On Thu, Nov 05, 2015 at 03:13:01AM -0700, Jan Beulich wrote:
> >>> On 05.11.15 at 10:57, <shuai.r...@linux.intel.com> wrote:
> > Ok. So alternative_input will not used here (means use the way
> > xrstor in Patch 8)? Or put the XSTATE_FIXUP into alternative_input ?
> > Which one is ok to you ?
>
> The latter, if necessary by extending alternative_input() accordingly
> (or provide a second, more flexible variant if need be; iirc Linux has
> gained a couple of variants over the years).
>
Ok,I will introduct 2 new macro to handle this.
1. #define ASM_OUTPUT2(a...) a
The first one(Porting from linux) is used when more than one output parameter
is needed.
2. #define alternative_io_fixup(oldinstr, newinstr, feature, fixup, output,
input...) \
asm volatile (ALTERNATIVE(oldinstr, newinstr, feature) "\n" \
fixup \
: output : "i"(0), ## input)
The second is to handle alternaive asm with fixup.
Then xrstor side will be:
alternative_io_fixup("1: "".byte 0x48,0x0f,0xae,0x2f",
".byte 0x48,0x0f,0xc7,0x1f",
X86_FEATURE_XSAVES,
XSTATE_FIXUP,
ASM_OUTPUT2("+&D" (ptr), "+&a" (lmask)),
"m" (*ptr), "g" (lmask), "d" (hmask),
"m" (xsave_cntxt_size) : "ecx");
Also, the position of the operand used in XSTATE_FIXUP will
changed.
Any suggestions of this ?
Thanks
> 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
