Eric Lee / smarc-fsl-linux-kernel

Download as
Browse Code ยป

Commit e2852ae825dba5ebc159788720baec1a28a57125

Authored by Tejun Heo
Committed by Linus Torvalds
1 parent 066a9be6c0
Exists in master and in 39 other branches 8mp-imx_5.4.70_2.3.0, 8qm-imx_5.4.70_2.3.0, emb_imx_lf-5.15.y, emb_lf-6.1.y, imx_3.0.35_4.1.0, imx_3.10.17_1.0.1_ga, imx_3.10.53_1.1.0_ga, imx_3.14.28_1.0.0_ga, imx_4.1.15_1.0.0_ga, pitx_8mp_lf-5.10.y, rt-smarc-imx_4.1.15_1.0.0_ga, rt_linux_5.15.71, smarc-8m-android-11.0.0_2.0.0, smarc-imx6_4.14.98_2.0.0_ga, smarc-imx6_4.9.88_2.0.0_ga, smarc-imx7_4.14.98_2.0.0_ga, smarc-imx7_4.9.11_1.0.0_ga, smarc-imx7_4.9.88_2.0.0_ga, smarc-imx_3.10.53_1.1.0_ga, smarc-imx_3.14.28_1.0.0_ga, smarc-imx_4.1.15_1.0.0_ga, smarc-imx_4.9.11_1.0.0_ga, smarc-imx_4.9.51_imx8m_ga, smarc-imx_4.9.88_2.0.0_ga, smarc-m6.0.1_2.1.0-ga, smarc-n7.1.2_2.0.0-ga, smarc-rel_imx_4.1.15_1.2.0_ga, smarc_8m_00d0_imx_4.14.98_2.0.0_ga, smarc_8m_imx_4.14.78_1.0.0_ga, smarc_8m_imx_4.14.98_2.0.0_ga, smarc_8m_imx_4.19.35_1.1.0, smarc_8mm_imx_4.14.78_1.0.0_ga, smarc_8mm_imx_4.14.98_2.0.0_ga, smarc_8mm_imx_4.19.35_1.1.0, smarc_8mm_imx_5.4.24_2.1.0, smarc_8mp_lf-5.10.y, smarc_8mq_imx_5.4.24_2.1.0, smarc_8mq_lf-5.10.y, smarc_imx_lf-5.15.y

percpu_counter: add debugobj support 

All percpu counters are linked to a global list on initialization and
removed from it on destruction.  The list is walked during CPU up/down.
If a percpu counter is freed without being properly destroyed, the system
will oops only on the next CPU up/down making it pretty nasty to track
down.  This patch adds debugobj support for percpu counters so that such
problems can be found easily.

As percpu counters don't make sense on stack and can't be statically
initialized, debugobj support is pretty simple.  It's initialized and
activated on counter initialization, and deactivatd and destroyed on
counter destruction.  With this patch applied, the bug fixed by commit
602586a83b719df0fbd94196a1359ed35aeb2df3 (shmem: put_super must
percpu_counter_destroy) triggers the following warning on tmpfs unmount
and the system won't oops on the next cpu up/down operation.

 ------------[ cut here ]------------
 WARNING: at lib/debugobjects.c:259 debug_print_object+0x5c/0x70()
 Hardware name: Bochs
 ODEBUG: free active (active state 0) object type: percpu_counter
 Modules linked in:
 Pid: 3999, comm: umount Not tainted 2.6.36-rc2-work+ #5
 Call Trace:
  [<ffffffff81083f7f>] warn_slowpath_common+0x7f/0xc0
  [<ffffffff81084076>] warn_slowpath_fmt+0x46/0x50
  [<ffffffff813b45cc>] debug_print_object+0x5c/0x70
  [<ffffffff813b50e5>] debug_check_no_obj_freed+0x125/0x210
  [<ffffffff811577d3>] kfree+0xb3/0x2f0
  [<ffffffff81132edd>] shmem_put_super+0x1d/0x30
  [<ffffffff81162e96>] generic_shutdown_super+0x56/0xe0
  [<ffffffff81162f86>] kill_anon_super+0x16/0x60
  [<ffffffff81162ff7>] kill_litter_super+0x27/0x30
  [<ffffffff81163295>] deactivate_locked_super+0x45/0x60
  [<ffffffff81163cfa>] deactivate_super+0x4a/0x70
  [<ffffffff8117d446>] mntput_no_expire+0x86/0xe0
  [<ffffffff8117df7f>] sys_umount+0x6f/0x360
  [<ffffffff8103f01b>] system_call_fastpath+0x16/0x1b
 ---[ end trace cce2a341ba3611a7 ]---

Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Thomas Gleixner <tglxlinutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Showing 2 changed files with 56 additions and 0 deletions Inline Diff

1 1
2 config PRINTK_TIME 2 config PRINTK_TIME
3 bool "Show timing information on printks" 3 bool "Show timing information on printks"
4 depends on PRINTK 4 depends on PRINTK
5 help 5 help
6 Selecting this option causes timing information to be 6 Selecting this option causes timing information to be
7 included in printk output. This allows you to measure 7 included in printk output. This allows you to measure
8 the interval between kernel operations, including bootup 8 the interval between kernel operations, including bootup
9 operations. This is useful for identifying long delays 9 operations. This is useful for identifying long delays
10 in kernel startup. 10 in kernel startup.
11 11
12 config ENABLE_WARN_DEPRECATED 12 config ENABLE_WARN_DEPRECATED
13 bool "Enable __deprecated logic" 13 bool "Enable __deprecated logic"
14 default y 14 default y
15 help 15 help
16 Enable the __deprecated logic in the kernel build. 16 Enable the __deprecated logic in the kernel build.
17 Disable this to suppress the "warning: 'foo' is deprecated 17 Disable this to suppress the "warning: 'foo' is deprecated
18 (declared at kernel/power/somefile.c:1234)" messages. 18 (declared at kernel/power/somefile.c:1234)" messages.
19 19
20 config ENABLE_MUST_CHECK 20 config ENABLE_MUST_CHECK
21 bool "Enable __must_check logic" 21 bool "Enable __must_check logic"
22 default y 22 default y
23 help 23 help
24 Enable the __must_check logic in the kernel build. Disable this to 24 Enable the __must_check logic in the kernel build. Disable this to
25 suppress the "warning: ignoring return value of 'foo', declared with 25 suppress the "warning: ignoring return value of 'foo', declared with
26 attribute warn_unused_result" messages. 26 attribute warn_unused_result" messages.
27 27
28 config FRAME_WARN 28 config FRAME_WARN
29 int "Warn for stack frames larger than (needs gcc 4.4)" 29 int "Warn for stack frames larger than (needs gcc 4.4)"
30 range 0 8192 30 range 0 8192
31 default 1024 if !64BIT 31 default 1024 if !64BIT
32 default 2048 if 64BIT 32 default 2048 if 64BIT
33 help 33 help
34 Tell gcc to warn at build time for stack frames larger than this. 34 Tell gcc to warn at build time for stack frames larger than this.
35 Setting this too low will cause a lot of warnings. 35 Setting this too low will cause a lot of warnings.
36 Setting it to 0 disables the warning. 36 Setting it to 0 disables the warning.
37 Requires gcc 4.4 37 Requires gcc 4.4
38 38
39 config MAGIC_SYSRQ 39 config MAGIC_SYSRQ
40 bool "Magic SysRq key" 40 bool "Magic SysRq key"
41 depends on !UML 41 depends on !UML
42 help 42 help
43 If you say Y here, you will have some control over the system even 43 If you say Y here, you will have some control over the system even
44 if the system crashes for example during kernel debugging (e.g., you 44 if the system crashes for example during kernel debugging (e.g., you
45 will be able to flush the buffer cache to disk, reboot the system 45 will be able to flush the buffer cache to disk, reboot the system
46 immediately or dump some status information). This is accomplished 46 immediately or dump some status information). This is accomplished
47 by pressing various keys while holding SysRq (Alt+PrintScreen). It 47 by pressing various keys while holding SysRq (Alt+PrintScreen). It
48 also works on a serial console (on PC hardware at least), if you 48 also works on a serial console (on PC hardware at least), if you
49 send a BREAK and then within 5 seconds a command keypress. The 49 send a BREAK and then within 5 seconds a command keypress. The
50 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y 50 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
51 unless you really know what this hack does. 51 unless you really know what this hack does.
52 52
53 config STRIP_ASM_SYMS 53 config STRIP_ASM_SYMS
54 bool "Strip assembler-generated symbols during link" 54 bool "Strip assembler-generated symbols during link"
55 default n 55 default n
56 help 56 help
57 Strip internal assembler-generated symbols during a link (symbols 57 Strip internal assembler-generated symbols during a link (symbols
58 that look like '.Lxxx') so they don't pollute the output of 58 that look like '.Lxxx') so they don't pollute the output of
59 get_wchan() and suchlike. 59 get_wchan() and suchlike.
60 60
61 config UNUSED_SYMBOLS 61 config UNUSED_SYMBOLS
62 bool "Enable unused/obsolete exported symbols" 62 bool "Enable unused/obsolete exported symbols"
63 default y if X86 63 default y if X86
64 help 64 help
65 Unused but exported symbols make the kernel needlessly bigger. For 65 Unused but exported symbols make the kernel needlessly bigger. For
66 that reason most of these unused exports will soon be removed. This 66 that reason most of these unused exports will soon be removed. This
67 option is provided temporarily to provide a transition period in case 67 option is provided temporarily to provide a transition period in case
68 some external kernel module needs one of these symbols anyway. If you 68 some external kernel module needs one of these symbols anyway. If you
69 encounter such a case in your module, consider if you are actually 69 encounter such a case in your module, consider if you are actually
70 using the right API. (rationale: since nobody in the kernel is using 70 using the right API. (rationale: since nobody in the kernel is using
71 this in a module, there is a pretty good chance it's actually the 71 this in a module, there is a pretty good chance it's actually the
72 wrong interface to use). If you really need the symbol, please send a 72 wrong interface to use). If you really need the symbol, please send a
73 mail to the linux kernel mailing list mentioning the symbol and why 73 mail to the linux kernel mailing list mentioning the symbol and why
74 you really need it, and what the merge plan to the mainline kernel for 74 you really need it, and what the merge plan to the mainline kernel for
75 your module is. 75 your module is.
76 76
77 config DEBUG_FS 77 config DEBUG_FS
78 bool "Debug Filesystem" 78 bool "Debug Filesystem"
79 help 79 help
80 debugfs is a virtual file system that kernel developers use to put 80 debugfs is a virtual file system that kernel developers use to put
81 debugging files into. Enable this option to be able to read and 81 debugging files into. Enable this option to be able to read and
82 write to these files. 82 write to these files.
83 83
84 For detailed documentation on the debugfs API, see 84 For detailed documentation on the debugfs API, see
85 Documentation/DocBook/filesystems. 85 Documentation/DocBook/filesystems.
86 86
87 If unsure, say N. 87 If unsure, say N.
88 88
89 config HEADERS_CHECK 89 config HEADERS_CHECK
90 bool "Run 'make headers_check' when building vmlinux" 90 bool "Run 'make headers_check' when building vmlinux"
91 depends on !UML 91 depends on !UML
92 help 92 help
93 This option will extract the user-visible kernel headers whenever 93 This option will extract the user-visible kernel headers whenever
94 building the kernel, and will run basic sanity checks on them to 94 building the kernel, and will run basic sanity checks on them to
95 ensure that exported files do not attempt to include files which 95 ensure that exported files do not attempt to include files which
96 were not exported, etc. 96 were not exported, etc.
97 97
98 If you're making modifications to header files which are 98 If you're making modifications to header files which are
99 relevant for userspace, say 'Y', and check the headers 99 relevant for userspace, say 'Y', and check the headers
100 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in 100 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
101 your build tree), to make sure they're suitable. 101 your build tree), to make sure they're suitable.
102 102
103 config DEBUG_SECTION_MISMATCH 103 config DEBUG_SECTION_MISMATCH
104 bool "Enable full Section mismatch analysis" 104 bool "Enable full Section mismatch analysis"
105 depends on UNDEFINED || (BLACKFIN) 105 depends on UNDEFINED || (BLACKFIN)
106 default y 106 default y
107 # This option is on purpose disabled for now. 107 # This option is on purpose disabled for now.
108 # It will be enabled when we are down to a reasonable number 108 # It will be enabled when we are down to a reasonable number
109 # of section mismatch warnings (< 10 for an allyesconfig build) 109 # of section mismatch warnings (< 10 for an allyesconfig build)
110 help 110 help
111 The section mismatch analysis checks if there are illegal 111 The section mismatch analysis checks if there are illegal
112 references from one section to another section. 112 references from one section to another section.
113 Linux will during link or during runtime drop some sections 113 Linux will during link or during runtime drop some sections
114 and any use of code/data previously in these sections will 114 and any use of code/data previously in these sections will
115 most likely result in an oops. 115 most likely result in an oops.
116 In the code functions and variables are annotated with 116 In the code functions and variables are annotated with
117 __init, __devinit etc. (see full list in include/linux/init.h) 117 __init, __devinit etc. (see full list in include/linux/init.h)
118 which results in the code/data being placed in specific sections. 118 which results in the code/data being placed in specific sections.
119 The section mismatch analysis is always done after a full 119 The section mismatch analysis is always done after a full
120 kernel build but enabling this option will in addition 120 kernel build but enabling this option will in addition
121 do the following: 121 do the following:
122 - Add the option -fno-inline-functions-called-once to gcc 122 - Add the option -fno-inline-functions-called-once to gcc
123 When inlining a function annotated __init in a non-init 123 When inlining a function annotated __init in a non-init
124 function we would lose the section information and thus 124 function we would lose the section information and thus
125 the analysis would not catch the illegal reference. 125 the analysis would not catch the illegal reference.
126 This option tells gcc to inline less but will also 126 This option tells gcc to inline less but will also
127 result in a larger kernel. 127 result in a larger kernel.
128 - Run the section mismatch analysis for each module/built-in.o 128 - Run the section mismatch analysis for each module/built-in.o
129 When we run the section mismatch analysis on vmlinux.o we 129 When we run the section mismatch analysis on vmlinux.o we
130 lose valueble information about where the mismatch was 130 lose valueble information about where the mismatch was
131 introduced. 131 introduced.
132 Running the analysis for each module/built-in.o file 132 Running the analysis for each module/built-in.o file
133 will tell where the mismatch happens much closer to the 133 will tell where the mismatch happens much closer to the
134 source. The drawback is that we will report the same 134 source. The drawback is that we will report the same
135 mismatch at least twice. 135 mismatch at least twice.
136 - Enable verbose reporting from modpost to help solving 136 - Enable verbose reporting from modpost to help solving
137 the section mismatches reported. 137 the section mismatches reported.
138 138
139 config DEBUG_KERNEL 139 config DEBUG_KERNEL
140 bool "Kernel debugging" 140 bool "Kernel debugging"
141 help 141 help
142 Say Y here if you are developing drivers or trying to debug and 142 Say Y here if you are developing drivers or trying to debug and
143 identify kernel problems. 143 identify kernel problems.
144 144
145 config DEBUG_SHIRQ 145 config DEBUG_SHIRQ
146 bool "Debug shared IRQ handlers" 146 bool "Debug shared IRQ handlers"
147 depends on DEBUG_KERNEL && GENERIC_HARDIRQS 147 depends on DEBUG_KERNEL && GENERIC_HARDIRQS
148 help 148 help
149 Enable this to generate a spurious interrupt as soon as a shared 149 Enable this to generate a spurious interrupt as soon as a shared
150 interrupt handler is registered, and just before one is deregistered. 150 interrupt handler is registered, and just before one is deregistered.
151 Drivers ought to be able to handle interrupts coming in at those 151 Drivers ought to be able to handle interrupts coming in at those
152 points; some don't and need to be caught. 152 points; some don't and need to be caught.
153 153
154 config LOCKUP_DETECTOR 154 config LOCKUP_DETECTOR
155 bool "Detect Hard and Soft Lockups" 155 bool "Detect Hard and Soft Lockups"
156 depends on DEBUG_KERNEL && !S390 156 depends on DEBUG_KERNEL && !S390
157 help 157 help
158 Say Y here to enable the kernel to act as a watchdog to detect 158 Say Y here to enable the kernel to act as a watchdog to detect
159 hard and soft lockups. 159 hard and soft lockups.
160 160
161 Softlockups are bugs that cause the kernel to loop in kernel 161 Softlockups are bugs that cause the kernel to loop in kernel
162 mode for more than 60 seconds, without giving other tasks a 162 mode for more than 60 seconds, without giving other tasks a
163 chance to run. The current stack trace is displayed upon 163 chance to run. The current stack trace is displayed upon
164 detection and the system will stay locked up. 164 detection and the system will stay locked up.
165 165
166 Hardlockups are bugs that cause the CPU to loop in kernel mode 166 Hardlockups are bugs that cause the CPU to loop in kernel mode
167 for more than 60 seconds, without letting other interrupts have a 167 for more than 60 seconds, without letting other interrupts have a
168 chance to run. The current stack trace is displayed upon detection 168 chance to run. The current stack trace is displayed upon detection
169 and the system will stay locked up. 169 and the system will stay locked up.
170 170
171 The overhead should be minimal. A periodic hrtimer runs to 171 The overhead should be minimal. A periodic hrtimer runs to
172 generate interrupts and kick the watchdog task every 10-12 seconds. 172 generate interrupts and kick the watchdog task every 10-12 seconds.
173 An NMI is generated every 60 seconds or so to check for hardlockups. 173 An NMI is generated every 60 seconds or so to check for hardlockups.
174 174
175 config HARDLOCKUP_DETECTOR 175 config HARDLOCKUP_DETECTOR
176 def_bool LOCKUP_DETECTOR && PERF_EVENTS && HAVE_PERF_EVENTS_NMI 176 def_bool LOCKUP_DETECTOR && PERF_EVENTS && HAVE_PERF_EVENTS_NMI
177 177
178 config BOOTPARAM_SOFTLOCKUP_PANIC 178 config BOOTPARAM_SOFTLOCKUP_PANIC
179 bool "Panic (Reboot) On Soft Lockups" 179 bool "Panic (Reboot) On Soft Lockups"
180 depends on LOCKUP_DETECTOR 180 depends on LOCKUP_DETECTOR
181 help 181 help
182 Say Y here to enable the kernel to panic on "soft lockups", 182 Say Y here to enable the kernel to panic on "soft lockups",
183 which are bugs that cause the kernel to loop in kernel 183 which are bugs that cause the kernel to loop in kernel
184 mode for more than 60 seconds, without giving other tasks a 184 mode for more than 60 seconds, without giving other tasks a
185 chance to run. 185 chance to run.
186 186
187 The panic can be used in combination with panic_timeout, 187 The panic can be used in combination with panic_timeout,
188 to cause the system to reboot automatically after a 188 to cause the system to reboot automatically after a
189 lockup has been detected. This feature is useful for 189 lockup has been detected. This feature is useful for
190 high-availability systems that have uptime guarantees and 190 high-availability systems that have uptime guarantees and
191 where a lockup must be resolved ASAP. 191 where a lockup must be resolved ASAP.
192 192
193 Say N if unsure. 193 Say N if unsure.
194 194
195 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE 195 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
196 int 196 int
197 depends on LOCKUP_DETECTOR 197 depends on LOCKUP_DETECTOR
198 range 0 1 198 range 0 1
199 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC 199 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
200 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC 200 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
201 201
202 config DETECT_HUNG_TASK 202 config DETECT_HUNG_TASK
203 bool "Detect Hung Tasks" 203 bool "Detect Hung Tasks"
204 depends on DEBUG_KERNEL 204 depends on DEBUG_KERNEL
205 default DETECT_SOFTLOCKUP 205 default DETECT_SOFTLOCKUP
206 help 206 help
207 Say Y here to enable the kernel to detect "hung tasks", 207 Say Y here to enable the kernel to detect "hung tasks",
208 which are bugs that cause the task to be stuck in 208 which are bugs that cause the task to be stuck in
209 uninterruptible "D" state indefinitiley. 209 uninterruptible "D" state indefinitiley.
210 210
211 When a hung task is detected, the kernel will print the 211 When a hung task is detected, the kernel will print the
212 current stack trace (which you should report), but the 212 current stack trace (which you should report), but the
213 task will stay in uninterruptible state. If lockdep is 213 task will stay in uninterruptible state. If lockdep is
214 enabled then all held locks will also be reported. This 214 enabled then all held locks will also be reported. This
215 feature has negligible overhead. 215 feature has negligible overhead.
216 216
217 config BOOTPARAM_HUNG_TASK_PANIC 217 config BOOTPARAM_HUNG_TASK_PANIC
218 bool "Panic (Reboot) On Hung Tasks" 218 bool "Panic (Reboot) On Hung Tasks"
219 depends on DETECT_HUNG_TASK 219 depends on DETECT_HUNG_TASK
220 help 220 help
221 Say Y here to enable the kernel to panic on "hung tasks", 221 Say Y here to enable the kernel to panic on "hung tasks",
222 which are bugs that cause the kernel to leave a task stuck 222 which are bugs that cause the kernel to leave a task stuck
223 in uninterruptible "D" state. 223 in uninterruptible "D" state.
224 224
225 The panic can be used in combination with panic_timeout, 225 The panic can be used in combination with panic_timeout,
226 to cause the system to reboot automatically after a 226 to cause the system to reboot automatically after a
227 hung task has been detected. This feature is useful for 227 hung task has been detected. This feature is useful for
228 high-availability systems that have uptime guarantees and 228 high-availability systems that have uptime guarantees and
229 where a hung tasks must be resolved ASAP. 229 where a hung tasks must be resolved ASAP.
230 230
231 Say N if unsure. 231 Say N if unsure.
232 232
233 config BOOTPARAM_HUNG_TASK_PANIC_VALUE 233 config BOOTPARAM_HUNG_TASK_PANIC_VALUE
234 int 234 int
235 depends on DETECT_HUNG_TASK 235 depends on DETECT_HUNG_TASK
236 range 0 1 236 range 0 1
237 default 0 if !BOOTPARAM_HUNG_TASK_PANIC 237 default 0 if !BOOTPARAM_HUNG_TASK_PANIC
238 default 1 if BOOTPARAM_HUNG_TASK_PANIC 238 default 1 if BOOTPARAM_HUNG_TASK_PANIC
239 239
240 config SCHED_DEBUG 240 config SCHED_DEBUG
241 bool "Collect scheduler debugging info" 241 bool "Collect scheduler debugging info"
242 depends on DEBUG_KERNEL && PROC_FS 242 depends on DEBUG_KERNEL && PROC_FS
243 default y 243 default y
244 help 244 help
245 If you say Y here, the /proc/sched_debug file will be provided 245 If you say Y here, the /proc/sched_debug file will be provided
246 that can help debug the scheduler. The runtime overhead of this 246 that can help debug the scheduler. The runtime overhead of this
247 option is minimal. 247 option is minimal.
248 248
249 config SCHEDSTATS 249 config SCHEDSTATS
250 bool "Collect scheduler statistics" 250 bool "Collect scheduler statistics"
251 depends on DEBUG_KERNEL && PROC_FS 251 depends on DEBUG_KERNEL && PROC_FS
252 help 252 help
253 If you say Y here, additional code will be inserted into the 253 If you say Y here, additional code will be inserted into the
254 scheduler and related routines to collect statistics about 254 scheduler and related routines to collect statistics about
255 scheduler behavior and provide them in /proc/schedstat. These 255 scheduler behavior and provide them in /proc/schedstat. These
256 stats may be useful for both tuning and debugging the scheduler 256 stats may be useful for both tuning and debugging the scheduler
257 If you aren't debugging the scheduler or trying to tune a specific 257 If you aren't debugging the scheduler or trying to tune a specific
258 application, you can say N to avoid the very slight overhead 258 application, you can say N to avoid the very slight overhead
259 this adds. 259 this adds.
260 260
261 config TIMER_STATS 261 config TIMER_STATS
262 bool "Collect kernel timers statistics" 262 bool "Collect kernel timers statistics"
263 depends on DEBUG_KERNEL && PROC_FS 263 depends on DEBUG_KERNEL && PROC_FS
264 help 264 help
265 If you say Y here, additional code will be inserted into the 265 If you say Y here, additional code will be inserted into the
266 timer routines to collect statistics about kernel timers being 266 timer routines to collect statistics about kernel timers being
267 reprogrammed. The statistics can be read from /proc/timer_stats. 267 reprogrammed. The statistics can be read from /proc/timer_stats.
268 The statistics collection is started by writing 1 to /proc/timer_stats, 268 The statistics collection is started by writing 1 to /proc/timer_stats,
269 writing 0 stops it. This feature is useful to collect information 269 writing 0 stops it. This feature is useful to collect information
270 about timer usage patterns in kernel and userspace. This feature 270 about timer usage patterns in kernel and userspace. This feature
271 is lightweight if enabled in the kernel config but not activated 271 is lightweight if enabled in the kernel config but not activated
272 (it defaults to deactivated on bootup and will only be activated 272 (it defaults to deactivated on bootup and will only be activated
273 if some application like powertop activates it explicitly). 273 if some application like powertop activates it explicitly).
274 274
275 config DEBUG_OBJECTS 275 config DEBUG_OBJECTS
276 bool "Debug object operations" 276 bool "Debug object operations"
277 depends on DEBUG_KERNEL 277 depends on DEBUG_KERNEL
278 help 278 help
279 If you say Y here, additional code will be inserted into the 279 If you say Y here, additional code will be inserted into the
280 kernel to track the life time of various objects and validate 280 kernel to track the life time of various objects and validate
281 the operations on those objects. 281 the operations on those objects.
282 282
283 config DEBUG_OBJECTS_SELFTEST 283 config DEBUG_OBJECTS_SELFTEST
284 bool "Debug objects selftest" 284 bool "Debug objects selftest"
285 depends on DEBUG_OBJECTS 285 depends on DEBUG_OBJECTS
286 help 286 help
287 This enables the selftest of the object debug code. 287 This enables the selftest of the object debug code.
288 288
289 config DEBUG_OBJECTS_FREE 289 config DEBUG_OBJECTS_FREE
290 bool "Debug objects in freed memory" 290 bool "Debug objects in freed memory"
291 depends on DEBUG_OBJECTS 291 depends on DEBUG_OBJECTS
292 help 292 help
293 This enables checks whether a k/v free operation frees an area 293 This enables checks whether a k/v free operation frees an area
294 which contains an object which has not been deactivated 294 which contains an object which has not been deactivated
295 properly. This can make kmalloc/kfree-intensive workloads 295 properly. This can make kmalloc/kfree-intensive workloads
296 much slower. 296 much slower.
297 297
298 config DEBUG_OBJECTS_TIMERS 298 config DEBUG_OBJECTS_TIMERS
299 bool "Debug timer objects" 299 bool "Debug timer objects"
300 depends on DEBUG_OBJECTS 300 depends on DEBUG_OBJECTS
301 help 301 help
302 If you say Y here, additional code will be inserted into the 302 If you say Y here, additional code will be inserted into the
303 timer routines to track the life time of timer objects and 303 timer routines to track the life time of timer objects and
304 validate the timer operations. 304 validate the timer operations.
305 305
306 config DEBUG_OBJECTS_WORK 306 config DEBUG_OBJECTS_WORK
307 bool "Debug work objects" 307 bool "Debug work objects"
308 depends on DEBUG_OBJECTS 308 depends on DEBUG_OBJECTS
309 help 309 help
310 If you say Y here, additional code will be inserted into the 310 If you say Y here, additional code will be inserted into the
311 work queue routines to track the life time of work objects and 311 work queue routines to track the life time of work objects and
312 validate the work operations. 312 validate the work operations.
313 313
314 config DEBUG_OBJECTS_RCU_HEAD 314 config DEBUG_OBJECTS_RCU_HEAD
315 bool "Debug RCU callbacks objects" 315 bool "Debug RCU callbacks objects"
316 depends on DEBUG_OBJECTS && PREEMPT 316 depends on DEBUG_OBJECTS && PREEMPT
317 help 317 help
318 Enable this to turn on debugging of RCU list heads (call_rcu() usage). 318 Enable this to turn on debugging of RCU list heads (call_rcu() usage).
319 319
320 config DEBUG_OBJECTS_PERCPU_COUNTER
321 bool "Debug percpu counter objects"
322 depends on DEBUG_OBJECTS
323 help
324 If you say Y here, additional code will be inserted into the
325 percpu counter routines to track the life time of percpu counter
326 objects and validate the percpu counter operations.
327
320 config DEBUG_OBJECTS_ENABLE_DEFAULT 328 config DEBUG_OBJECTS_ENABLE_DEFAULT
321 int "debug_objects bootup default value (0-1)" 329 int "debug_objects bootup default value (0-1)"
322 range 0 1 330 range 0 1
323 default "1" 331 default "1"
324 depends on DEBUG_OBJECTS 332 depends on DEBUG_OBJECTS
325 help 333 help
326 Debug objects boot parameter default value 334 Debug objects boot parameter default value
327 335
328 config DEBUG_SLAB 336 config DEBUG_SLAB
329 bool "Debug slab memory allocations" 337 bool "Debug slab memory allocations"
330 depends on DEBUG_KERNEL && SLAB && !KMEMCHECK 338 depends on DEBUG_KERNEL && SLAB && !KMEMCHECK
331 help 339 help
332 Say Y here to have the kernel do limited verification on memory 340 Say Y here to have the kernel do limited verification on memory
333 allocation as well as poisoning memory on free to catch use of freed 341 allocation as well as poisoning memory on free to catch use of freed
334 memory. This can make kmalloc/kfree-intensive workloads much slower. 342 memory. This can make kmalloc/kfree-intensive workloads much slower.
335 343
336 config DEBUG_SLAB_LEAK 344 config DEBUG_SLAB_LEAK
337 bool "Memory leak debugging" 345 bool "Memory leak debugging"
338 depends on DEBUG_SLAB 346 depends on DEBUG_SLAB
339 347
340 config SLUB_DEBUG_ON 348 config SLUB_DEBUG_ON
341 bool "SLUB debugging on by default" 349 bool "SLUB debugging on by default"
342 depends on SLUB && SLUB_DEBUG && !KMEMCHECK 350 depends on SLUB && SLUB_DEBUG && !KMEMCHECK
343 default n 351 default n
344 help 352 help
345 Boot with debugging on by default. SLUB boots by default with 353 Boot with debugging on by default. SLUB boots by default with
346 the runtime debug capabilities switched off. Enabling this is 354 the runtime debug capabilities switched off. Enabling this is
347 equivalent to specifying the "slub_debug" parameter on boot. 355 equivalent to specifying the "slub_debug" parameter on boot.
348 There is no support for more fine grained debug control like 356 There is no support for more fine grained debug control like
349 possible with slub_debug=xxx. SLUB debugging may be switched 357 possible with slub_debug=xxx. SLUB debugging may be switched
350 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying 358 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
351 "slub_debug=-". 359 "slub_debug=-".
352 360
353 config SLUB_STATS 361 config SLUB_STATS
354 default n 362 default n
355 bool "Enable SLUB performance statistics" 363 bool "Enable SLUB performance statistics"
356 depends on SLUB && SYSFS 364 depends on SLUB && SYSFS
357 help 365 help
358 SLUB statistics are useful to debug SLUBs allocation behavior in 366 SLUB statistics are useful to debug SLUBs allocation behavior in
359 order find ways to optimize the allocator. This should never be 367 order find ways to optimize the allocator. This should never be
360 enabled for production use since keeping statistics slows down 368 enabled for production use since keeping statistics slows down
361 the allocator by a few percentage points. The slabinfo command 369 the allocator by a few percentage points. The slabinfo command
362 supports the determination of the most active slabs to figure 370 supports the determination of the most active slabs to figure
363 out which slabs are relevant to a particular load. 371 out which slabs are relevant to a particular load.
364 Try running: slabinfo -DA 372 Try running: slabinfo -DA
365 373
366 config DEBUG_KMEMLEAK 374 config DEBUG_KMEMLEAK
367 bool "Kernel memory leak detector" 375 bool "Kernel memory leak detector"
368 depends on DEBUG_KERNEL && EXPERIMENTAL && !MEMORY_HOTPLUG && \ 376 depends on DEBUG_KERNEL && EXPERIMENTAL && !MEMORY_HOTPLUG && \
369 (X86 || ARM || PPC || S390 || SPARC64 || SUPERH || MICROBLAZE) 377 (X86 || ARM || PPC || S390 || SPARC64 || SUPERH || MICROBLAZE)
370 378
371 select DEBUG_FS if SYSFS 379 select DEBUG_FS if SYSFS
372 select STACKTRACE if STACKTRACE_SUPPORT 380 select STACKTRACE if STACKTRACE_SUPPORT
373 select KALLSYMS 381 select KALLSYMS
374 select CRC32 382 select CRC32
375 help 383 help
376 Say Y here if you want to enable the memory leak 384 Say Y here if you want to enable the memory leak
377 detector. The memory allocation/freeing is traced in a way 385 detector. The memory allocation/freeing is traced in a way
378 similar to the Boehm's conservative garbage collector, the 386 similar to the Boehm's conservative garbage collector, the
379 difference being that the orphan objects are not freed but 387 difference being that the orphan objects are not freed but
380 only shown in /sys/kernel/debug/kmemleak. Enabling this 388 only shown in /sys/kernel/debug/kmemleak. Enabling this
381 feature will introduce an overhead to memory 389 feature will introduce an overhead to memory
382 allocations. See Documentation/kmemleak.txt for more 390 allocations. See Documentation/kmemleak.txt for more
383 details. 391 details.
384 392
385 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances 393 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
386 of finding leaks due to the slab objects poisoning. 394 of finding leaks due to the slab objects poisoning.
387 395
388 In order to access the kmemleak file, debugfs needs to be 396 In order to access the kmemleak file, debugfs needs to be
389 mounted (usually at /sys/kernel/debug). 397 mounted (usually at /sys/kernel/debug).
390 398
391 config DEBUG_KMEMLEAK_EARLY_LOG_SIZE 399 config DEBUG_KMEMLEAK_EARLY_LOG_SIZE
392 int "Maximum kmemleak early log entries" 400 int "Maximum kmemleak early log entries"
393 depends on DEBUG_KMEMLEAK 401 depends on DEBUG_KMEMLEAK
394 range 200 40000 402 range 200 40000
395 default 400 403 default 400
396 help 404 help
397 Kmemleak must track all the memory allocations to avoid 405 Kmemleak must track all the memory allocations to avoid
398 reporting false positives. Since memory may be allocated or 406 reporting false positives. Since memory may be allocated or
399 freed before kmemleak is initialised, an early log buffer is 407 freed before kmemleak is initialised, an early log buffer is
400 used to store these actions. If kmemleak reports "early log 408 used to store these actions. If kmemleak reports "early log
401 buffer exceeded", please increase this value. 409 buffer exceeded", please increase this value.
402 410
403 config DEBUG_KMEMLEAK_TEST 411 config DEBUG_KMEMLEAK_TEST
404 tristate "Simple test for the kernel memory leak detector" 412 tristate "Simple test for the kernel memory leak detector"
405 depends on DEBUG_KMEMLEAK 413 depends on DEBUG_KMEMLEAK
406 help 414 help
407 Say Y or M here to build a test for the kernel memory leak 415 Say Y or M here to build a test for the kernel memory leak
408 detector. This option enables a module that explicitly leaks 416 detector. This option enables a module that explicitly leaks
409 memory. 417 memory.
410 418
411 If unsure, say N. 419 If unsure, say N.
412 420
413 config DEBUG_KMEMLEAK_DEFAULT_OFF 421 config DEBUG_KMEMLEAK_DEFAULT_OFF
414 bool "Default kmemleak to off" 422 bool "Default kmemleak to off"
415 depends on DEBUG_KMEMLEAK 423 depends on DEBUG_KMEMLEAK
416 help 424 help
417 Say Y here to disable kmemleak by default. It can then be enabled 425 Say Y here to disable kmemleak by default. It can then be enabled
418 on the command line via kmemleak=on. 426 on the command line via kmemleak=on.
419 427
420 config DEBUG_PREEMPT 428 config DEBUG_PREEMPT
421 bool "Debug preemptible kernel" 429 bool "Debug preemptible kernel"
422 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT 430 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
423 default y 431 default y
424 help 432 help
425 If you say Y here then the kernel will use a debug variant of the 433 If you say Y here then the kernel will use a debug variant of the
426 commonly used smp_processor_id() function and will print warnings 434 commonly used smp_processor_id() function and will print warnings
427 if kernel code uses it in a preemption-unsafe way. Also, the kernel 435 if kernel code uses it in a preemption-unsafe way. Also, the kernel
428 will detect preemption count underflows. 436 will detect preemption count underflows.
429 437
430 config DEBUG_RT_MUTEXES 438 config DEBUG_RT_MUTEXES
431 bool "RT Mutex debugging, deadlock detection" 439 bool "RT Mutex debugging, deadlock detection"
432 depends on DEBUG_KERNEL && RT_MUTEXES 440 depends on DEBUG_KERNEL && RT_MUTEXES
433 help 441 help
434 This allows rt mutex semantics violations and rt mutex related 442 This allows rt mutex semantics violations and rt mutex related
435 deadlocks (lockups) to be detected and reported automatically. 443 deadlocks (lockups) to be detected and reported automatically.
436 444
437 config DEBUG_PI_LIST 445 config DEBUG_PI_LIST
438 bool 446 bool
439 default y 447 default y
440 depends on DEBUG_RT_MUTEXES 448 depends on DEBUG_RT_MUTEXES
441 449
442 config RT_MUTEX_TESTER 450 config RT_MUTEX_TESTER
443 bool "Built-in scriptable tester for rt-mutexes" 451 bool "Built-in scriptable tester for rt-mutexes"
444 depends on DEBUG_KERNEL && RT_MUTEXES 452 depends on DEBUG_KERNEL && RT_MUTEXES
445 help 453 help
446 This option enables a rt-mutex tester. 454 This option enables a rt-mutex tester.
447 455
448 config DEBUG_SPINLOCK 456 config DEBUG_SPINLOCK
449 bool "Spinlock and rw-lock debugging: basic checks" 457 bool "Spinlock and rw-lock debugging: basic checks"
450 depends on DEBUG_KERNEL 458 depends on DEBUG_KERNEL
451 help 459 help
452 Say Y here and build SMP to catch missing spinlock initialization 460 Say Y here and build SMP to catch missing spinlock initialization
453 and certain other kinds of spinlock errors commonly made. This is 461 and certain other kinds of spinlock errors commonly made. This is
454 best used in conjunction with the NMI watchdog so that spinlock 462 best used in conjunction with the NMI watchdog so that spinlock
455 deadlocks are also debuggable. 463 deadlocks are also debuggable.
456 464
457 config DEBUG_MUTEXES 465 config DEBUG_MUTEXES
458 bool "Mutex debugging: basic checks" 466 bool "Mutex debugging: basic checks"
459 depends on DEBUG_KERNEL 467 depends on DEBUG_KERNEL
460 help 468 help
461 This feature allows mutex semantics violations to be detected and 469 This feature allows mutex semantics violations to be detected and
462 reported. 470 reported.
463 471
464 config BKL 472 config BKL
465 bool "Big Kernel Lock" if (SMP || PREEMPT) 473 bool "Big Kernel Lock" if (SMP || PREEMPT)
466 default y 474 default y
467 help 475 help
468 This is the traditional lock that is used in old code instead 476 This is the traditional lock that is used in old code instead
469 of proper locking. All drivers that use the BKL should depend 477 of proper locking. All drivers that use the BKL should depend
470 on this symbol. 478 on this symbol.
471 Say Y here unless you are working on removing the BKL. 479 Say Y here unless you are working on removing the BKL.
472 480
473 config DEBUG_LOCK_ALLOC 481 config DEBUG_LOCK_ALLOC
474 bool "Lock debugging: detect incorrect freeing of live locks" 482 bool "Lock debugging: detect incorrect freeing of live locks"
475 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 483 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
476 select DEBUG_SPINLOCK 484 select DEBUG_SPINLOCK
477 select DEBUG_MUTEXES 485 select DEBUG_MUTEXES
478 select LOCKDEP 486 select LOCKDEP
479 help 487 help
480 This feature will check whether any held lock (spinlock, rwlock, 488 This feature will check whether any held lock (spinlock, rwlock,
481 mutex or rwsem) is incorrectly freed by the kernel, via any of the 489 mutex or rwsem) is incorrectly freed by the kernel, via any of the
482 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(), 490 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
483 vfree(), etc.), whether a live lock is incorrectly reinitialized via 491 vfree(), etc.), whether a live lock is incorrectly reinitialized via
484 spin_lock_init()/mutex_init()/etc., or whether there is any lock 492 spin_lock_init()/mutex_init()/etc., or whether there is any lock
485 held during task exit. 493 held during task exit.
486 494
487 config PROVE_LOCKING 495 config PROVE_LOCKING
488 bool "Lock debugging: prove locking correctness" 496 bool "Lock debugging: prove locking correctness"
489 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 497 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
490 select LOCKDEP 498 select LOCKDEP
491 select DEBUG_SPINLOCK 499 select DEBUG_SPINLOCK
492 select DEBUG_MUTEXES 500 select DEBUG_MUTEXES
493 select DEBUG_LOCK_ALLOC 501 select DEBUG_LOCK_ALLOC
494 select TRACE_IRQFLAGS 502 select TRACE_IRQFLAGS
495 default n 503 default n
496 help 504 help
497 This feature enables the kernel to prove that all locking 505 This feature enables the kernel to prove that all locking
498 that occurs in the kernel runtime is mathematically 506 that occurs in the kernel runtime is mathematically
499 correct: that under no circumstance could an arbitrary (and 507 correct: that under no circumstance could an arbitrary (and
500 not yet triggered) combination of observed locking 508 not yet triggered) combination of observed locking
501 sequences (on an arbitrary number of CPUs, running an 509 sequences (on an arbitrary number of CPUs, running an
502 arbitrary number of tasks and interrupt contexts) cause a 510 arbitrary number of tasks and interrupt contexts) cause a
503 deadlock. 511 deadlock.
504 512
505 In short, this feature enables the kernel to report locking 513 In short, this feature enables the kernel to report locking
506 related deadlocks before they actually occur. 514 related deadlocks before they actually occur.
507 515
508 The proof does not depend on how hard and complex a 516 The proof does not depend on how hard and complex a
509 deadlock scenario would be to trigger: how many 517 deadlock scenario would be to trigger: how many
510 participant CPUs, tasks and irq-contexts would be needed 518 participant CPUs, tasks and irq-contexts would be needed
511 for it to trigger. The proof also does not depend on 519 for it to trigger. The proof also does not depend on
512 timing: if a race and a resulting deadlock is possible 520 timing: if a race and a resulting deadlock is possible
513 theoretically (no matter how unlikely the race scenario 521 theoretically (no matter how unlikely the race scenario
514 is), it will be proven so and will immediately be 522 is), it will be proven so and will immediately be
515 reported by the kernel (once the event is observed that 523 reported by the kernel (once the event is observed that
516 makes the deadlock theoretically possible). 524 makes the deadlock theoretically possible).
517 525
518 If a deadlock is impossible (i.e. the locking rules, as 526 If a deadlock is impossible (i.e. the locking rules, as
519 observed by the kernel, are mathematically correct), the 527 observed by the kernel, are mathematically correct), the
520 kernel reports nothing. 528 kernel reports nothing.
521 529
522 NOTE: this feature can also be enabled for rwlocks, mutexes 530 NOTE: this feature can also be enabled for rwlocks, mutexes
523 and rwsems - in which case all dependencies between these 531 and rwsems - in which case all dependencies between these
524 different locking variants are observed and mapped too, and 532 different locking variants are observed and mapped too, and
525 the proof of observed correctness is also maintained for an 533 the proof of observed correctness is also maintained for an
526 arbitrary combination of these separate locking variants. 534 arbitrary combination of these separate locking variants.
527 535
528 For more details, see Documentation/lockdep-design.txt. 536 For more details, see Documentation/lockdep-design.txt.
529 537
530 config PROVE_RCU 538 config PROVE_RCU
531 bool "RCU debugging: prove RCU correctness" 539 bool "RCU debugging: prove RCU correctness"
532 depends on PROVE_LOCKING 540 depends on PROVE_LOCKING
533 default n 541 default n
534 help 542 help
535 This feature enables lockdep extensions that check for correct 543 This feature enables lockdep extensions that check for correct
536 use of RCU APIs. This is currently under development. Say Y 544 use of RCU APIs. This is currently under development. Say Y
537 if you want to debug RCU usage or help work on the PROVE_RCU 545 if you want to debug RCU usage or help work on the PROVE_RCU
538 feature. 546 feature.
539 547
540 Say N if you are unsure. 548 Say N if you are unsure.
541 549
542 config PROVE_RCU_REPEATEDLY 550 config PROVE_RCU_REPEATEDLY
543 bool "RCU debugging: don't disable PROVE_RCU on first splat" 551 bool "RCU debugging: don't disable PROVE_RCU on first splat"
544 depends on PROVE_RCU 552 depends on PROVE_RCU
545 default n 553 default n
546 help 554 help
547 By itself, PROVE_RCU will disable checking upon issuing the 555 By itself, PROVE_RCU will disable checking upon issuing the
548 first warning (or "splat"). This feature prevents such 556 first warning (or "splat"). This feature prevents such
549 disabling, allowing multiple RCU-lockdep warnings to be printed 557 disabling, allowing multiple RCU-lockdep warnings to be printed
550 on a single reboot. 558 on a single reboot.
551 559
552 Say Y to allow multiple RCU-lockdep warnings per boot. 560 Say Y to allow multiple RCU-lockdep warnings per boot.
553 561
554 Say N if you are unsure. 562 Say N if you are unsure.
555 563
556 config SPARSE_RCU_POINTER 564 config SPARSE_RCU_POINTER
557 bool "RCU debugging: sparse-based checks for pointer usage" 565 bool "RCU debugging: sparse-based checks for pointer usage"
558 default n 566 default n
559 help 567 help
560 This feature enables the __rcu sparse annotation for 568 This feature enables the __rcu sparse annotation for
561 RCU-protected pointers. This annotation will cause sparse 569 RCU-protected pointers. This annotation will cause sparse
562 to flag any non-RCU used of annotated pointers. This can be 570 to flag any non-RCU used of annotated pointers. This can be
563 helpful when debugging RCU usage. Please note that this feature 571 helpful when debugging RCU usage. Please note that this feature
564 is not intended to enforce code cleanliness; it is instead merely 572 is not intended to enforce code cleanliness; it is instead merely
565 a debugging aid. 573 a debugging aid.
566 574
567 Say Y to make sparse flag questionable use of RCU-protected pointers 575 Say Y to make sparse flag questionable use of RCU-protected pointers
568 576
569 Say N if you are unsure. 577 Say N if you are unsure.
570 578
571 config LOCKDEP 579 config LOCKDEP
572 bool 580 bool
573 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 581 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
574 select STACKTRACE 582 select STACKTRACE
575 select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE 583 select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE
576 select KALLSYMS 584 select KALLSYMS
577 select KALLSYMS_ALL 585 select KALLSYMS_ALL
578 586
579 config LOCK_STAT 587 config LOCK_STAT
580 bool "Lock usage statistics" 588 bool "Lock usage statistics"
581 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 589 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
582 select LOCKDEP 590 select LOCKDEP
583 select DEBUG_SPINLOCK 591 select DEBUG_SPINLOCK
584 select DEBUG_MUTEXES 592 select DEBUG_MUTEXES
585 select DEBUG_LOCK_ALLOC 593 select DEBUG_LOCK_ALLOC
586 default n 594 default n
587 help 595 help
588 This feature enables tracking lock contention points 596 This feature enables tracking lock contention points
589 597
590 For more details, see Documentation/lockstat.txt 598 For more details, see Documentation/lockstat.txt
591 599
592 This also enables lock events required by "perf lock", 600 This also enables lock events required by "perf lock",
593 subcommand of perf. 601 subcommand of perf.
594 If you want to use "perf lock", you also need to turn on 602 If you want to use "perf lock", you also need to turn on
595 CONFIG_EVENT_TRACING. 603 CONFIG_EVENT_TRACING.
596 604
597 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events. 605 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
598 (CONFIG_LOCKDEP defines "acquire" and "release" events.) 606 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
599 607
600 config DEBUG_LOCKDEP 608 config DEBUG_LOCKDEP
601 bool "Lock dependency engine debugging" 609 bool "Lock dependency engine debugging"
602 depends on DEBUG_KERNEL && LOCKDEP 610 depends on DEBUG_KERNEL && LOCKDEP
603 help 611 help
604 If you say Y here, the lock dependency engine will do 612 If you say Y here, the lock dependency engine will do
605 additional runtime checks to debug itself, at the price 613 additional runtime checks to debug itself, at the price
606 of more runtime overhead. 614 of more runtime overhead.
607 615
608 config TRACE_IRQFLAGS 616 config TRACE_IRQFLAGS
609 bool 617 bool
610 help 618 help
611 Enables hooks to interrupt enabling and disabling for 619 Enables hooks to interrupt enabling and disabling for
612 either tracing or lock debugging. 620 either tracing or lock debugging.
613 621
614 config DEBUG_SPINLOCK_SLEEP 622 config DEBUG_SPINLOCK_SLEEP
615 bool "Spinlock debugging: sleep-inside-spinlock checking" 623 bool "Spinlock debugging: sleep-inside-spinlock checking"
616 depends on DEBUG_KERNEL 624 depends on DEBUG_KERNEL
617 help 625 help
618 If you say Y here, various routines which may sleep will become very 626 If you say Y here, various routines which may sleep will become very
619 noisy if they are called with a spinlock held. 627 noisy if they are called with a spinlock held.
620 628
621 config DEBUG_LOCKING_API_SELFTESTS 629 config DEBUG_LOCKING_API_SELFTESTS
622 bool "Locking API boot-time self-tests" 630 bool "Locking API boot-time self-tests"
623 depends on DEBUG_KERNEL 631 depends on DEBUG_KERNEL
624 help 632 help
625 Say Y here if you want the kernel to run a short self-test during 633 Say Y here if you want the kernel to run a short self-test during
626 bootup. The self-test checks whether common types of locking bugs 634 bootup. The self-test checks whether common types of locking bugs
627 are detected by debugging mechanisms or not. (if you disable 635 are detected by debugging mechanisms or not. (if you disable
628 lock debugging then those bugs wont be detected of course.) 636 lock debugging then those bugs wont be detected of course.)
629 The following locking APIs are covered: spinlocks, rwlocks, 637 The following locking APIs are covered: spinlocks, rwlocks,
630 mutexes and rwsems. 638 mutexes and rwsems.
631 639
632 config STACKTRACE 640 config STACKTRACE
633 bool 641 bool
634 depends on STACKTRACE_SUPPORT 642 depends on STACKTRACE_SUPPORT
635 643
636 config DEBUG_KOBJECT 644 config DEBUG_KOBJECT
637 bool "kobject debugging" 645 bool "kobject debugging"
638 depends on DEBUG_KERNEL 646 depends on DEBUG_KERNEL
639 help 647 help
640 If you say Y here, some extra kobject debugging messages will be sent 648 If you say Y here, some extra kobject debugging messages will be sent
641 to the syslog. 649 to the syslog.
642 650
643 config DEBUG_HIGHMEM 651 config DEBUG_HIGHMEM
644 bool "Highmem debugging" 652 bool "Highmem debugging"
645 depends on DEBUG_KERNEL && HIGHMEM 653 depends on DEBUG_KERNEL && HIGHMEM
646 help 654 help
647 This options enables addition error checking for high memory systems. 655 This options enables addition error checking for high memory systems.
648 Disable for production systems. 656 Disable for production systems.
649 657
650 config DEBUG_BUGVERBOSE 658 config DEBUG_BUGVERBOSE
651 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EMBEDDED 659 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EMBEDDED
652 depends on BUG 660 depends on BUG
653 depends on ARM || AVR32 || M32R || M68K || SPARC32 || SPARC64 || \ 661 depends on ARM || AVR32 || M32R || M68K || SPARC32 || SPARC64 || \
654 FRV || SUPERH || GENERIC_BUG || BLACKFIN || MN10300 662 FRV || SUPERH || GENERIC_BUG || BLACKFIN || MN10300
655 default y 663 default y
656 help 664 help
657 Say Y here to make BUG() panics output the file name and line number 665 Say Y here to make BUG() panics output the file name and line number
658 of the BUG call as well as the EIP and oops trace. This aids 666 of the BUG call as well as the EIP and oops trace. This aids
659 debugging but costs about 70-100K of memory. 667 debugging but costs about 70-100K of memory.
660 668
661 config DEBUG_INFO 669 config DEBUG_INFO
662 bool "Compile the kernel with debug info" 670 bool "Compile the kernel with debug info"
663 depends on DEBUG_KERNEL 671 depends on DEBUG_KERNEL
664 help 672 help
665 If you say Y here the resulting kernel image will include 673 If you say Y here the resulting kernel image will include
666 debugging info resulting in a larger kernel image. 674 debugging info resulting in a larger kernel image.
667 This adds debug symbols to the kernel and modules (gcc -g), and 675 This adds debug symbols to the kernel and modules (gcc -g), and
668 is needed if you intend to use kernel crashdump or binary object 676 is needed if you intend to use kernel crashdump or binary object
669 tools like crash, kgdb, LKCD, gdb, etc on the kernel. 677 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
670 Say Y here only if you plan to debug the kernel. 678 Say Y here only if you plan to debug the kernel.
671 679
672 If unsure, say N. 680 If unsure, say N.
673 681
674 config DEBUG_INFO_REDUCED 682 config DEBUG_INFO_REDUCED
675 bool "Reduce debugging information" 683 bool "Reduce debugging information"
676 depends on DEBUG_INFO 684 depends on DEBUG_INFO
677 help 685 help
678 If you say Y here gcc is instructed to generate less debugging 686 If you say Y here gcc is instructed to generate less debugging
679 information for structure types. This means that tools that 687 information for structure types. This means that tools that
680 need full debugging information (like kgdb or systemtap) won't 688 need full debugging information (like kgdb or systemtap) won't
681 be happy. But if you merely need debugging information to 689 be happy. But if you merely need debugging information to
682 resolve line numbers there is no loss. Advantage is that 690 resolve line numbers there is no loss. Advantage is that
683 build directory object sizes shrink dramatically over a full 691 build directory object sizes shrink dramatically over a full
684 DEBUG_INFO build and compile times are reduced too. 692 DEBUG_INFO build and compile times are reduced too.
685 Only works with newer gcc versions. 693 Only works with newer gcc versions.
686 694
687 config DEBUG_VM 695 config DEBUG_VM
688 bool "Debug VM" 696 bool "Debug VM"
689 depends on DEBUG_KERNEL 697 depends on DEBUG_KERNEL
690 help 698 help
691 Enable this to turn on extended checks in the virtual-memory system 699 Enable this to turn on extended checks in the virtual-memory system
692 that may impact performance. 700 that may impact performance.
693 701
694 If unsure, say N. 702 If unsure, say N.
695 703
696 config DEBUG_VIRTUAL 704 config DEBUG_VIRTUAL
697 bool "Debug VM translations" 705 bool "Debug VM translations"
698 depends on DEBUG_KERNEL && X86 706 depends on DEBUG_KERNEL && X86
699 help 707 help
700 Enable some costly sanity checks in virtual to page code. This can 708 Enable some costly sanity checks in virtual to page code. This can
701 catch mistakes with virt_to_page() and friends. 709 catch mistakes with virt_to_page() and friends.
702 710
703 If unsure, say N. 711 If unsure, say N.
704 712
705 config DEBUG_NOMMU_REGIONS 713 config DEBUG_NOMMU_REGIONS
706 bool "Debug the global anon/private NOMMU mapping region tree" 714 bool "Debug the global anon/private NOMMU mapping region tree"
707 depends on DEBUG_KERNEL && !MMU 715 depends on DEBUG_KERNEL && !MMU
708 help 716 help
709 This option causes the global tree of anonymous and private mapping 717 This option causes the global tree of anonymous and private mapping
710 regions to be regularly checked for invalid topology. 718 regions to be regularly checked for invalid topology.
711 719
712 config DEBUG_WRITECOUNT 720 config DEBUG_WRITECOUNT
713 bool "Debug filesystem writers count" 721 bool "Debug filesystem writers count"
714 depends on DEBUG_KERNEL 722 depends on DEBUG_KERNEL
715 help 723 help
716 Enable this to catch wrong use of the writers count in struct 724 Enable this to catch wrong use of the writers count in struct
717 vfsmount. This will increase the size of each file struct by 725 vfsmount. This will increase the size of each file struct by
718 32 bits. 726 32 bits.
719 727
720 If unsure, say N. 728 If unsure, say N.
721 729
722 config DEBUG_MEMORY_INIT 730 config DEBUG_MEMORY_INIT
723 bool "Debug memory initialisation" if EMBEDDED 731 bool "Debug memory initialisation" if EMBEDDED
724 default !EMBEDDED 732 default !EMBEDDED
725 help 733 help
726 Enable this for additional checks during memory initialisation. 734 Enable this for additional checks during memory initialisation.
727 The sanity checks verify aspects of the VM such as the memory model 735 The sanity checks verify aspects of the VM such as the memory model
728 and other information provided by the architecture. Verbose 736 and other information provided by the architecture. Verbose
729 information will be printed at KERN_DEBUG loglevel depending 737 information will be printed at KERN_DEBUG loglevel depending
730 on the mminit_loglevel= command-line option. 738 on the mminit_loglevel= command-line option.
731 739
732 If unsure, say Y 740 If unsure, say Y
733 741
734 config DEBUG_LIST 742 config DEBUG_LIST
735 bool "Debug linked list manipulation" 743 bool "Debug linked list manipulation"
736 depends on DEBUG_KERNEL 744 depends on DEBUG_KERNEL
737 help 745 help
738 Enable this to turn on extended checks in the linked-list 746 Enable this to turn on extended checks in the linked-list
739 walking routines. 747 walking routines.
740 748
741 If unsure, say N. 749 If unsure, say N.
742 750
743 config DEBUG_SG 751 config DEBUG_SG
744 bool "Debug SG table operations" 752 bool "Debug SG table operations"
745 depends on DEBUG_KERNEL 753 depends on DEBUG_KERNEL
746 help 754 help
747 Enable this to turn on checks on scatter-gather tables. This can 755 Enable this to turn on checks on scatter-gather tables. This can
748 help find problems with drivers that do not properly initialize 756 help find problems with drivers that do not properly initialize
749 their sg tables. 757 their sg tables.
750 758
751 If unsure, say N. 759 If unsure, say N.
752 760
753 config DEBUG_NOTIFIERS 761 config DEBUG_NOTIFIERS
754 bool "Debug notifier call chains" 762 bool "Debug notifier call chains"
755 depends on DEBUG_KERNEL 763 depends on DEBUG_KERNEL
756 help 764 help
757 Enable this to turn on sanity checking for notifier call chains. 765 Enable this to turn on sanity checking for notifier call chains.
758 This is most useful for kernel developers to make sure that 766 This is most useful for kernel developers to make sure that
759 modules properly unregister themselves from notifier chains. 767 modules properly unregister themselves from notifier chains.
760 This is a relatively cheap check but if you care about maximum 768 This is a relatively cheap check but if you care about maximum
761 performance, say N. 769 performance, say N.
762 770
763 config DEBUG_CREDENTIALS 771 config DEBUG_CREDENTIALS
764 bool "Debug credential management" 772 bool "Debug credential management"
765 depends on DEBUG_KERNEL 773 depends on DEBUG_KERNEL
766 help 774 help
767 Enable this to turn on some debug checking for credential 775 Enable this to turn on some debug checking for credential
768 management. The additional code keeps track of the number of 776 management. The additional code keeps track of the number of
769 pointers from task_structs to any given cred struct, and checks to 777 pointers from task_structs to any given cred struct, and checks to
770 see that this number never exceeds the usage count of the cred 778 see that this number never exceeds the usage count of the cred
771 struct. 779 struct.
772 780
773 Furthermore, if SELinux is enabled, this also checks that the 781 Furthermore, if SELinux is enabled, this also checks that the
774 security pointer in the cred struct is never seen to be invalid. 782 security pointer in the cred struct is never seen to be invalid.
775 783
776 If unsure, say N. 784 If unsure, say N.
777 785
778 # 786 #
779 # Select this config option from the architecture Kconfig, if it 787 # Select this config option from the architecture Kconfig, if it
780 # it is preferred to always offer frame pointers as a config 788 # it is preferred to always offer frame pointers as a config
781 # option on the architecture (regardless of KERNEL_DEBUG): 789 # option on the architecture (regardless of KERNEL_DEBUG):
782 # 790 #
783 config ARCH_WANT_FRAME_POINTERS 791 config ARCH_WANT_FRAME_POINTERS
784 bool 792 bool
785 help 793 help
786 794
787 config FRAME_POINTER 795 config FRAME_POINTER
788 bool "Compile the kernel with frame pointers" 796 bool "Compile the kernel with frame pointers"
789 depends on DEBUG_KERNEL && \ 797 depends on DEBUG_KERNEL && \
790 (CRIS || M68K || M68KNOMMU || FRV || UML || \ 798 (CRIS || M68K || M68KNOMMU || FRV || UML || \
791 AVR32 || SUPERH || BLACKFIN || MN10300) || \ 799 AVR32 || SUPERH || BLACKFIN || MN10300) || \
792 ARCH_WANT_FRAME_POINTERS 800 ARCH_WANT_FRAME_POINTERS
793 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS 801 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
794 help 802 help
795 If you say Y here the resulting kernel image will be slightly 803 If you say Y here the resulting kernel image will be slightly
796 larger and slower, but it gives very useful debugging information 804 larger and slower, but it gives very useful debugging information
797 in case of kernel bugs. (precise oopses/stacktraces/warnings) 805 in case of kernel bugs. (precise oopses/stacktraces/warnings)
798 806
799 config BOOT_PRINTK_DELAY 807 config BOOT_PRINTK_DELAY
800 bool "Delay each boot printk message by N milliseconds" 808 bool "Delay each boot printk message by N milliseconds"
801 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY 809 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
802 help 810 help
803 This build option allows you to read kernel boot messages 811 This build option allows you to read kernel boot messages
804 by inserting a short delay after each one. The delay is 812 by inserting a short delay after each one. The delay is
805 specified in milliseconds on the kernel command line, 813 specified in milliseconds on the kernel command line,
806 using "boot_delay=N". 814 using "boot_delay=N".
807 815
808 It is likely that you would also need to use "lpj=M" to preset 816 It is likely that you would also need to use "lpj=M" to preset
809 the "loops per jiffie" value. 817 the "loops per jiffie" value.
810 See a previous boot log for the "lpj" value to use for your 818 See a previous boot log for the "lpj" value to use for your
811 system, and then set "lpj=M" before setting "boot_delay=N". 819 system, and then set "lpj=M" before setting "boot_delay=N".
812 NOTE: Using this option may adversely affect SMP systems. 820 NOTE: Using this option may adversely affect SMP systems.
813 I.e., processors other than the first one may not boot up. 821 I.e., processors other than the first one may not boot up.
814 BOOT_PRINTK_DELAY also may cause DETECT_SOFTLOCKUP to detect 822 BOOT_PRINTK_DELAY also may cause DETECT_SOFTLOCKUP to detect
815 what it believes to be lockup conditions. 823 what it believes to be lockup conditions.
816 824
817 config RCU_TORTURE_TEST 825 config RCU_TORTURE_TEST
818 tristate "torture tests for RCU" 826 tristate "torture tests for RCU"
819 depends on DEBUG_KERNEL 827 depends on DEBUG_KERNEL
820 default n 828 default n
821 help 829 help
822 This option provides a kernel module that runs torture tests 830 This option provides a kernel module that runs torture tests
823 on the RCU infrastructure. The kernel module may be built 831 on the RCU infrastructure. The kernel module may be built
824 after the fact on the running kernel to be tested, if desired. 832 after the fact on the running kernel to be tested, if desired.
825 833
826 Say Y here if you want RCU torture tests to be built into 834 Say Y here if you want RCU torture tests to be built into
827 the kernel. 835 the kernel.
828 Say M if you want the RCU torture tests to build as a module. 836 Say M if you want the RCU torture tests to build as a module.
829 Say N if you are unsure. 837 Say N if you are unsure.
830 838
831 config RCU_TORTURE_TEST_RUNNABLE 839 config RCU_TORTURE_TEST_RUNNABLE
832 bool "torture tests for RCU runnable by default" 840 bool "torture tests for RCU runnable by default"
833 depends on RCU_TORTURE_TEST = y 841 depends on RCU_TORTURE_TEST = y
834 default n 842 default n
835 help 843 help
836 This option provides a way to build the RCU torture tests 844 This option provides a way to build the RCU torture tests
837 directly into the kernel without them starting up at boot 845 directly into the kernel without them starting up at boot
838 time. You can use /proc/sys/kernel/rcutorture_runnable 846 time. You can use /proc/sys/kernel/rcutorture_runnable
839 to manually override this setting. This /proc file is 847 to manually override this setting. This /proc file is
840 available only when the RCU torture tests have been built 848 available only when the RCU torture tests have been built
841 into the kernel. 849 into the kernel.
842 850
843 Say Y here if you want the RCU torture tests to start during 851 Say Y here if you want the RCU torture tests to start during
844 boot (you probably don't). 852 boot (you probably don't).
845 Say N here if you want the RCU torture tests to start only 853 Say N here if you want the RCU torture tests to start only
846 after being manually enabled via /proc. 854 after being manually enabled via /proc.
847 855
848 config RCU_CPU_STALL_DETECTOR 856 config RCU_CPU_STALL_DETECTOR
849 bool "Check for stalled CPUs delaying RCU grace periods" 857 bool "Check for stalled CPUs delaying RCU grace periods"
850 depends on TREE_RCU || TREE_PREEMPT_RCU 858 depends on TREE_RCU || TREE_PREEMPT_RCU
851 default y 859 default y
852 help 860 help
853 This option causes RCU to printk information on which 861 This option causes RCU to printk information on which
854 CPUs are delaying the current grace period, but only when 862 CPUs are delaying the current grace period, but only when
855 the grace period extends for excessive time periods. 863 the grace period extends for excessive time periods.
856 864
857 Say N if you want to disable such checks. 865 Say N if you want to disable such checks.
858 866
859 Say Y if you are unsure. 867 Say Y if you are unsure.
860 868
861 config RCU_CPU_STALL_TIMEOUT 869 config RCU_CPU_STALL_TIMEOUT
862 int "RCU CPU stall timeout in seconds" 870 int "RCU CPU stall timeout in seconds"
863 depends on RCU_CPU_STALL_DETECTOR 871 depends on RCU_CPU_STALL_DETECTOR
864 range 3 300 872 range 3 300
865 default 60 873 default 60
866 help 874 help
867 If a given RCU grace period extends more than the specified 875 If a given RCU grace period extends more than the specified
868 number of seconds, a CPU stall warning is printed. If the 876 number of seconds, a CPU stall warning is printed. If the
869 RCU grace period persists, additional CPU stall warnings are 877 RCU grace period persists, additional CPU stall warnings are
870 printed at more widely spaced intervals. 878 printed at more widely spaced intervals.
871 879
872 config RCU_CPU_STALL_DETECTOR_RUNNABLE 880 config RCU_CPU_STALL_DETECTOR_RUNNABLE
873 bool "RCU CPU stall checking starts automatically at boot" 881 bool "RCU CPU stall checking starts automatically at boot"
874 depends on RCU_CPU_STALL_DETECTOR 882 depends on RCU_CPU_STALL_DETECTOR
875 default y 883 default y
876 help 884 help
877 If set, start checking for RCU CPU stalls immediately on 885 If set, start checking for RCU CPU stalls immediately on
878 boot. Otherwise, RCU CPU stall checking must be manually 886 boot. Otherwise, RCU CPU stall checking must be manually
879 enabled. 887 enabled.
880 888
881 Say Y if you are unsure. 889 Say Y if you are unsure.
882 890
883 Say N if you wish to suppress RCU CPU stall checking during boot. 891 Say N if you wish to suppress RCU CPU stall checking during boot.
884 892
885 config RCU_CPU_STALL_VERBOSE 893 config RCU_CPU_STALL_VERBOSE
886 bool "Print additional per-task information for RCU_CPU_STALL_DETECTOR" 894 bool "Print additional per-task information for RCU_CPU_STALL_DETECTOR"
887 depends on RCU_CPU_STALL_DETECTOR && TREE_PREEMPT_RCU 895 depends on RCU_CPU_STALL_DETECTOR && TREE_PREEMPT_RCU
888 default y 896 default y
889 help 897 help
890 This option causes RCU to printk detailed per-task information 898 This option causes RCU to printk detailed per-task information
891 for any tasks that are stalling the current RCU grace period. 899 for any tasks that are stalling the current RCU grace period.
892 900
893 Say N if you are unsure. 901 Say N if you are unsure.
894 902
895 Say Y if you want to enable such checks. 903 Say Y if you want to enable such checks.
896 904
897 config KPROBES_SANITY_TEST 905 config KPROBES_SANITY_TEST
898 bool "Kprobes sanity tests" 906 bool "Kprobes sanity tests"
899 depends on DEBUG_KERNEL 907 depends on DEBUG_KERNEL
900 depends on KPROBES 908 depends on KPROBES
901 default n 909 default n
902 help 910 help
903 This option provides for testing basic kprobes functionality on 911 This option provides for testing basic kprobes functionality on
904 boot. A sample kprobe, jprobe and kretprobe are inserted and 912 boot. A sample kprobe, jprobe and kretprobe are inserted and
905 verified for functionality. 913 verified for functionality.
906 914
907 Say N if you are unsure. 915 Say N if you are unsure.
908 916
909 config BACKTRACE_SELF_TEST 917 config BACKTRACE_SELF_TEST
910 tristate "Self test for the backtrace code" 918 tristate "Self test for the backtrace code"
911 depends on DEBUG_KERNEL 919 depends on DEBUG_KERNEL
912 default n 920 default n
913 help 921 help
914 This option provides a kernel module that can be used to test 922 This option provides a kernel module that can be used to test
915 the kernel stack backtrace code. This option is not useful 923 the kernel stack backtrace code. This option is not useful
916 for distributions or general kernels, but only for kernel 924 for distributions or general kernels, but only for kernel
917 developers working on architecture code. 925 developers working on architecture code.
918 926
919 Note that if you want to also test saved backtraces, you will 927 Note that if you want to also test saved backtraces, you will
920 have to enable STACKTRACE as well. 928 have to enable STACKTRACE as well.
921 929
922 Say N if you are unsure. 930 Say N if you are unsure.
923 931
924 config DEBUG_BLOCK_EXT_DEVT 932 config DEBUG_BLOCK_EXT_DEVT
925 bool "Force extended block device numbers and spread them" 933 bool "Force extended block device numbers and spread them"
926 depends on DEBUG_KERNEL 934 depends on DEBUG_KERNEL
927 depends on BLOCK 935 depends on BLOCK
928 default n 936 default n
929 help 937 help
930 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON 938 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
931 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT 939 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
932 YOU ARE DOING. Distros, please enable this and fix whatever 940 YOU ARE DOING. Distros, please enable this and fix whatever
933 is broken. 941 is broken.
934 942
935 Conventionally, block device numbers are allocated from 943 Conventionally, block device numbers are allocated from
936 predetermined contiguous area. However, extended block area 944 predetermined contiguous area. However, extended block area
937 may introduce non-contiguous block device numbers. This 945 may introduce non-contiguous block device numbers. This
938 option forces most block device numbers to be allocated from 946 option forces most block device numbers to be allocated from
939 the extended space and spreads them to discover kernel or 947 the extended space and spreads them to discover kernel or
940 userland code paths which assume predetermined contiguous 948 userland code paths which assume predetermined contiguous
941 device number allocation. 949 device number allocation.
942 950
943 Note that turning on this debug option shuffles all the 951 Note that turning on this debug option shuffles all the
944 device numbers for all IDE and SCSI devices including libata 952 device numbers for all IDE and SCSI devices including libata
945 ones, so root partition specified using device number 953 ones, so root partition specified using device number
946 directly (via rdev or root=MAJ:MIN) won't work anymore. 954 directly (via rdev or root=MAJ:MIN) won't work anymore.
947 Textual device names (root=/dev/sdXn) will continue to work. 955 Textual device names (root=/dev/sdXn) will continue to work.
948 956
949 Say N if you are unsure. 957 Say N if you are unsure.
950 958
951 config DEBUG_FORCE_WEAK_PER_CPU 959 config DEBUG_FORCE_WEAK_PER_CPU
952 bool "Force weak per-cpu definitions" 960 bool "Force weak per-cpu definitions"
953 depends on DEBUG_KERNEL 961 depends on DEBUG_KERNEL
954 help 962 help
955 s390 and alpha require percpu variables in modules to be 963 s390 and alpha require percpu variables in modules to be
956 defined weak to work around addressing range issue which 964 defined weak to work around addressing range issue which
957 puts the following two restrictions on percpu variable 965 puts the following two restrictions on percpu variable
958 definitions. 966 definitions.
959 967
960 1. percpu symbols must be unique whether static or not 968 1. percpu symbols must be unique whether static or not
961 2. percpu variables can't be defined inside a function 969 2. percpu variables can't be defined inside a function
962 970
963 To ensure that generic code follows the above rules, this 971 To ensure that generic code follows the above rules, this
964 option forces all percpu variables to be defined as weak. 972 option forces all percpu variables to be defined as weak.
965 973
966 config LKDTM 974 config LKDTM
967 tristate "Linux Kernel Dump Test Tool Module" 975 tristate "Linux Kernel Dump Test Tool Module"
968 depends on DEBUG_FS 976 depends on DEBUG_FS
969 depends on BLOCK 977 depends on BLOCK
970 default n 978 default n
971 help 979 help
972 This module enables testing of the different dumping mechanisms by 980 This module enables testing of the different dumping mechanisms by
973 inducing system failures at predefined crash points. 981 inducing system failures at predefined crash points.
974 If you don't need it: say N 982 If you don't need it: say N
975 Choose M here to compile this code as a module. The module will be 983 Choose M here to compile this code as a module. The module will be
976 called lkdtm. 984 called lkdtm.
977 985
978 Documentation on how to use the module can be found in 986 Documentation on how to use the module can be found in
979 Documentation/fault-injection/provoke-crashes.txt 987 Documentation/fault-injection/provoke-crashes.txt
980 988
981 config CPU_NOTIFIER_ERROR_INJECT 989 config CPU_NOTIFIER_ERROR_INJECT
982 tristate "CPU notifier error injection module" 990 tristate "CPU notifier error injection module"
983 depends on HOTPLUG_CPU && DEBUG_KERNEL 991 depends on HOTPLUG_CPU && DEBUG_KERNEL
984 help 992 help
985 This option provides a kernel module that can be used to test 993 This option provides a kernel module that can be used to test
986 the error handling of the cpu notifiers 994 the error handling of the cpu notifiers
987 995
988 To compile this code as a module, choose M here: the module will 996 To compile this code as a module, choose M here: the module will
989 be called cpu-notifier-error-inject. 997 be called cpu-notifier-error-inject.
990 998
991 If unsure, say N. 999 If unsure, say N.
992 1000
993 config FAULT_INJECTION 1001 config FAULT_INJECTION
994 bool "Fault-injection framework" 1002 bool "Fault-injection framework"
995 depends on DEBUG_KERNEL 1003 depends on DEBUG_KERNEL
996 help 1004 help
997 Provide fault-injection framework. 1005 Provide fault-injection framework.
998 For more details, see Documentation/fault-injection/. 1006 For more details, see Documentation/fault-injection/.
999 1007
1000 config FAILSLAB 1008 config FAILSLAB
1001 bool "Fault-injection capability for kmalloc" 1009 bool "Fault-injection capability for kmalloc"
1002 depends on FAULT_INJECTION 1010 depends on FAULT_INJECTION
1003 depends on SLAB || SLUB 1011 depends on SLAB || SLUB
1004 help 1012 help
1005 Provide fault-injection capability for kmalloc. 1013 Provide fault-injection capability for kmalloc.
1006 1014
1007 config FAIL_PAGE_ALLOC 1015 config FAIL_PAGE_ALLOC
1008 bool "Fault-injection capabilitiy for alloc_pages()" 1016 bool "Fault-injection capabilitiy for alloc_pages()"
1009 depends on FAULT_INJECTION 1017 depends on FAULT_INJECTION
1010 help 1018 help
1011 Provide fault-injection capability for alloc_pages(). 1019 Provide fault-injection capability for alloc_pages().
1012 1020
1013 config FAIL_MAKE_REQUEST 1021 config FAIL_MAKE_REQUEST
1014 bool "Fault-injection capability for disk IO" 1022 bool "Fault-injection capability for disk IO"
1015 depends on FAULT_INJECTION && BLOCK 1023 depends on FAULT_INJECTION && BLOCK
1016 help 1024 help
1017 Provide fault-injection capability for disk IO. 1025 Provide fault-injection capability for disk IO.
1018 1026
1019 config FAIL_IO_TIMEOUT 1027 config FAIL_IO_TIMEOUT
1020 bool "Fault-injection capability for faking disk interrupts" 1028 bool "Fault-injection capability for faking disk interrupts"
1021 depends on FAULT_INJECTION && BLOCK 1029 depends on FAULT_INJECTION && BLOCK
1022 help 1030 help
1023 Provide fault-injection capability on end IO handling. This 1031 Provide fault-injection capability on end IO handling. This
1024 will make the block layer "forget" an interrupt as configured, 1032 will make the block layer "forget" an interrupt as configured,
1025 thus exercising the error handling. 1033 thus exercising the error handling.
1026 1034
1027 Only works with drivers that use the generic timeout handling, 1035 Only works with drivers that use the generic timeout handling,
1028 for others it wont do anything. 1036 for others it wont do anything.
1029 1037
1030 config FAULT_INJECTION_DEBUG_FS 1038 config FAULT_INJECTION_DEBUG_FS
1031 bool "Debugfs entries for fault-injection capabilities" 1039 bool "Debugfs entries for fault-injection capabilities"
1032 depends on FAULT_INJECTION && SYSFS && DEBUG_FS 1040 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1033 help 1041 help
1034 Enable configuration of fault-injection capabilities via debugfs. 1042 Enable configuration of fault-injection capabilities via debugfs.
1035 1043
1036 config FAULT_INJECTION_STACKTRACE_FILTER 1044 config FAULT_INJECTION_STACKTRACE_FILTER
1037 bool "stacktrace filter for fault-injection capabilities" 1045 bool "stacktrace filter for fault-injection capabilities"
1038 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT 1046 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1039 depends on !X86_64 1047 depends on !X86_64
1040 select STACKTRACE 1048 select STACKTRACE
1041 select FRAME_POINTER if !PPC && !S390 && !MICROBLAZE 1049 select FRAME_POINTER if !PPC && !S390 && !MICROBLAZE
1042 help 1050 help
1043 Provide stacktrace filter for fault-injection capabilities 1051 Provide stacktrace filter for fault-injection capabilities
1044 1052
1045 config LATENCYTOP 1053 config LATENCYTOP
1046 bool "Latency measuring infrastructure" 1054 bool "Latency measuring infrastructure"
1047 depends on HAVE_LATENCYTOP_SUPPORT 1055 depends on HAVE_LATENCYTOP_SUPPORT
1048 depends on DEBUG_KERNEL 1056 depends on DEBUG_KERNEL
1049 depends on STACKTRACE_SUPPORT 1057 depends on STACKTRACE_SUPPORT
1050 depends on PROC_FS 1058 depends on PROC_FS
1051 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE 1059 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE
1052 select KALLSYMS 1060 select KALLSYMS
1053 select KALLSYMS_ALL 1061 select KALLSYMS_ALL
1054 select STACKTRACE 1062 select STACKTRACE
1055 select SCHEDSTATS 1063 select SCHEDSTATS
1056 select SCHED_DEBUG 1064 select SCHED_DEBUG
1057 help 1065 help
1058 Enable this option if you want to use the LatencyTOP tool 1066 Enable this option if you want to use the LatencyTOP tool
1059 to find out which userspace is blocking on what kernel operations. 1067 to find out which userspace is blocking on what kernel operations.
1060 1068
1061 config SYSCTL_SYSCALL_CHECK 1069 config SYSCTL_SYSCALL_CHECK
1062 bool "Sysctl checks" 1070 bool "Sysctl checks"
1063 depends on SYSCTL 1071 depends on SYSCTL
1064 ---help--- 1072 ---help---
1065 sys_sysctl uses binary paths that have been found challenging 1073 sys_sysctl uses binary paths that have been found challenging
1066 to properly maintain and use. This enables checks that help 1074 to properly maintain and use. This enables checks that help
1067 you to keep things correct. 1075 you to keep things correct.
1068 1076
1069 source mm/Kconfig.debug 1077 source mm/Kconfig.debug
1070 source kernel/trace/Kconfig 1078 source kernel/trace/Kconfig
1071 1079
1072 config PROVIDE_OHCI1394_DMA_INIT 1080 config PROVIDE_OHCI1394_DMA_INIT
1073 bool "Remote debugging over FireWire early on boot" 1081 bool "Remote debugging over FireWire early on boot"
1074 depends on PCI && X86 1082 depends on PCI && X86
1075 help 1083 help
1076 If you want to debug problems which hang or crash the kernel early 1084 If you want to debug problems which hang or crash the kernel early
1077 on boot and the crashing machine has a FireWire port, you can use 1085 on boot and the crashing machine has a FireWire port, you can use
1078 this feature to remotely access the memory of the crashed machine 1086 this feature to remotely access the memory of the crashed machine
1079 over FireWire. This employs remote DMA as part of the OHCI1394 1087 over FireWire. This employs remote DMA as part of the OHCI1394
1080 specification which is now the standard for FireWire controllers. 1088 specification which is now the standard for FireWire controllers.
1081 1089
1082 With remote DMA, you can monitor the printk buffer remotely using 1090 With remote DMA, you can monitor the printk buffer remotely using
1083 firescope and access all memory below 4GB using fireproxy from gdb. 1091 firescope and access all memory below 4GB using fireproxy from gdb.
1084 Even controlling a kernel debugger is possible using remote DMA. 1092 Even controlling a kernel debugger is possible using remote DMA.
1085 1093
1086 Usage: 1094 Usage:
1087 1095
1088 If ohci1394_dma=early is used as boot parameter, it will initialize 1096 If ohci1394_dma=early is used as boot parameter, it will initialize
1089 all OHCI1394 controllers which are found in the PCI config space. 1097 all OHCI1394 controllers which are found in the PCI config space.
1090 1098
1091 As all changes to the FireWire bus such as enabling and disabling 1099 As all changes to the FireWire bus such as enabling and disabling
1092 devices cause a bus reset and thereby disable remote DMA for all 1100 devices cause a bus reset and thereby disable remote DMA for all
1093 devices, be sure to have the cable plugged and FireWire enabled on 1101 devices, be sure to have the cable plugged and FireWire enabled on
1094 the debugging host before booting the debug target for debugging. 1102 the debugging host before booting the debug target for debugging.
1095 1103
1096 This code (~1k) is freed after boot. By then, the firewire stack 1104 This code (~1k) is freed after boot. By then, the firewire stack
1097 in charge of the OHCI-1394 controllers should be used instead. 1105 in charge of the OHCI-1394 controllers should be used instead.
1098 1106
1099 See Documentation/debugging-via-ohci1394.txt for more information. 1107 See Documentation/debugging-via-ohci1394.txt for more information.
1100 1108
1101 config FIREWIRE_OHCI_REMOTE_DMA 1109 config FIREWIRE_OHCI_REMOTE_DMA
1102 bool "Remote debugging over FireWire with firewire-ohci" 1110 bool "Remote debugging over FireWire with firewire-ohci"
1103 depends on FIREWIRE_OHCI 1111 depends on FIREWIRE_OHCI
1104 help 1112 help
1105 This option lets you use the FireWire bus for remote debugging 1113 This option lets you use the FireWire bus for remote debugging
1106 with help of the firewire-ohci driver. It enables unfiltered 1114 with help of the firewire-ohci driver. It enables unfiltered
1107 remote DMA in firewire-ohci. 1115 remote DMA in firewire-ohci.
1108 See Documentation/debugging-via-ohci1394.txt for more information. 1116 See Documentation/debugging-via-ohci1394.txt for more information.
1109 1117
1110 If unsure, say N. 1118 If unsure, say N.
1111 1119
1112 config BUILD_DOCSRC 1120 config BUILD_DOCSRC
1113 bool "Build targets in Documentation/ tree" 1121 bool "Build targets in Documentation/ tree"
1114 depends on HEADERS_CHECK 1122 depends on HEADERS_CHECK
1115 help 1123 help
1116 This option attempts to build objects from the source files in the 1124 This option attempts to build objects from the source files in the
1117 kernel Documentation/ tree. 1125 kernel Documentation/ tree.
1118 1126
1119 Say N if you are unsure. 1127 Say N if you are unsure.
1120 1128
1121 config DYNAMIC_DEBUG 1129 config DYNAMIC_DEBUG
1122 bool "Enable dynamic printk() support" 1130 bool "Enable dynamic printk() support"
1123 default n 1131 default n
1124 depends on PRINTK 1132 depends on PRINTK
1125 depends on DEBUG_FS 1133 depends on DEBUG_FS
1126 help 1134 help
1127 1135
1128 Compiles debug level messages into the kernel, which would not 1136 Compiles debug level messages into the kernel, which would not
1129 otherwise be available at runtime. These messages can then be 1137 otherwise be available at runtime. These messages can then be
1130 enabled/disabled based on various levels of scope - per source file, 1138 enabled/disabled based on various levels of scope - per source file,
1131 function, module, format string, and line number. This mechanism 1139 function, module, format string, and line number. This mechanism
1132 implicitly enables all pr_debug() and dev_dbg() calls. The impact of 1140 implicitly enables all pr_debug() and dev_dbg() calls. The impact of
1133 this compile option is a larger kernel text size of about 2%. 1141 this compile option is a larger kernel text size of about 2%.
1134 1142
1135 Usage: 1143 Usage:
1136 1144
1137 Dynamic debugging is controlled via the 'dynamic_debug/control' file, 1145 Dynamic debugging is controlled via the 'dynamic_debug/control' file,
1138 which is contained in the 'debugfs' filesystem. Thus, the debugfs 1146 which is contained in the 'debugfs' filesystem. Thus, the debugfs
1139 filesystem must first be mounted before making use of this feature. 1147 filesystem must first be mounted before making use of this feature.
1140 We refer the control file as: <debugfs>/dynamic_debug/control. This 1148 We refer the control file as: <debugfs>/dynamic_debug/control. This
1141 file contains a list of the debug statements that can be enabled. The 1149 file contains a list of the debug statements that can be enabled. The
1142 format for each line of the file is: 1150 format for each line of the file is:
1143 1151
1144 filename:lineno [module]function flags format 1152 filename:lineno [module]function flags format
1145 1153
1146 filename : source file of the debug statement 1154 filename : source file of the debug statement
1147 lineno : line number of the debug statement 1155 lineno : line number of the debug statement
1148 module : module that contains the debug statement 1156 module : module that contains the debug statement
1149 function : function that contains the debug statement 1157 function : function that contains the debug statement
1150 flags : 'p' means the line is turned 'on' for printing 1158 flags : 'p' means the line is turned 'on' for printing
1151 format : the format used for the debug statement 1159 format : the format used for the debug statement
1152 1160
1153 From a live system: 1161 From a live system:
1154 1162
1155 nullarbor:~ # cat <debugfs>/dynamic_debug/control 1163 nullarbor:~ # cat <debugfs>/dynamic_debug/control
1156 # filename:lineno [module]function flags format 1164 # filename:lineno [module]function flags format
1157 fs/aio.c:222 [aio]__put_ioctx - "__put_ioctx:\040freeing\040%p\012" 1165 fs/aio.c:222 [aio]__put_ioctx - "__put_ioctx:\040freeing\040%p\012"
1158 fs/aio.c:248 [aio]ioctx_alloc - "ENOMEM:\040nr_events\040too\040high\012" 1166 fs/aio.c:248 [aio]ioctx_alloc - "ENOMEM:\040nr_events\040too\040high\012"
1159 fs/aio.c:1770 [aio]sys_io_cancel - "calling\040cancel\012" 1167 fs/aio.c:1770 [aio]sys_io_cancel - "calling\040cancel\012"
1160 1168
1161 Example usage: 1169 Example usage:
1162 1170
1163 // enable the message at line 1603 of file svcsock.c 1171 // enable the message at line 1603 of file svcsock.c
1164 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' > 1172 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
1165 <debugfs>/dynamic_debug/control 1173 <debugfs>/dynamic_debug/control
1166 1174
1167 // enable all the messages in file svcsock.c 1175 // enable all the messages in file svcsock.c
1168 nullarbor:~ # echo -n 'file svcsock.c +p' > 1176 nullarbor:~ # echo -n 'file svcsock.c +p' >
1169 <debugfs>/dynamic_debug/control 1177 <debugfs>/dynamic_debug/control
1170 1178
1171 // enable all the messages in the NFS server module 1179 // enable all the messages in the NFS server module
1172 nullarbor:~ # echo -n 'module nfsd +p' > 1180 nullarbor:~ # echo -n 'module nfsd +p' >
1173 <debugfs>/dynamic_debug/control 1181 <debugfs>/dynamic_debug/control
1174 1182
1175 // enable all 12 messages in the function svc_process() 1183 // enable all 12 messages in the function svc_process()
1176 nullarbor:~ # echo -n 'func svc_process +p' > 1184 nullarbor:~ # echo -n 'func svc_process +p' >
1177 <debugfs>/dynamic_debug/control 1185 <debugfs>/dynamic_debug/control
1178 1186
1179 // disable all 12 messages in the function svc_process() 1187 // disable all 12 messages in the function svc_process()
1180 nullarbor:~ # echo -n 'func svc_process -p' > 1188 nullarbor:~ # echo -n 'func svc_process -p' >
1181 <debugfs>/dynamic_debug/control 1189 <debugfs>/dynamic_debug/control
1182 1190
1183 See Documentation/dynamic-debug-howto.txt for additional information. 1191 See Documentation/dynamic-debug-howto.txt for additional information.
1184 1192
1185 config DMA_API_DEBUG 1193 config DMA_API_DEBUG
1186 bool "Enable debugging of DMA-API usage" 1194 bool "Enable debugging of DMA-API usage"
1187 depends on HAVE_DMA_API_DEBUG 1195 depends on HAVE_DMA_API_DEBUG
1188 help 1196 help
1189 Enable this option to debug the use of the DMA API by device drivers. 1197 Enable this option to debug the use of the DMA API by device drivers.
1190 With this option you will be able to detect common bugs in device 1198 With this option you will be able to detect common bugs in device
1191 drivers like double-freeing of DMA mappings or freeing mappings that 1199 drivers like double-freeing of DMA mappings or freeing mappings that
1192 were never allocated. 1200 were never allocated.
1193 This option causes a performance degredation. Use only if you want 1201 This option causes a performance degredation. Use only if you want
1194 to debug device drivers. If unsure, say N. 1202 to debug device drivers. If unsure, say N.
1195 1203
1196 config ATOMIC64_SELFTEST 1204 config ATOMIC64_SELFTEST
1197 bool "Perform an atomic64_t self-test at boot" 1205 bool "Perform an atomic64_t self-test at boot"
1198 help 1206 help
1199 Enable this option to test the atomic64_t functions at boot. 1207 Enable this option to test the atomic64_t functions at boot.
1200 1208
1201 If unsure, say N. 1209 If unsure, say N.
1202 1210
1203 source "samples/Kconfig" 1211 source "samples/Kconfig"
1204 1212
1205 source "lib/Kconfig.kgdb" 1213 source "lib/Kconfig.kgdb"
1206 1214
1207 source "lib/Kconfig.kmemcheck" 1215 source "lib/Kconfig.kmemcheck"
1208 1216
lib/percpu_counter.c
Diff comments   View file @ e2852ae
1 /* 1 /*
2 * Fast batching percpu counters. 2 * Fast batching percpu counters.
3 */ 3 */
4 4
5 #include <linux/percpu_counter.h> 5 #include <linux/percpu_counter.h>
6 #include <linux/notifier.h> 6 #include <linux/notifier.h>
7 #include <linux/mutex.h> 7 #include <linux/mutex.h>
8 #include <linux/init.h> 8 #include <linux/init.h>
9 #include <linux/cpu.h> 9 #include <linux/cpu.h>
10 #include <linux/module.h> 10 #include <linux/module.h>
11 #include <linux/debugobjects.h>
11 12
12 static LIST_HEAD(percpu_counters); 13 static LIST_HEAD(percpu_counters);
13 static DEFINE_MUTEX(percpu_counters_lock); 14 static DEFINE_MUTEX(percpu_counters_lock);
14 15
16 #ifdef CONFIG_DEBUG_OBJECTS_PERCPU_COUNTER
17
18 static struct debug_obj_descr percpu_counter_debug_descr;
19
20 static int percpu_counter_fixup_free(void *addr, enum debug_obj_state state)
21 {
22 struct percpu_counter *fbc = addr;
23
24 switch (state) {
25 case ODEBUG_STATE_ACTIVE:
26 percpu_counter_destroy(fbc);
27 debug_object_free(fbc, &percpu_counter_debug_descr);
28 return 1;
29 default:
30 return 0;
31 }
32 }
33
34 static struct debug_obj_descr percpu_counter_debug_descr = {
35 .name = "percpu_counter",
36 .fixup_free = percpu_counter_fixup_free,
37 };
38
39 static inline void debug_percpu_counter_activate(struct percpu_counter *fbc)
40 {
41 debug_object_init(fbc, &percpu_counter_debug_descr);
42 debug_object_activate(fbc, &percpu_counter_debug_descr);
43 }
44
45 static inline void debug_percpu_counter_deactivate(struct percpu_counter *fbc)
46 {
47 debug_object_deactivate(fbc, &percpu_counter_debug_descr);
48 debug_object_free(fbc, &percpu_counter_debug_descr);
49 }
50
51 #else /* CONFIG_DEBUG_OBJECTS_PERCPU_COUNTER */
52 static inline void debug_percpu_counter_activate(struct percpu_counter *fbc)
53 { }
54 static inline void debug_percpu_counter_deactivate(struct percpu_counter *fbc)
55 { }
56 #endif /* CONFIG_DEBUG_OBJECTS_PERCPU_COUNTER */
57
15 void percpu_counter_set(struct percpu_counter *fbc, s64 amount) 58 void percpu_counter_set(struct percpu_counter *fbc, s64 amount)
16 { 59 {
17 int cpu; 60 int cpu;
18 61
19 spin_lock(&fbc->lock); 62 spin_lock(&fbc->lock);
20 for_each_possible_cpu(cpu) { 63 for_each_possible_cpu(cpu) {
21 s32 *pcount = per_cpu_ptr(fbc->counters, cpu); 64 s32 *pcount = per_cpu_ptr(fbc->counters, cpu);
22 *pcount = 0; 65 *pcount = 0;
23 } 66 }
24 fbc->count = amount; 67 fbc->count = amount;
25 spin_unlock(&fbc->lock); 68 spin_unlock(&fbc->lock);
26 } 69 }
27 EXPORT_SYMBOL(percpu_counter_set); 70 EXPORT_SYMBOL(percpu_counter_set);
28 71
29 void __percpu_counter_add(struct percpu_counter *fbc, s64 amount, s32 batch) 72 void __percpu_counter_add(struct percpu_counter *fbc, s64 amount, s32 batch)
30 { 73 {
31 s64 count; 74 s64 count;
32 s32 *pcount; 75 s32 *pcount;
33 int cpu = get_cpu(); 76 int cpu = get_cpu();
34 77
35 pcount = per_cpu_ptr(fbc->counters, cpu); 78 pcount = per_cpu_ptr(fbc->counters, cpu);
36 count = *pcount + amount; 79 count = *pcount + amount;
37 if (count >= batch || count <= -batch) { 80 if (count >= batch || count <= -batch) {
38 spin_lock(&fbc->lock); 81 spin_lock(&fbc->lock);
39 fbc->count += count; 82 fbc->count += count;
40 *pcount = 0; 83 *pcount = 0;
41 spin_unlock(&fbc->lock); 84 spin_unlock(&fbc->lock);
42 } else { 85 } else {
43 *pcount = count; 86 *pcount = count;
44 } 87 }
45 put_cpu(); 88 put_cpu();
46 } 89 }
47 EXPORT_SYMBOL(__percpu_counter_add); 90 EXPORT_SYMBOL(__percpu_counter_add);
48 91
49 /* 92 /*
50 * Add up all the per-cpu counts, return the result. This is a more accurate 93 * Add up all the per-cpu counts, return the result. This is a more accurate
51 * but much slower version of percpu_counter_read_positive() 94 * but much slower version of percpu_counter_read_positive()
52 */ 95 */
53 s64 __percpu_counter_sum(struct percpu_counter *fbc) 96 s64 __percpu_counter_sum(struct percpu_counter *fbc)
54 { 97 {
55 s64 ret; 98 s64 ret;
56 int cpu; 99 int cpu;
57 100
58 spin_lock(&fbc->lock); 101 spin_lock(&fbc->lock);
59 ret = fbc->count; 102 ret = fbc->count;
60 for_each_online_cpu(cpu) { 103 for_each_online_cpu(cpu) {
61 s32 *pcount = per_cpu_ptr(fbc->counters, cpu); 104 s32 *pcount = per_cpu_ptr(fbc->counters, cpu);
62 ret += *pcount; 105 ret += *pcount;
63 } 106 }
64 spin_unlock(&fbc->lock); 107 spin_unlock(&fbc->lock);
65 return ret; 108 return ret;
66 } 109 }
67 EXPORT_SYMBOL(__percpu_counter_sum); 110 EXPORT_SYMBOL(__percpu_counter_sum);
68 111
69 int __percpu_counter_init(struct percpu_counter *fbc, s64 amount, 112 int __percpu_counter_init(struct percpu_counter *fbc, s64 amount,
70 struct lock_class_key *key) 113 struct lock_class_key *key)
71 { 114 {
72 spin_lock_init(&fbc->lock); 115 spin_lock_init(&fbc->lock);
73 lockdep_set_class(&fbc->lock, key); 116 lockdep_set_class(&fbc->lock, key);
74 fbc->count = amount; 117 fbc->count = amount;
75 fbc->counters = alloc_percpu(s32); 118 fbc->counters = alloc_percpu(s32);
76 if (!fbc->counters) 119 if (!fbc->counters)
77 return -ENOMEM; 120 return -ENOMEM;
121
122 debug_percpu_counter_activate(fbc);
123
78 #ifdef CONFIG_HOTPLUG_CPU 124 #ifdef CONFIG_HOTPLUG_CPU
79 INIT_LIST_HEAD(&fbc->list); 125 INIT_LIST_HEAD(&fbc->list);
80 mutex_lock(&percpu_counters_lock); 126 mutex_lock(&percpu_counters_lock);
81 list_add(&fbc->list, &percpu_counters); 127 list_add(&fbc->list, &percpu_counters);
82 mutex_unlock(&percpu_counters_lock); 128 mutex_unlock(&percpu_counters_lock);
83 #endif 129 #endif
84 return 0; 130 return 0;
85 } 131 }
86 EXPORT_SYMBOL(__percpu_counter_init); 132 EXPORT_SYMBOL(__percpu_counter_init);
87 133
88 void percpu_counter_destroy(struct percpu_counter *fbc) 134 void percpu_counter_destroy(struct percpu_counter *fbc)
89 { 135 {
90 if (!fbc->counters) 136 if (!fbc->counters)
91 return; 137 return;
138
139 debug_percpu_counter_deactivate(fbc);
92 140
93 #ifdef CONFIG_HOTPLUG_CPU 141 #ifdef CONFIG_HOTPLUG_CPU
94 mutex_lock(&percpu_counters_lock); 142 mutex_lock(&percpu_counters_lock);
95 list_del(&fbc->list); 143 list_del(&fbc->list);
96 mutex_unlock(&percpu_counters_lock); 144 mutex_unlock(&percpu_counters_lock);
97 #endif 145 #endif
98 free_percpu(fbc->counters); 146 free_percpu(fbc->counters);
99 fbc->counters = NULL; 147 fbc->counters = NULL;
100 } 148 }
101 EXPORT_SYMBOL(percpu_counter_destroy); 149 EXPORT_SYMBOL(percpu_counter_destroy);
102 150
103 int percpu_counter_batch __read_mostly = 32; 151 int percpu_counter_batch __read_mostly = 32;
104 EXPORT_SYMBOL(percpu_counter_batch); 152 EXPORT_SYMBOL(percpu_counter_batch);
105 153
106 static void compute_batch_value(void) 154 static void compute_batch_value(void)
107 { 155 {
108 int nr = num_online_cpus(); 156 int nr = num_online_cpus();
109 157
110 percpu_counter_batch = max(32, nr*2); 158 percpu_counter_batch = max(32, nr*2);
111 } 159 }
112 160
113 static int __cpuinit percpu_counter_hotcpu_callback(struct notifier_block *nb, 161 static int __cpuinit percpu_counter_hotcpu_callback(struct notifier_block *nb,
114 unsigned long action, void *hcpu) 162 unsigned long action, void *hcpu)
115 { 163 {
116 #ifdef CONFIG_HOTPLUG_CPU 164 #ifdef CONFIG_HOTPLUG_CPU
117 unsigned int cpu; 165 unsigned int cpu;
118 struct percpu_counter *fbc; 166 struct percpu_counter *fbc;
119 167
120 compute_batch_value(); 168 compute_batch_value();
121 if (action != CPU_DEAD) 169 if (action != CPU_DEAD)
122 return NOTIFY_OK; 170 return NOTIFY_OK;
123 171
124 cpu = (unsigned long)hcpu; 172 cpu = (unsigned long)hcpu;
125 mutex_lock(&percpu_counters_lock); 173 mutex_lock(&percpu_counters_lock);
126 list_for_each_entry(fbc, &percpu_counters, list) { 174 list_for_each_entry(fbc, &percpu_counters, list) {
127 s32 *pcount; 175 s32 *pcount;
128 unsigned long flags; 176 unsigned long flags;
129 177
130 spin_lock_irqsave(&fbc->lock, flags); 178 spin_lock_irqsave(&fbc->lock, flags);
131 pcount = per_cpu_ptr(fbc->counters, cpu); 179 pcount = per_cpu_ptr(fbc->counters, cpu);
132 fbc->count += *pcount; 180 fbc->count += *pcount;
133 *pcount = 0; 181 *pcount = 0;
134 spin_unlock_irqrestore(&fbc->lock, flags); 182 spin_unlock_irqrestore(&fbc->lock, flags);
135 } 183 }
136 mutex_unlock(&percpu_counters_lock); 184 mutex_unlock(&percpu_counters_lock);
137 #endif 185 #endif
138 return NOTIFY_OK; 186 return NOTIFY_OK;
139 } 187 }
140 188
141 /* 189 /*
142 * Compare counter against given value. 190 * Compare counter against given value.
143 * Return 1 if greater, 0 if equal and -1 if less 191 * Return 1 if greater, 0 if equal and -1 if less
144 */ 192 */
145 int percpu_counter_compare(struct percpu_counter *fbc, s64 rhs) 193 int percpu_counter_compare(struct percpu_counter *fbc, s64 rhs)
146 { 194 {
147 s64 count; 195 s64 count;
148 196
149 count = percpu_counter_read(fbc); 197 count = percpu_counter_read(fbc);
150 /* Check to see if rough count will be sufficient for comparison */ 198 /* Check to see if rough count will be sufficient for comparison */
151 if (abs(count - rhs) > (percpu_counter_batch*num_online_cpus())) { 199 if (abs(count - rhs) > (percpu_counter_batch*num_online_cpus())) {
152 if (count > rhs) 200 if (count > rhs)
153 return 1; 201 return 1;
154 else 202 else
155 return -1; 203 return -1;
156 } 204 }
157 /* Need to use precise count */ 205 /* Need to use precise count */
158 count = percpu_counter_sum(fbc); 206 count = percpu_counter_sum(fbc);
159 if (count > rhs) 207 if (count > rhs)
160 return 1; 208 return 1;
161 else if (count < rhs) 209 else if (count < rhs)
162 return -1; 210 return -1;
163 else 211 else
164 return 0; 212 return 0;
165 } 213 }
166 EXPORT_SYMBOL(percpu_counter_compare); 214 EXPORT_SYMBOL(percpu_counter_compare);
167 215
168 static int __init percpu_counter_startup(void) 216 static int __init percpu_counter_startup(void)
169 { 217 {
170 compute_batch_value(); 218 compute_batch_value();
171 hotcpu_notifier(percpu_counter_hotcpu_callback, 0); 219 hotcpu_notifier(percpu_counter_hotcpu_callback, 0);
172 return 0; 220 return 0;
173 } 221 }
174 module_init(percpu_counter_startup); 222 module_init(percpu_counter_startup);
175 223