RE: [PATCH V2 08/23] xen/ioreq: Introduce ioreq_params to abstract accesses to arch.hvm.params

[Paul Durrant <xadimgnik@xxxxxxxxx>]

> -----Original Message-----
> From: Oleksandr Tyshchenko <olekstysh@xxxxxxxxx>
> Sent: 15 October 2020 17:44
> To: xen-devel@xxxxxxxxxxxxxxxxxxxx
> Cc: Oleksandr Tyshchenko <oleksandr_tyshchenko@xxxxxxxx>; Paul Durrant 
> <paul@xxxxxxx>; Jan Beulich
> <jbeulich@xxxxxxxx>; Andrew Cooper <andrew.cooper3@xxxxxxxxxx>; Roger Pau 
> Monné
> <roger.pau@xxxxxxxxxx>; Wei Liu <wl@xxxxxxx>; Julien Grall <julien@xxxxxxx>; 
> Stefano Stabellini
> <sstabellini@xxxxxxxxxx>; Julien Grall <julien.grall@xxxxxxx>
> Subject: [PATCH V2 08/23] xen/ioreq: Introduce ioreq_params to abstract 
> accesses to arch.hvm.params
> 
> From: Oleksandr Tyshchenko <oleksandr_tyshchenko@xxxxxxxx>
> 
> We don't want to move HVM params field out of *arch.hvm* in this particular
> case as although it stores a few IOREQ params, it is not a (completely)
> IOREQ stuff and is specific to the architecture. Instead, abstract
> accesses by the proposed macro.
> 
> This is a follow up action to reduce layering violation in the common code.
> 
> Signed-off-by: Oleksandr Tyshchenko <oleksandr_tyshchenko@xxxxxxxx>
> CC: Julien Grall <julien.grall@xxxxxxx>
> 

Keeping the 'legacy' magic page code under an x86 ioreq.c would avoid the need 
for this patch.

  Paul

> ---
> Please note, this is a split/cleanup/hardening of Julien's PoC:
> "Add support for Guest IO forwarding to a device emulator"
> 
> Changes V1 -> V2:
>    - new patch
> ---
>  xen/common/ioreq.c               | 4 ++--
>  xen/include/asm-x86/hvm/domain.h | 2 ++
>  2 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/xen/common/ioreq.c b/xen/common/ioreq.c
> index 7f91bc2..a07f1d7 100644
> --- a/xen/common/ioreq.c
> +++ b/xen/common/ioreq.c
> @@ -223,7 +223,7 @@ static gfn_t hvm_alloc_legacy_ioreq_gfn(struct 
> ioreq_server *s)
>      for ( i = HVM_PARAM_IOREQ_PFN; i <= HVM_PARAM_BUFIOREQ_PFN; i++ )
>      {
>          if ( !test_and_clear_bit(i, &d->ioreq_gfn.legacy_mask) )
> -            return _gfn(d->arch.hvm.params[i]);
> +            return _gfn(ioreq_params(d, i));
>      }
> 
>      return INVALID_GFN;
> @@ -255,7 +255,7 @@ static bool hvm_free_legacy_ioreq_gfn(struct ioreq_server 
> *s,
> 
>      for ( i = HVM_PARAM_IOREQ_PFN; i <= HVM_PARAM_BUFIOREQ_PFN; i++ )
>      {
> -        if ( gfn_eq(gfn, _gfn(d->arch.hvm.params[i])) )
> +        if ( gfn_eq(gfn, _gfn(ioreq_params(d, i))) )
>               break;
>      }
>      if ( i > HVM_PARAM_BUFIOREQ_PFN )
> diff --git a/xen/include/asm-x86/hvm/domain.h 
> b/xen/include/asm-x86/hvm/domain.h
> index 5d60737..c3af339 100644
> --- a/xen/include/asm-x86/hvm/domain.h
> +++ b/xen/include/asm-x86/hvm/domain.h
> @@ -63,6 +63,8 @@ struct hvm_pi_ops {
>      void (*vcpu_block)(struct vcpu *);
>  };
> 
> +#define ioreq_params(d, i) ((d)->arch.hvm.params[i])
> +
>  struct hvm_domain {
>      /* Cached CF8 for guest PCI config cycles */
>      uint32_t                pci_cf8;
> --
> 2.7.4