Re: [RFC net-next 1/3] net: marvell: prestera: Add Switchdev driver for Prestera family ASIC device 98DX325x (AC3x)

From: Vadym Kochan
Date: Fri Feb 28 2020 - 03:06:09 EST


Hi Jiri,

On Thu, Feb 27, 2020 at 03:22:59PM +0100, Jiri Pirko wrote:
> Tue, Feb 25, 2020 at 05:30:54PM CET, vadym.kochan@xxxxxxxxxxx wrote:
> >Marvell Prestera 98DX326x integrates up to 24 ports of 1GbE with 8
> >ports of 10GbE uplinks or 2 ports of 40Gbps stacking for a largely
> >wireless SMB deployment.
> >
> >This driver implementation includes only L1 & basic L2 support.
> >
> >The core Prestera switching logic is implemented in prestera.c, there is
> >an intermediate hw layer between core logic and firmware. It is
> >implemented in prestera_hw.c, the purpose of it is to encapsulate hw
> >related logic, in future there is a plan to support more devices with
> >different HW related configurations.
> >
> >The following Switchdev features are supported:
> >
> > - VLAN-aware bridge offloading
> > - VLAN-unaware bridge offloading
> > - FDB offloading (learning, ageing)
> > - Switchport configuration
> >
> >Signed-off-by: Vadym Kochan <vadym.kochan@xxxxxxxxxxx>
> >Signed-off-by: Andrii Savka <andrii.savka@xxxxxxxxxxx>
> >Signed-off-by: Oleksandr Mazur <oleksandr.mazur@xxxxxxxxxxx>
> >Signed-off-by: Serhiy Boiko <serhiy.boiko@xxxxxxxxxxx>
> >Signed-off-by: Serhiy Pshyk <serhiy.pshyk@xxxxxxxxxxx>
> >Signed-off-by: Taras Chornyi <taras.chornyi@xxxxxxxxxxx>
> >Signed-off-by: Volodymyr Mytnyk <volodymyr.mytnyk@xxxxxxxxxxx>
> >---

[SNIP]

> >+};
> >+
> >+struct mvsw_msg_cmd {
> >+ u32 type;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_ret {
> >+ struct mvsw_msg_cmd cmd;
> >+ u32 status;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_common_request {
> >+ struct mvsw_msg_cmd cmd;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_common_response {
> >+ struct mvsw_msg_ret ret;
> >+} __packed __aligned(4);
> >+
> >+union mvsw_msg_switch_param {
> >+ u32 ageing_timeout;
> >+};
> >+
> >+struct mvsw_msg_switch_attr_cmd {
> >+ struct mvsw_msg_cmd cmd;
> >+ union mvsw_msg_switch_param param;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_switch_init_ret {
> >+ struct mvsw_msg_ret ret;
> >+ u32 port_count;
> >+ u32 mtu_max;
> >+ u8 switch_id;
> >+ u8 mac[ETH_ALEN];
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_port_autoneg_param {
> >+ u64 link_mode;
> >+ u8 enable;
> >+ u8 fec;
> >+};
> >+
> >+struct mvsw_msg_port_cap_param {
> >+ u64 link_mode;
> >+ u8 type;
> >+ u8 fec;
> >+ u8 transceiver;
> >+};
> >+
> >+union mvsw_msg_port_param {
> >+ u8 admin_state;
> >+ u8 oper_state;
> >+ u32 mtu;
> >+ u8 mac[ETH_ALEN];
> >+ u8 accept_frm_type;
> >+ u8 learning;
> >+ u32 speed;
> >+ u8 flood;
> >+ u32 link_mode;
> >+ u8 type;
> >+ u8 duplex;
> >+ u8 fec;
> >+ u8 mdix;
> >+ struct mvsw_msg_port_autoneg_param autoneg;
> >+ struct mvsw_msg_port_cap_param cap;
> >+};
> >+
> >+struct mvsw_msg_port_attr_cmd {
> >+ struct mvsw_msg_cmd cmd;
> >+ u32 attr;
> >+ u32 port;
> >+ u32 dev;
> >+ union mvsw_msg_port_param param;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_port_attr_ret {
> >+ struct mvsw_msg_ret ret;
> >+ union mvsw_msg_port_param param;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_port_stats_ret {
> >+ struct mvsw_msg_ret ret;
> >+ u64 stats[MVSW_PORT_CNT_MAX];
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_port_info_cmd {
> >+ struct mvsw_msg_cmd cmd;
> >+ u32 port;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_port_info_ret {
> >+ struct mvsw_msg_ret ret;
> >+ u32 hw_id;
> >+ u32 dev_id;
> >+ u16 fp_id;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_vlan_cmd {
> >+ struct mvsw_msg_cmd cmd;
> >+ u32 port;
> >+ u32 dev;
> >+ u16 vid;
> >+ u8 is_member;
> >+ u8 is_tagged;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_fdb_cmd {
> >+ struct mvsw_msg_cmd cmd;
> >+ u32 port;
> >+ u32 dev;
> >+ u8 mac[ETH_ALEN];
> >+ u16 vid;
> >+ u8 dynamic;
> >+ u32 flush_mode;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_event {
> >+ u16 type;
> >+ u16 id;
> >+} __packed __aligned(4);
> >+
> >+union mvsw_msg_event_fdb_param {
> >+ u8 mac[ETH_ALEN];
> >+};
> >+
> >+struct mvsw_msg_event_fdb {
> >+ struct mvsw_msg_event id;
> >+ u32 port_id;
> >+ u32 vid;
> >+ union mvsw_msg_event_fdb_param param;
> >+} __packed __aligned(4);
> >+
> >+union mvsw_msg_event_port_param {
> >+ u32 oper_state;
> >+};
> >+
> >+struct mvsw_msg_event_port {
> >+ struct mvsw_msg_event id;
> >+ u32 port_id;
> >+ union mvsw_msg_event_port_param param;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_bridge_cmd {
> >+ struct mvsw_msg_cmd cmd;
> >+ u32 port;
> >+ u32 dev;
> >+ u16 bridge;
> >+} __packed __aligned(4);
> >+
> >+struct mvsw_msg_bridge_ret {
> >+ struct mvsw_msg_ret ret;
> >+ u16 bridge;
> >+} __packed __aligned(4);
> >+
> >+#define fw_check_resp(_response) \
> >+({ \
> >+ int __er = 0; \
> >+ typeof(_response) __r = (_response); \
> >+ if (__r->ret.cmd.type != MVSW_MSG_TYPE_ACK) \
> >+ __er = -EBADE; \
> >+ else if (__r->ret.status != MVSW_MSG_ACK_OK) \
> >+ __er = -EINVAL; \
> >+ (__er); \
> >+})
> >+
> >+#define __fw_send_req_resp(_switch, _type, _request, _response, _wait) \
>
> Please try to avoid doing functions in macros like this one and the
> previous one.
>
>
> >+({ \
> >+ int __e; \
> >+ typeof(_switch) __sw = (_switch); \
> >+ typeof(_request) __req = (_request); \
> >+ typeof(_response) __resp = (_response); \
> >+ __req->cmd.type = (_type); \
> >+ __e = __sw->dev->send_req(__sw->dev, \
> >+ (u8 *)__req, sizeof(*__req), \
> >+ (u8 *)__resp, sizeof(*__resp), \
> >+ _wait); \
> >+ if (!__e) \
> >+ __e = fw_check_resp(__resp); \
> >+ (__e); \
> >+})
> >+
> >+#define fw_send_req_resp(_sw, _t, _req, _resp) \
> >+ __fw_send_req_resp(_sw, _t, _req, _resp, 0)
> >+
> >+#define fw_send_req_resp_wait(_sw, _t, _req, _resp, _wait) \
> >+ __fw_send_req_resp(_sw, _t, _req, _resp, _wait)
> >+
> >+#define fw_send_req(_sw, _t, _req) \
>
> This should be function, not define

Yeah, I understand your point, but here was the reason:

all packed structs which defined here in prestera_hw.c
are used for transmission request/return to/from the firmware, each of
the struct requires to have req/ret member (depends if it is request or
return from the firmware), and the purpose of the macro is to avoid case
when someone can forget to add req/ret member to the new such structure.

>
>
> >+({ \
> >+ struct mvsw_msg_common_response __re; \
> >+ (fw_send_req_resp(_sw, _t, _req, &__re)); \
> >+})
> >+

Regards,
Vadym Kochan,