Merge branch 'sverk/enc_atom-opt'

* sverk/enc_atom-opt:
  erts: Optimize atom encoding to use memcpy for pure ascii
  erts: Refactor erts_atom_get to use ErtsAtomEncoding

6 files changed, 40 insertions, 38 deletions

diff --git a/erts/emulator/beam/atom.c b/erts/emulator/beam/atom.c
index 259b5e49c1..b69f979397 100644
--- a/erts/emulator/beam/atom.c
+++ b/erts/emulator/beam/atom.c
@@ -367,7 +367,7 @@ int atom_table_sz(void)
 }
 
 int
-erts_atom_get(const char *name, int len, Eterm* ap, int is_latin1)
+erts_atom_get(const char *name, int len, Eterm* ap, ErtsAtomEncoding enc)
 {
     byte utf8_copy[MAX_ATOM_SZ_FROM_LATIN1];
     Atom a;
@@ -376,7 +376,7 @@ erts_atom_get(const char *name, int len, Eterm* ap, int is_latin1)
 
     a.len = (Sint16) len;
     a.name = (byte *)name;
-    if (is_latin1) {
+    if (enc == ERTS_ATOM_ENC_LATIN1) {
 	latin1_to_utf8(utf8_copy, (const byte**)&a.name, &len);
 	a.len = (Sint16) len;
     }
diff --git a/erts/emulator/beam/atom.h b/erts/emulator/beam/atom.h
index 1d3c4cb922..5904ae0f7e 100644
--- a/erts/emulator/beam/atom.h
+++ b/erts/emulator/beam/atom.h
@@ -139,7 +139,7 @@ int atom_static_put(byte*, int);
 void init_atom_table(void);
 void atom_info(int, void *);
 void dump_atoms(int, void *);
-int erts_atom_get(const char* name, int len, Eterm* ap, int is_latin1);
+int erts_atom_get(const char* name, int len, Eterm* ap, ErtsAtomEncoding enc);
 void erts_atom_get_text_space_sizes(Uint *reserved, Uint *used);
 #endif
 
diff --git a/erts/emulator/beam/bif.c b/erts/emulator/beam/bif.c
index 16df37a0fb..46f76624a5 100644
--- a/erts/emulator/beam/bif.c
+++ b/erts/emulator/beam/bif.c
@@ -2670,7 +2670,7 @@ BIF_RETTYPE list_to_existing_atom_1(BIF_ALIST_1)
     } else {
 	Eterm a;
 	
-	if (erts_atom_get(buf, i, &a, 1)) {
+	if (erts_atom_get(buf, i, &a, ERTS_ATOM_ENC_LATIN1)) {
 	    erts_free(ERTS_ALC_T_TMP, (void *) buf);
 	    BIF_RET(a);
 	} else {
diff --git a/erts/emulator/beam/erl_nif.c b/erts/emulator/beam/erl_nif.c
index fb612d8f19..068f904b76 100644
--- a/erts/emulator/beam/erl_nif.c
+++ b/erts/emulator/beam/erl_nif.c
@@ -984,7 +984,7 @@ int enif_make_existing_atom_len(ErlNifEnv* env, const char* name, size_t len,
 				ERL_NIF_TERM* atom, ErlNifCharEncoding encoding)
 {
     ASSERT(encoding == ERL_NIF_LATIN1);
-    return erts_atom_get(name, len, atom, 1);
+    return erts_atom_get(name, len, atom, ERTS_ATOM_ENC_LATIN1);
 }
 
 ERL_NIF_TERM enif_make_tuple(ErlNifEnv* env, unsigned cnt, ...)
@@ -1653,7 +1653,7 @@ BIF_RETTYPE load_nif_2(BIF_ALIST_2)
 	for (i=0; i < entry->num_of_funcs && ret==am_ok; i++) {
 	    BeamInstr** code_pp;
 	    ErlNifFunc* f = &entry->funcs[i];
-	    if (!erts_atom_get(f->name, sys_strlen(f->name), &f_atom, 1)
+	    if (!erts_atom_get(f->name, sys_strlen(f->name), &f_atom, ERTS_ATOM_ENC_LATIN1)
 		|| (code_pp = get_func_pp(mod->curr.code, f_atom, f->arity))==NULL) {
 		ret = load_nif_error(BIF_P,bad_lib,"Function not found %T:%s/%u",
 				     mod_atom, f->name, f->arity);
@@ -1756,7 +1756,7 @@ BIF_RETTYPE load_nif_2(BIF_ALIST_2)
 	for (i=0; i < entry->num_of_funcs; i++)
 	{
 	    BeamInstr* code_ptr;
-	    erts_atom_get(entry->funcs[i].name, sys_strlen(entry->funcs[i].name), &f_atom, 1); 
+	    erts_atom_get(entry->funcs[i].name, sys_strlen(entry->funcs[i].name), &f_atom, ERTS_ATOM_ENC_LATIN1); 
 	    code_ptr = *get_func_pp(mod->curr.code, f_atom, entry->funcs[i].arity);
 	    
 	    if (code_ptr[1] == 0) {
diff --git a/erts/emulator/beam/erl_unicode.c b/erts/emulator/beam/erl_unicode.c
index 58202976cf..99108af937 100644
--- a/erts/emulator/beam/erl_unicode.c
+++ b/erts/emulator/beam/erl_unicode.c
@@ -1905,7 +1905,7 @@ binary_to_atom(Process* proc, Eterm bin, Eterm enc, int must_exist)
 	    if (is_non_value(a))
 		goto badarg;
 	    BIF_RET(a);
-	} else if (erts_atom_get((char *)bytes, bin_size, &a, 1)) {
+	} else if (erts_atom_get((char *)bytes, bin_size, &a, ERTS_ATOM_ENC_LATIN1)) {
 	    erts_free_aligned_binary_bytes(temp_alloc);
 	    BIF_RET(a);
 	} else {
@@ -1940,7 +1940,7 @@ binary_to_atom(Process* proc, Eterm bin, Eterm enc, int must_exist)
 				ERTS_ATOM_ENC_UTF8,
 				0);
 	}
-	else if (!erts_atom_get((char*)bytes, bin_size, &res, 0)) {
+	else if (!erts_atom_get((char*)bytes, bin_size, &res, ERTS_ATOM_ENC_UTF8)) {
 	    goto badarg;
 	}
 	erts_free_aligned_binary_bytes(temp_alloc);
diff --git a/erts/emulator/beam/external.c b/erts/emulator/beam/external.c
index 546e5d6ce0..5ce0d97c74 100644
--- a/erts/emulator/beam/external.c
+++ b/erts/emulator/beam/external.c
@@ -1455,8 +1455,8 @@ enc_atom(ErtsAtomCacheMap *acmp, Eterm atom, byte *ep, Uint32 dflags)
     iix = get_iix_acache_map(acmp, atom, dflags);
     if (iix < 0) {
 	Atom *a = atom_tab(atom_val(atom));
+	len = a->len;
 	if (utf8_atoms || a->latin1_chars < 0) {
-	    len = a->len;
 	    if (len > 255) {
 		*ep++ = ATOM_UTF8_EXT;
 		put_int16(len, ep);
@@ -1472,15 +1472,25 @@ enc_atom(ErtsAtomCacheMap *acmp, Eterm atom, byte *ep, Uint32 dflags)
 	else {
 	    if (a->latin1_chars <= 255 && (dflags & DFLAG_SMALL_ATOM_TAGS)) {
 		*ep++ = SMALL_ATOM_EXT;
-		len = erts_utf8_to_latin1(ep+1, a->name, a->len);
-		ASSERT(len == a->latin1_chars);
+		if (len == a->latin1_chars) {
+		    sys_memcpy(ep+1, a->name, len);
+		}
+		else {
+		    len = erts_utf8_to_latin1(ep+1, a->name, len);
+		    ASSERT(len == a->latin1_chars);
+		}
 		put_int8(len, ep);
 		ep++;
 	    }
 	    else {
 		*ep++ = ATOM_EXT;
-		len = erts_utf8_to_latin1(ep+2, a->name, a->len);
-		ASSERT(len == a->latin1_chars);
+		if (len == a->latin1_chars) {
+		    sys_memcpy(ep+2, a->name, len);
+		}
+		else {
+		    len = erts_utf8_to_latin1(ep+2, a->name, len);
+		    ASSERT(len == a->latin1_chars);
+		}
 		put_int16(len, ep);
 		ep += 2;
 	    }	    
@@ -1524,7 +1534,8 @@ static byte*
 dec_atom(ErtsDistExternal *edep, byte* ep, Eterm* objp)
 {
     Uint len;
-    int n, is_latin1;
+    int n;
+    ErtsAtomEncoding char_enc;
 
     switch (*ep++) {
     case ATOM_CACHE_REF:
@@ -1540,34 +1551,29 @@ dec_atom(ErtsDistExternal *edep, byte* ep, Eterm* objp)
     case ATOM_EXT:
 	len = get_int16(ep),
 	ep += 2;
-	is_latin1 = 1;
+	char_enc = ERTS_ATOM_ENC_LATIN1;
         goto dec_atom_common;
     case SMALL_ATOM_EXT:
 	len = get_int8(ep);
 	ep++;
-	is_latin1 = 1;
+	char_enc = ERTS_ATOM_ENC_LATIN1;
 	goto dec_atom_common;
     case ATOM_UTF8_EXT:
 	len = get_int16(ep),
 	ep += 2;
-	is_latin1 = 0;
+	char_enc = ERTS_ATOM_ENC_UTF8;
 	goto dec_atom_common;
     case SMALL_ATOM_UTF8_EXT:
 	len = get_int8(ep),
 	ep++;
-	is_latin1 = 0;
+	char_enc = ERTS_ATOM_ENC_UTF8;
     dec_atom_common:
         if (edep && (edep->flags & ERTS_DIST_EXT_BTT_SAFE)) {
-	    if (!erts_atom_get((char*)ep, len, objp, is_latin1)) {
+	    if (!erts_atom_get((char*)ep, len, objp, char_enc)) {
                 goto error;
 	    }
         } else {
-	    Eterm atom = erts_atom_put(ep,
-				       len,
-				       (is_latin1
-					? ERTS_ATOM_ENC_LATIN1
-					: ERTS_ATOM_ENC_UTF8),
-				       0);
+	    Eterm atom = erts_atom_put(ep, len, char_enc, 0);
 	    if (is_non_value(atom))
 		goto error;
             *objp = atom;
@@ -2175,7 +2181,8 @@ static byte*
 dec_term(ErtsDistExternal *edep, Eterm** hpp, byte* ep, ErlOffHeap* off_heap, Eterm* objp)
 {
     Eterm* hp_saved = *hpp;
-    int n, is_latin1;
+    int n;
+    ErtsAtomEncoding char_enc;
     register Eterm* hp = *hpp;	/* Please don't take the address of hp */
     Eterm* next = objp;
 
@@ -2261,34 +2268,29 @@ dec_term(ErtsDistExternal *edep, Eterm** hpp, byte* ep, ErlOffHeap* off_heap, Et
 	case ATOM_EXT:
 	    n = get_int16(ep);
 	    ep += 2;
-	    is_latin1 = 1;
+	    char_enc = ERTS_ATOM_ENC_LATIN1;
 	    goto dec_term_atom_common;
 	case SMALL_ATOM_EXT:
 	    n = get_int8(ep);
 	    ep++;
-	    is_latin1 = 1;
+	    char_enc = ERTS_ATOM_ENC_LATIN1;
 	    goto dec_term_atom_common;
 	case ATOM_UTF8_EXT:
 	    n = get_int16(ep);
 	    ep += 2;
-	    is_latin1 = 0;
+	    char_enc = ERTS_ATOM_ENC_UTF8;
 	    goto dec_term_atom_common;
 	case SMALL_ATOM_UTF8_EXT:
 	    n = get_int8(ep);
 	    ep++;
-	    is_latin1 = 0;
+	    char_enc = ERTS_ATOM_ENC_UTF8;
 dec_term_atom_common:
 	    if (edep && (edep->flags & ERTS_DIST_EXT_BTT_SAFE)) {
-		if (!erts_atom_get((char*)ep, n, objp, is_latin1)) {
+		if (!erts_atom_get((char*)ep, n, objp, char_enc)) {
 		    goto error;
 		}
 	    } else {
-		Eterm atom = erts_atom_put(ep,
-					   n,
-					   (is_latin1
-					    ? ERTS_ATOM_ENC_LATIN1
-					    : ERTS_ATOM_ENC_UTF8),
-					   0);
+		Eterm atom = erts_atom_put(ep, n, char_enc, 0);
 		if (is_non_value(atom))
 		    goto error;
 	        *objp = atom;