[PATCH v2 2/2] x86/fpu: Initialise FTW in xstate_alloc_save_area()

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

Right now, xstate_alloc_save_area() leaves both XSTATE_BV and FTW clear.  This
causes a difference in behaviour between FXRSTOR and XRSTOR.

Switch to using using XSTATE's idea of initial configuration which will behave
the same even on pre-XSAVE hardware.  Expand the comment to explain why.

Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>
---
CC: Jan Beulich <JBeulich@xxxxxxxx>
CC: Roger Pau Monné <roger.pau@xxxxxxxxxx>
CC: Ross Lagerwall <ross.lagerwall@xxxxxxxxxx>

v2:
 * Rewrite the commmit message and comment.
---
 xen/arch/x86/xstate.c | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/xen/arch/x86/xstate.c b/xen/arch/x86/xstate.c
index e990abc9d18c..defe9b3f0cbe 100644
--- a/xen/arch/x86/xstate.c
+++ b/xen/arch/x86/xstate.c
@@ -550,11 +550,22 @@ int xstate_alloc_save_area(struct vcpu *v)
         return -ENOMEM;
 
     /*
-     * Set the memory image to default values, but don't force the context
-     * to be loaded from memory (i.e. keep save_area->xsave_hdr.xstate_bv
-     * clear).
+     * We're creating a vCPU, so conceptually we should be choosing the
+     * architectural #RESET values.
+     *
+     * However for historical reasons of configuring the external
+     * co-processor, FCW's #RESET state is different to what F(N)INIT and
+     * XSTATE consider the "initial configuration".
+     *
+     * Guests won't care about the difference; all software tends to executes
+     * FNINIT very early during setup.
+     *
+     * Use XSTATE's idea of initial configuration.  This allows XSTATE_BV to
+     * remain clear and for CPUs to use the INIT optimisation where
+     * applicable.
      */
     save_area->fpu_sse.fcw = FCW_DEFAULT;
+    save_area->fpu_sse.ftw = FXSAVE_FTW_RESET;
     save_area->fpu_sse.mxcsr = MXCSR_DEFAULT;
 
     v->arch.xsave_area = save_area;
-- 
2.39.5