This series introduces crypto(CPT) drivers(PF & VF) for Marvell OcteonTX2
CN96XX Soc.
OcteonTX2 SOC's resource virtualization unit (RVU) supports multiple
physical and virtual functions. Each of the PF/VF's functionality is
determined by what kind of resources are attached to it. When the CPT
block is attached to a VF, it can function as a security device.
The following document provides an overview of the hardware and
different drivers for the OcteonTX2 SOC:
https://www.kernel.org/doc/Documentation/networking/device_drivers/marvell/octeontx2.rst
The CPT PF driver is responsible for:
- Forwarding messages to/from VFs from/to admin function(AF),
- Enabling/disabling VFs,
- Loading/unloading microcode (creation/deletion of engine groups).
The CPT VF driver works as a crypto offload device.
This patch series includes:
- Patch to update existing Marvell sources to support the CPT driver.
- Patch that adds mailbox messages to the admin function (AF) driver,
to configure CPT HW registers.
- CPT PF driver patches that include AF<=>PF<=>VF mailbox communication,
sriov_configure, and firmware load to the acceleration engines.
- CPT VF driver patches that include VF<=>PF mailbox communication and
crypto offload support through the kernel cryptographic API.
This series is tested with CRYPTO_EXTRA_TESTS enabled and
CRYPTO_DISABLE_TESTS disabled.
Changes since v7:
* Removed writable entries in debugfs.
* Dropped IPsec support.
Changes since v6:
* Removed driver version.
Changes since v4:
* Rebased the patches onto net-next tree with base
'commit bc081a693a56 ("Merge branch 'Offload-tc-vlan-mangle-to-mscc_ocelot-switch'")'
Changes since v3:
* Splitup the patches into smaller patches with more informartion.
Changes since v2:
* Fixed C=1 warnings.
* Added code to exit CPT VF driver gracefully.
* Moved OcteonTx2 asm code to a header file under include/linux/soc/
Changes since v1:
* Moved Makefile changes from patch4 to patch2 and patch3.
Srujana Challa (12):
octeontx2-pf: move lmt flush to include/linux/soc
octeontx2-af: add mailbox interface for CPT
octeontx2-af: add debugfs entries for CPT block
drivers: crypto: add Marvell OcteonTX2 CPT PF driver
crypto: octeontx2: add mailbox communication with AF
crypto: octeontx2: enable SR-IOV and mailbox communication with VF
crypto: octeontx2: load microcode and create engine groups
crypto: octeontx2: add LF framework
crypto: octeontx2: add support to get engine capabilities
crypto: octeontx2: add virtual function driver support
crypto: octeontx2: add support to process the crypto request
crypto: octeontx2: register with linux crypto framework
MAINTAINERS | 2 +
drivers/crypto/marvell/Kconfig | 14 +
drivers/crypto/marvell/Makefile | 1 +
drivers/crypto/marvell/octeontx2/Makefile | 10 +
.../marvell/octeontx2/otx2_cpt_common.h | 123 ++
.../marvell/octeontx2/otx2_cpt_hw_types.h | 464 +++++
.../marvell/octeontx2/otx2_cpt_mbox_common.c | 202 ++
.../marvell/octeontx2/otx2_cpt_reqmgr.h | 197 ++
drivers/crypto/marvell/octeontx2/otx2_cptlf.c | 426 +++++
drivers/crypto/marvell/octeontx2/otx2_cptlf.h | 351 ++++
drivers/crypto/marvell/octeontx2/otx2_cptpf.h | 52 +
.../marvell/octeontx2/otx2_cptpf_main.c | 570 ++++++
.../marvell/octeontx2/otx2_cptpf_mbox.c | 331 ++++
.../marvell/octeontx2/otx2_cptpf_ucode.c | 1533 +++++++++++++++
.../marvell/octeontx2/otx2_cptpf_ucode.h | 162 ++
drivers/crypto/marvell/octeontx2/otx2_cptvf.h | 28 +
.../marvell/octeontx2/otx2_cptvf_algs.c | 1665 +++++++++++++++++
.../marvell/octeontx2/otx2_cptvf_algs.h | 170 ++
.../marvell/octeontx2/otx2_cptvf_main.c | 401 ++++
.../marvell/octeontx2/otx2_cptvf_mbox.c | 139 ++
.../marvell/octeontx2/otx2_cptvf_reqmgr.c | 539 ++++++
.../ethernet/marvell/octeontx2/af/Makefile | 3 +-
.../net/ethernet/marvell/octeontx2/af/mbox.h | 33 +
.../net/ethernet/marvell/octeontx2/af/rvu.c | 2 +-
.../net/ethernet/marvell/octeontx2/af/rvu.h | 2 +
.../ethernet/marvell/octeontx2/af/rvu_cpt.c | 229 +++
.../marvell/octeontx2/af/rvu_debugfs.c | 304 +++
.../ethernet/marvell/octeontx2/af/rvu_reg.h | 63 +-
.../marvell/octeontx2/nic/otx2_common.h | 13 +-
include/linux/soc/marvell/octeontx2/asm.h | 29 +
30 files changed, 8038 insertions(+), 20 deletions(-)
create mode 100644 drivers/crypto/marvell/octeontx2/Makefile
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_hw_types.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_reqmgr.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptlf.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptlf.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_reqmgr.c
create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c
create mode 100644 include/linux/soc/marvell/octeontx2/asm.h
--
2.28.0
CPT includes microcoded GigaCypher symmetric engines(SEs), IPsec
symmetric engines(IEs), and asymmetric engines (AEs).
Each engine receives CPT instructions from the engine groups it has
subscribed to. This patch loads microcode, configures three engine
groups(one for SEs, one for IEs and one for AEs), and configures
all engines.
Signed-off-by: Suheil Chandran <[email protected]>
Signed-off-by: Srujana Challa <[email protected]>
---
drivers/crypto/marvell/octeontx2/Makefile | 2 +-
.../marvell/octeontx2/otx2_cpt_common.h | 42 +
.../marvell/octeontx2/otx2_cpt_mbox_common.c | 77 +
drivers/crypto/marvell/octeontx2/otx2_cptpf.h | 3 +
.../marvell/octeontx2/otx2_cptpf_main.c | 73 +
.../marvell/octeontx2/otx2_cptpf_mbox.c | 46 +-
.../marvell/octeontx2/otx2_cptpf_ucode.c | 1368 +++++++++++++++++
.../marvell/octeontx2/otx2_cptpf_ucode.h | 161 ++
8 files changed, 1770 insertions(+), 2 deletions(-)
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.h
diff --git a/drivers/crypto/marvell/octeontx2/Makefile b/drivers/crypto/marvell/octeontx2/Makefile
index 8c8262e94f78..3c4155446296 100644
--- a/drivers/crypto/marvell/octeontx2/Makefile
+++ b/drivers/crypto/marvell/octeontx2/Makefile
@@ -2,6 +2,6 @@
obj-$(CONFIG_CRYPTO_DEV_OCTEONTX2_CPT) += octeontx2-cpt.o
octeontx2-cpt-objs := otx2_cptpf_main.o otx2_cptpf_mbox.o \
- otx2_cpt_mbox_common.o
+ otx2_cpt_mbox_common.o otx2_cptpf_ucode.o
ccflags-y += -I$(srctree)/drivers/net/ethernet/marvell/octeontx2/af
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
index 277c7c7f95cf..ae16dc102459 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
+++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
@@ -18,6 +18,37 @@
#define OTX2_CPT_RVU_FUNC_ADDR_S(blk, slot, offs) \
(((blk) << 20) | ((slot) << 12) | (offs))
+#define OTX2_CPT_INVALID_CRYPTO_ENG_GRP 0xFF
+#define OTX2_CPT_NAME_LENGTH 64
+
+#define BAD_OTX2_CPT_ENG_TYPE OTX2_CPT_MAX_ENG_TYPES
+
+enum otx2_cpt_eng_type {
+ OTX2_CPT_AE_TYPES = 1,
+ OTX2_CPT_SE_TYPES = 2,
+ OTX2_CPT_IE_TYPES = 3,
+ OTX2_CPT_MAX_ENG_TYPES,
+};
+
+/* Take mbox id from end of CPT mbox range in AF (range 0xA00 - 0xBFF) */
+#define MBOX_MSG_GET_ENG_GRP_NUM 0xBFF
+
+/*
+ * Message request and response to get engine group number
+ * which has attached a given type of engines (SE, AE, IE)
+ * This messages are only used between CPT PF <=> CPT VF
+ */
+struct otx2_cpt_egrp_num_msg {
+ struct mbox_msghdr hdr;
+ u8 eng_type;
+};
+
+struct otx2_cpt_egrp_num_rsp {
+ struct mbox_msghdr hdr;
+ u8 eng_type;
+ u8 eng_grp_num;
+};
+
static inline void otx2_cpt_write64(void __iomem *reg_base, u64 blk, u64 slot,
u64 offs, u64 val)
{
@@ -34,4 +65,15 @@ static inline u64 otx2_cpt_read64(void __iomem *reg_base, u64 blk, u64 slot,
int otx2_cpt_send_ready_msg(struct otx2_mbox *mbox, struct pci_dev *pdev);
int otx2_cpt_send_mbox_msg(struct otx2_mbox *mbox, struct pci_dev *pdev);
+
+int otx2_cpt_send_af_reg_requests(struct otx2_mbox *mbox,
+ struct pci_dev *pdev);
+int otx2_cpt_add_read_af_reg(struct otx2_mbox *mbox,
+ struct pci_dev *pdev, u64 reg, u64 *val);
+int otx2_cpt_add_write_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
+ u64 reg, u64 val);
+int otx2_cpt_read_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
+ u64 reg, u64 *val);
+int otx2_cpt_write_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
+ u64 reg, u64 val);
#endif /* __OTX2_CPT_COMMON_H */
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
index a122483b5976..ef1291c4881b 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
+++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
@@ -35,3 +35,80 @@ int otx2_cpt_send_ready_msg(struct otx2_mbox *mbox, struct pci_dev *pdev)
return otx2_cpt_send_mbox_msg(mbox, pdev);
}
+
+int otx2_cpt_send_af_reg_requests(struct otx2_mbox *mbox, struct pci_dev *pdev)
+{
+ return otx2_cpt_send_mbox_msg(mbox, pdev);
+}
+
+int otx2_cpt_add_read_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
+ u64 reg, u64 *val)
+{
+ struct cpt_rd_wr_reg_msg *reg_msg;
+
+ reg_msg = (struct cpt_rd_wr_reg_msg *)
+ otx2_mbox_alloc_msg_rsp(mbox, 0, sizeof(*reg_msg),
+ sizeof(*reg_msg));
+ if (reg_msg == NULL) {
+ dev_err(&pdev->dev, "RVU MBOX failed to get message.\n");
+ return -EFAULT;
+ }
+
+ reg_msg->hdr.id = MBOX_MSG_CPT_RD_WR_REGISTER;
+ reg_msg->hdr.sig = OTX2_MBOX_REQ_SIG;
+ reg_msg->hdr.pcifunc = 0;
+
+ reg_msg->is_write = 0;
+ reg_msg->reg_offset = reg;
+ reg_msg->ret_val = val;
+
+ return 0;
+}
+
+int otx2_cpt_add_write_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
+ u64 reg, u64 val)
+{
+ struct cpt_rd_wr_reg_msg *reg_msg;
+
+ reg_msg = (struct cpt_rd_wr_reg_msg *)
+ otx2_mbox_alloc_msg_rsp(mbox, 0, sizeof(*reg_msg),
+ sizeof(*reg_msg));
+ if (reg_msg == NULL) {
+ dev_err(&pdev->dev, "RVU MBOX failed to get message.\n");
+ return -EFAULT;
+ }
+
+ reg_msg->hdr.id = MBOX_MSG_CPT_RD_WR_REGISTER;
+ reg_msg->hdr.sig = OTX2_MBOX_REQ_SIG;
+ reg_msg->hdr.pcifunc = 0;
+
+ reg_msg->is_write = 1;
+ reg_msg->reg_offset = reg;
+ reg_msg->val = val;
+
+ return 0;
+}
+
+int otx2_cpt_read_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
+ u64 reg, u64 *val)
+{
+ int ret;
+
+ ret = otx2_cpt_add_read_af_reg(mbox, pdev, reg, val);
+ if (ret)
+ return ret;
+
+ return otx2_cpt_send_mbox_msg(mbox, pdev);
+}
+
+int otx2_cpt_write_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
+ u64 reg, u64 val)
+{
+ int ret;
+
+ ret = otx2_cpt_add_write_af_reg(mbox, pdev, reg, val);
+ if (ret)
+ return ret;
+
+ return otx2_cpt_send_mbox_msg(mbox, pdev);
+}
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf.h b/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
index c7c687cb091e..528a3975b9c2 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
@@ -6,6 +6,7 @@
#define __OTX2_CPTPF_H
#include "otx2_cpt_common.h"
+#include "otx2_cptpf_ucode.h"
struct otx2_cptpf_dev;
struct otx2_cptvf_info {
@@ -22,6 +23,8 @@ struct otx2_cptpf_dev {
void __iomem *vfpf_mbox_base; /* VF-PF mbox start address */
struct pci_dev *pdev; /* PCI device handle */
struct otx2_cptvf_info vf[OTX2_CPT_MAX_VFS_NUM];
+ struct otx2_cpt_eng_grps eng_grps;/* Engine groups information */
+
/* AF <=> PF mbox */
struct otx2_mbox afpf_mbox;
struct work_struct afpf_mbox_work;
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
index adff12adfdd7..165f431a5256 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
@@ -4,6 +4,7 @@
#include <linux/firmware.h>
#include "otx2_cpt_hw_types.h"
#include "otx2_cpt_common.h"
+#include "otx2_cptpf_ucode.h"
#include "otx2_cptpf.h"
#include "rvu_reg.h"
@@ -323,6 +324,59 @@ static int cpt_is_pf_usable(struct otx2_cptpf_dev *cptpf)
return 0;
}
+static int cptpf_device_reset(struct otx2_cptpf_dev *cptpf)
+{
+ int timeout = 10, ret;
+ u64 reg = 0;
+
+ ret = otx2_cpt_write_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_BLK_RST, 0x1);
+ if (ret)
+ return ret;
+
+ do {
+ ret = otx2_cpt_read_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_BLK_RST, ®);
+ if (ret)
+ return ret;
+
+ if (!((reg >> 63) & 0x1))
+ break;
+
+ usleep_range(10000, 20000);
+ if (timeout-- < 0)
+ return -EBUSY;
+ } while (1);
+
+ return ret;
+}
+
+static int cptpf_device_init(struct otx2_cptpf_dev *cptpf)
+{
+ union otx2_cptx_af_constants1 af_cnsts1 = {0};
+ int ret = 0;
+
+ /* Reset the CPT PF device */
+ ret = cptpf_device_reset(cptpf);
+ if (ret)
+ return ret;
+
+ /* Get number of SE, IE and AE engines */
+ ret = otx2_cpt_read_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_CONSTANTS1, &af_cnsts1.u);
+ if (ret)
+ return ret;
+
+ cptpf->eng_grps.avail.max_se_cnt = af_cnsts1.s.se;
+ cptpf->eng_grps.avail.max_ie_cnt = af_cnsts1.s.ie;
+ cptpf->eng_grps.avail.max_ae_cnt = af_cnsts1.s.ae;
+
+ /* Disable all cores */
+ ret = otx2_cpt_disable_all_cores(cptpf);
+
+ return ret;
+}
+
static int cptpf_sriov_disable(struct pci_dev *pdev)
{
struct otx2_cptpf_dev *cptpf = pci_get_drvdata(pdev);
@@ -354,6 +408,10 @@ static int cptpf_sriov_enable(struct pci_dev *pdev, int numvfs)
if (ret)
goto free_mbox;
+ ret = otx2_cpt_create_eng_grps(cptpf->pdev, &cptpf->eng_grps);
+ if (ret)
+ goto disable_intr;
+
cptpf->enabled_vfs = numvfs;
ret = pci_enable_sriov(pdev, numvfs);
if (ret)
@@ -448,7 +506,19 @@ static int otx2_cptpf_probe(struct pci_dev *pdev,
cptpf->max_vfs = pci_sriov_get_totalvfs(pdev);
+ /* Initialize CPT PF device */
+ err = cptpf_device_init(cptpf);
+ if (err)
+ goto unregister_intr;
+
+ /* Initialize engine groups */
+ err = otx2_cpt_init_eng_grps(pdev, &cptpf->eng_grps);
+ if (err)
+ goto unregister_intr;
+
return 0;
+unregister_intr:
+ cptpf_disable_afpf_mbox_intr(cptpf);
destroy_afpf_mbox:
cptpf_afpf_mbox_destroy(cptpf);
clear_drvdata:
@@ -462,7 +532,10 @@ static void otx2_cptpf_remove(struct pci_dev *pdev)
if (!cptpf)
return;
+
cptpf_sriov_disable(pdev);
+ /* Cleanup engine groups */
+ otx2_cpt_cleanup_eng_grps(pdev, &cptpf->eng_grps);
/* Disable AF-PF mailbox interrupt */
cptpf_disable_afpf_mbox_intr(cptpf);
/* Destroy AF-PF mbox */
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
index 398f1c6402d9..8852a7e5e035 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
@@ -35,6 +35,29 @@ static int forward_to_af(struct otx2_cptpf_dev *cptpf,
return 0;
}
+static int handle_msg_get_eng_grp_num(struct otx2_cptpf_dev *cptpf,
+ struct otx2_cptvf_info *vf,
+ struct mbox_msghdr *req)
+{
+ struct otx2_cpt_egrp_num_msg *grp_req;
+ struct otx2_cpt_egrp_num_rsp *rsp;
+
+ grp_req = (struct otx2_cpt_egrp_num_msg *)req;
+ rsp = (struct otx2_cpt_egrp_num_rsp *)
+ otx2_mbox_alloc_msg(&cptpf->vfpf_mbox, vf->vf_id, sizeof(*rsp));
+ if (!rsp)
+ return -ENOMEM;
+
+ rsp->hdr.id = MBOX_MSG_GET_ENG_GRP_NUM;
+ rsp->hdr.sig = OTX2_MBOX_RSP_SIG;
+ rsp->hdr.pcifunc = req->pcifunc;
+ rsp->eng_type = grp_req->eng_type;
+ rsp->eng_grp_num = otx2_cpt_get_eng_grp(&cptpf->eng_grps,
+ grp_req->eng_type);
+
+ return 0;
+}
+
static int cptpf_handle_vf_req(struct otx2_cptpf_dev *cptpf,
struct otx2_cptvf_info *vf,
struct mbox_msghdr *req, int size)
@@ -45,7 +68,15 @@ static int cptpf_handle_vf_req(struct otx2_cptpf_dev *cptpf,
if (req->sig != OTX2_MBOX_REQ_SIG)
goto inval_msg;
- return forward_to_af(cptpf, vf, req, size);
+ switch (req->id) {
+ case MBOX_MSG_GET_ENG_GRP_NUM:
+ err = handle_msg_get_eng_grp_num(cptpf, vf, req);
+ break;
+ default:
+ err = forward_to_af(cptpf, vf, req, size);
+ break;
+ }
+ return err;
inval_msg:
otx2_reply_invalid_msg(&cptpf->vfpf_mbox, vf->vf_id, 0, req->id);
@@ -148,6 +179,7 @@ static void process_afpf_mbox_msg(struct otx2_cptpf_dev *cptpf,
struct mbox_msghdr *msg)
{
struct device *dev = &cptpf->pdev->dev;
+ struct cpt_rd_wr_reg_msg *rsp_rd_wr;
if (msg->id >= MBOX_MSG_MAX) {
dev_err(dev, "MBOX msg with unknown ID %d\n", msg->id);
@@ -164,6 +196,18 @@ static void process_afpf_mbox_msg(struct otx2_cptpf_dev *cptpf,
cptpf->pf_id = (msg->pcifunc >> RVU_PFVF_PF_SHIFT) &
RVU_PFVF_PF_MASK;
break;
+ case MBOX_MSG_CPT_RD_WR_REGISTER:
+ rsp_rd_wr = (struct cpt_rd_wr_reg_msg *)msg;
+ if (msg->rc) {
+ dev_err(dev, "Reg %llx rd/wr(%d) failed %d\n",
+ rsp_rd_wr->reg_offset, rsp_rd_wr->is_write,
+ msg->rc);
+ return;
+ }
+ if (!rsp_rd_wr->is_write)
+ *rsp_rd_wr->ret_val = rsp_rd_wr->val;
+ break;
+
default:
dev_err(dev,
"Unsupported msg %d received.\n", msg->id);
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c
new file mode 100644
index 000000000000..190937b6515b
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c
@@ -0,0 +1,1368 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2020 Marvell. */
+
+#include <linux/ctype.h>
+#include <linux/firmware.h>
+#include "otx2_cptpf_ucode.h"
+#include "otx2_cpt_common.h"
+#include "otx2_cptpf.h"
+#include "rvu_reg.h"
+
+#define CSR_DELAY 30
+/* Tar archive defines */
+#define TAR_MAGIC "ustar"
+#define TAR_MAGIC_LEN 6
+#define TAR_BLOCK_LEN 512
+#define REGTYPE '0'
+#define AREGTYPE '\0'
+
+#define LOADFVC_RLEN 8
+#define LOADFVC_MAJOR_OP 0x01
+#define LOADFVC_MINOR_OP 0x08
+
+/* tar header as defined in POSIX 1003.1-1990. */
+struct tar_hdr_t {
+ char name[100];
+ char mode[8];
+ char uid[8];
+ char gid[8];
+ char size[12];
+ char mtime[12];
+ char chksum[8];
+ char typeflag;
+ char linkname[100];
+ char magic[6];
+ char version[2];
+ char uname[32];
+ char gname[32];
+ char devmajor[8];
+ char devminor[8];
+ char prefix[155];
+};
+
+struct tar_blk_t {
+ union {
+ struct tar_hdr_t hdr;
+ char block[TAR_BLOCK_LEN];
+ };
+};
+
+struct tar_arch_info_t {
+ struct list_head ucodes;
+ const struct firmware *fw;
+};
+
+static struct otx2_cpt_bitmap get_cores_bmap(struct device *dev,
+ struct otx2_cpt_eng_grp_info *eng_grp)
+{
+ struct otx2_cpt_bitmap bmap = { {0} };
+ bool found = false;
+ int i;
+
+ if (eng_grp->g->engs_num > OTX2_CPT_MAX_ENGINES) {
+ dev_err(dev, "unsupported number of engines %d on octeontx2\n",
+ eng_grp->g->engs_num);
+ return bmap;
+ }
+
+ for (i = 0; i < OTX2_CPT_MAX_ETYPES_PER_GRP; i++) {
+ if (eng_grp->engs[i].type) {
+ bitmap_or(bmap.bits, bmap.bits,
+ eng_grp->engs[i].bmap,
+ eng_grp->g->engs_num);
+ bmap.size = eng_grp->g->engs_num;
+ found = true;
+ }
+ }
+
+ if (!found)
+ dev_err(dev, "No engines reserved for engine group %d\n",
+ eng_grp->idx);
+ return bmap;
+}
+
+static int is_eng_type(int val, int eng_type)
+{
+ return val & (1 << eng_type);
+}
+
+static int is_2nd_ucode_used(struct otx2_cpt_eng_grp_info *eng_grp)
+{
+ if (eng_grp->ucode[1].type)
+ return true;
+ else
+ return false;
+}
+
+static void set_ucode_filename(struct otx2_cpt_ucode *ucode,
+ const char *filename)
+{
+ strlcpy(ucode->filename, filename, OTX2_CPT_NAME_LENGTH);
+}
+
+static char *get_eng_type_str(int eng_type)
+{
+ char *str = "unknown";
+
+ switch (eng_type) {
+ case OTX2_CPT_SE_TYPES:
+ str = "SE";
+ break;
+
+ case OTX2_CPT_IE_TYPES:
+ str = "IE";
+ break;
+
+ case OTX2_CPT_AE_TYPES:
+ str = "AE";
+ break;
+ }
+ return str;
+}
+
+static char *get_ucode_type_str(int ucode_type)
+{
+ char *str = "unknown";
+
+ switch (ucode_type) {
+ case (1 << OTX2_CPT_SE_TYPES):
+ str = "SE";
+ break;
+
+ case (1 << OTX2_CPT_IE_TYPES):
+ str = "IE";
+ break;
+
+ case (1 << OTX2_CPT_AE_TYPES):
+ str = "AE";
+ break;
+
+ case (1 << OTX2_CPT_SE_TYPES | 1 << OTX2_CPT_IE_TYPES):
+ str = "SE+IPSEC";
+ break;
+ }
+ return str;
+}
+
+static int get_ucode_type(struct device *dev,
+ struct otx2_cpt_ucode_hdr *ucode_hdr,
+ int *ucode_type)
+{
+ struct otx2_cptpf_dev *cptpf = dev_get_drvdata(dev);
+ char ver_str_prefix[OTX2_CPT_UCODE_VER_STR_SZ];
+ char tmp_ver_str[OTX2_CPT_UCODE_VER_STR_SZ];
+ struct pci_dev *pdev = cptpf->pdev;
+ int i, val = 0;
+ u8 nn;
+
+ strlcpy(tmp_ver_str, ucode_hdr->ver_str, OTX2_CPT_UCODE_VER_STR_SZ);
+ for (i = 0; i < strlen(tmp_ver_str); i++)
+ tmp_ver_str[i] = tolower(tmp_ver_str[i]);
+
+ sprintf(ver_str_prefix, "ocpt-%02d", pdev->revision);
+ if (!strnstr(tmp_ver_str, ver_str_prefix, OTX2_CPT_UCODE_VER_STR_SZ))
+ return -EINVAL;
+
+ nn = ucode_hdr->ver_num.nn;
+ if (strnstr(tmp_ver_str, "se-", OTX2_CPT_UCODE_VER_STR_SZ) &&
+ (nn == OTX2_CPT_SE_UC_TYPE1 || nn == OTX2_CPT_SE_UC_TYPE2 ||
+ nn == OTX2_CPT_SE_UC_TYPE3))
+ val |= 1 << OTX2_CPT_SE_TYPES;
+ if (strnstr(tmp_ver_str, "ie-", OTX2_CPT_UCODE_VER_STR_SZ) &&
+ (nn == OTX2_CPT_IE_UC_TYPE1 || nn == OTX2_CPT_IE_UC_TYPE2 ||
+ nn == OTX2_CPT_IE_UC_TYPE3))
+ val |= 1 << OTX2_CPT_IE_TYPES;
+ if (strnstr(tmp_ver_str, "ae", OTX2_CPT_UCODE_VER_STR_SZ) &&
+ nn == OTX2_CPT_AE_UC_TYPE)
+ val |= 1 << OTX2_CPT_AE_TYPES;
+
+ *ucode_type = val;
+
+ if (!val)
+ return -EINVAL;
+
+ return 0;
+}
+
+static int is_mem_zero(const char *ptr, int size)
+{
+ int i;
+
+ for (i = 0; i < size; i++) {
+ if (ptr[i])
+ return 0;
+ }
+ return 1;
+}
+
+static int __write_ucode_base(struct otx2_cptpf_dev *cptpf, int eng,
+ dma_addr_t dma_addr)
+{
+ return otx2_cpt_write_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_EXEX_UCODE_BASE(eng),
+ (u64)dma_addr);
+}
+
+static int cpt_set_ucode_base(struct otx2_cpt_eng_grp_info *eng_grp, void *obj)
+{
+ struct otx2_cptpf_dev *cptpf = obj;
+ struct otx2_cpt_engs_rsvd *engs;
+ dma_addr_t dma_addr;
+ int i, bit, ret;
+
+ /* Set PF number for microcode fetches */
+ ret = otx2_cpt_write_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_PF_FUNC,
+ cptpf->pf_id << RVU_PFVF_PF_SHIFT);
+ if (ret)
+ return ret;
+
+ for (i = 0; i < OTX2_CPT_MAX_ETYPES_PER_GRP; i++) {
+ engs = &eng_grp->engs[i];
+ if (!engs->type)
+ continue;
+
+ dma_addr = engs->ucode->dma;
+
+ /*
+ * Set UCODE_BASE only for the cores which are not used,
+ * other cores should have already valid UCODE_BASE set
+ */
+ for_each_set_bit(bit, engs->bmap, eng_grp->g->engs_num)
+ if (!eng_grp->g->eng_ref_cnt[bit]) {
+ ret = __write_ucode_base(cptpf, bit, dma_addr);
+ if (ret)
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static int cpt_detach_and_disable_cores(struct otx2_cpt_eng_grp_info *eng_grp,
+ void *obj)
+{
+ struct otx2_cptpf_dev *cptpf = obj;
+ struct otx2_cpt_bitmap bmap;
+ int i, timeout = 10;
+ int busy, ret;
+ u64 reg = 0;
+
+ bmap = get_cores_bmap(&cptpf->pdev->dev, eng_grp);
+ if (!bmap.size)
+ return -EINVAL;
+
+ /* Detach the cores from group */
+ for_each_set_bit(i, bmap.bits, bmap.size) {
+ ret = otx2_cpt_read_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_EXEX_CTL2(i), ®);
+ if (ret)
+ return ret;
+
+ if (reg & (1ull << eng_grp->idx)) {
+ eng_grp->g->eng_ref_cnt[i]--;
+ reg &= ~(1ull << eng_grp->idx);
+
+ ret = otx2_cpt_write_af_reg(&cptpf->afpf_mbox,
+ cptpf->pdev,
+ CPT_AF_EXEX_CTL2(i), reg);
+ if (ret)
+ return ret;
+ }
+ }
+
+ /* Wait for cores to become idle */
+ do {
+ busy = 0;
+ usleep_range(10000, 20000);
+ if (timeout-- < 0)
+ return -EBUSY;
+
+ for_each_set_bit(i, bmap.bits, bmap.size) {
+ ret = otx2_cpt_read_af_reg(&cptpf->afpf_mbox,
+ cptpf->pdev,
+ CPT_AF_EXEX_STS(i), ®);
+ if (ret)
+ return ret;
+
+ if (reg & 0x1) {
+ busy = 1;
+ break;
+ }
+ }
+ } while (busy);
+
+ /* Disable the cores only if they are not used anymore */
+ for_each_set_bit(i, bmap.bits, bmap.size) {
+ if (!eng_grp->g->eng_ref_cnt[i]) {
+ ret = otx2_cpt_write_af_reg(&cptpf->afpf_mbox,
+ cptpf->pdev,
+ CPT_AF_EXEX_CTL(i), 0x0);
+ if (ret)
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
+static int cpt_attach_and_enable_cores(struct otx2_cpt_eng_grp_info *eng_grp,
+ void *obj)
+{
+ struct otx2_cptpf_dev *cptpf = obj;
+ struct otx2_cpt_bitmap bmap;
+ u64 reg = 0;
+ int i, ret;
+
+ bmap = get_cores_bmap(&cptpf->pdev->dev, eng_grp);
+ if (!bmap.size)
+ return -EINVAL;
+
+ /* Attach the cores to the group */
+ for_each_set_bit(i, bmap.bits, bmap.size) {
+ ret = otx2_cpt_read_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_EXEX_CTL2(i), ®);
+ if (ret)
+ return ret;
+
+ if (!(reg & (1ull << eng_grp->idx))) {
+ eng_grp->g->eng_ref_cnt[i]++;
+ reg |= 1ull << eng_grp->idx;
+
+ ret = otx2_cpt_write_af_reg(&cptpf->afpf_mbox,
+ cptpf->pdev,
+ CPT_AF_EXEX_CTL2(i), reg);
+ if (ret)
+ return ret;
+ }
+ }
+
+ /* Enable the cores */
+ for_each_set_bit(i, bmap.bits, bmap.size) {
+ ret = otx2_cpt_add_write_af_reg(&cptpf->afpf_mbox,
+ cptpf->pdev,
+ CPT_AF_EXEX_CTL(i), 0x1);
+ if (ret)
+ return ret;
+ }
+ ret = otx2_cpt_send_af_reg_requests(&cptpf->afpf_mbox, cptpf->pdev);
+
+ return ret;
+}
+
+static int process_tar_file(struct device *dev,
+ struct tar_arch_info_t *tar_arch, char *filename,
+ const u8 *data, int size)
+{
+ struct tar_ucode_info_t *tar_ucode_info;
+ struct otx2_cpt_ucode_hdr *ucode_hdr;
+ int ucode_type, ucode_size;
+
+ /*
+ * If size is less than microcode header size then don't report
+ * an error because it might not be microcode file, just process
+ * next file from archive
+ */
+ if (size < sizeof(struct otx2_cpt_ucode_hdr))
+ return 0;
+
+ ucode_hdr = (struct otx2_cpt_ucode_hdr *) data;
+ /*
+ * If microcode version can't be found don't report an error
+ * because it might not be microcode file, just process next file
+ */
+ if (get_ucode_type(dev, ucode_hdr, &ucode_type))
+ return 0;
+
+ ucode_size = ntohl(ucode_hdr->code_length) * 2;
+ if (!ucode_size || (size < round_up(ucode_size, 16) +
+ sizeof(struct otx2_cpt_ucode_hdr) + OTX2_CPT_UCODE_SIGN_LEN)) {
+ dev_err(dev, "Ucode %s invalid size\n", filename);
+ return -EINVAL;
+ }
+
+ tar_ucode_info = kzalloc(sizeof(*tar_ucode_info), GFP_KERNEL);
+ if (!tar_ucode_info)
+ return -ENOMEM;
+
+ tar_ucode_info->ucode_ptr = data;
+ set_ucode_filename(&tar_ucode_info->ucode, filename);
+ memcpy(tar_ucode_info->ucode.ver_str, ucode_hdr->ver_str,
+ OTX2_CPT_UCODE_VER_STR_SZ);
+ tar_ucode_info->ucode.ver_num = ucode_hdr->ver_num;
+ tar_ucode_info->ucode.type = ucode_type;
+ tar_ucode_info->ucode.size = ucode_size;
+ list_add_tail(&tar_ucode_info->list, &tar_arch->ucodes);
+
+ return 0;
+}
+
+static void release_tar_archive(struct tar_arch_info_t *tar_arch)
+{
+ struct tar_ucode_info_t *curr, *temp;
+
+ if (!tar_arch)
+ return;
+
+ list_for_each_entry_safe(curr, temp, &tar_arch->ucodes, list) {
+ list_del(&curr->list);
+ kfree(curr);
+ }
+
+ if (tar_arch->fw)
+ release_firmware(tar_arch->fw);
+ kfree(tar_arch);
+}
+
+static struct tar_ucode_info_t *get_uc_from_tar_archive(
+ struct tar_arch_info_t *tar_arch,
+ int ucode_type)
+{
+ struct tar_ucode_info_t *curr;
+
+ list_for_each_entry(curr, &tar_arch->ucodes, list) {
+ if (!is_eng_type(curr->ucode.type, ucode_type))
+ continue;
+
+ return curr;
+ }
+ return NULL;
+}
+
+static void print_tar_dbg_info(struct tar_arch_info_t *tar_arch,
+ char *tar_filename)
+{
+ struct tar_ucode_info_t *curr;
+
+ pr_debug("Tar archive filename %s\n", tar_filename);
+ pr_debug("Tar archive pointer %p, size %ld\n", tar_arch->fw->data,
+ tar_arch->fw->size);
+ list_for_each_entry(curr, &tar_arch->ucodes, list) {
+ pr_debug("Ucode filename %s\n", curr->ucode.filename);
+ pr_debug("Ucode version string %s\n", curr->ucode.ver_str);
+ pr_debug("Ucode version %d.%d.%d.%d\n",
+ curr->ucode.ver_num.nn, curr->ucode.ver_num.xx,
+ curr->ucode.ver_num.yy, curr->ucode.ver_num.zz);
+ pr_debug("Ucode type (%d) %s\n", curr->ucode.type,
+ get_ucode_type_str(curr->ucode.type));
+ pr_debug("Ucode size %d\n", curr->ucode.size);
+ pr_debug("Ucode ptr %p\n", curr->ucode_ptr);
+ }
+}
+
+static struct tar_arch_info_t *load_tar_archive(struct device *dev,
+ char *tar_filename)
+{
+ struct tar_arch_info_t *tar_arch = NULL;
+ struct tar_blk_t *tar_blk;
+ unsigned int cur_size;
+ size_t tar_offs = 0;
+ size_t tar_size;
+ int ret;
+
+ tar_arch = kzalloc(sizeof(struct tar_arch_info_t), GFP_KERNEL);
+ if (!tar_arch)
+ return NULL;
+
+ INIT_LIST_HEAD(&tar_arch->ucodes);
+
+ /* Load tar archive */
+ ret = request_firmware(&tar_arch->fw, tar_filename, dev);
+ if (ret)
+ goto release_tar_arch;
+
+ if (tar_arch->fw->size < TAR_BLOCK_LEN) {
+ dev_err(dev, "Invalid tar archive %s\n", tar_filename);
+ goto release_tar_arch;
+ }
+
+ tar_size = tar_arch->fw->size;
+ tar_blk = (struct tar_blk_t *) tar_arch->fw->data;
+ if (strncmp(tar_blk->hdr.magic, TAR_MAGIC, TAR_MAGIC_LEN - 1)) {
+ dev_err(dev, "Unsupported format of tar archive %s\n",
+ tar_filename);
+ goto release_tar_arch;
+ }
+
+ while (1) {
+ /* Read current file size */
+ ret = kstrtouint(tar_blk->hdr.size, 8, &cur_size);
+ if (ret)
+ goto release_tar_arch;
+
+ if (tar_offs + cur_size > tar_size ||
+ tar_offs + 2 * TAR_BLOCK_LEN > tar_size) {
+ dev_err(dev, "Invalid tar archive %s\n", tar_filename);
+ goto release_tar_arch;
+ }
+
+ tar_offs += TAR_BLOCK_LEN;
+ if (tar_blk->hdr.typeflag == REGTYPE ||
+ tar_blk->hdr.typeflag == AREGTYPE) {
+ ret = process_tar_file(dev, tar_arch,
+ tar_blk->hdr.name,
+ &tar_arch->fw->data[tar_offs],
+ cur_size);
+ if (ret)
+ goto release_tar_arch;
+ }
+
+ tar_offs += (cur_size/TAR_BLOCK_LEN) * TAR_BLOCK_LEN;
+ if (cur_size % TAR_BLOCK_LEN)
+ tar_offs += TAR_BLOCK_LEN;
+
+ /* Check for the end of the archive */
+ if (tar_offs + 2 * TAR_BLOCK_LEN > tar_size) {
+ dev_err(dev, "Invalid tar archive %s\n", tar_filename);
+ goto release_tar_arch;
+ }
+
+ if (is_mem_zero(&tar_arch->fw->data[tar_offs],
+ 2 * TAR_BLOCK_LEN))
+ break;
+
+ /* Read next block from tar archive */
+ tar_blk = (struct tar_blk_t *) &tar_arch->fw->data[tar_offs];
+ }
+
+ print_tar_dbg_info(tar_arch, tar_filename);
+ return tar_arch;
+release_tar_arch:
+ release_tar_archive(tar_arch);
+ return NULL;
+}
+
+static struct otx2_cpt_engs_rsvd *find_engines_by_type(
+ struct otx2_cpt_eng_grp_info *eng_grp,
+ int eng_type)
+{
+ int i;
+
+ for (i = 0; i < OTX2_CPT_MAX_ETYPES_PER_GRP; i++) {
+ if (!eng_grp->engs[i].type)
+ continue;
+
+ if (eng_grp->engs[i].type == eng_type)
+ return &eng_grp->engs[i];
+ }
+ return NULL;
+}
+
+static int eng_grp_has_eng_type(struct otx2_cpt_eng_grp_info *eng_grp,
+ int eng_type)
+{
+ struct otx2_cpt_engs_rsvd *engs;
+
+ engs = find_engines_by_type(eng_grp, eng_type);
+
+ return (engs != NULL ? 1 : 0);
+}
+
+static int update_engines_avail_count(struct device *dev,
+ struct otx2_cpt_engs_available *avail,
+ struct otx2_cpt_engs_rsvd *engs, int val)
+{
+ switch (engs->type) {
+ case OTX2_CPT_SE_TYPES:
+ avail->se_cnt += val;
+ break;
+
+ case OTX2_CPT_IE_TYPES:
+ avail->ie_cnt += val;
+ break;
+
+ case OTX2_CPT_AE_TYPES:
+ avail->ae_cnt += val;
+ break;
+
+ default:
+ dev_err(dev, "Invalid engine type %d\n", engs->type);
+ return -EINVAL;
+ }
+ return 0;
+}
+
+static int update_engines_offset(struct device *dev,
+ struct otx2_cpt_engs_available *avail,
+ struct otx2_cpt_engs_rsvd *engs)
+{
+ switch (engs->type) {
+ case OTX2_CPT_SE_TYPES:
+ engs->offset = 0;
+ break;
+
+ case OTX2_CPT_IE_TYPES:
+ engs->offset = avail->max_se_cnt;
+ break;
+
+ case OTX2_CPT_AE_TYPES:
+ engs->offset = avail->max_se_cnt + avail->max_ie_cnt;
+ break;
+
+ default:
+ dev_err(dev, "Invalid engine type %d\n", engs->type);
+ return -EINVAL;
+ }
+ return 0;
+}
+
+static int release_engines(struct device *dev,
+ struct otx2_cpt_eng_grp_info *grp)
+{
+ int i, ret = 0;
+
+ for (i = 0; i < OTX2_CPT_MAX_ETYPES_PER_GRP; i++) {
+ if (!grp->engs[i].type)
+ continue;
+
+ if (grp->engs[i].count > 0) {
+ ret = update_engines_avail_count(dev, &grp->g->avail,
+ &grp->engs[i],
+ grp->engs[i].count);
+ if (ret)
+ return ret;
+ }
+
+ grp->engs[i].type = 0;
+ grp->engs[i].count = 0;
+ grp->engs[i].offset = 0;
+ grp->engs[i].ucode = NULL;
+ bitmap_zero(grp->engs[i].bmap, grp->g->engs_num);
+ }
+ return 0;
+}
+
+static int do_reserve_engines(struct device *dev,
+ struct otx2_cpt_eng_grp_info *grp,
+ struct otx2_cpt_engines *req_engs)
+{
+ struct otx2_cpt_engs_rsvd *engs = NULL;
+ int i, ret;
+
+ for (i = 0; i < OTX2_CPT_MAX_ETYPES_PER_GRP; i++) {
+ if (!grp->engs[i].type) {
+ engs = &grp->engs[i];
+ break;
+ }
+ }
+
+ if (!engs)
+ return -ENOMEM;
+
+ engs->type = req_engs->type;
+ engs->count = req_engs->count;
+
+ ret = update_engines_offset(dev, &grp->g->avail, engs);
+ if (ret)
+ return ret;
+
+ if (engs->count > 0) {
+ ret = update_engines_avail_count(dev, &grp->g->avail, engs,
+ -engs->count);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+static int check_engines_availability(struct device *dev,
+ struct otx2_cpt_eng_grp_info *grp,
+ struct otx2_cpt_engines *req_eng)
+{
+ int avail_cnt = 0;
+
+ switch (req_eng->type) {
+ case OTX2_CPT_SE_TYPES:
+ avail_cnt = grp->g->avail.se_cnt;
+ break;
+
+ case OTX2_CPT_IE_TYPES:
+ avail_cnt = grp->g->avail.ie_cnt;
+ break;
+
+ case OTX2_CPT_AE_TYPES:
+ avail_cnt = grp->g->avail.ae_cnt;
+ break;
+
+ default:
+ dev_err(dev, "Invalid engine type %d\n", req_eng->type);
+ return -EINVAL;
+ }
+
+ if (avail_cnt < req_eng->count) {
+ dev_err(dev,
+ "Error available %s engines %d < than requested %d\n",
+ get_eng_type_str(req_eng->type),
+ avail_cnt, req_eng->count);
+ return -EBUSY;
+ }
+ return 0;
+}
+
+static int reserve_engines(struct device *dev,
+ struct otx2_cpt_eng_grp_info *grp,
+ struct otx2_cpt_engines *req_engs, int ucodes_cnt)
+{
+ int i, ret = 0;
+
+ /* Validate if a number of requested engines are available */
+ for (i = 0; i < ucodes_cnt; i++) {
+ ret = check_engines_availability(dev, grp, &req_engs[i]);
+ if (ret)
+ return ret;
+ }
+
+ /* Reserve requested engines for this engine group */
+ for (i = 0; i < ucodes_cnt; i++) {
+ ret = do_reserve_engines(dev, grp, &req_engs[i]);
+ if (ret)
+ return ret;
+ }
+ return 0;
+}
+
+static void ucode_unload(struct device *dev, struct otx2_cpt_ucode *ucode)
+{
+ if (ucode->va) {
+ dma_free_coherent(dev, ucode->size, ucode->va, ucode->dma);
+ ucode->va = NULL;
+ ucode->dma = 0;
+ ucode->size = 0;
+ }
+
+ memset(&ucode->ver_str, 0, OTX2_CPT_UCODE_VER_STR_SZ);
+ memset(&ucode->ver_num, 0, sizeof(struct otx2_cpt_ucode_ver_num));
+ set_ucode_filename(ucode, "");
+ ucode->type = 0;
+}
+
+static int copy_ucode_to_dma_mem(struct device *dev,
+ struct otx2_cpt_ucode *ucode,
+ const u8 *ucode_data)
+{
+ u32 i;
+
+ /* Allocate DMAable space */
+ ucode->va = dma_alloc_coherent(dev, ucode->size, &ucode->dma,
+ GFP_KERNEL);
+ if (!ucode->va)
+ return -ENOMEM;
+
+ memcpy(ucode->va, ucode_data + sizeof(struct otx2_cpt_ucode_hdr),
+ ucode->size);
+
+ /* Byte swap 64-bit */
+ for (i = 0; i < (ucode->size / 8); i++)
+ cpu_to_be64s(&((u64 *)ucode->va)[i]);
+ /* Ucode needs 16-bit swap */
+ for (i = 0; i < (ucode->size / 2); i++)
+ cpu_to_be16s(&((u16 *)ucode->va)[i]);
+ return 0;
+}
+
+static int enable_eng_grp(struct otx2_cpt_eng_grp_info *eng_grp,
+ void *obj)
+{
+ int ret;
+
+ /* Point microcode to each core of the group */
+ ret = cpt_set_ucode_base(eng_grp, obj);
+ if (ret)
+ return ret;
+
+ /* Attach the cores to the group and enable them */
+ ret = cpt_attach_and_enable_cores(eng_grp, obj);
+
+ return ret;
+}
+
+static int disable_eng_grp(struct device *dev,
+ struct otx2_cpt_eng_grp_info *eng_grp,
+ void *obj)
+{
+ int i, ret;
+
+ /* Disable all engines used by this group */
+ ret = cpt_detach_and_disable_cores(eng_grp, obj);
+ if (ret)
+ return ret;
+
+ /* Unload ucode used by this engine group */
+ ucode_unload(dev, &eng_grp->ucode[0]);
+ ucode_unload(dev, &eng_grp->ucode[1]);
+
+ for (i = 0; i < OTX2_CPT_MAX_ETYPES_PER_GRP; i++) {
+ if (!eng_grp->engs[i].type)
+ continue;
+
+ eng_grp->engs[i].ucode = &eng_grp->ucode[0];
+ }
+
+ /* Clear UCODE_BASE register for each engine used by this group */
+ ret = cpt_set_ucode_base(eng_grp, obj);
+
+ return ret;
+}
+
+static void setup_eng_grp_mirroring(struct otx2_cpt_eng_grp_info *dst_grp,
+ struct otx2_cpt_eng_grp_info *src_grp)
+{
+ /* Setup fields for engine group which is mirrored */
+ src_grp->mirror.is_ena = false;
+ src_grp->mirror.idx = 0;
+ src_grp->mirror.ref_count++;
+
+ /* Setup fields for mirroring engine group */
+ dst_grp->mirror.is_ena = true;
+ dst_grp->mirror.idx = src_grp->idx;
+ dst_grp->mirror.ref_count = 0;
+}
+
+static void remove_eng_grp_mirroring(struct otx2_cpt_eng_grp_info *dst_grp)
+{
+ struct otx2_cpt_eng_grp_info *src_grp;
+
+ if (!dst_grp->mirror.is_ena)
+ return;
+
+ src_grp = &dst_grp->g->grp[dst_grp->mirror.idx];
+
+ src_grp->mirror.ref_count--;
+ dst_grp->mirror.is_ena = false;
+ dst_grp->mirror.idx = 0;
+ dst_grp->mirror.ref_count = 0;
+}
+
+static void update_requested_engs(struct otx2_cpt_eng_grp_info *mirror_eng_grp,
+ struct otx2_cpt_engines *engs, int engs_cnt)
+{
+ struct otx2_cpt_engs_rsvd *mirrored_engs;
+ int i;
+
+ for (i = 0; i < engs_cnt; i++) {
+ mirrored_engs = find_engines_by_type(mirror_eng_grp,
+ engs[i].type);
+ if (!mirrored_engs)
+ continue;
+
+ /*
+ * If mirrored group has this type of engines attached then
+ * there are 3 scenarios possible:
+ * 1) mirrored_engs.count == engs[i].count then all engines
+ * from mirrored engine group will be shared with this engine
+ * group
+ * 2) mirrored_engs.count > engs[i].count then only a subset of
+ * engines from mirrored engine group will be shared with this
+ * engine group
+ * 3) mirrored_engs.count < engs[i].count then all engines
+ * from mirrored engine group will be shared with this group
+ * and additional engines will be reserved for exclusively use
+ * by this engine group
+ */
+ engs[i].count -= mirrored_engs->count;
+ }
+}
+
+static struct otx2_cpt_eng_grp_info *find_mirrored_eng_grp(
+ struct otx2_cpt_eng_grp_info *grp)
+{
+ struct otx2_cpt_eng_grps *eng_grps = grp->g;
+ int i;
+
+ for (i = 0; i < OTX2_CPT_MAX_ENGINE_GROUPS; i++) {
+ if (!eng_grps->grp[i].is_enabled)
+ continue;
+ if (eng_grps->grp[i].ucode[0].type &&
+ eng_grps->grp[i].ucode[1].type)
+ continue;
+ if (grp->idx == i)
+ continue;
+ if (!strncasecmp(eng_grps->grp[i].ucode[0].ver_str,
+ grp->ucode[0].ver_str,
+ OTX2_CPT_UCODE_VER_STR_SZ))
+ return &eng_grps->grp[i];
+ }
+
+ return NULL;
+}
+
+static struct otx2_cpt_eng_grp_info *find_unused_eng_grp(
+ struct otx2_cpt_eng_grps *eng_grps)
+{
+ int i;
+
+ for (i = 0; i < OTX2_CPT_MAX_ENGINE_GROUPS; i++) {
+ if (!eng_grps->grp[i].is_enabled)
+ return &eng_grps->grp[i];
+ }
+ return NULL;
+}
+
+static int eng_grp_update_masks(struct device *dev,
+ struct otx2_cpt_eng_grp_info *eng_grp)
+{
+ struct otx2_cpt_engs_rsvd *engs, *mirrored_engs;
+ struct otx2_cpt_bitmap tmp_bmap = { {0} };
+ int i, j, cnt, max_cnt;
+ int bit;
+
+ for (i = 0; i < OTX2_CPT_MAX_ETYPES_PER_GRP; i++) {
+ engs = &eng_grp->engs[i];
+ if (!engs->type)
+ continue;
+ if (engs->count <= 0)
+ continue;
+
+ switch (engs->type) {
+ case OTX2_CPT_SE_TYPES:
+ max_cnt = eng_grp->g->avail.max_se_cnt;
+ break;
+
+ case OTX2_CPT_IE_TYPES:
+ max_cnt = eng_grp->g->avail.max_ie_cnt;
+ break;
+
+ case OTX2_CPT_AE_TYPES:
+ max_cnt = eng_grp->g->avail.max_ae_cnt;
+ break;
+
+ default:
+ dev_err(dev, "Invalid engine type %d\n", engs->type);
+ return -EINVAL;
+ }
+
+ cnt = engs->count;
+ WARN_ON(engs->offset + max_cnt > OTX2_CPT_MAX_ENGINES);
+ bitmap_zero(tmp_bmap.bits, eng_grp->g->engs_num);
+ for (j = engs->offset; j < engs->offset + max_cnt; j++) {
+ if (!eng_grp->g->eng_ref_cnt[j]) {
+ bitmap_set(tmp_bmap.bits, j, 1);
+ cnt--;
+ if (!cnt)
+ break;
+ }
+ }
+
+ if (cnt)
+ return -ENOSPC;
+
+ bitmap_copy(engs->bmap, tmp_bmap.bits, eng_grp->g->engs_num);
+ }
+
+ if (!eng_grp->mirror.is_ena)
+ return 0;
+
+ for (i = 0; i < OTX2_CPT_MAX_ETYPES_PER_GRP; i++) {
+ engs = &eng_grp->engs[i];
+ if (!engs->type)
+ continue;
+
+ mirrored_engs = find_engines_by_type(
+ &eng_grp->g->grp[eng_grp->mirror.idx],
+ engs->type);
+ WARN_ON(!mirrored_engs && engs->count <= 0);
+ if (!mirrored_engs)
+ continue;
+
+ bitmap_copy(tmp_bmap.bits, mirrored_engs->bmap,
+ eng_grp->g->engs_num);
+ if (engs->count < 0) {
+ bit = find_first_bit(mirrored_engs->bmap,
+ eng_grp->g->engs_num);
+ bitmap_clear(tmp_bmap.bits, bit, -engs->count);
+ }
+ bitmap_or(engs->bmap, engs->bmap, tmp_bmap.bits,
+ eng_grp->g->engs_num);
+ }
+ return 0;
+}
+
+static int delete_engine_group(struct device *dev,
+ struct otx2_cpt_eng_grp_info *eng_grp)
+{
+ int ret;
+
+ if (!eng_grp->is_enabled)
+ return 0;
+
+ if (eng_grp->mirror.ref_count)
+ return -EINVAL;
+
+ /* Removing engine group mirroring if enabled */
+ remove_eng_grp_mirroring(eng_grp);
+
+ /* Disable engine group */
+ ret = disable_eng_grp(dev, eng_grp, eng_grp->g->obj);
+ if (ret)
+ return ret;
+
+ /* Release all engines held by this engine group */
+ ret = release_engines(dev, eng_grp);
+ if (ret)
+ return ret;
+
+ eng_grp->is_enabled = false;
+
+ return 0;
+}
+
+static void update_ucode_ptrs(struct otx2_cpt_eng_grp_info *eng_grp)
+{
+ struct otx2_cpt_ucode *ucode;
+
+ if (eng_grp->mirror.is_ena)
+ ucode = &eng_grp->g->grp[eng_grp->mirror.idx].ucode[0];
+ else
+ ucode = &eng_grp->ucode[0];
+ WARN_ON(!eng_grp->engs[0].type);
+ eng_grp->engs[0].ucode = ucode;
+
+ if (eng_grp->engs[1].type) {
+ if (is_2nd_ucode_used(eng_grp))
+ eng_grp->engs[1].ucode = &eng_grp->ucode[1];
+ else
+ eng_grp->engs[1].ucode = ucode;
+ }
+}
+
+static int create_engine_group(struct device *dev,
+ struct otx2_cpt_eng_grps *eng_grps,
+ struct otx2_cpt_engines *engs, int ucodes_cnt,
+ void *ucode_data[], int is_print)
+{
+ struct otx2_cpt_eng_grp_info *mirrored_eng_grp;
+ struct tar_ucode_info_t *tar_ucode_info;
+ struct otx2_cpt_eng_grp_info *eng_grp;
+ int i, ret = 0;
+
+ /* Find engine group which is not used */
+ eng_grp = find_unused_eng_grp(eng_grps);
+ if (!eng_grp) {
+ dev_err(dev, "Error all engine groups are being used\n");
+ return -ENOSPC;
+ }
+ /* Load ucode */
+ for (i = 0; i < ucodes_cnt; i++) {
+ tar_ucode_info = (struct tar_ucode_info_t *) ucode_data[i];
+ eng_grp->ucode[i] = tar_ucode_info->ucode;
+ ret = copy_ucode_to_dma_mem(dev, &eng_grp->ucode[i],
+ tar_ucode_info->ucode_ptr);
+ if (ret)
+ goto unload_ucode;
+ }
+
+ /* Check if this group mirrors another existing engine group */
+ mirrored_eng_grp = find_mirrored_eng_grp(eng_grp);
+ if (mirrored_eng_grp) {
+ /* Setup mirroring */
+ setup_eng_grp_mirroring(eng_grp, mirrored_eng_grp);
+
+ /*
+ * Update count of requested engines because some
+ * of them might be shared with mirrored group
+ */
+ update_requested_engs(mirrored_eng_grp, engs, ucodes_cnt);
+ }
+ ret = reserve_engines(dev, eng_grp, engs, ucodes_cnt);
+ if (ret)
+ goto unload_ucode;
+
+ /* Update ucode pointers used by engines */
+ update_ucode_ptrs(eng_grp);
+
+ /* Update engine masks used by this group */
+ ret = eng_grp_update_masks(dev, eng_grp);
+ if (ret)
+ goto release_engs;
+
+ /* Enable engine group */
+ ret = enable_eng_grp(eng_grp, eng_grps->obj);
+ if (ret)
+ goto release_engs;
+
+ /*
+ * If this engine group mirrors another engine group
+ * then we need to unload ucode as we will use ucode
+ * from mirrored engine group
+ */
+ if (eng_grp->mirror.is_ena)
+ ucode_unload(dev, &eng_grp->ucode[0]);
+
+ eng_grp->is_enabled = true;
+
+ if (!is_print)
+ return 0;
+
+ if (mirrored_eng_grp)
+ dev_info(dev,
+ "Engine_group%d: reuse microcode %s from group %d\n",
+ eng_grp->idx, mirrored_eng_grp->ucode[0].ver_str,
+ mirrored_eng_grp->idx);
+ else
+ dev_info(dev, "Engine_group%d: microcode loaded %s\n",
+ eng_grp->idx, eng_grp->ucode[0].ver_str);
+ if (is_2nd_ucode_used(eng_grp))
+ dev_info(dev, "Engine_group%d: microcode loaded %s\n",
+ eng_grp->idx, eng_grp->ucode[1].ver_str);
+
+ return 0;
+
+release_engs:
+ release_engines(dev, eng_grp);
+unload_ucode:
+ ucode_unload(dev, &eng_grp->ucode[0]);
+ ucode_unload(dev, &eng_grp->ucode[1]);
+ return ret;
+}
+
+static void delete_engine_grps(struct pci_dev *pdev,
+ struct otx2_cpt_eng_grps *eng_grps)
+{
+ int i;
+
+ /* First delete all mirroring engine groups */
+ for (i = 0; i < OTX2_CPT_MAX_ENGINE_GROUPS; i++)
+ if (eng_grps->grp[i].mirror.is_ena)
+ delete_engine_group(&pdev->dev, &eng_grps->grp[i]);
+
+ /* Delete remaining engine groups */
+ for (i = 0; i < OTX2_CPT_MAX_ENGINE_GROUPS; i++)
+ delete_engine_group(&pdev->dev, &eng_grps->grp[i]);
+}
+
+int otx2_cpt_get_eng_grp(struct otx2_cpt_eng_grps *eng_grps, int eng_type)
+{
+
+ int eng_grp_num = OTX2_CPT_INVALID_CRYPTO_ENG_GRP;
+ struct otx2_cpt_eng_grp_info *grp;
+ int i;
+
+ for (i = 0; i < OTX2_CPT_MAX_ENGINE_GROUPS; i++) {
+ grp = &eng_grps->grp[i];
+ if (!grp->is_enabled)
+ continue;
+
+ if (eng_type == OTX2_CPT_SE_TYPES) {
+ if (eng_grp_has_eng_type(grp, eng_type) &&
+ !eng_grp_has_eng_type(grp, OTX2_CPT_IE_TYPES)) {
+ eng_grp_num = i;
+ break;
+ }
+ } else {
+ if (eng_grp_has_eng_type(grp, eng_type)) {
+ eng_grp_num = i;
+ break;
+ }
+ }
+ }
+ return eng_grp_num;
+}
+
+int otx2_cpt_create_eng_grps(struct pci_dev *pdev,
+ struct otx2_cpt_eng_grps *eng_grps)
+{
+ struct tar_ucode_info_t *tar_info[OTX2_CPT_MAX_ETYPES_PER_GRP] = { };
+ struct otx2_cpt_engines engs[OTX2_CPT_MAX_ETYPES_PER_GRP] = { {0} };
+ struct tar_arch_info_t *tar_arch = NULL;
+ char tar_filename[OTX2_CPT_NAME_LENGTH];
+ int ret;
+
+ /*
+ * We don't create engine groups if it was already
+ * made (when user enabled VFs for the first time)
+ */
+ if (eng_grps->is_grps_created)
+ return 0;
+
+ sprintf(tar_filename, "cpt%02d-mc.tar", pdev->revision);
+
+ tar_arch = load_tar_archive(&pdev->dev, tar_filename);
+ if (!tar_arch)
+ return -EINVAL;
+
+ /*
+ * Create engine group with SE engines for kernel
+ * crypto functionality (symmetric crypto)
+ */
+ tar_info[0] = get_uc_from_tar_archive(tar_arch, OTX2_CPT_SE_TYPES);
+ if (tar_info[0] == NULL) {
+ dev_err(&pdev->dev, "Unable to find SE firmware in %s\n",
+ tar_filename);
+ ret = -EINVAL;
+ goto release_tar_arch;
+ }
+ engs[0].type = OTX2_CPT_SE_TYPES;
+ engs[0].count = eng_grps->avail.max_se_cnt;
+
+ ret = create_engine_group(&pdev->dev, eng_grps, engs, 1,
+ (void **) tar_info, 1);
+ if (ret)
+ goto release_tar_arch;
+
+ /*
+ * Create engine group with SE+IE engines for IPSec.
+ * All SE engines will be shared with engine group 0.
+ */
+ tar_info[0] = get_uc_from_tar_archive(tar_arch, OTX2_CPT_SE_TYPES);
+ tar_info[1] = get_uc_from_tar_archive(tar_arch, OTX2_CPT_IE_TYPES);
+
+ if (tar_info[1] == NULL) {
+ dev_err(&pdev->dev, "Unable to find IE firmware in %s\n",
+ tar_filename);
+ ret = -EINVAL;
+ goto delete_eng_grp;
+ }
+ engs[0].type = OTX2_CPT_SE_TYPES;
+ engs[0].count = eng_grps->avail.max_se_cnt;
+ engs[1].type = OTX2_CPT_IE_TYPES;
+ engs[1].count = eng_grps->avail.max_ie_cnt;
+
+ ret = create_engine_group(&pdev->dev, eng_grps, engs, 2,
+ (void **) tar_info, 1);
+ if (ret)
+ goto delete_eng_grp;
+
+ /*
+ * Create engine group with AE engines for asymmetric
+ * crypto functionality.
+ */
+ tar_info[0] = get_uc_from_tar_archive(tar_arch, OTX2_CPT_AE_TYPES);
+ if (tar_info[0] == NULL) {
+ dev_err(&pdev->dev, "Unable to find AE firmware in %s\n",
+ tar_filename);
+ ret = -EINVAL;
+ goto delete_eng_grp;
+ }
+ engs[0].type = OTX2_CPT_AE_TYPES;
+ engs[0].count = eng_grps->avail.max_ae_cnt;
+
+ ret = create_engine_group(&pdev->dev, eng_grps, engs, 1,
+ (void **) tar_info, 1);
+ if (ret)
+ goto delete_eng_grp;
+
+ eng_grps->is_grps_created = true;
+
+ release_tar_archive(tar_arch);
+ return 0;
+
+delete_eng_grp:
+ delete_engine_grps(pdev, eng_grps);
+release_tar_arch:
+ release_tar_archive(tar_arch);
+ return ret;
+}
+
+int otx2_cpt_disable_all_cores(struct otx2_cptpf_dev *cptpf)
+{
+ int i, ret, busy, total_cores;
+ int timeout = 10;
+ u64 reg = 0;
+
+ total_cores = cptpf->eng_grps.avail.max_se_cnt +
+ cptpf->eng_grps.avail.max_ie_cnt +
+ cptpf->eng_grps.avail.max_ae_cnt;
+
+ /* Disengage the cores from groups */
+ for (i = 0; i < total_cores; i++) {
+ ret = otx2_cpt_add_write_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_EXEX_CTL2(i), 0x0);
+ if (ret)
+ return ret;
+
+ cptpf->eng_grps.eng_ref_cnt[i] = 0;
+ }
+ ret = otx2_cpt_send_af_reg_requests(&cptpf->afpf_mbox, cptpf->pdev);
+ if (ret)
+ return ret;
+
+ /* Wait for cores to become idle */
+ do {
+ busy = 0;
+ usleep_range(10000, 20000);
+ if (timeout-- < 0)
+ return -EBUSY;
+
+ for (i = 0; i < total_cores; i++) {
+ ret = otx2_cpt_read_af_reg(&cptpf->afpf_mbox,
+ cptpf->pdev,
+ CPT_AF_EXEX_STS(i), ®);
+ if (ret)
+ return ret;
+
+ if (reg & 0x1) {
+ busy = 1;
+ break;
+ }
+ }
+ } while (busy);
+
+ /* Disable the cores */
+ for (i = 0; i < total_cores; i++) {
+ ret = otx2_cpt_add_write_af_reg(&cptpf->afpf_mbox, cptpf->pdev,
+ CPT_AF_EXEX_CTL(i), 0x0);
+ if (ret)
+ return ret;
+ }
+ return otx2_cpt_send_af_reg_requests(&cptpf->afpf_mbox, cptpf->pdev);
+}
+
+void otx2_cpt_cleanup_eng_grps(struct pci_dev *pdev,
+ struct otx2_cpt_eng_grps *eng_grps)
+{
+ struct otx2_cpt_eng_grp_info *grp;
+ int i, j;
+
+ delete_engine_grps(pdev, eng_grps);
+ /* Release memory */
+ for (i = 0; i < OTX2_CPT_MAX_ENGINE_GROUPS; i++) {
+ grp = &eng_grps->grp[i];
+ for (j = 0; j < OTX2_CPT_MAX_ETYPES_PER_GRP; j++) {
+ kfree(grp->engs[j].bmap);
+ grp->engs[j].bmap = NULL;
+ }
+ }
+}
+
+int otx2_cpt_init_eng_grps(struct pci_dev *pdev,
+ struct otx2_cpt_eng_grps *eng_grps)
+{
+ struct otx2_cpt_eng_grp_info *grp;
+ int i, j, ret;
+
+ eng_grps->obj = pci_get_drvdata(pdev);
+ eng_grps->avail.se_cnt = eng_grps->avail.max_se_cnt;
+ eng_grps->avail.ie_cnt = eng_grps->avail.max_ie_cnt;
+ eng_grps->avail.ae_cnt = eng_grps->avail.max_ae_cnt;
+
+ eng_grps->engs_num = eng_grps->avail.max_se_cnt +
+ eng_grps->avail.max_ie_cnt +
+ eng_grps->avail.max_ae_cnt;
+ if (eng_grps->engs_num > OTX2_CPT_MAX_ENGINES) {
+ dev_err(&pdev->dev,
+ "Number of engines %d > than max supported %d\n",
+ eng_grps->engs_num, OTX2_CPT_MAX_ENGINES);
+ ret = -EINVAL;
+ goto cleanup_eng_grps;
+ }
+
+ for (i = 0; i < OTX2_CPT_MAX_ENGINE_GROUPS; i++) {
+ grp = &eng_grps->grp[i];
+ grp->g = eng_grps;
+ grp->idx = i;
+
+ for (j = 0; j < OTX2_CPT_MAX_ETYPES_PER_GRP; j++) {
+ grp->engs[j].bmap =
+ kcalloc(BITS_TO_LONGS(eng_grps->engs_num),
+ sizeof(long), GFP_KERNEL);
+ if (!grp->engs[j].bmap) {
+ ret = -ENOMEM;
+ goto cleanup_eng_grps;
+ }
+ }
+ }
+ return 0;
+cleanup_eng_grps:
+ otx2_cpt_cleanup_eng_grps(pdev, eng_grps);
+ return ret;
+}
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.h b/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.h
new file mode 100644
index 000000000000..ab7a1461b095
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.h
@@ -0,0 +1,161 @@
+/* SPDX-License-Identifier: GPL-2.0-only
+ * Copyright (C) 2020 Marvell.
+ */
+
+#ifndef __OTX2_CPTPF_UCODE_H
+#define __OTX2_CPTPF_UCODE_H
+
+#include <linux/pci.h>
+#include <linux/types.h>
+#include <linux/module.h>
+#include "otx2_cpt_hw_types.h"
+#include "otx2_cpt_common.h"
+
+/*
+ * On OcteonTX2 platform IPSec ucode can use both IE and SE engines therefore
+ * IE and SE engines can be attached to the same engine group.
+ */
+#define OTX2_CPT_MAX_ETYPES_PER_GRP 2
+
+/* CPT ucode signature size */
+#define OTX2_CPT_UCODE_SIGN_LEN 256
+
+/* Microcode version string length */
+#define OTX2_CPT_UCODE_VER_STR_SZ 44
+
+/* Maximum number of supported engines/cores on OcteonTX2 platform */
+#define OTX2_CPT_MAX_ENGINES 128
+
+#define OTX2_CPT_ENGS_BITMASK_LEN BITS_TO_LONGS(OTX2_CPT_MAX_ENGINES)
+
+/* Microcode types */
+enum otx2_cpt_ucode_type {
+ OTX2_CPT_AE_UC_TYPE = 1, /* AE-MAIN */
+ OTX2_CPT_SE_UC_TYPE1 = 20,/* SE-MAIN - combination of 21 and 22 */
+ OTX2_CPT_SE_UC_TYPE2 = 21,/* Fast Path IPSec + AirCrypto */
+ OTX2_CPT_SE_UC_TYPE3 = 22,/*
+ * Hash + HMAC + FlexiCrypto + RNG +
+ * Full Feature IPSec + AirCrypto + Kasumi
+ */
+ OTX2_CPT_IE_UC_TYPE1 = 30, /* IE-MAIN - combination of 31 and 32 */
+ OTX2_CPT_IE_UC_TYPE2 = 31, /* Fast Path IPSec */
+ OTX2_CPT_IE_UC_TYPE3 = 32, /*
+ * Hash + HMAC + FlexiCrypto + RNG +
+ * Full Future IPSec
+ */
+};
+
+struct otx2_cpt_bitmap {
+ unsigned long bits[OTX2_CPT_ENGS_BITMASK_LEN];
+ int size;
+};
+
+struct otx2_cpt_engines {
+ int type;
+ int count;
+};
+
+/* Microcode version number */
+struct otx2_cpt_ucode_ver_num {
+ u8 nn;
+ u8 xx;
+ u8 yy;
+ u8 zz;
+};
+
+struct otx2_cpt_ucode_hdr {
+ struct otx2_cpt_ucode_ver_num ver_num;
+ u8 ver_str[OTX2_CPT_UCODE_VER_STR_SZ];
+ __be32 code_length;
+ u32 padding[3];
+};
+
+struct otx2_cpt_ucode {
+ u8 ver_str[OTX2_CPT_UCODE_VER_STR_SZ];/*
+ * ucode version in readable
+ * format
+ */
+ struct otx2_cpt_ucode_ver_num ver_num;/* ucode version number */
+ char filename[OTX2_CPT_NAME_LENGTH];/* ucode filename */
+ dma_addr_t dma; /* phys address of ucode image */
+ void *va; /* virt address of ucode image */
+ u32 size; /* ucode image size */
+ int type; /* ucode image type SE, IE, AE or SE+IE */
+};
+
+struct tar_ucode_info_t {
+ struct list_head list;
+ struct otx2_cpt_ucode ucode;/* microcode information */
+ const u8 *ucode_ptr; /* pointer to microcode in tar archive */
+};
+
+/* Maximum and current number of engines available for all engine groups */
+struct otx2_cpt_engs_available {
+ int max_se_cnt;
+ int max_ie_cnt;
+ int max_ae_cnt;
+ int se_cnt;
+ int ie_cnt;
+ int ae_cnt;
+};
+
+/* Engines reserved to an engine group */
+struct otx2_cpt_engs_rsvd {
+ int type; /* engine type */
+ int count; /* number of engines attached */
+ int offset; /* constant offset of engine type in the bitmap */
+ unsigned long *bmap; /* attached engines bitmap */
+ struct otx2_cpt_ucode *ucode; /* ucode used by these engines */
+};
+
+struct otx2_cpt_mirror_info {
+ int is_ena; /*
+ * is mirroring enabled, it is set only for engine
+ * group which mirrors another engine group
+ */
+ int idx; /*
+ * index of engine group which is mirrored by this
+ * group, set only for engine group which mirrors
+ * another group
+ */
+ int ref_count; /*
+ * number of times this engine group is mirrored by
+ * other groups, this is set only for engine group
+ * which is mirrored by other group(s)
+ */
+};
+
+struct otx2_cpt_eng_grp_info {
+ struct otx2_cpt_eng_grps *g; /* pointer to engine_groups structure */
+ /* engines attached */
+ struct otx2_cpt_engs_rsvd engs[OTX2_CPT_MAX_ETYPES_PER_GRP];
+ /* ucodes information */
+ struct otx2_cpt_ucode ucode[OTX2_CPT_MAX_ETYPES_PER_GRP];
+ /* engine group mirroring information */
+ struct otx2_cpt_mirror_info mirror;
+ int idx; /* engine group index */
+ bool is_enabled; /*
+ * is engine group enabled, engine group is enabled
+ * when it has engines attached and ucode loaded
+ */
+};
+
+struct otx2_cpt_eng_grps {
+ struct otx2_cpt_eng_grp_info grp[OTX2_CPT_MAX_ENGINE_GROUPS];
+ struct otx2_cpt_engs_available avail;
+ void *obj; /* device specific data */
+ int engs_num; /* total number of engines supported */
+ u8 eng_ref_cnt[OTX2_CPT_MAX_ENGINES];/* engines reference count */
+ bool is_grps_created; /* Is the engine groups are already created */
+};
+struct otx2_cptpf_dev;
+int otx2_cpt_init_eng_grps(struct pci_dev *pdev,
+ struct otx2_cpt_eng_grps *eng_grps);
+void otx2_cpt_cleanup_eng_grps(struct pci_dev *pdev,
+ struct otx2_cpt_eng_grps *eng_grps);
+int otx2_cpt_create_eng_grps(struct pci_dev *pdev,
+ struct otx2_cpt_eng_grps *eng_grps);
+int otx2_cpt_disable_all_cores(struct otx2_cptpf_dev *cptpf);
+int otx2_cpt_get_eng_grp(struct otx2_cpt_eng_grps *eng_grps, int eng_type);
+
+#endif /* __OTX2_CPTPF_UCODE_H */
--
2.28.0
CPT RVU Local Functions(LFs) needs to be attached to the
PF/VF to submit the instructions to CPT.
This patch adds the interface to initialize and attach
the LFs. It also adds interface to register the LF's
interrupts.
Signed-off-by: Suheil Chandran <[email protected]>
Signed-off-by: Srujana Challa <[email protected]>
---
drivers/crypto/marvell/octeontx2/Makefile | 2 +-
.../marvell/octeontx2/otx2_cpt_common.h | 4 +
.../marvell/octeontx2/otx2_cpt_mbox_common.c | 56 +++
drivers/crypto/marvell/octeontx2/otx2_cptlf.c | 426 ++++++++++++++++++
drivers/crypto/marvell/octeontx2/otx2_cptlf.h | 282 ++++++++++++
drivers/crypto/marvell/octeontx2/otx2_cptpf.h | 2 +
.../marvell/octeontx2/otx2_cptpf_mbox.c | 8 +
7 files changed, 779 insertions(+), 1 deletion(-)
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptlf.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptlf.h
diff --git a/drivers/crypto/marvell/octeontx2/Makefile b/drivers/crypto/marvell/octeontx2/Makefile
index 3c4155446296..e47a55961bb8 100644
--- a/drivers/crypto/marvell/octeontx2/Makefile
+++ b/drivers/crypto/marvell/octeontx2/Makefile
@@ -2,6 +2,6 @@
obj-$(CONFIG_CRYPTO_DEV_OCTEONTX2_CPT) += octeontx2-cpt.o
octeontx2-cpt-objs := otx2_cptpf_main.o otx2_cptpf_mbox.o \
- otx2_cpt_mbox_common.o otx2_cptpf_ucode.o
+ otx2_cpt_mbox_common.o otx2_cptpf_ucode.o otx2_cptlf.o
ccflags-y += -I$(srctree)/drivers/net/ethernet/marvell/octeontx2/af
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
index ae16dc102459..d5576f5d3b90 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
+++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
@@ -76,4 +76,8 @@ int otx2_cpt_read_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
u64 reg, u64 *val);
int otx2_cpt_write_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
u64 reg, u64 val);
+struct otx2_cptlfs_info;
+int otx2_cpt_attach_rscrs_msg(struct otx2_cptlfs_info *lfs);
+int otx2_cpt_detach_rsrcs_msg(struct otx2_cptlfs_info *lfs);
+
#endif /* __OTX2_CPT_COMMON_H */
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
index ef1291c4881b..0933031ac827 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
+++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
@@ -2,6 +2,7 @@
/* Copyright (C) 2020 Marvell. */
#include "otx2_cpt_common.h"
+#include "otx2_cptlf.h"
int otx2_cpt_send_mbox_msg(struct otx2_mbox *mbox, struct pci_dev *pdev)
{
@@ -112,3 +113,58 @@ int otx2_cpt_write_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
return otx2_cpt_send_mbox_msg(mbox, pdev);
}
+
+int otx2_cpt_attach_rscrs_msg(struct otx2_cptlfs_info *lfs)
+{
+ struct otx2_mbox *mbox = lfs->mbox;
+ struct rsrc_attach *req;
+ int ret;
+
+ req = (struct rsrc_attach *)
+ otx2_mbox_alloc_msg_rsp(mbox, 0, sizeof(*req),
+ sizeof(struct msg_rsp));
+ if (req == NULL) {
+ dev_err(&lfs->pdev->dev, "RVU MBOX failed to get message.\n");
+ return -EFAULT;
+ }
+
+ req->hdr.id = MBOX_MSG_ATTACH_RESOURCES;
+ req->hdr.sig = OTX2_MBOX_REQ_SIG;
+ req->hdr.pcifunc = 0;
+ req->cptlfs = lfs->lfs_num;
+ ret = otx2_cpt_send_mbox_msg(mbox, lfs->pdev);
+ if (ret)
+ return ret;
+
+ if (!lfs->are_lfs_attached)
+ ret = -EINVAL;
+
+ return ret;
+}
+
+int otx2_cpt_detach_rsrcs_msg(struct otx2_cptlfs_info *lfs)
+{
+ struct otx2_mbox *mbox = lfs->mbox;
+ struct rsrc_detach *req;
+ int ret;
+
+ req = (struct rsrc_detach *)
+ otx2_mbox_alloc_msg_rsp(mbox, 0, sizeof(*req),
+ sizeof(struct msg_rsp));
+ if (req == NULL) {
+ dev_err(&lfs->pdev->dev, "RVU MBOX failed to get message.\n");
+ return -EFAULT;
+ }
+
+ req->hdr.id = MBOX_MSG_DETACH_RESOURCES;
+ req->hdr.sig = OTX2_MBOX_REQ_SIG;
+ req->hdr.pcifunc = 0;
+ ret = otx2_cpt_send_mbox_msg(mbox, lfs->pdev);
+ if (ret)
+ return ret;
+
+ if (lfs->are_lfs_attached)
+ ret = -EINVAL;
+
+ return ret;
+}
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptlf.c b/drivers/crypto/marvell/octeontx2/otx2_cptlf.c
new file mode 100644
index 000000000000..4069e4802072
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptlf.c
@@ -0,0 +1,426 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2020 Marvell. */
+
+#include "otx2_cpt_common.h"
+#include "otx2_cptlf.h"
+#include "rvu_reg.h"
+
+#define CPT_TIMER_HOLD 0x03F
+#define CPT_COUNT_HOLD 32
+
+static void cptlf_do_set_done_time_wait(struct otx2_cptlf_info *lf,
+ int time_wait)
+{
+ union otx2_cptx_lf_done_wait done_wait;
+
+ done_wait.u = otx2_cpt_read64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_DONE_WAIT);
+ done_wait.s.time_wait = time_wait;
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_DONE_WAIT, done_wait.u);
+}
+
+static void cptlf_do_set_done_num_wait(struct otx2_cptlf_info *lf, int num_wait)
+{
+ union otx2_cptx_lf_done_wait done_wait;
+
+ done_wait.u = otx2_cpt_read64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_DONE_WAIT);
+ done_wait.s.num_wait = num_wait;
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_DONE_WAIT, done_wait.u);
+}
+
+static void cptlf_set_done_time_wait(struct otx2_cptlfs_info *lfs,
+ int time_wait)
+{
+ int slot;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++)
+ cptlf_do_set_done_time_wait(&lfs->lf[slot], time_wait);
+}
+
+static void cptlf_set_done_num_wait(struct otx2_cptlfs_info *lfs, int num_wait)
+{
+ int slot;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++)
+ cptlf_do_set_done_num_wait(&lfs->lf[slot], num_wait);
+}
+
+static int cptlf_set_pri(struct otx2_cptlf_info *lf, int pri)
+{
+ struct otx2_cptlfs_info *lfs = lf->lfs;
+ union otx2_cptx_af_lf_ctrl lf_ctrl;
+ int ret;
+
+ ret = otx2_cpt_read_af_reg(lfs->mbox, lfs->pdev,
+ CPT_AF_LFX_CTL(lf->slot),
+ &lf_ctrl.u);
+ if (ret)
+ return ret;
+
+ lf_ctrl.s.pri = pri ? 1 : 0;
+
+ ret = otx2_cpt_write_af_reg(lfs->mbox, lfs->pdev,
+ CPT_AF_LFX_CTL(lf->slot),
+ lf_ctrl.u);
+ return ret;
+}
+
+static int cptlf_set_eng_grps_mask(struct otx2_cptlf_info *lf,
+ int eng_grps_mask)
+{
+ struct otx2_cptlfs_info *lfs = lf->lfs;
+ union otx2_cptx_af_lf_ctrl lf_ctrl;
+ int ret;
+
+ ret = otx2_cpt_read_af_reg(lfs->mbox, lfs->pdev,
+ CPT_AF_LFX_CTL(lf->slot),
+ &lf_ctrl.u);
+ if (ret)
+ return ret;
+
+ lf_ctrl.s.grp = eng_grps_mask;
+
+ ret = otx2_cpt_write_af_reg(lfs->mbox, lfs->pdev,
+ CPT_AF_LFX_CTL(lf->slot),
+ lf_ctrl.u);
+ return ret;
+}
+
+static int cptlf_set_grp_and_pri(struct otx2_cptlfs_info *lfs,
+ int eng_grp_mask, int pri)
+{
+ int slot, ret = 0;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++) {
+ ret = cptlf_set_pri(&lfs->lf[slot], pri);
+ if (ret)
+ return ret;
+
+ ret = cptlf_set_eng_grps_mask(&lfs->lf[slot], eng_grp_mask);
+ if (ret)
+ return ret;
+ }
+ return ret;
+}
+
+static void cptlf_hw_init(struct otx2_cptlfs_info *lfs)
+{
+ /* Disable instruction queues */
+ otx2_cptlf_disable_iqueues(lfs);
+
+ /* Set instruction queues base addresses */
+ otx2_cptlf_set_iqueues_base_addr(lfs);
+
+ /* Set instruction queues sizes */
+ otx2_cptlf_set_iqueues_size(lfs);
+
+ /* Set done interrupts time wait */
+ cptlf_set_done_time_wait(lfs, CPT_TIMER_HOLD);
+
+ /* Set done interrupts num wait */
+ cptlf_set_done_num_wait(lfs, CPT_COUNT_HOLD);
+
+ /* Enable instruction queues */
+ otx2_cptlf_enable_iqueues(lfs);
+}
+
+static void cptlf_hw_cleanup(struct otx2_cptlfs_info *lfs)
+{
+ /* Disable instruction queues */
+ otx2_cptlf_disable_iqueues(lfs);
+}
+
+static void cptlf_set_misc_intrs(struct otx2_cptlfs_info *lfs, u8 enable)
+{
+ union otx2_cptx_lf_misc_int_ena_w1s irq_misc = { .u = 0x0 };
+ u64 reg = enable ? OTX2_CPT_LF_MISC_INT_ENA_W1S :
+ OTX2_CPT_LF_MISC_INT_ENA_W1C;
+ int slot;
+
+ irq_misc.s.fault = 0x1;
+ irq_misc.s.hwerr = 0x1;
+ irq_misc.s.irde = 0x1;
+ irq_misc.s.nqerr = 0x1;
+ irq_misc.s.nwrp = 0x1;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++)
+ otx2_cpt_write64(lfs->reg_base, BLKADDR_CPT0, slot, reg,
+ irq_misc.u);
+}
+
+static void cptlf_enable_intrs(struct otx2_cptlfs_info *lfs)
+{
+ int slot;
+
+ /* Enable done interrupts */
+ for (slot = 0; slot < lfs->lfs_num; slot++)
+ otx2_cpt_write64(lfs->reg_base, BLKADDR_CPT0, slot,
+ OTX2_CPT_LF_DONE_INT_ENA_W1S, 0x1);
+ /* Enable Misc interrupts */
+ cptlf_set_misc_intrs(lfs, true);
+}
+
+static void cptlf_disable_intrs(struct otx2_cptlfs_info *lfs)
+{
+ int slot;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++)
+ otx2_cpt_write64(lfs->reg_base, BLKADDR_CPT0, slot,
+ OTX2_CPT_LF_DONE_INT_ENA_W1C, 0x1);
+ cptlf_set_misc_intrs(lfs, false);
+}
+
+static inline int cptlf_read_done_cnt(struct otx2_cptlf_info *lf)
+{
+ union otx2_cptx_lf_done irq_cnt;
+
+ irq_cnt.u = otx2_cpt_read64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_DONE);
+ return irq_cnt.s.done;
+}
+
+static irqreturn_t cptlf_misc_intr_handler(int __always_unused irq, void *arg)
+{
+ union otx2_cptx_lf_misc_int irq_misc, irq_misc_ack;
+ struct otx2_cptlf_info *lf = arg;
+ struct device *dev;
+
+ dev = &lf->lfs->pdev->dev;
+ irq_misc.u = otx2_cpt_read64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_MISC_INT);
+ irq_misc_ack.u = 0x0;
+
+ if (irq_misc.s.fault) {
+ dev_err(dev, "Memory error detected while executing CPT_INST_S, LF %d.\n",
+ lf->slot);
+ irq_misc_ack.s.fault = 0x1;
+
+ } else if (irq_misc.s.hwerr) {
+ dev_err(dev, "HW error from an engine executing CPT_INST_S, LF %d.",
+ lf->slot);
+ irq_misc_ack.s.hwerr = 0x1;
+
+ } else if (irq_misc.s.nwrp) {
+ dev_err(dev, "SMMU fault while writing CPT_RES_S to CPT_INST_S[RES_ADDR], LF %d.\n",
+ lf->slot);
+ irq_misc_ack.s.nwrp = 0x1;
+
+ } else if (irq_misc.s.irde) {
+ dev_err(dev, "Memory error when accessing instruction memory queue CPT_LF_Q_BASE[ADDR].\n");
+ irq_misc_ack.s.irde = 0x1;
+
+ } else if (irq_misc.s.nqerr) {
+ dev_err(dev, "Error enqueuing an instruction received at CPT_LF_NQ.\n");
+ irq_misc_ack.s.nqerr = 0x1;
+
+ } else {
+ dev_err(dev, "Unhandled interrupt in CPT LF %d\n", lf->slot);
+ return IRQ_NONE;
+ }
+
+ /* Acknowledge interrupts */
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_MISC_INT, irq_misc_ack.u);
+
+ return IRQ_HANDLED;
+}
+
+static irqreturn_t cptlf_done_intr_handler(int irq, void *arg)
+{
+ union otx2_cptx_lf_done_wait done_wait;
+ struct otx2_cptlf_info *lf = arg;
+ int irq_cnt;
+
+ /* Read the number of completed requests */
+ irq_cnt = cptlf_read_done_cnt(lf);
+ if (irq_cnt) {
+ done_wait.u = otx2_cpt_read64(lf->lfs->reg_base, BLKADDR_CPT0,
+ lf->slot, OTX2_CPT_LF_DONE_WAIT);
+ /* Acknowledge the number of completed requests */
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_DONE_ACK, irq_cnt);
+
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_DONE_WAIT, done_wait.u);
+ if (unlikely(!lf->wqe)) {
+ dev_err(&lf->lfs->pdev->dev, "No work for LF %d\n",
+ lf->slot);
+ return IRQ_NONE;
+ }
+
+ /* Schedule processing of completed requests */
+ tasklet_hi_schedule(&lf->wqe->work);
+ }
+ return IRQ_HANDLED;
+}
+
+void otx2_cptlf_unregister_interrupts(struct otx2_cptlfs_info *lfs)
+{
+ int i, offs, vector;
+
+ for (i = 0; i < lfs->lfs_num; i++) {
+ for (offs = 0; offs < OTX2_CPT_LF_MSIX_VECTORS; offs++) {
+ if (!lfs->lf[i].is_irq_reg[offs])
+ continue;
+
+ vector = pci_irq_vector(lfs->pdev,
+ lfs->lf[i].msix_offset + offs);
+ free_irq(vector, &lfs->lf[i]);
+ lfs->lf[i].is_irq_reg[offs] = false;
+ }
+ }
+ cptlf_disable_intrs(lfs);
+}
+
+static int cptlf_do_register_interrrupts(struct otx2_cptlfs_info *lfs,
+ int lf_num, int irq_offset,
+ irq_handler_t handler)
+{
+ int ret, vector;
+
+ vector = pci_irq_vector(lfs->pdev, lfs->lf[lf_num].msix_offset +
+ irq_offset);
+ ret = request_irq(vector, handler, 0,
+ lfs->lf[lf_num].irq_name[irq_offset],
+ &lfs->lf[lf_num]);
+ if (ret)
+ return ret;
+
+ lfs->lf[lf_num].is_irq_reg[irq_offset] = true;
+
+ return ret;
+}
+
+int otx2_cptlf_register_interrupts(struct otx2_cptlfs_info *lfs)
+{
+ int irq_offs, ret, i;
+
+ for (i = 0; i < lfs->lfs_num; i++) {
+ irq_offs = OTX2_CPT_LF_INT_VEC_E_MISC;
+ snprintf(lfs->lf[i].irq_name[irq_offs], 32, "CPTLF Misc%d", i);
+ ret = cptlf_do_register_interrrupts(lfs, i, irq_offs,
+ cptlf_misc_intr_handler);
+ if (ret)
+ goto free_irq;
+
+ irq_offs = OTX2_CPT_LF_INT_VEC_E_DONE;
+ snprintf(lfs->lf[i].irq_name[irq_offs], 32, "OTX2_CPTLF Done%d",
+ i);
+ ret = cptlf_do_register_interrrupts(lfs, i, irq_offs,
+ cptlf_done_intr_handler);
+ if (ret)
+ goto free_irq;
+ }
+ cptlf_enable_intrs(lfs);
+ return 0;
+free_irq:
+ otx2_cptlf_unregister_interrupts(lfs);
+ return ret;
+}
+
+void otx2_cptlf_free_irqs_affinity(struct otx2_cptlfs_info *lfs)
+{
+ int slot, offs;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++) {
+ for (offs = 0; offs < OTX2_CPT_LF_MSIX_VECTORS; offs++)
+ irq_set_affinity_hint(pci_irq_vector(lfs->pdev,
+ lfs->lf[slot].msix_offset +
+ offs), NULL);
+ if (lfs->lf[slot].affinity_mask)
+ free_cpumask_var(lfs->lf[slot].affinity_mask);
+ }
+}
+
+int otx2_cptlf_set_irqs_affinity(struct otx2_cptlfs_info *lfs)
+{
+ struct otx2_cptlf_info *lf = lfs->lf;
+ int slot, offs, ret;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++) {
+ if (!zalloc_cpumask_var(&lf[slot].affinity_mask, GFP_KERNEL)) {
+ dev_err(&lfs->pdev->dev,
+ "cpumask allocation failed for LF %d", slot);
+ ret = -ENOMEM;
+ goto free_affinity_mask;
+ }
+
+ cpumask_set_cpu(cpumask_local_spread(slot,
+ dev_to_node(&lfs->pdev->dev)),
+ lf[slot].affinity_mask);
+
+ for (offs = 0; offs < OTX2_CPT_LF_MSIX_VECTORS; offs++) {
+ ret = irq_set_affinity_hint(pci_irq_vector(lfs->pdev,
+ lf[slot].msix_offset + offs),
+ lf[slot].affinity_mask);
+ if (ret)
+ goto free_affinity_mask;
+ }
+ }
+ return 0;
+free_affinity_mask:
+ otx2_cptlf_free_irqs_affinity(lfs);
+ return ret;
+}
+
+int otx2_cptlf_init(struct otx2_cptlfs_info *lfs, u8 eng_grp_mask, int pri,
+ int lfs_num)
+{
+ int slot, ret;
+
+ if (!lfs->pdev || !lfs->reg_base)
+ return -EINVAL;
+
+ lfs->lfs_num = lfs_num;
+ for (slot = 0; slot < lfs->lfs_num; slot++) {
+ lfs->lf[slot].lfs = lfs;
+ lfs->lf[slot].slot = slot;
+ lfs->lf[slot].lmtline = lfs->reg_base +
+ OTX2_CPT_RVU_FUNC_ADDR_S(BLKADDR_LMT, slot,
+ OTX2_CPT_LMT_LF_LMTLINEX(0));
+ lfs->lf[slot].ioreg = lfs->reg_base +
+ OTX2_CPT_RVU_FUNC_ADDR_S(BLKADDR_CPT0, slot,
+ OTX2_CPT_LF_NQX(0));
+ }
+ /* Send request to attach LFs */
+ ret = otx2_cpt_attach_rscrs_msg(lfs);
+ if (ret)
+ goto clear_lfs_num;
+
+ ret = otx2_cpt_alloc_instruction_queues(lfs);
+ if (ret) {
+ dev_err(&lfs->pdev->dev,
+ "Allocating instruction queues failed\n");
+ goto detach_rsrcs;
+ }
+ cptlf_hw_init(lfs);
+ /*
+ * Allow each LF to execute requests destined to any of 8 engine
+ * groups and set queue priority of each LF to high
+ */
+ ret = cptlf_set_grp_and_pri(lfs, eng_grp_mask, pri);
+ if (ret)
+ goto free_iq;
+
+ return 0;
+free_iq:
+ otx2_cpt_free_instruction_queues(lfs);
+ cptlf_hw_cleanup(lfs);
+detach_rsrcs:
+ otx2_cpt_detach_rsrcs_msg(lfs);
+clear_lfs_num:
+ lfs->lfs_num = 0;
+ return ret;
+}
+
+void otx2_cptlf_shutdown(struct otx2_cptlfs_info *lfs)
+{
+ lfs->lfs_num = 0;
+ /* Cleanup LFs hardware side */
+ cptlf_hw_cleanup(lfs);
+ /* Send request to detach LFs */
+ otx2_cpt_detach_rsrcs_msg(lfs);
+}
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptlf.h b/drivers/crypto/marvell/octeontx2/otx2_cptlf.h
new file mode 100644
index 000000000000..be1315ddfb7b
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptlf.h
@@ -0,0 +1,282 @@
+/* SPDX-License-Identifier: GPL-2.0-only
+ * Copyright (C) 2020 Marvell.
+ */
+#ifndef __OTX2_CPTLF_H
+#define __OTX2_CPTLF_H
+
+#include <mbox.h>
+#include <rvu.h>
+#include "otx2_cpt_common.h"
+
+/*
+ * CPT instruction and pending queues user requested length in CPT_INST_S msgs
+ */
+#define OTX2_CPT_USER_REQUESTED_QLEN_MSGS 8200
+
+/*
+ * CPT instruction queue size passed to HW is in units of 40*CPT_INST_S
+ * messages.
+ */
+#define OTX2_CPT_SIZE_DIV40 (OTX2_CPT_USER_REQUESTED_QLEN_MSGS/40)
+
+/*
+ * CPT instruction and pending queues length in CPT_INST_S messages
+ */
+#define OTX2_CPT_INST_QLEN_MSGS ((OTX2_CPT_SIZE_DIV40 - 1) * 40)
+
+/* CPT instruction queue length in bytes */
+#define OTX2_CPT_INST_QLEN_BYTES (OTX2_CPT_SIZE_DIV40 * 40 * \
+ OTX2_CPT_INST_SIZE)
+
+/* CPT instruction group queue length in bytes */
+#define OTX2_CPT_INST_GRP_QLEN_BYTES (OTX2_CPT_SIZE_DIV40 * 16)
+
+/* CPT FC length in bytes */
+#define OTX2_CPT_Q_FC_LEN 128
+
+/* CPT instruction queue alignment */
+#define OTX2_CPT_INST_Q_ALIGNMENT 128
+
+/* Mask which selects all engine groups */
+#define OTX2_CPT_ALL_ENG_GRPS_MASK 0xFF
+
+/* Maximum LFs supported in OcteonTX2 for CPT */
+#define OTX2_CPT_MAX_LFS_NUM 64
+
+/* Queue priority */
+#define OTX2_CPT_QUEUE_HI_PRIO 0x1
+#define OTX2_CPT_QUEUE_LOW_PRIO 0x0
+
+enum otx2_cptlf_state {
+ OTX2_CPTLF_IN_RESET,
+ OTX2_CPTLF_STARTED,
+};
+
+struct otx2_cpt_inst_queue {
+ u8 *vaddr;
+ u8 *real_vaddr;
+ dma_addr_t dma_addr;
+ dma_addr_t real_dma_addr;
+ u32 size;
+};
+
+struct otx2_cptlfs_info;
+struct otx2_cptlf_wqe {
+ struct tasklet_struct work;
+ struct otx2_cptlfs_info *lfs;
+ u8 lf_num;
+};
+
+struct otx2_cptlf_info {
+ struct otx2_cptlfs_info *lfs; /* Ptr to cptlfs_info struct */
+ void __iomem *lmtline; /* Address of LMTLINE */
+ void __iomem *ioreg; /* LMTLINE send register */
+ int msix_offset; /* MSI-X interrupts offset */
+ cpumask_var_t affinity_mask; /* IRQs affinity mask */
+ u8 irq_name[OTX2_CPT_LF_MSIX_VECTORS][32];/* Interrupts name */
+ u8 is_irq_reg[OTX2_CPT_LF_MSIX_VECTORS]; /* Is interrupt registered */
+ u8 slot; /* Slot number of this LF */
+
+ struct otx2_cpt_inst_queue iqueue;/* Instruction queue */
+ struct otx2_cptlf_wqe *wqe; /* Tasklet work info */
+};
+
+struct otx2_cptlfs_info {
+ /* Registers start address of VF/PF LFs are attached to */
+ void __iomem *reg_base;
+ struct pci_dev *pdev; /* Device LFs are attached to */
+ struct otx2_cptlf_info lf[OTX2_CPT_MAX_LFS_NUM];
+ struct otx2_mbox *mbox;
+ u8 are_lfs_attached; /* Whether CPT LFs are attached */
+ u8 lfs_num; /* Number of CPT LFs */
+ atomic_t state; /* LF's state. started/reset */
+};
+
+static inline void otx2_cpt_free_instruction_queues(
+ struct otx2_cptlfs_info *lfs)
+{
+ struct otx2_cpt_inst_queue *iq;
+ int i;
+
+ for (i = 0; i < lfs->lfs_num; i++) {
+ iq = &lfs->lf[i].iqueue;
+ if (iq->real_vaddr)
+ dma_free_coherent(&lfs->pdev->dev,
+ iq->size,
+ iq->real_vaddr,
+ iq->real_dma_addr);
+ iq->real_vaddr = NULL;
+ iq->vaddr = NULL;
+ }
+}
+
+static inline int otx2_cpt_alloc_instruction_queues(
+ struct otx2_cptlfs_info *lfs)
+{
+ struct otx2_cpt_inst_queue *iq;
+ int ret = 0, i;
+
+ if (!lfs->lfs_num)
+ return -EINVAL;
+
+ for (i = 0; i < lfs->lfs_num; i++) {
+ iq = &lfs->lf[i].iqueue;
+ iq->size = OTX2_CPT_INST_QLEN_BYTES +
+ OTX2_CPT_Q_FC_LEN +
+ OTX2_CPT_INST_GRP_QLEN_BYTES +
+ OTX2_CPT_INST_Q_ALIGNMENT;
+ iq->real_vaddr = dma_alloc_coherent(&lfs->pdev->dev, iq->size,
+ &iq->real_dma_addr, GFP_KERNEL);
+ if (!iq->real_vaddr) {
+ ret = -ENOMEM;
+ goto error;
+ }
+ iq->vaddr = iq->real_vaddr + OTX2_CPT_INST_GRP_QLEN_BYTES;
+ iq->dma_addr = iq->real_dma_addr + OTX2_CPT_INST_GRP_QLEN_BYTES;
+
+ /* Align pointers */
+ iq->vaddr = PTR_ALIGN(iq->vaddr, OTX2_CPT_INST_Q_ALIGNMENT);
+ iq->dma_addr = PTR_ALIGN(iq->dma_addr,
+ OTX2_CPT_INST_Q_ALIGNMENT);
+ }
+ return 0;
+error:
+ otx2_cpt_free_instruction_queues(lfs);
+ return ret;
+}
+
+static inline void otx2_cptlf_set_iqueues_base_addr(
+ struct otx2_cptlfs_info *lfs)
+{
+ union otx2_cptx_lf_q_base lf_q_base;
+ int slot;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++) {
+ lf_q_base.u = lfs->lf[slot].iqueue.dma_addr;
+ otx2_cpt_write64(lfs->reg_base, BLKADDR_CPT0, slot,
+ OTX2_CPT_LF_Q_BASE, lf_q_base.u);
+ }
+}
+
+static inline void otx2_cptlf_do_set_iqueue_size(struct otx2_cptlf_info *lf)
+{
+ union otx2_cptx_lf_q_size lf_q_size = { .u = 0x0 };
+
+ lf_q_size.s.size_div40 = OTX2_CPT_SIZE_DIV40;
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_Q_SIZE, lf_q_size.u);
+}
+
+static inline void otx2_cptlf_set_iqueues_size(struct otx2_cptlfs_info *lfs)
+{
+ int slot;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++)
+ otx2_cptlf_do_set_iqueue_size(&lfs->lf[slot]);
+}
+
+static inline void otx2_cptlf_do_disable_iqueue(struct otx2_cptlf_info *lf)
+{
+ union otx2_cptx_lf_ctl lf_ctl = { .u = 0x0 };
+ union otx2_cptx_lf_inprog lf_inprog;
+ int timeout = 20;
+
+ /* Disable instructions enqueuing */
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_CTL, lf_ctl.u);
+
+ /* Wait for instruction queue to become empty */
+ do {
+ lf_inprog.u = otx2_cpt_read64(lf->lfs->reg_base, BLKADDR_CPT0,
+ lf->slot, OTX2_CPT_LF_INPROG);
+ if (!lf_inprog.s.inflight)
+ break;
+
+ usleep_range(10000, 20000);
+ if (timeout-- < 0) {
+ dev_err(&lf->lfs->pdev->dev,
+ "Error LF %d is still busy.\n", lf->slot);
+ break;
+ }
+
+ } while (1);
+
+ /*
+ * Disable executions in the LF's queue,
+ * the queue should be empty at this point
+ */
+ lf_inprog.s.eena = 0x0;
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_INPROG, lf_inprog.u);
+}
+
+static inline void otx2_cptlf_disable_iqueues(struct otx2_cptlfs_info *lfs)
+{
+ int slot;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++)
+ otx2_cptlf_do_disable_iqueue(&lfs->lf[slot]);
+}
+
+static inline void otx2_cptlf_set_iqueue_enq(struct otx2_cptlf_info *lf,
+ bool enable)
+{
+ union otx2_cptx_lf_ctl lf_ctl;
+
+ lf_ctl.u = otx2_cpt_read64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_CTL);
+
+ /* Set iqueue's enqueuing */
+ lf_ctl.s.ena = enable ? 0x1 : 0x0;
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_CTL, lf_ctl.u);
+}
+
+static inline void otx2_cptlf_enable_iqueue_enq(struct otx2_cptlf_info *lf)
+{
+ otx2_cptlf_set_iqueue_enq(lf, true);
+}
+
+static inline void otx2_cptlf_set_iqueue_exec(struct otx2_cptlf_info *lf,
+ bool enable)
+{
+ union otx2_cptx_lf_inprog lf_inprog;
+
+ lf_inprog.u = otx2_cpt_read64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_INPROG);
+
+ /* Set iqueue's execution */
+ lf_inprog.s.eena = enable ? 0x1 : 0x0;
+ otx2_cpt_write64(lf->lfs->reg_base, BLKADDR_CPT0, lf->slot,
+ OTX2_CPT_LF_INPROG, lf_inprog.u);
+}
+
+static inline void otx2_cptlf_enable_iqueue_exec(struct otx2_cptlf_info *lf)
+{
+ otx2_cptlf_set_iqueue_exec(lf, true);
+}
+
+static inline void otx2_cptlf_disable_iqueue_exec(struct otx2_cptlf_info *lf)
+{
+ otx2_cptlf_set_iqueue_exec(lf, false);
+}
+
+static inline void otx2_cptlf_enable_iqueues(struct otx2_cptlfs_info *lfs)
+{
+ int slot;
+
+ for (slot = 0; slot < lfs->lfs_num; slot++) {
+ otx2_cptlf_enable_iqueue_exec(&lfs->lf[slot]);
+ otx2_cptlf_enable_iqueue_enq(&lfs->lf[slot]);
+ }
+}
+
+int otx2_cptlf_init(struct otx2_cptlfs_info *lfs, u8 eng_grp_msk, int pri,
+ int lfs_num);
+void otx2_cptlf_shutdown(struct otx2_cptlfs_info *lfs);
+int otx2_cptlf_register_interrupts(struct otx2_cptlfs_info *lfs);
+void otx2_cptlf_unregister_interrupts(struct otx2_cptlfs_info *lfs);
+void otx2_cptlf_free_irqs_affinity(struct otx2_cptlfs_info *lfs);
+int otx2_cptlf_set_irqs_affinity(struct otx2_cptlfs_info *lfs);
+
+#endif /* __OTX2_CPTLF_H */
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf.h b/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
index 528a3975b9c2..15561fe50bc0 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
@@ -7,6 +7,7 @@
#include "otx2_cpt_common.h"
#include "otx2_cptpf_ucode.h"
+#include "otx2_cptlf.h"
struct otx2_cptpf_dev;
struct otx2_cptvf_info {
@@ -24,6 +25,7 @@ struct otx2_cptpf_dev {
struct pci_dev *pdev; /* PCI device handle */
struct otx2_cptvf_info vf[OTX2_CPT_MAX_VFS_NUM];
struct otx2_cpt_eng_grps eng_grps;/* Engine groups information */
+ struct otx2_cptlfs_info lfs; /* CPT LFs attached to this PF */
/* AF <=> PF mbox */
struct otx2_mbox afpf_mbox;
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
index 8852a7e5e035..c9bb6b14876c 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
@@ -207,6 +207,14 @@ static void process_afpf_mbox_msg(struct otx2_cptpf_dev *cptpf,
if (!rsp_rd_wr->is_write)
*rsp_rd_wr->ret_val = rsp_rd_wr->val;
break;
+ case MBOX_MSG_ATTACH_RESOURCES:
+ if (!msg->rc)
+ cptpf->lfs.are_lfs_attached = 1;
+ break;
+ case MBOX_MSG_DETACH_RESOURCES:
+ if (!msg->rc)
+ cptpf->lfs.are_lfs_attached = 0;
+ break;
default:
dev_err(dev,
--
2.28.0
Add entries to debugfs at /sys/kernel/debug/octeontx2/cpt.
cpt_pc: dump cpt performance HW registers.
Usage:
cat /sys/kernel/debug/octeontx2/cpt/cpt_pc
cpt_ae_sts: show cpt asymmetric engines current state
Usage:
cat /sys/kernel/debug/octeontx2/cpt/cpt_ae_sts
cpt_se_sts: show cpt symmetric engines current state
Usage:
cat /sys/kernel/debug/octeontx2/cpt/cpt_se_sts
cpt_engines_info: dump cpt engine control registers.
Usage:
cat /sys/kernel/debug/octeontx2/cpt/cpt_engines_info
cpt_lfs_info: dump cpt lfs control registers.
Usage:
cat /sys/kernel/debug/octeontx2/cpt/cpt_lfs_info
cpt_err_info: dump cpt error registers.
Usage:
cat /sys/kernel/debug/octeontx2/cpt/cpt_err_info
Signed-off-by: Suheil Chandran <[email protected]>
Signed-off-by: Srujana Challa <[email protected]>
---
.../net/ethernet/marvell/octeontx2/af/rvu.h | 1 +
.../marvell/octeontx2/af/rvu_debugfs.c | 304 ++++++++++++++++++
2 files changed, 305 insertions(+)
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
index c37e106d7006..ba18171c87d6 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
@@ -50,6 +50,7 @@ struct rvu_debugfs {
struct dentry *npa;
struct dentry *nix;
struct dentry *npc;
+ struct dentry *cpt;
struct dump_ctx npa_aura_ctx;
struct dump_ctx npa_pool_ctx;
struct dump_ctx nix_cq_ctx;
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
index 77adad4adb1b..24354bfb4e94 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
@@ -1676,6 +1676,309 @@ static void rvu_dbg_npc_init(struct rvu *rvu)
debugfs_remove_recursive(rvu->rvu_dbg.npc);
}
+/* CPT debugfs APIs */
+static int rvu_dbg_cpt_ae_sts_display(struct seq_file *filp, void *unused)
+{
+ struct rvu *rvu = filp->private;
+ u64 busy_sts = 0, free_sts = 0;
+ u32 e_min = 0, e_max = 0, e, i;
+ u16 max_ses, max_ies, max_aes;
+ int blkaddr;
+ u64 reg;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
+ if (blkaddr < 0)
+ return -ENODEV;
+
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_CONSTANTS1);
+ max_ses = reg & 0xffff;
+ max_ies = (reg >> 16) & 0xffff;
+ max_aes = (reg >> 32) & 0xffff;
+
+ e_min = max_ses + max_ies;
+ e_max = max_ses + max_ies + max_aes;
+
+ for (e = e_min, i = 0; e < e_max; e++, i++) {
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_STS(e));
+ if (reg & 0x1)
+ busy_sts |= 1ULL << i;
+
+ if (reg & 0x2)
+ free_sts |= 1ULL << i;
+ }
+ seq_printf(filp, "FREE STS : 0x%016llx\n", free_sts);
+ seq_printf(filp, "BUSY STS : 0x%016llx\n", busy_sts);
+
+ return 0;
+}
+
+RVU_DEBUG_SEQ_FOPS(cpt_ae_sts, cpt_ae_sts_display, NULL);
+
+static int rvu_dbg_cpt_se_sts_display(struct seq_file *filp, void *unused)
+{
+ struct rvu *rvu = filp->private;
+ u64 busy_sts = 0, free_sts = 0;
+ u32 e_min = 0, e_max = 0, e;
+ u16 max_ses;
+ int blkaddr;
+ u64 reg;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
+ if (blkaddr < 0)
+ return -ENODEV;
+
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_CONSTANTS1);
+ max_ses = reg & 0xffff;
+
+ e_min = 0;
+ e_max = max_ses;
+
+ for (e = e_min; e < e_max; e++) {
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_STS(e));
+ if (reg & 0x1)
+ busy_sts |= 1ULL << e;
+
+ if (reg & 0x2)
+ free_sts |= 1ULL << e;
+ }
+ seq_printf(filp, "FREE STS : 0x%016llx\n", free_sts);
+ seq_printf(filp, "BUSY STS : 0x%016llx\n", busy_sts);
+
+ return 0;
+}
+
+RVU_DEBUG_SEQ_FOPS(cpt_se_sts, cpt_se_sts_display, NULL);
+
+static int rvu_dbg_cpt_ie_sts_display(struct seq_file *filp, void *unused)
+{
+ struct rvu *rvu = filp->private;
+ u64 busy_sts = 0, free_sts = 0;
+ u32 e_min = 0, e_max = 0, e, i;
+ u16 max_ses, max_ies;
+ int blkaddr;
+ u64 reg;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
+ if (blkaddr < 0)
+ return -ENODEV;
+
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_CONSTANTS1);
+ max_ses = reg & 0xffff;
+ max_ies = (reg >> 16) & 0xffff;
+
+ e_min = max_ses;
+ e_max = max_ses + max_ies;
+
+ for (e = e_min, i = 0; e < e_max; e++, i++) {
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_STS(e));
+ if (reg & 0x1)
+ busy_sts |= 1ULL << i;
+
+ if (reg & 0x2)
+ free_sts |= 1ULL << i;
+ }
+ seq_printf(filp, "FREE STS : 0x%016llx\n", free_sts);
+ seq_printf(filp, "BUSY STS : 0x%016llx\n", busy_sts);
+
+ return 0;
+}
+
+RVU_DEBUG_SEQ_FOPS(cpt_ie_sts, cpt_ie_sts_display, NULL);
+
+static int rvu_dbg_cpt_engines_info_display(struct seq_file *filp, void *unused)
+{
+ struct rvu *rvu = filp->private;
+ u16 max_ses, max_ies, max_aes;
+ u32 e_max, e;
+ int blkaddr;
+ u64 reg;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
+ if (blkaddr < 0)
+ return -ENODEV;
+
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_CONSTANTS1);
+ max_ses = reg & 0xffff;
+ max_ies = (reg >> 16) & 0xffff;
+ max_aes = (reg >> 32) & 0xffff;
+
+ e_max = max_ses + max_ies + max_aes;
+
+ seq_puts(filp, "===========================================\n");
+ for (e = 0; e < e_max; e++) {
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_CTL2(e));
+ seq_printf(filp, "CPT Engine[%u] Group Enable 0x%02llx\n", e,
+ reg & 0xff);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_ACTIVE(e));
+ seq_printf(filp, "CPT Engine[%u] Active Info 0x%llx\n", e,
+ reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_CTL(e));
+ seq_printf(filp, "CPT Engine[%u] Control 0x%llx\n", e,
+ reg);
+ seq_puts(filp, "===========================================\n");
+ }
+ return 0;
+}
+
+RVU_DEBUG_SEQ_FOPS(cpt_engines_info, cpt_engines_info_display, NULL);
+
+static int rvu_dbg_cpt_lfs_info_display(struct seq_file *filp, void *unused)
+{
+ struct rvu *rvu = filp->private;
+ struct rvu_hwinfo *hw = rvu->hw;
+ struct rvu_block *block;
+ int blkaddr;
+ u64 reg;
+ u32 lf;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
+ if (blkaddr < 0)
+ return -ENODEV;
+
+ block = &hw->block[blkaddr];
+ if (!block->lf.bmap)
+ return -ENODEV;
+
+ seq_puts(filp, "===========================================\n");
+ for (lf = 0; lf < block->lf.max; lf++) {
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_LFX_CTL(lf));
+ seq_printf(filp, "CPT Lf[%u] CTL 0x%llx\n", lf, reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_LFX_CTL2(lf));
+ seq_printf(filp, "CPT Lf[%u] CTL2 0x%llx\n", lf, reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_LFX_PTR_CTL(lf));
+ seq_printf(filp, "CPT Lf[%u] PTR_CTL 0x%llx\n", lf, reg);
+ reg = rvu_read64(rvu, blkaddr, block->lfcfg_reg |
+ (lf << block->lfshift));
+ seq_printf(filp, "CPT Lf[%u] CFG 0x%llx\n", lf, reg);
+ seq_puts(filp, "===========================================\n");
+ }
+ return 0;
+}
+
+RVU_DEBUG_SEQ_FOPS(cpt_lfs_info, cpt_lfs_info_display, NULL);
+
+static int rvu_dbg_cpt_err_info_display(struct seq_file *filp, void *unused)
+{
+ struct rvu *rvu = filp->private;
+ u64 reg0, reg1;
+ int blkaddr;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
+ if (blkaddr < 0)
+ return -ENODEV;
+
+ reg0 = rvu_read64(rvu, blkaddr, CPT_AF_FLTX_INT(0));
+ reg1 = rvu_read64(rvu, blkaddr, CPT_AF_FLTX_INT(1));
+ seq_printf(filp, "CPT_AF_FLTX_INT: 0x%llx 0x%llx\n", reg0, reg1);
+ reg0 = rvu_read64(rvu, blkaddr, CPT_AF_PSNX_EXE(0));
+ reg1 = rvu_read64(rvu, blkaddr, CPT_AF_PSNX_EXE(1));
+ seq_printf(filp, "CPT_AF_PSNX_EXE: 0x%llx 0x%llx\n", reg0, reg1);
+ reg0 = rvu_read64(rvu, blkaddr, CPT_AF_PSNX_LF(0));
+ seq_printf(filp, "CPT_AF_PSNX_LF: 0x%llx\n", reg0);
+ reg0 = rvu_read64(rvu, blkaddr, CPT_AF_RVU_INT);
+ seq_printf(filp, "CPT_AF_RVU_INT: 0x%llx\n", reg0);
+ reg0 = rvu_read64(rvu, blkaddr, CPT_AF_RAS_INT);
+ seq_printf(filp, "CPT_AF_RAS_INT: 0x%llx\n", reg0);
+ reg0 = rvu_read64(rvu, blkaddr, CPT_AF_EXE_ERR_INFO);
+ seq_printf(filp, "CPT_AF_EXE_ERR_INFO: 0x%llx\n", reg0);
+
+ return 0;
+}
+
+RVU_DEBUG_SEQ_FOPS(cpt_err_info, cpt_err_info_display, NULL);
+
+static int rvu_dbg_cpt_pc_display(struct seq_file *filp, void *unused)
+{
+ struct rvu *rvu;
+ int blkaddr;
+ u64 reg;
+
+ rvu = filp->private;
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
+ if (blkaddr < 0)
+ return -ENODEV;
+
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_INST_REQ_PC);
+ seq_printf(filp, "CPT instruction requests %llu\n", reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_INST_LATENCY_PC);
+ seq_printf(filp, "CPT instruction latency %llu\n", reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_RD_REQ_PC);
+ seq_printf(filp, "CPT NCB read requests %llu\n", reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_RD_LATENCY_PC);
+ seq_printf(filp, "CPT NCB read latency %llu\n", reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_RD_UC_PC);
+ seq_printf(filp, "CPT read requests caused by UC fills %llu\n", reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_ACTIVE_CYCLES_PC);
+ seq_printf(filp, "CPT active cycles pc %llu\n", reg);
+ reg = rvu_read64(rvu, blkaddr, CPT_AF_CPTCLK_CNT);
+ seq_printf(filp, "CPT clock count pc %llu\n", reg);
+
+ return 0;
+}
+
+RVU_DEBUG_SEQ_FOPS(cpt_pc, cpt_pc_display, NULL);
+
+static void rvu_dbg_cpt_init(struct rvu *rvu)
+{
+ const struct device *dev = &rvu->pdev->dev;
+ struct dentry *pfile;
+
+ if (!is_block_implemented(rvu->hw, BLKADDR_CPT0))
+ return;
+
+ rvu->rvu_dbg.cpt = debugfs_create_dir("cpt", rvu->rvu_dbg.root);
+ if (!rvu->rvu_dbg.cpt)
+ return;
+
+ pfile = debugfs_create_file("cpt_pc", 0600,
+ rvu->rvu_dbg.cpt, rvu,
+ &rvu_dbg_cpt_pc_fops);
+ if (!pfile)
+ goto create_failed;
+
+ pfile = debugfs_create_file("cpt_ae_sts", 0600,
+ rvu->rvu_dbg.cpt, rvu,
+ &rvu_dbg_cpt_ae_sts_fops);
+ if (!pfile)
+ goto create_failed;
+
+ pfile = debugfs_create_file("cpt_se_sts", 0600,
+ rvu->rvu_dbg.cpt, rvu,
+ &rvu_dbg_cpt_se_sts_fops);
+ if (!pfile)
+ goto create_failed;
+
+ pfile = debugfs_create_file("cpt_ie_sts", 0600,
+ rvu->rvu_dbg.cpt, rvu,
+ &rvu_dbg_cpt_ie_sts_fops);
+ if (!pfile)
+ goto create_failed;
+
+ pfile = debugfs_create_file("cpt_engines_info", 0600,
+ rvu->rvu_dbg.cpt, rvu,
+ &rvu_dbg_cpt_engines_info_fops);
+ if (!pfile)
+ goto create_failed;
+
+ pfile = debugfs_create_file("cpt_lfs_info", 0600,
+ rvu->rvu_dbg.cpt, rvu,
+ &rvu_dbg_cpt_lfs_info_fops);
+ if (!pfile)
+ goto create_failed;
+
+ pfile = debugfs_create_file("cpt_err_info", 0600,
+ rvu->rvu_dbg.cpt, rvu,
+ &rvu_dbg_cpt_err_info_fops);
+ if (!pfile)
+ goto create_failed;
+
+ return;
+
+create_failed:
+ dev_err(dev, "Failed to create debugfs dir/file for CPT\n");
+ debugfs_remove_recursive(rvu->rvu_dbg.cpt);
+}
+
void rvu_dbg_init(struct rvu *rvu)
{
struct device *dev = &rvu->pdev->dev;
@@ -1695,6 +1998,7 @@ void rvu_dbg_init(struct rvu *rvu)
rvu_dbg_nix_init(rvu);
rvu_dbg_cgx_init(rvu);
rvu_dbg_npc_init(rvu);
+ rvu_dbg_cpt_init(rvu);
return;
--
2.28.0
On OcteonTX2 platform CPT instruction enqueue and NIX
packet send are only possible via LMTST operations which
uses LDEOR instruction. This patch moves lmt flush
function from OcteonTX2 nic driver to include/linux/soc
since it will be used by OcteonTX2 CPT and NIC driver for
LMTST.
Signed-off-by: Suheil Chandran <[email protected]>
Signed-off-by: Srujana Challa <[email protected]>
---
MAINTAINERS | 2 ++
.../marvell/octeontx2/nic/otx2_common.h | 13 +--------
include/linux/soc/marvell/octeontx2/asm.h | 29 +++++++++++++++++++
3 files changed, 32 insertions(+), 12 deletions(-)
create mode 100644 include/linux/soc/marvell/octeontx2/asm.h
diff --git a/MAINTAINERS b/MAINTAINERS
index f281f8077de0..bfb7f3e45d33 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -10476,6 +10476,7 @@ M: Srujana Challa <[email protected]>
L: [email protected]
S: Maintained
F: drivers/crypto/marvell/
+F: include/linux/soc/marvell/octeontx2/
MARVELL GIGABIT ETHERNET DRIVERS (skge/sky2)
M: Mirko Lindner <[email protected]>
@@ -10548,6 +10549,7 @@ M: hariprasad <[email protected]>
L: [email protected]
S: Supported
F: drivers/net/ethernet/marvell/octeontx2/nic/
+F: include/linux/soc/marvell/octeontx2/
MARVELL OCTEONTX2 RVU ADMIN FUNCTION DRIVER
M: Sunil Goutham <[email protected]>
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
index d6253f2a414d..2e8288445dcb 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
@@ -16,6 +16,7 @@
#include <linux/net_tstamp.h>
#include <linux/ptp_clock_kernel.h>
#include <linux/timecounter.h>
+#include <linux/soc/marvell/octeontx2/asm.h>
#include <mbox.h>
#include "otx2_reg.h"
@@ -421,21 +422,9 @@ static inline u64 otx2_atomic64_add(u64 incr, u64 *ptr)
return result;
}
-static inline u64 otx2_lmt_flush(uint64_t addr)
-{
- u64 result = 0;
-
- __asm__ volatile(".cpu generic+lse\n"
- "ldeor xzr,%x[rf],[%[rs]]"
- : [rf]"=r"(result)
- : [rs]"r"(addr));
- return result;
-}
-
#else
#define otx2_write128(lo, hi, addr)
#define otx2_atomic64_add(incr, ptr) ({ *ptr += incr; })
-#define otx2_lmt_flush(addr) ({ 0; })
#endif
/* Alloc pointer from pool/aura */
diff --git a/include/linux/soc/marvell/octeontx2/asm.h b/include/linux/soc/marvell/octeontx2/asm.h
new file mode 100644
index 000000000000..ae2279fe830a
--- /dev/null
+++ b/include/linux/soc/marvell/octeontx2/asm.h
@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: GPL-2.0-only
+ * Copyright (C) 2020 Marvell.
+ */
+
+#ifndef __SOC_OTX2_ASM_H
+#define __SOC_OTX2_ASM_H
+
+#if defined(CONFIG_ARM64)
+/*
+ * otx2_lmt_flush is used for LMT store operation.
+ * On octeontx2 platform CPT instruction enqueue and
+ * NIX packet send are only possible via LMTST
+ * operations and it uses LDEOR instruction targeting
+ * the coprocessor address.
+ */
+#define otx2_lmt_flush(ioaddr) \
+({ \
+ u64 result = 0; \
+ __asm__ volatile(".cpu generic+lse\n" \
+ "ldeor xzr, %x[rf], [%[rs]]" \
+ : [rf]"=r" (result) \
+ : [rs]"r" (ioaddr)); \
+ (result); \
+})
+#else
+#define otx2_lmt_flush(ioaddr) ({ 0; })
+#endif
+
+#endif /* __SOC_OTX2_ASM_H */
--
2.28.0
In the resource virtualization unit (RVU) each of the PF and AF
(admin function) share a 64KB of reserved memory region for
communication. This patch initializes PF <=> AF mailbox IRQs,
registers handlers for processing these communication messages.
Signed-off-by: Suheil Chandran <[email protected]>
Signed-off-by: Srujana Challa <[email protected]>
---
drivers/crypto/marvell/octeontx2/Makefile | 3 +-
.../marvell/octeontx2/otx2_cpt_common.h | 4 +
.../marvell/octeontx2/otx2_cpt_mbox_common.c | 37 +++++++
drivers/crypto/marvell/octeontx2/otx2_cptpf.h | 12 +++
.../marvell/octeontx2/otx2_cptpf_main.c | 101 +++++++++++++++++-
.../marvell/octeontx2/otx2_cptpf_mbox.c | 80 ++++++++++++++
6 files changed, 235 insertions(+), 2 deletions(-)
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
diff --git a/drivers/crypto/marvell/octeontx2/Makefile b/drivers/crypto/marvell/octeontx2/Makefile
index db763ad46a91..8c8262e94f78 100644
--- a/drivers/crypto/marvell/octeontx2/Makefile
+++ b/drivers/crypto/marvell/octeontx2/Makefile
@@ -1,6 +1,7 @@
# SPDX-License-Identifier: GPL-2.0-only
obj-$(CONFIG_CRYPTO_DEV_OCTEONTX2_CPT) += octeontx2-cpt.o
-octeontx2-cpt-objs := otx2_cptpf_main.o
+octeontx2-cpt-objs := otx2_cptpf_main.o otx2_cptpf_mbox.o \
+ otx2_cpt_mbox_common.o
ccflags-y += -I$(srctree)/drivers/net/ethernet/marvell/octeontx2/af
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
index eff4ffa58dc4..b677f8c7e724 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
+++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
@@ -12,6 +12,7 @@
#include <linux/crypto.h>
#include "otx2_cpt_hw_types.h"
#include "rvu.h"
+#include "mbox.h"
#define OTX2_CPT_RVU_FUNC_ADDR_S(blk, slot, offs) \
(((blk) << 20) | ((slot) << 12) | (offs))
@@ -29,4 +30,7 @@ static inline u64 otx2_cpt_read64(void __iomem *reg_base, u64 blk, u64 slot,
return readq_relaxed(reg_base +
OTX2_CPT_RVU_FUNC_ADDR_S(blk, slot, offs));
}
+
+int otx2_cpt_send_ready_msg(struct otx2_mbox *mbox, struct pci_dev *pdev);
+int otx2_cpt_send_mbox_msg(struct otx2_mbox *mbox, struct pci_dev *pdev);
#endif /* __OTX2_CPT_COMMON_H */
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
new file mode 100644
index 000000000000..a122483b5976
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2020 Marvell. */
+
+#include "otx2_cpt_common.h"
+
+int otx2_cpt_send_mbox_msg(struct otx2_mbox *mbox, struct pci_dev *pdev)
+{
+ int ret;
+
+ otx2_mbox_msg_send(mbox, 0);
+ ret = otx2_mbox_wait_for_rsp(mbox, 0);
+ if (ret == -EIO) {
+ dev_err(&pdev->dev, "RVU MBOX timeout.\n");
+ return ret;
+ } else if (ret) {
+ dev_err(&pdev->dev, "RVU MBOX error: %d.\n", ret);
+ return -EFAULT;
+ }
+ return ret;
+}
+
+int otx2_cpt_send_ready_msg(struct otx2_mbox *mbox, struct pci_dev *pdev)
+{
+ struct mbox_msghdr *req;
+
+ req = otx2_mbox_alloc_msg_rsp(mbox, 0, sizeof(*req),
+ sizeof(struct ready_msg_rsp));
+ if (req == NULL) {
+ dev_err(&pdev->dev, "RVU MBOX failed to get message.\n");
+ return -EFAULT;
+ }
+ req->id = MBOX_MSG_READY;
+ req->sig = OTX2_MBOX_REQ_SIG;
+ req->pcifunc = 0;
+
+ return otx2_cpt_send_mbox_msg(mbox, pdev);
+}
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf.h b/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
index 84cdc8cc2c15..87fe4c6838e5 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf.h
@@ -5,9 +5,21 @@
#ifndef __OTX2_CPTPF_H
#define __OTX2_CPTPF_H
+#include "otx2_cpt_common.h"
+
struct otx2_cptpf_dev {
void __iomem *reg_base; /* CPT PF registers start address */
+ void __iomem *afpf_mbox_base; /* PF-AF mbox start address */
struct pci_dev *pdev; /* PCI device handle */
+ /* AF <=> PF mbox */
+ struct otx2_mbox afpf_mbox;
+ struct work_struct afpf_mbox_work;
+ struct workqueue_struct *afpf_mbox_wq;
+
+ u8 pf_id; /* RVU PF number */
};
+irqreturn_t otx2_cptpf_afpf_mbox_intr(int irq, void *arg);
+void otx2_cptpf_afpf_mbox_handler(struct work_struct *work);
+
#endif /* __OTX2_CPTPF_H */
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
index 25ac8d5ff7ee..5b96dbe461f2 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
@@ -10,6 +10,74 @@
#define OTX2_CPT_DRV_NAME "octeontx2-cpt"
#define OTX2_CPT_DRV_STRING "Marvell OcteonTX2 CPT Physical Function Driver"
+static void cptpf_disable_afpf_mbox_intr(struct otx2_cptpf_dev *cptpf)
+{
+ /* Disable AF-PF interrupt */
+ otx2_cpt_write64(cptpf->reg_base, BLKADDR_RVUM, 0, RVU_PF_INT_ENA_W1C,
+ 0x1ULL);
+ /* Clear interrupt if any */
+ otx2_cpt_write64(cptpf->reg_base, BLKADDR_RVUM, 0, RVU_PF_INT, 0x1ULL);
+}
+
+static int cptpf_register_afpf_mbox_intr(struct otx2_cptpf_dev *cptpf)
+{
+ struct pci_dev *pdev = cptpf->pdev;
+ struct device *dev = &pdev->dev;
+ int ret, irq;
+
+ irq = pci_irq_vector(pdev, RVU_PF_INT_VEC_AFPF_MBOX);
+ /* Register AF-PF mailbox interrupt handler */
+ ret = devm_request_irq(dev, irq, otx2_cptpf_afpf_mbox_intr, 0,
+ "CPTAFPF Mbox", cptpf);
+ if (ret) {
+ dev_err(dev,
+ "IRQ registration failed for PFAF mbox irq\n");
+ return ret;
+ }
+ /* Clear interrupt if any, to avoid spurious interrupts */
+ otx2_cpt_write64(cptpf->reg_base, BLKADDR_RVUM, 0, RVU_PF_INT, 0x1ULL);
+ /* Enable AF-PF interrupt */
+ otx2_cpt_write64(cptpf->reg_base, BLKADDR_RVUM, 0, RVU_PF_INT_ENA_W1S,
+ 0x1ULL);
+
+ ret = otx2_cpt_send_ready_msg(&cptpf->afpf_mbox, cptpf->pdev);
+ if (ret) {
+ dev_warn(dev,
+ "AF not responding to mailbox, deferring probe\n");
+ cptpf_disable_afpf_mbox_intr(cptpf);
+ return -EPROBE_DEFER;
+ }
+ return 0;
+}
+
+static int cptpf_afpf_mbox_init(struct otx2_cptpf_dev *cptpf)
+{
+ int err;
+
+ cptpf->afpf_mbox_wq = alloc_workqueue("cpt_afpf_mailbox",
+ WQ_UNBOUND | WQ_HIGHPRI |
+ WQ_MEM_RECLAIM, 1);
+ if (!cptpf->afpf_mbox_wq)
+ return -ENOMEM;
+
+ err = otx2_mbox_init(&cptpf->afpf_mbox, cptpf->afpf_mbox_base,
+ cptpf->pdev, cptpf->reg_base, MBOX_DIR_PFAF, 1);
+ if (err)
+ goto error;
+
+ INIT_WORK(&cptpf->afpf_mbox_work, otx2_cptpf_afpf_mbox_handler);
+ return 0;
+error:
+ destroy_workqueue(cptpf->afpf_mbox_wq);
+ return err;
+}
+
+static void cptpf_afpf_mbox_destroy(struct otx2_cptpf_dev *cptpf)
+{
+ destroy_workqueue(cptpf->afpf_mbox_wq);
+ otx2_mbox_destroy(&cptpf->afpf_mbox);
+}
+
static int cpt_is_pf_usable(struct otx2_cptpf_dev *cptpf)
{
u64 rev;
@@ -33,6 +101,7 @@ static int otx2_cptpf_probe(struct pci_dev *pdev,
const struct pci_device_id *ent)
{
struct device *dev = &pdev->dev;
+ resource_size_t offset, size;
struct otx2_cptpf_dev *cptpf;
int err;
@@ -69,7 +138,34 @@ static int otx2_cptpf_probe(struct pci_dev *pdev,
if (err)
goto clear_drvdata;
+ offset = pci_resource_start(pdev, PCI_MBOX_BAR_NUM);
+ size = pci_resource_len(pdev, PCI_MBOX_BAR_NUM);
+ /* Map AF-PF mailbox memory */
+ cptpf->afpf_mbox_base = devm_ioremap_wc(dev, offset, size);
+ if (!cptpf->afpf_mbox_base) {
+ dev_err(&pdev->dev, "Unable to map BAR4\n");
+ err = -ENODEV;
+ goto clear_drvdata;
+ }
+ err = pci_alloc_irq_vectors(pdev, RVU_PF_INT_VEC_CNT,
+ RVU_PF_INT_VEC_CNT, PCI_IRQ_MSIX);
+ if (err < 0) {
+ dev_err(dev, "Request for %d msix vectors failed\n",
+ RVU_PF_INT_VEC_CNT);
+ goto clear_drvdata;
+ }
+ /* Initialize AF-PF mailbox */
+ err = cptpf_afpf_mbox_init(cptpf);
+ if (err)
+ goto clear_drvdata;
+ /* Register mailbox interrupt */
+ err = cptpf_register_afpf_mbox_intr(cptpf);
+ if (err)
+ goto destroy_afpf_mbox;
+
return 0;
+destroy_afpf_mbox:
+ cptpf_afpf_mbox_destroy(cptpf);
clear_drvdata:
pci_set_drvdata(pdev, NULL);
return err;
@@ -81,7 +177,10 @@ static void otx2_cptpf_remove(struct pci_dev *pdev)
if (!cptpf)
return;
-
+ /* Disable AF-PF mailbox interrupt */
+ cptpf_disable_afpf_mbox_intr(cptpf);
+ /* Destroy AF-PF mbox */
+ cptpf_afpf_mbox_destroy(cptpf);
pci_set_drvdata(pdev, NULL);
}
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c b/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
new file mode 100644
index 000000000000..0a8bd46b5686
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
@@ -0,0 +1,80 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2020 Marvell. */
+
+#include "otx2_cpt_common.h"
+#include "otx2_cptpf.h"
+#include "rvu_reg.h"
+
+irqreturn_t otx2_cptpf_afpf_mbox_intr(int __always_unused irq, void *arg)
+{
+ struct otx2_cptpf_dev *cptpf = arg;
+ u64 intr;
+
+ /* Read the interrupt bits */
+ intr = otx2_cpt_read64(cptpf->reg_base, BLKADDR_RVUM, 0, RVU_PF_INT);
+
+ if (intr & 0x1ULL) {
+ /* Schedule work queue function to process the MBOX request */
+ queue_work(cptpf->afpf_mbox_wq, &cptpf->afpf_mbox_work);
+ /* Clear and ack the interrupt */
+ otx2_cpt_write64(cptpf->reg_base, BLKADDR_RVUM, 0, RVU_PF_INT,
+ 0x1ULL);
+ }
+ return IRQ_HANDLED;
+}
+
+static void process_afpf_mbox_msg(struct otx2_cptpf_dev *cptpf,
+ struct mbox_msghdr *msg)
+{
+ struct device *dev = &cptpf->pdev->dev;
+
+ if (msg->id >= MBOX_MSG_MAX) {
+ dev_err(dev, "MBOX msg with unknown ID %d\n", msg->id);
+ return;
+ }
+ if (msg->sig != OTX2_MBOX_RSP_SIG) {
+ dev_err(dev, "MBOX msg with wrong signature %x, ID %d\n",
+ msg->sig, msg->id);
+ return;
+ }
+
+ switch (msg->id) {
+ case MBOX_MSG_READY:
+ cptpf->pf_id = (msg->pcifunc >> RVU_PFVF_PF_SHIFT) &
+ RVU_PFVF_PF_MASK;
+ break;
+ default:
+ dev_err(dev,
+ "Unsupported msg %d received.\n", msg->id);
+ break;
+ }
+}
+
+/* Handle mailbox messages received from AF */
+void otx2_cptpf_afpf_mbox_handler(struct work_struct *work)
+{
+ struct otx2_cptpf_dev *cptpf;
+ struct otx2_mbox *afpf_mbox;
+ struct otx2_mbox_dev *mdev;
+ struct mbox_hdr *rsp_hdr;
+ struct mbox_msghdr *msg;
+ int offset, i;
+
+ cptpf = container_of(work, struct otx2_cptpf_dev, afpf_mbox_work);
+ afpf_mbox = &cptpf->afpf_mbox;
+ mdev = &afpf_mbox->dev[0];
+ /* Sync mbox data into memory */
+ smp_wmb();
+
+ rsp_hdr = (struct mbox_hdr *)(mdev->mbase + afpf_mbox->rx_start);
+ offset = ALIGN(sizeof(*rsp_hdr), MBOX_MSG_ALIGN);
+
+ for (i = 0; i < rsp_hdr->num_msgs; i++) {
+ msg = (struct mbox_msghdr *)(mdev->mbase + afpf_mbox->rx_start +
+ offset);
+ process_afpf_mbox_msg(cptpf, msg);
+ offset = msg->next_msgoff;
+ mdev->msgs_acked++;
+ }
+ otx2_mbox_reset(afpf_mbox, 0);
+}
--
2.28.0
Add support for the Marvell OcteonTX2 CPT virtual function
driver. This patch includes probe, PCI specific initialization
and interrupt handling.
Signed-off-by: Suheil Chandran <[email protected]>
Signed-off-by: Srujana Challa <[email protected]>
---
drivers/crypto/marvell/octeontx2/Makefile | 4 +-
.../marvell/octeontx2/otx2_cpt_common.h | 1 +
.../marvell/octeontx2/otx2_cpt_mbox_common.c | 32 +++
drivers/crypto/marvell/octeontx2/otx2_cptvf.h | 28 +++
.../marvell/octeontx2/otx2_cptvf_main.c | 194 ++++++++++++++++++
.../marvell/octeontx2/otx2_cptvf_mbox.c | 113 ++++++++++
6 files changed, 371 insertions(+), 1 deletion(-)
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf.h
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c
create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c
diff --git a/drivers/crypto/marvell/octeontx2/Makefile b/drivers/crypto/marvell/octeontx2/Makefile
index e47a55961bb8..ef6fb2ab3571 100644
--- a/drivers/crypto/marvell/octeontx2/Makefile
+++ b/drivers/crypto/marvell/octeontx2/Makefile
@@ -1,7 +1,9 @@
# SPDX-License-Identifier: GPL-2.0-only
-obj-$(CONFIG_CRYPTO_DEV_OCTEONTX2_CPT) += octeontx2-cpt.o
+obj-$(CONFIG_CRYPTO_DEV_OCTEONTX2_CPT) += octeontx2-cpt.o octeontx2-cptvf.o
octeontx2-cpt-objs := otx2_cptpf_main.o otx2_cptpf_mbox.o \
otx2_cpt_mbox_common.o otx2_cptpf_ucode.o otx2_cptlf.o
+octeontx2-cptvf-objs := otx2_cptvf_main.o otx2_cptvf_mbox.o otx2_cptlf.o \
+ otx2_cpt_mbox_common.o
ccflags-y += -I$(srctree)/drivers/net/ethernet/marvell/octeontx2/af
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
index 705a0503b962..ca220178e518 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
+++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
@@ -115,5 +115,6 @@ int otx2_cpt_write_af_reg(struct otx2_mbox *mbox, struct pci_dev *pdev,
struct otx2_cptlfs_info;
int otx2_cpt_attach_rscrs_msg(struct otx2_cptlfs_info *lfs);
int otx2_cpt_detach_rsrcs_msg(struct otx2_cptlfs_info *lfs);
+int otx2_cpt_msix_offset_msg(struct otx2_cptlfs_info *lfs);
#endif /* __OTX2_CPT_COMMON_H */
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
index 0933031ac827..51cb6404ded7 100644
--- a/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
+++ b/drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
@@ -168,3 +168,35 @@ int otx2_cpt_detach_rsrcs_msg(struct otx2_cptlfs_info *lfs)
return ret;
}
+
+int otx2_cpt_msix_offset_msg(struct otx2_cptlfs_info *lfs)
+{
+ struct otx2_mbox *mbox = lfs->mbox;
+ struct pci_dev *pdev = lfs->pdev;
+ struct mbox_msghdr *req;
+ int ret, i;
+
+ req = otx2_mbox_alloc_msg_rsp(mbox, 0, sizeof(*req),
+ sizeof(struct msix_offset_rsp));
+ if (req == NULL) {
+ dev_err(&pdev->dev, "RVU MBOX failed to get message.\n");
+ return -EFAULT;
+ }
+
+ req->id = MBOX_MSG_MSIX_OFFSET;
+ req->sig = OTX2_MBOX_REQ_SIG;
+ req->pcifunc = 0;
+ ret = otx2_cpt_send_mbox_msg(mbox, pdev);
+ if (ret)
+ return ret;
+
+ for (i = 0; i < lfs->lfs_num; i++) {
+ if (lfs->lf[i].msix_offset == MSIX_VECTOR_INVALID) {
+ dev_err(&pdev->dev,
+ "Invalid msix offset %d for LF %d\n",
+ lfs->lf[i].msix_offset, i);
+ return -EINVAL;
+ }
+ }
+ return ret;
+}
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptvf.h b/drivers/crypto/marvell/octeontx2/otx2_cptvf.h
new file mode 100644
index 000000000000..4b01eb9d9f70
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptvf.h
@@ -0,0 +1,28 @@
+/* SPDX-License-Identifier: GPL-2.0-only
+ * Copyright (C) 2020 Marvell.
+ */
+
+#ifndef __OTX2_CPTVF_H
+#define __OTX2_CPTVF_H
+
+#include "mbox.h"
+#include "otx2_cptlf.h"
+
+struct otx2_cptvf_dev {
+ void __iomem *reg_base; /* Register start address */
+ void __iomem *pfvf_mbox_base; /* PF-VF mbox start address */
+ struct pci_dev *pdev; /* PCI device handle */
+ struct otx2_cptlfs_info lfs; /* CPT LFs attached to this VF */
+ u8 vf_id; /* Virtual function index */
+
+ /* PF <=> VF mbox */
+ struct otx2_mbox pfvf_mbox;
+ struct work_struct pfvf_mbox_work;
+ struct workqueue_struct *pfvf_mbox_wq;
+};
+
+irqreturn_t otx2_cptvf_pfvf_mbox_intr(int irq, void *arg);
+void otx2_cptvf_pfvf_mbox_handler(struct work_struct *work);
+int otx2_cptvf_send_eng_grp_num_msg(struct otx2_cptvf_dev *cptvf, int eng_type);
+
+#endif /* __OTX2_CPTVF_H */
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c b/drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c
new file mode 100644
index 000000000000..b90b265f0f42
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c
@@ -0,0 +1,194 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2020 Marvell. */
+
+#include "otx2_cpt_common.h"
+#include "otx2_cptvf.h"
+#include <rvu_reg.h>
+
+#define OTX2_CPTVF_DRV_NAME "octeontx2-cptvf"
+
+static void cptvf_enable_pfvf_mbox_intrs(struct otx2_cptvf_dev *cptvf)
+{
+ /* Clear interrupt if any */
+ otx2_cpt_write64(cptvf->reg_base, BLKADDR_RVUM, 0, OTX2_RVU_VF_INT,
+ 0x1ULL);
+
+ /* Enable PF-VF interrupt */
+ otx2_cpt_write64(cptvf->reg_base, BLKADDR_RVUM, 0,
+ OTX2_RVU_VF_INT_ENA_W1S, 0x1ULL);
+}
+
+static void cptvf_disable_pfvf_mbox_intrs(struct otx2_cptvf_dev *cptvf)
+{
+ /* Disable PF-VF interrupt */
+ otx2_cpt_write64(cptvf->reg_base, BLKADDR_RVUM, 0,
+ OTX2_RVU_VF_INT_ENA_W1C, 0x1ULL);
+
+ /* Clear interrupt if any */
+ otx2_cpt_write64(cptvf->reg_base, BLKADDR_RVUM, 0, OTX2_RVU_VF_INT,
+ 0x1ULL);
+}
+
+static int cptvf_register_interrupts(struct otx2_cptvf_dev *cptvf)
+{
+ int ret, irq;
+ u32 num_vec;
+
+ num_vec = pci_msix_vec_count(cptvf->pdev);
+ if (num_vec <= 0)
+ return -EINVAL;
+
+ /* Enable MSI-X */
+ ret = pci_alloc_irq_vectors(cptvf->pdev, num_vec, num_vec,
+ PCI_IRQ_MSIX);
+ if (ret < 0) {
+ dev_err(&cptvf->pdev->dev,
+ "Request for %d msix vectors failed\n", num_vec);
+ return ret;
+ }
+ irq = pci_irq_vector(cptvf->pdev, OTX2_CPT_VF_INT_VEC_E_MBOX);
+ /* Register VF<=>PF mailbox interrupt handler */
+ ret = devm_request_irq(&cptvf->pdev->dev, irq,
+ otx2_cptvf_pfvf_mbox_intr, 0,
+ "CPTPFVF Mbox", cptvf);
+ if (ret)
+ return ret;
+ /* Enable PF-VF mailbox interrupts */
+ cptvf_enable_pfvf_mbox_intrs(cptvf);
+
+ ret = otx2_cpt_send_ready_msg(&cptvf->pfvf_mbox, cptvf->pdev);
+ if (ret) {
+ dev_warn(&cptvf->pdev->dev,
+ "PF not responding to mailbox, deferring probe\n");
+ cptvf_disable_pfvf_mbox_intrs(cptvf);
+ return -EPROBE_DEFER;
+ }
+ return 0;
+}
+
+static int cptvf_pfvf_mbox_init(struct otx2_cptvf_dev *cptvf)
+{
+ int ret;
+
+ cptvf->pfvf_mbox_wq = alloc_workqueue("cpt_pfvf_mailbox",
+ WQ_UNBOUND | WQ_HIGHPRI |
+ WQ_MEM_RECLAIM, 1);
+ if (!cptvf->pfvf_mbox_wq)
+ return -ENOMEM;
+
+ ret = otx2_mbox_init(&cptvf->pfvf_mbox, cptvf->pfvf_mbox_base,
+ cptvf->pdev, cptvf->reg_base, MBOX_DIR_VFPF, 1);
+ if (ret)
+ goto free_wqe;
+
+ INIT_WORK(&cptvf->pfvf_mbox_work, otx2_cptvf_pfvf_mbox_handler);
+ return 0;
+free_wqe:
+ destroy_workqueue(cptvf->pfvf_mbox_wq);
+ return ret;
+}
+
+static void cptvf_pfvf_mbox_destroy(struct otx2_cptvf_dev *cptvf)
+{
+ destroy_workqueue(cptvf->pfvf_mbox_wq);
+ otx2_mbox_destroy(&cptvf->pfvf_mbox);
+}
+
+static int otx2_cptvf_probe(struct pci_dev *pdev,
+ const struct pci_device_id *ent)
+{
+ struct device *dev = &pdev->dev;
+ resource_size_t offset, size;
+ struct otx2_cptvf_dev *cptvf;
+ int ret;
+
+ cptvf = devm_kzalloc(dev, sizeof(*cptvf), GFP_KERNEL);
+ if (!cptvf)
+ return -ENOMEM;
+
+ ret = pcim_enable_device(pdev);
+ if (ret) {
+ dev_err(dev, "Failed to enable PCI device\n");
+ goto clear_drvdata;
+ }
+
+ ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(48));
+ if (ret) {
+ dev_err(dev, "Unable to get usable DMA configuration\n");
+ goto clear_drvdata;
+ }
+ /* Map VF's configuration registers */
+ ret = pcim_iomap_regions_request_all(pdev, 1 << PCI_PF_REG_BAR_NUM,
+ OTX2_CPTVF_DRV_NAME);
+ if (ret) {
+ dev_err(dev, "Couldn't get PCI resources 0x%x\n", ret);
+ goto clear_drvdata;
+ }
+ pci_set_master(pdev);
+ pci_set_drvdata(pdev, cptvf);
+ cptvf->pdev = pdev;
+
+ cptvf->reg_base = pcim_iomap_table(pdev)[PCI_PF_REG_BAR_NUM];
+
+ offset = pci_resource_start(pdev, PCI_MBOX_BAR_NUM);
+ size = pci_resource_len(pdev, PCI_MBOX_BAR_NUM);
+ /* Map PF-VF mailbox memory */
+ cptvf->pfvf_mbox_base = devm_ioremap_wc(dev, offset, size);
+ if (!cptvf->pfvf_mbox_base) {
+ dev_err(&pdev->dev, "Unable to map BAR4\n");
+ ret = -ENODEV;
+ goto clear_drvdata;
+ }
+ /* Initialize PF<=>VF mailbox */
+ ret = cptvf_pfvf_mbox_init(cptvf);
+ if (ret)
+ goto clear_drvdata;
+
+ /* Register interrupts */
+ ret = cptvf_register_interrupts(cptvf);
+ if (ret)
+ goto destroy_pfvf_mbox;
+
+ return 0;
+destroy_pfvf_mbox:
+ cptvf_pfvf_mbox_destroy(cptvf);
+clear_drvdata:
+ pci_set_drvdata(pdev, NULL);
+
+ return ret;
+}
+
+static void otx2_cptvf_remove(struct pci_dev *pdev)
+{
+ struct otx2_cptvf_dev *cptvf = pci_get_drvdata(pdev);
+
+ if (!cptvf) {
+ dev_err(&pdev->dev, "Invalid CPT VF device.\n");
+ return;
+ }
+ /* Disable PF-VF mailbox interrupt */
+ cptvf_disable_pfvf_mbox_intrs(cptvf);
+ /* Destroy PF-VF mbox */
+ cptvf_pfvf_mbox_destroy(cptvf);
+ pci_set_drvdata(pdev, NULL);
+}
+
+/* Supported devices */
+static const struct pci_device_id otx2_cptvf_id_table[] = {
+ {PCI_VDEVICE(CAVIUM, OTX2_CPT_PCI_VF_DEVICE_ID), 0},
+ { 0, } /* end of table */
+};
+
+static struct pci_driver otx2_cptvf_pci_driver = {
+ .name = OTX2_CPTVF_DRV_NAME,
+ .id_table = otx2_cptvf_id_table,
+ .probe = otx2_cptvf_probe,
+ .remove = otx2_cptvf_remove,
+};
+
+module_pci_driver(otx2_cptvf_pci_driver);
+
+MODULE_AUTHOR("Marvell");
+MODULE_DESCRIPTION("Marvell OcteonTX2 CPT Virtual Function Driver");
+MODULE_LICENSE("GPL v2");
+MODULE_DEVICE_TABLE(pci, otx2_cptvf_id_table);
diff --git a/drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c b/drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c
new file mode 100644
index 000000000000..417099a86742
--- /dev/null
+++ b/drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c
@@ -0,0 +1,113 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2020 Marvell. */
+
+#include "otx2_cpt_common.h"
+#include "otx2_cptvf.h"
+#include <rvu_reg.h>
+
+irqreturn_t otx2_cptvf_pfvf_mbox_intr(int __always_unused irq, void *arg)
+{
+ struct otx2_cptvf_dev *cptvf = arg;
+ u64 intr;
+
+ /* Read the interrupt bits */
+ intr = otx2_cpt_read64(cptvf->reg_base, BLKADDR_RVUM, 0,
+ OTX2_RVU_VF_INT);
+
+ if (intr & 0x1ULL) {
+ /* Schedule work queue function to process the MBOX request */
+ queue_work(cptvf->pfvf_mbox_wq, &cptvf->pfvf_mbox_work);
+ /* Clear and ack the interrupt */
+ otx2_cpt_write64(cptvf->reg_base, BLKADDR_RVUM, 0,
+ OTX2_RVU_VF_INT, 0x1ULL);
+ }
+ return IRQ_HANDLED;
+}
+
+static void process_pfvf_mbox_mbox_msg(struct otx2_cptvf_dev *cptvf,
+ struct mbox_msghdr *msg)
+{
+ struct otx2_cptlfs_info *lfs = &cptvf->lfs;
+ struct cpt_rd_wr_reg_msg *rsp_reg;
+ struct msix_offset_rsp *rsp_msix;
+ int i;
+
+ if (msg->id >= MBOX_MSG_MAX) {
+ dev_err(&cptvf->pdev->dev,
+ "MBOX msg with unknown ID %d\n", msg->id);
+ return;
+ }
+ if (msg->sig != OTX2_MBOX_RSP_SIG) {
+ dev_err(&cptvf->pdev->dev,
+ "MBOX msg with wrong signature %x, ID %d\n",
+ msg->sig, msg->id);
+ return;
+ }
+ switch (msg->id) {
+ case MBOX_MSG_READY:
+ cptvf->vf_id = ((msg->pcifunc >> RVU_PFVF_FUNC_SHIFT)
+ & RVU_PFVF_FUNC_MASK) - 1;
+ break;
+ case MBOX_MSG_ATTACH_RESOURCES:
+ /* Check if resources were successfully attached */
+ if (!msg->rc)
+ lfs->are_lfs_attached = 1;
+ break;
+ case MBOX_MSG_DETACH_RESOURCES:
+ /* Check if resources were successfully detached */
+ if (!msg->rc)
+ lfs->are_lfs_attached = 0;
+ break;
+ case MBOX_MSG_MSIX_OFFSET:
+ rsp_msix = (struct msix_offset_rsp *) msg;
+ for (i = 0; i < rsp_msix->cptlfs; i++)
+ lfs->lf[i].msix_offset = rsp_msix->cptlf_msixoff[i];
+ break;
+ case MBOX_MSG_CPT_RD_WR_REGISTER:
+ rsp_reg = (struct cpt_rd_wr_reg_msg *) msg;
+ if (msg->rc) {
+ dev_err(&cptvf->pdev->dev,
+ "Reg %llx rd/wr(%d) failed %d\n",
+ rsp_reg->reg_offset, rsp_reg->is_write,
+ msg->rc);
+ return;
+ }
+ if (!rsp_reg->is_write)
+ *rsp_reg->ret_val = rsp_reg->val;
+ break;
+ default:
+ dev_err(&cptvf->pdev->dev, "Unsupported msg %d received.\n",
+ msg->id);
+ break;
+ }
+}
+
+void otx2_cptvf_pfvf_mbox_handler(struct work_struct *work)
+{
+ struct otx2_cptvf_dev *cptvf;
+ struct otx2_mbox *pfvf_mbox;
+ struct otx2_mbox_dev *mdev;
+ struct mbox_hdr *rsp_hdr;
+ struct mbox_msghdr *msg;
+ int offset, i;
+
+ /* sync with mbox memory region */
+ smp_rmb();
+
+ cptvf = container_of(work, struct otx2_cptvf_dev, pfvf_mbox_work);
+ pfvf_mbox = &cptvf->pfvf_mbox;
+ mdev = &pfvf_mbox->dev[0];
+ rsp_hdr = (struct mbox_hdr *)(mdev->mbase + pfvf_mbox->rx_start);
+ if (rsp_hdr->num_msgs == 0)
+ return;
+ offset = ALIGN(sizeof(struct mbox_hdr), MBOX_MSG_ALIGN);
+
+ for (i = 0; i < rsp_hdr->num_msgs; i++) {
+ msg = (struct mbox_msghdr *)(mdev->mbase + pfvf_mbox->rx_start +
+ offset);
+ process_pfvf_mbox_mbox_msg(cptvf, msg);
+ offset = msg->next_msgoff;
+ mdev->msgs_acked++;
+ }
+ otx2_mbox_reset(pfvf_mbox, 0);
+}
--
2.28.0
On OcteonTX2 SoC, the admin function (AF) is the only one with all
priviliges to configure HW and alloc resources, PFs and it's VFs
have to request AF via mailbox for all their needs. This patch adds
a mailbox interface for CPT PFs and VFs to allocate resources
for cryptography.
Signed-off-by: Suheil Chandran <[email protected]>
Signed-off-by: Srujana Challa <[email protected]>
---
.../ethernet/marvell/octeontx2/af/Makefile | 3 +-
.../net/ethernet/marvell/octeontx2/af/mbox.h | 33 +++
.../net/ethernet/marvell/octeontx2/af/rvu.c | 2 +-
.../net/ethernet/marvell/octeontx2/af/rvu.h | 1 +
.../ethernet/marvell/octeontx2/af/rvu_cpt.c | 229 ++++++++++++++++++
.../ethernet/marvell/octeontx2/af/rvu_reg.h | 63 ++++-
6 files changed, 323 insertions(+), 8 deletions(-)
create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/Makefile b/drivers/net/ethernet/marvell/octeontx2/af/Makefile
index 2f7a861d0c7b..d258b9a8cefb 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/Makefile
+++ b/drivers/net/ethernet/marvell/octeontx2/af/Makefile
@@ -9,4 +9,5 @@ obj-$(CONFIG_OCTEONTX2_AF) += octeontx2_af.o
octeontx2_mbox-y := mbox.o rvu_trace.o
octeontx2_af-y := cgx.o rvu.o rvu_cgx.o rvu_npa.o rvu_nix.o \
- rvu_reg.o rvu_npc.o rvu_debugfs.o ptp.o
+ rvu_reg.o rvu_npc.o rvu_debugfs.o ptp.o \
+ rvu_cpt.o
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/mbox.h b/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
index 263a21129416..56b29fb49037 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
@@ -158,6 +158,11 @@ M(NPA_HWCTX_DISABLE, 0x403, npa_hwctx_disable, hwctx_disable_req, msg_rsp)\
/* SSO/SSOW mbox IDs (range 0x600 - 0x7FF) */ \
/* TIM mbox IDs (range 0x800 - 0x9FF) */ \
/* CPT mbox IDs (range 0xA00 - 0xBFF) */ \
+M(CPT_LF_ALLOC, 0xA00, cpt_lf_alloc, cpt_lf_alloc_req_msg, \
+ msg_rsp) \
+M(CPT_LF_FREE, 0xA01, cpt_lf_free, msg_req, msg_rsp) \
+M(CPT_RD_WR_REGISTER, 0xA02, cpt_rd_wr_register, cpt_rd_wr_reg_msg, \
+ cpt_rd_wr_reg_msg) \
/* NPC mbox IDs (range 0x6000 - 0x7FFF) */ \
M(NPC_MCAM_ALLOC_ENTRY, 0x6000, npc_mcam_alloc_entry, npc_mcam_alloc_entry_req,\
npc_mcam_alloc_entry_rsp) \
@@ -881,4 +886,32 @@ struct ptp_rsp {
u64 clk;
};
+/* CPT mailbox error codes
+ * Range 901 - 1000.
+ */
+enum cpt_af_status {
+ CPT_AF_ERR_PARAM = -901,
+ CPT_AF_ERR_GRP_INVALID = -902,
+ CPT_AF_ERR_LF_INVALID = -903,
+ CPT_AF_ERR_ACCESS_DENIED = -904,
+ CPT_AF_ERR_SSO_PF_FUNC_INVALID = -905,
+ CPT_AF_ERR_NIX_PF_FUNC_INVALID = -906
+};
+
+/* CPT mbox message formats */
+struct cpt_rd_wr_reg_msg {
+ struct mbox_msghdr hdr;
+ u64 reg_offset;
+ u64 *ret_val;
+ u64 val;
+ u8 is_write;
+};
+
+struct cpt_lf_alloc_req_msg {
+ struct mbox_msghdr hdr;
+ u16 nix_pf_func;
+ u16 sso_pf_func;
+ u16 eng_grpmsk;
+};
+
#endif /* MBOX_H */
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu.c
index e1f918960730..3e24897ed9b5 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu.c
@@ -1025,7 +1025,7 @@ int rvu_mbox_handler_ready(struct rvu *rvu, struct msg_req *req,
/* Get current count of a RVU block's LF/slots
* provisioned to a given RVU func.
*/
-static u16 rvu_get_rsrc_mapcount(struct rvu_pfvf *pfvf, int blktype)
+u16 rvu_get_rsrc_mapcount(struct rvu_pfvf *pfvf, int blktype)
{
switch (blktype) {
case BLKTYPE_NPA:
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
index 90eed3160915..c37e106d7006 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
@@ -429,6 +429,7 @@ int rvu_get_lf(struct rvu *rvu, struct rvu_block *block, u16 pcifunc, u16 slot);
int rvu_lf_reset(struct rvu *rvu, struct rvu_block *block, int lf);
int rvu_get_blkaddr(struct rvu *rvu, int blktype, u16 pcifunc);
int rvu_poll_reg(struct rvu *rvu, u64 block, u64 offset, u64 mask, bool zero);
+u16 rvu_get_rsrc_mapcount(struct rvu_pfvf *pfvf, int blktype);
/* RVU HW reg validation */
enum regmap_block {
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c
new file mode 100644
index 000000000000..e53d0a62f507
--- /dev/null
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c
@@ -0,0 +1,229 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2020 Marvell. */
+
+#include <linux/pci.h>
+#include "rvu_struct.h"
+#include "rvu_reg.h"
+#include "mbox.h"
+#include "rvu.h"
+
+/* CPT PF device id */
+#define PCI_DEVID_OTX2_CPT_PF 0xA0FD
+
+/* Maximum supported microcode groups */
+#define CPT_MAX_ENGINE_GROUPS 8
+
+static int get_cpt_pf_num(struct rvu *rvu)
+{
+ int i, domain_nr, cpt_pf_num = -1;
+ struct pci_dev *pdev;
+
+ domain_nr = pci_domain_nr(rvu->pdev->bus);
+ for (i = 0; i < rvu->hw->total_pfs; i++) {
+ pdev = pci_get_domain_bus_and_slot(domain_nr, i + 1, 0);
+ if (!pdev)
+ continue;
+
+ if (pdev->device == PCI_DEVID_OTX2_CPT_PF) {
+ cpt_pf_num = i;
+ put_device(&pdev->dev);
+ break;
+ }
+ put_device(&pdev->dev);
+ }
+ return cpt_pf_num;
+}
+
+static bool is_cpt_pf(struct rvu *rvu, u16 pcifunc)
+{
+ int cpt_pf_num = get_cpt_pf_num(rvu);
+
+ if (rvu_get_pf(pcifunc) != cpt_pf_num)
+ return false;
+ if (pcifunc & RVU_PFVF_FUNC_MASK)
+ return false;
+
+ return true;
+}
+
+static bool is_cpt_vf(struct rvu *rvu, u16 pcifunc)
+{
+ int cpt_pf_num = get_cpt_pf_num(rvu);
+
+ if (rvu_get_pf(pcifunc) != cpt_pf_num)
+ return false;
+ if (!(pcifunc & RVU_PFVF_FUNC_MASK))
+ return false;
+
+ return true;
+}
+
+int rvu_mbox_handler_cpt_lf_alloc(struct rvu *rvu,
+ struct cpt_lf_alloc_req_msg *req,
+ struct msg_rsp *rsp)
+{
+ u16 pcifunc = req->hdr.pcifunc;
+ struct rvu_block *block;
+ int cptlf, blkaddr;
+ int num_lfs, slot;
+ u64 val;
+
+ if (req->eng_grpmsk == 0x0)
+ return CPT_AF_ERR_GRP_INVALID;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, pcifunc);
+ if (blkaddr < 0)
+ return blkaddr;
+
+ block = &rvu->hw->block[blkaddr];
+ num_lfs = rvu_get_rsrc_mapcount(rvu_get_pfvf(rvu, pcifunc),
+ block->type);
+ if (!num_lfs)
+ return CPT_AF_ERR_LF_INVALID;
+
+ /* Check if requested 'CPTLF <=> NIXLF' mapping is valid */
+ if (req->nix_pf_func) {
+ /* If default, use 'this' CPTLF's PFFUNC */
+ if (req->nix_pf_func == RVU_DEFAULT_PF_FUNC)
+ req->nix_pf_func = pcifunc;
+ if (!is_pffunc_map_valid(rvu, req->nix_pf_func, BLKTYPE_NIX))
+ return CPT_AF_ERR_NIX_PF_FUNC_INVALID;
+ }
+
+ /* Check if requested 'CPTLF <=> SSOLF' mapping is valid */
+ if (req->sso_pf_func) {
+ /* If default, use 'this' CPTLF's PFFUNC */
+ if (req->sso_pf_func == RVU_DEFAULT_PF_FUNC)
+ req->sso_pf_func = pcifunc;
+ if (!is_pffunc_map_valid(rvu, req->sso_pf_func, BLKTYPE_SSO))
+ return CPT_AF_ERR_SSO_PF_FUNC_INVALID;
+ }
+
+ for (slot = 0; slot < num_lfs; slot++) {
+ cptlf = rvu_get_lf(rvu, block, pcifunc, slot);
+ if (cptlf < 0)
+ return CPT_AF_ERR_LF_INVALID;
+
+ /* Set CPT LF group and priority */
+ val = (u64)req->eng_grpmsk << 48 | 1;
+ rvu_write64(rvu, blkaddr, CPT_AF_LFX_CTL(cptlf), val);
+
+ /* Set CPT LF NIX_PF_FUNC and SSO_PF_FUNC */
+ val = (u64)req->nix_pf_func << 48 |
+ (u64)req->sso_pf_func << 32;
+ rvu_write64(rvu, blkaddr, CPT_AF_LFX_CTL2(cptlf), val);
+ }
+
+ return 0;
+}
+
+int rvu_mbox_handler_cpt_lf_free(struct rvu *rvu, struct msg_req *req,
+ struct msg_rsp *rsp)
+{
+ u16 pcifunc = req->hdr.pcifunc;
+ struct rvu_block *block;
+ int cptlf, blkaddr;
+ int num_lfs, slot;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, pcifunc);
+ if (blkaddr < 0)
+ return blkaddr;
+
+ block = &rvu->hw->block[blkaddr];
+ num_lfs = rvu_get_rsrc_mapcount(rvu_get_pfvf(rvu, pcifunc),
+ block->type);
+ if (!num_lfs)
+ return CPT_AF_ERR_LF_INVALID;
+
+ for (slot = 0; slot < num_lfs; slot++) {
+ cptlf = rvu_get_lf(rvu, block, pcifunc, slot);
+ if (cptlf < 0)
+ return CPT_AF_ERR_LF_INVALID;
+
+ /* Reset CPT LF group and priority */
+ rvu_write64(rvu, blkaddr, CPT_AF_LFX_CTL(cptlf), 0x0);
+ /* Reset CPT LF NIX_PF_FUNC and SSO_PF_FUNC */
+ rvu_write64(rvu, blkaddr, CPT_AF_LFX_CTL2(cptlf), 0x0);
+ }
+
+ return 0;
+}
+
+int rvu_mbox_handler_cpt_rd_wr_register(struct rvu *rvu,
+ struct cpt_rd_wr_reg_msg *req,
+ struct cpt_rd_wr_reg_msg *rsp)
+{
+ int blkaddr, num_lfs, offs, lf;
+ struct rvu_block *block;
+
+ blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
+ if (blkaddr < 0)
+ return blkaddr;
+
+ /* This message is accepted only if sent from CPT PF/VF */
+ if (!is_cpt_pf(rvu, req->hdr.pcifunc) &&
+ !is_cpt_vf(rvu, req->hdr.pcifunc))
+ return CPT_AF_ERR_ACCESS_DENIED;
+
+ rsp->reg_offset = req->reg_offset;
+ rsp->ret_val = req->ret_val;
+ rsp->is_write = req->is_write;
+
+ /* Registers that can be accessed from PF/VF */
+ if ((req->reg_offset & 0xFF000) == CPT_AF_LFX_CTL(0) ||
+ (req->reg_offset & 0xFF000) == CPT_AF_LFX_CTL2(0)) {
+ offs = req->reg_offset & 0xFFF;
+ if (offs % 8)
+ return CPT_AF_ERR_ACCESS_DENIED;
+ lf = offs >> 3;
+ block = &rvu->hw->block[blkaddr];
+ num_lfs = rvu_get_rsrc_mapcount(rvu_get_pfvf(rvu,
+ req->hdr.pcifunc),
+ block->type);
+ if (lf >= num_lfs)
+ /* Slot is not valid for that PF/VF */
+ return CPT_AF_ERR_ACCESS_DENIED;
+
+ /* Need to translate CPT LF slot to global number because
+ * VFs use local numbering from 0 to number of LFs - 1
+ */
+ lf = rvu_get_lf(rvu, &rvu->hw->block[blkaddr],
+ req->hdr.pcifunc, lf);
+ if (lf < 0)
+ return CPT_AF_ERR_ACCESS_DENIED;
+
+ req->reg_offset &= 0xFF000;
+ req->reg_offset += lf << 3;
+ rsp->reg_offset = req->reg_offset;
+ } else if (!(req->hdr.pcifunc & RVU_PFVF_FUNC_MASK)) {
+ /* Registers that can be accessed from PF */
+ switch (req->reg_offset & 0xFF000) {
+ case CPT_AF_PF_FUNC:
+ case CPT_AF_BLK_RST:
+ case CPT_AF_CONSTANTS1:
+ if (req->reg_offset & 0xFFF)
+ return CPT_AF_ERR_ACCESS_DENIED;
+ break;
+
+ case CPT_AF_EXEX_STS(0):
+ case CPT_AF_EXEX_CTL(0):
+ case CPT_AF_EXEX_CTL2(0):
+ case CPT_AF_EXEX_UCODE_BASE(0):
+ offs = req->reg_offset & 0xFFF;
+ if ((offs % 8) || (offs >> 3) > 127)
+ return CPT_AF_ERR_ACCESS_DENIED;
+ break;
+ default:
+ return CPT_AF_ERR_ACCESS_DENIED;
+ }
+ } else {
+ return CPT_AF_ERR_ACCESS_DENIED;
+ }
+
+ if (req->is_write)
+ rvu_write64(rvu, blkaddr, req->reg_offset, req->val);
+ else
+ rsp->val = rvu_read64(rvu, blkaddr, req->reg_offset);
+
+ return 0;
+}
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
index 7ca599b973c0..807b1c1a9d85 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
@@ -429,12 +429,63 @@
#define TIM_AF_LF_RST (0x20)
/* CPT */
-#define CPT_AF_CONSTANTS0 (0x0000)
-#define CPT_PRIV_LFX_CFG (0x41000)
-#define CPT_PRIV_LFX_INT_CFG (0x43000)
-#define CPT_AF_RVU_LF_CFG_DEBUG (0x45000)
-#define CPT_AF_LF_RST (0x44000)
-#define CPT_AF_BLK_RST (0x46000)
+#define CPT_AF_CONSTANTS0 (0x0000)
+#define CPT_AF_CONSTANTS1 (0x1000)
+#define CPT_AF_DIAG (0x3000)
+#define CPT_AF_ECO (0x4000)
+#define CPT_AF_FLTX_INT(a) (0xa000ull | (u64)(a) << 3)
+#define CPT_AF_FLTX_INT_W1S(a) (0xb000ull | (u64)(a) << 3)
+#define CPT_AF_FLTX_INT_ENA_W1C(a) (0xc000ull | (u64)(a) << 3)
+#define CPT_AF_FLTX_INT_ENA_W1S(a) (0xd000ull | (u64)(a) << 3)
+#define CPT_AF_PSNX_EXE(a) (0xe000ull | (u64)(a) << 3)
+#define CPT_AF_PSNX_EXE_W1S(a) (0xf000ull | (u64)(a) << 3)
+#define CPT_AF_PSNX_LF(a) (0x10000ull | (u64)(a) << 3)
+#define CPT_AF_PSNX_LF_W1S(a) (0x11000ull | (u64)(a) << 3)
+#define CPT_AF_EXEX_CTL2(a) (0x12000ull | (u64)(a) << 3)
+#define CPT_AF_EXEX_STS(a) (0x13000ull | (u64)(a) << 3)
+#define CPT_AF_EXE_ERR_INFO (0x14000)
+#define CPT_AF_EXEX_ACTIVE(a) (0x16000ull | (u64)(a) << 3)
+#define CPT_AF_INST_REQ_PC (0x17000)
+#define CPT_AF_INST_LATENCY_PC (0x18000)
+#define CPT_AF_RD_REQ_PC (0x19000)
+#define CPT_AF_RD_LATENCY_PC (0x1a000)
+#define CPT_AF_RD_UC_PC (0x1b000)
+#define CPT_AF_ACTIVE_CYCLES_PC (0x1c000)
+#define CPT_AF_EXE_DBG_CTL (0x1d000)
+#define CPT_AF_EXE_DBG_DATA (0x1e000)
+#define CPT_AF_EXE_REQ_TIMER (0x1f000)
+#define CPT_AF_EXEX_CTL(a) (0x20000ull | (u64)(a) << 3)
+#define CPT_AF_EXE_PERF_CTL (0x21000)
+#define CPT_AF_EXE_DBG_CNTX(a) (0x22000ull | (u64)(a) << 3)
+#define CPT_AF_EXE_PERF_EVENT_CNT (0x23000)
+#define CPT_AF_EXE_EPCI_INBX_CNT(a) (0x24000ull | (u64)(a) << 3)
+#define CPT_AF_EXE_EPCI_OUTBX_CNT(a) (0x25000ull | (u64)(a) << 3)
+#define CPT_AF_EXEX_UCODE_BASE(a) (0x26000ull | (u64)(a) << 3)
+#define CPT_AF_LFX_CTL(a) (0x27000ull | (u64)(a) << 3)
+#define CPT_AF_LFX_CTL2(a) (0x29000ull | (u64)(a) << 3)
+#define CPT_AF_CPTCLK_CNT (0x2a000)
+#define CPT_AF_PF_FUNC (0x2b000)
+#define CPT_AF_LFX_PTR_CTL(a) (0x2c000ull | (u64)(a) << 3)
+#define CPT_AF_GRPX_THR(a) (0x2d000ull | (u64)(a) << 3)
+#define CPT_AF_CTL (0x2e000ull)
+#define CPT_AF_XEX_THR(a) (0x2f000ull | (u64)(a) << 3)
+#define CPT_PRIV_LFX_CFG (0x41000)
+#define CPT_PRIV_AF_INT_CFG (0x42000)
+#define CPT_PRIV_LFX_INT_CFG (0x43000)
+#define CPT_AF_LF_RST (0x44000)
+#define CPT_AF_RVU_LF_CFG_DEBUG (0x45000)
+#define CPT_AF_BLK_RST (0x46000)
+#define CPT_AF_RVU_INT (0x47000)
+#define CPT_AF_RVU_INT_W1S (0x47008)
+#define CPT_AF_RVU_INT_ENA_W1S (0x47010)
+#define CPT_AF_RVU_INT_ENA_W1C (0x47018)
+#define CPT_AF_RAS_INT (0x47020)
+#define CPT_AF_RAS_INT_W1S (0x47028)
+#define CPT_AF_RAS_INT_ENA_W1S (0x47030)
+#define CPT_AF_RAS_INT_ENA_W1C (0x47038)
+
+#define CPT_AF_LF_CTL2_SHIFT 3
+#define CPT_AF_LF_SSO_PF_FUNC_SHIFT 32
#define NPC_AF_BLK_RST (0x00040)
--
2.28.0
On Wed, Oct 28, 2020 at 10:44 PM Srujana Challa <[email protected]> wrote:
>
> On OcteonTX2 SoC, the admin function (AF) is the only one with all
> priviliges to configure HW and alloc resources, PFs and it's VFs
> have to request AF via mailbox for all their needs. This patch adds
> a mailbox interface for CPT PFs and VFs to allocate resources
> for cryptography.
>
> Signed-off-by: Suheil Chandran <[email protected]>
> Signed-off-by: Srujana Challa <[email protected]>
> ---
> .../ethernet/marvell/octeontx2/af/Makefile | 3 +-
> .../net/ethernet/marvell/octeontx2/af/mbox.h | 33 +++
> .../net/ethernet/marvell/octeontx2/af/rvu.c | 2 +-
> .../net/ethernet/marvell/octeontx2/af/rvu.h | 1 +
> .../ethernet/marvell/octeontx2/af/rvu_cpt.c | 229 ++++++++++++++++++
> .../ethernet/marvell/octeontx2/af/rvu_reg.h | 63 ++++-
> 6 files changed, 323 insertions(+), 8 deletions(-)
> create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c
> diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
> index 7ca599b973c0..807b1c1a9d85 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
> +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
> @@ -429,12 +429,63 @@
> #define TIM_AF_LF_RST (0x20)
>
> /* CPT */
> -#define CPT_AF_CONSTANTS0 (0x0000)
> -#define CPT_PRIV_LFX_CFG (0x41000)
> -#define CPT_PRIV_LFX_INT_CFG (0x43000)
> -#define CPT_AF_RVU_LF_CFG_DEBUG (0x45000)
> -#define CPT_AF_LF_RST (0x44000)
> -#define CPT_AF_BLK_RST (0x46000)
> +#define CPT_AF_CONSTANTS0 (0x0000)
> +#define CPT_AF_CONSTANTS1 (0x1000)
> +#define CPT_AF_DIAG (0x3000)
> +#define CPT_AF_ECO (0x4000)
> +#define CPT_AF_FLTX_INT(a) (0xa000ull | (u64)(a) << 3)
> +#define CPT_AF_FLTX_INT_W1S(a) (0xb000ull | (u64)(a) << 3)
> +#define CPT_AF_FLTX_INT_ENA_W1C(a) (0xc000ull | (u64)(a) << 3)
> +#define CPT_AF_FLTX_INT_ENA_W1S(a) (0xd000ull | (u64)(a) << 3)
> +#define CPT_AF_PSNX_EXE(a) (0xe000ull | (u64)(a) << 3)
> +#define CPT_AF_PSNX_EXE_W1S(a) (0xf000ull | (u64)(a) << 3)
> +#define CPT_AF_PSNX_LF(a) (0x10000ull | (u64)(a) << 3)
> +#define CPT_AF_PSNX_LF_W1S(a) (0x11000ull | (u64)(a) << 3)
> +#define CPT_AF_EXEX_CTL2(a) (0x12000ull | (u64)(a) << 3)
> +#define CPT_AF_EXEX_STS(a) (0x13000ull | (u64)(a) << 3)
> +#define CPT_AF_EXE_ERR_INFO (0x14000)
> +#define CPT_AF_EXEX_ACTIVE(a) (0x16000ull | (u64)(a) << 3)
> +#define CPT_AF_INST_REQ_PC (0x17000)
> +#define CPT_AF_INST_LATENCY_PC (0x18000)
> +#define CPT_AF_RD_REQ_PC (0x19000)
> +#define CPT_AF_RD_LATENCY_PC (0x1a000)
> +#define CPT_AF_RD_UC_PC (0x1b000)
> +#define CPT_AF_ACTIVE_CYCLES_PC (0x1c000)
> +#define CPT_AF_EXE_DBG_CTL (0x1d000)
> +#define CPT_AF_EXE_DBG_DATA (0x1e000)
> +#define CPT_AF_EXE_REQ_TIMER (0x1f000)
> +#define CPT_AF_EXEX_CTL(a) (0x20000ull | (u64)(a) << 3)
> +#define CPT_AF_EXE_PERF_CTL (0x21000)
> +#define CPT_AF_EXE_DBG_CNTX(a) (0x22000ull | (u64)(a) << 3)
> +#define CPT_AF_EXE_PERF_EVENT_CNT (0x23000)
> +#define CPT_AF_EXE_EPCI_INBX_CNT(a) (0x24000ull | (u64)(a) << 3)
> +#define CPT_AF_EXE_EPCI_OUTBX_CNT(a) (0x25000ull | (u64)(a) << 3)
> +#define CPT_AF_EXEX_UCODE_BASE(a) (0x26000ull | (u64)(a) << 3)
> +#define CPT_AF_LFX_CTL(a) (0x27000ull | (u64)(a) << 3)
> +#define CPT_AF_LFX_CTL2(a) (0x29000ull | (u64)(a) << 3)
> +#define CPT_AF_CPTCLK_CNT (0x2a000)
> +#define CPT_AF_PF_FUNC (0x2b000)
> +#define CPT_AF_LFX_PTR_CTL(a) (0x2c000ull | (u64)(a) << 3)
> +#define CPT_AF_GRPX_THR(a) (0x2d000ull | (u64)(a) << 3)
> +#define CPT_AF_CTL (0x2e000ull)
> +#define CPT_AF_XEX_THR(a) (0x2f000ull | (u64)(a) << 3)
> +#define CPT_PRIV_LFX_CFG (0x41000)
> +#define CPT_PRIV_AF_INT_CFG (0x42000)
> +#define CPT_PRIV_LFX_INT_CFG (0x43000)
> +#define CPT_AF_LF_RST (0x44000)
> +#define CPT_AF_RVU_LF_CFG_DEBUG (0x45000)
> +#define CPT_AF_BLK_RST (0x46000)
> +#define CPT_AF_RVU_INT (0x47000)
> +#define CPT_AF_RVU_INT_W1S (0x47008)
> +#define CPT_AF_RVU_INT_ENA_W1S (0x47010)
> +#define CPT_AF_RVU_INT_ENA_W1C (0x47018)
> +#define CPT_AF_RAS_INT (0x47020)
> +#define CPT_AF_RAS_INT_W1S (0x47028)
> +#define CPT_AF_RAS_INT_ENA_W1S (0x47030)
> +#define CPT_AF_RAS_INT_ENA_W1C (0x47038)
> +
> +#define CPT_AF_LF_CTL2_SHIFT 3
> +#define CPT_AF_LF_SSO_PF_FUNC_SHIFT 32
>
> #define NPC_AF_BLK_RST (0x00040)
Most of these are not related to this patch. Used only in the
follow-on debug interface one?
On Wed, Oct 28, 2020 at 10:22 PM Srujana Challa <[email protected]> wrote:
>
> Add entries to debugfs at /sys/kernel/debug/octeontx2/cpt.
>
> cpt_pc: dump cpt performance HW registers.
> Usage:
> cat /sys/kernel/debug/octeontx2/cpt/cpt_pc
>
> cpt_ae_sts: show cpt asymmetric engines current state
> Usage:
> cat /sys/kernel/debug/octeontx2/cpt/cpt_ae_sts
>
> cpt_se_sts: show cpt symmetric engines current state
> Usage:
> cat /sys/kernel/debug/octeontx2/cpt/cpt_se_sts
>
> cpt_engines_info: dump cpt engine control registers.
> Usage:
> cat /sys/kernel/debug/octeontx2/cpt/cpt_engines_info
>
> cpt_lfs_info: dump cpt lfs control registers.
> Usage:
> cat /sys/kernel/debug/octeontx2/cpt/cpt_lfs_info
>
> cpt_err_info: dump cpt error registers.
> Usage:
> cat /sys/kernel/debug/octeontx2/cpt/cpt_err_info
>
> Signed-off-by: Suheil Chandran <[email protected]>
> Signed-off-by: Srujana Challa <[email protected]>
> ---
> .../net/ethernet/marvell/octeontx2/af/rvu.h | 1 +
> .../marvell/octeontx2/af/rvu_debugfs.c | 304 ++++++++++++++++++
> 2 files changed, 305 insertions(+)
>
> diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
> index c37e106d7006..ba18171c87d6 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
> +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
> @@ -50,6 +50,7 @@ struct rvu_debugfs {
> struct dentry *npa;
> struct dentry *nix;
> struct dentry *npc;
> + struct dentry *cpt;
> struct dump_ctx npa_aura_ctx;
> struct dump_ctx npa_pool_ctx;
> struct dump_ctx nix_cq_ctx;
> diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
> index 77adad4adb1b..24354bfb4e94 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
> +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
> @@ -1676,6 +1676,309 @@ static void rvu_dbg_npc_init(struct rvu *rvu)
> debugfs_remove_recursive(rvu->rvu_dbg.npc);
> }
>
> +/* CPT debugfs APIs */
> +static int rvu_dbg_cpt_ae_sts_display(struct seq_file *filp, void *unused)
> +{
> + struct rvu *rvu = filp->private;
> + u64 busy_sts = 0, free_sts = 0;
> + u32 e_min = 0, e_max = 0, e, i;
> + u16 max_ses, max_ies, max_aes;
> + int blkaddr;
> + u64 reg;
> +
> + blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
> + if (blkaddr < 0)
> + return -ENODEV;
> +
> + reg = rvu_read64(rvu, blkaddr, CPT_AF_CONSTANTS1);
> + max_ses = reg & 0xffff;
> + max_ies = (reg >> 16) & 0xffff;
> + max_aes = (reg >> 32) & 0xffff;
> +
> + e_min = max_ses + max_ies;
> + e_max = max_ses + max_ies + max_aes;
> +
> + for (e = e_min, i = 0; e < e_max; e++, i++) {
> + reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_STS(e));
> + if (reg & 0x1)
> + busy_sts |= 1ULL << i;
> +
> + if (reg & 0x2)
> + free_sts |= 1ULL << i;
> + }
> + seq_printf(filp, "FREE STS : 0x%016llx\n", free_sts);
> + seq_printf(filp, "BUSY STS : 0x%016llx\n", busy_sts);
> +
> + return 0;
> +}
> +
> +RVU_DEBUG_SEQ_FOPS(cpt_ae_sts, cpt_ae_sts_display, NULL);
> +
> +static int rvu_dbg_cpt_se_sts_display(struct seq_file *filp, void *unused)
> +{
> + struct rvu *rvu = filp->private;
> + u64 busy_sts = 0, free_sts = 0;
> + u32 e_min = 0, e_max = 0, e;
> + u16 max_ses;
> + int blkaddr;
> + u64 reg;
> +
> + blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
> + if (blkaddr < 0)
> + return -ENODEV;
> +
> + reg = rvu_read64(rvu, blkaddr, CPT_AF_CONSTANTS1);
> + max_ses = reg & 0xffff;
> +
> + e_min = 0;
> + e_max = max_ses;
> +
> + for (e = e_min; e < e_max; e++) {
> + reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_STS(e));
> + if (reg & 0x1)
> + busy_sts |= 1ULL << e;
> +
> + if (reg & 0x2)
> + free_sts |= 1ULL << e;
> + }
> + seq_printf(filp, "FREE STS : 0x%016llx\n", free_sts);
> + seq_printf(filp, "BUSY STS : 0x%016llx\n", busy_sts);
> +
> + return 0;
> +}
> +
> +RVU_DEBUG_SEQ_FOPS(cpt_se_sts, cpt_se_sts_display, NULL);
> +
> +static int rvu_dbg_cpt_ie_sts_display(struct seq_file *filp, void *unused)
> +{
> + struct rvu *rvu = filp->private;
> + u64 busy_sts = 0, free_sts = 0;
> + u32 e_min = 0, e_max = 0, e, i;
> + u16 max_ses, max_ies;
> + int blkaddr;
> + u64 reg;
> +
> + blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_CPT, 0);
> + if (blkaddr < 0)
> + return -ENODEV;
> +
> + reg = rvu_read64(rvu, blkaddr, CPT_AF_CONSTANTS1);
> + max_ses = reg & 0xffff;
> + max_ies = (reg >> 16) & 0xffff;
> +
> + e_min = max_ses;
> + e_max = max_ses + max_ies;
> +
> + for (e = e_min, i = 0; e < e_max; e++, i++) {
> + reg = rvu_read64(rvu, blkaddr, CPT_AF_EXEX_STS(e));
> + if (reg & 0x1)
> + busy_sts |= 1ULL << i;
> +
> + if (reg & 0x2)
> + free_sts |= 1ULL << i;
> + }
> + seq_printf(filp, "FREE STS : 0x%016llx\n", free_sts);
> + seq_printf(filp, "BUSY STS : 0x%016llx\n", busy_sts);
> +
> + return 0;
> +}
The above three are very similar. Could they use a single helper?
On Wed, Oct 28, 2020 at 6:50 PM Srujana Challa <[email protected]> wrote:
>
> This series introduces crypto(CPT) drivers(PF & VF) for Marvell OcteonTX2
> CN96XX Soc.
>
> OcteonTX2 SOC's resource virtualization unit (RVU) supports multiple
> physical and virtual functions. Each of the PF/VF's functionality is
> determined by what kind of resources are attached to it. When the CPT
> block is attached to a VF, it can function as a security device.
> The following document provides an overview of the hardware and
> different drivers for the OcteonTX2 SOC:
> https://www.kernel.org/doc/Documentation/networking/device_drivers/marvell/octeontx2.rst
>
> The CPT PF driver is responsible for:
> - Forwarding messages to/from VFs from/to admin function(AF),
> - Enabling/disabling VFs,
> - Loading/unloading microcode (creation/deletion of engine groups).
>
> The CPT VF driver works as a crypto offload device.
>
> This patch series includes:
> - Patch to update existing Marvell sources to support the CPT driver.
> - Patch that adds mailbox messages to the admin function (AF) driver,
> to configure CPT HW registers.
> - CPT PF driver patches that include AF<=>PF<=>VF mailbox communication,
> sriov_configure, and firmware load to the acceleration engines.
> - CPT VF driver patches that include VF<=>PF mailbox communication and
> crypto offload support through the kernel cryptographic API.
>
> This series is tested with CRYPTO_EXTRA_TESTS enabled and
> CRYPTO_DISABLE_TESTS disabled.
>
> Changes since v7:
> * Removed writable entries in debugfs.
> * Dropped IPsec support.
> Changes since v6:
> * Removed driver version.
> Changes since v4:
> * Rebased the patches onto net-next tree with base
> 'commit bc081a693a56 ("Merge branch 'Offload-tc-vlan-mangle-to-mscc_ocelot-switch'")'
> Changes since v3:
> * Splitup the patches into smaller patches with more informartion.
> Changes since v2:
> * Fixed C=1 warnings.
> * Added code to exit CPT VF driver gracefully.
> * Moved OcteonTx2 asm code to a header file under include/linux/soc/
> Changes since v1:
> * Moved Makefile changes from patch4 to patch2 and patch3.
>
> Srujana Challa (12):
> octeontx2-pf: move lmt flush to include/linux/soc
> octeontx2-af: add mailbox interface for CPT
> octeontx2-af: add debugfs entries for CPT block
> drivers: crypto: add Marvell OcteonTX2 CPT PF driver
> crypto: octeontx2: add mailbox communication with AF
> crypto: octeontx2: enable SR-IOV and mailbox communication with VF
> crypto: octeontx2: load microcode and create engine groups
> crypto: octeontx2: add LF framework
> crypto: octeontx2: add support to get engine capabilities
> crypto: octeontx2: add virtual function driver support
> crypto: octeontx2: add support to process the crypto request
> crypto: octeontx2: register with linux crypto framework
>
> MAINTAINERS | 2 +
> drivers/crypto/marvell/Kconfig | 14 +
> drivers/crypto/marvell/Makefile | 1 +
> drivers/crypto/marvell/octeontx2/Makefile | 10 +
> .../marvell/octeontx2/otx2_cpt_common.h | 123 ++
> .../marvell/octeontx2/otx2_cpt_hw_types.h | 464 +++++
> .../marvell/octeontx2/otx2_cpt_mbox_common.c | 202 ++
> .../marvell/octeontx2/otx2_cpt_reqmgr.h | 197 ++
> drivers/crypto/marvell/octeontx2/otx2_cptlf.c | 426 +++++
> drivers/crypto/marvell/octeontx2/otx2_cptlf.h | 351 ++++
> drivers/crypto/marvell/octeontx2/otx2_cptpf.h | 52 +
> .../marvell/octeontx2/otx2_cptpf_main.c | 570 ++++++
> .../marvell/octeontx2/otx2_cptpf_mbox.c | 331 ++++
> .../marvell/octeontx2/otx2_cptpf_ucode.c | 1533 +++++++++++++++
> .../marvell/octeontx2/otx2_cptpf_ucode.h | 162 ++
> drivers/crypto/marvell/octeontx2/otx2_cptvf.h | 28 +
> .../marvell/octeontx2/otx2_cptvf_algs.c | 1665 +++++++++++++++++
> .../marvell/octeontx2/otx2_cptvf_algs.h | 170 ++
> .../marvell/octeontx2/otx2_cptvf_main.c | 401 ++++
> .../marvell/octeontx2/otx2_cptvf_mbox.c | 139 ++
> .../marvell/octeontx2/otx2_cptvf_reqmgr.c | 539 ++++++
> .../ethernet/marvell/octeontx2/af/Makefile | 3 +-
> .../net/ethernet/marvell/octeontx2/af/mbox.h | 33 +
> .../net/ethernet/marvell/octeontx2/af/rvu.c | 2 +-
> .../net/ethernet/marvell/octeontx2/af/rvu.h | 2 +
> .../ethernet/marvell/octeontx2/af/rvu_cpt.c | 229 +++
> .../marvell/octeontx2/af/rvu_debugfs.c | 304 +++
> .../ethernet/marvell/octeontx2/af/rvu_reg.h | 63 +-
> .../marvell/octeontx2/nic/otx2_common.h | 13 +-
> include/linux/soc/marvell/octeontx2/asm.h | 29 +
> 30 files changed, 8038 insertions(+), 20 deletions(-)
> create mode 100644 drivers/crypto/marvell/octeontx2/Makefile
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_hw_types.h
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_reqmgr.h
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptlf.c
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptlf.h
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf.h
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.h
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf.h
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.c
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.h
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c
> create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_reqmgr.c
> create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c
> create mode 100644 include/linux/soc/marvell/octeontx2/asm.h
For netdrv review. I don't have a lot of detailed feedback. Overall it
looks sensible to me.
It is obviously a big beast of a patchset. Looking at previous review
comments, the documentation is now a lot better, which helps follow
the thread.
The point about parsing tar files remains open. In general error-prone
parsing is better left to userspace. In practice this is not a lot of
code, nor terribly complex.
The other point is that some files are very close to their oncteontx
counterparts. With minor cleanups, such as using in-place cpu_to_be64s
swap. Those cleanups probably make sense to the older device, too. One
option would be to apply them there first, then copy the files and use
coccinelle for mass renaming. But frankly that is a lot of extra work
to arrive at the same state. I just used --color-words=. --no-index to
perform an iterative review, which works well enough.
On Sat, 31 Oct 2020 15:28:59 -0400 Willem de Bruijn wrote:
> The point about parsing tar files remains open. In general error-prone
> parsing is better left to userspace.
The tar files have to go. Srujana said they have 3 files to load,
just load them individually.
> Subject: Re: [PATCH v8,net-next,00/12] Add Support for Marvell OcteonTX2
>
> On Sat, 31 Oct 2020 15:28:59 -0400 Willem de Bruijn wrote:
> > The point about parsing tar files remains open. In general error-prone
> > parsing is better left to userspace.
>
> The tar files have to go. Srujana said they have 3 files to load,
> just load them individually.
Okay, I will do the changes and submit next version.
> On Wed, Oct 28, 2020 at 6:50 PM Srujana Challa <[email protected]> wrote:
> >
> > This series introduces crypto(CPT) drivers(PF & VF) for Marvell OcteonTX2
> > CN96XX Soc.
> >
> > OcteonTX2 SOC's resource virtualization unit (RVU) supports multiple
> > physical and virtual functions. Each of the PF/VF's functionality is
> > determined by what kind of resources are attached to it. When the CPT
> > block is attached to a VF, it can function as a security device.
> > The following document provides an overview of the hardware and
> > different drivers for the OcteonTX2 SOC:
> > https://urldefense.proofpoint.com/v2/url?u=https-
> 3A__http://www.kernel.org_doc_Documentation_networking_device-
> 5Fdrivers_marvell_octeontx2.rst&d=DwIBaQ&c=nKjWec2b6R0mOyPaz7xtfQ&r=Fj
> 4OoD5hcKFpANhTWdwQzjT1Jpf7veC5263T47JVpnc&m=NJ9fKbMEy2o3y4ca5r-
> kkUR9gadv3CZqF3CJSDvXpnM&s=ScxKfEYpF9bGu2bsz15tQSKhM58NJpMe2-
> 7sxjT_yiI&e=
> >
> > The CPT PF driver is responsible for:
> > - Forwarding messages to/from VFs from/to admin function(AF),
> > - Enabling/disabling VFs,
> > - Loading/unloading microcode (creation/deletion of engine groups).
> >
> > The CPT VF driver works as a crypto offload device.
> >
> > This patch series includes:
> > - Patch to update existing Marvell sources to support the CPT driver.
> > - Patch that adds mailbox messages to the admin function (AF) driver,
> > to configure CPT HW registers.
> > - CPT PF driver patches that include AF<=>PF<=>VF mailbox communication,
> > sriov_configure, and firmware load to the acceleration engines.
> > - CPT VF driver patches that include VF<=>PF mailbox communication and
> > crypto offload support through the kernel cryptographic API.
> >
> > This series is tested with CRYPTO_EXTRA_TESTS enabled and
> > CRYPTO_DISABLE_TESTS disabled.
> >
> > Changes since v7:
> > * Removed writable entries in debugfs.
> > * Dropped IPsec support.
> > Changes since v6:
> > * Removed driver version.
> > Changes since v4:
> > * Rebased the patches onto net-next tree with base
> > 'commit bc081a693a56 ("Merge branch 'Offload-tc-vlan-mangle-to-
> mscc_ocelot-switch'")'
> > Changes since v3:
> > * Splitup the patches into smaller patches with more informartion.
> > Changes since v2:
> > * Fixed C=1 warnings.
> > * Added code to exit CPT VF driver gracefully.
> > * Moved OcteonTx2 asm code to a header file under include/linux/soc/
> > Changes since v1:
> > * Moved Makefile changes from patch4 to patch2 and patch3.
> >
> > Srujana Challa (12):
> > octeontx2-pf: move lmt flush to include/linux/soc
> > octeontx2-af: add mailbox interface for CPT
> > octeontx2-af: add debugfs entries for CPT block
> > drivers: crypto: add Marvell OcteonTX2 CPT PF driver
> > crypto: octeontx2: add mailbox communication with AF
> > crypto: octeontx2: enable SR-IOV and mailbox communication with VF
> > crypto: octeontx2: load microcode and create engine groups
> > crypto: octeontx2: add LF framework
> > crypto: octeontx2: add support to get engine capabilities
> > crypto: octeontx2: add virtual function driver support
> > crypto: octeontx2: add support to process the crypto request
> > crypto: octeontx2: register with linux crypto framework
> >
> > MAINTAINERS | 2 +
> > drivers/crypto/marvell/Kconfig | 14 +
> > drivers/crypto/marvell/Makefile | 1 +
> > drivers/crypto/marvell/octeontx2/Makefile | 10 +
> > .../marvell/octeontx2/otx2_cpt_common.h | 123 ++
> > .../marvell/octeontx2/otx2_cpt_hw_types.h | 464 +++++
> > .../marvell/octeontx2/otx2_cpt_mbox_common.c | 202 ++
> > .../marvell/octeontx2/otx2_cpt_reqmgr.h | 197 ++
> > drivers/crypto/marvell/octeontx2/otx2_cptlf.c | 426 +++++
> > drivers/crypto/marvell/octeontx2/otx2_cptlf.h | 351 ++++
> > drivers/crypto/marvell/octeontx2/otx2_cptpf.h | 52 +
> > .../marvell/octeontx2/otx2_cptpf_main.c | 570 ++++++
> > .../marvell/octeontx2/otx2_cptpf_mbox.c | 331 ++++
> > .../marvell/octeontx2/otx2_cptpf_ucode.c | 1533 +++++++++++++++
> > .../marvell/octeontx2/otx2_cptpf_ucode.h | 162 ++
> > drivers/crypto/marvell/octeontx2/otx2_cptvf.h | 28 +
> > .../marvell/octeontx2/otx2_cptvf_algs.c | 1665 +++++++++++++++++
> > .../marvell/octeontx2/otx2_cptvf_algs.h | 170 ++
> > .../marvell/octeontx2/otx2_cptvf_main.c | 401 ++++
> > .../marvell/octeontx2/otx2_cptvf_mbox.c | 139 ++
> > .../marvell/octeontx2/otx2_cptvf_reqmgr.c | 539 ++++++
> > .../ethernet/marvell/octeontx2/af/Makefile | 3 +-
> > .../net/ethernet/marvell/octeontx2/af/mbox.h | 33 +
> > .../net/ethernet/marvell/octeontx2/af/rvu.c | 2 +-
> > .../net/ethernet/marvell/octeontx2/af/rvu.h | 2 +
> > .../ethernet/marvell/octeontx2/af/rvu_cpt.c | 229 +++
> > .../marvell/octeontx2/af/rvu_debugfs.c | 304 +++
> > .../ethernet/marvell/octeontx2/af/rvu_reg.h | 63 +-
> > .../marvell/octeontx2/nic/otx2_common.h | 13 +-
> > include/linux/soc/marvell/octeontx2/asm.h | 29 +
> > 30 files changed, 8038 insertions(+), 20 deletions(-)
> > create mode 100644 drivers/crypto/marvell/octeontx2/Makefile
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_common.h
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_hw_types.h
> > create mode 100644
> drivers/crypto/marvell/octeontx2/otx2_cpt_mbox_common.c
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cpt_reqmgr.h
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptlf.c
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptlf.h
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf.h
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_main.c
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_mbox.c
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.c
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptpf_ucode.h
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf.h
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.c
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.h
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_main.c
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_mbox.c
> > create mode 100644 drivers/crypto/marvell/octeontx2/otx2_cptvf_reqmgr.c
> > create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c
> > create mode 100644 include/linux/soc/marvell/octeontx2/asm.h
>
> For netdrv review. I don't have a lot of detailed feedback. Overall it
> looks sensible to me.
>
> It is obviously a big beast of a patchset. Looking at previous review
> comments, the documentation is now a lot better, which helps follow
> the thread.
>
> The point about parsing tar files remains open. In general error-prone
> parsing is better left to userspace. In practice this is not a lot of
> code, nor terribly complex.
>
> The other point is that some files are very close to their oncteontx
> counterparts. With minor cleanups, such as using in-place cpu_to_be64s
> swap. Those cleanups probably make sense to the older device, too. One
> option would be to apply them there first, then copy the files and use
> coccinelle for mass renaming. But frankly that is a lot of extra work
> to arrive at the same state. I just used --color-words=. --no-index to
> perform an iterative review, which works well enough.
Thank you for the suggestion. The 2 devices have different capabilities and
our intent is to keep both the drivers separate without having a mutual code
dependency.
And our future hardware version will deviate even more, which reduces the
amount of common code.
> On Wed, Oct 28, 2020 at 10:44 PM Srujana Challa <[email protected]> wrote:
> >
> > On OcteonTX2 SoC, the admin function (AF) is the only one with all
> > priviliges to configure HW and alloc resources, PFs and it's VFs
> > have to request AF via mailbox for all their needs. This patch adds
> > a mailbox interface for CPT PFs and VFs to allocate resources
> > for cryptography.
> >
> > Signed-off-by: Suheil Chandran <[email protected]>
> > Signed-off-by: Srujana Challa <[email protected]>
> > ---
> > .../ethernet/marvell/octeontx2/af/Makefile | 3 +-
> > .../net/ethernet/marvell/octeontx2/af/mbox.h | 33 +++
> > .../net/ethernet/marvell/octeontx2/af/rvu.c | 2 +-
> > .../net/ethernet/marvell/octeontx2/af/rvu.h | 1 +
> > .../ethernet/marvell/octeontx2/af/rvu_cpt.c | 229 ++++++++++++++++++
> > .../ethernet/marvell/octeontx2/af/rvu_reg.h | 63 ++++-
> > 6 files changed, 323 insertions(+), 8 deletions(-)
> > create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/rvu_cpt.c
>
> > diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
> b/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
> > index 7ca599b973c0..807b1c1a9d85 100644
> > --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
> > +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
> > @@ -429,12 +429,63 @@
> > #define TIM_AF_LF_RST (0x20)
> >
> > /* CPT */
> > -#define CPT_AF_CONSTANTS0 (0x0000)
> > -#define CPT_PRIV_LFX_CFG (0x41000)
> > -#define CPT_PRIV_LFX_INT_CFG (0x43000)
> > -#define CPT_AF_RVU_LF_CFG_DEBUG (0x45000)
> > -#define CPT_AF_LF_RST (0x44000)
> > -#define CPT_AF_BLK_RST (0x46000)
> > +#define CPT_AF_CONSTANTS0 (0x0000)
> > +#define CPT_AF_CONSTANTS1 (0x1000)
> > +#define CPT_AF_DIAG (0x3000)
> > +#define CPT_AF_ECO (0x4000)
> > +#define CPT_AF_FLTX_INT(a) (0xa000ull | (u64)(a) << 3)
> > +#define CPT_AF_FLTX_INT_W1S(a) (0xb000ull | (u64)(a) << 3)
> > +#define CPT_AF_FLTX_INT_ENA_W1C(a) (0xc000ull | (u64)(a) << 3)
> > +#define CPT_AF_FLTX_INT_ENA_W1S(a) (0xd000ull | (u64)(a) << 3)
> > +#define CPT_AF_PSNX_EXE(a) (0xe000ull | (u64)(a) << 3)
> > +#define CPT_AF_PSNX_EXE_W1S(a) (0xf000ull | (u64)(a) << 3)
> > +#define CPT_AF_PSNX_LF(a) (0x10000ull | (u64)(a) << 3)
> > +#define CPT_AF_PSNX_LF_W1S(a) (0x11000ull | (u64)(a) << 3)
> > +#define CPT_AF_EXEX_CTL2(a) (0x12000ull | (u64)(a) << 3)
> > +#define CPT_AF_EXEX_STS(a) (0x13000ull | (u64)(a) << 3)
> > +#define CPT_AF_EXE_ERR_INFO (0x14000)
> > +#define CPT_AF_EXEX_ACTIVE(a) (0x16000ull | (u64)(a) << 3)
> > +#define CPT_AF_INST_REQ_PC (0x17000)
> > +#define CPT_AF_INST_LATENCY_PC (0x18000)
> > +#define CPT_AF_RD_REQ_PC (0x19000)
> > +#define CPT_AF_RD_LATENCY_PC (0x1a000)
> > +#define CPT_AF_RD_UC_PC (0x1b000)
> > +#define CPT_AF_ACTIVE_CYCLES_PC (0x1c000)
> > +#define CPT_AF_EXE_DBG_CTL (0x1d000)
> > +#define CPT_AF_EXE_DBG_DATA (0x1e000)
> > +#define CPT_AF_EXE_REQ_TIMER (0x1f000)
> > +#define CPT_AF_EXEX_CTL(a) (0x20000ull | (u64)(a) << 3)
> > +#define CPT_AF_EXE_PERF_CTL (0x21000)
> > +#define CPT_AF_EXE_DBG_CNTX(a) (0x22000ull | (u64)(a) << 3)
> > +#define CPT_AF_EXE_PERF_EVENT_CNT (0x23000)
> > +#define CPT_AF_EXE_EPCI_INBX_CNT(a) (0x24000ull | (u64)(a) << 3)
> > +#define CPT_AF_EXE_EPCI_OUTBX_CNT(a) (0x25000ull | (u64)(a) << 3)
> > +#define CPT_AF_EXEX_UCODE_BASE(a) (0x26000ull | (u64)(a) << 3)
> > +#define CPT_AF_LFX_CTL(a) (0x27000ull | (u64)(a) << 3)
> > +#define CPT_AF_LFX_CTL2(a) (0x29000ull | (u64)(a) << 3)
> > +#define CPT_AF_CPTCLK_CNT (0x2a000)
> > +#define CPT_AF_PF_FUNC (0x2b000)
> > +#define CPT_AF_LFX_PTR_CTL(a) (0x2c000ull | (u64)(a) << 3)
> > +#define CPT_AF_GRPX_THR(a) (0x2d000ull | (u64)(a) << 3)
> > +#define CPT_AF_CTL (0x2e000ull)
> > +#define CPT_AF_XEX_THR(a) (0x2f000ull | (u64)(a) << 3)
> > +#define CPT_PRIV_LFX_CFG (0x41000)
> > +#define CPT_PRIV_AF_INT_CFG (0x42000)
> > +#define CPT_PRIV_LFX_INT_CFG (0x43000)
> > +#define CPT_AF_LF_RST (0x44000)
> > +#define CPT_AF_RVU_LF_CFG_DEBUG (0x45000)
> > +#define CPT_AF_BLK_RST (0x46000)
> > +#define CPT_AF_RVU_INT (0x47000)
> > +#define CPT_AF_RVU_INT_W1S (0x47008)
> > +#define CPT_AF_RVU_INT_ENA_W1S (0x47010)
> > +#define CPT_AF_RVU_INT_ENA_W1C (0x47018)
> > +#define CPT_AF_RAS_INT (0x47020)
> > +#define CPT_AF_RAS_INT_W1S (0x47028)
> > +#define CPT_AF_RAS_INT_ENA_W1S (0x47030)
> > +#define CPT_AF_RAS_INT_ENA_W1C (0x47038)
> > +
> > +#define CPT_AF_LF_CTL2_SHIFT 3
> > +#define CPT_AF_LF_SSO_PF_FUNC_SHIFT 32
> >
> > #define NPC_AF_BLK_RST (0x00040)
>
> Most of these are not related to this patch. Used only in the
> follow-on debug interface one?
Added all hardware CPT AF register defines here. I will add this
point to the patch description in next version.
Thanks.