i387.h 10.1 KB
Newer Older
R
Roland McGrath 已提交
1 2 3 4 5 6 7 8 9
/*
 * Copyright (C) 1994 Linus Torvalds
 *
 * Pentium III FXSR, SSE support
 * General FPU state handling cleanups
 *	Gareth Hughes <gareth@valinux.com>, May 2000
 * x86-64 work by Andi Kleen 2002
 */

H
H. Peter Anvin 已提交
10 11
#ifndef _ASM_X86_I387_H
#define _ASM_X86_I387_H
R
Roland McGrath 已提交
12 13 14 15

#include <linux/sched.h>
#include <linux/kernel_stat.h>
#include <linux/regset.h>
16
#include <linux/hardirq.h>
17
#include <asm/asm.h>
R
Roland McGrath 已提交
18 19 20 21
#include <asm/processor.h>
#include <asm/sigcontext.h>
#include <asm/user.h>
#include <asm/uaccess.h>
22
#include <asm/xsave.h>
R
Roland McGrath 已提交
23

24
extern unsigned int sig_xstate_size;
R
Roland McGrath 已提交
25 26
extern void fpu_init(void);
extern void mxcsr_feature_mask_init(void);
27
extern int init_fpu(struct task_struct *child);
R
Roland McGrath 已提交
28
extern asmlinkage void math_state_restore(void);
29
extern void init_thread_xstate(void);
30
extern int dump_fpu(struct pt_regs *, struct user_i387_struct *);
R
Roland McGrath 已提交
31 32 33 34 35

extern user_regset_active_fn fpregs_active, xfpregs_active;
extern user_regset_get_fn fpregs_get, xfpregs_get, fpregs_soft_get;
extern user_regset_set_fn fpregs_set, xfpregs_set, fpregs_soft_set;

36
extern struct _fpx_sw_bytes fx_sw_reserved;
R
Roland McGrath 已提交
37
#ifdef CONFIG_IA32_EMULATION
38
extern unsigned int sig_xstate_ia32_size;
39
extern struct _fpx_sw_bytes fx_sw_reserved_ia32;
R
Roland McGrath 已提交
40
struct _fpstate_ia32;
41 42 43
struct _xstate_ia32;
extern int save_i387_xstate_ia32(void __user *buf);
extern int restore_i387_xstate_ia32(void __user *buf);
R
Roland McGrath 已提交
44 45
#endif

46 47
#define X87_FSW_ES (1 << 7)	/* Exception Summary */

R
Roland McGrath 已提交
48 49 50 51 52 53 54
#ifdef CONFIG_X86_64

/* Ignore delayed exceptions from user space */
static inline void tolerant_fwait(void)
{
	asm volatile("1: fwait\n"
		     "2:\n"
55
		     _ASM_EXTABLE(1b, 2b));
R
Roland McGrath 已提交
56 57
}

58
static inline int fxrstor_checking(struct i387_fxsave_struct *fx)
R
Roland McGrath 已提交
59 60 61 62 63 64 65 66 67
{
	int err;

	asm volatile("1:  rex64/fxrstor (%[fx])\n\t"
		     "2:\n"
		     ".section .fixup,\"ax\"\n"
		     "3:  movl $-1,%[err]\n"
		     "    jmp  2b\n"
		     ".previous\n"
68
		     _ASM_EXTABLE(1b, 3b)
R
Roland McGrath 已提交
69
		     : [err] "=r" (err)
J
Jiri Slaby 已提交
70
#if 0 /* See comment in fxsave() below. */
R
Roland McGrath 已提交
71 72 73 74 75 76 77 78 79 80 81 82
		     : [fx] "r" (fx), "m" (*fx), "0" (0));
#else
		     : [fx] "cdaSDb" (fx), "m" (*fx), "0" (0));
#endif
	return err;
}

/* AMD CPUs don't save/restore FDP/FIP/FOP unless an exception
   is pending. Clear the x87 state here by setting it to fixed
   values. The kernel data segment can be sometimes 0 and sometimes
   new user value. Both should be ok.
   Use the PDA as safe address because it should be already in L1. */
83
static inline void clear_fpu_state(struct task_struct *tsk)
R
Roland McGrath 已提交
84
{
85 86 87 88 89 90 91 92 93 94
	struct xsave_struct *xstate = &tsk->thread.xstate->xsave;
	struct i387_fxsave_struct *fx = &tsk->thread.xstate->fxsave;

	/*
	 * xsave header may indicate the init state of the FP.
	 */
	if ((task_thread_info(tsk)->status & TS_XSAVE) &&
	    !(xstate->xsave_hdr.xstate_bv & XSTATE_FP))
		return;

R
Roland McGrath 已提交
95
	if (unlikely(fx->swd & X87_FSW_ES))
96
		asm volatile("fnclex");
R
Roland McGrath 已提交
97
	alternative_input(ASM_NOP8 ASM_NOP2,
98 99 100
			  "    emms\n"		/* clear stack tags */
			  "    fildl %%gs:0",	/* load to clear state */
			  X86_FEATURE_FXSAVE_LEAK);
R
Roland McGrath 已提交
101 102
}

103
static inline int fxsave_user(struct i387_fxsave_struct __user *fx)
R
Roland McGrath 已提交
104 105 106 107 108 109 110 111 112
{
	int err;

	asm volatile("1:  rex64/fxsave (%[fx])\n\t"
		     "2:\n"
		     ".section .fixup,\"ax\"\n"
		     "3:  movl $-1,%[err]\n"
		     "    jmp  2b\n"
		     ".previous\n"
113
		     _ASM_EXTABLE(1b, 3b)
R
Roland McGrath 已提交
114
		     : [err] "=r" (err), "=m" (*fx)
J
Jiri Slaby 已提交
115
#if 0 /* See comment in fxsave() below. */
R
Roland McGrath 已提交
116 117 118 119
		     : [fx] "r" (fx), "0" (0));
#else
		     : [fx] "cdaSDb" (fx), "0" (0));
#endif
120 121
	if (unlikely(err) &&
	    __clear_user(fx, sizeof(struct i387_fxsave_struct)))
R
Roland McGrath 已提交
122 123 124 125 126
		err = -EFAULT;
	/* No need to clear here because the caller clears USED_MATH */
	return err;
}

127
static inline void fxsave(struct task_struct *tsk)
R
Roland McGrath 已提交
128 129 130 131 132 133 134 135 136
{
	/* Using "rex64; fxsave %0" is broken because, if the memory operand
	   uses any extended registers for addressing, a second REX prefix
	   will be generated (to the assembler, rex64 followed by semicolon
	   is a separate instruction), and hence the 64-bitness is lost. */
#if 0
	/* Using "fxsaveq %0" would be the ideal choice, but is only supported
	   starting with gas 2.16. */
	__asm__ __volatile__("fxsaveq %0"
137
			     : "=m" (tsk->thread.xstate->fxsave));
R
Roland McGrath 已提交
138 139 140 141 142 143
#elif 0
	/* Using, as a workaround, the properly prefixed form below isn't
	   accepted by any binutils version so far released, complaining that
	   the same type of prefix is used twice if an extended register is
	   needed for addressing (fix submitted to mainline 2005-11-21). */
	__asm__ __volatile__("rex64/fxsave %0"
144
			     : "=m" (tsk->thread.xstate->fxsave));
R
Roland McGrath 已提交
145 146 147
#else
	/* This, however, we can work around by forcing the compiler to select
	   an addressing mode that doesn't require extended registers. */
148 149 150
	__asm__ __volatile__("rex64/fxsave (%1)"
			     : "=m" (tsk->thread.xstate->fxsave)
			     : "cdaSDb" (&tsk->thread.xstate->fxsave));
R
Roland McGrath 已提交
151
#endif
152 153 154 155 156 157 158 159 160 161
}

static inline void __save_init_fpu(struct task_struct *tsk)
{
	if (task_thread_info(tsk)->status & TS_XSAVE)
		xsave(tsk);
	else
		fxsave(tsk);

	clear_fpu_state(tsk);
R
Roland McGrath 已提交
162 163 164 165 166
	task_thread_info(tsk)->status &= ~TS_USEDFPU;
}

#else  /* CONFIG_X86_32 */

167 168 169 170 171 172 173
#ifdef CONFIG_MATH_EMULATION
extern void finit_task(struct task_struct *tsk);
#else
static inline void finit_task(struct task_struct *tsk)
{
}
#endif
174

R
Roland McGrath 已提交
175 176 177 178 179
static inline void tolerant_fwait(void)
{
	asm volatile("fnclex ; fwait");
}

J
Jiri Slaby 已提交
180 181
/* perform fxrstor iff the processor has extended states, otherwise frstor */
static inline int fxrstor_checking(struct i387_fxsave_struct *fx)
R
Roland McGrath 已提交
182 183 184 185 186 187 188 189 190
{
	/*
	 * The "nop" is needed to make the instructions the same
	 * length.
	 */
	alternative_input(
		"nop ; frstor %1",
		"fxrstor %1",
		X86_FEATURE_FXSR,
J
Jiri Slaby 已提交
191 192
		"m" (*fx));

193
	return 0;
R
Roland McGrath 已提交
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
}

/* We need a safe address that is cheap to find and that is already
   in L1 during context switch. The best choices are unfortunately
   different for UP and SMP */
#ifdef CONFIG_SMP
#define safe_address (__per_cpu_offset[0])
#else
#define safe_address (kstat_cpu(0).cpustat.user)
#endif

/*
 * These must be called with preempt disabled
 */
static inline void __save_init_fpu(struct task_struct *tsk)
{
210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230
	if (task_thread_info(tsk)->status & TS_XSAVE) {
		struct xsave_struct *xstate = &tsk->thread.xstate->xsave;
		struct i387_fxsave_struct *fx = &tsk->thread.xstate->fxsave;

		xsave(tsk);

		/*
		 * xsave header may indicate the init state of the FP.
		 */
		if (!(xstate->xsave_hdr.xstate_bv & XSTATE_FP))
			goto end;

		if (unlikely(fx->swd & X87_FSW_ES))
			asm volatile("fnclex");

		/*
		 * we can do a simple return here or be paranoid :)
		 */
		goto clear_state;
	}

R
Roland McGrath 已提交
231 232 233 234 235 236 237
	/* Use more nops than strictly needed in case the compiler
	   varies code */
	alternative_input(
		"fnsave %[fx] ;fwait;" GENERIC_NOP8 GENERIC_NOP4,
		"fxsave %[fx]\n"
		"bt $7,%[fsw] ; jnc 1f ; fnclex\n1:",
		X86_FEATURE_FXSR,
238 239
		[fx] "m" (tsk->thread.xstate->fxsave),
		[fsw] "m" (tsk->thread.xstate->fxsave.swd) : "memory");
240
clear_state:
R
Roland McGrath 已提交
241 242 243 244 245 246 247 248 249
	/* AMD K7/K8 CPUs don't save/restore FDP/FIP/FOP unless an exception
	   is pending.  Clear the x87 state here by setting it to fixed
	   values. safe_address is a random variable that should be in L1 */
	alternative_input(
		GENERIC_NOP8 GENERIC_NOP2,
		"emms\n\t"	  	/* clear stack tags */
		"fildl %[addr]", 	/* set F?P to defined value */
		X86_FEATURE_FXSAVE_LEAK,
		[addr] "m" (safe_address));
250
end:
R
Roland McGrath 已提交
251 252 253
	task_thread_info(tsk)->status &= ~TS_USEDFPU;
}

254 255
#endif	/* CONFIG_X86_64 */

J
Jiri Slaby 已提交
256 257 258 259 260 261 262 263
static inline int restore_fpu_checking(struct task_struct *tsk)
{
	if (task_thread_info(tsk)->status & TS_XSAVE)
		return xrstor_checking(&tsk->thread.xstate->xsave);
	else
		return fxrstor_checking(&tsk->thread.xstate->fxsave);
}

R
Roland McGrath 已提交
264 265 266
/*
 * Signal frame handlers...
 */
267 268
extern int save_i387_xstate(void __user *buf);
extern int restore_i387_xstate(void __user *buf);
R
Roland McGrath 已提交
269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303

static inline void __unlazy_fpu(struct task_struct *tsk)
{
	if (task_thread_info(tsk)->status & TS_USEDFPU) {
		__save_init_fpu(tsk);
		stts();
	} else
		tsk->fpu_counter = 0;
}

static inline void __clear_fpu(struct task_struct *tsk)
{
	if (task_thread_info(tsk)->status & TS_USEDFPU) {
		tolerant_fwait();
		task_thread_info(tsk)->status &= ~TS_USEDFPU;
		stts();
	}
}

static inline void kernel_fpu_begin(void)
{
	struct thread_info *me = current_thread_info();
	preempt_disable();
	if (me->status & TS_USEDFPU)
		__save_init_fpu(me->task);
	else
		clts();
}

static inline void kernel_fpu_end(void)
{
	stts();
	preempt_enable();
}

304 305 306
/*
 * Some instructions like VIA's padlock instructions generate a spurious
 * DNA fault but don't modify SSE registers. And these instructions
307 308
 * get used from interrupt context as well. To prevent these kernel instructions
 * in interrupt context interacting wrongly with other user/kernel fpu usage, we
309 310 311 312 313
 * should use them only in the context of irq_ts_save/restore()
 */
static inline int irq_ts_save(void)
{
	/*
314 315 316
	 * If in process context and not atomic, we can take a spurious DNA fault.
	 * Otherwise, doing clts() in process context requires disabling preemption
	 * or some heavy lifting like kernel_fpu_begin()
317
	 */
318
	if (!in_atomic())
319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334
		return 0;

	if (read_cr0() & X86_CR0_TS) {
		clts();
		return 1;
	}

	return 0;
}

static inline void irq_ts_restore(int TS_state)
{
	if (TS_state)
		stts();
}

R
Roland McGrath 已提交
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380
#ifdef CONFIG_X86_64

static inline void save_init_fpu(struct task_struct *tsk)
{
	__save_init_fpu(tsk);
	stts();
}

#define unlazy_fpu	__unlazy_fpu
#define clear_fpu	__clear_fpu

#else  /* CONFIG_X86_32 */

/*
 * These disable preemption on their own and are safe
 */
static inline void save_init_fpu(struct task_struct *tsk)
{
	preempt_disable();
	__save_init_fpu(tsk);
	stts();
	preempt_enable();
}

static inline void unlazy_fpu(struct task_struct *tsk)
{
	preempt_disable();
	__unlazy_fpu(tsk);
	preempt_enable();
}

static inline void clear_fpu(struct task_struct *tsk)
{
	preempt_disable();
	__clear_fpu(tsk);
	preempt_enable();
}

#endif	/* CONFIG_X86_64 */

/*
 * i387 state interaction
 */
static inline unsigned short get_fpu_cwd(struct task_struct *tsk)
{
	if (cpu_has_fxsr) {
381
		return tsk->thread.xstate->fxsave.cwd;
R
Roland McGrath 已提交
382
	} else {
S
Suresh Siddha 已提交
383
		return (unsigned short)tsk->thread.xstate->fsave.cwd;
R
Roland McGrath 已提交
384 385 386 387 388 389
	}
}

static inline unsigned short get_fpu_swd(struct task_struct *tsk)
{
	if (cpu_has_fxsr) {
390
		return tsk->thread.xstate->fxsave.swd;
R
Roland McGrath 已提交
391
	} else {
S
Suresh Siddha 已提交
392
		return (unsigned short)tsk->thread.xstate->fsave.swd;
R
Roland McGrath 已提交
393 394 395 396 397 398
	}
}

static inline unsigned short get_fpu_mxcsr(struct task_struct *tsk)
{
	if (cpu_has_xmm) {
399
		return tsk->thread.xstate->fxsave.mxcsr;
R
Roland McGrath 已提交
400 401 402 403 404
	} else {
		return MXCSR_DEFAULT;
	}
}

H
H. Peter Anvin 已提交
405
#endif /* _ASM_X86_I387_H */