Hi Shameer,
On 3/21/19 11:47 AM, Shameer Kolothum wrote:
> This initializes the GED device with base memory and irq,
> configures ged memory hotplug event and builds the
> corresponding aml code.
>
> GED irq routing to Guest is also enabled.
I am not clear about why would need gsi and virt_gsi_handler for this patch.
Memory hotplug
> should now work.
Time to be confident? ;-)
>
> Signed-off-by: Shameer Kolothum <shameerali.kolothum.th...@huawei.com>
> ---
> hw/acpi/generic_event_device.c | 18 ++++++++++++++++++
> hw/arm/virt-acpi-build.c | 9 +++++++++
> hw/arm/virt.c | 30 +++++++++++++++++++++++++-----
> include/hw/arm/virt.h | 2 ++
> 4 files changed, 54 insertions(+), 5 deletions(-)
>
> diff --git a/hw/acpi/generic_event_device.c b/hw/acpi/generic_event_device.c
> index 9deaa33..02d5e66 100644
> --- a/hw/acpi/generic_event_device.c
> +++ b/hw/acpi/generic_event_device.c
> @@ -200,6 +200,23 @@ static void acpi_ged_event(GEDState *ged_st, uint32_t
> ged_irq_sel)
> qemu_irq_pulse(ged_st->gsi[ged_st->irq]);
> }
>
> +static void acpi_ged_init(MemoryRegion *as, DeviceState *dev, GEDState
> *ged_st)
> +{
> + VirtAcpiState *s = VIRT_ACPI(dev);
> +
> + assert(!ged_io_base && !ged_events && !ged_events_size);
> +
> + ged_io_base = s->ged_base;
> + ged_events = s->ged_events;
> + ged_events_size = s->ged_events_size;
Are you obliged to use those global variables? the ged device handle can
be accessed from vms. build_ged_aml could be passed those values through
args, whose value would be extracted from VirtAcpiState?
> + ged_st->irq = s->ged_irq;
> + ged_st->gsi = s->gsi;
> + qemu_mutex_init(&ged_st->lock);
> + memory_region_init_io(&ged_st->io, OBJECT(dev), &ged_ops, ged_st,
> + "acpi-ged-event", ACPI_GED_REG_LEN);
> + memory_region_add_subregion(as, ged_io_base, &ged_st->io);
This time you are not stuck with the hotplug framework; couldn't you do
here:
memory_region_init_io(&ged_st->io, OBJECT(dev), &ged_ops, ged_st,
> + "acpi-ged-event", ACPI_GED_REG_LEN);
sysbus_init_mmio(SYS_BUS_DEVICE(dev), &ged_st->io);
and in virt create_ged
sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, vms->memmap[VIRT_ACPI_GED].base);
You wouldn't need to use a property for the ged_io_base and use a static
variable?
> +}
> +
> static void virt_device_plug_cb(HotplugHandler *hotplug_dev,
> DeviceState *dev, Error **errp)
> {
> @@ -242,6 +259,7 @@ static void virt_device_realize(DeviceState *dev, Error
> **errp)
> acpi_memory_hotplug_init(get_system_memory(), OBJECT(dev),
> &s->memhp_state,
> s->memhp_base);
> + acpi_ged_init(get_system_memory(), dev, &s->ged_state);
> }
> }
>
> diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
> index 1887531..116e9c9 100644
> --- a/hw/arm/virt-acpi-build.c
> +++ b/hw/arm/virt-acpi-build.c
> @@ -41,6 +41,7 @@
> #include "hw/hw.h"
> #include "hw/acpi/aml-build.h"
> #include "hw/acpi/memory_hotplug.h"
> +#include "hw/acpi/generic_event_device.h"
> #include "hw/pci/pcie_host.h"
> #include "hw/pci/pci.h"
> #include "hw/arm/virt.h"
> @@ -50,6 +51,13 @@
> #define ARM_SPI_BASE 32
> #define ACPI_POWER_BUTTON_DEVICE "PWRB"
>
> +static void acpi_dsdt_add_ged(Aml *scope, VirtMachineState *vms)
> +{
> + int irq = vms->irqmap[VIRT_ACPI_GED] + ARM_SPI_BASE;
> +
> + build_ged_aml(scope, "\\_SB."GED_DEVICE, irq, AML_SYSTEM_MEMORY);
> +}
> +
> static void acpi_dsdt_add_memory_hotplug(Aml *scope, MachineState *ms)
> {
> uint32_t nr_mem = ms->ram_slots;
> @@ -758,6 +766,7 @@ build_dsdt(GArray *table_data, BIOSLinker *linker,
> VirtMachineState *vms)
> */
> scope = aml_scope("\\_SB");
> acpi_dsdt_add_memory_hotplug(scope, MACHINE(vms));
> + acpi_dsdt_add_ged(scope, vms);
> acpi_dsdt_add_cpus(scope, vms->smp_cpus);
> acpi_dsdt_add_uart(scope, &memmap[VIRT_UART],
> (irqmap[VIRT_UART] + ARM_SPI_BASE));
> diff --git a/hw/arm/virt.c b/hw/arm/virt.c
> index b602151..e3f8aa7 100644
> --- a/hw/arm/virt.c
> +++ b/hw/arm/virt.c
> @@ -63,6 +63,7 @@
> #include "target/arm/internals.h"
> #include "hw/mem/pc-dimm.h"
> #include "hw/mem/nvdimm.h"
> +#include "hw/acpi/generic_event_device.h"
>
> #define DEFINE_VIRT_MACHINE_LATEST(major, minor, latest) \
> static void virt_##major##_##minor##_class_init(ObjectClass *oc, \
> @@ -134,6 +135,7 @@ static const MemMapEntry base_memmap[] = {
> [VIRT_SECURE_UART] = { 0x09040000, 0x00001000 },
> [VIRT_SMMU] = { 0x09050000, 0x00020000 },
> [VIRT_PCDIMM_ACPI] = { 0x09070000, 0x00010000 },
> + [VIRT_ACPI_GED] = { 0x09080000, 0x00010000 },
> [VIRT_MMIO] = { 0x0a000000, 0x00000200 },
> /* ...repeating for a total of NUM_VIRTIO_TRANSPORTS, each of that size
> */
> [VIRT_PLATFORM_BUS] = { 0x0c000000, 0x02000000 },
> @@ -169,6 +171,7 @@ static const int a15irqmap[] = {
> [VIRT_PCIE] = 3, /* ... to 6 */
> [VIRT_GPIO] = 7,
> [VIRT_SECURE_UART] = 8,
> + [VIRT_ACPI_GED] = 9,
> [VIRT_MMIO] = 16, /* ...to 16 + NUM_VIRTIO_TRANSPORTS - 1 */
> [VIRT_GIC_V2M] = 48, /* ...to 48 + NUM_GICV2M_SPIS - 1 */
> [VIRT_SMMU] = 74, /* ...to 74 + NUM_SMMU_IRQS - 1 */
> @@ -184,6 +187,13 @@ static const char *valid_cpus[] = {
> ARM_CPU_TYPE_NAME("max"),
> };
>
> +static GedEvent ged_events[] = {
> + {
> + .selector = ACPI_GED_IRQ_SEL_MEM,
> + .event = GED_MEMORY_HOTPLUG,
> + },
> +};
> +
> static bool cpu_type_valid(const char *cpu)
> {
> int i;
> @@ -524,6 +534,11 @@ static DeviceState *create_virt_acpi(VirtMachineState
> *vms)
> dev = qdev_create(NULL, "virt-acpi");
> qdev_prop_set_uint64(dev, "memhp_base",
> vms->memmap[VIRT_PCDIMM_ACPI].base);
> + qdev_prop_set_ptr(dev, "gsi", vms->gsi);
> + qdev_prop_set_uint64(dev, "ged_base", vms->memmap[VIRT_ACPI_GED].base);
> + qdev_prop_set_uint32(dev, "ged_irq", vms->irqmap[VIRT_ACPI_GED]);
can't you avoid using this ged_irq prop as well? How is it different
from the other sysbus devices which do not rely on those props?
> + qdev_prop_set_ptr(dev, "ged_events", ged_events);
> + qdev_prop_set_uint32(dev, "ged_events_size", ARRAY_SIZE(ged_events));
> qdev_init_nofail(dev);
>
> return dev;
> @@ -568,6 +583,12 @@ static void create_v2m(VirtMachineState *vms, qemu_irq
> *pic)
> fdt_add_v2m_gic_node(vms);
> }
>
> +static void virt_gsi_handler(void *opaque, int n, int level)
> +{
> + qemu_irq *gic_irq = opaque;
> + qemu_set_irq(gic_irq[n], level);
> +}
> +
> static void create_gic(VirtMachineState *vms, qemu_irq *pic)
> {
> /* We create a standalone GIC */
> @@ -683,6 +704,8 @@ static void create_gic(VirtMachineState *vms, qemu_irq
> *pic)
> pic[i] = qdev_get_gpio_in(gicdev, i);
> }
>
> + vms->gsi = qemu_allocate_irqs(virt_gsi_handler, pic, NUM_IRQS);
> +
> fdt_add_gic_node(vms);
>
> if (type == 3 && vms->its) {
> @@ -1431,7 +1454,7 @@ static void machvirt_init(MachineState *machine)
> VirtMachineClass *vmc = VIRT_MACHINE_GET_CLASS(machine);
> MachineClass *mc = MACHINE_GET_CLASS(machine);
> const CPUArchIdList *possible_cpus;
> - qemu_irq pic[NUM_IRQS];
> + qemu_irq *pic;
> MemoryRegion *sysmem = get_system_memory();
> MemoryRegion *secure_sysmem = NULL;
> int n, virt_max_cpus;
> @@ -1627,6 +1650,7 @@ static void machvirt_init(MachineState *machine)
>
> create_flash(vms, sysmem, secure_sysmem ? secure_sysmem : sysmem);
>
> + pic = g_new0(qemu_irq, NUM_IRQS);
> create_gic(vms, pic);
>
> fdt_add_pmu_nodes(vms);
> @@ -1842,10 +1866,6 @@ static void virt_memory_pre_plug(HotplugHandler
> *hotplug_dev, DeviceState *dev,
> {
> const bool is_nvdimm = object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM);
>
> - if (dev->hotplugged) {
> - error_setg(errp, "memory hotplug is not supported");
> - }
> -
> if (is_nvdimm) {
> error_setg(errp, "nvdimm is not yet supported");
> return;
> diff --git a/include/hw/arm/virt.h b/include/hw/arm/virt.h
> index 14b2e0a..850296a 100644
> --- a/include/hw/arm/virt.h
> +++ b/include/hw/arm/virt.h
> @@ -78,6 +78,7 @@ enum {
> VIRT_SECURE_UART,
> VIRT_SECURE_MEM,
> VIRT_PCDIMM_ACPI,
> + VIRT_ACPI_GED,
> VIRT_LOWMEMMAP_LAST,
> };
>
> @@ -135,6 +136,7 @@ typedef struct {
> int psci_conduit;
> hwaddr highest_gpa;
> DeviceState *acpi;
> + qemu_irq *gsi;
> } VirtMachineState;
>
> #define VIRT_ECAM_ID(high) (high ? VIRT_HIGH_PCIE_ECAM : VIRT_PCIE_ECAM)
>
Thanks
Eric
