Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751423AbdINGyJ (ORCPT ); Thu, 14 Sep 2017 02:54:09 -0400 Received: from mail-pg0-f67.google.com ([74.125.83.67]:35762 "EHLO mail-pg0-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750962AbdINGyG (ORCPT ); Thu, 14 Sep 2017 02:54:06 -0400 X-Google-Smtp-Source: AOwi7QDHwWiFzxsUWBq4oLDZPUvy9rbpN9F+3MWAfG26bca4M031qE7AOm5QYWrMMJGSFn9gzGJZDA== Date: Thu, 14 Sep 2017 15:54:02 +0900 From: Stafford Horne To: Marc Zyngier Cc: LKML , Openrisc , Stefan Kristiansson , Thomas Gleixner , Jason Cooper , Rob Herring , Mark Rutland , Jonas Bonn , devicetree@vger.kernel.org Subject: Re: [PATCH v2 06/14] irqchip: add initial support for ompic Message-ID: <20170914065402.GU2609@lianli.shorne-pla.net> References: <20170910064926.5874-1-shorne@gmail.com> <20170910064926.5874-7-shorne@gmail.com> <86h8w6see4.fsf@arm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <86h8w6see4.fsf@arm.com> User-Agent: Mutt/1.8.3 (2017-05-23) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 13185 Lines: 388 On Wed, Sep 13, 2017 at 06:21:39PM +0100, Marc Zyngier wrote: > On Sun, Sep 10 2017 at 3:49:18 pm BST, Stafford Horne wrote: > > From: Stefan Kristiansson > > > > IPI driver for the Open Multi-Processor Interrupt Controller (ompic) as > > described in the Multicore support section of the OpenRISC 1.2 > > proposed architecture specification: > > > > https://github.com/stffrdhrn/doc/raw/arch-1.2-proposal/openrisc-arch-1.2-rev0.pdf > > > > Each OpenRISC core contains a full interrupt controller which is used in > > the SMP architecture for interrupt balancing. This IPI device, the > > ompic, is the only external device required for enabling SMP on > > OpenRISC. > > > > Pending ops are stored in a memory bit mask which can allow multiple > > pending operations to be set and serviced at a time. This is mostly > > borrowed from the alpha IPI implementation. > > > > Signed-off-by: Stefan Kristiansson > > [shorne@gmail.com: converted ops to bitmask, wrote commit message] > > Signed-off-by: Stafford Horne > > --- > > > > Changes since v1 > > - Added openrisc, prefix > > - Clarified 8 bytes per cpu > > - Removed #interrupt-cells as this will not be an irq parent > > - Changed ops to be percpu > > - Added DTS and intialization failure validations > > > > .../interrupt-controller/openrisc,ompic.txt | 19 ++ > > arch/openrisc/Kconfig | 1 + > > drivers/irqchip/Kconfig | 3 + > > drivers/irqchip/Makefile | 1 + > > drivers/irqchip/irq-ompic.c | 205 +++++++++++++++++++++ > > 5 files changed, 229 insertions(+) > > create mode 100644 Documentation/devicetree/bindings/interrupt-controller/openrisc,ompic.txt > > create mode 100644 drivers/irqchip/irq-ompic.c > > > > diff --git a/Documentation/devicetree/bindings/interrupt-controller/openrisc,ompic.txt b/Documentation/devicetree/bindings/interrupt-controller/openrisc,ompic.txt > > new file mode 100644 > > index 000000000000..346e6042d62f > > --- /dev/null > > +++ b/Documentation/devicetree/bindings/interrupt-controller/openrisc,ompic.txt > > @@ -0,0 +1,19 @@ > > +Open Multi-Processor Interrupt Controller > > + > > +Required properties: > > + > > +- compatible : This should be "openrisc,ompic" > > +- reg : Specifies base physical address and size of the register space. The > > + size is based on the number of cores the controller has been configured > > + to handle, this should be set to 8 bytes per cpu core. > > +- interrupt-controller : Identifies the node as an interrupt controller > > +- interrupts : Specifies the interrupt line to which the ompic is wired. > > + > > +Example: > > + > > +ompic: ompic { > > + compatible = "openrisc,ompic"; > > + reg = <0x98000000 16>; > > + interrupt-controller; > > + interrupts = <1>; > > +}; > > diff --git a/arch/openrisc/Kconfig b/arch/openrisc/Kconfig > > index b49acda5e8f4..34eb4e90f56c 100644 > > --- a/arch/openrisc/Kconfig > > +++ b/arch/openrisc/Kconfig > > @@ -30,6 +30,7 @@ config OPENRISC > > select NO_BOOTMEM > > select ARCH_USE_QUEUED_SPINLOCKS > > select ARCH_USE_QUEUED_RWLOCKS > > + select OMPIC if SMP > > > > config CPU_BIG_ENDIAN > > def_bool y > > diff --git a/drivers/irqchip/Kconfig b/drivers/irqchip/Kconfig > > index f1fd5f44d1d4..0e4c96c90b86 100644 > > --- a/drivers/irqchip/Kconfig > > +++ b/drivers/irqchip/Kconfig > > @@ -145,6 +145,9 @@ config CLPS711X_IRQCHIP > > select SPARSE_IRQ > > default y > > > > +config OMPIC > > + bool > > + > > config OR1K_PIC > > bool > > select IRQ_DOMAIN > > diff --git a/drivers/irqchip/Makefile b/drivers/irqchip/Makefile > > index e88d856cc09c..123047d7a20d 100644 > > --- a/drivers/irqchip/Makefile > > +++ b/drivers/irqchip/Makefile > > @@ -17,6 +17,7 @@ obj-$(CONFIG_DW_APB_ICTL) += irq-dw-apb-ictl.o > > obj-$(CONFIG_METAG) += irq-metag-ext.o > > obj-$(CONFIG_METAG_PERFCOUNTER_IRQS) += irq-metag.o > > obj-$(CONFIG_CLPS711X_IRQCHIP) += irq-clps711x.o > > +obj-$(CONFIG_OMPIC) += irq-ompic.o > > obj-$(CONFIG_OR1K_PIC) += irq-or1k-pic.o > > obj-$(CONFIG_ORION_IRQCHIP) += irq-orion.o > > obj-$(CONFIG_OMAP_IRQCHIP) += irq-omap-intc.o > > diff --git a/drivers/irqchip/irq-ompic.c b/drivers/irqchip/irq-ompic.c > > new file mode 100644 > > index 000000000000..cd2616b6639b > > --- /dev/null > > +++ b/drivers/irqchip/irq-ompic.c > > @@ -0,0 +1,205 @@ > > +/* > > + * Open Multi-Processor Interrupt Controller driver > > + * > > + * Copyright (C) 2014 Stefan Kristiansson > > + * Copyright (C) 2017 Stafford Horne > > + * > > + * This file is licensed under the terms of the GNU General Public License > > + * version 2. This program is licensed "as is" without any warranty of any > > + * kind, whether express or implied. > > + * > > + * The ompic device handles IPI communication because cores in mulicore > > + * OpenRISC systems. > > Should the above read "between cores"? Yes, it should be, I am bad with these kind of typos. > > + * > > + * Registers > > + * > > + * For each CPU the ompic has 2 registers. The control register for sending > > + * and acking IPIs and the status register for receiving IPIs. The register > > + * layouts are as follows: > > + * > > + * Control register > > + * +---------+---------+----------+---------+ > > + * | 31 | 30 | 29 .. 16 | 15 .. 0 | > > + * ----------+---------+----------+---------- > > + * | IRQ ACK | IRQ GEN | DST CORE | DATA | > > + * +---------+---------+----------+---------+ > > + * > > + * Status register > > + * +----------+-------------+----------+---------+ > > + * | 31 | 30 | 29 .. 16 | 15 .. 0 | > > + * -----------+-------------+----------+---------+ > > + * | Reserved | IRQ Pending | SRC CORE | DATA | > > + * +----------+-------------+----------+---------+ > > + * > > + * Architecture > > + * > > + * - The ompic generates a level interrupt to the CPU PIC when a message is > > + * ready. Messages are delivered via the memory bus. > > + * - The ompic does not have any interrupt input lines. > > + * - The ompic is wired to the same irq line on each core. > > + * - Devices are wired to the same irq line on each core. > > + * > > + * +---------+ +---------+ > > + * | CPU | | CPU | > > + * | Core 0 |<==\ (memory access) /==>| Core 1 | > > + * | [ PIC ]| | | | [ PIC ]| > > + * +----^-^--+ | | +----^-^--+ > > + * | | v v | | > > + * <====|=|=================================|=|==> (memory bus) > > + * | | ^ ^ | | > > + * (ipi | +------|---------+--------|-------|-+ (device irq) > > + * irq | | | | | > > + * core0)| +------|---------|--------|-------+ (ipi irq core1) > > + * | | | | | > > + * +----o-o-+ | +--------+ | > > + * | ompic |<===/ | Device |<===/ > > + * | IPI | +--------+ > > + * +--------+* > > + * > > + */ > > + > > +#include > > +#include > > +#include > > +#include > > +#include > > +#include > > +#include > > + > > +#include > > + > > +#define OMPIC_CPUBYTES 8 > > +#define OMPIC_CTRL(cpu) (0x0 + (cpu * OMPIC_CPUBYTES)) > > +#define OMPIC_STAT(cpu) (0x4 + (cpu * OMPIC_CPUBYTES)) > > + > > +#define OMPIC_CTRL_IRQ_ACK (1 << 31) > > +#define OMPIC_CTRL_IRQ_GEN (1 << 30) > > +#define OMPIC_CTRL_DST(cpu) (((cpu) & 0x3fff) << 16) > > + > > +#define OMPIC_STAT_IRQ_PENDING (1 << 30) > > + > > +#define OMPIC_DATA(x) ((x) & 0xffff) > > + > > +DEFINE_PER_CPU(unsigned long, ops); > > + > > +static void __iomem *ompic_base; > > + > > +static inline u32 ompic_readreg(void __iomem *base, loff_t offset) > > +{ > > + return ioread32be(base + offset); > > +} > > + > > +static void ompic_writereg(void __iomem *base, loff_t offset, u32 data) > > +{ > > + iowrite32be(data, base + offset); > > +} > > + > > +void ompic_raise_softirq(const struct cpumask *mask, unsigned int ipi_msg) > > Please make this static. OK > > +{ > > + unsigned int dst_cpu; > > + unsigned int src_cpu = smp_processor_id(); > > + > > + for_each_cpu(dst_cpu, mask) { > > + set_bit(ipi_msg, &per_cpu(ops, dst_cpu)); > > + > > + /* > > + * On OpenRISC the atomic set_bit() call implies a memory > > + * barrier. Otherwise we would need: smp_wmb(); paired > > + * with the read in ompic_ipi_handler. > > + */ > > One last question on this, because the architecture document is terribly > unclear: If you have CPU0 doing an atomic operation A0, CPU1 seeing A0 > and doeing another atomic A1 (the set_bit above) followed by an IPI to > CPU2, is CPU2 *guaranteed* to observe both A0 *and* A1? Because that's > required by the IPI semantics, and you wouldn't see that kind of issue > with only two CPUs. Could you suggest an architecture document that makes this case clear? I believe this will not be a problem, but: 1. If this needs to be clear in the architecture document I can propose changes. 2. To be clear is this the scenario you mean.. CASE1 - A0 and A1 are to different locations? A0 - writes to some unrelated global location? CPU0 CPU1 CPU2 A0:atomic store (global) A1:set_bit (ops[CPU2]) IPI read (A0,A1) OR CASE2 - A0 and A1 are to the same location. A0 - writes to the same location as A1 CPU0 CPU1 CPU2 A0:set_bit (ops[CPU2]) A1:set_bit (ops[CPU2]) IPI read (A0,A1) IPI OR - something else? In both cases CPU2 would be able to see the results of both atomic operations. All, cpus in the OpenRISC system snoop for memory writes to enable cash coherency, so each CPU would see each write once it is synced to memory (there is a single memory bus). This is not limited to atomic operations, but the atomic operations provide a syncrhonization point accross all CPUs. > > + > > + ompic_writereg(ompic_base, OMPIC_CTRL(src_cpu), > > + OMPIC_CTRL_IRQ_GEN | > > + OMPIC_CTRL_DST(dst_cpu) | > > + OMPIC_DATA(1)); > > + } > > +} > > + > > +irqreturn_t ompic_ipi_handler(int irq, void *dev_id) > > This should be static. OK > > +{ > > + unsigned int cpu = smp_processor_id(); > > + unsigned long *pending_ops = &per_cpu(ops, cpu); > > + unsigned long ops; > > + > > + ompic_writereg(ompic_base, OMPIC_CTRL(cpu), OMPIC_CTRL_IRQ_ACK); > > + while ((ops = xchg(pending_ops, 0)) != 0) { > > + > > + /* > > + * On OpenRISC the atomic xchg() call implies a memory > > + * barrier. Otherwise we may need an smp_rmb(); paired > > + * with the write in ompic_raise_softirq. > > + */ > > + > > + do { > > + unsigned long ipi_msg; > > + > > + ipi_msg = __ffs(ops); > > + ops &= ~(1UL << ipi_msg); > > + > > + handle_IPI(ipi_msg); > > + } while (ops); > > + } > > + > > + return IRQ_HANDLED; > > +} > > + > > +static struct irqaction ompi_ipi_irqaction = { > > + .handler = ompic_ipi_handler, > > + .flags = IRQF_PERCPU, > > + .name = "ompic_ipi", > > +}; > > + > > +int __init ompic_of_init(struct device_node *node, struct device_node *parent) > > static again. OK > > +{ > > + struct resource res; > > + int irq; > > + int ret; > > + > > + /* Validate the DT */ > > + if (ompic_base) { > > + pr_err("ompic: duplicate ompic's are not supported"); > > + return -EEXIST; > > + } > > + > > + if (of_address_to_resource(node, 0, &res)) { > > + pr_err("ompic: reg property requires an address and size"); > > + return -EINVAL; > > + } > > + > > + if (resource_size(&res) < (num_possible_cpus() * OMPIC_CPUBYTES)) { > > + pr_err("ompic: reg size, currently %d must be at least %d", > > + resource_size(&res), > > + (num_possible_cpus() * OMPIC_CPUBYTES)); > > + return -EINVAL; > > + } > > + > > + /* Setup the device */ > > + ompic_base = ioremap(res.start, resource_size(&res)); > > + if (IS_ERR(ompic_base)) { > > + pr_err("ompic: unable to map registers"); > > + return PTR_ERR(ompic_base); > > + } > > + > > + irq = irq_of_parse_and_map(node, 0); > > + if (irq <= 0) { > > + pr_err("ompic: unable to parse device irq"); > > + ret = -EINVAL; > > + goto out_unmap; > > + } > > + > > + ret = setup_irq(irq, &ompi_ipi_irqaction); > > + if (ret) > > + goto out_irq_disp; > > + > > + set_smp_cross_call(ompic_raise_softirq); > > + > > + return 0; > > + > > +out_irq_disp: > > + irq_dispose_mapping(irq); > > +out_unmap: > > + iounmap(ompic_base); > > + ompic_base = NULL; > > + return ret; > > +} > > +IRQCHIP_DECLARE(ompic, "openrisc,ompic", ompic_of_init); > Thanks for your time! -Stafford ps: Frank Zappa rocks :)