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:   Mon, 26 Jul 2021 13:45:40 +0100
From:   Matthew Wilcox <willy@infradead.org>
To:     Andrew Morton <akpm@linux-foundation.org>
Cc:     syzbot <syzbot+bb4c69145b4a52b40b27@syzkaller.appspotmail.com>,
        linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        syzkaller-bugs@googlegroups.com
Subject: Re: [syzbot] BUG: sleeping function called from invalid context in
 folio_copy
Message-ID: <YP6udKVlejbnTOiw@casper.infradead.org>
References: <000000000000f84cad05c7ee4778@google.com>
 <20210725221834.35e077daff0005f3d78d08d7@linux-foundation.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20210725221834.35e077daff0005f3d78d08d7@linux-foundation.org>
Precedence: bulk

On Sun, Jul 25, 2021 at 10:18:34PM -0700, Andrew Morton wrote:
> (Cc Matthew)
> 
> (It might not be a folio thing at all - I haven't checked yet)

Already fixed, thanks.  folio_copy() had a cond_resched() in it,
even if it was only copying a single page.  Now it only calls
cond_resched() if it's copying more than one page.

We discussed this in the folio call on Friday, because we're leaving a
bit of a landmine here for whoever tries to make multi-page folios work
with buffer_heads.  There's not much we can do about it for now; it'll
just be one of the many problems to be dealt with.  Hopefully filesystems
will convert to iomap instead of trying to make buffer_heads support
multiple pages.