Re: [PATCH] RFC: Use memops for mem paging, sharing, and access, instead of domctls

["Andres Lagar-Cavilla" <andres@lagarcavilla.org>]

> At 08:04 -0800 on 12 Jan (1326355474), Andres Lagar-Cavilla wrote:
>> > On Wed, Jan 11, Andres Lagar-Cavilla wrote:
>> >
>> >> +++ b/xen/include/public/domctl.h
>> >
>> >> +/* Use for teardown/setup of helper<->hypervisor interface for
>> paging,
>> >> + * access and sharing.*/
>> >>  struct xen_domctl_mem_event_op {
>> >>      uint32_t       op;           /* XEN_DOMCTL_MEM_EVENT_OP_*_* */
>> >>      uint32_t       mode;         /* XEN_DOMCTL_MEM_EVENT_OP_* */
>> >>
>> >> -    union {
>> >> -        /* OP_ENABLE IN:  Virtual address of shared page */
>> >> -        uint64_aligned_t shared_addr;
>> >> -        /* PAGING_PREP IN: buffer to immediately fill page in */
>> >> -        uint64_aligned_t buffer;
>> >> -    } u;
>> >> +    uint64_aligned_t shared_addr;  /* IN:  Virtual address of shared
>> >> page */
>> >>      uint64_aligned_t ring_addr;    /* IN:  Virtual address of ring
>> page
>> >> */
>> >>
>> >> -    /* Other OPs */
>> >> -    uint64_aligned_t gfn;          /* IN:  gfn of page being
>> operated
>> >> on */
>> >> +    /* For binary backwards compatibility */
>> >> +    uint64_aligned_t pad;
>> >>  };
>> >
>> > Assuming this struct is routed through libxc, and libxc gets a new
>> > SONAME for every release, doesnt this mean that every old binary has
>> to
>> > be recompiled anyway for the new release?
>> > If so, the padding is not needed.
>> Agreed, basically. Waiting to hear from tools maintainers about best
>> approach to libxc.
>>
>> It seems that there aren't that many users relying on a fixed ABI, so we
>> can (still, until 4.2) change things. But obviously I want to be
>> careful.
>
> Ping?

I have working code. Except that it blue-screened W7 once, so it's been
put in the back-burner for next week.

I can repost for review. Ok, I will.

I have not heard any objections thus far to the ABI change.

Andres
>
> Tim.
>