| Message ID | 20190924084201.107958-2-anup.patel@wdc.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | RTC support for QEMU RISC-V virt machine | expand |
On Tue, 24 Sep 2019 at 09:45, Anup Patel <Anup.Patel@wdc.com> wrote: > > This patch adds model for Google Goldfish virtual platform RTC device. > > We will be adding Goldfish RTC device to the QEMU RISC-V virt machine > for providing real date-time to Guest Linux. The corresponding Linux > driver for Goldfish RTC device is already available in upstream Linux. > > For now, VM migration support is not available for Goldfish RTC device > but it will be added later when we implement VM migration for KVM RISC-V. > > Signed-off-by: Anup Patel <anup.patel@wdc.com> > --- > + > +static Property goldfish_rtc_properties[] = { > + DEFINE_PROP_UINT64("tick-offset", GoldfishRTCState, tick_offset, 0), > + DEFINE_PROP_UINT64("alarm-next", GoldfishRTCState, alarm_next, 0), > + DEFINE_PROP_UINT32("alarm-running", GoldfishRTCState, alarm_running, 0), > + DEFINE_PROP_UINT32("irq-pending", GoldfishRTCState, irq_pending, 0), > + DEFINE_PROP_UINT32("irq-enabled", GoldfishRTCState, irq_enabled, 0), > + DEFINE_PROP_END_OF_LIST(), > +}; What are all these properties trying to do ? > + > +static void goldfish_rtc_class_init(ObjectClass *klass, void *data) > +{ > + DeviceClass *dc = DEVICE_CLASS(klass); > + dc->props = goldfish_rtc_properties; Missing reset function. If you don't want to implement migration support now you should at least put in something that block migration. (Personally I prefer to just write the vmstate, it's as easy as writing the code to block migrations.) thanks -- PMM
> -----Original Message----- > From: Peter Maydell <peter.maydell@linaro.org> > Sent: Tuesday, September 24, 2019 3:21 PM > To: Anup Patel <Anup.Patel@wdc.com> > Cc: Palmer Dabbelt <palmer@sifive.com>; Alistair Francis > <Alistair.Francis@wdc.com>; Sagar Karandikar <sagark@eecs.berkeley.edu>; > Bastian Koppelmann <kbastian@mail.uni-paderborn.de>; Atish Patra > <Atish.Patra@wdc.com>; qemu-riscv@nongnu.org; qemu- > devel@nongnu.org; Anup Patel <anup@brainfault.org> > Subject: Re: [PATCH 1/2] hw: timer: Add Goldfish RTC device > > On Tue, 24 Sep 2019 at 09:45, Anup Patel <Anup.Patel@wdc.com> wrote: > > > > This patch adds model for Google Goldfish virtual platform RTC device. > > > > We will be adding Goldfish RTC device to the QEMU RISC-V virt machine > > for providing real date-time to Guest Linux. The corresponding Linux > > driver for Goldfish RTC device is already available in upstream Linux. > > > > For now, VM migration support is not available for Goldfish RTC device > > but it will be added later when we implement VM migration for KVM RISC- > V. > > > > Signed-off-by: Anup Patel <anup.patel@wdc.com> > > --- > > + > > +static Property goldfish_rtc_properties[] = { > > + DEFINE_PROP_UINT64("tick-offset", GoldfishRTCState, tick_offset, 0), > > + DEFINE_PROP_UINT64("alarm-next", GoldfishRTCState, alarm_next, 0), > > + DEFINE_PROP_UINT32("alarm-running", GoldfishRTCState, > alarm_running, 0), > > + DEFINE_PROP_UINT32("irq-pending", GoldfishRTCState, irq_pending, > 0), > > + DEFINE_PROP_UINT32("irq-enabled", GoldfishRTCState, irq_enabled, > 0), > > + DEFINE_PROP_END_OF_LIST(), > > +}; > > What are all these properties trying to do ? Argh, I forgot to remove these before sending. I will drop these in next revision. > > > + > > +static void goldfish_rtc_class_init(ObjectClass *klass, void *data) { > > + DeviceClass *dc = DEVICE_CLASS(klass); > > + dc->props = goldfish_rtc_properties; > > Missing reset function. Sure, I will add it. I got confused because I saw reset function in HPET but not in PL031. > > If you don't want to implement migration support now you should at least > put in something that block migration. > (Personally I prefer to just write the vmstate, it's as easy as writing the code > to block migrations.) Okay, I will add vmstate. Is there a way to unit test the vmstate part ? OR Are you fine if we test-n-fix it later ? > > thanks > -- PMM Regards, Anup
On Tue, 24 Sep 2019 at 12:17, Anup Patel <Anup.Patel@wdc.com> wrote: > > > > > -----Original Message----- > > From: Peter Maydell <peter.maydell@linaro.org> > > Sent: Tuesday, September 24, 2019 3:21 PM > > To: Anup Patel <Anup.Patel@wdc.com> > > Cc: Palmer Dabbelt <palmer@sifive.com>; Alistair Francis > > <Alistair.Francis@wdc.com>; Sagar Karandikar <sagark@eecs.berkeley.edu>; > > Bastian Koppelmann <kbastian@mail.uni-paderborn.de>; Atish Patra > > <Atish.Patra@wdc.com>; qemu-riscv@nongnu.org; qemu- > > devel@nongnu.org; Anup Patel <anup@brainfault.org> > > Subject: Re: [PATCH 1/2] hw: timer: Add Goldfish RTC device > > > > On Tue, 24 Sep 2019 at 09:45, Anup Patel <Anup.Patel@wdc.com> wrote: > > > > > > This patch adds model for Google Goldfish virtual platform RTC device. > > > > > > We will be adding Goldfish RTC device to the QEMU RISC-V virt machine > > > for providing real date-time to Guest Linux. The corresponding Linux > > > driver for Goldfish RTC device is already available in upstream Linux. > > > > > > For now, VM migration support is not available for Goldfish RTC device > > > but it will be added later when we implement VM migration for KVM RISC- > > V. > > > > > > Signed-off-by: Anup Patel <anup.patel@wdc.com> > > > --- > > > + > > > +static Property goldfish_rtc_properties[] = { > > > + DEFINE_PROP_UINT64("tick-offset", GoldfishRTCState, tick_offset, 0), > > > + DEFINE_PROP_UINT64("alarm-next", GoldfishRTCState, alarm_next, 0), > > > + DEFINE_PROP_UINT32("alarm-running", GoldfishRTCState, > > alarm_running, 0), > > > + DEFINE_PROP_UINT32("irq-pending", GoldfishRTCState, irq_pending, > > 0), > > > + DEFINE_PROP_UINT32("irq-enabled", GoldfishRTCState, irq_enabled, > > 0), > > > + DEFINE_PROP_END_OF_LIST(), > > > +}; > > > > What are all these properties trying to do ? > > Argh, I forgot to remove these before sending. > > I will drop these in next revision. > > > > > > + > > > +static void goldfish_rtc_class_init(ObjectClass *klass, void *data) { > > > + DeviceClass *dc = DEVICE_CLASS(klass); > > > + dc->props = goldfish_rtc_properties; > > > > Missing reset function. > > Sure, I will add it. I got confused because I saw reset function in HPET > but not in PL031. Yeah, the lack of reset is a bug in the pl031. I did have a draft set of patches where I tried to fix that, but I got stuck because it was a bit unclear what I ought to be resetting. In a real hardware pl031 there is no persistent storage of the RTC value -- it's just a 1Hz counter, really, and guest firmware will write to it on startup (from some other real RTC). In QEMU we use it as a sort of "RTC for a VM", and back it with the QEMU RTC clock, so it doesn't behave quite as the hardware does but more like "view that you'd see from a combination of h/w and firmware behaviour". TLDR: don't use the pl031 code as a good model of how to do an RTC, it has some definite flaws. x86 or ppc RTC handling is probably a better place to look. Another random question -- is there an existing RTC device we could put in the riscv board rather than adding a new model of this goldfish device? Put another way, what are the merits of using goldfish in particular? We should probably document this device in docs/specs since it's a pure-virtual device. (You have a link to the URL for the GOLDFISH-VIRTUAL-HARDWARE.TXT for the android emulator, but that doesn't seem to match the implementation here -- the doc says the alarm registers are non-functional but this code does something with them.) > > If you don't want to implement migration support now you should at least > > put in something that block migration. > > (Personally I prefer to just write the vmstate, it's as easy as writing the code > > to block migrations.) > > Okay, I will add vmstate. > > Is there a way to unit test the vmstate part ? > OR > Are you fine if we test-n-fix it later ? Test and fix later is fine. That said, I've just remembered that for an RTC migration is potentially a tricky area because you typically don't want a vmsave-vmload to give you the same RTC value after load that there was on store, you want it to still be following the host RTC. We got that wrong on the PL031 and had to put in some ugly workarounds to maintain migration compatibility when we fixed it. So maybe it would be better to mark as non-migratable for now. thanks -- PMM
> -----Original Message----- > From: Peter Maydell <peter.maydell@linaro.org> > Sent: Tuesday, September 24, 2019 5:02 PM > To: Anup Patel <Anup.Patel@wdc.com> > Cc: Palmer Dabbelt <palmer@sifive.com>; Alistair Francis > <Alistair.Francis@wdc.com>; Sagar Karandikar <sagark@eecs.berkeley.edu>; > Bastian Koppelmann <kbastian@mail.uni-paderborn.de>; Atish Patra > <Atish.Patra@wdc.com>; qemu-riscv@nongnu.org; qemu- > devel@nongnu.org; Anup Patel <anup@brainfault.org> > Subject: Re: [PATCH 1/2] hw: timer: Add Goldfish RTC device > > On Tue, 24 Sep 2019 at 12:17, Anup Patel <Anup.Patel@wdc.com> wrote: > > > > > > > > > -----Original Message----- > > > From: Peter Maydell <peter.maydell@linaro.org> > > > Sent: Tuesday, September 24, 2019 3:21 PM > > > To: Anup Patel <Anup.Patel@wdc.com> > > > Cc: Palmer Dabbelt <palmer@sifive.com>; Alistair Francis > > > <Alistair.Francis@wdc.com>; Sagar Karandikar > > > <sagark@eecs.berkeley.edu>; Bastian Koppelmann > > > <kbastian@mail.uni-paderborn.de>; Atish Patra > <Atish.Patra@wdc.com>; > > > qemu-riscv@nongnu.org; qemu- devel@nongnu.org; Anup Patel > > > <anup@brainfault.org> > > > Subject: Re: [PATCH 1/2] hw: timer: Add Goldfish RTC device > > > > > > On Tue, 24 Sep 2019 at 09:45, Anup Patel <Anup.Patel@wdc.com> wrote: > > > > > > > > This patch adds model for Google Goldfish virtual platform RTC device. > > > > > > > > We will be adding Goldfish RTC device to the QEMU RISC-V virt > > > > machine for providing real date-time to Guest Linux. The > > > > corresponding Linux driver for Goldfish RTC device is already available in > upstream Linux. > > > > > > > > For now, VM migration support is not available for Goldfish RTC > > > > device but it will be added later when we implement VM migration > > > > for KVM RISC- > > > V. > > > > > > > > Signed-off-by: Anup Patel <anup.patel@wdc.com> > > > > > --- > > > > + > > > > +static Property goldfish_rtc_properties[] = { > > > > + DEFINE_PROP_UINT64("tick-offset", GoldfishRTCState, tick_offset, > 0), > > > > + DEFINE_PROP_UINT64("alarm-next", GoldfishRTCState, alarm_next, > 0), > > > > + DEFINE_PROP_UINT32("alarm-running", GoldfishRTCState, > > > alarm_running, 0), > > > > + DEFINE_PROP_UINT32("irq-pending", GoldfishRTCState, > > > > + irq_pending, > > > 0), > > > > + DEFINE_PROP_UINT32("irq-enabled", GoldfishRTCState, > > > > + irq_enabled, > > > 0), > > > > + DEFINE_PROP_END_OF_LIST(), > > > > +}; > > > > > > What are all these properties trying to do ? > > > > Argh, I forgot to remove these before sending. > > > > I will drop these in next revision. > > > > > > > > > + > > > > +static void goldfish_rtc_class_init(ObjectClass *klass, void *data) { > > > > + DeviceClass *dc = DEVICE_CLASS(klass); > > > > + dc->props = goldfish_rtc_properties; > > > > > > Missing reset function. > > > > Sure, I will add it. I got confused because I saw reset function in > > HPET but not in PL031. > > Yeah, the lack of reset is a bug in the pl031. I did have a draft set of patches > where I tried to fix that, but I got stuck because it was a bit unclear what I > ought to be resetting. In a real hardware pl031 there is no persistent storage > of the RTC value -- it's just a 1Hz counter, really, and guest firmware will write > to it on startup (from some other real RTC). In QEMU we use it as a sort of > "RTC for a VM", and back it with the QEMU RTC clock, so it doesn't behave > quite as the hardware does but more like "view that you'd see from a > combination of h/w and firmware behaviour". > > TLDR: don't use the pl031 code as a good model of how to do an RTC, it has > some definite flaws. x86 or ppc RTC handling is probably a better place to > look. > > Another random question -- is there an existing RTC device we could put in > the riscv board rather than adding a new model of this goldfish device? Put > another way, what are the merits of using goldfish in particular? As of now, there is no RTC device used across SOCs. Atleast nothing that I am aware of. I just wanted a very simple RTC for RISC-V virt machine and found goldfish RTC to be really useful. Although other Goldfish devices are not very attractive and we generally have an equivalent VirtIO device. Ideally, we should have a VirtIO RTC thingy but not sure if RTC fits in VirtIO ring programming model. > > We should probably document this device in docs/specs since it's a pure- > virtual device. (You have a link to the URL for the GOLDFISH-VIRTUAL- > HARDWARE.TXT for the android emulator, but that doesn't seem to match > the implementation here -- the doc says the alarm registers are non- > functional but this code does something with > them.) The implementation matches corresponding driver available in upstream Linux. We should certainly have a more complete document under docs/sepcs (like you suggested) for this device. > > > > If you don't want to implement migration support now you should at > > > least put in something that block migration. > > > (Personally I prefer to just write the vmstate, it's as easy as > > > writing the code to block migrations.) > > > > Okay, I will add vmstate. > > > > Is there a way to unit test the vmstate part ? > > OR > > Are you fine if we test-n-fix it later ? > > Test and fix later is fine. That said, I've just remembered that for an RTC > migration is potentially a tricky area because you typically don't want a > vmsave-vmload to give you the same RTC value after load that there was on > store, you want it to still be following the host RTC. > We got that wrong on the PL031 and had to put in some ugly workarounds to > maintain migration compatibility when we fixed it. > So maybe it would be better to mark as non-migratable for now. Yes, I saw the workaround in PL031 and the backward compatibility bits. This is first version of Goldfish RTC vmstate so we don't need the PL031 backward compatibility bits. I will try implement this in v2 and you can have a look. Regards, Anup
diff --git a/hw/timer/Kconfig b/hw/timer/Kconfig index eefc95f35e..bd1831fbe8 100644 --- a/hw/timer/Kconfig +++ b/hw/timer/Kconfig @@ -58,3 +58,6 @@ config CMSDK_APB_TIMER config CMSDK_APB_DUALTIMER bool select PTIMER + +config GOLDFISH_RTC + bool diff --git a/hw/timer/Makefile.objs b/hw/timer/Makefile.objs index 123d92c969..5dc6f880af 100644 --- a/hw/timer/Makefile.objs +++ b/hw/timer/Makefile.objs @@ -47,3 +47,4 @@ common-obj-$(CONFIG_SUN4V_RTC) += sun4v-rtc.o common-obj-$(CONFIG_CMSDK_APB_TIMER) += cmsdk-apb-timer.o common-obj-$(CONFIG_CMSDK_APB_DUALTIMER) += cmsdk-apb-dualtimer.o common-obj-$(CONFIG_MSF2) += mss-timer.o +common-obj-$(CONFIG_GOLDFISH_RTC) += goldfish_rtc.o diff --git a/hw/timer/goldfish_rtc.c b/hw/timer/goldfish_rtc.c new file mode 100644 index 0000000000..a3349d7cb4 --- /dev/null +++ b/hw/timer/goldfish_rtc.c @@ -0,0 +1,221 @@ +/* + * Goldfish virtual platform RTC + * + * Copyright (C) 2019 Western Digital Corporation or its affiliates. + * + * For more details on Google Goldfish virtual platform refer: + * https://android.googlesource.com/platform/external/qemu/+/master/docs/GOLDFISH-VIRTUAL-HARDWARE.TXT + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2 or later, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#include "qemu/osdep.h" +#include "qemu-common.h" +#include "hw/timer/goldfish_rtc.h" +#include "hw/irq.h" +#include "hw/qdev-properties.h" +#include "hw/sysbus.h" +#include "qemu/timer.h" +#include "sysemu/sysemu.h" +#include "qemu/cutils.h" +#include "qemu/log.h" + +#define RTC_TIME_LOW 0x00 +#define RTC_TIME_HIGH 0x04 +#define RTC_ALARM_LOW 0x08 +#define RTC_ALARM_HIGH 0x0c +#define RTC_IRQ_ENABLED 0x10 +#define RTC_CLEAR_ALARM 0x14 +#define RTC_ALARM_STATUS 0x18 +#define RTC_CLEAR_INTERRUPT 0x1c + +static void goldfish_rtc_update(GoldfishRTCState *s) +{ + qemu_set_irq(s->irq, (s->irq_pending & s->irq_enabled) ? 1 : 0); +} + +static void goldfish_rtc_interrupt(void *opaque) +{ + GoldfishRTCState *s = (GoldfishRTCState *)opaque; + + s->alarm_running = 0; + s->irq_pending = 1; + goldfish_rtc_update(s); +} + +static uint64_t goldfish_rtc_get_count(GoldfishRTCState *s) +{ + return s->tick_offset + (uint64_t)qemu_clock_get_ns(rtc_clock); +} + +static void goldfish_rtc_clear_alarm(GoldfishRTCState *s) +{ + timer_del(s->timer); + s->alarm_running = 0; +} + +static void goldfish_rtc_set_alarm(GoldfishRTCState *s) +{ + uint64_t ticks = goldfish_rtc_get_count(s); + uint64_t event = s->alarm_next; + + if (event <= ticks) { + timer_del(s->timer); + goldfish_rtc_interrupt(s); + } else { + int64_t now = qemu_clock_get_ns(rtc_clock); + timer_mod(s->timer, now + (event - ticks)); + s->alarm_running = 1; + } +} + +static uint64_t goldfish_rtc_read(void *opaque, hwaddr offset, + unsigned size) +{ + GoldfishRTCState *s = (GoldfishRTCState *)opaque; + uint64_t r; + + switch (offset) { + case RTC_TIME_LOW: + r = goldfish_rtc_get_count(s) & 0xffffffff; + break; + case RTC_TIME_HIGH: + r = goldfish_rtc_get_count(s) >> 32; + break; + case RTC_ALARM_LOW: + r = s->alarm_next & 0xffffffff; + break; + case RTC_ALARM_HIGH: + r = s->alarm_next >> 32; + break; + case RTC_IRQ_ENABLED: + r = s->irq_enabled; + break; + case RTC_ALARM_STATUS: + r = s->alarm_running; + break; + default: + qemu_log_mask(LOG_GUEST_ERROR, + "goldfish_rtc_read: Bad offset 0x%x\n", (int)offset); + r = 0; + break; + } + + return r; +} + +static void goldfish_rtc_write(void *opaque, hwaddr offset, + uint64_t value, unsigned size) +{ + uint64_t current_tick, new_tick; + GoldfishRTCState *s = (GoldfishRTCState *)opaque; + + switch (offset) { + case RTC_TIME_LOW: + current_tick = goldfish_rtc_get_count(s); + new_tick = current_tick & (0xffffffffULL << 32); + new_tick |= value; + s->tick_offset += new_tick - current_tick; + break; + case RTC_TIME_HIGH: + current_tick = goldfish_rtc_get_count(s); + new_tick = current_tick & 0xffffffffULL; + new_tick |= (value << 32); + s->tick_offset += new_tick - current_tick; + break; + case RTC_ALARM_LOW: + s->alarm_next &= (0xffffffffULL << 32); + s->alarm_next |= value; + goldfish_rtc_set_alarm(s); + break; + case RTC_ALARM_HIGH: + s->alarm_next &= 0xffffffffULL; + s->alarm_next |= (value << 32); + break; + case RTC_IRQ_ENABLED: + s->irq_enabled = (uint32_t)(value & 0x1); + goldfish_rtc_update(s); + break; + case RTC_CLEAR_ALARM: + goldfish_rtc_clear_alarm(s); + break; + case RTC_CLEAR_INTERRUPT: + s->irq_pending = 0; + goldfish_rtc_update(s); + break; + default: + qemu_log_mask(LOG_GUEST_ERROR, + "goldfish_rtc_write: Bad offset 0x%x\n", (int)offset); + break; + } +} + +static const MemoryRegionOps goldfish_rtc_ops = { + .read = goldfish_rtc_read, + .write = goldfish_rtc_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4 + } +}; + +static void goldfish_rtc_init(Object *obj) +{ + GoldfishRTCState *s = GOLDFISH_RTC(obj); + SysBusDevice *dev = SYS_BUS_DEVICE(obj); + struct tm tm; + + memory_region_init_io(&s->iomem, obj, &goldfish_rtc_ops, s, + "goldfish_rtc", 0x1000); + sysbus_init_mmio(dev, &s->iomem); + + sysbus_init_irq(dev, &s->irq); + + qemu_get_timedate(&tm, 0); + s->tick_offset = mktimegm(&tm); + s->tick_offset *= NANOSECONDS_PER_SECOND; + s->tick_offset -= qemu_clock_get_ns(rtc_clock); + + s->timer = timer_new_ns(rtc_clock, goldfish_rtc_interrupt, s); +} + +static Property goldfish_rtc_properties[] = { + DEFINE_PROP_UINT64("tick-offset", GoldfishRTCState, tick_offset, 0), + DEFINE_PROP_UINT64("alarm-next", GoldfishRTCState, alarm_next, 0), + DEFINE_PROP_UINT32("alarm-running", GoldfishRTCState, alarm_running, 0), + DEFINE_PROP_UINT32("irq-pending", GoldfishRTCState, irq_pending, 0), + DEFINE_PROP_UINT32("irq-enabled", GoldfishRTCState, irq_enabled, 0), + DEFINE_PROP_END_OF_LIST(), +}; + +static void goldfish_rtc_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + dc->props = goldfish_rtc_properties; +} + +static const TypeInfo goldfish_rtc_info = { + .name = TYPE_GOLDFISH_RTC, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(GoldfishRTCState), + .instance_init = goldfish_rtc_init, + .class_init = goldfish_rtc_class_init, +}; + +static void goldfish_rtc_register_types(void) +{ + type_register_static(&goldfish_rtc_info); +} + +type_init(goldfish_rtc_register_types) diff --git a/include/hw/timer/goldfish_rtc.h b/include/hw/timer/goldfish_rtc.h new file mode 100644 index 0000000000..f96a5f5158 --- /dev/null +++ b/include/hw/timer/goldfish_rtc.h @@ -0,0 +1,45 @@ +/* + * Goldfish virtual platform RTC + * + * Copyright (C) 2019 Western Digital Corporation or its affiliates. + * + * For more details on Google Goldfish virtual platform refer: + * https://android.googlesource.com/platform/external/qemu/+/master/docs/GOLDFISH-VIRTUAL-HARDWARE.TXT + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2 or later, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#ifndef HW_TIMER_GOLDFISH_RTC_H +#define HW_TIMER_GOLDFISH_RTC_H + +#include "hw/sysbus.h" + +#define TYPE_GOLDFISH_RTC "goldfish_rtc" +#define GOLDFISH_RTC(obj) \ +OBJECT_CHECK(GoldfishRTCState, (obj), TYPE_GOLDFISH_RTC) + +typedef struct GoldfishRTCState { + SysBusDevice parent_obj; + + MemoryRegion iomem; + QEMUTimer *timer; + qemu_irq irq; + + uint64_t tick_offset; + uint64_t alarm_next; + uint32_t alarm_running; + uint32_t irq_pending; + uint32_t irq_enabled; +} GoldfishRTCState; + +#endif
This patch adds model for Google Goldfish virtual platform RTC device. We will be adding Goldfish RTC device to the QEMU RISC-V virt machine for providing real date-time to Guest Linux. The corresponding Linux driver for Goldfish RTC device is already available in upstream Linux. For now, VM migration support is not available for Goldfish RTC device but it will be added later when we implement VM migration for KVM RISC-V. Signed-off-by: Anup Patel <anup.patel@wdc.com> --- hw/timer/Kconfig | 3 + hw/timer/Makefile.objs | 1 + hw/timer/goldfish_rtc.c | 221 ++++++++++++++++++++++++++++++++ include/hw/timer/goldfish_rtc.h | 45 +++++++ 4 files changed, 270 insertions(+) create mode 100644 hw/timer/goldfish_rtc.c create mode 100644 include/hw/timer/goldfish_rtc.h