Re: [PATCH] psi: Treat ksm swapping in copy as memstall

From: Johannes Weiner <hannes@cmpxchg.org>
To: CGEL <cgel.zte@gmail.com>
Cc: akpm@linux-foundation.org, sfr@canb.auug.org.au,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Yang Yang <yang.yang29@zte.com.cn>
Subject: Re: [PATCH] psi: Treat ksm swapping in copy as memstall
Date: Thu, 27 Jan 2022 20:29:08 -0500	[thread overview]
Message-ID: <YfNG5KAog6fI0kYu@cmpxchg.org> (raw)
In-Reply-To: <61ea820f.1c69fb81.e79d5.09c9@mx.google.com>

On Fri, Jan 21, 2022 at 09:51:08AM +0000, CGEL wrote:
>  Wed, Jan 19, 2022 at 07:58:23AM -0500, Johannes Weiner wrote:
> > On Wed, Jan 19, 2022 at 06:13:54AM +0000, CGEL wrote:
> > > I did a test, when we use zram, it takes longer time for ksm copying than
> > > swap_readpage(). Ksm copying average takes 147263ns, swap_readpage()
> > > average takes 55639ns. So I think this patch is reasonable.
> > 
> > Ok, that sounds reasonable to me as well. Please add the
> > PageWorkingset() check and resubmit the patch. Thanks!
> I am a litte confused about adding PageWorkingset(), since I
> think ksm_might_need_to_copy() memstall is like swap_readpage()
> memstall and swap_readpage() doesn't add PageWorkingset().

That's actually a bug! It should do that.

That psi_memstall_enter() in swap_readpage() was added by commit
937790699be9c8100e5358625e7dfa8b32bd33f2. It's for catching the
scenarios that don't go through submit_bio(). When you look at
submit_bio(), it counts stalls only if we have workingset pages:

	/*
	 * If we're reading data that is part of the userspace workingset, count
	 * submission time as memory stall.  When the device is congested, or
	 * the submitting cgroup IO-throttled, submission can be a significant
	 * part of overall IO time.
	 */
	if (unlikely(bio_op(bio) == REQ_OP_READ &&
	    bio_flagged(bio, BIO_WORKINGSET))) {
		unsigned long pflags;

		psi_memstall_enter(&pflags);
		submit_bio_noacct(bio);
		psi_memstall_leave(&pflags);
		return;
	}

I hope that clarifies it. I'll send a patch to fix up swap_readpage().

Thanks