linux-kernel - [PATCH v2 1/6] sched/stats: Print domain name in /proc/schedstat

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20241122084452.1064968-2-swapnil.sapkal@amd.com>
Date: Fri, 22 Nov 2024 08:44:47 +0000
From: Swapnil Sapkal <swapnil.sapkal@....com>
To: <peterz@...radead.org>, <mingo@...hat.com>, <acme@...nel.org>,
	<namhyung@...nel.org>, <irogers@...gle.com>, <james.clark@....com>
CC: <ravi.bangoria@....com>, <yu.c.chen@...el.com>, <mark.rutland@....com>,
	<alexander.shishkin@...ux.intel.com>, <jolsa@...nel.org>,
	<rostedt@...dmis.org>, <vincent.guittot@...aro.org>,
	<adrian.hunter@...el.com>, <kan.liang@...ux.intel.com>,
	<gautham.shenoy@....com>, <kprateek.nayak@....com>, <juri.lelli@...hat.com>,
	<yangjihong@...edance.com>, <void@...ifault.com>, <tj@...nel.org>,
	<vineethr@...ux.ibm.com>, <linux-kernel@...r.kernel.org>,
	<linux-perf-users@...r.kernel.org>, <santosh.shukla@....com>,
	<ananth.narayan@....com>, <sandipan.das@....com>, James Clark
	<james.clark@...aro.org>
Subject: [PATCH v2 1/6] sched/stats: Print domain name in /proc/schedstat

From: K Prateek Nayak <kprateek.nayak@....com>

Currently, there does not exist a straightforward way to extract the
names of the sched domains and match them to the per-cpu domain entry in
/proc/schedstat other than looking at the debugfs files which are only
visible after enabling "verbose" debug after commit 34320745dfc9
("sched/debug: Put sched/domains files under the verbose flag")

Since tools like `perf sched schedstat` require displaying per-domain
information in user friendly manner, display the names of sched domain,
alongside their level in /proc/schedstat if CONFIG_SCHED_DEBUG is enabled.

Domain names also makes the /proc/schedstat data unambiguous when some
of the cpus are offline. For example, on a 128 cpus AMD Zen3 machine
where CPU0 and CPU64 are SMT siblings and CPU64 is offline:

Before:
    cpu0 ...
    domain0 ...
    domain1 ...
    cpu1 ...
    domain0 ...
    domain1 ...
    domain2 ...

After:
    cpu0 ...
    domain0:MC ...
    domain1:PKG ...
    cpu1 ...
    domain0:SMT ...
    domain1:MC ...
    domain2:PKG ...

schedstat version has not been updated since this change merely adds
additional information to the domain name field and does not add a new
field altogether.

Signed-off-by: K Prateek Nayak <kprateek.nayak@....com>
Signed-off-by: Ravi Bangoria <ravi.bangoria@....com>
Tested-by: James Clark <james.clark@...aro.org>
---
 Documentation/scheduler/sched-stats.rst | 8 ++++++--
 kernel/sched/stats.c                    | 6 +++++-
 2 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/Documentation/scheduler/sched-stats.rst b/Documentation/scheduler/sched-stats.rst
index 7c2b16c4729d..b60a3e7bc108 100644
--- a/Documentation/scheduler/sched-stats.rst
+++ b/Documentation/scheduler/sched-stats.rst
@@ -6,6 +6,8 @@ Version 16 of schedstats changed the order of definitions within
 'enum cpu_idle_type', which changed the order of [CPU_MAX_IDLE_TYPES]
 columns in show_schedstat(). In particular the position of CPU_IDLE
 and __CPU_NOT_IDLE changed places. The size of the array is unchanged.
+With CONFIG_SCHED_DEBUG enabled, the domain field can also print the
+name of the corresponding sched domain.
 
 Version 15 of schedstats dropped counters for some sched_yield:
 yld_exp_empty, yld_act_empty and yld_both_empty. Otherwise, it is
@@ -71,9 +73,11 @@ Domain statistics
 -----------------
 One of these is produced per domain for each cpu described. (Note that if
 CONFIG_SMP is not defined, *no* domains are utilized and these lines
-will not appear in the output.)
+will not appear in the output. [:<name>] is an optional extension to the domain
+field that prints the name of the corresponding sched domain. It can appear in
+schedstat version 16 and above, and requires CONFIG_SCHED_DEBUG.)
 
-domain<N> <cpumask> 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
+domain<N>[:<name>] <cpumask> 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
 
 The first field is a bit mask indicating what cpus this domain operates over.
 
diff --git a/kernel/sched/stats.c b/kernel/sched/stats.c
index eb0cdcd4d921..bd4ed737e894 100644
--- a/kernel/sched/stats.c
+++ b/kernel/sched/stats.c
@@ -138,7 +138,11 @@ static int show_schedstat(struct seq_file *seq, void *v)
 		for_each_domain(cpu, sd) {
 			enum cpu_idle_type itype;
 
-			seq_printf(seq, "domain%d %*pb", dcount++,
+			seq_printf(seq, "domain%d", dcount++);
+#ifdef CONFIG_SCHED_DEBUG
+			seq_printf(seq, ":%s", sd->name);
+#endif
+			seq_printf(seq, " %*pb",
 				   cpumask_pr_args(sched_domain_span(sd)));
 			for (itype = 0; itype < CPU_MAX_IDLE_TYPES; itype++) {
 				seq_printf(seq, " %u %u %u %u %u %u %u %u",
-- 
2.43.0