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

Re: [Xen-devel] [PATCH] x86/HVM: avoid pointer wraparound in bufioreq handling



On Tue, 2015-06-16 at 07:44 +0100, Jan Beulich wrote:
> >>> On 15.06.15 at 16:30, <JBeulich@xxxxxxxx> wrote:
> > The number of slots per page being 511 (i.e. not a power of two) means
> > that the (32-bit) read and write indexes going beyond 2^32 will likely
> > disturb operation. Extend I/O req server creation so the caller can
> > indicate that it is using suitable atomic accesses where needed (not
> > all accesses to the two pointers really need to be atomic), allowing
> > the hypervisor to atomically canonicalize both pointers when both have
> > gone through at least one cycle.
> > 
> > Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
> 
> No matter that it's just a single line change, I realized that I
> forgot to Cc the tools maintainers. While a v2 will be needed (see
> the reply just sent to Andrew) I'd still appreciate input (if any) to
> limit the number of revisions needed.

For such a simple toolstack side change which just reflects the
underlying hcall interface I have no real opinion so far as the tools
side goes, but it would be good to update the comments in xenctrl.h too.
With that done for the tools change:
        Acked-by: Ian Campbell <ian.campbell@xxxxxxxxxx>

For the hypercall interface level, I wonder if handle_bufioreq is still
an appropriate name given its no longer treated as a boolean flag?
bufioreq_type or something perhaps?

Ian.


_______________________________________________
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®.