On Sat, Jun 15, 2024 at 12:31:49AM -0700, Christoph Hellwig wrote: > On Sat, Jun 15, 2024 at 07:55:28AM +0100, Al Viro wrote: > > It is an inode of _some_ dentry; it's most definitely not that > > of the argument named 'dentry'. > > No need to explain it here, the point was that this belongs into a > useful commit message. Seeing that fsdevel is archived, it might be worth spelling out, actually... Anyway, yes, something like "correct the inline descriptions of vfs_mkdir() et.al." ought to go into commit message. And it really ought to cover not just vfs_mkdir() - other similar comments from the same commit (6521f8917082 "namei: prepare for idmapped mounts") have similar issues. "Create a device node or file" for vfs_mknod() probably ought to be "Create a device node or other special file", while we are at it...
