RE: [Xen-ia64-devel] Question about vmx_ivt.S

["Xu, Anthony" <anthony.xu@xxxxxxxxx>]

Hi Tristan,

Yes the mf is necessary,
Hardware can speculatively load vhpt entries from VTHP page.
We had spent a lot of effort to identity this issue.

Anthony






>-----Original Message-----
>From: xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx [mailto:xen-ia64-devel-
>bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Tristan Gingold
>Sent: 2007年9月30日 23:26
>To: Xen-ia64-devel
>Subject: [Xen-ia64-devel] Question about vmx_ivt.S
>
>Hi,
>
>for both itlb and dtlb handlers, while swapping first and found entries:
>    st8 [r16] = r29, VLE_ITIR_OFFSET - VLE_TITAG_OFFSET
>    st8 [r28] = r22, VLE_ITIR_OFFSET - VLE_TITAG_OFFSET
>    extr.u r19 = r27, 56, 4
>    mf
>    ;;
>    ld8 r29 = [r16]
>    ld8 r22 = [r28]
>    dep r27 = r0, r27, 56, 4
>    dep r25 = r19, r25, 56, 4
>
>Is the mf (memory fence) necessary ?  The VHPT is per-cpu.
>
>Tristan.
>
>_______________________________________________
>Xen-ia64-devel mailing list
>Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
>http://lists.xensource.com/xen-ia64-devel

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel