nsphinx.addnodesdocument)}( rawsourcechildren]( translations LanguagesNode)}(hhh](h pending_xref)}(hhh]docutils.nodesTextChinese (Simplified)}parenthsba attributes}(ids]classes]names]dupnames]backrefs] refdomainstdreftypedoc reftarget'/translations/zh_CN/trace/ftrace-designmodnameN classnameN refexplicitutagnamehhh ubh)}(hhh]hChinese (Traditional)}hh2sbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/zh_TW/trace/ftrace-designmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hItalian}hhFsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/it_IT/trace/ftrace-designmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hJapanese}hhZsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/ja_JP/trace/ftrace-designmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hKorean}hhnsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/ko_KR/trace/ftrace-designmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hSpanish}hhsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/sp_SP/trace/ftrace-designmodnameN classnameN refexplicituh1hhh ubeh}(h]h ]h"]h$]h&]current_languageEnglishuh1h hh _documenthsourceNlineNubhsection)}(hhh](htitle)}(hFunction Tracer Designh]hFunction Tracer Design}(hhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhhhA/var/lib/git/docbuild/linux/Documentation/trace/ftrace-design.rsthKubh field_list)}(hhh]hfield)}(hhh](h field_name)}(hAuthorh]hAuthor}(hhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhhhKubh field_body)}(hMike Frysinger h]h paragraph)}(hMike Frysingerh]hMike Frysinger}(hhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhhubah}(h]h ]h"]h$]h&]uh1hhhubeh}(h]h ]h"]h$]h&]uh1hhhhKhhhhubah}(h]h ]h"]h$]h&]uh1hhhhhhhhKubhcaution)}(heThis document is out of date. Some of the description below doesn't match current implementation now.h]h)}(heThis document is out of date. Some of the description below doesn't match current implementation now.h]hgThis document is out of date. Some of the description below doesn’t match current implementation now.}(hhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhhubah}(h]h ]h"]h$]h&]uh1hhhhhhhhNubh)}(hhh](h)}(h Introductionh]h Introduction}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhK ubh)}(hHere we will cover the architecture pieces that the common function tracing code relies on for proper functioning. Things are broken down into increasing complexity so that you can start simple and at least get basic functionality.h]hHere we will cover the architecture pieces that the common function tracing code relies on for proper functioning. Things are broken down into increasing complexity so that you can start simple and at least get basic functionality.}(hj$hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubh)}(hNote that this focuses on architecture implementation details only. If you want more explanation of a feature in terms of common code, review the common ftrace.txt file.h]hNote that this focuses on architecture implementation details only. If you want more explanation of a feature in terms of common code, review the common ftrace.txt file.}(hj2hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubh)}(hIdeally, everyone who wishes to retain performance while supporting tracing in their kernel should make it all the way to dynamic ftrace support.h]hIdeally, everyone who wishes to retain performance while supporting tracing in their kernel should make it all the way to dynamic ftrace support.}(hj@hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubeh}(h] introductionah ]h"] introductionah$]h&]uh1hhhhhhhhK ubh)}(hhh](h)}(h Prerequisitesh]h Prerequisites}(hjYhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjVhhhhhKubhdefinition_list)}(hhh]hdefinition_list_item)}(hFtrace relies on these features being implemented: - STACKTRACE_SUPPORT - implement save_stack_trace() - TRACE_IRQFLAGS_SUPPORT - implement include/asm/irqflags.h h](hterm)}(h2Ftrace relies on these features being implemented:h]h2Ftrace relies on these features being implemented:}(hjthhhNhNubah}(h]h ]h"]h$]h&]uh1jrhhhK hjnubh definition)}(hhh]h bullet_list)}(hhh](h list_item)}(h1STACKTRACE_SUPPORT - implement save_stack_trace()h]h)}(hjh]h1STACKTRACE_SUPPORT - implement save_stack_trace()}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(h;TRACE_IRQFLAGS_SUPPORT - implement include/asm/irqflags.h h]h)}(h9TRACE_IRQFLAGS_SUPPORT - implement include/asm/irqflags.hh]h9TRACE_IRQFLAGS_SUPPORT - implement include/asm/irqflags.h}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]bullet-uh1jhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjnubeh}(h]h ]h"]h$]h&]uh1jlhhhK hjiubah}(h]h ]h"]h$]h&]uh1jghjVhhhNhNubeh}(h] prerequisitesah ]h"] prerequisitesah$]h&]uh1hhhhhhhhKubh)}(hhh](h)}(hHAVE_FUNCTION_TRACERh]hHAVE_FUNCTION_TRACER}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhK#ubh)}(hDYou will need to implement the mcount and the ftrace_stub functions.h]hDYou will need to implement the mcount and the ftrace_stub functions.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK%hjhhubh)}(hThe exact mcount symbol name will depend on your toolchain. Some call it "mcount", "_mcount", or even "__mcount". You can probably figure it out by running something like::h]hThe exact mcount symbol name will depend on your toolchain. Some call it “mcount”, “_mcount”, or even “__mcount”. You can probably figure it out by running something like:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK'hjhhubh literal_block)}(hO$ echo 'main(){}' | gcc -x c -S -o - - -pg | grep mcount call mcounth]hO$ echo 'main(){}' | gcc -x c -S -o - - -pg | grep mcount call mcount}hjsbah}(h]h ]h"]h$]h&] xml:spacepreserveuh1j hhhK+hjhhubh)}(hpWe'll make the assumption below that the symbol is "mcount" just to keep things nice and simple in the examples.h]hvWe’ll make the assumption below that the symbol is “mcount” just to keep things nice and simple in the examples.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK.hjhhubh)}(hX5Keep in mind that the ABI that is in effect inside of the mcount function is *highly* architecture/toolchain specific. We cannot help you in this regard, sorry. Dig up some old documentation and/or find someone more familiar than you to bang ideas off of. Typically, register usage (argument/scratch/etc...) is a major issue at this point, especially in relation to the location of the mcount call (before/after function prologue). You might also want to look at how glibc has implemented the mcount function for your architecture. It might be (semi-)relevant.h](hMKeep in mind that the ABI that is in effect inside of the mcount function is }(hj,hhhNhNubhemphasis)}(h*highly*h]hhighly}(hj6hhhNhNubah}(h]h ]h"]h$]h&]uh1j4hj,ubhX architecture/toolchain specific. We cannot help you in this regard, sorry. Dig up some old documentation and/or find someone more familiar than you to bang ideas off of. Typically, register usage (argument/scratch/etc...) is a major issue at this point, especially in relation to the location of the mcount call (before/after function prologue). You might also want to look at how glibc has implemented the mcount function for your architecture. It might be (semi-)relevant.}(hj,hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhK1hjhhubh)}(hXThe mcount function should check the function pointer ftrace_trace_function to see if it is set to ftrace_stub. If it is, there is nothing for you to do, so return immediately. If it isn't, then call that function in the same way the mcount function normally calls __mcount_internal -- the first argument is the "frompc" while the second argument is the "selfpc" (adjusted to remove the size of the mcount call that is embedded in the function).h]hXThe mcount function should check the function pointer ftrace_trace_function to see if it is set to ftrace_stub. If it is, there is nothing for you to do, so return immediately. If it isn’t, then call that function in the same way the mcount function normally calls __mcount_internal -- the first argument is the “frompc” while the second argument is the “selfpc” (adjusted to remove the size of the mcount call that is embedded in the function).}(hjNhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK:hjhhubh)}(hFor example, if the function foo() calls bar(), when the bar() function calls mcount(), the arguments mcount() will pass to the tracer are:h]hFor example, if the function foo() calls bar(), when the bar() function calls mcount(), the arguments mcount() will pass to the tracer are:}(hj\hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKAhjhhubh block_quote)}(hz- "frompc" - the address bar() will use to return to foo() - "selfpc" - the address bar() (with mcount() size adjustment) h]j)}(hhh](j)}(h8"frompc" - the address bar() will use to return to foo()h]h)}(hjuh]h<“frompc” - the address bar() will use to return to foo()}(hjwhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKDhjsubah}(h]h ]h"]h$]h&]uh1jhjpubj)}(h="selfpc" - the address bar() (with mcount() size adjustment) h]h)}(h<"selfpc" - the address bar() (with mcount() size adjustment)h]h@“selfpc” - the address bar() (with mcount() size adjustment)}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKEhjubah}(h]h ]h"]h$]h&]uh1jhjpubeh}(h]h ]h"]h$]h&]jjuh1jhhhKDhjlubah}(h]h ]h"]h$]h&]uh1jjhhhKDhjhhubh)}(hXAlso keep in mind that this mcount function will be called *a lot*, so optimizing for the default case of no tracer will help the smooth running of your system when tracing is disabled. So the start of the mcount function is typically the bare minimum with checking things before returning. That also means the code flow should usually be kept linear (i.e. no branching in the nop case). This is of course an optimization and not a hard requirement.h](h;Also keep in mind that this mcount function will be called }(hjhhhNhNubj5)}(h*a lot*h]ha lot}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1j4hjubhX, so optimizing for the default case of no tracer will help the smooth running of your system when tracing is disabled. So the start of the mcount function is typically the bare minimum with checking things before returning. That also means the code flow should usually be kept linear (i.e. no branching in the nop case). This is of course an optimization and not a hard requirement.}(hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKGhjhhubh)}(hhHere is some pseudo code that should help (these functions should actually be implemented in assembly)::h]hgHere is some pseudo code that should help (these functions should actually be implemented in assembly):}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKNhjhhubj )}(hX{void ftrace_stub(void) { return; } void mcount(void) { /* save any bare state needed in order to do initial checking */ extern void (*ftrace_trace_function)(unsigned long, unsigned long); if (ftrace_trace_function != ftrace_stub) goto do_trace; /* restore any bare state */ return; do_trace: /* save all state needed by the ABI (see paragraph above) */ unsigned long frompc = ...; unsigned long selfpc = - MCOUNT_INSN_SIZE; ftrace_trace_function(frompc, selfpc); /* restore all state needed by the ABI */ }h]hX{void ftrace_stub(void) { return; } void mcount(void) { /* save any bare state needed in order to do initial checking */ extern void (*ftrace_trace_function)(unsigned long, unsigned long); if (ftrace_trace_function != ftrace_stub) goto do_trace; /* restore any bare state */ return; do_trace: /* save all state needed by the ABI (see paragraph above) */ unsigned long frompc = ...; unsigned long selfpc = - MCOUNT_INSN_SIZE; ftrace_trace_function(frompc, selfpc); /* restore all state needed by the ABI */ }}hjsbah}(h]h ]h"]h$]h&]jjuh1j hhhKQhjhhubh)}(h.Don't forget to export mcount for modules ! ::h]h-Don’t forget to export mcount for modules !}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKmhjhhubj )}(h0extern void mcount(void); EXPORT_SYMBOL(mcount);h]h0extern void mcount(void); EXPORT_SYMBOL(mcount);}hjsbah}(h]h ]h"]h$]h&]jjuh1j hhhKphjhhubeh}(h]have-function-tracerah ]h"]have_function_tracerah$]h&]uh1hhhhhhhhK#ubh)}(hhh](h)}(hHAVE_FUNCTION_GRAPH_TRACERh]hHAVE_FUNCTION_GRAPH_TRACER}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhKuubh)}(hDeep breath ... time to do some real work. Here you will need to update the mcount function to check ftrace graph function pointers, as well as implement some functions to save (hijack) and restore the return address.h]hDeep breath ... time to do some real work. Here you will need to update the mcount function to check ftrace graph function pointers, as well as implement some functions to save (hijack) and restore the return address.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKwhjhhubh)}(hXThe mcount function should check the function pointers ftrace_graph_return (compare to ftrace_stub) and ftrace_graph_entry (compare to ftrace_graph_entry_stub). If either of those is not set to the relevant stub function, call the arch-specific function ftrace_graph_caller which in turn calls the arch-specific function prepare_ftrace_return. Neither of these function names is strictly required, but you should use them anyway to stay consistent across the architecture ports -- easier to compare & contrast things.h]hXThe mcount function should check the function pointers ftrace_graph_return (compare to ftrace_stub) and ftrace_graph_entry (compare to ftrace_graph_entry_stub). If either of those is not set to the relevant stub function, call the arch-specific function ftrace_graph_caller which in turn calls the arch-specific function prepare_ftrace_return. Neither of these function names is strictly required, but you should use them anyway to stay consistent across the architecture ports -- easier to compare & contrast things.}(hj-hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK{hjhhubh)}(hX3The arguments to prepare_ftrace_return are slightly different than what are passed to ftrace_trace_function. The second argument "selfpc" is the same, but the first argument should be a pointer to the "frompc". Typically this is located on the stack. This allows the function to hijack the return address temporarily to have it point to the arch-specific function return_to_handler. That function will simply call the common ftrace_return_to_handler function and that will return the original return address with which you can return to the original call site.h]hX;The arguments to prepare_ftrace_return are slightly different than what are passed to ftrace_trace_function. The second argument “selfpc” is the same, but the first argument should be a pointer to the “frompc”. Typically this is located on the stack. This allows the function to hijack the return address temporarily to have it point to the arch-specific function return_to_handler. That function will simply call the common ftrace_return_to_handler function and that will return the original return address with which you can return to the original call site.}(hj;hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubh)}(h(Here is the updated mcount pseudo code::h]h'Here is the updated mcount pseudo code:}(hjIhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubj )}(hXvoid mcount(void) { ... if (ftrace_trace_function != ftrace_stub) goto do_trace; +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + extern void (*ftrace_graph_return)(...); + extern void (*ftrace_graph_entry)(...); + if (ftrace_graph_return != ftrace_stub || + ftrace_graph_entry != ftrace_graph_entry_stub) + ftrace_graph_caller(); +#endif /* restore any bare state */ ...h]hXvoid mcount(void) { ... if (ftrace_trace_function != ftrace_stub) goto do_trace; +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + extern void (*ftrace_graph_return)(...); + extern void (*ftrace_graph_entry)(...); + if (ftrace_graph_return != ftrace_stub || + ftrace_graph_entry != ftrace_graph_entry_stub) + ftrace_graph_caller(); +#endif /* restore any bare state */ ...}hjWsbah}(h]h ]h"]h$]h&]jjuh1j hhhKhjhhubh)}(hKHere is the pseudo code for the new ftrace_graph_caller assembly function::h]hJHere is the pseudo code for the new ftrace_graph_caller assembly function:}(hjehhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubj )}(hX#ifdef CONFIG_FUNCTION_GRAPH_TRACER void ftrace_graph_caller(void) { /* save all state needed by the ABI */ unsigned long *frompc = &...; unsigned long selfpc = - MCOUNT_INSN_SIZE; /* passing frame pointer up is optional -- see below */ prepare_ftrace_return(frompc, selfpc, frame_pointer); /* restore all state needed by the ABI */ } #endifh]hX#ifdef CONFIG_FUNCTION_GRAPH_TRACER void ftrace_graph_caller(void) { /* save all state needed by the ABI */ unsigned long *frompc = &...; unsigned long selfpc = - MCOUNT_INSN_SIZE; /* passing frame pointer up is optional -- see below */ prepare_ftrace_return(frompc, selfpc, frame_pointer); /* restore all state needed by the ABI */ } #endif}hjssbah}(h]h ]h"]h$]h&]jjuh1j hhhKhjhhubh)}(hXLFor information on how to implement prepare_ftrace_return(), simply look at the x86 version (the frame pointer passing is optional; see the next section for more information). The only architecture-specific piece in it is the setup of the fault recovery table (the asm(...) code). The rest should be the same across architectures.h]hXLFor information on how to implement prepare_ftrace_return(), simply look at the x86 version (the frame pointer passing is optional; see the next section for more information). The only architecture-specific piece in it is the setup of the fault recovery table (the asm(...) code). The rest should be the same across architectures.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubh)}(hXKHere is the pseudo code for the new return_to_handler assembly function. Note that the ABI that applies here is different from what applies to the mcount code. Since you are returning from a function (after the epilogue), you might be able to skimp on things saved/restored (usually just registers used to pass return values). ::h]hXHHere is the pseudo code for the new return_to_handler assembly function. Note that the ABI that applies here is different from what applies to the mcount code. Since you are returning from a function (after the epilogue), you might be able to skimp on things saved/restored (usually just registers used to pass return values).}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubj )}(hXh#ifdef CONFIG_FUNCTION_GRAPH_TRACER void return_to_handler(void) { /* save all state needed by the ABI (see paragraph above) */ void (*original_return_point)(void) = ftrace_return_to_handler(); /* restore all state needed by the ABI */ /* this is usually either a return or a jump */ original_return_point(); } #endifh]hXh#ifdef CONFIG_FUNCTION_GRAPH_TRACER void return_to_handler(void) { /* save all state needed by the ABI (see paragraph above) */ void (*original_return_point)(void) = ftrace_return_to_handler(); /* restore all state needed by the ABI */ /* this is usually either a return or a jump */ original_return_point(); } #endif}hjsbah}(h]h ]h"]h$]h&]jjuh1j hhhKhjhhubeh}(h]have-function-graph-tracerah ]h"]have_function_graph_tracerah$]h&]uh1hhhhhhhhKuubh)}(hhh](h)}(hHAVE_FUNCTION_GRAPH_FP_TESTh]hHAVE_FUNCTION_GRAPH_FP_TEST}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhKubh)}(hXrAn arch may pass in a unique value (frame pointer) to both the entering and exiting of a function. On exit, the value is compared and if it does not match, then it will panic the kernel. This is largely a sanity check for bad code generation with gcc. If gcc for your port sanely updates the frame pointer under different optimization levels, then ignore this option.h]hXrAn arch may pass in a unique value (frame pointer) to both the entering and exiting of a function. On exit, the value is compared and if it does not match, then it will panic the kernel. This is largely a sanity check for bad code generation with gcc. If gcc for your port sanely updates the frame pointer under different optimization levels, then ignore this option.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubh)}(hX/However, adding support for it isn't terribly difficult. In your assembly code that calls prepare_ftrace_return(), pass the frame pointer as the 3rd argument. Then in the C version of that function, do what the x86 port does and pass it along to ftrace_push_return_trace() instead of a stub value of 0.h]hX1However, adding support for it isn’t terribly difficult. In your assembly code that calls prepare_ftrace_return(), pass the frame pointer as the 3rd argument. Then in the C version of that function, do what the x86 port does and pass it along to ftrace_push_return_trace() instead of a stub value of 0.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubh)}(hOSimilarly, when you call ftrace_return_to_handler(), pass it the frame pointer.h]hOSimilarly, when you call ftrace_return_to_handler(), pass it the frame pointer.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubeh}(h]have-function-graph-fp-testah ]h"]have_function_graph_fp_testah$]h&]uh1hhhhhhhhKubh)}(hhh](h)}(hHAVE_SYSCALL_TRACEPOINTSh]hHAVE_SYSCALL_TRACEPOINTS}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhKubh)}(h@You need very few things to get the syscalls tracing in an arch.h]h@You need very few things to get the syscalls tracing in an arch.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubjk)}(hXZ- Support HAVE_ARCH_TRACEHOOK (see arch/Kconfig). - Have a NR_syscalls variable in that provides the number of syscalls supported by the arch. - Support the TIF_SYSCALL_TRACEPOINT thread flags. - Put the trace_sys_enter() and trace_sys_exit() tracepoints calls from ptrace in the ptrace syscalls tracing path. - If the system call table on this arch is more complicated than a simple array of addresses of the system calls, implement an arch_syscall_addr to return the address of a given system call. - If the symbol names of the system calls do not match the function names on this arch, define ARCH_HAS_SYSCALL_MATCH_SYM_NAME in asm/ftrace.h and implement arch_syscall_match_sym_name with the appropriate logic to return true if the function name corresponds with the symbol name. - Tag this arch as HAVE_SYSCALL_TRACEPOINTS. h]j)}(hhh](j)}(h/Support HAVE_ARCH_TRACEHOOK (see arch/Kconfig).h]h)}(hjh]h/Support HAVE_ARCH_TRACEHOOK (see arch/Kconfig).}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(hiHave a NR_syscalls variable in that provides the number of syscalls supported by the arch.h]h)}(hiHave a NR_syscalls variable in that provides the number of syscalls supported by the arch.h]hiHave a NR_syscalls variable in that provides the number of syscalls supported by the arch.}(hj7hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhj3ubah}(h]h ]h"]h$]h&]uh1jhjubj)}(h0Support the TIF_SYSCALL_TRACEPOINT thread flags.h]h)}(hjMh]h0Support the TIF_SYSCALL_TRACEPOINT thread flags.}(hjOhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjKubah}(h]h ]h"]h$]h&]uh1jhjubj)}(hqPut the trace_sys_enter() and trace_sys_exit() tracepoints calls from ptrace in the ptrace syscalls tracing path.h]h)}(hqPut the trace_sys_enter() and trace_sys_exit() tracepoints calls from ptrace in the ptrace syscalls tracing path.h]hqPut the trace_sys_enter() and trace_sys_exit() tracepoints calls from ptrace in the ptrace syscalls tracing path.}(hjfhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjbubah}(h]h ]h"]h$]h&]uh1jhjubj)}(hIf the system call table on this arch is more complicated than a simple array of addresses of the system calls, implement an arch_syscall_addr to return the address of a given system call.h]h)}(hIf the system call table on this arch is more complicated than a simple array of addresses of the system calls, implement an arch_syscall_addr to return the address of a given system call.h]hIf the system call table on this arch is more complicated than a simple array of addresses of the system calls, implement an arch_syscall_addr to return the address of a given system call.}(hj~hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjzubah}(h]h ]h"]h$]h&]uh1jhjubj)}(hXIf the symbol names of the system calls do not match the function names on this arch, define ARCH_HAS_SYSCALL_MATCH_SYM_NAME in asm/ftrace.h and implement arch_syscall_match_sym_name with the appropriate logic to return true if the function name corresponds with the symbol name.h]h)}(hXIf the symbol names of the system calls do not match the function names on this arch, define ARCH_HAS_SYSCALL_MATCH_SYM_NAME in asm/ftrace.h and implement arch_syscall_match_sym_name with the appropriate logic to return true if the function name corresponds with the symbol name.h]hXIf the symbol names of the system calls do not match the function names on this arch, define ARCH_HAS_SYSCALL_MATCH_SYM_NAME in asm/ftrace.h and implement arch_syscall_match_sym_name with the appropriate logic to return true if the function name corresponds with the symbol name.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(h,Tag this arch as HAVE_SYSCALL_TRACEPOINTS. h]h)}(h*Tag this arch as HAVE_SYSCALL_TRACEPOINTS.h]h*Tag this arch as HAVE_SYSCALL_TRACEPOINTS.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]jjuh1jhhhKhjubah}(h]h ]h"]h$]h&]uh1jjhhhKhjhhubeh}(h]have-syscall-tracepointsah ]h"]have_syscall_tracepointsah$]h&]uh1hhhhhhhhKubh)}(hhh](h)}(hHAVE_FTRACE_MCOUNT_RECORDh]hHAVE_FTRACE_MCOUNT_RECORD}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhKubh)}(hSee scripts/recordmcount.pl for more info. Just fill in the arch-specific details for how to locate the addresses of mcount call sites via objdump. This option doesn't make much sense without also implementing dynamic ftrace.h]hSee scripts/recordmcount.pl for more info. Just fill in the arch-specific details for how to locate the addresses of mcount call sites via objdump. This option doesn’t make much sense without also implementing dynamic ftrace.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubeh}(h]have-ftrace-mcount-recordah ]h"]have_ftrace_mcount_recordah$]h&]uh1hhhhhhhhKubh)}(hhh](h)}(hHAVE_DYNAMIC_FTRACEh]hHAVE_DYNAMIC_FTRACE}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhKubh)}(h|You will first need HAVE_FTRACE_MCOUNT_RECORD and HAVE_FUNCTION_TRACER, so scroll your reader back up if you got over eager.h]h|You will first need HAVE_FTRACE_MCOUNT_RECORD and HAVE_FUNCTION_TRACER, so scroll your reader back up if you got over eager.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubjh)}(hhh]jm)}(hXOnce those are out of the way, you will need to implement: - asm/ftrace.h: - MCOUNT_ADDR - ftrace_call_adjust() - struct dyn_arch_ftrace{} - asm code: - mcount() (new stub) - ftrace_caller() - ftrace_call() - ftrace_stub() - C code: - ftrace_dyn_arch_init() - ftrace_make_nop() - ftrace_make_call() - ftrace_update_ftrace_func() h](js)}(h:Once those are out of the way, you will need to implement:h]h:Once those are out of the way, you will need to implement:}(hj#hhhNhNubah}(h]h ]h"]h$]h&]uh1jrhhhM hjubj)}(hhh]j)}(hhh](j)}(h_asm/ftrace.h: - MCOUNT_ADDR - ftrace_call_adjust() - struct dyn_arch_ftrace{}h]jh)}(hhh]jm)}(hMasm/ftrace.h: - MCOUNT_ADDR - ftrace_call_adjust() - struct dyn_arch_ftrace{}h](js)}(h asm/ftrace.h:h]h asm/ftrace.h:}(hjBhhhNhNubah}(h]h ]h"]h$]h&]uh1jrhhhMhj>ubj)}(hhh]j)}(hhh](j)}(h MCOUNT_ADDRh]h)}(hjXh]h MCOUNT_ADDR}(hjZhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjVubah}(h]h ]h"]h$]h&]uh1jhjSubj)}(hftrace_call_adjust()h]h)}(hjoh]hftrace_call_adjust()}(hjqhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjmubah}(h]h ]h"]h$]h&]uh1jhjSubj)}(hstruct dyn_arch_ftrace{}h]h)}(hjh]hstruct dyn_arch_ftrace{}}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjubah}(h]h ]h"]h$]h&]uh1jhjSubeh}(h]h ]h"]h$]h&]jjuh1jhhhMhjPubah}(h]h ]h"]h$]h&]uh1jhj>ubeh}(h]h ]h"]h$]h&]uh1jlhhhMhj;ubah}(h]h ]h"]h$]h&]uh1jghj7ubah}(h]h ]h"]h$]h&]uh1jhj4ubj)}(hiasm code: - mcount() (new stub) - ftrace_caller() - ftrace_call() - ftrace_stub()h]jh)}(hhh]jm)}(hQasm code: - mcount() (new stub) - ftrace_caller() - ftrace_call() - ftrace_stub()h](js)}(h asm code:h]h asm code:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jrhhhMhjubj)}(hhh]j)}(hhh](j)}(hmcount() (new stub)h]h)}(hjh]hmcount() (new stub)}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(hftrace_caller()h]h)}(hjh]hftrace_caller()}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(h ftrace_call()h]h)}(hjh]h ftrace_call()}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(h ftrace_stub()h]h)}(hjh]h ftrace_stub()}(hj!hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]jjuh1jhhhMhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jlhhhMhjubah}(h]h ]h"]h$]h&]uh1jghjubah}(h]h ]h"]h$]h&]uh1jhj4ubj)}(hC code: - ftrace_dyn_arch_init() - ftrace_make_nop() - ftrace_make_call() - ftrace_update_ftrace_func() h]jh)}(hhh]jm)}(hhC code: - ftrace_dyn_arch_init() - ftrace_make_nop() - ftrace_make_call() - ftrace_update_ftrace_func() h](js)}(hC code:h]hC code:}(hj]hhhNhNubah}(h]h ]h"]h$]h&]uh1jrhhhM hjYubj)}(hhh]j)}(hhh](j)}(hftrace_dyn_arch_init()h]h)}(hjsh]hftrace_dyn_arch_init()}(hjuhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM hjqubah}(h]h ]h"]h$]h&]uh1jhjnubj)}(hftrace_make_nop()h]h)}(hjh]hftrace_make_nop()}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM hjubah}(h]h ]h"]h$]h&]uh1jhjnubj)}(hftrace_make_call()h]h)}(hjh]hftrace_make_call()}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM hjubah}(h]h ]h"]h$]h&]uh1jhjnubj)}(hftrace_update_ftrace_func() h]h)}(hftrace_update_ftrace_func()h]hftrace_update_ftrace_func()}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM hjubah}(h]h ]h"]h$]h&]uh1jhjnubeh}(h]h ]h"]h$]h&]jjuh1jhhhM hjkubah}(h]h ]h"]h$]h&]uh1jhjYubeh}(h]h ]h"]h$]h&]uh1jlhhhM hjVubah}(h]h ]h"]h$]h&]uh1jghjRubah}(h]h ]h"]h$]h&]uh1jhj4ubeh}(h]h ]h"]h$]h&]jjuh1jhhhMhj1ubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jlhhhM hjubah}(h]h ]h"]h$]h&]uh1jghjhhhNhNubh)}(hGFirst you will need to fill out some arch details in your asm/ftrace.h.h]hGFirst you will need to fill out some arch details in your asm/ftrace.h.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjhhubh)}(hDDefine MCOUNT_ADDR as the address of your mcount symbol similar to::h]hCDefine MCOUNT_ADDR as the address of your mcount symbol similar to:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjhhubj )}(h+#define MCOUNT_ADDR ((unsigned long)mcount)h]h+#define MCOUNT_ADDR ((unsigned long)mcount)}hj sbah}(h]h ]h"]h$]h&]jjuh1j hhhMhjhhubh)}(hHSince no one else will have a decl for that function, you will need to::h]hGSince no one else will have a decl for that function, you will need to:}(hj.hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjhhubj )}(hextern void mcount(void);h]hextern void mcount(void);}hj<sbah}(h]h ]h"]h$]h&]jjuh1j hhhMhjhhubh)}(hoYou will also need the helper function ftrace_call_adjust(). Most people will be able to stub it out like so::h]hnYou will also need the helper function ftrace_call_adjust(). Most people will be able to stub it out like so:}(hjJhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjhhubj )}(h[static inline unsigned long ftrace_call_adjust(unsigned long addr) { return addr; }h]h[static inline unsigned long ftrace_call_adjust(unsigned long addr) { return addr; }}hjXsbah}(h]h ]h"]h$]h&]jjuh1j hhhMhjhhubh)}(h
h]h
}(hjfhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM!hjhhubh)}(hLastly you will need the custom dyn_arch_ftrace structure. If you need some extra state when runtime patching arbitrary call sites, this is the place. For now though, create an empty struct::h]hLastly you will need the custom dyn_arch_ftrace structure. If you need some extra state when runtime patching arbitrary call sites, this is the place. For now though, create an empty struct:}(hjthhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM#hjhhubj )}(h>struct dyn_arch_ftrace { /* No extra data needed */ };h]h>struct dyn_arch_ftrace { /* No extra data needed */ };}hjsbah}(h]h ]h"]h$]h&]jjuh1j hhhM'hjhhubh)}(hXSWith the header out of the way, we can fill out the assembly code. While we did already create a mcount() function earlier, dynamic ftrace only wants a stub function. This is because the mcount() will only be used during boot and then all references to it will be patched out never to return. Instead, the guts of the old mcount() will be used to create a new ftrace_caller() function. Because the two are hard to merge, it will most likely be a lot easier to have two separate definitions split up by #ifdefs. Same goes for the ftrace_stub() as that will now be inlined in ftrace_caller().h]hXSWith the header out of the way, we can fill out the assembly code. While we did already create a mcount() function earlier, dynamic ftrace only wants a stub function. This is because the mcount() will only be used during boot and then all references to it will be patched out never to return. Instead, the guts of the old mcount() will be used to create a new ftrace_caller() function. Because the two are hard to merge, it will most likely be a lot easier to have two separate definitions split up by #ifdefs. Same goes for the ftrace_stub() as that will now be inlined in ftrace_caller().}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM+hjhhubh)}(hrBefore we get confused anymore, let's check out some pseudo code so you can implement your own stuff in assembly::h]hsBefore we get confused anymore, let’s check out some pseudo code so you can implement your own stuff in assembly:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM4hjhhubj )}(hXuvoid mcount(void) { return; } void ftrace_caller(void) { /* save all state needed by the ABI (see paragraph above) */ unsigned long frompc = ...; unsigned long selfpc = - MCOUNT_INSN_SIZE; ftrace_call: ftrace_stub(frompc, selfpc); /* restore all state needed by the ABI */ ftrace_stub: return; }h]hXuvoid mcount(void) { return; } void ftrace_caller(void) { /* save all state needed by the ABI (see paragraph above) */ unsigned long frompc = ...; unsigned long selfpc = - MCOUNT_INSN_SIZE; ftrace_call: ftrace_stub(frompc, selfpc); /* restore all state needed by the ABI */ ftrace_stub: return; }}hjsbah}(h]h ]h"]h$]h&]jjuh1j hhhM7hjhhubh)}(hXThis might look a little odd at first, but keep in mind that we will be runtime patching multiple things. First, only functions that we actually want to trace will be patched to call ftrace_caller(). Second, since we only have one tracer active at a time, we will patch the ftrace_caller() function itself to call the specific tracer in question. That is the point of the ftrace_call label.h]hXThis might look a little odd at first, but keep in mind that we will be runtime patching multiple things. First, only functions that we actually want to trace will be patched to call ftrace_caller(). Second, since we only have one tracer active at a time, we will patch the ftrace_caller() function itself to call the specific tracer in question. That is the point of the ftrace_call label.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMLhjhhubh)}(hWith that in mind, let's move on to the C code that will actually be doing the runtime patching. You'll need a little knowledge of your arch's opcodes in order to make it through the next section.h]hWith that in mind, let’s move on to the C code that will actually be doing the runtime patching. You’ll need a little knowledge of your arch’s opcodes in order to make it through the next section.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMRhjhhubh)}(hEvery arch has an init callback function. If you need to do something early on to initialize some state, this is the time to do that. Otherwise, this simple function below should be sufficient for most people::h]hEvery arch has an init callback function. If you need to do something early on to initialize some state, this is the time to do that. Otherwise, this simple function below should be sufficient for most people:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMVhjhhubj )}(h;int __init ftrace_dyn_arch_init(void) { return 0; }h]h;int __init ftrace_dyn_arch_init(void) { return 0; }}hjsbah}(h]h ]h"]h$]h&]jjuh1j hhhMZhjhhubh)}(hXThere are two functions that are used to do runtime patching of arbitrary functions. The first is used to turn the mcount call site into a nop (which is what helps us retain runtime performance when not tracing). The second is used to turn the mcount call site into a call to an arbitrary location (but typically that is ftracer_caller()). See the general function definition in linux/ftrace.h for the functions::h]hXThere are two functions that are used to do runtime patching of arbitrary functions. The first is used to turn the mcount call site into a nop (which is what helps us retain runtime performance when not tracing). The second is used to turn the mcount call site into a call to an arbitrary location (but typically that is ftracer_caller()). See the general function definition in linux/ftrace.h for the functions:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM_hjhhubj )}(h$ftrace_make_nop() ftrace_make_call()h]h$ftrace_make_nop() ftrace_make_call()}hjsbah}(h]h ]h"]h$]h&]jjuh1j hhhMfhjhhubh)}(h}The rec->ip value is the address of the mcount call site that was collected by the scripts/recordmcount.pl during build time.h]h}The rec->ip value is the address of the mcount call site that was collected by the scripts/recordmcount.pl during build time.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMihjhhubh)}(hXThe last function is used to do runtime patching of the active tracer. This will be modifying the assembly code at the location of the ftrace_call symbol inside of the ftrace_caller() function. So you should have sufficient padding at that location to support the new function calls you'll be inserting. Some people will be using a "call" type instruction while others will be using a "branch" type instruction. Specifically, the function is::h]hXThe last function is used to do runtime patching of the active tracer. This will be modifying the assembly code at the location of the ftrace_call symbol inside of the ftrace_caller() function. So you should have sufficient padding at that location to support the new function calls you’ll be inserting. Some people will be using a “call” type instruction while others will be using a “branch” type instruction. Specifically, the function is:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMlhjhhubj )}(hftrace_update_ftrace_func()h]hftrace_update_ftrace_func()}hj*sbah}(h]h ]h"]h$]h&]jjuh1j hhhMshjhhubeh}(h]have-dynamic-ftraceah ]h"]have_dynamic_ftraceah$]h&]uh1hhhhhhhhKubh)}(hhh](h)}(h0HAVE_DYNAMIC_FTRACE + HAVE_FUNCTION_GRAPH_TRACERh]h0HAVE_DYNAMIC_FTRACE + HAVE_FUNCTION_GRAPH_TRACER}(hjChhhNhNubah}(h]h ]h"]h$]h&]uh1hhj@hhhhhMwubh)}(hjThe function grapher needs a few tweaks in order to work with dynamic ftrace. Basically, you will need to:h]hjThe function grapher needs a few tweaks in order to work with dynamic ftrace. Basically, you will need to:}(hjQhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMyhj@hhubjk)}(h- update: - ftrace_caller() - ftrace_graph_call() - ftrace_graph_caller() - implement: - ftrace_enable_ftrace_graph_caller() - ftrace_disable_ftrace_graph_caller() h]j)}(hhh](j)}(hYupdate: - ftrace_caller() - ftrace_graph_call() - ftrace_graph_caller()h]jh)}(hhh]jm)}(hGupdate: - ftrace_caller() - ftrace_graph_call() - ftrace_graph_caller()h](js)}(hupdate:h]hupdate:}(hjqhhhNhNubah}(h]h ]h"]h$]h&]uh1jrhhhM~hjmubj)}(hhh]j)}(hhh](j)}(hftrace_caller()h]h)}(hjh]hftrace_caller()}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM}hjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(hftrace_graph_call()h]h)}(hjh]hftrace_graph_call()}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM~hjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(hftrace_graph_caller()h]h)}(hjh]hftrace_graph_caller()}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]jjuh1jhhhM}hjubah}(h]h ]h"]h$]h&]uh1jhjmubeh}(h]h ]h"]h$]h&]uh1jlhhhM~hjjubah}(h]h ]h"]h$]h&]uh1jghjfubah}(h]h ]h"]h$]h&]uh1jhjcubj)}(hdimplement: - ftrace_enable_ftrace_graph_caller() - ftrace_disable_ftrace_graph_caller() h]jh)}(hhh]jm)}(hXimplement: - ftrace_enable_ftrace_graph_caller() - ftrace_disable_ftrace_graph_caller() h](js)}(h implement:h]h implement:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jrhhhMhjubj)}(hhh]j)}(hhh](j)}(h#ftrace_enable_ftrace_graph_caller()h]h)}(hj h]h#ftrace_enable_ftrace_graph_caller()}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhj ubj)}(h%ftrace_disable_ftrace_graph_caller() h]h)}(h$ftrace_disable_ftrace_graph_caller()h]h$ftrace_disable_ftrace_graph_caller()}(hj" hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhj ubeh}(h]h ]h"]h$]h&]jjuh1jhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jlhhhMhjubah}(h]h ]h"]h$]h&]uh1jghjubah}(h]h ]h"]h$]h&]uh1jhjcubeh}(h]h ]h"]h$]h&]jjuh1jhhhM|hj_ubah}(h]h ]h"]h$]h&]uh1jjhhhM|hj@hhubh)}(h
h]h
}(hj` hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj@hhubh)}(h Quick notes:h]h Quick notes:}(hjn hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj@hhubjk)}(hX- add a nop stub after the ftrace_call location named ftrace_graph_call; stub needs to be large enough to support a call to ftrace_graph_caller() - update ftrace_graph_caller() to work with being called by the new ftrace_caller() since some semantics may have changed - ftrace_enable_ftrace_graph_caller() will runtime patch the ftrace_graph_call location with a call to ftrace_graph_caller() - ftrace_disable_ftrace_graph_caller() will runtime patch the ftrace_graph_call location with nopsh]j)}(hhh](j)}(hadd a nop stub after the ftrace_call location named ftrace_graph_call; stub needs to be large enough to support a call to ftrace_graph_caller()h]h)}(hadd a nop stub after the ftrace_call location named ftrace_graph_call; stub needs to be large enough to support a call to ftrace_graph_caller()h]hadd a nop stub after the ftrace_call location named ftrace_graph_call; stub needs to be large enough to support a call to ftrace_graph_caller()}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhj ubj)}(hwupdate ftrace_graph_caller() to work with being called by the new ftrace_caller() since some semantics may have changedh]h)}(hwupdate ftrace_graph_caller() to work with being called by the new ftrace_caller() since some semantics may have changedh]hwupdate ftrace_graph_caller() to work with being called by the new ftrace_caller() since some semantics may have changed}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhj ubj)}(hzftrace_enable_ftrace_graph_caller() will runtime patch the ftrace_graph_call location with a call to ftrace_graph_caller()h]h)}(hzftrace_enable_ftrace_graph_caller() will runtime patch the ftrace_graph_call location with a call to ftrace_graph_caller()h]hzftrace_enable_ftrace_graph_caller() will runtime patch the ftrace_graph_call location with a call to ftrace_graph_caller()}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhj ubj)}(h`ftrace_disable_ftrace_graph_caller() will runtime patch the ftrace_graph_call location with nopsh]h)}(h`ftrace_disable_ftrace_graph_caller() will runtime patch the ftrace_graph_call location with nopsh]h`ftrace_disable_ftrace_graph_caller() will runtime patch the ftrace_graph_call location with nops}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhj ubeh}(h]h ]h"]h$]h&]jjuh1jhhhMhj| ubah}(h]h ]h"]h$]h&]uh1jjhhhMhj@hhubeh}(h].have-dynamic-ftrace-have-function-graph-tracerah ]h"]0have_dynamic_ftrace + have_function_graph_tracerah$]h&]uh1hhhhhhhhMwubeh}(h]function-tracer-designah ]h"]function tracer designah$]h&]uh1hhhhhhhhKubeh}(h]h ]h"]h$]h&]sourcehuh1hcurrent_sourceN current_lineNsettingsdocutils.frontendValues)}(hN generatorN datestampN source_linkN source_urlN toc_backlinksentryfootnote_backlinksK sectnum_xformKstrip_commentsNstrip_elements_with_classesN strip_classesN report_levelK halt_levelKexit_status_levelKdebugNwarning_streamN tracebackinput_encoding utf-8-siginput_encoding_error_handlerstrictoutput_encodingutf-8output_encoding_error_handlerj" error_encodingutf-8error_encoding_error_handlerbackslashreplace language_codeenrecord_dependenciesNconfigN id_prefixhauto_id_prefixid dump_settingsNdump_internalsNdump_transformsNdump_pseudo_xmlNexpose_internalsNstrict_visitorN_disable_configN_sourceh _destinationN _config_files]7/var/lib/git/docbuild/linux/Documentation/docutils.confafile_insertion_enabled raw_enabledKline_length_limitM'pep_referencesN pep_base_urlhttps://peps.python.org/pep_file_url_templatepep-%04drfc_referencesN rfc_base_url&https://datatracker.ietf.org/doc/html/ tab_widthKtrim_footnote_reference_spacesyntax_highlightlong smart_quotessmartquotes_locales]character_level_inline_markupdoctitle_xform docinfo_xformKsectsubtitle_xform image_loadinglinkembed_stylesheetcloak_email_addressessection_self_linkenvNubreporterNindirect_targets]substitution_defs}substitution_names}refnames}refids}nameids}(j j jSjPjjj jjjjjjjjjj=j:j j u nametypes}(j jSjj jjjjj=j uh}(j hjPjjjVjjjjjjjjjjj:jj j@u footnote_refs} citation_refs} autofootnotes]autofootnote_refs]symbol_footnotes]symbol_footnote_refs] footnotes] citations]autofootnote_startKsymbol_footnote_startK id_counter collectionsCounter}Rparse_messages]hsystem_message)}(hhh]h)}(hfPossible title underline, too short for the title. Treating it as ordinary text because it's so short.h]hhPossible title underline, too short for the title. Treating it as ordinary text because it’s so short.}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubah}(h]h ]h"]h$]h&]levelKtypeINFOlineKnsourcehuh1j hjhhhhhKnubatransform_messages] transformerN include_log] decorationNhhub.