x86: Introduce 'struct fpu' and related API

Currently all fpu state access is through tsk->thread.xstate.  Since we wish
to generalize fpu access to non-task contexts, wrap the state in a new
'struct fpu' and convert existing access to use an fpu API.

Signal frame handlers are not converted to the API since they will remain
task context only things.

Signed-off-by: Avi Kivity <avi@redhat.com>
Acked-by: Suresh Siddha <suresh.b.siddha@intel.com>
LKML-Reference: <1273135546-29690-3-git-send-email-avi@redhat.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>

5 files changed, 63 insertions, 66 deletions

diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c
index 14ca1dc7a70..86cef6b3225 100644
--- a/arch/x86/kernel/i387.c
+++ b/arch/x86/kernel/i387.c
@@ -107,57 +107,57 @@ void __cpuinit fpu_init(void)
 }
 #endif	/* CONFIG_X86_64 */
 
-/*
- * The _current_ task is using the FPU for the first time
- * so initialize it and set the mxcsr to its default
- * value at reset if we support XMM instructions and then
- * remeber the current task has used the FPU.
- */
-int init_fpu(struct task_struct *tsk)
+static void fpu_finit(struct fpu *fpu)
 {
-	if (tsk_used_math(tsk)) {
-		if (HAVE_HWFP && tsk == current)
-			unlazy_fpu(tsk);
-		return 0;
-	}
-
-	/*
-	 * Memory allocation at the first usage of the FPU and other state.
-	 */
-	if (!tsk->thread.xstate) {
-		tsk->thread.xstate = kmem_cache_alloc(task_xstate_cachep,
-						      GFP_KERNEL);
-		if (!tsk->thread.xstate)
-			return -ENOMEM;
-	}
-
 #ifdef CONFIG_X86_32
 	if (!HAVE_HWFP) {
-		memset(tsk->thread.xstate, 0, xstate_size);
-		finit_task(tsk);
-		set_stopped_child_used_math(tsk);
-		return 0;
+		finit_soft_fpu(&fpu->state->soft);
+		return;
 	}
 #endif
 
 	if (cpu_has_fxsr) {
-		struct i387_fxsave_struct *fx = &tsk->thread.xstate->fxsave;
+		struct i387_fxsave_struct *fx = &fpu->state->fxsave;
 
 		memset(fx, 0, xstate_size);
 		fx->cwd = 0x37f;
 		if (cpu_has_xmm)
 			fx->mxcsr = MXCSR_DEFAULT;
 	} else {
-		struct i387_fsave_struct *fp = &tsk->thread.xstate->fsave;
+		struct i387_fsave_struct *fp = &fpu->state->fsave;
 		memset(fp, 0, xstate_size);
 		fp->cwd = 0xffff037fu;
 		fp->swd = 0xffff0000u;
 		fp->twd = 0xffffffffu;
 		fp->fos = 0xffff0000u;
 	}
+}
+
+/*
+ * The _current_ task is using the FPU for the first time
+ * so initialize it and set the mxcsr to its default
+ * value at reset if we support XMM instructions and then
+ * remeber the current task has used the FPU.
+ */
+int init_fpu(struct task_struct *tsk)
+{
+	int ret;
+
+	if (tsk_used_math(tsk)) {
+		if (HAVE_HWFP && tsk == current)
+			unlazy_fpu(tsk);
+		return 0;
+	}
+
 	/*
-	 * Only the device not available exception or ptrace can call init_fpu.
+	 * Memory allocation at the first usage of the FPU and other state.
 	 */
+	ret = fpu_alloc(&tsk->thread.fpu);
+	if (ret)
+		return ret;
+
+	fpu_finit(&tsk->thread.fpu);
+
 	set_stopped_child_used_math(tsk);
 	return 0;
 }
@@ -191,7 +191,7 @@ int xfpregs_get(struct task_struct *target, const struct user_regset *regset,
 		return ret;
 
 	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &target->thread.xstate->fxsave, 0, -1);
+				   &target->thread.fpu.state->fxsave, 0, -1);
 }
 
 int xfpregs_set(struct task_struct *target, const struct user_regset *regset,
@@ -208,19 +208,19 @@ int xfpregs_set(struct task_struct *target, const struct user_regset *regset,
 		return ret;
 
 	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
-				 &target->thread.xstate->fxsave, 0, -1);
+				 &target->thread.fpu.state->fxsave, 0, -1);
 
 	/*
 	 * mxcsr reserved bits must be masked to zero for security reasons.
 	 */
-	target->thread.xstate->fxsave.mxcsr &= mxcsr_feature_mask;
+	target->thread.fpu.state->fxsave.mxcsr &= mxcsr_feature_mask;
 
 	/*
 	 * update the header bits in the xsave header, indicating the
 	 * presence of FP and SSE state.
 	 */
 	if (cpu_has_xsave)
-		target->thread.xstate->xsave.xsave_hdr.xstate_bv |= XSTATE_FPSSE;
+		target->thread.fpu.state->xsave.xsave_hdr.xstate_bv |= XSTATE_FPSSE;
 
 	return ret;
 }
@@ -243,14 +243,14 @@ int xstateregs_get(struct task_struct *target, const struct user_regset *regset,
 	 * memory layout in the thread struct, so that we can copy the entire
 	 * xstateregs to the user using one user_regset_copyout().
 	 */
-	memcpy(&target->thread.xstate->fxsave.sw_reserved,
+	memcpy(&target->thread.fpu.state->fxsave.sw_reserved,
 	       xstate_fx_sw_bytes, sizeof(xstate_fx_sw_bytes));
 
 	/*
 	 * Copy the xstate memory layout.
 	 */
 	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.xstate->xsave, 0, -1);
+				  &target->thread.fpu.state->xsave, 0, -1);
 	return ret;
 }
 
@@ -269,14 +269,14 @@ int xstateregs_set(struct task_struct *target, const struct user_regset *regset,
 		return ret;
 
 	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
-				 &target->thread.xstate->xsave, 0, -1);
+				 &target->thread.fpu.state->xsave, 0, -1);
 
 	/*
 	 * mxcsr reserved bits must be masked to zero for security reasons.
 	 */
-	target->thread.xstate->fxsave.mxcsr &= mxcsr_feature_mask;
+	target->thread.fpu.state->fxsave.mxcsr &= mxcsr_feature_mask;
 
-	xsave_hdr = &target->thread.xstate->xsave.xsave_hdr;
+	xsave_hdr = &target->thread.fpu.state->xsave.xsave_hdr;
 
 	xsave_hdr->xstate_bv &= pcntxt_mask;
 	/*
@@ -362,7 +362,7 @@ static inline u32 twd_fxsr_to_i387(struct i387_fxsave_struct *fxsave)
 static void
 convert_from_fxsr(struct user_i387_ia32_struct *env, struct task_struct *tsk)
 {
-	struct i387_fxsave_struct *fxsave = &tsk->thread.xstate->fxsave;
+	struct i387_fxsave_struct *fxsave = &tsk->thread.fpu.state->fxsave;
 	struct _fpreg *to = (struct _fpreg *) &env->st_space[0];
 	struct _fpxreg *from = (struct _fpxreg *) &fxsave->st_space[0];
 	int i;
@@ -402,7 +402,7 @@ static void convert_to_fxsr(struct task_struct *tsk,
 			    const struct user_i387_ia32_struct *env)
 
 {
-	struct i387_fxsave_struct *fxsave = &tsk->thread.xstate->fxsave;
+	struct i387_fxsave_struct *fxsave = &tsk->thread.fpu.state->fxsave;
 	struct _fpreg *from = (struct _fpreg *) &env->st_space[0];
 	struct _fpxreg *to = (struct _fpxreg *) &fxsave->st_space[0];
 	int i;
@@ -442,7 +442,7 @@ int fpregs_get(struct task_struct *target, const struct user_regset *regset,
 
 	if (!cpu_has_fxsr) {
 		return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					   &target->thread.xstate->fsave, 0,
+					   &target->thread.fpu.state->fsave, 0,
 					   -1);
 	}
 
@@ -472,7 +472,7 @@ int fpregs_set(struct task_struct *target, const struct user_regset *regset,
 
 	if (!cpu_has_fxsr) {
 		return user_regset_copyin(&pos, &count, &kbuf, &ubuf,
-					  &target->thread.xstate->fsave, 0, -1);
+					  &target->thread.fpu.state->fsave, 0, -1);
 	}
 
 	if (pos > 0 || count < sizeof(env))
@@ -487,7 +487,7 @@ int fpregs_set(struct task_struct *target, const struct user_regset *regset,
 	 * presence of FP.
 	 */
 	if (cpu_has_xsave)
-		target->thread.xstate->xsave.xsave_hdr.xstate_bv |= XSTATE_FP;
+		target->thread.fpu.state->xsave.xsave_hdr.xstate_bv |= XSTATE_FP;
 	return ret;
 }
 
@@ -498,7 +498,7 @@ int fpregs_set(struct task_struct *target, const struct user_regset *regset,
 static inline int save_i387_fsave(struct _fpstate_ia32 __user *buf)
 {
 	struct task_struct *tsk = current;
-	struct i387_fsave_struct *fp = &tsk->thread.xstate->fsave;
+	struct i387_fsave_struct *fp = &tsk->thread.fpu.state->fsave;
 
 	fp->status = fp->swd;
 	if (__copy_to_user(buf, fp, sizeof(struct i387_fsave_struct)))
@@ -509,7 +509,7 @@ static inline int save_i387_fsave(struct _fpstate_ia32 __user *buf)
 static int save_i387_fxsave(struct _fpstate_ia32 __user *buf)
 {
 	struct task_struct *tsk = current;
-	struct i387_fxsave_struct *fx = &tsk->thread.xstate->fxsave;
+	struct i387_fxsave_struct *fx = &tsk->thread.fpu.state->fxsave;
 	struct user_i387_ia32_struct env;
 	int err = 0;
 
@@ -544,7 +544,7 @@ static int save_i387_xsave(void __user *buf)
 	 * header as well as change any contents in the memory layout.
 	 * xrestore as part of sigreturn will capture all the changes.
 	 */
-	tsk->thread.xstate->xsave.xsave_hdr.xstate_bv |= XSTATE_FPSSE;
+	tsk->thread.fpu.state->xsave.xsave_hdr.xstate_bv |= XSTATE_FPSSE;
 
 	if (save_i387_fxsave(fx) < 0)
 		return -1;
@@ -596,7 +596,7 @@ static inline int restore_i387_fsave(struct _fpstate_ia32 __user *buf)
 {
 	struct task_struct *tsk = current;
 
-	return __copy_from_user(&tsk->thread.xstate->fsave, buf,
+	return __copy_from_user(&tsk->thread.fpu.state->fsave, buf,
 				sizeof(struct i387_fsave_struct));
 }
 
@@ -607,10 +607,10 @@ static int restore_i387_fxsave(struct _fpstate_ia32 __user *buf,
 	struct user_i387_ia32_struct env;
 	int err;
 
-	err = __copy_from_user(&tsk->thread.xstate->fxsave, &buf->_fxsr_env[0],
+	err = __copy_from_user(&tsk->thread.fpu.state->fxsave, &buf->_fxsr_env[0],
 			       size);
 	/* mxcsr reserved bits must be masked to zero for security reasons */
-	tsk->thread.xstate->fxsave.mxcsr &= mxcsr_feature_mask;
+	tsk->thread.fpu.state->fxsave.mxcsr &= mxcsr_feature_mask;
 	if (err || __copy_from_user(&env, buf, sizeof(env)))
 		return 1;
 	convert_to_fxsr(tsk, &env);
@@ -626,7 +626,7 @@ static int restore_i387_xsave(void __user *buf)
 	struct i387_fxsave_struct __user *fx =
 		(struct i387_fxsave_struct __user *) &fx_user->_fxsr_env[0];
 	struct xsave_hdr_struct *xsave_hdr =
-				&current->thread.xstate->xsave.xsave_hdr;
+				&current->thread.fpu.state->xsave.xsave_hdr;
 	u64 mask;
 	int err;
 
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index 28ad9f4d8b9..f18fd9c1524 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -32,25 +32,22 @@ struct kmem_cache *task_xstate_cachep;
 
 int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src)
 {
+	int ret;
+
 	*dst = *src;
-	if (src->thread.xstate) {
-		dst->thread.xstate = kmem_cache_alloc(task_xstate_cachep,
-						      GFP_KERNEL);
-		if (!dst->thread.xstate)
-			return -ENOMEM;
-		WARN_ON((unsigned long)dst->thread.xstate & 15);
-		memcpy(dst->thread.xstate, src->thread.xstate, xstate_size);
+	if (fpu_allocated(&src->thread.fpu)) {
+		memset(&dst->thread.fpu, 0, sizeof(dst->thread.fpu));
+		ret = fpu_alloc(&dst->thread.fpu);
+		if (ret)
+			return ret;
+		fpu_copy(&dst->thread.fpu, &src->thread.fpu);
 	}
 	return 0;
 }
 
 void free_thread_xstate(struct task_struct *tsk)
 {
-	if (tsk->thread.xstate) {
-		kmem_cache_free(task_xstate_cachep, tsk->thread.xstate);
-		tsk->thread.xstate = NULL;
-	}
-
+	fpu_free(&tsk->thread.fpu);
 	WARN(tsk->thread.ds_ctx, "leaking DS context\n");
 }
 
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index f6c62667e30..0a7a4f5bbaa 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -317,7 +317,7 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 
 	/* we're going to use this soon, after a few expensive things */
 	if (preload_fpu)
-		prefetch(next->xstate);
+		prefetch(next->fpu.state);
 
 	/*
 	 * Reload esp0.
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 17cb3295cbf..979215f5198 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -396,7 +396,7 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 
 	/* we're going to use this soon, after a few expensive things */
 	if (preload_fpu)
-		prefetch(next->xstate);
+		prefetch(next->fpu.state);
 
 	/*
 	 * Reload esp0, LDT and the page table pointer:
diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c
index c1b0a11033a..37e68fc5e24 100644
--- a/arch/x86/kernel/xsave.c
+++ b/arch/x86/kernel/xsave.c
@@ -109,7 +109,7 @@ int save_i387_xstate(void __user *buf)
 		task_thread_info(tsk)->status &= ~TS_USEDFPU;
 		stts();
 	} else {
-		if (__copy_to_user(buf, &tsk->thread.xstate->fxsave,
+		if (__copy_to_user(buf, &tsk->thread.fpu.state->fxsave,
 				   xstate_size))
 			return -1;
 	}