€•nD      Œsphinx.addnodes”Œdocument”“”)”}”(Œ	rawsource”Œ ”Œchildren”]”(Œtranslations”ŒLanguagesNode”“”)”}”(hhh]”(h Œpending_xref”“”)”}”(hhh]”Œdocutils.nodes”ŒText”“”ŒChinese (Simplified)”…””}”Œparent”hsbaŒ
attributes”}”(Œids”]”Œclasses”]”Œnames”]”Œdupnames”]”Œbackrefs”]”Œ	refdomain”Œstd”Œreftype”Œdoc”Œ	reftarget”Œ(/translations/zh_CN/trace/hwlat_detector”Œmodname”NŒ	classname”NŒrefexplicit”ˆuŒtagname”hhhubh)”}”(hhh]”hŒChinese (Traditional)”…””}”hh2sbah}”(h]”h ]”h"]”h$]”h&]”Œ	refdomain”h)Œreftype”h+Œ	reftarget”Œ(/translations/zh_TW/trace/hwlat_detector”Œmodname”NŒ	classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒItalian”…””}”hhFsbah}”(h]”h ]”h"]”h$]”h&]”Œ	refdomain”h)Œreftype”h+Œ	reftarget”Œ(/translations/it_IT/trace/hwlat_detector”Œmodname”NŒ	classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒJapanese”…””}”hhZsbah}”(h]”h ]”h"]”h$]”h&]”Œ	refdomain”h)Œreftype”h+Œ	reftarget”Œ(/translations/ja_JP/trace/hwlat_detector”Œmodname”NŒ	classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒKorean”…””}”hhnsbah}”(h]”h ]”h"]”h$]”h&]”Œ	refdomain”h)Œreftype”h+Œ	reftarget”Œ(/translations/ko_KR/trace/hwlat_detector”Œmodname”NŒ	classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒSpanish”…””}”hh‚sbah}”(h]”h ]”h"]”h$]”h&]”Œ	refdomain”h)Œreftype”h+Œ	reftarget”Œ(/translations/sp_SP/trace/hwlat_detector”Œmodname”NŒ	classname”NŒrefexplicit”ˆuh1hhhubeh}”(h]”h ]”h"]”h$]”h&]”Œcurrent_language”ŒEnglish”uh1h
hhŒ	_document”hŒsource”NŒline”NubhŒsection”“”)”}”(hhh]”(hŒtitle”“”)”}”(hŒHardware Latency Detector”h]”hŒHardware Latency Detector”…””}”(hh¨hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1h¦hh£hžhhŸŒB/var/lib/git/docbuild/linux/Documentation/trace/hwlat_detector.rst”h Kubh¢)”}”(hhh]”(h§)”}”(hŒIntroduction”h]”hŒIntroduction”…””}”(hhºhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1h¦hh·hžhhŸh¶h KubhŒ	paragraph”“”)”}”(hXË  The tracer hwlat_detector is a special purpose tracer that is used to
detect large system latencies induced by the behavior of certain underlying
hardware or firmware, independent of Linux itself. The code was developed
originally to detect SMIs (System Management Interrupts) on x86 systems,
however there is nothing x86 specific about this patchset. It was
originally written for use by the "RT" patch since the Real Time
kernel is highly latency sensitive.”h]”hXÏ  The tracer hwlat_detector is a special purpose tracer that is used to
detect large system latencies induced by the behavior of certain underlying
hardware or firmware, independent of Linux itself. The code was developed
originally to detect SMIs (System Management Interrupts) on x86 systems,
however there is nothing x86 specific about this patchset. It was
originally written for use by the â€œRTâ€ patch since the Real Time
kernel is highly latency sensitive.”…””}”(hhÊhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h Khh·hžhubhÉ)”}”(hX  SMIs are not serviced by the Linux kernel, which means that it does not
even know that they are occurring. SMIs are instead set up by BIOS code
and are serviced by BIOS code, usually for "critical" events such as
management of thermal sensors and fans. Sometimes though, SMIs are used for
other tasks and those tasks can spend an inordinate amount of time in the
handler (sometimes measured in milliseconds). Obviously this is a problem if
you are trying to keep event service latencies down in the microsecond range.”h]”hX	  SMIs are not serviced by the Linux kernel, which means that it does not
even know that they are occurring. SMIs are instead set up by BIOS code
and are serviced by BIOS code, usually for â€œcriticalâ€ events such as
management of thermal sensors and fans. Sometimes though, SMIs are used for
other tasks and those tasks can spend an inordinate amount of time in the
handler (sometimes measured in milliseconds). Obviously this is a problem if
you are trying to keep event service latencies down in the microsecond range.”…””}”(hhØhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h Khh·hžhubhÉ)”}”(hX©  The hardware latency detector works by hogging one of the cpus for configurable
amounts of time (with interrupts disabled), polling the CPU Time Stamp Counter
for some period, then looking for gaps in the TSC data. Any gap indicates a
time when the polling was interrupted and since the interrupts are disabled,
the only thing that could do that would be an SMI or other hardware hiccup
(or an NMI, but those can be tracked).”h]”hX©  The hardware latency detector works by hogging one of the cpus for configurable
amounts of time (with interrupts disabled), polling the CPU Time Stamp Counter
for some period, then looking for gaps in the TSC data. Any gap indicates a
time when the polling was interrupted and since the interrupts are disabled,
the only thing that could do that would be an SMI or other hardware hiccup
(or an NMI, but those can be tracked).”…””}”(hhæhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h Khh·hžhubhÉ)”}”(hŒÒNote that the hwlat detector should *NEVER* be used in a production environment.
It is intended to be run manually to determine if the hardware platform has a
problem with long system firmware service routines.”h]”(hŒ$Note that the hwlat detector should ”…””}”(hhôhžhhŸNh NubhŒemphasis”“”)”}”(hŒ*NEVER*”h]”hŒNEVER”…””}”(hhþhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hühhôubhŒ§ be used in a production environment.
It is intended to be run manually to determine if the hardware platform has a
problem with long system firmware service routines.”…””}”(hhôhžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h Khh·hžhubeh}”(h]”Œintroduction”ah ]”h"]”Œintroduction”ah$]”h&]”uh1h¡hh£hžhhŸh¶h Kubh¢)”}”(hhh]”(h§)”}”(hŒUsage”h]”hŒUsage”…””}”(hj!  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1h¦hj  hžhhŸh¶h K$ubhÉ)”}”(hŒýWrite the ASCII text "hwlat" into the current_tracer file of the tracing system
(mounted at /sys/kernel/tracing or /sys/kernel/tracing). It is possible to
redefine the threshold in microseconds (us) above which latency spikes will
be taken into account.”h]”hX  Write the ASCII text â€œhwlatâ€ into the current_tracer file of the tracing system
(mounted at /sys/kernel/tracing or /sys/kernel/tracing). It is possible to
redefine the threshold in microseconds (us) above which latency spikes will
be taken into account.”…””}”(hj/  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h K&hj  hžhubhÉ)”}”(hŒ	Example::”h]”hŒExample:”…””}”(hj=  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h K+hj  hžhubhŒliteral_block”“”)”}”(hŒa# echo hwlat > /sys/kernel/tracing/current_tracer
# echo 100 > /sys/kernel/tracing/tracing_thresh”h]”hŒa# echo hwlat > /sys/kernel/tracing/current_tracer
# echo 100 > /sys/kernel/tracing/tracing_thresh”…””}”hjM  sbah}”(h]”h ]”h"]”h$]”h&]”Œ	xml:space”Œpreserve”uh1jK  hŸh¶h K-hj  hžhubhÉ)”}”(hŒNThe /sys/kernel/tracing/hwlat_detector interface contains the following files:”h]”hŒNThe /sys/kernel/tracing/hwlat_detector interface contains the following files:”…””}”(hj]  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h K0hj  hžhubhŒblock_quote”“”)”}”(hŒ´- width - time period to sample with CPUs held (usecs)
          must be less than the total window size (enforced)
- window - total period of sampling, width being inside (usecs)
”h]”hŒbullet_list”“”)”}”(hhh]”(hŒ	list_item”“”)”}”(hŒowidth - time period to sample with CPUs held (usecs)
        must be less than the total window size (enforced)”h]”hŒdefinition_list”“”)”}”(hhh]”hŒdefinition_list_item”“”)”}”(hŒgwidth - time period to sample with CPUs held (usecs)
must be less than the total window size (enforced)”h]”(hŒterm”“”)”}”(hŒ4width - time period to sample with CPUs held (usecs)”h]”hŒ4width - time period to sample with CPUs held (usecs)”…””}”(hj‰  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j‡  hŸh¶h K2hjƒ  ubhŒ
definition”“”)”}”(hhh]”hÉ)”}”(hŒ2must be less than the total window size (enforced)”h]”hŒ2must be less than the total window size (enforced)”…””}”(hjœ  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h K3hj™  ubah}”(h]”h ]”h"]”h$]”h&]”uh1j—  hjƒ  ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j  hŸh¶h K2hj~  ubah}”(h]”h ]”h"]”h$]”h&]”uh1j|  hjx  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hjs  ubjw  )”}”(hŒ>window - total period of sampling, width being inside (usecs)
”h]”hÉ)”}”(hŒ=window - total period of sampling, width being inside (usecs)”h]”hŒ=window - total period of sampling, width being inside (usecs)”…””}”(hjÆ  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h K4hjÂ  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hjs  ubeh}”(h]”h ]”h"]”h$]”h&]”Œbullet”Œ-”uh1jq  hŸh¶h K2hjm  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jk  hŸh¶h K2hj  hžhubhÉ)”}”(hXz  By default the width is set to 500,000 and window to 1,000,000, meaning that
for every 1,000,000 usecs (1s) the hwlat detector will spin for 500,000 usecs
(0.5s). If tracing_thresh contains zero when hwlat tracer is enabled, it will
change to a default of 10 usecs. If any latencies that exceed the threshold is
observed then the data will be written to the tracing ring buffer.”h]”hXz  By default the width is set to 500,000 and window to 1,000,000, meaning that
for every 1,000,000 usecs (1s) the hwlat detector will spin for 500,000 usecs
(0.5s). If tracing_thresh contains zero when hwlat tracer is enabled, it will
change to a default of 10 usecs. If any latencies that exceed the threshold is
observed then the data will be written to the tracing ring buffer.”…””}”(hjè  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h K6hj  hžhubhÉ)”}”(hŒ£The minimum sleep time between periods is 1 millisecond. Even if width
is less than 1 millisecond apart from window, to allow the system to not
be totally starved.”h]”hŒ£The minimum sleep time between periods is 1 millisecond. Even if width
is less than 1 millisecond apart from window, to allow the system to not
be totally starved.”…””}”(hjö  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h K<hj  hžhubhÉ)”}”(hX-  If tracing_thresh was zero when hwlat detector was started, it will be set
back to zero if another tracer is loaded. Note, the last value in
tracing_thresh that hwlat detector had will be saved and this value will
be restored in tracing_thresh if it is still zero when hwlat detector is
started again.”h]”hX-  If tracing_thresh was zero when hwlat detector was started, it will be set
back to zero if another tracer is loaded. Note, the last value in
tracing_thresh that hwlat detector had will be saved and this value will
be restored in tracing_thresh if it is still zero when hwlat detector is
started again.”…””}”(hj  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h K@hj  hžhubhÉ)”}”(hŒEThe following tracing directory files are used by the hwlat_detector:”h]”hŒEThe following tracing directory files are used by the hwlat_detector:”…””}”(hj  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KFhj  hžhubhÉ)”}”(hŒin /sys/kernel/tracing:”h]”hŒin /sys/kernel/tracing:”…””}”(hj   hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KHhj  hžhubjl  )”}”(hXœ  - tracing_threshold    - minimum latency value to be considered (usecs)
- tracing_max_latency  - maximum hardware latency actually observed (usecs)
- tracing_cpumask      - the CPUs to move the hwlat thread across
- hwlat_detector/width - specified amount of time to spin within window (usecs)
- hwlat_detector/window        - amount of time between (width) runs (usecs)
- hwlat_detector/mode  - the thread mode
”h]”jr  )”}”(hhh]”(jw  )”}”(hŒEtracing_threshold    - minimum latency value to be considered (usecs)”h]”hÉ)”}”(hj7  h]”hŒEtracing_threshold    - minimum latency value to be considered (usecs)”…””}”(hj9  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KJhj5  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hj2  ubjw  )”}”(hŒItracing_max_latency  - maximum hardware latency actually observed (usecs)”h]”hÉ)”}”(hjN  h]”hŒItracing_max_latency  - maximum hardware latency actually observed (usecs)”…””}”(hjP  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KKhjL  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hj2  ubjw  )”}”(hŒ?tracing_cpumask      - the CPUs to move the hwlat thread across”h]”hÉ)”}”(hje  h]”hŒ?tracing_cpumask      - the CPUs to move the hwlat thread across”…””}”(hjg  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KLhjc  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hj2  ubjw  )”}”(hŒMhwlat_detector/width - specified amount of time to spin within window (usecs)”h]”hÉ)”}”(hj|  h]”hŒMhwlat_detector/width - specified amount of time to spin within window (usecs)”…””}”(hj~  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KMhjz  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hj2  ubjw  )”}”(hŒJhwlat_detector/window        - amount of time between (width) runs (usecs)”h]”hÉ)”}”(hj“  h]”hŒJhwlat_detector/window        - amount of time between (width) runs (usecs)”…””}”(hj•  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KNhj‘  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hj2  ubjw  )”}”(hŒ'hwlat_detector/mode  - the thread mode
”h]”hÉ)”}”(hŒ&hwlat_detector/mode  - the thread mode”h]”hŒ&hwlat_detector/mode  - the thread mode”…””}”(hj¬  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KOhj¨  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hj2  ubeh}”(h]”h ]”h"]”h$]”h&]”jà  já  uh1jq  hŸh¶h KJhj.  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jk  hŸh¶h KJhj  hžhubhÉ)”}”(hŒñBy default, one hwlat detector's kernel thread will migrate across each CPU
specified in cpumask at the beginning of a new window, in a round-robin
fashion. This behavior can be changed by changing the thread mode,
the available options are:”h]”hŒóBy default, one hwlat detectorâ€™s kernel thread will migrate across each CPU
specified in cpumask at the beginning of a new window, in a round-robin
fashion. This behavior can be changed by changing the thread mode,
the available options are:”…””}”(hjÌ  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KQhj  hžhubjl  )”}”(hŒ¬- none:        do not force migration
- round-robin: migrate across each CPU specified in cpumask [default]
- per-cpu:     create one thread for each cpu in tracing_cpumask”h]”jr  )”}”(hhh]”(jw  )”}”(hŒ#none:        do not force migration”h]”hÉ)”}”(hjã  h]”hŒ#none:        do not force migration”…””}”(hjå  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KVhjá  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hjÞ  ubjw  )”}”(hŒCround-robin: migrate across each CPU specified in cpumask [default]”h]”hÉ)”}”(hjú  h]”hŒCround-robin: migrate across each CPU specified in cpumask [default]”…””}”(hjü  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KWhjø  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hjÞ  ubjw  )”}”(hŒ>per-cpu:     create one thread for each cpu in tracing_cpumask”h]”hÉ)”}”(hj  h]”hŒ>per-cpu:     create one thread for each cpu in tracing_cpumask”…””}”(hj  hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÈhŸh¶h KXhj  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jv  hjÞ  ubeh}”(h]”h ]”h"]”h$]”h&]”jà  já  uh1jq  hŸh¶h KVhjÚ  ubah}”(h]”h ]”h"]”h$]”h&]”uh1jk  hŸh¶h KVhj  hžhubeh}”(h]”Œusage”ah ]”h"]”Œusage”ah$]”h&]”uh1h¡hh£hžhhŸh¶h K$ubeh}”(h]”Œhardware-latency-detector”ah ]”h"]”Œhardware latency detector”ah$]”h&]”uh1h¡hhhžhhŸh¶h Kubeh}”(h]”h ]”h"]”h$]”h&]”Œsource”h¶uh1hŒcurrent_source”NŒcurrent_line”NŒsettings”Œdocutils.frontend”ŒValues”“”)”}”(h¦NŒ	generator”NŒ	datestamp”NŒsource_link”NŒ
source_url”NŒtoc_backlinks”Œentry”Œfootnote_backlinks”KŒsectnum_xform”KŒstrip_comments”NŒstrip_elements_with_classes”NŒstrip_classes”NŒreport_level”KŒ
halt_level”KŒexit_status_level”KŒdebug”NŒwarning_stream”NŒ	traceback”ˆŒinput_encoding”Œ	utf-8-sig”Œinput_encoding_error_handler”Œstrict”Œoutput_encoding”Œutf-8”Œoutput_encoding_error_handler”je  Œerror_encoding”Œutf-8”Œerror_encoding_error_handler”Œbackslashreplace”Œlanguage_code”Œen”Œrecord_dependencies”NŒconfig”NŒ	id_prefix”hŒauto_id_prefix”Œid”Œdump_settings”NŒdump_internals”NŒdump_transforms”NŒdump_pseudo_xml”NŒexpose_internals”NŒstrict_visitor”NŒ_disable_config”NŒ_source”h¶Œ_destination”NŒ_config_files”]”Œ7/var/lib/git/docbuild/linux/Documentation/docutils.conf”aŒfile_insertion_enabled”ˆŒraw_enabled”KŒline_length_limit”M'Œpep_references”NŒpep_base_url”Œhttps://peps.python.org/”Œpep_file_url_template”Œpep-%04d”Œrfc_references”NŒrfc_base_url”Œ&https://datatracker.ietf.org/doc/html/”Œ	tab_width”KŒtrim_footnote_reference_space”‰Œsyntax_highlight”Œlong”Œsmart_quotes”ˆŒsmartquotes_locales”]”Œcharacter_level_inline_markup”‰Œdoctitle_xform”‰Œdocinfo_xform”KŒsectsubtitle_xform”‰Œimage_loading”Œlink”Œembed_stylesheet”‰Œcloak_email_addresses”ˆŒsection_self_link”‰Œenv”NubŒreporter”NŒindirect_targets”]”Œsubstitution_defs”}”Œsubstitution_names”}”Œrefnames”}”Œrefids”}”Œnameids”}”(j?  j<  j  j  j7  j4  uŒ	nametypes”}”(j?  ‰j  ‰j7  ‰uh}”(j<  h£j  h·j4  j  uŒfootnote_refs”}”Œcitation_refs”}”Œautofootnotes”]”Œautofootnote_refs”]”Œsymbol_footnotes”]”Œsymbol_footnote_refs”]”Œ	footnotes”]”Œ	citations”]”Œautofootnote_start”KŒsymbol_footnote_start”K Œ
id_counter”Œcollections”ŒCounter”“”}”…”R”Œparse_messages”]”Œtransform_messages”]”Œtransformer”NŒinclude_log”]”Œ
decoration”Nhžhub.