Jeff,
is 100% uniqeness is so much required for pipe inode numbers?
AFAIU, it is not that critical for pipefs (unlike smb, nfs etc.)
Thanks,
Kirill
> This converts pipefs to use the new scheme. Here we're calling iunique to get
> a unique i_ino value for the new inode, and then hashing it afterward. We
> call iunique with a max_reserved value of 1 to avoid collision with the root
> inode. Since the inode is now hashed, we need to take care that we end up in
> generic_delete_inode rather than generic_forget_inode or we'll create a nasty
> leak, so we clear_nlink when we destroy the pipe info.
>
> I'm not certain that this is the right place to add the clear_nlink, though
> it does seem to work. I'm open to suggestions on a better place to put
> this, or of a better way to make sure that we end up with i_nlink == 0 at
> iput time.
>
> Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
>
> diff --git a/fs/pipe.c b/fs/pipe.c
> index 68090e8..1d44ff0 100644
> --- a/fs/pipe.c
> +++ b/fs/pipe.c
> @@ -825,6 +825,7 @@ void free_pipe_info(struct inode *inode)
> {
> __free_pipe_info(inode->i_pipe);
> inode->i_pipe = NULL;
> + clear_nlink(inode);
> }
>
> static struct vfsmount *pipe_mnt __read_mostly;
> @@ -871,6 +872,8 @@ static struct inode * get_pipe_inode(void)
> inode->i_uid = current->fsuid;
> inode->i_gid = current->fsgid;
> inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
> + inode->i_ino = iunique(pipe_mnt->mnt_sb, 1);
> + insert_inode_hash(inode);
>
> return inode;
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
