From nobody Wed Apr 1 10:05:09 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 37A83449EA4; Tue, 31 Mar 2026 19:59:17 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1774987157; cv=none; b=PJX/7ql7AtKL4BnkBz0qcAt6/dr4aFKmyt67dTlTE5zZkfaiQ7GngPFYUcfQMujxdoBoSXAIUiTgW4WLORqgMtiGHoXh7YkCt8gHQ8MNT8QTnuysQyf3UL9k5peOk58/azWx46tsxxDArxcNty7Kt2GyTrlDp9G2VE+c2HuGBJ8= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1774987157; c=relaxed/simple; bh=AikkHmmAxNcJR+FHXE7h+S7D9pSLPgTAiDurSpCHMl4=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=i6FZIA5sbgDE82HqE5QZrnRIWH4rEtufNRkw8g+XdgYF0kCodE/7mec/LdhHmGhQ+Ye4d4RktRgKM6cDQssVx1m9jzKsWEyCcNRLUJXjM+J+nFpF2qH41L5jxJVCZGsjI/Y6uwZPpRZ3xHwL5FaPu8lKMUejWyAFWibA2FkUlvA= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=tWGMcvGj; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="tWGMcvGj" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 65382C2BCB2; Tue, 31 Mar 2026 19:59:16 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1774987157; bh=AikkHmmAxNcJR+FHXE7h+S7D9pSLPgTAiDurSpCHMl4=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=tWGMcvGjI67vIma0GclsYigRNhbx3ifC3PZOSIKfBo03WzBQQdjjj6w6SfuV2zUnm 52ncT+a9kDeYhvf2o/2Zk7zLfLnrsA8LeIqpub+hqrdfDFgMHcy3JFiCgUc+r5E4+W F7G9vaBcgUe1VUFBRLnClCP8ZrIZPNkxjXgixch/MOlpjE1bnOQfEXa/jWVUWfm3I8 bk6lhco0NJ9QqhzUjYrjdPE/X8LBMueOwZ7h4cbb1X71VXjK9EuJkdlKC+t9s77sLO 11TM5IjdP06I44B17bkyJQOMsD854QglhxX0n6k8ZKiaXa1sq98OW6FzUWvbzALUOc 2pHIJKqycAUGw== From: "Mario Limonciello (AMD)" To: "Gautham R . Shenoy" Cc: Perry Yuan , linux-kernel@vger.kernel.org (open list:X86 ARCHITECTURE (32-BIT AND 64-BIT)), linux-pm@vger.kernel.org (open list:CPU FREQUENCY SCALING FRAMEWORK), "Mario Limonciello (AMD)" Subject: [PATCH v7 1/5] cpufreq/amd-pstate: Add dynamic energy performance preference Date: Tue, 31 Mar 2026 14:59:03 -0500 Message-ID: <20260331195907.2557193-2-superm1@kernel.org> X-Mailer: git-send-email 2.43.0 In-Reply-To: <20260331195907.2557193-1-superm1@kernel.org> References: <20260331195907.2557193-1-superm1@kernel.org> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Dynamic energy performance preference changes the EPP profile based on whether the machine is running on AC or DC power. A notification chain from the power supply core is used to adjust EPP values on plug in or plug out events. When enabled, the driver exposes a sysfs toggle for dynamic EPP, blocks manual writes to energy_performance_preference, and keeps the policy in performance mode while it "owns" the EPP updates. For non-server systems: * the default EPP for AC mode is `performance`. * the default EPP for DC mode is `balance_performance`. For server systems dynamic EPP is mostly a no-op. Signed-off-by: Mario Limonciello (AMD) --- v6->v7: * Fix accidental casualty of floor perf from rebase (Gautham) * Adjust documentation (Gautham) --- Documentation/admin-guide/pm/amd-pstate.rst | 18 ++- drivers/cpufreq/Kconfig.x86 | 12 ++ drivers/cpufreq/amd-pstate.c | 131 +++++++++++++++++++- drivers/cpufreq/amd-pstate.h | 10 +- 4 files changed, 163 insertions(+), 8 deletions(-) diff --git a/Documentation/admin-guide/pm/amd-pstate.rst b/Documentation/ad= min-guide/pm/amd-pstate.rst index b43675b7f739b..bb1341763882b 100644 --- a/Documentation/admin-guide/pm/amd-pstate.rst +++ b/Documentation/admin-guide/pm/amd-pstate.rst @@ -325,7 +325,7 @@ and user can change current preference according to ene= rgy or performance needs Please get all support profiles list from ``energy_performance_available_preferences`` attribute, all the profiles a= re integer values defined between 0 to 255 when EPP feature is enabled by pla= tform -firmware, if EPP feature is disabled, driver will ignore the written value +firmware, but if the dynamic EPP feature is enabled, driver will block wri= tes. This attribute is read-write. =20 ``boost`` @@ -347,6 +347,22 @@ boost or `1` to enable it, for the respective CPU usin= g the sysfs path Other performance and frequency values can be read back from ``/sys/devices/system/cpu/cpuX/acpi_cppc/``, see :ref:`cppc_sysfs`. =20 +Dynamic energy performance profile +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D +The amd-pstate driver supports dynamically selecting the energy performance +profile based on whether the machine is running on AC or DC power. + +Whether this behavior is enabled by default depends on the kernel +config option `CONFIG_X86_AMD_PSTATE_DYNAMIC_EPP`. This behavior can also = be overridden +at runtime by the sysfs file ``/sys/devices/system/cpu/cpufreq/policyX/dyn= amic_epp``. + +When set to enabled, the driver will select a different energy performance +profile when the machine is running on battery or AC power. +When set to disabled, the driver will not change the energy performance pr= ofile +based on the power source and will not react to user desired power state. + +Attempting to manually write to the ``energy_performance_preference`` sysfs +file will fail when ``dynamic_epp`` is enabled. =20 ``amd-pstate`` vs ``acpi-cpufreq`` =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D diff --git a/drivers/cpufreq/Kconfig.x86 b/drivers/cpufreq/Kconfig.x86 index 2c5c228408bf2..cdaa8d858045a 100644 --- a/drivers/cpufreq/Kconfig.x86 +++ b/drivers/cpufreq/Kconfig.x86 @@ -68,6 +68,18 @@ config X86_AMD_PSTATE_DEFAULT_MODE For details, take a look at: . =20 +config X86_AMD_PSTATE_DYNAMIC_EPP + bool "AMD Processor P-State dynamic EPP support" + depends on X86_AMD_PSTATE + default n + help + Allow the kernel to dynamically change the energy performance + value from events like ACPI platform profile and AC adapter plug + events. + + This feature can also be changed at runtime, this configuration + option only sets the kernel default value behavior. + config X86_AMD_PSTATE_UT tristate "selftest for AMD Processor P-State driver" depends on X86 && ACPI_PROCESSOR diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c index f207252eb5f5f..d3c8eaab88286 100644 --- a/drivers/cpufreq/amd-pstate.c +++ b/drivers/cpufreq/amd-pstate.c @@ -36,6 +36,7 @@ #include #include #include +#include #include #include =20 @@ -86,6 +87,11 @@ static struct cpufreq_driver amd_pstate_driver; static struct cpufreq_driver amd_pstate_epp_driver; static int cppc_state =3D AMD_PSTATE_UNDEFINED; static bool amd_pstate_prefcore =3D true; +#ifdef CONFIG_X86_AMD_PSTATE_DYNAMIC_EPP +static bool dynamic_epp =3D CONFIG_X86_AMD_PSTATE_DYNAMIC_EPP; +#else +static bool dynamic_epp; +#endif static struct quirk_entry *quirks; =20 /* @@ -1155,6 +1161,74 @@ static void amd_pstate_cpu_exit(struct cpufreq_polic= y *policy) kfree(cpudata); } =20 +static int amd_pstate_get_balanced_epp(struct cpufreq_policy *policy) +{ + struct amd_cpudata *cpudata =3D policy->driver_data; + + if (power_supply_is_system_supplied()) + return cpudata->epp_default_ac; + else + return cpudata->epp_default_dc; +} + +static int amd_pstate_power_supply_notifier(struct notifier_block *nb, + unsigned long event, void *data) +{ + struct amd_cpudata *cpudata =3D container_of(nb, struct amd_cpudata, powe= r_nb); + struct cpufreq_policy *policy __free(put_cpufreq_policy) =3D cpufreq_cpu_= get(cpudata->cpu); + u8 epp; + int ret; + + if (event !=3D PSY_EVENT_PROP_CHANGED) + return NOTIFY_OK; + + epp =3D amd_pstate_get_balanced_epp(policy); + + ret =3D amd_pstate_set_epp(policy, epp); + if (ret) + pr_warn("Failed to set CPU %d EPP %u: %d\n", cpudata->cpu, epp, ret); + + return NOTIFY_OK; +} +static void amd_pstate_clear_dynamic_epp(struct cpufreq_policy *policy) +{ + struct amd_cpudata *cpudata =3D policy->driver_data; + + if (cpudata->power_nb.notifier_call) + power_supply_unreg_notifier(&cpudata->power_nb); + cpudata->dynamic_epp =3D false; +} + +static int amd_pstate_set_dynamic_epp(struct cpufreq_policy *policy) +{ + struct amd_cpudata *cpudata =3D policy->driver_data; + int ret; + u8 epp; + + policy->policy =3D CPUFREQ_POLICY_PERFORMANCE; + epp =3D amd_pstate_get_balanced_epp(policy); + ret =3D amd_pstate_set_epp(policy, epp); + if (ret) + return ret; + + /* only enable notifier if things will actually change */ + if (cpudata->epp_default_ac !=3D cpudata->epp_default_dc) { + cpudata->power_nb.notifier_call =3D amd_pstate_power_supply_notifier; + ret =3D power_supply_reg_notifier(&cpudata->power_nb); + if (ret) + goto cleanup; + } + + cpudata->dynamic_epp =3D true; + + return 0; + +cleanup: + amd_pstate_clear_dynamic_epp(policy); + + return ret; +} + /* Sysfs attributes */ =20 /* @@ -1244,14 +1318,19 @@ static ssize_t store_energy_performance_preference( ssize_t ret; u8 epp; =20 + if (cpudata->dynamic_epp) { + pr_debug("EPP cannot be set when dynamic EPP is enabled\n"); + return -EBUSY; + } + ret =3D sysfs_match_string(energy_perf_strings, buf); if (ret < 0) return -EINVAL; =20 - if (!ret) - epp =3D cpudata->epp_default; - else + if (ret) epp =3D epp_values[ret]; + else + epp =3D amd_pstate_get_balanced_epp(policy); =20 if (epp > 0 && policy->policy =3D=3D CPUFREQ_POLICY_PERFORMANCE) { pr_debug("EPP cannot be set under performance policy\n"); @@ -1259,6 +1338,8 @@ static ssize_t store_energy_performance_preference( } =20 ret =3D amd_pstate_set_epp(policy, epp); + if (ret) + return ret; =20 return ret ? ret : count; } @@ -1620,12 +1701,40 @@ static ssize_t prefcore_show(struct device *dev, return sysfs_emit(buf, "%s\n", str_enabled_disabled(amd_pstate_prefcore)); } =20 +static ssize_t dynamic_epp_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + return sysfs_emit(buf, "%s\n", str_enabled_disabled(dynamic_epp)); +} + +static ssize_t dynamic_epp_store(struct device *a, struct device_attribute= *b, + const char *buf, size_t count) +{ + bool enabled; + int ret; + + ret =3D kstrtobool(buf, &enabled); + if (ret) + return ret; + + if (dynamic_epp =3D=3D enabled) + return -EINVAL; + + /* reinitialize with desired dynamic EPP value */ + dynamic_epp =3D enabled; + ret =3D amd_pstate_change_driver_mode(cppc_state); + + return ret ? ret : count; +} + static DEVICE_ATTR_RW(status); static DEVICE_ATTR_RO(prefcore); +static DEVICE_ATTR_RW(dynamic_epp); =20 static struct attribute *pstate_global_attributes[] =3D { &dev_attr_status.attr, &dev_attr_prefcore.attr, + &dev_attr_dynamic_epp.attr, NULL }; =20 @@ -1715,13 +1824,17 @@ static int amd_pstate_epp_cpu_init(struct cpufreq_p= olicy *policy) if (amd_pstate_acpi_pm_profile_server() || amd_pstate_acpi_pm_profile_undefined()) { policy->policy =3D CPUFREQ_POLICY_PERFORMANCE; - cpudata->epp_default =3D amd_pstate_get_epp(cpudata); + cpudata->epp_default_ac =3D cpudata->epp_default_dc =3D amd_pstate_get_e= pp(cpudata); } else { policy->policy =3D CPUFREQ_POLICY_POWERSAVE; - cpudata->epp_default =3D AMD_CPPC_EPP_BALANCE_PERFORMANCE; + cpudata->epp_default_ac =3D AMD_CPPC_EPP_PERFORMANCE; + cpudata->epp_default_dc =3D AMD_CPPC_EPP_BALANCE_PERFORMANCE; } =20 - ret =3D amd_pstate_set_epp(policy, cpudata->epp_default); + if (dynamic_epp) + ret =3D amd_pstate_set_dynamic_epp(policy); + else + ret =3D amd_pstate_set_epp(policy, amd_pstate_get_balanced_epp(policy)); if (ret) goto free_cpudata1; =20 @@ -1753,6 +1866,8 @@ static void amd_pstate_epp_cpu_exit(struct cpufreq_po= licy *policy) amd_pstate_update_perf(policy, perf.bios_min_perf, 0U, 0U, 0U, false); amd_pstate_set_floor_perf(policy, cpudata->bios_floor_perf); =20 + if (cpudata->dynamic_epp) + amd_pstate_clear_dynamic_epp(policy); kfree(cpudata); policy->driver_data =3D NULL; } @@ -1790,6 +1905,10 @@ static int amd_pstate_epp_set_policy(struct cpufreq_= policy *policy) if (!policy->cpuinfo.max_freq) return -ENODEV; =20 + /* policy can't be changed to powersave policy while dynamic epp is enabl= ed */ + if (policy->policy =3D=3D CPUFREQ_POLICY_POWERSAVE && cpudata->dynamic_ep= p) + return -EBUSY; + cpudata->policy =3D policy->policy; =20 ret =3D amd_pstate_epp_update_limit(policy, true); diff --git a/drivers/cpufreq/amd-pstate.h b/drivers/cpufreq/amd-pstate.h index 32b8b26ce388f..d929ae3163b3d 100644 --- a/drivers/cpufreq/amd-pstate.h +++ b/drivers/cpufreq/amd-pstate.h @@ -85,6 +85,11 @@ struct amd_aperf_mperf { * AMD P-State driver supports preferred core featue. * @epp_cached: Cached CPPC energy-performance preference value * @policy: Cpufreq policy value + * @suspended: If CPU core if offlined + * @epp_default_ac: Default EPP value for AC power source + * @epp_default_dc: Default EPP value for DC power source + * @dynamic_epp: Whether dynamic EPP is enabled + * @power_nb: Notifier block for power events * * The amd_cpudata is key private data for each CPU thread in AMD P-State,= and * represents all the attributes and goals that AMD P-State requests at ru= ntime. @@ -118,7 +123,10 @@ struct amd_cpudata { /* EPP feature related attributes*/ u32 policy; bool suspended; - u8 epp_default; + u8 epp_default_ac; + u8 epp_default_dc; + bool dynamic_epp; + struct notifier_block power_nb; }; =20 /* --=20 2.43.0