From patchwork Wed Mar 16 17:18:27 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Peter Maydell X-Patchwork-Id: 8603061 Return-Path: X-Original-To: patchwork-qemu-devel@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork1.web.kernel.org (Postfix) with ESMTP id 74CB49F294 for ; Wed, 16 Mar 2016 17:27:55 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 095EF20212 for ; Wed, 16 Mar 2016 17:27:54 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 72A4720172 for ; Wed, 16 Mar 2016 17:27:52 +0000 (UTC) Received: from localhost ([::1]:57732 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1agFEd-0005oA-O5 for patchwork-qemu-devel@patchwork.kernel.org; Wed, 16 Mar 2016 13:27:51 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:35414) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1agF5r-0002C3-7Z for qemu-devel@nongnu.org; Wed, 16 Mar 2016 13:18:56 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1agF5o-0001Ex-TA for qemu-devel@nongnu.org; Wed, 16 Mar 2016 13:18:47 -0400 Received: from orth.archaic.org.uk ([2001:8b0:1d0::2]:56203) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1agF5o-0001Bd-It for qemu-devel@nongnu.org; Wed, 16 Mar 2016 13:18:44 -0400 Received: from pm215 by orth.archaic.org.uk with local (Exim 4.84_2) (envelope-from ) id 1agF5n-0000yJ-B9 for qemu-devel@nongnu.org; Wed, 16 Mar 2016 17:18:43 +0000 From: Peter Maydell To: qemu-devel@nongnu.org Date: Wed, 16 Mar 2016 17:18:27 +0000 Message-Id: <1458148715-16864-14-git-send-email-peter.maydell@linaro.org> X-Mailer: git-send-email 1.9.1 In-Reply-To: <1458148715-16864-1-git-send-email-peter.maydell@linaro.org> References: <1458148715-16864-1-git-send-email-peter.maydell@linaro.org> X-detected-operating-system: by eggs.gnu.org: GNU/Linux 2.2.x-3.x [generic] X-Received-From: 2001:8b0:1d0::2 Subject: [Qemu-devel] [PULL 13/21] hw/intc: Add (new) ASPEED VIC device model X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: Andrew Jeffery Implement a basic ASPEED VIC device model for the AST2400 SoC[1], with enough functionality to boot an aspeed_defconfig Linux kernel. The model implements the 'new' (revised) register set: While the hardware exposes both the new and legacy register sets, accesses to the model's legacy register set will not be serviced (however the access will be logged). [1] http://www.aspeedtech.com/products.php?fPath=20&rId=376 Signed-off-by: Andrew Jeffery Message-id: 1458096317-25223-3-git-send-email-andrew@aj.id.au Reviewed-by: Peter Maydell Signed-off-by: Peter Maydell --- hw/intc/Makefile.objs | 1 + hw/intc/aspeed_vic.c | 339 +++++++++++++++++++++++++++++++++++++++++++ include/hw/intc/aspeed_vic.h | 48 ++++++ trace-events | 7 + 4 files changed, 395 insertions(+) create mode 100644 hw/intc/aspeed_vic.c create mode 100644 include/hw/intc/aspeed_vic.h diff --git a/hw/intc/Makefile.objs b/hw/intc/Makefile.objs index 6a13a39..0e47f0f 100644 --- a/hw/intc/Makefile.objs +++ b/hw/intc/Makefile.objs @@ -31,3 +31,4 @@ obj-$(CONFIG_XICS_KVM) += xics_kvm.o obj-$(CONFIG_ALLWINNER_A10_PIC) += allwinner-a10-pic.o obj-$(CONFIG_S390_FLIC) += s390_flic.o obj-$(CONFIG_S390_FLIC_KVM) += s390_flic_kvm.o +obj-$(CONFIG_ASPEED_SOC) += aspeed_vic.o diff --git a/hw/intc/aspeed_vic.c b/hw/intc/aspeed_vic.c new file mode 100644 index 0000000..19a0ff7 --- /dev/null +++ b/hw/intc/aspeed_vic.c @@ -0,0 +1,339 @@ +/* + * ASPEED Interrupt Controller (New) + * + * Andrew Jeffery + * + * Copyright 2015, 2016 IBM Corp. + * + * This code is licensed under the GPL version 2 or later. See + * the COPYING file in the top-level directory. + */ + +/* The hardware exposes two register sets, a legacy set and a 'new' set. The + * model implements the 'new' register set, and logs warnings on accesses to + * the legacy IO space. + * + * The hardware uses 32bit registers to manage 51 IRQs, with low and high + * registers for each conceptual register. The device model's implementation + * uses 64bit data types to store both low and high register values (in the one + * member), but must cope with access offset values in multiples of 4 passed to + * the callbacks. As such the read() and write() implementations process the + * provided offset to understand whether the access is requesting the lower or + * upper 32 bits of the 64bit member. + * + * Additionally, the "Interrupt Enable", "Edge Status" and "Software Interrupt" + * fields have separate "enable"/"status" and "clear" registers, where set bits + * are written to one or the other to change state (avoiding a + * read-modify-write sequence). + */ + +#include "qemu/osdep.h" +#include +#include "hw/intc/aspeed_vic.h" +#include "qemu/bitops.h" +#include "trace.h" + +#define AVIC_NEW_BASE_OFFSET 0x80 + +#define AVIC_L_MASK 0xFFFFFFFFU +#define AVIC_H_MASK 0x0007FFFFU +#define AVIC_EVENT_W_MASK (0x78000ULL << 32) + +static void aspeed_vic_update(AspeedVICState *s) +{ + uint64_t new = (s->raw & s->enable); + uint64_t flags; + + flags = new & s->select; + trace_aspeed_vic_update_fiq(!!flags); + qemu_set_irq(s->fiq, !!flags); + + flags = new & ~s->select; + trace_aspeed_vic_update_irq(!!flags); + qemu_set_irq(s->irq, !!flags); +} + +static void aspeed_vic_set_irq(void *opaque, int irq, int level) +{ + uint64_t irq_mask; + bool raise; + AspeedVICState *s = (AspeedVICState *)opaque; + + if (irq > ASPEED_VIC_NR_IRQS) { + qemu_log_mask(LOG_GUEST_ERROR, "%s: Invalid interrupt number: %d\n", + __func__, irq); + return; + } + + trace_aspeed_vic_set_irq(irq, level); + + irq_mask = BIT(irq); + if (s->sense & irq_mask) { + /* level-triggered */ + if (s->event & irq_mask) { + /* high-sensitive */ + raise = level; + } else { + /* low-sensitive */ + raise = !level; + } + s->raw = deposit64(s->raw, irq, 1, raise); + } else { + uint64_t old_level = s->level & irq_mask; + + /* edge-triggered */ + if (s->dual_edge & irq_mask) { + raise = (!!old_level) != (!!level); + } else { + if (s->event & irq_mask) { + /* rising-sensitive */ + raise = !old_level && level; + } else { + /* falling-sensitive */ + raise = old_level && !level; + } + } + if (raise) { + s->raw = deposit64(s->raw, irq, 1, raise); + } + } + s->level = deposit64(s->level, irq, 1, level); + aspeed_vic_update(s); +} + +static uint64_t aspeed_vic_read(void *opaque, hwaddr offset, unsigned size) +{ + uint64_t val; + const bool high = !!(offset & 0x4); + hwaddr n_offset = (offset & ~0x4); + AspeedVICState *s = (AspeedVICState *)opaque; + + if (offset < AVIC_NEW_BASE_OFFSET) { + qemu_log_mask(LOG_UNIMP, "%s: Ignoring read from legacy registers " + "at 0x%" HWADDR_PRIx "[%u]\n", __func__, offset, size); + return 0; + } + + n_offset -= AVIC_NEW_BASE_OFFSET; + + switch (n_offset) { + case 0x0: /* IRQ Status */ + val = s->raw & ~s->select & s->enable; + break; + case 0x08: /* FIQ Status */ + val = s->raw & s->select & s->enable; + break; + case 0x10: /* Raw Interrupt Status */ + val = s->raw; + break; + case 0x18: /* Interrupt Selection */ + val = s->select; + break; + case 0x20: /* Interrupt Enable */ + val = s->enable; + break; + case 0x30: /* Software Interrupt */ + val = s->trigger; + break; + case 0x40: /* Interrupt Sensitivity */ + val = s->sense; + break; + case 0x48: /* Interrupt Both Edge Trigger Control */ + val = s->dual_edge; + break; + case 0x50: /* Interrupt Event */ + val = s->event; + break; + case 0x60: /* Edge Triggered Interrupt Status */ + val = s->raw & ~s->sense; + break; + /* Illegal */ + case 0x28: /* Interrupt Enable Clear */ + case 0x38: /* Software Interrupt Clear */ + case 0x58: /* Edge Triggered Interrupt Clear */ + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Read of write-only register with offset 0x%" + HWADDR_PRIx "\n", __func__, offset); + val = 0; + break; + default: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Bad register at offset 0x%" HWADDR_PRIx "\n", + __func__, offset); + val = 0; + break; + } + if (high) { + val = extract64(val, 32, 19); + } + trace_aspeed_vic_read(offset, size, val); + return val; +} + +static void aspeed_vic_write(void *opaque, hwaddr offset, uint64_t data, + unsigned size) +{ + const bool high = !!(offset & 0x4); + hwaddr n_offset = (offset & ~0x4); + AspeedVICState *s = (AspeedVICState *)opaque; + + if (offset < AVIC_NEW_BASE_OFFSET) { + qemu_log_mask(LOG_UNIMP, + "%s: Ignoring write to legacy registers at 0x%" + HWADDR_PRIx "[%u] <- 0x%" PRIx64 "\n", __func__, offset, + size, data); + return; + } + + n_offset -= AVIC_NEW_BASE_OFFSET; + trace_aspeed_vic_write(offset, size, data); + + /* Given we have members using separate enable/clear registers, deposit64() + * isn't quite the tool for the job. Instead, relocate the incoming bits to + * the required bit offset based on the provided access address + */ + if (high) { + data &= AVIC_H_MASK; + data <<= 32; + } else { + data &= AVIC_L_MASK; + } + + switch (n_offset) { + case 0x18: /* Interrupt Selection */ + /* Register has deposit64() semantics - overwrite requested 32 bits */ + if (high) { + s->select &= AVIC_L_MASK; + } else { + s->select &= ((uint64_t) AVIC_H_MASK) << 32; + } + s->select |= data; + break; + case 0x20: /* Interrupt Enable */ + s->enable |= data; + break; + case 0x28: /* Interrupt Enable Clear */ + s->enable &= ~data; + break; + case 0x30: /* Software Interrupt */ + qemu_log_mask(LOG_UNIMP, "%s: Software interrupts unavailable. " + "IRQs requested: 0x%016" PRIx64 "\n", __func__, data); + break; + case 0x38: /* Software Interrupt Clear */ + qemu_log_mask(LOG_UNIMP, "%s: Software interrupts unavailable. " + "IRQs to be cleared: 0x%016" PRIx64 "\n", __func__, data); + break; + case 0x50: /* Interrupt Event */ + /* Register has deposit64() semantics - overwrite the top four valid + * IRQ bits, as only the top four IRQs (GPIOs) can change their event + * type */ + if (high) { + s->event &= ~AVIC_EVENT_W_MASK; + s->event |= (data & AVIC_EVENT_W_MASK); + } else { + qemu_log_mask(LOG_GUEST_ERROR, + "Ignoring invalid write to interrupt event register"); + } + break; + case 0x58: /* Edge Triggered Interrupt Clear */ + s->raw &= ~(data & ~s->sense); + break; + case 0x00: /* IRQ Status */ + case 0x08: /* FIQ Status */ + case 0x10: /* Raw Interrupt Status */ + case 0x40: /* Interrupt Sensitivity */ + case 0x48: /* Interrupt Both Edge Trigger Control */ + case 0x60: /* Edge Triggered Interrupt Status */ + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Write of read-only register with offset 0x%" + HWADDR_PRIx "\n", __func__, offset); + break; + + default: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: Bad register at offset 0x%" HWADDR_PRIx "\n", + __func__, offset); + break; + } + aspeed_vic_update(s); +} + +static const MemoryRegionOps aspeed_vic_ops = { + .read = aspeed_vic_read, + .write = aspeed_vic_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid.min_access_size = 4, + .valid.max_access_size = 4, + .valid.unaligned = false, +}; + +static void aspeed_vic_reset(DeviceState *dev) +{ + AspeedVICState *s = ASPEED_VIC(dev); + + s->level = 0; + s->raw = 0; + s->select = 0; + s->enable = 0; + s->trigger = 0; + s->sense = 0x1F07FFF8FFFFULL; + s->dual_edge = 0xF800070000ULL; + s->event = 0x5F07FFF8FFFFULL; +} + +#define AVIC_IO_REGION_SIZE 0x20000 + +static void aspeed_vic_realize(DeviceState *dev, Error **errp) +{ + SysBusDevice *sbd = SYS_BUS_DEVICE(dev); + AspeedVICState *s = ASPEED_VIC(dev); + + memory_region_init_io(&s->iomem, OBJECT(s), &aspeed_vic_ops, s, + TYPE_ASPEED_VIC, AVIC_IO_REGION_SIZE); + + sysbus_init_mmio(sbd, &s->iomem); + + qdev_init_gpio_in(dev, aspeed_vic_set_irq, ASPEED_VIC_NR_IRQS); + sysbus_init_irq(sbd, &s->irq); + sysbus_init_irq(sbd, &s->fiq); +} + +static const VMStateDescription vmstate_aspeed_vic = { + .name = "aspeed.new-vic", + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_UINT64(level, AspeedVICState), + VMSTATE_UINT64(raw, AspeedVICState), + VMSTATE_UINT64(select, AspeedVICState), + VMSTATE_UINT64(enable, AspeedVICState), + VMSTATE_UINT64(trigger, AspeedVICState), + VMSTATE_UINT64(sense, AspeedVICState), + VMSTATE_UINT64(dual_edge, AspeedVICState), + VMSTATE_UINT64(event, AspeedVICState), + VMSTATE_END_OF_LIST() + } +}; + +static void aspeed_vic_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + dc->realize = aspeed_vic_realize; + dc->reset = aspeed_vic_reset; + dc->desc = "ASPEED Interrupt Controller (New)"; + dc->vmsd = &vmstate_aspeed_vic; +} + +static const TypeInfo aspeed_vic_info = { + .name = TYPE_ASPEED_VIC, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(AspeedVICState), + .class_init = aspeed_vic_class_init, +}; + +static void aspeed_vic_register_types(void) +{ + type_register_static(&aspeed_vic_info); +} + +type_init(aspeed_vic_register_types); diff --git a/include/hw/intc/aspeed_vic.h b/include/hw/intc/aspeed_vic.h new file mode 100644 index 0000000..107ff17 --- /dev/null +++ b/include/hw/intc/aspeed_vic.h @@ -0,0 +1,48 @@ +/* + * ASPEED Interrupt Controller (New) + * + * Andrew Jeffery + * + * Copyright 2016 IBM Corp. + * + * This code is licensed under the GPL version 2 or later. See + * the COPYING file in the top-level directory. + * + * Need to add SVIC and CVIC support + */ +#ifndef ASPEED_VIC_H +#define ASPEED_VIC_H + +#include "hw/sysbus.h" + +#define TYPE_ASPEED_VIC "aspeed.vic" +#define ASPEED_VIC(obj) OBJECT_CHECK(AspeedVICState, (obj), TYPE_ASPEED_VIC) + +#define ASPEED_VIC_NR_IRQS 51 + +typedef struct AspeedVICState { + /*< private >*/ + SysBusDevice parent_obj; + + /*< public >*/ + MemoryRegion iomem; + qemu_irq irq; + qemu_irq fiq; + + uint64_t level; + uint64_t raw; + uint64_t select; + uint64_t enable; + uint64_t trigger; + + /* 0=edge, 1=level */ + uint64_t sense; + + /* 0=single-edge, 1=dual-edge */ + uint64_t dual_edge; + + /* 0=low-sensitive/falling-edge, 1=high-sensitive/rising-edge */ + uint64_t event; +} AspeedVICState; + +#endif /* ASPEED_VIC_H */ diff --git a/trace-events b/trace-events index e589230..d494de1 100644 --- a/trace-events +++ b/trace-events @@ -1901,3 +1901,10 @@ aspeed_timer_ctrl_pulse_enable(uint8_t i, bool enable) "Timer %" PRIu8 ": %d" aspeed_timer_set_ctrl2(uint32_t value) "Value: 0x%" PRIx32 aspeed_timer_set_value(int timer, int reg, uint32_t value) "Timer %d register %d: 0x%" PRIx32 aspeed_timer_read(uint64_t offset, unsigned size, uint64_t value) "From 0x%" PRIx64 ": of size %u: 0x%" PRIx64 + +# hw/intc/aspeed_vic.c +aspeed_vic_set_irq(int irq, int level) "Enabling IRQ %d: %d" +aspeed_vic_update_fiq(int flags) "Raising FIQ: %d" +aspeed_vic_update_irq(int flags) "Raising IRQ: %d" +aspeed_vic_read(uint64_t offset, unsigned size, uint32_t value) "From 0x%" PRIx64 " of size %u: 0x%" PRIx32 +aspeed_vic_write(uint64_t offset, unsigned size, uint32_t data) "To 0x%" PRIx64 " of size %u: 0x%" PRIx32