sphinx.addnodesdocument)}( rawsourcechildren]( translations LanguagesNode)}(hhh](h pending_xref)}(hhh]docutils.nodesTextChinese (Simplified)}(hhparenthuba attributes}(ids]classes]names]dupnames]backrefs] refdomainstdreftypedoc reftarget-/translations/zh_CN/admin-guide/perf-securitymodnameN classnameN refexplicitutagnamehhh ubh)}(hhh]hChinese (Traditional)}(hhhh2ubah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget-/translations/zh_TW/admin-guide/perf-securitymodnameN classnameN refexplicituh1hhh ubh)}(hhh]hItalian}(hhhhFubah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget-/translations/it_IT/admin-guide/perf-securitymodnameN classnameN refexplicituh1hhh ubh)}(hhh]hJapanese}(hhhhZubah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget-/translations/ja_JP/admin-guide/perf-securitymodnameN classnameN refexplicituh1hhh ubh)}(hhh]hKorean}(hhhhnubah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget-/translations/ko_KR/admin-guide/perf-securitymodnameN classnameN refexplicituh1hhh ubh)}(hhh]hSpanish}(hhhhubah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget-/translations/sp_SP/admin-guide/perf-securitymodnameN classnameN refexplicituh1hhh ubeh}(h]h ]h"]h$]h&]current_languageEnglishuh1h hh _documenthsourceNlineNubhtarget)}(h.. _perf_security:h]h}(h]h ]h"]h$]h&]refid perf-securityuh1hhKhhhhhG/var/lib/git/docbuild/linux/Documentation/admin-guide/perf-security.rstubhsection)}(hhh](htitle)}(hPerf events and tool securityh]hPerf events and tool security}(hhhhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhhhhhKubh)}(hhh](h)}(hOverviewh]hOverview}(hhhhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhhhhhKubh paragraph)}(hX9Usage of Performance Counters for Linux (perf_events) [1]_ , [2]_ , [3]_ can impose a considerable risk of leaking sensitive data accessed by monitored processes. The data leakage is possible both in scenarios of direct usage of perf_events system call API [2]_ and over data files generated by Perf tool user mode utility (Perf) [3]_ , [4]_ . The risk depends on the nature of data that perf_events performance monitoring units (PMU) [2]_ and Perf collect and expose for performance analysis. Collected system and performance data may be split into several categories:h](h6Usage of Performance Counters for Linux (perf_events) }(h6Usage of Performance Counters for Linux (perf_events) hhhhhNhNubhfootnote_reference)}(h[1]_h]h1}(hhhhhhhNhNubah}(h]id1ah ]h"]h$]h&]hid32docnameadmin-guide/perf-securityuh1hhh،resolvedKubh , }(h , hhhhhNhNubh)}(h[2]_h]h2}(hhhhhhhNhNubah}(h]id2ah ]h"]h$]h&]hid33hhuh1hhhhKubh , }(h , hhubh)}(h[3]_h]h3}(hhhjhhhNhNubah}(h]id3ah ]h"]h$]h&]hid34hhuh1hhhhKubh can impose a considerable risk of leaking sensitive data accessed by monitored processes. The data leakage is possible both in scenarios of direct usage of perf_events system call API }(h can impose a considerable risk of leaking sensitive data accessed by monitored processes. The data leakage is possible both in scenarios of direct usage of perf_events system call API hhhhhNhNubh)}(h[2]_h]h2}(hhhj%hhhNhNubah}(h]id4ah ]h"]h$]h&]hj hhuh1hhhhKubhE and over data files generated by Perf tool user mode utility (Perf) }(hE and over data files generated by Perf tool user mode utility (Perf) hhhhhNhNubh)}(h[3]_h]h3}(hhhj9hhhNhNubah}(h]id5ah ]h"]h$]h&]hjhhuh1hhhhKubh , }(hhhhubh)}(h[4]_h]h4}(hhhjLhhhNhNubah}(h]id6ah ]h"]h$]h&]hid35hhuh1hhhhKubh^ . The risk depends on the nature of data that perf_events performance monitoring units (PMU) }(h^ . The risk depends on the nature of data that perf_events performance monitoring units (PMU) hhhhhNhNubh)}(h[2]_h]h2}(hhhjahhhNhNubah}(h]id7ah ]h"]h$]h&]hj hhuh1hhhhKubh and Perf collect and expose for performance analysis. Collected system and performance data may be split into several categories:}(h and Perf collect and expose for performance analysis. Collected system and performance data may be split into several categories:hhhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhK hhhhubhenumerated_list)}(hhh](h list_item)}(hX'System hardware and software configuration data, for example: a CPU model and its cache configuration, an amount of available memory and its topology, used kernel and Perf versions, performance monitoring setup including experiment time, events configuration, Perf command line parameters, etc. h]h)}(hX&System hardware and software configuration data, for example: a CPU model and its cache configuration, an amount of available memory and its topology, used kernel and Perf versions, performance monitoring setup including experiment time, events configuration, Perf command line parameters, etc.h]hX&System hardware and software configuration data, for example: a CPU model and its cache configuration, an amount of available memory and its topology, used kernel and Perf versions, performance monitoring setup including experiment time, events configuration, Perf command line parameters, etc.}(hjhjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhj}hhhhhNubj)}(hUser and kernel module paths and their load addresses with sizes, process and thread names with their PIDs and TIDs, timestamps for captured hardware and software events. h]h)}(hUser and kernel module paths and their load addresses with sizes, process and thread names with their PIDs and TIDs, timestamps for captured hardware and software events.h]hUser and kernel module paths and their load addresses with sizes, process and thread names with their PIDs and TIDs, timestamps for captured hardware and software events.}(hjhjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhj}hhhhhNubj)}(hXContent of kernel software counters (e.g., for context switches, page faults, CPU migrations), architectural hardware performance counters (PMC) [8]_ and machine specific registers (MSR) [9]_ that provide execution metrics for various monitored parts of the system (e.g., memory controller (IMC), interconnect (QPI/UPI) or peripheral (PCIe) uncore counters) without direct attribution to any execution context state. h]h)}(hXContent of kernel software counters (e.g., for context switches, page faults, CPU migrations), architectural hardware performance counters (PMC) [8]_ and machine specific registers (MSR) [9]_ that provide execution metrics for various monitored parts of the system (e.g., memory controller (IMC), interconnect (QPI/UPI) or peripheral (PCIe) uncore counters) without direct attribution to any execution context state.h](hContent of kernel software counters (e.g., for context switches, page faults, CPU migrations), architectural hardware performance counters (PMC) }(hContent of kernel software counters (e.g., for context switches, page faults, CPU migrations), architectural hardware performance counters (PMC) hjhhhNhNubh)}(h[8]_h]h8}(hhhjhhhNhNubah}(h]id8ah ]h"]h$]h&]hid39hhuh1hhjhKubh& and machine specific registers (MSR) }(h& and machine specific registers (MSR) hjhhhNhNubh)}(h[9]_h]h9}(hhhjhhhNhNubah}(h]id9ah ]h"]h$]h&]hid40hhuh1hhjhKubh that provide execution metrics for various monitored parts of the system (e.g., memory controller (IMC), interconnect (QPI/UPI) or peripheral (PCIe) uncore counters) without direct attribution to any execution context state.}(h that provide execution metrics for various monitored parts of the system (e.g., memory controller (IMC), interconnect (QPI/UPI) or peripheral (PCIe) uncore counters) without direct attribution to any execution context state.hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhj}hhhhhNubj)}(hContent of architectural execution context registers (e.g., RIP, RSP, RBP on x86_64), process user and kernel space memory addresses and data, content of various architectural MSRs that capture data from this category. h]h)}(hContent of architectural execution context registers (e.g., RIP, RSP, RBP on x86_64), process user and kernel space memory addresses and data, content of various architectural MSRs that capture data from this category.h]hContent of architectural execution context registers (e.g., RIP, RSP, RBP on x86_64), process user and kernel space memory addresses and data, content of various architectural MSRs that capture data from this category.}(hjhjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK%hjubah}(h]h ]h"]h$]h&]uh1jhj}hhhhhNubeh}(h]h ]h"]h$]h&]enumtypearabicprefixhsuffix.uh1j{hhhhhhhKubh)}(hXData that belong to the fourth category can potentially contain sensitive process data. If PMUs in some monitoring modes capture values of execution context registers or data from process memory then access to such monitoring modes requires to be ordered and secured properly. So, perf_events performance monitoring and observability operations are the subject for security access control management [5]_ .h](hXData that belong to the fourth category can potentially contain sensitive process data. If PMUs in some monitoring modes capture values of execution context registers or data from process memory then access to such monitoring modes requires to be ordered and secured properly. So, perf_events performance monitoring and observability operations are the subject for security access control management }(hXData that belong to the fourth category can potentially contain sensitive process data. If PMUs in some monitoring modes capture values of execution context registers or data from process memory then access to such monitoring modes requires to be ordered and secured properly. So, perf_events performance monitoring and observability operations are the subject for security access control management hjhhhNhNubh)}(h[5]_h]h5}(hhhj!hhhNhNubah}(h]id10ah ]h"]h$]h&]hid36hhuh1hhjhKubh .}(h .hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhK*hhhhubeh}(h]overviewah ]h"]overviewah$]h&]uh1hhhhhhhhKubh)}(hhh](h)}(hperf_events access controlh]hperf_events access control}(hjIhjGhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjDhhhhhK2ubh)}(hXTo perform security checks, the Linux implementation splits processes into two categories [6]_ : a) privileged processes (whose effective user ID is 0, referred to as superuser or root), and b) unprivileged processes (whose effective UID is nonzero). Privileged processes bypass all kernel security permission checks so perf_events performance monitoring is fully available to privileged processes without access, scope and resource restrictions.h](hZTo perform security checks, the Linux implementation splits processes into two categories }(hZTo perform security checks, the Linux implementation splits processes into two categories hjUhhhNhNubh)}(h[6]_h]h6}(hhhj^hhhNhNubah}(h]id11ah ]h"]h$]h&]hid37hhuh1hhjUhKubhX` : a) privileged processes (whose effective user ID is 0, referred to as superuser or root), and b) unprivileged processes (whose effective UID is nonzero). Privileged processes bypass all kernel security permission checks so perf_events performance monitoring is fully available to privileged processes without access, scope and resource restrictions.}(hX` : a) privileged processes (whose effective user ID is 0, referred to as superuser or root), and b) unprivileged processes (whose effective UID is nonzero). Privileged processes bypass all kernel security permission checks so perf_events performance monitoring is fully available to privileged processes without access, scope and resource restrictions.hjUhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhK4hjDhhubh)}(hUnprivileged processes are subject to a full security permission check based on the process's credentials [5]_ (usually: effective UID, effective GID, and supplementary group list).h](hlUnprivileged processes are subject to a full security permission check based on the process’s credentials }(hjUnprivileged processes are subject to a full security permission check based on the process's credentials hjyhhhNhNubh)}(h[5]_h]h5}(hhhjhhhNhNubah}(h]id12ah ]h"]h$]h&]hj0hhuh1hhjyhKubhG (usually: effective UID, effective GID, and supplementary group list).}(hG (usually: effective UID, effective GID, and supplementary group list).hjyhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhK=0: *scope* includes per-process and system wide performance monitoring but excludes raw tracepoints and ftrace function tracepoints monitoring. CPU and system events happened when executing either in user or in kernel space can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK [6]_ capability. h](j)}(h>=0:h]h>=0:}(hjhjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhKhjubj)}(hhh]h)}(hX*scope* includes per-process and system wide performance monitoring but excludes raw tracepoints and ftrace function tracepoints monitoring. CPU and system events happened when executing either in user or in kernel space can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK [6]_ capability.h](j)}(h*scope*h]hscope}(hhhjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhjubhX includes per-process and system wide performance monitoring but excludes raw tracepoints and ftrace function tracepoints monitoring. CPU and system events happened when executing either in user or in kernel space can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK }(hX includes per-process and system wide performance monitoring but excludes raw tracepoints and ftrace function tracepoints monitoring. CPU and system events happened when executing either in user or in kernel space can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK hjhhhNhNubh)}(h[6]_h]h6}(hhhjhhhNhNubah}(h]id26ah ]h"]h$]h&]hjmhhuh1hhjhKubh capability.}(h capability.hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhKhjhhubj)}(hXt>=1: *scope* includes per-process performance monitoring only and excludes system wide performance monitoring. CPU and system events happened when executing either in user or in kernel space can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK capability. h](j)}(h>=1:h]h>=1:}(hjhjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhKhjubj)}(hhh]h)}(hXn*scope* includes per-process performance monitoring only and excludes system wide performance monitoring. CPU and system events happened when executing either in user or in kernel space can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK capability.h](j)}(h*scope*h]hscope}(hhhjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhjubhXg includes per-process performance monitoring only and excludes system wide performance monitoring. CPU and system events happened when executing either in user or in kernel space can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK capability.}(hXg includes per-process performance monitoring only and excludes system wide performance monitoring. CPU and system events happened when executing either in user or in kernel space can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK capability.hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhKhjhhubj)}(hX5>=2: *scope* includes per-process performance monitoring only. CPU and system events happened when executing in user space only can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK capability. h](j)}(h>=2:h]h>=2:}(hjhjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhMhjubj)}(hhh]h)}(hX/*scope* includes per-process performance monitoring only. CPU and system events happened when executing in user space only can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK capability.h](j)}(h*scope*h]hscope}(hhhj+hhhNhNubah}(h]h ]h"]h$]h&]uh1jhj'ubhX( includes per-process performance monitoring only. CPU and system events happened when executing in user space only can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK capability.}(hX( includes per-process performance monitoring only. CPU and system events happened when executing in user space only can be monitored and captured for later analysis. Per-user per-cpu perf_event_mlock_kb locking limit is imposed but ignored for unprivileged processes with CAP_IPC_LOCK capability.hj'hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhj$ubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhMhjhhubeh}(h]h ]h"]h$]h&]uh1jhjhhhhhNubeh}(h]unprivileged-usersah ]h"]unprivileged usersah$]h&]uh1hhhhhhhhKubh)}(hhh](h)}(hResource controlh]hResource control}(hjchjahhhNhNubah}(h]h ]h"]h$]h&]uh1hhj^hhhhhMubh)}(hhh](h)}(hOpen file descriptorsh]hOpen file descriptors}(hjthjrhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjohhhhhM ubh)}(hXThe perf_events system call API [2]_ allocates file descriptors for every configured PMU event. Open file descriptors are a per-process accountable resource governed by the RLIMIT_NOFILE [11]_ limit (ulimit -n), which is usually derived from the login shell process. When configuring Perf collection for a long list of events on a large server system, this limit can be easily hit preventing required monitoring configuration. RLIMIT_NOFILE limit can be increased on per-user basis modifying content of the limits.conf file [12]_ . Ordinarily, a Perf sampling session (perf record) requires an amount of open perf_event file descriptors that is not less than the number of monitored events multiplied by the number of monitored CPUs.h](h The perf_events system call API }(h The perf_events system call API hjhhhNhNubh)}(h[2]_h]h2}(hhhjhhhNhNubah}(h]id27ah ]h"]h$]h&]hj hhuh1hhjhKubh allocates file descriptors for every configured PMU event. Open file descriptors are a per-process accountable resource governed by the RLIMIT_NOFILE }(h allocates file descriptors for every configured PMU event. Open file descriptors are a per-process accountable resource governed by the RLIMIT_NOFILE hjhhhNhNubh)}(h[11]_h]h11}(hhhjhhhNhNubah}(h]id28ah ]h"]h$]h&]hid42hhuh1hhjhKubhXL limit (ulimit -n), which is usually derived from the login shell process. When configuring Perf collection for a long list of events on a large server system, this limit can be easily hit preventing required monitoring configuration. RLIMIT_NOFILE limit can be increased on per-user basis modifying content of the limits.conf file }(hXL limit (ulimit -n), which is usually derived from the login shell process. When configuring Perf collection for a long list of events on a large server system, this limit can be easily hit preventing required monitoring configuration. RLIMIT_NOFILE limit can be increased on per-user basis modifying content of the limits.conf file hjhhhNhNubh)}(h[12]_h]h12}(hhhjhhhNhNubah}(h]id29ah ]h"]h$]h&]hid43hhuh1hhjhKubh . Ordinarily, a Perf sampling session (perf record) requires an amount of open perf_event file descriptors that is not less than the number of monitored events multiplied by the number of monitored CPUs.}(h . Ordinarily, a Perf sampling session (perf record) requires an amount of open perf_event file descriptors that is not less than the number of monitored events multiplied by the number of monitored CPUs.hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhM hjohhubeh}(h]open-file-descriptorsah ]h"]open file descriptorsah$]h&]uh1hhj^hhhhhM ubh)}(hhh](h)}(hMemory allocationh]hMemory allocation}(hjhjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhMubh)}(hXThe amount of memory available to user processes for capturing performance monitoring data is governed by the perf_event_mlock_kb [2]_ setting. This perf_event specific resource setting defines overall per-cpu limits of memory allowed for mapping by the user processes to execute performance monitoring. The setting essentially extends the RLIMIT_MEMLOCK [11]_ limit, but only for memory regions mapped specifically for capturing monitored performance events and related data.h](hThe amount of memory available to user processes for capturing performance monitoring data is governed by the perf_event_mlock_kb }(hThe amount of memory available to user processes for capturing performance monitoring data is governed by the perf_event_mlock_kb hjhhhNhNubh)}(h[2]_h]h2}(hhhjhhhNhNubah}(h]id30ah ]h"]h$]h&]hj hhuh1hhjhKubh setting. This perf_event specific resource setting defines overall per-cpu limits of memory allowed for mapping by the user processes to execute performance monitoring. The setting essentially extends the RLIMIT_MEMLOCK }(h setting. This perf_event specific resource setting defines overall per-cpu limits of memory allowed for mapping by the user processes to execute performance monitoring. The setting essentially extends the RLIMIT_MEMLOCK hjhhhNhNubh)}(h[11]_h]h11}(hhhjhhhNhNubah}(h]id31ah ]h"]h$]h&]hjhhuh1hhjhKubht limit, but only for memory regions mapped specifically for capturing monitored performance events and related data.}(ht limit, but only for memory regions mapped specifically for capturing monitored performance events and related data.hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMhjhhubh)}(hXFor example, if a machine has eight cores and perf_event_mlock_kb limit is set to 516 KiB, then a user process is provided with 516 KiB * 8 = 4128 KiB of memory above the RLIMIT_MEMLOCK limit (ulimit -l) for perf_event mmap buffers. In particular, this means that, if the user wants to start two or more performance monitoring processes, the user is required to manually distribute the available 4128 KiB between the monitoring processes, for example, using the --mmap-pages Perf record mode option. Otherwise, the first started performance monitoring process allocates all available 4128 KiB and the other processes will fail to proceed due to the lack of memory.h]hXFor example, if a machine has eight cores and perf_event_mlock_kb limit is set to 516 KiB, then a user process is provided with 516 KiB * 8 = 4128 KiB of memory above the RLIMIT_MEMLOCK limit (ulimit -l) for perf_event mmap buffers. In particular, this means that, if the user wants to start two or more performance monitoring processes, the user is required to manually distribute the available 4128 KiB between the monitoring processes, for example, using the --mmap-pages Perf record mode option. Otherwise, the first started performance monitoring process allocates all available 4128 KiB and the other processes will fail to proceed due to the lack of memory.}(hjhjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM"hjhhubh)}(hXDRLIMIT_MEMLOCK and perf_event_mlock_kb resource constraints are ignored for processes with the CAP_IPC_LOCK capability. Thus, perf_events/Perf privileged users can be provided with memory above the constraints for perf_events/Perf performance monitoring purpose by providing the Perf executable with CAP_IPC_LOCK capability.h]hXDRLIMIT_MEMLOCK and perf_event_mlock_kb resource constraints are ignored for processes with the CAP_IPC_LOCK capability. Thus, perf_events/Perf privileged users can be provided with memory above the constraints for perf_events/Perf performance monitoring purpose by providing the Perf executable with CAP_IPC_LOCK capability.}(hj-hj+hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM-hjhhubeh}(h]memory-allocationah ]h"]memory allocationah$]h&]uh1hhj^hhhhhMubeh}(h]resource-controlah ]h"]resource controlah$]h&]uh1hhhhhhhhMubh)}(hhh](h)}(h Bibliographyh]h Bibliography}(hjNhjLhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjIhhhhhM4ubhfootnote)}(h%``_h](hlabel)}(h1h]h1}(hhhjbhhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj\ubh)}(hj^h](h reference)}(hj^h]h https://lwn.net/Articles/337493/}(h https://lwn.net/Articles/337493/hjuhhhNhNubah}(h]h ]h"]h$]h&]name https://lwn.net/Articles/337493/refurijuh1jshjpubh)}(h"h]h}(h]https-lwn-net-articles-337493ah ]h"] https://lwn.net/articles/337493/ah$]h&]refurijuh1h referencedKhjpubeh}(h]h ]h"]h$]h&]uh1hhhhM6hj\ubeh}(h]hah ]h"]1ah$]h&]hahhuh1jZhhhM6hjIhhhKubj[)}(h@``_h](ja)}(h2h]h2}(hhhjhhhNhNubah}(h]h ]h"]h$]h&]uh1j`hjubh)}(hjh](jt)}(hjh]h;http://man7.org/linux/man-pages/man2/perf_event_open.2.html}(h;http://man7.org/linux/man-pages/man2/perf_event_open.2.htmlhjhhhNhNubah}(h]h ]h"]h$]h&]name;http://man7.org/linux/man-pages/man2/perf_event_open.2.htmljjuh1jshjubh)}(h=h]h}(h]9http-man7-org-linux-man-pages-man2-perf-event-open-2-htmlah ]h"];http://man7.org/linux/man-pages/man2/perf_event_open.2.htmlah$]h&]refurijuh1hjKhjubeh}(h]h ]h"]h$]h&]uh1hhhhM7hjubeh}(h]j ah ]h"]2ah$]h&](jj/jkjj@jjehhuh1jZhhhM7hjIhhhKubj[)}(h=``_h](ja)}(h3h]h3}(hhhjhhhNhNubah}(h]h ]h"]h$]h&]uh1j`hjubh)}(hjh](jt)}(hjh]h8http://web.eece.maine.edu/~vweaver/projects/perf_events/}(h8http://web.eece.maine.edu/~vweaver/projects/perf_events/hjhhhNhNubah}(h]h ]h"]h$]h&]name8http://web.eece.maine.edu/~vweaver/projects/perf_events/jj uh1jshjubh)}(h:h]h}(h]4http-web-eece-maine-edu-vweaver-projects-perf-eventsah ]h"]8http://web.eece.maine.edu/~vweaver/projects/perf_events/ah$]h&]refurij uh1hjKhjubeh}(h]h ]h"]h$]h&]uh1hhhhM8hjubeh}(h]jah ]h"]3ah$]h&](jjCehhuh1jZhhhM8hjIhhhKubj[)}(h5``_h](ja)}(h4h]h4}(hhhj# hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj! h](jt)}(hj! h]h0https://perf.wiki.kernel.org/index.php/Main_Page}(h0https://perf.wiki.kernel.org/index.php/Main_Pagehj4 hhhNhNubah}(h]h ]h"]h$]h&]name0https://perf.wiki.kernel.org/index.php/Main_PagejjC uh1jshj1 ubh)}(h2h]h}(h].https-perf-wiki-kernel-org-index-php-main-pageah ]h"]0https://perf.wiki.kernel.org/index.php/main_pageah$]h&]refurijC uh1hjKhj1 ubeh}(h]h ]h"]h$]h&]uh1hhhhM9hj ubeh}(h]j[ah ]h"]4ah$]h&]jVahhuh1jZhhhM9hjIhhhKubj[)}(hE``_h](ja)}(h5h]h5}(hhhjb hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj^ ubh)}(hj` h](jt)}(hj` h]h@https://www.kernel.org/doc/html/latest/security/credentials.html}(h@https://www.kernel.org/doc/html/latest/security/credentials.htmlhjs hhhNhNubah}(h]h ]h"]h$]h&]name@https://www.kernel.org/doc/html/latest/security/credentials.htmljj uh1jshjp ubh)}(hBh]h}(h]>https-www-kernel-org-doc-html-latest-security-credentials-htmlah ]h"]@https://www.kernel.org/doc/html/latest/security/credentials.htmlah$]h&]refurij uh1hjKhjp ubeh}(h]h ]h"]h$]h&]uh1hhhhM:hj^ ubeh}(h]j0ah ]h"]5ah$]h&](j+jehhuh1jZhhhM:hjIhhhKubj[)}(h=``_h](ja)}(h6h]h6}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h8http://man7.org/linux/man-pages/man7/capabilities.7.html}(h8http://man7.org/linux/man-pages/man7/capabilities.7.htmlhj hhhNhNubah}(h]h ]h"]h$]h&]name8http://man7.org/linux/man-pages/man7/capabilities.7.htmljj uh1jshj ubh)}(h:h]h}(h]6http-man7-org-linux-man-pages-man7-capabilities-7-htmlah ]h"]8http://man7.org/linux/man-pages/man7/capabilities.7.htmlah$]h&]refurij uh1hjKhj ubeh}(h]h ]h"]h$]h&]uh1hhhhM;hj ubeh}(h]jmah ]h"]6ah$]h&](jhjjzjjujehhuh1jZhhhM;hjIhhhKubj[)}(h7``_h](ja)}(h7h]h7}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h2http://man7.org/linux/man-pages/man2/ptrace.2.html}(h2http://man7.org/linux/man-pages/man2/ptrace.2.htmlhj hhhNhNubah}(h]h ]h"]h$]h&]name2http://man7.org/linux/man-pages/man2/ptrace.2.htmljj uh1jshj ubh)}(h4h]h}(h]0http-man7-org-linux-man-pages-man2-ptrace-2-htmlah ]h"]2http://man7.org/linux/man-pages/man2/ptrace.2.htmlah$]h&]refurij uh1hjKhj ubeh}(h]h ]h"]h$]h&]uh1hhhhM<hj ubeh}(h]j4ah ]h"]7ah$]h&]j/ahhuh1jZhhhM<hjIhhhKubj[)}(h?``_h](ja)}(h8h]h8}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h:https://en.wikipedia.org/wiki/Hardware_performance_counter}(h:https://en.wikipedia.org/wiki/Hardware_performance_counterhj0 hhhNhNubah}(h]h ]h"]h$]h&]name:https://en.wikipedia.org/wiki/Hardware_performance_counterjj? uh1jshj- ubh)}(h<h]h}(h]8https-en-wikipedia-org-wiki-hardware-performance-counterah ]h"]:https://en.wikipedia.org/wiki/hardware_performance_counterah$]h&]refurij? uh1hjKhj- ubeh}(h]h ]h"]h$]h&]uh1hhhhM=hj ubeh}(h]jah ]h"]8ah$]h&]jahhuh1jZhhhM=hjIhhhKubj[)}(h:``_h](ja)}(h9h]h9}(hhhj^ hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hjZ ubh)}(hj\ h](jt)}(hj\ h]h5https://en.wikipedia.org/wiki/Model-specific_register}(h5https://en.wikipedia.org/wiki/Model-specific_registerhjo hhhNhNubah}(h]h ]h"]h$]h&]name5https://en.wikipedia.org/wiki/Model-specific_registerjj~ uh1jshjl ubh)}(h7h]h}(h]3https-en-wikipedia-org-wiki-model-specific-registerah ]h"]5https://en.wikipedia.org/wiki/model-specific_registerah$]h&]refurij~ uh1hjKhjl ubeh}(h]h ]h"]h$]h&]uh1hhhhM>hjZ ubeh}(h]jah ]h"]9ah$]h&]jahhuh1jZhhhM>hjIhhhKubj[)}(h4``_h](ja)}(h10h]h10}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h/http://man7.org/linux/man-pages/man5/acl.5.html}(h/http://man7.org/linux/man-pages/man5/acl.5.htmlhj hhhNhNubah}(h]h ]h"]h$]h&]name/http://man7.org/linux/man-pages/man5/acl.5.htmljj uh1jshj ubh)}(h1h]h}(h]-http-man7-org-linux-man-pages-man5-acl-5-htmlah ]h"]/http://man7.org/linux/man-pages/man5/acl.5.htmlah$]h&]refurij uh1hjKhj ubeh}(h]h ]h"]h$]h&]uh1hhhhM?hj ubeh}(h]jah ]h"]10ah$]h&]jahhuh1jZhhhM?hjIhhhKubj[)}(h:``_h](ja)}(h11h]h11}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h5http://man7.org/linux/man-pages/man2/getrlimit.2.html}(h5http://man7.org/linux/man-pages/man2/getrlimit.2.htmlhj hhhNhNubah}(h]h ]h"]h$]h&]name5http://man7.org/linux/man-pages/man2/getrlimit.2.htmljj uh1jshj ubh)}(h7h]h}(h]3http-man7-org-linux-man-pages-man2-getrlimit-2-htmlah ]h"]5http://man7.org/linux/man-pages/man2/getrlimit.2.htmlah$]h&]refurij uh1hjKhj ubeh}(h]h ]h"]h$]h&]uh1hhhhM@hj ubeh}(h]jah ]h"]11ah$]h&](jj ehhuh1jZhhhM@hjIhhhKubj[)}(h<``_h](ja)}(h12h]h12}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h7http://man7.org/linux/man-pages/man5/limits.conf.5.html}(h7http://man7.org/linux/man-pages/man5/limits.conf.5.htmlhj, hhhNhNubah}(h]h ]h"]h$]h&]name7http://man7.org/linux/man-pages/man5/limits.conf.5.htmljj; uh1jshj) ubh)}(h9h]h}(h]5http-man7-org-linux-man-pages-man5-limits-conf-5-htmlah ]h"]7http://man7.org/linux/man-pages/man5/limits.conf.5.htmlah$]h&]refurij; uh1hjKhj) ubeh}(h]h ]h"]h$]h&]uh1hhhhMAhj ubeh}(h]jah ]h"]12ah$]h&]jahhuh1jZhhhMAhjIhhhKubj[)}(h/``_h](ja)}(h13h]h13}(hhhjZ hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hjV ubh)}(hjX h](jt)}(hjX h]h*https://sites.google.com/site/fullycapable}(h*https://sites.google.com/site/fullycapablehjk hhhNhNubah}(h]h ]h"]h$]h&]name*https://sites.google.com/site/fullycapablejjz uh1jshjh ubh)}(h,h]h}(h](https-sites-google-com-site-fullycapableah ]h"]*https://sites.google.com/site/fullycapableah$]h&]refurijz uh1hjKhjh ubeh}(h]h ]h"]h$]h&]uh1hhhhMBhjV ubeh}(h]jah ]h"]13ah$]h&]jahhuh1jZhhhMBhjIhhhKubj[)}(h7``_h](ja)}(h14h]h14}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h2http://man7.org/linux/man-pages/man8/auditd.8.html}(h2http://man7.org/linux/man-pages/man8/auditd.8.htmlhj hhhNhNubah}(h]h ]h"]h$]h&]name2http://man7.org/linux/man-pages/man8/auditd.8.htmljj uh1jshj ubh)}(h4h]h}(h]0http-man7-org-linux-man-pages-man8-auditd-8-htmlah ]h"]2http://man7.org/linux/man-pages/man8/auditd.8.htmlah$]h&]refurij uh1hjKhj ubeh}(h]h ]h"]h$]h&]uh1hhhhMChj ubeh}(h]jah ]h"]14ah$]h&]j ahhuh1jZhhhMChjIhhhKubj[)}(h6``_h](ja)}(h15h]h15}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h1https://man7.org/linux/man-pages/man8/sudo.8.html}(h1https://man7.org/linux/man-pages/man8/sudo.8.htmlhj hhhNhNubah}(h]h ]h"]h$]h&]name1https://man7.org/linux/man-pages/man8/sudo.8.htmljj uh1jshj ubh)}(h3h]h}(h]/https-man7-org-linux-man-pages-man8-sudo-8-htmlah ]h"]1https://man7.org/linux/man-pages/man8/sudo.8.htmlah$]h&]refurij uh1hjKhj ubeh}(h]h ]h"]h$]h&]uh1hhhhMDhj ubeh}(h]jah ]h"]15ah$]h&]jahhuh1jZhhhMDhjIhhhKubj[)}(h;``_h](ja)}(h16h]h16}(hhhj hhhNhNubah}(h]h ]h"]h$]h&]uh1j`hj ubh)}(hj h](jt)}(hj h]h6https://git.kernel.org/pub/scm/libs/libcap/libcap.git/}(h6https://git.kernel.org/pub/scm/libs/libcap/libcap.git/hj( hhhNhNubah}(h]h ]h"]h$]h&]name6https://git.kernel.org/pub/scm/libs/libcap/libcap.git/jj7 uh1jshj% ubh)}(h8h]h}(h]3https-git-kernel-org-pub-scm-libs-libcap-libcap-gitah ]h"]6https://git.kernel.org/pub/scm/libs/libcap/libcap.git/ah$]h&]refurij7 uh1hjKhj% ubeh}(h]h ]h"]h$]h&]uh1hhhhMEhj ubeh}(h]jDah ]h"]16ah$]h&](j?jehhuh1jZhhhMEhjIhhhKubeh}(h] bibliographyah ]h"] bibliographyah$]h&]uh1hhhhhhhhM4ubeh}(h](perf-events-and-tool-securityheh ]h"](perf events and tool security perf_securityeh$]h&]uh1hhhhhhhhKexpect_referenced_by_name}j` hsexpect_referenced_by_id}hhsubeh}(h]h ]h"]h$]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.confapep_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_spacefile_insertion_enabled raw_enabledKline_length_limitM'syntax_highlightlong smart_quotessmartquotes_locales]character_level_inline_markupdoctitle_xform docinfo_xformKsectsubtitle_xform image_loadinglinkembed_stylesheetcloak_email_addressessection_self_link embed_imagesenvNubreporterNindirect_targets]substitution_defs}substitution_names}refnames}(1]ha2](hj%jajj6jje3](jj9e4]jLa8]ja9]ja5](j!je6](j^jjpjjkje13]ja14]ja7]j%a10]ja15]ja16](j5je11](jje12]jaurefids}h]hasnameids}(j` hj_ j\ jAj>jSjPjjj[jXjFjCjjj>j;jW jT jhjjjj jjj jj j j[ j[jM jJ j j0j j j jmj j j j4j j jW jjI jF j jj j j jj j j jj j jS jjE jB j jj j j jj j j jj j jO jDjA j> u nametypes}(j` j_ NjANjSNjNj[NjFNjNj>NjW Njjjjj j j[ jM j j j j j j jW jI j j j j j j jS jE j j j j j j jO jA uh}(hhj\ hj>hhhjhjjj/j%jCj9jVjLjkjajjjjj+j!jPjDjhj^jjjjjjj jj/j%jjVjzjpjjjjjjj?j5jjjujkjXjjjj@j6jjjCj^jjojjjjjjj;jjjj jjT jIhj\jjj jjjjjj j j[j jJ jD j0j^ j j jmj j j j4j j j jj jF j@ jjZ j j jj j j jj j j jj jB j< jjV j j{ jj j j jj j j jDj j> j8 u footnote_refs}(j ]haj ](hj%jajj6jjej ](jj9ej ]jLaj ]jaj ]jaj ](j!jej ](j^jjpjjkjej ]jaj ]jaj ]j%aj ]jaj ]jaj ](j5jej ](jjej ]jau citation_refs} autofootnotes]autofootnote_refs]symbol_footnotes]symbol_footnote_refs] footnotes](j\jjj j^ j j j jZ j j j jV j j j e citations]autofootnote_startKsymbol_footnote_startK id_counter collectionsCounter}j K/sRparse_messages](hsystem_message)}(hhh]h)}(h:Enumerated list start value not ordinal-1: "2" (ordinal 2)h]h>Enumerated list start value not ordinal-1: “2” (ordinal 2)}(hhhj# hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubah}(h]h ]h"]h$]h&]levelKtypeINFOsourcehlineKuh1j hjVhhhhhKubj )}(hhh]h)}(h:Enumerated list start value not ordinal-1: "2" (ordinal 2)h]h>Enumerated list start value not ordinal-1: “2” (ordinal 2)}(hhhj? hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj< ubah}(h]h ]h"]h$]h&]levelKtypej9 sourcehlineKuh1j hjVhhhhhKubj )}(hhh]h)}(h:Enumerated list start value not ordinal-1: "3" (ordinal 3)h]h>Enumerated list start value not ordinal-1: “3” (ordinal 3)}(hhhjZ hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjW ubah}(h]h ]h"]h$]h&]levelKtypej9 sourcehlineKuh1j hjVhhhhhKubetransform_messages]j )}(hhh]h)}(hhh]h3Hyperlink target "perf-security" is not referenced.}(hhhjw ubah}(h]h ]h"]h$]h&]uh1hhjt ubah}(h]h ]h"]h$]h&]levelKtypej9 sourcehlineKuh1j uba transformerN include_log] decorationNhhub.