Implement clone() C library function properly.

Only provide an implementation for ARM at the moment, since
it requires specific assembly fragments (the standard syscall
stubs cannot be used because the child returns in a different
stack).

libc/arch-x86/bionic/clone.S

@@ -48,3 +48,8 @@ __pthread_clone:
         popl    %ecx
         popl    %ebx
         ret
+
+/* XXX: TODO: Add __bionic_clone here
+ *            See bionic/bionic_clone.c and arch-arm/bionic/clone.S
+ *            for more details...
+ */

libc/arch-x86/syscalls.mk

@@ -5,7 +5,7 @@ syscall_src += arch-x86/syscalls/_exit_thread.S
 syscall_src += arch-x86/syscalls/__fork.S
 syscall_src += arch-x86/syscalls/_waitpid.S
 syscall_src += arch-x86/syscalls/waitid.S
-syscall_src += arch-x86/syscalls/__clone.S
+syscall_src += arch-x86/syscalls/__sys_clone.S
 syscall_src += arch-x86/syscalls/execve.S
 syscall_src += arch-x86/syscalls/setuid.S
 syscall_src += arch-x86/syscalls/getuid.S

libc/arch-x86/syscalls/__sys_clone.S

@@ -2,19 +2,21 @@
 #include <sys/linux-syscalls.h>
 
     .text
-    .type __clone, @function
-    .globl __clone
+    .type __sys_clone, @function
+    .globl __sys_clone
     .align 4
 
-__clone:
+__sys_clone:
     pushl   %ebx
     pushl   %ecx
     pushl   %edx
     pushl   %esi
-    mov     20(%esp), %ebx
-    mov     24(%esp), %ecx
-    mov     28(%esp), %edx
-    mov     32(%esp), %esi
+    pushl   %edi
+    mov     24(%esp), %ebx
+    mov     28(%esp), %ecx
+    mov     32(%esp), %edx
+    mov     36(%esp), %esi
+    mov     40(%esp), %edi
     movl    $__NR_clone, %eax
     int     $0x80
     cmpl    $-129, %eax
@@ -25,6 +27,7 @@ __clone:
     addl    $4, %esp
     orl     $-1, %eax
 1:
+    popl    %edi
     popl    %esi
     popl    %edx
     popl    %ecx