netdev - Re: [PATCH net-next v1] net/9p/trans_fd: check the return value of parse

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 21 Sep 2022 15:12:38 -0700
From:   Li Zhong <floridsleeves@...il.com>
To:     asmadeus@...ewreck.org
Cc:     netdev@...r.kernel.org, v9fs-developer@...ts.sourceforge.net,
        pabeni@...hat.com, kuba@...nel.org, edumazet@...gle.com,
        davem@...emloft.net, linux_oss@...debyte.com, lucho@...kov.net,
        ericvh@...il.com
Subject: Re: [PATCH net-next v1] net/9p/trans_fd: check the return value of parse_opts

On Wed, Sep 21, 2022 at 2:23 PM <asmadeus@...ewreck.org> wrote:
>
> Li Zhong wrote on Wed, Sep 21, 2022 at 02:09:21PM -0700:
> > parse_opts() could fail when there is error parsing mount options into
> > p9_fd_opts structure due to allocation failure. In that case opts will
> > contain invalid data.
>
> In practice opts->rfd/wfd is set to ~0 before the failure modes so they
> will contain exactly what we want them to contain: something that'll
> fail the check below.
>
> It is however cleared like this so I'll queue this patch in 9p tree when
> I have a moment, but I'll clarify the commit message to say this is
> NO-OP : please feel free to send a v2 if you want to put your own words
> in there; otherwise it'll be something like below:
> ----
> net/9p: clarify trans_fd parse_opt failure handling
>
> This parse_opts will set invalid opts.rfd/wfd in case of failure which
> we already check, but it is not clear for readers that parse_opts error
> are handled in p9_fd_create: clarify this by explicitly checking the
> return value.
> ----
>

Thanks for the patient reply! I agree that the check on
opts.rfd/wft against ~0 will prevent error even if it fails
memory allocation. But currently the error log is
'insufficient options', which is kind of misleading and the
error code returned is -ENOPROTOOPT instead of -ENOMEM, which
I guess would be better if we distinguish between them.

>
> Also, in practice args != null doesn't seem to be checked before (the
> parse_opt() in client.c allows it) so keeping the error message common
> might be better?
> (allocation failure will print its own messages anyway and doesn't need
> checking)
>
> >
> > Signed-off-by: Li Zhong <floridsleeves@...il.com>
> > ---
> >  net/9p/trans_fd.c | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/net/9p/trans_fd.c b/net/9p/trans_fd.c
> > index e758978b44be..11ae64c1a24b 100644
> > --- a/net/9p/trans_fd.c
> > +++ b/net/9p/trans_fd.c
> > @@ -1061,7 +1061,9 @@ p9_fd_create(struct p9_client *client, const char *addr, char *args)
> >       int err;
> >       struct p9_fd_opts opts;
> >
> > -     parse_opts(args, &opts);
> > +     err = parse_opts(args, &opts);
> > +     if (err < 0)
> > +             return err;
> >       client->trans_opts.fd.rfd = opts.rfd;
> >       client->trans_opts.fd.wfd = opts.wfd;
> >