diff options
Diffstat (limited to 'erts/doc/src')
| -rw-r--r-- | erts/doc/src/counters.xml | 36 | ||||
| -rw-r--r-- | erts/doc/src/erl_nif.xml | 69 | ||||
| -rw-r--r-- | erts/doc/src/notes.xml | 20 | ||||
| -rw-r--r-- | erts/doc/src/persistent_term.xml | 2 |
4 files changed, 97 insertions, 30 deletions
diff --git a/erts/doc/src/counters.xml b/erts/doc/src/counters.xml index 85eedfdadc..ba4a22759f 100644 --- a/erts/doc/src/counters.xml +++ b/erts/doc/src/counters.xml @@ -85,14 +85,22 @@ </item> <tag><c>write_concurrency</c></tag> <item><p>This is an optimization to achieve very efficient concurrent - write operations at the expense of potential read inconsistency and memory - consumption per counter.</p> + <seealso marker="#add/3"><c>add</c></seealso> and <seealso + marker="#sub/3"><c>sub</c></seealso> operations at the expense of potential read + inconsistency and memory consumption per counter.</p> <p>Read operations may see sequentially inconsistent results with regard to concurrent write operations. Even if write operation A is done sequencially before write operation B, a concurrent reader may see any combination of A and B, including only B. A read operation is only guaranteed to see all writes done sequentially before the read. No writes are ever lost, but will eventually all be seen.</p> + <p>The typical use case for <c>write_concurrency</c> is when + concurrent calls to <seealso marker="#add/3"><c>add</c></seealso> and + <seealso marker="#sub/3"><c>sub</c></seealso> toward the same counters + are very frequent, while calls to <seealso marker="#get/2"><c>get</c> + </seealso> and <seealso marker="#put/3"><c>put</c></seealso> are much + less frequent. The lack of absolute read consistency must also be + acceptable.</p> </item> </taglist> </desc> @@ -110,7 +118,8 @@ <name name="add" arity="3"/> <fsummary>Add to counter</fsummary> <desc> - <p>Add <c><anno>Incr</anno></c> to counter.</p> + <p>Add <c><anno>Incr</anno></c> to counter at index + <c><anno>Ix</anno></c>.</p> </desc> </func> @@ -118,7 +127,26 @@ <name name="sub" arity="3"/> <fsummary>Subtract from counter</fsummary> <desc> - <p>Subtract <c><anno>Decr</anno></c> from counter.</p> + <p>Subtract <c><anno>Decr</anno></c> from counter at index + <c><anno>Ix</anno></c>.</p> + </desc> + </func> + + <func> + <name name="put" arity="3"/> + <fsummary>Set counter to value</fsummary> + <desc> + <p>Write <c><anno>Value</anno></c> to counter at index + <c><anno>Ix</anno></c>.</p> + <note> + <p>Despite its name, the <c>write_concurrency</c> optimization does not + improve <c>put</c>. A call to <c>put</c> is a relative heavy + operation compared to the very lightweight and scalable <seealso + marker="#add/3"><c>add</c></seealso> and <seealso marker="#sub/3"> + <c>sub</c></seealso>. The cost for a <c>put</c> with + <c>write_concurrency</c> is lika a <seealso marker="#get/2"><c>get</c> + </seealso> plus a <c>put</c> without <c>write_concurrency</c>.</p> + </note> </desc> </func> diff --git a/erts/doc/src/erl_nif.xml b/erts/doc/src/erl_nif.xml index 190ec12d0e..bbc12b0a56 100644 --- a/erts/doc/src/erl_nif.xml +++ b/erts/doc/src/erl_nif.xml @@ -293,7 +293,7 @@ return term;</code> arguments. When you write to a shared state either through static variables or <seealso marker="#enif_priv_data"> <c>enif_priv_data</c></seealso>, you need to supply your own explicit - synchronization. This includes terms in process-independent + synchronization. This includes terms in process independent environments that are shared between threads. Resource objects also require synchronization if you treat them as mutable.</p> <p>The library initialization callbacks <c>load</c> and @@ -596,7 +596,7 @@ int writeiovec(ErlNifEnv *env, ERL_NIF_TERM term, ERL_NIF_TERM *tail, <c>--enable-static-nifs</c>, you must define <c>STATIC_ERLANG_NIF</c> before the <c>ERL_NIF_INIT</c> declaration.</p> </item> - <tag><marker id="load"/><c>int (*load)(ErlNifEnv* env, void** priv_data, + <tag><marker id="load"/><c>int (*load)(ErlNifEnv* caller_env, void** priv_data, ERL_NIF_TERM load_info)</c></tag> <item> <p><c>load</c> is called when the NIF library is loaded @@ -612,7 +612,7 @@ int writeiovec(ErlNifEnv *env, ERL_NIF_TERM term, ERL_NIF_TERM *tail, anything other than <c>0</c>. <c>load</c> can be <c>NULL</c> if initialization is not needed.</p> </item> - <tag><marker id="upgrade"/><c>int (*upgrade)(ErlNifEnv* env, void** + <tag><marker id="upgrade"/><c>int (*upgrade)(ErlNifEnv* caller_env, void** priv_data, void** old_priv_data, ERL_NIF_TERM load_info)</c></tag> <item> <p><c>upgrade</c> is called when the NIF library is loaded @@ -626,7 +626,7 @@ int writeiovec(ErlNifEnv *env, ERL_NIF_TERM term, ERL_NIF_TERM *tail, <p>The library fails to load if <c>upgrade</c> returns anything other than <c>0</c> or if <c>upgrade</c> is <c>NULL</c>.</p> </item> - <tag><marker id="unload"/><c>void (*unload)(ErlNifEnv* env, void* + <tag><marker id="unload"/><c>void (*unload)(ErlNifEnv* caller_env, void* priv_data)</c></tag> <item> <p><c>unload</c> is called when the module code that @@ -654,27 +654,41 @@ int writeiovec(ErlNifEnv *env, ERL_NIF_TERM term, ERL_NIF_TERM *tail, <p><c>ErlNifEnv</c> represents an environment that can host Erlang terms. All terms in an environment are valid as long as the environment is valid. <c>ErlNifEnv</c> is an opaque type; pointers to - it can only be passed on to API functions. Two types of environments + it can only be passed on to API functions. Three types of environments exist:</p> <taglist> - <tag>Process-bound environment</tag> + <tag>Process bound environment</tag> <item> <p>Passed as the first argument to all NIFs. All function arguments passed to a NIF belong to that environment. The return value from a NIF must also be a term belonging to the same environment.</p> - <p>A process-bound environment contains transient information + <p>A process bound environment contains transient information about the calling Erlang process. The environment is only valid in the thread where it was supplied as argument until the NIF returns. It is thus useless and dangerous to store pointers to - process-bound environments between NIF calls.</p> + process bound environments between NIF calls.</p> </item> - <tag>Process-independent environment</tag> + <tag>Callback environment</tag> + <item> + <p>Passed as the first argument to all the non-NIF callback functions + (<seealso marker="#load"><c>load</c></seealso>, + <seealso marker="#upgrade"><c>upgrade</c></seealso>, + <seealso marker="#unload"><c>unload</c></seealso>, + <seealso marker="#ErlNifResourceDtor"><c>dtor</c></seealso>, + <seealso marker="#ErlNifResourceDown"><c>down</c></seealso> and + <seealso marker="#ErlNifResourceStop"><c>stop</c></seealso>). + Works like a process bound environment but with a temporary + pseudo process that "terminates" when the callback has + returned. Terms may be created in this environment but they will + only be accessible during the callback.</p> + </item> + <tag>Process independent environment</tag> <item> <p>Created by calling <seealso marker="#enif_alloc_env"> <c>enif_alloc_env</c></seealso>. This environment can be used to store terms between NIF calls and to send terms with <seealso marker="#enif_send"><c>enif_send</c></seealso>. A - process-independent environment with all its terms is valid until + process independent environment with all its terms is valid until you explicitly invalidate it with <seealso marker="#enif_free_env"><c>enif_free_env</c></seealso> or <c>enif_send</c>.</p> @@ -799,7 +813,7 @@ typedef struct { <tag><marker id="ErlNifResourceDtor"/><c>ErlNifResourceDtor</c></tag> <item> <code type="none"> -typedef void ErlNifResourceDtor(ErlNifEnv* env, void* obj);</code> +typedef void ErlNifResourceDtor(ErlNifEnv* caller_env, void* obj);</code> <p>The function prototype of a resource destructor function.</p> <p>The <c>obj</c> argument is a pointer to the resource. The only allowed use for the resource in the destructor is to access its @@ -809,7 +823,7 @@ typedef void ErlNifResourceDtor(ErlNifEnv* env, void* obj);</code> <tag><marker id="ErlNifResourceDown"/><c>ErlNifResourceDown</c></tag> <item> <code type="none"> -typedef void ErlNifResourceDown(ErlNifEnv* env, void* obj, ErlNifPid* pid, ErlNifMonitor* mon);</code> +typedef void ErlNifResourceDown(ErlNifEnv* caller_env, void* obj, ErlNifPid* pid, ErlNifMonitor* mon);</code> <p>The function prototype of a resource down function, called on the behalf of <seealso marker="#enif_monitor_process"> enif_monitor_process</seealso>. <c>obj</c> is the resource, <c>pid</c> @@ -820,7 +834,7 @@ typedef void ErlNifResourceDown(ErlNifEnv* env, void* obj, ErlNifPid* pid, ErlNi <tag><marker id="ErlNifResourceStop"/><c>ErlNifResourceStop</c></tag> <item> <code type="none"> -typedef void ErlNifResourceStop(ErlNifEnv* env, void* obj, ErlNifEvent event, int is_direct_call);</code> +typedef void ErlNifResourceStop(ErlNifEnv* caller_env, void* obj, ErlNifEvent event, int is_direct_call);</code> <p>The function prototype of a resource stop function, called on the behalf of <seealso marker="#enif_select"> enif_select</seealso>. <c>obj</c> is the resource, <c>event</c> is OS event, @@ -987,7 +1001,7 @@ typedef struct { <name><ret>ErlNifEnv *</ret><nametext>enif_alloc_env()</nametext></name> <fsummary>Create a new environment.</fsummary> <desc> - <p>Allocates a new process-independent environment. The environment can + <p>Allocates a new process independent environment. The environment can be used to hold terms that are not bound to any process. Such terms can later be copied to a process environment with <seealso marker="#enif_make_copy"><c>enif_make_copy</c></seealso> or @@ -1211,14 +1225,17 @@ typedef struct { </func> <func> - <name><ret>int</ret><nametext>enif_demonitor_process(ErlNifEnv* env, void* obj, + <name><ret>int</ret><nametext>enif_demonitor_process(ErlNifEnv* caller_env, void* obj, const ErlNifMonitor* mon)</nametext></name> <fsummary>Cancel a process monitor.</fsummary> <desc> <marker id="enif_demonitor_process"></marker> <p>Cancels a monitor created earlier with <seealso marker="#enif_monitor_process"> <c>enif_monitor_process</c></seealso>. Argument <c>obj</c> is a pointer - to the resource holding the monitor and <c>*mon</c> identifies the monitor.</p> + to the resource holding the monitor and <c>*mon</c> identifies the + monitor.</p> + <p>Argument <c>caller_env</c> is the environment of the calling process + or callback. Must only be NULL if calling from a custom thread.</p> <p>Returns <c>0</c> if the monitor was successfully identified and removed. Returns a non-zero value if the monitor could not be identified, which means it was either</p> @@ -2268,7 +2285,7 @@ enif_inspect_iovec(env, max_elements, term, &tail, &iovec); between nodes.</p> <list type="bulleted"> <item> - <p>Two resource terms will compare equal iff they + <p>Two resource terms will compare equal if and only if they would yield the same resource object pointer when passed to <seealso marker="#enif_get_resource"><c>enif_get_resource</c></seealso>.</p> </item> @@ -2572,7 +2589,7 @@ enif_map_iterator_destroy(env, &iter);</code> </func> <func> - <name><ret>int</ret><nametext>enif_monitor_process(ErlNifEnv* env, void* obj, + <name><ret>int</ret><nametext>enif_monitor_process(ErlNifEnv* caller_env, void* obj, const ErlNifPid* target_pid, ErlNifMonitor* mon)</nametext></name> <fsummary>Monitor a process from a resource.</fsummary> <desc> @@ -2593,6 +2610,8 @@ enif_map_iterator_destroy(env, &iter);</code> <seealso marker="#enif_compare_monitors"><c>enif_compare_monitors</c></seealso>. A monitor is automatically removed when it triggers or when the resource is deallocated.</p> + <p>Argument <c>caller_env</c> is the environment of the calling process + or callback. Must only be NULL if calling from a custom thread.</p> <p>Returns <c>0</c> on success, < 0 if no <c>down</c> callback is provided, and > 0 if the process is no longer alive.</p> <p>This function is only thread-safe when the emulator with SMP support @@ -2768,7 +2787,7 @@ enif_map_iterator_destroy(env, &iter);</code> <item>The port ID of the receiving port. The port ID is to refer to a port on the local node.</item> <tag><c>msg_env</c></tag> - <item>The environment of the message term. Can be a process-independent + <item>The environment of the message term. Can be a process independent environment allocated with <seealso marker="#enif_alloc_env"> <c>enif_alloc_env</c></seealso> or <c>NULL</c>.</item> <tag><c>msg</c></tag> @@ -3124,26 +3143,26 @@ if (retval & ERL_NIF_SELECT_STOP_CALLED) { <p>Initializes the <seealso marker="#ErlNifPid"><c>ErlNifPid</c></seealso> variable at <c>*pid</c> to represent the calling process.</p> <p>Returns <c>pid</c> if successful, or NULL if <c>caller_env</c> is not - a <seealso marker="#ErlNifEnv">process-bound environment</seealso>.</p> + a <seealso marker="#ErlNifEnv">process bound environment</seealso>.</p> </desc> </func> <func> - <name><ret>int</ret><nametext>enif_send(ErlNifEnv* env, ErlNifPid* to_pid, + <name><ret>int</ret><nametext>enif_send(ErlNifEnv* caller_env, ErlNifPid* to_pid, ErlNifEnv* msg_env, ERL_NIF_TERM msg)</nametext></name> <fsummary>Send a message to a process.</fsummary> <desc> <p>Sends a message to a process.</p> <taglist> - <tag><c>env</c></tag> - <item>The environment of the calling process. Must be <c>NULL</c> - only if calling from a created thread.</item> + <tag><c>caller_env</c></tag> + <item>The environment of the calling process or callback. Must be <c>NULL</c> + only if calling from a custom thread not spawned by ERTS.</item> <tag><c>*to_pid</c></tag> <item>The pid of the receiving process. The pid is to refer to a process on the local node.</item> <tag><c>msg_env</c></tag> <item>The environment of the message term. Must be a - process-independent environment allocated with + process independent environment allocated with <seealso marker="#enif_alloc_env"><c>enif_alloc_env</c></seealso> or NULL.</item> <tag><c>msg</c></tag> diff --git a/erts/doc/src/notes.xml b/erts/doc/src/notes.xml index f384adcf52..2a823d9fe7 100644 --- a/erts/doc/src/notes.xml +++ b/erts/doc/src/notes.xml @@ -31,6 +31,26 @@ </header> <p>This document describes the changes made to the ERTS application.</p> +<section><title>Erts 10.1.3</title> + + <section><title>Improvements and New Features</title> + <list> + <item> + <p>Added an optional <c>./configure</c> flag to compile + the emulator with spectre mitigation: + <c>--with-spectre-mitigation</c></p> + <p>Note that this requires a recent version of GCC with + support for spectre mitigation and the + <c>--mindirect-branch=thunk</c> flag, such as + <c>8.1</c>.</p> + <p> + Own Id: OTP-15430 Aux Id: ERIERL-237 </p> + </item> + </list> + </section> + +</section> + <section><title>Erts 10.1.2</title> <section><title>Fixed Bugs and Malfunctions</title> diff --git a/erts/doc/src/persistent_term.xml b/erts/doc/src/persistent_term.xml index d2a138d65f..29a6c67051 100644 --- a/erts/doc/src/persistent_term.xml +++ b/erts/doc/src/persistent_term.xml @@ -71,7 +71,7 @@ scan of their heaps for the term that has been deleted. While such scan is relatively light-weight, if there are many processes, the system can become less responsive until all - process have scanned theirs heaps.</p></item> + process have scanned their heaps.</p></item> <item><p>If the deleted term (or any part of it) is still used by a process, that process will do a major (fullsweep) garbage |