Re: patch for 2.1.51 d_umount

Bill Hawes (whawes@star.net)
Tue, 19 Aug 1997 15:33:25 -0400

Linus Torvalds wrote:
> In this case, I don't see why initrd couldn't just increment the count on
> its own root inode, so that doing two dput()'s would be fine.. That way
> the normal code would stay normal.

I went to do this, and realized there's a slight problem -- d_umount
also checks the root d_count as a condition for unmounting. So the
count has to stay at 1.

I agree with your principle of keeping special cases separate, but this
seems like the exception; when unmounting the root, there is no covered
directory, so we have to check for it.

Regards,
Bill