Store cacheline size during perf record in header, so
that cacheline size can be used for other features, like
sort.

V6: Namhyung feedback and tests
V5: Namhyung feedback
V4: Ian feedback
V3: Rebase off perf-tools-next round two
V2: Rebase off perf-tools-next

Follow-up patch from message ID "aYZiQk6Uftzlb_JV@x1"

Testing:
- Built perf
- Ran record + report with feat enabled
- Ran record + report with feat disabled

Testing example with feat enabled:
$ perf record ./Example
$ perf report --header-only | grep -C 3 cacheline
CPU_DOMAIN_INFO info available, use -I to display
e_machine : 62
e_flags : 0
cacheline size: 64
missing features: TRACING_DATA BUILD_ID BRANCH_STACK GROUP_DESC AUXTRACE \
STAT CLOCKID DIR_FORMAT COMPRESSED CLOCK_DATA
========

Signed-off-by: Ricky Ringler <ricky.ringler@proton.me>
---
 tools/perf/builtin-inject.c |  1 +
 tools/perf/util/env.h       |  1 +
 tools/perf/util/header.c    | 33 +++++++++++++++++++++++++++++
 tools/perf/util/header.h    |  1 +
 tools/perf/util/sort.c      | 41 +++++++++++++++++++++++++++----------
 5 files changed, 66 insertions(+), 11 deletions(-)

diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index 5b29f4296861..11ac7c8c4be3 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -2134,6 +2134,7 @@ static bool keep_feat(struct perf_inject *inject, int feat)
 	case HEADER_HYBRID_TOPOLOGY:
 	case HEADER_PMU_CAPS:
 	case HEADER_CPU_DOMAIN_INFO:
+	case HEADER_CLN_SIZE:
 		return true;
 	/* Information that can be updated */
 	case HEADER_BUILD_ID:
diff --git a/tools/perf/util/env.h b/tools/perf/util/env.h
index a4501cbca375..c7052ac1f856 100644
--- a/tools/perf/util/env.h
+++ b/tools/perf/util/env.h
@@ -112,6 +112,7 @@ struct perf_env {
 	struct cpu_cache_level	*caches;
 	struct cpu_domain_map	**cpu_domain;
 	int			 caches_cnt;
+	unsigned int		cln_size;
 	u32			comp_ratio;
 	u32			comp_ver;
 	u32			comp_type;
diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
index 9142a8ba4019..4d852bd4ca9a 100644
--- a/tools/perf/util/header.c
+++ b/tools/perf/util/header.c
@@ -54,6 +54,7 @@
 #include "bpf-event.h"
 #include "bpf-utils.h"
 #include "clockid.h"
+#include "cacheline.h"
 
 #include <linux/ctype.h>
 #include <internal/lib.h>
@@ -1304,6 +1305,22 @@ static int write_cache(struct feat_fd *ff,
 	return ret;
 }
 
+#define DEFAULT_CACHELINE_SIZE = 64
+
+static int write_cln_size(struct feat_fd *ff,
+		       struct evlist *evlist __maybe_unused)
+{
+	int cln_size = cacheline_size();
+
+
+	if(!cln_size)
+		cln_size = DEFAULT_CACHELINE_SIZE;
+
+	ff->ph->env.cln_size = cln_size;
+
+	return do_write(ff, &cln_size, sizeof(cln_size));
+}
+
 static int write_stat(struct feat_fd *ff __maybe_unused,
 		      struct evlist *evlist __maybe_unused)
 {
@@ -2261,6 +2278,11 @@ static void print_cache(struct feat_fd *ff, FILE *fp __maybe_unused)
 	}
 }
 
+static void print_cln_size(struct feat_fd *ff, FILE *fp)
+{
+	fprintf(fp, "# cacheline size: %u\n", ff->ph->env.cln_size);
+}
+
 static void print_compressed(struct feat_fd *ff, FILE *fp)
 {
 	fprintf(fp, "# compressed : %s, level = %d, ratio = %d\n",
@@ -3154,6 +3176,16 @@ static int process_cache(struct feat_fd *ff, void *data __maybe_unused)
 	return -1;
 }
 
+static int process_cln_size(struct feat_fd *ff, void *data __maybe_unused)
+{
+	struct perf_env *env = &ff->ph->env;
+
+	if (do_read_u32(ff, &env->cln_size))
+		return -1;
+
+	return 0;
+}
+
 static int process_sample_time(struct feat_fd *ff, void *data __maybe_unused)
 {
 	struct perf_session *session;
@@ -3763,6 +3795,7 @@ const struct perf_header_feature_ops feat_ops[HEADER_LAST_FEATURE] = {
 	FEAT_OPR(PMU_CAPS,	pmu_caps,	false),
 	FEAT_OPR(CPU_DOMAIN_INFO,	cpu_domain_info,	true),
 	FEAT_OPR(E_MACHINE,	e_machine,	false),
+	FEAT_OPR(CLN_SIZE,	cln_size,	false),
 };
 
 struct header_print_data {
diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h
index cc40ac796f52..be315040727f 100644
--- a/tools/perf/util/header.h
+++ b/tools/perf/util/header.h
@@ -55,6 +55,7 @@ enum {
 	HEADER_PMU_CAPS,
 	HEADER_CPU_DOMAIN_INFO,
 	HEADER_E_MACHINE,
+	HEADER_CLN_SIZE,
 	HEADER_LAST_FEATURE,
 	HEADER_FEAT_BITS	= 256,
 };
diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index 42d5cd7ef4e2..5f617cf03d5d 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -30,6 +30,7 @@
 #include "time-utils.h"
 #include "cgroup.h"
 #include "machine.h"
+#include "session.h"
 #include "trace-event.h"
 #include <linux/kernel.h>
 #include <linux/string.h>
@@ -2474,7 +2475,30 @@ struct sort_entry sort_type_offset = {
 
 /* --sort typecln */
 
-#define DEFAULT_CACHELINE_SIZE 64
+static int
+hist_entry__cln_size(struct hist_entry *he)
+{
+	int ret = 0;
+
+	if (he && he->hists) {
+		struct evsel *evsel = hists_to_evsel(he->hists);
+
+
+		if (evsel) {
+			struct perf_session *session = evsel__session(evsel);
+
+			ret = session->header.env.cln_size;
+		}
+	}
+
+	if (!ret || ret < 1) {
+		int default_cacheline_size = 64; // avoid div/0 later
+
+		ret = default_cacheline_size;
+	}
+
+	return ret;
+}
 
 static int64_t
 sort__typecln_sort(struct hist_entry *left, struct hist_entry *right)
@@ -2482,11 +2506,9 @@ sort__typecln_sort(struct hist_entry *left, struct hist_entry *right)
 	struct annotated_data_type *left_type = left->mem_type;
 	struct annotated_data_type *right_type = right->mem_type;
 	int64_t left_cln, right_cln;
+	int64_t cln_size_left = hist_entry__cln_size(left);
+	int64_t cln_size_right = hist_entry__cln_size(right);
 	int64_t ret;
-	int cln_size = cacheline_size();
-
-	if (cln_size == 0)
-		cln_size = DEFAULT_CACHELINE_SIZE;
 
 	if (!left_type) {
 		sort__type_init(left);
@@ -2502,8 +2524,8 @@ sort__typecln_sort(struct hist_entry *left, struct hist_entry *right)
 	if (ret)
 		return ret;
 
-	left_cln = left->mem_type_off / cln_size;
-	right_cln = right->mem_type_off / cln_size;
+	left_cln = left->mem_type_off / cln_size_left;
+	right_cln = right->mem_type_off / cln_size_right;
 	return left_cln - right_cln;
 }
 
@@ -2511,10 +2533,7 @@ static int hist_entry__typecln_snprintf(struct hist_entry *he, char *bf,
 				     size_t size, unsigned int width __maybe_unused)
 {
 	struct annotated_data_type *he_type = he->mem_type;
-	int cln_size = cacheline_size();
-
-	if (cln_size == 0)
-		cln_size = DEFAULT_CACHELINE_SIZE;
+	int cln_size = hist_entry__cln_size(he);
 
 	return repsep_snprintf(bf, size, "%s: cache-line %d", he_type->self.type_name,
 			       he->mem_type_off / cln_size);
-- 
2.53.0