Re: [PATCH 1/6] seq_file: add seq_read_iter

[Al Viro]

On Sat, Nov 14, 2020 at 07:00:25AM +0000, Al Viro wrote:
> On Fri, Nov 13, 2020 at 11:19:34PM -0700, Nathan Chancellor wrote:
> 
> > Assuming so, I have attached the output both with and without the
> > WARN_ON. Looks like mountinfo is what is causing the error?
> 
> Cute...  FWIW, on #origin + that commit with fix folded in I don't
> see anything unusual in reads from mountinfo ;-/  OTOH, they'd
> obviously been... creative with readv(2) arguments, so it would
> be very interesting to see what it is they are passing to it.
> 
> I'm half-asleep right now; will try to cook something to gather
> that information tomorrow morning.  'Later...

OK, so let's do this: fix in seq_read_iter() + in do_loop_readv_writev()
(on entry) the following (racy as hell, but will do for debugging):

	bool weird = false;

	if (unlikely(memcmp(file->f_path.dentry->d_name.name, "mountinfo", 10))) {
		int i;

		for (i = 0; i < iter->nr_segs; i++)
			if (!iter->iov[i].iov_len)
				weird = true;
		if (weird) {
			printk(KERN_ERR "[%s]: weird readv on %p4D (%ld) ",
				current->comm, filp, (long)filp->f_pos);
			for (i = 0; i < iter->nr_segs; i++)
				printk(KERN_CONT "%c%zd", i ? ':' : '<',
					iter->iov[i].iov_len);
			printk(KERN_CONT "> ");
		}
	}
and in the end (just before return)
	if (weird)
		printk(KERN_CONT "-> %zd\n", ret);

Preferably along with the results of cat /proc/<whatever it is>/mountinfo both
on that and on the working kernel...