Re: [PATCH net-next v12 4/5] net: rnpgbe: Add basic mbx_fw support
flat view
From: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Date: 2025-09-17 10:45:49
Also in:
linux-doc, linux-hardening, lkml
On 16/09/2025 12:29, Dong Yibo wrote:
Add fundamental firmware (FW) communication operations via PF-FW mailbox, including: - FW sync (via HW info query with retries) - HW reset (post FW command to reset hardware) - MAC address retrieval (request FW for port-specific MAC) - Power management (powerup/powerdown notification to FW) Signed-off-by: Dong Yibo <dong100@mucse.com>
Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev> small nits below
quoted hunk ↗ jump to hunk
+static void build_get_hw_info_req(struct mbx_fw_cmd_req *req) +{ + req->flags = 0; + req->opcode = cpu_to_le16(GET_HW_INFO); + req->datalen = cpu_to_le16(MUCSE_MBX_REQ_HDR_LEN); + req->reply_lo = 0; + req->reply_hi = 0; +}
All these build*() functions re-init flags and reply to 0, but all mbx_fw_cmd_req are zero-inited on the stack. Might be better clean things assignments, but no strong opinion because the code is explicit If you will think of refactoring this part, it might be a good idea to avoid build*() functions at all and do proper initialization of mbx_fw_cmd_req in callers?
quoted hunk ↗ jump to hunk
+ +/** + * mucse_mbx_get_info - Get hw info from fw + * @hw: pointer to the HW structure + * + * mucse_mbx_get_info tries to get hw info from hw. + * + * Return: 0 on success, negative errno on failure + **/ +static int mucse_mbx_get_info(struct mucse_hw *hw) +{ + struct mbx_fw_cmd_reply reply = {}; + struct mbx_fw_cmd_req req = {};
something like:
struct mbx_fw_cmd_req req =
{
.opcode = cpu_to_le16(GET_HW_INFO),
.datalen = cpu_to_le16(MUCSE_MBX_REQ_HDR_LEN),
}