Skip to contentSkip to conversation

x86/fpu changes for v4.4

GIT PULL Ingo Molnar 1 participant 1 message Nov 3, 2015

Subsystems (5)
x86 Kselftest Crypto API KVM x86 +1 more subsystems
lore
Share & tools
lore
#0 lore raw

Linus,

Please pull the latest x86-fpu-for-linus git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-fpu-for-linus

# HEAD: 158ecc39185b885420e5136b803b29be2bbec7fb x86/fpu: Fixup uninitialized feature_name warning

There are two main areas of changes:

- Rework of the extended FPU state code to robustify the kernel's usage of cpuid provided xstate sizes - and related changes. (Dave Hansen)

- math emulation enhancements: new modern FPU instructions support, with testcases, plus cleanups. (Denys Vlasnko)

Thanks,

Ingo

------------------>
Borislav Petkov (1):
x86/fpu: Fixup uninitialized feature_name warning

Dave Hansen (15):
x86/fpu: Print xfeature buffer size in decimal x86/fpu: Move XSAVE-disabling code to a helper x86/fpu: Remove XSTATE_RESERVE
x86/fpu: Remove partial LWP support definitions x86/fpu: Rename XSAVE macros
x86/fpu: Rename XFEATURES_NR_MAX
x86/fpu: Rework XSTATE_* macros to remove magic '2' x86/fpu: Remove 'xfeature_nr'
x86/fpu: Add xfeature_enabled() helper instead of test_bit() x86/fpu/mpx: Rework MPX 'xstate' types
x86/fpu: Rework YMM definition
x86/fpu: Add C structures for AVX-512 state components x86/fpu: Correct and check XSAVE xstate size calculations x86/fpu: Check to ensure increasing-offset xstate offsets x86/fpu: Check CPU-provided sizes against struct declarations

Denys Vlasenko (7):
x86/fpu/math-emu: Remove !NO_UNDOC_CODE x86/fpu/math-emu, selftests: Add tests for FCMOV and FCOMI insns x86/fpu/math-emu: Remove define layer for undocumented opcodes x86/fpu/math-emu: Add support for F[U]COMI[P] insns x86/fpu/math-emu: Add support for FCMOVcc insns x86/fpu/math-emu, selftests: Add test for FISTTP instructions x86/fpu/math-emu: Add support for FISTTP instructions

 arch/x86/crypto/camellia_aesni_avx2_glue.c |   3 +-
 arch/x86/crypto/camellia_aesni_avx_glue.c  |   3 +-
 arch/x86/crypto/cast5_avx_glue.c           |   3 +-
 arch/x86/crypto/cast6_avx_glue.c           |   3 +-
 arch/x86/crypto/chacha20_glue.c            |   2 +-
 arch/x86/crypto/poly1305_glue.c            |   2 +-
 arch/x86/crypto/serpent_avx2_glue.c        |   3 +-
 arch/x86/crypto/serpent_avx_glue.c         |   3 +-
 arch/x86/crypto/sha1_ssse3_glue.c          |   2 +-
 arch/x86/crypto/sha256_ssse3_glue.c        |   2 +-
 arch/x86/crypto/sha512_ssse3_glue.c        |   2 +-
 arch/x86/crypto/twofish_avx_glue.c         |   2 +-
 arch/x86/include/asm/fpu/types.h           | 148 +++++++----
 arch/x86/include/asm/fpu/xstate.h          |  15 +-
 arch/x86/include/asm/trace/mpx.h           |   7 +-
 arch/x86/kernel/fpu/init.c                 |  18 +-
 arch/x86/kernel/fpu/regset.c               |   4 +-
 arch/x86/kernel/fpu/signal.c               |   6 +-
 arch/x86/kernel/fpu/xstate.c               | 398 +++++++++++++++++++++++++----
 arch/x86/kernel/traps.c                    |   4 +-
 arch/x86/kvm/cpuid.c                       |   4 +-
 arch/x86/kvm/x86.c                         |  27 +-
 arch/x86/kvm/x86.h                         |   6 +-
 arch/x86/math-emu/fpu_aux.c                |  70 +++++
 arch/x86/math-emu/fpu_entry.c              |  96 +++----
 arch/x86/math-emu/fpu_proto.h              |  12 +
 arch/x86/math-emu/load_store.c             |  63 ++++-
 arch/x86/math-emu/reg_compare.c            | 128 ++++++++++
 arch/x86/mm/mpx.c                          |  15 +-
 tools/testing/selftests/x86/Makefile       |   5 +-
 tools/testing/selftests/x86/test_FCMOV.c   |  93 +++++++
 tools/testing/selftests/x86/test_FCOMI.c   | 331 ++++++++++++++++++++++++
 tools/testing/selftests/x86/test_FISTTP.c  | 137 ++++++++++
 33 files changed, 1374 insertions(+), 243 deletions(-)
 create mode 100644 tools/testing/selftests/x86/test_FCMOV.c
 create mode 100644 tools/testing/selftests/x86/test_FCOMI.c
 create mode 100644 tools/testing/selftests/x86/test_FISTTP.c
arch/x86/crypto/camellia_aesni_avx2_glue.c +2 -1
index 4c65c70e628b..d84456924563 100644
@@ -567,7 +567,8 @@ static int __init camellia_aesni_init(void)
 		return -ENODEV;
 	}
 
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, &feature_name)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM,
+				&feature_name)) {
 		pr_info("CPU feature '%s' is not supported.\n", feature_name);
 		return -ENODEV;
 	}
arch/x86/crypto/camellia_aesni_avx_glue.c +2 -1
index 80a0e4389c9a..12e729bfe71b 100644
@@ -554,7 +554,8 @@ static int __init camellia_aesni_init(void)
 {
 	const char *feature_name;
 
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, &feature_name)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM,
+				&feature_name)) {
 		pr_info("CPU feature '%s' is not supported.\n", feature_name);
 		return -ENODEV;
 	}
arch/x86/crypto/cast5_avx_glue.c +2 -1
index be00aa48b2b5..8648158f3916 100644
@@ -469,7 +469,8 @@ static int __init cast5_init(void)
 {
 	const char *feature_name;
 
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, &feature_name)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM,
+				&feature_name)) {
 		pr_info("CPU feature '%s' is not supported.\n", feature_name);
 		return -ENODEV;
 	}
arch/x86/crypto/cast6_avx_glue.c +2 -1
index 5dbba7224221..fca459578c35 100644
@@ -591,7 +591,8 @@ static int __init cast6_init(void)
 {
 	const char *feature_name;
 
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, &feature_name)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM,
+				&feature_name)) {
 		pr_info("CPU feature '%s' is not supported.\n", feature_name);
 		return -ENODEV;
 	}
arch/x86/crypto/chacha20_glue.c +1 -1
index effe2160b7c5..722bacea040e 100644
@@ -130,7 +130,7 @@ static int __init chacha20_simd_mod_init(void)
 
 #ifdef CONFIG_AS_AVX2
 	chacha20_use_avx2 = cpu_has_avx && cpu_has_avx2 &&
-			    cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, NULL);
+			    cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM, NULL);
 #endif
 	return crypto_register_alg(&alg);
 }
arch/x86/crypto/poly1305_glue.c +1 -1
index f7170d764f32..4264a3d59589 100644
@@ -184,7 +184,7 @@ static int __init poly1305_simd_mod_init(void)
 
 #ifdef CONFIG_AS_AVX2
 	poly1305_use_avx2 = cpu_has_avx && cpu_has_avx2 &&
-			    cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, NULL);
+			    cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM, NULL);
 	alg.descsize = sizeof(struct poly1305_simd_desc_ctx);
 	if (poly1305_use_avx2)
 		alg.descsize += 10 * sizeof(u32);
arch/x86/crypto/serpent_avx2_glue.c +2 -1
index 7d838dc4d888..6d198342e2de 100644
@@ -542,7 +542,8 @@ static int __init init(void)
 		pr_info("AVX2 instructions are not detected.\n");
 		return -ENODEV;
 	}
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, &feature_name)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM,
+				&feature_name)) {
 		pr_info("CPU feature '%s' is not supported.\n", feature_name);
 		return -ENODEV;
 	}
arch/x86/crypto/serpent_avx_glue.c +2 -1
index da7dafc9b16d..5dc37026c7ce 100644
@@ -597,7 +597,8 @@ static int __init serpent_init(void)
 {
 	const char *feature_name;
 
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, &feature_name)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM,
+				&feature_name)) {
 		pr_info("CPU feature '%s' is not supported.\n", feature_name);
 		return -ENODEV;
 	}
arch/x86/crypto/sha1_ssse3_glue.c +1 -1
index 7c48e8b20848..00212c32d4db 100644
@@ -121,7 +121,7 @@ static struct shash_alg alg = {
 #ifdef CONFIG_AS_AVX
 static bool __init avx_usable(void)
 {
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, NULL)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM, NULL)) {
 		if (cpu_has_avx)
 			pr_info("AVX detected but unusable.\n");
 		return false;
arch/x86/crypto/sha256_ssse3_glue.c +1 -1
index f8097fc0d1d1..0e0e85aea634 100644
@@ -130,7 +130,7 @@ static struct shash_alg algs[] = { {
 #ifdef CONFIG_AS_AVX
 static bool __init avx_usable(void)
 {
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, NULL)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM, NULL)) {
 		if (cpu_has_avx)
 			pr_info("AVX detected but unusable.\n");
 		return false;
arch/x86/crypto/sha512_ssse3_glue.c +1 -1
index 2edad7b81870..0c8c38c101ac 100644
@@ -129,7 +129,7 @@ static struct shash_alg algs[] = { {
 #ifdef CONFIG_AS_AVX
 static bool __init avx_usable(void)
 {
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, NULL)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM, NULL)) {
 		if (cpu_has_avx)
 			pr_info("AVX detected but unusable.\n");
 		return false;
arch/x86/crypto/twofish_avx_glue.c +1 -1
index c2bd0ce718ee..b7a3904b953c 100644
@@ -558,7 +558,7 @@ static int __init twofish_init(void)
 {
 	const char *feature_name;
 
-	if (!cpu_has_xfeatures(XSTATE_SSE | XSTATE_YMM, &feature_name)) {
+	if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM, &feature_name)) {
 		pr_info("CPU feature '%s' is not supported.\n", feature_name);
 		return -ENODEV;
 	}
arch/x86/include/asm/fpu/types.h +103 -45

195 lines not on this page · raw message

arch/x86/include/asm/fpu/xstate.h +10 -5
index 4656b25bb9a7..3a6c89b70307 100644
@@ -6,7 +6,7 @@
 #include <linux/uaccess.h>
 
 /* Bit 63 of XCR0 is reserved for future expansion */
-#define XSTATE_EXTEND_MASK	(~(XSTATE_FPSSE | (1ULL << 63)))
+#define XFEATURE_MASK_EXTEND	(~(XFEATURE_MASK_FPSSE | (1ULL << 63)))
 
 #define XSTATE_CPUID		0x0000000d
 
@@ -19,14 +19,18 @@
 #define XSAVE_YMM_OFFSET    (XSAVE_HDR_SIZE + XSAVE_HDR_OFFSET)
 
 /* Supported features which support lazy state saving */
-#define XSTATE_LAZY	(XSTATE_FP | XSTATE_SSE | XSTATE_YMM		      \
-			| XSTATE_OPMASK | XSTATE_ZMM_Hi256 | XSTATE_Hi16_ZMM)
+#define XFEATURE_MASK_LAZY	(XFEATURE_MASK_FP | \
+				 XFEATURE_MASK_SSE | \
+				 XFEATURE_MASK_YMM | \
+				 XFEATURE_MASK_OPMASK |	\
+				 XFEATURE_MASK_ZMM_Hi256 | \
+				 XFEATURE_MASK_Hi16_ZMM)
 
 /* Supported features which require eager state saving */
-#define XSTATE_EAGER	(XSTATE_BNDREGS | XSTATE_BNDCSR)
+#define XFEATURE_MASK_EAGER	(XFEATURE_MASK_BNDREGS | XFEATURE_MASK_BNDCSR)
 
 /* All currently supported features */
-#define XCNTXT_MASK	(XSTATE_LAZY | XSTATE_EAGER)
+#define XCNTXT_MASK	(XFEATURE_MASK_LAZY | XFEATURE_MASK_EAGER)
 
 #ifdef CONFIG_X86_64
 #define REX_PREFIX	"0x48, "
@@ -40,6 +44,7 @@ extern u64 xstate_fx_sw_bytes[USER_XSTATE_FX_SW_WORDS];
 
 extern void update_regset_xstate_info(unsigned int size, u64 xstate_mask);
 
+void fpu__xstate_clear_all_cpu_caps(void);
 void *get_xsave_addr(struct xregs_state *xsave, int xstate);
 const void *get_xsave_field_ptr(int xstate_field);
 
arch/x86/include/asm/trace/mpx.h +4 -3
index 173dd3ba108c..0f492fc50bce 100644
@@ -11,7 +11,7 @@
 TRACE_EVENT(mpx_bounds_register_exception,
 
 	TP_PROTO(void *addr_referenced,
-		 const struct bndreg *bndreg),
+		 const struct mpx_bndreg *bndreg),
 	TP_ARGS(addr_referenced, bndreg),
 
 	TP_STRUCT__entry(
@@ -44,7 +44,7 @@ TRACE_EVENT(mpx_bounds_register_exception,
 
 TRACE_EVENT(bounds_exception_mpx,
 
-	TP_PROTO(const struct bndcsr *bndcsr),
+	TP_PROTO(const struct mpx_bndcsr *bndcsr),
 	TP_ARGS(bndcsr),
 
 	TP_STRUCT__entry(
@@ -116,7 +116,8 @@ TRACE_EVENT(mpx_new_bounds_table,
 /*
  * This gets used outside of MPX-specific code, so we need a stub.
  */
-static inline void trace_bounds_exception_mpx(const struct bndcsr *bndcsr)
+static inline
+void trace_bounds_exception_mpx(const struct mpx_bndcsr *bndcsr)
 {
 }
 
arch/x86/kernel/fpu/init.c +4 -14
index d14e9ac3235a..be39b5fde4b9 100644
@@ -290,11 +290,11 @@ static void __init fpu__init_system_ctx_switch(void)
 	if (cpu_has_xsaveopt && eagerfpu != DISABLE)
 		eagerfpu = ENABLE;
 
-	if (xfeatures_mask & XSTATE_EAGER) {
+	if (xfeatures_mask & XFEATURE_MASK_EAGER) {
 		if (eagerfpu == DISABLE) {
 			pr_err("x86/fpu: eagerfpu switching disabled, disabling the following xstate features: 0x%llx.\n",
-			       xfeatures_mask & XSTATE_EAGER);
-			xfeatures_mask &= ~XSTATE_EAGER;
+			       xfeatures_mask & XFEATURE_MASK_EAGER);
+			xfeatures_mask &= ~XFEATURE_MASK_EAGER;
 		} else {
 			eagerfpu = ENABLE;
 		}
@@ -354,17 +354,7 @@ static int __init x86_noxsave_setup(char *s)
 	if (strlen(s))
 		return 0;
 
-	setup_clear_cpu_cap(X86_FEATURE_XSAVE);
-	setup_clear_cpu_cap(X86_FEATURE_XSAVEOPT);
-	setup_clear_cpu_cap(X86_FEATURE_XSAVEC);
-	setup_clear_cpu_cap(X86_FEATURE_XSAVES);
-	setup_clear_cpu_cap(X86_FEATURE_AVX);
-	setup_clear_cpu_cap(X86_FEATURE_AVX2);
-	setup_clear_cpu_cap(X86_FEATURE_AVX512F);
-	setup_clear_cpu_cap(X86_FEATURE_AVX512PF);
-	setup_clear_cpu_cap(X86_FEATURE_AVX512ER);
-	setup_clear_cpu_cap(X86_FEATURE_AVX512CD);
-	setup_clear_cpu_cap(X86_FEATURE_MPX);
+	fpu__xstate_clear_all_cpu_caps();
 
 	return 1;
 }
arch/x86/kernel/fpu/regset.c +2 -2
index dc60810c1c74..0bc3490420c5 100644
@@ -66,7 +66,7 @@ int xfpregs_set(struct task_struct *target, const struct user_regset *regset,
 	 * presence of FP and SSE state.
 	 */
 	if (cpu_has_xsave)
-		fpu->state.xsave.header.xfeatures |= XSTATE_FPSSE;
+		fpu->state.xsave.header.xfeatures |= XFEATURE_MASK_FPSSE;
 
 	return ret;
 }
@@ -326,7 +326,7 @@ int fpregs_set(struct task_struct *target, const struct user_regset *regset,
 	 * presence of FP.
 	 */
 	if (cpu_has_xsave)
-		fpu->state.xsave.header.xfeatures |= XSTATE_FP;
+		fpu->state.xsave.header.xfeatures |= XFEATURE_MASK_FP;
 	return ret;
 }
 
arch/x86/kernel/fpu/signal.c +3 -3
index 50ec9af1bd51..eb032677f939 100644
@@ -107,7 +107,7 @@ static inline int save_xstate_epilog(void __user *buf, int ia32_frame)
 	 * header as well as change any contents in the memory layout.
 	 * xrestore as part of sigreturn will capture all the changes.
 	 */
-	xfeatures |= XSTATE_FPSSE;
+	xfeatures |= XFEATURE_MASK_FPSSE;
 
 	err |= __put_user(xfeatures, (__u32 *)&x->header.xfeatures);
 
@@ -207,7 +207,7 @@ sanitize_restored_xstate(struct task_struct *tsk,
 		 * layout and not enabled by the OS.
 		 */
 		if (fx_only)
-			header->xfeatures = XSTATE_FPSSE;
+			header->xfeatures = XFEATURE_MASK_FPSSE;
 		else
 			header->xfeatures &= (xfeatures_mask & xfeatures);
 	}
@@ -230,7 +230,7 @@ static inline int copy_user_to_fpregs_zeroing(void __user *buf, u64 xbv, int fx_
 {
 	if (use_xsave()) {
 		if ((unsigned long)buf % 64 || fx_only) {
-			u64 init_bv = xfeatures_mask & ~XSTATE_FPSSE;
+			u64 init_bv = xfeatures_mask & ~XFEATURE_MASK_FPSSE;
 			copy_kernel_to_xregs(&init_fpstate.xsave, init_bv);
 			return copy_user_to_fxregs(buf);
 		} else {
arch/x86/kernel/fpu/xstate.c +345 -53

511 lines not on this page · raw message

arch/x86/kernel/traps.c +2 -2
index 346eec73f7db..ade185a46b1d 100644
@@ -361,7 +361,7 @@ dotraplinkage void do_double_fault(struct pt_regs *regs, long error_code)
 
 dotraplinkage void do_bounds(struct pt_regs *regs, long error_code)
 {
-	const struct bndcsr *bndcsr;
+	const struct mpx_bndcsr *bndcsr;
 	siginfo_t *info;
 
 	RCU_LOCKDEP_WARN(!rcu_is_watching(), "entry code didn't wake RCU");
@@ -384,7 +384,7 @@ dotraplinkage void do_bounds(struct pt_regs *regs, long error_code)
 	 * which is all zeros which indicates MPX was not
 	 * responsible for the exception.
 	 */
-	bndcsr = get_xsave_field_ptr(XSTATE_BNDCSR);
+	bndcsr = get_xsave_field_ptr(XFEATURE_MASK_BNDCSR);
 	if (!bndcsr)
 		goto exit_trap;
 
arch/x86/kvm/cpuid.c +2 -2
index 2fbea2544f24..156441bcaac8 100644
@@ -30,7 +30,7 @@ static u32 xstate_required_size(u64 xstate_bv, bool compacted)
 	int feature_bit = 0;
 	u32 ret = XSAVE_HDR_SIZE + XSAVE_HDR_OFFSET;
 
-	xstate_bv &= XSTATE_EXTEND_MASK;
+	xstate_bv &= XFEATURE_MASK_EXTEND;
 	while (xstate_bv) {
 		if (xstate_bv & 0x1) {
 		        u32 eax, ebx, ecx, edx, offset;
@@ -51,7 +51,7 @@ u64 kvm_supported_xcr0(void)
 	u64 xcr0 = KVM_SUPPORTED_XCR0 & host_xcr0;
 
 	if (!kvm_x86_ops->mpx_supported())
-		xcr0 &= ~(XSTATE_BNDREGS | XSTATE_BNDCSR);
+		xcr0 &= ~(XFEATURE_MASK_BNDREGS | XFEATURE_MASK_BNDCSR);
 
 	return xcr0;
 }
arch/x86/kvm/x86.c +14 -13
index a60bdbccff51..2d4e54db49af 100644
@@ -662,9 +662,9 @@ static int __kvm_set_xcr(struct kvm_vcpu *vcpu, u32 index, u64 xcr)
 	/* Only support XCR_XFEATURE_ENABLED_MASK(xcr0) now  */
 	if (index != XCR_XFEATURE_ENABLED_MASK)
 		return 1;
-	if (!(xcr0 & XSTATE_FP))
+	if (!(xcr0 & XFEATURE_MASK_FP))
 		return 1;
-	if ((xcr0 & XSTATE_YMM) && !(xcr0 & XSTATE_SSE))
+	if ((xcr0 & XFEATURE_MASK_YMM) && !(xcr0 & XFEATURE_MASK_SSE))
 		return 1;
 
 	/*
@@ -672,23 +672,24 @@ static int __kvm_set_xcr(struct kvm_vcpu *vcpu, u32 index, u64 xcr)
 	 * saving.  However, xcr0 bit 0 is always set, even if the
 	 * emulated CPU does not support XSAVE (see fx_init).
 	 */
-	valid_bits = vcpu->arch.guest_supported_xcr0 | XSTATE_FP;
+	valid_bits = vcpu->arch.guest_supported_xcr0 | XFEATURE_MASK_FP;
 	if (xcr0 & ~valid_bits)
 		return 1;
 
-	if ((!(xcr0 & XSTATE_BNDREGS)) != (!(xcr0 & XSTATE_BNDCSR)))
+	if ((!(xcr0 & XFEATURE_MASK_BNDREGS)) !=
+	    (!(xcr0 & XFEATURE_MASK_BNDCSR)))
 		return 1;
 
-	if (xcr0 & XSTATE_AVX512) {
-		if (!(xcr0 & XSTATE_YMM))
+	if (xcr0 & XFEATURE_MASK_AVX512) {
+		if (!(xcr0 & XFEATURE_MASK_YMM))
 			return 1;
-		if ((xcr0 & XSTATE_AVX512) != XSTATE_AVX512)
+		if ((xcr0 & XFEATURE_MASK_AVX512) != XFEATURE_MASK_AVX512)
 			return 1;
 	}
 	kvm_put_guest_xcr0(vcpu);
 	vcpu->arch.xcr0 = xcr0;
 
-	if ((xcr0 ^ old_xcr0) & XSTATE_EXTEND_MASK)
+	if ((xcr0 ^ old_xcr0) & XFEATURE_MASK_EXTEND)
 		kvm_update_cpuid(vcpu);
 	return 0;
 }
@@ -2906,7 +2907,7 @@ static void fill_xsave(u8 *dest, struct kvm_vcpu *vcpu)
 	 * Copy each region from the possibly compacted offset to the
 	 * non-compacted offset.
 	 */
-	valid = xstate_bv & ~XSTATE_FPSSE;
+	valid = xstate_bv & ~XFEATURE_MASK_FPSSE;
 	while (valid) {
 		u64 feature = valid & -valid;
 		int index = fls64(feature) - 1;
@@ -2944,7 +2945,7 @@ static void load_xsave(struct kvm_vcpu *vcpu, u8 *src)
 	 * Copy each region from the non-compacted offset to the
 	 * possibly compacted offset.
 	 */
-	valid = xstate_bv & ~XSTATE_FPSSE;
+	valid = xstate_bv & ~XFEATURE_MASK_FPSSE;
 	while (valid) {
 		u64 feature = valid & -valid;
 		int index = fls64(feature) - 1;
@@ -2972,7 +2973,7 @@ static void kvm_vcpu_ioctl_x86_get_xsave(struct kvm_vcpu *vcpu,
 			&vcpu->arch.guest_fpu.state.fxsave,
 			sizeof(struct fxregs_state));
 		*(u64 *)&guest_xsave->region[XSAVE_HDR_OFFSET / sizeof(u32)] =
-			XSTATE_FPSSE;
+			XFEATURE_MASK_FPSSE;
 	}
 }
 
@@ -2992,7 +2993,7 @@ static int kvm_vcpu_ioctl_x86_set_xsave(struct kvm_vcpu *vcpu,
 			return -EINVAL;
 		load_xsave(vcpu, (u8 *)guest_xsave->region);
 	} else {
-		if (xstate_bv & ~XSTATE_FPSSE)
+		if (xstate_bv & ~XFEATURE_MASK_FPSSE)
 			return -EINVAL;
 		memcpy(&vcpu->arch.guest_fpu.state.fxsave,
 			guest_xsave->region, sizeof(struct fxregs_state));
@@ -7001,7 +7002,7 @@ static void fx_init(struct kvm_vcpu *vcpu)
 	/*
 	 * Ensure guest xcr0 is valid for loading
 	 */
-	vcpu->arch.xcr0 = XSTATE_FP;
+	vcpu->arch.xcr0 = XFEATURE_MASK_FP;
 
 	vcpu->arch.cr0 |= X86_CR0_ET;
 }
arch/x86/kvm/x86.h +3 -3
index 2f822cd886c2..f2afa5fe48a6 100644
@@ -180,9 +180,9 @@ int kvm_mtrr_get_msr(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata);
 bool kvm_mtrr_check_gfn_range_consistency(struct kvm_vcpu *vcpu, gfn_t gfn,
 					  int page_num);
 
-#define KVM_SUPPORTED_XCR0     (XSTATE_FP | XSTATE_SSE | XSTATE_YMM \
-				| XSTATE_BNDREGS | XSTATE_BNDCSR \
-				| XSTATE_AVX512)
+#define KVM_SUPPORTED_XCR0     (XFEATURE_MASK_FP | XFEATURE_MASK_SSE \
+				| XFEATURE_MASK_YMM | XFEATURE_MASK_BNDREGS \
+				| XFEATURE_MASK_BNDCSR | XFEATURE_MASK_AVX512)
 extern u64 host_xcr0;
 
 extern u64 kvm_supported_xcr0(void);
arch/x86/math-emu/fpu_aux.c +70 -0

76 lines not on this page · raw message

arch/x86/math-emu/fpu_entry.c +31 -65

116 lines not on this page · raw message

arch/x86/math-emu/fpu_proto.h +12 -0
index 9779df436b7d..caff438b9c1d 100644
@@ -46,6 +46,14 @@ extern void fstsw_(void);
 extern void fp_nop(void);
 extern void fld_i_(void);
 extern void fxch_i(void);
+extern void fcmovb(void);
+extern void fcmove(void);
+extern void fcmovbe(void);
+extern void fcmovu(void);
+extern void fcmovnb(void);
+extern void fcmovne(void);
+extern void fcmovnbe(void);
+extern void fcmovnu(void);
 extern void ffree_(void);
 extern void ffreep(void);
 extern void fst_i_(void);
@@ -108,6 +116,10 @@ extern void fcompp(void);
 extern void fucom_(void);
 extern void fucomp(void);
 extern void fucompp(void);
+extern void fcomi_(void);
+extern void fcomip(void);
+extern void fucomi_(void);
+extern void fucomip(void);
 /* reg_constant.c */
 extern void fconst(void);
 /* reg_ld_str.c */
arch/x86/math-emu/load_store.c +51 -12

115 lines not on this page · raw message

arch/x86/math-emu/reg_compare.c +128 -0

143 lines not on this page · raw message

arch/x86/mm/mpx.c +8 -7
index 134948b0926f..b0ae85f90f10 100644
@@ -237,7 +237,8 @@ static int mpx_insn_decode(struct insn *insn,
  */
 siginfo_t *mpx_generate_siginfo(struct pt_regs *regs)
 {
-	const struct bndreg *bndregs, *bndreg;
+	const struct mpx_bndreg_state *bndregs;
+	const struct mpx_bndreg *bndreg;
 	siginfo_t *info = NULL;
 	struct insn insn;
 	uint8_t bndregno;
@@ -258,13 +259,13 @@ siginfo_t *mpx_generate_siginfo(struct pt_regs *regs)
 		goto err_out;
 	}
 	/* get bndregs field from current task's xsave area */
-	bndregs = get_xsave_field_ptr(XSTATE_BNDREGS);
+	bndregs = get_xsave_field_ptr(XFEATURE_MASK_BNDREGS);
 	if (!bndregs) {
 		err = -EINVAL;
 		goto err_out;
 	}
 	/* now go select the individual register in the set of 4 */
-	bndreg = &bndregs[bndregno];
+	bndreg = &bndregs->bndreg[bndregno];
 
 	info = kzalloc(sizeof(*info), GFP_KERNEL);
 	if (!info) {
@@ -306,7 +307,7 @@ siginfo_t *mpx_generate_siginfo(struct pt_regs *regs)
 
 static __user void *mpx_get_bounds_dir(void)
 {
-	const struct bndcsr *bndcsr;
+	const struct mpx_bndcsr *bndcsr;
 
 	if (!cpu_feature_enabled(X86_FEATURE_MPX))
 		return MPX_INVALID_BOUNDS_DIR;
@@ -315,7 +316,7 @@ static __user void *mpx_get_bounds_dir(void)
 	 * The bounds directory pointer is stored in a register
 	 * only accessible if we first do an xsave.
 	 */
-	bndcsr = get_xsave_field_ptr(XSTATE_BNDCSR);
+	bndcsr = get_xsave_field_ptr(XFEATURE_MASK_BNDCSR);
 	if (!bndcsr)
 		return MPX_INVALID_BOUNDS_DIR;
 
@@ -489,10 +490,10 @@ static int allocate_bt(struct mm_struct *mm, long __user *bd_entry)
 static int do_mpx_bt_fault(void)
 {
 	unsigned long bd_entry, bd_base;
-	const struct bndcsr *bndcsr;
+	const struct mpx_bndcsr *bndcsr;
 	struct mm_struct *mm = current->mm;
 
-	bndcsr = get_xsave_field_ptr(XSTATE_BNDCSR);
+	bndcsr = get_xsave_field_ptr(XFEATURE_MASK_BNDCSR);
 	if (!bndcsr)
 		return -EINVAL;
 	/*
tools/testing/selftests/x86/Makefile +3 -2
index 29089b24d18b..7145b3d9030c 100644
@@ -5,7 +5,8 @@ include ../lib.mk
 .PHONY: all all_32 all_64 warn_32bit_failure clean
 
 TARGETS_C_BOTHBITS := single_step_syscall sysret_ss_attrs ldt_gdt syscall_nt
-TARGETS_C_32BIT_ONLY := entry_from_vm86 syscall_arg_fault sigreturn
+TARGETS_C_32BIT_ONLY := entry_from_vm86 syscall_arg_fault sigreturn \
+			test_FCMOV test_FCOMI test_FISTTP
 
 TARGETS_C_32BIT_ALL := $(TARGETS_C_BOTHBITS) $(TARGETS_C_32BIT_ONLY)
 BINARIES_32 := $(TARGETS_C_32BIT_ALL:%=%_32)
@@ -35,7 +36,7 @@ all_64: $(BINARIES_64)
 	$(RM) $(BINARIES_32) $(BINARIES_64)
 
 $(TARGETS_C_32BIT_ALL:%=%_32): %_32: %.c
-	$(CC) -m32 -o $@ $(CFLAGS) $(EXTRA_CFLAGS) $^ -lrt -ldl
+	$(CC) -m32 -o $@ $(CFLAGS) $(EXTRA_CFLAGS) $^ -lrt -ldl -lm
 
 $(TARGETS_C_BOTHBITS:%=%_64): %_64: %.c
 	$(CC) -m64 -o $@ $(CFLAGS) $(EXTRA_CFLAGS) $^ -lrt -ldl
tools/testing/selftests/x86/test_FCMOV.c new file +93 -0

93 lines not on this page · raw message

tools/testing/selftests/x86/test_FCOMI.c new file +331 -0

331 lines not on this page · raw message

tools/testing/selftests/x86/test_FISTTP.c new file +137 -0

137 lines not on this page · raw message