Blame view
mm/memory-failure.c
48.4 KB
6a46079cf HWPOISON: The hig... |
1 2 3 4 5 6 7 8 9 |
/* * Copyright (C) 2008, 2009 Intel Corporation * Authors: Andi Kleen, Fengguang Wu * * This software may be redistributed and/or modified under the terms of * the GNU General Public License ("GPL") version 2 only as published by the * Free Software Foundation. * * High level machine check handler. Handles pages reported by the |
1c80b990a HWPOISON: Improve... |
10 |
* hardware as being corrupted usually due to a multi-bit ECC memory or cache |
6a46079cf HWPOISON: The hig... |
11 |
* failure. |
1c80b990a HWPOISON: Improve... |
12 13 14 |
* * In addition there is a "soft offline" entry point that allows stop using * not-yet-corrupted-by-suspicious pages without killing anything. |
6a46079cf HWPOISON: The hig... |
15 16 |
* * Handles page cache pages in various states. The tricky part |
1c80b990a HWPOISON: Improve... |
17 18 19 20 21 22 |
* here is that we can access any page asynchronously in respect to * other VM users, because memory failures could happen anytime and * anywhere. This could violate some of their assumptions. This is why * this code has to be extremely careful. Generally it tries to use * normal locking rules, as in get the standard locks, even if that means * the error handling takes potentially a long time. |
e0de78dfb mm, hwpoison: add... |
23 24 25 26 27 28 29 30 |
* * It can be very tempting to add handling for obscure cases here. * In general any code for handling new cases should only be added iff: * - You know how to test it. * - You have a test that can be added to mce-test * https://git.kernel.org/cgit/utils/cpu/mce/mce-test.git/ * - The case actually shows up as a frequent (top 10) page state in * tools/vm/page-types when running a real workload. |
1c80b990a HWPOISON: Improve... |
31 32 33 34 35 36 37 |
* * There are several operations here with exponential complexity because * of unsuitable VM data structures. For example the operation to map back * from RMAP chains to processes has to walk the complete process list and * has non linear complexity with the number. But since memory corruptions * are rare we hope to get away with this. This avoids impacting the core * VM. |
6a46079cf HWPOISON: The hig... |
38 |
*/ |
6a46079cf HWPOISON: The hig... |
39 40 41 |
#include <linux/kernel.h> #include <linux/mm.h> #include <linux/page-flags.h> |
478c5ffc0 HWPOISON: add pag... |
42 |
#include <linux/kernel-page-flags.h> |
3f07c0144 sched/headers: Pr... |
43 |
#include <linux/sched/signal.h> |
299300258 sched/headers: Pr... |
44 |
#include <linux/sched/task.h> |
01e00f880 HWPOISON: fix oop... |
45 |
#include <linux/ksm.h> |
6a46079cf HWPOISON: The hig... |
46 |
#include <linux/rmap.h> |
b9e15bafd mm: Add export.h ... |
47 |
#include <linux/export.h> |
6a46079cf HWPOISON: The hig... |
48 49 50 |
#include <linux/pagemap.h> #include <linux/swap.h> #include <linux/backing-dev.h> |
facb6011f HWPOISON: Add sof... |
51 |
#include <linux/migrate.h> |
facb6011f HWPOISON: Add sof... |
52 |
#include <linux/suspend.h> |
5a0e3ad6a include cleanup: ... |
53 |
#include <linux/slab.h> |
bf998156d KVM: Avoid killin... |
54 |
#include <linux/swapops.h> |
7af446a84 HWPOISON, hugetlb... |
55 |
#include <linux/hugetlb.h> |
20d6c96b5 mem-hotplug: intr... |
56 |
#include <linux/memory_hotplug.h> |
5db8a73a8 mm/memory-failure... |
57 |
#include <linux/mm_inline.h> |
ea8f5fb8a HWPoison: add mem... |
58 |
#include <linux/kfifo.h> |
a5f651090 mm: hwpoison: rat... |
59 |
#include <linux/ratelimit.h> |
6a46079cf HWPOISON: The hig... |
60 |
#include "internal.h" |
97f0b1345 tracing: add trac... |
61 |
#include "ras/ras_event.h" |
6a46079cf HWPOISON: The hig... |
62 63 64 65 |
int sysctl_memory_failure_early_kill __read_mostly = 0; int sysctl_memory_failure_recovery __read_mostly = 1; |
293c07e31 memory-failure: u... |
66 |
atomic_long_t num_poisoned_pages __read_mostly = ATOMIC_LONG_INIT(0); |
6a46079cf HWPOISON: The hig... |
67 |
|
27df5068e HWPOISON: Add PRO... |
68 |
#if defined(CONFIG_HWPOISON_INJECT) || defined(CONFIG_HWPOISON_INJECT_MODULE) |
1bfe5febe HWPOISON: add an ... |
69 |
u32 hwpoison_filter_enable = 0; |
7c116f2b0 HWPOISON: add fs/... |
70 71 |
u32 hwpoison_filter_dev_major = ~0U; u32 hwpoison_filter_dev_minor = ~0U; |
478c5ffc0 HWPOISON: add pag... |
72 73 |
u64 hwpoison_filter_flags_mask; u64 hwpoison_filter_flags_value; |
1bfe5febe HWPOISON: add an ... |
74 |
EXPORT_SYMBOL_GPL(hwpoison_filter_enable); |
7c116f2b0 HWPOISON: add fs/... |
75 76 |
EXPORT_SYMBOL_GPL(hwpoison_filter_dev_major); EXPORT_SYMBOL_GPL(hwpoison_filter_dev_minor); |
478c5ffc0 HWPOISON: add pag... |
77 78 |
EXPORT_SYMBOL_GPL(hwpoison_filter_flags_mask); EXPORT_SYMBOL_GPL(hwpoison_filter_flags_value); |
7c116f2b0 HWPOISON: add fs/... |
79 80 81 82 83 84 85 86 87 88 89 |
static int hwpoison_filter_dev(struct page *p) { struct address_space *mapping; dev_t dev; if (hwpoison_filter_dev_major == ~0U && hwpoison_filter_dev_minor == ~0U) return 0; /* |
1c80b990a HWPOISON: Improve... |
90 |
* page_mapping() does not accept slab pages. |
7c116f2b0 HWPOISON: add fs/... |
91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 |
*/ if (PageSlab(p)) return -EINVAL; mapping = page_mapping(p); if (mapping == NULL || mapping->host == NULL) return -EINVAL; dev = mapping->host->i_sb->s_dev; if (hwpoison_filter_dev_major != ~0U && hwpoison_filter_dev_major != MAJOR(dev)) return -EINVAL; if (hwpoison_filter_dev_minor != ~0U && hwpoison_filter_dev_minor != MINOR(dev)) return -EINVAL; return 0; } |
478c5ffc0 HWPOISON: add pag... |
109 110 111 112 113 114 115 116 117 118 119 |
static int hwpoison_filter_flags(struct page *p) { if (!hwpoison_filter_flags_mask) return 0; if ((stable_page_flags(p) & hwpoison_filter_flags_mask) == hwpoison_filter_flags_value) return 0; else return -EINVAL; } |
4fd466eb4 HWPOISON: add mem... |
120 121 122 123 124 125 126 127 128 129 |
/* * This allows stress tests to limit test scope to a collection of tasks * by putting them under some memcg. This prevents killing unrelated/important * processes such as /sbin/init. Note that the target task may share clean * pages with init (eg. libc text), which is harmless. If the target task * share _dirty_ pages with another task B, the test scheme must make sure B * is also included in the memcg. At last, due to race conditions this filter * can only guarantee that the page either belongs to the memcg tasks, or is * a freed page. */ |
94a59fb36 hwpoison: use pag... |
130 |
#ifdef CONFIG_MEMCG |
4fd466eb4 HWPOISON: add mem... |
131 132 133 134 |
u64 hwpoison_filter_memcg; EXPORT_SYMBOL_GPL(hwpoison_filter_memcg); static int hwpoison_filter_task(struct page *p) { |
4fd466eb4 HWPOISON: add mem... |
135 136 |
if (!hwpoison_filter_memcg) return 0; |
94a59fb36 hwpoison: use pag... |
137 |
if (page_cgroup_ino(p) != hwpoison_filter_memcg) |
4fd466eb4 HWPOISON: add mem... |
138 139 140 141 142 143 144 |
return -EINVAL; return 0; } #else static int hwpoison_filter_task(struct page *p) { return 0; } #endif |
7c116f2b0 HWPOISON: add fs/... |
145 146 |
int hwpoison_filter(struct page *p) { |
1bfe5febe HWPOISON: add an ... |
147 148 |
if (!hwpoison_filter_enable) return 0; |
7c116f2b0 HWPOISON: add fs/... |
149 150 |
if (hwpoison_filter_dev(p)) return -EINVAL; |
478c5ffc0 HWPOISON: add pag... |
151 152 |
if (hwpoison_filter_flags(p)) return -EINVAL; |
4fd466eb4 HWPOISON: add mem... |
153 154 |
if (hwpoison_filter_task(p)) return -EINVAL; |
7c116f2b0 HWPOISON: add fs/... |
155 156 |
return 0; } |
27df5068e HWPOISON: Add PRO... |
157 158 159 160 161 162 |
#else int hwpoison_filter(struct page *p) { return 0; } #endif |
7c116f2b0 HWPOISON: add fs/... |
163 |
EXPORT_SYMBOL_GPL(hwpoison_filter); |
6a46079cf HWPOISON: The hig... |
164 |
/* |
7329bbeb9 HWPOISON: Add cod... |
165 166 167 |
* Send all the processes who have the page mapped a signal. * ``action optional'' if they are not immediately affected by the error * ``action required'' if error happened in current execution context |
6a46079cf HWPOISON: The hig... |
168 |
*/ |
7329bbeb9 HWPOISON: Add cod... |
169 170 |
static int kill_proc(struct task_struct *t, unsigned long addr, int trapno, unsigned long pfn, struct page *page, int flags) |
6a46079cf HWPOISON: The hig... |
171 172 173 |
{ struct siginfo si; int ret; |
495367c05 mm/memory-failure... |
174 175 176 |
pr_err("Memory failure: %#lx: Killing %s:%d due to hardware memory corruption ", pfn, t->comm, t->pid); |
6a46079cf HWPOISON: The hig... |
177 178 |
si.si_signo = SIGBUS; si.si_errno = 0; |
6a46079cf HWPOISON: The hig... |
179 180 181 182 |
si.si_addr = (void *)addr; #ifdef __ARCH_SI_TRAPNO si.si_trapno = trapno; #endif |
f9121153f mm/hwpoison: don'... |
183 |
si.si_addr_lsb = compound_order(compound_head(page)) + PAGE_SHIFT; |
7329bbeb9 HWPOISON: Add cod... |
184 |
|
a70ffcac7 mm/memory-failure... |
185 |
if ((flags & MF_ACTION_REQUIRED) && t->mm == current->mm) { |
7329bbeb9 HWPOISON: Add cod... |
186 |
si.si_code = BUS_MCEERR_AR; |
a70ffcac7 mm/memory-failure... |
187 |
ret = force_sig_info(SIGBUS, &si, current); |
7329bbeb9 HWPOISON: Add cod... |
188 189 190 191 192 193 194 195 196 197 |
} else { /* * Don't use force here, it's convenient if the signal * can be temporarily blocked. * This could cause a loop when the user sets SIGBUS * to SIG_IGN, but hopefully no one will do that? */ si.si_code = BUS_MCEERR_AO; ret = send_sig_info(SIGBUS, &si, t); /* synchronous? */ } |
6a46079cf HWPOISON: The hig... |
198 |
if (ret < 0) |
495367c05 mm/memory-failure... |
199 200 |
pr_info("Memory failure: Error sending signal to %s:%d: %d ", |
1170532bb mm: convert print... |
201 |
t->comm, t->pid, ret); |
6a46079cf HWPOISON: The hig... |
202 203 204 205 |
return ret; } /* |
588f9ce6c HWPOISON: Be more... |
206 207 208 |
* When a unknown page type is encountered drain as many buffers as possible * in the hope to turn the page into a LRU or free page, which we can handle. */ |
facb6011f HWPOISON: Add sof... |
209 |
void shake_page(struct page *p, int access) |
588f9ce6c HWPOISON: Be more... |
210 |
{ |
8bcb74de7 mm: hwpoison: cal... |
211 212 |
if (PageHuge(p)) return; |
588f9ce6c HWPOISON: Be more... |
213 214 215 216 |
if (!PageSlab(p)) { lru_add_drain_all(); if (PageLRU(p)) return; |
c05543293 mm, memory_hotplu... |
217 |
drain_all_pages(page_zone(p)); |
588f9ce6c HWPOISON: Be more... |
218 219 220 |
if (PageLRU(p) || is_free_buddy_page(p)) return; } |
facb6011f HWPOISON: Add sof... |
221 |
|
588f9ce6c HWPOISON: Be more... |
222 |
/* |
6b4f7799c mm: vmscan: invok... |
223 224 |
* Only call shrink_node_slabs here (which would also shrink * other caches) if access is not potentially fatal. |
588f9ce6c HWPOISON: Be more... |
225 |
*/ |
cb731d6c6 vmscan: per memor... |
226 227 |
if (access) drop_slab_node(page_to_nid(p)); |
588f9ce6c HWPOISON: Be more... |
228 229 230 231 |
} EXPORT_SYMBOL_GPL(shake_page); /* |
6a46079cf HWPOISON: The hig... |
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 |
* Kill all processes that have a poisoned page mapped and then isolate * the page. * * General strategy: * Find all processes having the page mapped and kill them. * But we keep a page reference around so that the page is not * actually freed yet. * Then stash the page away * * There's no convenient way to get back to mapped processes * from the VMAs. So do a brute-force search over all * running processes. * * Remember that machine checks are not common (or rather * if they are common you have other problems), so this shouldn't * be a performance issue. * * Also there are some races possible while we get from the * error detection to actually handle it. */ struct to_kill { struct list_head nd; struct task_struct *tsk; unsigned long addr; |
9033ae164 HWPOISON: Turn ad... |
257 |
char addr_valid; |
6a46079cf HWPOISON: The hig... |
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 |
}; /* * Failure handling: if we can't find or can't kill a process there's * not much we can do. We just print a message and ignore otherwise. */ /* * Schedule a process for later kill. * Uses GFP_ATOMIC allocations to avoid potential recursions in the VM. * TBD would GFP_NOIO be enough? */ static void add_to_kill(struct task_struct *tsk, struct page *p, struct vm_area_struct *vma, struct list_head *to_kill, struct to_kill **tkc) { struct to_kill *tk; if (*tkc) { tk = *tkc; *tkc = NULL; } else { tk = kmalloc(sizeof(struct to_kill), GFP_ATOMIC); if (!tk) { |
495367c05 mm/memory-failure... |
283 284 |
pr_err("Memory failure: Out of memory while machine check handling "); |
6a46079cf HWPOISON: The hig... |
285 286 287 288 289 290 291 292 293 294 295 296 297 |
return; } } tk->addr = page_address_in_vma(p, vma); tk->addr_valid = 1; /* * In theory we don't have to kill when the page was * munmaped. But it could be also a mremap. Since that's * likely very rare kill anyways just out of paranoia, but use * a SIGKILL because the error is not contained anymore. */ if (tk->addr == -EFAULT) { |
495367c05 mm/memory-failure... |
298 299 |
pr_info("Memory failure: Unable to find user space address %lx in %s ", |
6a46079cf HWPOISON: The hig... |
300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 |
page_to_pfn(p), tsk->comm); tk->addr_valid = 0; } get_task_struct(tsk); tk->tsk = tsk; list_add_tail(&tk->nd, to_kill); } /* * Kill the processes that have been collected earlier. * * Only do anything when DOIT is set, otherwise just free the list * (this is used for clean pages which do not need killing) * Also when FAIL is set do a force kill because something went * wrong earlier. */ |
6751ed65d x86/mce: Fix sigi... |
316 |
static void kill_procs(struct list_head *to_kill, int forcekill, int trapno, |
666e5a406 mm: make ttu's re... |
317 |
bool fail, struct page *page, unsigned long pfn, |
7329bbeb9 HWPOISON: Add cod... |
318 |
int flags) |
6a46079cf HWPOISON: The hig... |
319 320 321 322 |
{ struct to_kill *tk, *next; list_for_each_entry_safe (tk, next, to_kill, nd) { |
6751ed65d x86/mce: Fix sigi... |
323 |
if (forcekill) { |
6a46079cf HWPOISON: The hig... |
324 |
/* |
af901ca18 tree-wide: fix as... |
325 |
* In case something went wrong with munmapping |
6a46079cf HWPOISON: The hig... |
326 327 |
* make sure the process doesn't catch the * signal and then access the memory. Just kill it. |
6a46079cf HWPOISON: The hig... |
328 329 |
*/ if (fail || tk->addr_valid == 0) { |
495367c05 mm/memory-failure... |
330 331 |
pr_err("Memory failure: %#lx: forcibly killing %s:%d because of failure to unmap corrupted page ", |
1170532bb mm: convert print... |
332 |
pfn, tk->tsk->comm, tk->tsk->pid); |
0783205e1 mm: hwpoison: use... |
333 334 |
do_send_sig_info(SIGKILL, SEND_SIG_PRIV, tk->tsk, PIDTYPE_PID); |
6a46079cf HWPOISON: The hig... |
335 336 337 338 339 340 341 342 |
} /* * In theory the process could have mapped * something else on the address in-between. We could * check for that, but we need to tell the * process anyways. */ |
7329bbeb9 HWPOISON: Add cod... |
343 344 |
else if (kill_proc(tk->tsk, tk->addr, trapno, pfn, page, flags) < 0) |
495367c05 mm/memory-failure... |
345 346 |
pr_err("Memory failure: %#lx: Cannot send advisory machine check signal to %s:%d ", |
1170532bb mm: convert print... |
347 |
pfn, tk->tsk->comm, tk->tsk->pid); |
6a46079cf HWPOISON: The hig... |
348 349 350 351 352 |
} put_task_struct(tk->tsk); kfree(tk); } } |
3ba08129e mm/memory-failure... |
353 354 355 356 357 358 359 360 361 |
/* * Find a dedicated thread which is supposed to handle SIGBUS(BUS_MCEERR_AO) * on behalf of the thread group. Return task_struct of the (first found) * dedicated thread if found, and return NULL otherwise. * * We already hold read_lock(&tasklist_lock) in the caller, so we don't * have to call rcu_read_lock/unlock() in this function. */ static struct task_struct *find_early_kill_thread(struct task_struct *tsk) |
6a46079cf HWPOISON: The hig... |
362 |
{ |
3ba08129e mm/memory-failure... |
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 |
struct task_struct *t; for_each_thread(tsk, t) if ((t->flags & PF_MCE_PROCESS) && (t->flags & PF_MCE_EARLY)) return t; return NULL; } /* * Determine whether a given process is "early kill" process which expects * to be signaled when some page under the process is hwpoisoned. * Return task_struct of the dedicated thread (main thread unless explicitly * specified) if the process is "early kill," and otherwise returns NULL. */ static struct task_struct *task_early_kill(struct task_struct *tsk, int force_early) { struct task_struct *t; |
6a46079cf HWPOISON: The hig... |
381 |
if (!tsk->mm) |
3ba08129e mm/memory-failure... |
382 |
return NULL; |
74614de17 mm/memory-failure... |
383 |
if (force_early) |
3ba08129e mm/memory-failure... |
384 385 386 387 388 389 390 |
return tsk; t = find_early_kill_thread(tsk); if (t) return t; if (sysctl_memory_failure_early_kill) return tsk; return NULL; |
6a46079cf HWPOISON: The hig... |
391 392 393 394 395 396 |
} /* * Collect processes when the error hit an anonymous page. */ static void collect_procs_anon(struct page *page, struct list_head *to_kill, |
74614de17 mm/memory-failure... |
397 |
struct to_kill **tkc, int force_early) |
6a46079cf HWPOISON: The hig... |
398 399 400 401 |
{ struct vm_area_struct *vma; struct task_struct *tsk; struct anon_vma *av; |
bf181b9f9 mm anon rmap: rep... |
402 |
pgoff_t pgoff; |
6a46079cf HWPOISON: The hig... |
403 |
|
4fc3f1d66 mm/rmap, migratio... |
404 |
av = page_lock_anon_vma_read(page); |
6a46079cf HWPOISON: The hig... |
405 |
if (av == NULL) /* Not actually mapped anymore */ |
9b679320a mm/memory-failure... |
406 |
return; |
a0f7a756c mm/rmap.c: fix pg... |
407 |
pgoff = page_to_pgoff(page); |
9b679320a mm/memory-failure... |
408 |
read_lock(&tasklist_lock); |
6a46079cf HWPOISON: The hig... |
409 |
for_each_process (tsk) { |
5beb49305 mm: change anon_v... |
410 |
struct anon_vma_chain *vmac; |
3ba08129e mm/memory-failure... |
411 |
struct task_struct *t = task_early_kill(tsk, force_early); |
5beb49305 mm: change anon_v... |
412 |
|
3ba08129e mm/memory-failure... |
413 |
if (!t) |
6a46079cf HWPOISON: The hig... |
414 |
continue; |
bf181b9f9 mm anon rmap: rep... |
415 416 |
anon_vma_interval_tree_foreach(vmac, &av->rb_root, pgoff, pgoff) { |
5beb49305 mm: change anon_v... |
417 |
vma = vmac->vma; |
6a46079cf HWPOISON: The hig... |
418 419 |
if (!page_mapped_in_vma(page, vma)) continue; |
3ba08129e mm/memory-failure... |
420 421 |
if (vma->vm_mm == t->mm) add_to_kill(t, page, vma, to_kill, tkc); |
6a46079cf HWPOISON: The hig... |
422 423 |
} } |
6a46079cf HWPOISON: The hig... |
424 |
read_unlock(&tasklist_lock); |
4fc3f1d66 mm/rmap, migratio... |
425 |
page_unlock_anon_vma_read(av); |
6a46079cf HWPOISON: The hig... |
426 427 428 429 430 431 |
} /* * Collect processes when the error hit a file mapped page. */ static void collect_procs_file(struct page *page, struct list_head *to_kill, |
74614de17 mm/memory-failure... |
432 |
struct to_kill **tkc, int force_early) |
6a46079cf HWPOISON: The hig... |
433 434 435 |
{ struct vm_area_struct *vma; struct task_struct *tsk; |
6a46079cf HWPOISON: The hig... |
436 |
struct address_space *mapping = page->mapping; |
d28eb9c86 mm/memory-failure... |
437 |
i_mmap_lock_read(mapping); |
9b679320a mm/memory-failure... |
438 |
read_lock(&tasklist_lock); |
6a46079cf HWPOISON: The hig... |
439 |
for_each_process(tsk) { |
a0f7a756c mm/rmap.c: fix pg... |
440 |
pgoff_t pgoff = page_to_pgoff(page); |
3ba08129e mm/memory-failure... |
441 |
struct task_struct *t = task_early_kill(tsk, force_early); |
6a46079cf HWPOISON: The hig... |
442 |
|
3ba08129e mm/memory-failure... |
443 |
if (!t) |
6a46079cf HWPOISON: The hig... |
444 |
continue; |
6b2dbba8b mm: replace vma p... |
445 |
vma_interval_tree_foreach(vma, &mapping->i_mmap, pgoff, |
6a46079cf HWPOISON: The hig... |
446 447 448 449 450 451 452 453 |
pgoff) { /* * Send early kill signal to tasks where a vma covers * the page but the corrupted page is not necessarily * mapped it in its pte. * Assume applications who requested early kill want * to be informed of all such data corruptions. */ |
3ba08129e mm/memory-failure... |
454 455 |
if (vma->vm_mm == t->mm) add_to_kill(t, page, vma, to_kill, tkc); |
6a46079cf HWPOISON: The hig... |
456 457 |
} } |
6a46079cf HWPOISON: The hig... |
458 |
read_unlock(&tasklist_lock); |
d28eb9c86 mm/memory-failure... |
459 |
i_mmap_unlock_read(mapping); |
6a46079cf HWPOISON: The hig... |
460 461 462 463 464 465 466 467 |
} /* * Collect the processes who have the corrupted page mapped to kill. * This is done in two steps for locking reasons. * First preallocate one tokill structure outside the spin locks, * so that we can kill at least one process reasonably reliable. */ |
74614de17 mm/memory-failure... |
468 469 |
static void collect_procs(struct page *page, struct list_head *tokill, int force_early) |
6a46079cf HWPOISON: The hig... |
470 471 472 473 474 475 476 477 478 479 |
{ struct to_kill *tk; if (!page->mapping) return; tk = kmalloc(sizeof(struct to_kill), GFP_NOIO); if (!tk) return; if (PageAnon(page)) |
74614de17 mm/memory-failure... |
480 |
collect_procs_anon(page, tokill, &tk, force_early); |
6a46079cf HWPOISON: The hig... |
481 |
else |
74614de17 mm/memory-failure... |
482 |
collect_procs_file(page, tokill, &tk, force_early); |
6a46079cf HWPOISON: The hig... |
483 484 |
kfree(tk); } |
6a46079cf HWPOISON: The hig... |
485 |
static const char *action_name[] = { |
cc637b170 memory-failure: e... |
486 487 488 489 |
[MF_IGNORED] = "Ignored", [MF_FAILED] = "Failed", [MF_DELAYED] = "Delayed", [MF_RECOVERED] = "Recovered", |
64d37a2ba mm/memory-failure... |
490 491 492 |
}; static const char * const action_page_types[] = { |
cc637b170 memory-failure: e... |
493 494 495 496 497 498 499 |
[MF_MSG_KERNEL] = "reserved kernel page", [MF_MSG_KERNEL_HIGH_ORDER] = "high-order kernel page", [MF_MSG_SLAB] = "kernel slab page", [MF_MSG_DIFFERENT_COMPOUND] = "different compound page after locking", [MF_MSG_POISONED_HUGE] = "huge page already hardware poisoned", [MF_MSG_HUGE] = "huge page", [MF_MSG_FREE_HUGE] = "free huge page", |
b16a6af97 mm: hwpoison: dis... |
500 |
[MF_MSG_NON_PMD_HUGE] = "non-pmd-sized huge page", |
cc637b170 memory-failure: e... |
501 502 503 504 505 506 507 508 509 510 511 512 513 |
[MF_MSG_UNMAP_FAILED] = "unmapping failed page", [MF_MSG_DIRTY_SWAPCACHE] = "dirty swapcache page", [MF_MSG_CLEAN_SWAPCACHE] = "clean swapcache page", [MF_MSG_DIRTY_MLOCKED_LRU] = "dirty mlocked LRU page", [MF_MSG_CLEAN_MLOCKED_LRU] = "clean mlocked LRU page", [MF_MSG_DIRTY_UNEVICTABLE_LRU] = "dirty unevictable LRU page", [MF_MSG_CLEAN_UNEVICTABLE_LRU] = "clean unevictable LRU page", [MF_MSG_DIRTY_LRU] = "dirty LRU page", [MF_MSG_CLEAN_LRU] = "clean LRU page", [MF_MSG_TRUNCATED_LRU] = "already truncated LRU page", [MF_MSG_BUDDY] = "free buddy page", [MF_MSG_BUDDY_2ND] = "free buddy page (2nd try)", [MF_MSG_UNKNOWN] = "unknown page", |
64d37a2ba mm/memory-failure... |
514 |
}; |
6a46079cf HWPOISON: The hig... |
515 |
/* |
dc2a1cbf7 HWPOISON: introdu... |
516 517 518 519 520 521 522 523 524 525 526 527 528 529 |
* XXX: It is possible that a page is isolated from LRU cache, * and then kept in swap cache or failed to remove from page cache. * The page count will stop it from being freed by unpoison. * Stress tests should be aware of this memory leak problem. */ static int delete_from_lru_cache(struct page *p) { if (!isolate_lru_page(p)) { /* * Clear sensible page flags, so that the buddy system won't * complain when the page is unpoison-and-freed. */ ClearPageActive(p); ClearPageUnevictable(p); |
18365225f hwpoison, memcg: ... |
530 531 532 533 534 535 |
/* * Poisoned page might never drop its ref count to 0 so we have * to uncharge it manually from its memcg. */ mem_cgroup_uncharge(p); |
dc2a1cbf7 HWPOISON: introdu... |
536 537 538 |
/* * drop the page count elevated by isolate_lru_page() */ |
09cbfeaf1 mm, fs: get rid o... |
539 |
put_page(p); |
dc2a1cbf7 HWPOISON: introdu... |
540 541 542 543 |
return 0; } return -EIO; } |
78bb92034 mm: hwpoison: dis... |
544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 |
static int truncate_error_page(struct page *p, unsigned long pfn, struct address_space *mapping) { int ret = MF_FAILED; if (mapping->a_ops->error_remove_page) { int err = mapping->a_ops->error_remove_page(mapping, p); if (err != 0) { pr_info("Memory failure: %#lx: Failed to punch page: %d ", pfn, err); } else if (page_has_private(p) && !try_to_release_page(p, GFP_NOIO)) { pr_info("Memory failure: %#lx: failed to release buffers ", pfn); } else { ret = MF_RECOVERED; } } else { /* * If the file system doesn't support it just invalidate * This fails on dirty or anything with private pages */ if (invalidate_inode_page(p)) ret = MF_RECOVERED; else pr_info("Memory failure: %#lx: Failed to invalidate ", pfn); } return ret; } |
dc2a1cbf7 HWPOISON: introdu... |
579 |
/* |
6a46079cf HWPOISON: The hig... |
580 581 582 583 584 585 |
* Error hit kernel page. * Do nothing, try to be lucky and not touch this instead. For a few cases we * could be more sophisticated. */ static int me_kernel(struct page *p, unsigned long pfn) { |
cc637b170 memory-failure: e... |
586 |
return MF_IGNORED; |
6a46079cf HWPOISON: The hig... |
587 588 589 590 591 592 593 |
} /* * Page in unknown state. Do nothing. */ static int me_unknown(struct page *p, unsigned long pfn) { |
495367c05 mm/memory-failure... |
594 595 |
pr_err("Memory failure: %#lx: Unknown page state ", pfn); |
cc637b170 memory-failure: e... |
596 |
return MF_FAILED; |
6a46079cf HWPOISON: The hig... |
597 598 599 |
} /* |
6a46079cf HWPOISON: The hig... |
600 601 602 603 |
* Clean (or cleaned) page cache page. */ static int me_pagecache_clean(struct page *p, unsigned long pfn) { |
6a46079cf HWPOISON: The hig... |
604 |
struct address_space *mapping; |
dc2a1cbf7 HWPOISON: introdu... |
605 |
delete_from_lru_cache(p); |
6a46079cf HWPOISON: The hig... |
606 607 608 609 610 |
/* * For anonymous pages we're done the only reference left * should be the one m_f() holds. */ if (PageAnon(p)) |
cc637b170 memory-failure: e... |
611 |
return MF_RECOVERED; |
6a46079cf HWPOISON: The hig... |
612 613 614 615 616 617 618 619 620 621 622 623 624 |
/* * Now truncate the page in the page cache. This is really * more like a "temporary hole punch" * Don't do this for block devices when someone else * has a reference, because it could be file system metadata * and that's not safe to truncate. */ mapping = page_mapping(p); if (!mapping) { /* * Page has been teared down in the meanwhile */ |
cc637b170 memory-failure: e... |
625 |
return MF_FAILED; |
6a46079cf HWPOISON: The hig... |
626 627 628 629 630 631 632 |
} /* * Truncation is a bit tricky. Enable it per file system for now. * * Open: to take i_mutex or not for this? Right now we don't. */ |
78bb92034 mm: hwpoison: dis... |
633 |
return truncate_error_page(p, pfn, mapping); |
6a46079cf HWPOISON: The hig... |
634 635 636 |
} /* |
549543dff mm, memory-failur... |
637 |
* Dirty pagecache page |
6a46079cf HWPOISON: The hig... |
638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 |
* Issues: when the error hit a hole page the error is not properly * propagated. */ static int me_pagecache_dirty(struct page *p, unsigned long pfn) { struct address_space *mapping = page_mapping(p); SetPageError(p); /* TBD: print more information about the file. */ if (mapping) { /* * IO error will be reported by write(), fsync(), etc. * who check the mapping. * This way the application knows that something went * wrong with its dirty file data. * * There's one open issue: * * The EIO will be only reported on the next IO * operation and then cleared through the IO map. * Normally Linux has two mechanisms to pass IO error * first through the AS_EIO flag in the address space * and then through the PageError flag in the page. * Since we drop pages on memory failure handling the * only mechanism open to use is through AS_AIO. * * This has the disadvantage that it gets cleared on * the first operation that returns an error, while * the PageError bit is more sticky and only cleared * when the page is reread or dropped. If an * application assumes it will always get error on * fsync, but does other operations on the fd before |
25985edce Fix common misspe... |
670 |
* and the page is dropped between then the error |
6a46079cf HWPOISON: The hig... |
671 672 673 674 675 676 677 678 679 680 681 |
* will not be properly reported. * * This can already happen even without hwpoisoned * pages: first on metadata IO errors (which only * report through AS_EIO) or when the page is dropped * at the wrong time. * * So right now we assume that the application DTRT on * the first EIO, but we're not worse than other parts * of the kernel. */ |
af21bfaf7 mm: fix mapping_s... |
682 |
mapping_set_error(mapping, -EIO); |
6a46079cf HWPOISON: The hig... |
683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 |
} return me_pagecache_clean(p, pfn); } /* * Clean and dirty swap cache. * * Dirty swap cache page is tricky to handle. The page could live both in page * cache and swap cache(ie. page is freshly swapped in). So it could be * referenced concurrently by 2 types of PTEs: * normal PTEs and swap PTEs. We try to handle them consistently by calling * try_to_unmap(TTU_IGNORE_HWPOISON) to convert the normal PTEs to swap PTEs, * and then * - clear dirty bit to prevent IO * - remove from LRU * - but keep in the swap cache, so that when we return to it on * a later page fault, we know the application is accessing * corrupted data and shall be killed (we installed simple * interception code in do_swap_page to catch it). * * Clean swap cache pages can be directly isolated. A later page fault will * bring in the known good data from disk. */ static int me_swapcache_dirty(struct page *p, unsigned long pfn) { |
6a46079cf HWPOISON: The hig... |
709 710 711 |
ClearPageDirty(p); /* Trigger EIO in shmem: */ ClearPageUptodate(p); |
dc2a1cbf7 HWPOISON: introdu... |
712 |
if (!delete_from_lru_cache(p)) |
cc637b170 memory-failure: e... |
713 |
return MF_DELAYED; |
dc2a1cbf7 HWPOISON: introdu... |
714 |
else |
cc637b170 memory-failure: e... |
715 |
return MF_FAILED; |
6a46079cf HWPOISON: The hig... |
716 717 718 719 |
} static int me_swapcache_clean(struct page *p, unsigned long pfn) { |
6a46079cf HWPOISON: The hig... |
720 |
delete_from_swap_cache(p); |
e43c3afb3 HWPOISON: return ... |
721 |
|
dc2a1cbf7 HWPOISON: introdu... |
722 |
if (!delete_from_lru_cache(p)) |
cc637b170 memory-failure: e... |
723 |
return MF_RECOVERED; |
dc2a1cbf7 HWPOISON: introdu... |
724 |
else |
cc637b170 memory-failure: e... |
725 |
return MF_FAILED; |
6a46079cf HWPOISON: The hig... |
726 727 728 729 730 |
} /* * Huge pages. Needs work. * Issues: |
93f70f900 HWPOISON, hugetlb... |
731 732 |
* - Error on hugepage is contained in hugepage unit (not in raw page unit.) * To narrow down kill region to one page, we need to break up pmd. |
6a46079cf HWPOISON: The hig... |
733 734 735 |
*/ static int me_huge_page(struct page *p, unsigned long pfn) { |
6de2b1aab HWPOISON, hugetlb... |
736 |
int res = 0; |
93f70f900 HWPOISON, hugetlb... |
737 |
struct page *hpage = compound_head(p); |
78bb92034 mm: hwpoison: dis... |
738 |
struct address_space *mapping; |
2491ffee9 mm/memory-failure... |
739 740 741 |
if (!PageHuge(hpage)) return MF_DELAYED; |
78bb92034 mm: hwpoison: dis... |
742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 |
mapping = page_mapping(hpage); if (mapping) { res = truncate_error_page(hpage, pfn, mapping); } else { unlock_page(hpage); /* * migration entry prevents later access on error anonymous * hugepage, so we can free and dissolve it into buddy to * save healthy subpages. */ if (PageAnon(hpage)) put_page(hpage); dissolve_free_huge_page(p); res = MF_RECOVERED; lock_page(hpage); |
93f70f900 HWPOISON, hugetlb... |
757 |
} |
78bb92034 mm: hwpoison: dis... |
758 759 |
return res; |
6a46079cf HWPOISON: The hig... |
760 761 762 763 764 765 766 767 768 |
} /* * Various page states we can handle. * * A page state is defined by its current page->flags bits. * The table matches them in order and calls the right handler. * * This is quite tricky because we can access page at any time |
25985edce Fix common misspe... |
769 |
* in its live cycle, so all accesses have to be extremely careful. |
6a46079cf HWPOISON: The hig... |
770 771 772 773 774 775 |
* * This is not complete. More states could be added. * For any missing state don't attempt recovery. */ #define dirty (1UL << PG_dirty) |
6326fec11 mm: Use owner_pri... |
776 |
#define sc ((1UL << PG_swapcache) | (1UL << PG_swapbacked)) |
6a46079cf HWPOISON: The hig... |
777 778 779 780 |
#define unevict (1UL << PG_unevictable) #define mlock (1UL << PG_mlocked) #define writeback (1UL << PG_writeback) #define lru (1UL << PG_lru) |
6a46079cf HWPOISON: The hig... |
781 |
#define head (1UL << PG_head) |
6a46079cf HWPOISON: The hig... |
782 |
#define slab (1UL << PG_slab) |
6a46079cf HWPOISON: The hig... |
783 784 785 786 787 |
#define reserved (1UL << PG_reserved) static struct page_state { unsigned long mask; unsigned long res; |
cc637b170 memory-failure: e... |
788 |
enum mf_action_page_type type; |
6a46079cf HWPOISON: The hig... |
789 790 |
int (*action)(struct page *p, unsigned long pfn); } error_states[] = { |
cc637b170 memory-failure: e... |
791 |
{ reserved, reserved, MF_MSG_KERNEL, me_kernel }, |
95d01fc66 HWPOISON: remove ... |
792 793 794 795 |
/* * free pages are specially detected outside this table: * PG_buddy pages only make a small fraction of all free pages. */ |
6a46079cf HWPOISON: The hig... |
796 797 798 799 800 801 |
/* * Could in theory check if slab page is free or if we can drop * currently unused objects without touching them. But just * treat it as standard kernel for now. */ |
cc637b170 memory-failure: e... |
802 |
{ slab, slab, MF_MSG_SLAB, me_kernel }, |
6a46079cf HWPOISON: The hig... |
803 |
|
cc637b170 memory-failure: e... |
804 |
{ head, head, MF_MSG_HUGE, me_huge_page }, |
6a46079cf HWPOISON: The hig... |
805 |
|
cc637b170 memory-failure: e... |
806 807 |
{ sc|dirty, sc|dirty, MF_MSG_DIRTY_SWAPCACHE, me_swapcache_dirty }, { sc|dirty, sc, MF_MSG_CLEAN_SWAPCACHE, me_swapcache_clean }, |
6a46079cf HWPOISON: The hig... |
808 |
|
cc637b170 memory-failure: e... |
809 810 |
{ mlock|dirty, mlock|dirty, MF_MSG_DIRTY_MLOCKED_LRU, me_pagecache_dirty }, { mlock|dirty, mlock, MF_MSG_CLEAN_MLOCKED_LRU, me_pagecache_clean }, |
6a46079cf HWPOISON: The hig... |
811 |
|
cc637b170 memory-failure: e... |
812 813 |
{ unevict|dirty, unevict|dirty, MF_MSG_DIRTY_UNEVICTABLE_LRU, me_pagecache_dirty }, { unevict|dirty, unevict, MF_MSG_CLEAN_UNEVICTABLE_LRU, me_pagecache_clean }, |
5f4b9fc5c HWPOISON: change ... |
814 |
|
cc637b170 memory-failure: e... |
815 816 |
{ lru|dirty, lru|dirty, MF_MSG_DIRTY_LRU, me_pagecache_dirty }, { lru|dirty, lru, MF_MSG_CLEAN_LRU, me_pagecache_clean }, |
6a46079cf HWPOISON: The hig... |
817 818 819 820 |
/* * Catchall entry: must be at end. */ |
cc637b170 memory-failure: e... |
821 |
{ 0, 0, MF_MSG_UNKNOWN, me_unknown }, |
6a46079cf HWPOISON: The hig... |
822 |
}; |
2326c467d HWPOISON: Undefin... |
823 824 825 826 827 828 |
#undef dirty #undef sc #undef unevict #undef mlock #undef writeback #undef lru |
2326c467d HWPOISON: Undefin... |
829 |
#undef head |
2326c467d HWPOISON: Undefin... |
830 831 |
#undef slab #undef reserved |
ff604cf6d mm: hwpoison: fix... |
832 833 834 835 |
/* * "Dirty/Clean" indication is not 100% accurate due to the possibility of * setting PG_dirty outside page lock. See also comment above set_page_dirty(). */ |
cc3e2af42 memory-failure: c... |
836 837 |
static void action_result(unsigned long pfn, enum mf_action_page_type type, enum mf_result result) |
6a46079cf HWPOISON: The hig... |
838 |
{ |
97f0b1345 tracing: add trac... |
839 |
trace_memory_failure_event(pfn, type, result); |
495367c05 mm/memory-failure... |
840 841 |
pr_err("Memory failure: %#lx: recovery action for %s: %s ", |
64d37a2ba mm/memory-failure... |
842 |
pfn, action_page_types[type], action_name[result]); |
6a46079cf HWPOISON: The hig... |
843 844 845 |
} static int page_action(struct page_state *ps, struct page *p, |
bd1ce5f91 HWPOISON: avoid g... |
846 |
unsigned long pfn) |
6a46079cf HWPOISON: The hig... |
847 848 |
{ int result; |
7456b0405 HWPOISON: fix inv... |
849 |
int count; |
6a46079cf HWPOISON: The hig... |
850 851 |
result = ps->action(p, pfn); |
7456b0405 HWPOISON: fix inv... |
852 |
|
bd1ce5f91 HWPOISON: avoid g... |
853 |
count = page_count(p) - 1; |
cc637b170 memory-failure: e... |
854 |
if (ps->action == me_swapcache_dirty && result == MF_DELAYED) |
138ce286e HWPOISON: return ... |
855 |
count--; |
78bb92034 mm: hwpoison: dis... |
856 |
if (count > 0) { |
495367c05 mm/memory-failure... |
857 858 |
pr_err("Memory failure: %#lx: %s still referenced by %d users ", |
64d37a2ba mm/memory-failure... |
859 |
pfn, action_page_types[ps->type], count); |
cc637b170 memory-failure: e... |
860 |
result = MF_FAILED; |
138ce286e HWPOISON: return ... |
861 |
} |
64d37a2ba mm/memory-failure... |
862 |
action_result(pfn, ps->type, result); |
6a46079cf HWPOISON: The hig... |
863 864 865 866 867 |
/* Could do more checks here if page looks ok */ /* * Could adjust zone counters here to correct for the missing page. */ |
cc637b170 memory-failure: e... |
868 |
return (result == MF_RECOVERED || result == MF_DELAYED) ? 0 : -EBUSY; |
6a46079cf HWPOISON: The hig... |
869 |
} |
ead07f6a8 mm/memory-failure... |
870 871 872 873 874 875 876 877 878 879 |
/** * get_hwpoison_page() - Get refcount for memory error handling: * @page: raw error page (hit by memory error) * * Return: return 0 if failed to grab the refcount, otherwise true (some * non-zero value.) */ int get_hwpoison_page(struct page *page) { struct page *head = compound_head(page); |
4e41a30c6 mm: hwpoison: adj... |
880 |
if (!PageHuge(head) && PageTransHuge(head)) { |
98ed2b005 mm/memory-failure... |
881 882 883 884 885 886 887 |
/* * Non anonymous thp exists only in allocation/free time. We * can't handle such a case correctly, so let's give it up. * This should be better than triggering BUG_ON when kernel * tries to touch the "partially handled" page. */ if (!PageAnon(head)) { |
495367c05 mm/memory-failure... |
888 889 |
pr_err("Memory failure: %#lx: non anonymous thp ", |
98ed2b005 mm/memory-failure... |
890 891 892 |
page_to_pfn(page)); return 0; } |
ead07f6a8 mm/memory-failure... |
893 |
} |
c2e7e00b7 mm/memory-failure... |
894 895 896 |
if (get_page_unless_zero(head)) { if (head == compound_head(page)) return 1; |
495367c05 mm/memory-failure... |
897 898 899 |
pr_info("Memory failure: %#lx cannot catch tail ", page_to_pfn(page)); |
c2e7e00b7 mm/memory-failure... |
900 901 902 903 |
put_page(head); } return 0; |
ead07f6a8 mm/memory-failure... |
904 905 |
} EXPORT_SYMBOL_GPL(get_hwpoison_page); |
6a46079cf HWPOISON: The hig... |
906 907 908 909 |
/* * Do all that is necessary to remove user space mappings. Unmap * the pages and send SIGBUS to the processes if the data was dirty. */ |
666e5a406 mm: make ttu's re... |
910 |
static bool hwpoison_user_mappings(struct page *p, unsigned long pfn, |
54b9dd14d mm/memory-failure... |
911 |
int trapno, int flags, struct page **hpagep) |
6a46079cf HWPOISON: The hig... |
912 |
{ |
a128ca71f mm: delete unnece... |
913 |
enum ttu_flags ttu = TTU_IGNORE_MLOCK | TTU_IGNORE_ACCESS; |
6a46079cf HWPOISON: The hig... |
914 915 |
struct address_space *mapping; LIST_HEAD(tokill); |
666e5a406 mm: make ttu's re... |
916 |
bool unmap_success; |
6751ed65d x86/mce: Fix sigi... |
917 |
int kill = 1, forcekill; |
54b9dd14d mm/memory-failure... |
918 |
struct page *hpage = *hpagep; |
286c469a9 mm: hwpoison: cal... |
919 |
bool mlocked = PageMlocked(hpage); |
6a46079cf HWPOISON: The hig... |
920 |
|
93a9eb39f hwpoison: fix hug... |
921 922 923 924 925 |
/* * Here we are interested only in user-mapped pages, so skip any * other types of pages. */ if (PageReserved(p) || PageSlab(p)) |
666e5a406 mm: make ttu's re... |
926 |
return true; |
93a9eb39f hwpoison: fix hug... |
927 |
if (!(PageLRU(hpage) || PageHuge(p))) |
666e5a406 mm: make ttu's re... |
928 |
return true; |
6a46079cf HWPOISON: The hig... |
929 |
|
6a46079cf HWPOISON: The hig... |
930 931 932 933 |
/* * This check implies we don't kill processes if their pages * are in the swap cache early. Those are always late kills. */ |
7af446a84 HWPOISON, hugetlb... |
934 |
if (!page_mapped(hpage)) |
666e5a406 mm: make ttu's re... |
935 |
return true; |
1668bfd5b HWPOISON: abort o... |
936 |
|
52089b14c hwpoison: call ac... |
937 |
if (PageKsm(p)) { |
495367c05 mm/memory-failure... |
938 939 |
pr_err("Memory failure: %#lx: can't handle KSM pages. ", pfn); |
666e5a406 mm: make ttu's re... |
940 |
return false; |
52089b14c hwpoison: call ac... |
941 |
} |
6a46079cf HWPOISON: The hig... |
942 943 |
if (PageSwapCache(p)) { |
495367c05 mm/memory-failure... |
944 945 946 |
pr_err("Memory failure: %#lx: keeping poisoned page in swap cache ", pfn); |
6a46079cf HWPOISON: The hig... |
947 948 949 950 951 952 |
ttu |= TTU_IGNORE_HWPOISON; } /* * Propagate the dirty bit from PTEs to struct page first, because we * need this to decide if we should kill or just drop the page. |
db0480b3a HWPOISON: comment... |
953 954 |
* XXX: the dirty test could be racy: set_page_dirty() may not always * be called inside page lock (it's recommended but not enforced). |
6a46079cf HWPOISON: The hig... |
955 |
*/ |
7af446a84 HWPOISON, hugetlb... |
956 |
mapping = page_mapping(hpage); |
6751ed65d x86/mce: Fix sigi... |
957 |
if (!(flags & MF_MUST_KILL) && !PageDirty(hpage) && mapping && |
7af446a84 HWPOISON, hugetlb... |
958 959 960 |
mapping_cap_writeback_dirty(mapping)) { if (page_mkclean(hpage)) { SetPageDirty(hpage); |
6a46079cf HWPOISON: The hig... |
961 962 963 |
} else { kill = 0; ttu |= TTU_IGNORE_HWPOISON; |
495367c05 mm/memory-failure... |
964 965 |
pr_info("Memory failure: %#lx: corrupted page was clean: dropped without side effects ", |
6a46079cf HWPOISON: The hig... |
966 967 968 |
pfn); } } |
a6d30ddda thp: fix the wron... |
969 |
/* |
6a46079cf HWPOISON: The hig... |
970 971 972 973 974 975 976 977 |
* First collect all the processes that have the page * mapped in dirty form. This has to be done before try_to_unmap, * because ttu takes the rmap data structures down. * * Error handling: We ignore errors here because * there's nothing that can be done. */ if (kill) |
415c64c14 mm/memory-failure... |
978 |
collect_procs(hpage, &tokill, flags & MF_ACTION_REQUIRED); |
6a46079cf HWPOISON: The hig... |
979 |
|
666e5a406 mm: make ttu's re... |
980 981 |
unmap_success = try_to_unmap(hpage, ttu); if (!unmap_success) |
495367c05 mm/memory-failure... |
982 983 |
pr_err("Memory failure: %#lx: failed to unmap page (mapcount=%d) ", |
1170532bb mm: convert print... |
984 |
pfn, page_mapcount(hpage)); |
a6d30ddda thp: fix the wron... |
985 |
|
6a46079cf HWPOISON: The hig... |
986 |
/* |
286c469a9 mm: hwpoison: cal... |
987 988 989 990 991 992 993 |
* try_to_unmap() might put mlocked page in lru cache, so call * shake_page() again to ensure that it's flushed. */ if (mlocked) shake_page(hpage, 0); /* |
6a46079cf HWPOISON: The hig... |
994 995 996 |
* Now that the dirty bit has been propagated to the * struct page and all unmaps done we can decide if * killing is needed or not. Only kill when the page |
6751ed65d x86/mce: Fix sigi... |
997 998 |
* was dirty or the process is not restartable, * otherwise the tokill list is merely |
6a46079cf HWPOISON: The hig... |
999 1000 1001 1002 |
* freed. When there was a problem unmapping earlier * use a more force-full uncatchable kill to prevent * any accesses to the poisoned memory. */ |
415c64c14 mm/memory-failure... |
1003 |
forcekill = PageDirty(hpage) || (flags & MF_MUST_KILL); |
666e5a406 mm: make ttu's re... |
1004 |
kill_procs(&tokill, forcekill, trapno, !unmap_success, p, pfn, flags); |
1668bfd5b HWPOISON: abort o... |
1005 |
|
666e5a406 mm: make ttu's re... |
1006 |
return unmap_success; |
6a46079cf HWPOISON: The hig... |
1007 |
} |
0348d2ebe mm: hwpoison: int... |
1008 1009 |
static int identify_page_state(unsigned long pfn, struct page *p, unsigned long page_flags) |
761ad8d7c mm: hwpoison: int... |
1010 1011 |
{ struct page_state *ps; |
0348d2ebe mm: hwpoison: int... |
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 |
/* * The first check uses the current page flags which may not have any * relevant information. The second check with the saved page flags is * carried out only if the first check can't determine the page status. */ for (ps = error_states;; ps++) if ((p->flags & ps->mask) == ps->res) break; page_flags |= (p->flags & (1UL << PG_dirty)); if (!ps->mask) for (ps = error_states;; ps++) if ((page_flags & ps->mask) == ps->res) break; return page_action(ps, p, pfn); } static int memory_failure_hugetlb(unsigned long pfn, int trapno, int flags) { |
761ad8d7c mm: hwpoison: int... |
1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 |
struct page *p = pfn_to_page(pfn); struct page *head = compound_head(p); int res; unsigned long page_flags; if (TestSetPageHWPoison(head)) { pr_err("Memory failure: %#lx: already hardware poisoned ", pfn); return 0; } num_poisoned_pages_inc(); if (!(flags & MF_COUNT_INCREASED) && !get_hwpoison_page(p)) { /* * Check "filter hit" and "race with other subpage." */ lock_page(head); if (PageHWPoison(head)) { if ((hwpoison_filter(p) && TestClearPageHWPoison(p)) || (p != head && TestSetPageHWPoison(head))) { num_poisoned_pages_dec(); unlock_page(head); return 0; } } unlock_page(head); dissolve_free_huge_page(p); action_result(pfn, MF_MSG_FREE_HUGE, MF_DELAYED); return 0; } lock_page(head); page_flags = head->flags; if (!PageHWPoison(head)) { pr_err("Memory failure: %#lx: just unpoisoned ", pfn); num_poisoned_pages_dec(); unlock_page(head); put_hwpoison_page(head); return 0; } |
b16a6af97 mm: hwpoison: dis... |
1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 |
/* * TODO: hwpoison for pud-sized hugetlb doesn't work right now, so * simply disable it. In order to make it work properly, we need * make sure that: * - conversion of a pud that maps an error hugetlb into hwpoison * entry properly works, and * - other mm code walking over page table is aware of pud-aligned * hwpoison entries. */ if (huge_page_size(page_hstate(head)) > PMD_SIZE) { action_result(pfn, MF_MSG_NON_PMD_HUGE, MF_IGNORED); res = -EBUSY; goto out; } |
761ad8d7c mm: hwpoison: int... |
1091 1092 1093 1094 1095 |
if (!hwpoison_user_mappings(p, pfn, trapno, flags, &head)) { action_result(pfn, MF_MSG_UNMAP_FAILED, MF_IGNORED); res = -EBUSY; goto out; } |
0348d2ebe mm: hwpoison: int... |
1096 |
res = identify_page_state(pfn, p, page_flags); |
761ad8d7c mm: hwpoison: int... |
1097 1098 1099 1100 |
out: unlock_page(head); return res; } |
cd42f4a3b HWPOISON: Clean u... |
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 |
/** * memory_failure - Handle memory failure of a page. * @pfn: Page Number of the corrupted page * @trapno: Trap number reported in the signal to user space. * @flags: fine tune action taken * * This function is called by the low level machine check code * of an architecture when it detects hardware memory corruption * of a page. It tries its best to recover, which includes * dropping pages, killing processes etc. * * The function is primarily of use for corruptions that * happen outside the current execution context (e.g. when * detected by a background scrubber) * * Must run in process context (e.g. a work queue) with interrupts * enabled and no spinlocks hold. */ int memory_failure(unsigned long pfn, int trapno, int flags) |
6a46079cf HWPOISON: The hig... |
1120 |
{ |
6a46079cf HWPOISON: The hig... |
1121 |
struct page *p; |
7af446a84 HWPOISON, hugetlb... |
1122 |
struct page *hpage; |
415c64c14 mm/memory-failure... |
1123 |
struct page *orig_head; |
6a46079cf HWPOISON: The hig... |
1124 |
int res; |
524fca1e7 HWPOISON: fix mis... |
1125 |
unsigned long page_flags; |
6a46079cf HWPOISON: The hig... |
1126 1127 1128 1129 1130 |
if (!sysctl_memory_failure_recovery) panic("Memory failure from trap %d on page %lx", trapno, pfn); if (!pfn_valid(pfn)) { |
495367c05 mm/memory-failure... |
1131 1132 1133 |
pr_err("Memory failure: %#lx: memory outside kernel control ", pfn); |
a7560fc80 HWPOISON: return ... |
1134 |
return -ENXIO; |
6a46079cf HWPOISON: The hig... |
1135 1136 1137 |
} p = pfn_to_page(pfn); |
761ad8d7c mm: hwpoison: int... |
1138 1139 |
if (PageHuge(p)) return memory_failure_hugetlb(pfn, trapno, flags); |
6a46079cf HWPOISON: The hig... |
1140 |
if (TestSetPageHWPoison(p)) { |
495367c05 mm/memory-failure... |
1141 1142 1143 |
pr_err("Memory failure: %#lx: already hardware poisoned ", pfn); |
6a46079cf HWPOISON: The hig... |
1144 1145 |
return 0; } |
761ad8d7c mm: hwpoison: int... |
1146 |
orig_head = hpage = compound_head(p); |
b37ff71cc mm: hwpoison: cha... |
1147 |
num_poisoned_pages_inc(); |
6a46079cf HWPOISON: The hig... |
1148 1149 1150 1151 1152 |
/* * We need/can do nothing about count=0 pages. * 1) it's a free page, and therefore in safe hand: * prep_new_page() will be the gate keeper. |
761ad8d7c mm: hwpoison: int... |
1153 |
* 2) it's part of a non-compound high order page. |
6a46079cf HWPOISON: The hig... |
1154 1155 1156 1157 1158 1159 |
* Implies some kernel user: cannot stop them from * R/W the page; let's pray that the page has been * used and will be freed some time later. * In fact it's dangerous to directly bump up page count from 0, * that may make page_freeze_refs()/page_unfreeze_refs() mismatch. */ |
ead07f6a8 mm/memory-failure... |
1160 |
if (!(flags & MF_COUNT_INCREASED) && !get_hwpoison_page(p)) { |
8d22ba1b7 HWPOISON: detect ... |
1161 |
if (is_free_buddy_page(p)) { |
cc637b170 memory-failure: e... |
1162 |
action_result(pfn, MF_MSG_BUDDY, MF_DELAYED); |
8d22ba1b7 HWPOISON: detect ... |
1163 1164 |
return 0; } else { |
cc637b170 memory-failure: e... |
1165 |
action_result(pfn, MF_MSG_KERNEL_HIGH_ORDER, MF_IGNORED); |
8d22ba1b7 HWPOISON: detect ... |
1166 1167 |
return -EBUSY; } |
6a46079cf HWPOISON: The hig... |
1168 |
} |
761ad8d7c mm: hwpoison: int... |
1169 |
if (PageTransHuge(hpage)) { |
c3901e722 mm: hwpoison: fix... |
1170 1171 1172 1173 |
lock_page(p); if (!PageAnon(p) || unlikely(split_huge_page(p))) { unlock_page(p); if (!PageAnon(p)) |
495367c05 mm/memory-failure... |
1174 1175 1176 |
pr_err("Memory failure: %#lx: non anonymous thp ", pfn); |
7f6bf39bb mm/hwpoison: fix ... |
1177 |
else |
495367c05 mm/memory-failure... |
1178 1179 1180 |
pr_err("Memory failure: %#lx: thp split failed ", pfn); |
ead07f6a8 mm/memory-failure... |
1181 |
if (TestClearPageHWPoison(p)) |
b37ff71cc mm: hwpoison: cha... |
1182 |
num_poisoned_pages_dec(); |
665d9da7f mm/hwpoison: repl... |
1183 |
put_hwpoison_page(p); |
415c64c14 mm/memory-failure... |
1184 1185 |
return -EBUSY; } |
c3901e722 mm: hwpoison: fix... |
1186 |
unlock_page(p); |
415c64c14 mm/memory-failure... |
1187 1188 1189 |
VM_BUG_ON_PAGE(!page_count(p), p); hpage = compound_head(p); } |
6a46079cf HWPOISON: The hig... |
1190 |
/* |
e43c3afb3 HWPOISON: return ... |
1191 1192 |
* We ignore non-LRU pages for good reasons. * - PG_locked is only well defined for LRU pages and a few others |
48c935ad8 page-flags: defin... |
1193 |
* - to avoid races with __SetPageLocked() |
e43c3afb3 HWPOISON: return ... |
1194 1195 1196 1197 |
* - to avoid races with __SetPageSlab*() (and more non-atomic ops) * The check (unnecessarily) ignores LRU pages being isolated and * walked by the page reclaim code, however that's not a big loss. */ |
8bcb74de7 mm: hwpoison: cal... |
1198 1199 1200 1201 1202 1203 1204 1205 |
shake_page(p, 0); /* shake_page could have turned it free. */ if (!PageLRU(p) && is_free_buddy_page(p)) { if (flags & MF_COUNT_INCREASED) action_result(pfn, MF_MSG_BUDDY, MF_DELAYED); else action_result(pfn, MF_MSG_BUDDY_2ND, MF_DELAYED); return 0; |
e43c3afb3 HWPOISON: return ... |
1206 |
} |
e43c3afb3 HWPOISON: return ... |
1207 |
|
761ad8d7c mm: hwpoison: int... |
1208 |
lock_page(p); |
847ce401d HWPOISON: Add unp... |
1209 1210 |
/* |
f37d4298a hwpoison: fix rac... |
1211 1212 1213 |
* The page could have changed compound pages during the locking. * If this happens just bail out. */ |
415c64c14 mm/memory-failure... |
1214 |
if (PageCompound(p) && compound_head(p) != orig_head) { |
cc637b170 memory-failure: e... |
1215 |
action_result(pfn, MF_MSG_DIFFERENT_COMPOUND, MF_IGNORED); |
f37d4298a hwpoison: fix rac... |
1216 1217 1218 1219 1220 |
res = -EBUSY; goto out; } /* |
524fca1e7 HWPOISON: fix mis... |
1221 1222 1223 1224 1225 1226 |
* We use page flags to determine what action should be taken, but * the flags can be modified by the error containment action. One * example is an mlocked page, where PG_mlocked is cleared by * page_remove_rmap() in try_to_unmap_one(). So to determine page status * correctly, we save a copy of the page flags at this time. */ |
7258ae5c5 mm/memory-failure... |
1227 1228 1229 1230 |
if (PageHuge(p)) page_flags = hpage->flags; else page_flags = p->flags; |
524fca1e7 HWPOISON: fix mis... |
1231 1232 |
/* |
847ce401d HWPOISON: Add unp... |
1233 1234 1235 |
* unpoison always clear PG_hwpoison inside page lock */ if (!PageHWPoison(p)) { |
495367c05 mm/memory-failure... |
1236 1237 |
pr_err("Memory failure: %#lx: just unpoisoned ", pfn); |
b37ff71cc mm: hwpoison: cha... |
1238 |
num_poisoned_pages_dec(); |
761ad8d7c mm: hwpoison: int... |
1239 1240 |
unlock_page(p); put_hwpoison_page(p); |
a09233f3e mm/memory-failure... |
1241 |
return 0; |
847ce401d HWPOISON: Add unp... |
1242 |
} |
7c116f2b0 HWPOISON: add fs/... |
1243 1244 |
if (hwpoison_filter(p)) { if (TestClearPageHWPoison(p)) |
b37ff71cc mm: hwpoison: cha... |
1245 |
num_poisoned_pages_dec(); |
761ad8d7c mm: hwpoison: int... |
1246 1247 |
unlock_page(p); put_hwpoison_page(p); |
7c116f2b0 HWPOISON: add fs/... |
1248 1249 |
return 0; } |
847ce401d HWPOISON: Add unp... |
1250 |
|
761ad8d7c mm: hwpoison: int... |
1251 |
if (!PageTransTail(p) && !PageLRU(p)) |
0bc1f8b06 hwpoison: fix the... |
1252 |
goto identify_page_state; |
7013febc8 HWPOISON, hugetlb... |
1253 |
/* |
6edd6cc66 mm/memory-failure... |
1254 1255 1256 |
* It's very difficult to mess with pages currently under IO * and in many cases impossible, so we just avoid it here. */ |
6a46079cf HWPOISON: The hig... |
1257 1258 1259 1260 |
wait_on_page_writeback(p); /* * Now take care of user space mappings. |
e64a782fe mm: change __remo... |
1261 |
* Abort on fail: __delete_from_page_cache() assumes unmapped page. |
54b9dd14d mm/memory-failure... |
1262 1263 1264 |
* * When the raw error page is thp tail page, hpage points to the raw * page after thp split. |
6a46079cf HWPOISON: The hig... |
1265 |
*/ |
666e5a406 mm: make ttu's re... |
1266 |
if (!hwpoison_user_mappings(p, pfn, trapno, flags, &hpage)) { |
cc637b170 memory-failure: e... |
1267 |
action_result(pfn, MF_MSG_UNMAP_FAILED, MF_IGNORED); |
1668bfd5b HWPOISON: abort o... |
1268 1269 1270 |
res = -EBUSY; goto out; } |
6a46079cf HWPOISON: The hig... |
1271 1272 1273 1274 |
/* * Torn down by someone else? */ |
dc2a1cbf7 HWPOISON: introdu... |
1275 |
if (PageLRU(p) && !PageSwapCache(p) && p->mapping == NULL) { |
cc637b170 memory-failure: e... |
1276 |
action_result(pfn, MF_MSG_TRUNCATED_LRU, MF_IGNORED); |
d95ea51e3 HWPOISON: make se... |
1277 |
res = -EBUSY; |
6a46079cf HWPOISON: The hig... |
1278 1279 |
goto out; } |
0bc1f8b06 hwpoison: fix the... |
1280 |
identify_page_state: |
0348d2ebe mm: hwpoison: int... |
1281 |
res = identify_page_state(pfn, p, page_flags); |
6a46079cf HWPOISON: The hig... |
1282 |
out: |
761ad8d7c mm: hwpoison: int... |
1283 |
unlock_page(p); |
6a46079cf HWPOISON: The hig... |
1284 1285 |
return res; } |
cd42f4a3b HWPOISON: Clean u... |
1286 |
EXPORT_SYMBOL_GPL(memory_failure); |
847ce401d HWPOISON: Add unp... |
1287 |
|
ea8f5fb8a HWPoison: add mem... |
1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 |
#define MEMORY_FAILURE_FIFO_ORDER 4 #define MEMORY_FAILURE_FIFO_SIZE (1 << MEMORY_FAILURE_FIFO_ORDER) struct memory_failure_entry { unsigned long pfn; int trapno; int flags; }; struct memory_failure_cpu { DECLARE_KFIFO(fifo, struct memory_failure_entry, MEMORY_FAILURE_FIFO_SIZE); spinlock_t lock; struct work_struct work; }; static DEFINE_PER_CPU(struct memory_failure_cpu, memory_failure_cpu); /** * memory_failure_queue - Schedule handling memory failure of a page. * @pfn: Page Number of the corrupted page * @trapno: Trap number reported in the signal to user space. * @flags: Flags for memory failure handling * * This function is called by the low level hardware error handler * when it detects hardware memory corruption of a page. It schedules * the recovering of error page, including dropping pages, killing * processes etc. * * The function is primarily of use for corruptions that * happen outside the current execution context (e.g. when * detected by a background scrubber) * * Can run in IRQ context. */ void memory_failure_queue(unsigned long pfn, int trapno, int flags) { struct memory_failure_cpu *mf_cpu; unsigned long proc_flags; struct memory_failure_entry entry = { .pfn = pfn, .trapno = trapno, .flags = flags, }; mf_cpu = &get_cpu_var(memory_failure_cpu); spin_lock_irqsave(&mf_cpu->lock, proc_flags); |
498d319bb kfifo API type sa... |
1335 |
if (kfifo_put(&mf_cpu->fifo, entry)) |
ea8f5fb8a HWPoison: add mem... |
1336 1337 |
schedule_work_on(smp_processor_id(), &mf_cpu->work); else |
8e33a52fa treewide: Fix pri... |
1338 1339 |
pr_err("Memory failure: buffer overflow when queuing memory failure at %#lx ", |
ea8f5fb8a HWPoison: add mem... |
1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 |
pfn); spin_unlock_irqrestore(&mf_cpu->lock, proc_flags); put_cpu_var(memory_failure_cpu); } EXPORT_SYMBOL_GPL(memory_failure_queue); static void memory_failure_work_func(struct work_struct *work) { struct memory_failure_cpu *mf_cpu; struct memory_failure_entry entry = { 0, }; unsigned long proc_flags; int gotten; |
7c8e0181e mm: replace __get... |
1352 |
mf_cpu = this_cpu_ptr(&memory_failure_cpu); |
ea8f5fb8a HWPoison: add mem... |
1353 1354 1355 1356 1357 1358 |
for (;;) { spin_lock_irqsave(&mf_cpu->lock, proc_flags); gotten = kfifo_get(&mf_cpu->fifo, &entry); spin_unlock_irqrestore(&mf_cpu->lock, proc_flags); if (!gotten) break; |
cf870c70a mce: acpi/apei: S... |
1359 1360 1361 1362 |
if (entry.flags & MF_SOFT_OFFLINE) soft_offline_page(pfn_to_page(entry.pfn), entry.flags); else memory_failure(entry.pfn, entry.trapno, entry.flags); |
ea8f5fb8a HWPoison: add mem... |
1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 |
} } static int __init memory_failure_init(void) { struct memory_failure_cpu *mf_cpu; int cpu; for_each_possible_cpu(cpu) { mf_cpu = &per_cpu(memory_failure_cpu, cpu); spin_lock_init(&mf_cpu->lock); INIT_KFIFO(mf_cpu->fifo); INIT_WORK(&mf_cpu->work, memory_failure_work_func); } return 0; } core_initcall(memory_failure_init); |
a5f651090 mm: hwpoison: rat... |
1381 1382 1383 1384 1385 |
#define unpoison_pr_info(fmt, pfn, rs) \ ({ \ if (__ratelimit(rs)) \ pr_info(fmt, pfn); \ }) |
847ce401d HWPOISON: Add unp... |
1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 |
/** * unpoison_memory - Unpoison a previously poisoned page * @pfn: Page number of the to be unpoisoned page * * Software-unpoison a page that has been poisoned by * memory_failure() earlier. * * This is only done on the software-level, so it only works * for linux injected failures, not real hardware failures * * Returns 0 for success, otherwise -errno. */ int unpoison_memory(unsigned long pfn) { struct page *page; struct page *p; int freeit = 0; |
a5f651090 mm: hwpoison: rat... |
1403 1404 |
static DEFINE_RATELIMIT_STATE(unpoison_rs, DEFAULT_RATELIMIT_INTERVAL, DEFAULT_RATELIMIT_BURST); |
847ce401d HWPOISON: Add unp... |
1405 1406 1407 1408 1409 1410 1411 1412 |
if (!pfn_valid(pfn)) return -ENXIO; p = pfn_to_page(pfn); page = compound_head(p); if (!PageHWPoison(p)) { |
495367c05 mm/memory-failure... |
1413 1414 |
unpoison_pr_info("Unpoison: Page was already unpoisoned %#lx ", |
a5f651090 mm: hwpoison: rat... |
1415 |
pfn, &unpoison_rs); |
847ce401d HWPOISON: Add unp... |
1416 1417 |
return 0; } |
230ac719c mm/hwpoison: don'... |
1418 |
if (page_count(page) > 1) { |
495367c05 mm/memory-failure... |
1419 1420 |
unpoison_pr_info("Unpoison: Someone grabs the hwpoison page %#lx ", |
a5f651090 mm: hwpoison: rat... |
1421 |
pfn, &unpoison_rs); |
230ac719c mm/hwpoison: don'... |
1422 1423 1424 1425 |
return 0; } if (page_mapped(page)) { |
495367c05 mm/memory-failure... |
1426 1427 |
unpoison_pr_info("Unpoison: Someone maps the hwpoison page %#lx ", |
a5f651090 mm: hwpoison: rat... |
1428 |
pfn, &unpoison_rs); |
230ac719c mm/hwpoison: don'... |
1429 1430 1431 1432 |
return 0; } if (page_mapping(page)) { |
495367c05 mm/memory-failure... |
1433 1434 |
unpoison_pr_info("Unpoison: the hwpoison page has non-NULL mapping %#lx ", |
a5f651090 mm: hwpoison: rat... |
1435 |
pfn, &unpoison_rs); |
230ac719c mm/hwpoison: don'... |
1436 1437 |
return 0; } |
0cea3fdc4 mm/hwpoison: fix ... |
1438 1439 1440 1441 1442 |
/* * unpoison_memory() can encounter thp only when the thp is being * worked by memory_failure() and the page lock is not held yet. * In such case, we yield to memory_failure() and make unpoison fail. */ |
e76d30e20 mm/hwpoison: fix ... |
1443 |
if (!PageHuge(page) && PageTransHuge(page)) { |
495367c05 mm/memory-failure... |
1444 1445 |
unpoison_pr_info("Unpoison: Memory failure is now running on %#lx ", |
a5f651090 mm: hwpoison: rat... |
1446 |
pfn, &unpoison_rs); |
ead07f6a8 mm/memory-failure... |
1447 |
return 0; |
0cea3fdc4 mm/hwpoison: fix ... |
1448 |
} |
ead07f6a8 mm/memory-failure... |
1449 |
if (!get_hwpoison_page(p)) { |
847ce401d HWPOISON: Add unp... |
1450 |
if (TestClearPageHWPoison(p)) |
8e30456b6 mm/hwpoison: intr... |
1451 |
num_poisoned_pages_dec(); |
495367c05 mm/memory-failure... |
1452 1453 |
unpoison_pr_info("Unpoison: Software-unpoisoned free page %#lx ", |
a5f651090 mm: hwpoison: rat... |
1454 |
pfn, &unpoison_rs); |
847ce401d HWPOISON: Add unp... |
1455 1456 |
return 0; } |
7eaceacca block: remove per... |
1457 |
lock_page(page); |
847ce401d HWPOISON: Add unp... |
1458 1459 1460 1461 1462 1463 |
/* * This test is racy because PG_hwpoison is set outside of page lock. * That's acceptable because that won't trigger kernel panic. Instead, * the PG_hwpoison page will be caught and isolated on the entrance to * the free buddy page pool. */ |
c9fbdd5f1 HWPOISON, hugetlb... |
1464 |
if (TestClearPageHWPoison(page)) { |
495367c05 mm/memory-failure... |
1465 1466 |
unpoison_pr_info("Unpoison: Software-unpoisoned page %#lx ", |
a5f651090 mm: hwpoison: rat... |
1467 |
pfn, &unpoison_rs); |
b37ff71cc mm: hwpoison: cha... |
1468 |
num_poisoned_pages_dec(); |
847ce401d HWPOISON: Add unp... |
1469 1470 1471 |
freeit = 1; } unlock_page(page); |
665d9da7f mm/hwpoison: repl... |
1472 |
put_hwpoison_page(page); |
3ba5eebc4 mm/memory-failure... |
1473 |
if (freeit && !(pfn == my_zero_pfn(0) && page_count(p) == 1)) |
665d9da7f mm/hwpoison: repl... |
1474 |
put_hwpoison_page(page); |
847ce401d HWPOISON: Add unp... |
1475 1476 1477 1478 |
return 0; } EXPORT_SYMBOL(unpoison_memory); |
facb6011f HWPOISON: Add sof... |
1479 1480 1481 |
static struct page *new_page(struct page *p, unsigned long private, int **x) { |
12686d153 HWPOISON: Try to ... |
1482 |
int nid = page_to_nid(p); |
94310cbca mm/madvise: enabl... |
1483 |
|
ef77ba5ce mm, hugetlb, soft... |
1484 |
return new_page_nodemask(p, nid, &node_states[N_MEMORY]); |
facb6011f HWPOISON: Add sof... |
1485 1486 1487 1488 1489 1490 1491 1492 |
} /* * Safely get reference count of an arbitrary page. * Returns 0 for a free page, -EIO for a zero refcount page * that is not free, and 1 for any other page type. * For 1 the page is returned with increased page count, otherwise not. */ |
af8fae7c0 mm/memory-failure... |
1493 |
static int __get_any_page(struct page *p, unsigned long pfn, int flags) |
facb6011f HWPOISON: Add sof... |
1494 1495 1496 1497 1498 1499 1500 |
{ int ret; if (flags & MF_COUNT_INCREASED) return 1; /* |
d950b9588 HWPOISON, hugetlb... |
1501 1502 1503 |
* When the target page is a free hugepage, just remove it * from free hugepage list. */ |
ead07f6a8 mm/memory-failure... |
1504 |
if (!get_hwpoison_page(p)) { |
d950b9588 HWPOISON, hugetlb... |
1505 |
if (PageHuge(p)) { |
71dd0b8ae mm/memory_failure... |
1506 1507 |
pr_info("%s: %#lx free huge page ", __func__, pfn); |
af8fae7c0 mm/memory-failure... |
1508 |
ret = 0; |
d950b9588 HWPOISON, hugetlb... |
1509 |
} else if (is_free_buddy_page(p)) { |
71dd0b8ae mm/memory_failure... |
1510 1511 |
pr_info("%s: %#lx free buddy page ", __func__, pfn); |
facb6011f HWPOISON: Add sof... |
1512 1513 |
ret = 0; } else { |
71dd0b8ae mm/memory_failure... |
1514 1515 1516 |
pr_info("%s: %#lx: unknown zero refcount page type %lx ", __func__, pfn, p->flags); |
facb6011f HWPOISON: Add sof... |
1517 1518 1519 1520 1521 1522 |
ret = -EIO; } } else { /* Not a free page */ ret = 1; } |
facb6011f HWPOISON: Add sof... |
1523 1524 |
return ret; } |
af8fae7c0 mm/memory-failure... |
1525 1526 1527 |
static int get_any_page(struct page *page, unsigned long pfn, int flags) { int ret = __get_any_page(page, pfn, flags); |
85fbe5d1b HWPOISON: soft of... |
1528 1529 |
if (ret == 1 && !PageHuge(page) && !PageLRU(page) && !__PageMovable(page)) { |
af8fae7c0 mm/memory-failure... |
1530 1531 1532 |
/* * Try to free it. */ |
665d9da7f mm/hwpoison: repl... |
1533 |
put_hwpoison_page(page); |
af8fae7c0 mm/memory-failure... |
1534 1535 1536 1537 1538 1539 |
shake_page(page, 1); /* * Did it turn free? */ ret = __get_any_page(page, pfn, 0); |
d96b339f4 mm: soft-offline:... |
1540 |
if (ret == 1 && !PageLRU(page)) { |
4f32be677 mm/hwpoison: fix ... |
1541 |
/* Drop page reference which is from __get_any_page() */ |
665d9da7f mm/hwpoison: repl... |
1542 |
put_hwpoison_page(page); |
82a2481e8 mm/memory-failure... |
1543 1544 1545 |
pr_info("soft_offline: %#lx: unknown non LRU page type %lx (%pGp) ", pfn, page->flags, &page->flags); |
af8fae7c0 mm/memory-failure... |
1546 1547 1548 1549 1550 |
return -EIO; } } return ret; } |
d950b9588 HWPOISON, hugetlb... |
1551 1552 1553 1554 1555 |
static int soft_offline_huge_page(struct page *page, int flags) { int ret; unsigned long pfn = page_to_pfn(page); struct page *hpage = compound_head(page); |
b8ec1cee5 mm: soft-offline:... |
1556 |
LIST_HEAD(pagelist); |
d950b9588 HWPOISON, hugetlb... |
1557 |
|
af8fae7c0 mm/memory-failure... |
1558 1559 1560 1561 1562 |
/* * This double-check of PageHWPoison is to avoid the race with * memory_failure(). See also comment in __soft_offline_page(). */ lock_page(hpage); |
0ebff32c3 memory-failure: f... |
1563 |
if (PageHWPoison(hpage)) { |
af8fae7c0 mm/memory-failure... |
1564 |
unlock_page(hpage); |
665d9da7f mm/hwpoison: repl... |
1565 |
put_hwpoison_page(hpage); |
0ebff32c3 memory-failure: f... |
1566 1567 |
pr_info("soft offline: %#lx hugepage already poisoned ", pfn); |
af8fae7c0 mm/memory-failure... |
1568 |
return -EBUSY; |
0ebff32c3 memory-failure: f... |
1569 |
} |
af8fae7c0 mm/memory-failure... |
1570 |
unlock_page(hpage); |
d950b9588 HWPOISON, hugetlb... |
1571 |
|
bcc542223 mm: hugetlb: intr... |
1572 |
ret = isolate_huge_page(hpage, &pagelist); |
036138080 mm/hwpoison: fix ... |
1573 1574 1575 1576 |
/* * get_any_page() and isolate_huge_page() takes a refcount each, * so need to drop one here. */ |
665d9da7f mm/hwpoison: repl... |
1577 |
put_hwpoison_page(hpage); |
036138080 mm/hwpoison: fix ... |
1578 |
if (!ret) { |
bcc542223 mm: hugetlb: intr... |
1579 1580 1581 1582 |
pr_info("soft offline: %#lx hugepage failed to isolate ", pfn); return -EBUSY; } |
68711a746 mm, migration: ad... |
1583 |
ret = migrate_pages(&pagelist, new_page, NULL, MPOL_MF_MOVE_ALL, |
b8ec1cee5 mm: soft-offline:... |
1584 |
MIGRATE_SYNC, MR_MEMORY_FAILURE); |
d950b9588 HWPOISON, hugetlb... |
1585 |
if (ret) { |
82a2481e8 mm/memory-failure... |
1586 1587 1588 |
pr_info("soft offline: %#lx: migration failed %d, type %lx (%pGp) ", pfn, ret, page->flags, &page->flags); |
30809f559 mm/migrate: fix r... |
1589 1590 |
if (!list_empty(&pagelist)) putback_movable_pages(&pagelist); |
b8ec1cee5 mm: soft-offline:... |
1591 1592 |
if (ret > 0) ret = -EIO; |
af8fae7c0 mm/memory-failure... |
1593 |
} else { |
b37ff71cc mm: hwpoison: cha... |
1594 |
if (PageHuge(page)) |
c3114a84f mm: hugetlb: soft... |
1595 |
dissolve_free_huge_page(page); |
d950b9588 HWPOISON, hugetlb... |
1596 |
} |
d950b9588 HWPOISON, hugetlb... |
1597 1598 |
return ret; } |
af8fae7c0 mm/memory-failure... |
1599 1600 1601 1602 |
static int __soft_offline_page(struct page *page, int flags) { int ret; unsigned long pfn = page_to_pfn(page); |
facb6011f HWPOISON: Add sof... |
1603 |
|
facb6011f HWPOISON: Add sof... |
1604 |
/* |
af8fae7c0 mm/memory-failure... |
1605 1606 1607 1608 |
* Check PageHWPoison again inside page lock because PageHWPoison * is set by memory_failure() outside page lock. Note that * memory_failure() also double-checks PageHWPoison inside page lock, * so there's no race between soft_offline_page() and memory_failure(). |
facb6011f HWPOISON: Add sof... |
1609 |
*/ |
0ebff32c3 memory-failure: f... |
1610 1611 |
lock_page(page); wait_on_page_writeback(page); |
af8fae7c0 mm/memory-failure... |
1612 1613 |
if (PageHWPoison(page)) { unlock_page(page); |
665d9da7f mm/hwpoison: repl... |
1614 |
put_hwpoison_page(page); |
af8fae7c0 mm/memory-failure... |
1615 1616 1617 1618 |
pr_info("soft offline: %#lx page already poisoned ", pfn); return -EBUSY; } |
facb6011f HWPOISON: Add sof... |
1619 1620 1621 1622 1623 1624 |
/* * Try to invalidate first. This should work for * non dirty unmapped page cache pages. */ ret = invalidate_inode_page(page); unlock_page(page); |
facb6011f HWPOISON: Add sof... |
1625 |
/* |
facb6011f HWPOISON: Add sof... |
1626 1627 1628 |
* RED-PEN would be better to keep it isolated here, but we * would need to fix isolation locking first. */ |
facb6011f HWPOISON: Add sof... |
1629 |
if (ret == 1) { |
665d9da7f mm/hwpoison: repl... |
1630 |
put_hwpoison_page(page); |
fb46e7352 HWPOISON: Convert... |
1631 1632 |
pr_info("soft_offline: %#lx: invalidated ", pfn); |
af8fae7c0 mm/memory-failure... |
1633 |
SetPageHWPoison(page); |
8e30456b6 mm/hwpoison: intr... |
1634 |
num_poisoned_pages_inc(); |
af8fae7c0 mm/memory-failure... |
1635 |
return 0; |
facb6011f HWPOISON: Add sof... |
1636 1637 1638 1639 1640 1641 1642 |
} /* * Simple invalidation didn't work. * Try to migrate to a new page instead. migrate.c * handles a large number of cases for us. */ |
85fbe5d1b HWPOISON: soft of... |
1643 1644 1645 1646 |
if (PageLRU(page)) ret = isolate_lru_page(page); else ret = isolate_movable_page(page, ISOLATE_UNEVICTABLE); |
bd486285f mem-hwpoison: fix... |
1647 1648 1649 1650 |
/* * Drop page reference which is came from get_any_page() * successful isolate_lru_page() already took another one. */ |
665d9da7f mm/hwpoison: repl... |
1651 |
put_hwpoison_page(page); |
facb6011f HWPOISON: Add sof... |
1652 1653 |
if (!ret) { LIST_HEAD(pagelist); |
85fbe5d1b HWPOISON: soft of... |
1654 1655 1656 1657 1658 1659 1660 1661 |
/* * After isolated lru page, the PageLRU will be cleared, * so use !__PageMovable instead for LRU page's mapping * cannot have PAGE_MAPPING_MOVABLE. */ if (!__PageMovable(page)) inc_node_page_state(page, NR_ISOLATED_ANON + page_is_file_cache(page)); |
facb6011f HWPOISON: Add sof... |
1662 |
list_add(&page->lru, &pagelist); |
68711a746 mm, migration: ad... |
1663 |
ret = migrate_pages(&pagelist, new_page, NULL, MPOL_MF_MOVE_ALL, |
9c620e2bc mm: remove offlin... |
1664 |
MIGRATE_SYNC, MR_MEMORY_FAILURE); |
facb6011f HWPOISON: Add sof... |
1665 |
if (ret) { |
85fbe5d1b HWPOISON: soft of... |
1666 1667 |
if (!list_empty(&pagelist)) putback_movable_pages(&pagelist); |
59c82b70d mm/migrate: remov... |
1668 |
|
82a2481e8 mm/memory-failure... |
1669 1670 1671 |
pr_info("soft offline: %#lx: migration failed %d, type %lx (%pGp) ", pfn, ret, page->flags, &page->flags); |
facb6011f HWPOISON: Add sof... |
1672 1673 1674 1675 |
if (ret > 0) ret = -EIO; } } else { |
82a2481e8 mm/memory-failure... |
1676 1677 1678 |
pr_info("soft offline: %#lx: isolation failed: %d, page count %d, type %lx (%pGp) ", pfn, ret, page_count(page), page->flags, &page->flags); |
facb6011f HWPOISON: Add sof... |
1679 |
} |
facb6011f HWPOISON: Add sof... |
1680 1681 |
return ret; } |
86e057734 mm/hwpoison: drop... |
1682 |
|
acc14dc4b mm: soft-offline:... |
1683 1684 1685 1686 1687 1688 1689 |
static int soft_offline_in_use_page(struct page *page, int flags) { int ret; struct page *hpage = compound_head(page); if (!PageHuge(page) && PageTransHuge(hpage)) { lock_page(hpage); |
98fd1ef42 mm: soft-offline:... |
1690 1691 1692 1693 1694 1695 1696 1697 1698 |
if (!PageAnon(hpage) || unlikely(split_huge_page(hpage))) { unlock_page(hpage); if (!PageAnon(hpage)) pr_info("soft offline: %#lx: non anonymous thp ", page_to_pfn(page)); else pr_info("soft offline: %#lx: thp split failed ", page_to_pfn(page)); put_hwpoison_page(hpage); |
acc14dc4b mm: soft-offline:... |
1699 1700 |
return -EBUSY; } |
98fd1ef42 mm: soft-offline:... |
1701 |
unlock_page(hpage); |
acc14dc4b mm: soft-offline:... |
1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 |
get_hwpoison_page(page); put_hwpoison_page(hpage); } if (PageHuge(page)) ret = soft_offline_huge_page(page, flags); else ret = __soft_offline_page(page, flags); return ret; } static void soft_offline_free_page(struct page *page) { |
b37ff71cc mm: hwpoison: cha... |
1716 |
struct page *head = compound_head(page); |
acc14dc4b mm: soft-offline:... |
1717 |
|
b37ff71cc mm: hwpoison: cha... |
1718 1719 1720 |
if (!TestSetPageHWPoison(head)) { num_poisoned_pages_inc(); if (PageHuge(head)) |
d4a3a60b3 mm: soft-offline:... |
1721 |
dissolve_free_huge_page(page); |
acc14dc4b mm: soft-offline:... |
1722 1723 |
} } |
86e057734 mm/hwpoison: drop... |
1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 |
/** * soft_offline_page - Soft offline a page. * @page: page to offline * @flags: flags. Same as memory_failure(). * * Returns 0 on success, otherwise negated errno. * * Soft offline a page, by migration or invalidation, * without killing anything. This is for the case when * a page is not corrupted yet (so it's still valid to access), * but has had a number of corrected errors and is better taken * out. * * The actual policy on when to do that is maintained by * user space. * * This should never impact any application or cause data loss, * however it might take some time. * * This is not a 100% solution for all memory, but tries to be * ``good enough'' for the majority of memory. */ int soft_offline_page(struct page *page, int flags) { int ret; unsigned long pfn = page_to_pfn(page); |
86e057734 mm/hwpoison: drop... |
1750 1751 1752 1753 |
if (PageHWPoison(page)) { pr_info("soft offline: %#lx page already poisoned ", pfn); |
1e0e635be mm/hwpoison: fix ... |
1754 |
if (flags & MF_COUNT_INCREASED) |
665d9da7f mm/hwpoison: repl... |
1755 |
put_hwpoison_page(page); |
86e057734 mm/hwpoison: drop... |
1756 1757 |
return -EBUSY; } |
86e057734 mm/hwpoison: drop... |
1758 |
|
bfc8c9013 mem-hotplug: impl... |
1759 |
get_online_mems(); |
86e057734 mm/hwpoison: drop... |
1760 |
ret = get_any_page(page, pfn, flags); |
bfc8c9013 mem-hotplug: impl... |
1761 |
put_online_mems(); |
4e41a30c6 mm: hwpoison: adj... |
1762 |
|
acc14dc4b mm: soft-offline:... |
1763 1764 1765 1766 |
if (ret > 0) ret = soft_offline_in_use_page(page, flags); else if (ret == 0) soft_offline_free_page(page); |
4e41a30c6 mm: hwpoison: adj... |
1767 |
|
86e057734 mm/hwpoison: drop... |
1768 1769 |
return ret; } |