bugfixes, change types.h to be linux abi, stat system call

src/sys/syscalls/close.c

@@ -5,7 +5,6 @@
 #include <arch/spinlock.h>
 
 syscallret syscall_close(int ifd){
-
         syscallret retv;
         retv.ret = -1;
 

src/sys/syscalls/open.c

@@ -69,7 +69,7 @@ syscallret syscall_open(const char* pathname, int flags, mode_t mode){
 	fd->flags = flags + 1; // 1 is added to make O_RDONLY etc easier to lookup
 	fd->offset = 0;
 	
-	vnode_t* file;
+	vnode_t* file = NULL;
 
 	size_t ret = vfs_open(&file, 
 		*name == '/' ? proc->root : proc->cwd,

src/sys/syscalls/read.c

@@ -57,6 +57,8 @@ syscallret syscall_read(int ifd, void* buff, size_t count){
 
 	retv.errno = 0;
 	retv.ret = readc;
+
+	free(kbuff);
 	
 	return retv;
 }

src/sys/syscalls/stat.c

@@ -0,0 +1,51 @@
+#include <kernel/syscalls.h>
+#include <sys/stat.h>
+#include <kernel/vmm.h>
+#include <errno.h>
+#include <kernel/vfs.h>
+#include <kernel/sched.h>
+#include <arch/cls.h>
+#include <kernel/alloc.h>
+#include <string.h>
+
+syscallret syscall_stat(const char* path, stat* st){
+	
+	syscallret retv;
+	retv.ret = -1;
+
+	if(st > USER_SPACE_END){
+		retv.errno = EFAULT;
+		return retv;
+	}
+
+	char* kpath = alloc(strlen(path)); // XXX use proper user strlen func
+	
+	if(!path){
+		retv.errno = ENOMEM;
+		return retv;
+	}
+
+	strcpy(kpath, path); // XXX use user strcpy func
+
+	vnode_t* file;
+	
+	proc_t* proc = arch_getcls()->thread->proc;
+
+	int ret = vfs_open(&file, *kpath == '/' ? proc->root : proc->cwd, kpath);
+
+	if(ret){
+		retv.errno = ret;
+		goto _ret;
+	}
+
+	stat stbuff = file->st;
+
+	vfs_close(file);
+
+	memcpy(st, &stbuff, sizeof(stat)); // XXX use user memcpy
+
+	_ret:
+	free(kpath);
+	return retv;
+
+}