powerpc: Detection and scheduler optimization for POWER9 bigcore
PATCH v2 2/2 powerpc: Enable CPU_FTR_ASYM_SMT for interleaved big-cores
In reply to Gautham R. ShenoyFrom: "Gautham R. Shenoy" <[email protected]>
A pair of IBM POWER9 SMT4 cores can be fused together to form a big-core with 8 SMT threads. This can be discovered via the "ibm,thread-groups" CPU property in the device tree which will indicate which group of threads that share the L1 cache, translation cache and instruction data flow. If there are multiple such group of threads, then the core is a big-core.
Furthermore, if the thread-ids of the threads of the big-core can be obtained by interleaving the thread-ids of the thread-groups (component small core), then such a big-core is called an interleaved big-core.
Eg: Threads in the pair of component SMT4 cores of an interleaved big-core are numbered {0,2,4,6} and {1,3,5,7} respectively.
The SMT4 cores forming a big-core are more or less independent units. Thus when multiple tasks are scheduled to run on the fused core, we get the best performance when the tasks are spread across the pair of SMT4 cores.
This patch enables CPU_FTR_ASYM_SMT bit in the cpu-features on detecting the presence of interleaved big-cores at boot up. This will will bias the load-balancing of tasks on smaller numbered threads, which will automatically result in spreading the tasks uniformly across the associated pair of SMT4 cores.
--- arch/powerpc/kernel/setup-common.c | 67 +++++++++++++++++++++++++++++++++++++- 1 file changed, 66 insertions(+), 1 deletion(-)
arch/powerpc/kernel/setup-common.c +66 -1
@@ -537,6 +537,56 @@ int get_cpu_thread_group_start(int cpu, struct thread_groups *tg)return -1 ; } +/* +* check_interleaved_big_core - Checks if the thread group tg +* corresponds to a big-core whose threads are interleavings of the +* threads of the component small cores. +* +* @tg: A thread-group struct for the core. +* +* Returns true if the core is a interleaved big-core. +* Returns false otherwise. +*/ +static inline bool check_interleaved_big_core (struct thread_groups *tg) +{ +int nr_groups; +int threads_per_group; +int cur_cpu, next_cpu, i, j; + + nr_groups = tg->nr_groups; + threads_per_group = tg->threads_per_group; + +if (tg->property !=1 ) +return false; + +if (nr_groups <2 || threads_per_group <2 ) +return false; + +/* +* In case of an interleaved big-core, the thread-ids of the +* big-core can be obtained by interleaving the the thread-ids +* of the component small +* +* Eg: On a 8-thread big-core with two SMT4 small cores, the +* threads of the two component small cores will be +* {0, 2, 4, 6} and {1, 3, 5, 7}. +*/ +for (i =0 ; i < nr_groups; i++) { +int group_start = i * threads_per_group; + +for (j =0 ; j < threads_per_group -1 ; j++) { +int cur_idx = group_start + j; + + cur_cpu = tg->thread_list[cur_idx]; + next_cpu = tg->thread_list[cur_idx +1 ]; +if (next_cpu != cur_cpu + nr_groups) +return false; + } + } + +return true; +} +/** * setup_cpu_maps - initialize the following cpu maps: * cpu_possible_mask
@@ -560,6 +610,7 @@ void __init smp_setup_cpu_maps(void)struct device_node *dn;int cpu =0 ;int nthreads =1 ; +bool has_interleaved_big_core = true;DBG ("smp_setup_cpu_maps()\n" );
@@ -613,6 +664,12 @@ void __init smp_setup_cpu_maps(void)if (parse_thread_groups (dn, &tg) || tg.nr_groups <1 || tg.property !=1 ) { has_big_cores = false; + has_interleaved_big_core = false; + } + +if (has_interleaved_big_core) { + has_interleaved_big_core = +check_interleaved_big_core (&tg); }if (cpu >= nr_cpu_ids) {
@@ -669,7 +726,15 @@ void __init smp_setup_cpu_maps(void) vdso_data->processorCount =num_present_cpus ();#endif /* CONFIG_PPC64 */ -/* Initialize CPU <=> thread mapping/ +if (has_interleaved_big_core) { +int key =__builtin_ctzl (CPU_FTR_ASYM_SMT ); + + cur_cpu_spec->cpu_features |=CPU_FTR_ASYM_SMT ; +static_branch_enable (&cpu_feature_keys[key]); +pr_info ("Detected interleaved big-cores\n" ); + } + +/* Initialize CPU <=> thread mapping/ * * WARNING: We assume that the number of threads is the same for * every CPU in the system. If that is not the case, then some code
Signature · 1 line
1.9.4
PATCH v2 1/2 powerpc: Detect the presence of big-cores via "ibm,thread-groups"
In reply to Gautham R. ShenoyFrom: "Gautham R. Shenoy" <[email protected]>
On IBM POWER9, the device tree exposes a property array identifed by "ibm,thread-groups" which will indicate which groups of threads share a particular set of resources.
As of today we only have one form of grouping identifying the group of threads in the core that share the L1 cache, translation cache and instruction data flow.
This patch defines the helper function to parse the contents of "ibm,thread-groups" and a new structure to contain the parsed output.
The patch also creates the sysfs file named "small_core_siblings" that returns the physical ids of the threads in the core that share the L1 cache, translation cache and instruction data flow.
--- Documentation/ABI/testing/sysfs-devices-system-cpu | 8 ++ arch/powerpc/include/asm/cputhreads.h | 22 +++++ arch/powerpc/kernel/setup-common.c | 110 +++++++++++++++++++++ arch/powerpc/kernel/sysfs.c | 35 +++++++ 4 files changed, 175 insertions(+)
Documentation/ABI/testing/sysfs-devices-system-cpu +8 -0
@@ -487,3 +487,11 @@ Description: Information about CPU vulnerabilities "Not affected" CPU is not affected by the vulnerability "Vulnerable" CPU is affected and no mitigation in effect "Mitigation: $M" CPU is affected and mitigation $M is in effect + +What: /sys/devices/system/cpu/cpu[0-9]+/small_core_sibings +Date: 03-Jul-2018 +KernelVersion: v4.18.0 +Contact: Gautham R. Shenoy <[email protected]> +Description: List of Physical ids of CPUs which share the the L1 cache, + translation cache and instruction data-flow with this CPU. +Values: Comma separated list of decimal integers.
arch/powerpc/include/asm/cputhreads.h +22 -0
@@ -23,11 +23,13 @@extern int threads_per_core;extern int threads_per_subcore;extern int threads_shift; +extern bool has_big_cores;extern cpumask_t threads_core_mask;#else #define threads_per_core 1#define threads_per_subcore 1#define threads_shift 0 +#define has_big_cores 0#define threads_core_mask (*get_cpu_mask(0))#endif
@@ -69,12 +71,32 @@ static inline cpumask_t cpu_online_cores_map(void)return cpu_thread_mask_to_cores (cpu_online_mask); } +#define MAX_THREAD_LIST_SIZE 8 +struct thread_groups { +unsigned int property; +unsigned int nr_groups; +unsigned int threads_per_group; +unsigned int thread_list[MAX_THREAD_LIST_SIZE ]; +}; +#ifdef CONFIG_SMP int cpu_core_index_of_thread (int cpu);int cpu_first_thread_of_core (int core); +int parse_thread_groups (struct device_node *dn,struct thread_groups *tg); +int get_cpu_thread_group_start (int cpu,struct thread_groups *tg);#else static inline int cpu_core_index_of_thread (int cpu) {return cpu; }static inline int cpu_first_thread_of_core (int core) {return core; } +static inline int parse_thread_groups (struct device_node *dn, +struct thread_groups *tg) +{ +return -ENODATA ; +} + +static inline int get_cpu_thread_group_start (int cpu,struct thread_groups *tg) +{ +return -1 ; +}#endif static inline int cpu_thread_in_core (int cpu)
arch/powerpc/kernel/setup-common.c +110 -0
138 lines not on this page · raw message
arch/powerpc/kernel/sysfs.c +35 -0
@@ -18,6 +18,7 @@#include <asm/smp.h> #include <asm/pmc.h> #include <asm/firmware.h> +#include <asm/cputhreads.h> #include "cacheinfo.h" #include "setup.h"
@@ -1025,6 +1026,33 @@ static ssize_t show_physical_id(struct device *dev, } static DEVICE_ATTR(physical_id, 0444, show_physical_id, NULL); +static ssize_t show_small_core_siblings (struct device *dev, +struct device_attribute *attr, +char *buf) +{ + struct cpu *cpu = container_of(dev, struct cpu, dev); +struct device_node *dn =of_get_cpu_node (cpu->dev.id,NULL ); +struct thread_groups tg; +int i, j; +ssize_t ret =0 ; + +if (parse_thread_groups (dn, &tg)) +return -ENODATA ; + + i =get_cpu_thread_group_start (cpu->dev.id, &tg); + +if (i ==-1 ) +return -ENODATA ; + +for (j =0 ; j < tg.threads_per_group -1 ; j++) + ret +=sprintf (buf + ret,"%d," , tg.thread_list[i + j]); + + ret +=sprintf (buf + ret,"%d\n" , tg.thread_list[i + j]); + +return ret; +} +static DEVICE_ATTR(small_core_siblings, 0444, show_small_core_siblings, NULL); +static int __init topology_init (void ) {int cpu, r;
@@ -1048,6 +1076,13 @@ static int __init topology_init(void)register_cpu (c, cpu);device_create_file (&c->dev, &dev_attr_physical_id); + +if (has_big_cores) { +const struct device_attribute *attr = + &dev_attr_small_core_siblings; + +device_create_file (&c->dev, attr); + } } } r = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "powerpc/topology:online",
Signature · 1 line
1.9.4
v2 cover powerpc: Detection and scheduler optimization for POWER9 bigcore
From: "Gautham R. Shenoy" <[email protected]>
Hi,
This is the second iteration of the patchset to add support for big-core on POWER9.
The earlier version can be found here: https://lkml.org/lkml/2018/5/11/245.
The changes from the previous version:
- Added comments explaining the "ibm,thread-groups" device tree property.
- Uses cleaner device-tree parsing functions to parse the u32 arrays.
- Adds a sysfs file listing the small-core siblings for every CPU.
- Enables the scheduler optimization by setting the CPU_FTR_ASYM_SMT bit in the cur_cpu_spec->cpu_features on detecting the presence of interleaved big-core.
- Handles the corner case where there is only a single thread-group or when there is a single thread in a thread-group.
Description:
~~~~~~~~~~~~~~~~~~~~
A pair of IBM POWER9 SMT4 cores can be fused together to form a big-core with 8 SMT threads. This can be discovered via the "ibm,thread-groups" CPU property in the device tree which will indicate which group of threads that share the L1 cache, translation cache and instruction data flow. If there are multiple such group of threads, then the core is a big-core. Furthermore, the thread-ids of such a big-core is obtained by interleaving the thread-ids of the component SMT4 cores.
Eg: Threads in the pair of component SMT4 cores of an interleaved big-core are numbered {0,2,4,6} and {1,3,5,7} respectively.
On such a big-core, when multiple tasks are scheduled to run on the big-core, we get the best performance when the tasks are spread across the pair of SMT4 cores.
The Linux scheduler supports a flag called "SD_ASYM_PACKING" which when set in the SMT sched-domain, biases the load-balancing of the tasks on the smaller numbered threads in the core. On an big-core whose threads are interleavings of the threads of the small cores, enabling SD_ASYM_PACKING in the SMT sched-domain automatically results in spreading the tasks uniformly across the associated pair of SMT4 cores, thereby yielding better performance.
This patchset contains two patches which on detecting the presence of interleaved big-cores will enable the the CPU_FTR_ASYM_SMT bit in the cur_cpu_spec->cpu_feature.
Patch 1: adds support to detect the presence of
big-cores and reports the small-core siblings of each CPU X via the sysfs file "/sys/devices/system/cpu/cpuX/big_core_siblings".
Patch 2: checks if the thread-ids of the component small-cores are interleaved, in which case we enable the the CPU_FTR_ASYM_SMT bit in the cur_cpu_spec->cpu_features which results in the SD_ASYM_PACKING flag being set at the SMT level sched-domain.
Results:
~~~~~~~~~~~~~~~~~
Experimental results for ebizzy with 2 threads, bound to a single big-core show a marked improvement with this patchset over the 4.18-rc3 vanilla kernel.
The result of 100 such runs for 4.18-rc3 kernel and the 4.18-rc3 + big-core-patches are as follows
4.18-rc3:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
records/s : # samples : Histogram
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[0 - 1000000] : 0 : # [1000000 - 2000000] : 3 : # [2000000 - 3000000] : 16 : #### [3000000 - 4000000] : 11 : ### [4000000 - 5000000] : 0 : # [5000000 - 6000000] : 70 : ###############
4.18-rc3 + big-core-patches
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
records/s : # samples : Histogram
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[0 - 1000000] : 0 : # [1000000 - 2000000] : 0 : # [2000000 - 3000000] : 6 : ## [3000000 - 4000000] : 0 : # [4000000 - 5000000] : 2 : # [5000000 - 6000000] : 92 : ###################
Gautham R. Shenoy (2):
powerpc: Detect the presence of big-cores via "ibm,thread-groups"
powerpc: Enable CPU_FTR_ASYM_SMT for interleaved big-cores
Documentation/ABI/testing/sysfs-devices-system-cpu | 8 + arch/powerpc/include/asm/cputhreads.h | 22 +++ arch/powerpc/kernel/setup-common.c | 177 ++++++++++++++++++++- arch/powerpc/kernel/sysfs.c | 35 ++++ 4 files changed, 241 insertions(+), 1 deletion(-)
Signature · 1 line
1.9.4
Re: [v2 PATCH 1/2] powerpc: Detect the presence of big-cores via "ibm, thread-groups"
In reply to Gautham R. ShenoyQuoted 30 lines · Gautham R. Shenoy
From: "Gautham R. Shenoy" <[email protected]>
On IBM POWER9, the device tree exposes a property array identifed by "ibm,thread-groups" which will indicate which groups of threads share a particular set of resources.
As of today we only have one form of grouping identifying the group of threads in the core that share the L1 cache, translation cache and instruction data flow.
This patch defines the helper function to parse the contents of "ibm,thread-groups" and a new structure to contain the parsed output.
The patch also creates the sysfs file named "
small_core_siblings" that returns the physical ids of the threads in the core that share the L1 cache, translation cache and instruction data flow.Signed-off-by Gautham R. Shenoy <[email protected]>--- Documentation/ABI/testing/sysfs-devices-system-cpu | 8 ++ arch/powerpc/include/asm/cputhreads.h | 22 +++++ arch/powerpc/kernel/setup-common.c | 110 +++++++++++++++++++++ arch/powerpc/kernel/sysfs.c | 35 +++++++ 4 files changed, 175 insertions(+)Documentation/ABI/testing/sysfs-devices-system-cpu +0
-0
Documentation/ABI/testing/sysfs-devices-system-cpu Description: Information about CPU vulnerabilities · line 491
Quoted 5 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +2
-0@@ -487,3 +487,11 @@ Description: Information about CPU vulnerabilities "Not affected" CPU is not affected by the vulnerability "Vulnerable" CPU is affected and no mitigation in effect "Mitigation: $M" CPU is affected and mitigation $M is in effect + +What: /sys/devices/system/cpu/cpu[0-9]+/small_core_sibings
s/small_core_sibings/small_core_siblings
By the way, big_core_siblings was mentioned in the introductory email.
Quoted 75 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +6
-0+Date: 03-Jul-2018 +KernelVersion: v4.18.0 +Contact: Gautham R. Shenoy <[email protected]> +Description: List of Physical ids of CPUs which share the the L1 cache, + translation cache and instruction data-flow with this CPU. +Values: Comma separated list of decimal integers.arch/powerpc/include/asm/cputhreads.h +22
-0@@ -23,11 +23,13 @@extern int threads_per_core;extern int threads_per_subcore;extern int threads_shift; +extern bool has_big_cores;extern cpumask_t threads_core_mask;#else #define threads_per_core 1#define threads_per_subcore 1#define threads_shift 0 +#define has_big_cores 0#define threads_core_mask (*get_cpu_mask(0))#endif @@ -69,12 +71,32 @@ static inline cpumask_t cpu_online_cores_map(void)return cpu_thread_mask_to_cores (cpu_online_mask); } +#define MAX_THREAD_LIST_SIZE 8 +struct thread_groups { +unsigned int property; +unsigned int nr_groups; +unsigned int threads_per_group; +unsigned int thread_list[MAX_THREAD_LIST_SIZE ]; +}; +#ifdef CONFIG_SMP int cpu_core_index_of_thread (int cpu);int cpu_first_thread_of_core (int core); +int parse_thread_groups (struct device_node *dn,struct thread_groups *tg); +int get_cpu_thread_group_start (int cpu,struct thread_groups *tg);#else static inline int cpu_core_index_of_thread (int cpu) {return cpu; }static inline int cpu_first_thread_of_core (int core) {return core; } +static inline int parse_thread_groups (struct device_node *dn, +struct thread_groups *tg) +{ +return -ENODATA ; +} + +static inline int get_cpu_thread_group_start (int cpu,struct thread_groups *tg) +{ +return -1 ; +}#endif static inline int cpu_thread_in_core (int cpu)arch/powerpc/kernel/setup-common.c +2
-0@@ -402,10 +402,12 @@ void __init check_for_initrd(void)#ifdef CONFIG_SMP int threads_per_core, threads_per_subcore, threads_shift; +bool has_big_cores = true;cpumask_t threads_core_mask;EXPORT_SYMBOL_GPL (threads_per_core);EXPORT_SYMBOL_GPL (threads_per_subcore);EXPORT_SYMBOL_GPL (threads_shift); +EXPORT_SYMBOL_GPL (has_big_cores);EXPORT_SYMBOL_GPL (threads_core_mask);static void __init cpu_init_thread_core_maps (int tpc)
arch/powerpc/kernel/setup-common.c static void __init cpu_init_thread_core_maps(int tpc) · line 442
Quoted 8 lines
arch/powerpc/kernel/setup-common.c +5
-0@@ -433,6 +435,108 @@ static void __init cpu_init_thread_core_maps(int tpc)u32 *cpu_to_phys_id =NULL ; +/* +* parse_thread_groups: Parses the "ibm,thread-groups" device tree +* property for the CPU device node dn and stores +* the parsed output in the thread_groups +* structure tg.
Perhaps document the arguments of this function, as done in the second patch?
Quoted 61 lines
arch/powerpc/kernel/setup-common.c +61
-0+* +* ibm,thread-groups[0..N-1] array defines which group of threads in +* the CPU-device node can be grouped together based on the property. +* +* ibm,thread-groups[0] tells us the property based on which the +* threads are being grouped together. If this value is 1, it implies +* that the threads in the same group share L1, translation cache. +* +* ibm,thread-groups[1] tells us how many such thread groups exist. +* +* ibm,thread-groups[2] tells us the number of threads in each such +* group. +* +* ibm,thread-groups[3..N-1] is the list of threads identified by +* "ibm,ppc-interrupt-server#s" arranged as per their membership in +* the grouping. +* +* Example: If ibm,thread-groups = [1,2,4,5,6,7,8,9,10,11,12] it +* implies that there are 2 groups of 4 threads each, where each group +* of threads share L1, translation cache. +* +* The "ibm,ppc-interrupt-server#s" of the first group is {5,6,7,8} +* and the "ibm,ppc-interrupt-server#s" of the second group is {9, 10, +* 11, 12} structure +* +* Returns 0 on success, -EINVAL if the property does not exist, +* -ENODATA if property does not have a value, and -EOVERFLOW if the +* property data isn't large enough. +*/ +int parse_thread_groups (struct device_node *dn, +struct thread_groups *tg) +{ +unsigned int nr_groups, threads_per_group, property; +int i; +u32 thread_group_array[3 +MAX_THREAD_LIST_SIZE ]; +u32 *thread_list; +size_t total_threads; +int ret; + + ret =of_property_read_u32_array (dn,"ibm,thread-groups" , + thread_group_array,3 ); + +if (ret) +return ret; + + property = thread_group_array[0 ]; + nr_groups = thread_group_array[1 ]; + threads_per_group = thread_group_array[2 ]; + total_threads = nr_groups * threads_per_group; + + ret =of_property_read_u32_array (dn,"ibm,thread-groups" , + thread_group_array, +3 + total_threads); +if (ret) +return ret; + + thread_list = &thread_group_array[3 ]; + +for (i =0 ; i < total_threads; i++) + tg->thread_list[i] = thread_list[i]; +
arch/powerpc/kernel/setup-common.c static void __init cpu_init_thread_core_maps(int tpc) · line 513
Quoted 10 lines
arch/powerpc/kernel/setup-common.c +10
-0+ tg->property = property; + tg->nr_groups = nr_groups; + tg->threads_per_group = threads_per_group; + +return 0 ; +} + +/* +* get_cpu_thread_group_start : Searches the thread group in tg->thread_list +* that @cpu belongs to.
Same here.
Quoted 30 lines
arch/powerpc/kernel/setup-common.c +26
-0+* +* Returns the index to tg->thread_list that points to the the start +* of the thread_group that @cpu belongs to. +* +* Returns -1 if cpu doesn't belong to any of the groups pointed +* to by tg->thread_list. +*/ +int get_cpu_thread_group_start (int cpu,struct thread_groups *tg) +{ +int hw_cpu_id =get_hard_smp_processor_id (cpu); +int i, j; + +for (i =0 ; i < tg->nr_groups; i++) { +int group_start = i * tg->threads_per_group; + +for (j =0 ; j < tg->threads_per_group; j++) { +int idx = group_start + j; + +if (tg->thread_list[idx] == hw_cpu_id) +return group_start; + } + } + +return -1 ; +} +/** * setup_cpu_maps - initialize the following cpu maps: * cpu_possible_mask
arch/powerpc/kernel/setup-common.c void __init smp_setup_cpu_maps(void) · line 574
Quoted 4 lines
arch/powerpc/kernel/setup-common.c +1
-0@@ -467,6 +571,7 @@ void __init smp_setup_cpu_maps(void)const __be32 *intserv;__be32 cpu_be;int j, len; +struct thread_groups tg = {.nr_groups =0 };
We assume has_big_cores = true but here we initialize .nr_groups otherwise. It's kind of contradictory.
What if has_big_cores is assumed false and members of tg are initialized with zeroes?
arch/powerpc/kernel/setup-common.c void __init smp_setup_cpu_maps(void) · line 617
Quoted 12 lines
arch/powerpc/kernel/setup-common.c +5
-0DBG (" * %pOF...\n" , dn);@@ -505,6 +610,11 @@ void __init smp_setup_cpu_maps(void) cpu++; } +if (parse_thread_groups (dn, &tg) || + tg.nr_groups <1 || tg.property !=1 ) { + has_big_cores = false; + } +
parse_thread_groups() returns before setting tg.property if property doesn't exist. Are we confident that tg.property won't contain any garbage that could lead to a false positive here? Shouldn't we also initialize .property when declaring tg?
What if this logic is encapsulated in a function? For example:
has_big_cores = dt_has_big_cores(dn, &tg);
Quoted 66 lines
arch/powerpc/kernel/setup-common.c +0
-0if (cpu >= nr_cpu_ids) {of_node_put (dn);break ;arch/powerpc/kernel/sysfs.c +35
-0@@ -18,6 +18,7 @@#include <asm/smp.h> #include <asm/pmc.h> #include <asm/firmware.h> +#include <asm/cputhreads.h> #include "cacheinfo.h" #include "setup.h" @@ -1025,6 +1026,33 @@ static ssize_t show_physical_id(struct device *dev, } static DEVICE_ATTR(physical_id, 0444, show_physical_id, NULL); +static ssize_t show_small_core_siblings (struct device *dev, +struct device_attribute *attr, +char *buf) +{ + struct cpu *cpu = container_of(dev, struct cpu, dev); +struct device_node *dn =of_get_cpu_node (cpu->dev.id,NULL ); +struct thread_groups tg; +int i, j; +ssize_t ret =0 ; + +if (parse_thread_groups (dn, &tg)) +return -ENODATA ; + + i =get_cpu_thread_group_start (cpu->dev.id, &tg); + +if (i ==-1 ) +return -ENODATA ; + +for (j =0 ; j < tg.threads_per_group -1 ; j++) + ret +=sprintf (buf + ret,"%d," , tg.thread_list[i + j]); + + ret +=sprintf (buf + ret,"%d\n" , tg.thread_list[i + j]); + +return ret; +} +static DEVICE_ATTR(small_core_siblings, 0444, show_small_core_siblings, NULL); +static int __init topology_init (void ) {int cpu, r;@@ -1048,6 +1076,13 @@ static int __init topology_init(void)register_cpu (c, cpu);device_create_file (&c->dev, &dev_attr_physical_id); + +if (has_big_cores) { +const struct device_attribute *attr = + &dev_attr_small_core_siblings; + +device_create_file (&c->dev, attr); + } } } r = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "powerpc/topology:online",Signature · 1 line
1.9.4
Cheers
Murilo
Re: [v2 PATCH 2/2] powerpc: Enable CPU_FTR_ASYM_SMT for interleaved big-cores
In reply to Gautham R. ShenoyQuoted 117 lines · Gautham R. Shenoy
From: "Gautham R. Shenoy" <[email protected]>
A pair of IBM POWER9 SMT4 cores can be fused together to form a big-core with 8 SMT threads. This can be discovered via the "ibm,thread-groups" CPU property in the device tree which will indicate which group of threads that share the L1 cache, translation cache and instruction data flow. If there are multiple such group of threads, then the core is a big-core.
Furthermore, if the thread-ids of the threads of the big-core can be obtained by interleaving the thread-ids of the thread-groups (component small core), then such a big-core is called an interleaved big-core.
Eg: Threads in the pair of component SMT4 cores of an interleaved big-core are numbered {0,2,4,6} and {1,3,5,7} respectively.
The SMT4 cores forming a big-core are more or less independent units. Thus when multiple tasks are scheduled to run on the fused core, we get the best performance when the tasks are spread across the pair of SMT4 cores.
This patch enables
CPU_FTR_ASYM_SMTbit in the cpu-features on detecting the presence of interleaved big-cores at boot up. This will will bias the load-balancing of tasks on smaller numbered threads, which will automatically result in spreading the tasks uniformly across the associated pair of SMT4 cores.Signed-off-by Gautham R. Shenoy <[email protected]>--- arch/powerpc/kernel/setup-common.c | 67 +++++++++++++++++++++++++++++++++++++- 1 file changed, 66 insertions(+), 1 deletion(-)arch/powerpc/kernel/setup-common.c +57
-0@@ -537,6 +537,56 @@ int get_cpu_thread_group_start(int cpu, struct thread_groups *tg)return -1 ; } +/* +* check_interleaved_big_core - Checks if the thread group tg +* corresponds to a big-core whose threads are interleavings of the +* threads of the component small cores. +* +* @tg: A thread-group struct for the core. +* +* Returns true if the core is a interleaved big-core. +* Returns false otherwise. +*/ +static inline bool check_interleaved_big_core (struct thread_groups *tg) +{ +int nr_groups; +int threads_per_group; +int cur_cpu, next_cpu, i, j; + + nr_groups = tg->nr_groups; + threads_per_group = tg->threads_per_group; + +if (tg->property !=1 ) +return false; + +if (nr_groups <2 || threads_per_group <2 ) +return false; + +/* +* In case of an interleaved big-core, the thread-ids of the +* big-core can be obtained by interleaving the the thread-ids +* of the component small +* +* Eg: On a 8-thread big-core with two SMT4 small cores, the +* threads of the two component small cores will be +* {0, 2, 4, 6} and {1, 3, 5, 7}. +*/ +for (i =0 ; i < nr_groups; i++) { +int group_start = i * threads_per_group; + +for (j =0 ; j < threads_per_group -1 ; j++) { +int cur_idx = group_start + j; + + cur_cpu = tg->thread_list[cur_idx]; + next_cpu = tg->thread_list[cur_idx +1 ]; +if (next_cpu != cur_cpu + nr_groups) +return false; + } + } + +return true; +} +/** * setup_cpu_maps - initialize the following cpu maps: * cpu_possible_mask @@ -560,6 +610,7 @@ void __init smp_setup_cpu_maps(void)struct device_node *dn;int cpu =0 ;int nthreads =1 ; +bool has_interleaved_big_core = true;DBG ("smp_setup_cpu_maps()\n" );@@ -613,6 +664,12 @@ void __init smp_setup_cpu_maps(void)if (parse_thread_groups (dn, &tg) || tg.nr_groups <1 || tg.property !=1 ) { has_big_cores = false; + has_interleaved_big_core = false; + } + +if (has_interleaved_big_core) { + has_interleaved_big_core = +check_interleaved_big_core (&tg); }if (cpu >= nr_cpu_ids) {@@ -669,7 +726,15 @@ void __init smp_setup_cpu_maps(void) vdso_data->processorCount =num_present_cpus ();
arch/powerpc/kernel/setup-common.c void __init smp_setup_cpu_maps(void) · line 735
Quoted 10 lines
arch/powerpc/kernel/setup-common.c +7
-1#endif /* CONFIG_PPC64 */ -/* Initialize CPU <=> thread mapping/ +if (has_interleaved_big_core) { +int key =__builtin_ctzl (CPU_FTR_ASYM_SMT ); + + cur_cpu_spec->cpu_features |=CPU_FTR_ASYM_SMT ; +static_branch_enable (&cpu_feature_keys[key]); +pr_info ("Detected interleaved big-cores\n" ); + }
Shouldn't we use cpu_has_feature(CPU_FTR_ASYM_SMT) before setting it?
Quoted 8 lines
arch/powerpc/kernel/setup-common.c +2
-0+ +/* Initialize CPU <=> thread mapping/ * * WARNING: We assume that the number of threads is the same for * every CPU in the system. If that is not the case, then some code Signature · 1 line
1.9.4
Signature · 1 line
Murilo
Re: [v2 PATCH 2/2] powerpc: Enable CPU_FTR_ASYM_SMT for interleaved big-cores
Hi Murilo,
Thanks for the review.
On Tue, Jul 03, 2018 at 02:53:46PM -0300, Murilo Opsfelder Araujo wrote:
[..snip..]
Quoted 11 lines
arch/powerpc/kernel/setup-common.c void __init smp_setup_cpu_maps(void) · line 735
Quoted 8 lines
arch/powerpc/kernel/setup-common.c +7
-1-/* Initialize CPU <=> thread mapping/ +if (has_interleaved_big_core) { +int key =__builtin_ctzl (CPU_FTR_ASYM_SMT ); + + cur_cpu_spec->cpu_features |=CPU_FTR_ASYM_SMT ; +static_branch_enable (&cpu_feature_keys[key]); +pr_info ("Detected interleaved big-cores\n" ); + }Shouldn't we use
cpu_has_feature(CPU_FTR_ASYM_SMT)before settingQuoted 1 line
it?
Are you suggesting that we do the following?
if (has_interleaved_big_core &&
!cpu_has_feature(CPU_FTR_ASYM_SMT)) {
...
}Currently CPU_FTR_ASYM_SMT is set at compile time for only POWER7 where running the tasks on lower numbered threads give us the benefit of SMT thread folding. Interleaved big core is a feature introduced only on POWER9. Thus, we know that CPU_FTR_ASYM_SMT is not set in cpu_features at this point.
Quoted 12 lines
Quoted 9 lines
arch/powerpc/kernel/setup-common.c +2
-0+ +/* Initialize CPU <=> thread mapping/ * * WARNING: We assume that the number of threads is the same for * every CPU in the system. If that is not the case, then some code Signature · 1 line
1.9.4Signature · 1 line
Murilo
--
Thanks and Regards
gautham.
Re: [v2 PATCH 1/2] powerpc: Detect the presence of big-cores via "ibm, thread-groups"
In reply to Murilo Opsfelder AraujoHello Murilo,
Thanks for reviewing the patch. Replies inline.
Quoted 26 lines · Murilo Opsfelder Araujo
Quoted 25 lines · Gautham R. Shenoy
From: "Gautham R. Shenoy" <[email protected]>
On IBM POWER9, the device tree exposes a property array identifed by "ibm,thread-groups" which will indicate which groups of threads share a particular set of resources.
As of today we only have one form of grouping identifying the group of threads in the core that share the L1 cache, translation cache and instruction data flow.
This patch defines the helper function to parse the contents of "ibm,thread-groups" and a new structure to contain the parsed output.
The patch also creates the sysfs file named "
small_core_siblings" that returns the physical ids of the threads in the core that share the L1 cache, translation cache and instruction data flow.Signed-off-by Gautham R. Shenoy <[email protected]>--- Documentation/ABI/testing/sysfs-devices-system-cpu | 8 ++ arch/powerpc/include/asm/cputhreads.h | 22 +++++ arch/powerpc/kernel/setup-common.c | 110 +++++++++++++++++++++ arch/powerpc/kernel/sysfs.c | 35 +++++++ 4 files changed, 175 insertions(+)
Quoted 12 lines
Documentation/ABI/testing/sysfs-devices-system-cpu Description: Information about CPU vulnerabilities · line 491
Quoted 10 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +2
-0@@ -487,3 +487,11 @@ Description: Information about CPU vulnerabilities "Not affected" CPU is not affected by the vulnerability "Vulnerable" CPU is affected and no mitigation in effect "Mitigation: $M" CPU is affected and mitigation $M is in effect + +What: /sys/devices/system/cpu/cpu[0-9]+/small_core_sibingss/small_core_sibings/small_core_siblings
Nice catch! Will fix this.
Quoted 2 lines
By the way,
big_core_siblingswas mentioned in the introductory
email.
It should be small_core_siblings in the introductory e-mail. My bad.
Quoted 7 lines
Quoted 7 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +6
-0+Date: 03-Jul-2018 +KernelVersion: v4.18.0 +Contact: Gautham R. Shenoy <[email protected]> +Description: List of Physical ids of CPUs which share the the L1 cache, + translation cache and instruction data-flow with this CPU. +Values: Comma separated list of decimal integers.
[..snip..]
Quoted 8 lines
Quoted 5 lines
+/*
+ *parse_thread_groups: Parses the "ibm,thread-groups" device tree+ * property for the CPU device node dn and stores + * the parsed output in the thread_groups + * structure tg.Perhaps document the arguments of this function, as done in the second patch?
Will do this. Thanks.
Quoted 66 lines
Quoted 66 lines
+ *
+ * ibm,thread-groups[0..N-1] array defines which group of threads in
+ * the CPU-device node can be grouped together based on the property.
+ *
+ * ibm,thread-groups[0] tells us the property based on which the
+ * threads are being grouped together. If this value is 1, it implies
+ * that the threads in the same group share L1, translation cache.
+ *
+ * ibm,thread-groups[1] tells us how many such thread groups exist.
+ *
+ * ibm,thread-groups[2] tells us the number of threads in each such
+ * group.
+ *
+ * ibm,thread-groups[3..N-1] is the list of threads identified by
+ * "ibm,ppc-interrupt-server#s" arranged as per their membership in
+ * the grouping.
+ *
+ * Example: If ibm,thread-groups = [1,2,4,5,6,7,8,9,10,11,12] it
+ * implies that there are 2 groups of 4 threads each, where each group
+ * of threads share L1, translation cache.
+ *
+ * The "ibm,ppc-interrupt-server#s" of the first group is {5,6,7,8}
+ * and the "ibm,ppc-interrupt-server#s" of the second group is {9, 10,
+ * 11, 12} structure
+ *
+ * Returns 0 on success,-EINVALif the property does not exist,
+ *-ENODATAif property does not have a value, and-EOVERFLOWif the
+ * property data isn't large enough.
+ */Documentation/ABI/testing/sysfs-devices-system-cpu +36
-0+int parse_thread_groups(struct device_node *dn, + struct thread_groups *tg) +{ + unsigned int nr_groups, threads_per_group, property; + int i; + u32 thread_group_array[3 + MAX_THREAD_LIST_SIZE]; + u32 *thread_list; + size_t total_threads; + int ret; + + ret = of_property_read_u32_array(dn, "ibm,thread-groups", + thread_group_array, 3); + + if (ret) + return ret; + + property = thread_group_array[0]; + nr_groups = thread_group_array[1]; + threads_per_group = thread_group_array[2]; + total_threads = nr_groups * threads_per_group; + + ret = of_property_read_u32_array(dn, "ibm,thread-groups", + thread_group_array, + 3 + total_threads); + if (ret) + return ret; + + thread_list = &thread_group_array[3]; + + for (i = 0 ; i < total_threads; i++) + tg->thread_list[i] = thread_list[i]; + + tg->property = property; + tg->nr_groups = nr_groups; + tg->threads_per_group = threads_per_group; +
Quoted 8 lines
Documentation/ABI/testing/sysfs-devices-system-cpu
Quoted 6 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +6
-0+ return 0; +} + +/* + * get_cpu_thread_group_start : Searches the thread group in tg->thread_list + * that @cpu belongs to.Same here.
Sure.
Quoted 30 lines
Quoted 30 lines
+ *
+ * Returns the index totg->thread_listthat points to the the start
+ * of thethread_groupthat @cpu belongs to.
+ *
+ * Returns -1 if cpu doesn't belong to any of the groups pointed
+ * to bytg->thread_list.
+ */Documentation/ABI/testing/sysfs-devices-system-cpu +19
-0+int get_cpu_thread_group_start(int cpu, struct thread_groups *tg) +{ + int hw_cpu_id = get_hard_smp_processor_id(cpu); + int i, j; + + for (i = 0; i < tg->nr_groups; i++) { + int group_start = i * tg->threads_per_group; + + for (j = 0; j < tg->threads_per_group; j++) { + int idx = group_start + j; + + if (tg->thread_list[idx] == hw_cpu_id) + return group_start; + } + } + + return -1; +} + /** * setup_cpu_maps - initialize the following cpu maps: * cpu_possible_mask
Quoted 8 lines
Documentation/ABI/testing/sysfs-devices-system-cpu void __init smp_setup_cpu_maps(void) · line 574
Quoted 5 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +1
-0@@ -467,6 +571,7 @@ void __init smp_setup_cpu_maps(void) const __be32 *intserv; __be32 cpu_be; int j, len; + struct thread_groups tg = {.nr_groups = 0};We assume
has_big_cores= true but here we initialize .nr_groups otherwise. It's kind of contradictory.
.nr_groups is being initialized to some sane value here. Perhaps I should move the initializations of tg.nr_groups and tg.property inside parse_thread_groups.
Quoted 3 lines
What if
has_big_coresis assumed false and members of tg are initialized with zeroes?
The idea here is that after parsing all the CPU nodes, the variable "has_big_cores" continues to remain to true if all the CPU nodes are big cores. Even if one of them isn't a big core (not sure if this is possible in practise) then we want to set it to false.
Hence we start with the assumption that has_big_cores is true, and switch it on finding even one core that is not a big-core.
But I got to know that this is an overkill since if the component small core is bad, the entire big-core is disabled. Thus it might be sufficient to just check for one CPU node, if it is a big core or not, and set the variable from "false" to "true".
Quoted 8 lines
Quoted 8 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +0
-0DBG(" * %pOF...\n", dn);@@ -505,6 +610,11 @@ void __init smp_setup_cpu_maps(void) cpu++; }
Quoted 10 lines
Documentation/ABI/testing/sysfs-devices-system-cpu void __init smp_setup_cpu_maps(void) · line 617
Quoted 5 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +5
-0+ if (parse_thread_groups(dn, &tg) || + tg.nr_groups < 1 || tg.property != 1) { + has_big_cores = false; + } +
parse_thread_groups()returns before setting tg.property if property doesn't exist. Are we confident that tg.property won't contain any garbage that could lead to a false positive here? Shouldn't we also initialize .property when declaring tg?
Yes we should. Will move the initializations to parse_thread_groups.
Quoted 4 lines
What if this logic is encapsulated in a function? For example:
has_big_cores = dt_has_big_cores(dn, &tg);
Good idea.
Quoted 5 lines
arch/powerpc/kernel/sysfs.c
Quoted 5 lines
Documentation/ABI/testing/sysfs-devices-system-cpu +0
-0if (cpu >= nr_cpu_ids) { of_node_put(dn); break;arch/powerpc/kernel/sysfs.c +0
-0
[..snip..]
Will address these changes in the subsequent patch series.
Quoted 3 lines
Cheers
Murilo
--
Thanks and Regards
gautham.
Re: [v2 PATCH 2/2] powerpc: Enable CPU_FTR_ASYM_SMT for interleaved big-cores
In reply to Gautham R ShenoyQuoted 27 lines · Gautham R Shenoy
Hi Murilo,
Thanks for the review.
On Tue, Jul 03, 2018 at 02:53:46PM -0300, Murilo Opsfelder Araujo wrote:
[..snip..]Quoted 11 lines
arch/powerpc/kernel/setup-common.c void __init smp_setup_cpu_maps(void) · line 735
Quoted 8 lines
arch/powerpc/kernel/setup-common.c +7
-1-/* Initialize CPU <=> thread mapping/ +if (has_interleaved_big_core) { +int key =__builtin_ctzl (CPU_FTR_ASYM_SMT ); + + cur_cpu_spec->cpu_features |=CPU_FTR_ASYM_SMT ; +static_branch_enable (&cpu_feature_keys[key]); +pr_info ("Detected interleaved big-cores\n" ); + }Shouldn't we use
cpu_has_feature(CPU_FTR_ASYM_SMT)before settingQuoted 1 line
it?
Are you suggesting that we do the following?
if (has_interleaved_big_core && !cpu_has_feature(CPU_FTR_ASYM_SMT)) { ... }
Quoted 5 lines
Currently
CPU_FTR_ASYM_SMTis set at compile time for only POWER7 where running the tasks on lower numbered threads give us the benefit of SMT thread folding. Interleaved big core is a feature introduced only on POWER9. Thus, we know thatCPU_FTR_ASYM_SMTis not set incpu_featuresat this point.
Since we're setting CPU_FTR_ASYM_SMT, it doesn't make sense to use cpu_has_feature(CPU_FTR_ASYM_SMT). I thought cpu_has_feature() held all available features (not necessarily enabled) that we could check before setting or enabling such feature. I think I misread it. Sorry.
Quoted 18 lines
Quoted 13 lines
Quoted 10 lines
arch/powerpc/kernel/setup-common.c +2
-0+ +/* Initialize CPU <=> thread mapping/ * * WARNING: We assume that the number of threads is the same for * every CPU in the system. If that is not the case, then some code Signature · 1 line
1.9.4Signature · 1 line
Murilo--
Thanks and Regards
gautham.
Signature · 1 line
Murilo