bcachefs: Fix for partial buffered writes

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

diff --git a/fs/bcachefs/fs-io.c b/fs/bcachefs/fs-io.c
index 791e72ce1c29081aeefc5ebc262e45e624108bb0..18356cbe07949221ddda5d3fb119955e5b8da5e1 100644 (file)
--- a/fs/bcachefs/fs-io.c
+++ b/fs/bcachefs/fs-io.c
@@ -756,8 +756,8 @@ static void bch2_set_page_dirty(struct bch_fs *c,
        struct bch_page_state *s = bch2_page_state(page);
        unsigned i, dirty_sectors = 0;
 
-       WARN_ON(page_offset(page) + offset + len >
-               round_up(i_size_read(&inode->v), block_bytes(c)));
+       WARN_ON((u64) page_offset(page) + offset + len >
+               round_up((u64) i_size_read(&inode->v), block_bytes(c)));
 
        for (i = round_down(offset, block_bytes(c)) >> 9;
             i < round_up(offset + len, block_bytes(c)) >> 9;
@@ -1704,14 +1704,6 @@ retry_reservation:
        if (!copied)
                goto out;
 
-       nr_pages_copied = DIV_ROUND_UP(offset + copied, PAGE_SIZE);
-       inode->ei_last_dirtied = (unsigned long) current;
-
-       spin_lock(&inode->v.i_lock);
-       if (pos + copied > inode->v.i_size)
-               i_size_write(&inode->v, pos + copied);
-       spin_unlock(&inode->v.i_lock);
-
        if (copied < len &&
            ((offset + copied) & (PAGE_SIZE - 1))) {
                struct page *page = pages[(offset + copied) >> PAGE_SHIFT];
@@ -1722,6 +1714,11 @@ retry_reservation:
                }
        }
 
+       spin_lock(&inode->v.i_lock);
+       if (pos + copied > inode->v.i_size)
+               i_size_write(&inode->v, pos + copied);
+       spin_unlock(&inode->v.i_lock);
+
        while (set_dirty < copied) {
                struct page *page = pages[(offset + set_dirty) >> PAGE_SHIFT];
                unsigned pg_offset = (offset + set_dirty) & (PAGE_SIZE - 1);
@@ -1737,6 +1734,9 @@ retry_reservation:
 
                set_dirty += pg_len;
        }
+
+       nr_pages_copied = DIV_ROUND_UP(offset + copied, PAGE_SIZE);
+       inode->ei_last_dirtied = (unsigned long) current;
 out:
        for (i = nr_pages_copied; i < nr_pages; i++) {
                unlock_page(pages[i]);