1 /*
2  * Copyright(c) 2016 Intel Corporation.
3  *
4  * This file is provided under a dual BSD/GPLv2 license.  When using or
5  * redistributing this file, you may do so under either license.
6  *
7  * GPL LICENSE SUMMARY
8  *
9  * This program is free software; you can redistribute it and/or modify
10  * it under the terms of version 2 of the GNU General Public License as
11  * published by the Free Software Foundation.
12  *
13  * This program is distributed in the hope that it will be useful, but
14  * WITHOUT ANY WARRANTY; without even the implied warranty of
15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
16  * General Public License for more details.
17  *
18  * BSD LICENSE
19  *
20  * Redistribution and use in source and binary forms, with or without
21  * modification, are permitted provided that the following conditions
22  * are met:
23  *
24  *  - Redistributions of source code must retain the above copyright
25  *    notice, this list of conditions and the following disclaimer.
26  *  - Redistributions in binary form must reproduce the above copyright
27  *    notice, this list of conditions and the following disclaimer in
28  *    the documentation and/or other materials provided with the
29  *    distribution.
30  *  - Neither the name of Intel Corporation nor the names of its
31  *    contributors may be used to endorse or promote products derived
32  *    from this software without specific prior written permission.
33  *
34  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
35  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
36  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
37  * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
38  * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
39  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
40  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
41  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
42  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
43  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
44  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
45  *
46  */
47 #include <linux/list.h>
48 #include <linux/rculist.h>
49 #include <linux/mmu_notifier.h>
50 #include <linux/interval_tree_generic.h>
51 
52 #include "mmu_rb.h"
53 #include "trace.h"
54 
55 struct mmu_rb_handler {
56 	struct mmu_notifier mn;
57 	struct rb_root root;
58 	void *ops_arg;
59 	spinlock_t lock;        /* protect the RB tree */
60 	struct mmu_rb_ops *ops;
61 	struct mm_struct *mm;
62 	struct list_head lru_list;
63 	struct work_struct del_work;
64 	struct list_head del_list;
65 	struct workqueue_struct *wq;
66 };
67 
68 static unsigned long mmu_node_start(struct mmu_rb_node *);
69 static unsigned long mmu_node_last(struct mmu_rb_node *);
70 static inline void mmu_notifier_page(struct mmu_notifier *, struct mm_struct *,
71 				     unsigned long);
72 static inline void mmu_notifier_range_start(struct mmu_notifier *,
73 					    struct mm_struct *,
74 					    unsigned long, unsigned long);
75 static void mmu_notifier_mem_invalidate(struct mmu_notifier *,
76 					struct mm_struct *,
77 					unsigned long, unsigned long);
78 static struct mmu_rb_node *__mmu_rb_search(struct mmu_rb_handler *,
79 					   unsigned long, unsigned long);
80 static void do_remove(struct mmu_rb_handler *handler,
81 		      struct list_head *del_list);
82 static void handle_remove(struct work_struct *work);
83 
84 static const struct mmu_notifier_ops mn_opts = {
85 	.invalidate_page = mmu_notifier_page,
86 	.invalidate_range_start = mmu_notifier_range_start,
87 };
88 
89 INTERVAL_TREE_DEFINE(struct mmu_rb_node, node, unsigned long, __last,
90 		     mmu_node_start, mmu_node_last, static, __mmu_int_rb);
91 
92 static unsigned long mmu_node_start(struct mmu_rb_node *node)
93 {
94 	return node->addr & PAGE_MASK;
95 }
96 
97 static unsigned long mmu_node_last(struct mmu_rb_node *node)
98 {
99 	return PAGE_ALIGN(node->addr + node->len) - 1;
100 }
101 
102 int hfi1_mmu_rb_register(void *ops_arg, struct mm_struct *mm,
103 			 struct mmu_rb_ops *ops,
104 			 struct workqueue_struct *wq,
105 			 struct mmu_rb_handler **handler)
106 {
107 	struct mmu_rb_handler *handlr;
108 	int ret;
109 
110 	handlr = kmalloc(sizeof(*handlr), GFP_KERNEL);
111 	if (!handlr)
112 		return -ENOMEM;
113 
114 	handlr->root = RB_ROOT;
115 	handlr->ops = ops;
116 	handlr->ops_arg = ops_arg;
117 	INIT_HLIST_NODE(&handlr->mn.hlist);
118 	spin_lock_init(&handlr->lock);
119 	handlr->mn.ops = &mn_opts;
120 	handlr->mm = mm;
121 	INIT_WORK(&handlr->del_work, handle_remove);
122 	INIT_LIST_HEAD(&handlr->del_list);
123 	INIT_LIST_HEAD(&handlr->lru_list);
124 	handlr->wq = wq;
125 
126 	ret = mmu_notifier_register(&handlr->mn, handlr->mm);
127 	if (ret) {
128 		kfree(handlr);
129 		return ret;
130 	}
131 
132 	*handler = handlr;
133 	return 0;
134 }
135 
136 void hfi1_mmu_rb_unregister(struct mmu_rb_handler *handler)
137 {
138 	struct mmu_rb_node *rbnode;
139 	struct rb_node *node;
140 	unsigned long flags;
141 	struct list_head del_list;
142 
143 	/* Unregister first so we don't get any more notifications. */
144 	mmu_notifier_unregister(&handler->mn, handler->mm);
145 
146 	/*
147 	 * Make sure the wq delete handler is finished running.  It will not
148 	 * be triggered once the mmu notifiers are unregistered above.
149 	 */
150 	flush_work(&handler->del_work);
151 
152 	INIT_LIST_HEAD(&del_list);
153 
154 	spin_lock_irqsave(&handler->lock, flags);
155 	while ((node = rb_first(&handler->root))) {
156 		rbnode = rb_entry(node, struct mmu_rb_node, node);
157 		rb_erase(node, &handler->root);
158 		/* move from LRU list to delete list */
159 		list_move(&rbnode->list, &del_list);
160 	}
161 	spin_unlock_irqrestore(&handler->lock, flags);
162 
163 	do_remove(handler, &del_list);
164 
165 	kfree(handler);
166 }
167 
168 int hfi1_mmu_rb_insert(struct mmu_rb_handler *handler,
169 		       struct mmu_rb_node *mnode)
170 {
171 	struct mmu_rb_node *node;
172 	unsigned long flags;
173 	int ret = 0;
174 
175 	spin_lock_irqsave(&handler->lock, flags);
176 	hfi1_cdbg(MMU, "Inserting node addr 0x%llx, len %u", mnode->addr,
177 		  mnode->len);
178 	node = __mmu_rb_search(handler, mnode->addr, mnode->len);
179 	if (node) {
180 		ret = -EINVAL;
181 		goto unlock;
182 	}
183 	__mmu_int_rb_insert(mnode, &handler->root);
184 	list_add(&mnode->list, &handler->lru_list);
185 
186 	ret = handler->ops->insert(handler->ops_arg, mnode);
187 	if (ret) {
188 		__mmu_int_rb_remove(mnode, &handler->root);
189 		list_del(&mnode->list); /* remove from LRU list */
190 	}
191 unlock:
192 	spin_unlock_irqrestore(&handler->lock, flags);
193 	return ret;
194 }
195 
196 /* Caller must hold handler lock */
197 static struct mmu_rb_node *__mmu_rb_search(struct mmu_rb_handler *handler,
198 					   unsigned long addr,
199 					   unsigned long len)
200 {
201 	struct mmu_rb_node *node = NULL;
202 
203 	hfi1_cdbg(MMU, "Searching for addr 0x%llx, len %u", addr, len);
204 	if (!handler->ops->filter) {
205 		node = __mmu_int_rb_iter_first(&handler->root, addr,
206 					       (addr + len) - 1);
207 	} else {
208 		for (node = __mmu_int_rb_iter_first(&handler->root, addr,
209 						    (addr + len) - 1);
210 		     node;
211 		     node = __mmu_int_rb_iter_next(node, addr,
212 						   (addr + len) - 1)) {
213 			if (handler->ops->filter(node, addr, len))
214 				return node;
215 		}
216 	}
217 	return node;
218 }
219 
220 struct mmu_rb_node *hfi1_mmu_rb_extract(struct mmu_rb_handler *handler,
221 					unsigned long addr, unsigned long len)
222 {
223 	struct mmu_rb_node *node;
224 	unsigned long flags;
225 
226 	spin_lock_irqsave(&handler->lock, flags);
227 	node = __mmu_rb_search(handler, addr, len);
228 	if (node) {
229 		__mmu_int_rb_remove(node, &handler->root);
230 		list_del(&node->list); /* remove from LRU list */
231 	}
232 	spin_unlock_irqrestore(&handler->lock, flags);
233 
234 	return node;
235 }
236 
237 void hfi1_mmu_rb_evict(struct mmu_rb_handler *handler, void *evict_arg)
238 {
239 	struct mmu_rb_node *rbnode, *ptr;
240 	struct list_head del_list;
241 	unsigned long flags;
242 	bool stop = false;
243 
244 	INIT_LIST_HEAD(&del_list);
245 
246 	spin_lock_irqsave(&handler->lock, flags);
247 	list_for_each_entry_safe_reverse(rbnode, ptr, &handler->lru_list,
248 					 list) {
249 		if (handler->ops->evict(handler->ops_arg, rbnode, evict_arg,
250 					&stop)) {
251 			__mmu_int_rb_remove(rbnode, &handler->root);
252 			/* move from LRU list to delete list */
253 			list_move(&rbnode->list, &del_list);
254 		}
255 		if (stop)
256 			break;
257 	}
258 	spin_unlock_irqrestore(&handler->lock, flags);
259 
260 	while (!list_empty(&del_list)) {
261 		rbnode = list_first_entry(&del_list, struct mmu_rb_node, list);
262 		list_del(&rbnode->list);
263 		handler->ops->remove(handler->ops_arg, rbnode);
264 	}
265 }
266 
267 /*
268  * It is up to the caller to ensure that this function does not race with the
269  * mmu invalidate notifier which may be calling the users remove callback on
270  * 'node'.
271  */
272 void hfi1_mmu_rb_remove(struct mmu_rb_handler *handler,
273 			struct mmu_rb_node *node)
274 {
275 	unsigned long flags;
276 
277 	/* Validity of handler and node pointers has been checked by caller. */
278 	hfi1_cdbg(MMU, "Removing node addr 0x%llx, len %u", node->addr,
279 		  node->len);
280 	spin_lock_irqsave(&handler->lock, flags);
281 	__mmu_int_rb_remove(node, &handler->root);
282 	list_del(&node->list); /* remove from LRU list */
283 	spin_unlock_irqrestore(&handler->lock, flags);
284 
285 	handler->ops->remove(handler->ops_arg, node);
286 }
287 
288 static inline void mmu_notifier_page(struct mmu_notifier *mn,
289 				     struct mm_struct *mm, unsigned long addr)
290 {
291 	mmu_notifier_mem_invalidate(mn, mm, addr, addr + PAGE_SIZE);
292 }
293 
294 static inline void mmu_notifier_range_start(struct mmu_notifier *mn,
295 					    struct mm_struct *mm,
296 					    unsigned long start,
297 					    unsigned long end)
298 {
299 	mmu_notifier_mem_invalidate(mn, mm, start, end);
300 }
301 
302 static void mmu_notifier_mem_invalidate(struct mmu_notifier *mn,
303 					struct mm_struct *mm,
304 					unsigned long start, unsigned long end)
305 {
306 	struct mmu_rb_handler *handler =
307 		container_of(mn, struct mmu_rb_handler, mn);
308 	struct rb_root *root = &handler->root;
309 	struct mmu_rb_node *node, *ptr = NULL;
310 	unsigned long flags;
311 	bool added = false;
312 
313 	spin_lock_irqsave(&handler->lock, flags);
314 	for (node = __mmu_int_rb_iter_first(root, start, end - 1);
315 	     node; node = ptr) {
316 		/* Guard against node removal. */
317 		ptr = __mmu_int_rb_iter_next(node, start, end - 1);
318 		hfi1_cdbg(MMU, "Invalidating node addr 0x%llx, len %u",
319 			  node->addr, node->len);
320 		if (handler->ops->invalidate(handler->ops_arg, node)) {
321 			__mmu_int_rb_remove(node, root);
322 			/* move from LRU list to delete list */
323 			list_move(&node->list, &handler->del_list);
324 			added = true;
325 		}
326 	}
327 	spin_unlock_irqrestore(&handler->lock, flags);
328 
329 	if (added)
330 		queue_work(handler->wq, &handler->del_work);
331 }
332 
333 /*
334  * Call the remove function for the given handler and the list.  This
335  * is expected to be called with a delete list extracted from handler.
336  * The caller should not be holding the handler lock.
337  */
338 static void do_remove(struct mmu_rb_handler *handler,
339 		      struct list_head *del_list)
340 {
341 	struct mmu_rb_node *node;
342 
343 	while (!list_empty(del_list)) {
344 		node = list_first_entry(del_list, struct mmu_rb_node, list);
345 		list_del(&node->list);
346 		handler->ops->remove(handler->ops_arg, node);
347 	}
348 }
349 
350 /*
351  * Work queue function to remove all nodes that have been queued up to
352  * be removed.  The key feature is that mm->mmap_sem is not being held
353  * and the remove callback can sleep while taking it, if needed.
354  */
355 static void handle_remove(struct work_struct *work)
356 {
357 	struct mmu_rb_handler *handler = container_of(work,
358 						struct mmu_rb_handler,
359 						del_work);
360 	struct list_head del_list;
361 	unsigned long flags;
362 
363 	/* remove anything that is queued to get removed */
364 	spin_lock_irqsave(&handler->lock, flags);
365 	list_replace_init(&handler->del_list, &del_list);
366 	spin_unlock_irqrestore(&handler->lock, flags);
367 
368 	do_remove(handler, &del_list);
369 }
370