Re: [Xen-devel] [PATCH RFC] tools/libxc, xen/x86: Added xc_set_mem_access_sparse()

[Razvan Cojocaru <rcojocaru@xxxxxxxxxxxxxxx>]

On 08/26/16 23:33, Tamas K Lengyel wrote:
> On Fri, Aug 26, 2016 at 12:11 AM, Razvan Cojocaru
> <rcojocaru@xxxxxxxxxxxxxxx> wrote:
>> Currently it is only possible to set mem_access restrictions only for
>> a contiguous range of GFNs (or, as a particular case, for a single GFN).
>> This patch introduces a new libxc function taking an array of GFNs.
>> The alternative would be to set each page in turn, using a userspace-HV
>> roundtrip for each call, and triggering a TLB flush per page set.
> 
> I think this is a useful addition.

Thanks!

>> Signed-off-by: Razvan Cojocaru <rcojocaru@xxxxxxxxxxxxxxx>
>> ---
>>  tools/libxc/include/xenctrl.h |  4 +++
>>  tools/libxc/xc_mem_access.c   | 32 +++++++++++++++++++++++
>>  xen/arch/x86/hvm/hvm.c        |  2 +-
>>  xen/arch/x86/mm/p2m.c         | 59 
>> ++++++++++++++++++++++++++++++-------------
>>  xen/common/compat/memory.c    |  1 -
>>  xen/common/mem_access.c       | 13 +++++++++-
>>  xen/include/public/memory.h   |  6 +++++
>>  xen/include/xen/p2m-common.h  |  6 ++---
>>  8 files changed, 100 insertions(+), 23 deletions(-)
>>
>> diff --git a/tools/libxc/include/xenctrl.h b/tools/libxc/include/xenctrl.h
>> index 560ce7b..ac84908 100644
>> --- a/tools/libxc/include/xenctrl.h
>> +++ b/tools/libxc/include/xenctrl.h
>> @@ -2125,6 +2125,10 @@ int xc_set_mem_access(xc_interface *xch, domid_t 
>> domain_id,
>>                        xenmem_access_t access, uint64_t first_pfn,
>>                        uint32_t nr);
>>
>> +int xc_set_mem_access_sparse(xc_interface *xch, domid_t domain_id,
>> +                             xenmem_access_t access, xen_pfn_t *pages,
>> +                             uint32_t nr);
>> +
> 
> I guess adding a comment here would be nice for people just looking at
> the header to be able to tell that pages = array, nr = array size.
> Also, wouldn't it make sense to make "access" an array as well, so you
> could one-set multiple pages to different permissions?

Sure, I'll add a comment (actually even rename that parameter to size).

Yes, making access an array as well does make sense, I'll look into
that. I could then just iterate over both arrays and basically use
p2m_set_mem_access() for each single element (with p2m_lock() /
p2m_unlock() around the loop to prevent TLB flushes on each iteration).


Thanks,
Razvan

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