On Tue, Jun 04, 2024 at 10:35:43AM -0700, Ivan Babrou wrote: > On Mon, Jun 3, 2024 at 8:40 PM Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote: > > > > ... when the only way to get to it is via ->getattr() in > > proc_fd_inode_operations? Note that proc_fd_inode_operations > > has ->lookup() in it; it _can't_ be ->i_op of a non-directory. > > > > Am I missing something here? > > It's been two years, but I think I was just extra cautious. Does anyone have objections against the following? [PATCH] proc_fd_getattr(): don't bother with S_ISDIR() check that thing is callable only as ->i_op->getattr() instance and only for directory inodes (/proc/*/fd and /proc/*/task/*/fd) Signed-off-by: Al Viro <viro@xxxxxxxxxxxxxxxxxx> --- diff --git a/fs/proc/fd.c b/fs/proc/fd.c index 586bbc84ca04..bda7abcf29fa 100644 --- a/fs/proc/fd.c +++ b/fs/proc/fd.c @@ -357,18 +357,9 @@ static int proc_fd_getattr(struct mnt_idmap *idmap, u32 request_mask, unsigned int query_flags) { struct inode *inode = d_inode(path->dentry); - int rv = 0; generic_fillattr(&nop_mnt_idmap, request_mask, inode, stat); - - /* If it's a directory, put the number of open fds there */ - if (S_ISDIR(inode->i_mode)) { - rv = proc_readfd_count(inode, &stat->size); - if (rv < 0) - return rv; - } - - return rv; + return proc_readfd_count(inode, &stat->size); } const struct inode_operations proc_fd_inode_operations = {