Free space of overwritten files in LittleFS

Fixes esp8266#7426

LittleFS doesn't update the on-flash data structures when a file is
reopened as O_TRUNC until the file is closed.  This means the space of
the original, inaccessible file cannot be used, causing OOS errors in
cases when a large file is being overwritten.

Explicitly call the file sync operation to update the on-flash metadata
as soon as a file is opened.  For normal files it's a no-op, but for
O_TRUNC modes it will free the space, allowing full overwrite of large
files.

Author: earlephilhower

libraries/LittleFS/src/LittleFS.cpp

@@ -87,6 +87,7 @@ FileImplPtr LittleFSImpl::open(const char* path, OpenMode openMode, AccessMode a
         // a directory whose name we are carrying around but which cannot be read or written
         return std::make_shared<LittleFSFileImpl>(this, path, nullptr, flags, creation);
     } else if (rc == 0) {
+        lfs_file_sync(&_lfs, fd.get());
         return std::make_shared<LittleFSFileImpl>(this, path, fd, flags, creation);
     } else {
         DEBUGV("LittleFSDirImpl::openFile: rc=%d fd=%p path=`%s` openMode=%d accessMode=%d err=%d\n",