[PATCH V4 0/4] perf: Support Panther Lake uncore

[PATCH V4 0/4] perf: Support Panther Lake uncore

[kan.liang]

From: Kan Liang <kan.liang@linux.intel.com>

Changes since V3:
- __uncore_max_dies -> uncore_max_dies()
- Add reviewed-by from Dapeng

Changes since V2:
- Rebase on the later tip.git
  rdmsrl_safe_on_cpu -> rdmsrq_safe_on_cpu

Starting from the Panther Lake, the uncore discovery table mechanism is
also introduced into the client platforms. The difference is that the
portal of the global discovery table is retrieved from an MSR, rather
than MMIO space.

The first two patches provide a generic support for the extension of the
uncore discovery table mechanism.
The rest patches enable the uncore PMU on Panther Lake.

Kan Liang (4):
  perf/x86/intel/uncore: Support MSR portal for discovery tables
  perf/x86/intel/uncore: Support customized MMIO map size
  perf/x86/intel/uncore: Add Panther Lake support
  perf/x86/intel/uncore: Add iMC freerunning for Panther Lake

 arch/x86/events/intel/uncore.c           |  7 ++
 arch/x86/events/intel/uncore.h           |  2 +
 arch/x86/events/intel/uncore_discovery.c | 89 ++++++++++++++++++------
 arch/x86/events/intel/uncore_discovery.h |  7 ++
 arch/x86/events/intel/uncore_snb.c       | 79 +++++++++++++++++++++
 arch/x86/events/intel/uncore_snbep.c     |  4 +-
 6 files changed, 166 insertions(+), 22 deletions(-)

-- 
2.38.1

[PATCH V4 1/4] perf/x86/intel/uncore: Support MSR portal for discovery tables

[kan.liang]

From: Kan Liang <kan.liang@linux.intel.com>

Starting from the Panther Lake, the discovery table mechanism is also
supported in client platforms. The difference is that the portal of the
global discovery table is retrieved from an MSR.

The layout of discovery tables are the same as the server platforms.
Factor out __parse_discovery_table() to parse discover tables.

The uncore PMON is Die scope. Need to parse the discovery tables for
each die.

Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore_discovery.c | 87 ++++++++++++++++++------
 arch/x86/events/intel/uncore_discovery.h |  3 +
 2 files changed, 70 insertions(+), 20 deletions(-)

diff --git a/arch/x86/events/intel/uncore_discovery.c b/arch/x86/events/intel/uncore_discovery.c
index 571e44b49691..8680f66c3e34 100644
--- a/arch/x86/events/intel/uncore_discovery.c
+++ b/arch/x86/events/intel/uncore_discovery.c
@@ -273,32 +273,15 @@ uncore_ignore_unit(struct uncore_unit_discovery *unit, int *ignore)
 	return false;
 }
 
-static int parse_discovery_table(struct pci_dev *dev, int die,
-				 u32 bar_offset, bool *parsed,
-				 int *ignore)
+static int __parse_discovery_table(resource_size_t addr, int die,
+				   bool *parsed, int *ignore)
 {
 	struct uncore_global_discovery global;
 	struct uncore_unit_discovery unit;
 	void __iomem *io_addr;
-	resource_size_t addr;
 	unsigned long size;
-	u32 val;
 	int i;
 
-	pci_read_config_dword(dev, bar_offset, &val);
-
-	if (val & ~PCI_BASE_ADDRESS_MEM_MASK & ~PCI_BASE_ADDRESS_MEM_TYPE_64)
-		return -EINVAL;
-
-	addr = (resource_size_t)(val & PCI_BASE_ADDRESS_MEM_MASK);
-#ifdef CONFIG_PHYS_ADDR_T_64BIT
-	if ((val & PCI_BASE_ADDRESS_MEM_TYPE_MASK) == PCI_BASE_ADDRESS_MEM_TYPE_64) {
-		u32 val2;
-
-		pci_read_config_dword(dev, bar_offset + 4, &val2);
-		addr |= ((resource_size_t)val2) << 32;
-	}
-#endif
 	size = UNCORE_DISCOVERY_GLOBAL_MAP_SIZE;
 	io_addr = ioremap(addr, size);
 	if (!io_addr)
@@ -341,7 +324,32 @@ static int parse_discovery_table(struct pci_dev *dev, int die,
 	return 0;
 }
 
-bool intel_uncore_has_discovery_tables(int *ignore)
+static int parse_discovery_table(struct pci_dev *dev, int die,
+				 u32 bar_offset, bool *parsed,
+				 int *ignore)
+{
+	resource_size_t addr;
+	u32 val;
+
+	pci_read_config_dword(dev, bar_offset, &val);
+
+	if (val & ~PCI_BASE_ADDRESS_MEM_MASK & ~PCI_BASE_ADDRESS_MEM_TYPE_64)
+		return -EINVAL;
+
+	addr = (resource_size_t)(val & PCI_BASE_ADDRESS_MEM_MASK);
+#ifdef CONFIG_PHYS_ADDR_T_64BIT
+	if ((val & PCI_BASE_ADDRESS_MEM_TYPE_MASK) == PCI_BASE_ADDRESS_MEM_TYPE_64) {
+		u32 val2;
+
+		pci_read_config_dword(dev, bar_offset + 4, &val2);
+		addr |= ((resource_size_t)val2) << 32;
+	}
+#endif
+
+	return __parse_discovery_table(addr, die, parsed, ignore);
+}
+
+static bool intel_uncore_has_discovery_tables_pci(int *ignore)
 {
 	u32 device, val, entry_id, bar_offset;
 	int die, dvsec = 0, ret = true;
@@ -390,6 +398,45 @@ bool intel_uncore_has_discovery_tables(int *ignore)
 	return ret;
 }
 
+static bool intel_uncore_has_discovery_tables_msr(int *ignore)
+{
+	unsigned long *die_mask;
+	bool parsed = false;
+	int cpu, die;
+	u64 base;
+
+	die_mask = kcalloc(BITS_TO_LONGS(uncore_max_dies()),
+			   sizeof(unsigned long), GFP_KERNEL);
+	if (!die_mask)
+		return false;
+
+	cpus_read_lock();
+	for_each_online_cpu(cpu) {
+		die = topology_logical_die_id(cpu);
+		if (__test_and_set_bit(die, die_mask))
+			continue;
+
+		if (rdmsrq_safe_on_cpu(cpu, UNCORE_DISCOVERY_MSR, &base))
+			continue;
+
+		if (!base)
+			continue;
+
+		__parse_discovery_table(base, die, &parsed, ignore);
+	}
+
+	cpus_read_unlock();
+
+	kfree(die_mask);
+	return parsed;
+}
+
+bool intel_uncore_has_discovery_tables(int *ignore)
+{
+	return intel_uncore_has_discovery_tables_msr(ignore) ||
+	       intel_uncore_has_discovery_tables_pci(ignore);
+}
+
 void intel_uncore_clear_discovery_tables(void)
 {
 	struct intel_uncore_discovery_type *type, *next;
diff --git a/arch/x86/events/intel/uncore_discovery.h b/arch/x86/events/intel/uncore_discovery.h
index 0e94aa7db8e7..690f737e6837 100644
--- a/arch/x86/events/intel/uncore_discovery.h
+++ b/arch/x86/events/intel/uncore_discovery.h
@@ -1,5 +1,8 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 
+/* Store the full address of the global discovery table */
+#define UNCORE_DISCOVERY_MSR			0x201e
+
 /* Generic device ID of a discovery table device */
 #define UNCORE_DISCOVERY_TABLE_DEVICE		0x09a7
 /* Capability ID for a discovery table device */
-- 
2.38.1

[PATCH V4 2/4] perf/x86/intel/uncore: Support customized MMIO map size

[kan.liang]

From: Kan Liang <kan.liang@linux.intel.com>

For a server platform, the MMIO map size is always 0x4000. However, a
client platform may have a smaller map size.

Make the map size customizable.

Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore_discovery.c | 2 +-
 arch/x86/events/intel/uncore_snbep.c     | 2 ++
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/x86/events/intel/uncore_discovery.c b/arch/x86/events/intel/uncore_discovery.c
index 8680f66c3e34..142cf714bfe2 100644
--- a/arch/x86/events/intel/uncore_discovery.c
+++ b/arch/x86/events/intel/uncore_discovery.c
@@ -650,7 +650,7 @@ void intel_generic_uncore_mmio_init_box(struct intel_uncore_box *box)
 	}
 
 	addr = unit->addr;
-	box->io_addr = ioremap(addr, UNCORE_GENERIC_MMIO_SIZE);
+	box->io_addr = ioremap(addr, type->mmio_map_size);
 	if (!box->io_addr) {
 		pr_warn("Uncore type %d box %d: ioremap error for 0x%llx.\n",
 			type->type_id, unit->id, (unsigned long long)addr);
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index 76d96df1475a..2f5c2eb1ce0c 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -6408,6 +6408,8 @@ static void uncore_type_customized_copy(struct intel_uncore_type *to_type,
 		to_type->get_topology = from_type->get_topology;
 	if (from_type->cleanup_mapping)
 		to_type->cleanup_mapping = from_type->cleanup_mapping;
+	if (from_type->mmio_map_size)
+		to_type->mmio_map_size = from_type->mmio_map_size;
 }
 
 static struct intel_uncore_type **
-- 
2.38.1

[PATCH V4 3/4] perf/x86/intel/uncore: Add Panther Lake support

[kan.liang]

From: Kan Liang <kan.liang@linux.intel.com>

The Panther Lake supports CBOX, MC, sNCU, and HBO uncore PMON.

The CBOX is similar to Lunar Lake. The only difference is the number of
CBOX.

The other three uncore PMON can be retrieved from the discovery table.
The global control register resides in the sNCU. The global freeze bit
is set by default. It must be cleared before monitoring any uncore
counters.

Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore.c           |  7 +++
 arch/x86/events/intel/uncore.h           |  2 +
 arch/x86/events/intel/uncore_discovery.h |  4 ++
 arch/x86/events/intel/uncore_snb.c       | 71 ++++++++++++++++++++++++
 arch/x86/events/intel/uncore_snbep.c     |  2 +-
 5 files changed, 85 insertions(+), 1 deletion(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 5811e172f721..8a4f5a86e18d 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -1806,6 +1806,12 @@ static const struct intel_uncore_init_fun lnl_uncore_init __initconst = {
 	.mmio_init = lnl_uncore_mmio_init,
 };
 
+static const struct intel_uncore_init_fun ptl_uncore_init __initconst = {
+	.cpu_init = ptl_uncore_cpu_init,
+	.mmio_init = ptl_uncore_mmio_init,
+	.use_discovery = true,
+};
+
 static const struct intel_uncore_init_fun icx_uncore_init __initconst = {
 	.cpu_init = icx_uncore_cpu_init,
 	.pci_init = icx_uncore_pci_init,
@@ -1887,6 +1893,7 @@ static const struct x86_cpu_id intel_uncore_match[] __initconst = {
 	X86_MATCH_VFM(INTEL_ARROWLAKE_U,	&mtl_uncore_init),
 	X86_MATCH_VFM(INTEL_ARROWLAKE_H,	&mtl_uncore_init),
 	X86_MATCH_VFM(INTEL_LUNARLAKE_M,	&lnl_uncore_init),
+	X86_MATCH_VFM(INTEL_PANTHERLAKE_L,	&ptl_uncore_init),
 	X86_MATCH_VFM(INTEL_SAPPHIRERAPIDS_X,	&spr_uncore_init),
 	X86_MATCH_VFM(INTEL_EMERALDRAPIDS_X,	&spr_uncore_init),
 	X86_MATCH_VFM(INTEL_GRANITERAPIDS_X,	&gnr_uncore_init),
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 3dcb88c0ecfa..d8815fff7588 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -612,10 +612,12 @@ void tgl_uncore_cpu_init(void);
 void adl_uncore_cpu_init(void);
 void lnl_uncore_cpu_init(void);
 void mtl_uncore_cpu_init(void);
+void ptl_uncore_cpu_init(void);
 void tgl_uncore_mmio_init(void);
 void tgl_l_uncore_mmio_init(void);
 void adl_uncore_mmio_init(void);
 void lnl_uncore_mmio_init(void);
+void ptl_uncore_mmio_init(void);
 int snb_pci2phy_map_init(int devid);
 
 /* uncore_snbep.c */
diff --git a/arch/x86/events/intel/uncore_discovery.h b/arch/x86/events/intel/uncore_discovery.h
index 690f737e6837..dff75c98e22f 100644
--- a/arch/x86/events/intel/uncore_discovery.h
+++ b/arch/x86/events/intel/uncore_discovery.h
@@ -171,3 +171,7 @@ bool intel_generic_uncore_assign_hw_event(struct perf_event *event,
 					  struct intel_uncore_box *box);
 void uncore_find_add_unit(struct intel_uncore_discovery_unit *node,
 			  struct rb_root *root, u16 *num_units);
+struct intel_uncore_type **
+uncore_get_uncores(enum uncore_access_type type_id, int num_extra,
+		   struct intel_uncore_type **extra, int max_num_types,
+		   struct intel_uncore_type **uncores);
diff --git a/arch/x86/events/intel/uncore_snb.c b/arch/x86/events/intel/uncore_snb.c
index edb7fd50efe0..817dddc1454d 100644
--- a/arch/x86/events/intel/uncore_snb.c
+++ b/arch/x86/events/intel/uncore_snb.c
@@ -1854,3 +1854,74 @@ void lnl_uncore_mmio_init(void)
 }
 
 /* end of Lunar Lake MMIO uncore support */
+
+/* Panther Lake uncore support */
+
+#define UNCORE_PTL_MAX_NUM_UNCORE_TYPES		42
+#define UNCORE_PTL_TYPE_IMC			6
+#define UNCORE_PTL_TYPE_SNCU			34
+#define UNCORE_PTL_TYPE_HBO			41
+
+#define PTL_UNCORE_GLOBAL_CTL_OFFSET		0x380
+
+static struct intel_uncore_type ptl_uncore_imc = {
+	.name			= "imc",
+	.mmio_map_size		= 0xf00,
+};
+
+static void ptl_uncore_sncu_init_box(struct intel_uncore_box *box)
+{
+	intel_generic_uncore_mmio_init_box(box);
+
+	/* Clear the global freeze bit */
+	if (box->io_addr)
+		writel(0, box->io_addr + PTL_UNCORE_GLOBAL_CTL_OFFSET);
+}
+
+static struct intel_uncore_ops ptl_uncore_sncu_ops = {
+	.init_box		= ptl_uncore_sncu_init_box,
+	.exit_box		= uncore_mmio_exit_box,
+	.disable_box		= intel_generic_uncore_mmio_disable_box,
+	.enable_box		= intel_generic_uncore_mmio_enable_box,
+	.disable_event		= intel_generic_uncore_mmio_disable_event,
+	.enable_event		= intel_generic_uncore_mmio_enable_event,
+	.read_counter		= uncore_mmio_read_counter,
+};
+
+static struct intel_uncore_type ptl_uncore_sncu = {
+	.name			= "sncu",
+	.ops			= &ptl_uncore_sncu_ops,
+	.mmio_map_size		= 0xf00,
+};
+
+static struct intel_uncore_type ptl_uncore_hbo = {
+	.name			= "hbo",
+	.mmio_map_size		= 0xf00,
+};
+
+static struct intel_uncore_type *ptl_uncores[UNCORE_PTL_MAX_NUM_UNCORE_TYPES] = {
+	[UNCORE_PTL_TYPE_IMC] = &ptl_uncore_imc,
+	[UNCORE_PTL_TYPE_SNCU] = &ptl_uncore_sncu,
+	[UNCORE_PTL_TYPE_HBO] = &ptl_uncore_hbo,
+};
+
+void ptl_uncore_mmio_init(void)
+{
+	uncore_mmio_uncores = uncore_get_uncores(UNCORE_ACCESS_MMIO, 0, NULL,
+						 UNCORE_PTL_MAX_NUM_UNCORE_TYPES,
+						 ptl_uncores);
+}
+
+static struct intel_uncore_type *ptl_msr_uncores[] = {
+	&mtl_uncore_cbox,
+	NULL
+};
+
+void ptl_uncore_cpu_init(void)
+{
+	mtl_uncore_cbox.num_boxes = 6;
+	mtl_uncore_cbox.ops = &lnl_uncore_msr_ops;
+	uncore_msr_uncores = ptl_msr_uncores;
+}
+
+/* end of Panther Lake uncore support */
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index 2f5c2eb1ce0c..b19c262d47f5 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -6412,7 +6412,7 @@ static void uncore_type_customized_copy(struct intel_uncore_type *to_type,
 		to_type->mmio_map_size = from_type->mmio_map_size;
 }
 
-static struct intel_uncore_type **
+struct intel_uncore_type **
 uncore_get_uncores(enum uncore_access_type type_id, int num_extra,
 		   struct intel_uncore_type **extra, int max_num_types,
 		   struct intel_uncore_type **uncores)
-- 
2.38.1

[PATCH V4 4/4] perf/x86/intel/uncore: Add iMC freerunning for Panther Lake

[kan.liang]

From: Kan Liang <kan.liang@linux.intel.com>

PTL uncore imc freerunning counters are the same as the previous HW.

Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore_snb.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/arch/x86/events/intel/uncore_snb.c b/arch/x86/events/intel/uncore_snb.c
index 817dddc1454d..a8c4afc9cdff 100644
--- a/arch/x86/events/intel/uncore_snb.c
+++ b/arch/x86/events/intel/uncore_snb.c
@@ -1905,9 +1905,17 @@ static struct intel_uncore_type *ptl_uncores[UNCORE_PTL_MAX_NUM_UNCORE_TYPES] =
 	[UNCORE_PTL_TYPE_HBO] = &ptl_uncore_hbo,
 };
 
+#define UNCORE_PTL_MMIO_EXTRA_UNCORES		1
+
+static struct intel_uncore_type *ptl_mmio_extra_uncores[UNCORE_PTL_MMIO_EXTRA_UNCORES] = {
+	&adl_uncore_imc_free_running,
+};
+
 void ptl_uncore_mmio_init(void)
 {
-	uncore_mmio_uncores = uncore_get_uncores(UNCORE_ACCESS_MMIO, 0, NULL,
+	uncore_mmio_uncores = uncore_get_uncores(UNCORE_ACCESS_MMIO,
+						 UNCORE_PTL_MMIO_EXTRA_UNCORES,
+						 ptl_mmio_extra_uncores,
 						 UNCORE_PTL_MAX_NUM_UNCORE_TYPES,
 						 ptl_uncores);
 }
-- 
2.38.1