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: <20160929100538.GA3590@OptiPlex>
Date:   Thu, 29 Sep 2016 18:05:38 +0800
From:   Geliang Tang <geliangtang@...il.com>
To:     Kees Cook <keescook@...omium.org>
Cc:     Anton Vorontsov <anton@...msg.org>,
        Colin Cross <ccross@...roid.com>,
        Tony Luck <tony.luck@...el.com>,
        Geliang Tang <geliangtang@...il.com>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] ramoops: add pdata NULL check to ramoops_probe

On Wed, Sep 28, 2016 at 12:45:13PM -0700, Kees Cook wrote:
> On Wed, Sep 28, 2016 at 3:32 AM, Geliang Tang <geliangtang@...il.com> wrote:
> > When the pdata is NULL, ramoops_probe() segfaults. So this patch adds
> > a NULL check to it.
> 
> While I don't mind the check, is this even possible? A device
> triggering a ramoops probe should already have a platform_data
> (excepting the DT case which is already covered). Is there a situation
> you can create to trigger this Oops?
> 

Hi Kees,

This did happen on my device recently.

We use platform_device_register() or platform_device_register_data() to set the
ramoops_platform_data into dev->platform_data. If this ramoops_platform_data did
not set successfully or if we pass a NULL parameter to platform_device_register(),
we will get a NULL pdata in ramoops_probe(). This will trigger a kernel Oops.

Here is a test for this. If we set dummy_data to NULL in ramoops_register_dummy(),
we will get the Oops:

@@ -747,6 +755,7 @@ static void ramoops_register_dummy(void)
	*/
	dummy_data->ecc_info.ecc_size = ramoops_ecc == 1 ? 16 : ramoops_ecc;

+       dummy_data = NULL;
	dummy = platform_device_register_data(NULL, "ramoops", -1,
				dummy_data, sizeof(struct ramoops_platform_data));
	if (IS_ERR(dummy)) {

So I think this pdata NULL check is useful.

Thanks.

-Geliang

> >
> > Signed-off-by: Geliang Tang <geliangtang@...il.com>
> > ---
> >  fs/pstore/ram.c | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/fs/pstore/ram.c b/fs/pstore/ram.c
> > index 6ad831b..dd9832d 100644
> > --- a/fs/pstore/ram.c
> > +++ b/fs/pstore/ram.c
> > @@ -576,6 +576,9 @@ static int ramoops_probe(struct platform_device *pdev)
> >         if (cxt->max_dump_cnt)
> >                 goto fail_out;
> >
> > +       if (!pdata)
> > +               goto fail_out;
> > +
> >         if (!pdata->mem_size || (!pdata->record_size && !pdata->console_size &&
> >                         !pdata->ftrace_size && !pdata->pmsg_size)) {
> >                 pr_err("The memory size and the record/console size must be "
> > --
> > 2.7.4
> >

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ