| 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
| ||
|
Date: Thu, 12 Nov 2020 11:40:33 +0800
From: Yun-hao Chung <howardchung@...gle.com>
To: Marcel Holtmann <marcel@...tmann.org>
Cc: linux-bluetooth <linux-bluetooth@...r.kernel.org>,
Luiz Augusto von Dentz <luiz.dentz@...il.com>,
Alain Michaud <alainm@...omium.org>,
Manish Mandlik <mmandlik@...omium.org>,
Miao-chen Chou <mcchou@...omium.org>,
"David S. Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>,
Johan Hedberg <johan.hedberg@...il.com>,
LKML <linux-kernel@...r.kernel.org>,
netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH v9 5/6] Bluetooth: Refactor read default sys config for
various types
On Wed, Nov 11, 2020 at 7:20 PM Marcel Holtmann <marcel@...tmann.org> wrote:
>
> Hi Howard,
>
> > Refactor read default system configuration function so that it's capable
> > of returning different types than u16
> >
> > Signed-off-by: Howard Chung <howardchung@...gle.com>
> > ---
> >
> > (no changes since v8)
> >
> > Changes in v8:
> > - Update the commit title and message
> >
> > net/bluetooth/mgmt_config.c | 140 +++++++++++++++++++++---------------
> > 1 file changed, 84 insertions(+), 56 deletions(-)
> >
> > diff --git a/net/bluetooth/mgmt_config.c b/net/bluetooth/mgmt_config.c
> > index 2d3ad288c78ac..282fbf82f3192 100644
> > --- a/net/bluetooth/mgmt_config.c
> > +++ b/net/bluetooth/mgmt_config.c
> > @@ -11,72 +11,100 @@
> > #include "mgmt_util.h"
> > #include "mgmt_config.h"
> >
> > -#define HDEV_PARAM_U16(_param_code_, _param_name_) \
> > -{ \
> > - { cpu_to_le16(_param_code_), sizeof(__u16) }, \
> > - { cpu_to_le16(hdev->_param_name_) } \
> > -}
> > +#define HDEV_PARAM_U16(_param_name_) \
> > + struct {\
> > + struct mgmt_tlv entry; \
> > + __le16 value; \
> > + } __packed _param_name_
> >
> > -#define HDEV_PARAM_U16_JIFFIES_TO_MSECS(_param_code_, _param_name_) \
> > -{ \
> > - { cpu_to_le16(_param_code_), sizeof(__u16) }, \
> > - { cpu_to_le16(jiffies_to_msecs(hdev->_param_name_)) } \
> > -}
> > +#define TLV_SET_U16(_param_code_, _param_name_) \
> > + { \
> > + { cpu_to_le16(_param_code_), sizeof(__u16) }, \
> > + cpu_to_le16(hdev->_param_name_) \
> > + }
> > +
> > +#define TLV_SET_U16_JIFFIES_TO_MSECS(_param_code_, _param_name_) \
> > + { \
> > + { cpu_to_le16(_param_code_), sizeof(__u16) }, \
> > + cpu_to_le16(jiffies_to_msecs(hdev->_param_name_)) \
> > + }
> >
> > int read_def_system_config(struct sock *sk, struct hci_dev *hdev, void *data,
> > u16 data_len)
> > {
> > - struct {
> > - struct mgmt_tlv entry;
> > - union {
> > - /* This is a simplification for now since all values
> > - * are 16 bits. In the future, this code may need
> > - * refactoring to account for variable length values
> > - * and properly calculate the required buffer size.
> > - */
> > - __le16 value;
> > - };
> > - } __packed params[] = {
> > + int ret;
> > + struct mgmt_rp_read_def_system_config {
> > /* Please see mgmt-api.txt for documentation of these values */
> > - HDEV_PARAM_U16(0x0000, def_page_scan_type),
> > - HDEV_PARAM_U16(0x0001, def_page_scan_int),
> > - HDEV_PARAM_U16(0x0002, def_page_scan_window),
> > - HDEV_PARAM_U16(0x0003, def_inq_scan_type),
> > - HDEV_PARAM_U16(0x0004, def_inq_scan_int),
> > - HDEV_PARAM_U16(0x0005, def_inq_scan_window),
> > - HDEV_PARAM_U16(0x0006, def_br_lsto),
> > - HDEV_PARAM_U16(0x0007, def_page_timeout),
> > - HDEV_PARAM_U16(0x0008, sniff_min_interval),
> > - HDEV_PARAM_U16(0x0009, sniff_max_interval),
> > - HDEV_PARAM_U16(0x000a, le_adv_min_interval),
> > - HDEV_PARAM_U16(0x000b, le_adv_max_interval),
> > - HDEV_PARAM_U16(0x000c, def_multi_adv_rotation_duration),
> > - HDEV_PARAM_U16(0x000d, le_scan_interval),
> > - HDEV_PARAM_U16(0x000e, le_scan_window),
> > - HDEV_PARAM_U16(0x000f, le_scan_int_suspend),
> > - HDEV_PARAM_U16(0x0010, le_scan_window_suspend),
> > - HDEV_PARAM_U16(0x0011, le_scan_int_discovery),
> > - HDEV_PARAM_U16(0x0012, le_scan_window_discovery),
> > - HDEV_PARAM_U16(0x0013, le_scan_int_adv_monitor),
> > - HDEV_PARAM_U16(0x0014, le_scan_window_adv_monitor),
> > - HDEV_PARAM_U16(0x0015, le_scan_int_connect),
> > - HDEV_PARAM_U16(0x0016, le_scan_window_connect),
> > - HDEV_PARAM_U16(0x0017, le_conn_min_interval),
> > - HDEV_PARAM_U16(0x0018, le_conn_max_interval),
> > - HDEV_PARAM_U16(0x0019, le_conn_latency),
> > - HDEV_PARAM_U16(0x001a, le_supv_timeout),
> > - HDEV_PARAM_U16_JIFFIES_TO_MSECS(0x001b,
> > - def_le_autoconnect_timeout),
> > - HDEV_PARAM_U16(0x001d, advmon_allowlist_duration),
> > - HDEV_PARAM_U16(0x001e, advmon_no_filter_duration),
> > + HDEV_PARAM_U16(def_page_scan_type);
> > + HDEV_PARAM_U16(def_page_scan_int);
> > + HDEV_PARAM_U16(def_page_scan_window);
> > + HDEV_PARAM_U16(def_inq_scan_type);
> > + HDEV_PARAM_U16(def_inq_scan_int);
> > + HDEV_PARAM_U16(def_inq_scan_window);
> > + HDEV_PARAM_U16(def_br_lsto);
> > + HDEV_PARAM_U16(def_page_timeout);
> > + HDEV_PARAM_U16(sniff_min_interval);
> > + HDEV_PARAM_U16(sniff_max_interval);
> > + HDEV_PARAM_U16(le_adv_min_interval);
> > + HDEV_PARAM_U16(le_adv_max_interval);
> > + HDEV_PARAM_U16(def_multi_adv_rotation_duration);
> > + HDEV_PARAM_U16(le_scan_interval);
> > + HDEV_PARAM_U16(le_scan_window);
> > + HDEV_PARAM_U16(le_scan_int_suspend);
> > + HDEV_PARAM_U16(le_scan_window_suspend);
> > + HDEV_PARAM_U16(le_scan_int_discovery);
> > + HDEV_PARAM_U16(le_scan_window_discovery);
> > + HDEV_PARAM_U16(le_scan_int_adv_monitor);
> > + HDEV_PARAM_U16(le_scan_window_adv_monitor);
> > + HDEV_PARAM_U16(le_scan_int_connect);
> > + HDEV_PARAM_U16(le_scan_window_connect);
> > + HDEV_PARAM_U16(le_conn_min_interval);
> > + HDEV_PARAM_U16(le_conn_max_interval);
> > + HDEV_PARAM_U16(le_conn_latency);
> > + HDEV_PARAM_U16(le_supv_timeout);
> > + HDEV_PARAM_U16(def_le_autoconnect_timeout);
> > + HDEV_PARAM_U16(advmon_allowlist_duration);
> > + HDEV_PARAM_U16(advmon_no_filter_duration);
> > + } __packed rp = {
> > + TLV_SET_U16(0x0000, def_page_scan_type),
> > + TLV_SET_U16(0x0001, def_page_scan_int),
> > + TLV_SET_U16(0x0002, def_page_scan_window),
> > + TLV_SET_U16(0x0003, def_inq_scan_type),
> > + TLV_SET_U16(0x0004, def_inq_scan_int),
> > + TLV_SET_U16(0x0005, def_inq_scan_window),
> > + TLV_SET_U16(0x0006, def_br_lsto),
> > + TLV_SET_U16(0x0007, def_page_timeout),
> > + TLV_SET_U16(0x0008, sniff_min_interval),
> > + TLV_SET_U16(0x0009, sniff_max_interval),
> > + TLV_SET_U16(0x000a, le_adv_min_interval),
> > + TLV_SET_U16(0x000b, le_adv_max_interval),
> > + TLV_SET_U16(0x000c, def_multi_adv_rotation_duration),
> > + TLV_SET_U16(0x000d, le_scan_interval),
> > + TLV_SET_U16(0x000e, le_scan_window),
> > + TLV_SET_U16(0x000f, le_scan_int_suspend),
> > + TLV_SET_U16(0x0010, le_scan_window_suspend),
> > + TLV_SET_U16(0x0011, le_scan_int_discovery),
> > + TLV_SET_U16(0x0012, le_scan_window_discovery),
> > + TLV_SET_U16(0x0013, le_scan_int_adv_monitor),
> > + TLV_SET_U16(0x0014, le_scan_window_adv_monitor),
> > + TLV_SET_U16(0x0015, le_scan_int_connect),
> > + TLV_SET_U16(0x0016, le_scan_window_connect),
> > + TLV_SET_U16(0x0017, le_conn_min_interval),
> > + TLV_SET_U16(0x0018, le_conn_max_interval),
> > + TLV_SET_U16(0x0019, le_conn_latency),
> > + TLV_SET_U16(0x001a, le_supv_timeout),
> > + TLV_SET_U16_JIFFIES_TO_MSECS(0x001b,
> > + def_le_autoconnect_timeout),
> > + TLV_SET_U16(0x001d, advmon_allowlist_duration),
> > + TLV_SET_U16(0x001e, advmon_no_filter_duration),
> > };
> > - struct mgmt_rp_read_def_system_config *rp = (void *)params;
> >
> > bt_dev_dbg(hdev, "sock %p", sk);
> >
> > - return mgmt_cmd_complete(sk, hdev->id,
> > - MGMT_OP_READ_DEF_SYSTEM_CONFIG,
> > - 0, rp, sizeof(params));
> > + ret = mgmt_cmd_complete(sk, hdev->id,
> > + MGMT_OP_READ_DEF_SYSTEM_CONFIG,
> > + 0, &rp, sizeof(rp));
> > + return ret;
> > }
> >
>
> frankly I would prefer if we do the re-factoring first and only then add new
> parameters.
Yes. This patch only does the refactor stuff and the next patch is the one that
adds the new parameter.
>
> While I am looking at this, I am also really confused about this
> JIFFIES_TO_MSEC business. We should actually not store anything in jiffies
> or use it in an API. Is there a good reason to keep storing things in
> jiffies internally?
>
I don't know about this, but it looks like it should be in a different
patch right?
> Regards
>
> Marcel
>
Powered by blists - more mailing lists