2019-10-08 12:34:58

by Matteo Croce

Subject: [PATCH] staging: vchiq: don't leak kernel address

Since commit ad67b74d2469d9b8 ("printk: hash addresses printed with %p"),
an obfuscated kernel pointer is printed at boot:

vchiq: vchiq_init_state: slot_zero = (____ptrval____)

Remove the the print completely, as it's useless without the address.

Signed-off-by: Matteo Croce <[email protected]>
---
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.c | 3 ---
1 file changed, 3 deletions(-)

diff --git a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.c b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.c
index 56a23a297fa4..084cd4b0f07c 100644
--- a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.c
+++ b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.c
@@ -2132,9 +2132,6 @@ vchiq_init_state(struct vchiq_state *state, struct vchiq_slot_zero *slot_zero)
char threadname[16];
int i;

- vchiq_log_warning(vchiq_core_log_level,
- "%s: slot_zero = %pK", __func__, slot_zero);
-
if (vchiq_states[0]) {
pr_err("%s: VCHIQ state already initialized\n", __func__);
return VCHIQ_ERROR;
--
2.21.0

2019-10-08 13:19:15

by Dan Carpenter

Subject: Re: [PATCH] staging: vchiq: don't leak kernel address

The subject doesn't match the patch. It should just be "remove useless
printk".

regards,
dan carpenter

2019-10-08 14:25:12

by Matteo Croce

Subject: Re: [PATCH] staging: vchiq: don't leak kernel address

On Tue, Oct 8, 2019 at 3:16 PM Dan Carpenter <[email protected]> wrote:
>
> The subject doesn't match the patch. It should just be "remove useless
> printk".
>
> regards,
> dan carpenter
>

Well, it avoids leaking an address by removing an useless printk.
It seems that GKH already picked the patch in his staging tree, but
I'm fine with both subjects, really,

Greg?

--
Matteo Croce
per aspera ad upstream

2019-10-08 14:28:30

by Dan Carpenter

Subject: Re: [PATCH] staging: vchiq: don't leak kernel address

On Tue, Oct 08, 2019 at 04:21:54PM +0200, Matteo Croce wrote:
> On Tue, Oct 8, 2019 at 3:16 PM Dan Carpenter <[email protected]> wrote:
> >
> > The subject doesn't match the patch. It should just be "remove useless
> > printk".
> >
> > regards,
> > dan carpenter
> >
>
> Well, it avoids leaking an address by removing an useless printk.
> It seems that GKH already picked the patch in his staging tree, but
> I'm fine with both subjects, really,

The address wasn't leaked because it was already %pK. The subject
says there is an info leak security problem, when the opposite is true.

regards,
dan carpenter

2019-10-08 15:16:36

by Greg Kroah-Hartman

Subject: Re: [PATCH] staging: vchiq: don't leak kernel address

On Tue, Oct 08, 2019 at 05:25:17PM +0300, Dan Carpenter wrote:
> On Tue, Oct 08, 2019 at 04:21:54PM +0200, Matteo Croce wrote:
> > On Tue, Oct 8, 2019 at 3:16 PM Dan Carpenter <[email protected]> wrote:
> > >
> > > The subject doesn't match the patch. It should just be "remove useless
> > > printk".
> > >
> > > regards,
> > > dan carpenter
> > >
> >
> > Well, it avoids leaking an address by removing an useless printk.
> > It seems that GKH already picked the patch in his staging tree, but
> > I'm fine with both subjects, really,
>
> The address wasn't leaked because it was already %pK. The subject
> says there is an info leak security problem, when the opposite is true.

I've edited the subject line now.

thanks,

greg k-h