From: Artem Bityutskiy <dedekind1@gmail.com>
To: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Linux PM Mailing List <linux-pm@vger.kernel.org>
Subject: [PATCH 3/3] intel_idle: Drop C-states redundant when PC6 is disabled
Date: Sat, 25 Apr 2026 10:25:32 +0300 [thread overview]
Message-ID: <20260425072532.358365-5-dedekind1@gmail.com> (raw)
In-Reply-To: <20260425072532.358365-1-dedekind1@gmail.com>
From: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
On modern Xeon platforms, such as Granite Rapids, Sierra Forest, and
Clearwater Forest, there are two flavors of requestable C6 states: C6
and C6P. C6 allows only core C6 (CC6), while C6P allows both CC6 and
package C6 (PC6). PC6 saves more power but also has a higher exit
latency, so many users disable it in BIOS.
When PC6 is disabled, C6P becomes identical to C6 — the CPU treats C6P
requests as C6 requests. Exposing both C6 and C6P to user space in this
situation is confusing: two states with the same name look different but
behave the same. It also adds unnecessary overhead to the cpuidle
subsystem, which is a fast path: the governor evaluates every registered
state on idle entry.
Drop C-states that are redundant when PC6 is disabled by marking them
with CPUIDLE_FLAG_UNUSABLE, which causes cpuidle to exclude them when
registering idle states.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
---
drivers/idle/intel_idle.c | 53 +++++++++++++++++++++++++++++++++++++++
1 file changed, 53 insertions(+)
diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
index 562f8e27256e7..f052564eb308b 100644
--- a/drivers/idle/intel_idle.c
+++ b/drivers/idle/intel_idle.c
@@ -2087,6 +2087,53 @@ static void __init spr_idle_state_table_update(void)
}
}
+/**
+ * drop_pc6_redundant_cstates() - Drop C-states redundant when PC6 is disabled.
+ * @states: Idle states table to modify.
+ *
+ * When PC6 is disabled in BIOS, C-states that exist solely to enable PC6
+ * entry (such as C6P or C6SP) become identical to shallower C-states like
+ * C6, and are therefore redundant. Should be called only on systems with
+ * multiple C6 flavors.
+ */
+static void __init drop_pc6_redundant_cstates(struct cpuidle_state *states)
+{
+ int count;
+
+ if (!skx_is_pc6_disabled())
+ /* PC6 is not disabled, nothing to do */
+ return;
+
+ for (count = 0; states[count].enter; count++)
+ continue;
+
+ if (count < 2) {
+ pr_debug("Too few idle states to drop PC6-redundant states\n");
+ return;
+ }
+
+ /*
+ * Sanity check: At this point all platforms with multiple C6 flavors
+ * use the CPUIDLE_FLAG_PARTIAL_HINT_MATCH flag. And the last state in
+ * the table is the one that becomes redundant when PC6 is disabled.
+ */
+ if (!(states[count - 1].flags & CPUIDLE_FLAG_PARTIAL_HINT_MATCH)) {
+ pr_debug("Can't drop PC6-redundant states: unexpected flags\n");
+ return;
+ }
+
+ /*
+ * On all current platforms with multiple C6 flavors, there is only one
+ * C-state that becomes redundant when PC6 is disabled. This state is
+ * the last one in the table. Drop it by marking it with
+ * CPUIDLE_FLAG_UNUSABLE so that cpuidle excludes it when registering
+ * idle states.
+ */
+ pr_info("Dropping idle state %s because PC6 is disabled\n",
+ states[count - 1].name);
+ states[count - 1].flags |= CPUIDLE_FLAG_UNUSABLE;
+}
+
/**
* byt_cht_auto_demotion_disable - Disable Bay/Cherry Trail auto-demotion.
*/
@@ -2176,6 +2223,12 @@ static void __init intel_idle_init_cstates_icpu(struct cpuidle_driver *drv)
case INTEL_ATOM_AIRMONT:
byt_cht_auto_demotion_disable();
break;
+ case INTEL_GRANITERAPIDS_D:
+ case INTEL_GRANITERAPIDS_X:
+ case INTEL_ATOM_CRESTMONT_X:
+ case INTEL_ATOM_DARKMONT_X:
+ drop_pc6_redundant_cstates(cpuidle_state_table);
+ break;
}
for (cstate = 0; cstate < CPUIDLE_STATE_MAX; ++cstate) {
--
2.53.0
prev parent reply other threads:[~2026-04-25 7:26 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-25 7:25 [PATCH 0/3] intel_idle: Drop C-states redundant when PC6 is disabled Artem Bityutskiy
2026-04-25 7:25 ` [PATCH 1/3] intel_idle: Add constants for MSR_PKG_CST_CONFIG_CONTROL Artem Bityutskiy
2026-04-25 7:25 ` [PATCH] intel_idle: Add Panther Lake C-states table Artem Bityutskiy
2026-04-25 7:27 ` Artem Bityutskiy
2026-04-25 7:25 ` [PATCH 2/3] intel_idle: Introduce a helper for checking PC6 Artem Bityutskiy
2026-04-25 7:25 ` Artem Bityutskiy [this message]
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=20260425072532.358365-5-dedekind1@gmail.com \
--to=dedekind1@gmail.com \
--cc=linux-pm@vger.kernel.org \
--cc=rafael@kernel.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