On Wed, Mar 23, 2016 at 09:28:03AM +0800, Shuai Ruan wrote: > On Tue, Mar 22, 2016 at 04:53:02AM -0600, Jan Beulich wrote: > > >>> On 16.03.16 at 13:12, <shuai.r...@linux.intel.com> wrote: > > > > Don't you need to use xcomp_bv here? That's what "Extended > > Region of an XSAVE Area" in SDM Vol 1 suggests to me. > > > "OPERATION OF XRSTORS" and "OPERATION OF XSAVES" in SDM Vol1. > > For xsaves: > "Execution of XSAVES performs the init optimization to reduce the amount > of data written to memory." > > For xrstors: > "XRSTORS updates state component i based on the value of bit i in the > XSTATE_BV field of the XSAVE header". > Ignore this. You are right. I misunderstand some XSAVES/XRSTORS behaviors in SDM . In the previous versions, I use xcom_bv to caculate comp_offset (but changing to xstate_bv in this version:( ). Thanks. > > > /* Set XSTATE_BV and XCOMP_BV. */ > > > xsave->xsave_hdr.xstate_bv = xstate_bv; > > > xsave->xsave_hdr.xcomp_bv = v->arch.xcr0_accum | > > > XSTATE_COMPACTION_ENABLED; > > > + setup_xstate_comp(xstate_comp_offsets, xstate_bv); > > > > Same here then I think. > > > > Jan > > > > _______________________________________________ > > Xen-devel mailing list > > Xen-devel@lists.xen.org > > http://lists.xen.org/xen-devel > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xen.org > http://lists.xen.org/xen-devel
_______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org http://lists.xen.org/xen-devel
