powercap: intel_rapl: Fixes and new platform enabling
Subsystems (2)
Zhang Rui
reviewercover powercap: intel_rapl: Fixes and new platform enabling
Patch 1/6 fixes a real bug when MMIO RAPL driver is probed on platforms that are not listed in current CPU model list. IMO, it should be considered as stable material.
Patch 2/6 fixes a potential racing issue, but I have not reproduced it yet.
Patch 3/6 ~ 4/6 fix a problem that TPMI RAPL driver probes disabled System (Psys) RAPL Domains.
Patch 5/6 and 6/6 are two simple CPU model updates to support MSR RAPL on Arrowlake and Lunarlake platforms.
thanks,
rui
----------------------------------------------------------------
Sumeet Pawnikar (1):
powercap: intel_rapl: add support for Arrow Lake
Zhang Rui (5):
powercap: intel_rapl: Fix a NULL pointer reference bug
powercap: intel_rapl: Fix locking for TPMI RAPL
powercap: intel_rapl_tpmi: Fix a register bug
powercap: intel_rapl_tpmi: Fix System Domain probing
powercap: intel_rapl: Add support for LNL-M paltform
drivers/powercap/intel_rapl_common.c | 36 ++++++++++++++++++++-- drivers/powercap/intel_rapl_msr.c | 8 ++--- drivers/powercap/intel_rapl_tpmi.c | 15 +++++++++ .../intel/int340x_thermal/processor_thermal_rapl.c | 8 ++--- include/linux/intel_rapl.h | 6 ++++ 5 files changed, 62 insertions(+), 11 deletions(-)
Zhang Rui
reviewerPATCH 1/6 powercap: intel_rapl: Fix a NULL pointer dereference bug
A NULL pointer dereference bug is triggered when probing MMIO RAPL driver on platforms with CPU ID not listed in intel_rapl_common CPU model list.
This is because the intel_rapl_common module still probes on such platforms even if 'defaults_msr' is not set after commit 1488ac990ac8 ("powercap: intel_rapl: Allow probing without CPUID match"). Thus the MMIO RAPL rp->priv->defaults is NULL when registering to RAPL framework.
Fix the problem by adding sanity check to ensure rp->priv->rapl_defaults is always valid.
--- drivers/powercap/intel_rapl_common.c | 5 +++++ 1 file changed, 5 insertions(+)
drivers/powercap/intel_rapl_common.c +5 -0
@@ -759,6 +759,11 @@ static int rapl_config(struct rapl_package *rp)default :return -EINVAL ; } + +/* defaults_msr can be NULL on unsupported platforms */ +if (!rp->priv->defaults || !rp->priv->rpi) +return -ENODEV ; +return 0 ; }
Signature · 1 line
2.34.1
Zhang Rui
reviewerPATCH 2/6 powercap: intel_rapl: Fix locking for TPMI RAPL
In reply to Zhang RuiRAPL framework uses CPU hotplug lock to protect rapl_packages list and rp->lead_cpu to guarantee that
1. the RAPL package device is not unprobed and freed
2. the cached rp->lead_cpu is always valid
for operations like powercap sysfs accesses.
Current RAPL APIs assume they are called from CPU hotplug callbacks which hold the CPU hotplug lock. But TPMI RAPL driver invokes the APIs in the driver' .probe() function without CPU hotplug lock held.
Fix the problem by providing both locked and lockless version RAPL APIs.
--- drivers/powercap/intel_rapl_common.c | 29 +++++++++++++++++-- drivers/powercap/intel_rapl_msr.c | 8 ++--- .../int340x_thermal/processor_thermal_rapl.c | 8 ++--- include/linux/intel_rapl.h | 6 ++++ 4 files changed, 40 insertions(+), 11 deletions(-)
drivers/powercap/intel_rapl_common.c +26 -3
@@ -5,6 +5,7 @@*/ #define pr_fmt (fmt) KBUILD_MODNAME ": " fmt +#include <linux/cleanup.h> #include <linux/kernel.h> #include <linux/module.h> #include <linux/list.h>
@@ -1504,7 +1505,7 @@ static int rapl_detect_domains(struct rapl_package *rp) }/* called from CPU hotplug notifier, hotplug lock held */ -void rapl_remove_package (struct rapl_package *rp) +void rapl_remove_package_cpuslocked (struct rapl_package *rp) {struct rapl_domain *rd, *rd_package =NULL ;
@@ -1533,10 +1534,18 @@ void rapl_remove_package(struct rapl_package *rp)list_del (&rp->plist);kfree (rp); } +EXPORT_SYMBOL_GPL (rapl_remove_package_cpuslocked); + +void rapl_remove_package (struct rapl_package *rp) +{ +guard (cpus_read_lock)(); +rapl_remove_package_cpuslocked (rp); +}EXPORT_SYMBOL_GPL (rapl_remove_package);/* caller to ensure CPU hotplug lock is held */ -struct rapl_package *rapl_find_package_domain (int id,struct rapl_if_priv *priv,bool id_is_cpu) +struct rapl_package *rapl_find_package_domain_cpuslocked (int id,struct rapl_if_priv *priv, +bool id_is_cpu) {struct rapl_package *rp;int uid;
@@ -1554,10 +1563,17 @@ struct rapl_package *rapl_find_package_domain(int id, struct rapl_if_priv *priv,return NULL ; } +EXPORT_SYMBOL_GPL (rapl_find_package_domain_cpuslocked); + +struct rapl_package *rapl_find_package_domain (int id,struct rapl_if_priv *priv,bool id_is_cpu) +{ +guard (cpus_read_lock)(); +return rapl_find_package_domain_cpuslocked (id, priv, id_is_cpu); +}EXPORT_SYMBOL_GPL (rapl_find_package_domain);/* called from CPU hotplug notifier, hotplug lock held */ -struct rapl_package *rapl_add_package (int id,struct rapl_if_priv *priv,bool id_is_cpu) +struct rapl_package *rapl_add_package_cpuslocked (int id,struct rapl_if_priv *priv,bool id_is_cpu) {struct rapl_package *rp;int ret;
@@ -1603,6 +1619,13 @@ struct rapl_package *rapl_add_package(int id, struct rapl_if_priv *priv, bool idkfree (rp);return ERR_PTR (ret); } +EXPORT_SYMBOL_GPL (rapl_add_package_cpuslocked); + +struct rapl_package *rapl_add_package (int id,struct rapl_if_priv *priv,bool id_is_cpu) +{ +guard (cpus_read_lock)(); +return rapl_add_package_cpuslocked (id, priv, id_is_cpu); +}EXPORT_SYMBOL_GPL (rapl_add_package);static void power_limit_state_save (void )
drivers/powercap/intel_rapl_msr.c +4 -4
@@ -73,9 +73,9 @@ static int rapl_cpu_online(unsigned int cpu) {struct rapl_package *rp; - rp =rapl_find_package_domain (cpu, rapl_msr_priv, true); + rp =rapl_find_package_domain_cpuslocked (cpu, rapl_msr_priv, true);if (!rp) { - rp =rapl_add_package (cpu, rapl_msr_priv, true); + rp =rapl_add_package_cpuslocked (cpu, rapl_msr_priv, true);if (IS_ERR (rp))return PTR_ERR (rp); }
@@ -88,14 +88,14 @@ static int rapl_cpu_down_prep(unsigned int cpu)struct rapl_package *rp;int lead_cpu; - rp =rapl_find_package_domain (cpu, rapl_msr_priv, true); + rp =rapl_find_package_domain_cpuslocked (cpu, rapl_msr_priv, true);if (!rp)return 0 ;cpumask_clear_cpu (cpu, &rp->cpumask); lead_cpu =cpumask_first (&rp->cpumask);if (lead_cpu >= nr_cpu_ids) -rapl_remove_package (rp); +rapl_remove_package_cpuslocked (rp);else if (rp->lead_cpu == cpu) rp->lead_cpu = lead_cpu;return 0 ;
drivers/thermal/intel/int340x_thermal/processor_thermal_rapl.c +4 -4
@@ -27,9 +27,9 @@ static int rapl_mmio_cpu_online(unsigned int cpu)if (topology_physical_package_id (cpu))return 0 ; - rp =rapl_find_package_domain (cpu, &rapl_mmio_priv, true); + rp =rapl_find_package_domain_cpuslocked (cpu, &rapl_mmio_priv, true);if (!rp) { - rp =rapl_add_package (cpu, &rapl_mmio_priv, true); + rp =rapl_add_package_cpuslocked (cpu, &rapl_mmio_priv, true);if (IS_ERR (rp))return PTR_ERR (rp); }
@@ -42,14 +42,14 @@ static int rapl_mmio_cpu_down_prep(unsigned int cpu)struct rapl_package *rp;int lead_cpu; - rp =rapl_find_package_domain (cpu, &rapl_mmio_priv, true); + rp =rapl_find_package_domain_cpuslocked (cpu, &rapl_mmio_priv, true);if (!rp)return 0 ;cpumask_clear_cpu (cpu, &rp->cpumask); lead_cpu =cpumask_first (&rp->cpumask);if (lead_cpu >= nr_cpu_ids) -rapl_remove_package (rp); +rapl_remove_package_cpuslocked (rp);else if (rp->lead_cpu == cpu) rp->lead_cpu = lead_cpu;return 0 ;
include/linux/intel_rapl.h +6 -0
@@ -178,6 +178,12 @@ struct rapl_package {struct rapl_if_priv *priv; }; +struct rapl_package *rapl_find_package_domain_cpuslocked (int id,struct rapl_if_priv *priv, +bool id_is_cpu); +struct rapl_package *rapl_add_package_cpuslocked (int id,struct rapl_if_priv *priv, +bool id_is_cpu); +void rapl_remove_package_cpuslocked (struct rapl_package *rp); +struct rapl_package *rapl_find_package_domain (int id,struct rapl_if_priv *priv,bool id_is_cpu);struct rapl_package *rapl_add_package (int id,struct rapl_if_priv *priv,bool id_is_cpu);void rapl_remove_package (struct rapl_package *rp);
Signature · 1 line
2.34.1
Zhang Rui
reviewerPATCH 3/6 powercap: intel_rapl_tpmi: Fix a register bug
In reply to Zhang RuiAdd the missing Domain Info register. This also fixes the bogus definition of the Interrupt register.
Neither of these two registers was used previously.
--- drivers/powercap/intel_rapl_tpmi.c | 1 + 1 file changed, 1 insertion(+)
drivers/powercap/intel_rapl_tpmi.c +1 -0
@@ -40,6 +40,7 @@ enum tpmi_rapl_register {TPMI_RAPL_REG_ENERGY_STATUS ,TPMI_RAPL_REG_PERF_STATUS ,TPMI_RAPL_REG_POWER_INFO , +TPMI_RAPL_REG_DOMAIN_INFO ,TPMI_RAPL_REG_INTERRUPT , TPMI_RAPL_REG_MAX = 15, };
Signature · 1 line
2.34.1
Zhang Rui
reviewerPATCH 4/6 powercap: intel_rapl_tpmi: Fix System Domain probing
In reply to Zhang RuiOnly domain root packages can enumerate System (Psys) domain. Whether a package is domain root or not is described in the Bit 0 of the Domain Info register.
Add support for Domain Info register and fix the System domain probing accordingly.
--- drivers/powercap/intel_rapl_tpmi.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)
drivers/powercap/intel_rapl_tpmi.c +14 -0
@@ -131,6 +131,12 @@ static void trp_release(struct tpmi_rapl_package *trp)mutex_unlock (&tpmi_rapl_lock); } +/* +* Bit 0 of TPMI_RAPL_REG_DOMAIN_INFO indicates if the current package is a domain +* root or not. Only domain root packages can enumerate System (Psys) Domain. +*/ +#define TPMI_RAPL_DOMAIN_ROOT BIT(0) +static int parse_one_domain (struct tpmi_rapl_package *trp,u32 offset) {u8 tpmi_domain_version;
@@ -140,6 +146,7 @@ static int parse_one_domain(struct tpmi_rapl_package *trp, u32 offset)enum rapl_domain_reg_id reg_id;int tpmi_domain_size, tpmi_domain_flags;u64 tpmi_domain_header =readq (trp->base + offset); +u64 tpmi_domain_info;/* Domain Parent bits are ignored for now */ tpmi_domain_version = tpmi_domain_header &0xff ;
@@ -170,6 +177,13 @@ static int parse_one_domain(struct tpmi_rapl_package *trp, u32 offset) domain_type =RAPL_DOMAIN_PACKAGE ;break ;case TPMI_RAPL_DOMAIN_SYSTEM : +if (!(tpmi_domain_flags &BIT (TPMI_RAPL_REG_DOMAIN_INFO ))) { +pr_warn (FW_BUG "System domain must support Domain Info register\n" ); +return -ENODEV ; + } + tpmi_domain_info =readq (trp->base + offset +TPMI_RAPL_REG_DOMAIN_INFO ); +if (!(tpmi_domain_info &TPMI_RAPL_DOMAIN_ROOT )) +return 0 ; domain_type =RAPL_DOMAIN_PLATFORM ;break ;case TPMI_RAPL_DOMAIN_MEMORY :
Signature · 1 line
2.34.1
Zhang Rui
reviewerPATCH 5/6 powercap: intel_rapl: Add support for LNL-M paltform
In reply to Zhang RuiAdd support for LNL-M platform.
--- drivers/powercap/intel_rapl_common.c | 1 + 1 file changed, 1 insertion(+)
drivers/powercap/intel_rapl_common.c +1 -0
@@ -1262,6 +1262,7 @@ static const struct x86_cpu_id rapl_ids[] __initconst = {X86_MATCH_INTEL_FAM6_MODEL (METEORLAKE_L , &rapl_defaults_core),X86_MATCH_INTEL_FAM6_MODEL (SAPPHIRERAPIDS_X , &rapl_defaults_spr_server),X86_MATCH_INTEL_FAM6_MODEL (EMERALDRAPIDS_X , &rapl_defaults_spr_server), +X86_MATCH_INTEL_FAM6_MODEL (LUNARLAKE_M , &rapl_defaults_core),X86_MATCH_INTEL_FAM6_MODEL (LAKEFIELD , &rapl_defaults_core), X86_MATCH_INTEL_FAM6_MODEL(ATOM_SILVERMONT, &rapl_defaults_byt),
Signature · 1 line
2.34.1
Zhang Rui
reviewerPATCH 6/6 powercap: intel_rapl: add support for Arrow Lake
In reply to Zhang RuiFrom: Sumeet Pawnikar <[email protected]>
Add support for LNL-M platform.
--- drivers/powercap/intel_rapl_common.c | 1 + 1 file changed, 1 insertion(+)
drivers/powercap/intel_rapl_common.c +1 -0
@@ -1263,6 +1263,7 @@ static const struct x86_cpu_id rapl_ids[] __initconst = {X86_MATCH_INTEL_FAM6_MODEL (SAPPHIRERAPIDS_X , &rapl_defaults_spr_server),X86_MATCH_INTEL_FAM6_MODEL (EMERALDRAPIDS_X , &rapl_defaults_spr_server),X86_MATCH_INTEL_FAM6_MODEL (LUNARLAKE_M , &rapl_defaults_core), +X86_MATCH_INTEL_FAM6_MODEL (ARROWLAKE , &rapl_defaults_core),X86_MATCH_INTEL_FAM6_MODEL (LAKEFIELD , &rapl_defaults_core), X86_MATCH_INTEL_FAM6_MODEL(ATOM_SILVERMONT, &rapl_defaults_byt),
Signature · 1 line
2.34.1
Rafael J. Wysocki
maintainerQuoted 28 lines · Zhang Rui
Patch 1/6 fixes a real bug when MMIO RAPL driver is probed on platforms that are not listed in current CPU model list. IMO, it should be considered as stable material.
Patch 2/6 fixes a potential racing issue, but I have not reproduced it yet.
Patch 3/6 ~ 4/6 fix a problem that TPMI RAPL driver probes disabled System (Psys) RAPL Domains.
Patch 5/6 and 6/6 are two simple CPU model updates to support MSR RAPL on Arrowlake and Lunarlake platforms.
thanks,
rui----------------------------------------------------------------
Sumeet Pawnikar (1):
powercap:intel_rapl: add support for Arrow LakeZhang Rui (5):
powercap:intel_rapl: Fix a NULL pointer reference bug
powercap:intel_rapl: Fix locking for TPMI RAPL
powercap:intel_rapl_tpmi: Fix a register bug
powercap:intel_rapl_tpmi: Fix System Domain probing
powercap:intel_rapl: Add support for LNL-M paltform
Quoted 6 lines
drivers/powercap/intel_rapl_common.c | 36 ++++++++++++++++++++-- drivers/powercap/intel_rapl_msr.c | 8 ++--- drivers/powercap/intel_rapl_tpmi.c | 15 +++++++++ .../intel/int340x_thermal/processor_thermal_rapl.c | 8 ++--- include/linux/intel_rapl.h | 6 ++++ 5 files changed, 62 insertions(+), 11 deletions(-)
All applied as 6.9 material with some minor changes in the subject and changelogs of the last two patches.
Thanks!