use repne scas[dq] for searching for l_entry

Makefile

@@ -37,9 +37,9 @@ ASFLAGS += -DUSE_INTERP -DALIGN_STACK
 NASM    ?= nasm
 PYTHON3 ?= python3
 
-all: $(BINDIR)/hello-crt $(BINDIR)/sdl-crt
+all: $(BINDIR)/hello-crt $(BINDIR)/sdl-crt $(BINDIR)/flag-crt
 
-LIBS += $(shell pkg-config --libs sdl2) -lX11 #-lGL
+LIBS += $(filter-out -pthread,$(shell pkg-config --libs sdl2)) -lX11 #-lGL
 
 clean:
 	@$(RM) -vrf $(OBJDIR) $(BINDIR)

src/loader32.asm

@@ -29,16 +29,23 @@ _smol_start:
 %endif
 
         mov ebx, eax
-        mov esi, eax
-.looper:
-      lodsd
-        cmp dword eax, _smol_start
-        jne short .looper
-        sub esi, ebx
-        sub esi, LM_ENTRY_OFFSET_BASE+4 ; +4: take inc-after from lodsd into acct
+;       mov esi, eax
+;.looper:
+;     lodsd
+;       cmp dword eax, _smol_start
+;       jne short .looper
+;       sub esi, ebx
+;       sub esi, LM_ENTRY_OFFSET_BASE+4 ; +4: take inc-after from lodsd into acct
+        mov edi, eax
+       push -1
+        pop ecx
+        mov eax, _smol_start
+repne scasd
+        sub edi, ebx
+        sub edi, LM_ENTRY_OFFSET_BASE+4
 
        xchg ebp, ebx
-       xchg ebx, esi
+       xchg ebx, edi;esi
         mov esi, _symbols
 
 link: ; (struct link_map *root, char *symtable)

src/loader64.asm

@@ -31,30 +31,39 @@ _smol_start:
     mov r12, [rsp -  8]        ; return address of _dl_init
     mov r11d, dword [r12 - 20] ; decode part of 'mov rdi, [rel _rtld_global]'
     mov r12, [r12 + r11 - 16]  ; ???
-        ; struct link_map* root = r12
-   ;mov r12, rdi
 %endif
+        ; struct link_map* root = r12
 %ifdef SKIP_ENTRIES
     mov r12, [r12 + L_NEXT_OFF] ; skip this binary
     mov r12, [r12 + L_NEXT_OFF] ; skip the vdso
 %endif
 
-    mov rsi, r12
-
+;   mov rsi, r12
             ; size_t* field = (size_t*)root;
             ; for (; *field != _smol_start; ++field) ;
-    .next_off:
-      lodsq
-        cmp rax, _smol_start
-        jne short .next_off
+;   .next_off:
+;     lodsq
+;       cmp rax, _smol_start
+;       jne short .next_off
 
         ; // rbx = offsetof(struct link_map* rsi, l_entry) - DEFAULT_OFFSET
         ; rbx = field - root - offsetof(struct link_map, l_entry)
-    sub rsi, r12
-    sub rsi, LF_ENTRY_OFF+8
-   xchg rbx, rsi
-
-    mov esi, _symbols
+;   sub rsi, r12
+;   sub rsi, LF_ENTRY_OFF+8
+;  xchg rbx, rsi
+
+        mov rdi, r12
+       push -1
+        pop rcx
+       ;mov rax, _smol_start
+        lea rax, [rel _smol_start]
+repne scasq
+        sub rdi, r12
+        sub rdi, LF_ENTRY_OFF+8
+       xchg rbx, rdi
+
+   ;mov esi, _symbols
+    lea esi, [rel _symbols]
 
             ; for (rsi = (uint8_t*)_symbols; *rsi; ++rsi) {
      .next_needed:

src/flag.c → test/flag.c

+/* "Tiny X11 Trans Flag" by Suricrasia Online, edited a bit to make testing
+ * easier */
 #include<X11/X.h>
 #include<X11/Xlib.h>
 #include<X11/Xutil.h>
@@ -12,8 +14,9 @@
 static Atom wmDeleteMessage;
 #endif
 
-__attribute__((__used__))
-void _start() {
+/*__attribute__((__used__))
+void _start() {*/
+int main() {
   //initialize the window
   Display* dpy = XOpenDisplay(NULL);