Received-SPF: pass (google.com: domain of linux-ext4-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <c6fb0951-a472-dbb4-1970-fe9cece5d182@huawei.com>
 <CAD+ocbwkQ4rMYhiOm4msnBH65vh6Pm25ZkPsC2pD0sFy68bPgA@mail.gmail.com> <YDfYC+xUal5EdibL@mit.edu>
In-Reply-To: <YDfYC+xUal5EdibL@mit.edu>
From:   harshad shirwadkar <harshadshirwadkar@gmail.com>
Date:   Thu, 25 Feb 2021 09:13:15 -0800
Message-ID: <CAD+ocbyebaRPYws8==NuPS=v-bjZVSoPYc2AM=mvRVaYs9eZXQ@mail.gmail.com>
Subject: Re: [PATCH] debugfs: fix memory leak problem in read_list()
To:     "Theodore Ts'o" <tytso@mit.edu>
Cc:     Zhiqiang Liu <liuzhiqiang26@huawei.com>,
        Ext4 Developers List <linux-ext4@vger.kernel.org>,
        linfeilong <linfeilong@huawei.com>,
        lihaotian <lihaotian9@huawei.com>,
        "lijinlin (A)" <lijinlin3@huawei.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Thanks for the clarification, I think my repo was a bit stale.

- Harshad

On Thu, Feb 25, 2021 at 9:02 AM Theodore Ts'o <tytso@mit.edu> wrote:
>
> On Thu, Feb 25, 2021 at 07:51:09AM -0800, harshad shirwadkar wrote:
> > On Sat, Feb 20, 2021 at 12:41 AM Zhiqiang Liu <liuzhiqiang26@huawei.com> wrote:
> > >
> > >
> > > In read_list func, if strtoull() fails in while loop,
> > > we will return the error code directly. Then, memory of
> > > variable lst will be leaked without setting to *list.
> > >
> > > Signed-off-by: Zhiqiang Liu <liuzhiqiang26@huawei.com>
> > > Signed-off-by: linfeilong <linfeilong@huawei.com>
> > > ---
> > >  debugfs/util.c | 12 ++++++++----
> > >  1 file changed, 8 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/debugfs/util.c b/debugfs/util.c
> > > index be6b550e..9e880548 100644
> > > --- a/debugfs/util.c
> > > +++ b/debugfs/util.c
> > > @@ -530,12 +530,16 @@ errcode_t read_list(char *str, blk64_t **list, size_t *len)
> > >
> > >                 errno = 0;
> > >                 y = x = strtoull(tok, &e, 0);
> > > -               if (errno)
> > > -                       return errno;
> > > +               if (errno) {
> > > +                       retval = errno;
> > > +                       break;
> > > +               }
> > Shouldn't we have `goto err;` here instead of break? strtoull failure
> > here indicates that no valid value was found, so instead of returning
> > the allocated memory, we should just free the memory and return error.
>
> As of commit 462c424500a5 ("debugfs: fix memory allocation failures
> when parsing journal_write arguments") there is no longer the err:
> goto target.  The goal is to move to a model where the caller is
> exclusively responsible for freeing any allocated memory, since if
> realloc() has gotten into the act, the memory pointed to in *list
> would have been freed by realloc().  The fix is to make sure *list is
> updated before we return.  This also allows the caller to have access
> to the list of numbers parsed before we ran into an error.
>
> So the Zhiqiang's patch is correc, and I will apply it.
>
>                                          - Ted