Fixed file buffers falling out-of-date when bypassed during writes

This can happend if we read some data into our buffer (or write+flush+seek
in some weird pattern) and then do a bypassing write that overlaps the
buffer.

The solution here is to make sure the buffer is cleared when doing
bypassing writes.

In theory, we could be a bit smarter by checking and only clearing when
overlap occurs, or even updating the buffer with the new contents like
we do in the bd cache layer, but the value would probably be minimal.
File writes are already expected to clobber buffered reads.

lfs.c

@@ -10478,7 +10478,13 @@ lfs_ssize_t lfsr_file_write(lfs_t *lfs, lfsr_file_t *file,
         // strictly necessary, but enforces a more intuitive write order
         // and avoids weird cases with low-level write heuristics
         //
+        // oh, and also avoids issues with out-of-date buffers
+        //
         if (!unflushed_ && size >= lfs->cfg->cache_size) {
+            // clear buffer to avoid out-of-date data
+            buffer_pos_ = 0;
+            buffer_size_ = 0;
+
             err = lfsr_ftree_flush(lfs, &file->mdir, &ftree_,
                     pos_, buffer_, size);
             if (err) {