diff --git a/aarch64.c b/aarch64.c
index e153cd46..59162f20 100644
--- a/aarch64.c
+++ b/aarch64.c
@@ -59,6 +59,10 @@ static void co_init() {
   #endif
 }
 
+const char* co_method() {
+  return "aarch64";
+}
+
 cothread_t co_active() {
   if(!co_active_handle) co_active_handle = &co_active_buffer;
   return co_active_handle;
@@ -73,7 +77,8 @@ cothread_t co_derive(void* memory, unsigned int size, void (*entrypoint)(void))
   if(!co_active_handle) co_active_handle = &co_active_buffer;
 
   if(handle = (unsigned long*)memory) {
-    unsigned long* p = (unsigned long*)((unsigned char*)handle + size);
+    unsigned int offset = (size & ~15);
+    unsigned long* p = (unsigned long*)((unsigned char*)handle + offset);
     handle[19] = (unsigned long)p;           /* x29 (frame pointer) */
     handle[20] = (unsigned long)p;           /* x30 (stack pointer) */
     handle[21] = (unsigned long)entrypoint;  /* x31 (link register) */
@@ -83,23 +88,9 @@ cothread_t co_derive(void* memory, unsigned int size, void (*entrypoint)(void))
 }
 
 cothread_t co_create(unsigned int size, void (*entrypoint)(void)) {
-  unsigned long* handle;
-  if(!co_swap) {
-    co_init();
-    co_swap = (void (*)(cothread_t, cothread_t))co_swap_function;
-  }
-  if(!co_active_handle) co_active_handle = &co_active_buffer;
-  size += 256;
-  size &= ~15;
-
-  if(handle = (unsigned long*)malloc(size)) {
-    unsigned long* p = (unsigned long*)((unsigned char*)handle + size);
-    handle[19] = (unsigned long)p;           /* x29 (frame pointer) */
-    handle[20] = (unsigned long)p;           /* x30 (stack pointer) */
-    handle[21] = (unsigned long)entrypoint;  /* x31 (link register) */
-  }
-
-  return handle;
+  void* memory = malloc(size);
+  if(!memory) return (cothread_t)0;
+  return co_derive(memory, size, entrypoint);
 }
 
 void co_delete(cothread_t handle) {
diff --git a/amd64.c b/amd64.c
index 664730a1..e9aef889 100644
--- a/amd64.c
+++ b/amd64.c
@@ -115,6 +115,10 @@ static void crash() {
   assert(0);  /* called only if cothread_t entrypoint returns */
 }
 
+const char* co_method() {
+  return "amd64";
+}
+
 cothread_t co_active() {
   if(!co_active_handle) co_active_handle = &co_active_buffer;
   return co_active_handle;
@@ -140,22 +144,9 @@ cothread_t co_derive(void* memory, unsigned int size, void (*entrypoint)(void))
 }
 
 cothread_t co_create(unsigned int size, void (*entrypoint)(void)) {
-  cothread_t handle;
-  if(!co_swap) {
-    co_init();
-    co_swap = (void (*)(cothread_t, cothread_t))co_swap_function;
-  }
-  if(!co_active_handle) co_active_handle = &co_active_buffer;
-
-  if(handle = (cothread_t)malloc(size)) {
-    unsigned int offset = (size & ~15) - 32;
-    long long *p = (long long*)((char*)handle + offset);  /* seek to top of stack */
-    *--p = (long long)crash;                              /* crash if entrypoint returns */
-    *--p = (long long)entrypoint;                         /* start of function */
-    *(long long*)handle = (long long)p;                   /* stack pointer */
-  }
-
-  return handle;
+  void* memory = malloc(size);
+  if(!memory) return (cothread_t)0;
+  return co_derive(memory, size, entrypoint);
 }
 
 void co_delete(cothread_t handle) {
diff --git a/arm.c b/arm.c
index 20c95310..8d872bf5 100644
--- a/arm.c
+++ b/arm.c
@@ -35,6 +35,10 @@ static void co_init() {
   #endif
 }
 
+const char* co_method() {
+  return "arm";
+}
+
 cothread_t co_active() {
   if(!co_active_handle) co_active_handle = &co_active_buffer;
   return co_active_handle;
@@ -49,7 +53,8 @@ cothread_t co_derive(void* memory, unsigned int size, void (*entrypoint)(void))
   if(!co_active_handle) co_active_handle = &co_active_buffer;
 
   if(handle = (unsigned long*)memory) {
-    unsigned long* p = (unsigned long*)((unsigned char*)handle + size);
+    unsigned int offset = (size & ~15);
+    unsigned long* p = (unsigned long*)((unsigned char*)handle + offset);
     handle[8] = (unsigned long)p;
     handle[9] = (unsigned long)entrypoint;
   }
@@ -58,22 +63,9 @@ cothread_t co_derive(void* memory, unsigned int size, void (*entrypoint)(void))
 }
 
 cothread_t co_create(unsigned int size, void (*entrypoint)(void)) {
-  unsigned long* handle;
-  if(!co_swap) {
-    co_init();
-    co_swap = (void (*)(cothread_t, cothread_t))co_swap_function;
-  }
-  if(!co_active_handle) co_active_handle = &co_active_buffer;
-  size += 256;
-  size &= ~15;
-
-  if(handle = (unsigned long*)malloc(size)) {
-    unsigned long* p = (unsigned long*)((unsigned char*)handle + size);
-    handle[8] = (unsigned long)p;
-    handle[9] = (unsigned long)entrypoint;
-  }
-
-  return handle;
+  void* memory = malloc(size);
+  if(!memory) return (cothread_t)0;
+  return co_derive(memory, size, entrypoint);
 }
 
 void co_delete(cothread_t handle) {
diff --git a/fiber.c b/fiber.c
index bdf4dd4e..f508b0f8 100644
--- a/fiber.c
+++ b/fiber.c
@@ -16,6 +16,10 @@ static void __stdcall co_thunk(void* coentry) {
   ((void (*)(void))coentry)();
 }
 
+const char* co_method() {
+  return "fiber";
+}
+
 cothread_t co_active() {
   if(!co_active_) {
     ConvertThreadToFiber(0);
diff --git a/libco.c b/libco.c
index f9c79bf4..21fe4cab 100644
--- a/libco.c
+++ b/libco.c
@@ -1,7 +1,7 @@
 #if defined(__clang__)
   #pragma clang diagnostic ignored "-Wparentheses"
 
-  //placing code in section(text) does not mark it executable with Clang.
+  /* placing code in section(text) does not mark it executable with Clang. */
   #undef  LIBCO_MPROTECT
   #define LIBCO_MPROTECT
 #endif
diff --git a/libco.h b/libco.h
index a436957c..0b94c2e9 100644
--- a/libco.h
+++ b/libco.h
@@ -1,5 +1,5 @@
 /*
-  libco v20 (2019-10-14)
+  libco v20 (2019-10-16)
   author: byuu
   license: ISC
 */
@@ -13,6 +13,7 @@ extern "C" {
 
 typedef void* cothread_t;
 
+const char* co_method();
 cothread_t co_active();
 cothread_t co_derive(void*, unsigned int, void (*)(void));
 cothread_t co_create(unsigned int, void (*)(void));
diff --git a/ppc.c b/ppc.c
index 969fcec2..314997c8 100644
--- a/ppc.c
+++ b/ppc.c
@@ -413,6 +413,10 @@ static void co_init_(void) {
   co_active_handle = co_create_(state_size, (uintptr_t)&co_switch);
 }
 
+const char* co_method() {
+  return "ppc";
+}
+
 cothread_t co_active() {
   if(!co_active_handle) co_init_();
 
diff --git a/ppc64v2.c b/ppc64v2.c
index b1f153d4..fac464a6 100644
--- a/ppc64v2.c
+++ b/ppc64v2.c
@@ -221,6 +221,10 @@ __asm__(
   ".size swap_context, .-swap_context\n"
 );
 
+const char* co_method() {
+  return "ppc64v2";
+}
+
 cothread_t co_active() {
   if(!co_active_handle) {
     co_active_handle = (struct ppc64_context*)malloc(MIN_STACK + sizeof(struct ppc64_context));
@@ -255,11 +259,9 @@ cothread_t co_derive(void* memory, unsigned int size, void (*coentry)(void)) {
 }
 
 cothread_t co_create(unsigned int size, void (*coentry)(void)) {
-  size_t total = MAX(size, MIN_STACK) + sizeof(struct ppc64_context);
-  void* memory = malloc(total);
-
+  void* memory = malloc(size);
   if(!memory) return (cothread_t)0;
-  return co_derive(memory, total, coentry);
+  return co_derive(memory, size, coentry);
 }
 
 void co_delete(cothread_t handle) {
diff --git a/sjlj.c b/sjlj.c
index d99572a3..b4faf17b 100644
--- a/sjlj.c
+++ b/sjlj.c
@@ -33,6 +33,10 @@ static void springboard(int ignored) {
   }
 }
 
+const char* co_method() {
+  return "sjlj";
+}
+
 cothread_t co_active() {
   if(!co_running) co_running = &co_primary;
   return (cothread_t)co_running;
diff --git a/ucontext.c b/ucontext.c
index edf513d4..49fa976c 100644
--- a/ucontext.c
+++ b/ucontext.c
@@ -26,6 +26,10 @@ extern "C" {
 static thread_local ucontext_t co_primary;
 static thread_local ucontext_t* co_running = 0;
 
+const char* co_module() {
+  return "ucontext";
+}
+
 cothread_t co_active() {
   if(!co_running) co_running = &co_primary;
   return (cothread_t)co_running;
diff --git a/x86.c b/x86.c
index 3eb2e21e..8effa0d4 100644
--- a/x86.c
+++ b/x86.c
@@ -69,6 +69,10 @@ static void crash() {
   assert(0);  /* called only if cothread_t entrypoint returns */
 }
 
+const char* co_method() {
+  return "x86";
+}
+
 cothread_t co_active() {
   if(!co_active_handle) co_active_handle = &co_active_buffer;
   return co_active_handle;
@@ -94,22 +98,9 @@ cothread_t co_derive(void* memory, unsigned int size, void (*entrypoint)(void))
 }
 
 cothread_t co_create(unsigned int size, void (*entrypoint)(void)) {
-  cothread_t handle;
-  if(!co_swap) {
-    co_init();
-    co_swap = (void (fastcall*)(cothread_t, cothread_t))co_swap_function;
-  }
-  if(!co_active_handle) co_active_handle = &co_active_buffer;
-
-  if(handle = (cothread_t)malloc(size)) {
-    unsigned int offset = (size & ~15) - 32;
-    long *p = (long*)((char*)handle + offset);  /* seek to top of stack */
-    *--p = (long)crash;                         /* crash if entrypoint returns */
-    *--p = (long)entrypoint;                    /* start of function */
-    *(long*)handle = (long)p;                   /* stack pointer */
-  }
-
-  return handle;
+  void* memory = malloc(size);
+  if(!memory) return (cothread_t)0;
+  return co_derive(memory, size, entrypoint);
 }
 
 void co_delete(cothread_t handle) {