On Mon, 17 Jun 2024 22:49:47 +0300, Alex Shumsky wrote: > Fix btrfs_read/read_and_truncate_page write out of bounds of destination > buffer. Old behavior break bootstd malloc'd buffers of exact file size. > Previously this OOB write have not been noticed because distroboot usually > read files into huge static memory areas. > > Signed-off-by: Alex Shumsky <alexthr...@gmail.com> > --- > > fs/btrfs/inode.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c > index 4691612eda..b51f578b49 100644 > --- a/fs/btrfs/inode.c > +++ b/fs/btrfs/inode.c > @@ -640,7 +640,7 @@ static int read_and_truncate_page(struct btrfs_path *path, > extent_type = btrfs_file_extent_type(leaf, fi); > if (extent_type == BTRFS_FILE_EXTENT_INLINE) { > ret = btrfs_read_extent_inline(path, fi, buf); > - memcpy(dest, buf + page_off, min(page_len, ret)); > + memcpy(dest, buf + page_off, min(min(page_len, ret), len));
This looks (still) broken for the compressed inline extent error paths, where ret=-errno . nit: please prefix your patch subject with "uboot" so that casual linux-btrfs readers like me don't get confused by matching kernel file paths.