initial support for dirty schedulers and dirty NIFs

Add initial support for dirty schedulers.

There are two types of dirty schedulers: CPU schedulers and I/O
schedulers. By default, there are as many dirty CPU schedulers as there are
normal schedulers and as many dirty CPU schedulers online as normal
schedulers online. There are 10 dirty I/O schedulers (similar to the choice
of 10 as the default for async threads).

By default, dirty schedulers are disabled and conditionally compiled
out. To enable them, you must pass --enable-dirty-schedulers to the
top-level configure script when building Erlang/OTP.

Current dirty scheduler support requires the emulator to be built with SMP
support. This restriction will be lifted in the future.

You can specify the number of dirty schedulers with the command-line
options +SDcpu (for dirty CPU schedulers) and +SDio (for dirty I/O
schedulers). The +SDcpu option is similar to the +S option in that it takes
two numbers separated by a colon: C1:C2, where C1 specifies the number of
dirty schedulers available and C2 specifies the number of dirty schedulers
online. The +SDPcpu option allows numbers of dirty CPU schedulers available
and dirty CPU schedulers online to be specified as percentages, similar to
the existing +SP option for normal schedulers. The number of dirty CPU
schedulers created and dirty CPU schedulers online may not exceed the
number of normal schedulers created and normal schedulers online,
respectively. The +SDio option takes only a single number specifying the
number of dirty I/O schedulers available and online. There is no support
yet for programmatically changing at run time the number of dirty CPU
schedulers online via erlang:system_flag/2. Also, changing the number of
normal schedulers online via erlang:system_flag(schedulers_online,
NewSchedulersOnline) should ensure that there are no more dirty CPU
schedulers than normal schedulers, but this is not yet implemented. You can
retrieve the number of dirty schedulers by passing dirty_cpu_schedulers,
dirty_cpu_schedulers_online, or dirty_io_schedulers to
erlang:system_info/1.

Currently only NIFs are able to access dirty scheduler
functionality. Neither drivers nor BIFs currently support dirty
schedulers. This restriction will be addressed in the future.

If dirty scheduler support is present in the runtime, the initial status
line Erlang prints before presenting its interactive prompt will include
the indicator "[ds:C1:C2:I]" where "ds" indicates "dirty schedulers", "C1"
indicates the number of dirty CPU schedulers available, "C2" indicates the
number of dirty CPU schedulers online, and "I" indicates the number of
dirty I/O schedulers.

Document The dirty NIF API in the erl_nif man page. The API closely follows
Rickard Green's presentation slides from his talk "Future Extensions to the
Native Interface", presented at the 2011 Erlang Factory held in the San
Francisco Bay Area. Rickard's slides are available online at
http://bit.ly/1m34UHB .

Document the new erl command-line options, the additions to
erlang:system_info/1, and also add the erlang:system_flag/2 dirty scheduler
documentation even though it's not yet implemented.

To determine whether the dirty NIF API is available, native code can check
to see whether the C preprocessor macro ERL_NIF_DIRTY_SCHEDULER_SUPPORT is
defined. To check if dirty schedulers are available at run time, native
code can call the boolean enif_have_dirty_schedulers() function, and Erlang
code can call erlang:system_info(dirty_cpu_schedulers), which raises
badarg if no dirty scheduler support is available.

Add a simple dirty NIF test to the emulator NIF suite.

1 files changed, 151 insertions, 0 deletions

diff --git a/erts/emulator/beam/erl_nif.c b/erts/emulator/beam/erl_nif.c
index dc285b3cf7..e1e213c4eb 100644
--- a/erts/emulator/beam/erl_nif.c
+++ b/erts/emulator/beam/erl_nif.c
@@ -34,6 +34,7 @@
 #include "beam_bp.h"
 #include "erl_thr_progress.h"
 #include "dtrace-wrapper.h"
+#include "erl_process.h"
 #if defined(USE_DYNAMIC_TRACE) && (defined(USE_DTRACE) || defined(USE_SYSTEMTAP))
 #define HAVE_USE_DTRACE 1
 #endif
@@ -1451,6 +1452,156 @@ int enif_consume_timeslice(ErlNifEnv* env, int percent)
     return ERTS_BIF_REDS_LEFT(env->proc) == 0;
 }
 
+#ifdef ERTS_DIRTY_SCHEDULERS
+
+static void
+alloc_proc_psd(Process* proc, Export **ep)
+{
+    int i;
+    if (!*ep) {
+	*ep = erts_alloc(ERTS_ALC_T_PSD, sizeof(Export));
+	sys_memset((void*) *ep, 0, sizeof(Export));
+	for (i=0; i<ERTS_NUM_CODE_IX; i++) {
+	    (*ep)->addressv[i] = &(*ep)->code[3];
+	}
+	(*ep)->code[3] = (BeamInstr) em_call_nif;
+    }
+    (void) ERTS_PROC_SET_DIRTY_SCHED_TRAP_EXPORT(proc, ERTS_PROC_LOCK_MAIN, *ep);
+}
+
+static ERL_NIF_TERM
+execute_dirty_nif_finalizer(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[])
+{
+    Eterm* reg = ERTS_PROC_GET_SCHDATA(env->proc)->x_reg_array;
+    ERL_NIF_TERM result = (ERL_NIF_TERM) reg[0];
+    typedef ERL_NIF_TERM (*FinalizerFP)(ErlNifEnv*, ERL_NIF_TERM);
+    FinalizerFP fp;
+#if HAVE_INT64 && SIZEOF_LONG != 8
+    ASSERT(sizeof(fp) <= sizeof(ErlNifUInt64));
+    enif_get_uint64(env, reg[1], (ErlNifUInt64 *) &fp);
+#else
+    ASSERT(sizeof(fp) <= sizeof(unsigned long));
+    enif_get_ulong(env, reg[1], (unsigned long *) &fp);
+#endif
+    return (*fp)(env, result);
+}
+
+#endif /* ERTS_DIRTY_SCHEDULERS */
+
+#ifdef ERL_NIF_DIRTY_SCHEDULER_SUPPORT
+
+ERL_NIF_TERM
+enif_schedule_dirty_nif(ErlNifEnv* env, int flags,
+			ERL_NIF_TERM (*fp)(ErlNifEnv*, int, const ERL_NIF_TERM[]),
+			int argc, const ERL_NIF_TERM argv[])
+{
+#ifdef USE_THREADS
+    erts_aint32_t state, n, a;
+    Process* proc = env->proc;
+    Eterm* reg = ERTS_PROC_GET_SCHDATA(proc)->x_reg_array;
+    Export* ep = NULL;
+    int i;
+
+    int chkflgs = (flags & (ERL_NIF_DIRTY_JOB_IO_BOUND|ERL_NIF_DIRTY_JOB_CPU_BOUND));
+    if (chkflgs != ERL_NIF_DIRTY_JOB_IO_BOUND && chkflgs != ERL_NIF_DIRTY_JOB_CPU_BOUND)
+	return enif_make_badarg(env);
+
+    a = erts_smp_atomic32_read_acqb(&proc->state);
+    while (1) {
+	n = state = a;
+	if (chkflgs == ERL_NIF_DIRTY_JOB_CPU_BOUND)
+	    n |= ERTS_PSFLG_DIRTY_CPU_PROC;
+	else
+	    n |= ERTS_PSFLG_DIRTY_IO_PROC;
+	a = erts_smp_atomic32_cmpxchg_mb(&proc->state, n, state);
+	if (a == state)
+	    break;
+    }
+    if (!(ep = ERTS_PROC_GET_DIRTY_SCHED_TRAP_EXPORT(proc)))
+	alloc_proc_psd(proc, &ep);
+    ERTS_VBUMP_ALL_REDS(proc);
+    ep->code[2] = argc;
+    for (i = 0; i < argc; i++) {
+	reg[i] = (Eterm) argv[i];
+    }
+    proc->i = (BeamInstr*) ep->addressv[0];
+    ep->code[4] = (BeamInstr) fp;
+    proc->freason = TRAP;
+
+    return THE_NON_VALUE;
+#else
+    return (*fp)(env, argc, argv);
+#endif
+}
+
+ERL_NIF_TERM
+enif_schedule_dirty_nif_finalizer(ErlNifEnv* env, ERL_NIF_TERM result,
+				  ERL_NIF_TERM (*fp)(ErlNifEnv*, ERL_NIF_TERM))
+{
+#ifdef USE_THREADS
+    erts_aint32_t state, n, a;
+    Process* proc = env->proc;
+    Eterm* reg = ERTS_PROC_GET_SCHDATA(proc)->x_reg_array;
+    Export* ep;
+
+    a = erts_smp_atomic32_read_acqb(&proc->state);
+    while (1) {
+	n = state = a;
+	if (!(n & (ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q|ERTS_PSFLG_DIRTY_IO_PROC_IN_Q)))
+	    break;
+	n &= ~(ERTS_PSFLG_DIRTY_CPU_PROC|ERTS_PSFLG_DIRTY_IO_PROC
+	       |ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q|ERTS_PSFLG_DIRTY_IO_PROC_IN_Q);
+	a = erts_smp_atomic32_cmpxchg_mb(&proc->state, n, state);
+	if (a == state)
+	    break;
+    }
+    if (!(ep = ERTS_PROC_GET_DIRTY_SCHED_TRAP_EXPORT(proc)))
+	alloc_proc_psd(proc, &ep);
+    ERTS_VBUMP_ALL_REDS(proc);
+    ep->code[2] = 2;
+    reg[0] = (Eterm) result;
+#if HAVE_INT64 && SIZEOF_LONG != 8
+    ASSERT(sizeof(fp) <= sizeof(ErlNifUInt64));
+    reg[1] = (Eterm) enif_make_uint64(env, (ErlNifUInt64) fp);
+#else
+    ASSERT(sizeof(fp) <= sizeof(unsigned long));
+    reg[1] = (Eterm) enif_make_ulong(env, (unsigned long) fp);
+#endif
+    proc->i = (BeamInstr*) ep->addressv[0];
+    ep->code[4] = (BeamInstr) execute_dirty_nif_finalizer;
+    proc->freason = TRAP;
+
+    return THE_NON_VALUE;
+#else
+    return (*fp)(env, result);
+#endif
+}
+
+/* A simple finalizer that just returns its result argument */
+ERL_NIF_TERM
+enif_dirty_nif_finalizer(ErlNifEnv* env, ERL_NIF_TERM result)
+{
+    return result;
+}
+
+int
+enif_is_on_dirty_scheduler(ErlNifEnv* env)
+{
+    return ERTS_SCHEDULER_IS_DIRTY(env->proc->scheduler_data);
+}
+
+int
+enif_have_dirty_schedulers()
+{
+#ifdef USE_THREADS
+    return 1;
+#else
+    return 0;
+#endif
+}
+
+#endif /* ERL_NIF_DIRTY_SCHEDULER_SUPPORT */
+
 /***************************************************************************
  **                              load_nif/2                               **
  ***************************************************************************/