1 /*-
2 * Copyright (c) 2010 Isilon Systems, Inc.
3 * Copyright (c) 2010 iX Systems, Inc.
4 * Copyright (c) 2010 Panasas, Inc.
5 * Copyright (c) 2013-2017 Mellanox Technologies, Ltd.
6 * Copyright (c) 2015 François Tigeot
7 * Copyright (c) 2015 Matthew Dillon <dillon@backplane.com>
8 * All rights reserved.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 * notice unmodified, this list of conditions, and the following
15 * disclaimer.
16 * 2. Redistributions in binary form must reproduce the above copyright
17 * notice, this list of conditions and the following disclaimer in the
18 * documentation and/or other materials provided with the distribution.
19 *
20 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
21 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
22 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
23 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
24 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
25 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
29 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30 */
31 #ifndef _LINUXKPI_LINUX_MM_H_
32 #define _LINUXKPI_LINUX_MM_H_
33
34 #include <linux/spinlock.h>
35 #include <linux/gfp.h>
36 #include <linux/kernel.h>
37 #include <linux/mm_types.h>
38 #include <linux/pfn.h>
39 #include <linux/list.h>
40 #include <linux/mmap_lock.h>
41 #include <linux/shrinker.h>
42 #include <linux/page.h>
43
44 #include <asm/pgtable.h>
45
46 #define PAGE_ALIGN(x) ALIGN(x, PAGE_SIZE)
47
48 /*
49 * Make sure our LinuxKPI defined virtual memory flags don't conflict
50 * with the ones defined by FreeBSD:
51 */
52 CTASSERT((VM_PROT_ALL & -(1 << 8)) == 0);
53
54 #define VM_READ VM_PROT_READ
55 #define VM_WRITE VM_PROT_WRITE
56 #define VM_EXEC VM_PROT_EXECUTE
57
58 #define VM_PFNINTERNAL (1 << 8) /* FreeBSD private flag to vm_insert_pfn() */
59 #define VM_MIXEDMAP (1 << 9)
60 #define VM_NORESERVE (1 << 10)
61 #define VM_PFNMAP (1 << 11)
62 #define VM_IO (1 << 12)
63 #define VM_MAYWRITE (1 << 13)
64 #define VM_DONTCOPY (1 << 14)
65 #define VM_DONTEXPAND (1 << 15)
66 #define VM_DONTDUMP (1 << 16)
67 #define VM_SHARED (1 << 17)
68
69 #define VMA_MAX_PREFAULT_RECORD 1
70
71 #define FOLL_WRITE (1 << 0)
72 #define FOLL_FORCE (1 << 1)
73
74 #define VM_FAULT_OOM (1 << 0)
75 #define VM_FAULT_SIGBUS (1 << 1)
76 #define VM_FAULT_MAJOR (1 << 2)
77 #define VM_FAULT_WRITE (1 << 3)
78 #define VM_FAULT_HWPOISON (1 << 4)
79 #define VM_FAULT_HWPOISON_LARGE (1 << 5)
80 #define VM_FAULT_SIGSEGV (1 << 6)
81 #define VM_FAULT_NOPAGE (1 << 7)
82 #define VM_FAULT_LOCKED (1 << 8)
83 #define VM_FAULT_RETRY (1 << 9)
84 #define VM_FAULT_FALLBACK (1 << 10)
85
86 #define VM_FAULT_ERROR (VM_FAULT_OOM | VM_FAULT_SIGBUS | VM_FAULT_SIGSEGV | \
87 VM_FAULT_HWPOISON |VM_FAULT_HWPOISON_LARGE | VM_FAULT_FALLBACK)
88
89 #define FAULT_FLAG_WRITE (1 << 0)
90 #define FAULT_FLAG_MKWRITE (1 << 1)
91 #define FAULT_FLAG_ALLOW_RETRY (1 << 2)
92 #define FAULT_FLAG_RETRY_NOWAIT (1 << 3)
93 #define FAULT_FLAG_KILLABLE (1 << 4)
94 #define FAULT_FLAG_TRIED (1 << 5)
95 #define FAULT_FLAG_USER (1 << 6)
96 #define FAULT_FLAG_REMOTE (1 << 7)
97 #define FAULT_FLAG_INSTRUCTION (1 << 8)
98
99 #define fault_flag_allow_retry_first(flags) \
100 (((flags) & (FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_TRIED)) == FAULT_FLAG_ALLOW_RETRY)
101
102 typedef int (*pte_fn_t)(linux_pte_t *, unsigned long addr, void *data);
103
104 struct vm_area_struct {
105 vm_offset_t vm_start;
106 vm_offset_t vm_end;
107 vm_offset_t vm_pgoff;
108 pgprot_t vm_page_prot;
109 unsigned long vm_flags;
110 struct mm_struct *vm_mm;
111 void *vm_private_data;
112 const struct vm_operations_struct *vm_ops;
113 struct linux_file *vm_file;
114
115 /* internal operation */
116 vm_paddr_t vm_pfn; /* PFN for memory map */
117 vm_size_t vm_len; /* length for memory map */
118 vm_pindex_t vm_pfn_first;
119 int vm_pfn_count;
120 int *vm_pfn_pcount;
121 vm_object_t vm_obj;
122 vm_map_t vm_cached_map;
123 TAILQ_ENTRY(vm_area_struct) vm_entry;
124 };
125
126 struct vm_fault {
127 unsigned int flags;
128 pgoff_t pgoff;
129 union {
130 /* user-space address */
131 void *virtual_address; /* < 4.11 */
132 unsigned long address; /* >= 4.11 */
133 };
134 struct page *page;
135 struct vm_area_struct *vma;
136 };
137
138 struct vm_operations_struct {
139 void (*open) (struct vm_area_struct *);
140 void (*close) (struct vm_area_struct *);
141 int (*fault) (struct vm_area_struct *, struct vm_fault *);
142 int (*access) (struct vm_area_struct *, unsigned long, void *, int, int);
143 };
144
145 struct sysinfo {
146 uint64_t totalram; /* Total usable main memory size */
147 uint64_t freeram; /* Available memory size */
148 uint64_t totalhigh; /* Total high memory size */
149 uint64_t freehigh; /* Available high memory size */
150 uint32_t mem_unit; /* Memory unit size in bytes */
151 };
152
153 static inline struct page *
virt_to_head_page(const void * p)154 virt_to_head_page(const void *p)
155 {
156
157 return (virt_to_page(p));
158 }
159
160 /*
161 * Compute log2 of the power of two rounded up count of pages
162 * needed for size bytes.
163 */
164 static inline int
get_order(unsigned long size)165 get_order(unsigned long size)
166 {
167 int order;
168
169 size = (size - 1) >> PAGE_SHIFT;
170 order = 0;
171 while (size) {
172 order++;
173 size >>= 1;
174 }
175 return (order);
176 }
177
178 static inline void *
lowmem_page_address(struct page * page)179 lowmem_page_address(struct page *page)
180 {
181 return (page_address(page));
182 }
183
184 /*
185 * This only works via memory map operations.
186 */
187 static inline int
io_remap_pfn_range(struct vm_area_struct * vma,unsigned long addr,unsigned long pfn,unsigned long size,vm_memattr_t prot)188 io_remap_pfn_range(struct vm_area_struct *vma,
189 unsigned long addr, unsigned long pfn, unsigned long size,
190 vm_memattr_t prot)
191 {
192 vma->vm_page_prot = prot;
193 vma->vm_pfn = pfn;
194 vma->vm_len = size;
195
196 return (0);
197 }
198
199 vm_fault_t
200 lkpi_vmf_insert_pfn_prot_locked(struct vm_area_struct *vma, unsigned long addr,
201 unsigned long pfn, pgprot_t prot);
202
203 static inline vm_fault_t
vmf_insert_pfn_prot(struct vm_area_struct * vma,unsigned long addr,unsigned long pfn,pgprot_t prot)204 vmf_insert_pfn_prot(struct vm_area_struct *vma, unsigned long addr,
205 unsigned long pfn, pgprot_t prot)
206 {
207 vm_fault_t ret;
208
209 VM_OBJECT_WLOCK(vma->vm_obj);
210 ret = lkpi_vmf_insert_pfn_prot_locked(vma, addr, pfn, prot);
211 VM_OBJECT_WUNLOCK(vma->vm_obj);
212
213 return (ret);
214 }
215 #define vmf_insert_pfn_prot(...) \
216 _Static_assert(false, \
217 "This function is always called in a loop. Consider using the locked version")
218
219 static inline int
apply_to_page_range(struct mm_struct * mm,unsigned long address,unsigned long size,pte_fn_t fn,void * data)220 apply_to_page_range(struct mm_struct *mm, unsigned long address,
221 unsigned long size, pte_fn_t fn, void *data)
222 {
223 return (-ENOTSUP);
224 }
225
226 int zap_vma_ptes(struct vm_area_struct *vma, unsigned long address,
227 unsigned long size);
228
229 int lkpi_remap_pfn_range(struct vm_area_struct *vma,
230 unsigned long start_addr, unsigned long start_pfn, unsigned long size,
231 pgprot_t prot);
232
233 static inline int
remap_pfn_range(struct vm_area_struct * vma,unsigned long addr,unsigned long pfn,unsigned long size,pgprot_t prot)234 remap_pfn_range(struct vm_area_struct *vma, unsigned long addr,
235 unsigned long pfn, unsigned long size, pgprot_t prot)
236 {
237 return (lkpi_remap_pfn_range(vma, addr, pfn, size, prot));
238 }
239
240 static inline unsigned long
vma_pages(struct vm_area_struct * vma)241 vma_pages(struct vm_area_struct *vma)
242 {
243 return ((vma->vm_end - vma->vm_start) >> PAGE_SHIFT);
244 }
245
246 #define offset_in_page(off) ((unsigned long)(off) & (PAGE_SIZE - 1))
247
248 static inline void
set_page_dirty(struct page * page)249 set_page_dirty(struct page *page)
250 {
251 vm_page_dirty(page);
252 }
253
254 static inline void
mark_page_accessed(struct page * page)255 mark_page_accessed(struct page *page)
256 {
257 vm_page_reference(page);
258 }
259
260 static inline void
get_page(struct page * page)261 get_page(struct page *page)
262 {
263 vm_page_wire(page);
264 }
265
266 extern long
267 get_user_pages(unsigned long start, unsigned long nr_pages,
268 unsigned int gup_flags, struct page **,
269 struct vm_area_struct **);
270
271 static inline long
pin_user_pages(unsigned long start,unsigned long nr_pages,unsigned int gup_flags,struct page ** pages,struct vm_area_struct ** vmas)272 pin_user_pages(unsigned long start, unsigned long nr_pages,
273 unsigned int gup_flags, struct page **pages,
274 struct vm_area_struct **vmas)
275 {
276 return get_user_pages(start, nr_pages, gup_flags, pages, vmas);
277 }
278
279 extern int
280 __get_user_pages_fast(unsigned long start, int nr_pages, int write,
281 struct page **);
282
283 static inline int
pin_user_pages_fast(unsigned long start,int nr_pages,unsigned int gup_flags,struct page ** pages)284 pin_user_pages_fast(unsigned long start, int nr_pages,
285 unsigned int gup_flags, struct page **pages)
286 {
287 return __get_user_pages_fast(
288 start, nr_pages, !!(gup_flags & FOLL_WRITE), pages);
289 }
290
291 extern long
292 get_user_pages_remote(struct task_struct *, struct mm_struct *,
293 unsigned long start, unsigned long nr_pages,
294 unsigned int gup_flags, struct page **,
295 struct vm_area_struct **);
296
297 static inline long
pin_user_pages_remote(struct task_struct * task,struct mm_struct * mm,unsigned long start,unsigned long nr_pages,unsigned int gup_flags,struct page ** pages,struct vm_area_struct ** vmas)298 pin_user_pages_remote(struct task_struct *task, struct mm_struct *mm,
299 unsigned long start, unsigned long nr_pages,
300 unsigned int gup_flags, struct page **pages,
301 struct vm_area_struct **vmas)
302 {
303 return get_user_pages_remote(
304 task, mm, start, nr_pages, gup_flags, pages, vmas);
305 }
306
307 static inline void
put_page(struct page * page)308 put_page(struct page *page)
309 {
310 vm_page_unwire(page, PQ_ACTIVE);
311 }
312
313 #define unpin_user_page(page) put_page(page)
314 #define unpin_user_pages(pages, npages) release_pages(pages, npages)
315
316 #define copy_highpage(to, from) pmap_copy_page(from, to)
317
318 static inline pgprot_t
vm_get_page_prot(unsigned long vm_flags)319 vm_get_page_prot(unsigned long vm_flags)
320 {
321 return (vm_flags & VM_PROT_ALL);
322 }
323
324 static inline struct page *
vmalloc_to_page(const void * addr)325 vmalloc_to_page(const void *addr)
326 {
327 vm_paddr_t paddr;
328
329 paddr = pmap_kextract((vm_offset_t)addr);
330 return (PHYS_TO_VM_PAGE(paddr));
331 }
332
333 static inline int
trylock_page(struct page * page)334 trylock_page(struct page *page)
335 {
336 return (vm_page_trylock(page));
337 }
338
339 static inline void
unlock_page(struct page * page)340 unlock_page(struct page *page)
341 {
342
343 vm_page_unlock(page);
344 }
345
346 extern int is_vmalloc_addr(const void *addr);
347 void si_meminfo(struct sysinfo *si);
348
349 static inline unsigned long
totalram_pages(void)350 totalram_pages(void)
351 {
352 return ((unsigned long)physmem);
353 }
354
355 #define unmap_mapping_range(...) lkpi_unmap_mapping_range(__VA_ARGS__)
356 void lkpi_unmap_mapping_range(void *obj, loff_t const holebegin __unused,
357 loff_t const holelen, int even_cows __unused);
358
359 #define PAGE_ALIGNED(p) __is_aligned(p, PAGE_SIZE)
360
361 void vma_set_file(struct vm_area_struct *vma, struct linux_file *file);
362
363 static inline void
might_alloc(gfp_t gfp_mask __unused)364 might_alloc(gfp_t gfp_mask __unused)
365 {
366 }
367
368 #define is_cow_mapping(flags) (false)
369
370 #endif /* _LINUXKPI_LINUX_MM_H_ */
371