From 8d565748b6035eeda18895c213396a4c9fac6a4c Mon Sep 17 00:00:00 2001 From: Zenghui Yu Date: Sun, 10 Feb 2019 05:24:10 +0000 Subject: irqchip/gic-v3-its: Avoid parsing _indirect_ twice for Device table In current logic, its_parse_indirect_baser() will be invoked twice when allocating Device tables. Add a *break* to omit the unnecessary and annoying (might be ...) invoking. Fixes: 32bd44dc19de ("irqchip/gic-v3-its: Fix the incorrect parsing of VCPU table size") Cc: stable@vger.kernel.org Signed-off-by: Zenghui Yu Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-gic-v3-its.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c index db20e992a40f..9f529a6aeaae 100644 --- a/drivers/irqchip/irq-gic-v3-its.c +++ b/drivers/irqchip/irq-gic-v3-its.c @@ -1946,6 +1946,8 @@ static int its_alloc_tables(struct its_node *its) indirect = its_parse_indirect_baser(its, baser, psz, &order, its->device_ids); + break; + case GITS_BASER_TYPE_VCPU: indirect = its_parse_indirect_baser(its, baser, psz, &order, -- cgit v1.2.3 From 9e543e22e204722357fe43c4769bb22329e65381 Mon Sep 17 00:00:00 2001 From: Jiaxun Yang Date: Fri, 1 Feb 2019 14:22:35 +0800 Subject: irqchip: Add driver for Loongson-1 interrupt controller This controller appeared on Loongson-1 family MCUs including Loongson-1B and Loongson-1C. Signed-off-by: Jiaxun Yang Signed-off-by: Marc Zyngier --- drivers/irqchip/Kconfig | 9 +++ drivers/irqchip/Makefile | 1 + drivers/irqchip/irq-ls1x.c | 192 +++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 202 insertions(+) create mode 100644 drivers/irqchip/irq-ls1x.c (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/Kconfig b/drivers/irqchip/Kconfig index 3d1e60779078..5dcb5456cd14 100644 --- a/drivers/irqchip/Kconfig +++ b/drivers/irqchip/Kconfig @@ -406,6 +406,15 @@ config IMX_IRQSTEER help Support for the i.MX IRQSTEER interrupt multiplexer/remapper. +config LS1X_IRQ + bool "Loongson-1 Interrupt Controller" + depends on MACH_LOONGSON32 + default y + select IRQ_DOMAIN + select GENERIC_IRQ_CHIP + help + Support for the Loongson-1 platform Interrupt Controller. + endmenu config SIFIVE_PLIC diff --git a/drivers/irqchip/Makefile b/drivers/irqchip/Makefile index c93713d24b86..7acd0e36d0b4 100644 --- a/drivers/irqchip/Makefile +++ b/drivers/irqchip/Makefile @@ -94,3 +94,4 @@ obj-$(CONFIG_CSKY_APB_INTC) += irq-csky-apb-intc.o obj-$(CONFIG_SIFIVE_PLIC) += irq-sifive-plic.o obj-$(CONFIG_IMX_IRQSTEER) += irq-imx-irqsteer.o obj-$(CONFIG_MADERA_IRQ) += irq-madera.o +obj-$(CONFIG_LS1X_IRQ) += irq-ls1x.o diff --git a/drivers/irqchip/irq-ls1x.c b/drivers/irqchip/irq-ls1x.c new file mode 100644 index 000000000000..86b72fbd3b45 --- /dev/null +++ b/drivers/irqchip/irq-ls1x.c @@ -0,0 +1,192 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2019, Jiaxun Yang + * Loongson-1 platform IRQ support + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#define LS_REG_INTC_STATUS 0x00 +#define LS_REG_INTC_EN 0x04 +#define LS_REG_INTC_SET 0x08 +#define LS_REG_INTC_CLR 0x0c +#define LS_REG_INTC_POL 0x10 +#define LS_REG_INTC_EDGE 0x14 + +/** + * struct ls1x_intc_priv - private ls1x-intc data. + * @domain: IRQ domain. + * @intc_base: IO Base of intc registers. + */ + +struct ls1x_intc_priv { + struct irq_domain *domain; + void __iomem *intc_base; +}; + + +static void ls1x_chained_handle_irq(struct irq_desc *desc) +{ + struct ls1x_intc_priv *priv = irq_desc_get_handler_data(desc); + struct irq_chip *chip = irq_desc_get_chip(desc); + u32 pending; + + chained_irq_enter(chip, desc); + pending = readl(priv->intc_base + LS_REG_INTC_STATUS) & + readl(priv->intc_base + LS_REG_INTC_EN); + + if (!pending) + spurious_interrupt(); + + while (pending) { + int bit = __ffs(pending); + + generic_handle_irq(irq_find_mapping(priv->domain, bit)); + pending &= ~BIT(bit); + } + + chained_irq_exit(chip, desc); +} + +static void ls_intc_set_bit(struct irq_chip_generic *gc, + unsigned int offset, + u32 mask, bool set) +{ + if (set) + writel(readl(gc->reg_base + offset) | mask, + gc->reg_base + offset); + else + writel(readl(gc->reg_base + offset) & ~mask, + gc->reg_base + offset); +} + +static int ls_intc_set_type(struct irq_data *data, unsigned int type) +{ + struct irq_chip_generic *gc = irq_data_get_irq_chip_data(data); + u32 mask = data->mask; + + switch (type) { + case IRQ_TYPE_LEVEL_HIGH: + ls_intc_set_bit(gc, LS_REG_INTC_EDGE, mask, false); + ls_intc_set_bit(gc, LS_REG_INTC_POL, mask, true); + break; + case IRQ_TYPE_LEVEL_LOW: + ls_intc_set_bit(gc, LS_REG_INTC_EDGE, mask, false); + ls_intc_set_bit(gc, LS_REG_INTC_POL, mask, false); + break; + case IRQ_TYPE_EDGE_RISING: + ls_intc_set_bit(gc, LS_REG_INTC_EDGE, mask, true); + ls_intc_set_bit(gc, LS_REG_INTC_POL, mask, true); + break; + case IRQ_TYPE_EDGE_FALLING: + ls_intc_set_bit(gc, LS_REG_INTC_EDGE, mask, true); + ls_intc_set_bit(gc, LS_REG_INTC_POL, mask, false); + break; + default: + return -EINVAL; + } + + irqd_set_trigger_type(data, type); + return irq_setup_alt_chip(data, type); +} + + +static int __init ls1x_intc_of_init(struct device_node *node, + struct device_node *parent) +{ + struct irq_chip_generic *gc; + struct irq_chip_type *ct; + struct ls1x_intc_priv *priv; + int parent_irq, err = 0; + + priv = kzalloc(sizeof(*priv), GFP_KERNEL); + if (!priv) + return -ENOMEM; + + priv->intc_base = of_iomap(node, 0); + if (!priv->intc_base) { + err = -ENODEV; + goto out_free_priv; + } + + parent_irq = irq_of_parse_and_map(node, 0); + if (!parent_irq) { + pr_err("ls1x-irq: unable to get parent irq\n"); + err = -ENODEV; + goto out_iounmap; + } + + /* Set up an IRQ domain */ + priv->domain = irq_domain_add_linear(node, 32, &irq_generic_chip_ops, + NULL); + if (!priv->domain) { + pr_err("ls1x-irq: cannot add IRQ domain\n"); + goto out_iounmap; + } + + err = irq_alloc_domain_generic_chips(priv->domain, 32, 2, + node->full_name, handle_level_irq, + IRQ_NOREQUEST | IRQ_NOPROBE | IRQ_NOAUTOEN, 0, + IRQ_GC_INIT_MASK_CACHE); + if (err) { + pr_err("ls1x-irq: unable to register IRQ domain\n"); + goto out_free_domain; + } + + /* Mask all irqs */ + writel(0x0, priv->intc_base + LS_REG_INTC_EN); + + /* Ack all irqs */ + writel(0xffffffff, priv->intc_base + LS_REG_INTC_CLR); + + /* Set all irqs to high level triggered */ + writel(0xffffffff, priv->intc_base + LS_REG_INTC_POL); + + gc = irq_get_domain_generic_chip(priv->domain, 0); + + gc->reg_base = priv->intc_base; + + ct = gc->chip_types; + ct[0].type = IRQ_TYPE_LEVEL_MASK; + ct[0].regs.mask = LS_REG_INTC_EN; + ct[0].regs.ack = LS_REG_INTC_CLR; + ct[0].chip.irq_unmask = irq_gc_mask_set_bit; + ct[0].chip.irq_mask = irq_gc_mask_clr_bit; + ct[0].chip.irq_ack = irq_gc_ack_set_bit; + ct[0].chip.irq_set_type = ls_intc_set_type; + ct[0].handler = handle_level_irq; + + ct[1].type = IRQ_TYPE_EDGE_BOTH; + ct[1].regs.mask = LS_REG_INTC_EN; + ct[1].regs.ack = LS_REG_INTC_CLR; + ct[1].chip.irq_unmask = irq_gc_mask_set_bit; + ct[1].chip.irq_mask = irq_gc_mask_clr_bit; + ct[1].chip.irq_ack = irq_gc_ack_set_bit; + ct[1].chip.irq_set_type = ls_intc_set_type; + ct[1].handler = handle_edge_irq; + + irq_set_chained_handler_and_data(parent_irq, + ls1x_chained_handle_irq, priv); + + return 0; + +out_free_domain: + irq_domain_remove(priv->domain); +out_iounmap: + iounmap(priv->intc_base); +out_free_priv: + kfree(priv); + + return err; +} + +IRQCHIP_DECLARE(ls1x_intc, "loongson,ls1x-intc", ls1x_intc_of_init); -- cgit v1.2.3 From 518bfe84ec417318b2470652cdb27978ddfeaa59 Mon Sep 17 00:00:00 2001 From: Aaro Koskinen Date: Wed, 6 Feb 2019 23:26:08 +0200 Subject: irqchip/i8259: Fix shutdown order by moving syscore_ops registration When using cpufreq on Loongson 2F MIPS platform, "poweroff" command gets frequently stuck in syscore_shutdown(). The reason is that i8259A_shutdown() gets called before cpufreq_suspend(), and if we have pending work then irq_work_sync() in cpufreq_dbs_governor_stop() gets stuck forever as we have all interrupts masked already. irq-i8259 is registering syscore_ops using device_initcall(), while cpufreq uses core_initcall(). Fix the shutdown order simply by registering the irq syscore_ops during the early IRQ init instead of using a separate initcall at later stage. Signed-off-by: Aaro Koskinen Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-i8259.c | 9 +-------- 1 file changed, 1 insertion(+), 8 deletions(-) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-i8259.c b/drivers/irqchip/irq-i8259.c index b0d4aab1a58c..d000870d9b6b 100644 --- a/drivers/irqchip/irq-i8259.c +++ b/drivers/irqchip/irq-i8259.c @@ -225,14 +225,6 @@ static struct syscore_ops i8259_syscore_ops = { .shutdown = i8259A_shutdown, }; -static int __init i8259A_init_sysfs(void) -{ - register_syscore_ops(&i8259_syscore_ops); - return 0; -} - -device_initcall(i8259A_init_sysfs); - static void init_8259A(int auto_eoi) { unsigned long flags; @@ -332,6 +324,7 @@ struct irq_domain * __init __init_i8259_irqs(struct device_node *node) panic("Failed to add i8259 IRQ domain"); setup_irq(I8259A_IRQ_BASE + PIC_CASCADE_IR, &irq2); + register_syscore_ops(&i8259_syscore_ops); return domain; } -- cgit v1.2.3 From fc03acaeab358c008a194b78daa10e78401376a8 Mon Sep 17 00:00:00 2001 From: Atish Patra Date: Tue, 12 Feb 2019 03:10:11 -0800 Subject: irqchip/irq-sifive-plic: Check and continue in case of an invalid cpuid. riscv_hartid_to_cpuid can return invalid cpuid for a hart that is present in DT but was never brought up. Print the appropriate warning message and continue. Signed-off-by: Atish Patra Reviewed-by: Anup Patel Reviewed-by: Christoph Hellwig Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-sifive-plic.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c index 357e9daf94ae..254ecd76e8be 100644 --- a/drivers/irqchip/irq-sifive-plic.c +++ b/drivers/irqchip/irq-sifive-plic.c @@ -237,6 +237,11 @@ static int __init plic_init(struct device_node *node, } cpu = riscv_hartid_to_cpuid(hartid); + if (cpu < 0) { + pr_warn("Invalid cpuid for context %d\n", i); + continue; + } + handler = per_cpu_ptr(&plic_handlers, cpu); handler->present = true; handler->ctxid = i; -- cgit v1.2.3 From 86c7cbf1e8d1d4f4f60e229fdc2a5b21c09c29a3 Mon Sep 17 00:00:00 2001 From: Anup Patel Date: Tue, 12 Feb 2019 18:22:43 +0530 Subject: irqchip/sifive-plic: Pre-compute context hart base and enable base This patch does following optimizations: 1. Pre-compute hart base for each context handler 2. Pre-compute enable base for each context handler 3. Have enable lock for each context handler instead of global plic_toggle_lock Signed-off-by: Anup Patel Reviewed-by: Christoph Hellwig Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-sifive-plic.c | 47 +++++++++++++++++---------------------- 1 file changed, 21 insertions(+), 26 deletions(-) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c index 254ecd76e8be..715ef7b3f42d 100644 --- a/drivers/irqchip/irq-sifive-plic.c +++ b/drivers/irqchip/irq-sifive-plic.c @@ -59,37 +59,28 @@ static void __iomem *plic_regs; struct plic_handler { bool present; - int ctxid; + void __iomem *hart_base; + /* + * Protect mask operations on the registers given that we can't + * assume atomic memory operations work on them. + */ + raw_spinlock_t enable_lock; + void __iomem *enable_base; }; static DEFINE_PER_CPU(struct plic_handler, plic_handlers); -static inline void __iomem *plic_hart_offset(int ctxid) -{ - return plic_regs + CONTEXT_BASE + ctxid * CONTEXT_PER_HART; -} - -static inline u32 __iomem *plic_enable_base(int ctxid) -{ - return plic_regs + ENABLE_BASE + ctxid * ENABLE_PER_HART; -} - -/* - * Protect mask operations on the registers given that we can't assume that - * atomic memory operations work on them. - */ -static DEFINE_RAW_SPINLOCK(plic_toggle_lock); - -static inline void plic_toggle(int ctxid, int hwirq, int enable) +static inline void plic_toggle(struct plic_handler *handler, + int hwirq, int enable) { - u32 __iomem *reg = plic_enable_base(ctxid) + (hwirq / 32); + u32 __iomem *reg = handler->enable_base + (hwirq / 32) * sizeof(u32); u32 hwirq_mask = 1 << (hwirq % 32); - raw_spin_lock(&plic_toggle_lock); + raw_spin_lock(&handler->enable_lock); if (enable) writel(readl(reg) | hwirq_mask, reg); else writel(readl(reg) & ~hwirq_mask, reg); - raw_spin_unlock(&plic_toggle_lock); + raw_spin_unlock(&handler->enable_lock); } static inline void plic_irq_toggle(struct irq_data *d, int enable) @@ -101,7 +92,7 @@ static inline void plic_irq_toggle(struct irq_data *d, int enable) struct plic_handler *handler = per_cpu_ptr(&plic_handlers, cpu); if (handler->present) - plic_toggle(handler->ctxid, d->hwirq, enable); + plic_toggle(handler, d->hwirq, enable); } } @@ -150,7 +141,7 @@ static struct irq_domain *plic_irqdomain; static void plic_handle_irq(struct pt_regs *regs) { struct plic_handler *handler = this_cpu_ptr(&plic_handlers); - void __iomem *claim = plic_hart_offset(handler->ctxid) + CONTEXT_CLAIM; + void __iomem *claim = handler->hart_base + CONTEXT_CLAIM; irq_hw_number_t hwirq; WARN_ON_ONCE(!handler->present); @@ -244,12 +235,16 @@ static int __init plic_init(struct device_node *node, handler = per_cpu_ptr(&plic_handlers, cpu); handler->present = true; - handler->ctxid = i; + handler->hart_base = + plic_regs + CONTEXT_BASE + i * CONTEXT_PER_HART; + raw_spin_lock_init(&handler->enable_lock); + handler->enable_base = + plic_regs + ENABLE_BASE + i * ENABLE_PER_HART; /* priority must be > threshold to trigger an interrupt */ - writel(0, plic_hart_offset(i) + CONTEXT_THRESHOLD); + writel(0, handler->hart_base + CONTEXT_THRESHOLD); for (hwirq = 1; hwirq <= nr_irqs; hwirq++) - plic_toggle(i, hwirq, 0); + plic_toggle(handler, hwirq, 0); nr_mapped++; } -- cgit v1.2.3 From 3fecb5aac2888814884881317ef139437338d8d0 Mon Sep 17 00:00:00 2001 From: Anup Patel Date: Tue, 12 Feb 2019 18:22:44 +0530 Subject: irqchip/sifive-plic: Add warning in plic_init() if handler already present We have two enteries (one for M-mode and another for S-mode) in the interrupts-extended DT property of PLIC DT node for each HART. It is expected that firmware/bootloader will set M-mode HWIRQ line of each HART to 0xffffffff (i.e. -1) in interrupts-extended DT property because Linux runs in S-mode only. If firmware/bootloader is buggy then it will not correctly update interrupts-extended DT property which might result in a plic_handler configured twice. This patch adds a warning in plic_init() if a plic_handler is already marked present. This warning provides us a hint about incorrectly updated interrupts-extended DT property. Signed-off-by: Anup Patel Reviewed-by: Christoph Hellwig Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-sifive-plic.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c index 715ef7b3f42d..b8721b4f8b2f 100644 --- a/drivers/irqchip/irq-sifive-plic.c +++ b/drivers/irqchip/irq-sifive-plic.c @@ -234,6 +234,11 @@ static int __init plic_init(struct device_node *node, } handler = per_cpu_ptr(&plic_handlers, cpu); + if (handler->present) { + pr_warn("handler already present for context %d.\n", i); + continue; + } + handler->present = true; handler->hart_base = plic_regs + CONTEXT_BASE + i * CONTEXT_PER_HART; -- cgit v1.2.3 From 6adfe8d2f5b353529d5a3a7842b764afbcd122e1 Mon Sep 17 00:00:00 2001 From: Anup Patel Date: Tue, 12 Feb 2019 18:22:45 +0530 Subject: irqchip/sifive-plic: Differentiate between PLIC handler and context We explicitly differentiate between PLIC handler and context because PLIC context is for given mode of HART whereas PLIC handler is per-CPU software construct meant for handling interrupts from a particular PLIC context. To achieve this differentiation, we rename "nr_handlers" to "nr_contexts" and "nr_mapped" to "nr_handlers" in plic_init(). Signed-off-by: Anup Patel Reviewed-by: Christoph Hellwig Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-sifive-plic.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c index b8721b4f8b2f..c5cc7e137c08 100644 --- a/drivers/irqchip/irq-sifive-plic.c +++ b/drivers/irqchip/irq-sifive-plic.c @@ -177,7 +177,7 @@ static int plic_find_hart_id(struct device_node *node) static int __init plic_init(struct device_node *node, struct device_node *parent) { - int error = 0, nr_handlers, nr_mapped = 0, i; + int error = 0, nr_contexts, nr_handlers = 0, i; u32 nr_irqs; if (plic_regs) { @@ -194,10 +194,10 @@ static int __init plic_init(struct device_node *node, if (WARN_ON(!nr_irqs)) goto out_iounmap; - nr_handlers = of_irq_count(node); - if (WARN_ON(!nr_handlers)) + nr_contexts = of_irq_count(node); + if (WARN_ON(!nr_contexts)) goto out_iounmap; - if (WARN_ON(nr_handlers < num_possible_cpus())) + if (WARN_ON(nr_contexts < num_possible_cpus())) goto out_iounmap; error = -ENOMEM; @@ -206,7 +206,7 @@ static int __init plic_init(struct device_node *node, if (WARN_ON(!plic_irqdomain)) goto out_iounmap; - for (i = 0; i < nr_handlers; i++) { + for (i = 0; i < nr_contexts; i++) { struct of_phandle_args parent; struct plic_handler *handler; irq_hw_number_t hwirq; @@ -250,11 +250,11 @@ static int __init plic_init(struct device_node *node, writel(0, handler->hart_base + CONTEXT_THRESHOLD); for (hwirq = 1; hwirq <= nr_irqs; hwirq++) plic_toggle(handler, hwirq, 0); - nr_mapped++; + nr_handlers++; } - pr_info("mapped %d interrupts to %d (out of %d) handlers.\n", - nr_irqs, nr_mapped, nr_handlers); + pr_info("mapped %d interrupts with %d handlers for %d contexts.\n", + nr_irqs, nr_handlers, nr_contexts); set_handle_irq(plic_handle_irq); return 0; -- cgit v1.2.3 From cc9f04f9a84f745949e325661550ed14bd0ff322 Mon Sep 17 00:00:00 2001 From: Anup Patel Date: Tue, 12 Feb 2019 18:22:46 +0530 Subject: irqchip/sifive-plic: Implement irq_set_affinity() for SMP host Currently on SMP host, all CPUs take external interrupts routed via PLIC. All CPUs will try to claim a given external interrupt but only one of them will succeed while other CPUs would simply resume whatever they were doing before. This means if we have N CPUs then for every external interrupt N-1 CPUs will always fail to claim it and waste their CPU time. Instead of above, external interrupts should be taken by only one CPU and we should have provision to explicitly specify IRQ affinity from kernel-space or user-space. This patch provides irq_set_affinity() implementation for PLIC driver. It also updates irq_enable() such that PLIC interrupts are only enabled for one of CPUs specified in IRQ affinity mask. With this patch in-place, we can change IRQ affinity at any-time from user-space using procfs. Example: / # cat /proc/interrupts CPU0 CPU1 CPU2 CPU3 8: 44 0 0 0 SiFive PLIC 8 virtio0 10: 48 0 0 0 SiFive PLIC 10 ttyS0 IPI0: 55 663 58 363 Rescheduling interrupts IPI1: 0 1 3 16 Function call interrupts / # / # / # echo 4 > /proc/irq/10/smp_affinity / # / # cat /proc/interrupts CPU0 CPU1 CPU2 CPU3 8: 45 0 0 0 SiFive PLIC 8 virtio0 10: 160 0 17 0 SiFive PLIC 10 ttyS0 IPI0: 68 693 77 410 Rescheduling interrupts IPI1: 0 2 3 16 Function call interrupts Signed-off-by: Anup Patel Reviewed-by: Christoph Hellwig Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-sifive-plic.c | 45 +++++++++++++++++++++++++++++++++------ 1 file changed, 39 insertions(+), 6 deletions(-) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c index c5cc7e137c08..cf755964f2f8 100644 --- a/drivers/irqchip/irq-sifive-plic.c +++ b/drivers/irqchip/irq-sifive-plic.c @@ -83,29 +83,59 @@ static inline void plic_toggle(struct plic_handler *handler, raw_spin_unlock(&handler->enable_lock); } -static inline void plic_irq_toggle(struct irq_data *d, int enable) +static inline void plic_irq_toggle(const struct cpumask *mask, + int hwirq, int enable) { int cpu; - writel(enable, plic_regs + PRIORITY_BASE + d->hwirq * PRIORITY_PER_ID); - for_each_cpu(cpu, irq_data_get_affinity_mask(d)) { + writel(enable, plic_regs + PRIORITY_BASE + hwirq * PRIORITY_PER_ID); + for_each_cpu(cpu, mask) { struct plic_handler *handler = per_cpu_ptr(&plic_handlers, cpu); if (handler->present) - plic_toggle(handler, d->hwirq, enable); + plic_toggle(handler, hwirq, enable); } } static void plic_irq_enable(struct irq_data *d) { - plic_irq_toggle(d, 1); + unsigned int cpu = cpumask_any_and(irq_data_get_affinity_mask(d), + cpu_online_mask); + if (WARN_ON_ONCE(cpu >= nr_cpu_ids)) + return; + plic_irq_toggle(cpumask_of(cpu), d->hwirq, 1); } static void plic_irq_disable(struct irq_data *d) { - plic_irq_toggle(d, 0); + plic_irq_toggle(cpu_possible_mask, d->hwirq, 0); } +#ifdef CONFIG_SMP +static int plic_set_affinity(struct irq_data *d, + const struct cpumask *mask_val, bool force) +{ + unsigned int cpu; + + if (force) + cpu = cpumask_first(mask_val); + else + cpu = cpumask_any_and(mask_val, cpu_online_mask); + + if (cpu >= nr_cpu_ids) + return -EINVAL; + + if (!irqd_irq_disabled(d)) { + plic_irq_toggle(cpu_possible_mask, d->hwirq, 0); + plic_irq_toggle(cpumask_of(cpu), d->hwirq, 1); + } + + irq_data_update_effective_affinity(d, cpumask_of(cpu)); + + return IRQ_SET_MASK_OK_DONE; +} +#endif + static struct irq_chip plic_chip = { .name = "SiFive PLIC", /* @@ -114,6 +144,9 @@ static struct irq_chip plic_chip = { */ .irq_enable = plic_irq_enable, .irq_disable = plic_irq_disable, +#ifdef CONFIG_SMP + .irq_set_affinity = plic_set_affinity, +#endif }; static int plic_irqdomain_map(struct irq_domain *d, unsigned int irq, -- cgit v1.2.3 From 539d378242858c58f9e57b54e57be1f7f1204ad4 Mon Sep 17 00:00:00 2001 From: Shanker Donthineni Date: Mon, 14 Jan 2019 09:50:19 +0000 Subject: irqchip/gicv3-its: Use NUMA aware memory allocation for ITS tables The NUMA node information is visible to ITS driver but not being used other than handling hardware errata. ITS/GICR hardware accesses to the local NUMA node is usually quicker than the remote NUMA node. How slow the remote NUMA accesses are depends on the implementation details. This patch allocates memory for ITS management tables and command queue from the corresponding NUMA node using the appropriate NUMA aware functions. This change improves the performance of the ITS tables read latency on systems where it has more than one ITS block, and with the slower inter node accesses. Apache Web server benchmarking using ab tool on a HiSilicon D06 board with multiple numa mem nodes shows Time per request and Transfer rate improvements of ~3.6% with this patch. Signed-off-by: Shanker Donthineni Signed-off-by: Hanjun Guo Signed-off-by: Shameer Kolothum Reviewed-by: Ganapatrao Kulkarni Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-gic-v3-its.c | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c index 9f529a6aeaae..fb7157188294 100644 --- a/drivers/irqchip/irq-gic-v3-its.c +++ b/drivers/irqchip/irq-gic-v3-its.c @@ -1737,6 +1737,7 @@ static int its_setup_baser(struct its_node *its, struct its_baser *baser, u64 type = GITS_BASER_TYPE(val); u64 baser_phys, tmp; u32 alloc_pages; + struct page *page; void *base; retry_alloc_baser: @@ -1749,10 +1750,11 @@ retry_alloc_baser: order = get_order(GITS_BASER_PAGES_MAX * psz); } - base = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order); - if (!base) + page = alloc_pages_node(its->numa_node, GFP_KERNEL | __GFP_ZERO, order); + if (!page) return -ENOMEM; + base = (void *)page_address(page); baser_phys = virt_to_phys(base); /* Check if the physical address of the memory is above 48bits */ @@ -2238,7 +2240,8 @@ static struct its_baser *its_get_baser(struct its_node *its, u32 type) return NULL; } -static bool its_alloc_table_entry(struct its_baser *baser, u32 id) +static bool its_alloc_table_entry(struct its_node *its, + struct its_baser *baser, u32 id) { struct page *page; u32 esz, idx; @@ -2258,7 +2261,8 @@ static bool its_alloc_table_entry(struct its_baser *baser, u32 id) /* Allocate memory for 2nd level table */ if (!table[idx]) { - page = alloc_pages(GFP_KERNEL | __GFP_ZERO, get_order(baser->psz)); + page = alloc_pages_node(its->numa_node, GFP_KERNEL | __GFP_ZERO, + get_order(baser->psz)); if (!page) return false; @@ -2289,7 +2293,7 @@ static bool its_alloc_device_table(struct its_node *its, u32 dev_id) if (!baser) return (ilog2(dev_id) < its->device_ids); - return its_alloc_table_entry(baser, dev_id); + return its_alloc_table_entry(its, baser, dev_id); } static bool its_alloc_vpe_table(u32 vpe_id) @@ -2313,7 +2317,7 @@ static bool its_alloc_vpe_table(u32 vpe_id) if (!baser) return false; - if (!its_alloc_table_entry(baser, vpe_id)) + if (!its_alloc_table_entry(its, baser, vpe_id)) return false; } @@ -2347,7 +2351,7 @@ static struct its_device *its_create_device(struct its_node *its, u32 dev_id, nr_ites = max(2, nvecs); sz = nr_ites * its->ite_size; sz = max(sz, ITS_ITT_ALIGN) + ITS_ITT_ALIGN - 1; - itt = kzalloc(sz, GFP_KERNEL); + itt = kzalloc_node(sz, GFP_KERNEL, its->numa_node); if (alloc_lpis) { lpi_map = its_lpi_alloc(nvecs, &lpi_base, &nr_lpis); if (lpi_map) @@ -3488,6 +3492,7 @@ static int __init its_probe_one(struct resource *res, void __iomem *its_base; u32 val, ctlr; u64 baser, tmp, typer; + struct page *page; int err; its_base = ioremap(res->start, resource_size(res)); @@ -3543,12 +3548,13 @@ static int __init its_probe_one(struct resource *res, its->numa_node = numa_node; - its->cmd_base = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, - get_order(ITS_CMD_QUEUE_SZ)); - if (!its->cmd_base) { + page = alloc_pages_node(its->numa_node, GFP_KERNEL | __GFP_ZERO, + get_order(ITS_CMD_QUEUE_SZ)); + if (!page) { err = -ENOMEM; goto out_free_its; } + its->cmd_base = (void *)page_address(page); its->cmd_write = its->cmd_base; its->fwnode_handle = handle; its->get_msi_base = its_irq_get_msi_base; -- cgit v1.2.3 From 33517881ede742107f416533b8c3e4abc56763da Mon Sep 17 00:00:00 2001 From: Doug Berger Date: Wed, 20 Feb 2019 14:15:28 -0800 Subject: irqchip/brcmstb-l2: Use _irqsave locking variants in non-interrupt code Using the irq_gc_lock/irq_gc_unlock functions in the suspend and resume functions creates the opportunity for a deadlock during suspend, resume, and shutdown. Using the irq_gc_lock_irqsave/ irq_gc_unlock_irqrestore variants prevents this possible deadlock. Cc: stable@vger.kernel.org Fixes: 7f646e92766e2 ("irqchip: brcmstb-l2: Add Broadcom Set Top Box Level-2 interrupt controller") Signed-off-by: Doug Berger Signed-off-by: Florian Fainelli [maz: tidied up $SUBJECT] Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-brcmstb-l2.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-brcmstb-l2.c b/drivers/irqchip/irq-brcmstb-l2.c index 0e65f609352e..83364fedbf0a 100644 --- a/drivers/irqchip/irq-brcmstb-l2.c +++ b/drivers/irqchip/irq-brcmstb-l2.c @@ -129,8 +129,9 @@ static void brcmstb_l2_intc_suspend(struct irq_data *d) struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); struct brcmstb_l2_intc_data *b = gc->private; + unsigned long flags; - irq_gc_lock(gc); + irq_gc_lock_irqsave(gc, flags); /* Save the current mask */ b->saved_mask = irq_reg_readl(gc, ct->regs.mask); @@ -139,7 +140,7 @@ static void brcmstb_l2_intc_suspend(struct irq_data *d) irq_reg_writel(gc, ~gc->wake_active, ct->regs.disable); irq_reg_writel(gc, gc->wake_active, ct->regs.enable); } - irq_gc_unlock(gc); + irq_gc_unlock_irqrestore(gc, flags); } static void brcmstb_l2_intc_resume(struct irq_data *d) @@ -147,8 +148,9 @@ static void brcmstb_l2_intc_resume(struct irq_data *d) struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); struct brcmstb_l2_intc_data *b = gc->private; + unsigned long flags; - irq_gc_lock(gc); + irq_gc_lock_irqsave(gc, flags); if (ct->chip.irq_ack) { /* Clear unmasked non-wakeup interrupts */ irq_reg_writel(gc, ~b->saved_mask & ~gc->wake_active, @@ -158,7 +160,7 @@ static void brcmstb_l2_intc_resume(struct irq_data *d) /* Restore the saved mask */ irq_reg_writel(gc, b->saved_mask, ct->regs.disable); irq_reg_writel(gc, ~b->saved_mask, ct->regs.enable); - irq_gc_unlock(gc); + irq_gc_unlock_irqrestore(gc, flags); } static int __init brcmstb_l2_intc_of_init(struct device_node *np, -- cgit v1.2.3 From deb904e45b4e32517f91512db5c50457004313d2 Mon Sep 17 00:00:00 2001 From: Aisheng Dong Date: Wed, 20 Feb 2019 11:40:47 +0000 Subject: irqchip/imx-irqsteer: Change to use reg_num instead of irq_group One group can manage 64 interrupts by using two registers (e.g. STATUS/SET). However, the integrated irqsteer may support only 32 interrupts which needs only one register in a group. But the current driver assume there's a mininum of two registers in a group which result in a wrong register map for 32 interrupts per channel irqsteer. Let's use the reg_num caculated by interrupts per channel instead of irq_group to cover this case. Cc: Rob Herring Cc: Shawn Guo Reviewed-by: Lucas Stach Signed-off-by: Dong Aisheng Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-imx-irqsteer.c | 35 +++++++++++++++++++---------------- 1 file changed, 19 insertions(+), 16 deletions(-) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-imx-irqsteer.c b/drivers/irqchip/irq-imx-irqsteer.c index 5b3f1d735685..67ed86250cbf 100644 --- a/drivers/irqchip/irq-imx-irqsteer.c +++ b/drivers/irqchip/irq-imx-irqsteer.c @@ -13,7 +13,7 @@ #include #include -#define CTRL_STRIDE_OFF(_t, _r) (_t * 8 * _r) +#define CTRL_STRIDE_OFF(_t, _r) (_t * 4 * _r) #define CHANCTRL 0x0 #define CHANMASK(n, t) (CTRL_STRIDE_OFF(t, 0) + 0x4 * (n) + 0x4) #define CHANSET(n, t) (CTRL_STRIDE_OFF(t, 1) + 0x4 * (n) + 0x4) @@ -26,7 +26,7 @@ struct irqsteer_data { struct clk *ipg_clk; int irq; raw_spinlock_t lock; - int irq_groups; + int reg_num; int channel; struct irq_domain *domain; u32 *saved_reg; @@ -35,7 +35,7 @@ struct irqsteer_data { static int imx_irqsteer_get_reg_index(struct irqsteer_data *data, unsigned long irqnum) { - return (data->irq_groups * 2 - irqnum / 32 - 1); + return (data->reg_num - irqnum / 32 - 1); } static void imx_irqsteer_irq_unmask(struct irq_data *d) @@ -46,9 +46,9 @@ static void imx_irqsteer_irq_unmask(struct irq_data *d) u32 val; raw_spin_lock_irqsave(&data->lock, flags); - val = readl_relaxed(data->regs + CHANMASK(idx, data->irq_groups)); + val = readl_relaxed(data->regs + CHANMASK(idx, data->reg_num)); val |= BIT(d->hwirq % 32); - writel_relaxed(val, data->regs + CHANMASK(idx, data->irq_groups)); + writel_relaxed(val, data->regs + CHANMASK(idx, data->reg_num)); raw_spin_unlock_irqrestore(&data->lock, flags); } @@ -60,9 +60,9 @@ static void imx_irqsteer_irq_mask(struct irq_data *d) u32 val; raw_spin_lock_irqsave(&data->lock, flags); - val = readl_relaxed(data->regs + CHANMASK(idx, data->irq_groups)); + val = readl_relaxed(data->regs + CHANMASK(idx, data->reg_num)); val &= ~BIT(d->hwirq % 32); - writel_relaxed(val, data->regs + CHANMASK(idx, data->irq_groups)); + writel_relaxed(val, data->regs + CHANMASK(idx, data->reg_num)); raw_spin_unlock_irqrestore(&data->lock, flags); } @@ -94,13 +94,13 @@ static void imx_irqsteer_irq_handler(struct irq_desc *desc) chained_irq_enter(irq_desc_get_chip(desc), desc); - for (i = 0; i < data->irq_groups * 64; i += 32) { + for (i = 0; i < data->reg_num * 32; i += 32) { int idx = imx_irqsteer_get_reg_index(data, i); unsigned long irqmap; int pos, virq; irqmap = readl_relaxed(data->regs + - CHANSTATUS(idx, data->irq_groups)); + CHANSTATUS(idx, data->reg_num)); for_each_set_bit(pos, &irqmap, 32) { virq = irq_find_mapping(data->domain, pos + i); @@ -146,12 +146,15 @@ static int imx_irqsteer_probe(struct platform_device *pdev) raw_spin_lock_init(&data->lock); - of_property_read_u32(np, "fsl,irq-groups", &data->irq_groups); + of_property_read_u32(np, "fsl,num-irqs", &data->reg_num); of_property_read_u32(np, "fsl,channel", &data->channel); + /* one register bit map represents 32 input interrupts */ + data->reg_num /= 32; + if (IS_ENABLED(CONFIG_PM_SLEEP)) { data->saved_reg = devm_kzalloc(&pdev->dev, - sizeof(u32) * data->irq_groups * 2, + sizeof(u32) * data->reg_num, GFP_KERNEL); if (!data->saved_reg) return -ENOMEM; @@ -166,7 +169,7 @@ static int imx_irqsteer_probe(struct platform_device *pdev) /* steer all IRQs into configured channel */ writel_relaxed(BIT(data->channel), data->regs + CHANCTRL); - data->domain = irq_domain_add_linear(np, data->irq_groups * 64, + data->domain = irq_domain_add_linear(np, data->reg_num * 32, &imx_irqsteer_domain_ops, data); if (!data->domain) { dev_err(&pdev->dev, "failed to create IRQ domain\n"); @@ -199,9 +202,9 @@ static void imx_irqsteer_save_regs(struct irqsteer_data *data) { int i; - for (i = 0; i < data->irq_groups * 2; i++) + for (i = 0; i < data->reg_num; i++) data->saved_reg[i] = readl_relaxed(data->regs + - CHANMASK(i, data->irq_groups)); + CHANMASK(i, data->reg_num)); } static void imx_irqsteer_restore_regs(struct irqsteer_data *data) @@ -209,9 +212,9 @@ static void imx_irqsteer_restore_regs(struct irqsteer_data *data) int i; writel_relaxed(BIT(data->channel), data->regs + CHANCTRL); - for (i = 0; i < data->irq_groups * 2; i++) + for (i = 0; i < data->reg_num; i++) writel_relaxed(data->saved_reg[i], - data->regs + CHANMASK(i, data->irq_groups)); + data->regs + CHANMASK(i, data->reg_num)); } static int imx_irqsteer_suspend(struct device *dev) -- cgit v1.2.3 From 28528fca4908142bd1a3247956cba56c9c667d71 Mon Sep 17 00:00:00 2001 From: Aisheng Dong Date: Wed, 20 Feb 2019 11:40:51 +0000 Subject: irqchip/imx-irqsteer: Add multi output interrupts support One irqsteer channel can support up to 8 output interrupts. Cc: Marc Zyngier Cc: Lucas Stach Cc: Shawn Guo Reviewed-by: Lucas Stach Signed-off-by: Dong Aisheng Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-imx-irqsteer.c | 88 +++++++++++++++++++++++++++++--------- 1 file changed, 68 insertions(+), 20 deletions(-) (limited to 'drivers/irqchip') diff --git a/drivers/irqchip/irq-imx-irqsteer.c b/drivers/irqchip/irq-imx-irqsteer.c index 67ed86250cbf..d1098f4da6a4 100644 --- a/drivers/irqchip/irq-imx-irqsteer.c +++ b/drivers/irqchip/irq-imx-irqsteer.c @@ -10,6 +10,7 @@ #include #include #include +#include #include #include @@ -21,10 +22,13 @@ #define CHAN_MINTDIS(t) (CTRL_STRIDE_OFF(t, 3) + 0x4) #define CHAN_MASTRSTAT(t) (CTRL_STRIDE_OFF(t, 3) + 0x8) +#define CHAN_MAX_OUTPUT_INT 0x8 + struct irqsteer_data { void __iomem *regs; struct clk *ipg_clk; - int irq; + int irq[CHAN_MAX_OUTPUT_INT]; + int irq_count; raw_spinlock_t lock; int reg_num; int channel; @@ -87,23 +91,47 @@ static const struct irq_domain_ops imx_irqsteer_domain_ops = { .xlate = irq_domain_xlate_onecell, }; +static int imx_irqsteer_get_hwirq_base(struct irqsteer_data *data, u32 irq) +{ + int i; + + for (i = 0; i < data->irq_count; i++) { + if (data->irq[i] == irq) + return i * 64; + } + + return -EINVAL; +} + static void imx_irqsteer_irq_handler(struct irq_desc *desc) { struct irqsteer_data *data = irq_desc_get_handler_data(desc); - int i; + int hwirq; + int irq, i; chained_irq_enter(irq_desc_get_chip(desc), desc); - for (i = 0; i < data->reg_num * 32; i += 32) { - int idx = imx_irqsteer_get_reg_index(data, i); + irq = irq_desc_get_irq(desc); + hwirq = imx_irqsteer_get_hwirq_base(data, irq); + if (hwirq < 0) { + pr_warn("%s: unable to get hwirq base for irq %d\n", + __func__, irq); + return; + } + + for (i = 0; i < 2; i++, hwirq += 32) { + int idx = imx_irqsteer_get_reg_index(data, hwirq); unsigned long irqmap; int pos, virq; + if (hwirq >= data->reg_num * 32) + break; + irqmap = readl_relaxed(data->regs + CHANSTATUS(idx, data->reg_num)); for_each_set_bit(pos, &irqmap, 32) { - virq = irq_find_mapping(data->domain, pos + i); + virq = irq_find_mapping(data->domain, pos + hwirq); if (virq) generic_handle_irq(virq); } @@ -117,7 +145,8 @@ static int imx_irqsteer_probe(struct platform_device *pdev) struct device_node *np = pdev->dev.of_node; struct irqsteer_data *data; struct resource *res; - int ret; + u32 irqs_num; + int i, ret; data = devm_kzalloc(&pdev->dev, sizeof(*data), GFP_KERNEL); if (!data) @@ -130,12 +159,6 @@ static int imx_irqsteer_probe(struct platform_device *pdev) return PTR_ERR(data->regs); } - data->irq = platform_get_irq(pdev, 0); - if (data->irq <= 0) { - dev_err(&pdev->dev, "failed to get irq\n"); - return -ENODEV; - } - data->ipg_clk = devm_clk_get(&pdev->dev, "ipg"); if (IS_ERR(data->ipg_clk)) { ret = PTR_ERR(data->ipg_clk); @@ -146,11 +169,15 @@ static int imx_irqsteer_probe(struct platform_device *pdev) raw_spin_lock_init(&data->lock); - of_property_read_u32(np, "fsl,num-irqs", &data->reg_num); + of_property_read_u32(np, "fsl,num-irqs", &irqs_num); of_property_read_u32(np, "fsl,channel", &data->channel); - /* one register bit map represents 32 input interrupts */ - data->reg_num /= 32; + /* + * There is one output irq for each group of 64 inputs. + * One register bit map can represent 32 input interrupts. + */ + data->irq_count = DIV_ROUND_UP(irqs_num, 64); + data->reg_num = irqs_num / 32; if (IS_ENABLED(CONFIG_PM_SLEEP)) { data->saved_reg = devm_kzalloc(&pdev->dev, @@ -173,23 +200,44 @@ static int imx_irqsteer_probe(struct platform_device *pdev) &imx_irqsteer_domain_ops, data); if (!data->domain) { dev_err(&pdev->dev, "failed to create IRQ domain\n"); - clk_disable_unprepare(data->ipg_clk); - return -ENOMEM; + ret = -ENOMEM; + goto out; } - irq_set_chained_handler_and_data(data->irq, imx_irqsteer_irq_handler, - data); + if (!data->irq_count || data->irq_count > CHAN_MAX_OUTPUT_INT) { + ret = -EINVAL; + goto out; + } + + for (i = 0; i < data->irq_count; i++) { + data->irq[i] = irq_of_parse_and_map(np, i); + if (!data->irq[i]) { + ret = -EINVAL; + goto out; + } + + irq_set_chained_handler_and_data(data->irq[i], + imx_irqsteer_irq_handler, + data); + } platform_set_drvdata(pdev, data); return 0; +out: + clk_disable_unprepare(data->ipg_clk); + return ret; } static int imx_irqsteer_remove(struct platform_device *pdev) { struct irqsteer_data *irqsteer_data = platform_get_drvdata(pdev); + int i; + + for (i = 0; i < irqsteer_data->irq_count; i++) + irq_set_chained_handler_and_data(irqsteer_data->irq[i], + NULL, NULL); - irq_set_chained_handler_and_data(irqsteer_data->irq, NULL, NULL); irq_domain_remove(irqsteer_data->domain); clk_disable_unprepare(irqsteer_data->ipg_clk); -- cgit v1.2.3