Merge branch 'x86-fpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 FPU updates from Ingo Molnar:
 "This tree contains two main changes:

   - The big FPU code rewrite: wide reaching cleanups and reorganization
     that pulls all the FPU code together into a clean base in
     arch/x86/fpu/.

     The resulting code is leaner and faster, and much easier to
     understand.  This enables future work to further simplify the FPU
     code (such as removing lazy FPU restores).

     By its nature these changes have a substantial regression risk: FPU
     code related bugs are long lived, because races are often subtle
     and bugs mask as user-space failures that are difficult to track
     back to kernel side backs.  I'm aware of no unfixed (or even
     suspected) FPU related regression so far.

   - MPX support rework/fixes.  As this is still not a released CPU
     feature, there were some buglets in the code - should be much more
     robust now (Dave Hansen)"

* 'x86-fpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (250 commits)
  x86/fpu: Fix double-increment in setup_xstate_features()
  x86/mpx: Allow 32-bit binaries on 64-bit kernels again
  x86/mpx: Do not count MPX VMAs as neighbors when unmapping
  x86/mpx: Rewrite the unmap code
  x86/mpx: Support 32-bit binaries on 64-bit kernels
  x86/mpx: Use 32-bit-only cmpxchg() for 32-bit apps
  x86/mpx: Introduce new 'directory entry' to 'addr' helper function
  x86/mpx: Add temporary variable to reduce masking
  x86: Make is_64bit_mm() widely available
  x86/mpx: Trace allocation of new bounds tables
  x86/mpx: Trace the attempts to find bounds tables
  x86/mpx: Trace entry to bounds exception paths
  x86/mpx: Trace #BR exceptions
  x86/mpx: Introduce a boot-time disable flag
  x86/mpx: Restrict the mmap() size check to bounds tables
  x86/mpx: Remove redundant MPX_BNDCFG_ADDR_MASK
  x86/mpx: Clean up the code by not passing a task pointer around when unnecessary
  x86/mpx: Use the new get_xsave_field_ptr()API
  x86/fpu/xstate: Wrap get_xsave_addr() to make it safer
  x86/fpu/xstate: Fix up bad get_xsave_addr() assumptions
  ...

1 files changed, 28 insertions, 10 deletions

diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c
index 1ea14fd53933..206996c1669d 100644
--- a/arch/x86/kernel/signal.c
+++ b/arch/x86/kernel/signal.c
@@ -26,8 +26,8 @@
 
 #include <asm/processor.h>
 #include <asm/ucontext.h>
-#include <asm/i387.h>
-#include <asm/fpu-internal.h>
+#include <asm/fpu/internal.h>
+#include <asm/fpu/signal.h>
 #include <asm/vdso.h>
 #include <asm/mce.h>
 #include <asm/sighandling.h>
@@ -103,7 +103,7 @@ int restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc)
 		get_user_ex(buf, &sc->fpstate);
 	} get_user_catch(err);
 
-	err |= restore_xstate_sig(buf, config_enabled(CONFIG_X86_32));
+	err |= fpu__restore_sig(buf, config_enabled(CONFIG_X86_32));
 
 	force_iret();
 
@@ -199,6 +199,7 @@ get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size,
 	unsigned long sp = regs->sp;
 	unsigned long buf_fx = 0;
 	int onsigstack = on_sig_stack(sp);
+	struct fpu *fpu = &current->thread.fpu;
 
 	/* redzone */
 	if (config_enabled(CONFIG_X86_64))
@@ -218,9 +219,9 @@ get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size,
 		}
 	}
 
-	if (used_math()) {
-		sp = alloc_mathframe(sp, config_enabled(CONFIG_X86_32),
-				     &buf_fx, &math_size);
+	if (fpu->fpstate_active) {
+		sp = fpu__alloc_mathframe(sp, config_enabled(CONFIG_X86_32),
+					  &buf_fx, &math_size);
 		*fpstate = (void __user *)sp;
 	}
 
@@ -234,8 +235,8 @@ get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size,
 		return (void __user *)-1L;
 
 	/* save i387 and extended state */
-	if (used_math() &&
-	    save_xstate_sig(*fpstate, (void __user *)buf_fx, math_size) < 0)
+	if (fpu->fpstate_active &&
+	    copy_fpstate_to_sigframe(*fpstate, (void __user *)buf_fx, math_size) < 0)
 		return (void __user *)-1L;
 
 	return (void __user *)sp;
@@ -593,6 +594,22 @@ badframe:
 	return 0;
 }
 
+static inline int is_ia32_compat_frame(void)
+{
+	return config_enabled(CONFIG_IA32_EMULATION) &&
+	       test_thread_flag(TIF_IA32);
+}
+
+static inline int is_ia32_frame(void)
+{
+	return config_enabled(CONFIG_X86_32) || is_ia32_compat_frame();
+}
+
+static inline int is_x32_frame(void)
+{
+	return config_enabled(CONFIG_X86_X32_ABI) && test_thread_flag(TIF_X32);
+}
+
 static int
 setup_rt_frame(struct ksignal *ksig, struct pt_regs *regs)
 {
@@ -617,6 +634,7 @@ static void
 handle_signal(struct ksignal *ksig, struct pt_regs *regs)
 {
 	bool stepping, failed;
+	struct fpu *fpu = &current->thread.fpu;
 
 	/* Are we from a system call? */
 	if (syscall_get_nr(current, regs) >= 0) {
@@ -665,8 +683,8 @@ handle_signal(struct ksignal *ksig, struct pt_regs *regs)
 		/*
 		 * Ensure the signal handler starts with the new fpu state.
 		 */
-		if (used_math())
-			fpu_reset_state(current);
+		if (fpu->fpstate_active)
+			fpu__clear(fpu);
 	}
 	signal_setup_done(failed, ksig, stepping);
 }