Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220429004244.3557063-1-goldstein.w.n@gmail.com> <8f8d07c1-9276-df86-f1dc-3d272d4ab91d@kernel.dk>
In-Reply-To: <8f8d07c1-9276-df86-f1dc-3d272d4ab91d@kernel.dk>
From:   Noah Goldstein <goldstein.w.n@gmail.com>
Date:   Fri, 29 Apr 2022 10:36:57 -0500
Message-ID: <CAFUsyfK-Mo76PNBHmvUnavHgemgb92g2muqWJjPKv5T7TE-rhA@mail.gmail.com>
Subject: Re: [PATCH v1] io_uring: Fix memory leak if file setup fails.
To:     Jens Axboe <axboe@kernel.dk>
Cc:     Pavel Begunkov <asml.silence@gmail.com>,
        "open list:IO_URING" <io-uring@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Apr 29, 2022 at 7:47 AM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 4/28/22 6:42 PM, Noah Goldstein wrote:
> > If `get_unused_fd_flags` files fails (either in setting up `ctx` as
> > `tctx->last` or `get_unused_fd_flags`) `ctx` will never be freed.
>
> There's a comment there telling you why, the fput will end up
> releasing it just like it would when an application closes it.

I see. Thought the 'it' refered to in the comment was the file, not
ctx.

>
> > I very well may be missing something (or there may be a double
> > free if the failure is after `get_unused_fd_flags`) but looks
> > to me to be a memory leak.
>
> Have you tried synthetically reproducing the two failures you're
> thinking of and tracing cleanup?
>
> --
> Jens Axboe
>