[PATCH 5/7] soc/fman: Add the FMan parser and KeyGen FLIB(s)

From: Emil Medve
Date: Thu Mar 05 2015 - 00:32:43 EST


From: Igal Liberman <Igal.Liberman@xxxxxxxxxxxxx>

Signed-off-by: Igal Liberman <Igal.Liberman@xxxxxxxxxxxxx>
---
drivers/soc/fsl/fman/Kconfig | 6 +
drivers/soc/fsl/fman/Makefile | 1 +
drivers/soc/fsl/fman/pcd/Makefile | 13 +
drivers/soc/fsl/fman/pcd/fman_kg.c | 850 ++++++++++++++++++++++++++++++++++++
drivers/soc/fsl/fman/pcd/fman_prs.c | 127 ++++++
5 files changed, 997 insertions(+)
create mode 100644 drivers/soc/fsl/fman/pcd/Makefile
create mode 100644 drivers/soc/fsl/fman/pcd/fman_kg.c
create mode 100644 drivers/soc/fsl/fman/pcd/fman_prs.c

diff --git a/drivers/soc/fsl/fman/Kconfig b/drivers/soc/fsl/fman/Kconfig
index c1c258c..2559ee9 100644
--- a/drivers/soc/fsl/fman/Kconfig
+++ b/drivers/soc/fsl/fman/Kconfig
@@ -20,4 +20,10 @@ config FSL_FMAN_MAC
help
Freescale DPAA FMan MAC support

+config FSL_FMAN_PCD
+ bool "FMan PCD"
+ default n
+ help
+ Freescale DPAA FMan Parse, Classify and Distribute support
+
endif # FSL_FMAN
diff --git a/drivers/soc/fsl/fman/Makefile b/drivers/soc/fsl/fman/Makefile
index ffaabd9..dc37f35 100644
--- a/drivers/soc/fsl/fman/Makefile
+++ b/drivers/soc/fsl/fman/Makefile
@@ -12,5 +12,6 @@ fsl_fman-objs := fman.o

obj-$(CONFIG_FSL_FMAN_PORT) += port/
obj-$(CONFIG_FSL_FMAN_MAC) += mac/
+obj-$(CONFIG_FSL_FMAN_PCD) += pcd/

endif
diff --git a/drivers/soc/fsl/fman/pcd/Makefile b/drivers/soc/fsl/fman/pcd/Makefile
new file mode 100644
index 0000000..fa40367
--- /dev/null
+++ b/drivers/soc/fsl/fman/pcd/Makefile
@@ -0,0 +1,13 @@
+ccflags-y += -DVERSION=\"\"
+
+ifeq ($(CONFIG_FSL_FMAN_PCD),y)
+
+FMAN = $(srctree)/drivers/soc/fsl/fman
+
+ccflags-y += -I$(FMAN)/flib
+
+obj-$(CONFIG_FSL_FMAN_PCD) += fsl_fman_pcd.o
+
+fsl_fman_pcd-objs := fman_kg.o fman_prs.o
+
+endif
diff --git a/drivers/soc/fsl/fman/pcd/fman_kg.c b/drivers/soc/fsl/fman/pcd/fman_kg.c
new file mode 100644
index 0000000..f420db6
--- /dev/null
+++ b/drivers/soc/fsl/fman/pcd/fman_kg.c
@@ -0,0 +1,850 @@
+/*
+ * Copyright 2008-2015 Freescale Semiconductor Inc.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * * Neither the name of Freescale Semiconductor nor the
+ * names of its contributors may be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ *
+ * ALTERNATIVELY, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") as published by the Free Software
+ * Foundation, either version 2 of that License or (at your option) any
+ * later version.
+ *
+ * THIS SOFTWARE IS PROVIDED BY Freescale Semiconductor ``AS IS'' AND ANY
+ * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL Freescale Semiconductor BE LIABLE FOR ANY
+ * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "fsl_fman_kg.h"
+
+/* static functions */
+
+static uint32_t build_ar_bind_scheme(uint8_t hwport_id, bool write)
+{
+ uint32_t rw;
+
+ rw = write ? (uint32_t)FM_KG_KGAR_WRITE : (uint32_t)FM_KG_KGAR_READ;
+
+ return (uint32_t)(FM_KG_KGAR_GO | rw | FM_PCD_KG_KGAR_SEL_PORT_ENTRY |
+ hwport_id | FM_PCD_KG_KGAR_SEL_PORT_WSEL_SP);
+}
+
+static void clear_pe_all_scheme(struct fman_kg_regs __iomem *regs,
+ uint8_t hwport_id)
+{
+ uint32_t ar;
+
+ fman_kg_write_sp(regs, 0xffffffff, 0);
+
+ ar = build_ar_bind_scheme(hwport_id, true);
+ fman_kg_write_ar_wait(regs, ar);
+}
+
+static uint32_t build_ar_bind_cls_plan(uint8_t hwport_id, bool write)
+{
+ uint32_t rw;
+
+ rw = write ? (uint32_t)FM_KG_KGAR_WRITE : (uint32_t)FM_KG_KGAR_READ;
+
+ return (uint32_t)(FM_KG_KGAR_GO | rw | FM_PCD_KG_KGAR_SEL_PORT_ENTRY |
+ hwport_id | FM_PCD_KG_KGAR_SEL_PORT_WSEL_CPP);
+}
+
+static void clear_pe_all_cls_plan(struct fman_kg_regs __iomem *regs,
+ uint8_t hwport_id)
+{
+ uint32_t ar;
+
+ fman_kg_write_cpp(regs, 0);
+
+ ar = build_ar_bind_cls_plan(hwport_id, true);
+ fman_kg_write_ar_wait(regs, ar);
+}
+
+static uint8_t get_gen_ht_code(enum fman_kg_gen_extract_src src,
+ bool no_validation, uint8_t *offset)
+{
+ int code;
+
+ switch (src) {
+ case E_FMAN_KG_GEN_EXTRACT_ETH:
+ code = no_validation ? 0x73 : 0x3;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_ETYPE:
+ code = no_validation ? 0x77 : 0x7;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_SNAP:
+ code = no_validation ? 0x74 : 0x4;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_VLAN_TCI_1:
+ code = no_validation ? 0x75 : 0x5;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_VLAN_TCI_N:
+ code = no_validation ? 0x76 : 0x6;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_PPPOE:
+ code = no_validation ? 0x78 : 0x8;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_MPLS_1:
+ code = no_validation ? 0x79 : 0x9;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_MPLS_2:
+ code = no_validation ? FM_KG_SCH_GEN_HT_INVALID : 0x19;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_MPLS_3:
+ code = no_validation ? FM_KG_SCH_GEN_HT_INVALID : 0x29;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_MPLS_N:
+ code = no_validation ? 0x7a : 0xa;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_IPV4_1:
+ code = no_validation ? 0x7b : 0xb;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_IPV6_1:
+ code = no_validation ? 0x7b : 0x1b;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_IPV4_2:
+ code = no_validation ? 0x7c : 0xc;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_IPV6_2:
+ code = no_validation ? 0x7c : 0x1c;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_MINENCAP:
+ code = no_validation ? 0x7c : 0x2c;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_IP_PID:
+ code = no_validation ? 0x72 : 0x2;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_GRE:
+ code = no_validation ? 0x7d : 0xd;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_TCP:
+ code = no_validation ? 0x7e : 0xe;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_UDP:
+ code = no_validation ? 0x7e : 0x1e;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_SCTP:
+ code = no_validation ? 0x7e : 0x3e;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_DCCP:
+ code = no_validation ? 0x7e : 0x4e;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_IPSEC_AH:
+ code = no_validation ? 0x7e : 0x2e;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_IPSEC_ESP:
+ code = no_validation ? 0x7e : 0x6e;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_SHIM_1:
+ code = 0x70;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_SHIM_2:
+ code = 0x71;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_FROM_DFLT:
+ code = 0x10;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_FROM_FRAME_START:
+ code = 0x40;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_FROM_PARSE_RESULT:
+ code = 0x20;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_FROM_END_OF_PARSE:
+ code = 0x7f;
+ break;
+
+ case E_FMAN_KG_GEN_EXTRACT_FROM_FQID:
+ code = 0x20;
+ *offset += 0x20;
+ break;
+
+ default:
+ code = FM_KG_SCH_GEN_HT_INVALID;
+ }
+
+ return (uint8_t)code;
+}
+
+static uint32_t build_ar_scheme(uint8_t scheme,
+ uint8_t hwport_id,
+ bool update_counter, bool write)
+{
+ uint32_t rw;
+
+ rw = (uint32_t)(write ? FM_KG_KGAR_WRITE : FM_KG_KGAR_READ);
+
+ return (uint32_t)(FM_KG_KGAR_GO | rw | FM_KG_KGAR_SEL_SCHEME_ENTRY |
+ hwport_id |
+ ((uint32_t)scheme << FM_KG_KGAR_NUM_SHIFT) |
+ (update_counter ? FM_KG_KGAR_SCM_WSEL_UPDATE_CNT :
+ 0));
+}
+
+static uint32_t build_ar_cls_plan(uint8_t grp,
+ uint8_t entries_mask,
+ uint8_t hwport_id, bool write)
+{
+ uint32_t rw;
+
+ rw = (uint32_t)(write ? FM_KG_KGAR_WRITE : FM_KG_KGAR_READ);
+
+ return (uint32_t)(FM_KG_KGAR_GO | rw |
+ FM_PCD_KG_KGAR_SEL_CLS_PLAN_ENTRY | hwport_id |
+ ((uint32_t)grp << FM_KG_KGAR_NUM_SHIFT) |
+ ((uint32_t)entries_mask << FM_KG_KGAR_WSEL_SHIFT));
+}
+
+int fman_kg_write_ar_wait(struct fman_kg_regs __iomem *regs, uint32_t fmkg_ar)
+{
+ iowrite32be(fmkg_ar, &regs->fmkg_ar);
+ /* Wait for GO to be idle and read error */
+ while ((fmkg_ar = ioread32be(&regs->fmkg_ar)) & FM_KG_KGAR_GO)
+ ;
+ if (fmkg_ar & FM_PCD_KG_KGAR_ERR)
+ return -EINVAL;
+ return 0;
+}
+
+void fman_kg_write_sp(struct fman_kg_regs __iomem *regs, uint32_t sp, bool add)
+{
+ struct fman_kg_pe_regs __iomem *kgpe_regs;
+ uint32_t tmp;
+
+ kgpe_regs = (struct fman_kg_pe_regs __iomem *)&regs->fmkg_indirect[0];
+ tmp = ioread32be(&kgpe_regs->fmkg_pe_sp);
+
+ if (add)
+ tmp |= sp;
+ else /* clear */
+ tmp &= ~sp;
+
+ iowrite32be(tmp, &kgpe_regs->fmkg_pe_sp);
+}
+
+void fman_kg_write_cpp(struct fman_kg_regs __iomem *regs, uint32_t cpp)
+{
+ struct fman_kg_pe_regs __iomem *kgpe_regs;
+
+ kgpe_regs = (struct fman_kg_pe_regs __iomem *)&regs->fmkg_indirect[0];
+
+ iowrite32be(cpp, &kgpe_regs->fmkg_pe_cpp);
+}
+
+void fman_kg_get_event(struct fman_kg_regs __iomem *regs,
+ uint32_t *event, uint32_t *scheme_idx)
+{
+ uint32_t mask, force;
+
+ *event = ioread32be(&regs->fmkg_eer);
+ mask = ioread32be(&regs->fmkg_eeer);
+ *scheme_idx = ioread32be(&regs->fmkg_seer);
+ *scheme_idx &= ioread32be(&regs->fmkg_seeer);
+
+ *event &= mask;
+
+ /* clear the forced events */
+ force = ioread32be(&regs->fmkg_feer);
+ if (force & *event)
+ iowrite32be(force & ~*event, &regs->fmkg_feer);
+
+ iowrite32be(*event, &regs->fmkg_eer);
+ iowrite32be(*scheme_idx, &regs->fmkg_seer);
+}
+
+void fman_kg_init(struct fman_kg_regs __iomem *regs,
+ uint32_t exceptions, uint32_t dflt_nia)
+{
+ uint32_t tmp;
+ int i;
+
+ iowrite32be(FM_EX_KG_DOUBLE_ECC | FM_EX_KG_KEYSIZE_OVERFLOW,
+ &regs->fmkg_eer);
+
+ tmp = 0;
+ if (exceptions & FM_EX_KG_DOUBLE_ECC)
+ tmp |= FM_EX_KG_DOUBLE_ECC;
+
+ if (exceptions & FM_EX_KG_KEYSIZE_OVERFLOW)
+ tmp |= FM_EX_KG_KEYSIZE_OVERFLOW;
+
+ iowrite32be(tmp, &regs->fmkg_eeer);
+ iowrite32be(0, &regs->fmkg_fdor);
+ iowrite32be(0, &regs->fmkg_gdv0r);
+ iowrite32be(0, &regs->fmkg_gdv1r);
+ iowrite32be(dflt_nia, &regs->fmkg_gcr);
+
+ /* Clear binding between ports to schemes and classification plans
+ * so that all ports are not bound to any scheme/classification plan
+ */
+ for (i = 0; i < FMAN_MAX_NUM_OF_HW_PORTS; i++) {
+ clear_pe_all_scheme(regs, (uint8_t)i);
+ clear_pe_all_cls_plan(regs, (uint8_t)i);
+ }
+}
+
+void fman_kg_enable_scheme_interrupts(struct fman_kg_regs __iomem *regs)
+{
+ /* enable and enable all scheme interrupts */
+ iowrite32be(0xFFFFFFFF, &regs->fmkg_seer);
+ iowrite32be(0xFFFFFFFF, &regs->fmkg_seeer);
+}
+
+void fman_kg_enable(struct fman_kg_regs __iomem *regs)
+{
+ iowrite32be(ioread32be(&regs->fmkg_gcr) | FM_KG_KGGCR_EN,
+ &regs->fmkg_gcr);
+}
+
+void fman_kg_disable(struct fman_kg_regs __iomem *regs)
+{
+ iowrite32be(ioread32be(&regs->fmkg_gcr) & ~FM_KG_KGGCR_EN,
+ &regs->fmkg_gcr);
+}
+
+void fman_kg_set_data_after_prs(struct fman_kg_regs __iomem *regs,
+ uint8_t offset)
+{
+ iowrite32be(offset, &regs->fmkg_fdor);
+}
+
+void fman_kg_set_dflt_val(struct fman_kg_regs __iomem *regs,
+ uint8_t def_id, uint32_t val)
+{
+ if (def_id == 0)
+ iowrite32be(val, &regs->fmkg_gdv0r);
+ else
+ iowrite32be(val, &regs->fmkg_gdv1r);
+}
+
+void fman_kg_set_exception(struct fman_kg_regs __iomem *regs,
+ uint32_t exception, bool enable)
+{
+ uint32_t tmp;
+
+ tmp = ioread32be(&regs->fmkg_eeer);
+
+ if (enable)
+ tmp |= exception;
+ else
+ tmp &= ~exception;
+
+ iowrite32be(tmp, &regs->fmkg_eeer);
+}
+
+void fman_kg_get_exception(struct fman_kg_regs __iomem *regs,
+ uint32_t *events, uint32_t *scheme_ids, bool clear)
+{
+ uint32_t mask;
+
+ *events = ioread32be(&regs->fmkg_eer);
+ mask = ioread32be(&regs->fmkg_eeer);
+ *events &= mask;
+
+ *scheme_ids = 0;
+
+ if (*events & FM_EX_KG_KEYSIZE_OVERFLOW) {
+ *scheme_ids = ioread32be(&regs->fmkg_seer);
+ mask = ioread32be(&regs->fmkg_seeer);
+ *scheme_ids &= mask;
+ }
+
+ if (clear) {
+ iowrite32be(*scheme_ids, &regs->fmkg_seer);
+ iowrite32be(*events, &regs->fmkg_eer);
+ }
+}
+
+void fman_kg_get_capture(struct fman_kg_regs __iomem *regs,
+ struct fman_kg_ex_ecc_attr *ecc_attr, bool clear)
+{
+ uint32_t tmp;
+
+ tmp = ioread32be(&regs->fmkg_serc);
+
+ if (tmp & KG_FMKG_SERC_CAP) {
+ /* Captured data is valid */
+ ecc_attr->valid = true;
+ ecc_attr->double_ecc =
+ (bool)((tmp & KG_FMKG_SERC_CET) ? true : false);
+ ecc_attr->single_ecc_count =
+ (uint8_t)((tmp & KG_FMKG_SERC_CNT_MSK) >>
+ KG_FMKG_SERC_CNT_SHIFT);
+ ecc_attr->addr = (uint16_t)(tmp & KG_FMKG_SERC_ADDR_MSK);
+
+ if (clear)
+ iowrite32be(KG_FMKG_SERC_CAP, &regs->fmkg_serc);
+ } else {
+ /* No ECC error is captured */
+ ecc_attr->valid = false;
+ }
+}
+
+int fman_kg_build_scheme(struct fman_kg_scheme_params *params,
+ struct fman_kg_scheme_regs *scheme_regs)
+{
+ struct fman_kg_extract_params *extract_params;
+ struct fman_kg_gen_extract_params *gen_params;
+ uint32_t tmp_reg, i, select, mask, fqb;
+ uint8_t offset, shift, ht;
+
+ /* Zero out all registers so no need to care about unused ones */
+ memset(scheme_regs, 0, sizeof(struct fman_kg_scheme_regs));
+
+ /* Mode register */
+ tmp_reg = fm_kg_build_nia(params->next_engine,
+ params->next_engine_action);
+ if (tmp_reg == KG_NIA_INVALID)
+ return -EINVAL;
+
+ if (params->next_engine == E_FMAN_PCD_PLCR) {
+ tmp_reg |= FMAN_KG_SCH_MODE_NIA_PLCR;
+ } else if (params->next_engine == E_FMAN_PCD_CC) {
+ tmp_reg |= (uint32_t)params->cc_params.base_offset <<
+ FMAN_KG_SCH_MODE_CCOBASE_SHIFT;
+ }
+
+ tmp_reg |= FMAN_KG_SCH_MODE_EN;
+ scheme_regs->kgse_mode = tmp_reg;
+
+ /* Match vector */
+ scheme_regs->kgse_mv = params->match_vector;
+
+ extract_params = &params->extract_params;
+
+ /* Scheme default values registers */
+ scheme_regs->kgse_dv0 = extract_params->def_scheme_0;
+ scheme_regs->kgse_dv1 = extract_params->def_scheme_1;
+
+ /* Extract Known Fields Command register */
+ scheme_regs->kgse_ekfc = extract_params->known_fields;
+
+ /* Entry Extract Known Default Value register */
+ tmp_reg = 0;
+ tmp_reg |= extract_params->known_fields_def.mac_addr <<
+ FMAN_KG_SCH_DEF_MAC_ADDR_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.vlan_tci <<
+ FMAN_KG_SCH_DEF_VLAN_TCI_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.etype <<
+ FMAN_KG_SCH_DEF_ETYPE_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.ppp_sid <<
+ FMAN_KG_SCH_DEF_PPP_SID_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.ppp_pid <<
+ FMAN_KG_SCH_DEF_PPP_PID_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.mpls <<
+ FMAN_KG_SCH_DEF_MPLS_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.ip_addr <<
+ FMAN_KG_SCH_DEF_IP_ADDR_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.ptype <<
+ FMAN_KG_SCH_DEF_PTYPE_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.ip_tos_tc <<
+ FMAN_KG_SCH_DEF_IP_TOS_TC_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.ipv6_fl <<
+ FMAN_KG_SCH_DEF_IPV6_FL_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.ipsec_spi <<
+ FMAN_KG_SCH_DEF_IPSEC_SPI_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.l4_port <<
+ FMAN_KG_SCH_DEF_L4_PORT_SHIFT;
+ tmp_reg |= extract_params->known_fields_def.tcp_flg <<
+ FMAN_KG_SCH_DEF_TCP_FLG_SHIFT;
+
+ scheme_regs->kgse_ekdv = tmp_reg;
+
+ /* Generic extract registers */
+ if (extract_params->gen_extract_num > FM_KG_NUM_OF_GENERIC_REGS)
+ return -EINVAL;
+
+ for (i = 0; i < extract_params->gen_extract_num; i++) {
+ gen_params = extract_params->gen_extract + i;
+
+ tmp_reg = FMAN_KG_SCH_GEN_VALID;
+ tmp_reg |= (uint32_t)gen_params->def_val <<
+ FMAN_KG_SCH_GEN_DEF_SHIFT;
+
+ if (gen_params->type == E_FMAN_KG_HASH_EXTRACT) {
+ if ((gen_params->extract > FMAN_KG_SCH_GEN_SIZE_MAX) ||
+ (gen_params->extract == 0)) {
+ return -EINVAL;
+ }
+ } else {
+ tmp_reg |= FMAN_KG_SCH_GEN_OR;
+ }
+
+ tmp_reg |= (uint32_t)gen_params->extract <<
+ FMAN_KG_SCH_GEN_SIZE_SHIFT;
+ tmp_reg |= (uint32_t)gen_params->mask <<
+ FMAN_KG_SCH_GEN_MASK_SHIFT;
+
+ offset = gen_params->offset;
+ ht = get_gen_ht_code(gen_params->src,
+ gen_params->no_validation, &offset);
+ tmp_reg |= (uint32_t)ht << FMAN_KG_SCH_GEN_HT_SHIFT;
+ tmp_reg |= offset;
+
+ scheme_regs->kgse_gec[i] = tmp_reg;
+ }
+
+ /* Masks registers */
+ if (extract_params->masks_num > FM_KG_EXTRACT_MASKS_NUM)
+ return -EINVAL;
+
+ select = 0;
+ mask = 0;
+ fqb = 0;
+ for (i = 0; i < extract_params->masks_num; i++) {
+ /* MCSx fields */
+ KG_GET_MASK_SEL_SHIFT(shift, i);
+ if (extract_params->masks[i].is_known) {
+ /* Mask known field */
+ select |= extract_params->masks[i].field_or_gen_idx <<
+ shift;
+ } else {
+ /* Mask generic extract */
+ select |= (extract_params->masks[i].field_or_gen_idx +
+ FM_KG_MASK_SEL_GEN_BASE) << shift;
+ }
+
+ /* MOx fields - spread between se_bmch and se_fqb registers */
+ KG_GET_MASK_OFFSET_SHIFT(shift, i);
+ if (i < 2) {
+ select |= (uint32_t)extract_params->masks[i].offset <<
+ shift;
+ } else {
+ fqb |= (uint32_t)extract_params->masks[i].offset <<
+ shift;
+ }
+
+ /* BMx fields */
+ KG_GET_MASK_SHIFT(shift, i);
+ mask |= (uint32_t)extract_params->masks[i].mask << shift;
+ }
+
+ /* Finish with rest of BMx fileds -
+ * don't mask bits for unused masks by setting
+ * corresponding BMx field = 0xFF
+ */
+ for (i = extract_params->masks_num; i < FM_KG_EXTRACT_MASKS_NUM; i++) {
+ KG_GET_MASK_SHIFT(shift, i);
+ mask |= 0xFF << shift;
+ }
+
+ scheme_regs->kgse_bmch = select;
+ scheme_regs->kgse_bmcl = mask;
+
+ /* Finish with FQB register initialization.
+ * Check fqid is 24-bit value.
+ */
+ if (params->base_fqid & ~0x00FFFFFF)
+ return -EINVAL;
+
+ fqb |= params->base_fqid;
+ scheme_regs->kgse_fqb = fqb;
+
+ /* Hash Configuration register */
+ tmp_reg = 0;
+ if (params->hash_params.use_hash) {
+ /* Check hash mask is 24-bit value */
+ if (params->hash_params.mask & ~0x00FFFFFF)
+ return -EINVAL;
+
+ /* Hash function produces 64-bit value, 24 bits of that
+ * are used to generate fq_id and policer profile.
+ * Thus, maximal shift is 40 bits to allow 24 bits out of 64.
+ */
+ if (params->hash_params.shift_r > FMAN_KG_SCH_HASH_HSHIFT_MAX)
+ return -EINVAL;
+
+ tmp_reg |= params->hash_params.mask;
+ tmp_reg |= (uint32_t)params->hash_params.shift_r <<
+ FMAN_KG_SCH_HASH_HSHIFT_SHIFT;
+
+ if (params->hash_params.sym)
+ tmp_reg |= FMAN_KG_SCH_HASH_SYM;
+ }
+
+ if (params->bypass_fqid_gen)
+ tmp_reg |= FMAN_KG_SCH_HASH_NO_FQID_GEN;
+
+ scheme_regs->kgse_hc = tmp_reg;
+
+ /* Policer Profile register */
+ if (params->policer_params.bypass_pp_gen) {
+ tmp_reg = FMAN_KG_SCH_PP_NO_GEN;
+ } else {
+ /* Lower 8 bits of 24-bits extracted from hash result
+ * are used for policer profile generation.
+ * That leaves maximum shift value = 23.
+ */
+ if (params->policer_params.shift > FMAN_KG_SCH_PP_SHIFT_MAX)
+ return -EINVAL;
+
+ tmp_reg = params->policer_params.base;
+ tmp_reg |= ((uint32_t)params->policer_params.shift <<
+ FMAN_KG_SCH_PP_SH_SHIFT) & FMAN_KG_SCH_PP_SH_MASK;
+ tmp_reg |= ((uint32_t)params->policer_params.shift <<
+ FMAN_KG_SCH_PP_SL_SHIFT) & FMAN_KG_SCH_PP_SL_MASK;
+ tmp_reg |= (uint32_t)params->policer_params.mask <<
+ FMAN_KG_SCH_PP_MASK_SHIFT;
+ }
+
+ scheme_regs->kgse_ppc = tmp_reg;
+
+ /* Coarse Classification Bit Select register */
+ if (params->next_engine == E_FMAN_PCD_CC)
+ scheme_regs->kgse_ccbs = params->cc_params.qlcv_bits_sel;
+
+ /* Packets Counter register */
+ if (params->update_counter)
+ scheme_regs->kgse_spc = params->counter_value;
+ return 0;
+}
+
+int fman_kg_write_scheme(struct fman_kg_regs __iomem *regs,
+ uint8_t scheme_id,
+ uint8_t hwport_id,
+ struct fman_kg_scheme_regs *scheme_regs,
+ bool update_counter)
+{
+ struct fman_kg_scheme_regs __iomem *kgse_regs;
+ uint32_t tmp_reg;
+ int err, i;
+
+ /* Write indirect scheme registers */
+ kgse_regs =
+ (struct fman_kg_scheme_regs __iomem *)&regs->fmkg_indirect[0];
+
+ iowrite32be(scheme_regs->kgse_mode, &kgse_regs->kgse_mode);
+ iowrite32be(scheme_regs->kgse_ekfc, &kgse_regs->kgse_ekfc);
+ iowrite32be(scheme_regs->kgse_ekdv, &kgse_regs->kgse_ekdv);
+ iowrite32be(scheme_regs->kgse_bmch, &kgse_regs->kgse_bmch);
+ iowrite32be(scheme_regs->kgse_bmcl, &kgse_regs->kgse_bmcl);
+ iowrite32be(scheme_regs->kgse_fqb, &kgse_regs->kgse_fqb);
+ iowrite32be(scheme_regs->kgse_hc, &kgse_regs->kgse_hc);
+ iowrite32be(scheme_regs->kgse_ppc, &kgse_regs->kgse_ppc);
+ iowrite32be(scheme_regs->kgse_spc, &kgse_regs->kgse_spc);
+ iowrite32be(scheme_regs->kgse_dv0, &kgse_regs->kgse_dv0);
+ iowrite32be(scheme_regs->kgse_dv1, &kgse_regs->kgse_dv1);
+ iowrite32be(scheme_regs->kgse_ccbs, &kgse_regs->kgse_ccbs);
+ iowrite32be(scheme_regs->kgse_mv, &kgse_regs->kgse_mv);
+
+ for (i = 0; i < FM_KG_NUM_OF_GENERIC_REGS; i++)
+ iowrite32be(scheme_regs->kgse_gec[i], &kgse_regs->kgse_gec[i]);
+
+ /* Write AR (Action register) */
+ tmp_reg = build_ar_scheme(scheme_id, hwport_id, update_counter, true);
+ err = fman_kg_write_ar_wait(regs, tmp_reg);
+ return err;
+}
+
+int fman_kg_delete_scheme(struct fman_kg_regs __iomem *regs,
+ uint8_t scheme_id, uint8_t hwport_id)
+{
+ struct fman_kg_scheme_regs __iomem *kgse_regs;
+ uint32_t tmp_reg;
+ int err, i;
+
+ kgse_regs =
+ (struct fman_kg_scheme_regs __iomem *)&regs->fmkg_indirect[0];
+
+ /* Clear all registers including enable bit in mode register */
+ for (i = 0; i < (sizeof(struct fman_kg_scheme_regs)) / 4; ++i)
+ iowrite32be(0, ((uint32_t __iomem *)kgse_regs + i));
+
+ /* Write AR (Action register) */
+ tmp_reg = build_ar_scheme(scheme_id, hwport_id, false, true);
+ err = fman_kg_write_ar_wait(regs, tmp_reg);
+ return err;
+}
+
+int fman_kg_get_scheme_counter(struct fman_kg_regs __iomem *regs,
+ uint8_t scheme_id,
+ uint8_t hwport_id, uint32_t *counter)
+{
+ struct fman_kg_scheme_regs __iomem *kgse_regs;
+ uint32_t tmp_reg;
+ int err;
+
+ kgse_regs =
+ (struct fman_kg_scheme_regs __iomem *)&regs->fmkg_indirect[0];
+
+ tmp_reg = build_ar_scheme(scheme_id, hwport_id, true, false);
+ err = fman_kg_write_ar_wait(regs, tmp_reg);
+
+ if (err != 0)
+ return err;
+
+ *counter = ioread32be(&kgse_regs->kgse_spc);
+
+ return 0;
+}
+
+int fman_kg_set_scheme_counter(struct fman_kg_regs __iomem *regs,
+ uint8_t scheme_id,
+ uint8_t hwport_id, uint32_t counter)
+{
+ struct fman_kg_scheme_regs __iomem *kgse_regs;
+ uint32_t tmp_reg;
+ int err;
+
+ kgse_regs =
+ (struct fman_kg_scheme_regs __iomem *)&regs->fmkg_indirect[0];
+
+ tmp_reg = build_ar_scheme(scheme_id, hwport_id, true, false);
+
+ err = fman_kg_write_ar_wait(regs, tmp_reg);
+ if (err != 0)
+ return err;
+
+ /* Keygen indirect access memory contains all scheme_id registers
+ * by now. Change only counter value.
+ */
+ iowrite32be(counter, &kgse_regs->kgse_spc);
+
+ /* Write back scheme registers */
+ tmp_reg = build_ar_scheme(scheme_id, hwport_id, true, true);
+ err = fman_kg_write_ar_wait(regs, tmp_reg);
+
+ return err;
+}
+
+uint32_t fman_kg_get_schemes_total_counter(struct fman_kg_regs __iomem *regs)
+{
+ return ioread32be(&regs->fmkg_tpc);
+}
+
+int fman_kg_build_cls_plan(struct fman_kg_cls_plan_params *params,
+ struct fman_kg_cp_regs *cls_plan_regs)
+{
+ uint8_t entries_set, entry_bit;
+ int i;
+
+ /* Zero out all group's register */
+ memset(cls_plan_regs, 0, sizeof(struct fman_kg_cp_regs));
+
+ /* Go over all classification entries in params->entries_mask and
+ * configure the corresponding cpe register
+ */
+ entries_set = params->entries_mask;
+ for (i = 0; entries_set; i++) {
+ entry_bit = (uint8_t)(0x80 >> i);
+ if ((entry_bit & entries_set) == 0)
+ continue;
+ entries_set ^= entry_bit;
+ cls_plan_regs->kgcpe[i] = params->mask_vector[i];
+ }
+
+ return 0;
+}
+
+int fman_kg_write_cls_plan(struct fman_kg_regs __iomem *regs,
+ uint8_t grp_id,
+ uint8_t entries_mask,
+ uint8_t hwport_id,
+ struct fman_kg_cp_regs __iomem *cls_plan_regs)
+{
+ struct fman_kg_cp_regs __iomem *kgcpe_regs;
+ uint32_t tmp_reg;
+ int i, err;
+
+ /* Check group index is valid and the group isn't empty */
+ if (grp_id >= FM_KG_CLS_PLAN_GRPS_NUM)
+ return -EINVAL;
+
+ /* Write indirect classification plan registers */
+ kgcpe_regs = (struct fman_kg_cp_regs __iomem *)&regs->fmkg_indirect[0];
+
+ for (i = 0; i < FM_KG_NUM_CLS_PLAN_ENTR; i++)
+ iowrite32be(cls_plan_regs->kgcpe[i], &kgcpe_regs->kgcpe[i]);
+
+ tmp_reg = build_ar_cls_plan(grp_id, entries_mask, hwport_id, true);
+ err = fman_kg_write_ar_wait(regs, tmp_reg);
+ return err;
+}
+
+int fman_kg_write_bind_schemes(struct fman_kg_regs __iomem *regs,
+ uint8_t hwport_id, uint32_t schemes)
+{
+ struct fman_kg_pe_regs __iomem *kg_pe_regs;
+ uint32_t tmp_reg;
+ int err;
+
+ kg_pe_regs = (struct fman_kg_pe_regs __iomem *)&regs->fmkg_indirect[0];
+
+ iowrite32be(schemes, &kg_pe_regs->fmkg_pe_sp);
+
+ tmp_reg = build_ar_bind_scheme(hwport_id, true);
+ err = fman_kg_write_ar_wait(regs, tmp_reg);
+ return err;
+}
+
+int fman_kg_build_bind_cls_plans(uint8_t grp_base,
+ uint8_t grp_mask, uint32_t *bind_cls_plans)
+{
+ /* Check grp_base and grp_mask are 5-bits values */
+ if ((grp_base & ~0x0000001F) || (grp_mask & ~0x0000001F))
+ return -EINVAL;
+
+ *bind_cls_plans =
+ (uint32_t)((grp_mask << FMAN_KG_PE_CPP_MASK_SHIFT) | grp_base);
+ return 0;
+}
+
+int fman_kg_write_bind_cls_plans(struct fman_kg_regs __iomem *regs,
+ uint8_t hwport_id, uint32_t bind_cls_plans)
+{
+ struct fman_kg_pe_regs __iomem *kg_pe_regs;
+ uint32_t tmp_reg;
+ int err;
+
+ kg_pe_regs = (struct fman_kg_pe_regs __iomem *)&regs->fmkg_indirect[0];
+
+ iowrite32be(bind_cls_plans, &kg_pe_regs->fmkg_pe_cpp);
+
+ tmp_reg = build_ar_bind_cls_plan(hwport_id, true);
+ err = fman_kg_write_ar_wait(regs, tmp_reg);
+ return err;
+}
diff --git a/drivers/soc/fsl/fman/pcd/fman_prs.c b/drivers/soc/fsl/fman/pcd/fman_prs.c
new file mode 100644
index 0000000..e4bb20f
--- /dev/null
+++ b/drivers/soc/fsl/fman/pcd/fman_prs.c
@@ -0,0 +1,127 @@
+/*
+ * Copyright 2008-2015 Freescale Semiconductor Inc.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * * Neither the name of Freescale Semiconductor nor the
+ * names of its contributors may be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ *
+ * ALTERNATIVELY, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") as published by the Free Software
+ * Foundation, either version 2 of that License or (at your option) any
+ * later version.
+ *
+ * THIS SOFTWARE IS PROVIDED BY Freescale Semiconductor ``AS IS'' AND ANY
+ * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL Freescale Semiconductor BE LIABLE FOR ANY
+ * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "fsl_fman_prs.h"
+
+uint32_t fman_prs_get_err_event(struct fman_prs_regs __iomem *regs,
+ uint32_t ev_mask)
+{
+ return ioread32be(&regs->fmpr_perr) & ev_mask;
+}
+
+uint32_t fman_prs_get_err_ev_mask(struct fman_prs_regs __iomem *regs)
+{
+ return ioread32be(&regs->fmpr_perer);
+}
+
+void fman_prs_ack_err_event(struct fman_prs_regs __iomem *regs, uint32_t event)
+{
+ iowrite32be(event, &regs->fmpr_perr);
+}
+
+uint32_t fman_prs_get_expt_event(struct fman_prs_regs __iomem *regs,
+ uint32_t ev_mask)
+{
+ return ioread32be(&regs->fmpr_pevr) & ev_mask;
+}
+
+uint32_t fman_prs_get_expt_ev_mask(struct fman_prs_regs __iomem *regs)
+{
+ return ioread32be(&regs->fmpr_pever);
+}
+
+void fman_prs_ack_expt_event(struct fman_prs_regs __iomem *regs, uint32_t event)
+{
+ iowrite32be(event, &regs->fmpr_pevr);
+}
+
+void fman_prs_defconfig(struct fman_prs_cfg *cfg)
+{
+ cfg->port_id_stat = 0;
+ cfg->max_prs_cyc_lim = DEFAULT_MAX_PRS_CYC_LIM;
+ cfg->prs_exceptions = 0x03000000;
+}
+
+int fman_prs_init(struct fman_prs_regs __iomem *regs, struct fman_prs_cfg *cfg)
+{
+ uint32_t tmp;
+
+ iowrite32be(cfg->max_prs_cyc_lim, &regs->fmpr_rpclim);
+ iowrite32be((FM_PCD_PRS_SINGLE_ECC | FM_PCD_PRS_PORT_IDLE_STS),
+ &regs->fmpr_pevr);
+
+ if (cfg->prs_exceptions & FM_PCD_EX_PRS_SINGLE_ECC)
+ iowrite32be(FM_PCD_PRS_SINGLE_ECC, &regs->fmpr_pever);
+ else
+ iowrite32be(0, &regs->fmpr_pever);
+
+ iowrite32be(FM_PCD_PRS_DOUBLE_ECC, &regs->fmpr_perr);
+
+ tmp = 0;
+ if (cfg->prs_exceptions & FM_PCD_EX_PRS_DOUBLE_ECC)
+ tmp |= FM_PCD_PRS_DOUBLE_ECC;
+ iowrite32be(tmp, &regs->fmpr_perer);
+
+ iowrite32be(cfg->port_id_stat, &regs->fmpr_ppsc);
+
+ return 0;
+}
+
+void fman_prs_enable(struct fman_prs_regs __iomem *regs)
+{
+ uint32_t tmp;
+
+ tmp = ioread32be(&regs->fmpr_rpimac) | FM_PCD_PRS_RPIMAC_EN;
+ iowrite32be(tmp, &regs->fmpr_rpimac);
+}
+
+void fman_prs_disable(struct fman_prs_regs __iomem *regs)
+{
+ uint32_t tmp;
+
+ tmp = ioread32be(&regs->fmpr_rpimac) & ~FM_PCD_PRS_RPIMAC_EN;
+ iowrite32be(tmp, &regs->fmpr_rpimac);
+}
+
+void fman_prs_set_stst_port_msk(struct fman_prs_regs __iomem *regs,
+ uint32_t pid_msk)
+{
+ iowrite32be(pid_msk, &regs->fmpr_ppsc);
+}
+
+void fman_prs_set_stst(struct fman_prs_regs __iomem *regs, bool enable)
+{
+ if (enable)
+ iowrite32be(FM_PCD_PRS_PPSC_ALL_PORTS, &regs->fmpr_ppsc);
+ else
+ iowrite32be(0, &regs->fmpr_ppsc);
+}
--
2.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/