From: Dapeng Mi <dapeng1.mi@linux.intel.com>
To: Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Namhyung Kim <namhyung@kernel.org>,
Ian Rogers <irogers@google.com>,
Adrian Hunter <adrian.hunter@intel.com>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>,
Kan Liang <kan.liang@linux.intel.com>,
Andi Kleen <ak@linux.intel.com>,
Eranian Stephane <eranian@google.com>
Cc: linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
Dapeng Mi <dapeng1.mi@intel.com>,
Dapeng Mi <dapeng1.mi@linux.intel.com>
Subject: [Patch v2 11/24] perf/x86/intel: Factor out common functions to process PEBS groups
Date: Tue, 18 Feb 2025 15:28:05 +0000 [thread overview]
Message-ID: <20250218152818.158614-12-dapeng1.mi@linux.intel.com> (raw)
In-Reply-To: <20250218152818.158614-1-dapeng1.mi@linux.intel.com>
Adaptive PEBS and arch-PEBS share lots of same code to process these
PEBS groups, like basic, GPR and meminfo groups. Extract these shared
code to common functions to avoid duplicated code.
Signed-off-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
---
arch/x86/events/intel/ds.c | 239 ++++++++++++++++++-------------------
1 file changed, 119 insertions(+), 120 deletions(-)
diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c
index f3c0e509c531..65eaba3aa48d 100644
--- a/arch/x86/events/intel/ds.c
+++ b/arch/x86/events/intel/ds.c
@@ -2068,6 +2068,91 @@ static inline void __setup_pebs_counter_group(struct cpu_hw_events *cpuc,
#define PEBS_LATENCY_MASK 0xffff
+static inline void __setup_perf_sample_data(struct perf_event *event,
+ struct pt_regs *iregs,
+ struct perf_sample_data *data)
+{
+ perf_sample_data_init(data, 0, event->hw.last_period);
+ data->period = event->hw.last_period;
+
+ /*
+ * We must however always use iregs for the unwinder to stay sane; the
+ * record BP,SP,IP can point into thin air when the record is from a
+ * previous PMI context or an (I)RET happened between the record and
+ * PMI.
+ */
+ perf_sample_save_callchain(data, event, iregs);
+}
+
+static inline void __setup_pebs_basic_group(struct perf_event *event,
+ struct pt_regs *regs,
+ struct perf_sample_data *data,
+ u64 sample_type, u64 ip,
+ u64 tsc, u16 retire)
+{
+ /* The ip in basic is EventingIP */
+ set_linear_ip(regs, ip);
+ regs->flags = PERF_EFLAGS_EXACT;
+ setup_pebs_time(event, data, tsc);
+
+ if (sample_type & PERF_SAMPLE_WEIGHT_STRUCT)
+ data->weight.var3_w = retire;
+}
+
+static inline void __setup_pebs_gpr_group(struct perf_event *event,
+ struct pt_regs *regs,
+ struct pebs_gprs *gprs,
+ u64 sample_type)
+{
+ if (event->attr.precise_ip < 2) {
+ set_linear_ip(regs, gprs->ip);
+ regs->flags &= ~PERF_EFLAGS_EXACT;
+ }
+
+ if (sample_type & PERF_SAMPLE_REGS_INTR)
+ adaptive_pebs_save_regs(regs, gprs);
+}
+
+static inline void __setup_pebs_meminfo_group(struct perf_event *event,
+ struct perf_sample_data *data,
+ u64 sample_type, u64 latency,
+ u16 instr_latency, u64 address,
+ u64 aux, u64 tsx_tuning, u64 ax)
+{
+ if (sample_type & PERF_SAMPLE_WEIGHT_TYPE) {
+ u64 tsx_latency = intel_get_tsx_weight(tsx_tuning);
+
+ data->weight.var2_w = instr_latency;
+
+ /*
+ * Although meminfo::latency is defined as a u64,
+ * only the lower 32 bits include the valid data
+ * in practice on Ice Lake and earlier platforms.
+ */
+ if (sample_type & PERF_SAMPLE_WEIGHT)
+ data->weight.full = latency ?: tsx_latency;
+ else
+ data->weight.var1_dw = (u32)latency ?: tsx_latency;
+
+ data->sample_flags |= PERF_SAMPLE_WEIGHT_TYPE;
+ }
+
+ if (sample_type & PERF_SAMPLE_DATA_SRC) {
+ data->data_src.val = get_data_src(event, aux);
+ data->sample_flags |= PERF_SAMPLE_DATA_SRC;
+ }
+
+ if (sample_type & PERF_SAMPLE_ADDR_TYPE) {
+ data->addr = address;
+ data->sample_flags |= PERF_SAMPLE_ADDR;
+ }
+
+ if (sample_type & PERF_SAMPLE_TRANSACTION) {
+ data->txn = intel_get_tsx_transaction(tsx_tuning, ax);
+ data->sample_flags |= PERF_SAMPLE_TRANSACTION;
+ }
+}
+
/*
* With adaptive PEBS the layout depends on what fields are configured.
*/
@@ -2077,12 +2162,14 @@ static void setup_pebs_adaptive_sample_data(struct perf_event *event,
struct pt_regs *regs)
{
struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events);
+ u64 sample_type = event->attr.sample_type;
struct pebs_basic *basic = __pebs;
void *next_record = basic + 1;
- u64 sample_type, format_group;
struct pebs_meminfo *meminfo = NULL;
struct pebs_gprs *gprs = NULL;
struct x86_perf_regs *perf_regs;
+ u64 format_group;
+ u16 retire;
if (basic == NULL)
return;
@@ -2090,32 +2177,17 @@ static void setup_pebs_adaptive_sample_data(struct perf_event *event,
perf_regs = container_of(regs, struct x86_perf_regs, regs);
perf_regs->xmm_regs = NULL;
- sample_type = event->attr.sample_type;
format_group = basic->format_group;
- perf_sample_data_init(data, 0, event->hw.last_period);
- data->period = event->hw.last_period;
- setup_pebs_time(event, data, basic->tsc);
-
- /*
- * We must however always use iregs for the unwinder to stay sane; the
- * record BP,SP,IP can point into thin air when the record is from a
- * previous PMI context or an (I)RET happened between the record and
- * PMI.
- */
- perf_sample_save_callchain(data, event, iregs);
+ __setup_perf_sample_data(event, iregs, data);
*regs = *iregs;
- /* The ip in basic is EventingIP */
- set_linear_ip(regs, basic->ip);
- regs->flags = PERF_EFLAGS_EXACT;
- if (sample_type & PERF_SAMPLE_WEIGHT_STRUCT) {
- if (x86_pmu.flags & PMU_FL_RETIRE_LATENCY)
- data->weight.var3_w = basic->retire_latency;
- else
- data->weight.var3_w = 0;
- }
+ /* basic group */
+ retire = x86_pmu.flags & PMU_FL_RETIRE_LATENCY ?
+ basic->retire_latency : 0;
+ __setup_pebs_basic_group(event, regs, data, sample_type,
+ basic->ip, basic->tsc, retire);
/*
* The record for MEMINFO is in front of GP
@@ -2131,54 +2203,20 @@ static void setup_pebs_adaptive_sample_data(struct perf_event *event,
gprs = next_record;
next_record = gprs + 1;
- if (event->attr.precise_ip < 2) {
- set_linear_ip(regs, gprs->ip);
- regs->flags &= ~PERF_EFLAGS_EXACT;
- }
-
- if (sample_type & PERF_SAMPLE_REGS_INTR)
- adaptive_pebs_save_regs(regs, gprs);
+ __setup_pebs_gpr_group(event, regs, gprs, sample_type);
}
if (format_group & PEBS_DATACFG_MEMINFO) {
- if (sample_type & PERF_SAMPLE_WEIGHT_TYPE) {
- u64 latency = x86_pmu.flags & PMU_FL_INSTR_LATENCY ?
- meminfo->cache_latency : meminfo->mem_latency;
-
- if (x86_pmu.flags & PMU_FL_INSTR_LATENCY)
- data->weight.var2_w = meminfo->instr_latency;
-
- /*
- * Although meminfo::latency is defined as a u64,
- * only the lower 32 bits include the valid data
- * in practice on Ice Lake and earlier platforms.
- */
- if (sample_type & PERF_SAMPLE_WEIGHT) {
- data->weight.full = latency ?:
- intel_get_tsx_weight(meminfo->tsx_tuning);
- } else {
- data->weight.var1_dw = (u32)latency ?:
- intel_get_tsx_weight(meminfo->tsx_tuning);
- }
-
- data->sample_flags |= PERF_SAMPLE_WEIGHT_TYPE;
- }
-
- if (sample_type & PERF_SAMPLE_DATA_SRC) {
- data->data_src.val = get_data_src(event, meminfo->aux);
- data->sample_flags |= PERF_SAMPLE_DATA_SRC;
- }
+ u64 latency = x86_pmu.flags & PMU_FL_INSTR_LATENCY ?
+ meminfo->cache_latency : meminfo->mem_latency;
+ u64 instr_latency = x86_pmu.flags & PMU_FL_INSTR_LATENCY ?
+ meminfo->instr_latency : 0;
+ u64 ax = gprs ? gprs->ax : 0;
- if (sample_type & PERF_SAMPLE_ADDR_TYPE) {
- data->addr = meminfo->address;
- data->sample_flags |= PERF_SAMPLE_ADDR;
- }
-
- if (sample_type & PERF_SAMPLE_TRANSACTION) {
- data->txn = intel_get_tsx_transaction(meminfo->tsx_tuning,
- gprs ? gprs->ax : 0);
- data->sample_flags |= PERF_SAMPLE_TRANSACTION;
- }
+ __setup_pebs_meminfo_group(event, data, sample_type, latency,
+ instr_latency, meminfo->address,
+ meminfo->aux, meminfo->tsx_tuning,
+ ax);
}
if (format_group & PEBS_DATACFG_XMMS) {
@@ -2241,13 +2279,13 @@ static void setup_arch_pebs_sample_data(struct perf_event *event,
struct pt_regs *regs)
{
struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events);
+ u64 sample_type = event->attr.sample_type;
struct arch_pebs_header *header = NULL;
struct arch_pebs_aux *meminfo = NULL;
struct arch_pebs_gprs *gprs = NULL;
struct x86_perf_regs *perf_regs;
void *next_record;
void *at = __pebs;
- u64 sample_type;
if (at == NULL)
return;
@@ -2255,18 +2293,7 @@ static void setup_arch_pebs_sample_data(struct perf_event *event,
perf_regs = container_of(regs, struct x86_perf_regs, regs);
perf_regs->xmm_regs = NULL;
- sample_type = event->attr.sample_type;
- perf_sample_data_init(data, 0, event->hw.last_period);
- data->period = event->hw.last_period;
-
- /*
- * We must however always use iregs for the unwinder to stay sane; the
- * record BP,SP,IP can point into thin air when the record is from a
- * previous PMI context or an (I)RET happened between the record and
- * PMI.
- */
- if (sample_type & PERF_SAMPLE_CALLCHAIN)
- perf_sample_save_callchain(data, event, iregs);
+ __setup_perf_sample_data(event, iregs, data);
*regs = *iregs;
@@ -2275,16 +2302,14 @@ static void setup_arch_pebs_sample_data(struct perf_event *event,
next_record = at + sizeof(struct arch_pebs_header);
if (header->basic) {
struct arch_pebs_basic *basic = next_record;
+ u16 retire = 0;
- /* The ip in basic is EventingIP */
- set_linear_ip(regs, basic->ip);
- regs->flags = PERF_EFLAGS_EXACT;
- setup_pebs_time(event, data, basic->tsc);
+ next_record = basic + 1;
if (sample_type & PERF_SAMPLE_WEIGHT_STRUCT)
- data->weight.var3_w = basic->valid ? basic->retire : 0;
-
- next_record = basic + 1;
+ retire = basic->valid ? basic->retire : 0;
+ __setup_pebs_basic_group(event, regs, data, sample_type,
+ basic->ip, basic->tsc, retire);
}
/*
@@ -2301,44 +2326,18 @@ static void setup_arch_pebs_sample_data(struct perf_event *event,
gprs = next_record;
next_record = gprs + 1;
- if (event->attr.precise_ip < 2) {
- set_linear_ip(regs, gprs->ip);
- regs->flags &= ~PERF_EFLAGS_EXACT;
- }
-
- if (sample_type & PERF_SAMPLE_REGS_INTR)
- adaptive_pebs_save_regs(regs, (struct pebs_gprs *)gprs);
+ __setup_pebs_gpr_group(event, regs, (struct pebs_gprs *)gprs,
+ sample_type);
}
if (header->aux) {
- if (sample_type & PERF_SAMPLE_WEIGHT_TYPE) {
- u16 latency = meminfo->cache_latency;
- u64 tsx_latency = intel_get_tsx_weight(meminfo->tsx_tuning);
+ u64 ax = gprs ? gprs->ax : 0;
- data->weight.var2_w = meminfo->instr_latency;
-
- if (sample_type & PERF_SAMPLE_WEIGHT)
- data->weight.full = latency ?: tsx_latency;
- else
- data->weight.var1_dw = latency ?: (u32)tsx_latency;
- data->sample_flags |= PERF_SAMPLE_WEIGHT_TYPE;
- }
-
- if (sample_type & PERF_SAMPLE_DATA_SRC) {
- data->data_src.val = get_data_src(event, meminfo->aux);
- data->sample_flags |= PERF_SAMPLE_DATA_SRC;
- }
-
- if (sample_type & PERF_SAMPLE_ADDR_TYPE) {
- data->addr = meminfo->address;
- data->sample_flags |= PERF_SAMPLE_ADDR;
- }
-
- if (sample_type & PERF_SAMPLE_TRANSACTION) {
- data->txn = intel_get_tsx_transaction(meminfo->tsx_tuning,
- gprs ? gprs->ax : 0);
- data->sample_flags |= PERF_SAMPLE_TRANSACTION;
- }
+ __setup_pebs_meminfo_group(event, data, sample_type,
+ meminfo->cache_latency,
+ meminfo->instr_latency,
+ meminfo->address, meminfo->aux,
+ meminfo->tsx_tuning, ax);
}
if (header->xmm) {
--
2.40.1
next prev parent reply other threads:[~2025-02-18 8:14 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-18 15:27 [Patch v2 00/24] Arch-PEBS and PMU supports for Clearwater Forest and Panther Lake Dapeng Mi
2025-02-18 15:27 ` [Patch v2 01/24] perf/x86: Add dynamic constraint Dapeng Mi
2025-02-18 15:27 ` [Patch v2 02/24] perf/x86/intel: Add Panther Lake support Dapeng Mi
2025-02-18 15:27 ` [Patch v2 03/24] perf/x86/intel: Add PMU support for Clearwater Forest Dapeng Mi
2025-02-18 15:27 ` [Patch v2 04/24] perf/x86/intel: Parse CPUID archPerfmonExt leaves for non-hybrid CPUs Dapeng Mi
2025-02-18 15:27 ` [Patch v2 05/24] perf/x86/intel: Decouple BTS initialization from PEBS initialization Dapeng Mi
2025-02-18 15:28 ` [Patch v2 06/24] perf/x86/intel: Rename x86_pmu.pebs to x86_pmu.ds_pebs Dapeng Mi
2025-02-18 15:28 ` [Patch v2 07/24] perf/x86/intel: Introduce pairs of PEBS static calls Dapeng Mi
2025-02-18 15:28 ` [Patch v2 08/24] perf/x86/intel: Initialize architectural PEBS Dapeng Mi
2025-02-18 15:28 ` [Patch v2 09/24] perf/x86/intel/ds: Factor out common PEBS processing code to functions Dapeng Mi
2025-02-18 15:28 ` [Patch v2 10/24] perf/x86/intel: Process arch-PEBS records or record fragments Dapeng Mi
2025-02-25 10:39 ` Peter Zijlstra
2025-02-25 11:00 ` Peter Zijlstra
2025-02-26 5:20 ` Mi, Dapeng
2025-02-26 9:35 ` Peter Zijlstra
2025-02-26 15:45 ` Liang, Kan
2025-02-27 2:04 ` Mi, Dapeng
2025-02-25 20:42 ` Andi Kleen
2025-02-26 2:54 ` Mi, Dapeng
2025-02-18 15:28 ` Dapeng Mi [this message]
2025-02-25 11:02 ` [Patch v2 11/24] perf/x86/intel: Factor out common functions to process PEBS groups Peter Zijlstra
2025-02-26 5:24 ` Mi, Dapeng
2025-02-18 15:28 ` [Patch v2 12/24] perf/x86/intel: Allocate arch-PEBS buffer and initialize PEBS_BASE MSR Dapeng Mi
2025-02-25 11:18 ` Peter Zijlstra
2025-02-26 5:48 ` Mi, Dapeng
2025-02-26 9:46 ` Peter Zijlstra
2025-02-27 2:05 ` Mi, Dapeng
2025-02-25 11:25 ` Peter Zijlstra
2025-02-26 6:19 ` Mi, Dapeng
2025-02-26 9:48 ` Peter Zijlstra
2025-02-27 2:09 ` Mi, Dapeng
2025-02-18 15:28 ` [Patch v2 13/24] perf/x86/intel: Update dyn_constranit base on PEBS event precise level Dapeng Mi
2025-02-27 14:06 ` Liang, Kan
2025-03-05 1:41 ` Mi, Dapeng
2025-02-18 15:28 ` [Patch v2 14/24] perf/x86/intel: Setup PEBS data configuration and enable legacy groups Dapeng Mi
2025-02-18 15:28 ` [Patch v2 15/24] perf/x86/intel: Add SSP register support for arch-PEBS Dapeng Mi
2025-02-25 11:52 ` Peter Zijlstra
2025-02-26 6:56 ` Mi, Dapeng
2025-02-25 11:54 ` Peter Zijlstra
2025-02-25 20:44 ` Andi Kleen
2025-02-27 6:29 ` Mi, Dapeng
2025-02-18 15:28 ` [Patch v2 16/24] perf/x86/intel: Add counter group " Dapeng Mi
2025-02-18 15:28 ` [Patch v2 17/24] perf/core: Support to capture higher width vector registers Dapeng Mi
2025-02-25 20:32 ` Peter Zijlstra
2025-02-26 7:55 ` Mi, Dapeng
2025-02-18 15:28 ` [Patch v2 18/24] perf/x86/intel: Support arch-PEBS vector registers group capturing Dapeng Mi
2025-02-25 15:32 ` Peter Zijlstra
2025-02-26 8:08 ` Mi, Dapeng
2025-02-27 6:40 ` Mi, Dapeng
2025-03-04 3:08 ` Mi, Dapeng
2025-03-04 16:26 ` Liang, Kan
2025-03-05 1:34 ` Mi, Dapeng
2025-02-18 15:28 ` [Patch v2 19/24] perf tools: Support to show SSP register Dapeng Mi
2025-02-18 15:28 ` [Patch v2 20/24] perf tools: Enhance arch__intr/user_reg_mask() helpers Dapeng Mi
2025-02-18 15:28 ` [Patch v2 21/24] perf tools: Enhance sample_regs_user/intr to capture more registers Dapeng Mi
2025-02-18 15:28 ` [Patch v2 22/24] perf tools: Support to capture more vector registers (x86/Intel) Dapeng Mi
2025-02-18 15:28 ` [Patch v2 23/24] perf tools/tests: Add vector registers PEBS sampling test Dapeng Mi
2025-02-18 15:28 ` [Patch v2 24/24] perf tools: Fix incorrect --user-regs comments Dapeng Mi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250218152818.158614-12-dapeng1.mi@linux.intel.com \
--to=dapeng1.mi@linux.intel.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=ak@linux.intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=dapeng1.mi@intel.com \
--cc=eranian@google.com \
--cc=irogers@google.com \
--cc=kan.liang@linux.intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox