Hi Shannon, On 03/08/2016 10:50, Shannon Zhao wrote: > > > On 2016/8/3 15:22, Auger Eric wrote: >> Hi Shannon, >> >> On 03/08/2016 02:56, Shannon Zhao wrote: >>>> Hi Eric, >>>> >>>> On 2016/8/3 2:07, Eric Auger wrote: >>>>>> This patch exposes the GICv3 ITS to the ACPI guest. The ITS structure >>>>>> is added to the MADT table. >>>>>> >>>>>> Signed-off-by: Eric Auger <eric.au...@redhat.com> >>>>>> >>>>>> --- >>>>>> >>>>>> v5: new >>>>>> >>>>>> Tested with Tomasz' kernel series on guest side: >>>>>> - [PATCH V7 0/8] Introduce ACPI world to ITS, >>>>>> https://lkml.org/lkml/2016/6/20/321 >>>>>> - for running PCIe on the guest (virtio-pci-net or vhost-net) >>>>>> the following series is also needed, although not directly ITS: >>>>>> Support for ARM64 ACPI based PCI host controller, >>>>>> https://lwn.net/Articles/690995/ >>>>>> --- >>>>>> hw/arm/virt-acpi-build.c | 7 +++++++ >>>>>> include/hw/acpi/acpi-defs.h | 13 ++++++++++++- >>>>>> 2 files changed, 19 insertions(+), 1 deletion(-) >>>>>> >>>>>> diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c >>>>>> index 28fc59c..6cfedff 100644 >>>>>> --- a/hw/arm/virt-acpi-build.c >>>>>> +++ b/hw/arm/virt-acpi-build.c >>>>>> @@ -546,6 +546,7 @@ build_madt(GArray *table_data, BIOSLinker *linker, >>>>>> VirtGuestInfo *guest_info) >>>>>> } >>>>>> >>>>>> if (guest_info->gic_version == 3) { >>>>>> + AcpiMadtGicIts *gic_its; >>>>>> AcpiMadtGenericRedistributor *gicr = acpi_data_push(table_data, >>>>>> sizeof *gicr); >>>>>> >>>>>> @@ -553,6 +554,12 @@ build_madt(GArray *table_data, BIOSLinker *linker, >>>>>> VirtGuestInfo *guest_info) >>>>>> gicr->length = sizeof(*gicr); >>>>>> gicr->base_address = cpu_to_le64(memmap[VIRT_GIC_REDIST].base); >>>>>> gicr->range_length = cpu_to_le32(memmap[VIRT_GIC_REDIST].size); >>>>>> + >>>>>> + gic_its = acpi_data_push(table_data, sizeof *gic_its); >>>>>> + gic_its->type = ACPI_APIC_ITS_STRUCTURE; >>>>>> + gic_its->length = sizeof(*gic_its); >>>>>> + gic_its->gic_its_id = 0; >>>>>> + gic_its->base_address = cpu_to_le64(memmap[VIRT_GIC_ITS].base); >>>> Since for TCG it doesn't support ITS yet, it should check here using >>>> its_class_name(). >>>> >>>>>> } else { >>>>>> gic_msi = acpi_data_push(table_data, sizeof *gic_msi); >>>>>> gic_msi->type = ACPI_APIC_GENERIC_MSI_FRAME; >>>>>> diff --git a/include/hw/acpi/acpi-defs.h b/include/hw/acpi/acpi-defs.h >>>>>> index 41c1d95..ba3be1e 100644 >>>>>> --- a/include/hw/acpi/acpi-defs.h >>>>>> +++ b/include/hw/acpi/acpi-defs.h >>>>>> @@ -294,7 +294,8 @@ typedef struct AcpiMultipleApicTable >>>>>> AcpiMultipleApicTable; >>>>>> #define ACPI_APIC_GENERIC_DISTRIBUTOR 12 >>>>>> #define ACPI_APIC_GENERIC_MSI_FRAME 13 >>>>>> #define ACPI_APIC_GENERIC_REDISTRIBUTOR 14 >>>>>> -#define ACPI_APIC_RESERVED 15 /* 15 and greater are >>>>>> reserved */ >>>>>> +#define ACPI_APIC_ITS_STRUCTURE 15 >>>> use ACPI_APIC_GENERIC_TRANSLATOR instead. >>>> >>>>>> +#define ACPI_APIC_RESERVED 16 /* 16 and greater are >>>>>> reserved */ >>>>>> >>>>>> /* >>>>>> * MADT sub-structures (Follow MULTIPLE_APIC_DESCRIPTION_TABLE) >>>>>> @@ -386,6 +387,16 @@ struct AcpiMadtGenericMsiFrame { >>>>>> >>>>>> typedef struct AcpiMadtGenericMsiFrame AcpiMadtGenericMsiFrame; >>>>>> >>>>>> +struct AcpiMadtGicIts { >>>>>> + ACPI_SUB_HEADER_DEF >>>>>> + uint16_t reserved; >>>>>> + uint32_t gic_its_id; >>>>>> + uint64_t base_address; >>>>>> + uint32_t reserved2; >>>>>> +} QEMU_PACKED; >>>>>> + >>>>>> +typedef struct AcpiMadtGicIts AcpiMadtGicIts; >>>>>> + >>>> Define it like below to respect the name in linux kernel and also keep >>>> consistent with other structures. >>>> >>>> +struct AcpiMadtGenericTranslator { >>>> + ACPI_SUB_HEADER_DEF >>>> + uint16_t reserved; >>>> + uint32_t translation_id; >>>> + uint64_t base_address; >>>> + uint32_t reserved2; >>>> +} QEMU_PACKED; >>>> + >>>> +typedef struct AcpiMadtGenericTranslator AcpiMadtGenericTranslator; >>>> >>>> BTW, you could have a look at [1] which I sent before. >>>> >>>> [1] https://lists.gnu.org/archive/html/qemu-devel/2015-11/msg06282.html >> Hum sorry, please apologize. I was not aware of those. Do you want to >> respin or shall I respin mine taking into account your comments and >> changing the author. Besides, thanks for the review. > Never mind. I think you can respin yours. OK thanks.
Eric > > Thanks, >
