diff mbox

[RESEND,RFC,6/6] hw/arm/sysbus-fdt: enable amd-xgbe dynamic instantiation

Message ID 1447946528-1533-7-git-send-email-eric.auger@linaro.org
State New
Headers show

Commit Message

Eric Auger Nov. 19, 2015, 3:22 p.m. UTC
This patch allows the instantiation of the vfio-amd-xgbe device
from the QEMU command line (-device vfio-amd-xgbe,host="<device>").

The guest is exposed with a device tree node that combines the description
of both XGBE and PHY (representation supported from 4.2 onwards kernel):
Documentation/devicetree/bindings/net/amd-xgbe.txt.

There are 5 register regions, 6 interrupts including 4 optional
edge-sensitive per-channel interrupts.

Property values are inherited from host device tree. It is mandated
host uses device tree, dtc binary is installed, and the host also uses a
combined XGBE/PHY representation (>= 4.2 host kernel).

2 clock nodes (dma and ptp) also are created. It is mandated those clocks
are fixed on host side.

Signed-off-by: Eric Auger <eric.auger@linaro.org>
---
 hw/arm/sysbus-fdt.c | 171 ++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 167 insertions(+), 4 deletions(-)

Comments

Alex Bennée Nov. 26, 2015, 5:14 p.m. UTC | #1
Eric Auger <eric.auger@linaro.org> writes:

> This patch allows the instantiation of the vfio-amd-xgbe device
> from the QEMU command line (-device vfio-amd-xgbe,host="<device>").
>
> The guest is exposed with a device tree node that combines the description
> of both XGBE and PHY (representation supported from 4.2 onwards kernel):
> Documentation/devicetree/bindings/net/amd-xgbe.txt.
>
> There are 5 register regions, 6 interrupts including 4 optional
> edge-sensitive per-channel interrupts.
>
> Property values are inherited from host device tree. It is mandated
> host uses device tree, dtc binary is installed, and the host also uses a
> combined XGBE/PHY representation (>= 4.2 host kernel).

I might ping you later about this as I'm running 4.3 but can't see the
device. However my board could be a very old rev.

>
> 2 clock nodes (dma and ptp) also are created. It is mandated those clocks
> are fixed on host side.
>
> Signed-off-by: Eric Auger <eric.auger@linaro.org>
> ---
>  hw/arm/sysbus-fdt.c | 171 ++++++++++++++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 167 insertions(+), 4 deletions(-)
>
> diff --git a/hw/arm/sysbus-fdt.c b/hw/arm/sysbus-fdt.c
> index 22e5801..b13d380 100644
> --- a/hw/arm/sysbus-fdt.c
> +++ b/hw/arm/sysbus-fdt.c
> @@ -22,6 +22,7 @@
>   */
>
>  #include <libfdt.h>
> +#include <linux/vfio.h>
>  #include "hw/arm/sysbus-fdt.h"
>  #include "qemu/error-report.h"
>  #include "sysemu/device_tree.h"
> @@ -29,8 +30,10 @@
>  #include "sysemu/sysemu.h"
>  #include "hw/vfio/vfio-platform.h"
>  #include "hw/vfio/vfio-calxeda-xgmac.h"
> +#include "hw/vfio/vfio-amd-xgbe.h"
>  #include "hw/arm/fdt.h"
>
> +
>  /*
>   * internal struct that contains the information to create dynamic
>   * sysbus device node
> @@ -120,10 +123,7 @@ static HostProperty clock_inherited_properties[] = {
>   * host_phandle: phandle of the clock in host device tree
>   * guest_phandle: phandle to assign to the guest node
>   */
> -int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
> -                                uint32_t host_phandle,
> -                                uint32_t guest_phandle);
> -int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
> +static int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
>                                  uint32_t host_phandle,
>                                  uint32_t guest_phandle)
>  {

This looks like it should be squashed with 5/6.

> @@ -166,6 +166,21 @@ out:
>      return ret;
>  }
>
> +/**
> + * sysfs_to_dt_name
> + *
> + * convert the name found in sysfs into the node name
> + * for instance e0900000.xgmac is converted into xgmac@e0900000
> + */
> +static char *sysfs_to_dt_name(const char *sysfs_name)
> +{
> +    gchar **substrings =  g_strsplit(sysfs_name, ".", 2);
> +    char *dt_name;
> +
> +    dt_name = g_strdup_printf("%s@%s", substrings[1], substrings[0]);

You need to g_strfreev(substrings) once you are done.

> +    return dt_name;
> +}
> +
>  /* Device Specific Code */
>
>  /**
> @@ -233,9 +248,157 @@ fail_reg:
>      return ret;
>  }
>
> +
> +/* AMD xgbe properties whose values are copied/pasted from host */
> +static HostProperty amd_xgbe_inherited_properties[] = {
> +    {"compatible", 0},
> +    {"dma-coherent", 1},
> +    {"amd,per-channel-interrupt", 1},
> +    {"phy-mode", 0},
> +    {"mac-address", 1},
> +    {"amd,speed-set", 0},
> +    {"amd,serdes-blwc", 1},
> +    {"amd,serdes-cdr-rate", 1},
> +    {"amd,serdes-pq-skew", 1},
> +    {"amd,serdes-tx-amp", 1},
> +    {"amd,serdes-dfe-tap-config", 1},
> +    {"amd,serdes-dfe-tap-enable", 1},
> +    {"clock-names", 0},
> +};
> +
> +/**
> + * add_amd_xgbe_fdt_node
> + *
> + * Generates the combined xgbe/phy node following kernel >=4.2
> + * binding documentation:
> + * Documentation/devicetree/bindings/net/amd-xgbe.txt:
> + * Also 2 clock nodes are created (dma and ptp)
> + */
> +static int add_amd_xgbe_fdt_node(SysBusDevice *sbdev, void *opaque)
> +{
> +    PlatformBusFDTData *data = opaque;
> +    PlatformBusDevice *pbus = data->pbus;
> +    VFIOPlatformDevice *vdev = VFIO_PLATFORM_DEVICE(sbdev);
> +    VFIODevice *vbasedev = &vdev->vbasedev;
> +    VFIOINTp *intp;
> +    const char *parent_node = data->pbus_node_name;
> +    char *node_path, *nodename, *dt_name;
> +    void *guest_fdt = data->fdt, *host_fdt;
> +    const void *r;
> +    int i, ret = -1, prop_len;
> +    uint32_t *irq_attr, *reg_attr, *host_clock_phandles;
> +    uint64_t mmio_base, irq_number;
> +    uint32_t guest_clock_phandles[2];
> +
> +    host_fdt = load_device_tree_from_sysfs();
> +    if (!host_fdt) {
> +        goto stop;
> +    }
> +    dt_name = sysfs_to_dt_name(vbasedev->name);
> +    ret = qemu_fdt_node_path(host_fdt, dt_name, vdev->compat, &node_path);
> +    g_free(dt_name);
> +
> +    if (ret) {
> +        goto stop;
> +    }
> +
> +    /* generate nodes for DMA_CLK and PTP_CLK */
> +    r = qemu_fdt_getprop(host_fdt, node_path, "clocks", &prop_len);
> +    if (prop_len != 8) {
> +        goto stop;
> +    }
> +    host_clock_phandles = (uint32_t *)r;
> +    host_clock_phandles[0] = be32_to_cpu(host_clock_phandles[0]);
> +    host_clock_phandles[1] = be32_to_cpu(host_clock_phandles[1]);
> +    guest_clock_phandles[0] = qemu_fdt_alloc_phandle(guest_fdt);
> +    guest_clock_phandles[1] = qemu_fdt_alloc_phandle(guest_fdt);
> +
> +    ret = fdt_build_clock_node(host_fdt, guest_fdt,
> +                               host_clock_phandles[0],
> +                               guest_clock_phandles[0]);

Hmm a little comment about the endianess here would be worthwhile. It
seems odd to have a fdt function return stuff in be32 form but need it
in host cpu form in other calls.

As you only use host_clock_phandles[] once you could push the endian
conversion to the call rather than the in-place swizzle.:

        ret = fdt_build_clock_node(host_fdt, guest_fdt,
                                   be32_to_cpu(host_clock_phandles[0]),
                                   guest_clock_phandles[0]);


> +    if (ret) {
> +        goto stop;
> +    }
> +
> +    ret = fdt_build_clock_node(host_fdt, guest_fdt,
> +                               host_clock_phandles[1],
> +                               guest_clock_phandles[1]);
> +    if (ret) {
> +        goto stop;
> +    }
> +
> +    /* combined XGBE/PHY node */
> +    mmio_base = platform_bus_get_mmio_addr(pbus, sbdev, 0);
> +    nodename = g_strdup_printf("%s/%s@%" PRIx64, parent_node,
> +                               vbasedev->name, mmio_base);
> +    qemu_fdt_add_subnode(guest_fdt, nodename);
> +
> +    inherit_properties(amd_xgbe_inherited_properties,
> +                       ARRAY_SIZE(amd_xgbe_inherited_properties),
> +                       host_fdt, guest_fdt,
> +                       node_path, nodename);
> +
> +    qemu_fdt_setprop_cells(guest_fdt, nodename, "clocks",
> +                           guest_clock_phandles[0],
> +                           guest_clock_phandles[1]);
> +
> +    reg_attr = g_new(uint32_t, vbasedev->num_regions * 2);
> +    for (i = 0; i < vbasedev->num_regions; i++) {
> +        mmio_base = platform_bus_get_mmio_addr(pbus, sbdev, i);
> +        reg_attr[2 * i] = cpu_to_be32(mmio_base);
> +        reg_attr[2 * i + 1] = cpu_to_be32(
> +                                memory_region_size(&vdev->regions[i]->mem));
> +    }
> +    ret = qemu_fdt_setprop(guest_fdt, nodename, "reg", reg_attr,
> +                           vbasedev->num_regions * 2 * sizeof(uint32_t));
> +    if (ret) {
> +        error_report("could not set reg property of node %s", nodename);
> +        goto fail_reg;
> +    }
> +
> +    irq_attr = g_new(uint32_t, vbasedev->num_irqs * 3);
> +    for (i = 0; i < vbasedev->num_irqs; i++) {
> +        irq_number = platform_bus_get_irqn(pbus, sbdev , i)
> +                         + data->irq_start;
> +        irq_attr[3 * i] = cpu_to_be32(GIC_FDT_IRQ_TYPE_SPI);
> +        irq_attr[3 * i + 1] = cpu_to_be32(irq_number);
> +        /*
> +          * General device interrupt and PCS auto-negociation interrupts are
> +          * level-sensitive while the 4 per-channel interrupts are edge
> +          * sensitive
> +          */
> +        QLIST_FOREACH(intp, &vdev->intp_list, next) {
> +            if (intp->pin == i) {
> +                break;
> +            }
> +        }
> +        if (intp->flags & VFIO_IRQ_INFO_AUTOMASKED) {
> +            irq_attr[3 * i + 2] = cpu_to_be32(GIC_FDT_IRQ_FLAGS_LEVEL_HI);
> +        } else {
> +            irq_attr[3 * i + 2] = cpu_to_be32(GIC_FDT_IRQ_FLAGS_EDGE_LO_HI);
> +        }
> +    }
> +    ret = qemu_fdt_setprop(guest_fdt, nodename, "interrupts",
> +                     irq_attr, vbasedev->num_irqs * 3 * sizeof(uint32_t));
> +    if (ret) {
> +        error_report("could not set interrupts property of node %s",
> +                     nodename);
> +    }
> +    g_free(host_fdt);
> +    g_free(node_path);
> +    g_free(irq_attr);
> +fail_reg:
> +    g_free(reg_attr);
> +    g_free(nodename);
> +    return ret;
> +stop:
> +    exit(1);
> +}
> +
>  /* list of supported dynamic sysbus devices */
>  static const NodeCreationPair add_fdt_node_functions[] = {
>      {TYPE_VFIO_CALXEDA_XGMAC, add_calxeda_midway_xgmac_fdt_node},
> +    {TYPE_VFIO_AMD_XGBE, add_amd_xgbe_fdt_node},
>      {"", NULL}, /* last element */
>  };


--
Alex Bennée
Eric Auger Dec. 3, 2015, 4:17 p.m. UTC | #2
On 11/26/2015 06:14 PM, Alex Bennée wrote:
> 
> Eric Auger <eric.auger@linaro.org> writes:
> 
>> This patch allows the instantiation of the vfio-amd-xgbe device
>> from the QEMU command line (-device vfio-amd-xgbe,host="<device>").
>>
>> The guest is exposed with a device tree node that combines the description
>> of both XGBE and PHY (representation supported from 4.2 onwards kernel):
>> Documentation/devicetree/bindings/net/amd-xgbe.txt.
>>
>> There are 5 register regions, 6 interrupts including 4 optional
>> edge-sensitive per-channel interrupts.
>>
>> Property values are inherited from host device tree. It is mandated
>> host uses device tree, dtc binary is installed, and the host also uses a
>> combined XGBE/PHY representation (>= 4.2 host kernel).
> 
> I might ping you later about this as I'm running 4.3 but can't see the
> device. However my board could be a very old rev.
On my side it is in the smb directory.

/sys/firmware/devicetree/base/smb/xgmac@*
> 
>>
>> 2 clock nodes (dma and ptp) also are created. It is mandated those clocks
>> are fixed on host side.
>>
>> Signed-off-by: Eric Auger <eric.auger@linaro.org>
>> ---
>>  hw/arm/sysbus-fdt.c | 171 ++++++++++++++++++++++++++++++++++++++++++++++++++--
>>  1 file changed, 167 insertions(+), 4 deletions(-)
>>
>> diff --git a/hw/arm/sysbus-fdt.c b/hw/arm/sysbus-fdt.c
>> index 22e5801..b13d380 100644
>> --- a/hw/arm/sysbus-fdt.c
>> +++ b/hw/arm/sysbus-fdt.c
>> @@ -22,6 +22,7 @@
>>   */
>>
>>  #include <libfdt.h>
>> +#include <linux/vfio.h>
>>  #include "hw/arm/sysbus-fdt.h"
>>  #include "qemu/error-report.h"
>>  #include "sysemu/device_tree.h"
>> @@ -29,8 +30,10 @@
>>  #include "sysemu/sysemu.h"
>>  #include "hw/vfio/vfio-platform.h"
>>  #include "hw/vfio/vfio-calxeda-xgmac.h"
>> +#include "hw/vfio/vfio-amd-xgbe.h"
>>  #include "hw/arm/fdt.h"
>>
>> +
>>  /*
>>   * internal struct that contains the information to create dynamic
>>   * sysbus device node
>> @@ -120,10 +123,7 @@ static HostProperty clock_inherited_properties[] = {
>>   * host_phandle: phandle of the clock in host device tree
>>   * guest_phandle: phandle to assign to the guest node
>>   */
>> -int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
>> -                                uint32_t host_phandle,
>> -                                uint32_t guest_phandle);
>> -int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
>> +static int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
>>                                  uint32_t host_phandle,
>>                                  uint32_t guest_phandle)
>>  {
> 
> This looks like it should be squashed with 5/6.
Sure ;-)
> 
>> @@ -166,6 +166,21 @@ out:
>>      return ret;
>>  }
>>
>> +/**
>> + * sysfs_to_dt_name
>> + *
>> + * convert the name found in sysfs into the node name
>> + * for instance e0900000.xgmac is converted into xgmac@e0900000
>> + */
>> +static char *sysfs_to_dt_name(const char *sysfs_name)
>> +{
>> +    gchar **substrings =  g_strsplit(sysfs_name, ".", 2);
>> +    char *dt_name;
>> +
>> +    dt_name = g_strdup_printf("%s@%s", substrings[1], substrings[0]);
> 
> You need to g_strfreev(substrings) once you are done.
sure!
> 
>> +    return dt_name;
>> +}
>> +
>>  /* Device Specific Code */
>>
>>  /**
>> @@ -233,9 +248,157 @@ fail_reg:
>>      return ret;
>>  }
>>
>> +
>> +/* AMD xgbe properties whose values are copied/pasted from host */
>> +static HostProperty amd_xgbe_inherited_properties[] = {
>> +    {"compatible", 0},
>> +    {"dma-coherent", 1},
>> +    {"amd,per-channel-interrupt", 1},
>> +    {"phy-mode", 0},
>> +    {"mac-address", 1},
>> +    {"amd,speed-set", 0},
>> +    {"amd,serdes-blwc", 1},
>> +    {"amd,serdes-cdr-rate", 1},
>> +    {"amd,serdes-pq-skew", 1},
>> +    {"amd,serdes-tx-amp", 1},
>> +    {"amd,serdes-dfe-tap-config", 1},
>> +    {"amd,serdes-dfe-tap-enable", 1},
>> +    {"clock-names", 0},
>> +};
>> +
>> +/**
>> + * add_amd_xgbe_fdt_node
>> + *
>> + * Generates the combined xgbe/phy node following kernel >=4.2
>> + * binding documentation:
>> + * Documentation/devicetree/bindings/net/amd-xgbe.txt:
>> + * Also 2 clock nodes are created (dma and ptp)
>> + */
>> +static int add_amd_xgbe_fdt_node(SysBusDevice *sbdev, void *opaque)
>> +{
>> +    PlatformBusFDTData *data = opaque;
>> +    PlatformBusDevice *pbus = data->pbus;
>> +    VFIOPlatformDevice *vdev = VFIO_PLATFORM_DEVICE(sbdev);
>> +    VFIODevice *vbasedev = &vdev->vbasedev;
>> +    VFIOINTp *intp;
>> +    const char *parent_node = data->pbus_node_name;
>> +    char *node_path, *nodename, *dt_name;
>> +    void *guest_fdt = data->fdt, *host_fdt;
>> +    const void *r;
>> +    int i, ret = -1, prop_len;
>> +    uint32_t *irq_attr, *reg_attr, *host_clock_phandles;
>> +    uint64_t mmio_base, irq_number;
>> +    uint32_t guest_clock_phandles[2];
>> +
>> +    host_fdt = load_device_tree_from_sysfs();
>> +    if (!host_fdt) {
>> +        goto stop;
>> +    }
>> +    dt_name = sysfs_to_dt_name(vbasedev->name);
>> +    ret = qemu_fdt_node_path(host_fdt, dt_name, vdev->compat, &node_path);
>> +    g_free(dt_name);
>> +
>> +    if (ret) {
>> +        goto stop;
>> +    }
>> +
>> +    /* generate nodes for DMA_CLK and PTP_CLK */
>> +    r = qemu_fdt_getprop(host_fdt, node_path, "clocks", &prop_len);
>> +    if (prop_len != 8) {
>> +        goto stop;
>> +    }
>> +    host_clock_phandles = (uint32_t *)r;
>> +    host_clock_phandles[0] = be32_to_cpu(host_clock_phandles[0]);
>> +    host_clock_phandles[1] = be32_to_cpu(host_clock_phandles[1]);
>> +    guest_clock_phandles[0] = qemu_fdt_alloc_phandle(guest_fdt);
>> +    guest_clock_phandles[1] = qemu_fdt_alloc_phandle(guest_fdt);
>> +
>> +    ret = fdt_build_clock_node(host_fdt, guest_fdt,
>> +                               host_clock_phandles[0],
>> +                               guest_clock_phandles[0]);
> 
> Hmm a little comment about the endianess here would be worthwhile. It
> seems odd to have a fdt function return stuff in be32 form but need it
> in host cpu form in other calls.
> 
> As you only use host_clock_phandles[] once you could push the endian
> conversion to the call rather than the in-place swizzle.:
OK

Thanks!

Eric
> 
>         ret = fdt_build_clock_node(host_fdt, guest_fdt,
>                                    be32_to_cpu(host_clock_phandles[0]),
>                                    guest_clock_phandles[0]);
> 
> 
>> +    if (ret) {
>> +        goto stop;
>> +    }
>> +
>> +    ret = fdt_build_clock_node(host_fdt, guest_fdt,
>> +                               host_clock_phandles[1],
>> +                               guest_clock_phandles[1]);
>> +    if (ret) {
>> +        goto stop;
>> +    }
>> +
>> +    /* combined XGBE/PHY node */
>> +    mmio_base = platform_bus_get_mmio_addr(pbus, sbdev, 0);
>> +    nodename = g_strdup_printf("%s/%s@%" PRIx64, parent_node,
>> +                               vbasedev->name, mmio_base);
>> +    qemu_fdt_add_subnode(guest_fdt, nodename);
>> +
>> +    inherit_properties(amd_xgbe_inherited_properties,
>> +                       ARRAY_SIZE(amd_xgbe_inherited_properties),
>> +                       host_fdt, guest_fdt,
>> +                       node_path, nodename);
>> +
>> +    qemu_fdt_setprop_cells(guest_fdt, nodename, "clocks",
>> +                           guest_clock_phandles[0],
>> +                           guest_clock_phandles[1]);
>> +
>> +    reg_attr = g_new(uint32_t, vbasedev->num_regions * 2);
>> +    for (i = 0; i < vbasedev->num_regions; i++) {
>> +        mmio_base = platform_bus_get_mmio_addr(pbus, sbdev, i);
>> +        reg_attr[2 * i] = cpu_to_be32(mmio_base);
>> +        reg_attr[2 * i + 1] = cpu_to_be32(
>> +                                memory_region_size(&vdev->regions[i]->mem));
>> +    }
>> +    ret = qemu_fdt_setprop(guest_fdt, nodename, "reg", reg_attr,
>> +                           vbasedev->num_regions * 2 * sizeof(uint32_t));
>> +    if (ret) {
>> +        error_report("could not set reg property of node %s", nodename);
>> +        goto fail_reg;
>> +    }
>> +
>> +    irq_attr = g_new(uint32_t, vbasedev->num_irqs * 3);
>> +    for (i = 0; i < vbasedev->num_irqs; i++) {
>> +        irq_number = platform_bus_get_irqn(pbus, sbdev , i)
>> +                         + data->irq_start;
>> +        irq_attr[3 * i] = cpu_to_be32(GIC_FDT_IRQ_TYPE_SPI);
>> +        irq_attr[3 * i + 1] = cpu_to_be32(irq_number);
>> +        /*
>> +          * General device interrupt and PCS auto-negociation interrupts are
>> +          * level-sensitive while the 4 per-channel interrupts are edge
>> +          * sensitive
>> +          */
>> +        QLIST_FOREACH(intp, &vdev->intp_list, next) {
>> +            if (intp->pin == i) {
>> +                break;
>> +            }
>> +        }
>> +        if (intp->flags & VFIO_IRQ_INFO_AUTOMASKED) {
>> +            irq_attr[3 * i + 2] = cpu_to_be32(GIC_FDT_IRQ_FLAGS_LEVEL_HI);
>> +        } else {
>> +            irq_attr[3 * i + 2] = cpu_to_be32(GIC_FDT_IRQ_FLAGS_EDGE_LO_HI);
>> +        }
>> +    }
>> +    ret = qemu_fdt_setprop(guest_fdt, nodename, "interrupts",
>> +                     irq_attr, vbasedev->num_irqs * 3 * sizeof(uint32_t));
>> +    if (ret) {
>> +        error_report("could not set interrupts property of node %s",
>> +                     nodename);
>> +    }
>> +    g_free(host_fdt);
>> +    g_free(node_path);
>> +    g_free(irq_attr);
>> +fail_reg:
>> +    g_free(reg_attr);
>> +    g_free(nodename);
>> +    return ret;
>> +stop:
>> +    exit(1);
>> +}
>> +
>>  /* list of supported dynamic sysbus devices */
>>  static const NodeCreationPair add_fdt_node_functions[] = {
>>      {TYPE_VFIO_CALXEDA_XGMAC, add_calxeda_midway_xgmac_fdt_node},
>> +    {TYPE_VFIO_AMD_XGBE, add_amd_xgbe_fdt_node},
>>      {"", NULL}, /* last element */
>>  };
> 
> 
> --
> Alex Bennée
>
diff mbox

Patch

diff --git a/hw/arm/sysbus-fdt.c b/hw/arm/sysbus-fdt.c
index 22e5801..b13d380 100644
--- a/hw/arm/sysbus-fdt.c
+++ b/hw/arm/sysbus-fdt.c
@@ -22,6 +22,7 @@ 
  */
 
 #include <libfdt.h>
+#include <linux/vfio.h>
 #include "hw/arm/sysbus-fdt.h"
 #include "qemu/error-report.h"
 #include "sysemu/device_tree.h"
@@ -29,8 +30,10 @@ 
 #include "sysemu/sysemu.h"
 #include "hw/vfio/vfio-platform.h"
 #include "hw/vfio/vfio-calxeda-xgmac.h"
+#include "hw/vfio/vfio-amd-xgbe.h"
 #include "hw/arm/fdt.h"
 
+
 /*
  * internal struct that contains the information to create dynamic
  * sysbus device node
@@ -120,10 +123,7 @@  static HostProperty clock_inherited_properties[] = {
  * host_phandle: phandle of the clock in host device tree
  * guest_phandle: phandle to assign to the guest node
  */
-int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
-                                uint32_t host_phandle,
-                                uint32_t guest_phandle);
-int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
+static int fdt_build_clock_node(void *host_fdt, void *guest_fdt,
                                 uint32_t host_phandle,
                                 uint32_t guest_phandle)
 {
@@ -166,6 +166,21 @@  out:
     return ret;
 }
 
+/**
+ * sysfs_to_dt_name
+ *
+ * convert the name found in sysfs into the node name
+ * for instance e0900000.xgmac is converted into xgmac@e0900000
+ */
+static char *sysfs_to_dt_name(const char *sysfs_name)
+{
+    gchar **substrings =  g_strsplit(sysfs_name, ".", 2);
+    char *dt_name;
+
+    dt_name = g_strdup_printf("%s@%s", substrings[1], substrings[0]);
+    return dt_name;
+}
+
 /* Device Specific Code */
 
 /**
@@ -233,9 +248,157 @@  fail_reg:
     return ret;
 }
 
+
+/* AMD xgbe properties whose values are copied/pasted from host */
+static HostProperty amd_xgbe_inherited_properties[] = {
+    {"compatible", 0},
+    {"dma-coherent", 1},
+    {"amd,per-channel-interrupt", 1},
+    {"phy-mode", 0},
+    {"mac-address", 1},
+    {"amd,speed-set", 0},
+    {"amd,serdes-blwc", 1},
+    {"amd,serdes-cdr-rate", 1},
+    {"amd,serdes-pq-skew", 1},
+    {"amd,serdes-tx-amp", 1},
+    {"amd,serdes-dfe-tap-config", 1},
+    {"amd,serdes-dfe-tap-enable", 1},
+    {"clock-names", 0},
+};
+
+/**
+ * add_amd_xgbe_fdt_node
+ *
+ * Generates the combined xgbe/phy node following kernel >=4.2
+ * binding documentation:
+ * Documentation/devicetree/bindings/net/amd-xgbe.txt:
+ * Also 2 clock nodes are created (dma and ptp)
+ */
+static int add_amd_xgbe_fdt_node(SysBusDevice *sbdev, void *opaque)
+{
+    PlatformBusFDTData *data = opaque;
+    PlatformBusDevice *pbus = data->pbus;
+    VFIOPlatformDevice *vdev = VFIO_PLATFORM_DEVICE(sbdev);
+    VFIODevice *vbasedev = &vdev->vbasedev;
+    VFIOINTp *intp;
+    const char *parent_node = data->pbus_node_name;
+    char *node_path, *nodename, *dt_name;
+    void *guest_fdt = data->fdt, *host_fdt;
+    const void *r;
+    int i, ret = -1, prop_len;
+    uint32_t *irq_attr, *reg_attr, *host_clock_phandles;
+    uint64_t mmio_base, irq_number;
+    uint32_t guest_clock_phandles[2];
+
+    host_fdt = load_device_tree_from_sysfs();
+    if (!host_fdt) {
+        goto stop;
+    }
+    dt_name = sysfs_to_dt_name(vbasedev->name);
+    ret = qemu_fdt_node_path(host_fdt, dt_name, vdev->compat, &node_path);
+    g_free(dt_name);
+
+    if (ret) {
+        goto stop;
+    }
+
+    /* generate nodes for DMA_CLK and PTP_CLK */
+    r = qemu_fdt_getprop(host_fdt, node_path, "clocks", &prop_len);
+    if (prop_len != 8) {
+        goto stop;
+    }
+    host_clock_phandles = (uint32_t *)r;
+    host_clock_phandles[0] = be32_to_cpu(host_clock_phandles[0]);
+    host_clock_phandles[1] = be32_to_cpu(host_clock_phandles[1]);
+    guest_clock_phandles[0] = qemu_fdt_alloc_phandle(guest_fdt);
+    guest_clock_phandles[1] = qemu_fdt_alloc_phandle(guest_fdt);
+
+    ret = fdt_build_clock_node(host_fdt, guest_fdt,
+                               host_clock_phandles[0],
+                               guest_clock_phandles[0]);
+    if (ret) {
+        goto stop;
+    }
+
+    ret = fdt_build_clock_node(host_fdt, guest_fdt,
+                               host_clock_phandles[1],
+                               guest_clock_phandles[1]);
+    if (ret) {
+        goto stop;
+    }
+
+    /* combined XGBE/PHY node */
+    mmio_base = platform_bus_get_mmio_addr(pbus, sbdev, 0);
+    nodename = g_strdup_printf("%s/%s@%" PRIx64, parent_node,
+                               vbasedev->name, mmio_base);
+    qemu_fdt_add_subnode(guest_fdt, nodename);
+
+    inherit_properties(amd_xgbe_inherited_properties,
+                       ARRAY_SIZE(amd_xgbe_inherited_properties),
+                       host_fdt, guest_fdt,
+                       node_path, nodename);
+
+    qemu_fdt_setprop_cells(guest_fdt, nodename, "clocks",
+                           guest_clock_phandles[0],
+                           guest_clock_phandles[1]);
+
+    reg_attr = g_new(uint32_t, vbasedev->num_regions * 2);
+    for (i = 0; i < vbasedev->num_regions; i++) {
+        mmio_base = platform_bus_get_mmio_addr(pbus, sbdev, i);
+        reg_attr[2 * i] = cpu_to_be32(mmio_base);
+        reg_attr[2 * i + 1] = cpu_to_be32(
+                                memory_region_size(&vdev->regions[i]->mem));
+    }
+    ret = qemu_fdt_setprop(guest_fdt, nodename, "reg", reg_attr,
+                           vbasedev->num_regions * 2 * sizeof(uint32_t));
+    if (ret) {
+        error_report("could not set reg property of node %s", nodename);
+        goto fail_reg;
+    }
+
+    irq_attr = g_new(uint32_t, vbasedev->num_irqs * 3);
+    for (i = 0; i < vbasedev->num_irqs; i++) {
+        irq_number = platform_bus_get_irqn(pbus, sbdev , i)
+                         + data->irq_start;
+        irq_attr[3 * i] = cpu_to_be32(GIC_FDT_IRQ_TYPE_SPI);
+        irq_attr[3 * i + 1] = cpu_to_be32(irq_number);
+        /*
+          * General device interrupt and PCS auto-negociation interrupts are
+          * level-sensitive while the 4 per-channel interrupts are edge
+          * sensitive
+          */
+        QLIST_FOREACH(intp, &vdev->intp_list, next) {
+            if (intp->pin == i) {
+                break;
+            }
+        }
+        if (intp->flags & VFIO_IRQ_INFO_AUTOMASKED) {
+            irq_attr[3 * i + 2] = cpu_to_be32(GIC_FDT_IRQ_FLAGS_LEVEL_HI);
+        } else {
+            irq_attr[3 * i + 2] = cpu_to_be32(GIC_FDT_IRQ_FLAGS_EDGE_LO_HI);
+        }
+    }
+    ret = qemu_fdt_setprop(guest_fdt, nodename, "interrupts",
+                     irq_attr, vbasedev->num_irqs * 3 * sizeof(uint32_t));
+    if (ret) {
+        error_report("could not set interrupts property of node %s",
+                     nodename);
+    }
+    g_free(host_fdt);
+    g_free(node_path);
+    g_free(irq_attr);
+fail_reg:
+    g_free(reg_attr);
+    g_free(nodename);
+    return ret;
+stop:
+    exit(1);
+}
+
 /* list of supported dynamic sysbus devices */
 static const NodeCreationPair add_fdt_node_functions[] = {
     {TYPE_VFIO_CALXEDA_XGMAC, add_calxeda_midway_xgmac_fdt_node},
+    {TYPE_VFIO_AMD_XGBE, add_amd_xgbe_fdt_node},
     {"", NULL}, /* last element */
 };