1GETTING STARTED WITH KMEMCHECK
2==============================
3
4Vegard Nossum <vegardno@ifi.uio.no>
5
6
7Contents
8========
90. Introduction
101. Downloading
112. Configuring and compiling
123. How to use
133.1. Booting
143.2. Run-time enable/disable
153.3. Debugging
163.4. Annotating false positives
174. Reporting errors
185. Technical description
19
20
210. Introduction
22===============
23
24kmemcheck is a debugging feature for the Linux Kernel. More specifically, it
25is a dynamic checker that detects and warns about some uses of uninitialized
26memory.
27
28Userspace programmers might be familiar with Valgrind's memcheck. The main
29difference between memcheck and kmemcheck is that memcheck works for userspace
30programs only, and kmemcheck works for the kernel only. The implementations
31are of course vastly different. Because of this, kmemcheck is not as accurate
32as memcheck, but it turns out to be good enough in practice to discover real
33programmer errors that the compiler is not able to find through static
34analysis.
35
36Enabling kmemcheck on a kernel will probably slow it down to the extent that
37the machine will not be usable for normal workloads such as e.g. an
38interactive desktop. kmemcheck will also cause the kernel to use about twice
39as much memory as normal. For this reason, kmemcheck is strictly a debugging
40feature.
41
42
431. Downloading
44==============
45
46As of version 2.6.31-rc1, kmemcheck is included in the mainline kernel.
47
48
492. Configuring and compiling
50============================
51
52kmemcheck only works for the x86 (both 32- and 64-bit) platform. A number of
53configuration variables must have specific settings in order for the kmemcheck
54menu to even appear in "menuconfig". These are:
55
56  o CONFIG_CC_OPTIMIZE_FOR_SIZE=n
57
58	This option is located under "General setup" / "Optimize for size".
59
60	Without this, gcc will use certain optimizations that usually lead to
61	false positive warnings from kmemcheck. An example of this is a 16-bit
62	field in a struct, where gcc may load 32 bits, then discard the upper
63	16 bits. kmemcheck sees only the 32-bit load, and may trigger a
64	warning for the upper 16 bits (if they're uninitialized).
65
66  o CONFIG_SLAB=y or CONFIG_SLUB=y
67
68	This option is located under "General setup" / "Choose SLAB
69	allocator".
70
71  o CONFIG_FUNCTION_TRACER=n
72
73	This option is located under "Kernel hacking" / "Tracers" / "Kernel
74	Function Tracer"
75
76	When function tracing is compiled in, gcc emits a call to another
77	function at the beginning of every function. This means that when the
78	page fault handler is called, the ftrace framework will be called
79	before kmemcheck has had a chance to handle the fault. If ftrace then
80	modifies memory that was tracked by kmemcheck, the result is an
81	endless recursive page fault.
82
83  o CONFIG_DEBUG_PAGEALLOC=n
84
85	This option is located under "Kernel hacking" / "Memory Debugging"
86	 / "Debug page memory allocations".
87
88In addition, I highly recommend turning on CONFIG_DEBUG_INFO=y. This is also
89located under "Kernel hacking". With this, you will be able to get line number
90information from the kmemcheck warnings, which is extremely valuable in
91debugging a problem. This option is not mandatory, however, because it slows
92down the compilation process and produces a much bigger kernel image.
93
94Now the kmemcheck menu should be visible (under "Kernel hacking" / "Memory
95Debugging" / "kmemcheck: trap use of uninitialized memory"). Here follows
96a description of the kmemcheck configuration variables:
97
98  o CONFIG_KMEMCHECK
99
100	This must be enabled in order to use kmemcheck at all...
101
102  o CONFIG_KMEMCHECK_[DISABLED | ENABLED | ONESHOT]_BY_DEFAULT
103
104	This option controls the status of kmemcheck at boot-time. "Enabled"
105	will enable kmemcheck right from the start, "disabled" will boot the
106	kernel as normal (but with the kmemcheck code compiled in, so it can
107	be enabled at run-time after the kernel has booted), and "one-shot" is
108	a special mode which will turn kmemcheck off automatically after
109	detecting the first use of uninitialized memory.
110
111	If you are using kmemcheck to actively debug a problem, then you
112	probably want to choose "enabled" here.
113
114	The one-shot mode is mostly useful in automated test setups because it
115	can prevent floods of warnings and increase the chances of the machine
116	surviving in case something is really wrong. In other cases, the one-
117	shot mode could actually be counter-productive because it would turn
118	itself off at the very first error -- in the case of a false positive
119	too -- and this would come in the way of debugging the specific
120	problem you were interested in.
121
122	If you would like to use your kernel as normal, but with a chance to
123	enable kmemcheck in case of some problem, it might be a good idea to
124	choose "disabled" here. When kmemcheck is disabled, most of the run-
125	time overhead is not incurred, and the kernel will be almost as fast
126	as normal.
127
128  o CONFIG_KMEMCHECK_QUEUE_SIZE
129
130	Select the maximum number of error reports to store in an internal
131	(fixed-size) buffer. Since errors can occur virtually anywhere and in
132	any context, we need a temporary storage area which is guaranteed not
133	to generate any other page faults when accessed. The queue will be
134	emptied as soon as a tasklet may be scheduled. If the queue is full,
135	new error reports will be lost.
136
137	The default value of 64 is probably fine. If some code produces more
138	than 64 errors within an irqs-off section, then the code is likely to
139	produce many, many more, too, and these additional reports seldom give
140	any more information (the first report is usually the most valuable
141	anyway).
142
143	This number might have to be adjusted if you are not using serial
144	console or similar to capture the kernel log. If you are using the
145	"dmesg" command to save the log, then getting a lot of kmemcheck
146	warnings might overflow the kernel log itself, and the earlier reports
147	will get lost in that way instead. Try setting this to 10 or so on
148	such a setup.
149
150  o CONFIG_KMEMCHECK_SHADOW_COPY_SHIFT
151
152	Select the number of shadow bytes to save along with each entry of the
153	error-report queue. These bytes indicate what parts of an allocation
154	are initialized, uninitialized, etc. and will be displayed when an
155	error is detected to help the debugging of a particular problem.
156
157	The number entered here is actually the logarithm of the number of
158	bytes that will be saved. So if you pick for example 5 here, kmemcheck
159	will save 2^5 = 32 bytes.
160
161	The default value should be fine for debugging most problems. It also
162	fits nicely within 80 columns.
163
164  o CONFIG_KMEMCHECK_PARTIAL_OK
165
166	This option (when enabled) works around certain GCC optimizations that
167	produce 32-bit reads from 16-bit variables where the upper 16 bits are
168	thrown away afterwards.
169
170	The default value (enabled) is recommended. This may of course hide
171	some real errors, but disabling it would probably produce a lot of
172	false positives.
173
174  o CONFIG_KMEMCHECK_BITOPS_OK
175
176	This option silences warnings that would be generated for bit-field
177	accesses where not all the bits are initialized at the same time. This
178	may also hide some real bugs.
179
180	This option is probably obsolete, or it should be replaced with
181	the kmemcheck-/bitfield-annotations for the code in question. The
182	default value is therefore fine.
183
184Now compile the kernel as usual.
185
186
1873. How to use
188=============
189
1903.1. Booting
191============
192
193First some information about the command-line options. There is only one
194option specific to kmemcheck, and this is called "kmemcheck". It can be used
195to override the default mode as chosen by the CONFIG_KMEMCHECK_*_BY_DEFAULT
196option. Its possible settings are:
197
198  o kmemcheck=0 (disabled)
199  o kmemcheck=1 (enabled)
200  o kmemcheck=2 (one-shot mode)
201
202If SLUB debugging has been enabled in the kernel, it may take precedence over
203kmemcheck in such a way that the slab caches which are under SLUB debugging
204will not be tracked by kmemcheck. In order to ensure that this doesn't happen
205(even though it shouldn't by default), use SLUB's boot option "slub_debug",
206like this: slub_debug=-
207
208In fact, this option may also be used for fine-grained control over SLUB vs.
209kmemcheck. For example, if the command line includes "kmemcheck=1
210slub_debug=,dentry", then SLUB debugging will be used only for the "dentry"
211slab cache, and with kmemcheck tracking all the other caches. This is advanced
212usage, however, and is not generally recommended.
213
214
2153.2. Run-time enable/disable
216============================
217
218When the kernel has booted, it is possible to enable or disable kmemcheck at
219run-time. WARNING: This feature is still experimental and may cause false
220positive warnings to appear. Therefore, try not to use this. If you find that
221it doesn't work properly (e.g. you see an unreasonable amount of warnings), I
222will be happy to take bug reports.
223
224Use the file /proc/sys/kernel/kmemcheck for this purpose, e.g.:
225
226	$ echo 0 > /proc/sys/kernel/kmemcheck # disables kmemcheck
227
228The numbers are the same as for the kmemcheck= command-line option.
229
230
2313.3. Debugging
232==============
233
234A typical report will look something like this:
235
236WARNING: kmemcheck: Caught 32-bit read from uninitialized memory (ffff88003e4a2024)
23780000000000000000000000000000000000000000088ffff0000000000000000
238 i i i i u u u u i i i i i i i i u u u u u u u u u u u u u u u u
239         ^
240
241Pid: 1856, comm: ntpdate Not tainted 2.6.29-rc5 #264 945P-A
242RIP: 0010:[<ffffffff8104ede8>]  [<ffffffff8104ede8>] __dequeue_signal+0xc8/0x190
243RSP: 0018:ffff88003cdf7d98  EFLAGS: 00210002
244RAX: 0000000000000030 RBX: ffff88003d4ea968 RCX: 0000000000000009
245RDX: ffff88003e5d6018 RSI: ffff88003e5d6024 RDI: ffff88003cdf7e84
246RBP: ffff88003cdf7db8 R08: ffff88003e5d6000 R09: 0000000000000000
247R10: 0000000000000080 R11: 0000000000000000 R12: 000000000000000e
248R13: ffff88003cdf7e78 R14: ffff88003d530710 R15: ffff88003d5a98c8
249FS:  0000000000000000(0000) GS:ffff880001982000(0063) knlGS:00000
250CS:  0010 DS: 002b ES: 002b CR0: 0000000080050033
251CR2: ffff88003f806ea0 CR3: 000000003c036000 CR4: 00000000000006a0
252DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
253DR3: 0000000000000000 DR6: 00000000ffff4ff0 DR7: 0000000000000400
254 [<ffffffff8104f04e>] dequeue_signal+0x8e/0x170
255 [<ffffffff81050bd8>] get_signal_to_deliver+0x98/0x390
256 [<ffffffff8100b87d>] do_notify_resume+0xad/0x7d0
257 [<ffffffff8100c7b5>] int_signal+0x12/0x17
258 [<ffffffffffffffff>] 0xffffffffffffffff
259
260The single most valuable information in this report is the RIP (or EIP on 32-
261bit) value. This will help us pinpoint exactly which instruction that caused
262the warning.
263
264If your kernel was compiled with CONFIG_DEBUG_INFO=y, then all we have to do
265is give this address to the addr2line program, like this:
266
267	$ addr2line -e vmlinux -i ffffffff8104ede8
268	arch/x86/include/asm/string_64.h:12
269	include/asm-generic/siginfo.h:287
270	kernel/signal.c:380
271	kernel/signal.c:410
272
273The "-e vmlinux" tells addr2line which file to look in. IMPORTANT: This must
274be the vmlinux of the kernel that produced the warning in the first place! If
275not, the line number information will almost certainly be wrong.
276
277The "-i" tells addr2line to also print the line numbers of inlined functions.
278In this case, the flag was very important, because otherwise, it would only
279have printed the first line, which is just a call to memcpy(), which could be
280called from a thousand places in the kernel, and is therefore not very useful.
281These inlined functions would not show up in the stack trace above, simply
282because the kernel doesn't load the extra debugging information. This
283technique can of course be used with ordinary kernel oopses as well.
284
285In this case, it's the caller of memcpy() that is interesting, and it can be
286found in include/asm-generic/siginfo.h, line 287:
287
288281 static inline void copy_siginfo(struct siginfo *to, struct siginfo *from)
289282 {
290283         if (from->si_code < 0)
291284                 memcpy(to, from, sizeof(*to));
292285         else
293286                 /* _sigchld is currently the largest know union member */
294287                 memcpy(to, from, __ARCH_SI_PREAMBLE_SIZE + sizeof(from->_sifields._sigchld));
295288 }
296
297Since this was a read (kmemcheck usually warns about reads only, though it can
298warn about writes to unallocated or freed memory as well), it was probably the
299"from" argument which contained some uninitialized bytes. Following the chain
300of calls, we move upwards to see where "from" was allocated or initialized,
301kernel/signal.c, line 380:
302
303359 static void collect_signal(int sig, struct sigpending *list, siginfo_t *info)
304360 {
305...
306367         list_for_each_entry(q, &list->list, list) {
307368                 if (q->info.si_signo == sig) {
308369                         if (first)
309370                                 goto still_pending;
310371                         first = q;
311...
312377         if (first) {
313378 still_pending:
314379                 list_del_init(&first->list);
315380                 copy_siginfo(info, &first->info);
316381                 __sigqueue_free(first);
317...
318392         }
319393 }
320
321Here, it is &first->info that is being passed on to copy_siginfo(). The
322variable "first" was found on a list -- passed in as the second argument to
323collect_signal(). We  continue our journey through the stack, to figure out
324where the item on "list" was allocated or initialized. We move to line 410:
325
326395 static int __dequeue_signal(struct sigpending *pending, sigset_t *mask,
327396                         siginfo_t *info)
328397 {
329...
330410                 collect_signal(sig, pending, info);
331...
332414 }
333
334Now we need to follow the "pending" pointer, since that is being passed on to
335collect_signal() as "list". At this point, we've run out of lines from the
336"addr2line" output. Not to worry, we just paste the next addresses from the
337kmemcheck stack dump, i.e.:
338
339 [<ffffffff8104f04e>] dequeue_signal+0x8e/0x170
340 [<ffffffff81050bd8>] get_signal_to_deliver+0x98/0x390
341 [<ffffffff8100b87d>] do_notify_resume+0xad/0x7d0
342 [<ffffffff8100c7b5>] int_signal+0x12/0x17
343
344	$ addr2line -e vmlinux -i ffffffff8104f04e ffffffff81050bd8 \
345		ffffffff8100b87d ffffffff8100c7b5
346	kernel/signal.c:446
347	kernel/signal.c:1806
348	arch/x86/kernel/signal.c:805
349	arch/x86/kernel/signal.c:871
350	arch/x86/kernel/entry_64.S:694
351
352Remember that since these addresses were found on the stack and not as the
353RIP value, they actually point to the _next_ instruction (they are return
354addresses). This becomes obvious when we look at the code for line 446:
355
356422 int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
357423 {
358...
359431                 signr = __dequeue_signal(&tsk->signal->shared_pending,
360432                                          mask, info);
361433                 /*
362434                  * itimer signal ?
363435                  *
364436                  * itimers are process shared and we restart periodic
365437                  * itimers in the signal delivery path to prevent DoS
366438                  * attacks in the high resolution timer case. This is
367439                  * compliant with the old way of self restarting
368440                  * itimers, as the SIGALRM is a legacy signal and only
369441                  * queued once. Changing the restart behaviour to
370442                  * restart the timer in the signal dequeue path is
371443                  * reducing the timer noise on heavy loaded !highres
372444                  * systems too.
373445                  */
374446                 if (unlikely(signr == SIGALRM)) {
375...
376489 }
377
378So instead of looking at 446, we should be looking at 431, which is the line
379that executes just before 446. Here we see that what we are looking for is
380&tsk->signal->shared_pending.
381
382Our next task is now to figure out which function that puts items on this
383"shared_pending" list. A crude, but efficient tool, is git grep:
384
385	$ git grep -n 'shared_pending' kernel/
386	...
387	kernel/signal.c:828:    pending = group ? &t->signal->shared_pending : &t->pending;
388	kernel/signal.c:1339:   pending = group ? &t->signal->shared_pending : &t->pending;
389	...
390
391There were more results, but none of them were related to list operations,
392and these were the only assignments. We inspect the line numbers more closely
393and find that this is indeed where items are being added to the list:
394
395816 static int send_signal(int sig, struct siginfo *info, struct task_struct *t,
396817                         int group)
397818 {
398...
399828         pending = group ? &t->signal->shared_pending : &t->pending;
400...
401851         q = __sigqueue_alloc(t, GFP_ATOMIC, (sig < SIGRTMIN &&
402852                                              (is_si_special(info) ||
403853                                               info->si_code >= 0)));
404854         if (q) {
405855                 list_add_tail(&q->list, &pending->list);
406...
407890 }
408
409and:
410
4111309 int send_sigqueue(struct sigqueue *q, struct task_struct *t, int group)
4121310 {
413....
4141339         pending = group ? &t->signal->shared_pending : &t->pending;
4151340         list_add_tail(&q->list, &pending->list);
416....
4171347 }
418
419In the first case, the list element we are looking for, "q", is being returned
420from the function __sigqueue_alloc(), which looks like an allocation function.
421Let's take a look at it:
422
423187 static struct sigqueue *__sigqueue_alloc(struct task_struct *t, gfp_t flags,
424188                                          int override_rlimit)
425189 {
426190         struct sigqueue *q = NULL;
427191         struct user_struct *user;
428192 
429193         /*
430194          * We won't get problems with the target's UID changing under us
431195          * because changing it requires RCU be used, and if t != current, the
432196          * caller must be holding the RCU readlock (by way of a spinlock) and
433197          * we use RCU protection here
434198          */
435199         user = get_uid(__task_cred(t)->user);
436200         atomic_inc(&user->sigpending);
437201         if (override_rlimit ||
438202             atomic_read(&user->sigpending) <=
439203                         t->signal->rlim[RLIMIT_SIGPENDING].rlim_cur)
440204                 q = kmem_cache_alloc(sigqueue_cachep, flags);
441205         if (unlikely(q == NULL)) {
442206                 atomic_dec(&user->sigpending);
443207                 free_uid(user);
444208         } else {
445209                 INIT_LIST_HEAD(&q->list);
446210                 q->flags = 0;
447211                 q->user = user;
448212         }
449213 
450214         return q;
451215 }
452
453We see that this function initializes q->list, q->flags, and q->user. It seems
454that now is the time to look at the definition of "struct sigqueue", e.g.:
455
45614 struct sigqueue {
45715         struct list_head list;
45816         int flags;
45917         siginfo_t info;
46018         struct user_struct *user;
46119 };
462
463And, you might remember, it was a memcpy() on &first->info that caused the
464warning, so this makes perfect sense. It also seems reasonable to assume that
465it is the caller of __sigqueue_alloc() that has the responsibility of filling
466out (initializing) this member.
467
468But just which fields of the struct were uninitialized? Let's look at
469kmemcheck's report again:
470
471WARNING: kmemcheck: Caught 32-bit read from uninitialized memory (ffff88003e4a2024)
47280000000000000000000000000000000000000000088ffff0000000000000000
473 i i i i u u u u i i i i i i i i u u u u u u u u u u u u u u u u
474         ^
475
476These first two lines are the memory dump of the memory object itself, and the
477shadow bytemap, respectively. The memory object itself is in this case
478&first->info. Just beware that the start of this dump is NOT the start of the
479object itself! The position of the caret (^) corresponds with the address of
480the read (ffff88003e4a2024).
481
482The shadow bytemap dump legend is as follows:
483
484  i - initialized
485  u - uninitialized
486  a - unallocated (memory has been allocated by the slab layer, but has not
487      yet been handed off to anybody)
488  f - freed (memory has been allocated by the slab layer, but has been freed
489      by the previous owner)
490
491In order to figure out where (relative to the start of the object) the
492uninitialized memory was located, we have to look at the disassembly. For
493that, we'll need the RIP address again:
494
495RIP: 0010:[<ffffffff8104ede8>]  [<ffffffff8104ede8>] __dequeue_signal+0xc8/0x190
496
497	$ objdump -d --no-show-raw-insn vmlinux | grep -C 8 ffffffff8104ede8:
498	ffffffff8104edc8:       mov    %r8,0x8(%r8)
499	ffffffff8104edcc:       test   %r10d,%r10d
500	ffffffff8104edcf:       js     ffffffff8104ee88 <__dequeue_signal+0x168>
501	ffffffff8104edd5:       mov    %rax,%rdx
502	ffffffff8104edd8:       mov    $0xc,%ecx
503	ffffffff8104eddd:       mov    %r13,%rdi
504	ffffffff8104ede0:       mov    $0x30,%eax
505	ffffffff8104ede5:       mov    %rdx,%rsi
506	ffffffff8104ede8:       rep movsl %ds:(%rsi),%es:(%rdi)
507	ffffffff8104edea:       test   $0x2,%al
508	ffffffff8104edec:       je     ffffffff8104edf0 <__dequeue_signal+0xd0>
509	ffffffff8104edee:       movsw  %ds:(%rsi),%es:(%rdi)
510	ffffffff8104edf0:       test   $0x1,%al
511	ffffffff8104edf2:       je     ffffffff8104edf5 <__dequeue_signal+0xd5>
512	ffffffff8104edf4:       movsb  %ds:(%rsi),%es:(%rdi)
513	ffffffff8104edf5:       mov    %r8,%rdi
514	ffffffff8104edf8:       callq  ffffffff8104de60 <__sigqueue_free>
515
516As expected, it's the "rep movsl" instruction from the memcpy() that causes
517the warning. We know about REP MOVSL that it uses the register RCX to count
518the number of remaining iterations. By taking a look at the register dump
519again (from the kmemcheck report), we can figure out how many bytes were left
520to copy:
521
522RAX: 0000000000000030 RBX: ffff88003d4ea968 RCX: 0000000000000009
523
524By looking at the disassembly, we also see that %ecx is being loaded with the
525value $0xc just before (ffffffff8104edd8), so we are very lucky. Keep in mind
526that this is the number of iterations, not bytes. And since this is a "long"
527operation, we need to multiply by 4 to get the number of bytes. So this means
528that the uninitialized value was encountered at 4 * (0xc - 0x9) = 12 bytes
529from the start of the object.
530
531We can now try to figure out which field of the "struct siginfo" that was not
532initialized. This is the beginning of the struct:
533
53440 typedef struct siginfo {
53541         int si_signo;
53642         int si_errno;
53743         int si_code;
53844                 
53945         union {
540..
54192         } _sifields;
54293 } siginfo_t;
543
544On 64-bit, the int is 4 bytes long, so it must the union member that has
545not been initialized. We can verify this using gdb:
546
547	$ gdb vmlinux
548	...
549	(gdb) p &((struct siginfo *) 0)->_sifields
550	$1 = (union {...} *) 0x10
551
552Actually, it seems that the union member is located at offset 0x10 -- which
553means that gcc has inserted 4 bytes of padding between the members si_code
554and _sifields. We can now get a fuller picture of the memory dump:
555
556         _----------------------------=> si_code
557        /        _--------------------=> (padding)
558       |        /        _------------=> _sifields(._kill._pid)
559       |       |        /        _----=> _sifields(._kill._uid)
560       |       |       |        / 
561-------|-------|-------|-------|
56280000000000000000000000000000000000000000088ffff0000000000000000
563 i i i i u u u u i i i i i i i i u u u u u u u u u u u u u u u u
564
565This allows us to realize another important fact: si_code contains the value
5660x80. Remember that x86 is little endian, so the first 4 bytes "80000000" are
567really the number 0x00000080. With a bit of research, we find that this is
568actually the constant SI_KERNEL defined in include/asm-generic/siginfo.h:
569
570144 #define SI_KERNEL       0x80            /* sent by the kernel from somewhere     */
571
572This macro is used in exactly one place in the x86 kernel: In send_signal()
573in kernel/signal.c:
574
575816 static int send_signal(int sig, struct siginfo *info, struct task_struct *t,
576817                         int group)
577818 {
578...
579828         pending = group ? &t->signal->shared_pending : &t->pending;
580...
581851         q = __sigqueue_alloc(t, GFP_ATOMIC, (sig < SIGRTMIN &&
582852                                              (is_si_special(info) ||
583853                                               info->si_code >= 0)));
584854         if (q) {
585855                 list_add_tail(&q->list, &pending->list);
586856                 switch ((unsigned long) info) {
587...
588865                 case (unsigned long) SEND_SIG_PRIV:
589866                         q->info.si_signo = sig;
590867                         q->info.si_errno = 0;
591868                         q->info.si_code = SI_KERNEL;
592869                         q->info.si_pid = 0;
593870                         q->info.si_uid = 0;
594871                         break;
595...
596890 }
597
598Not only does this match with the .si_code member, it also matches the place
599we found earlier when looking for where siginfo_t objects are enqueued on the
600"shared_pending" list.
601
602So to sum up: It seems that it is the padding introduced by the compiler
603between two struct fields that is uninitialized, and this gets reported when
604we do a memcpy() on the struct. This means that we have identified a false
605positive warning.
606
607Normally, kmemcheck will not report uninitialized accesses in memcpy() calls
608when both the source and destination addresses are tracked. (Instead, we copy
609the shadow bytemap as well). In this case, the destination address clearly
610was not tracked. We can dig a little deeper into the stack trace from above:
611
612	arch/x86/kernel/signal.c:805
613	arch/x86/kernel/signal.c:871
614	arch/x86/kernel/entry_64.S:694
615
616And we clearly see that the destination siginfo object is located on the
617stack:
618
619782 static void do_signal(struct pt_regs *regs)
620783 {
621784         struct k_sigaction ka;
622785         siginfo_t info;
623...
624804         signr = get_signal_to_deliver(&info, &ka, regs, NULL);
625...
626854 }
627
628And this &info is what eventually gets passed to copy_siginfo() as the
629destination argument.
630
631Now, even though we didn't find an actual error here, the example is still a
632good one, because it shows how one would go about to find out what the report
633was all about.
634
635
6363.4. Annotating false positives
637===============================
638
639There are a few different ways to make annotations in the source code that
640will keep kmemcheck from checking and reporting certain allocations. Here
641they are:
642
643  o __GFP_NOTRACK_FALSE_POSITIVE
644
645	This flag can be passed to kmalloc() or kmem_cache_alloc() (therefore
646	also to other functions that end up calling one of these) to indicate
647	that the allocation should not be tracked because it would lead to
648	a false positive report. This is a "big hammer" way of silencing
649	kmemcheck; after all, even if the false positive pertains to 
650	particular field in a struct, for example, we will now lose the
651	ability to find (real) errors in other parts of the same struct.
652
653	Example:
654
655	    /* No warnings will ever trigger on accessing any part of x */
656	    x = kmalloc(sizeof *x, GFP_KERNEL | __GFP_NOTRACK_FALSE_POSITIVE);
657
658  o kmemcheck_bitfield_begin(name)/kmemcheck_bitfield_end(name) and
659	kmemcheck_annotate_bitfield(ptr, name)
660
661	The first two of these three macros can be used inside struct
662	definitions to signal, respectively, the beginning and end of a
663	bitfield. Additionally, this will assign the bitfield a name, which
664	is given as an argument to the macros.
665
666	Having used these markers, one can later use
667	kmemcheck_annotate_bitfield() at the point of allocation, to indicate
668	which parts of the allocation is part of a bitfield.
669
670	Example:
671
672	    struct foo {
673		int x;
674
675		kmemcheck_bitfield_begin(flags);
676		int flag_a:1;
677		int flag_b:1;
678		kmemcheck_bitfield_end(flags);
679
680		int y;
681	    };
682
683	    struct foo *x = kmalloc(sizeof *x);
684
685	    /* No warnings will trigger on accessing the bitfield of x */
686	    kmemcheck_annotate_bitfield(x, flags);
687
688	Note that kmemcheck_annotate_bitfield() can be used even before the
689	return value of kmalloc() is checked -- in other words, passing NULL
690	as the first argument is legal (and will do nothing).
691
692
6934. Reporting errors
694===================
695
696As we have seen, kmemcheck will produce false positive reports. Therefore, it
697is not very wise to blindly post kmemcheck warnings to mailing lists and
698maintainers. Instead, I encourage maintainers and developers to find errors
699in their own code. If you get a warning, you can try to work around it, try
700to figure out if it's a real error or not, or simply ignore it. Most
701developers know their own code and will quickly and efficiently determine the
702root cause of a kmemcheck report. This is therefore also the most efficient
703way to work with kmemcheck.
704
705That said, we (the kmemcheck maintainers) will always be on the lookout for
706false positives that we can annotate and silence. So whatever you find,
707please drop us a note privately! Kernel configs and steps to reproduce (if
708available) are of course a great help too.
709
710Happy hacking!
711
712
7135. Technical description
714========================
715
716kmemcheck works by marking memory pages non-present. This means that whenever
717somebody attempts to access the page, a page fault is generated. The page
718fault handler notices that the page was in fact only hidden, and so it calls
719on the kmemcheck code to make further investigations.
720
721When the investigations are completed, kmemcheck "shows" the page by marking
722it present (as it would be under normal circumstances). This way, the
723interrupted code can continue as usual.
724
725But after the instruction has been executed, we should hide the page again, so
726that we can catch the next access too! Now kmemcheck makes use of a debugging
727feature of the processor, namely single-stepping. When the processor has
728finished the one instruction that generated the memory access, a debug
729exception is raised. From here, we simply hide the page again and continue
730execution, this time with the single-stepping feature turned off.
731
732kmemcheck requires some assistance from the memory allocator in order to work.
733The memory allocator needs to
734
735  1. Tell kmemcheck about newly allocated pages and pages that are about to
736     be freed. This allows kmemcheck to set up and tear down the shadow memory
737     for the pages in question. The shadow memory stores the status of each
738     byte in the allocation proper, e.g. whether it is initialized or
739     uninitialized.
740
741  2. Tell kmemcheck which parts of memory should be marked uninitialized.
742     There are actually a few more states, such as "not yet allocated" and
743     "recently freed".
744
745If a slab cache is set up using the SLAB_NOTRACK flag, it will never return
746memory that can take page faults because of kmemcheck.
747
748If a slab cache is NOT set up using the SLAB_NOTRACK flag, callers can still
749request memory with the __GFP_NOTRACK or __GFP_NOTRACK_FALSE_POSITIVE flags.
750This does not prevent the page faults from occurring, however, but marks the
751object in question as being initialized so that no warnings will ever be
752produced for this object.
753
754Currently, the SLAB and SLUB allocators are supported by kmemcheck.
755