From: Jason Andryuk <jandryuk@gmail.com>
To: xen-devel@lists.xenproject.org
Cc: Jason Andryuk <jandryuk@gmail.com>,
Ian Jackson <iwj@xenproject.org>, Wei Liu <wl@xen.org>
Subject: [PATCH 12/13] xenpm: Add set-cpufreq-hwp subcommand
Date: Mon, 3 May 2021 15:28:09 -0400 [thread overview]
Message-ID: <20210503192810.36084-13-jandryuk@gmail.com> (raw)
In-Reply-To: <20210503192810.36084-1-jandryuk@gmail.com>
set-cpufreq-hwp allows setting the Hardware P-State (HWP) parameters.
It can be run on all or just a single cpu. There are presets of
balance, powersave & performance. Those can be further tweaked by
param:val arguments as explained in the usage description.
Parameter names are just checked to the first 3 characters to shorten
typing.
Some options are hardware dependent, and ranges can be found in
get-cpufreq-para.
Signed-off-by: Jason Andryuk <jandryuk@gmail.com>
---
tools/misc/xenpm.c | 240 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 240 insertions(+)
diff --git a/tools/misc/xenpm.c b/tools/misc/xenpm.c
index a686f8f46e..d3bcaf3b58 100644
--- a/tools/misc/xenpm.c
+++ b/tools/misc/xenpm.c
@@ -67,6 +67,25 @@ void show_help(void)
" set-max-cstate <num>|'unlimited' [<num2>|'unlimited']\n"
" set the C-State limitation (<num> >= 0) and\n"
" optionally the C-sub-state limitation (<num2> >= 0)\n"
+ " set-cpufreq-hwp [cpuid] [balance|performance|powersave] <param:val>*\n"
+ " set Hardware P-State (HWP) parameters\n"
+ " optionally a preset of one of\n"
+ " balance|performance|powersave\n"
+ " an optional list of param:val arguments\n"
+ " minimum:N hw_lowest ... hw_highest\n"
+ " maximum:N hw_lowest ... hw_highest\n"
+ " desired:N hw_lowest ... hw_highest\n"
+ " Set explicit performance target.\n"
+ " non-zero disables auto-HWP mode.\n"
+ " energy_perf:0-255 (or 0-15)\n"
+ " energy/performance hint\n"
+ " lower favor performance\n"
+ " higher favor powersave\n"
+ " 127 (or 7) balance\n"
+ " act_window:N{,m,u}s range 0us-1270s\n"
+ " window for internal calculations.\n"
+ " 0 lets the hardware decide.\n"
+ " get-cpufreq-para returns hw_lowest/highest.\n"
" start [seconds] start collect Cx/Px statistics,\n"
" output after CTRL-C or SIGINT or several seconds.\n"
" enable-turbo-mode [cpuid] enable Turbo Mode for processors that support it.\n"
@@ -1309,6 +1328,226 @@ void disable_turbo_mode(int argc, char *argv[])
errno, strerror(errno));
}
+/*
+ * Parse activity_window:NNN{us,ms,s} and validate range.
+ *
+ * Activity window is a 7bit mantissa (0-127) with a 3bit exponent (0-7) base
+ * 10 in microseconds. So the range is 1 microsecond to 1270 seconds. A value
+ * of 0 lets the hardware autonomously select the window.
+ *
+ * Return 0 on success
+ * -1 on error
+ * 1 Not activity_window. i.e. try parsing as another argument
+ */
+static int parse_activity_window(xc_set_hwp_para_t *set_hwp, char *p)
+{
+ char *param = NULL, *val = NULL, *suffix = NULL;
+ unsigned int u;
+ unsigned int exponent = 0;
+ unsigned int multiplier = 1;
+ int ret;
+
+ ret = sscanf(p, "%m[a-z_A-Z]:%ms", ¶m, &val);
+ if ( ret != 2 )
+ {
+ return -1;
+ }
+
+ if ( strncasecmp(param, "act", 3) != 0 )
+ {
+ ret = 1;
+
+ goto out;
+ }
+
+ free(param);
+ param = NULL;
+
+ ret = sscanf(val, "%u%ms", &u, &suffix);
+ if ( ret != 1 && ret != 2 )
+ {
+ fprintf(stderr, "invalid activity window: %s\n", val);
+
+ ret = -1;
+
+ goto out;
+ }
+
+ if ( ret == 2 && suffix )
+ {
+ if ( strcasecmp(suffix, "s") == 0 )
+ {
+ multiplier = 1000 * 1000;
+ exponent = 6;
+ }
+ else if ( strcasecmp(suffix, "ms") == 0 )
+ {
+ multiplier = 1000;
+ exponent = 3;
+ }
+ else if ( strcasecmp(suffix, "us") == 0 )
+ {
+ multiplier = 1;
+ exponent = 0;
+ }
+ else
+ {
+ fprintf(stderr, "invalid activity window units: %s\n", suffix);
+
+ ret = -1;
+ goto out;
+ }
+ }
+
+ if ( u > 1270 * 1000 * 1000 / multiplier )
+ {
+ fprintf(stderr, "activity window %s too large\n", val);
+
+ ret = -1;
+ goto out;
+ }
+
+ /* looking for 7 bits of mantissa and 3 bits of exponent */
+ while ( u > 127 )
+ {
+ u /= 10;
+ exponent += 1;
+ }
+
+ set_hwp->activity_window = ( exponent & 0x7 ) << 7 | ( u & 0x7f );
+ set_hwp->set_params |= XEN_SYSCTL_HWP_SET_ACT_WINDOW;
+
+ ret = 0;
+
+ out:
+ free(suffix);
+ free(param);
+ free(val);
+
+ return ret;
+}
+
+static int parse_hwp_opts(xc_set_hwp_para_t *set_hwp, int *cpuid,
+ int argc, char *argv[])
+{
+ int i = 0;
+
+ if ( argc < 1 )
+ return -1;
+
+ if ( parse_cpuid_non_fatal(argv[i], cpuid) == 0 )
+ {
+ i++;
+ }
+
+ if ( i == argc )
+ return -1;
+
+ if ( strcasecmp(argv[i], "powersave") == 0 )
+ {
+ set_hwp->set_params = XEN_SYSCTL_HWP_SET_PRESET_POWERSAVE;
+ i++;
+ }
+ else if ( strcasecmp(argv[i], "performance") == 0 )
+ {
+ set_hwp->set_params = XEN_SYSCTL_HWP_SET_PRESET_PERFORMANCE;
+ i++;
+ }
+ else if ( strcasecmp(argv[i], "balance") == 0 )
+ {
+ set_hwp->set_params = XEN_SYSCTL_HWP_SET_PRESET_BALANCE;
+ i++;
+ }
+
+ for ( ; i < argc; i++)
+ {
+ unsigned int val;
+ char *param;
+ int ret;
+
+ ret = parse_activity_window(set_hwp, argv[i]);
+ switch ( ret )
+ {
+ case -1:
+ return -1;
+ case 0:
+ continue;
+ break;
+ case 1:
+ /* try other parsing */
+ break;
+ }
+
+ /* sscanf can't handle split on ':' for "%ms:%u' */
+ ret = sscanf(argv[i], "%m[a-zA-Z_]:%u", ¶m, &val);
+ if ( ret != 2 )
+ {
+ fprintf(stderr, "%s is an invalid hwp parameter.\n", argv[i]);
+ return -1;
+ }
+
+ if ( val > 255 )
+ {
+ fprintf(stderr, "%s value %u is out of range.\n", param, val);
+ return -1;
+ }
+
+ if ( strncasecmp(param, "min", 3) == 0 )
+ {
+ set_hwp->minimum = val;
+ set_hwp->set_params |= XEN_SYSCTL_HWP_SET_MINIMUM;
+ }
+ else if ( strncasecmp(param, "max", 3) == 0 )
+ {
+ set_hwp->maximum = val;
+ set_hwp->set_params |= XEN_SYSCTL_HWP_SET_MAXIMUM;
+ }
+ else if ( strncasecmp(param, "des", 3) == 0 )
+ {
+ set_hwp->desired = val;
+ set_hwp->set_params |= XEN_SYSCTL_HWP_SET_DESIRED;
+ }
+ else if ( strncasecmp(param, "ene", 3) == 0 )
+ {
+ set_hwp->energy_perf = val;
+ set_hwp->set_params |= XEN_SYSCTL_HWP_SET_ENERGY_PERF;
+ }
+ else
+ {
+ fprintf(stderr, "%s is an invalid parameter\n.", param);
+ return -1;
+ }
+
+ free(param);
+ }
+
+ return 0;
+}
+
+static void hwp_set_func(int argc, char *argv[])
+{
+ xc_set_hwp_para_t set_hwp = {};
+ int cpuid = -1;
+ int i = 0;
+
+ if ( parse_hwp_opts(&set_hwp, &cpuid, argc, argv) )
+ {
+ fprintf(stderr, "Missing, excess, or invalid argument(s)\n");
+ exit(EINVAL);
+ }
+
+ if ( cpuid != -1 )
+ {
+ i = cpuid;
+ max_cpu_nr = i + 1;
+ }
+
+ for ( ; i < max_cpu_nr; i++ )
+ if ( xc_set_cpufreq_hwp(xc_handle, i, &set_hwp) )
+ fprintf(stderr, "[CPU%d] failed to set hwp params (%d - %s)\n",
+ i, errno, strerror(errno));
+}
+
struct {
const char *name;
void (*function)(int argc, char *argv[]);
@@ -1319,6 +1558,7 @@ struct {
{ "get-cpufreq-average", cpufreq_func },
{ "start", start_gather_func },
{ "get-cpufreq-para", cpufreq_para_func },
+ { "set-cpufreq-hwp", hwp_set_func },
{ "set-scaling-maxfreq", scaling_max_freq_func },
{ "set-scaling-minfreq", scaling_min_freq_func },
{ "set-scaling-governor", scaling_governor_func },
--
2.30.2
next prev parent reply other threads:[~2021-05-03 19:35 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-03 19:27 [PATCH 00/13] Intel Hardware P-States (HWP) support Jason Andryuk
2021-05-03 19:27 ` [PATCH 01/13] cpufreq: Allow restricting to internal governors only Jason Andryuk
2021-05-26 13:18 ` Jan Beulich
2021-05-26 14:12 ` Jason Andryuk
2021-05-26 15:09 ` Jan Beulich
2021-05-26 16:44 ` Jason Andryuk
2021-05-03 19:27 ` [PATCH 02/13] cpufreq: Add perf_freq to cpuinfo Jason Andryuk
2021-05-26 13:24 ` Jan Beulich
2021-05-26 14:19 ` Jason Andryuk
2021-05-03 19:28 ` [PATCH 03/13] cpufreq: Export intel_feature_detect Jason Andryuk
2021-05-26 13:27 ` Jan Beulich
2021-05-26 14:44 ` Jason Andryuk
2021-05-26 15:11 ` Jan Beulich
2021-05-03 19:28 ` [PATCH 04/13] cpufreq: Add Hardware P-State (HWP) driver Jason Andryuk
2021-05-26 14:59 ` Jan Beulich
2021-05-27 7:23 ` Jan Beulich
2021-05-27 18:50 ` Jason Andryuk
2021-05-28 6:35 ` Jan Beulich
2021-06-03 11:55 ` Jason Andryuk
2021-06-04 6:39 ` Jan Beulich
2021-05-27 7:45 ` Jan Beulich
2021-05-03 19:28 ` [PATCH 05/13] xenpm: Change get-cpufreq-para output for internal Jason Andryuk
2021-05-26 15:21 ` Jan Beulich
2021-05-27 5:54 ` Jan Beulich
2021-05-03 19:28 ` [PATCH 06/13] cpufreq: Export HWP parameters to userspace Jason Andryuk
2021-05-27 7:55 ` Jan Beulich
2021-05-28 13:19 ` Jason Andryuk
2021-05-28 13:39 ` Jan Beulich
2021-05-03 19:28 ` [PATCH 07/13] libxc: Include hwp_para in definitions Jason Andryuk
2021-05-03 19:28 ` [PATCH 08/13] xenpm: Print HWP parameters Jason Andryuk
2021-05-27 8:02 ` Jan Beulich
2021-05-03 19:28 ` [PATCH 09/13] xen: Add SET_CPUFREQ_HWP xen_sysctl_pm_op Jason Andryuk
2021-05-27 8:33 ` Jan Beulich
2021-05-03 19:28 ` [PATCH 10/13] libxc: Add xc_set_cpufreq_hwp Jason Andryuk
2021-05-04 8:03 ` Jan Beulich
2021-05-04 11:31 ` Jason Andryuk
2021-05-27 9:45 ` Jan Beulich
2021-05-03 19:28 ` [PATCH 11/13] xenpm: Factor out a non-fatal cpuid_parse variant Jason Andryuk
2021-05-27 8:41 ` Jan Beulich
2021-05-03 19:28 ` Jason Andryuk [this message]
2021-05-27 9:46 ` [PATCH 12/13] xenpm: Add set-cpufreq-hwp subcommand Jan Beulich
2021-05-03 19:28 ` [PATCH 13/13] CHANGELOG: Add Intel HWP entry Jason Andryuk
2021-05-27 9:48 ` Jan Beulich
2021-05-20 14:57 ` [PATCH 00/13] Intel Hardware P-States (HWP) support Jan Beulich
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=20210503192810.36084-13-jandryuk@gmail.com \
--to=jandryuk@gmail.com \
--cc=iwj@xenproject.org \
--cc=wl@xen.org \
--cc=xen-devel@lists.xenproject.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;
as well as URLs for NNTP newsgroup(s).