Re: [PATCH V3 1/3] xen/mem_access: Support for memory-content hiding

[Razvan Cojocaru <rcojocaru@bitdefender.com>]

On 07/07/2015 04:15 PM, Lengyel, Tamas wrote:
> 
> 
> On Tue, Jul 7, 2015 at 9:09 AM, Razvan Cojocaru
> <rcojocaru@bitdefender.com <mailto:rcojocaru@bitdefender.com>> wrote:
> 
>     So VM_EVENT_FLAG_FOREIGN (1 << 1), and then MEM_ACCESS_EMULATE (1 << 6).
>     Now you're adding VM_EVENT_FLAG_TOGGLE_SINGLESTEP (1 << 2), and if we're
>     not very careful, slowly but surely the VM_EVENT_FLAG_ constants will
>     crawl towards (1 << 6) and start overlapping with MEM_ACCESS_EMULATE,
>     MEM_ACCESS_EMULATE_NOWRITE and so on, because they're not clearly
>     #defined right after the rest of the VM_EVENT_FLAG_ ones, are called
>     MEM_ACCESS_<something> just like the mem_access event specific flags,
>     and they use bit shifts relative to the MEM_ACCESS_ ones as well. This
>     is in part what has caused my confusion here.
> 
>     What do you think?
> 
>     Thanks,
>     Razvan
> 
> 
> Well, this is clearly an oversight that we need to fix. MEM_ACCESS_*
> flags should be only set on the mem_access flags field.

Right, if there are no objections, I'll just submit a small patch (in a
few minutes) that moves MEM_ACCESS_EMULATE and
MEM_ACCESS_EMULATE_NOWRITE up near the VM_EVENT_FLAG_ constants, and
renames them VM_EVENT_FLAG_EMULATE and VM_EVENT_FLAG_EMULATE_NOWRITE.

It seems like this would be a pretty straightforward change, so I assume
it's likely to go in fast.


Thanks,
Razvan