Re: [Qemu-devel] [RFC PATCH v3 13/18] spapr_pci: Enable DDW

[David Gibson <david@gibson.dropbear.id.au>]

[-- Attachment #1: Type: text/plain, Size: 3777 bytes --]

On Fri, Aug 29, 2014 at 08:12:17PM +1000, Alexey Kardashevskiy wrote:
> This implements DDW for emulated PHB.
> 
> This advertises the query/create/remove RTAS tokens in device tree.
> This does not advertise the reset RTAS token though, will be added later.
> 
> The "ddw" property is enabled by default on a PHB but for compatibility
> pseries-2.1 machine disables it.
> 
> Since QEMU does not implement any 64bit DMA capable device, this hack
> has been used to enable 64bit DMA on E1000:
> 
> diff --git a/hw/net/e1000.c b/hw/net/e1000.c
> index 0fc29a0..131f80a 100644
> --- a/hw/net/e1000.c
> +++ b/hw/net/e1000.c
> @@ -240,6 +240,7 @@ static const uint32_t mac_reg_init[] = {
>      [STATUS] =  0x80000000 | E1000_STATUS_GIO_MASTER_ENABLE |
>                  E1000_STATUS_ASDV | E1000_STATUS_MTXCKOK |
>                  E1000_STATUS_SPEED_1000 | E1000_STATUS_FD |
> +                E1000_STATUS_PCIX_MODE |
>                  E1000_STATUS_LU,
>      [MANC] =    E1000_MANC_EN_MNG2HOST | E1000_MANC_RCV_TCO_EN |
>                  E1000_MANC_ARP_EN | E1000_MANC_0298_EN |
> 
> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
> ---
> Changes:
> v3:
> * removed reset
> * windows_num is now 1 or bigger rather than 0-based value and it is only
> changed in PHB code, not in RTAS
> * added page mask check in create()
> 
> v2:
> * tested on hacked emulated E1000
> * implemented DDW reset on the PHB reset
> * spapr_pci_ddw_remove/spapr_pci_ddw_reset are public for reuse by VFIO
> ---
>  hw/ppc/spapr.c              |  9 +++++
>  hw/ppc/spapr_pci.c          | 94 +++++++++++++++++++++++++++++++++++++++++++++
>  include/hw/pci-host/spapr.h |  7 ++++
>  3 files changed, 110 insertions(+)
> 
> diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> index d2d3c27..663cb75 100644
> --- a/hw/ppc/spapr.c
> +++ b/hw/ppc/spapr.c
> @@ -1670,11 +1670,20 @@ static const TypeInfo spapr_machine_info = {
>  static void spapr_machine_2_1_class_init(ObjectClass *oc, void *data)
>  {
>      MachineClass *mc = MACHINE_CLASS(oc);
> +    static GlobalProperty compat_props[] = {
> +        {
> +            .driver   = TYPE_SPAPR_PCI_HOST_BRIDGE,
> +            .property = "ddw",
> +            .value    = stringify(off),
> +        },
> +        { /* end of list */ }
> +    };
>  
>      mc->name = "pseries-2.1";
>      mc->desc = "pSeries Logical Partition (PAPR compliant) v2.1";
>      mc->alias = "pseries";
>      mc->is_default = 1;
> +    mc->compat_props = compat_props;
>  }
>  
>  static const TypeInfo spapr_machine_2_1_info = {
> diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c
> index 2968b39..04ee1dc 100644
> --- a/hw/ppc/spapr_pci.c
> +++ b/hw/ppc/spapr_pci.c
> @@ -470,6 +470,76 @@ static const MemoryRegionOps spapr_msi_ops = {
>  };
>  
>  /*
> + * Dynamic DMA windows
> + */
> +static int spapr_pci_ddw_query(sPAPRPHBState *sphb,
> +                               uint32_t *windows_available,
> +                               uint32_t *page_size_mask)
> +{
> +    *windows_available = 1;
> +    *page_size_mask = DDW_PGSIZE_64K | DDW_PGSIZE_16M;
> +
> +    return 0;
> +}
> +
> +static int spapr_pci_ddw_create(sPAPRPHBState *sphb, uint32_t page_shift,
> +                                uint32_t window_shift, uint32_t liobn,
> +                                sPAPRTCETable **ptcet)
> +{

I think the callbacks should check the ddw_enabled parameter and fail
if ddw is disabled, rather than just relying on the guest not calling
them if they're not advertised.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]