Message-ID: <20260121143208.340549136@infradead.org>
Date: Wed, 21 Jan 2026 15:25:02 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: tglx@...utronix.de,
 mathieu.desnoyers@...icios.com
Cc: linux-kernel@...r.kernel.org,
 peterz@...radead.org,
 paulmck@...nel.org,
 boqun.feng@...il.com,
 corbet@....net,
 prakash.sangappa@...cle.com,
 vineethr@...ux.ibm.com,
 kprateek.nayak@....com,
 rostedt@...dmis.org,
 bigeasy@...utronix.de,
 arnd@...db.de,
 rdunlap@...radead.org,
 rongevarg@...il.com,
 longman@...hat.com
Subject: [PATCH 5/5] selftests/rseq: Add rseq slice histogram script

A script that processes trace-cmd data and generates a histogram of
rseq slice_ext durations for the recorded workload.

Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
---
 Documentation/userspace-api/rseq.rst            |    3 
 tools/testing/selftests/rseq/rseq-slice-hist.py |  132 ++++++++++++++++++++++++
 2 files changed, 135 insertions(+)

--- a/Documentation/userspace-api/rseq.rst
+++ b/Documentation/userspace-api/rseq.rst
@@ -83,6 +83,9 @@ determined by debugfs:rseq/slice_ext_nse
 is the minimum value. It can be incremented to 50 usecs, however doing so
 can/will affect the minimum scheduling latency.
 
+Any proposed changes to this default will have to come with a selftest and
+rseq-slice-hist.py output that shows the new value has merrit.
+
 The kernel indicates the grant by clearing rseq::slice_ctrl::request and
 setting rseq::slice_ctrl::granted to 1. If there is a reschedule of the
 thread after granting the extension, the kernel clears the granted bit to
--- /dev/null
+++ b/tools/testing/selftests/rseq/rseq-slice-hist.py
@@ -0,0 +1,132 @@
+#!/usr/bin/python3
+
+#
+# trace-cmd record -e hrtimer_start -e hrtimer_cancel -e hrtimer_expire_entry -- $cmd
+#
+
+from tracecmd import *
+
+def load_kallsyms(file_path='/proc/kallsyms'):
+    """
+    Parses /proc/kallsyms into a dictionary.
+    Returns: { address_int: symbol_name }
+    """
+    kallsyms_map = {}
+
+    try:
+        with open(file_path, 'r') as f:
+            for line in f:
+                # The format is: [address] [type] [name] [module]
+                parts = line.split()
+                if len(parts) < 3:
+                    continue
+
+                addr = int(parts[0], 16)
+                name = parts[2]
+
+                kallsyms_map[addr] = name
+
+    except PermissionError:
+        print(f"Error: Permission denied reading {file_path}. Try running with sudo.")
+    except FileNotFoundError:
+        print(f"Error: {file_path} not found.")
+
+    return kallsyms_map
+
+ksyms = load_kallsyms()
+
+# pending[timer_ptr] = {'ts': timestamp, 'comm': comm}
+pending = {}
+
+# histograms[comm][bucket] = count
+histograms = {}
+
+class OnlineHarmonicMean:
+    def __init__(self):
+        self.n = 0          # Count of elements
+        self.S = 0.0        # Cumulative sum of reciprocals
+
+    def update(self, x):
+        if x == 0:
+            raise ValueError("Harmonic mean is undefined for zero.")
+
+        self.n += 1
+        self.S += 1.0 / x
+        return self.n / self.S
+
+    @property
+    def mean(self):
+        return self.n / self.S if self.n > 0 else 0
+
+ohms = {}
+
+def handle_start(record):
+    func_name = ksyms[record.num_field("function")]
+    if "rseq_slice_expired" in func_name:
+        timer_ptr = record.num_field("hrtimer")
+        pending[timer_ptr] = {
+            'ts': record.ts,
+            'comm': record.comm
+        }
+    return None
+
+def handle_cancel(record):
+    timer_ptr = record.num_field("hrtimer")
+
+    if timer_ptr in pending:
+        start_data = pending.pop(timer_ptr)
+        duration_ns = record.ts - start_data['ts']
+        duration_us = duration_ns // 1000
+
+        comm = start_data['comm']
+
+        if comm not in ohms:
+            ohms[comm] = OnlineHarmonicMean()
+
+        ohms[comm].update(duration_ns)
+
+        if comm not in histograms:
+            histograms[comm] = {}
+
+        histograms[comm][duration_us] = histograms[comm].get(duration_us, 0) + 1
+    return None
+
+def handle_expire(record):
+    timer_ptr = record.num_field("hrtimer")
+
+    if timer_ptr in pending:
+        start_data = pending.pop(timer_ptr)
+        comm = start_data['comm']
+
+        if comm not in histograms:
+            histograms[comm] = {}
+
+        # Record -1 bucket for expired (failed to cancel)
+        histograms[comm][-1] = histograms[comm].get(-1, 0) + 1
+    return None
+
+if __name__ == "__main__":
+    t = Trace("trace.dat")
+    for cpu in range(0, t.cpus):
+        ev = t.read_event(cpu)
+        while ev:
+            if "hrtimer_start" in ev.name:
+                handle_start(ev)
+            if "hrtimer_cancel" in ev.name:
+                handle_cancel(ev)
+            if "hrtimer_expire_entry" in ev.name:
+                handle_expire(ev)
+
+            ev = t.read_event(cpu)
+
+    print("\n" + "="*40)
+    print("RSEQ SLICE HISTOGRAM (us)")
+    print("="*40)
+    for comm, buckets in histograms.items():
+        print(f"\nTask: {comm}    Mean: {ohms[comm].mean:.3f} ns")
+        print(f"  {'Latency (us)':<15} | {'Count'}")
+        print(f"  {'-'*30}")
+        # Sort buckets numerically, putting -1 at the top
+        for bucket in sorted(buckets.keys()):
+            label = "EXPIRED" if bucket == -1 else f"{bucket} us"
+            print(f"  {label:<15} | {buckets[bucket]}")

Powered by blists - more mailing lists