[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <202402070156.189B1D8@keescook>
Date: Wed, 7 Feb 2024 01:56:55 -0800
From: Kees Cook <keescook@...omium.org>
To: "Gustavo A. R. Silva" <gustavo@...eddedor.com>
Cc: Brian Norris <briannorris@...omium.org>, Kalle Valo <kvalo@...nel.org>,
Dmitry Antipov <dmantipov@...dex.ru>,
Johannes Berg <johannes.berg@...el.com>,
zuoqilin <zuoqilin@...ong.com>, Ruan Jinjie <ruanjinjie@...wei.com>,
Thomas Gleixner <tglx@...utronix.de>,
Christophe JAILLET <christophe.jaillet@...adoo.fr>,
"Gustavo A . R . Silva" <gustavoars@...nel.org>,
linux-wireless@...r.kernel.org, Dan Carpenter <error27@...il.com>,
Rafael Beims <rafael.beims@...adex.com>,
David Lin <yu-hao.lin@....com>, Lukas Wunner <lukas@...ner.de>,
Simon Horman <horms@...nel.org>, linux-kernel@...r.kernel.org,
linux-hardening@...r.kernel.org
Subject: Re: [PATCH v2] wifi: mwifiex: Refactor 1-element array into flexible
array in struct mwifiex_ie_types_chan_list_param_set
On Tue, Feb 06, 2024 at 02:32:32PM -0600, Gustavo A. R. Silva wrote:
>
>
> On 2/6/24 12:39, Kees Cook wrote:
> > struct mwifiex_ie_types_chan_list_param_set::chan_scan_param is treated
> > as a flexible array, so convert it into one so that it doesn't trip the
> > array bounds sanitizer[1]. Only once place was using sizeof() on the
> > whole struct (in 11n.c), so adjust it to follow the calculation pattern
> > used by scan.c to avoid including the trailing single element.
> >
> > Link: https://github.com/KSPP/linux/issues/51 [1]
> > Cc: Brian Norris <briannorris@...omium.org>
> > Cc: Kalle Valo <kvalo@...nel.org>
> > Cc: Dmitry Antipov <dmantipov@...dex.ru>
> > Cc: Johannes Berg <johannes.berg@...el.com>
> > Cc: zuoqilin <zuoqilin@...ong.com>
> > Cc: Ruan Jinjie <ruanjinjie@...wei.com>
> > Cc: Thomas Gleixner <tglx@...utronix.de>
> > Cc: Christophe JAILLET <christophe.jaillet@...adoo.fr>
> > Cc: Gustavo A. R. Silva <gustavoars@...nel.org>
> > Cc: linux-wireless@...r.kernel.org
> > Signed-off-by: Kees Cook <keescook@...omium.org>
> > ---
> > drivers/net/wireless/marvell/mwifiex/11n.c | 8 +++-----
> > drivers/net/wireless/marvell/mwifiex/fw.h | 2 +-
> > drivers/net/wireless/marvell/mwifiex/scan.c | 14 ++++++--------
> > 3 files changed, 10 insertions(+), 14 deletions(-)
> >
> > diff --git a/drivers/net/wireless/marvell/mwifiex/11n.c b/drivers/net/wireless/marvell/mwifiex/11n.c
> > index 90e401100898..9ed90da4dfcf 100644
> > --- a/drivers/net/wireless/marvell/mwifiex/11n.c
> > +++ b/drivers/net/wireless/marvell/mwifiex/11n.c
> > @@ -392,12 +392,10 @@ mwifiex_cmd_append_11n_tlv(struct mwifiex_private *priv,
> > chan_list =
> > (struct mwifiex_ie_types_chan_list_param_set *) *buffer;
> > - memset(chan_list, 0,
> > - sizeof(struct mwifiex_ie_types_chan_list_param_set));
> > + memset(chan_list, 0, struct_size(chan_list, chan_scan_param, 1));
> > chan_list->header.type = cpu_to_le16(TLV_TYPE_CHANLIST);
> > - chan_list->header.len = cpu_to_le16(
> > - sizeof(struct mwifiex_ie_types_chan_list_param_set) -
> > - sizeof(struct mwifiex_ie_types_header));
> > + chan_list->header.len =
> > + cpu_to_le16(sizeof(struct mwifiex_chan_scan_param_set));
> > chan_list->chan_scan_param[0].chan_number =
> > bss_desc->bcn_ht_oper->primary_chan;
> > chan_list->chan_scan_param[0].radio_type =
> This probably still needs a bit more work.
>
> There are a couple more instances of `sizeof()` that should probably be
> audited and adjusted:
>
> drivers/net/wireless/marvell/mwifiex/11n.c:414: *buffer += sizeof(struct mwifiex_ie_types_chan_list_param_set);
> drivers/net/wireless/marvell/mwifiex/11n.c:415: ret_len += sizeof(struct mwifiex_ie_types_chan_list_param_set);
Good call. I think this is the right delta:
- *buffer += sizeof(struct mwifiex_ie_types_chan_list_param_set);
- ret_len += sizeof(struct mwifiex_ie_types_chan_list_param_set);
+ *buffer += struct_size(chan_list, chan_scan_param, 1);
+ ret_len += struct_size(chan_list, chan_scan_param, 1);
I will send a v3.
--
Kees Cook
Powered by blists - more mailing lists