irqchip/gic-v3-its: Balance LPI affinity across CPUs
Subsystems (2)
Marc Zyngier
maintainerv4 cover irqchip/gic-v3-its: Balance LPI affinity across CPUs
When mapping a LPI, the ITS driver picks the first possible affinity, which is in most cases CPU0, assuming that if that's not suitable, someone will come and set the affinity to something more interesting.
It apparently isn't the case, and people complain of poor performance when many interrupts are glued to the same CPU. So let's place the interrupts by finding the "least loaded" CPU (that is, the one that has the fewer LPIs mapped to it). So called 'managed' interrupts are an interesting case where the affinity is actually dictated by the kernel itself, and we should honor this.
* From v3:
- Always pre-decrement/post-increment affinity to avoid useless changes of affinity (John)
- Don't use the node mask as a superset of the proposed affinity as the ACPI tables can't really describe this (John)
- Rebased on v5.7-rc5
* From v2:
- Split accounting from CPU selection
- Track managed and unmanaged interrupts separately
Marc Zyngier (2):
irqchip/gic-v3-its: Track LPI distribution on a per CPU basis irqchip/gic-v3-its: Balance initial LPI affinity across CPUs
drivers/irqchip/irq-gic-v3-its.c | 170 ++++++++++++++++++++++++++----- 1 file changed, 143 insertions(+), 27 deletions(-)
Signature · 1 line
2.26.2
Marc Zyngier
maintainerPATCH v4 1/2 irqchip/gic-v3-its: Track LPI distribution on a per CPU basis
In order to improve the distribution of LPIs among CPUs, let start by tracking the number of LPIs assigned to CPUs, both for managed and non-managed interrupts (as separate counters).
--- drivers/irqchip/irq-gic-v3-its.c | 49 ++++++++++++++++++++++++++++++-- 1 file changed, 46 insertions(+), 3 deletions(-)
drivers/irqchip/irq-gic-v3-its.c +46 -3
@@ -174,6 +174,13 @@ static struct {int next_victim; } vpe_proxy; +struct cpu_lpi_count { +atomic_t managed; +atomic_t unmanaged; +}; + +static DEFINE_PER_CPU(struct cpu_lpi_count, cpu_lpi_count); +static LIST_HEAD (its_nodes);static DEFINE_RAW_SPINLOCK (its_lock);static struct rdists *gic_rdists;
@@ -1510,6 +1517,30 @@ static void its_unmask_irq(struct irq_data *d)lpi_update_config (d,0 ,LPI_PROP_ENABLED ); } +static __maybe_unused u32 its_read_lpi_count (struct irq_data *d,int cpu) +{ +if (irqd_affinity_is_managed (d)) +return atomic_read (&per_cpu_ptr (&cpu_lpi_count, cpu)->managed); + +return atomic_read (&per_cpu_ptr (&cpu_lpi_count, cpu)->unmanaged); +} + +static void its_inc_lpi_count (struct irq_data *d,int cpu) +{ +if (irqd_affinity_is_managed (d)) +atomic_inc (&per_cpu_ptr (&cpu_lpi_count, cpu)->managed); +else +atomic_inc (&per_cpu_ptr (&cpu_lpi_count, cpu)->unmanaged); +} + +static void its_dec_lpi_count (struct irq_data *d,int cpu) +{ +if (irqd_affinity_is_managed (d)) +atomic_dec (&per_cpu_ptr (&cpu_lpi_count, cpu)->managed); +else +atomic_dec (&per_cpu_ptr (&cpu_lpi_count, cpu)->unmanaged); +} +static int its_set_affinity (struct irq_data *d,const struct cpumask *mask_val,bool force) {
@@ -1518,34 +1549,44 @@ static int its_set_affinity(struct irq_data *d, const struct cpumask *mask_val,struct its_device *its_dev =irq_data_get_irq_chip_data (d);struct its_collection *target_col;u32 id =its_get_event_id (d); +int prev_cpu;/* A forwarded interrupt should use irq_set_vcpu_affinity */ if (irqd_is_forwarded_to_vcpu (d))return -EINVAL ; + prev_cpu = its_dev->event_map.col_map[id]; +its_dec_lpi_count (d, prev_cpu); +/* lpi cannot be routed to a redistributor that is on a foreign node */ if (its_dev->its->flags &ITS_FLAGS_WORKAROUND_CAVIUM_23144 ) {if (its_dev->its->numa_node >=0 ) { cpu_mask =cpumask_of_node (its_dev->its->numa_node);if (!cpumask_intersects (mask_val, cpu_mask)) -return -EINVAL ; + goto err; } } cpu =cpumask_any_and (mask_val, cpu_mask);if (cpu >= nr_cpu_ids) -return -EINVAL ; + goto err;/* don't set the affinity when the target cpu is same as current one */ -if (cpu != its_dev->event_map.col_map[id]) { +if (cpu != prev_cpu) { target_col = &its_dev->its->collections[cpu];its_send_movi (its_dev, target_col, id); its_dev->event_map.col_map[id] = cpu;irq_data_update_effective_affinity (d,cpumask_of (cpu)); } +its_inc_lpi_count (d, cpu); +return IRQ_SET_MASK_OK_DONE ; + +err: +its_inc_lpi_count (d, prev_cpu); +return -EINVAL ; }static u64 its_irq_get_msi_base (struct its_device *its_dev)
@@ -3448,6 +3489,7 @@ static int its_irq_domain_activate(struct irq_domain *domain, cpu =cpumask_first (cpu_online_mask); } +its_inc_lpi_count (d, cpu); its_dev->event_map.col_map[event] = cpu;irq_data_update_effective_affinity (d,cpumask_of (cpu));
@@ -3462,6 +3504,7 @@ static void its_irq_domain_deactivate(struct irq_domain *domain,struct its_device *its_dev =irq_data_get_irq_chip_data (d);u32 event =its_get_event_id (d); +its_dec_lpi_count (d, its_dev->event_map.col_map[event]);/* Stop the delivery of interrupts */ its_send_discard (its_dev, event); }
Signature · 1 line
2.26.2
[tip: irq/core] irqchip/gic-v3-its: Track LPI distribution on a per CPU basis
The following commit has been merged into the irq/core branch of tip:
The following commit has been merged into the irq/core branch of tip:
Commit-ID: 2f13ff1d1d5c0257c97ea76b86a2d9c99c44a4b9 Gitweb: https://git.kernel.org/tip/2f13ff1d1d5c0257c97ea76b86a2d9c99c44a4b9 Author: Marc Zyngier <[email protected]> AuthorDate: Fri, 15 May 2020 17:57:51 +01:00 Committer: Marc Zyngier <[email protected]> CommitterDate: Mon, 18 May 2020 10:30:42 +01:00
irqchip/gic-v3-its: Track LPI distribution on a per CPU basis
In order to improve the distribution of LPIs among CPUs, let start by tracking the number of LPIs assigned to CPUs, both for managed and non-managed interrupts (as separate counters).
--- drivers/irqchip/irq-gic-v3-its.c | 49 +++++++++++++++++++++++++++++-- 1 file changed, 46 insertions(+), 3 deletions(-)
drivers/irqchip/irq-gic-v3-its.c +46 -3
@@ -174,6 +174,13 @@ static struct {int next_victim; } vpe_proxy; +struct cpu_lpi_count { +atomic_t managed; +atomic_t unmanaged; +}; + +static DEFINE_PER_CPU(struct cpu_lpi_count, cpu_lpi_count); +static LIST_HEAD (its_nodes);static DEFINE_RAW_SPINLOCK (its_lock);static struct rdists *gic_rdists;
@@ -1510,6 +1517,30 @@ static void its_unmask_irq(struct irq_data *d)lpi_update_config (d,0 ,LPI_PROP_ENABLED ); } +static __maybe_unused u32 its_read_lpi_count (struct irq_data *d,int cpu) +{ +if (irqd_affinity_is_managed (d)) +return atomic_read (&per_cpu_ptr (&cpu_lpi_count, cpu)->managed); + +return atomic_read (&per_cpu_ptr (&cpu_lpi_count, cpu)->unmanaged); +} + +static void its_inc_lpi_count (struct irq_data *d,int cpu) +{ +if (irqd_affinity_is_managed (d)) +atomic_inc (&per_cpu_ptr (&cpu_lpi_count, cpu)->managed); +else +atomic_inc (&per_cpu_ptr (&cpu_lpi_count, cpu)->unmanaged); +} + +static void its_dec_lpi_count (struct irq_data *d,int cpu) +{ +if (irqd_affinity_is_managed (d)) +atomic_dec (&per_cpu_ptr (&cpu_lpi_count, cpu)->managed); +else +atomic_dec (&per_cpu_ptr (&cpu_lpi_count, cpu)->unmanaged); +} +static int its_set_affinity (struct irq_data *d,const struct cpumask *mask_val,bool force) {
@@ -1518,34 +1549,44 @@ static int its_set_affinity(struct irq_data *d, const struct cpumask *mask_val,struct its_device *its_dev =irq_data_get_irq_chip_data (d);struct its_collection *target_col;u32 id =its_get_event_id (d); +int prev_cpu;/* A forwarded interrupt should use irq_set_vcpu_affinity */ if (irqd_is_forwarded_to_vcpu (d))return -EINVAL ; + prev_cpu = its_dev->event_map.col_map[id]; +its_dec_lpi_count (d, prev_cpu); +/* lpi cannot be routed to a redistributor that is on a foreign node */ if (its_dev->its->flags &ITS_FLAGS_WORKAROUND_CAVIUM_23144 ) {if (its_dev->its->numa_node >=0 ) { cpu_mask =cpumask_of_node (its_dev->its->numa_node);if (!cpumask_intersects (mask_val, cpu_mask)) -return -EINVAL ; + goto err; } } cpu =cpumask_any_and (mask_val, cpu_mask);if (cpu >= nr_cpu_ids) -return -EINVAL ; + goto err;/* don't set the affinity when the target cpu is same as current one */ -if (cpu != its_dev->event_map.col_map[id]) { +if (cpu != prev_cpu) { target_col = &its_dev->its->collections[cpu];its_send_movi (its_dev, target_col, id); its_dev->event_map.col_map[id] = cpu;irq_data_update_effective_affinity (d,cpumask_of (cpu)); } +its_inc_lpi_count (d, cpu); +return IRQ_SET_MASK_OK_DONE ; + +err: +its_inc_lpi_count (d, prev_cpu); +return -EINVAL ; }static u64 its_irq_get_msi_base (struct its_device *its_dev)
@@ -3448,6 +3489,7 @@ static int its_irq_domain_activate(struct irq_domain *domain, cpu =cpumask_first (cpu_online_mask); } +its_inc_lpi_count (d, cpu); its_dev->event_map.col_map[event] = cpu;irq_data_update_effective_affinity (d,cpumask_of (cpu));
@@ -3462,6 +3504,7 @@ static void its_irq_domain_deactivate(struct irq_domain *domain,struct its_device *its_dev =irq_data_get_irq_chip_data (d);u32 event =its_get_event_id (d); +its_dec_lpi_count (d, its_dev->event_map.col_map[event]);/* Stop the delivery of interrupts */ its_send_discard (its_dev, event); }
Marc Zyngier
maintainerPATCH v4 2/2 irqchip/gic-v3-its: Balance initial LPI affinity across CPUs
In reply to Marc ZyngierWhen mapping a LPI, the ITS driver picks the first possible affinity, which is in most cases CPU0, assuming that if that's not suitable, someone will come and set the affinity to something more interesting.
It apparently isn't the case, and people complain of poor performance when many interrupts are glued to the same CPU. So let's place the interrupts by finding the "least loaded" CPU (that is, the one that has the fewer LPIs mapped to it). So called 'managed' interrupts are an interesting case where the affinity is actually dictated by the kernel itself, and we should honor this.
--- drivers/irqchip/irq-gic-v3-its.c | 127 ++++++++++++++++++++++++------- 1 file changed, 100 insertions(+), 27 deletions(-)
drivers/irqchip/irq-gic-v3-its.c +100 -27
@@ -1541,15 +1541,104 @@ static void its_dec_lpi_count(struct irq_data *d, int cpu)atomic_dec (&per_cpu_ptr (&cpu_lpi_count, cpu)->unmanaged); } +static unsigned int cpumask_pick_least_loaded (struct irq_data *d, +const struct cpumask *cpu_mask) +{ +unsigned int cpu = nr_cpu_ids, tmp; +int count =S32_MAX ; + +for_each_cpu (tmp, cpu_mask) { +int this_count =its_read_lpi_count (d, tmp); +if (this_count < count) { + cpu = tmp; + count = this_count; + } + } + +return cpu; +} + +/* +* As suggested by Thomas Gleixner in: +* https://lore.kernel.org/r/[email protected] +*/ +static int its_select_cpu (struct irq_data *d, +const struct cpumask *aff_mask) +{ +struct its_device *its_dev =irq_data_get_irq_chip_data (d); +cpumask_var_t tmpmask; +int cpu, node; + +if (!alloc_cpumask_var (&tmpmask,GFP_KERNEL )) +return -ENOMEM ; + + node = its_dev->its->numa_node; + +if (!irqd_affinity_is_managed (d)) { +/* First try the NUMA node */ +if (node !=NUMA_NO_NODE ) { +/* +* Try the intersection of the affinity mask and the +* node mask (and the online mask, just to be safe). +*/ +cpumask_and (tmpmask,cpumask_of_node (node), aff_mask); +cpumask_and (tmpmask, tmpmask, cpu_online_mask); + +/* +* Ideally, we would check if the mask is empty, and +* try again on the full node here. +* +* But it turns out that the way API describes the +* affinity for ITSs only deals about memory, and +* not target CPUs, so it cannot describe a single +* ITS placed next to two NUMA nodes. +* +* Instead, just fallback on the online mask. This +* diverges from Thomas' suggestion above. +*/ + cpu =cpumask_pick_least_loaded (d, tmpmask); +if (cpu < nr_cpu_ids) + goto out; + +/* If we can't cross sockets, give up */ +if ((its_dev->its->flags &ITS_FLAGS_WORKAROUND_CAVIUM_23144 )) + goto out; + +/* If the above failed, expand the search */ + } + +/* Try the intersection of the affinity and online masks */ +cpumask_and (tmpmask, aff_mask, cpu_online_mask); + +/* If that doesn't fly, the online mask is the last resort */ +if (cpumask_empty (tmpmask)) +cpumask_copy (tmpmask, cpu_online_mask); + + cpu =cpumask_pick_least_loaded (d, tmpmask); + }else { +cpumask_and (tmpmask,irq_data_get_affinity_mask (d), cpu_online_mask); + +/* If we cannot cross sockets, limit the search to that node */ +if ((its_dev->its->flags &ITS_FLAGS_WORKAROUND_CAVIUM_23144 ) && + node !=NUMA_NO_NODE ) +cpumask_and (tmpmask, tmpmask,cpumask_of_node (node)); + + cpu =cpumask_pick_least_loaded (d, tmpmask); + } +out: +free_cpumask_var (tmpmask); + +pr_debug ("IRQ%d -> %*pbl CPU%d\n" , d->irq,cpumask_pr_args (aff_mask), cpu); +return cpu; +} +static int its_set_affinity (struct irq_data *d,const struct cpumask *mask_val,bool force) { -unsigned int cpu; -const struct cpumask *cpu_mask = cpu_online_mask;struct its_device *its_dev =irq_data_get_irq_chip_data (d);struct its_collection *target_col;u32 id =its_get_event_id (d); -int prev_cpu; +int cpu, prev_cpu;/* A forwarded interrupt should use irq_set_vcpu_affinity */ if (irqd_is_forwarded_to_vcpu (d))
@@ -1558,18 +1647,12 @@ static int its_set_affinity(struct irq_data *d, const struct cpumask *mask_val, prev_cpu = its_dev->event_map.col_map[id];its_dec_lpi_count (d, prev_cpu); -/* lpi cannot be routed to a redistributor that is on a foreign node */ -if (its_dev->its->flags &ITS_FLAGS_WORKAROUND_CAVIUM_23144 ) { -if (its_dev->its->numa_node >=0 ) { - cpu_mask =cpumask_of_node (its_dev->its->numa_node); -if (!cpumask_intersects (mask_val, cpu_mask)) - goto err; - } - } - - cpu =cpumask_any_and (mask_val, cpu_mask); +if (!force) + cpu =its_select_cpu (d, mask_val); +else + cpu =cpumask_pick_least_loaded (d, mask_val); -if (cpu >= nr_cpu_ids) +if (cpu <0 || cpu >= nr_cpu_ids) goto err;/* don't set the affinity when the target cpu is same as current one */
@@ -3473,21 +3556,11 @@ static int its_irq_domain_activate(struct irq_domain *domain, {struct its_device *its_dev =irq_data_get_irq_chip_data (d);u32 event =its_get_event_id (d); -const struct cpumask *cpu_mask = cpu_online_mask;int cpu; -/* get the cpu_mask of local node */ -if (its_dev->its->numa_node >=0 ) - cpu_mask =cpumask_of_node (its_dev->its->numa_node); - -/* Bind the LPI to the first possible CPU */ - cpu =cpumask_first_and (cpu_mask, cpu_online_mask); -if (cpu >= nr_cpu_ids) { -if (its_dev->its->flags &ITS_FLAGS_WORKAROUND_CAVIUM_23144 ) -return -EINVAL ; - - cpu =cpumask_first (cpu_online_mask); - } + cpu =its_select_cpu (d, cpu_online_mask); +if (cpu <0 || cpu >= nr_cpu_ids) +return -EINVAL ;its_inc_lpi_count (d, cpu); its_dev->event_map.col_map[event] = cpu;
Signature · 1 line
2.26.2
Quoted 5 lines · Marc Zyngier
When mapping a LPI, the ITS driver picks the first possible affinity, which is in most cases CPU0, assuming that if that's not suitable, someone will come and set the affinity to something more interesting.
Quoted 8 lines
It apparently isn't the case, and people complain of poor performance when many interrupts are glued to the same CPU. So let's place the interrupts by finding the "least loaded" CPU (that is, the one that has the fewer LPIs mapped to it). So called 'managed' interrupts are an interesting case where the affinity is actually dictated by the kernel itself, and we should honor this.
Cheers Marc, I ran these again and the figures look good:
NVMe with nvme.use_threaded_interrupts=1/0
Before: 1000K/1100K IOPS
After: 1100K/1400K IOPS
For some reason v5.7-rc5 looks more stable than rc4 (which I tested previously).
Quoted 18 lines
* From v3:
- Always pre-decrement/post-increment affinity to avoid useless changes of affinity (John)
- Don't use the node mask as a superset of the proposed affinity as the ACPI tables can't really describe this (John)
- Rebased on v5.7-rc5* From v2:
- Split accounting from CPU selection
- Track managed and unmanaged interrupts separatelyMarc Zyngier (2):
irqchip/gic-v3-its: Track LPI distribution on a per CPU basis irqchip/gic-v3-its: Balance initial LPI affinity across CPUsdrivers/irqchip/irq-gic-v3-its.c | 170 ++++++++++++++++++++++++++----- 1 file changed, 143 insertions(+), 27 deletions(-)