Re: [Xen-devel] [PATCH v2 REPOST 11/12] x86/hvm/ioreq: defer mapping gfns until they are actually requsted

[Roger Pau Monné <roger.pau@xxxxxxxxxx>]

On Tue, Aug 22, 2017 at 03:51:05PM +0100, Paul Durrant wrote:
> A subsequent patch will introduce a new scheme to allow an emulator to
> map IOREQ server pages directly from Xen rather than the guest P2M.
> 
> This patch lays the groundwork for that change by deferring mapping of
> gfns until their values are requested by an emulator. To that end, the
> pad field of the xen_dm_op_get_ioreq_server_info structure is re-purposed
> to a flags field and new flag, XEN_DMOP_no_gfns, defined which modifies the
> behaviour of XEN_DMOP_get_ioreq_server_info to allow the caller to avoid
> requesting the gfn values.
> 
> Signed-off-by: Paul Durrant <paul.durrant@xxxxxxxxxx>
> ---
> Cc: Ian Jackson <ian.jackson@xxxxxxxxxxxxx>
> Cc: Wei Liu <wei.liu2@xxxxxxxxxx>
> Cc: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
> Cc: George Dunlap <George.Dunlap@xxxxxxxxxxxxx>
> Cc: Jan Beulich <jbeulich@xxxxxxxx>
> Cc: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>
> Cc: Stefano Stabellini <sstabellini@xxxxxxxxxx>
> Cc: Tim Deegan <tim@xxxxxxx>
> ---
>  tools/libs/devicemodel/core.c                   |  8 +++++
>  tools/libs/devicemodel/include/xendevicemodel.h |  6 ++--
>  xen/arch/x86/hvm/dm.c                           |  9 +++--
>  xen/arch/x86/hvm/ioreq.c                        | 44 
> ++++++++++++++-----------
>  xen/include/asm-x86/hvm/domain.h                |  2 +-
>  xen/include/public/hvm/dm_op.h                  | 32 ++++++++++--------
>  6 files changed, 61 insertions(+), 40 deletions(-)
> 
> diff --git a/tools/libs/devicemodel/core.c b/tools/libs/devicemodel/core.c
> index fcb260d29b..907c894e77 100644
> --- a/tools/libs/devicemodel/core.c
> +++ b/tools/libs/devicemodel/core.c
> @@ -188,6 +188,14 @@ int xendevicemodel_get_ioreq_server_info(
>  
>      data->id = id;
>  
> +    /*
> +     * If the caller is not requesting gfn values then instruct the
> +     * hypercall not to retrieve them as this may cause them to be
> +     * mapped.
> +     */
> +    if (!ioreq_gfn && !bufioreq_gfn)
> +        data->flags = XEN_DMOP_no_gfns;

Since this is memset to 0 I would use |= here, in case someone adds a
new flag further up.

Also, seeing the code below, don't you need to retry on error in case
you are dealing with an old hypervisor version? (that will choke on pad
being set).

> diff --git a/xen/include/asm-x86/hvm/domain.h 
> b/xen/include/asm-x86/hvm/domain.h
> index 7b93d10209..b8bcd559a5 100644
> --- a/xen/include/asm-x86/hvm/domain.h
> +++ b/xen/include/asm-x86/hvm/domain.h
> @@ -70,7 +70,7 @@ struct hvm_ioreq_server {
>      evtchn_port_t          bufioreq_evtchn;
>      struct rangeset        *range[NR_IO_RANGE_TYPES];

I would place bufioreq_handling here in order to have a more compact
structure layout.

>      bool                   enabled;
> -    bool                   bufioreq_atomic;
> +    int                    bufioreq_handling;
>      bool                   is_default;
>  };
>  
> diff --git a/xen/include/public/hvm/dm_op.h b/xen/include/public/hvm/dm_op.h
> index 6bbab5fca3..9677bd74e7 100644
> --- a/xen/include/public/hvm/dm_op.h
> +++ b/xen/include/public/hvm/dm_op.h
> @@ -79,28 +79,34 @@ struct xen_dm_op_create_ioreq_server {
>   * XEN_DMOP_get_ioreq_server_info: Get all the information necessary to
>   *                                 access IOREQ Server <id>.
>   *
> - * The emulator needs to map the synchronous ioreq structures and buffered
> - * ioreq ring (if it exists) that Xen uses to request emulation. These are
> - * hosted in the target domain's gmfns <ioreq_gfn> and <bufioreq_gfn>
> - * respectively. In addition, if the IOREQ Server is handling buffered
> - * emulation requests, the emulator needs to bind to event channel
> - * <bufioreq_port> to listen for them. (The event channels used for
> - * synchronous emulation requests are specified in the per-CPU ioreq
> - * structures in <ioreq_gfn>).
> - * If the IOREQ Server is not handling buffered emulation requests then the
> - * values handed back in <bufioreq_gfn> and <bufioreq_port> will both be 0.
> + * If the IOREQ Server is handling buffered emulation requests, the
> + * emulator needs to bind to event channel <bufioreq_port> to listen for
> + * them. (The event channels used for synchronous emulation requests are
> + * specified in the per-CPU ioreq structures).
> + * In addition, if the XENMEM_acquire_resource memory op cannot be used,
> + * the emulator will need to map the synchronous ioreq structures and
> + * buffered ioreq ring (if it exists) from guest memory. If <flags> does
> + * not contain XEN_DMOP_no_gfns then these pages will be made available and
> + * the frame numbers passed back in gfns <ioreq_gfn> and <bufioreq_gfn>
> + * respectively. (If the IOREQ Server is not handling buffered emulation
> + * only <ioreq_gfn> will be valid).
>   */
>  #define XEN_DMOP_get_ioreq_server_info 2
>  
>  struct xen_dm_op_get_ioreq_server_info {
>      /* IN - server id */
>      ioservid_t id;
> -    uint16_t pad;
> +    /* IN - flags */
> +    uint16_t flags;

Don't you need to bump some version or similar to let the consumers
know this field is now available? Or that's done using a compile test?

> +#define _XEN_DMOP_no_gfns 0
> +#define XEN_DMOP_no_gfns (1u << _XEN_DMOP_no_gfns)
> +
>      /* OUT - buffered ioreq port */
>      evtchn_port_t bufioreq_port;
> -    /* OUT - sync ioreq gfn */
> +    /* OUT - sync ioreq gfn (see block comment above) */
>      uint64_aligned_t ioreq_gfn;
> -    /* OUT - buffered ioreq gfn */
> +    /* OUT - buffered ioreq gfn (see block comment above)*/
>      uint64_aligned_t bufioreq_gfn;
>  };
>  
> -- 
> 2.11.0
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxx
> https://lists.xen.org/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel