diff options
Diffstat (limited to 'lib/tools/doc/src')
-rw-r--r-- | lib/tools/doc/src/lcnt.xml | 104 | ||||
-rw-r--r-- | lib/tools/doc/src/lcnt_chapter.xml | 9 |
2 files changed, 92 insertions, 21 deletions
diff --git a/lib/tools/doc/src/lcnt.xml b/lib/tools/doc/src/lcnt.xml index 31e5c241e9..922c2ac0f4 100644 --- a/lib/tools/doc/src/lcnt.xml +++ b/lib/tools/doc/src/lcnt.xml @@ -109,14 +109,6 @@ statistics. If the server held any lock statistics data before the collect then that data is lost. </p> - <note> - <p> - When collection occurs the runtime system transitions to a single thread, - blocking all other threads. No other tasks will be scheduled during this - operation. Depending on the size of the data this might take a long time - (several seconds) and cause timeouts in the system. - </p> - </note> </desc> </func> @@ -322,24 +314,22 @@ <func> <name>apply(Fun) -> term()</name> <fsummary>Same as <c>apply(Fun, [])</c>.</fsummary> + <type> + <v>Fun = fun()</v> + </type> <desc> <p>Same as <c>apply(Fun, [])</c>.</p> </desc> </func> <func> <name>apply(Fun, Args) -> term()</name> - <fsummary>Clears counters, applies function and collects the profiling results.</fsummary> + <fsummary>Same as <c>apply(Module, Function, Args)</c>.</fsummary> <type> <v>Fun = fun()</v> <v>Args = [term()]</v> </type> <desc> - <p> Clears the lock counters and then setups the instrumentation to save all destroyed locks. - After setup the fun is called, passing the elements in <c>Args</c> as arguments. - When the fun returns the statistics are immediately collected to the server. After the - collection the instrumentation is returned to its previous behavior. - The result of the applied fun is returned. - </p> + <p>Same as <c>apply(Module, Function, Args)</c>.</p> </desc> </func> <func> @@ -357,6 +347,13 @@ collection the instrumentation is returned to its previous behavior. The result of the applied function is returned. </p> + <warning> + <p> + This function should only be used for micro-benchmarks; it sets <c>copy_save</c> + to <c>true</c> for the duration of the call, which can quickly lead to running + out of memory. + </p> + </warning> </desc> </func> @@ -429,6 +426,68 @@ <desc> <p>Clear the internal counters. Same as <c>lcnt:clear(Node)</c>.</p></desc> </func> + <func> + <name>rt_mask() -> [category_atom()]</name> + <fsummary>Same as <c>rt_mask(node())</c>.</fsummary> + <desc><p>Same as <c>rt_mask(node())</c>.</p></desc> + </func> + + <func> + <name>rt_mask(Node) -> [category_atom()]</name> + <fsummary>Returns the current lock category mask.</fsummary> + <type> + <v>Node = node()</v> + </type> + <desc> + <p> + Refer to <c>rt_mask/2</c> for a list of valid categories. All + categories are enabled by default. + </p> + </desc> + </func> + + <func> + <name>rt_mask(Categories) -> ok | {error, copy_save_enabled}</name> + <fsummary>Same as <c>rt_mask(node(), Categories)</c>.</fsummary> + <type> + <v>Categories = [atom()]</v> + </type> + <desc><p>Same as <c>rt_mask(node(), Categories)</c>.</p></desc> + </func> + + <func> + <name>rt_mask(Node, Categories) -> ok | {error, copy_save_enabled}</name> + <fsummary>Changes the lock category mask.</fsummary> + <type> + <v>Node = node()</v> + <v>Categories = [atom()]</v> + </type> + <desc> + <p> + Sets the lock category mask to the given categories. + </p> + <p> + This will fail if the <c>copy_save</c> option is enabled; see + <c>lcnt:rt_opt/2</c>. + </p> + <p>Valid categories are:</p> + <taglist> + <item><c>allocator</c></item> + <item><c>db</c> (ETS tables)</item> + <item><c>debug</c></item> + <item><c>distribution</c></item> + <item><c>generic</c></item> + <item><c>io</c></item> + <item><c>process</c></item> + <item><c>scheduler</c></item> + </taglist> + <p> + This list is subject to change at any time, as is the category any given lock + may belong to. + </p> + </desc> + </func> + <func> <name>rt_opt({Type, bool()}) -> bool()</name> <fsummary>Same as <c>rt_opt(node(), {Type, Opt})</c>.</fsummary> @@ -442,16 +501,25 @@ <v>Type = copy_save | process_locks</v> </type> <desc> - <p>Changes the lock counter behavior and returns the previous behaviour.</p> <p>Option description:</p> <taglist> <tag><c>{copy_save, bool()}</c></tag> - <item>Enable statistics saving from destroyed locks by copying. This might consume a lot of memory. + <item>Retains the statistics of destroyed locks. <br/>Default: <c>false</c> + <warning> + <p> + This option will use a lot of memory when enabled, which must be + reclaimed with <c>lcnt:rt_clear</c>. Note that it makes no distinction + between locks that were destroyed and locks for which counting was + disabled, so enabling this option will disable changes to the lock + category mask. + </p> + </warning> </item> <tag><c>{process_locks, bool()}</c></tag> - <item>Profile process locks. + <item>Profile process locks, equal to adding <c>process</c> to the lock category mask; + see <c>lcnt:rt_mask/2</c> <br/>Default: <c>true</c> </item> </taglist> diff --git a/lib/tools/doc/src/lcnt_chapter.xml b/lib/tools/doc/src/lcnt_chapter.xml index c73fcb31e0..24b58136aa 100644 --- a/lib/tools/doc/src/lcnt_chapter.xml +++ b/lib/tools/doc/src/lcnt_chapter.xml @@ -29,7 +29,7 @@ <approved>nobody</approved> <checked>no</checked> <date>2009-11-26</date> - <rev>PA1</rev> + <rev>PA2</rev> <file>lcnt_chapter.xml</file> </header> <p> @@ -97,8 +97,11 @@ ok ok </pre> <p> - Another way to to profile a specific function is to use <c>lcnt:apply/3</c> or <c>lcnt:apply/1</c> which does <c>lcnt:clear/0</c> before the function and <c>lcnt:collect/0</c> after its invocation. - It also sets <c>copy_save</c> to <c>true</c> for the duration of the function call + Another way to to profile a specific function is to use <c>lcnt:apply/3</c> or <c>lcnt:apply/1</c> + which does <c>lcnt:clear/0</c> before the function and <c>lcnt:collect/0</c> after its invocation. + This method should only be used in micro-benchmarks since it sets <c>copy_save</c> to <c>true</c> + for the duration of the function call, which may cause the emulator to run out of memory if + attempted under load. </p> <pre> Erlang R13B03 (erts-5.7.4) [source] [smp:8:8] [rq:8] [async-threads:0] [hipe] |