system.h 13.4 KB
Newer Older
1
2
#ifndef _ASM_X86_SYSTEM_H
#define _ASM_X86_SYSTEM_H
3
4

#include <asm/asm.h>
5
6
7
#include <asm/segment.h>
#include <asm/cpufeature.h>
#include <asm/cmpxchg.h>
Andi Kleen's avatar
Andi Kleen committed
8
#include <asm/nops.h>
9

10
#include <linux/kernel.h>
11
#include <linux/irqflags.h>
12

13
14
15
16
17
18
19
/* entries in ARCH_DLINFO: */
#ifdef CONFIG_IA32_EMULATION
# define AT_VECTOR_SIZE_ARCH 2
#else
# define AT_VECTOR_SIZE_ARCH 1
#endif

20
struct task_struct; /* one of the stranger aspects of C forward declarations */
21
22
struct task_struct *__switch_to(struct task_struct *prev,
				struct task_struct *next);
23
struct tss_struct;
24
25
void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p,
		      struct tss_struct *tss);
26

27
28
#ifdef CONFIG_X86_32

29
30
#ifdef CONFIG_CC_STACKPROTECTOR
#define __switch_canary							\
31
32
	"movl %P[task_canary](%[next]), %%ebx\n\t"			\
	"movl %%ebx, "__percpu_arg([stack_canary])"\n\t"
33
#define __switch_canary_oparam						\
34
	, [stack_canary] "=m" (per_cpu_var(stack_canary.canary))
35
36
37
38
39
40
41
42
#define __switch_canary_iparam						\
	, [task_canary] "i" (offsetof(struct task_struct, stack_canary))
#else	/* CC_STACKPROTECTOR */
#define __switch_canary
#define __switch_canary_oparam
#define __switch_canary_iparam
#endif	/* CC_STACKPROTECTOR */

43
44
45
46
/*
 * Saving eflags is important. It switches not only IOPL between tasks,
 * it also protects other tasks from NT leaking through sysenter etc.
 */
Ingo Molnar's avatar
Ingo Molnar committed
47
48
#define switch_to(prev, next, last)					\
do {									\
Ingo Molnar's avatar
Ingo Molnar committed
49
50
51
52
53
54
55
56
	/*								\
	 * Context-switching clobbers all registers, so we clobber	\
	 * them explicitly, via unused output variables.		\
	 * (EAX and EBP is not listed because EBP is saved/restored	\
	 * explicitly for wchan access and EAX is the return value of	\
	 * __switch_to())						\
	 */								\
	unsigned long ebx, ecx, edx, esi, edi;				\
Ingo Molnar's avatar
Ingo Molnar committed
57
									\
58
59
60
61
62
63
	asm volatile("pushfl\n\t"		/* save    flags */	\
		     "pushl %%ebp\n\t"		/* save    EBP   */	\
		     "movl %%esp,%[prev_sp]\n\t"	/* save    ESP   */ \
		     "movl %[next_sp],%%esp\n\t"	/* restore ESP   */ \
		     "movl $1f,%[prev_ip]\n\t"	/* save    EIP   */	\
		     "pushl %[next_ip]\n\t"	/* restore EIP   */	\
64
		     __switch_canary					\
65
66
67
68
		     "jmp __switch_to\n"	/* regparm call  */	\
		     "1:\t"						\
		     "popl %%ebp\n\t"		/* restore EBP   */	\
		     "popfl\n"			/* restore flags */	\
Ingo Molnar's avatar
Ingo Molnar committed
69
									\
70
71
72
73
		     /* output parameters */				\
		     : [prev_sp] "=m" (prev->thread.sp),		\
		       [prev_ip] "=m" (prev->thread.ip),		\
		       "=a" (last),					\
Ingo Molnar's avatar
Ingo Molnar committed
74
									\
75
76
77
78
		       /* clobbered output registers: */		\
		       "=b" (ebx), "=c" (ecx), "=d" (edx),		\
		       "=S" (esi), "=D" (edi)				\
		       							\
79
80
		       __switch_canary_oparam				\
									\
81
82
83
84
85
86
		       /* input parameters: */				\
		     : [next_sp]  "m" (next->thread.sp),		\
		       [next_ip]  "m" (next->thread.ip),		\
		       							\
		       /* regparm parameters for __switch_to(): */	\
		       [prev]     "a" (prev),				\
87
88
		       [next]     "d" (next)				\
									\
89
90
		       __switch_canary_iparam				\
									\
91
92
		     : /* reloaded segment registers */			\
			"memory");					\
93
94
} while (0)

95
96
97
98
/*
 * disable hlt during certain critical i/o operations
 */
#define HAVE_DISABLE_HLT
99
#else
100
101
102
103
104
105
106
107
108
109
110
#define __SAVE(reg, offset) "movq %%" #reg ",(14-" #offset ")*8(%%rsp)\n\t"
#define __RESTORE(reg, offset) "movq (14-" #offset ")*8(%%rsp),%%" #reg "\n\t"

/* frame pointer must be last for get_wchan */
#define SAVE_CONTEXT    "pushf ; pushq %%rbp ; movq %%rsi,%%rbp\n\t"
#define RESTORE_CONTEXT "movq %%rbp,%%rsi ; popq %%rbp ; popf\t"

#define __EXTRA_CLOBBER  \
	, "rcx", "rbx", "rdx", "r8", "r9", "r10", "r11", \
	  "r12", "r13", "r14", "r15"

111
112
113
#ifdef CONFIG_CC_STACKPROTECTOR
#define __switch_canary							  \
	"movq %P[task_canary](%%rsi),%%r8\n\t"				  \
114
115
116
117
118
	"movq %%r8,"__percpu_arg([gs_canary])"\n\t"
#define __switch_canary_oparam						  \
	, [gs_canary] "=m" (per_cpu_var(irq_stack_union.stack_canary))
#define __switch_canary_iparam						  \
	, [task_canary] "i" (offsetof(struct task_struct, stack_canary))
119
120
#else	/* CC_STACKPROTECTOR */
#define __switch_canary
121
122
#define __switch_canary_oparam
#define __switch_canary_iparam
123
124
#endif	/* CC_STACKPROTECTOR */

125
126
/* Save restore flags to clear handle leaking NT */
#define switch_to(prev, next, last) \
127
	asm volatile(SAVE_CONTEXT					  \
128
129
130
	     "movq %%rsp,%P[threadrsp](%[prev])\n\t" /* save RSP */	  \
	     "movq %P[threadrsp](%[next]),%%rsp\n\t" /* restore RSP */	  \
	     "call __switch_to\n\t"					  \
131
	     "movq "__percpu_arg([current_task])",%%rsi\n\t"		  \
132
	     __switch_canary						  \
133
134
	     "movq %P[thread_info](%%rsi),%%r8\n\t"			  \
	     "movq %%rax,%%rdi\n\t" 					  \
135
136
	     "testl  %[_tif_fork],%P[ti_flags](%%r8)\n\t"	  \
	     "jnz   ret_from_fork\n\t"					  \
137
138
	     RESTORE_CONTEXT						  \
	     : "=a" (last)					  	  \
139
	       __switch_canary_oparam					  \
140
141
142
	     : [next] "S" (next), [prev] "D" (prev),			  \
	       [threadrsp] "i" (offsetof(struct task_struct, thread.sp)), \
	       [ti_flags] "i" (offsetof(struct thread_info, flags)),	  \
143
	       [_tif_fork] "i" (_TIF_FORK),			  	  \
144
	       [thread_info] "i" (offsetof(struct task_struct, stack)),   \
145
	       [current_task] "m" (per_cpu_var(current_task))		  \
146
	       __switch_canary_iparam					  \
147
	     : "memory", "cc" __EXTRA_CLOBBER)
148
#endif
149
150
151

#ifdef __KERNEL__

152
extern void native_load_gs_index(unsigned);
153

154
155
156
157
158
159
/*
 * Load a segment. Fall back on loading the zero
 * segment if something goes wrong..
 */
#define loadsegment(seg, value)			\
	asm volatile("\n"			\
160
161
162
163
164
165
166
167
168
		     "1:\t"			\
		     "movl %k0,%%" #seg "\n"	\
		     "2:\n"			\
		     ".section .fixup,\"ax\"\n"	\
		     "3:\t"			\
		     "movl %k1, %%" #seg "\n\t"	\
		     "jmp 2b\n"			\
		     ".previous\n"		\
		     _ASM_EXTABLE(1b,3b)	\
169
		     : :"r" (value), "r" (0) : "memory")
170
171


172
173
174
/*
 * Save a segment register away
 */
175
#define savesegment(seg, value)				\
176
	asm("mov %%" #seg ",%0":"=r" (value) : : "memory")
177

Tejun Heo's avatar
Tejun Heo committed
178
179
180
181
/*
 * x86_32 user gs accessors.
 */
#ifdef CONFIG_X86_32
182
#ifdef CONFIG_X86_32_LAZY_GS
Tejun Heo's avatar
Tejun Heo committed
183
184
185
#define get_user_gs(regs)	(u16)({unsigned long v; savesegment(gs, v); v;})
#define set_user_gs(regs, v)	loadsegment(gs, (unsigned long)(v))
#define task_user_gs(tsk)	((tsk)->thread.gs)
186
187
188
189
190
191
192
193
194
195
#define lazy_save_gs(v)		savesegment(gs, (v))
#define lazy_load_gs(v)		loadsegment(gs, (v))
#else	/* X86_32_LAZY_GS */
#define get_user_gs(regs)	(u16)((regs)->gs)
#define set_user_gs(regs, v)	do { (regs)->gs = (v); } while (0)
#define task_user_gs(tsk)	(task_pt_regs(tsk)->gs)
#define lazy_save_gs(v)		do { } while (0)
#define lazy_load_gs(v)		do { } while (0)
#endif	/* X86_32_LAZY_GS */
#endif	/* X86_32 */
Tejun Heo's avatar
Tejun Heo committed
196

197
198
199
static inline unsigned long get_limit(unsigned long segment)
{
	unsigned long __limit;
200
201
	asm("lsll %1,%0" : "=r" (__limit) : "r" (segment));
	return __limit + 1;
202
}
203
204
205

static inline void native_clts(void)
{
206
	asm volatile("clts");
207
208
209
210
211
212
213
214
215
216
217
218
219
220
}

/*
 * Volatile isn't enough to prevent the compiler from reordering the
 * read/write functions for the control registers and messing everything up.
 * A memory clobber would solve the problem, but would prevent reordering of
 * all loads stores around it, which can hurt performance. Solution is to
 * use a variable and mimic reads and writes to it to enforce serialization
 */
static unsigned long __force_order;

static inline unsigned long native_read_cr0(void)
{
	unsigned long val;
221
	asm volatile("mov %%cr0,%0\n\t" : "=r" (val), "=m" (__force_order));
222
223
224
225
226
	return val;
}

static inline void native_write_cr0(unsigned long val)
{
227
	asm volatile("mov %0,%%cr0": : "r" (val), "m" (__force_order));
228
229
230
231
232
}

static inline unsigned long native_read_cr2(void)
{
	unsigned long val;
233
	asm volatile("mov %%cr2,%0\n\t" : "=r" (val), "=m" (__force_order));
234
235
236
237
238
	return val;
}

static inline void native_write_cr2(unsigned long val)
{
239
	asm volatile("mov %0,%%cr2": : "r" (val), "m" (__force_order));
240
241
242
243
244
}

static inline unsigned long native_read_cr3(void)
{
	unsigned long val;
245
	asm volatile("mov %%cr3,%0\n\t" : "=r" (val), "=m" (__force_order));
246
247
248
249
250
	return val;
}

static inline void native_write_cr3(unsigned long val)
{
251
	asm volatile("mov %0,%%cr3": : "r" (val), "m" (__force_order));
252
253
254
255
256
}

static inline unsigned long native_read_cr4(void)
{
	unsigned long val;
257
	asm volatile("mov %%cr4,%0\n\t" : "=r" (val), "=m" (__force_order));
258
259
260
261
262
263
264
265
266
	return val;
}

static inline unsigned long native_read_cr4_safe(void)
{
	unsigned long val;
	/* This could fault if %cr4 does not exist. In x86_64, a cr4 always
	 * exists, so it will never fail. */
#ifdef CONFIG_X86_32
267
268
	asm volatile("1: mov %%cr4, %0\n"
		     "2:\n"
269
		     _ASM_EXTABLE(1b, 2b)
270
		     : "=r" (val), "=m" (__force_order) : "0" (0));
271
272
273
274
275
276
277
278
#else
	val = native_read_cr4();
#endif
	return val;
}

static inline void native_write_cr4(unsigned long val)
{
279
	asm volatile("mov %0,%%cr4": : "r" (val), "m" (__force_order));
280
281
}

282
283
284
285
286
287
288
289
290
291
292
293
294
295
#ifdef CONFIG_X86_64
static inline unsigned long native_read_cr8(void)
{
	unsigned long cr8;
	asm volatile("movq %%cr8,%0" : "=r" (cr8));
	return cr8;
}

static inline void native_write_cr8(unsigned long val)
{
	asm volatile("movq %0,%%cr8" :: "r" (val) : "memory");
}
#endif

296
297
298
299
static inline void native_wbinvd(void)
{
	asm volatile("wbinvd": : :"memory");
}
300

301
302
303
304
305
306
307
308
309
310
311
312
313
#ifdef CONFIG_PARAVIRT
#include <asm/paravirt.h>
#else
#define read_cr0()	(native_read_cr0())
#define write_cr0(x)	(native_write_cr0(x))
#define read_cr2()	(native_read_cr2())
#define write_cr2(x)	(native_write_cr2(x))
#define read_cr3()	(native_read_cr3())
#define write_cr3(x)	(native_write_cr3(x))
#define read_cr4()	(native_read_cr4())
#define read_cr4_safe()	(native_read_cr4_safe())
#define write_cr4(x)	(native_write_cr4(x))
#define wbinvd()	(native_wbinvd())
314
#ifdef CONFIG_X86_64
315
316
#define read_cr8()	(native_read_cr8())
#define write_cr8(x)	(native_write_cr8(x))
317
#define load_gs_index   native_load_gs_index
318
319
#endif

320
321
322
323
324
/* Clear the 'TS' bit */
#define clts()		(native_clts())

#endif/* CONFIG_PARAVIRT */

325
#define stts() write_cr0(read_cr0() | X86_CR0_TS)
326

327
328
#endif /* __KERNEL__ */

329
static inline void clflush(volatile void *__p)
330
{
331
	asm volatile("clflush %0" : "+m" (*(volatile char __force *)__p));
332
333
}

334
#define nop() asm volatile ("nop")
335
336
337
338
339
340
341
342
343
344
345

void disable_hlt(void);
void enable_hlt(void);

void cpu_idle_wait(void);

extern unsigned long arch_align_stack(unsigned long sp);
extern void free_init_pages(char *what, unsigned long begin, unsigned long end);

void default_idle(void);

346
347
void stop_this_cpu(void *dummy);

348
349
350
351
352
353
354
/*
 * Force strict CPU ordering.
 * And yes, this is required on UP too when we're talking
 * to devices.
 */
#ifdef CONFIG_X86_32
/*
355
 * Some non-Intel clones support out of order store. wmb() ceases to be a
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
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
 * nop for these.
 */
#define mb() alternative("lock; addl $0,0(%%esp)", "mfence", X86_FEATURE_XMM2)
#define rmb() alternative("lock; addl $0,0(%%esp)", "lfence", X86_FEATURE_XMM2)
#define wmb() alternative("lock; addl $0,0(%%esp)", "sfence", X86_FEATURE_XMM)
#else
#define mb() 	asm volatile("mfence":::"memory")
#define rmb()	asm volatile("lfence":::"memory")
#define wmb()	asm volatile("sfence" ::: "memory")
#endif

/**
 * read_barrier_depends - Flush all pending reads that subsequents reads
 * depend on.
 *
 * No data-dependent reads from memory-like regions are ever reordered
 * over this barrier.  All reads preceding this primitive are guaranteed
 * to access memory (but not necessarily other CPUs' caches) before any
 * reads following this primitive that depend on the data return by
 * any of the preceding reads.  This primitive is much lighter weight than
 * rmb() on most CPUs, and is never heavier weight than is
 * rmb().
 *
 * These ordering constraints are respected by both the local CPU
 * and the compiler.
 *
 * Ordering is not guaranteed by anything other than these primitives,
 * not even by data dependencies.  See the documentation for
 * memory_barrier() for examples and URLs to more information.
 *
 * For example, the following code would force ordering (the initial
 * value of "a" is zero, "b" is one, and "p" is "&a"):
 *
 * <programlisting>
 *	CPU 0				CPU 1
 *
 *	b = 2;
 *	memory_barrier();
 *	p = &b;				q = p;
 *					read_barrier_depends();
 *					d = *q;
 * </programlisting>
 *
 * because the read of "*q" depends on the read of "p" and these
 * two reads are separated by a read_barrier_depends().  However,
 * the following code, with the same initial values for "a" and "b":
 *
 * <programlisting>
 *	CPU 0				CPU 1
 *
 *	a = 2;
 *	memory_barrier();
 *	b = 3;				y = b;
 *					read_barrier_depends();
 *					x = a;
 * </programlisting>
 *
 * does not enforce ordering, since there is no data dependency between
 * the read of "a" and the read of "b".  Therefore, on some CPUs, such
 * as Alpha, "y" could be set to 3 and "x" to 0.  Use rmb()
 * in cases like this where there are no data dependencies.
 **/

#define read_barrier_depends()	do { } while (0)

#ifdef CONFIG_SMP
#define smp_mb()	mb()
#ifdef CONFIG_X86_PPRO_FENCE
# define smp_rmb()	rmb()
#else
# define smp_rmb()	barrier()
#endif
#ifdef CONFIG_X86_OOSTORE
# define smp_wmb() 	wmb()
#else
# define smp_wmb()	barrier()
#endif
#define smp_read_barrier_depends()	read_barrier_depends()
434
#define set_mb(var, value) do { (void)xchg(&var, value); } while (0)
435
436
437
438
439
440
441
442
#else
#define smp_mb()	barrier()
#define smp_rmb()	barrier()
#define smp_wmb()	barrier()
#define smp_read_barrier_depends()	do { } while (0)
#define set_mb(var, value) do { var = value; barrier(); } while (0)
#endif

Andi Kleen's avatar
Andi Kleen committed
443
444
445
446
447
448
449
450
451
452
453
454
/*
 * Stop RDTSC speculation. This is needed when you need to use RDTSC
 * (or get_cycles or vread that possibly accesses the TSC) in a defined
 * code region.
 *
 * (Could use an alternative three way for this if there was one.)
 */
static inline void rdtsc_barrier(void)
{
	alternative(ASM_NOP3, "mfence", X86_FEATURE_MFENCE_RDTSC);
	alternative(ASM_NOP3, "lfence", X86_FEATURE_LFENCE_RDTSC);
}
455

456
#endif /* _ASM_X86_SYSTEM_H */