Commit 9498b080 authored by Shilpasri G Bhat's avatar Shilpasri G Bhat Committed by Tim Gardner

cpufreq: powernv: Remove cpu_to_chip_id() from hot-path

BugLink: http://bugs.launchpad.net/bugs/1555765

cpu_to_chip_id() does a DT walk through to find out the chip id by
taking a contended device tree lock. This adds an unnecessary overhead
in a hot path. So instead of calling cpu_to_chip_id() everytime cache
the chip ids for all cores in the array 'core_to_chip_map' and use it
in the hotpath.
Reported-by: default avatarAnton Blanchard <anton@samba.org>
Signed-off-by: default avatarShilpasri G Bhat <shilpa.bhat@linux.vnet.ibm.com>
Reviewed-by: default avatarGautham R. Shenoy <ego@linux.vnet.ibm.com>
Acked-by: default avatarViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: default avatarRafael J. Wysocki <rafael.j.wysocki@intel.com>
(cherry picked from linux-next commit 96c4726f)
Signed-off-by: default avatarTim Gardner <tim.gardner@canonical.com>
parent 4a8dbc12
...@@ -43,6 +43,7 @@ ...@@ -43,6 +43,7 @@
static struct cpufreq_frequency_table powernv_freqs[POWERNV_MAX_PSTATES+1]; static struct cpufreq_frequency_table powernv_freqs[POWERNV_MAX_PSTATES+1];
static bool rebooting, throttled, occ_reset; static bool rebooting, throttled, occ_reset;
static unsigned int *core_to_chip_map;
static struct chip { static struct chip {
unsigned int id; unsigned int id;
...@@ -313,13 +314,14 @@ static inline unsigned int get_nominal_index(void) ...@@ -313,13 +314,14 @@ static inline unsigned int get_nominal_index(void)
static void powernv_cpufreq_throttle_check(void *data) static void powernv_cpufreq_throttle_check(void *data)
{ {
unsigned int cpu = smp_processor_id(); unsigned int cpu = smp_processor_id();
unsigned int chip_id = core_to_chip_map[cpu_core_index_of_thread(cpu)];
unsigned long pmsr; unsigned long pmsr;
int pmsr_pmax, i; int pmsr_pmax, i;
pmsr = get_pmspr(SPRN_PMSR); pmsr = get_pmspr(SPRN_PMSR);
for (i = 0; i < nr_chips; i++) for (i = 0; i < nr_chips; i++)
if (chips[i].id == cpu_to_chip_id(cpu)) if (chips[i].id == chip_id)
break; break;
/* Check for Pmax Capping */ /* Check for Pmax Capping */
...@@ -559,19 +561,29 @@ static int init_chip_info(void) ...@@ -559,19 +561,29 @@ static int init_chip_info(void)
unsigned int chip[256]; unsigned int chip[256];
unsigned int cpu, i; unsigned int cpu, i;
unsigned int prev_chip_id = UINT_MAX; unsigned int prev_chip_id = UINT_MAX;
cpumask_t cpu_mask;
int ret = -ENOMEM;
for_each_possible_cpu(cpu) { core_to_chip_map = kcalloc(cpu_nr_cores(), sizeof(unsigned int),
GFP_KERNEL);
if (!core_to_chip_map)
goto out;
cpumask_copy(&cpu_mask, cpu_possible_mask);
for_each_cpu(cpu, &cpu_mask) {
unsigned int id = cpu_to_chip_id(cpu); unsigned int id = cpu_to_chip_id(cpu);
if (prev_chip_id != id) { if (prev_chip_id != id) {
prev_chip_id = id; prev_chip_id = id;
chip[nr_chips++] = id; chip[nr_chips++] = id;
} }
core_to_chip_map[cpu_core_index_of_thread(cpu)] = id;
cpumask_andnot(&cpu_mask, &cpu_mask, cpu_sibling_mask(cpu));
} }
chips = kmalloc_array(nr_chips, sizeof(struct chip), GFP_KERNEL); chips = kmalloc_array(nr_chips, sizeof(struct chip), GFP_KERNEL);
if (!chips) if (!chips)
return -ENOMEM; goto free_chip_map;
for (i = 0; i < nr_chips; i++) { for (i = 0; i < nr_chips; i++) {
chips[i].id = chip[i]; chips[i].id = chip[i];
...@@ -582,6 +594,10 @@ static int init_chip_info(void) ...@@ -582,6 +594,10 @@ static int init_chip_info(void)
} }
return 0; return 0;
free_chip_map:
kfree(core_to_chip_map);
out:
return ret;
} }
static int __init powernv_cpufreq_init(void) static int __init powernv_cpufreq_init(void)
...@@ -616,6 +632,7 @@ static void __exit powernv_cpufreq_exit(void) ...@@ -616,6 +632,7 @@ static void __exit powernv_cpufreq_exit(void)
opal_message_notifier_unregister(OPAL_MSG_OCC, opal_message_notifier_unregister(OPAL_MSG_OCC,
&powernv_cpufreq_opal_nb); &powernv_cpufreq_opal_nb);
kfree(chips); kfree(chips);
kfree(core_to_chip_map);
cpufreq_unregister_driver(&powernv_cpufreq_driver); cpufreq_unregister_driver(&powernv_cpufreq_driver);
} }
module_exit(powernv_cpufreq_exit); module_exit(powernv_cpufreq_exit);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment