summaryrefslogtreecommitdiffstats
path: root/arch/x86/kernel/step.c
blob: 5884dd485db85d86b345a7f8bc665f65f47b77fb (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
/*
 * x86 single-step support code, common to 32-bit and 64-bit.
 */
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/ptrace.h>

#ifdef CONFIG_X86_32
#include <linux/uaccess.h>

#include <asm/desc.h>

/*
 * Return EIP plus the CS segment base.  The segment limit is also
 * adjusted, clamped to the kernel/user address space (whichever is
 * appropriate), and returned in *eip_limit.
 *
 * The segment is checked, because it might have been changed by another
 * task between the original faulting instruction and here.
 *
 * If CS is no longer a valid code segment, or if EIP is beyond the
 * limit, or if it is a kernel address when CS is not a kernel segment,
 * then the returned value will be greater than *eip_limit.
 *
 * This is slow, but is very rarely executed.
 */
unsigned long get_segment_eip(struct pt_regs *regs,
					    unsigned long *eip_limit)
{
	unsigned long ip = regs->ip;
	unsigned seg = regs->cs & 0xffff;
	u32 seg_ar, seg_limit, base, *desc;

	/* Unlikely, but must come before segment checks. */
	if (unlikely(regs->flags & VM_MASK)) {
		base = seg << 4;
		*eip_limit = base + 0xffff;
		return base + (ip & 0xffff);
	}

	/* The standard kernel/user address space limit. */
	*eip_limit = user_mode(regs) ? USER_DS.seg : KERNEL_DS.seg;

	/* By far the most common cases. */
	if (likely(SEGMENT_IS_FLAT_CODE(seg)))
		return ip;

	/* Check the segment exists, is within the current LDT/GDT size,
	   that kernel/user (ring 0..3) has the appropriate privilege,
	   that it's a code segment, and get the limit. */
	__asm__("larl %3,%0; lsll %3,%1"
		 : "=&r" (seg_ar), "=r" (seg_limit) : "0" (0), "rm" (seg));
	if ((~seg_ar & 0x9800) || ip > seg_limit) {
		*eip_limit = 0;
		return 1;	 /* So that returned ip > *eip_limit. */
	}

	/* Get the GDT/LDT descriptor base.
	   When you look for races in this code remember that
	   LDT and other horrors are only used in user space. */
	if (seg & (1<<2)) {
		/* Must lock the LDT while reading it. */
		mutex_lock(&current->mm->context.lock);
		desc = current->mm->context.ldt;
		desc = (void *)desc + (seg & ~7);
	} else {
		/* Must disable preemption while reading the GDT. */
		desc = (u32 *)get_cpu_gdt_table(get_cpu());
		desc = (void *)desc + (seg & ~7);
	}

	/* Decode the code segment base from the descriptor */
	base = get_desc_base((struct desc_struct *)desc);

	if (seg & (1<<2))
		mutex_unlock(&current->mm->context.lock);
	else
		put_cpu();

	/* Adjust EIP and segment limit, and clamp at the kernel limit.
	   It's legitimate for segments to wrap at 0xffffffff. */
	seg_limit += base;
	if (seg_limit < *eip_limit && seg_limit >= base)
		*eip_limit = seg_limit;
	return ip + base;
}
#endif

#ifdef CONFIG_X86_32
static
#endif
unsigned long convert_rip_to_linear(struct task_struct *child, struct pt_regs *regs)
{
	unsigned long addr, seg;

	addr = regs->ip;
	seg = regs->cs & 0xffff;
	if (v8086_mode(regs)) {
		addr = (addr & 0xffff) + (seg << 4);
		return addr;
	}

	/*
	 * We'll assume that the code segments in the GDT
	 * are all zero-based. That is largely true: the
	 * TLS segments are used for data, and the PNPBIOS
	 * and APM bios ones we just ignore here.
	 */
	if ((seg & SEGMENT_TI_MASK) == SEGMENT_LDT) {
		u32 *desc;
		unsigned long base;

		seg &= ~7UL;

		mutex_lock(&child->mm->context.lock);
		if (unlikely((seg >> 3) >= child->mm->context.size))
			addr = -1L; /* bogus selector, access would fault */
		else {
			desc = child->mm->context.ldt + seg;
			base = ((desc[0] >> 16) |
				((desc[1] & 0xff) << 16) |
				(desc[1] & 0xff000000));

			/* 16-bit code segment? */
			if (!((desc[1] >> 22) & 1))
				addr &= 0xffff;
			addr += base;
		}
		mutex_unlock(&child->mm->context.lock);
	}

	return addr;
}

static int is_setting_trap_flag(struct task_struct *child, struct pt_regs *regs)
{
	int i, copied;
	unsigned char opcode[15];
	unsigned long addr = convert_rip_to_linear(child, regs);

	copied = access_process_vm(child, addr, opcode, sizeof(opcode), 0);
	for (i = 0; i < copied; i++) {
		switch (opcode[i]) {
		/* popf and iret */
		case 0x9d: case 0xcf:
			return 1;

			/* CHECKME: 64 65 */

		/* opcode and address size prefixes */
		case 0x66: case 0x67:
			continue;
		/* irrelevant prefixes (segment overrides and repeats) */
		case 0x26: case 0x2e:
		case 0x36: case 0x3e:
		case 0x64: case 0x65:
		case 0xf0: case 0xf2: case 0xf3:
			continue;

#ifdef CONFIG_X86_64
		case 0x40 ... 0x4f:
			if (regs->cs != __USER_CS)
				/* 32-bit mode: register increment */
				return 0;
			/* 64-bit mode: REX prefix */
			continue;
#endif

			/* CHECKME: f2, f3 */

		/*
		 * pushf: NOTE! We should probably not let
		 * the user see the TF bit being set. But
		 * it's more pain than it's worth to avoid
		 * it, and a debugger could emulate this
		 * all in user space if it _really_ cares.
		 */
		case 0x9c:
		default:
			return 0;
		}
	}
	return 0;
}

/*
 * Enable single-stepping.  Return nonzero if user mode is not using TF itself.
 */
static int enable_single_step(struct task_struct *child)
{
	struct pt_regs *regs = task_pt_regs(child);

	/*
	 * Always set TIF_SINGLESTEP - this guarantees that
	 * we single-step system calls etc..  This will also
	 * cause us to set TF when returning to user mode.
	 */
	set_tsk_thread_flag(child, TIF_SINGLESTEP);

	/*
	 * If TF was already set, don't do anything else
	 */
	if (regs->flags & X86_EFLAGS_TF)
		return 0;

	/* Set TF on the kernel stack.. */
	regs->flags |= X86_EFLAGS_TF;

	/*
	 * ..but if TF is changed by the instruction we will trace,
	 * don't mark it as being "us" that set it, so that we
	 * won't clear it by hand later.
	 */
	if (is_setting_trap_flag(child, regs))
		return 0;

	set_tsk_thread_flag(child, TIF_FORCED_TF);

	return 1;
}

/*
 * Install this value in MSR_IA32_DEBUGCTLMSR whenever child is running.
 */
static void write_debugctlmsr(struct task_struct *child, unsigned long val)
{
	child->thread.debugctlmsr = val;

	if (child != current)
		return;

#ifdef CONFIG_X86_64
	wrmsrl(MSR_IA32_DEBUGCTLMSR, val);
#else
	wrmsr(MSR_IA32_DEBUGCTLMSR, val, 0);
#endif
}

/*
 * Enable single or block step.
 */
static void enable_step(struct task_struct *child, bool block)
{
	/*
	 * Make sure block stepping (BTF) is not enabled unless it should be.
	 * Note that we don't try to worry about any is_setting_trap_flag()
	 * instructions after the first when using block stepping.
	 * So noone should try to use debugger block stepping in a program
	 * that uses user-mode single stepping itself.
	 */
	if (enable_single_step(child) && block) {
		set_tsk_thread_flag(child, TIF_DEBUGCTLMSR);
		write_debugctlmsr(child,
				  child->thread.debugctlmsr | DEBUGCTLMSR_BTF);
	} else {
	    write_debugctlmsr(child,
			      child->thread.debugctlmsr & ~TIF_DEBUGCTLMSR);

	    if (!child->thread.debugctlmsr)
		    clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR);
	}
}

void user_enable_single_step(struct task_struct *child)
{
	enable_step(child, 0);
}

void user_enable_block_step(struct task_struct *child)
{
	enable_step(child, 1);
}

void user_disable_single_step(struct task_struct *child)
{
	/*
	 * Make sure block stepping (BTF) is disabled.
	 */
	write_debugctlmsr(child,
			  child->thread.debugctlmsr & ~TIF_DEBUGCTLMSR);

	if (!child->thread.debugctlmsr)
		clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR);

	/* Always clear TIF_SINGLESTEP... */
	clear_tsk_thread_flag(child, TIF_SINGLESTEP);

	/* But touch TF only if it was set by us.. */
	if (test_and_clear_tsk_thread_flag(child, TIF_FORCED_TF))
		task_pt_regs(child)->flags &= ~X86_EFLAGS_TF;
}
OpenPOWER on IntegriCloud