1 ====================== 1 ======================
2 Function Tracer Design 2 Function Tracer Design
3 ====================== 3 ======================
4 4
5 :Author: Mike Frysinger 5 :Author: Mike Frysinger
6 6
7 .. caution:: 7 .. caution::
8 This document is out of date. Some of 8 This document is out of date. Some of the description below doesn't
9 match current implementation now. 9 match current implementation now.
10 10
11 Introduction 11 Introduction
12 ------------ 12 ------------
13 13
14 Here we will cover the architecture pieces tha 14 Here we will cover the architecture pieces that the common function tracing
15 code relies on for proper functioning. Things 15 code relies on for proper functioning. Things are broken down into increasing
16 complexity so that you can start simple and at 16 complexity so that you can start simple and at least get basic functionality.
17 17
18 Note that this focuses on architecture impleme 18 Note that this focuses on architecture implementation details only. If you
19 want more explanation of a feature in terms of 19 want more explanation of a feature in terms of common code, review the common
20 ftrace.txt file. 20 ftrace.txt file.
21 21
22 Ideally, everyone who wishes to retain perform 22 Ideally, everyone who wishes to retain performance while supporting tracing in
23 their kernel should make it all the way to dyn 23 their kernel should make it all the way to dynamic ftrace support.
24 24
25 25
26 Prerequisites 26 Prerequisites
27 ------------- 27 -------------
28 28
29 Ftrace relies on these features being implemen 29 Ftrace relies on these features being implemented:
30 - STACKTRACE_SUPPORT - implement save_stack_ 30 - STACKTRACE_SUPPORT - implement save_stack_trace()
31 - TRACE_IRQFLAGS_SUPPORT - implement include 31 - TRACE_IRQFLAGS_SUPPORT - implement include/asm/irqflags.h
32 32
33 33
34 HAVE_FUNCTION_TRACER 34 HAVE_FUNCTION_TRACER
35 -------------------- 35 --------------------
36 36
37 You will need to implement the mcount and the 37 You will need to implement the mcount and the ftrace_stub functions.
38 38
39 The exact mcount symbol name will depend on yo 39 The exact mcount symbol name will depend on your toolchain. Some call it
40 "mcount", "_mcount", or even "__mcount". You 40 "mcount", "_mcount", or even "__mcount". You can probably figure it out by
41 running something like:: 41 running something like::
42 42
43 $ echo 'main(){}' | gcc -x c -S -o - - 43 $ echo 'main(){}' | gcc -x c -S -o - - -pg | grep mcount
44 call mcount 44 call mcount
45 45
46 We'll make the assumption below that the symbo 46 We'll make the assumption below that the symbol is "mcount" just to keep things
47 nice and simple in the examples. 47 nice and simple in the examples.
48 48
49 Keep in mind that the ABI that is in effect in 49 Keep in mind that the ABI that is in effect inside of the mcount function is
50 *highly* architecture/toolchain specific. We 50 *highly* architecture/toolchain specific. We cannot help you in this regard,
51 sorry. Dig up some old documentation and/or f 51 sorry. Dig up some old documentation and/or find someone more familiar than
52 you to bang ideas off of. Typically, register 52 you to bang ideas off of. Typically, register usage (argument/scratch/etc...)
53 is a major issue at this point, especially in 53 is a major issue at this point, especially in relation to the location of the
54 mcount call (before/after function prologue). 54 mcount call (before/after function prologue). You might also want to look at
55 how glibc has implemented the mcount function 55 how glibc has implemented the mcount function for your architecture. It might
56 be (semi-)relevant. 56 be (semi-)relevant.
57 57
58 The mcount function should check the function 58 The mcount function should check the function pointer ftrace_trace_function
59 to see if it is set to ftrace_stub. If it is, 59 to see if it is set to ftrace_stub. If it is, there is nothing for you to do,
60 so return immediately. If it isn't, then call 60 so return immediately. If it isn't, then call that function in the same way
61 the mcount function normally calls __mcount_in 61 the mcount function normally calls __mcount_internal -- the first argument is
62 the "frompc" while the second argument is the 62 the "frompc" while the second argument is the "selfpc" (adjusted to remove the
63 size of the mcount call that is embedded in th 63 size of the mcount call that is embedded in the function).
64 64
65 For example, if the function foo() calls bar() 65 For example, if the function foo() calls bar(), when the bar() function calls
66 mcount(), the arguments mcount() will pass to 66 mcount(), the arguments mcount() will pass to the tracer are:
67 67
68 - "frompc" - the address bar() will use to r 68 - "frompc" - the address bar() will use to return to foo()
69 - "selfpc" - the address bar() (with mcount( 69 - "selfpc" - the address bar() (with mcount() size adjustment)
70 70
71 Also keep in mind that this mcount function wi 71 Also keep in mind that this mcount function will be called *a lot*, so
72 optimizing for the default case of no tracer w 72 optimizing for the default case of no tracer will help the smooth running of
73 your system when tracing is disabled. So the 73 your system when tracing is disabled. So the start of the mcount function is
74 typically the bare minimum with checking thing 74 typically the bare minimum with checking things before returning. That also
75 means the code flow should usually be kept lin 75 means the code flow should usually be kept linear (i.e. no branching in the nop
76 case). This is of course an optimization and 76 case). This is of course an optimization and not a hard requirement.
77 77
78 Here is some pseudo code that should help (the 78 Here is some pseudo code that should help (these functions should actually be
79 implemented in assembly):: 79 implemented in assembly)::
80 80
81 void ftrace_stub(void) 81 void ftrace_stub(void)
82 { 82 {
83 return; 83 return;
84 } 84 }
85 85
86 void mcount(void) 86 void mcount(void)
87 { 87 {
88 /* save any bare state needed 88 /* save any bare state needed in order to do initial checking */
89 89
90 extern void (*ftrace_trace_fun 90 extern void (*ftrace_trace_function)(unsigned long, unsigned long);
91 if (ftrace_trace_function != f 91 if (ftrace_trace_function != ftrace_stub)
92 goto do_trace; 92 goto do_trace;
93 93
94 /* restore any bare state */ 94 /* restore any bare state */
95 95
96 return; 96 return;
97 97
98 do_trace: 98 do_trace:
99 99
100 /* save all state needed by th 100 /* save all state needed by the ABI (see paragraph above) */
101 101
102 unsigned long frompc = ...; 102 unsigned long frompc = ...;
103 unsigned long selfpc = <return 103 unsigned long selfpc = <return address> - MCOUNT_INSN_SIZE;
104 ftrace_trace_function(frompc, 104 ftrace_trace_function(frompc, selfpc);
105 105
106 /* restore all state needed by 106 /* restore all state needed by the ABI */
107 } 107 }
108 108
109 Don't forget to export mcount for modules ! 109 Don't forget to export mcount for modules !
110 :: 110 ::
111 111
112 extern void mcount(void); 112 extern void mcount(void);
113 EXPORT_SYMBOL(mcount); 113 EXPORT_SYMBOL(mcount);
114 114
115 115
116 HAVE_FUNCTION_GRAPH_TRACER 116 HAVE_FUNCTION_GRAPH_TRACER
117 -------------------------- 117 --------------------------
118 118
119 Deep breath ... time to do some real work. He 119 Deep breath ... time to do some real work. Here you will need to update the
120 mcount function to check ftrace graph function 120 mcount function to check ftrace graph function pointers, as well as implement
121 some functions to save (hijack) and restore th 121 some functions to save (hijack) and restore the return address.
122 122
123 The mcount function should check the function 123 The mcount function should check the function pointers ftrace_graph_return
124 (compare to ftrace_stub) and ftrace_graph_entr 124 (compare to ftrace_stub) and ftrace_graph_entry (compare to
125 ftrace_graph_entry_stub). If either of those 125 ftrace_graph_entry_stub). If either of those is not set to the relevant stub
126 function, call the arch-specific function ftra 126 function, call the arch-specific function ftrace_graph_caller which in turn
127 calls the arch-specific function prepare_ftrac 127 calls the arch-specific function prepare_ftrace_return. Neither of these
128 function names is strictly required, but you s 128 function names is strictly required, but you should use them anyway to stay
129 consistent across the architecture ports -- ea 129 consistent across the architecture ports -- easier to compare & contrast
130 things. 130 things.
131 131
132 The arguments to prepare_ftrace_return are sli 132 The arguments to prepare_ftrace_return are slightly different than what are
133 passed to ftrace_trace_function. The second a 133 passed to ftrace_trace_function. The second argument "selfpc" is the same,
134 but the first argument should be a pointer to 134 but the first argument should be a pointer to the "frompc". Typically this is
135 located on the stack. This allows the functio 135 located on the stack. This allows the function to hijack the return address
136 temporarily to have it point to the arch-speci 136 temporarily to have it point to the arch-specific function return_to_handler.
137 That function will simply call the common ftra 137 That function will simply call the common ftrace_return_to_handler function and
138 that will return the original return address w 138 that will return the original return address with which you can return to the
139 original call site. 139 original call site.
140 140
141 Here is the updated mcount pseudo code:: 141 Here is the updated mcount pseudo code::
142 142
143 void mcount(void) 143 void mcount(void)
144 { 144 {
145 ... 145 ...
146 if (ftrace_trace_function != f 146 if (ftrace_trace_function != ftrace_stub)
147 goto do_trace; 147 goto do_trace;
148 148
149 +#ifdef CONFIG_FUNCTION_GRAPH_TRACER 149 +#ifdef CONFIG_FUNCTION_GRAPH_TRACER
150 + extern void (*ftrace_graph_ret 150 + extern void (*ftrace_graph_return)(...);
151 + extern void (*ftrace_graph_ent 151 + extern void (*ftrace_graph_entry)(...);
152 + if (ftrace_graph_return != ftr 152 + if (ftrace_graph_return != ftrace_stub ||
153 + ftrace_graph_entry != ftra 153 + ftrace_graph_entry != ftrace_graph_entry_stub)
154 + ftrace_graph_caller(); 154 + ftrace_graph_caller();
155 +#endif 155 +#endif
156 156
157 /* restore any bare state */ 157 /* restore any bare state */
158 ... 158 ...
159 159
160 Here is the pseudo code for the new ftrace_gra 160 Here is the pseudo code for the new ftrace_graph_caller assembly function::
161 161
162 #ifdef CONFIG_FUNCTION_GRAPH_TRACER 162 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
163 void ftrace_graph_caller(void) 163 void ftrace_graph_caller(void)
164 { 164 {
165 /* save all state needed by th 165 /* save all state needed by the ABI */
166 166
167 unsigned long *frompc = &...; 167 unsigned long *frompc = &...;
168 unsigned long selfpc = <return 168 unsigned long selfpc = <return address> - MCOUNT_INSN_SIZE;
169 /* passing frame pointer up is 169 /* passing frame pointer up is optional -- see below */
170 prepare_ftrace_return(frompc, 170 prepare_ftrace_return(frompc, selfpc, frame_pointer);
171 171
172 /* restore all state needed by 172 /* restore all state needed by the ABI */
173 } 173 }
174 #endif 174 #endif
175 175
176 For information on how to implement prepare_ft 176 For information on how to implement prepare_ftrace_return(), simply look at the
177 x86 version (the frame pointer passing is opti 177 x86 version (the frame pointer passing is optional; see the next section for
178 more information). The only architecture-spec 178 more information). The only architecture-specific piece in it is the setup of
179 the fault recovery table (the asm(...) code). 179 the fault recovery table (the asm(...) code). The rest should be the same
180 across architectures. 180 across architectures.
181 181
182 Here is the pseudo code for the new return_to_ 182 Here is the pseudo code for the new return_to_handler assembly function. Note
183 that the ABI that applies here is different fr 183 that the ABI that applies here is different from what applies to the mcount
184 code. Since you are returning from a function 184 code. Since you are returning from a function (after the epilogue), you might
185 be able to skimp on things saved/restored (usu 185 be able to skimp on things saved/restored (usually just registers used to pass
186 return values). 186 return values).
187 :: 187 ::
188 188
189 #ifdef CONFIG_FUNCTION_GRAPH_TRACER 189 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
190 void return_to_handler(void) 190 void return_to_handler(void)
191 { 191 {
192 /* save all state needed by th 192 /* save all state needed by the ABI (see paragraph above) */
193 193
194 void (*original_return_point)( 194 void (*original_return_point)(void) = ftrace_return_to_handler();
195 195
196 /* restore all state needed by 196 /* restore all state needed by the ABI */
197 197
198 /* this is usually either a re 198 /* this is usually either a return or a jump */
199 original_return_point(); 199 original_return_point();
200 } 200 }
201 #endif 201 #endif
202 202
203 203
204 HAVE_FUNCTION_GRAPH_FP_TEST 204 HAVE_FUNCTION_GRAPH_FP_TEST
205 --------------------------- 205 ---------------------------
206 206
207 An arch may pass in a unique value (frame poin 207 An arch may pass in a unique value (frame pointer) to both the entering and
208 exiting of a function. On exit, the value is 208 exiting of a function. On exit, the value is compared and if it does not
209 match, then it will panic the kernel. This is 209 match, then it will panic the kernel. This is largely a sanity check for bad
210 code generation with gcc. If gcc for your por 210 code generation with gcc. If gcc for your port sanely updates the frame
211 pointer under different optimization levels, t 211 pointer under different optimization levels, then ignore this option.
212 212
213 However, adding support for it isn't terribly 213 However, adding support for it isn't terribly difficult. In your assembly code
214 that calls prepare_ftrace_return(), pass the f 214 that calls prepare_ftrace_return(), pass the frame pointer as the 3rd argument.
215 Then in the C version of that function, do wha 215 Then in the C version of that function, do what the x86 port does and pass it
216 along to ftrace_push_return_trace() instead of 216 along to ftrace_push_return_trace() instead of a stub value of 0.
217 217
218 Similarly, when you call ftrace_return_to_hand 218 Similarly, when you call ftrace_return_to_handler(), pass it the frame pointer.
219 219
>> 220 HAVE_FUNCTION_GRAPH_RET_ADDR_PTR
>> 221 --------------------------------
>> 222
>> 223 An arch may pass in a pointer to the return address on the stack. This
>> 224 prevents potential stack unwinding issues where the unwinder gets out of
>> 225 sync with ret_stack and the wrong addresses are reported by
>> 226 ftrace_graph_ret_addr().
>> 227
>> 228 Adding support for it is easy: just define the macro in asm/ftrace.h and
>> 229 pass the return address pointer as the 'retp' argument to
>> 230 ftrace_push_return_trace().
>> 231
>> 232 HAVE_FTRACE_NMI_ENTER
>> 233 ---------------------
>> 234
>> 235 If you can't trace NMI functions, then skip this option.
>> 236
>> 237 <details to be filled>
>> 238
>> 239
220 HAVE_SYSCALL_TRACEPOINTS 240 HAVE_SYSCALL_TRACEPOINTS
221 ------------------------ 241 ------------------------
222 242
223 You need very few things to get the syscalls t 243 You need very few things to get the syscalls tracing in an arch.
224 244
225 - Support HAVE_ARCH_TRACEHOOK (see arch/Kcon 245 - Support HAVE_ARCH_TRACEHOOK (see arch/Kconfig).
226 - Have a NR_syscalls variable in <asm/unistd 246 - Have a NR_syscalls variable in <asm/unistd.h> that provides the number
227 of syscalls supported by the arch. 247 of syscalls supported by the arch.
228 - Support the TIF_SYSCALL_TRACEPOINT thread 248 - Support the TIF_SYSCALL_TRACEPOINT thread flags.
229 - Put the trace_sys_enter() and trace_sys_ex 249 - Put the trace_sys_enter() and trace_sys_exit() tracepoints calls from ptrace
230 in the ptrace syscalls tracing path. 250 in the ptrace syscalls tracing path.
231 - If the system call table on this arch is m 251 - If the system call table on this arch is more complicated than a simple array
232 of addresses of the system calls, implemen 252 of addresses of the system calls, implement an arch_syscall_addr to return
233 the address of a given system call. 253 the address of a given system call.
234 - If the symbol names of the system calls do 254 - If the symbol names of the system calls do not match the function names on
235 this arch, define ARCH_HAS_SYSCALL_MATCH_S 255 this arch, define ARCH_HAS_SYSCALL_MATCH_SYM_NAME in asm/ftrace.h and
236 implement arch_syscall_match_sym_name with 256 implement arch_syscall_match_sym_name with the appropriate logic to return
237 true if the function name corresponds with 257 true if the function name corresponds with the symbol name.
238 - Tag this arch as HAVE_SYSCALL_TRACEPOINTS. 258 - Tag this arch as HAVE_SYSCALL_TRACEPOINTS.
239 259
240 260
241 HAVE_FTRACE_MCOUNT_RECORD 261 HAVE_FTRACE_MCOUNT_RECORD
242 ------------------------- 262 -------------------------
243 263
244 See scripts/recordmcount.pl for more info. Ju 264 See scripts/recordmcount.pl for more info. Just fill in the arch-specific
245 details for how to locate the addresses of mco 265 details for how to locate the addresses of mcount call sites via objdump.
246 This option doesn't make much sense without al 266 This option doesn't make much sense without also implementing dynamic ftrace.
247 267
248 268
249 HAVE_DYNAMIC_FTRACE 269 HAVE_DYNAMIC_FTRACE
250 ------------------- 270 -------------------
251 271
252 You will first need HAVE_FTRACE_MCOUNT_RECORD 272 You will first need HAVE_FTRACE_MCOUNT_RECORD and HAVE_FUNCTION_TRACER, so
253 scroll your reader back up if you got over eag 273 scroll your reader back up if you got over eager.
254 274
255 Once those are out of the way, you will need t 275 Once those are out of the way, you will need to implement:
256 - asm/ftrace.h: 276 - asm/ftrace.h:
257 - MCOUNT_ADDR 277 - MCOUNT_ADDR
258 - ftrace_call_adjust() 278 - ftrace_call_adjust()
259 - struct dyn_arch_ftrace{} 279 - struct dyn_arch_ftrace{}
260 - asm code: 280 - asm code:
261 - mcount() (new stub) 281 - mcount() (new stub)
262 - ftrace_caller() 282 - ftrace_caller()
263 - ftrace_call() 283 - ftrace_call()
264 - ftrace_stub() 284 - ftrace_stub()
265 - C code: 285 - C code:
266 - ftrace_dyn_arch_init() 286 - ftrace_dyn_arch_init()
267 - ftrace_make_nop() 287 - ftrace_make_nop()
268 - ftrace_make_call() 288 - ftrace_make_call()
269 - ftrace_update_ftrace_func() 289 - ftrace_update_ftrace_func()
270 290
271 First you will need to fill out some arch deta 291 First you will need to fill out some arch details in your asm/ftrace.h.
272 292
273 Define MCOUNT_ADDR as the address of your mcou 293 Define MCOUNT_ADDR as the address of your mcount symbol similar to::
274 294
275 #define MCOUNT_ADDR ((unsigned long)mc 295 #define MCOUNT_ADDR ((unsigned long)mcount)
276 296
277 Since no one else will have a decl for that fu 297 Since no one else will have a decl for that function, you will need to::
278 298
279 extern void mcount(void); 299 extern void mcount(void);
280 300
281 You will also need the helper function ftrace_ 301 You will also need the helper function ftrace_call_adjust(). Most people
282 will be able to stub it out like so:: 302 will be able to stub it out like so::
283 303
284 static inline unsigned long ftrace_cal 304 static inline unsigned long ftrace_call_adjust(unsigned long addr)
285 { 305 {
286 return addr; 306 return addr;
287 } 307 }
288 308
289 <details to be filled> 309 <details to be filled>
290 310
291 Lastly you will need the custom dyn_arch_ftrac 311 Lastly you will need the custom dyn_arch_ftrace structure. If you need
292 some extra state when runtime patching arbitra 312 some extra state when runtime patching arbitrary call sites, this is the
293 place. For now though, create an empty struct 313 place. For now though, create an empty struct::
294 314
295 struct dyn_arch_ftrace { 315 struct dyn_arch_ftrace {
296 /* No extra data needed */ 316 /* No extra data needed */
297 }; 317 };
298 318
299 With the header out of the way, we can fill ou 319 With the header out of the way, we can fill out the assembly code. While we
300 did already create a mcount() function earlier 320 did already create a mcount() function earlier, dynamic ftrace only wants a
301 stub function. This is because the mcount() w 321 stub function. This is because the mcount() will only be used during boot
302 and then all references to it will be patched 322 and then all references to it will be patched out never to return. Instead,
303 the guts of the old mcount() will be used to c 323 the guts of the old mcount() will be used to create a new ftrace_caller()
304 function. Because the two are hard to merge, 324 function. Because the two are hard to merge, it will most likely be a lot
305 easier to have two separate definitions split 325 easier to have two separate definitions split up by #ifdefs. Same goes for
306 the ftrace_stub() as that will now be inlined 326 the ftrace_stub() as that will now be inlined in ftrace_caller().
307 327
308 Before we get confused anymore, let's check ou 328 Before we get confused anymore, let's check out some pseudo code so you can
309 implement your own stuff in assembly:: 329 implement your own stuff in assembly::
310 330
311 void mcount(void) 331 void mcount(void)
312 { 332 {
313 return; 333 return;
314 } 334 }
315 335
316 void ftrace_caller(void) 336 void ftrace_caller(void)
317 { 337 {
318 /* save all state needed by th 338 /* save all state needed by the ABI (see paragraph above) */
319 339
320 unsigned long frompc = ...; 340 unsigned long frompc = ...;
321 unsigned long selfpc = <return 341 unsigned long selfpc = <return address> - MCOUNT_INSN_SIZE;
322 342
323 ftrace_call: 343 ftrace_call:
324 ftrace_stub(frompc, selfpc); 344 ftrace_stub(frompc, selfpc);
325 345
326 /* restore all state needed by 346 /* restore all state needed by the ABI */
327 347
328 ftrace_stub: 348 ftrace_stub:
329 return; 349 return;
330 } 350 }
331 351
332 This might look a little odd at first, but kee 352 This might look a little odd at first, but keep in mind that we will be runtime
333 patching multiple things. First, only functio 353 patching multiple things. First, only functions that we actually want to trace
334 will be patched to call ftrace_caller(). Seco 354 will be patched to call ftrace_caller(). Second, since we only have one tracer
335 active at a time, we will patch the ftrace_cal 355 active at a time, we will patch the ftrace_caller() function itself to call the
336 specific tracer in question. That is the poin 356 specific tracer in question. That is the point of the ftrace_call label.
337 357
338 With that in mind, let's move on to the C code 358 With that in mind, let's move on to the C code that will actually be doing the
339 runtime patching. You'll need a little knowle 359 runtime patching. You'll need a little knowledge of your arch's opcodes in
340 order to make it through the next section. 360 order to make it through the next section.
341 361
342 Every arch has an init callback function. If 362 Every arch has an init callback function. If you need to do something early on
343 to initialize some state, this is the time to 363 to initialize some state, this is the time to do that. Otherwise, this simple
344 function below should be sufficient for most p 364 function below should be sufficient for most people::
345 365
346 int __init ftrace_dyn_arch_init(void) 366 int __init ftrace_dyn_arch_init(void)
347 { 367 {
348 return 0; 368 return 0;
349 } 369 }
350 370
351 There are two functions that are used to do ru 371 There are two functions that are used to do runtime patching of arbitrary
352 functions. The first is used to turn the mcou 372 functions. The first is used to turn the mcount call site into a nop (which
353 is what helps us retain runtime performance wh 373 is what helps us retain runtime performance when not tracing). The second is
354 used to turn the mcount call site into a call 374 used to turn the mcount call site into a call to an arbitrary location (but
355 typically that is ftracer_caller()). See the 375 typically that is ftracer_caller()). See the general function definition in
356 linux/ftrace.h for the functions:: 376 linux/ftrace.h for the functions::
357 377
358 ftrace_make_nop() 378 ftrace_make_nop()
359 ftrace_make_call() 379 ftrace_make_call()
360 380
361 The rec->ip value is the address of the mcount 381 The rec->ip value is the address of the mcount call site that was collected
362 by the scripts/recordmcount.pl during build ti 382 by the scripts/recordmcount.pl during build time.
363 383
364 The last function is used to do runtime patchi 384 The last function is used to do runtime patching of the active tracer. This
365 will be modifying the assembly code at the loc 385 will be modifying the assembly code at the location of the ftrace_call symbol
366 inside of the ftrace_caller() function. So yo 386 inside of the ftrace_caller() function. So you should have sufficient padding
367 at that location to support the new function c 387 at that location to support the new function calls you'll be inserting. Some
368 people will be using a "call" type instruction 388 people will be using a "call" type instruction while others will be using a
369 "branch" type instruction. Specifically, the 389 "branch" type instruction. Specifically, the function is::
370 390
371 ftrace_update_ftrace_func() 391 ftrace_update_ftrace_func()
372 392
373 393
374 HAVE_DYNAMIC_FTRACE + HAVE_FUNCTION_GRAPH_TRAC 394 HAVE_DYNAMIC_FTRACE + HAVE_FUNCTION_GRAPH_TRACER
375 ---------------------------------------------- 395 ------------------------------------------------
376 396
377 The function grapher needs a few tweaks in ord 397 The function grapher needs a few tweaks in order to work with dynamic ftrace.
378 Basically, you will need to: 398 Basically, you will need to:
379 399
380 - update: 400 - update:
381 - ftrace_caller() 401 - ftrace_caller()
382 - ftrace_graph_call() 402 - ftrace_graph_call()
383 - ftrace_graph_caller() 403 - ftrace_graph_caller()
384 - implement: 404 - implement:
385 - ftrace_enable_ftrace_graph_c 405 - ftrace_enable_ftrace_graph_caller()
386 - ftrace_disable_ftrace_graph_ 406 - ftrace_disable_ftrace_graph_caller()
387 407
388 <details to be filled> 408 <details to be filled>
389 409
390 Quick notes: 410 Quick notes:
391 411
392 - add a nop stub after the ftrace_call 412 - add a nop stub after the ftrace_call location named ftrace_graph_call;
393 stub needs to be large enough to sup 413 stub needs to be large enough to support a call to ftrace_graph_caller()
394 - update ftrace_graph_caller() to work 414 - update ftrace_graph_caller() to work with being called by the new
395 ftrace_caller() since some semantics 415 ftrace_caller() since some semantics may have changed
396 - ftrace_enable_ftrace_graph_caller() 416 - ftrace_enable_ftrace_graph_caller() will runtime patch the
397 ftrace_graph_call location with a ca 417 ftrace_graph_call location with a call to ftrace_graph_caller()
398 - ftrace_disable_ftrace_graph_caller() 418 - ftrace_disable_ftrace_graph_caller() will runtime patch the
399 ftrace_graph_call location with nops 419 ftrace_graph_call location with nops
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.