qed: Support management-based resource locking
Global locking can't properly be used to synchronize between different
PFs in all scenarios, as those instances might reside in different
logical partitions [e.g., when a PF is assigned via PDA to some VM].
The management firmware provides a generic infrastructure for
device locks. For each 'resource', it's guaranteed it could be acquired
by at most a single PF at any given time [or by management firmware].
This patch adds the necessary logic in qed for utilizing said
infrastructure, implementing lock/unlock internal APIs.
Signed-off-by: Tomer Tayar <Tomer.Tayar@cavium.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/drivers/net/ethernet/qlogic/qed/qed_mcp.c b/drivers/net/ethernet/qlogic/qed/qed_mcp.c
index 490619f..ddcbc24 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_mcp.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_mcp.c
@@ -2271,3 +2271,179 @@ int qed_mcp_initiate_pf_flr(struct qed_hwfn *p_hwfn, struct qed_ptt *p_ptt)
return qed_mcp_cmd(p_hwfn, p_ptt, DRV_MSG_CODE_INITIATE_PF_FLR, 0,
&mcp_resp, &mcp_param);
}
+
+static int qed_mcp_resource_cmd(struct qed_hwfn *p_hwfn,
+ struct qed_ptt *p_ptt,
+ u32 param, u32 *p_mcp_resp, u32 *p_mcp_param)
+{
+ int rc;
+
+ rc = qed_mcp_cmd(p_hwfn, p_ptt, DRV_MSG_CODE_RESOURCE_CMD, param,
+ p_mcp_resp, p_mcp_param);
+ if (rc)
+ return rc;
+
+ if (*p_mcp_resp == FW_MSG_CODE_UNSUPPORTED) {
+ DP_INFO(p_hwfn,
+ "The resource command is unsupported by the MFW\n");
+ return -EINVAL;
+ }
+
+ if (*p_mcp_param == RESOURCE_OPCODE_UNKNOWN_CMD) {
+ u8 opcode = QED_MFW_GET_FIELD(param, RESOURCE_CMD_REQ_OPCODE);
+
+ DP_NOTICE(p_hwfn,
+ "The resource command is unknown to the MFW [param 0x%08x, opcode %d]\n",
+ param, opcode);
+ return -EINVAL;
+ }
+
+ return rc;
+}
+
+int
+__qed_mcp_resc_lock(struct qed_hwfn *p_hwfn,
+ struct qed_ptt *p_ptt,
+ struct qed_resc_lock_params *p_params)
+{
+ u32 param = 0, mcp_resp, mcp_param;
+ u8 opcode;
+ int rc;
+
+ switch (p_params->timeout) {
+ case QED_MCP_RESC_LOCK_TO_DEFAULT:
+ opcode = RESOURCE_OPCODE_REQ;
+ p_params->timeout = 0;
+ break;
+ case QED_MCP_RESC_LOCK_TO_NONE:
+ opcode = RESOURCE_OPCODE_REQ_WO_AGING;
+ p_params->timeout = 0;
+ break;
+ default:
+ opcode = RESOURCE_OPCODE_REQ_W_AGING;
+ break;
+ }
+
+ QED_MFW_SET_FIELD(param, RESOURCE_CMD_REQ_RESC, p_params->resource);
+ QED_MFW_SET_FIELD(param, RESOURCE_CMD_REQ_OPCODE, opcode);
+ QED_MFW_SET_FIELD(param, RESOURCE_CMD_REQ_AGE, p_params->timeout);
+
+ DP_VERBOSE(p_hwfn,
+ QED_MSG_SP,
+ "Resource lock request: param 0x%08x [age %d, opcode %d, resource %d]\n",
+ param, p_params->timeout, opcode, p_params->resource);
+
+ /* Attempt to acquire the resource */
+ rc = qed_mcp_resource_cmd(p_hwfn, p_ptt, param, &mcp_resp, &mcp_param);
+ if (rc)
+ return rc;
+
+ /* Analyze the response */
+ p_params->owner = QED_MFW_GET_FIELD(mcp_param, RESOURCE_CMD_RSP_OWNER);
+ opcode = QED_MFW_GET_FIELD(mcp_param, RESOURCE_CMD_RSP_OPCODE);
+
+ DP_VERBOSE(p_hwfn,
+ QED_MSG_SP,
+ "Resource lock response: mcp_param 0x%08x [opcode %d, owner %d]\n",
+ mcp_param, opcode, p_params->owner);
+
+ switch (opcode) {
+ case RESOURCE_OPCODE_GNT:
+ p_params->b_granted = true;
+ break;
+ case RESOURCE_OPCODE_BUSY:
+ p_params->b_granted = false;
+ break;
+ default:
+ DP_NOTICE(p_hwfn,
+ "Unexpected opcode in resource lock response [mcp_param 0x%08x, opcode %d]\n",
+ mcp_param, opcode);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+int
+qed_mcp_resc_lock(struct qed_hwfn *p_hwfn,
+ struct qed_ptt *p_ptt, struct qed_resc_lock_params *p_params)
+{
+ u32 retry_cnt = 0;
+ int rc;
+
+ do {
+ /* No need for an interval before the first iteration */
+ if (retry_cnt) {
+ if (p_params->sleep_b4_retry) {
+ u16 retry_interval_in_ms =
+ DIV_ROUND_UP(p_params->retry_interval,
+ 1000);
+
+ msleep(retry_interval_in_ms);
+ } else {
+ udelay(p_params->retry_interval);
+ }
+ }
+
+ rc = __qed_mcp_resc_lock(p_hwfn, p_ptt, p_params);
+ if (rc)
+ return rc;
+
+ if (p_params->b_granted)
+ break;
+ } while (retry_cnt++ < p_params->retry_num);
+
+ return 0;
+}
+
+int
+qed_mcp_resc_unlock(struct qed_hwfn *p_hwfn,
+ struct qed_ptt *p_ptt,
+ struct qed_resc_unlock_params *p_params)
+{
+ u32 param = 0, mcp_resp, mcp_param;
+ u8 opcode;
+ int rc;
+
+ opcode = p_params->b_force ? RESOURCE_OPCODE_FORCE_RELEASE
+ : RESOURCE_OPCODE_RELEASE;
+ QED_MFW_SET_FIELD(param, RESOURCE_CMD_REQ_RESC, p_params->resource);
+ QED_MFW_SET_FIELD(param, RESOURCE_CMD_REQ_OPCODE, opcode);
+
+ DP_VERBOSE(p_hwfn, QED_MSG_SP,
+ "Resource unlock request: param 0x%08x [opcode %d, resource %d]\n",
+ param, opcode, p_params->resource);
+
+ /* Attempt to release the resource */
+ rc = qed_mcp_resource_cmd(p_hwfn, p_ptt, param, &mcp_resp, &mcp_param);
+ if (rc)
+ return rc;
+
+ /* Analyze the response */
+ opcode = QED_MFW_GET_FIELD(mcp_param, RESOURCE_CMD_RSP_OPCODE);
+
+ DP_VERBOSE(p_hwfn, QED_MSG_SP,
+ "Resource unlock response: mcp_param 0x%08x [opcode %d]\n",
+ mcp_param, opcode);
+
+ switch (opcode) {
+ case RESOURCE_OPCODE_RELEASED_PREVIOUS:
+ DP_INFO(p_hwfn,
+ "Resource unlock request for an already released resource [%d]\n",
+ p_params->resource);
+ /* Fallthrough */
+ case RESOURCE_OPCODE_RELEASED:
+ p_params->b_released = true;
+ break;
+ case RESOURCE_OPCODE_WRONG_OWNER:
+ p_params->b_released = false;
+ break;
+ default:
+ DP_NOTICE(p_hwfn,
+ "Unexpected opcode in resource unlock response [mcp_param 0x%08x, opcode %d]\n",
+ mcp_param, opcode);
+ return -EINVAL;
+ }
+
+ return 0;
+}