fix build when PSCI isn't selected
* Don't indirect migrate_info_up_cpu
* Fix ver to u32 in psci_get_version

Thanks,
Mark.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2015-May/341770.html

Quoted 13 lines

Mark Rutland (12):
arm/arm64: kvm: add missing PSCI include
arm64: smp_plat: add get_logical_index
arm64: smp: consistently use error codes
arm64: psci: remove unnecessary id indirection
arm64: psci: support unsigned return values
arm64: psci: account for Trusted OS instances
arm64: psci: kill psci_power_state
arm64: psci: remove ACPI coupling
arm64: psci: factor invocation code to drivers
drivers: psci: support native SMC{32,64} calls
ARM: migrate to common PSCI client code
MAINTAINERS: add PSCI entry

So, apart from some comments I had, the series looks fine.

How do you plan to merge this? I'm happy for it to go through the arm-soc tree (hopefully there are no conflicts with some of Lorenzo's patches for the unification of ACPI/DT CPU initialisation).

Quoted 3 lines

Unfortunately I think there are (minor, owing to PSCI headers reshuffling, but to be sorted out), let me know please the preferred way to merge the two sets.

I guess the best thing would be to have a stable branch with your ACPI/DT unification patches patches, atop of which this series can be rebased. Then your series can go through Catalin's tree and this via arm-soc without conflict.

This series unifies the 32-bit and 64-bit PSCI client code, moving the bulk of the FW invocation and probing out to a common locaiton in drivers/firmware. As part of this, the remaining edge cases for PSCI 0.2 (Trusted OSs rejecting CPU_OFF) are accounted for, maknig both 32-bit adn 64-bit clients more robust.

This results in a reasonable saving in terms of lines of code, and will allow for PSCI 1.0 support to be unified form the beginning, avoiding further duplication.

The series is based on v4.1-rc3.

Quoted 10 lines · Mark Rutland

This series unifies the 32-bit and 64-bit PSCI client code, moving the bulk of the FW invocation and probing out to a common locaiton in drivers/firmware. As part of this, the remaining edge cases for PSCI 0.2 (Trusted OSs rejecting CPU_OFF) are accounted for, maknig both 32-bit adn 64-bit clients more robust.

This results in a reasonable saving in terms of lines of code, and will allow for PSCI 1.0 support to be unified form the beginning, avoiding further duplication.

The series is based on v4.1-rc3.

Before I review this patch set, I applied this patch set on top of v4.1-rc3 (and adding some SBSA uart patches), compiled both ok with ACPI=y/n on ARM64, and this patch set:
- booted ok with SMP with ACPI on ARM64;
- after boot, I do cpu hotplug operations for CPUs (echo 0/1 > cpu*/online) , and it works pretty fine, so PSCI CPU_ON/OFF works.

For this patch set:

Tested-by Hanjun Guo <[email protected]>