[RESEND PATCH 0/6] Perf uncore support for Snow Ridge server

[RESEND PATCH 0/6] Perf uncore support for Snow Ridge server

[kan.liang]

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

The patch series intends to enable perf uncore support for Snow Ridge
server.

Here is the link for the uncore document.
https://cdrdv2.intel.com/v1/dl/getContent/611319

Patch 1: Fixes a generic issue for uncore free-running counter, which
also impacts the Snow Ridge server.

Patch 2-6: Perf uncore support for Snow Ridge server.

Kan Liang (6):
  perf/x86/intel/uncore: Handle invalid event coding for free-running
    counter
  perf/x86/intel/uncore: Add uncore support for Snow Ridge server
  perf/x86/intel/uncore: Extract codes of box ref/unref
  perf/x86/intel/uncore: Support MMIO type uncore blocks
  perf/x86/intel/uncore: Clean up client IMC
  perf/x86/intel/uncore: Add IMC uncore support for Snow Ridge

 arch/x86/events/intel/uncore.c       | 122 +++++--
 arch/x86/events/intel/uncore.h       |  41 ++-
 arch/x86/events/intel/uncore_snb.c   |  16 +-
 arch/x86/events/intel/uncore_snbep.c | 601 +++++++++++++++++++++++++++++++++++
 4 files changed, 737 insertions(+), 43 deletions(-)

-- 
2.7.4

[RESEND PATCH 1/6] perf/x86/intel/uncore: Handle invalid event coding for free-running counter

[kan.liang]

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

Counting with invalid event coding for free-running counter may cause
OOPs, e.g. uncore_iio_free_running_0/event=1/.

Current code only validate the event with free-running event format,
event=0xff,umask=0xXY. Non-free-running event format never be checked
for the PMU with free-running counters.

Add generic hw_config() to check and reject the invalid event coding
for free-running PMU.

Fixes: 0f519f0352e3 ("perf/x86/intel/uncore: Support IIO free-running counters on SKX")
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Cc: stable@vger.kernel.org
---
 arch/x86/events/intel/uncore.h       | 10 ++++++++++
 arch/x86/events/intel/uncore_snbep.c |  1 +
 2 files changed, 11 insertions(+)

diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 79eb2e2..28499e3 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -419,6 +419,16 @@ static inline bool is_freerunning_event(struct perf_event *event)
 	       (((cfg >> 8) & 0xff) >= UNCORE_FREERUNNING_UMASK_START);
 }
 
+/* Check and reject invalid config */
+static inline int uncore_freerunning_hw_config(struct intel_uncore_box *box,
+					       struct perf_event *event)
+{
+	if (is_freerunning_event(event))
+		return 0;
+
+	return -EINVAL;
+}
+
 static inline void uncore_disable_box(struct intel_uncore_box *box)
 {
 	if (box->pmu->type->ops->disable_box)
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index b10e043..8e4e8e4 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -3585,6 +3585,7 @@ static struct uncore_event_desc skx_uncore_iio_freerunning_events[] = {
 
 static struct intel_uncore_ops skx_uncore_iio_freerunning_ops = {
 	.read_counter		= uncore_msr_read_counter,
+	.hw_config		= uncore_freerunning_hw_config,
 };
 
 static struct attribute *skx_uncore_iio_freerunning_formats_attr[] = {
-- 
2.7.4

[tip:perf/core] perf/x86/intel/uncore: Handle invalid event coding for free-running counter

[tip-bot for Kan Liang]

Commit-ID:  543ac280b3576c0009e8c0fcd4d6bfc9978d7bd0
Gitweb:     https://git.kernel.org/tip/543ac280b3576c0009e8c0fcd4d6bfc9978d7bd0
Author:     Kan Liang <kan.liang@linux.intel.com>
AuthorDate: Tue, 30 Apr 2019 17:53:43 -0700
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Mon, 17 Jun 2019 12:36:17 +0200

perf/x86/intel/uncore: Handle invalid event coding for free-running counter

Counting with invalid event coding for free-running counter may cause
OOPs, e.g. uncore_iio_free_running_0/event=1/.

Current code only validate the event with free-running event format,
event=0xff,umask=0xXY. Non-free-running event format never be checked
for the PMU with free-running counters.

Add generic hw_config() to check and reject the invalid event coding
for free-running PMU.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: acme@kernel.org
Cc: eranian@google.com
Fixes: 0f519f0352e3 ("perf/x86/intel/uncore: Support IIO free-running counters on SKX")
Link: https://lkml.kernel.org/r/1556672028-119221-2-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/events/intel/uncore.h       | 10 ++++++++++
 arch/x86/events/intel/uncore_snbep.c |  1 +
 2 files changed, 11 insertions(+)

diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 33aba2504cb1..b3cad2b65766 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -419,6 +419,16 @@ static inline bool is_freerunning_event(struct perf_event *event)
 	       (((cfg >> 8) & 0xff) >= UNCORE_FREERUNNING_UMASK_START);
 }
 
+/* Check and reject invalid config */
+static inline int uncore_freerunning_hw_config(struct intel_uncore_box *box,
+					       struct perf_event *event)
+{
+	if (is_freerunning_event(event))
+		return 0;
+
+	return -EINVAL;
+}
+
 static inline void uncore_disable_box(struct intel_uncore_box *box)
 {
 	if (box->pmu->type->ops->disable_box)
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index bbe89bc589f9..fdb1a57ee1e5 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -3585,6 +3585,7 @@ static struct uncore_event_desc skx_uncore_iio_freerunning_events[] = {
 
 static struct intel_uncore_ops skx_uncore_iio_freerunning_ops = {
 	.read_counter		= uncore_msr_read_counter,
+	.hw_config		= uncore_freerunning_hw_config,
 };
 
 static struct attribute *skx_uncore_iio_freerunning_formats_attr[] = {

[RESEND PATCH 2/6] perf/x86/intel/uncore: Add uncore support for Snow Ridge server

[kan.liang]

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

The uncore subsystem on Snow Ridge is similar as previous SKX server.
The uncore units on Snow Ridge include Ubox, Chabox, IIO, IRP, M2PCIE,
PCU, M2M, PCIE3 and IMC.
- The config register encoding and pci device IDs are changed.
- For CHA, the umask_ext and filter_tid fields are changed.
- For IIO, the ch_mask and fc_mask fields are changed.
- For M2M, the mask_ext field is changed.
- Add new PCIe3 unit for PCIe3 root port which provides the interface
  between PCIe devices, plugged into the PCIe port, and the components
  (in M2IOSF).
- IMC can only be accessed via MMIO on Snow Ridge now. Current common
  code doesn't support it yet. IMC will be supported in following
  patches.
- There are 9 free running counters for IIO CLOCKS and bandwidth In.
- Full uncore event list is not published yet. Event constrain is not
  included in this patch. It will be added later separately.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore.c       |   6 +
 arch/x86/events/intel/uncore.h       |   2 +
 arch/x86/events/intel/uncore_snbep.c | 403 +++++++++++++++++++++++++++++++++++
 3 files changed, 411 insertions(+)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index fc40a14..ee23b50 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -1372,6 +1372,11 @@ static const struct intel_uncore_init_fun icl_uncore_init __initconst = {
 	.pci_init = skl_uncore_pci_init,
 };
 
+static const struct intel_uncore_init_fun snr_uncore_init __initconst = {
+	.cpu_init = snr_uncore_cpu_init,
+	.pci_init = snr_uncore_pci_init,
+};
+
 static const struct x86_cpu_id intel_uncore_match[] __initconst = {
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_NEHALEM_EP,	  nhm_uncore_init),
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_NEHALEM,	  nhm_uncore_init),
@@ -1399,6 +1404,7 @@ static const struct x86_cpu_id intel_uncore_match[] __initconst = {
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_KABYLAKE_MOBILE, skl_uncore_init),
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_KABYLAKE_DESKTOP, skl_uncore_init),
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_ICELAKE_MOBILE, icl_uncore_init),
+	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_ATOM_TREMONT_X, snr_uncore_init),
 	{},
 };
 
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 28499e3..5e97e5e 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -538,6 +538,8 @@ int knl_uncore_pci_init(void);
 void knl_uncore_cpu_init(void);
 int skx_uncore_pci_init(void);
 void skx_uncore_cpu_init(void);
+int snr_uncore_pci_init(void);
+void snr_uncore_cpu_init(void);
 
 /* uncore_nhmex.c */
 void nhmex_uncore_cpu_init(void);
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index 8e4e8e4..851f76f 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -324,12 +324,64 @@
 #define SKX_M2M_PCI_PMON_CTR0		0x200
 #define SKX_M2M_PCI_PMON_BOX_CTL	0x258
 
+/* SNR Ubox */
+#define SNR_U_MSR_PMON_CTR0			0x1f98
+#define SNR_U_MSR_PMON_CTL0			0x1f91
+#define SNR_U_MSR_PMON_UCLK_FIXED_CTL		0x1f93
+#define SNR_U_MSR_PMON_UCLK_FIXED_CTR		0x1f94
+
+/* SNR CHA */
+#define SNR_CHA_RAW_EVENT_MASK_EXT		0x3ffffff
+#define SNR_CHA_MSR_PMON_CTL0			0x1c01
+#define SNR_CHA_MSR_PMON_CTR0			0x1c08
+#define SNR_CHA_MSR_PMON_BOX_CTL		0x1c00
+#define SNR_C0_MSR_PMON_BOX_FILTER0		0x1c05
+
+
+/* SNR IIO */
+#define SNR_IIO_MSR_PMON_CTL0			0x1e08
+#define SNR_IIO_MSR_PMON_CTR0			0x1e01
+#define SNR_IIO_MSR_PMON_BOX_CTL		0x1e00
+#define SNR_IIO_MSR_OFFSET			0x10
+#define SNR_IIO_PMON_RAW_EVENT_MASK_EXT		0x7ffff
+
+/* SNR IRP */
+#define SNR_IRP0_MSR_PMON_CTL0			0x1ea8
+#define SNR_IRP0_MSR_PMON_CTR0			0x1ea1
+#define SNR_IRP0_MSR_PMON_BOX_CTL		0x1ea0
+#define SNR_IRP_MSR_OFFSET			0x10
+
+/* SNR M2PCIE */
+#define SNR_M2PCIE_MSR_PMON_CTL0		0x1e58
+#define SNR_M2PCIE_MSR_PMON_CTR0		0x1e51
+#define SNR_M2PCIE_MSR_PMON_BOX_CTL		0x1e50
+#define SNR_M2PCIE_MSR_OFFSET			0x10
+
+/* SNR PCU */
+#define SNR_PCU_MSR_PMON_CTL0			0x1ef1
+#define SNR_PCU_MSR_PMON_CTR0			0x1ef8
+#define SNR_PCU_MSR_PMON_BOX_CTL		0x1ef0
+#define SNR_PCU_MSR_PMON_BOX_FILTER		0x1efc
+
+/* SNR M2M */
+#define SNR_M2M_PCI_PMON_CTL0			0x468
+#define SNR_M2M_PCI_PMON_CTR0			0x440
+#define SNR_M2M_PCI_PMON_BOX_CTL		0x438
+#define SNR_M2M_PCI_PMON_UMASK_EXT		0xff
+
+/* SNR PCIE3 */
+#define SNR_PCIE3_PCI_PMON_CTL0			0x508
+#define SNR_PCIE3_PCI_PMON_CTR0			0x4e8
+#define SNR_PCIE3_PCI_PMON_BOX_CTL		0x4e4
+
 DEFINE_UNCORE_FORMAT_ATTR(event, event, "config:0-7");
 DEFINE_UNCORE_FORMAT_ATTR(event2, event, "config:0-6");
 DEFINE_UNCORE_FORMAT_ATTR(event_ext, event, "config:0-7,21");
 DEFINE_UNCORE_FORMAT_ATTR(use_occ_ctr, use_occ_ctr, "config:7");
 DEFINE_UNCORE_FORMAT_ATTR(umask, umask, "config:8-15");
 DEFINE_UNCORE_FORMAT_ATTR(umask_ext, umask, "config:8-15,32-43,45-55");
+DEFINE_UNCORE_FORMAT_ATTR(umask_ext2, umask, "config:8-15,32-57");
+DEFINE_UNCORE_FORMAT_ATTR(umask_ext3, umask, "config:8-15,32-39");
 DEFINE_UNCORE_FORMAT_ATTR(qor, qor, "config:16");
 DEFINE_UNCORE_FORMAT_ATTR(edge, edge, "config:18");
 DEFINE_UNCORE_FORMAT_ATTR(tid_en, tid_en, "config:19");
@@ -343,11 +395,14 @@ DEFINE_UNCORE_FORMAT_ATTR(occ_invert, occ_invert, "config:30");
 DEFINE_UNCORE_FORMAT_ATTR(occ_edge, occ_edge, "config:14-51");
 DEFINE_UNCORE_FORMAT_ATTR(occ_edge_det, occ_edge_det, "config:31");
 DEFINE_UNCORE_FORMAT_ATTR(ch_mask, ch_mask, "config:36-43");
+DEFINE_UNCORE_FORMAT_ATTR(ch_mask2, ch_mask, "config:36-47");
 DEFINE_UNCORE_FORMAT_ATTR(fc_mask, fc_mask, "config:44-46");
+DEFINE_UNCORE_FORMAT_ATTR(fc_mask2, fc_mask, "config:48-50");
 DEFINE_UNCORE_FORMAT_ATTR(filter_tid, filter_tid, "config1:0-4");
 DEFINE_UNCORE_FORMAT_ATTR(filter_tid2, filter_tid, "config1:0");
 DEFINE_UNCORE_FORMAT_ATTR(filter_tid3, filter_tid, "config1:0-5");
 DEFINE_UNCORE_FORMAT_ATTR(filter_tid4, filter_tid, "config1:0-8");
+DEFINE_UNCORE_FORMAT_ATTR(filter_tid5, filter_tid, "config1:0-9");
 DEFINE_UNCORE_FORMAT_ATTR(filter_cid, filter_cid, "config1:5");
 DEFINE_UNCORE_FORMAT_ATTR(filter_link, filter_link, "config1:5-8");
 DEFINE_UNCORE_FORMAT_ATTR(filter_link2, filter_link, "config1:6-8");
@@ -3968,3 +4023,351 @@ int skx_uncore_pci_init(void)
 }
 
 /* end of SKX uncore support */
+
+/* SNR uncore support */
+
+static struct intel_uncore_type snr_uncore_ubox = {
+	.name			= "ubox",
+	.num_counters		= 2,
+	.num_boxes		= 1,
+	.perf_ctr_bits		= 48,
+	.fixed_ctr_bits		= 48,
+	.perf_ctr		= SNR_U_MSR_PMON_CTR0,
+	.event_ctl		= SNR_U_MSR_PMON_CTL0,
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.fixed_ctr		= SNR_U_MSR_PMON_UCLK_FIXED_CTR,
+	.fixed_ctl		= SNR_U_MSR_PMON_UCLK_FIXED_CTL,
+	.ops			= &ivbep_uncore_msr_ops,
+	.format_group		= &ivbep_uncore_format_group,
+};
+
+static struct attribute *snr_uncore_cha_formats_attr[] = {
+	&format_attr_event.attr,
+	&format_attr_umask_ext2.attr,
+	&format_attr_edge.attr,
+	&format_attr_tid_en.attr,
+	&format_attr_inv.attr,
+	&format_attr_thresh8.attr,
+	&format_attr_filter_tid5.attr,
+	NULL,
+};
+static const struct attribute_group snr_uncore_chabox_format_group = {
+	.name = "format",
+	.attrs = snr_uncore_cha_formats_attr,
+};
+
+static int snr_cha_hw_config(struct intel_uncore_box *box, struct perf_event *event)
+{
+	struct hw_perf_event_extra *reg1 = &event->hw.extra_reg;
+
+	reg1->reg = SNR_C0_MSR_PMON_BOX_FILTER0 +
+		    box->pmu->type->msr_offset * box->pmu->pmu_idx;
+	reg1->config = event->attr.config1 & SKX_CHA_MSR_PMON_BOX_FILTER_TID;
+	reg1->idx = 0;
+
+	return 0;
+}
+
+static void snr_cha_enable_event(struct intel_uncore_box *box,
+				   struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+	struct hw_perf_event_extra *reg1 = &hwc->extra_reg;
+
+	if (reg1->idx != EXTRA_REG_NONE)
+		wrmsrl(reg1->reg, reg1->config);
+
+	wrmsrl(hwc->config_base, hwc->config | SNBEP_PMON_CTL_EN);
+}
+
+static struct intel_uncore_ops snr_uncore_chabox_ops = {
+	.init_box		= ivbep_uncore_msr_init_box,
+	.disable_box		= snbep_uncore_msr_disable_box,
+	.enable_box		= snbep_uncore_msr_enable_box,
+	.disable_event		= snbep_uncore_msr_disable_event,
+	.enable_event		= snr_cha_enable_event,
+	.read_counter		= uncore_msr_read_counter,
+	.hw_config		= snr_cha_hw_config,
+};
+
+static struct intel_uncore_type snr_uncore_chabox = {
+	.name			= "cha",
+	.num_counters		= 4,
+	.num_boxes		= 6,
+	.perf_ctr_bits		= 48,
+	.event_ctl		= SNR_CHA_MSR_PMON_CTL0,
+	.perf_ctr		= SNR_CHA_MSR_PMON_CTR0,
+	.box_ctl		= SNR_CHA_MSR_PMON_BOX_CTL,
+	.msr_offset		= HSWEP_CBO_MSR_OFFSET,
+	.event_mask		= HSWEP_S_MSR_PMON_RAW_EVENT_MASK,
+	.event_mask_ext		= SNR_CHA_RAW_EVENT_MASK_EXT,
+	.ops			= &snr_uncore_chabox_ops,
+	.format_group		= &snr_uncore_chabox_format_group,
+};
+
+static struct attribute *snr_uncore_iio_formats_attr[] = {
+	&format_attr_event.attr,
+	&format_attr_umask.attr,
+	&format_attr_edge.attr,
+	&format_attr_inv.attr,
+	&format_attr_thresh9.attr,
+	&format_attr_ch_mask2.attr,
+	&format_attr_fc_mask2.attr,
+	NULL,
+};
+
+static const struct attribute_group snr_uncore_iio_format_group = {
+	.name = "format",
+	.attrs = snr_uncore_iio_formats_attr,
+};
+
+static struct intel_uncore_type snr_uncore_iio = {
+	.name			= "iio",
+	.num_counters		= 4,
+	.num_boxes		= 5,
+	.perf_ctr_bits		= 48,
+	.event_ctl		= SNR_IIO_MSR_PMON_CTL0,
+	.perf_ctr		= SNR_IIO_MSR_PMON_CTR0,
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.event_mask_ext		= SNR_IIO_PMON_RAW_EVENT_MASK_EXT,
+	.box_ctl		= SNR_IIO_MSR_PMON_BOX_CTL,
+	.msr_offset		= SNR_IIO_MSR_OFFSET,
+	.ops			= &ivbep_uncore_msr_ops,
+	.format_group		= &snr_uncore_iio_format_group,
+};
+
+static struct intel_uncore_type snr_uncore_irp = {
+	.name			= "irp",
+	.num_counters		= 2,
+	.num_boxes		= 5,
+	.perf_ctr_bits		= 48,
+	.event_ctl		= SNR_IRP0_MSR_PMON_CTL0,
+	.perf_ctr		= SNR_IRP0_MSR_PMON_CTR0,
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.box_ctl		= SNR_IRP0_MSR_PMON_BOX_CTL,
+	.msr_offset		= SNR_IRP_MSR_OFFSET,
+	.ops			= &ivbep_uncore_msr_ops,
+	.format_group		= &ivbep_uncore_format_group,
+};
+
+static struct intel_uncore_type snr_uncore_m2pcie = {
+	.name		= "m2pcie",
+	.num_counters	= 4,
+	.num_boxes	= 5,
+	.perf_ctr_bits	= 48,
+	.event_ctl	= SNR_M2PCIE_MSR_PMON_CTL0,
+	.perf_ctr	= SNR_M2PCIE_MSR_PMON_CTR0,
+	.box_ctl	= SNR_M2PCIE_MSR_PMON_BOX_CTL,
+	.msr_offset	= SNR_M2PCIE_MSR_OFFSET,
+	.event_mask	= SNBEP_PMON_RAW_EVENT_MASK,
+	.ops		= &ivbep_uncore_msr_ops,
+	.format_group	= &ivbep_uncore_format_group,
+};
+
+static int snr_pcu_hw_config(struct intel_uncore_box *box, struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+	struct hw_perf_event_extra *reg1 = &hwc->extra_reg;
+	int ev_sel = hwc->config & SNBEP_PMON_CTL_EV_SEL_MASK;
+
+	if (ev_sel >= 0xb && ev_sel <= 0xe) {
+		reg1->reg = SNR_PCU_MSR_PMON_BOX_FILTER;
+		reg1->idx = ev_sel - 0xb;
+		reg1->config = event->attr.config1 & (0xff << reg1->idx);
+	}
+	return 0;
+}
+
+static struct intel_uncore_ops snr_uncore_pcu_ops = {
+	IVBEP_UNCORE_MSR_OPS_COMMON_INIT(),
+	.hw_config		= snr_pcu_hw_config,
+	.get_constraint		= snbep_pcu_get_constraint,
+	.put_constraint		= snbep_pcu_put_constraint,
+};
+
+static struct intel_uncore_type snr_uncore_pcu = {
+	.name			= "pcu",
+	.num_counters		= 4,
+	.num_boxes		= 1,
+	.perf_ctr_bits		= 48,
+	.perf_ctr		= SNR_PCU_MSR_PMON_CTR0,
+	.event_ctl		= SNR_PCU_MSR_PMON_CTL0,
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.box_ctl		= SNR_PCU_MSR_PMON_BOX_CTL,
+	.num_shared_regs	= 1,
+	.ops			= &snr_uncore_pcu_ops,
+	.format_group		= &skx_uncore_pcu_format_group,
+};
+
+enum perf_uncore_snr_iio_freerunning_type_id {
+	SNR_IIO_MSR_IOCLK,
+	SNR_IIO_MSR_BW_IN,
+
+	SNR_IIO_FREERUNNING_TYPE_MAX,
+};
+
+static struct freerunning_counters snr_iio_freerunning[] = {
+	[SNR_IIO_MSR_IOCLK]	= { 0x1eac, 0x1, 0x10, 1, 48 },
+	[SNR_IIO_MSR_BW_IN]	= { 0x1f00, 0x1, 0x10, 8, 48 },
+};
+
+static struct uncore_event_desc snr_uncore_iio_freerunning_events[] = {
+	/* Free-Running IIO CLOCKS Counter */
+	INTEL_UNCORE_EVENT_DESC(ioclk,			"event=0xff,umask=0x10"),
+	/* Free-Running IIO BANDWIDTH IN Counters */
+	INTEL_UNCORE_EVENT_DESC(bw_in_port0,		"event=0xff,umask=0x20"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port0.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port0.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port1,		"event=0xff,umask=0x21"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port1.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port1.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port2,		"event=0xff,umask=0x22"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port2.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port2.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port3,		"event=0xff,umask=0x23"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port3.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port3.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port4,		"event=0xff,umask=0x24"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port4.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port4.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port5,		"event=0xff,umask=0x25"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port5.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port5.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port6,		"event=0xff,umask=0x26"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port6.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port6.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port7,		"event=0xff,umask=0x27"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port7.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port7.unit,	"MiB"),
+	{ /* end: all zeroes */ },
+};
+
+static struct intel_uncore_type snr_uncore_iio_free_running = {
+	.name			= "iio_free_running",
+	.num_counters		= 9,
+	.num_boxes		= 5,
+	.num_freerunning_types	= SNR_IIO_FREERUNNING_TYPE_MAX,
+	.freerunning		= snr_iio_freerunning,
+	.ops			= &skx_uncore_iio_freerunning_ops,
+	.event_descs		= snr_uncore_iio_freerunning_events,
+	.format_group		= &skx_uncore_iio_freerunning_format_group,
+};
+
+static struct intel_uncore_type *snr_msr_uncores[] = {
+	&snr_uncore_ubox,
+	&snr_uncore_chabox,
+	&snr_uncore_iio,
+	&snr_uncore_irp,
+	&snr_uncore_m2pcie,
+	&snr_uncore_pcu,
+	&snr_uncore_iio_free_running,
+	NULL,
+};
+
+void snr_uncore_cpu_init(void)
+{
+	uncore_msr_uncores = snr_msr_uncores;
+}
+
+static void snr_m2m_uncore_pci_init_box(struct intel_uncore_box *box)
+{
+	struct pci_dev *pdev = box->pci_dev;
+	int box_ctl = uncore_pci_box_ctl(box);
+
+	__set_bit(UNCORE_BOX_FLAG_CTL_OFFS8, &box->flags);
+	pci_write_config_dword(pdev, box_ctl, IVBEP_PMON_BOX_CTL_INT);
+}
+
+static struct intel_uncore_ops snr_m2m_uncore_pci_ops = {
+	.init_box	= snr_m2m_uncore_pci_init_box,
+	.disable_box	= snbep_uncore_pci_disable_box,
+	.enable_box	= snbep_uncore_pci_enable_box,
+	.disable_event	= snbep_uncore_pci_disable_event,
+	.enable_event	= snbep_uncore_pci_enable_event,
+	.read_counter	= snbep_uncore_pci_read_counter,
+};
+
+static struct attribute *snr_m2m_uncore_formats_attr[] = {
+	&format_attr_event.attr,
+	&format_attr_umask_ext3.attr,
+	&format_attr_edge.attr,
+	&format_attr_inv.attr,
+	&format_attr_thresh8.attr,
+	NULL,
+};
+
+static const struct attribute_group snr_m2m_uncore_format_group = {
+	.name = "format",
+	.attrs = snr_m2m_uncore_formats_attr,
+};
+
+static struct intel_uncore_type snr_uncore_m2m = {
+	.name		= "m2m",
+	.num_counters   = 4,
+	.num_boxes	= 1,
+	.perf_ctr_bits	= 48,
+	.perf_ctr	= SNR_M2M_PCI_PMON_CTR0,
+	.event_ctl	= SNR_M2M_PCI_PMON_CTL0,
+	.event_mask	= SNBEP_PMON_RAW_EVENT_MASK,
+	.event_mask_ext	= SNR_M2M_PCI_PMON_UMASK_EXT,
+	.box_ctl	= SNR_M2M_PCI_PMON_BOX_CTL,
+	.ops		= &snr_m2m_uncore_pci_ops,
+	.format_group	= &snr_m2m_uncore_format_group,
+};
+
+static struct intel_uncore_type snr_uncore_pcie3 = {
+	.name		= "pcie3",
+	.num_counters	= 4,
+	.num_boxes	= 1,
+	.perf_ctr_bits	= 48,
+	.perf_ctr	= SNR_PCIE3_PCI_PMON_CTR0,
+	.event_ctl	= SNR_PCIE3_PCI_PMON_CTL0,
+	.event_mask	= SNBEP_PMON_RAW_EVENT_MASK,
+	.box_ctl	= SNR_PCIE3_PCI_PMON_BOX_CTL,
+	.ops		= &ivbep_uncore_pci_ops,
+	.format_group	= &ivbep_uncore_format_group,
+};
+
+enum {
+	SNR_PCI_UNCORE_M2M,
+	SNR_PCI_UNCORE_PCIE3,
+};
+
+static struct intel_uncore_type *snr_pci_uncores[] = {
+	[SNR_PCI_UNCORE_M2M]		= &snr_uncore_m2m,
+	[SNR_PCI_UNCORE_PCIE3]		= &snr_uncore_pcie3,
+	NULL,
+};
+
+static const struct pci_device_id snr_uncore_pci_ids[] = {
+	{ /* M2M */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x344a),
+		.driver_data = UNCORE_PCI_DEV_FULL_DATA(12, 0, SNR_PCI_UNCORE_M2M, 0),
+	},
+	{ /* PCIe3 */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x334a),
+		.driver_data = UNCORE_PCI_DEV_FULL_DATA(4, 0, SNR_PCI_UNCORE_PCIE3, 0),
+	},
+	{ /* end: all zeroes */ }
+};
+
+static struct pci_driver snr_uncore_pci_driver = {
+	.name		= "snr_uncore",
+	.id_table	= snr_uncore_pci_ids,
+};
+
+int snr_uncore_pci_init(void)
+{
+	/* SNR UBOX DID */
+	int ret = snbep_pci2phy_map_init(0x3460, SKX_CPUNODEID,
+					 SKX_GIDNIDMAP, true);
+
+	if (ret)
+		return ret;
+
+	uncore_pci_uncores = snr_pci_uncores;
+	uncore_pci_driver = &snr_uncore_pci_driver;
+	return 0;
+}
+
+/* end of SNR uncore support */
-- 
2.7.4

[tip:perf/core] perf/x86/intel/uncore: Add uncore support for Snow Ridge server

[tip-bot for Kan Liang]

Commit-ID:  210cc5f9db7a5c66b7ca6290b7d35cc7db7e9dbd
Gitweb:     https://git.kernel.org/tip/210cc5f9db7a5c66b7ca6290b7d35cc7db7e9dbd
Author:     Kan Liang <kan.liang@linux.intel.com>
AuthorDate: Tue, 30 Apr 2019 17:53:44 -0700
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Mon, 17 Jun 2019 12:36:18 +0200

perf/x86/intel/uncore: Add uncore support for Snow Ridge server

The uncore subsystem on Snow Ridge is similar as previous SKX server.
The uncore units on Snow Ridge include Ubox, Chabox, IIO, IRP, M2PCIE,
PCU, M2M, PCIE3 and IMC.

- The config register encoding and pci device IDs are changed.
- For CHA, the umask_ext and filter_tid fields are changed.
- For IIO, the ch_mask and fc_mask fields are changed.
- For M2M, the mask_ext field is changed.
- Add new PCIe3 unit for PCIe3 root port which provides the interface
  between PCIe devices, plugged into the PCIe port, and the components
  (in M2IOSF).
- IMC can only be accessed via MMIO on Snow Ridge now. Current common
  code doesn't support it yet. IMC will be supported in following
  patches.
- There are 9 free running counters for IIO CLOCKS and bandwidth In.
- Full uncore event list is not published yet. Event constrain is not
  included in this patch. It will be added later separately.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: acme@kernel.org
Cc: eranian@google.com
Link: https://lkml.kernel.org/r/1556672028-119221-3-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/events/intel/uncore.c       |   6 +
 arch/x86/events/intel/uncore.h       |   2 +
 arch/x86/events/intel/uncore_snbep.c | 403 +++++++++++++++++++++++++++++++++++
 3 files changed, 411 insertions(+)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index d37bb2c657b0..342c323e0f6a 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -1375,6 +1375,11 @@ static const struct intel_uncore_init_fun icl_uncore_init __initconst = {
 	.pci_init = skl_uncore_pci_init,
 };
 
+static const struct intel_uncore_init_fun snr_uncore_init __initconst = {
+	.cpu_init = snr_uncore_cpu_init,
+	.pci_init = snr_uncore_pci_init,
+};
+
 static const struct x86_cpu_id intel_uncore_match[] __initconst = {
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_NEHALEM_EP,	  nhm_uncore_init),
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_NEHALEM,	  nhm_uncore_init),
@@ -1404,6 +1409,7 @@ static const struct x86_cpu_id intel_uncore_match[] __initconst = {
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_ICELAKE_MOBILE, icl_uncore_init),
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_ICELAKE_NNPI, icl_uncore_init),
 	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_ICELAKE_DESKTOP, icl_uncore_init),
+	X86_UNCORE_MODEL_MATCH(INTEL_FAM6_ATOM_TREMONT_X, snr_uncore_init),
 	{},
 };
 
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index b3cad2b65766..b444ed35d6f7 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -538,6 +538,8 @@ int knl_uncore_pci_init(void);
 void knl_uncore_cpu_init(void);
 int skx_uncore_pci_init(void);
 void skx_uncore_cpu_init(void);
+int snr_uncore_pci_init(void);
+void snr_uncore_cpu_init(void);
 
 /* uncore_nhmex.c */
 void nhmex_uncore_cpu_init(void);
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index fdb1a57ee1e5..3d8752b37413 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -324,12 +324,64 @@
 #define SKX_M2M_PCI_PMON_CTR0		0x200
 #define SKX_M2M_PCI_PMON_BOX_CTL	0x258
 
+/* SNR Ubox */
+#define SNR_U_MSR_PMON_CTR0			0x1f98
+#define SNR_U_MSR_PMON_CTL0			0x1f91
+#define SNR_U_MSR_PMON_UCLK_FIXED_CTL		0x1f93
+#define SNR_U_MSR_PMON_UCLK_FIXED_CTR		0x1f94
+
+/* SNR CHA */
+#define SNR_CHA_RAW_EVENT_MASK_EXT		0x3ffffff
+#define SNR_CHA_MSR_PMON_CTL0			0x1c01
+#define SNR_CHA_MSR_PMON_CTR0			0x1c08
+#define SNR_CHA_MSR_PMON_BOX_CTL		0x1c00
+#define SNR_C0_MSR_PMON_BOX_FILTER0		0x1c05
+
+
+/* SNR IIO */
+#define SNR_IIO_MSR_PMON_CTL0			0x1e08
+#define SNR_IIO_MSR_PMON_CTR0			0x1e01
+#define SNR_IIO_MSR_PMON_BOX_CTL		0x1e00
+#define SNR_IIO_MSR_OFFSET			0x10
+#define SNR_IIO_PMON_RAW_EVENT_MASK_EXT		0x7ffff
+
+/* SNR IRP */
+#define SNR_IRP0_MSR_PMON_CTL0			0x1ea8
+#define SNR_IRP0_MSR_PMON_CTR0			0x1ea1
+#define SNR_IRP0_MSR_PMON_BOX_CTL		0x1ea0
+#define SNR_IRP_MSR_OFFSET			0x10
+
+/* SNR M2PCIE */
+#define SNR_M2PCIE_MSR_PMON_CTL0		0x1e58
+#define SNR_M2PCIE_MSR_PMON_CTR0		0x1e51
+#define SNR_M2PCIE_MSR_PMON_BOX_CTL		0x1e50
+#define SNR_M2PCIE_MSR_OFFSET			0x10
+
+/* SNR PCU */
+#define SNR_PCU_MSR_PMON_CTL0			0x1ef1
+#define SNR_PCU_MSR_PMON_CTR0			0x1ef8
+#define SNR_PCU_MSR_PMON_BOX_CTL		0x1ef0
+#define SNR_PCU_MSR_PMON_BOX_FILTER		0x1efc
+
+/* SNR M2M */
+#define SNR_M2M_PCI_PMON_CTL0			0x468
+#define SNR_M2M_PCI_PMON_CTR0			0x440
+#define SNR_M2M_PCI_PMON_BOX_CTL		0x438
+#define SNR_M2M_PCI_PMON_UMASK_EXT		0xff
+
+/* SNR PCIE3 */
+#define SNR_PCIE3_PCI_PMON_CTL0			0x508
+#define SNR_PCIE3_PCI_PMON_CTR0			0x4e8
+#define SNR_PCIE3_PCI_PMON_BOX_CTL		0x4e4
+
 DEFINE_UNCORE_FORMAT_ATTR(event, event, "config:0-7");
 DEFINE_UNCORE_FORMAT_ATTR(event2, event, "config:0-6");
 DEFINE_UNCORE_FORMAT_ATTR(event_ext, event, "config:0-7,21");
 DEFINE_UNCORE_FORMAT_ATTR(use_occ_ctr, use_occ_ctr, "config:7");
 DEFINE_UNCORE_FORMAT_ATTR(umask, umask, "config:8-15");
 DEFINE_UNCORE_FORMAT_ATTR(umask_ext, umask, "config:8-15,32-43,45-55");
+DEFINE_UNCORE_FORMAT_ATTR(umask_ext2, umask, "config:8-15,32-57");
+DEFINE_UNCORE_FORMAT_ATTR(umask_ext3, umask, "config:8-15,32-39");
 DEFINE_UNCORE_FORMAT_ATTR(qor, qor, "config:16");
 DEFINE_UNCORE_FORMAT_ATTR(edge, edge, "config:18");
 DEFINE_UNCORE_FORMAT_ATTR(tid_en, tid_en, "config:19");
@@ -343,11 +395,14 @@ DEFINE_UNCORE_FORMAT_ATTR(occ_invert, occ_invert, "config:30");
 DEFINE_UNCORE_FORMAT_ATTR(occ_edge, occ_edge, "config:14-51");
 DEFINE_UNCORE_FORMAT_ATTR(occ_edge_det, occ_edge_det, "config:31");
 DEFINE_UNCORE_FORMAT_ATTR(ch_mask, ch_mask, "config:36-43");
+DEFINE_UNCORE_FORMAT_ATTR(ch_mask2, ch_mask, "config:36-47");
 DEFINE_UNCORE_FORMAT_ATTR(fc_mask, fc_mask, "config:44-46");
+DEFINE_UNCORE_FORMAT_ATTR(fc_mask2, fc_mask, "config:48-50");
 DEFINE_UNCORE_FORMAT_ATTR(filter_tid, filter_tid, "config1:0-4");
 DEFINE_UNCORE_FORMAT_ATTR(filter_tid2, filter_tid, "config1:0");
 DEFINE_UNCORE_FORMAT_ATTR(filter_tid3, filter_tid, "config1:0-5");
 DEFINE_UNCORE_FORMAT_ATTR(filter_tid4, filter_tid, "config1:0-8");
+DEFINE_UNCORE_FORMAT_ATTR(filter_tid5, filter_tid, "config1:0-9");
 DEFINE_UNCORE_FORMAT_ATTR(filter_cid, filter_cid, "config1:5");
 DEFINE_UNCORE_FORMAT_ATTR(filter_link, filter_link, "config1:5-8");
 DEFINE_UNCORE_FORMAT_ATTR(filter_link2, filter_link, "config1:6-8");
@@ -3968,3 +4023,351 @@ int skx_uncore_pci_init(void)
 }
 
 /* end of SKX uncore support */
+
+/* SNR uncore support */
+
+static struct intel_uncore_type snr_uncore_ubox = {
+	.name			= "ubox",
+	.num_counters		= 2,
+	.num_boxes		= 1,
+	.perf_ctr_bits		= 48,
+	.fixed_ctr_bits		= 48,
+	.perf_ctr		= SNR_U_MSR_PMON_CTR0,
+	.event_ctl		= SNR_U_MSR_PMON_CTL0,
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.fixed_ctr		= SNR_U_MSR_PMON_UCLK_FIXED_CTR,
+	.fixed_ctl		= SNR_U_MSR_PMON_UCLK_FIXED_CTL,
+	.ops			= &ivbep_uncore_msr_ops,
+	.format_group		= &ivbep_uncore_format_group,
+};
+
+static struct attribute *snr_uncore_cha_formats_attr[] = {
+	&format_attr_event.attr,
+	&format_attr_umask_ext2.attr,
+	&format_attr_edge.attr,
+	&format_attr_tid_en.attr,
+	&format_attr_inv.attr,
+	&format_attr_thresh8.attr,
+	&format_attr_filter_tid5.attr,
+	NULL,
+};
+static const struct attribute_group snr_uncore_chabox_format_group = {
+	.name = "format",
+	.attrs = snr_uncore_cha_formats_attr,
+};
+
+static int snr_cha_hw_config(struct intel_uncore_box *box, struct perf_event *event)
+{
+	struct hw_perf_event_extra *reg1 = &event->hw.extra_reg;
+
+	reg1->reg = SNR_C0_MSR_PMON_BOX_FILTER0 +
+		    box->pmu->type->msr_offset * box->pmu->pmu_idx;
+	reg1->config = event->attr.config1 & SKX_CHA_MSR_PMON_BOX_FILTER_TID;
+	reg1->idx = 0;
+
+	return 0;
+}
+
+static void snr_cha_enable_event(struct intel_uncore_box *box,
+				   struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+	struct hw_perf_event_extra *reg1 = &hwc->extra_reg;
+
+	if (reg1->idx != EXTRA_REG_NONE)
+		wrmsrl(reg1->reg, reg1->config);
+
+	wrmsrl(hwc->config_base, hwc->config | SNBEP_PMON_CTL_EN);
+}
+
+static struct intel_uncore_ops snr_uncore_chabox_ops = {
+	.init_box		= ivbep_uncore_msr_init_box,
+	.disable_box		= snbep_uncore_msr_disable_box,
+	.enable_box		= snbep_uncore_msr_enable_box,
+	.disable_event		= snbep_uncore_msr_disable_event,
+	.enable_event		= snr_cha_enable_event,
+	.read_counter		= uncore_msr_read_counter,
+	.hw_config		= snr_cha_hw_config,
+};
+
+static struct intel_uncore_type snr_uncore_chabox = {
+	.name			= "cha",
+	.num_counters		= 4,
+	.num_boxes		= 6,
+	.perf_ctr_bits		= 48,
+	.event_ctl		= SNR_CHA_MSR_PMON_CTL0,
+	.perf_ctr		= SNR_CHA_MSR_PMON_CTR0,
+	.box_ctl		= SNR_CHA_MSR_PMON_BOX_CTL,
+	.msr_offset		= HSWEP_CBO_MSR_OFFSET,
+	.event_mask		= HSWEP_S_MSR_PMON_RAW_EVENT_MASK,
+	.event_mask_ext		= SNR_CHA_RAW_EVENT_MASK_EXT,
+	.ops			= &snr_uncore_chabox_ops,
+	.format_group		= &snr_uncore_chabox_format_group,
+};
+
+static struct attribute *snr_uncore_iio_formats_attr[] = {
+	&format_attr_event.attr,
+	&format_attr_umask.attr,
+	&format_attr_edge.attr,
+	&format_attr_inv.attr,
+	&format_attr_thresh9.attr,
+	&format_attr_ch_mask2.attr,
+	&format_attr_fc_mask2.attr,
+	NULL,
+};
+
+static const struct attribute_group snr_uncore_iio_format_group = {
+	.name = "format",
+	.attrs = snr_uncore_iio_formats_attr,
+};
+
+static struct intel_uncore_type snr_uncore_iio = {
+	.name			= "iio",
+	.num_counters		= 4,
+	.num_boxes		= 5,
+	.perf_ctr_bits		= 48,
+	.event_ctl		= SNR_IIO_MSR_PMON_CTL0,
+	.perf_ctr		= SNR_IIO_MSR_PMON_CTR0,
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.event_mask_ext		= SNR_IIO_PMON_RAW_EVENT_MASK_EXT,
+	.box_ctl		= SNR_IIO_MSR_PMON_BOX_CTL,
+	.msr_offset		= SNR_IIO_MSR_OFFSET,
+	.ops			= &ivbep_uncore_msr_ops,
+	.format_group		= &snr_uncore_iio_format_group,
+};
+
+static struct intel_uncore_type snr_uncore_irp = {
+	.name			= "irp",
+	.num_counters		= 2,
+	.num_boxes		= 5,
+	.perf_ctr_bits		= 48,
+	.event_ctl		= SNR_IRP0_MSR_PMON_CTL0,
+	.perf_ctr		= SNR_IRP0_MSR_PMON_CTR0,
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.box_ctl		= SNR_IRP0_MSR_PMON_BOX_CTL,
+	.msr_offset		= SNR_IRP_MSR_OFFSET,
+	.ops			= &ivbep_uncore_msr_ops,
+	.format_group		= &ivbep_uncore_format_group,
+};
+
+static struct intel_uncore_type snr_uncore_m2pcie = {
+	.name		= "m2pcie",
+	.num_counters	= 4,
+	.num_boxes	= 5,
+	.perf_ctr_bits	= 48,
+	.event_ctl	= SNR_M2PCIE_MSR_PMON_CTL0,
+	.perf_ctr	= SNR_M2PCIE_MSR_PMON_CTR0,
+	.box_ctl	= SNR_M2PCIE_MSR_PMON_BOX_CTL,
+	.msr_offset	= SNR_M2PCIE_MSR_OFFSET,
+	.event_mask	= SNBEP_PMON_RAW_EVENT_MASK,
+	.ops		= &ivbep_uncore_msr_ops,
+	.format_group	= &ivbep_uncore_format_group,
+};
+
+static int snr_pcu_hw_config(struct intel_uncore_box *box, struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+	struct hw_perf_event_extra *reg1 = &hwc->extra_reg;
+	int ev_sel = hwc->config & SNBEP_PMON_CTL_EV_SEL_MASK;
+
+	if (ev_sel >= 0xb && ev_sel <= 0xe) {
+		reg1->reg = SNR_PCU_MSR_PMON_BOX_FILTER;
+		reg1->idx = ev_sel - 0xb;
+		reg1->config = event->attr.config1 & (0xff << reg1->idx);
+	}
+	return 0;
+}
+
+static struct intel_uncore_ops snr_uncore_pcu_ops = {
+	IVBEP_UNCORE_MSR_OPS_COMMON_INIT(),
+	.hw_config		= snr_pcu_hw_config,
+	.get_constraint		= snbep_pcu_get_constraint,
+	.put_constraint		= snbep_pcu_put_constraint,
+};
+
+static struct intel_uncore_type snr_uncore_pcu = {
+	.name			= "pcu",
+	.num_counters		= 4,
+	.num_boxes		= 1,
+	.perf_ctr_bits		= 48,
+	.perf_ctr		= SNR_PCU_MSR_PMON_CTR0,
+	.event_ctl		= SNR_PCU_MSR_PMON_CTL0,
+	.event_mask		= SNBEP_PMON_RAW_EVENT_MASK,
+	.box_ctl		= SNR_PCU_MSR_PMON_BOX_CTL,
+	.num_shared_regs	= 1,
+	.ops			= &snr_uncore_pcu_ops,
+	.format_group		= &skx_uncore_pcu_format_group,
+};
+
+enum perf_uncore_snr_iio_freerunning_type_id {
+	SNR_IIO_MSR_IOCLK,
+	SNR_IIO_MSR_BW_IN,
+
+	SNR_IIO_FREERUNNING_TYPE_MAX,
+};
+
+static struct freerunning_counters snr_iio_freerunning[] = {
+	[SNR_IIO_MSR_IOCLK]	= { 0x1eac, 0x1, 0x10, 1, 48 },
+	[SNR_IIO_MSR_BW_IN]	= { 0x1f00, 0x1, 0x10, 8, 48 },
+};
+
+static struct uncore_event_desc snr_uncore_iio_freerunning_events[] = {
+	/* Free-Running IIO CLOCKS Counter */
+	INTEL_UNCORE_EVENT_DESC(ioclk,			"event=0xff,umask=0x10"),
+	/* Free-Running IIO BANDWIDTH IN Counters */
+	INTEL_UNCORE_EVENT_DESC(bw_in_port0,		"event=0xff,umask=0x20"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port0.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port0.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port1,		"event=0xff,umask=0x21"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port1.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port1.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port2,		"event=0xff,umask=0x22"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port2.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port2.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port3,		"event=0xff,umask=0x23"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port3.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port3.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port4,		"event=0xff,umask=0x24"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port4.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port4.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port5,		"event=0xff,umask=0x25"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port5.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port5.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port6,		"event=0xff,umask=0x26"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port6.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port6.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port7,		"event=0xff,umask=0x27"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port7.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(bw_in_port7.unit,	"MiB"),
+	{ /* end: all zeroes */ },
+};
+
+static struct intel_uncore_type snr_uncore_iio_free_running = {
+	.name			= "iio_free_running",
+	.num_counters		= 9,
+	.num_boxes		= 5,
+	.num_freerunning_types	= SNR_IIO_FREERUNNING_TYPE_MAX,
+	.freerunning		= snr_iio_freerunning,
+	.ops			= &skx_uncore_iio_freerunning_ops,
+	.event_descs		= snr_uncore_iio_freerunning_events,
+	.format_group		= &skx_uncore_iio_freerunning_format_group,
+};
+
+static struct intel_uncore_type *snr_msr_uncores[] = {
+	&snr_uncore_ubox,
+	&snr_uncore_chabox,
+	&snr_uncore_iio,
+	&snr_uncore_irp,
+	&snr_uncore_m2pcie,
+	&snr_uncore_pcu,
+	&snr_uncore_iio_free_running,
+	NULL,
+};
+
+void snr_uncore_cpu_init(void)
+{
+	uncore_msr_uncores = snr_msr_uncores;
+}
+
+static void snr_m2m_uncore_pci_init_box(struct intel_uncore_box *box)
+{
+	struct pci_dev *pdev = box->pci_dev;
+	int box_ctl = uncore_pci_box_ctl(box);
+
+	__set_bit(UNCORE_BOX_FLAG_CTL_OFFS8, &box->flags);
+	pci_write_config_dword(pdev, box_ctl, IVBEP_PMON_BOX_CTL_INT);
+}
+
+static struct intel_uncore_ops snr_m2m_uncore_pci_ops = {
+	.init_box	= snr_m2m_uncore_pci_init_box,
+	.disable_box	= snbep_uncore_pci_disable_box,
+	.enable_box	= snbep_uncore_pci_enable_box,
+	.disable_event	= snbep_uncore_pci_disable_event,
+	.enable_event	= snbep_uncore_pci_enable_event,
+	.read_counter	= snbep_uncore_pci_read_counter,
+};
+
+static struct attribute *snr_m2m_uncore_formats_attr[] = {
+	&format_attr_event.attr,
+	&format_attr_umask_ext3.attr,
+	&format_attr_edge.attr,
+	&format_attr_inv.attr,
+	&format_attr_thresh8.attr,
+	NULL,
+};
+
+static const struct attribute_group snr_m2m_uncore_format_group = {
+	.name = "format",
+	.attrs = snr_m2m_uncore_formats_attr,
+};
+
+static struct intel_uncore_type snr_uncore_m2m = {
+	.name		= "m2m",
+	.num_counters   = 4,
+	.num_boxes	= 1,
+	.perf_ctr_bits	= 48,
+	.perf_ctr	= SNR_M2M_PCI_PMON_CTR0,
+	.event_ctl	= SNR_M2M_PCI_PMON_CTL0,
+	.event_mask	= SNBEP_PMON_RAW_EVENT_MASK,
+	.event_mask_ext	= SNR_M2M_PCI_PMON_UMASK_EXT,
+	.box_ctl	= SNR_M2M_PCI_PMON_BOX_CTL,
+	.ops		= &snr_m2m_uncore_pci_ops,
+	.format_group	= &snr_m2m_uncore_format_group,
+};
+
+static struct intel_uncore_type snr_uncore_pcie3 = {
+	.name		= "pcie3",
+	.num_counters	= 4,
+	.num_boxes	= 1,
+	.perf_ctr_bits	= 48,
+	.perf_ctr	= SNR_PCIE3_PCI_PMON_CTR0,
+	.event_ctl	= SNR_PCIE3_PCI_PMON_CTL0,
+	.event_mask	= SNBEP_PMON_RAW_EVENT_MASK,
+	.box_ctl	= SNR_PCIE3_PCI_PMON_BOX_CTL,
+	.ops		= &ivbep_uncore_pci_ops,
+	.format_group	= &ivbep_uncore_format_group,
+};
+
+enum {
+	SNR_PCI_UNCORE_M2M,
+	SNR_PCI_UNCORE_PCIE3,
+};
+
+static struct intel_uncore_type *snr_pci_uncores[] = {
+	[SNR_PCI_UNCORE_M2M]		= &snr_uncore_m2m,
+	[SNR_PCI_UNCORE_PCIE3]		= &snr_uncore_pcie3,
+	NULL,
+};
+
+static const struct pci_device_id snr_uncore_pci_ids[] = {
+	{ /* M2M */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x344a),
+		.driver_data = UNCORE_PCI_DEV_FULL_DATA(12, 0, SNR_PCI_UNCORE_M2M, 0),
+	},
+	{ /* PCIe3 */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x334a),
+		.driver_data = UNCORE_PCI_DEV_FULL_DATA(4, 0, SNR_PCI_UNCORE_PCIE3, 0),
+	},
+	{ /* end: all zeroes */ }
+};
+
+static struct pci_driver snr_uncore_pci_driver = {
+	.name		= "snr_uncore",
+	.id_table	= snr_uncore_pci_ids,
+};
+
+int snr_uncore_pci_init(void)
+{
+	/* SNR UBOX DID */
+	int ret = snbep_pci2phy_map_init(0x3460, SKX_CPUNODEID,
+					 SKX_GIDNIDMAP, true);
+
+	if (ret)
+		return ret;
+
+	uncore_pci_uncores = snr_pci_uncores;
+	uncore_pci_driver = &snr_uncore_pci_driver;
+	return 0;
+}
+
+/* end of SNR uncore support */

[RESEND PATCH 3/6] perf/x86/intel/uncore: Extract codes of box ref/unref

[kan.liang]

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

For uncore box which can only be accessed by MSR, its reference
box->refcnt is updated in CPU hot plug. The uncore boxes needs to be
init/exit accordingly for the first/last CPU of a socket.
Starts from Snow Ridge server, a new type of uncore box is introduced,
which can only be accessed by MMIO. The driver needs to map/unmap
MMIO space for the first/last CPU of a socket.

Extract the codes of box ref/unref and init/exit for reuse later.

There is no functional change.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore.c | 55 +++++++++++++++++++++++++++---------------
 1 file changed, 36 insertions(+), 19 deletions(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index ee23b50..0b72ca5 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -1140,12 +1140,27 @@ static void uncore_change_context(struct intel_uncore_type **uncores,
 		uncore_change_type_ctx(*uncores, old_cpu, new_cpu);
 }
 
-static int uncore_event_cpu_offline(unsigned int cpu)
+static void uncore_box_unref(struct intel_uncore_type **types, int id)
 {
-	struct intel_uncore_type *type, **types = uncore_msr_uncores;
+	struct intel_uncore_type *type;
 	struct intel_uncore_pmu *pmu;
 	struct intel_uncore_box *box;
-	int i, pkg, target;
+	int i;
+
+	for (; *types; types++) {
+		type = *types;
+		pmu = type->pmus;
+		for (i = 0; i < type->num_boxes; i++, pmu++) {
+			box = pmu->boxes[id];
+			if (box && atomic_dec_return(&box->refcnt) == 0)
+				uncore_box_exit(box);
+		}
+	}
+}
+
+static int uncore_event_cpu_offline(unsigned int cpu)
+{
+	int pkg, target;
 
 	/* Check if exiting cpu is used for collecting uncore events */
 	if (!cpumask_test_and_clear_cpu(cpu, &uncore_cpu_mask))
@@ -1165,15 +1180,7 @@ static int uncore_event_cpu_offline(unsigned int cpu)
 unref:
 	/* Clear the references */
 	pkg = topology_logical_package_id(cpu);
-	for (; *types; types++) {
-		type = *types;
-		pmu = type->pmus;
-		for (i = 0; i < type->num_boxes; i++, pmu++) {
-			box = pmu->boxes[pkg];
-			if (box && atomic_dec_return(&box->refcnt) == 0)
-				uncore_box_exit(box);
-		}
-	}
+	uncore_box_unref(uncore_msr_uncores, pkg);
 	return 0;
 }
 
@@ -1215,16 +1222,15 @@ static int allocate_boxes(struct intel_uncore_type **types,
 	}
 	return -ENOMEM;
 }
-
-static int uncore_event_cpu_online(unsigned int cpu)
+static int uncore_box_ref(struct intel_uncore_type **types,
+			  int id, unsigned int cpu)
 {
-	struct intel_uncore_type *type, **types = uncore_msr_uncores;
+	struct intel_uncore_type *type;
 	struct intel_uncore_pmu *pmu;
 	struct intel_uncore_box *box;
-	int i, ret, pkg, target;
+	int i, ret;
 
-	pkg = topology_logical_package_id(cpu);
-	ret = allocate_boxes(types, pkg, cpu);
+	ret = allocate_boxes(types, id, cpu);
 	if (ret)
 		return ret;
 
@@ -1232,11 +1238,22 @@ static int uncore_event_cpu_online(unsigned int cpu)
 		type = *types;
 		pmu = type->pmus;
 		for (i = 0; i < type->num_boxes; i++, pmu++) {
-			box = pmu->boxes[pkg];
+			box = pmu->boxes[id];
 			if (box && atomic_inc_return(&box->refcnt) == 1)
 				uncore_box_init(box);
 		}
 	}
+	return 0;
+}
+
+static int uncore_event_cpu_online(unsigned int cpu)
+{
+	int ret, pkg, target;
+
+	pkg = topology_logical_package_id(cpu);
+	ret = uncore_box_ref(uncore_msr_uncores, pkg, cpu);
+	if (ret)
+		return ret;
 
 	/*
 	 * Check if there is an online cpu in the package
-- 
2.7.4

Re: [RESEND PATCH 3/6] perf/x86/intel/uncore: Extract codes of box ref/unref

[Peter Zijlstra]

On Tue, Apr 30, 2019 at 05:53:45PM -0700, kan.liang@linux.intel.com wrote:
> +static void uncore_box_unref(struct intel_uncore_type **types, int id)
>  {
> +	struct intel_uncore_type *type;
>  	struct intel_uncore_pmu *pmu;
>  	struct intel_uncore_box *box;
> +	int i;
> +
> +	for (; *types; types++) {
> +		type = *types;
> +		pmu = type->pmus;
> +		for (i = 0; i < type->num_boxes; i++, pmu++) {
> +			box = pmu->boxes[id];
> +			if (box && atomic_dec_return(&box->refcnt) == 0)
> +				uncore_box_exit(box);
> +		}
> +	}
> +}

> +static int uncore_box_ref(struct intel_uncore_type **types,
> +			  int id, unsigned int cpu)
>  {
> +	struct intel_uncore_type *type;
>  	struct intel_uncore_pmu *pmu;
>  	struct intel_uncore_box *box;
> +	int i, ret;
>  
> +	ret = allocate_boxes(types, id, cpu);
>  	if (ret)
>  		return ret;
>  
> @@ -1232,11 +1238,22 @@ static int uncore_event_cpu_online(unsigned int cpu)
>  		type = *types;
>  		pmu = type->pmus;
>  		for (i = 0; i < type->num_boxes; i++, pmu++) {
> +			box = pmu->boxes[id];
>  			if (box && atomic_inc_return(&box->refcnt) == 1)
>  				uncore_box_init(box);
>  		}
>  	}
> +	return 0;
> +}

This relies on all online/offline events to be globally serialized, and
they are. But that does make me wonder why we're using atomic_t.

Without the serialization there is an online-online race where the first
online sets the refcount to 1, lets the second online continue without
the first having completed the init().

Anyway, the code isn't wrong, just weird.

[tip:perf/core] perf/x86/intel/uncore: Factor out box ref/unref functions

[tip-bot for Kan Liang]

Commit-ID:  c8872d90e0a3651a096860d3241625ccfa1647e0
Gitweb:     https://git.kernel.org/tip/c8872d90e0a3651a096860d3241625ccfa1647e0
Author:     Kan Liang <kan.liang@linux.intel.com>
AuthorDate: Tue, 30 Apr 2019 17:53:45 -0700
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Mon, 17 Jun 2019 12:36:19 +0200

perf/x86/intel/uncore: Factor out box ref/unref functions

For uncore box which can only be accessed by MSR, its reference
box->refcnt is updated in CPU hot plug. The uncore boxes need to be
initalized and exited accordingly for the first/last CPU of a socket.

Starts from Snow Ridge server, a new type of uncore box is introduced,
which can only be accessed by MMIO. The driver needs to map/unmap
MMIO space for the first/last CPU of a socket.

Extract the codes of box ref/unref and init/exit for reuse later.

There is no functional change.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: acme@kernel.org
Cc: eranian@google.com
Link: https://lkml.kernel.org/r/1556672028-119221-4-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/events/intel/uncore.c | 54 ++++++++++++++++++++++++++++--------------
 1 file changed, 36 insertions(+), 18 deletions(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 342c323e0f6a..7b0c88903d47 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -1143,12 +1143,27 @@ static void uncore_change_context(struct intel_uncore_type **uncores,
 		uncore_change_type_ctx(*uncores, old_cpu, new_cpu);
 }
 
-static int uncore_event_cpu_offline(unsigned int cpu)
+static void uncore_box_unref(struct intel_uncore_type **types, int id)
 {
-	struct intel_uncore_type *type, **types = uncore_msr_uncores;
+	struct intel_uncore_type *type;
 	struct intel_uncore_pmu *pmu;
 	struct intel_uncore_box *box;
-	int i, die, target;
+	int i;
+
+	for (; *types; types++) {
+		type = *types;
+		pmu = type->pmus;
+		for (i = 0; i < type->num_boxes; i++, pmu++) {
+			box = pmu->boxes[id];
+			if (box && atomic_dec_return(&box->refcnt) == 0)
+				uncore_box_exit(box);
+		}
+	}
+}
+
+static int uncore_event_cpu_offline(unsigned int cpu)
+{
+	int die, target;
 
 	/* Check if exiting cpu is used for collecting uncore events */
 	if (!cpumask_test_and_clear_cpu(cpu, &uncore_cpu_mask))
@@ -1168,15 +1183,7 @@ static int uncore_event_cpu_offline(unsigned int cpu)
 unref:
 	/* Clear the references */
 	die = topology_logical_die_id(cpu);
-	for (; *types; types++) {
-		type = *types;
-		pmu = type->pmus;
-		for (i = 0; i < type->num_boxes; i++, pmu++) {
-			box = pmu->boxes[die];
-			if (box && atomic_dec_return(&box->refcnt) == 0)
-				uncore_box_exit(box);
-		}
-	}
+	uncore_box_unref(uncore_msr_uncores, die);
 	return 0;
 }
 
@@ -1219,15 +1226,15 @@ cleanup:
 	return -ENOMEM;
 }
 
-static int uncore_event_cpu_online(unsigned int cpu)
+static int uncore_box_ref(struct intel_uncore_type **types,
+			  int id, unsigned int cpu)
 {
-	struct intel_uncore_type *type, **types = uncore_msr_uncores;
+	struct intel_uncore_type *type;
 	struct intel_uncore_pmu *pmu;
 	struct intel_uncore_box *box;
-	int i, ret, die, target;
+	int i, ret;
 
-	die = topology_logical_die_id(cpu);
-	ret = allocate_boxes(types, die, cpu);
+	ret = allocate_boxes(types, id, cpu);
 	if (ret)
 		return ret;
 
@@ -1235,11 +1242,22 @@ static int uncore_event_cpu_online(unsigned int cpu)
 		type = *types;
 		pmu = type->pmus;
 		for (i = 0; i < type->num_boxes; i++, pmu++) {
-			box = pmu->boxes[die];
+			box = pmu->boxes[id];
 			if (box && atomic_inc_return(&box->refcnt) == 1)
 				uncore_box_init(box);
 		}
 	}
+	return 0;
+}
+
+static int uncore_event_cpu_online(unsigned int cpu)
+{
+	int ret, die, target;
+
+	die = topology_logical_die_id(cpu);
+	ret = uncore_box_ref(uncore_msr_uncores, die, cpu);
+	if (ret)
+		return ret;
 
 	/*
 	 * Check if there is an online cpu in the package

[RESEND PATCH 4/6] perf/x86/intel/uncore: Support MMIO type uncore blocks

[kan.liang]

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

A new MMIO type uncore box is introduced on Snow Ridge server. The
counters of MMIO type uncore box can only be accessed by MMIO.

Add a new uncore type, uncore_mmio_uncores, for MMIO type uncore blocks.

Support MMIO type uncore blocks in CPU hot plug. The MMIO space has to
be map/unmap for the first/last CPU. The context also need to be
migrated if the bind CPU changes.

Add mmio_init() to init and register PMUs for MMIO type uncore blocks.

Add a helper to calculate the box_ctl address.

The helpers which calculate ctl/ctr can be shared with PCI type uncore
blocks.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore.c | 51 ++++++++++++++++++++++++++++++++++++------
 arch/x86/events/intel/uncore.h | 21 ++++++++++++-----
 2 files changed, 60 insertions(+), 12 deletions(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 0b72ca5..3c00635 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -7,6 +7,7 @@
 static struct intel_uncore_type *empty_uncore[] = { NULL, };
 struct intel_uncore_type **uncore_msr_uncores = empty_uncore;
 struct intel_uncore_type **uncore_pci_uncores = empty_uncore;
+struct intel_uncore_type **uncore_mmio_uncores = empty_uncore;
 
 static bool pcidrv_registered;
 struct pci_driver *uncore_pci_driver;
@@ -1175,12 +1176,14 @@ static int uncore_event_cpu_offline(unsigned int cpu)
 		target = -1;
 
 	uncore_change_context(uncore_msr_uncores, cpu, target);
+	uncore_change_context(uncore_mmio_uncores, cpu, target);
 	uncore_change_context(uncore_pci_uncores, cpu, target);
 
 unref:
 	/* Clear the references */
 	pkg = topology_logical_package_id(cpu);
 	uncore_box_unref(uncore_msr_uncores, pkg);
+	uncore_box_unref(uncore_mmio_uncores, pkg);
 	return 0;
 }
 
@@ -1248,12 +1251,13 @@ static int uncore_box_ref(struct intel_uncore_type **types,
 
 static int uncore_event_cpu_online(unsigned int cpu)
 {
-	int ret, pkg, target;
+	int pkg, target, msr_ret, mmio_ret;
 
 	pkg = topology_logical_package_id(cpu);
-	ret = uncore_box_ref(uncore_msr_uncores, pkg, cpu);
-	if (ret)
-		return ret;
+	msr_ret = uncore_box_ref(uncore_msr_uncores, pkg, cpu);
+	mmio_ret = uncore_box_ref(uncore_mmio_uncores, pkg, cpu);
+	if (msr_ret && mmio_ret)
+		return -ENOMEM;
 
 	/*
 	 * Check if there is an online cpu in the package
@@ -1265,7 +1269,10 @@ static int uncore_event_cpu_online(unsigned int cpu)
 
 	cpumask_set_cpu(cpu, &uncore_cpu_mask);
 
-	uncore_change_context(uncore_msr_uncores, -1, cpu);
+	if (!msr_ret)
+		uncore_change_context(uncore_msr_uncores, -1, cpu);
+	if (!mmio_ret)
+		uncore_change_context(uncore_mmio_uncores, -1, cpu);
 	uncore_change_context(uncore_pci_uncores, -1, cpu);
 	return 0;
 }
@@ -1313,12 +1320,35 @@ static int __init uncore_cpu_init(void)
 	return ret;
 }
 
+static int __init uncore_mmio_init(void)
+{
+	struct intel_uncore_type **types = uncore_mmio_uncores;
+	int ret;
+
+	ret = uncore_types_init(types, true);
+	if (ret)
+		goto err;
+
+	for (; *types; types++) {
+		ret = type_pmu_register(*types);
+		if (ret)
+			goto err;
+	}
+	return 0;
+err:
+	uncore_types_exit(uncore_mmio_uncores);
+	uncore_mmio_uncores = empty_uncore;
+	return ret;
+}
+
+
 #define X86_UNCORE_MODEL_MATCH(model, init)	\
 	{ X86_VENDOR_INTEL, 6, model, X86_FEATURE_ANY, (unsigned long)&init }
 
 struct intel_uncore_init_fun {
 	void	(*cpu_init)(void);
 	int	(*pci_init)(void);
+	void	(*mmio_init)(void);
 };
 
 static const struct intel_uncore_init_fun nhm_uncore_init __initconst = {
@@ -1431,7 +1461,7 @@ static int __init intel_uncore_init(void)
 {
 	const struct x86_cpu_id *id;
 	struct intel_uncore_init_fun *uncore_init;
-	int pret = 0, cret = 0, ret;
+	int pret = 0, cret = 0, mret = 0, ret;
 
 	id = x86_match_cpu(intel_uncore_match);
 	if (!id)
@@ -1454,7 +1484,12 @@ static int __init intel_uncore_init(void)
 		cret = uncore_cpu_init();
 	}
 
-	if (cret && pret)
+	if (uncore_init->mmio_init) {
+		uncore_init->mmio_init();
+		mret = uncore_mmio_init();
+	}
+
+	if (cret && pret && mret)
 		return -ENODEV;
 
 	/* Install hotplug callbacks to setup the targets for each package */
@@ -1468,6 +1503,7 @@ static int __init intel_uncore_init(void)
 
 err:
 	uncore_types_exit(uncore_msr_uncores);
+	uncore_types_exit(uncore_mmio_uncores);
 	uncore_pci_exit();
 	return ret;
 }
@@ -1477,6 +1513,7 @@ static void __exit intel_uncore_exit(void)
 {
 	cpuhp_remove_state(CPUHP_AP_PERF_X86_UNCORE_ONLINE);
 	uncore_types_exit(uncore_msr_uncores);
+	uncore_types_exit(uncore_mmio_uncores);
 	uncore_pci_exit();
 }
 module_exit(intel_uncore_exit);
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 5e97e5e..426a490 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -56,7 +56,10 @@ struct intel_uncore_type {
 	unsigned fixed_ctr;
 	unsigned fixed_ctl;
 	unsigned box_ctl;
-	unsigned msr_offset;
+	union {
+		unsigned msr_offset;
+		unsigned mmio_offset;
+	};
 	unsigned num_shared_regs:8;
 	unsigned single_fixed:1;
 	unsigned pair_ctr_ctl:1;
@@ -190,6 +193,13 @@ static inline bool uncore_pmc_freerunning(int idx)
 	return idx == UNCORE_PMC_IDX_FREERUNNING;
 }
 
+static inline
+unsigned int uncore_mmio_box_ctl(struct intel_uncore_box *box)
+{
+	return box->pmu->type->box_ctl +
+	       box->pmu->type->mmio_offset * box->pmu->pmu_idx;
+}
+
 static inline unsigned uncore_pci_box_ctl(struct intel_uncore_box *box)
 {
 	return box->pmu->type->box_ctl;
@@ -330,7 +340,7 @@ unsigned uncore_msr_perf_ctr(struct intel_uncore_box *box, int idx)
 static inline
 unsigned uncore_fixed_ctl(struct intel_uncore_box *box)
 {
-	if (box->pci_dev)
+	if (box->pci_dev || box->io_addr)
 		return uncore_pci_fixed_ctl(box);
 	else
 		return uncore_msr_fixed_ctl(box);
@@ -339,7 +349,7 @@ unsigned uncore_fixed_ctl(struct intel_uncore_box *box)
 static inline
 unsigned uncore_fixed_ctr(struct intel_uncore_box *box)
 {
-	if (box->pci_dev)
+	if (box->pci_dev || box->io_addr)
 		return uncore_pci_fixed_ctr(box);
 	else
 		return uncore_msr_fixed_ctr(box);
@@ -348,7 +358,7 @@ unsigned uncore_fixed_ctr(struct intel_uncore_box *box)
 static inline
 unsigned uncore_event_ctl(struct intel_uncore_box *box, int idx)
 {
-	if (box->pci_dev)
+	if (box->pci_dev || box->io_addr)
 		return uncore_pci_event_ctl(box, idx);
 	else
 		return uncore_msr_event_ctl(box, idx);
@@ -357,7 +367,7 @@ unsigned uncore_event_ctl(struct intel_uncore_box *box, int idx)
 static inline
 unsigned uncore_perf_ctr(struct intel_uncore_box *box, int idx)
 {
-	if (box->pci_dev)
+	if (box->pci_dev || box->io_addr)
 		return uncore_pci_perf_ctr(box, idx);
 	else
 		return uncore_msr_perf_ctr(box, idx);
@@ -507,6 +517,7 @@ u64 uncore_shared_reg_config(struct intel_uncore_box *box, int idx);
 
 extern struct intel_uncore_type **uncore_msr_uncores;
 extern struct intel_uncore_type **uncore_pci_uncores;
+extern struct intel_uncore_type **uncore_mmio_uncores;
 extern struct pci_driver *uncore_pci_driver;
 extern raw_spinlock_t pci2phy_map_lock;
 extern struct list_head pci2phy_map_head;
-- 
2.7.4

[tip:perf/core] perf/x86/intel/uncore: Support MMIO type uncore blocks

[tip-bot for Kan Liang]

Commit-ID:  3da04b8a00dd6d39970b9e764b78c5dfb40ec013
Gitweb:     https://git.kernel.org/tip/3da04b8a00dd6d39970b9e764b78c5dfb40ec013
Author:     Kan Liang <kan.liang@linux.intel.com>
AuthorDate: Tue, 30 Apr 2019 17:53:46 -0700
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Mon, 17 Jun 2019 12:36:20 +0200

perf/x86/intel/uncore: Support MMIO type uncore blocks

A new MMIO type uncore box is introduced on Snow Ridge server. The
counters of MMIO type uncore box can only be accessed by MMIO.

Add a new uncore type, uncore_mmio_uncores, for MMIO type uncore blocks.

Support MMIO type uncore blocks in CPU hot plug. The MMIO space has to
be map/unmap for the first/last CPU. The context also need to be
migrated if the bind CPU changes.

Add mmio_init() to init and register PMUs for MMIO type uncore blocks.

Add a helper to calculate the box_ctl address.

The helpers which calculate ctl/ctr can be shared with PCI type uncore
blocks.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: acme@kernel.org
Cc: eranian@google.com
Link: https://lkml.kernel.org/r/1556672028-119221-5-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/events/intel/uncore.c | 51 ++++++++++++++++++++++++++++++++++++------
 arch/x86/events/intel/uncore.h | 21 ++++++++++++-----
 2 files changed, 60 insertions(+), 12 deletions(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 7b0c88903d47..442f30e5e49a 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -8,6 +8,7 @@
 static struct intel_uncore_type *empty_uncore[] = { NULL, };
 struct intel_uncore_type **uncore_msr_uncores = empty_uncore;
 struct intel_uncore_type **uncore_pci_uncores = empty_uncore;
+struct intel_uncore_type **uncore_mmio_uncores = empty_uncore;
 
 static bool pcidrv_registered;
 struct pci_driver *uncore_pci_driver;
@@ -1178,12 +1179,14 @@ static int uncore_event_cpu_offline(unsigned int cpu)
 		target = -1;
 
 	uncore_change_context(uncore_msr_uncores, cpu, target);
+	uncore_change_context(uncore_mmio_uncores, cpu, target);
 	uncore_change_context(uncore_pci_uncores, cpu, target);
 
 unref:
 	/* Clear the references */
 	die = topology_logical_die_id(cpu);
 	uncore_box_unref(uncore_msr_uncores, die);
+	uncore_box_unref(uncore_mmio_uncores, die);
 	return 0;
 }
 
@@ -1252,12 +1255,13 @@ static int uncore_box_ref(struct intel_uncore_type **types,
 
 static int uncore_event_cpu_online(unsigned int cpu)
 {
-	int ret, die, target;
+	int die, target, msr_ret, mmio_ret;
 
 	die = topology_logical_die_id(cpu);
-	ret = uncore_box_ref(uncore_msr_uncores, die, cpu);
-	if (ret)
-		return ret;
+	msr_ret = uncore_box_ref(uncore_msr_uncores, die, cpu);
+	mmio_ret = uncore_box_ref(uncore_mmio_uncores, die, cpu);
+	if (msr_ret && mmio_ret)
+		return -ENOMEM;
 
 	/*
 	 * Check if there is an online cpu in the package
@@ -1269,7 +1273,10 @@ static int uncore_event_cpu_online(unsigned int cpu)
 
 	cpumask_set_cpu(cpu, &uncore_cpu_mask);
 
-	uncore_change_context(uncore_msr_uncores, -1, cpu);
+	if (!msr_ret)
+		uncore_change_context(uncore_msr_uncores, -1, cpu);
+	if (!mmio_ret)
+		uncore_change_context(uncore_mmio_uncores, -1, cpu);
 	uncore_change_context(uncore_pci_uncores, -1, cpu);
 	return 0;
 }
@@ -1317,12 +1324,35 @@ err:
 	return ret;
 }
 
+static int __init uncore_mmio_init(void)
+{
+	struct intel_uncore_type **types = uncore_mmio_uncores;
+	int ret;
+
+	ret = uncore_types_init(types, true);
+	if (ret)
+		goto err;
+
+	for (; *types; types++) {
+		ret = type_pmu_register(*types);
+		if (ret)
+			goto err;
+	}
+	return 0;
+err:
+	uncore_types_exit(uncore_mmio_uncores);
+	uncore_mmio_uncores = empty_uncore;
+	return ret;
+}
+
+
 #define X86_UNCORE_MODEL_MATCH(model, init)	\
 	{ X86_VENDOR_INTEL, 6, model, X86_FEATURE_ANY, (unsigned long)&init }
 
 struct intel_uncore_init_fun {
 	void	(*cpu_init)(void);
 	int	(*pci_init)(void);
+	void	(*mmio_init)(void);
 };
 
 static const struct intel_uncore_init_fun nhm_uncore_init __initconst = {
@@ -1437,7 +1467,7 @@ static int __init intel_uncore_init(void)
 {
 	const struct x86_cpu_id *id;
 	struct intel_uncore_init_fun *uncore_init;
-	int pret = 0, cret = 0, ret;
+	int pret = 0, cret = 0, mret = 0, ret;
 
 	id = x86_match_cpu(intel_uncore_match);
 	if (!id)
@@ -1460,7 +1490,12 @@ static int __init intel_uncore_init(void)
 		cret = uncore_cpu_init();
 	}
 
-	if (cret && pret)
+	if (uncore_init->mmio_init) {
+		uncore_init->mmio_init();
+		mret = uncore_mmio_init();
+	}
+
+	if (cret && pret && mret)
 		return -ENODEV;
 
 	/* Install hotplug callbacks to setup the targets for each package */
@@ -1474,6 +1509,7 @@ static int __init intel_uncore_init(void)
 
 err:
 	uncore_types_exit(uncore_msr_uncores);
+	uncore_types_exit(uncore_mmio_uncores);
 	uncore_pci_exit();
 	return ret;
 }
@@ -1483,6 +1519,7 @@ static void __exit intel_uncore_exit(void)
 {
 	cpuhp_remove_state(CPUHP_AP_PERF_X86_UNCORE_ONLINE);
 	uncore_types_exit(uncore_msr_uncores);
+	uncore_types_exit(uncore_mmio_uncores);
 	uncore_pci_exit();
 }
 module_exit(intel_uncore_exit);
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index b444ed35d6f7..fc5f48816005 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -56,7 +56,10 @@ struct intel_uncore_type {
 	unsigned fixed_ctr;
 	unsigned fixed_ctl;
 	unsigned box_ctl;
-	unsigned msr_offset;
+	union {
+		unsigned msr_offset;
+		unsigned mmio_offset;
+	};
 	unsigned num_shared_regs:8;
 	unsigned single_fixed:1;
 	unsigned pair_ctr_ctl:1;
@@ -190,6 +193,13 @@ static inline bool uncore_pmc_freerunning(int idx)
 	return idx == UNCORE_PMC_IDX_FREERUNNING;
 }
 
+static inline
+unsigned int uncore_mmio_box_ctl(struct intel_uncore_box *box)
+{
+	return box->pmu->type->box_ctl +
+	       box->pmu->type->mmio_offset * box->pmu->pmu_idx;
+}
+
 static inline unsigned uncore_pci_box_ctl(struct intel_uncore_box *box)
 {
 	return box->pmu->type->box_ctl;
@@ -330,7 +340,7 @@ unsigned uncore_msr_perf_ctr(struct intel_uncore_box *box, int idx)
 static inline
 unsigned uncore_fixed_ctl(struct intel_uncore_box *box)
 {
-	if (box->pci_dev)
+	if (box->pci_dev || box->io_addr)
 		return uncore_pci_fixed_ctl(box);
 	else
 		return uncore_msr_fixed_ctl(box);
@@ -339,7 +349,7 @@ unsigned uncore_fixed_ctl(struct intel_uncore_box *box)
 static inline
 unsigned uncore_fixed_ctr(struct intel_uncore_box *box)
 {
-	if (box->pci_dev)
+	if (box->pci_dev || box->io_addr)
 		return uncore_pci_fixed_ctr(box);
 	else
 		return uncore_msr_fixed_ctr(box);
@@ -348,7 +358,7 @@ unsigned uncore_fixed_ctr(struct intel_uncore_box *box)
 static inline
 unsigned uncore_event_ctl(struct intel_uncore_box *box, int idx)
 {
-	if (box->pci_dev)
+	if (box->pci_dev || box->io_addr)
 		return uncore_pci_event_ctl(box, idx);
 	else
 		return uncore_msr_event_ctl(box, idx);
@@ -357,7 +367,7 @@ unsigned uncore_event_ctl(struct intel_uncore_box *box, int idx)
 static inline
 unsigned uncore_perf_ctr(struct intel_uncore_box *box, int idx)
 {
-	if (box->pci_dev)
+	if (box->pci_dev || box->io_addr)
 		return uncore_pci_perf_ctr(box, idx);
 	else
 		return uncore_msr_perf_ctr(box, idx);
@@ -507,6 +517,7 @@ u64 uncore_shared_reg_config(struct intel_uncore_box *box, int idx);
 
 extern struct intel_uncore_type **uncore_msr_uncores;
 extern struct intel_uncore_type **uncore_pci_uncores;
+extern struct intel_uncore_type **uncore_mmio_uncores;
 extern struct pci_driver *uncore_pci_driver;
 extern raw_spinlock_t pci2phy_map_lock;
 extern struct list_head pci2phy_map_head;

[RESEND PATCH 5/6] perf/x86/intel/uncore: Clean up client IMC

[kan.liang]

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

The client IMC block is accessed by MMIO. Current code uses an informal
way to access the block, which is not recommended.

Cleaning up the code by using __iomem annotation and the accessor
functions (read[lq]()).
Move exit_box() and read_counter() to generic code, which can be shared
with server later.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore.c     | 15 +++++++++++++++
 arch/x86/events/intel/uncore.h     |  6 +++++-
 arch/x86/events/intel/uncore_snb.c | 16 ++--------------
 3 files changed, 22 insertions(+), 15 deletions(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 3c00635..39b0f96 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -119,6 +119,21 @@ u64 uncore_msr_read_counter(struct intel_uncore_box *box, struct perf_event *eve
 	return count;
 }
 
+void uncore_mmio_exit_box(struct intel_uncore_box *box)
+{
+	if (box->io_addr)
+		iounmap(box->io_addr);
+}
+
+u64 uncore_mmio_read_counter(struct intel_uncore_box *box,
+			     struct perf_event *event)
+{
+	if (!box->io_addr)
+		return 0;
+
+	return readq(box->io_addr + event->hw.event_base);
+}
+
 /*
  * generic get constraint function for shared match/mask registers.
  */
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 426a490..738bed3 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -2,6 +2,7 @@
 #include <linux/slab.h>
 #include <linux/pci.h>
 #include <asm/apicdef.h>
+#include <linux/io-64-nonatomic-lo-hi.h>
 
 #include <linux/perf_event.h>
 #include "../perf_event.h"
@@ -128,7 +129,7 @@ struct intel_uncore_box {
 	struct hrtimer hrtimer;
 	struct list_head list;
 	struct list_head active_list;
-	void *io_addr;
+	void __iomem *io_addr;
 	struct intel_uncore_extra_reg shared_regs[0];
 };
 
@@ -502,6 +503,9 @@ static inline struct intel_uncore_box *uncore_event_to_box(struct perf_event *ev
 
 struct intel_uncore_box *uncore_pmu_to_box(struct intel_uncore_pmu *pmu, int cpu);
 u64 uncore_msr_read_counter(struct intel_uncore_box *box, struct perf_event *event);
+void uncore_mmio_exit_box(struct intel_uncore_box *box);
+u64 uncore_mmio_read_counter(struct intel_uncore_box *box,
+			     struct perf_event *event);
 void uncore_pmu_start_hrtimer(struct intel_uncore_box *box);
 void uncore_pmu_cancel_hrtimer(struct intel_uncore_box *box);
 void uncore_pmu_event_start(struct perf_event *event, int flags);
diff --git a/arch/x86/events/intel/uncore_snb.c b/arch/x86/events/intel/uncore_snb.c
index f843181..5d0ce4347 100644
--- a/arch/x86/events/intel/uncore_snb.c
+++ b/arch/x86/events/intel/uncore_snb.c
@@ -420,11 +420,6 @@ static void snb_uncore_imc_init_box(struct intel_uncore_box *box)
 	box->hrtimer_duration = UNCORE_SNB_IMC_HRTIMER_INTERVAL;
 }
 
-static void snb_uncore_imc_exit_box(struct intel_uncore_box *box)
-{
-	iounmap(box->io_addr);
-}
-
 static void snb_uncore_imc_enable_box(struct intel_uncore_box *box)
 {}
 
@@ -437,13 +432,6 @@ static void snb_uncore_imc_enable_event(struct intel_uncore_box *box, struct per
 static void snb_uncore_imc_disable_event(struct intel_uncore_box *box, struct perf_event *event)
 {}
 
-static u64 snb_uncore_imc_read_counter(struct intel_uncore_box *box, struct perf_event *event)
-{
-	struct hw_perf_event *hwc = &event->hw;
-
-	return (u64)*(unsigned int *)(box->io_addr + hwc->event_base);
-}
-
 /*
  * Keep the custom event_init() function compatible with old event
  * encoding for free running counters.
@@ -570,13 +558,13 @@ static struct pmu snb_uncore_imc_pmu = {
 
 static struct intel_uncore_ops snb_uncore_imc_ops = {
 	.init_box	= snb_uncore_imc_init_box,
-	.exit_box	= snb_uncore_imc_exit_box,
+	.exit_box	= uncore_mmio_exit_box,
 	.enable_box	= snb_uncore_imc_enable_box,
 	.disable_box	= snb_uncore_imc_disable_box,
 	.disable_event	= snb_uncore_imc_disable_event,
 	.enable_event	= snb_uncore_imc_enable_event,
 	.hw_config	= snb_uncore_imc_hw_config,
-	.read_counter	= snb_uncore_imc_read_counter,
+	.read_counter	= uncore_mmio_read_counter,
 };
 
 static struct intel_uncore_type snb_uncore_imc = {
-- 
2.7.4

[tip:perf/core] perf/x86/intel/uncore: Clean up client IMC

[tip-bot for Kan Liang]

Commit-ID:  07ce734dd8adc0f170d43c15a9b91b707a21b9d7
Gitweb:     https://git.kernel.org/tip/07ce734dd8adc0f170d43c15a9b91b707a21b9d7
Author:     Kan Liang <kan.liang@linux.intel.com>
AuthorDate: Tue, 30 Apr 2019 17:53:47 -0700
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Mon, 17 Jun 2019 12:36:21 +0200

perf/x86/intel/uncore: Clean up client IMC

The client IMC block is accessed by MMIO. Current code uses an informal
way to access the block, which is not recommended.

Clean up the code by using __iomem annotation and the accessor
functions (read[lq]()).

Move exit_box() and read_counter() to generic code, which can be shared
with the server code later.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: acme@kernel.org
Cc: eranian@google.com
Link: https://lkml.kernel.org/r/1556672028-119221-6-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/events/intel/uncore.c     | 15 +++++++++++++++
 arch/x86/events/intel/uncore.h     |  6 +++++-
 arch/x86/events/intel/uncore_snb.c | 16 ++--------------
 3 files changed, 22 insertions(+), 15 deletions(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 442f30e5e49a..8bb537143d86 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -120,6 +120,21 @@ u64 uncore_msr_read_counter(struct intel_uncore_box *box, struct perf_event *eve
 	return count;
 }
 
+void uncore_mmio_exit_box(struct intel_uncore_box *box)
+{
+	if (box->io_addr)
+		iounmap(box->io_addr);
+}
+
+u64 uncore_mmio_read_counter(struct intel_uncore_box *box,
+			     struct perf_event *event)
+{
+	if (!box->io_addr)
+		return 0;
+
+	return readq(box->io_addr + event->hw.event_base);
+}
+
 /*
  * generic get constraint function for shared match/mask registers.
  */
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index fc5f48816005..cdd9691365a1 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -2,6 +2,7 @@
 #include <linux/slab.h>
 #include <linux/pci.h>
 #include <asm/apicdef.h>
+#include <linux/io-64-nonatomic-lo-hi.h>
 
 #include <linux/perf_event.h>
 #include "../perf_event.h"
@@ -128,7 +129,7 @@ struct intel_uncore_box {
 	struct hrtimer hrtimer;
 	struct list_head list;
 	struct list_head active_list;
-	void *io_addr;
+	void __iomem *io_addr;
 	struct intel_uncore_extra_reg shared_regs[0];
 };
 
@@ -502,6 +503,9 @@ static inline struct intel_uncore_box *uncore_event_to_box(struct perf_event *ev
 
 struct intel_uncore_box *uncore_pmu_to_box(struct intel_uncore_pmu *pmu, int cpu);
 u64 uncore_msr_read_counter(struct intel_uncore_box *box, struct perf_event *event);
+void uncore_mmio_exit_box(struct intel_uncore_box *box);
+u64 uncore_mmio_read_counter(struct intel_uncore_box *box,
+			     struct perf_event *event);
 void uncore_pmu_start_hrtimer(struct intel_uncore_box *box);
 void uncore_pmu_cancel_hrtimer(struct intel_uncore_box *box);
 void uncore_pmu_event_start(struct perf_event *event, int flags);
diff --git a/arch/x86/events/intel/uncore_snb.c b/arch/x86/events/intel/uncore_snb.c
index b0ca4f88c6f2..dbaa1b088a30 100644
--- a/arch/x86/events/intel/uncore_snb.c
+++ b/arch/x86/events/intel/uncore_snb.c
@@ -428,11 +428,6 @@ static void snb_uncore_imc_init_box(struct intel_uncore_box *box)
 	box->hrtimer_duration = UNCORE_SNB_IMC_HRTIMER_INTERVAL;
 }
 
-static void snb_uncore_imc_exit_box(struct intel_uncore_box *box)
-{
-	iounmap(box->io_addr);
-}
-
 static void snb_uncore_imc_enable_box(struct intel_uncore_box *box)
 {}
 
@@ -445,13 +440,6 @@ static void snb_uncore_imc_enable_event(struct intel_uncore_box *box, struct per
 static void snb_uncore_imc_disable_event(struct intel_uncore_box *box, struct perf_event *event)
 {}
 
-static u64 snb_uncore_imc_read_counter(struct intel_uncore_box *box, struct perf_event *event)
-{
-	struct hw_perf_event *hwc = &event->hw;
-
-	return (u64)*(unsigned int *)(box->io_addr + hwc->event_base);
-}
-
 /*
  * Keep the custom event_init() function compatible with old event
  * encoding for free running counters.
@@ -578,13 +566,13 @@ static struct pmu snb_uncore_imc_pmu = {
 
 static struct intel_uncore_ops snb_uncore_imc_ops = {
 	.init_box	= snb_uncore_imc_init_box,
-	.exit_box	= snb_uncore_imc_exit_box,
+	.exit_box	= uncore_mmio_exit_box,
 	.enable_box	= snb_uncore_imc_enable_box,
 	.disable_box	= snb_uncore_imc_disable_box,
 	.disable_event	= snb_uncore_imc_disable_event,
 	.enable_event	= snb_uncore_imc_enable_event,
 	.hw_config	= snb_uncore_imc_hw_config,
-	.read_counter	= snb_uncore_imc_read_counter,
+	.read_counter	= uncore_mmio_read_counter,
 };
 
 static struct intel_uncore_type snb_uncore_imc = {

[RESEND PATCH 6/6] perf/x86/intel/uncore: Add IMC uncore support for Snow Ridge

[kan.liang]

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

IMC uncore unit can only be accessed via MMIO on Snow Ridge.
The MMIO space of IMC uncore is at the specified offsets from the
MEM0_BAR. Add snr_uncore_get_mc_dev() to locate the PCI device with
MMIO_BASE and MEM0_BAR register.

Add new ops to access the IMC registers via MMIO.

Add 3 new free running counters for clocks, read and write bandwidth.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
---
 arch/x86/events/intel/uncore.c       |   3 +-
 arch/x86/events/intel/uncore.h       |   2 +
 arch/x86/events/intel/uncore_snbep.c | 197 +++++++++++++++++++++++++++++++++++
 3 files changed, 201 insertions(+), 1 deletion(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 39b0f96..f5db8dd 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -28,7 +28,7 @@ struct event_constraint uncore_constraint_empty =
 
 MODULE_LICENSE("GPL");
 
-static int uncore_pcibus_to_physid(struct pci_bus *bus)
+int uncore_pcibus_to_physid(struct pci_bus *bus)
 {
 	struct pci2phy_map *map;
 	int phys_id = -1;
@@ -1437,6 +1437,7 @@ static const struct intel_uncore_init_fun icl_uncore_init __initconst = {
 static const struct intel_uncore_init_fun snr_uncore_init __initconst = {
 	.cpu_init = snr_uncore_cpu_init,
 	.pci_init = snr_uncore_pci_init,
+	.mmio_init = snr_uncore_mmio_init,
 };
 
 static const struct x86_cpu_id intel_uncore_match[] __initconst = {
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index 738bed3..57641bf 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -163,6 +163,7 @@ struct pci2phy_map {
 };
 
 struct pci2phy_map *__find_pci2phy_map(int segment);
+int uncore_pcibus_to_physid(struct pci_bus *bus);
 
 ssize_t uncore_event_show(struct kobject *kobj,
 			  struct kobj_attribute *attr, char *buf);
@@ -555,6 +556,7 @@ int skx_uncore_pci_init(void);
 void skx_uncore_cpu_init(void);
 int snr_uncore_pci_init(void);
 void snr_uncore_cpu_init(void);
+void snr_uncore_mmio_init(void);
 
 /* uncore_nhmex.c */
 void nhmex_uncore_cpu_init(void);
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index 851f76f..7420e8b6 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -374,6 +374,19 @@
 #define SNR_PCIE3_PCI_PMON_CTR0			0x4e8
 #define SNR_PCIE3_PCI_PMON_BOX_CTL		0x4e4
 
+/* SNR IMC */
+#define SNR_IMC_MMIO_PMON_FIXED_CTL		0x54
+#define SNR_IMC_MMIO_PMON_FIXED_CTR		0x38
+#define SNR_IMC_MMIO_PMON_CTL0			0x40
+#define SNR_IMC_MMIO_PMON_CTR0			0x8
+#define SNR_IMC_MMIO_PMON_BOX_CTL		0x22800
+#define SNR_IMC_MMIO_OFFSET			0x4000
+#define SNR_IMC_MMIO_SIZE			0x4000
+#define SNR_IMC_MMIO_BASE_OFFSET		0xd0
+#define SNR_IMC_MMIO_BASE_MASK			0x1FFFFFFF
+#define SNR_IMC_MMIO_MEM0_OFFSET		0xd8
+#define SNR_IMC_MMIO_MEM0_MASK			0x7FF
+
 DEFINE_UNCORE_FORMAT_ATTR(event, event, "config:0-7");
 DEFINE_UNCORE_FORMAT_ATTR(event2, event, "config:0-6");
 DEFINE_UNCORE_FORMAT_ATTR(event_ext, event, "config:0-7,21");
@@ -4370,4 +4383,188 @@ int snr_uncore_pci_init(void)
 	return 0;
 }
 
+static struct pci_dev *snr_uncore_get_mc_dev(int id)
+{
+	struct pci_dev *mc_dev = NULL;
+	int phys_id, pkg;
+
+	while (1) {
+		mc_dev = pci_get_device(PCI_VENDOR_ID_INTEL, 0x3451, mc_dev);
+		if (!mc_dev)
+			break;
+		phys_id = uncore_pcibus_to_physid(mc_dev->bus);
+		if (phys_id < 0)
+			continue;
+		pkg = topology_phys_to_logical_pkg(phys_id);
+		if (pkg < 0)
+			continue;
+		else if (pkg == id)
+			break;
+	}
+	return mc_dev;
+}
+
+static void snr_uncore_mmio_init_box(struct intel_uncore_box *box)
+{
+	struct pci_dev *pdev = snr_uncore_get_mc_dev(box->pkgid);
+	unsigned int box_ctl = uncore_mmio_box_ctl(box);
+	resource_size_t addr;
+	u32 pci_dword;
+
+	if (!pdev)
+		return;
+
+	pci_read_config_dword(pdev, SNR_IMC_MMIO_BASE_OFFSET, &pci_dword);
+	addr = (pci_dword & SNR_IMC_MMIO_BASE_MASK) << 23;
+
+	pci_read_config_dword(pdev, SNR_IMC_MMIO_MEM0_OFFSET, &pci_dword);
+	addr |= (pci_dword & SNR_IMC_MMIO_MEM0_MASK) << 12;
+
+	addr += box_ctl;
+
+	box->io_addr = ioremap(addr, SNR_IMC_MMIO_SIZE);
+	if (!box->io_addr)
+		return;
+
+	writel(IVBEP_PMON_BOX_CTL_INT, box->io_addr);
+}
+
+static void snr_uncore_mmio_disable_box(struct intel_uncore_box *box)
+{
+	u32 config;
+
+	if (!box->io_addr)
+		return;
+
+	config = readl(box->io_addr);
+	config |= SNBEP_PMON_BOX_CTL_FRZ;
+	writel(config, box->io_addr);
+}
+
+static void snr_uncore_mmio_enable_box(struct intel_uncore_box *box)
+{
+	u32 config;
+
+	if (!box->io_addr)
+		return;
+
+	config = readl(box->io_addr);
+	config &= ~SNBEP_PMON_BOX_CTL_FRZ;
+	writel(config, box->io_addr);
+}
+
+static void snr_uncore_mmio_enable_event(struct intel_uncore_box *box,
+					   struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+
+	if (!box->io_addr)
+		return;
+
+	writel(hwc->config | SNBEP_PMON_CTL_EN,
+	       box->io_addr + hwc->config_base);
+}
+
+static void snr_uncore_mmio_disable_event(struct intel_uncore_box *box,
+					    struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+
+	if (!box->io_addr)
+		return;
+
+	writel(hwc->config, box->io_addr + hwc->config_base);
+}
+
+static struct intel_uncore_ops snr_uncore_mmio_ops = {
+	.init_box	= snr_uncore_mmio_init_box,
+	.exit_box	= uncore_mmio_exit_box,
+	.disable_box	= snr_uncore_mmio_disable_box,
+	.enable_box	= snr_uncore_mmio_enable_box,
+	.disable_event	= snr_uncore_mmio_disable_event,
+	.enable_event	= snr_uncore_mmio_enable_event,
+	.read_counter	= uncore_mmio_read_counter,
+};
+
+static struct uncore_event_desc snr_uncore_imc_events[] = {
+	INTEL_UNCORE_EVENT_DESC(clockticks,      "event=0x00,umask=0x00"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_read,  "event=0x04,umask=0x0f"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_read.scale, "6.103515625e-5"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_read.unit, "MiB"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_write, "event=0x04,umask=0x30"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_write.scale, "6.103515625e-5"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_write.unit, "MiB"),
+	{ /* end: all zeroes */ },
+};
+
+static struct intel_uncore_type snr_uncore_imc = {
+	.name		= "imc",
+	.num_counters   = 4,
+	.num_boxes	= 2,
+	.perf_ctr_bits	= 48,
+	.fixed_ctr_bits	= 48,
+	.fixed_ctr	= SNR_IMC_MMIO_PMON_FIXED_CTR,
+	.fixed_ctl	= SNR_IMC_MMIO_PMON_FIXED_CTL,
+	.event_descs	= snr_uncore_imc_events,
+	.perf_ctr	= SNR_IMC_MMIO_PMON_CTR0,
+	.event_ctl	= SNR_IMC_MMIO_PMON_CTL0,
+	.event_mask	= SNBEP_PMON_RAW_EVENT_MASK,
+	.box_ctl	= SNR_IMC_MMIO_PMON_BOX_CTL,
+	.mmio_offset	= SNR_IMC_MMIO_OFFSET,
+	.ops		= &snr_uncore_mmio_ops,
+	.format_group	= &skx_uncore_format_group,
+};
+
+enum perf_uncore_snr_imc_freerunning_type_id {
+	SNR_IMC_DCLK,
+	SNR_IMC_DDR,
+
+	SNR_IMC_FREERUNNING_TYPE_MAX,
+};
+
+static struct freerunning_counters snr_imc_freerunning[] = {
+	[SNR_IMC_DCLK]	= { 0x22b0, 0x0, 0, 1, 48 },
+	[SNR_IMC_DDR]	= { 0x2290, 0x8, 0, 2, 48 },
+};
+
+static struct uncore_event_desc snr_uncore_imc_freerunning_events[] = {
+	INTEL_UNCORE_EVENT_DESC(dclk,		"event=0xff,umask=0x10"),
+
+	INTEL_UNCORE_EVENT_DESC(read,		"event=0xff,umask=0x20"),
+	INTEL_UNCORE_EVENT_DESC(read.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(read.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(write,		"event=0xff,umask=0x21"),
+	INTEL_UNCORE_EVENT_DESC(write.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(write.unit,	"MiB"),
+};
+
+static struct intel_uncore_ops snr_uncore_imc_freerunning_ops = {
+	.init_box	= snr_uncore_mmio_init_box,
+	.exit_box	= uncore_mmio_exit_box,
+	.read_counter	= uncore_mmio_read_counter,
+	.hw_config	= uncore_freerunning_hw_config,
+};
+
+static struct intel_uncore_type snr_uncore_imc_free_running = {
+	.name			= "imc_free_running",
+	.num_counters		= 3,
+	.num_boxes		= 1,
+	.num_freerunning_types	= SNR_IMC_FREERUNNING_TYPE_MAX,
+	.freerunning		= snr_imc_freerunning,
+	.ops			= &snr_uncore_imc_freerunning_ops,
+	.event_descs		= snr_uncore_imc_freerunning_events,
+	.format_group		= &skx_uncore_iio_freerunning_format_group,
+};
+
+static struct intel_uncore_type *snr_mmio_uncores[] = {
+	&snr_uncore_imc,
+	&snr_uncore_imc_free_running,
+	NULL,
+};
+
+void snr_uncore_mmio_init(void)
+{
+	uncore_mmio_uncores = snr_mmio_uncores;
+}
+
 /* end of SNR uncore support */
-- 
2.7.4

[tip:perf/core] perf/x86/intel/uncore: Add IMC uncore support for Snow Ridge

[tip-bot for Kan Liang]

Commit-ID:  ee49532b38dd084650bf715eabe7e3828fb8d275
Gitweb:     https://git.kernel.org/tip/ee49532b38dd084650bf715eabe7e3828fb8d275
Author:     Kan Liang <kan.liang@linux.intel.com>
AuthorDate: Tue, 30 Apr 2019 17:53:48 -0700
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Mon, 17 Jun 2019 12:36:22 +0200

perf/x86/intel/uncore: Add IMC uncore support for Snow Ridge

IMC uncore unit can only be accessed via MMIO on Snow Ridge.
The MMIO space of IMC uncore is at the specified offsets from the
MEM0_BAR. Add snr_uncore_get_mc_dev() to locate the PCI device with
MMIO_BASE and MEM0_BAR register.

Add new ops to access the IMC registers via MMIO.

Add 3 new free running counters for clocks, read and write bandwidth.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: acme@kernel.org
Cc: eranian@google.com
Link: https://lkml.kernel.org/r/1556672028-119221-7-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/events/intel/uncore.c       |   3 +-
 arch/x86/events/intel/uncore.h       |   2 +
 arch/x86/events/intel/uncore_snbep.c | 197 +++++++++++++++++++++++++++++++++++
 3 files changed, 201 insertions(+), 1 deletion(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 8bb537143d86..3694a5d0703d 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -29,7 +29,7 @@ struct event_constraint uncore_constraint_empty =
 
 MODULE_LICENSE("GPL");
 
-static int uncore_pcibus_to_physid(struct pci_bus *bus)
+int uncore_pcibus_to_physid(struct pci_bus *bus)
 {
 	struct pci2phy_map *map;
 	int phys_id = -1;
@@ -1441,6 +1441,7 @@ static const struct intel_uncore_init_fun icl_uncore_init __initconst = {
 static const struct intel_uncore_init_fun snr_uncore_init __initconst = {
 	.cpu_init = snr_uncore_cpu_init,
 	.pci_init = snr_uncore_pci_init,
+	.mmio_init = snr_uncore_mmio_init,
 };
 
 static const struct x86_cpu_id intel_uncore_match[] __initconst = {
diff --git a/arch/x86/events/intel/uncore.h b/arch/x86/events/intel/uncore.h
index cdd9691365a1..f36f7bebbc1b 100644
--- a/arch/x86/events/intel/uncore.h
+++ b/arch/x86/events/intel/uncore.h
@@ -163,6 +163,7 @@ struct pci2phy_map {
 };
 
 struct pci2phy_map *__find_pci2phy_map(int segment);
+int uncore_pcibus_to_physid(struct pci_bus *bus);
 
 ssize_t uncore_event_show(struct kobject *kobj,
 			  struct kobj_attribute *attr, char *buf);
@@ -555,6 +556,7 @@ int skx_uncore_pci_init(void);
 void skx_uncore_cpu_init(void);
 int snr_uncore_pci_init(void);
 void snr_uncore_cpu_init(void);
+void snr_uncore_mmio_init(void);
 
 /* uncore_nhmex.c */
 void nhmex_uncore_cpu_init(void);
diff --git a/arch/x86/events/intel/uncore_snbep.c b/arch/x86/events/intel/uncore_snbep.c
index 3d8752b37413..b10a5ec79e48 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -374,6 +374,19 @@
 #define SNR_PCIE3_PCI_PMON_CTR0			0x4e8
 #define SNR_PCIE3_PCI_PMON_BOX_CTL		0x4e4
 
+/* SNR IMC */
+#define SNR_IMC_MMIO_PMON_FIXED_CTL		0x54
+#define SNR_IMC_MMIO_PMON_FIXED_CTR		0x38
+#define SNR_IMC_MMIO_PMON_CTL0			0x40
+#define SNR_IMC_MMIO_PMON_CTR0			0x8
+#define SNR_IMC_MMIO_PMON_BOX_CTL		0x22800
+#define SNR_IMC_MMIO_OFFSET			0x4000
+#define SNR_IMC_MMIO_SIZE			0x4000
+#define SNR_IMC_MMIO_BASE_OFFSET		0xd0
+#define SNR_IMC_MMIO_BASE_MASK			0x1FFFFFFF
+#define SNR_IMC_MMIO_MEM0_OFFSET		0xd8
+#define SNR_IMC_MMIO_MEM0_MASK			0x7FF
+
 DEFINE_UNCORE_FORMAT_ATTR(event, event, "config:0-7");
 DEFINE_UNCORE_FORMAT_ATTR(event2, event, "config:0-6");
 DEFINE_UNCORE_FORMAT_ATTR(event_ext, event, "config:0-7,21");
@@ -4370,4 +4383,188 @@ int snr_uncore_pci_init(void)
 	return 0;
 }
 
+static struct pci_dev *snr_uncore_get_mc_dev(int id)
+{
+	struct pci_dev *mc_dev = NULL;
+	int phys_id, pkg;
+
+	while (1) {
+		mc_dev = pci_get_device(PCI_VENDOR_ID_INTEL, 0x3451, mc_dev);
+		if (!mc_dev)
+			break;
+		phys_id = uncore_pcibus_to_physid(mc_dev->bus);
+		if (phys_id < 0)
+			continue;
+		pkg = topology_phys_to_logical_pkg(phys_id);
+		if (pkg < 0)
+			continue;
+		else if (pkg == id)
+			break;
+	}
+	return mc_dev;
+}
+
+static void snr_uncore_mmio_init_box(struct intel_uncore_box *box)
+{
+	struct pci_dev *pdev = snr_uncore_get_mc_dev(box->dieid);
+	unsigned int box_ctl = uncore_mmio_box_ctl(box);
+	resource_size_t addr;
+	u32 pci_dword;
+
+	if (!pdev)
+		return;
+
+	pci_read_config_dword(pdev, SNR_IMC_MMIO_BASE_OFFSET, &pci_dword);
+	addr = (pci_dword & SNR_IMC_MMIO_BASE_MASK) << 23;
+
+	pci_read_config_dword(pdev, SNR_IMC_MMIO_MEM0_OFFSET, &pci_dword);
+	addr |= (pci_dword & SNR_IMC_MMIO_MEM0_MASK) << 12;
+
+	addr += box_ctl;
+
+	box->io_addr = ioremap(addr, SNR_IMC_MMIO_SIZE);
+	if (!box->io_addr)
+		return;
+
+	writel(IVBEP_PMON_BOX_CTL_INT, box->io_addr);
+}
+
+static void snr_uncore_mmio_disable_box(struct intel_uncore_box *box)
+{
+	u32 config;
+
+	if (!box->io_addr)
+		return;
+
+	config = readl(box->io_addr);
+	config |= SNBEP_PMON_BOX_CTL_FRZ;
+	writel(config, box->io_addr);
+}
+
+static void snr_uncore_mmio_enable_box(struct intel_uncore_box *box)
+{
+	u32 config;
+
+	if (!box->io_addr)
+		return;
+
+	config = readl(box->io_addr);
+	config &= ~SNBEP_PMON_BOX_CTL_FRZ;
+	writel(config, box->io_addr);
+}
+
+static void snr_uncore_mmio_enable_event(struct intel_uncore_box *box,
+					   struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+
+	if (!box->io_addr)
+		return;
+
+	writel(hwc->config | SNBEP_PMON_CTL_EN,
+	       box->io_addr + hwc->config_base);
+}
+
+static void snr_uncore_mmio_disable_event(struct intel_uncore_box *box,
+					    struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+
+	if (!box->io_addr)
+		return;
+
+	writel(hwc->config, box->io_addr + hwc->config_base);
+}
+
+static struct intel_uncore_ops snr_uncore_mmio_ops = {
+	.init_box	= snr_uncore_mmio_init_box,
+	.exit_box	= uncore_mmio_exit_box,
+	.disable_box	= snr_uncore_mmio_disable_box,
+	.enable_box	= snr_uncore_mmio_enable_box,
+	.disable_event	= snr_uncore_mmio_disable_event,
+	.enable_event	= snr_uncore_mmio_enable_event,
+	.read_counter	= uncore_mmio_read_counter,
+};
+
+static struct uncore_event_desc snr_uncore_imc_events[] = {
+	INTEL_UNCORE_EVENT_DESC(clockticks,      "event=0x00,umask=0x00"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_read,  "event=0x04,umask=0x0f"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_read.scale, "6.103515625e-5"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_read.unit, "MiB"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_write, "event=0x04,umask=0x30"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_write.scale, "6.103515625e-5"),
+	INTEL_UNCORE_EVENT_DESC(cas_count_write.unit, "MiB"),
+	{ /* end: all zeroes */ },
+};
+
+static struct intel_uncore_type snr_uncore_imc = {
+	.name		= "imc",
+	.num_counters   = 4,
+	.num_boxes	= 2,
+	.perf_ctr_bits	= 48,
+	.fixed_ctr_bits	= 48,
+	.fixed_ctr	= SNR_IMC_MMIO_PMON_FIXED_CTR,
+	.fixed_ctl	= SNR_IMC_MMIO_PMON_FIXED_CTL,
+	.event_descs	= snr_uncore_imc_events,
+	.perf_ctr	= SNR_IMC_MMIO_PMON_CTR0,
+	.event_ctl	= SNR_IMC_MMIO_PMON_CTL0,
+	.event_mask	= SNBEP_PMON_RAW_EVENT_MASK,
+	.box_ctl	= SNR_IMC_MMIO_PMON_BOX_CTL,
+	.mmio_offset	= SNR_IMC_MMIO_OFFSET,
+	.ops		= &snr_uncore_mmio_ops,
+	.format_group	= &skx_uncore_format_group,
+};
+
+enum perf_uncore_snr_imc_freerunning_type_id {
+	SNR_IMC_DCLK,
+	SNR_IMC_DDR,
+
+	SNR_IMC_FREERUNNING_TYPE_MAX,
+};
+
+static struct freerunning_counters snr_imc_freerunning[] = {
+	[SNR_IMC_DCLK]	= { 0x22b0, 0x0, 0, 1, 48 },
+	[SNR_IMC_DDR]	= { 0x2290, 0x8, 0, 2, 48 },
+};
+
+static struct uncore_event_desc snr_uncore_imc_freerunning_events[] = {
+	INTEL_UNCORE_EVENT_DESC(dclk,		"event=0xff,umask=0x10"),
+
+	INTEL_UNCORE_EVENT_DESC(read,		"event=0xff,umask=0x20"),
+	INTEL_UNCORE_EVENT_DESC(read.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(read.unit,	"MiB"),
+	INTEL_UNCORE_EVENT_DESC(write,		"event=0xff,umask=0x21"),
+	INTEL_UNCORE_EVENT_DESC(write.scale,	"3.814697266e-6"),
+	INTEL_UNCORE_EVENT_DESC(write.unit,	"MiB"),
+};
+
+static struct intel_uncore_ops snr_uncore_imc_freerunning_ops = {
+	.init_box	= snr_uncore_mmio_init_box,
+	.exit_box	= uncore_mmio_exit_box,
+	.read_counter	= uncore_mmio_read_counter,
+	.hw_config	= uncore_freerunning_hw_config,
+};
+
+static struct intel_uncore_type snr_uncore_imc_free_running = {
+	.name			= "imc_free_running",
+	.num_counters		= 3,
+	.num_boxes		= 1,
+	.num_freerunning_types	= SNR_IMC_FREERUNNING_TYPE_MAX,
+	.freerunning		= snr_imc_freerunning,
+	.ops			= &snr_uncore_imc_freerunning_ops,
+	.event_descs		= snr_uncore_imc_freerunning_events,
+	.format_group		= &skx_uncore_iio_freerunning_format_group,
+};
+
+static struct intel_uncore_type *snr_mmio_uncores[] = {
+	&snr_uncore_imc,
+	&snr_uncore_imc_free_running,
+	NULL,
+};
+
+void snr_uncore_mmio_init(void)
+{
+	uncore_mmio_uncores = snr_mmio_uncores;
+}
+
 /* end of SNR uncore support */

Re: [RESEND PATCH 0/6] Perf uncore support for Snow Ridge server

[Peter Zijlstra]

On Tue, Apr 30, 2019 at 05:53:42PM -0700, kan.liang@linux.intel.com wrote:
> From: Kan Liang <kan.liang@linux.intel.com>
> 
> The patch series intends to enable perf uncore support for Snow Ridge
> server.
> 
> Here is the link for the uncore document.
> https://cdrdv2.intel.com/v1/dl/getContent/611319
> 
> Patch 1: Fixes a generic issue for uncore free-running counter, which
> also impacts the Snow Ridge server.
> 
> Patch 2-6: Perf uncore support for Snow Ridge server.
> 
> Kan Liang (6):
>   perf/x86/intel/uncore: Handle invalid event coding for free-running
>     counter
>   perf/x86/intel/uncore: Add uncore support for Snow Ridge server
>   perf/x86/intel/uncore: Extract codes of box ref/unref
>   perf/x86/intel/uncore: Support MMIO type uncore blocks
>   perf/x86/intel/uncore: Clean up client IMC
>   perf/x86/intel/uncore: Add IMC uncore support for Snow Ridge
> 
>  arch/x86/events/intel/uncore.c       | 122 +++++--
>  arch/x86/events/intel/uncore.h       |  41 ++-
>  arch/x86/events/intel/uncore_snb.c   |  16 +-
>  arch/x86/events/intel/uncore_snbep.c | 601 +++++++++++++++++++++++++++++++++++
>  4 files changed, 737 insertions(+), 43 deletions(-)

Thanks!

Re: [RESEND PATCH 0/6] Perf uncore support for Snow Ridge server

[Peter Zijlstra]

On Tue, Apr 30, 2019 at 05:53:42PM -0700, kan.liang@linux.intel.com wrote:
> From: Kan Liang <kan.liang@linux.intel.com>
> 
> The patch series intends to enable perf uncore support for Snow Ridge
> server.
> 
> Here is the link for the uncore document.
> https://cdrdv2.intel.com/v1/dl/getContent/611319
> 
> Patch 1: Fixes a generic issue for uncore free-running counter, which
> also impacts the Snow Ridge server.
> 
> Patch 2-6: Perf uncore support for Snow Ridge server.
> 
> Kan Liang (6):
>   perf/x86/intel/uncore: Handle invalid event coding for free-running
>     counter
>   perf/x86/intel/uncore: Add uncore support for Snow Ridge server
>   perf/x86/intel/uncore: Extract codes of box ref/unref
>   perf/x86/intel/uncore: Support MMIO type uncore blocks
>   perf/x86/intel/uncore: Clean up client IMC
>   perf/x86/intel/uncore: Add IMC uncore support for Snow Ridge
> 
>  arch/x86/events/intel/uncore.c       | 122 +++++--
>  arch/x86/events/intel/uncore.h       |  41 ++-
>  arch/x86/events/intel/uncore_snb.c   |  16 +-
>  arch/x86/events/intel/uncore_snbep.c | 601 +++++++++++++++++++++++++++++++++++
>  4 files changed, 737 insertions(+), 43 deletions(-)

Kan, we had horrible conflicts between this set and the topology
s/pkg/die/ thing. I've attempted a rebase here:

  https://git.kernel.org/pub/scm/linux/kernel/git/peterz/queue.git/log/?h=perf/core

can you please double check?

Re: [RESEND PATCH 0/6] Perf uncore support for Snow Ridge server

[Liang, Kan]

On 6/3/2019 12:34 PM, Peter Zijlstra wrote:
> On Tue, Apr 30, 2019 at 05:53:42PM -0700, kan.liang@linux.intel.com wrote:
>> From: Kan Liang <kan.liang@linux.intel.com>
>>
>> The patch series intends to enable perf uncore support for Snow Ridge
>> server.
>>
>> Here is the link for the uncore document.
>> https://cdrdv2.intel.com/v1/dl/getContent/611319
>>
>> Patch 1: Fixes a generic issue for uncore free-running counter, which
>> also impacts the Snow Ridge server.
>>
>> Patch 2-6: Perf uncore support for Snow Ridge server.
>>
>> Kan Liang (6):
>>    perf/x86/intel/uncore: Handle invalid event coding for free-running
>>      counter
>>    perf/x86/intel/uncore: Add uncore support for Snow Ridge server
>>    perf/x86/intel/uncore: Extract codes of box ref/unref
>>    perf/x86/intel/uncore: Support MMIO type uncore blocks
>>    perf/x86/intel/uncore: Clean up client IMC
>>    perf/x86/intel/uncore: Add IMC uncore support for Snow Ridge
>>
>>   arch/x86/events/intel/uncore.c       | 122 +++++--
>>   arch/x86/events/intel/uncore.h       |  41 ++-
>>   arch/x86/events/intel/uncore_snb.c   |  16 +-
>>   arch/x86/events/intel/uncore_snbep.c | 601 +++++++++++++++++++++++++++++++++++
>>   4 files changed, 737 insertions(+), 43 deletions(-)
> 
> Kan, we had horrible conflicts between this set and the topology
> s/pkg/die/ thing. I've attempted a rebase here:
> 
>    https://git.kernel.org/pub/scm/linux/kernel/git/peterz/queue.git/log/?h=perf/core
> 
> can you please double check?

There are two errors. I think they have been reported by 0-day.
Rest of the patches looks good.

The first one is from ("perf/x86/intel/uncore: Support MMIO type uncore 
blocks"). The fix is as below.

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index baee051..8453e60 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -1201,7 +1201,7 @@ static int uncore_event_cpu_offline(unsigned int cpu)
  	/* Clear the references */
  	die = topology_logical_die_id(cpu);
  	uncore_box_unref(uncore_msr_uncores, die);
-	uncore_box_unref(uncore_mmio_uncores, pkg);
+	uncore_box_unref(uncore_mmio_uncores, die);
  	return 0;
  }


The other is from ("perf/x86/intel/uncore: Add IMC uncore support for 
Snow Ridge"). The fix is as below.

diff --git a/arch/x86/events/intel/uncore_snbep.c 
b/arch/x86/events/intel/uncore_snbep.c
index f4e04d8..b10a5ec 100644
--- a/arch/x86/events/intel/uncore_snbep.c
+++ b/arch/x86/events/intel/uncore_snbep.c
@@ -4406,7 +4406,7 @@ static struct pci_dev *snr_uncore_get_mc_dev(int id)

  static void snr_uncore_mmio_init_box(struct intel_uncore_box *box)
  {
-	struct pci_dev *pdev = snr_uncore_get_mc_dev(box->pkgid);
+	struct pci_dev *pdev = snr_uncore_get_mc_dev(box->dieid);
  	unsigned int box_ctl = uncore_mmio_box_ctl(box);
  	resource_size_t addr;
  	u32 pci_dword;




Thanks,
Kan