info.c 6.4 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242
  1. /*
  2. * Copyright (c) 2006 Oracle. All rights reserved.
  3. *
  4. * This software is available to you under a choice of one of two
  5. * licenses. You may choose to be licensed under the terms of the GNU
  6. * General Public License (GPL) Version 2, available from the file
  7. * COPYING in the main directory of this source tree, or the
  8. * OpenIB.org BSD license below:
  9. *
  10. * Redistribution and use in source and binary forms, with or
  11. * without modification, are permitted provided that the following
  12. * conditions are met:
  13. *
  14. * - Redistributions of source code must retain the above
  15. * copyright notice, this list of conditions and the following
  16. * disclaimer.
  17. *
  18. * - Redistributions in binary form must reproduce the above
  19. * copyright notice, this list of conditions and the following
  20. * disclaimer in the documentation and/or other materials
  21. * provided with the distribution.
  22. *
  23. * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
  24. * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  25. * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
  26. * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
  27. * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
  28. * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  29. * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  30. * SOFTWARE.
  31. *
  32. */
  33. #include <linux/percpu.h>
  34. #include <linux/seq_file.h>
  35. #include <linux/slab.h>
  36. #include <linux/proc_fs.h>
  37. #include <linux/export.h>
  38. #include "rds.h"
  39. /*
  40. * This file implements a getsockopt() call which copies a set of fixed
  41. * sized structs into a user-specified buffer as a means of providing
  42. * read-only information about RDS.
  43. *
  44. * For a given information source there are a given number of fixed sized
  45. * structs at a given time. The structs are only copied if the user-specified
  46. * buffer is big enough. The destination pages that make up the buffer
  47. * are pinned for the duration of the copy.
  48. *
  49. * This gives us the following benefits:
  50. *
  51. * - simple implementation, no copy "position" across multiple calls
  52. * - consistent snapshot of an info source
  53. * - atomic copy works well with whatever locking info source has
  54. * - one portable tool to get rds info across implementations
  55. * - long-lived tool can get info without allocating
  56. *
  57. * at the following costs:
  58. *
  59. * - info source copy must be pinned, may be "large"
  60. */
  61. struct rds_info_iterator {
  62. struct page **pages;
  63. void *addr;
  64. unsigned long offset;
  65. };
  66. static DEFINE_SPINLOCK(rds_info_lock);
  67. static rds_info_func rds_info_funcs[RDS_INFO_LAST - RDS_INFO_FIRST + 1];
  68. void rds_info_register_func(int optname, rds_info_func func)
  69. {
  70. int offset = optname - RDS_INFO_FIRST;
  71. BUG_ON(optname < RDS_INFO_FIRST || optname > RDS_INFO_LAST);
  72. spin_lock(&rds_info_lock);
  73. BUG_ON(rds_info_funcs[offset]);
  74. rds_info_funcs[offset] = func;
  75. spin_unlock(&rds_info_lock);
  76. }
  77. EXPORT_SYMBOL_GPL(rds_info_register_func);
  78. void rds_info_deregister_func(int optname, rds_info_func func)
  79. {
  80. int offset = optname - RDS_INFO_FIRST;
  81. BUG_ON(optname < RDS_INFO_FIRST || optname > RDS_INFO_LAST);
  82. spin_lock(&rds_info_lock);
  83. BUG_ON(rds_info_funcs[offset] != func);
  84. rds_info_funcs[offset] = NULL;
  85. spin_unlock(&rds_info_lock);
  86. }
  87. EXPORT_SYMBOL_GPL(rds_info_deregister_func);
  88. /*
  89. * Typically we hold an atomic kmap across multiple rds_info_copy() calls
  90. * because the kmap is so expensive. This must be called before using blocking
  91. * operations while holding the mapping and as the iterator is torn down.
  92. */
  93. void rds_info_iter_unmap(struct rds_info_iterator *iter)
  94. {
  95. if (iter->addr) {
  96. kunmap_atomic(iter->addr);
  97. iter->addr = NULL;
  98. }
  99. }
  100. /*
  101. * get_user_pages() called flush_dcache_page() on the pages for us.
  102. */
  103. void rds_info_copy(struct rds_info_iterator *iter, void *data,
  104. unsigned long bytes)
  105. {
  106. unsigned long this;
  107. while (bytes) {
  108. if (!iter->addr)
  109. iter->addr = kmap_atomic(*iter->pages);
  110. this = min(bytes, PAGE_SIZE - iter->offset);
  111. rdsdebug("page %p addr %p offset %lu this %lu data %p "
  112. "bytes %lu\n", *iter->pages, iter->addr,
  113. iter->offset, this, data, bytes);
  114. memcpy(iter->addr + iter->offset, data, this);
  115. data += this;
  116. bytes -= this;
  117. iter->offset += this;
  118. if (iter->offset == PAGE_SIZE) {
  119. kunmap_atomic(iter->addr);
  120. iter->addr = NULL;
  121. iter->offset = 0;
  122. iter->pages++;
  123. }
  124. }
  125. }
  126. EXPORT_SYMBOL_GPL(rds_info_copy);
  127. /*
  128. * @optval points to the userspace buffer that the information snapshot
  129. * will be copied into.
  130. *
  131. * @optlen on input is the size of the buffer in userspace. @optlen
  132. * on output is the size of the requested snapshot in bytes.
  133. *
  134. * This function returns -errno if there is a failure, particularly -ENOSPC
  135. * if the given userspace buffer was not large enough to fit the snapshot.
  136. * On success it returns the positive number of bytes of each array element
  137. * in the snapshot.
  138. */
  139. int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
  140. int __user *optlen)
  141. {
  142. struct rds_info_iterator iter;
  143. struct rds_info_lengths lens;
  144. unsigned long nr_pages = 0;
  145. unsigned long start;
  146. rds_info_func func;
  147. struct page **pages = NULL;
  148. int ret;
  149. int len;
  150. int total;
  151. if (get_user(len, optlen)) {
  152. ret = -EFAULT;
  153. goto out;
  154. }
  155. /* check for all kinds of wrapping and the like */
  156. start = (unsigned long)optval;
  157. if (len < 0 || len > INT_MAX - PAGE_SIZE + 1 || start + len < start) {
  158. ret = -EINVAL;
  159. goto out;
  160. }
  161. /* a 0 len call is just trying to probe its length */
  162. if (len == 0)
  163. goto call_func;
  164. nr_pages = (PAGE_ALIGN(start + len) - (start & PAGE_MASK))
  165. >> PAGE_SHIFT;
  166. pages = kmalloc_array(nr_pages, sizeof(struct page *), GFP_KERNEL);
  167. if (!pages) {
  168. ret = -ENOMEM;
  169. goto out;
  170. }
  171. ret = pin_user_pages_fast(start, nr_pages, FOLL_WRITE, pages);
  172. if (ret != nr_pages) {
  173. if (ret > 0)
  174. nr_pages = ret;
  175. else
  176. nr_pages = 0;
  177. ret = -EAGAIN; /* XXX ? */
  178. goto out;
  179. }
  180. rdsdebug("len %d nr_pages %lu\n", len, nr_pages);
  181. call_func:
  182. func = rds_info_funcs[optname - RDS_INFO_FIRST];
  183. if (!func) {
  184. ret = -ENOPROTOOPT;
  185. goto out;
  186. }
  187. iter.pages = pages;
  188. iter.addr = NULL;
  189. iter.offset = start & (PAGE_SIZE - 1);
  190. func(sock, len, &iter, &lens);
  191. BUG_ON(lens.each == 0);
  192. total = lens.nr * lens.each;
  193. rds_info_iter_unmap(&iter);
  194. if (total > len) {
  195. len = total;
  196. ret = -ENOSPC;
  197. } else {
  198. len = total;
  199. ret = lens.each;
  200. }
  201. if (put_user(len, optlen))
  202. ret = -EFAULT;
  203. out:
  204. if (pages)
  205. unpin_user_pages(pages, nr_pages);
  206. kfree(pages);
  207. return ret;
  208. }