Merge tag 'upstream/18.11-rc2' into 18.11.x

New upstream version 18.11-rc2

Change-Id: I5ef3516f5bef8875ef584d1e68b4893c21eb1f8e
Signed-off-by: Luca Boccassi <luca.boccassi@gmail.com>

1 files changed, 178 insertions, 0 deletions

diff --git a/drivers/crypto/octeontx/otx_cryptodev_mbox.c b/drivers/crypto/octeontx/otx_cryptodev_mbox.c
new file mode 100644
index 00000000..a8e51a8e
--- /dev/null
+++ b/drivers/crypto/octeontx/otx_cryptodev_mbox.c
@@ -0,0 +1,178 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2018 Cavium, Inc
+ */
+
+#include <unistd.h>
+
+#include "otx_cryptodev_hw_access.h"
+#include "otx_cryptodev_mbox.h"
+
+void
+otx_cpt_handle_mbox_intr(struct cpt_vf *cptvf)
+{
+	struct cpt_mbox mbx = {0, 0};
+
+	/*
+	 * MBOX[0] contains msg
+	 * MBOX[1] contains data
+	 */
+	mbx.msg  = CPT_READ_CSR(CPT_CSR_REG_BASE(cptvf),
+				CPTX_VFX_PF_MBOXX(0, 0, 0));
+	mbx.data = CPT_READ_CSR(CPT_CSR_REG_BASE(cptvf),
+				CPTX_VFX_PF_MBOXX(0, 0, 1));
+
+	CPT_LOG_DP_DEBUG("%s: Mailbox msg 0x%lx from PF",
+		    cptvf->dev_name, (unsigned int long)mbx.msg);
+	switch (mbx.msg) {
+	case OTX_CPT_MSG_READY:
+		{
+			otx_cpt_chipid_vfid_t cid;
+
+			cid.u64 = mbx.data;
+			cptvf->pf_acked = true;
+			cptvf->vfid = cid.s.vfid;
+			CPT_LOG_DP_DEBUG("%s: Received VFID %d chip_id %d",
+					 cptvf->dev_name,
+					 cptvf->vfid, cid.s.chip_id);
+		}
+		break;
+	case OTX_CPT_MSG_QBIND_GRP:
+		cptvf->pf_acked = true;
+		cptvf->vftype = mbx.data;
+		CPT_LOG_DP_DEBUG("%s: VF %d type %s group %d",
+				 cptvf->dev_name, cptvf->vfid,
+				 ((mbx.data == SE_TYPE) ? "SE" : "AE"),
+				 cptvf->vfgrp);
+		break;
+	case OTX_CPT_MBOX_MSG_TYPE_ACK:
+		cptvf->pf_acked = true;
+		break;
+	case OTX_CPT_MBOX_MSG_TYPE_NACK:
+		cptvf->pf_nacked = true;
+		break;
+	default:
+		CPT_LOG_DP_DEBUG("%s: Invalid msg from PF, msg 0x%lx",
+				 cptvf->dev_name, (unsigned int long)mbx.msg);
+		break;
+	}
+}
+
+/* Send a mailbox message to PF
+ * @vf: vf from which this message to be sent
+ * @mbx: Message to be sent
+ */
+static void
+otx_cpt_send_msg_to_pf(struct cpt_vf *cptvf, struct cpt_mbox *mbx)
+{
+	/* Writing mbox(1) causes interrupt */
+	CPT_WRITE_CSR(CPT_CSR_REG_BASE(cptvf),
+		      CPTX_VFX_PF_MBOXX(0, 0, 0), mbx->msg);
+	CPT_WRITE_CSR(CPT_CSR_REG_BASE(cptvf),
+		      CPTX_VFX_PF_MBOXX(0, 0, 1), mbx->data);
+}
+
+static int32_t
+otx_cpt_send_msg_to_pf_timeout(struct cpt_vf *cptvf, struct cpt_mbox *mbx)
+{
+	int timeout = OTX_CPT_MBOX_MSG_TIMEOUT;
+	int sleep_ms = 10;
+
+	cptvf->pf_acked = false;
+	cptvf->pf_nacked = false;
+
+	otx_cpt_send_msg_to_pf(cptvf, mbx);
+
+	/* Wait for previous message to be acked, timeout 2sec */
+	while (!cptvf->pf_acked) {
+		if (cptvf->pf_nacked)
+			return -EINVAL;
+		usleep(sleep_ms * 1000);
+		otx_cpt_poll_misc(cptvf);
+		if (cptvf->pf_acked)
+			break;
+		timeout -= sleep_ms;
+		if (!timeout) {
+			CPT_LOG_ERR("%s: PF didn't ack mbox msg %lx(vfid %u)",
+				    cptvf->dev_name,
+				    (unsigned int long)(mbx->msg & 0xFF),
+				    cptvf->vfid);
+			return -EBUSY;
+		}
+	}
+	return 0;
+}
+
+int
+otx_cpt_check_pf_ready(struct cpt_vf *cptvf)
+{
+	struct cpt_mbox mbx = {0, 0};
+
+	mbx.msg = OTX_CPT_MSG_READY;
+	if (otx_cpt_send_msg_to_pf_timeout(cptvf, &mbx)) {
+		CPT_LOG_ERR("%s: PF didn't respond to READY msg",
+			    cptvf->dev_name);
+		return 1;
+	}
+	return 0;
+}
+
+int
+otx_cpt_send_vq_size_msg(struct cpt_vf *cptvf)
+{
+	struct cpt_mbox mbx = {0, 0};
+
+	mbx.msg = OTX_CPT_MSG_QLEN;
+
+	mbx.data = cptvf->qsize;
+	if (otx_cpt_send_msg_to_pf_timeout(cptvf, &mbx)) {
+		CPT_LOG_ERR("%s: PF didn't respond to vq_size msg",
+			    cptvf->dev_name);
+		return 1;
+	}
+	return 0;
+}
+
+int
+otx_cpt_send_vf_grp_msg(struct cpt_vf *cptvf, uint32_t group)
+{
+	struct cpt_mbox mbx = {0, 0};
+
+	mbx.msg = OTX_CPT_MSG_QBIND_GRP;
+
+	/* Convey group of the VF */
+	mbx.data = group;
+	if (otx_cpt_send_msg_to_pf_timeout(cptvf, &mbx)) {
+		CPT_LOG_ERR("%s: PF didn't respond to vf_type msg",
+			    cptvf->dev_name);
+		return 1;
+	}
+	return 0;
+}
+
+int
+otx_cpt_send_vf_up(struct cpt_vf *cptvf)
+{
+	struct cpt_mbox mbx = {0, 0};
+
+	mbx.msg = OTX_CPT_MSG_VF_UP;
+	if (otx_cpt_send_msg_to_pf_timeout(cptvf, &mbx)) {
+		CPT_LOG_ERR("%s: PF didn't respond to UP msg",
+			    cptvf->dev_name);
+		return 1;
+	}
+	return 0;
+}
+
+int
+otx_cpt_send_vf_down(struct cpt_vf *cptvf)
+{
+	struct cpt_mbox mbx = {0, 0};
+
+	mbx.msg = OTX_CPT_MSG_VF_DOWN;
+	if (otx_cpt_send_msg_to_pf_timeout(cptvf, &mbx)) {
+		CPT_LOG_ERR("%s: PF didn't respond to DOWN msg",
+			    cptvf->dev_name);
+		return 1;
+	}
+	return 0;
+}