Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 1 Sep 2020 17:48:27 +0100
From:   Matthew Wilcox <willy@infradead.org>
To:     David Howells <dhowells@redhat.com>
Cc:     linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Subject: Re: [RFC PATCH 0/7] mm: Make more use of readahead_control
Message-ID: <20200901164827.GQ14765@casper.infradead.org>
References: <159897769535.405783.17587409235571100774.stgit@warthog.procyon.org.uk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <159897769535.405783.17587409235571100774.stgit@warthog.procyon.org.uk>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Sep 01, 2020 at 05:28:15PM +0100, David Howells wrote:
> Here's a set of patches to expand the use of the readahead_control struct,
> essentially from do_sync_mmap_readahead() down.

I like this.

> Note that I've been
> passing the number of pages to read in rac->_nr_pages, and then saving it
> and changing it certain points, e.g. page_cache_readahead_unbounded().

I do not like this.  You're essentially mutating the meaning of _nr_pages
as the ractl moves down the stack, and that's going to lead to bugs.
I'd keep it as a separate argument.

> Also there's an apparent minor bug in khugepaged.c that I've included a
> patch for: page_cache_sync_readahead() looks to be given the wrong size in
> collapse_file().

This needs to go in as an independent fix.  But you didn't actually cc Song.