Free Electrons

Embedded Linux Experts

  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
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
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
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
/*
 *  arch/s390/lib/uaccess_pt.c
 *
 *  User access functions based on page table walks for enhanced
 *  system layout without hardware support.
 *
 *    Copyright IBM Corp. 2006
 *    Author(s): Gerald Schaefer (gerald.schaefer@de.ibm.com)
 */

#include <linux/errno.h>
#include <linux/hardirq.h>
#include <linux/mm.h>
#include <asm/uaccess.h>
#include <asm/futex.h>
#include "uaccess.h"

static inline pte_t *follow_table(struct mm_struct *mm, unsigned long addr)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;

	pgd = pgd_offset(mm, addr);
	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
		return (pte_t *) 0x3a;

	pud = pud_offset(pgd, addr);
	if (pud_none(*pud) || unlikely(pud_bad(*pud)))
		return (pte_t *) 0x3b;

	pmd = pmd_offset(pud, addr);
	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
		return (pte_t *) 0x10;

	return pte_offset_map(pmd, addr);
}

static __always_inline size_t __user_copy_pt(unsigned long uaddr, void *kptr,
					     size_t n, int write_user)
{
	struct mm_struct *mm = current->mm;
	unsigned long offset, pfn, done, size;
	pte_t *pte;
	void *from, *to;

	done = 0;
retry:
	spin_lock(&mm->page_table_lock);
	do {
		pte = follow_table(mm, uaddr);
		if ((unsigned long) pte < 0x1000)
			goto fault;
		if (!pte_present(*pte)) {
			pte = (pte_t *) 0x11;
			goto fault;
		} else if (write_user && !pte_write(*pte)) {
			pte = (pte_t *) 0x04;
			goto fault;
		}

		pfn = pte_pfn(*pte);
		offset = uaddr & (PAGE_SIZE - 1);
		size = min(n - done, PAGE_SIZE - offset);
		if (write_user) {
			to = (void *)((pfn << PAGE_SHIFT) + offset);
			from = kptr + done;
		} else {
			from = (void *)((pfn << PAGE_SHIFT) + offset);
			to = kptr + done;
		}
		memcpy(to, from, size);
		done += size;
		uaddr += size;
	} while (done < n);
	spin_unlock(&mm->page_table_lock);
	return n - done;
fault:
	spin_unlock(&mm->page_table_lock);
	if (__handle_fault(uaddr, (unsigned long) pte, write_user))
		return n - done;
	goto retry;
}

/*
 * Do DAT for user address by page table walk, return kernel address.
 * This function needs to be called with current->mm->page_table_lock held.
 */
static __always_inline unsigned long __dat_user_addr(unsigned long uaddr)
{
	struct mm_struct *mm = current->mm;
	unsigned long pfn;
	pte_t *pte;
	int rc;

retry:
	pte = follow_table(mm, uaddr);
	if ((unsigned long) pte < 0x1000)
		goto fault;
	if (!pte_present(*pte)) {
		pte = (pte_t *) 0x11;
		goto fault;
	}

	pfn = pte_pfn(*pte);
	return (pfn << PAGE_SHIFT) + (uaddr & (PAGE_SIZE - 1));
fault:
	spin_unlock(&mm->page_table_lock);
	rc = __handle_fault(uaddr, (unsigned long) pte, 0);
	spin_lock(&mm->page_table_lock);
	if (!rc)
		goto retry;
	return 0;
}

size_t copy_from_user_pt(size_t n, const void __user *from, void *to)
{
	size_t rc;

	if (segment_eq(get_fs(), KERNEL_DS)) {
		memcpy(to, (void __kernel __force *) from, n);
		return 0;
	}
	rc = __user_copy_pt((unsigned long) from, to, n, 0);
	if (unlikely(rc))
		memset(to + n - rc, 0, rc);
	return rc;
}

size_t copy_to_user_pt(size_t n, void __user *to, const void *from)
{
	if (segment_eq(get_fs(), KERNEL_DS)) {
		memcpy((void __kernel __force *) to, from, n);
		return 0;
	}
	return __user_copy_pt((unsigned long) to, (void *) from, n, 1);
}

static size_t clear_user_pt(size_t n, void __user *to)
{
	long done, size, ret;

	if (segment_eq(get_fs(), KERNEL_DS)) {
		memset((void __kernel __force *) to, 0, n);
		return 0;
	}
	done = 0;
	do {
		if (n - done > PAGE_SIZE)
			size = PAGE_SIZE;
		else
			size = n - done;
		ret = __user_copy_pt((unsigned long) to + done,
				      &empty_zero_page, size, 1);
		done += size;
		if (ret)
			return ret + n - done;
	} while (done < n);
	return 0;
}

static size_t strnlen_user_pt(size_t count, const char __user *src)
{
	char *addr;
	unsigned long uaddr = (unsigned long) src;
	struct mm_struct *mm = current->mm;
	unsigned long offset, pfn, done, len;
	pte_t *pte;
	size_t len_str;

	if (segment_eq(get_fs(), KERNEL_DS))
		return strnlen((const char __kernel __force *) src, count) + 1;
	done = 0;
retry:
	spin_lock(&mm->page_table_lock);
	do {
		pte = follow_table(mm, uaddr);
		if ((unsigned long) pte < 0x1000)
			goto fault;
		if (!pte_present(*pte)) {
			pte = (pte_t *) 0x11;
			goto fault;
		}

		pfn = pte_pfn(*pte);
		offset = uaddr & (PAGE_SIZE-1);
		addr = (char *)(pfn << PAGE_SHIFT) + offset;
		len = min(count - done, PAGE_SIZE - offset);
		len_str = strnlen(addr, len);
		done += len_str;
		uaddr += len_str;
	} while ((len_str == len) && (done < count));
	spin_unlock(&mm->page_table_lock);
	return done + 1;
fault:
	spin_unlock(&mm->page_table_lock);
	if (__handle_fault(uaddr, (unsigned long) pte, 0))
		return 0;
	goto retry;
}

static size_t strncpy_from_user_pt(size_t count, const char __user *src,
				   char *dst)
{
	size_t n = strnlen_user_pt(count, src);

	if (!n)
		return -EFAULT;
	if (n > count)
		n = count;
	if (segment_eq(get_fs(), KERNEL_DS)) {
		memcpy(dst, (const char __kernel __force *) src, n);
		if (dst[n-1] == '\0')
			return n-1;
		else
			return n;
	}
	if (__user_copy_pt((unsigned long) src, dst, n, 0))
		return -EFAULT;
	if (dst[n-1] == '\0')
		return n-1;
	else
		return n;
}

static size_t copy_in_user_pt(size_t n, void __user *to,
			      const void __user *from)
{
	struct mm_struct *mm = current->mm;
	unsigned long offset_from, offset_to, offset_max, pfn_from, pfn_to,
		      uaddr, done, size, error_code;
	unsigned long uaddr_from = (unsigned long) from;
	unsigned long uaddr_to = (unsigned long) to;
	pte_t *pte_from, *pte_to;
	int write_user;

	if (segment_eq(get_fs(), KERNEL_DS)) {
		memcpy((void __force *) to, (void __force *) from, n);
		return 0;
	}
	done = 0;
retry:
	spin_lock(&mm->page_table_lock);
	do {
		write_user = 0;
		uaddr = uaddr_from;
		pte_from = follow_table(mm, uaddr_from);
		error_code = (unsigned long) pte_from;
		if (error_code < 0x1000)
			goto fault;
		if (!pte_present(*pte_from)) {
			error_code = 0x11;
			goto fault;
		}

		write_user = 1;
		uaddr = uaddr_to;
		pte_to = follow_table(mm, uaddr_to);
		error_code = (unsigned long) pte_to;
		if (error_code < 0x1000)
			goto fault;
		if (!pte_present(*pte_to)) {
			error_code = 0x11;
			goto fault;
		} else if (!pte_write(*pte_to)) {
			error_code = 0x04;
			goto fault;
		}

		pfn_from = pte_pfn(*pte_from);
		pfn_to = pte_pfn(*pte_to);
		offset_from = uaddr_from & (PAGE_SIZE-1);
		offset_to = uaddr_from & (PAGE_SIZE-1);
		offset_max = max(offset_from, offset_to);
		size = min(n - done, PAGE_SIZE - offset_max);

		memcpy((void *)(pfn_to << PAGE_SHIFT) + offset_to,
		       (void *)(pfn_from << PAGE_SHIFT) + offset_from, size);
		done += size;
		uaddr_from += size;
		uaddr_to += size;
	} while (done < n);
	spin_unlock(&mm->page_table_lock);
	return n - done;
fault:
	spin_unlock(&mm->page_table_lock);
	if (__handle_fault(uaddr, error_code, write_user))
		return n - done;
	goto retry;
}

#define __futex_atomic_op(insn, ret, oldval, newval, uaddr, oparg)	\
	asm volatile("0: l   %1,0(%6)\n"				\
		     "1: " insn						\
		     "2: cs  %1,%2,0(%6)\n"				\
		     "3: jl  1b\n"					\
		     "   lhi %0,0\n"					\
		     "4:\n"						\
		     EX_TABLE(0b,4b) EX_TABLE(2b,4b) EX_TABLE(3b,4b)	\
		     : "=d" (ret), "=&d" (oldval), "=&d" (newval),	\
		       "=m" (*uaddr)					\
		     : "0" (-EFAULT), "d" (oparg), "a" (uaddr),		\
		       "m" (*uaddr) : "cc" );

static int __futex_atomic_op_pt(int op, u32 __user *uaddr, int oparg, int *old)
{
	int oldval = 0, newval, ret;

	switch (op) {
	case FUTEX_OP_SET:
		__futex_atomic_op("lr %2,%5\n",
				  ret, oldval, newval, uaddr, oparg);
		break;
	case FUTEX_OP_ADD:
		__futex_atomic_op("lr %2,%1\nar %2,%5\n",
				  ret, oldval, newval, uaddr, oparg);
		break;
	case FUTEX_OP_OR:
		__futex_atomic_op("lr %2,%1\nor %2,%5\n",
				  ret, oldval, newval, uaddr, oparg);
		break;
	case FUTEX_OP_ANDN:
		__futex_atomic_op("lr %2,%1\nnr %2,%5\n",
				  ret, oldval, newval, uaddr, oparg);
		break;
	case FUTEX_OP_XOR:
		__futex_atomic_op("lr %2,%1\nxr %2,%5\n",
				  ret, oldval, newval, uaddr, oparg);
		break;
	default:
		ret = -ENOSYS;
	}
	if (ret == 0)
		*old = oldval;
	return ret;
}

int futex_atomic_op_pt(int op, u32 __user *uaddr, int oparg, int *old)
{
	int ret;

	if (segment_eq(get_fs(), KERNEL_DS))
		return __futex_atomic_op_pt(op, uaddr, oparg, old);
	spin_lock(&current->mm->page_table_lock);
	uaddr = (int __user *) __dat_user_addr((unsigned long) uaddr);
	if (!uaddr) {
		spin_unlock(&current->mm->page_table_lock);
		return -EFAULT;
	}
	get_page(virt_to_page(uaddr));
	spin_unlock(&current->mm->page_table_lock);
	ret = __futex_atomic_op_pt(op, uaddr, oparg, old);
	put_page(virt_to_page(uaddr));
	return ret;
}

static int __futex_atomic_cmpxchg_pt(u32 *uval, u32 __user *uaddr,
				     u32 oldval, u32 newval)
{
	int ret;

	asm volatile("0: cs   %1,%4,0(%5)\n"
		     "1: la   %0,0\n"
		     "2:\n"
		     EX_TABLE(0b,2b) EX_TABLE(1b,2b)
		     : "=d" (ret), "+d" (oldval), "=m" (*uaddr)
		     : "0" (-EFAULT), "d" (newval), "a" (uaddr), "m" (*uaddr)
		     : "cc", "memory" );
	*uval = oldval;
	return ret;
}

int futex_atomic_cmpxchg_pt(u32 *uval, u32 __user *uaddr,
			    u32 oldval, u32 newval)
{
	int ret;

	if (segment_eq(get_fs(), KERNEL_DS))
		return __futex_atomic_cmpxchg_pt(uval, uaddr, oldval, newval);
	spin_lock(&current->mm->page_table_lock);
	uaddr = (int __user *) __dat_user_addr((unsigned long) uaddr);
	if (!uaddr) {
		spin_unlock(&current->mm->page_table_lock);
		return -EFAULT;
	}
	get_page(virt_to_page(uaddr));
	spin_unlock(&current->mm->page_table_lock);
	ret = __futex_atomic_cmpxchg_pt(uval, uaddr, oldval, newval);
	put_page(virt_to_page(uaddr));
	return ret;
}

struct uaccess_ops uaccess_pt = {
	.copy_from_user		= copy_from_user_pt,
	.copy_from_user_small	= copy_from_user_pt,
	.copy_to_user		= copy_to_user_pt,
	.copy_to_user_small	= copy_to_user_pt,
	.copy_in_user		= copy_in_user_pt,
	.clear_user		= clear_user_pt,
	.strnlen_user		= strnlen_user_pt,
	.strncpy_from_user	= strncpy_from_user_pt,
	.futex_atomic_op	= futex_atomic_op_pt,
	.futex_atomic_cmpxchg	= futex_atomic_cmpxchg_pt,
};