Re: [PATCH v2 2/3] hw/acpi: Implement the SRAT GI affinity structure

[Alex Williamson <alex.williamson@redhat.com>]

On Sun, 8 Oct 2023 01:47:39 +0530
<ankita@nvidia.com> wrote:

> From: Ankit Agrawal <ankita@nvidia.com>
> 
> ACPI spec provides a scheme to associate "Generic Initiators" [1]
> (e.g. heterogeneous processors and accelerators, GPUs, and I/O devices with
> integrated compute or DMA engines GPUs) with Proximity Domains. This is
> achieved using Generic Initiator Affinity Structure in SRAT. During bootup,
> Linux kernel parse the ACPI SRAT to determine the PXM ids and create a NUMA
> node for each unique PXM ID encountered. Qemu currently do not implement
> these structures while building SRAT.
> 
> Add GI structures while building VM ACPI SRAT. The association between
> devices and PXM are stored using acpi-generic-initiator object. Lookup
> presence of all such objects and use them to build these structures.
> 
> The structure needs a PCI device handle [2] that consists of the device BDF.
> The vfio-pci-nohotplug device corresponding to the acpi-generic-initiator
> object is located to determine the BDF.
> 
> [1] ACPI Spec 6.5, Section 5.2.16.6
> [2] ACPI Spec 6.5, Table 5.66
> 
> Signed-off-by: Ankit Agrawal <ankita@nvidia.com>
> ---
>  hw/acpi/acpi-generic-initiator.c         | 78 ++++++++++++++++++++++++
>  hw/arm/virt-acpi-build.c                 |  3 +
>  hw/vfio/pci.c                            |  2 -
>  hw/vfio/pci.h                            |  2 +
>  include/hw/acpi/acpi-generic-initiator.h | 22 +++++++
>  5 files changed, 105 insertions(+), 2 deletions(-)
> 
> diff --git a/hw/acpi/acpi-generic-initiator.c b/hw/acpi/acpi-generic-initiator.c
> index 6406736090..1ae79639be 100644
> --- a/hw/acpi/acpi-generic-initiator.c
> +++ b/hw/acpi/acpi-generic-initiator.c
> @@ -72,3 +72,81 @@ static void acpi_generic_initiator_class_init(ObjectClass *oc, void *data)
>                                NULL, acpi_generic_initiator_set_node, NULL,
>                                NULL);
>  }
> +
> +static int acpi_generic_initiator_list(Object *obj, void *opaque)
> +{
> +    GSList **list = opaque;
> +
> +    if (object_dynamic_cast(obj, TYPE_ACPI_GENERIC_INITIATOR)) {
> +        *list = g_slist_append(*list, ACPI_GENERIC_INITIATOR(obj));
> +    }
> +
> +    object_child_foreach(obj, acpi_generic_initiator_list, opaque);
> +    return 0;
> +}
> +
> +/*
> + * Identify Generic Initiator objects and link them into the list which is
> + * returned to the caller.
> + *
> + * Note: it is the caller's responsibility to free the list to avoid
> + * memory leak.
> + */
> +static GSList *acpi_generic_initiator_get_list(void)
> +{
> +    GSList *list = NULL;
> +
> +    object_child_foreach(object_get_root(), acpi_generic_initiator_list, &list);
> +    return list;
> +}
> +
> +/*
> + * ACPI spec, Revision 6.5
> + * 5.2.16.6 Generic Initiator Affinity Structure
> + */
> +static void build_srat_generic_initiator_affinity(GArray *table_data, int node,
> +                                                  PCIDeviceHandle *handle,
> +                                                  GenericAffinityFlags flags)
> +{
> +    build_append_int_noprefix(table_data, 5, 1);     /* Type */
> +    build_append_int_noprefix(table_data, 32, 1);    /* Length */
> +    build_append_int_noprefix(table_data, 0, 1);     /* Reserved */
> +    build_append_int_noprefix(table_data, 1, 1);     /* Device Handle Type */
> +    build_append_int_noprefix(table_data, node, 4);  /* Proximity Domain */
> +    build_append_int_noprefix(table_data, handle->segment, 2);
> +    build_append_int_noprefix(table_data, handle->bdf, 2);
> +    build_append_int_noprefix(table_data, handle->res0, 4);
> +    build_append_int_noprefix(table_data, handle->res1, 8);

Why are we storing reserved fields in the PCIDeviceHandle?  This
function is already specific to building a PCI Device Handle, so we
could just loop build_append_byte() with a fixed zero value here.

> +    build_append_int_noprefix(table_data, flags, 4); /* Flags */
> +    build_append_int_noprefix(table_data, 0, 4);     /* Reserved */
> +}
> +
> +void build_srat_generic_initiator(GArray *table_data)
> +{
> +    GSList *gi_list, *list = acpi_generic_initiator_get_list();
> +    for (gi_list = list; gi_list; gi_list = gi_list->next) {
> +        AcpiGenericInitiator *gi = gi_list->data;
> +        Object *o;
> +        int count;
> +
> +        if (gi->node == MAX_NODES) {
> +            continue;
> +        }

Why do we have uninitialized AcpiGenericInitiator objects lingering?

> +
> +        o = object_resolve_path_type(gi->device, TYPE_VFIO_PCI_NOHOTPLUG, NULL);

TYPE_PCI_DEVICE?  Maybe you could check hotpluggable from the device
class, but certainly the generic code should not be dependent on being
a vfio-pci-nohotplug device.  The spec also supports an ACPI object
description, so should this be build_srat_generic_pci_initiator()?


> +        if (!o) {
> +            continue;
> +        }
> +
> +        for (count = 0; count < gi->node_count; count++) {
> +            PCIDeviceHandle dev_handle = {0};
> +            PCIDevice *pci_dev = PCI_DEVICE(o);
> +
> +            dev_handle.bdf = pci_dev->devfn;

Where does the bus part of the bdf get filled in?

> +            build_srat_generic_initiator_affinity(table_data,
> +                                                  gi->node + count, &dev_handle,
> +                                                  GEN_AFFINITY_ENABLED);

Seems like the code that built the AcpiGenericInitiator object should
supply the flags.  In fact the flag GEN_AFFINITY_ENABLED might be a
better indicator to populate the SRAT with the GI than the node value.
Thanks,

Alex

> +        }
> +    }
> +    g_slist_free(list);
> +}
> diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
> index 6b674231c2..7337d8076b 100644
> --- a/hw/arm/virt-acpi-build.c
> +++ b/hw/arm/virt-acpi-build.c
> @@ -58,6 +58,7 @@
>  #include "migration/vmstate.h"
>  #include "hw/acpi/ghes.h"
>  #include "hw/acpi/viot.h"
> +#include "hw/acpi/acpi-generic-initiator.h"
>  
>  #define ARM_SPI_BASE 32
>  
> @@ -558,6 +559,8 @@ build_srat(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
>          }
>      }
>  
> +    build_srat_generic_initiator(table_data);
> +
>      if (ms->nvdimms_state->is_enabled) {
>          nvdimm_build_srat(table_data);
>      }
> diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
> index a205c6b113..5e2a7c650a 100644
> --- a/hw/vfio/pci.c
> +++ b/hw/vfio/pci.c
> @@ -43,8 +43,6 @@
>  #include "migration/blocker.h"
>  #include "migration/qemu-file.h"
>  
> -#define TYPE_VFIO_PCI_NOHOTPLUG "vfio-pci-nohotplug"
> -
>  /* Protected by BQL */
>  static KVMRouteChange vfio_route_change;
>  
> diff --git a/hw/vfio/pci.h b/hw/vfio/pci.h
> index a2771b9ff3..74ac77a260 100644
> --- a/hw/vfio/pci.h
> +++ b/hw/vfio/pci.h
> @@ -118,6 +118,8 @@ typedef struct VFIOMSIXInfo {
>  #define TYPE_VFIO_PCI "vfio-pci"
>  OBJECT_DECLARE_SIMPLE_TYPE(VFIOPCIDevice, VFIO_PCI)
>  
> +#define TYPE_VFIO_PCI_NOHOTPLUG "vfio-pci-nohotplug"
> +
>  struct VFIOPCIDevice {
>      PCIDevice pdev;
>      VFIODevice vbasedev;
> diff --git a/include/hw/acpi/acpi-generic-initiator.h b/include/hw/acpi/acpi-generic-initiator.h
> index e67e6e23b1..e8e2670309 100644
> --- a/include/hw/acpi/acpi-generic-initiator.h
> +++ b/include/hw/acpi/acpi-generic-initiator.h
> @@ -27,4 +27,26 @@ typedef struct AcpiGenericInitiatorClass {
>          ObjectClass parent_class;
>  } AcpiGenericInitiatorClass;
>  
> +/*
> + * ACPI 6.5: Table 5-68 Flags - Generic Initiator
> + */
> +typedef enum {
> +    GEN_AFFINITY_NOFLAGS = 0,
> +    GEN_AFFINITY_ENABLED = (1 << 0),
> +    GEN_AFFINITY_ARCH_TRANS = (1 << 1),
> +} GenericAffinityFlags;
> +
> +/*
> + * ACPI 6.5: Table 5-66 Device Handle - PCI
> + * Device Handle definition
> + */
> +typedef struct PCIDeviceHandle {
> +    uint16_t segment;
> +    uint16_t bdf;
> +    uint32_t res0;
> +    uint64_t res1;
> +} PCIDeviceHandle;
> +
> +void build_srat_generic_initiator(GArray *table_data);
> +
>  #endif