[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Possible issue with x86_emulate when writing results back to memory



Thanks,

> > The attached patch is for discussion purposes only - if it is deemed
> acceptable I'll resubmit a proper patch request against unstable.
> 
> That seems like a plausible explanation.
> 
> The patch however needs some work.  As this function is identical, it
> should have a common implementation somewhere, possibly part of
> x86_emulate.h, and it would probably be better as:
> 

Not sure I want the generic HVM code to be dependent on x86_emulate... not sure 
it should be in hvm.c either

> To better match real hardware, it might be appropriate for
> "memcpy_atomic()" (name subject to improved suggestions) to use a while
> loop and issue 8 byte writes at a time, falling down to 4, 2 then 1 when
> reaching the end of the data to be copied.
> 

My concern here would be that the generic hvm routine __hvm_copy needs to use 
this when emulating instruction but not the rest of the time and I'd be 
concerned about the perf impact.

I'll noodle on a suitable single place to put this...

> ~Andrew

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


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.