[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Y6NTGjcM8aUE0M5w@localhost.localdomain>
Date: Wed, 21 Dec 2022 19:40:26 +0100
From: Michal Swiatkowski <michal.swiatkowski@...ux.intel.com>
To: Daniil Tatianin <d-tatianin@...dex-team.ru>
Cc: Shahed Shaikh <shshaikh@...vell.com>,
Manish Chopra <manishc@...vell.com>,
GR-Linux-NIC-Dev@...vell.com,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1] qlcnic: prevent ->dcb use-after-free on
qlcnic_dcb_enable() failure
On Wed, Dec 21, 2022 at 10:47:29AM +0300, Daniil Tatianin wrote:
> On 12/20/22 4:32 PM, Michal Swiatkowski wrote:
> > On Tue, Dec 20, 2022 at 03:56:49PM +0300, Daniil Tatianin wrote:
> > > adapter->dcb would get silently freed inside qlcnic_dcb_enable() in
> > > case qlcnic_dcb_attach() would return an error, which always happens
> > > under OOM conditions. This would lead to use-after-free because both
> > > of the existing callers invoke qlcnic_dcb_get_info() on the obtained
> > > pointer, which is potentially freed at that point.
> > >
> > > Propagate errors from qlcnic_dcb_enable(), and instead free the dcb
> > > pointer at callsite.
> > >
> > > Found by Linux Verification Center (linuxtesting.org) with the SVACE
> > > static analysis tool.
> > >
> > > Signed-off-by: Daniil Tatianin <d-tatianin@...dex-team.ru>
> >
> > Please add Fix tag and net as target (net-next is close till the end of
> > this year)
> >
> Sorry, I'll include a fixes tag.
> Could you please explain what I would have to do to add net as target?
Sorry, maybe it was bad wording. I meant to have PATCH net v2 in title.
For example by adding to git format-patch:
--subject-prefix="PATCH net v2"
> > > ---
> > > drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c | 9 ++++++++-
> > > drivers/net/ethernet/qlogic/qlcnic/qlcnic_dcb.h | 5 ++---
> > > drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c | 9 ++++++++-
> > > 3 files changed, 18 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c
> > > index dbb800769cb6..465f149d94d4 100644
> > > --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c
> > > +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c
> > > @@ -2505,7 +2505,14 @@ int qlcnic_83xx_init(struct qlcnic_adapter *adapter)
> > > goto disable_mbx_intr;
> > > qlcnic_83xx_clear_function_resources(adapter);
> > > - qlcnic_dcb_enable(adapter->dcb);
> > > +
> > > + err = qlcnic_dcb_enable(adapter->dcb);
> > > + if (err) {
> > > + qlcnic_clear_dcb_ops(adapter->dcb);
> > > + adapter->dcb = NULL;
> > > + goto disable_mbx_intr;
> > > + }
> >
> > Maybe I miss sth but it looks like there can be memory leak.
> > For example if error in attach happen after allocating of dcb->cfg.
> > Isn't it better to call qlcnic_dcb_free instead of qlcnic_clear_dcb_ops?
> I think you're right, if attach fails midway then we might leak cfg or
> something else.
> I'll use qlcnic_dcb_free() instead and completely remove
> qlcnic_clear_dcb_ops() as it
> seems to be unused and relies on dcb being in a very specific state.
Great, thanks
[...]
Powered by blists - more mailing lists