KVM: selftests: Add a fully functional "struct xstate" for x86

Add a working xstate data structure for the usage of AMX and potential
future usage on other xstate components. AMX selftest requires checking
both the xstate_bv and xcomp_bv. Existing code relies on pointer
arithmetics to fetch xstate_bv and does not support xcomp_bv.

So, add a working xstate data structure into processor.h for x86.

Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Mingwei Zhang <mizhang@google.com>
Link: https://lore.kernel.org/r/20230221163655.920289-3-mizhang@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>

Author: mzhang3579

tools/testing/selftests/kvm/include/x86_64/processor.h

@@ -48,6 +48,18 @@ extern bool host_cpu_is_amd;
 #define X86_CR4_SMAP		(1ul << 21)
 #define X86_CR4_PKE		(1ul << 22)
 
+struct xstate_header {
+	u64				xstate_bv;
+	u64				xcomp_bv;
+	u64				reserved[6];
+} __attribute__((packed));
+
+struct xstate {
+	u8				i387[512];
+	struct xstate_header		header;
+	u8				extended_state_area[0];
+} __attribute__ ((packed, aligned (64)));
+
 /* Note, these are ordered alphabetically to match kvm_cpuid_entry2.  Eww. */
 enum cpuid_output_regs {
 	KVM_CPUID_EAX,

tools/testing/selftests/kvm/x86_64/amx_test.c

@@ -41,10 +41,6 @@
 
 #define XSAVE_HDR_OFFSET		512
 
-struct xsave_data {
-	u8 area[XSAVE_SIZE];
-} __aligned(64);
-
 struct tile_config {
 	u8  palette_id;
 	u8  start_row;
@@ -103,13 +99,13 @@ static inline void __tilerelease(void)
 	asm volatile(".byte 0xc4, 0xe2, 0x78, 0x49, 0xc0" ::);
 }
 
-static inline void __xsavec(struct xsave_data *data, uint64_t rfbm)
+static inline void __xsavec(struct xstate *xstate, uint64_t rfbm)
 {
 	uint32_t rfbm_lo = rfbm;
 	uint32_t rfbm_hi = rfbm >> 32;
 
 	asm volatile("xsavec (%%rdi)"
-		     : : "D" (data), "a" (rfbm_lo), "d" (rfbm_hi)
+		     : : "D" (xstate), "a" (rfbm_lo), "d" (rfbm_hi)
 		     : "memory");
 }
 
@@ -158,16 +154,6 @@ static void set_tilecfg(struct tile_config *cfg)
 	}
 }
 
-static void set_xstatebv(void *data, uint64_t bv)
-{
-	*(uint64_t *)(data + XSAVE_HDR_OFFSET) = bv;
-}
-
-static u64 get_xstatebv(void *data)
-{
-	return *(u64 *)(data + XSAVE_HDR_OFFSET);
-}
-
 static void init_regs(void)
 {
 	uint64_t cr4, xcr0;
@@ -184,7 +170,7 @@ static void init_regs(void)
 
 static void __attribute__((__flatten__)) guest_code(struct tile_config *amx_cfg,
 						    struct tile_data *tiledata,
-						    struct xsave_data *xsave_data)
+						    struct xstate *xstate)
 {
 	init_regs();
 	check_cpuid_xsave();
@@ -205,9 +191,9 @@ static void __attribute__((__flatten__)) guest_code(struct tile_config *amx_cfg,
 	__tilerelease();
 	GUEST_SYNC(5);
 	/* bit 18 not in the XCOMP_BV after xsavec() */
-	set_xstatebv(xsave_data, XFEATURE_MASK_XTILEDATA);
-	__xsavec(xsave_data, XFEATURE_MASK_XTILEDATA);
-	GUEST_ASSERT((get_xstatebv(xsave_data) & XFEATURE_MASK_XTILEDATA) == 0);
+	xstate->header.xstate_bv = XFEATURE_MASK_XTILEDATA;
+	__xsavec(xstate, XFEATURE_MASK_XTILEDATA);
+	GUEST_ASSERT(!(xstate->header.xstate_bv & XFEATURE_MASK_XTILEDATA));
 
 	/* xfd=0x40000, disable amx tiledata */
 	wrmsr(MSR_IA32_XFD, XFEATURE_MASK_XTILEDATA);
@@ -243,7 +229,7 @@ int main(int argc, char *argv[])
 	struct kvm_vm *vm;
 	struct kvm_x86_state *state;
 	int xsave_restore_size;
-	vm_vaddr_t amx_cfg, tiledata, xsavedata;
+	vm_vaddr_t amx_cfg, tiledata, xstate;
 	struct ucall uc;
 	u32 amx_offset;
 	int stage, ret;
@@ -282,10 +268,10 @@ int main(int argc, char *argv[])
 	tiledata = vm_vaddr_alloc_pages(vm, 2);
 	memset(addr_gva2hva(vm, tiledata), rand() | 1, 2 * getpagesize());
 
-	/* xsave data for guest_code */
-	xsavedata = vm_vaddr_alloc_pages(vm, 3);
-	memset(addr_gva2hva(vm, xsavedata), 0, 3 * getpagesize());
-	vcpu_args_set(vcpu, 3, amx_cfg, tiledata, xsavedata);
+	/* XSAVE state for guest_code */
+	xstate = vm_vaddr_alloc_pages(vm, DIV_ROUND_UP(XSAVE_SIZE, PAGE_SIZE));
+	memset(addr_gva2hva(vm, xstate), 0, PAGE_SIZE * DIV_ROUND_UP(XSAVE_SIZE, PAGE_SIZE));
+	vcpu_args_set(vcpu, 3, amx_cfg, tiledata, xstate);
 
 	for (stage = 1; ; stage++) {
 		vcpu_run(vcpu);