linux-kernel - Re: security/loadpin: Allow to exclude specific file types

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]

Message-ID: <201905311330.EA6B6E5F@keescook>
Date:   Fri, 31 May 2019 13:33:02 -0700
From:   Kees Cook <keescook@...omium.org>
To:     Ke Wu <mikewu@...gle.com>
Cc:     Colin Ian King <colin.king@...onical.com>,
        James Morris <jamorris@...ux.microsoft.com>,
        "Serge E. Hallyn" <serge@...lyn.com>,
        linux-security-module@...r.kernel.org,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: security/loadpin: Allow to exclude specific file types

On Fri, May 31, 2019 at 11:03:17AM -0700, Ke Wu wrote:
> I think Coverity is correct. Note that it's the size of
> kernel_read_file_str (rather than exclude_read_files) doesn't equal to
> ignore_read_file_id.
> 
> This is because READING_MAX_ID is also an element in
> kernel_read_file_str, which makes the size of kernel_read_file_str to
> be READING_MAX_ID+1. I will send a new patch to fix the issue. Thanks
> for the  analysis!

Ah! Yes, I see now. I was looking at the wrong things. It should be
possible to just do:

> > >> 209                for (j = 0; j < ARRAY_SIZE(kernel_read_file_str); j++) {

for (j = 0; j < ARRAY_SIZE(ignore_read_file_id); j++)

and add a

BUILD_BUG_ON(ARRAY_SIZE(kernel_read_file_str) < ARRAY_SIZE(ignore_read_file_id))

for future robustness checking.

Thanks for looking at this more closely!

-Kees

> > >> 210                        if (strcmp(cur, kernel_read_file_str[j]) == 0) {
> > >> 211                                pr_info("excluding: %s\n",
> > >> 212                                        kernel_read_file_str[j]);
> > >>
> > >> CID 81977 (#1 of 1): Out-of-bounds write
> > >> overrun-local: Overrunning array ignore_read_file_id of 8 4-byte
> > >> elements at element index 8 (byte offset 35) using index j (which
> > >> evaluates to 8).
> > >>
> > >> 213                                ignore_read_file_id[j] = 1;
> > >>
> > >> According to Coverity ignore_read_file_id is an array of 8 integers.
> > >> However, ARRAY_SIZE(kernel_read_file_str) is 9, so we have an out of
> > >> bounds write on ignore_read_file[j] when j is 8.
> > >
> > > What am I missing? This doesn't fail the build:
> > >
> > > +       BUILD_BUG_ON(ARRAY_SIZE(exclude_read_files) !=
> > > +                    ARRAY_SIZE(ignore_read_file_id));
> > >
> > > They have the same number of elements.
> > >
> >
> > Yep, that's very true. I'll discuss this with Coverity as this seems
> > like a weird false positive.
> >
> > Apologies for the noise.
> >
> > Colin
> 
> 
> 
> -- 
> Ke Wu | Software Engineer | mikewu@...gle.com | Google Inc.

-- 
Kees Cook