[srv.init] Improve loader for non-aligned segments

The drv.uart ELF currently ends up with a segment vaddr starting at
0x215010, which includes .data and .bss. The old loader was mishandling
this in a few ways:

- Not zeroing out the leading 16 bytes, or the trailing .bss section
- Copying the segment data to the start of the page, so it was offset by
  -16 bytes.
- Mapping the VMA into the child program at the non-page-aligned
  address, which causes all sorts of trouble.

src/user/srv.init/loader.cpp

@@ -78,11 +78,17 @@ load_program(const module_program &prog, j6_handle_t sys, char *err_msg)
             return false;
         }
 
-        void *src = reinterpret_cast<void *>(prog.base_address + seg.offset);
+        uintptr_t start = prog.base_address + seg.offset;
-        void *dest = reinterpret_cast<void *>(load_addr);
+        size_t prelude = start & 0xfff;
-        memcpy(dest, src, seg.file_size);
+        size_t prologue = seg.mem_size - (prelude+seg.file_size);
 
-        res = j6_vma_map(sub_vma, proc, seg.vaddr);
+        uint8_t *src = reinterpret_cast<uint8_t *>(start);
+        uint8_t *dest = reinterpret_cast<uint8_t *>(load_addr);
+        memset(dest, 0, prelude);
+        memcpy(dest+prelude, src, seg.file_size);
+        memset(dest+prelude+seg.file_size, 0, prologue);
+
+        res = j6_vma_map(sub_vma, proc, seg.vaddr & ~0xfffull);
         if (res != j6_status_ok) {
             sprintf(err_msg, "  ** error loading program '%s': mapping sub vma to child: %lx", prog.filename, res);
             return false;