Merge tag 'x86-fpu-2021-11-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 fpu updates from Thomas Gleixner:

 - Cleanup of extable fixup handling to be more robust, which in turn
   allows to make the FPU exception fixups more robust as well.

 - Change the return code for signal frame related failures from
   explicit error codes to a boolean fail/success as that's all what the
   calling code evaluates.

 - A large refactoring of the FPU code to prepare for adding AMX
   support:

      - Distangle the public header maze and remove especially the
        misnomed kitchen sink internal.h which is despite it's name
        included all over the place.

      - Add a proper abstraction for the register buffer storage (struct
        fpstate) which allows to dynamically size the buffer at runtime
        by flipping the pointer to the buffer container from the default
        container which is embedded in task_struct::tread::fpu to a
        dynamically allocated container with a larger register buffer.

      - Convert the code over to the new fpstate mechanism.

      - Consolidate the KVM FPU handling by moving the FPU related code
        into the FPU core which removes the number of exports and avoids
        adding even more export when AMX has to be supported in KVM.
        This also removes duplicated code which was of course
        unnecessary different and incomplete in the KVM copy.

      - Simplify the KVM FPU buffer handling by utilizing the new
        fpstate container and just switching the buffer pointer from the
        user space buffer to the KVM guest buffer when entering
        vcpu_run() and flipping it back when leaving the function. This
        cuts the memory requirements of a vCPU for FPU buffers in half
        and avoids pointless memory copy operations.

        This also solves the so far unresolved problem of adding AMX
        support because the current FPU buffer handling of KVM inflicted
        a circular dependency between adding AMX support to the core and
        to KVM. With the new scheme of switching fpstate AMX support can
        be added to the core code without affecting KVM.

      - Replace various variables with proper data structures so the
        extra information required for adding dynamically enabled FPU
        features (AMX) can be added in one place

 - Add AMX (Advanced Matrix eXtensions) support (finally):

   AMX is a large XSTATE component which is going to be available with
   Saphire Rapids XEON CPUs. The feature comes with an extra MSR
   (MSR_XFD) which allows to trap the (first) use of an AMX related
   instruction, which has two benefits:

    1) It allows the kernel to control access to the feature

    2) It allows the kernel to dynamically allocate the large register
       state buffer instead of burdening every task with the the extra
       8K or larger state storage.

   It would have been great to gain this kind of control already with
   AVX512.

   The support comes with the following infrastructure components:

    1) arch_prctl() to
        - read the supported features (equivalent to XGETBV(0))
        - read the permitted features for a task
        - request permission for a dynamically enabled feature

       Permission is granted per process, inherited on fork() and
       cleared on exec(). The permission policy of the kernel is
       restricted to sigaltstack size validation, but the syscall
       obviously allows further restrictions via seccomp etc.

    2) A stronger sigaltstack size validation for sys_sigaltstack(2)
       which takes granted permissions and the potentially resulting
       larger signal frame into account. This mechanism can also be used
       to enforce factual sigaltstack validation independent of dynamic
       features to help with finding potential victims of the 2K
       sigaltstack size constant which is broken since AVX512 support
       was added.

    3) Exception handling for #NM traps to catch first use of a extended
       feature via a new cause MSR. If the exception was caused by the
       use of such a feature, the handler checks permission for that
       feature. If permission has not been granted, the handler sends a
       SIGILL like the #UD handler would do if the feature would have
       been disabled in XCR0. If permission has been granted, then a new
       fpstate which fits the larger buffer requirement is allocated.

       In the unlikely case that this allocation fails, the handler
       sends SIGSEGV to the task. That's not elegant, but unavoidable as
       the other discussed options of preallocation or full per task
       permissions come with their own set of horrors for kernel and/or
       userspace. So this is the lesser of the evils and SIGSEGV caused
       by unexpected memory allocation failures is not a fundamentally
       new concept either.

       When allocation succeeds, the fpstate properties are filled in to
       reflect the extended feature set and the resulting sizes, the
       fpu::fpstate pointer is updated accordingly and the trap is
       disarmed for this task permanently.

    4) Enumeration and size calculations

    5) Trap switching via MSR_XFD

       The XFD (eXtended Feature Disable) MSR is context switched with
       the same life time rules as the FPU register state itself. The
       mechanism is keyed off with a static key which is default
       disabled so !AMX equipped CPUs have zero overhead. On AMX enabled
       CPUs the overhead is limited by comparing the tasks XFD value
       with a per CPU shadow variable to avoid redundant MSR writes. In
       case of switching from a AMX using task to a non AMX using task
       or vice versa, the extra MSR write is obviously inevitable.

       All other places which need to be aware of the variable feature
       sets and resulting variable sizes are not affected at all because
       they retrieve the information (feature set, sizes) unconditonally
       from the fpstate properties.

    6) Enable the new AMX states

   Note, this is relatively new code despite the fact that AMX support
   is in the works for more than a year now.

   The big refactoring of the FPU code, which allowed to do a proper
   integration has been started exactly 3 weeks ago. Refactoring of the
   existing FPU code and of the original AMX patches took a week and has
   been subject to extensive review and testing. The only fallout which
   has not been caught in review and testing right away was restricted
   to AMX enabled systems, which is completely irrelevant for anyone
   outside Intel and their early access program. There might be dragons
   lurking as usual, but so far the fine grained refactoring has held up
   and eventual yet undetected fallout is bisectable and should be
   easily addressable before the 5.16 release. Famous last words...

   Many thanks to Chang Bae and Dave Hansen for working hard on this and
   also to the various test teams at Intel who reserved extra capacity
   to follow the rapid development of this closely which provides the
   confidence level required to offer this rather large update for
   inclusion into 5.16-rc1

* tag 'x86-fpu-2021-11-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (110 commits)
  Documentation/x86: Add documentation for using dynamic XSTATE features
  x86/fpu: Include vmalloc.h for vzalloc()
  selftests/x86/amx: Add context switch test
  selftests/x86/amx: Add test cases for AMX state management
  x86/fpu/amx: Enable the AMX feature in 64-bit mode
  x86/fpu: Add XFD handling for dynamic states
  x86/fpu: Calculate the default sizes independently
  x86/fpu/amx: Define AMX state components and have it used for boot-time checks
  x86/fpu/xstate: Prepare XSAVE feature table for gaps in state component numbers
  x86/fpu/xstate: Add fpstate_realloc()/free()
  x86/fpu/xstate: Add XFD #NM handler
  x86/fpu: Update XFD state where required
  x86/fpu: Add sanity checks for XFD
  x86/fpu: Add XFD state to fpstate
  x86/msr-index: Add MSRs for XFD
  x86/cpufeatures: Add eXtended Feature Disabling (XFD) feature bit
  x86/fpu: Reset permission and fpstate on exec()
  x86/fpu: Prepare fpu_clone() for dynamically enabled features
  x86/fpu/signal: Prepare for variable sigframe length
  x86/signal: Use fpu::__state_user_size for sigalt stack validation
  ...

2 files changed, 852 insertions, 1 deletions

diff --git a/tools/testing/selftests/x86/Makefile b/tools/testing/selftests/x86/Makefile
index b4142cd1c5c2..8a1f62ab3c8e 100644
--- a/tools/testing/selftests/x86/Makefile
+++ b/tools/testing/selftests/x86/Makefile
@@ -18,7 +18,7 @@ TARGETS_C_32BIT_ONLY := entry_from_vm86 test_syscall_vdso unwind_vdso \
 			test_FCMOV test_FCOMI test_FISTTP \
 			vdso_restorer
 TARGETS_C_64BIT_ONLY := fsgsbase sysret_rip syscall_numbering \
-			corrupt_xstate_header
+			corrupt_xstate_header amx
 # Some selftests require 32bit support enabled also on 64bit systems
 TARGETS_C_32BIT_NEEDED := ldt_gdt ptrace_syscall
 
diff --git a/tools/testing/selftests/x86/amx.c b/tools/testing/selftests/x86/amx.c
new file mode 100644
index 000000000000..3615ef4a48bb
--- /dev/null
+++ b/tools/testing/selftests/x86/amx.c
@@ -0,0 +1,851 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#define _GNU_SOURCE
+#include <err.h>
+#include <errno.h>
+#include <pthread.h>
+#include <setjmp.h>
+#include <stdio.h>
+#include <string.h>
+#include <stdbool.h>
+#include <unistd.h>
+#include <x86intrin.h>
+
+#include <sys/auxv.h>
+#include <sys/mman.h>
+#include <sys/shm.h>
+#include <sys/syscall.h>
+#include <sys/wait.h>
+
+#ifndef __x86_64__
+# error This test is 64-bit only
+#endif
+
+#define XSAVE_HDR_OFFSET	512
+#define XSAVE_HDR_SIZE		64
+
+struct xsave_buffer {
+	union {
+		struct {
+			char legacy[XSAVE_HDR_OFFSET];
+			char header[XSAVE_HDR_SIZE];
+			char extended[0];
+		};
+		char bytes[0];
+	};
+};
+
+static inline uint64_t xgetbv(uint32_t index)
+{
+	uint32_t eax, edx;
+
+	asm volatile("xgetbv;"
+		     : "=a" (eax), "=d" (edx)
+		     : "c" (index));
+	return eax + ((uint64_t)edx << 32);
+}
+
+static inline void cpuid(uint32_t *eax, uint32_t *ebx, uint32_t *ecx, uint32_t *edx)
+{
+	asm volatile("cpuid;"
+		     : "=a" (*eax), "=b" (*ebx), "=c" (*ecx), "=d" (*edx)
+		     : "0" (*eax), "2" (*ecx));
+}
+
+static inline void xsave(struct xsave_buffer *xbuf, uint64_t rfbm)
+{
+	uint32_t rfbm_lo = rfbm;
+	uint32_t rfbm_hi = rfbm >> 32;
+
+	asm volatile("xsave (%%rdi)"
+		     : : "D" (xbuf), "a" (rfbm_lo), "d" (rfbm_hi)
+		     : "memory");
+}
+
+static inline void xrstor(struct xsave_buffer *xbuf, uint64_t rfbm)
+{
+	uint32_t rfbm_lo = rfbm;
+	uint32_t rfbm_hi = rfbm >> 32;
+
+	asm volatile("xrstor (%%rdi)"
+		     : : "D" (xbuf), "a" (rfbm_lo), "d" (rfbm_hi));
+}
+
+/* err() exits and will not return */
+#define fatal_error(msg, ...)	err(1, "[FAIL]\t" msg, ##__VA_ARGS__)
+
+static void sethandler(int sig, void (*handler)(int, siginfo_t *, void *),
+		       int flags)
+{
+	struct sigaction sa;
+
+	memset(&sa, 0, sizeof(sa));
+	sa.sa_sigaction = handler;
+	sa.sa_flags = SA_SIGINFO | flags;
+	sigemptyset(&sa.sa_mask);
+	if (sigaction(sig, &sa, 0))
+		fatal_error("sigaction");
+}
+
+static void clearhandler(int sig)
+{
+	struct sigaction sa;
+
+	memset(&sa, 0, sizeof(sa));
+	sa.sa_handler = SIG_DFL;
+	sigemptyset(&sa.sa_mask);
+	if (sigaction(sig, &sa, 0))
+		fatal_error("sigaction");
+}
+
+#define XFEATURE_XTILECFG	17
+#define XFEATURE_XTILEDATA	18
+#define XFEATURE_MASK_XTILECFG	(1 << XFEATURE_XTILECFG)
+#define XFEATURE_MASK_XTILEDATA	(1 << XFEATURE_XTILEDATA)
+#define XFEATURE_MASK_XTILE	(XFEATURE_MASK_XTILECFG | XFEATURE_MASK_XTILEDATA)
+
+#define CPUID_LEAF1_ECX_XSAVE_MASK	(1 << 26)
+#define CPUID_LEAF1_ECX_OSXSAVE_MASK	(1 << 27)
+static inline void check_cpuid_xsave(void)
+{
+	uint32_t eax, ebx, ecx, edx;
+
+	/*
+	 * CPUID.1:ECX.XSAVE[bit 26] enumerates general
+	 * support for the XSAVE feature set, including
+	 * XGETBV.
+	 */
+	eax = 1;
+	ecx = 0;
+	cpuid(&eax, &ebx, &ecx, &edx);
+	if (!(ecx & CPUID_LEAF1_ECX_XSAVE_MASK))
+		fatal_error("cpuid: no CPU xsave support");
+	if (!(ecx & CPUID_LEAF1_ECX_OSXSAVE_MASK))
+		fatal_error("cpuid: no OS xsave support");
+}
+
+static uint32_t xbuf_size;
+
+static struct {
+	uint32_t xbuf_offset;
+	uint32_t size;
+} xtiledata;
+
+#define CPUID_LEAF_XSTATE		0xd
+#define CPUID_SUBLEAF_XSTATE_USER	0x0
+#define TILE_CPUID			0x1d
+#define TILE_PALETTE_ID			0x1
+
+static void check_cpuid_xtiledata(void)
+{
+	uint32_t eax, ebx, ecx, edx;
+
+	eax = CPUID_LEAF_XSTATE;
+	ecx = CPUID_SUBLEAF_XSTATE_USER;
+	cpuid(&eax, &ebx, &ecx, &edx);
+
+	/*
+	 * EBX enumerates the size (in bytes) required by the XSAVE
+	 * instruction for an XSAVE area containing all the user state
+	 * components corresponding to bits currently set in XCR0.
+	 *
+	 * Stash that off so it can be used to allocate buffers later.
+	 */
+	xbuf_size = ebx;
+
+	eax = CPUID_LEAF_XSTATE;
+	ecx = XFEATURE_XTILEDATA;
+
+	cpuid(&eax, &ebx, &ecx, &edx);
+	/*
+	 * eax: XTILEDATA state component size
+	 * ebx: XTILEDATA state component offset in user buffer
+	 */
+	if (!eax || !ebx)
+		fatal_error("xstate cpuid: invalid tile data size/offset: %d/%d",
+				eax, ebx);
+
+	xtiledata.size	      = eax;
+	xtiledata.xbuf_offset = ebx;
+}
+
+/* The helpers for managing XSAVE buffer and tile states: */
+
+struct xsave_buffer *alloc_xbuf(void)
+{
+	struct xsave_buffer *xbuf;
+
+	/* XSAVE buffer should be 64B-aligned. */
+	xbuf = aligned_alloc(64, xbuf_size);
+	if (!xbuf)
+		fatal_error("aligned_alloc()");
+	return xbuf;
+}
+
+static inline void clear_xstate_header(struct xsave_buffer *buffer)
+{
+	memset(&buffer->header, 0, sizeof(buffer->header));
+}
+
+static inline uint64_t get_xstatebv(struct xsave_buffer *buffer)
+{
+	/* XSTATE_BV is at the beginning of the header: */
+	return *(uint64_t *)&buffer->header;
+}
+
+static inline void set_xstatebv(struct xsave_buffer *buffer, uint64_t bv)
+{
+	/* XSTATE_BV is at the beginning of the header: */
+	*(uint64_t *)(&buffer->header) = bv;
+}
+
+static void set_rand_tiledata(struct xsave_buffer *xbuf)
+{
+	int *ptr = (int *)&xbuf->bytes[xtiledata.xbuf_offset];
+	int data;
+	int i;
+
+	/*
+	 * Ensure that 'data' is never 0.  This ensures that
+	 * the registers are never in their initial configuration
+	 * and thus never tracked as being in the init state.
+	 */
+	data = rand() | 1;
+
+	for (i = 0; i < xtiledata.size / sizeof(int); i++, ptr++)
+		*ptr = data;
+}
+
+struct xsave_buffer *stashed_xsave;
+
+static void init_stashed_xsave(void)
+{
+	stashed_xsave = alloc_xbuf();
+	if (!stashed_xsave)
+		fatal_error("failed to allocate stashed_xsave\n");
+	clear_xstate_header(stashed_xsave);
+}
+
+static void free_stashed_xsave(void)
+{
+	free(stashed_xsave);
+}
+
+/* See 'struct _fpx_sw_bytes' at sigcontext.h */
+#define SW_BYTES_OFFSET		464
+/* N.B. The struct's field name varies so read from the offset. */
+#define SW_BYTES_BV_OFFSET	(SW_BYTES_OFFSET + 8)
+
+static inline struct _fpx_sw_bytes *get_fpx_sw_bytes(void *buffer)
+{
+	return (struct _fpx_sw_bytes *)(buffer + SW_BYTES_OFFSET);
+}
+
+static inline uint64_t get_fpx_sw_bytes_features(void *buffer)
+{
+	return *(uint64_t *)(buffer + SW_BYTES_BV_OFFSET);
+}
+
+/* Work around printf() being unsafe in signals: */
+#define SIGNAL_BUF_LEN 1000
+char signal_message_buffer[SIGNAL_BUF_LEN];
+void sig_print(char *msg)
+{
+	int left = SIGNAL_BUF_LEN - strlen(signal_message_buffer) - 1;
+
+	strncat(signal_message_buffer, msg, left);
+}
+
+static volatile bool noperm_signaled;
+static int noperm_errs;
+/*
+ * Signal handler for when AMX is used but
+ * permission has not been obtained.
+ */
+static void handle_noperm(int sig, siginfo_t *si, void *ctx_void)
+{
+	ucontext_t *ctx = (ucontext_t *)ctx_void;
+	void *xbuf = ctx->uc_mcontext.fpregs;
+	struct _fpx_sw_bytes *sw_bytes;
+	uint64_t features;
+
+	/* Reset the signal message buffer: */
+	signal_message_buffer[0] = '\0';
+	sig_print("\tAt SIGILL handler,\n");
+
+	if (si->si_code != ILL_ILLOPC) {
+		noperm_errs++;
+		sig_print("[FAIL]\tInvalid signal code.\n");
+	} else {
+		sig_print("[OK]\tValid signal code (ILL_ILLOPC).\n");
+	}
+
+	sw_bytes = get_fpx_sw_bytes(xbuf);
+	/*
+	 * Without permission, the signal XSAVE buffer should not
+	 * have room for AMX register state (aka. xtiledata).
+	 * Check that the size does not overlap with where xtiledata
+	 * will reside.
+	 *
+	 * This also implies that no state components *PAST*
+	 * XTILEDATA (features >=19) can be present in the buffer.
+	 */
+	if (sw_bytes->xstate_size <= xtiledata.xbuf_offset) {
+		sig_print("[OK]\tValid xstate size\n");
+	} else {
+		noperm_errs++;
+		sig_print("[FAIL]\tInvalid xstate size\n");
+	}
+
+	features = get_fpx_sw_bytes_features(xbuf);
+	/*
+	 * Without permission, the XTILEDATA feature
+	 * bit should not be set.
+	 */
+	if ((features & XFEATURE_MASK_XTILEDATA) == 0) {
+		sig_print("[OK]\tValid xstate mask\n");
+	} else {
+		noperm_errs++;
+		sig_print("[FAIL]\tInvalid xstate mask\n");
+	}
+
+	noperm_signaled = true;
+	ctx->uc_mcontext.gregs[REG_RIP] += 3; /* Skip the faulting XRSTOR */
+}
+
+/* Return true if XRSTOR is successful; otherwise, false. */
+static inline bool xrstor_safe(struct xsave_buffer *xbuf, uint64_t mask)
+{
+	noperm_signaled = false;
+	xrstor(xbuf, mask);
+
+	/* Print any messages produced by the signal code: */
+	printf("%s", signal_message_buffer);
+	/*
+	 * Reset the buffer to make sure any future printing
+	 * only outputs new messages:
+	 */
+	signal_message_buffer[0] = '\0';
+
+	if (noperm_errs)
+		fatal_error("saw %d errors in noperm signal handler\n", noperm_errs);
+
+	return !noperm_signaled;
+}
+
+/*
+ * Use XRSTOR to populate the XTILEDATA registers with
+ * random data.
+ *
+ * Return true if successful; otherwise, false.
+ */
+static inline bool load_rand_tiledata(struct xsave_buffer *xbuf)
+{
+	clear_xstate_header(xbuf);
+	set_xstatebv(xbuf, XFEATURE_MASK_XTILEDATA);
+	set_rand_tiledata(xbuf);
+	return xrstor_safe(xbuf, XFEATURE_MASK_XTILEDATA);
+}
+
+/* Return XTILEDATA to its initial configuration. */
+static inline void init_xtiledata(void)
+{
+	clear_xstate_header(stashed_xsave);
+	xrstor_safe(stashed_xsave, XFEATURE_MASK_XTILEDATA);
+}
+
+enum expected_result { FAIL_EXPECTED, SUCCESS_EXPECTED };
+
+/* arch_prctl() and sigaltstack() test */
+
+#define ARCH_GET_XCOMP_PERM	0x1022
+#define ARCH_REQ_XCOMP_PERM	0x1023
+
+static void req_xtiledata_perm(void)
+{
+	syscall(SYS_arch_prctl, ARCH_REQ_XCOMP_PERM, XFEATURE_XTILEDATA);
+}
+
+static void validate_req_xcomp_perm(enum expected_result exp)
+{
+	unsigned long bitmask;
+	long rc;
+
+	rc = syscall(SYS_arch_prctl, ARCH_REQ_XCOMP_PERM, XFEATURE_XTILEDATA);
+	if (exp == FAIL_EXPECTED) {
+		if (rc) {
+			printf("[OK]\tARCH_REQ_XCOMP_PERM saw expected failure..\n");
+			return;
+		}
+
+		fatal_error("ARCH_REQ_XCOMP_PERM saw unexpected success.\n");
+	} else if (rc) {
+		fatal_error("ARCH_REQ_XCOMP_PERM saw unexpected failure.\n");
+	}
+
+	rc = syscall(SYS_arch_prctl, ARCH_GET_XCOMP_PERM, &bitmask);
+	if (rc) {
+		fatal_error("prctl(ARCH_GET_XCOMP_PERM) error: %ld", rc);
+	} else if (bitmask & XFEATURE_MASK_XTILE) {
+		printf("\tARCH_REQ_XCOMP_PERM is successful.\n");
+	}
+}
+
+static void validate_xcomp_perm(enum expected_result exp)
+{
+	bool load_success = load_rand_tiledata(stashed_xsave);
+
+	if (exp == FAIL_EXPECTED) {
+		if (load_success) {
+			noperm_errs++;
+			printf("[FAIL]\tLoad tiledata succeeded.\n");
+		} else {
+			printf("[OK]\tLoad tiledata failed.\n");
+		}
+	} else if (exp == SUCCESS_EXPECTED) {
+		if (load_success) {
+			printf("[OK]\tLoad tiledata succeeded.\n");
+		} else {
+			noperm_errs++;
+			printf("[FAIL]\tLoad tiledata failed.\n");
+		}
+	}
+}
+
+#ifndef AT_MINSIGSTKSZ
+#  define AT_MINSIGSTKSZ	51
+#endif
+
+static void *alloc_altstack(unsigned int size)
+{
+	void *altstack;
+
+	altstack = mmap(NULL, size, PROT_READ | PROT_WRITE,
+			MAP_PRIVATE | MAP_ANONYMOUS | MAP_STACK, -1, 0);
+
+	if (altstack == MAP_FAILED)
+		fatal_error("mmap() for altstack");
+
+	return altstack;
+}
+
+static void setup_altstack(void *addr, unsigned long size, enum expected_result exp)
+{
+	stack_t ss;
+	int rc;
+
+	memset(&ss, 0, sizeof(ss));
+	ss.ss_size = size;
+	ss.ss_sp = addr;
+
+	rc = sigaltstack(&ss, NULL);
+
+	if (exp == FAIL_EXPECTED) {
+		if (rc) {
+			printf("[OK]\tsigaltstack() failed.\n");
+		} else {
+			fatal_error("sigaltstack() succeeded unexpectedly.\n");
+		}
+	} else if (rc) {
+		fatal_error("sigaltstack()");
+	}
+}
+
+static void test_dynamic_sigaltstack(void)
+{
+	unsigned int small_size, enough_size;
+	unsigned long minsigstksz;
+	void *altstack;
+
+	minsigstksz = getauxval(AT_MINSIGSTKSZ);
+	printf("\tAT_MINSIGSTKSZ = %lu\n", minsigstksz);
+	/*
+	 * getauxval() itself can return 0 for failure or
+	 * success.  But, in this case, AT_MINSIGSTKSZ
+	 * will always return a >=0 value if implemented.
+	 * Just check for 0.
+	 */
+	if (minsigstksz == 0) {
+		printf("no support for AT_MINSIGSTKSZ, skipping sigaltstack tests\n");
+		return;
+	}
+
+	enough_size = minsigstksz * 2;
+
+	altstack = alloc_altstack(enough_size);
+	printf("\tAllocate memory for altstack (%u bytes).\n", enough_size);
+
+	/*
+	 * Try setup_altstack() with a size which can not fit
+	 * XTILEDATA.  ARCH_REQ_XCOMP_PERM should fail.
+	 */
+	small_size = minsigstksz - xtiledata.size;
+	printf("\tAfter sigaltstack() with small size (%u bytes).\n", small_size);
+	setup_altstack(altstack, small_size, SUCCESS_EXPECTED);
+	validate_req_xcomp_perm(FAIL_EXPECTED);
+
+	/*
+	 * Try setup_altstack() with a size derived from
+	 * AT_MINSIGSTKSZ.  It should be more than large enough
+	 * and thus ARCH_REQ_XCOMP_PERM should succeed.
+	 */
+	printf("\tAfter sigaltstack() with enough size (%u bytes).\n", enough_size);
+	setup_altstack(altstack, enough_size, SUCCESS_EXPECTED);
+	validate_req_xcomp_perm(SUCCESS_EXPECTED);
+
+	/*
+	 * Try to coerce setup_altstack() to again accept a
+	 * too-small altstack.  This ensures that big-enough
+	 * sigaltstacks can not shrink to a too-small value
+	 * once XTILEDATA permission is established.
+	 */
+	printf("\tThen, sigaltstack() with small size (%u bytes).\n", small_size);
+	setup_altstack(altstack, small_size, FAIL_EXPECTED);
+}
+
+static void test_dynamic_state(void)
+{
+	pid_t parent, child, grandchild;
+
+	parent = fork();
+	if (parent < 0) {
+		/* fork() failed */
+		fatal_error("fork");
+	} else if (parent > 0) {
+		int status;
+		/* fork() succeeded.  Now in the parent. */
+
+		wait(&status);
+		if (!WIFEXITED(status) || WEXITSTATUS(status))
+			fatal_error("arch_prctl test parent exit");
+		return;
+	}
+	/* fork() succeeded.  Now in the child . */
+
+	printf("[RUN]\tCheck ARCH_REQ_XCOMP_PERM around process fork() and sigaltack() test.\n");
+
+	printf("\tFork a child.\n");
+	child = fork();
+	if (child < 0) {
+		fatal_error("fork");
+	} else if (child > 0) {
+		int status;
+
+		wait(&status);
+		if (!WIFEXITED(status) || WEXITSTATUS(status))
+			fatal_error("arch_prctl test child exit");
+		_exit(0);
+	}
+
+	/*
+	 * The permission request should fail without an
+	 * XTILEDATA-compatible signal stack
+	 */
+	printf("\tTest XCOMP_PERM at child.\n");
+	validate_xcomp_perm(FAIL_EXPECTED);
+
+	/*
+	 * Set up an XTILEDATA-compatible signal stack and
+	 * also obtain permission to populate XTILEDATA.
+	 */
+	printf("\tTest dynamic sigaltstack at child:\n");
+	test_dynamic_sigaltstack();
+
+	/* Ensure that XTILEDATA can be populated. */
+	printf("\tTest XCOMP_PERM again at child.\n");
+	validate_xcomp_perm(SUCCESS_EXPECTED);
+
+	printf("\tFork a grandchild.\n");
+	grandchild = fork();
+	if (grandchild < 0) {
+		/* fork() failed */
+		fatal_error("fork");
+	} else if (!grandchild) {
+		/* fork() succeeded.  Now in the (grand)child. */
+		printf("\tTest XCOMP_PERM at grandchild.\n");
+
+		/*
+		 * Ensure that the grandchild inherited
+		 * permission and a compatible sigaltstack:
+		 */
+		validate_xcomp_perm(SUCCESS_EXPECTED);
+	} else {
+		int status;
+		/* fork() succeeded.  Now in the parent. */
+
+		wait(&status);
+		if (!WIFEXITED(status) || WEXITSTATUS(status))
+			fatal_error("fork test grandchild");
+	}
+
+	_exit(0);
+}
+
+/*
+ * Save current register state and compare it to @xbuf1.'
+ *
+ * Returns false if @xbuf1 matches the registers.
+ * Returns true  if @xbuf1 differs from the registers.
+ */
+static inline bool __validate_tiledata_regs(struct xsave_buffer *xbuf1)
+{
+	struct xsave_buffer *xbuf2;
+	int ret;
+
+	xbuf2 = alloc_xbuf();
+	if (!xbuf2)
+		fatal_error("failed to allocate XSAVE buffer\n");
+
+	xsave(xbuf2, XFEATURE_MASK_XTILEDATA);
+	ret = memcmp(&xbuf1->bytes[xtiledata.xbuf_offset],
+		     &xbuf2->bytes[xtiledata.xbuf_offset],
+		     xtiledata.size);
+
+	free(xbuf2);
+
+	if (ret == 0)
+		return false;
+	return true;
+}
+
+static inline void validate_tiledata_regs_same(struct xsave_buffer *xbuf)
+{
+	int ret = __validate_tiledata_regs(xbuf);
+
+	if (ret != 0)
+		fatal_error("TILEDATA registers changed");
+}
+
+static inline void validate_tiledata_regs_changed(struct xsave_buffer *xbuf)
+{
+	int ret = __validate_tiledata_regs(xbuf);
+
+	if (ret == 0)
+		fatal_error("TILEDATA registers did not change");
+}
+
+/* tiledata inheritance test */
+
+static void test_fork(void)
+{
+	pid_t child, grandchild;
+
+	child = fork();
+	if (child < 0) {
+		/* fork() failed */
+		fatal_error("fork");
+	} else if (child > 0) {
+		/* fork() succeeded.  Now in the parent. */
+		int status;
+
+		wait(&status);
+		if (!WIFEXITED(status) || WEXITSTATUS(status))
+			fatal_error("fork test child");
+		return;
+	}
+	/* fork() succeeded.  Now in the child. */
+	printf("[RUN]\tCheck tile data inheritance.\n\tBefore fork(), load tiledata\n");
+
+	load_rand_tiledata(stashed_xsave);
+
+	grandchild = fork();
+	if (grandchild < 0) {
+		/* fork() failed */
+		fatal_error("fork");
+	} else if (grandchild > 0) {
+		/* fork() succeeded.  Still in the first child. */
+		int status;
+
+		wait(&status);
+		if (!WIFEXITED(status) || WEXITSTATUS(status))
+			fatal_error("fork test grand child");
+		_exit(0);
+	}
+	/* fork() succeeded.  Now in the (grand)child. */
+
+	/*
+	 * TILEDATA registers are not preserved across fork().
+	 * Ensure that their value has changed:
+	 */
+	validate_tiledata_regs_changed(stashed_xsave);
+
+	_exit(0);
+}
+
+/* Context switching test */
+
+static struct _ctxtswtest_cfg {
+	unsigned int iterations;
+	unsigned int num_threads;
+} ctxtswtest_config;
+
+struct futex_info {
+	pthread_t thread;
+	int nr;
+	pthread_mutex_t mutex;
+	struct futex_info *next;
+};
+
+static void *check_tiledata(void *info)
+{
+	struct futex_info *finfo = (struct futex_info *)info;
+	struct xsave_buffer *xbuf;
+	int i;
+
+	xbuf = alloc_xbuf();
+	if (!xbuf)
+		fatal_error("unable to allocate XSAVE buffer");
+
+	/*
+	 * Load random data into 'xbuf' and then restore
+	 * it to the tile registers themselves.
+	 */
+	load_rand_tiledata(xbuf);
+	for (i = 0; i < ctxtswtest_config.iterations; i++) {
+		pthread_mutex_lock(&finfo->mutex);
+
+		/*
+		 * Ensure the register values have not
+		 * diverged from those recorded in 'xbuf'.
+		 */
+		validate_tiledata_regs_same(xbuf);
+
+		/* Load new, random values into xbuf and registers */
+		load_rand_tiledata(xbuf);
+
+		/*
+		 * The last thread's last unlock will be for
+		 * thread 0's mutex.  However, thread 0 will
+		 * have already exited the loop and the mutex
+		 * will already be unlocked.
+		 *
+		 * Because this is not an ERRORCHECK mutex,
+		 * that inconsistency will be silently ignored.
+		 */
+		pthread_mutex_unlock(&finfo->next->mutex);
+	}
+
+	free(xbuf);
+	/*
+	 * Return this thread's finfo, which is
+	 * a unique value for this thread.
+	 */
+	return finfo;
+}
+
+static int create_threads(int num, struct futex_info *finfo)
+{
+	int i;
+
+	for (i = 0; i < num; i++) {
+		int next_nr;
+
+		finfo[i].nr = i;
+		/*
+		 * Thread 'i' will wait on this mutex to
+		 * be unlocked.  Lock it immediately after
+		 * initialization:
+		 */
+		pthread_mutex_init(&finfo[i].mutex, NULL);
+		pthread_mutex_lock(&finfo[i].mutex);
+
+		next_nr = (i + 1) % num;
+		finfo[i].next = &finfo[next_nr];
+
+		if (pthread_create(&finfo[i].thread, NULL, check_tiledata, &finfo[i]))
+			fatal_error("pthread_create()");
+	}
+	return 0;
+}
+
+static void affinitize_cpu0(void)
+{
+	cpu_set_t cpuset;
+
+	CPU_ZERO(&cpuset);
+	CPU_SET(0, &cpuset);
+
+	if (sched_setaffinity(0, sizeof(cpuset), &cpuset) != 0)
+		fatal_error("sched_setaffinity to CPU 0");
+}
+
+static void test_context_switch(void)
+{
+	struct futex_info *finfo;
+	int i;
+
+	/* Affinitize to one CPU to force context switches */
+	affinitize_cpu0();
+
+	req_xtiledata_perm();
+
+	printf("[RUN]\tCheck tiledata context switches, %d iterations, %d threads.\n",
+	       ctxtswtest_config.iterations,
+	       ctxtswtest_config.num_threads);
+
+
+	finfo = malloc(sizeof(*finfo) * ctxtswtest_config.num_threads);
+	if (!finfo)
+		fatal_error("malloc()");
+
+	create_threads(ctxtswtest_config.num_threads, finfo);
+
+	/*
+	 * This thread wakes up thread 0
+	 * Thread 0 will wake up 1
+	 * Thread 1 will wake up 2
+	 * ...
+	 * the last thread will wake up 0
+	 *
+	 * ... this will repeat for the configured
+	 * number of iterations.
+	 */
+	pthread_mutex_unlock(&finfo[0].mutex);
+
+	/* Wait for all the threads to finish: */
+	for (i = 0; i < ctxtswtest_config.num_threads; i++) {
+		void *thread_retval;
+		int rc;
+
+		rc = pthread_join(finfo[i].thread, &thread_retval);
+
+		if (rc)
+			fatal_error("pthread_join() failed for thread %d err: %d\n",
+					i, rc);
+
+		if (thread_retval != &finfo[i])
+			fatal_error("unexpected thread retval for thread %d: %p\n",
+					i, thread_retval);
+
+	}
+
+	printf("[OK]\tNo incorrect case was found.\n");
+
+	free(finfo);
+}
+
+int main(void)
+{
+	/* Check hardware availability at first */
+	check_cpuid_xsave();
+	check_cpuid_xtiledata();
+
+	init_stashed_xsave();
+	sethandler(SIGILL, handle_noperm, 0);
+
+	test_dynamic_state();
+
+	/* Request permission for the following tests */
+	req_xtiledata_perm();
+
+	test_fork();
+
+	ctxtswtest_config.iterations = 10;
+	ctxtswtest_config.num_threads = 5;
+	test_context_switch();
+
+	clearhandler(SIGILL);
+	free_stashed_xsave();
+
+	return 0;
+}