diff options
Diffstat (limited to 'drivers/net/ethernet/cavium')
36 files changed, 5317 insertions, 1101 deletions
diff --git a/drivers/net/ethernet/cavium/Kconfig b/drivers/net/ethernet/cavium/Kconfig index 0ef232d3331e..92f411c9f0df 100644 --- a/drivers/net/ethernet/cavium/Kconfig +++ b/drivers/net/ethernet/cavium/Kconfig @@ -36,10 +36,20 @@ config THUNDER_NIC_BGX depends on 64BIT select PHYLIB select MDIO_THUNDER + select THUNDER_NIC_RGX ---help--- This driver supports programming and controlling of MAC interface from NIC physical function driver. +config THUNDER_NIC_RGX + tristate "Thunder MAC interface driver (RGX)" + depends on 64BIT + select PHYLIB + select MDIO_THUNDER + ---help--- + This driver supports configuring XCV block of RGX interface + present on CN81XX chip. + config LIQUIDIO tristate "Cavium LiquidIO support" depends on 64BIT @@ -48,7 +58,7 @@ config LIQUIDIO select LIBCRC32C ---help--- This driver supports Cavium LiquidIO Intelligent Server Adapters - based on CN66XX and CN68XX chips. + based on CN66XX, CN68XX and CN23XX chips. To compile this driver as a module, choose M here: the module will be called liquidio. This is recommended. diff --git a/drivers/net/ethernet/cavium/liquidio/Makefile b/drivers/net/ethernet/cavium/liquidio/Makefile index 2f366806835d..5a27b2a44039 100644 --- a/drivers/net/ethernet/cavium/liquidio/Makefile +++ b/drivers/net/ethernet/cavium/liquidio/Makefile @@ -3,14 +3,16 @@ # obj-$(CONFIG_LIQUIDIO) += liquidio.o -liquidio-objs := lio_main.o \ - lio_ethtool.o \ - request_manager.o \ - response_manager.o \ - octeon_device.o \ - cn66xx_device.o \ - cn68xx_device.o \ - octeon_mem_ops.o \ - octeon_droq.o \ - octeon_console.o \ - octeon_nic.o +liquidio-$(CONFIG_LIQUIDIO) += lio_ethtool.o \ + lio_core.o \ + request_manager.o \ + response_manager.o \ + octeon_device.o \ + cn66xx_device.o \ + cn68xx_device.o \ + cn23xx_pf_device.o \ + octeon_mem_ops.o \ + octeon_droq.o \ + octeon_nic.o + +liquidio-objs := lio_main.o octeon_console.o $(liquidio-y) diff --git a/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_device.c b/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_device.c new file mode 100644 index 000000000000..bddb198c0b74 --- /dev/null +++ b/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_device.c @@ -0,0 +1,1237 @@ +/********************************************************************** +* Author: Cavium, Inc. +* +* Contact: support@cavium.com +* Please include "LiquidIO" in the subject. +* +* Copyright (c) 2003-2015 Cavium, Inc. +* +* This file is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License, Version 2, as +* published by the Free Software Foundation. +* +* This file is distributed in the hope that it will be useful, but +* AS-IS and WITHOUT ANY WARRANTY; without even the implied warranty +* of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, TITLE, or +* NONINFRINGEMENT. See the GNU General Public License for more +* details. +* +* This file may also be available under a different license from Cavium. +* Contact Cavium, Inc. for more information +**********************************************************************/ + +#include <linux/pci.h> +#include <linux/netdevice.h> +#include <linux/vmalloc.h> +#include "liquidio_common.h" +#include "octeon_droq.h" +#include "octeon_iq.h" +#include "response_manager.h" +#include "octeon_device.h" +#include "cn23xx_pf_device.h" +#include "octeon_main.h" + +#define RESET_NOTDONE 0 +#define RESET_DONE 1 + +/* Change the value of SLI Packet Input Jabber Register to allow + * VXLAN TSO packets which can be 64424 bytes, exceeding the + * MAX_GSO_SIZE we supplied to the kernel + */ +#define CN23XX_INPUT_JABBER 64600 + +#define LIOLUT_RING_DISTRIBUTION 9 +const int liolut_num_vfs_to_rings_per_vf[LIOLUT_RING_DISTRIBUTION] = { + 0, 8, 4, 2, 2, 2, 1, 1, 1 +}; + +void cn23xx_dump_pf_initialized_regs(struct octeon_device *oct) +{ + int i = 0; + u32 regval = 0; + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + + /*In cn23xx_soft_reset*/ + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%llx\n", + "CN23XX_WIN_WR_MASK_REG", CVM_CAST64(CN23XX_WIN_WR_MASK_REG), + CVM_CAST64(octeon_read_csr64(oct, CN23XX_WIN_WR_MASK_REG))); + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "CN23XX_SLI_SCRATCH1", CVM_CAST64(CN23XX_SLI_SCRATCH1), + CVM_CAST64(octeon_read_csr64(oct, CN23XX_SLI_SCRATCH1))); + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "CN23XX_RST_SOFT_RST", CN23XX_RST_SOFT_RST, + lio_pci_readq(oct, CN23XX_RST_SOFT_RST)); + + /*In cn23xx_set_dpi_regs*/ + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "CN23XX_DPI_DMA_CONTROL", CN23XX_DPI_DMA_CONTROL, + lio_pci_readq(oct, CN23XX_DPI_DMA_CONTROL)); + + for (i = 0; i < 6; i++) { + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_DPI_DMA_ENG_ENB", i, + CN23XX_DPI_DMA_ENG_ENB(i), + lio_pci_readq(oct, CN23XX_DPI_DMA_ENG_ENB(i))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_DPI_DMA_ENG_BUF", i, + CN23XX_DPI_DMA_ENG_BUF(i), + lio_pci_readq(oct, CN23XX_DPI_DMA_ENG_BUF(i))); + } + + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", "CN23XX_DPI_CTL", + CN23XX_DPI_CTL, lio_pci_readq(oct, CN23XX_DPI_CTL)); + + /*In cn23xx_setup_pcie_mps and cn23xx_setup_pcie_mrrs */ + pci_read_config_dword(oct->pci_dev, CN23XX_CONFIG_PCIE_DEVCTL, ®val); + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "CN23XX_CONFIG_PCIE_DEVCTL", + CVM_CAST64(CN23XX_CONFIG_PCIE_DEVCTL), CVM_CAST64(regval)); + + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_DPI_SLI_PRTX_CFG", oct->pcie_port, + CN23XX_DPI_SLI_PRTX_CFG(oct->pcie_port), + lio_pci_readq(oct, CN23XX_DPI_SLI_PRTX_CFG(oct->pcie_port))); + + /*In cn23xx_specific_regs_setup */ + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_S2M_PORTX_CTL", oct->pcie_port, + CVM_CAST64(CN23XX_SLI_S2M_PORTX_CTL(oct->pcie_port)), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_S2M_PORTX_CTL(oct->pcie_port)))); + + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "CN23XX_SLI_RING_RST", CVM_CAST64(CN23XX_SLI_PKT_IOQ_RING_RST), + (u64)octeon_read_csr64(oct, CN23XX_SLI_PKT_IOQ_RING_RST)); + + /*In cn23xx_setup_global_mac_regs*/ + for (i = 0; i < CN23XX_MAX_MACS; i++) { + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_PKT_MAC_RINFO64", i, + CVM_CAST64(CN23XX_SLI_PKT_MAC_RINFO64(i, oct->pf_num)), + CVM_CAST64(octeon_read_csr64 + (oct, CN23XX_SLI_PKT_MAC_RINFO64 + (i, oct->pf_num)))); + } + + /*In cn23xx_setup_global_input_regs*/ + for (i = 0; i < CN23XX_MAX_INPUT_QUEUES; i++) { + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_IQ_PKT_CONTROL64", i, + CVM_CAST64(CN23XX_SLI_IQ_PKT_CONTROL64(i)), + CVM_CAST64(octeon_read_csr64 + (oct, CN23XX_SLI_IQ_PKT_CONTROL64(i)))); + } + + /*In cn23xx_setup_global_output_regs*/ + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "CN23XX_SLI_OQ_WMARK", CVM_CAST64(CN23XX_SLI_OQ_WMARK), + CVM_CAST64(octeon_read_csr64(oct, CN23XX_SLI_OQ_WMARK))); + + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_OQ_PKT_CONTROL", i, + CVM_CAST64(CN23XX_SLI_OQ_PKT_CONTROL(i)), + CVM_CAST64(octeon_read_csr( + oct, CN23XX_SLI_OQ_PKT_CONTROL(i)))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_OQ_PKT_INT_LEVELS", i, + CVM_CAST64(CN23XX_SLI_OQ_PKT_INT_LEVELS(i)), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_OQ_PKT_INT_LEVELS(i)))); + } + + /*In cn23xx_enable_interrupt and cn23xx_disable_interrupt*/ + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "cn23xx->intr_enb_reg64", + CVM_CAST64((long)(cn23xx->intr_enb_reg64)), + CVM_CAST64(readq(cn23xx->intr_enb_reg64))); + + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "cn23xx->intr_sum_reg64", + CVM_CAST64((long)(cn23xx->intr_sum_reg64)), + CVM_CAST64(readq(cn23xx->intr_sum_reg64))); + + /*In cn23xx_setup_iq_regs*/ + for (i = 0; i < CN23XX_MAX_INPUT_QUEUES; i++) { + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_IQ_BASE_ADDR64", i, + CVM_CAST64(CN23XX_SLI_IQ_BASE_ADDR64(i)), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_IQ_BASE_ADDR64(i)))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_IQ_SIZE", i, + CVM_CAST64(CN23XX_SLI_IQ_SIZE(i)), + CVM_CAST64(octeon_read_csr + (oct, CN23XX_SLI_IQ_SIZE(i)))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_IQ_DOORBELL", i, + CVM_CAST64(CN23XX_SLI_IQ_DOORBELL(i)), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_IQ_DOORBELL(i)))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_IQ_INSTR_COUNT64", i, + CVM_CAST64(CN23XX_SLI_IQ_INSTR_COUNT64(i)), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_IQ_INSTR_COUNT64(i)))); + } + + /*In cn23xx_setup_oq_regs*/ + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_OQ_BASE_ADDR64", i, + CVM_CAST64(CN23XX_SLI_OQ_BASE_ADDR64(i)), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_OQ_BASE_ADDR64(i)))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_OQ_SIZE", i, + CVM_CAST64(CN23XX_SLI_OQ_SIZE(i)), + CVM_CAST64(octeon_read_csr + (oct, CN23XX_SLI_OQ_SIZE(i)))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_OQ_BUFF_INFO_SIZE", i, + CVM_CAST64(CN23XX_SLI_OQ_BUFF_INFO_SIZE(i)), + CVM_CAST64(octeon_read_csr( + oct, CN23XX_SLI_OQ_BUFF_INFO_SIZE(i)))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_OQ_PKTS_SENT", i, + CVM_CAST64(CN23XX_SLI_OQ_PKTS_SENT(i)), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_OQ_PKTS_SENT(i)))); + dev_dbg(&oct->pci_dev->dev, "%s(%d)[%llx] : 0x%016llx\n", + "CN23XX_SLI_OQ_PKTS_CREDIT", i, + CVM_CAST64(CN23XX_SLI_OQ_PKTS_CREDIT(i)), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_OQ_PKTS_CREDIT(i)))); + } + + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "CN23XX_SLI_PKT_TIME_INT", + CVM_CAST64(CN23XX_SLI_PKT_TIME_INT), + CVM_CAST64(octeon_read_csr64(oct, CN23XX_SLI_PKT_TIME_INT))); + dev_dbg(&oct->pci_dev->dev, "%s[%llx] : 0x%016llx\n", + "CN23XX_SLI_PKT_CNT_INT", + CVM_CAST64(CN23XX_SLI_PKT_CNT_INT), + CVM_CAST64(octeon_read_csr64(oct, CN23XX_SLI_PKT_CNT_INT))); +} + +static int cn23xx_pf_soft_reset(struct octeon_device *oct) +{ + octeon_write_csr64(oct, CN23XX_WIN_WR_MASK_REG, 0xFF); + + dev_dbg(&oct->pci_dev->dev, "OCTEON[%d]: BIST enabled for CN23XX soft reset\n", + oct->octeon_id); + + octeon_write_csr64(oct, CN23XX_SLI_SCRATCH1, 0x1234ULL); + + /* Initiate chip-wide soft reset */ + lio_pci_readq(oct, CN23XX_RST_SOFT_RST); + lio_pci_writeq(oct, 1, CN23XX_RST_SOFT_RST); + + /* Wait for 100ms as Octeon resets. */ + mdelay(100); + + if (octeon_read_csr64(oct, CN23XX_SLI_SCRATCH1) == 0x1234ULL) { + dev_err(&oct->pci_dev->dev, "OCTEON[%d]: Soft reset failed\n", + oct->octeon_id); + return 1; + } + + dev_dbg(&oct->pci_dev->dev, "OCTEON[%d]: Reset completed\n", + oct->octeon_id); + + /* restore the reset value*/ + octeon_write_csr64(oct, CN23XX_WIN_WR_MASK_REG, 0xFF); + + return 0; +} + +static void cn23xx_enable_error_reporting(struct octeon_device *oct) +{ + u32 regval; + u32 uncorrectable_err_mask, corrtable_err_status; + + pci_read_config_dword(oct->pci_dev, CN23XX_CONFIG_PCIE_DEVCTL, ®val); + if (regval & CN23XX_CONFIG_PCIE_DEVCTL_MASK) { + uncorrectable_err_mask = 0; + corrtable_err_status = 0; + pci_read_config_dword(oct->pci_dev, + CN23XX_CONFIG_PCIE_UNCORRECT_ERR_MASK, + &uncorrectable_err_mask); + pci_read_config_dword(oct->pci_dev, + CN23XX_CONFIG_PCIE_CORRECT_ERR_STATUS, + &corrtable_err_status); + dev_err(&oct->pci_dev->dev, "PCI-E Fatal error detected;\n" + "\tdev_ctl_status_reg = 0x%08x\n" + "\tuncorrectable_error_mask_reg = 0x%08x\n" + "\tcorrectable_error_status_reg = 0x%08x\n", + regval, uncorrectable_err_mask, + corrtable_err_status); + } + + regval |= 0xf; /* Enable Link error reporting */ + + dev_dbg(&oct->pci_dev->dev, "OCTEON[%d]: Enabling PCI-E error reporting..\n", + oct->octeon_id); + pci_write_config_dword(oct->pci_dev, CN23XX_CONFIG_PCIE_DEVCTL, regval); +} + +static u32 cn23xx_coprocessor_clock(struct octeon_device *oct) +{ + /* Bits 29:24 of RST_BOOT[PNR_MUL] holds the ref.clock MULTIPLIER + * for SLI. + */ + + /* TBD: get the info in Hand-shake */ + return (((lio_pci_readq(oct, CN23XX_RST_BOOT) >> 24) & 0x3f) * 50); +} + +u32 cn23xx_pf_get_oq_ticks(struct octeon_device *oct, u32 time_intr_in_us) +{ + /* This gives the SLI clock per microsec */ + u32 oqticks_per_us = cn23xx_coprocessor_clock(oct); + + oct->pfvf_hsword.coproc_tics_per_us = oqticks_per_us; + + /* This gives the clock cycles per millisecond */ + oqticks_per_us *= 1000; + + /* This gives the oq ticks (1024 core clock cycles) per millisecond */ + oqticks_per_us /= 1024; + + /* time_intr is in microseconds. The next 2 steps gives the oq ticks + * corressponding to time_intr. + */ + oqticks_per_us *= time_intr_in_us; + oqticks_per_us /= 1000; + + return oqticks_per_us; +} + +static void cn23xx_setup_global_mac_regs(struct octeon_device *oct) +{ + u64 reg_val; + u16 mac_no = oct->pcie_port; + u16 pf_num = oct->pf_num; + + /* programming SRN and TRS for each MAC(0..3) */ + + dev_dbg(&oct->pci_dev->dev, "%s:Using pcie port %d\n", + __func__, mac_no); + /* By default, mapping all 64 IOQs to a single MACs */ + + reg_val = + octeon_read_csr64(oct, CN23XX_SLI_PKT_MAC_RINFO64(mac_no, pf_num)); + + if (oct->rev_id == OCTEON_CN23XX_REV_1_1) { + /* setting SRN <6:0> */ + reg_val = pf_num * CN23XX_MAX_RINGS_PER_PF_PASS_1_1; + } else { + /* setting SRN <6:0> */ + reg_val = pf_num * CN23XX_MAX_RINGS_PER_PF; + } + + /* setting TRS <23:16> */ + reg_val = reg_val | + (oct->sriov_info.trs << CN23XX_PKT_MAC_CTL_RINFO_TRS_BIT_POS); + /* write these settings to MAC register */ + octeon_write_csr64(oct, CN23XX_SLI_PKT_MAC_RINFO64(mac_no, pf_num), + reg_val); + + dev_dbg(&oct->pci_dev->dev, "SLI_PKT_MAC(%d)_PF(%d)_RINFO : 0x%016llx\n", + mac_no, pf_num, (u64)octeon_read_csr64 + (oct, CN23XX_SLI_PKT_MAC_RINFO64(mac_no, pf_num))); +} + +static int cn23xx_reset_io_queues(struct octeon_device *oct) +{ + int ret_val = 0; + u64 d64; + u32 q_no, srn, ern; + u32 loop = 1000; + + srn = oct->sriov_info.pf_srn; + ern = srn + oct->sriov_info.num_pf_rings; + + /*As per HRM reg description, s/w cant write 0 to ENB. */ + /*to make the queue off, need to set the RST bit. */ + + /* Reset the Enable bit for all the 64 IQs. */ + for (q_no = srn; q_no < ern; q_no++) { + /* set RST bit to 1. This bit applies to both IQ and OQ */ + d64 = octeon_read_csr64(oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no)); + d64 = d64 | CN23XX_PKT_INPUT_CTL_RST; + octeon_write_csr64(oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), d64); + } + + /*wait until the RST bit is clear or the RST and quite bits are set*/ + for (q_no = srn; q_no < ern; q_no++) { + u64 reg_val = octeon_read_csr64(oct, + CN23XX_SLI_IQ_PKT_CONTROL64(q_no)); + while ((READ_ONCE(reg_val) & CN23XX_PKT_INPUT_CTL_RST) && + !(READ_ONCE(reg_val) & CN23XX_PKT_INPUT_CTL_QUIET) && + loop--) { + WRITE_ONCE(reg_val, octeon_read_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no))); + } + if (!loop) { + dev_err(&oct->pci_dev->dev, + "clearing the reset reg failed or setting the quiet reg failed for qno: %u\n", + q_no); + return -1; + } + WRITE_ONCE(reg_val, READ_ONCE(reg_val) & + ~CN23XX_PKT_INPUT_CTL_RST); + octeon_write_csr64(oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), + READ_ONCE(reg_val)); + + WRITE_ONCE(reg_val, octeon_read_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no))); + if (READ_ONCE(reg_val) & CN23XX_PKT_INPUT_CTL_RST) { + dev_err(&oct->pci_dev->dev, + "clearing the reset failed for qno: %u\n", + q_no); + ret_val = -1; + } + } + + return ret_val; +} + +static int cn23xx_pf_setup_global_input_regs(struct octeon_device *oct) +{ + u32 q_no, ern, srn; + u64 pf_num; + u64 intr_threshold, reg_val; + struct octeon_instr_queue *iq; + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + + pf_num = oct->pf_num; + + srn = oct->sriov_info.pf_srn; + ern = srn + oct->sriov_info.num_pf_rings; + + if (cn23xx_reset_io_queues(oct)) + return -1; + + /** Set the MAC_NUM and PVF_NUM in IQ_PKT_CONTROL reg + * for all queues.Only PF can set these bits. + * bits 29:30 indicate the MAC num. + * bits 32:47 indicate the PVF num. + */ + for (q_no = 0; q_no < ern; q_no++) { + reg_val = oct->pcie_port << CN23XX_PKT_INPUT_CTL_MAC_NUM_POS; + reg_val |= pf_num << CN23XX_PKT_INPUT_CTL_PF_NUM_POS; + + octeon_write_csr64(oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), + reg_val); + } + + /* Select ES, RO, NS, RDSIZE,DPTR Fomat#0 for + * pf queues + */ + for (q_no = srn; q_no < ern; q_no++) { + void __iomem *inst_cnt_reg; + + iq = oct->instr_queue[q_no]; + if (iq) + inst_cnt_reg = iq->inst_cnt_reg; + else + inst_cnt_reg = (u8 *)oct->mmio[0].hw_addr + + CN23XX_SLI_IQ_INSTR_COUNT64(q_no); + + reg_val = + octeon_read_csr64(oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no)); + + reg_val |= CN23XX_PKT_INPUT_CTL_MASK; + + octeon_write_csr64(oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), + reg_val); + + /* Set WMARK level for triggering PI_INT */ + /* intr_threshold = CN23XX_DEF_IQ_INTR_THRESHOLD & */ + intr_threshold = CFG_GET_IQ_INTR_PKT(cn23xx->conf) & + CN23XX_PKT_IN_DONE_WMARK_MASK; + + writeq((readq(inst_cnt_reg) & + ~(CN23XX_PKT_IN_DONE_WMARK_MASK << + CN23XX_PKT_IN_DONE_WMARK_BIT_POS)) | + (intr_threshold << CN23XX_PKT_IN_DONE_WMARK_BIT_POS), + inst_cnt_reg); + } + return 0; +} + +static void cn23xx_pf_setup_global_output_regs(struct octeon_device *oct) +{ + u32 reg_val; + u32 q_no, ern, srn; + u64 time_threshold; + + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + + srn = oct->sriov_info.pf_srn; + ern = srn + oct->sriov_info.num_pf_rings; + + if (CFG_GET_IS_SLI_BP_ON(cn23xx->conf)) { + octeon_write_csr64(oct, CN23XX_SLI_OQ_WMARK, 32); + } else { + /** Set Output queue watermark to 0 to disable backpressure */ + octeon_write_csr64(oct, CN23XX_SLI_OQ_WMARK, 0); + } + + for (q_no = srn; q_no < ern; q_no++) { + reg_val = octeon_read_csr(oct, CN23XX_SLI_OQ_PKT_CONTROL(q_no)); + + /* set IPTR & DPTR */ + reg_val |= + (CN23XX_PKT_OUTPUT_CTL_IPTR | CN23XX_PKT_OUTPUT_CTL_DPTR); + + /* reset BMODE */ + reg_val &= ~(CN23XX_PKT_OUTPUT_CTL_BMODE); + + /* No Relaxed Ordering, No Snoop, 64-bit Byte swap + * for Output Queue ScatterList + * reset ROR_P, NSR_P + */ + reg_val &= ~(CN23XX_PKT_OUTPUT_CTL_ROR_P); + reg_val &= ~(CN23XX_PKT_OUTPUT_CTL_NSR_P); + +#ifdef __LITTLE_ENDIAN_BITFIELD + reg_val &= ~(CN23XX_PKT_OUTPUT_CTL_ES_P); +#else + reg_val |= (CN23XX_PKT_OUTPUT_CTL_ES_P); +#endif + /* No Relaxed Ordering, No Snoop, 64-bit Byte swap + * for Output Queue Data + * reset ROR, NSR + */ + reg_val &= ~(CN23XX_PKT_OUTPUT_CTL_ROR); + reg_val &= ~(CN23XX_PKT_OUTPUT_CTL_NSR); + /* set the ES bit */ + reg_val |= (CN23XX_PKT_OUTPUT_CTL_ES); + + /* write all the selected settings */ + octeon_write_csr(oct, CN23XX_SLI_OQ_PKT_CONTROL(q_no), reg_val); + + /* Enabling these interrupt in oct->fn_list.enable_interrupt() + * routine which called after IOQ init. + * Set up interrupt packet and time thresholds + * for all the OQs + */ + time_threshold = cn23xx_pf_get_oq_ticks( + oct, (u32)CFG_GET_OQ_INTR_TIME(cn23xx->conf)); + + octeon_write_csr64(oct, CN23XX_SLI_OQ_PKT_INT_LEVELS(q_no), + (CFG_GET_OQ_INTR_PKT(cn23xx->conf) | + (time_threshold << 32))); + } + + /** Setting the water mark level for pko back pressure **/ + writeq(0x40, (u8 *)oct->mmio[0].hw_addr + CN23XX_SLI_OQ_WMARK); + + /** Disabling setting OQs in reset when ring has no dorebells + * enabling this will cause of head of line blocking + */ + /* Do it only for pass1.1. and pass1.2 */ + if ((oct->rev_id == OCTEON_CN23XX_REV_1_0) || + (oct->rev_id == OCTEON_CN23XX_REV_1_1)) + writeq(readq((u8 *)oct->mmio[0].hw_addr + + CN23XX_SLI_GBL_CONTROL) | 0x2, + (u8 *)oct->mmio[0].hw_addr + CN23XX_SLI_GBL_CONTROL); + + /** Enable channel-level backpressure */ + if (oct->pf_num) + writeq(0xffffffffffffffffULL, + (u8 *)oct->mmio[0].hw_addr + CN23XX_SLI_OUT_BP_EN2_W1S); + else + writeq(0xffffffffffffffffULL, + (u8 *)oct->mmio[0].hw_addr + CN23XX_SLI_OUT_BP_EN_W1S); +} + +static int cn23xx_setup_pf_device_regs(struct octeon_device *oct) +{ + cn23xx_enable_error_reporting(oct); + + /* program the MAC(0..3)_RINFO before setting up input/output regs */ + cn23xx_setup_global_mac_regs(oct); + + if (cn23xx_pf_setup_global_input_regs(oct)) + return -1; + + cn23xx_pf_setup_global_output_regs(oct); + + /* Default error timeout value should be 0x200000 to avoid host hang + * when reads invalid register + */ + octeon_write_csr64(oct, CN23XX_SLI_WINDOW_CTL, + CN23XX_SLI_WINDOW_CTL_DEFAULT); + + /* set SLI_PKT_IN_JABBER to handle large VXLAN packets */ + octeon_write_csr64(oct, CN23XX_SLI_PKT_IN_JABBER, CN23XX_INPUT_JABBER); + return 0; +} + +static void cn23xx_setup_iq_regs(struct octeon_device *oct, u32 iq_no) +{ + struct octeon_instr_queue *iq = oct->instr_queue[iq_no]; + u64 pkt_in_done; + + iq_no += oct->sriov_info.pf_srn; + + /* Write the start of the input queue's ring and its size */ + octeon_write_csr64(oct, CN23XX_SLI_IQ_BASE_ADDR64(iq_no), + iq->base_addr_dma); + octeon_write_csr(oct, CN23XX_SLI_IQ_SIZE(iq_no), iq->max_count); + + /* Remember the doorbell & instruction count register addr + * for this queue + */ + iq->doorbell_reg = + (u8 *)oct->mmio[0].hw_addr + CN23XX_SLI_IQ_DOORBELL(iq_no); + iq->inst_cnt_reg = + (u8 *)oct->mmio[0].hw_addr + CN23XX_SLI_IQ_INSTR_COUNT64(iq_no); + dev_dbg(&oct->pci_dev->dev, "InstQ[%d]:dbell reg @ 0x%p instcnt_reg @ 0x%p\n", + iq_no, iq->doorbell_reg, iq->inst_cnt_reg); + + /* Store the current instruction counter (used in flush_iq + * calculation) + */ + pkt_in_done = readq(iq->inst_cnt_reg); + + if (oct->msix_on) { + /* Set CINT_ENB to enable IQ interrupt */ + writeq((pkt_in_done | CN23XX_INTR_CINT_ENB), + iq->inst_cnt_reg); + } else { + /* Clear the count by writing back what we read, but don't + * enable interrupts + */ + writeq(pkt_in_done, iq->inst_cnt_reg); + } + + iq->reset_instr_cnt = 0; +} + +static void cn23xx_setup_oq_regs(struct octeon_device *oct, u32 oq_no) +{ + u32 reg_val; + struct octeon_droq *droq = oct->droq[oq_no]; + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + u64 time_threshold; + u64 cnt_threshold; + + oq_no += oct->sriov_info.pf_srn; + + octeon_write_csr64(oct, CN23XX_SLI_OQ_BASE_ADDR64(oq_no), + droq->desc_ring_dma); + octeon_write_csr(oct, CN23XX_SLI_OQ_SIZE(oq_no), droq->max_count); + + octeon_write_csr(oct, CN23XX_SLI_OQ_BUFF_INFO_SIZE(oq_no), + (droq->buffer_size | (OCT_RH_SIZE << 16))); + + /* Get the mapped address of the pkt_sent and pkts_credit regs */ + droq->pkts_sent_reg = + (u8 *)oct->mmio[0].hw_addr + CN23XX_SLI_OQ_PKTS_SENT(oq_no); + droq->pkts_credit_reg = + (u8 *)oct->mmio[0].hw_addr + CN23XX_SLI_OQ_PKTS_CREDIT(oq_no); + + if (!oct->msix_on) { + /* Enable this output queue to generate Packet Timer Interrupt + */ + reg_val = + octeon_read_csr(oct, CN23XX_SLI_OQ_PKT_CONTROL(oq_no)); + reg_val |= CN23XX_PKT_OUTPUT_CTL_TENB; + octeon_write_csr(oct, CN23XX_SLI_OQ_PKT_CONTROL(oq_no), + reg_val); + + /* Enable this output queue to generate Packet Count Interrupt + */ + reg_val = + octeon_read_csr(oct, CN23XX_SLI_OQ_PKT_CONTROL(oq_no)); + reg_val |= CN23XX_PKT_OUTPUT_CTL_CENB; + octeon_write_csr(oct, CN23XX_SLI_OQ_PKT_CONTROL(oq_no), + reg_val); + } else { + time_threshold = cn23xx_pf_get_oq_ticks( + oct, (u32)CFG_GET_OQ_INTR_TIME(cn23xx->conf)); + cnt_threshold = (u32)CFG_GET_OQ_INTR_PKT(cn23xx->conf); + + octeon_write_csr64( + oct, CN23XX_SLI_OQ_PKT_INT_LEVELS(oq_no), + ((time_threshold << 32 | cnt_threshold))); + } +} + +static int cn23xx_enable_io_queues(struct octeon_device *oct) +{ + u64 reg_val; + u32 srn, ern, q_no; + u32 loop = 1000; + + srn = oct->sriov_info.pf_srn; + ern = srn + oct->num_iqs; + + for (q_no = srn; q_no < ern; q_no++) { + /* set the corresponding IQ IS_64B bit */ + if (oct->io_qmask.iq64B & BIT_ULL(q_no - srn)) { + reg_val = octeon_read_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no)); + reg_val = reg_val | CN23XX_PKT_INPUT_CTL_IS_64B; + octeon_write_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), reg_val); + } + + /* set the corresponding IQ ENB bit */ + if (oct->io_qmask.iq & BIT_ULL(q_no - srn)) { + /* IOQs are in reset by default in PEM2 mode, + * clearing reset bit + */ + reg_val = octeon_read_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no)); + + if (reg_val & CN23XX_PKT_INPUT_CTL_RST) { + while ((reg_val & CN23XX_PKT_INPUT_CTL_RST) && + !(reg_val & + CN23XX_PKT_INPUT_CTL_QUIET) && + loop--) { + reg_val = octeon_read_csr64( + oct, + CN23XX_SLI_IQ_PKT_CONTROL64(q_no)); + } + if (!loop) { + dev_err(&oct->pci_dev->dev, + "clearing the reset reg failed or setting the quiet reg failed for qno: %u\n", + q_no); + return -1; + } + reg_val = reg_val & ~CN23XX_PKT_INPUT_CTL_RST; + octeon_write_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), + reg_val); + + reg_val = octeon_read_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no)); + if (reg_val & CN23XX_PKT_INPUT_CTL_RST) { + dev_err(&oct->pci_dev->dev, + "clearing the reset failed for qno: %u\n", + q_no); + return -1; + } + } + reg_val = octeon_read_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no)); + reg_val = reg_val | CN23XX_PKT_INPUT_CTL_RING_ENB; + octeon_write_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), reg_val); + } + } + for (q_no = srn; q_no < ern; q_no++) { + u32 reg_val; + /* set the corresponding OQ ENB bit */ + if (oct->io_qmask.oq & BIT_ULL(q_no - srn)) { + reg_val = octeon_read_csr( + oct, CN23XX_SLI_OQ_PKT_CONTROL(q_no)); + reg_val = reg_val | CN23XX_PKT_OUTPUT_CTL_RING_ENB; + octeon_write_csr(oct, CN23XX_SLI_OQ_PKT_CONTROL(q_no), + reg_val); + } + } + return 0; +} + +static void cn23xx_disable_io_queues(struct octeon_device *oct) +{ + int q_no, loop; + u64 d64; + u32 d32; + u32 srn, ern; + + srn = oct->sriov_info.pf_srn; + ern = srn + oct->num_iqs; + + /*** Disable Input Queues. ***/ + for (q_no = srn; q_no < ern; q_no++) { + loop = HZ; + + /* start the Reset for a particular ring */ + WRITE_ONCE(d64, octeon_read_csr64( + oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no))); + WRITE_ONCE(d64, READ_ONCE(d64) & + (~(CN23XX_PKT_INPUT_CTL_RING_ENB))); + WRITE_ONCE(d64, READ_ONCE(d64) | CN23XX_PKT_INPUT_CTL_RST); + octeon_write_csr64(oct, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), + READ_ONCE(d64)); + + /* Wait until hardware indicates that the particular IQ + * is out of reset. + */ + WRITE_ONCE(d64, octeon_read_csr64( + oct, CN23XX_SLI_PKT_IOQ_RING_RST)); + while (!(READ_ONCE(d64) & BIT_ULL(q_no)) && loop--) { + WRITE_ONCE(d64, octeon_read_csr64( + oct, CN23XX_SLI_PKT_IOQ_RING_RST)); + schedule_timeout_uninterruptible(1); + } + + /* Reset the doorbell register for this Input Queue. */ + octeon_write_csr(oct, CN23XX_SLI_IQ_DOORBELL(q_no), 0xFFFFFFFF); + while (octeon_read_csr64(oct, CN23XX_SLI_IQ_DOORBELL(q_no)) && + loop--) { + schedule_timeout_uninterruptible(1); + } + } + + /*** Disable Output Queues. ***/ + for (q_no = srn; q_no < ern; q_no++) { + loop = HZ; + + /* Wait until hardware indicates that the particular IQ + * is out of reset.It given that SLI_PKT_RING_RST is + * common for both IQs and OQs + */ + WRITE_ONCE(d64, octeon_read_csr64( + oct, CN23XX_SLI_PKT_IOQ_RING_RST)); + while (!(READ_ONCE(d64) & BIT_ULL(q_no)) && loop--) { + WRITE_ONCE(d64, octeon_read_csr64( + oct, CN23XX_SLI_PKT_IOQ_RING_RST)); + schedule_timeout_uninterruptible(1); + } + + /* Reset the doorbell register for this Output Queue. */ + octeon_write_csr(oct, CN23XX_SLI_OQ_PKTS_CREDIT(q_no), + 0xFFFFFFFF); + while (octeon_read_csr64(oct, + CN23XX_SLI_OQ_PKTS_CREDIT(q_no)) && + loop--) { + schedule_timeout_uninterruptible(1); + } + + /* clear the SLI_PKT(0..63)_CNTS[CNT] reg value */ + WRITE_ONCE(d32, octeon_read_csr( + oct, CN23XX_SLI_OQ_PKTS_SENT(q_no))); + octeon_write_csr(oct, CN23XX_SLI_OQ_PKTS_SENT(q_no), + READ_ONCE(d32)); + } +} + +static u64 cn23xx_pf_msix_interrupt_handler(void *dev) +{ + struct octeon_ioq_vector *ioq_vector = (struct octeon_ioq_vector *)dev; + struct octeon_device *oct = ioq_vector->oct_dev; + u64 pkts_sent; + u64 ret = 0; + struct octeon_droq *droq = oct->droq[ioq_vector->droq_index]; + + dev_dbg(&oct->pci_dev->dev, "In %s octeon_dev @ %p\n", __func__, oct); + + if (!droq) { + dev_err(&oct->pci_dev->dev, "23XX bringup FIXME: oct pfnum:%d ioq_vector->ioq_num :%d droq is NULL\n", + oct->pf_num, ioq_vector->ioq_num); + return 0; + } + + pkts_sent = readq(droq->pkts_sent_reg); + + /* If our device has interrupted, then proceed. Also check + * for all f's if interrupt was triggered on an error + * and the PCI read fails. + */ + if (!pkts_sent || (pkts_sent == 0xFFFFFFFFFFFFFFFFULL)) + return ret; + + /* Write count reg in sli_pkt_cnts to clear these int.*/ + if ((pkts_sent & CN23XX_INTR_PO_INT) || + (pkts_sent & CN23XX_INTR_PI_INT)) { + if (pkts_sent & CN23XX_INTR_PO_INT) + ret |= MSIX_PO_INT; + } + + if (pkts_sent & CN23XX_INTR_PI_INT) + /* We will clear the count when we update the read_index. */ + ret |= MSIX_PI_INT; + + /* Never need to handle msix mbox intr for pf. They arrive on the last + * msix + */ + return ret; +} + +static irqreturn_t cn23xx_interrupt_handler(void *dev) +{ + struct octeon_device *oct = (struct octeon_device *)dev; + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + u64 intr64; + + dev_dbg(&oct->pci_dev->dev, "In %s octeon_dev @ %p\n", __func__, oct); + intr64 = readq(cn23xx->intr_sum_reg64); + + oct->int_status = 0; + + if (intr64 & CN23XX_INTR_ERR) + dev_err(&oct->pci_dev->dev, "OCTEON[%d]: Error Intr: 0x%016llx\n", + oct->octeon_id, CVM_CAST64(intr64)); + + if (oct->msix_on != LIO_FLAG_MSIX_ENABLED) { + if (intr64 & CN23XX_INTR_PKT_DATA) + oct->int_status |= OCT_DEV_INTR_PKT_DATA; + } + + if (intr64 & (CN23XX_INTR_DMA0_FORCE)) + oct->int_status |= OCT_DEV_INTR_DMA0_FORCE; + if (intr64 & (CN23XX_INTR_DMA1_FORCE)) + oct->int_status |= OCT_DEV_INTR_DMA1_FORCE; + + /* Clear the current interrupts */ + writeq(intr64, cn23xx->intr_sum_reg64); + + return IRQ_HANDLED; +} + +static void cn23xx_bar1_idx_setup(struct octeon_device *oct, u64 core_addr, + u32 idx, int valid) +{ + u64 bar1; + u64 reg_adr; + + if (!valid) { + reg_adr = lio_pci_readq( + oct, CN23XX_PEM_BAR1_INDEX_REG(oct->pcie_port, idx)); + WRITE_ONCE(bar1, reg_adr); + lio_pci_writeq(oct, (READ_ONCE(bar1) & 0xFFFFFFFEULL), + CN23XX_PEM_BAR1_INDEX_REG(oct->pcie_port, idx)); + reg_adr = lio_pci_readq( + oct, CN23XX_PEM_BAR1_INDEX_REG(oct->pcie_port, idx)); + WRITE_ONCE(bar1, reg_adr); + return; + } + + /* The PEM(0..3)_BAR1_INDEX(0..15)[ADDR_IDX]<23:4> stores + * bits <41:22> of the Core Addr + */ + lio_pci_writeq(oct, (((core_addr >> 22) << 4) | PCI_BAR1_MASK), + CN23XX_PEM_BAR1_INDEX_REG(oct->pcie_port, idx)); + + WRITE_ONCE(bar1, lio_pci_readq( + oct, CN23XX_PEM_BAR1_INDEX_REG(oct->pcie_port, idx))); +} + +static void cn23xx_bar1_idx_write(struct octeon_device *oct, u32 idx, u32 mask) +{ + lio_pci_writeq(oct, mask, + CN23XX_PEM_BAR1_INDEX_REG(oct->pcie_port, idx)); +} + +static u32 cn23xx_bar1_idx_read(struct octeon_device *oct, u32 idx) +{ + return (u32)lio_pci_readq( + oct, CN23XX_PEM_BAR1_INDEX_REG(oct->pcie_port, idx)); +} + +/* always call with lock held */ +static u32 cn23xx_update_read_index(struct octeon_instr_queue *iq) +{ + u32 new_idx; + u32 last_done; + u32 pkt_in_done = readl(iq->inst_cnt_reg); + + last_done = pkt_in_done - iq->pkt_in_done; + iq->pkt_in_done = pkt_in_done; + + /* Modulo of the new index with the IQ size will give us + * the new index. The iq->reset_instr_cnt is always zero for + * cn23xx, so no extra adjustments are needed. + */ + new_idx = (iq->octeon_read_index + + (u32)(last_done & CN23XX_PKT_IN_DONE_CNT_MASK)) % + iq->max_count; + + return new_idx; +} + +static void cn23xx_enable_pf_interrupt(struct octeon_device *oct, u8 intr_flag) +{ + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + u64 intr_val = 0; + + /* Divide the single write to multiple writes based on the flag. */ + /* Enable Interrupt */ + if (intr_flag == OCTEON_ALL_INTR) { + writeq(cn23xx->intr_mask64, cn23xx->intr_enb_reg64); + } else if (intr_flag & OCTEON_OUTPUT_INTR) { + intr_val = readq(cn23xx->intr_enb_reg64); + intr_val |= CN23XX_INTR_PKT_DATA; + writeq(intr_val, cn23xx->intr_enb_reg64); + } +} + +static void cn23xx_disable_pf_interrupt(struct octeon_device *oct, u8 intr_flag) +{ + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + u64 intr_val = 0; + + /* Disable Interrupts */ + if (intr_flag == OCTEON_ALL_INTR) { + writeq(0, cn23xx->intr_enb_reg64); + } else if (intr_flag & OCTEON_OUTPUT_INTR) { + intr_val = readq(cn23xx->intr_enb_reg64); + intr_val &= ~CN23XX_INTR_PKT_DATA; + writeq(intr_val, cn23xx->intr_enb_reg64); + } +} + +static void cn23xx_get_pcie_qlmport(struct octeon_device *oct) +{ + oct->pcie_port = (octeon_read_csr(oct, CN23XX_SLI_MAC_NUMBER)) & 0xff; + + dev_dbg(&oct->pci_dev->dev, "OCTEON: CN23xx uses PCIE Port %d\n", + oct->pcie_port); +} + +static void cn23xx_get_pf_num(struct octeon_device *oct) +{ + u32 fdl_bit = 0; + + /** Read Function Dependency Link reg to get the function number */ + pci_read_config_dword(oct->pci_dev, CN23XX_PCIE_SRIOV_FDL, &fdl_bit); + oct->pf_num = ((fdl_bit >> CN23XX_PCIE_SRIOV_FDL_BIT_POS) & + CN23XX_PCIE_SRIOV_FDL_MASK); +} + +static void cn23xx_setup_reg_address(struct octeon_device *oct) +{ + u8 __iomem *bar0_pciaddr = oct->mmio[0].hw_addr; + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + + oct->reg_list.pci_win_wr_addr_hi = + (u32 __iomem *)(bar0_pciaddr + CN23XX_WIN_WR_ADDR_HI); + oct->reg_list.pci_win_wr_addr_lo = + (u32 __iomem *)(bar0_pciaddr + CN23XX_WIN_WR_ADDR_LO); + oct->reg_list.pci_win_wr_addr = + (u64 __iomem *)(bar0_pciaddr + CN23XX_WIN_WR_ADDR64); + + oct->reg_list.pci_win_rd_addr_hi = + (u32 __iomem *)(bar0_pciaddr + CN23XX_WIN_RD_ADDR_HI); + oct->reg_list.pci_win_rd_addr_lo = + (u32 __iomem *)(bar0_pciaddr + CN23XX_WIN_RD_ADDR_LO); + oct->reg_list.pci_win_rd_addr = + (u64 __iomem *)(bar0_pciaddr + CN23XX_WIN_RD_ADDR64); + + oct->reg_list.pci_win_wr_data_hi = + (u32 __iomem *)(bar0_pciaddr + CN23XX_WIN_WR_DATA_HI); + oct->reg_list.pci_win_wr_data_lo = + (u32 __iomem *)(bar0_pciaddr + CN23XX_WIN_WR_DATA_LO); + oct->reg_list.pci_win_wr_data = + (u64 __iomem *)(bar0_pciaddr + CN23XX_WIN_WR_DATA64); + + oct->reg_list.pci_win_rd_data_hi = + (u32 __iomem *)(bar0_pciaddr + CN23XX_WIN_RD_DATA_HI); + oct->reg_list.pci_win_rd_data_lo = + (u32 __iomem *)(bar0_pciaddr + CN23XX_WIN_RD_DATA_LO); + oct->reg_list.pci_win_rd_data = + (u64 __iomem *)(bar0_pciaddr + CN23XX_WIN_RD_DATA64); + + cn23xx_get_pcie_qlmport(oct); + + cn23xx->intr_mask64 = CN23XX_INTR_MASK; + if (!oct->msix_on) + cn23xx->intr_mask64 |= CN23XX_INTR_PKT_TIME; + if (oct->rev_id >= OCTEON_CN23XX_REV_1_1) + cn23xx->intr_mask64 |= CN23XX_INTR_VF_MBOX; + + cn23xx->intr_sum_reg64 = + bar0_pciaddr + + CN23XX_SLI_MAC_PF_INT_SUM64(oct->pcie_port, oct->pf_num); + cn23xx->intr_enb_reg64 = + bar0_pciaddr + + CN23XX_SLI_MAC_PF_INT_ENB64(oct->pcie_port, oct->pf_num); +} + +static int cn23xx_sriov_config(struct octeon_device *oct) +{ + u32 total_rings; + struct octeon_cn23xx_pf *cn23xx = (struct octeon_cn23xx_pf *)oct->chip; + /* num_vfs is already filled for us */ + u32 pf_srn, num_pf_rings; + + cn23xx->conf = + (struct octeon_config *)oct_get_config_info(oct, LIO_23XX); + switch (oct->rev_id) { + case OCTEON_CN23XX_REV_1_0: + total_rings = CN23XX_MAX_RINGS_PER_PF_PASS_1_0; + break; + case OCTEON_CN23XX_REV_1_1: + total_rings = CN23XX_MAX_RINGS_PER_PF_PASS_1_1; + break; + default: + total_rings = CN23XX_MAX_RINGS_PER_PF; + break; + } + if (!oct->sriov_info.num_pf_rings) { + if (total_rings > num_present_cpus()) + num_pf_rings = num_present_cpus(); + else + num_pf_rings = total_rings; + } else { + num_pf_rings = oct->sriov_info.num_pf_rings; + + if (num_pf_rings > total_rings) { + dev_warn(&oct->pci_dev->dev, + "num_queues_per_pf requested %u is more than available rings. Reducing to %u\n", + num_pf_rings, total_rings); + num_pf_rings = total_rings; + } + } + + total_rings = num_pf_rings; + /* the first ring of the pf */ + pf_srn = total_rings - num_pf_rings; + + oct->sriov_info.trs = total_rings; + oct->sriov_info.pf_srn = pf_srn; + oct->sriov_info.num_pf_rings = num_pf_rings; + dev_dbg(&oct->pci_dev->dev, "trs:%d pf_srn:%d num_pf_rings:%d\n", + oct->sriov_info.trs, oct->sriov_info.pf_srn, + oct->sriov_info.num_pf_rings); + return 0; +} + +int setup_cn23xx_octeon_pf_device(struct octeon_device *oct) +{ + if (octeon_map_pci_barx(oct, 0, 0)) + return 1; + + if (octeon_map_pci_barx(oct, 1, MAX_BAR1_IOREMAP_SIZE)) { + dev_err(&oct->pci_dev->dev, "%s CN23XX BAR1 map failed\n", + __func__); + octeon_unmap_pci_barx(oct, 0); + return 1; + } + + cn23xx_get_pf_num(oct); + + if (cn23xx_sriov_config(oct)) { + octeon_unmap_pci_barx(oct, 0); + octeon_unmap_pci_barx(oct, 1); + return 1; + } + + octeon_write_csr64(oct, CN23XX_SLI_MAC_CREDIT_CNT, 0x3F802080802080ULL); + + oct->fn_list.setup_iq_regs = cn23xx_setup_iq_regs; + oct->fn_list.setup_oq_regs = cn23xx_setup_oq_regs; + oct->fn_list.process_interrupt_regs = cn23xx_interrupt_handler; + oct->fn_list.msix_interrupt_handler = cn23xx_pf_msix_interrupt_handler; + + oct->fn_list.soft_reset = cn23xx_pf_soft_reset; + oct->fn_list.setup_device_regs = cn23xx_setup_pf_device_regs; + oct->fn_list.update_iq_read_idx = cn23xx_update_read_index; + + oct->fn_list.bar1_idx_setup = cn23xx_bar1_idx_setup; + oct->fn_list.bar1_idx_write = cn23xx_bar1_idx_write; + oct->fn_list.bar1_idx_read = cn23xx_bar1_idx_read; + + oct->fn_list.enable_interrupt = cn23xx_enable_pf_interrupt; + oct->fn_list.disable_interrupt = cn23xx_disable_pf_interrupt; + + oct->fn_list.enable_io_queues = cn23xx_enable_io_queues; + oct->fn_list.disable_io_queues = cn23xx_disable_io_queues; + + cn23xx_setup_reg_address(oct); + + oct->coproc_clock_rate = 1000000ULL * cn23xx_coprocessor_clock(oct); + + return 0; +} + +int validate_cn23xx_pf_config_info(struct octeon_device *oct, + struct octeon_config *conf23xx) +{ + if (CFG_GET_IQ_MAX_Q(conf23xx) > CN23XX_MAX_INPUT_QUEUES) { + dev_err(&oct->pci_dev->dev, "%s: Num IQ (%d) exceeds Max (%d)\n", + __func__, CFG_GET_IQ_MAX_Q(conf23xx), + CN23XX_MAX_INPUT_QUEUES); + return 1; + } + + if (CFG_GET_OQ_MAX_Q(conf23xx) > CN23XX_MAX_OUTPUT_QUEUES) { + dev_err(&oct->pci_dev->dev, "%s: Num OQ (%d) exceeds Max (%d)\n", + __func__, CFG_GET_OQ_MAX_Q(conf23xx), + CN23XX_MAX_OUTPUT_QUEUES); + return 1; + } + + if (CFG_GET_IQ_INSTR_TYPE(conf23xx) != OCTEON_32BYTE_INSTR && + CFG_GET_IQ_INSTR_TYPE(conf23xx) != OCTEON_64BYTE_INSTR) { + dev_err(&oct->pci_dev->dev, "%s: Invalid instr type for IQ\n", + __func__); + return 1; + } + + if (!(CFG_GET_OQ_INFO_PTR(conf23xx)) || + !(CFG_GET_OQ_REFILL_THRESHOLD(conf23xx))) { + dev_err(&oct->pci_dev->dev, "%s: Invalid parameter for OQ\n", + __func__); + return 1; + } + + if (!(CFG_GET_OQ_INTR_TIME(conf23xx))) { + dev_err(&oct->pci_dev->dev, "%s: Invalid parameter for OQ\n", + __func__); + return 1; + } + + return 0; +} + +void cn23xx_dump_iq_regs(struct octeon_device *oct) +{ + u32 regval, q_no; + + dev_dbg(&oct->pci_dev->dev, "SLI_IQ_DOORBELL_0 [0x%x]: 0x%016llx\n", + CN23XX_SLI_IQ_DOORBELL(0), + CVM_CAST64(octeon_read_csr64 + (oct, CN23XX_SLI_IQ_DOORBELL(0)))); + + dev_dbg(&oct->pci_dev->dev, "SLI_IQ_BASEADDR_0 [0x%x]: 0x%016llx\n", + CN23XX_SLI_IQ_BASE_ADDR64(0), + CVM_CAST64(octeon_read_csr64 + (oct, CN23XX_SLI_IQ_BASE_ADDR64(0)))); + + dev_dbg(&oct->pci_dev->dev, "SLI_IQ_FIFO_RSIZE_0 [0x%x]: 0x%016llx\n", + CN23XX_SLI_IQ_SIZE(0), + CVM_CAST64(octeon_read_csr64(oct, CN23XX_SLI_IQ_SIZE(0)))); + + dev_dbg(&oct->pci_dev->dev, "SLI_CTL_STATUS [0x%x]: 0x%016llx\n", + CN23XX_SLI_CTL_STATUS, + CVM_CAST64(octeon_read_csr64(oct, CN23XX_SLI_CTL_STATUS))); + + for (q_no = 0; q_no < CN23XX_MAX_INPUT_QUEUES; q_no++) { + dev_dbg(&oct->pci_dev->dev, "SLI_PKT[%d]_INPUT_CTL [0x%x]: 0x%016llx\n", + q_no, CN23XX_SLI_IQ_PKT_CONTROL64(q_no), + CVM_CAST64(octeon_read_csr64 + (oct, + CN23XX_SLI_IQ_PKT_CONTROL64(q_no)))); + } + + pci_read_config_dword(oct->pci_dev, CN23XX_CONFIG_PCIE_DEVCTL, ®val); + dev_dbg(&oct->pci_dev->dev, "Config DevCtl [0x%x]: 0x%08x\n", + CN23XX_CONFIG_PCIE_DEVCTL, regval); + + dev_dbg(&oct->pci_dev->dev, "SLI_PRT[%d]_CFG [0x%llx]: 0x%016llx\n", + oct->pcie_port, CN23XX_DPI_SLI_PRTX_CFG(oct->pcie_port), + CVM_CAST64(lio_pci_readq( + oct, CN23XX_DPI_SLI_PRTX_CFG(oct->pcie_port)))); + + dev_dbg(&oct->pci_dev->dev, "SLI_S2M_PORT[%d]_CTL [0x%x]: 0x%016llx\n", + oct->pcie_port, CN23XX_SLI_S2M_PORTX_CTL(oct->pcie_port), + CVM_CAST64(octeon_read_csr64( + oct, CN23XX_SLI_S2M_PORTX_CTL(oct->pcie_port)))); +} + +int cn23xx_fw_loaded(struct octeon_device *oct) +{ + u64 val; + + val = octeon_read_csr64(oct, CN23XX_SLI_SCRATCH1); + return (val >> 1) & 1ULL; +} diff --git a/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_device.h b/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_device.h new file mode 100644 index 000000000000..21b5c9051967 --- /dev/null +++ b/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_device.h @@ -0,0 +1,59 @@ +/********************************************************************** +* Author: Cavium, Inc. +* +* Contact: support@cavium.com +* Please include "LiquidIO" in the subject. +* +* Copyright (c) 2003-2015 Cavium, Inc. +* +* This file is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License, Version 2, as +* published by the Free Software Foundation. +* +* This file is distributed in the hope that it will be useful, but +* AS-IS and WITHOUT ANY WARRANTY; without even the implied warranty +* of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, TITLE, or +* NONINFRINGEMENT. See the GNU General Public License for more +* details. +* +* This file may also be available under a different license from Cavium. +* Contact Cavium, Inc. for more information +**********************************************************************/ + +/*! \file cn23xx_device.h + * \brief Host Driver: Routines that perform CN23XX specific operations. +*/ + +#ifndef __CN23XX_PF_DEVICE_H__ +#define __CN23XX_PF_DEVICE_H__ + +#include "cn23xx_pf_regs.h" + +/* Register address and configuration for a CN23XX devices. + * If device specific changes need to be made then add a struct to include + * device specific fields as shown in the commented section + */ +struct octeon_cn23xx_pf { + /** PCI interrupt summary register */ + u8 __iomem *intr_sum_reg64; + + /** PCI interrupt enable register */ + u8 __iomem *intr_enb_reg64; + + /** The PCI interrupt mask used by interrupt handler */ + u64 intr_mask64; + + struct octeon_config *conf; +}; + +int setup_cn23xx_octeon_pf_device(struct octeon_device *oct); + +int validate_cn23xx_pf_config_info(struct octeon_device *oct, + struct octeon_config *conf23xx); + +u32 cn23xx_pf_get_oq_ticks(struct octeon_device *oct, u32 time_intr_in_us); + +void cn23xx_dump_pf_initialized_regs(struct octeon_device *oct); + +int cn23xx_fw_loaded(struct octeon_device *oct); +#endif diff --git a/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_regs.h b/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_regs.h new file mode 100644 index 000000000000..03d79d95ab75 --- /dev/null +++ b/drivers/net/ethernet/cavium/liquidio/cn23xx_pf_regs.h @@ -0,0 +1,604 @@ +/********************************************************************** +* Author: Cavium, Inc. +* +* Contact: support@cavium.com +* Please include "LiquidIO" in the subject. +* +* Copyright (c) 2003-2015 Cavium, Inc. +* +* This file is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License, Version 2, as +* published by the Free Software Foundation. +* +* This file is distributed in the hope that it will be useful, but +* AS-IS and WITHOUT ANY WARRANTY; without even the implied warranty +* of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, TITLE, or +* NONINFRINGEMENT. See the GNU General Public License for more +* details. +* +* This file may also be available under a different license from Cavium. +* Contact Cavium, Inc. for more information +**********************************************************************/ + +/*! \file cn23xx_regs.h + * \brief Host Driver: Register Address and Register Mask values for + * Octeon CN23XX devices. +*/ + +#ifndef __CN23XX_PF_REGS_H__ +#define __CN23XX_PF_REGS_H__ + +#define CN23XX_CONFIG_VENDOR_ID 0x00 +#define CN23XX_CONFIG_DEVICE_ID 0x02 + +#define CN23XX_CONFIG_XPANSION_BAR 0x38 + +#define CN23XX_CONFIG_MSIX_CAP 0x50 +#define CN23XX_CONFIG_MSIX_LMSI 0x54 +#define CN23XX_CONFIG_MSIX_UMSI 0x58 +#define CN23XX_CONFIG_MSIX_MSIMD 0x5C +#define CN23XX_CONFIG_MSIX_MSIMM 0x60 +#define CN23XX_CONFIG_MSIX_MSIMP 0x64 + +#define CN23XX_CONFIG_PCIE_CAP 0x70 +#define CN23XX_CONFIG_PCIE_DEVCAP 0x74 +#define CN23XX_CONFIG_PCIE_DEVCTL 0x78 +#define CN23XX_CONFIG_PCIE_LINKCAP 0x7C +#define CN23XX_CONFIG_PCIE_LINKCTL 0x80 +#define CN23XX_CONFIG_PCIE_SLOTCAP 0x84 +#define CN23XX_CONFIG_PCIE_SLOTCTL 0x88 +#define CN23XX_CONFIG_PCIE_DEVCTL2 0x98 +#define CN23XX_CONFIG_PCIE_LINKCTL2 0xA0 +#define CN23XX_CONFIG_PCIE_UNCORRECT_ERR_MASK 0x108 +#define CN23XX_CONFIG_PCIE_CORRECT_ERR_STATUS 0x110 +#define CN23XX_CONFIG_PCIE_DEVCTL_MASK 0x00040000 + +#define CN23XX_PCIE_SRIOV_FDL 0x188 +#define CN23XX_PCIE_SRIOV_FDL_BIT_POS 0x10 +#define CN23XX_PCIE_SRIOV_FDL_MASK 0xFF + +#define CN23XX_CONFIG_PCIE_FLTMSK 0x720 + +#define CN23XX_CONFIG_SRIOV_VFDEVID 0x190 + +#define CN23XX_CONFIG_SRIOV_BAR_START 0x19C +#define CN23XX_CONFIG_SRIOV_BARX(i) \ + (CN23XX_CONFIG_SRIOV_BAR_START + (i * 4)) +#define CN23XX_CONFIG_SRIOV_BAR_PF 0x08 +#define CN23XX_CONFIG_SRIOV_BAR_64BIT 0x04 +#define CN23XX_CONFIG_SRIOV_BAR_IO 0x01 + +/* ############## BAR0 Registers ################ */ + +#define CN23XX_SLI_CTL_PORT_START 0x286E0 +#define CN23XX_PORT_OFFSET 0x10 + +#define CN23XX_SLI_CTL_PORT(p) \ + (CN23XX_SLI_CTL_PORT_START + ((p) * CN23XX_PORT_OFFSET)) + +/* 2 scatch registers (64-bit) */ +#define CN23XX_SLI_WINDOW_CTL 0x282E0 +#define CN23XX_SLI_SCRATCH1 0x283C0 +#define CN23XX_SLI_SCRATCH2 0x283D0 +#define CN23XX_SLI_WINDOW_CTL_DEFAULT 0x200000ULL + +/* 1 registers (64-bit) - SLI_CTL_STATUS */ +#define CN23XX_SLI_CTL_STATUS 0x28570 + +/* SLI Packet Input Jabber Register (64 bit register) + * <31:0> for Byte count for limiting sizes of packet sizes + * that are allowed for sli packet inbound packets. + * the default value is 0xFA00(=64000). + */ +#define CN23XX_SLI_PKT_IN_JABBER 0x29170 +/* The input jabber is used to determine the TSO max size. + * Due to H/W limitation, this need to be reduced to 60000 + * in order to to H/W TSO and avoid the WQE malfarmation + * PKO_BUG_24989_WQE_LEN + */ +#define CN23XX_DEFAULT_INPUT_JABBER 0xEA60 /*60000*/ + +#define CN23XX_WIN_WR_ADDR_LO 0x20000 +#define CN23XX_WIN_WR_ADDR_HI 0x20004 +#define CN23XX_WIN_WR_ADDR64 CN23XX_WIN_WR_ADDR_LO + +#define CN23XX_WIN_RD_ADDR_LO 0x20010 +#define CN23XX_WIN_RD_ADDR_HI 0x20014 +#define CN23XX_WIN_RD_ADDR64 CN23XX_WIN_RD_ADDR_LO + +#define CN23XX_WIN_WR_DATA_LO 0x20020 +#define CN23XX_WIN_WR_DATA_HI 0x20024 +#define CN23XX_WIN_WR_DATA64 CN23XX_WIN_WR_DATA_LO + +#define CN23XX_WIN_RD_DATA_LO 0x20040 +#define CN23XX_WIN_RD_DATA_HI 0x20044 +#define CN23XX_WIN_RD_DATA64 CN23XX_WIN_RD_DATA_LO + +#define CN23XX_WIN_WR_MASK_LO 0x20030 +#define CN23XX_WIN_WR_MASK_HI 0x20034 +#define CN23XX_WIN_WR_MASK_REG CN23XX_WIN_WR_MASK_LO +#define CN23XX_SLI_MAC_CREDIT_CNT 0x23D70 + +/* 4 registers (64-bit) for mapping IOQs to MACs(PEMs)- + * SLI_PKT_MAC(0..3)_PF(0..1)_RINFO + */ +#define CN23XX_SLI_PKT_MAC_RINFO_START64 0x29030 + +/*1 register (64-bit) to determine whether IOQs are in reset. */ +#define CN23XX_SLI_PKT_IOQ_RING_RST 0x291E0 + +/* Each Input Queue register is at a 16-byte Offset in BAR0 */ +#define CN23XX_IQ_OFFSET 0x20000 + +#define CN23XX_MAC_RINFO_OFFSET 0x20 +#define CN23XX_PF_RINFO_OFFSET 0x10 + +#define CN23XX_SLI_PKT_MAC_RINFO64(mac, pf) \ + (CN23XX_SLI_PKT_MAC_RINFO_START64 + \ + ((mac) * CN23XX_MAC_RINFO_OFFSET) + \ + ((pf) * CN23XX_PF_RINFO_OFFSET)) + +/** mask for total rings, setting TRS to base */ +#define CN23XX_PKT_MAC_CTL_RINFO_TRS BIT_ULL(16) +/** mask for starting ring number: setting SRN <6:0> = 0x7F */ +#define CN23XX_PKT_MAC_CTL_RINFO_SRN (0x7F) + +/* Starting bit of the TRS field in CN23XX_SLI_PKT_MAC_RINFO64 register */ +#define CN23XX_PKT_MAC_CTL_RINFO_TRS_BIT_POS 16 +/* Starting bit of SRN field in CN23XX_SLI_PKT_MAC_RINFO64 register */ +#define CN23XX_PKT_MAC_CTL_RINFO_SRN_BIT_POS 0 +/* Starting bit of RPVF field in CN23XX_SLI_PKT_MAC_RINFO64 register */ +#define CN23XX_PKT_MAC_CTL_RINFO_RPVF_BIT_POS 32 +/* Starting bit of NVFS field in CN23XX_SLI_PKT_MAC_RINFO64 register */ +#define CN23XX_PKT_MAC_CTL_RINFO_NVFS_BIT_POS 48 + +/*###################### REQUEST QUEUE #########################*/ + +/* 64 registers for Input Queue Instr Count - SLI_PKT_IN_DONE0_CNTS */ +#define CN23XX_SLI_IQ_INSTR_COUNT_START64 0x10040 + +/* 64 registers for Input Queues Start Addr - SLI_PKT0_INSTR_BADDR */ +#define CN23XX_SLI_IQ_BASE_ADDR_START64 0x10010 + +/* 64 registers for Input Doorbell - SLI_PKT0_INSTR_BAOFF_DBELL */ +#define CN23XX_SLI_IQ_DOORBELL_START 0x10020 + +/* 64 registers for Input Queue size - SLI_PKT0_INSTR_FIFO_RSIZE */ +#define CN23XX_SLI_IQ_SIZE_START 0x10030 + +/* 64 registers (64-bit) - ES, RO, NS, Arbitration for Input Queue Data & + * gather list fetches. SLI_PKT(0..63)_INPUT_CONTROL. + */ +#define CN23XX_SLI_IQ_PKT_CONTROL_START64 0x10000 + +/*------- Request Queue Macros ---------*/ +#define CN23XX_SLI_IQ_PKT_CONTROL64(iq) \ + (CN23XX_SLI_IQ_PKT_CONTROL_START64 + ((iq) * CN23XX_IQ_OFFSET)) + +#define CN23XX_SLI_IQ_BASE_ADDR64(iq) \ + (CN23XX_SLI_IQ_BASE_ADDR_START64 + ((iq) * CN23XX_IQ_OFFSET)) + +#define CN23XX_SLI_IQ_SIZE(iq) \ + (CN23XX_SLI_IQ_SIZE_START + ((iq) * CN23XX_IQ_OFFSET)) + +#define CN23XX_SLI_IQ_DOORBELL(iq) \ + (CN23XX_SLI_IQ_DOORBELL_START + ((iq) * CN23XX_IQ_OFFSET)) + +#define CN23XX_SLI_IQ_INSTR_COUNT64(iq) \ + (CN23XX_SLI_IQ_INSTR_COUNT_START64 + ((iq) * CN23XX_IQ_OFFSET)) + +/*------------------ Masks ----------------*/ +#define CN23XX_PKT_INPUT_CTL_VF_NUM BIT_ULL(32) +#define CN23XX_PKT_INPUT_CTL_MAC_NUM BIT(29) +/* Number of instructions to be read in one MAC read request. + * setting to Max value(4) + */ +#define CN23XX_PKT_INPUT_CTL_RDSIZE (3 << 25) +#define CN23XX_PKT_INPUT_CTL_IS_64B BIT(24) +#define CN23XX_PKT_INPUT_CTL_RST BIT(23) +#define CN23XX_PKT_INPUT_CTL_QUIET BIT(28) +#define CN23XX_PKT_INPUT_CTL_RING_ENB BIT(22) +#define CN23XX_PKT_INPUT_CTL_DATA_NS BIT(8) +#define CN23XX_PKT_INPUT_CTL_DATA_ES_64B_SWAP BIT(6) +#define CN23XX_PKT_INPUT_CTL_DATA_RO BIT(5) +#define CN23XX_PKT_INPUT_CTL_USE_CSR BIT(4) +#define CN23XX_PKT_INPUT_CTL_GATHER_NS BIT(3) +#define CN23XX_PKT_INPUT_CTL_GATHER_ES_64B_SWAP (2) +#define CN23XX_PKT_INPUT_CTL_GATHER_RO (1) + +/** Rings per Virtual Function **/ +#define CN23XX_PKT_INPUT_CTL_RPVF_MASK (0x3F) +#define CN23XX_PKT_INPUT_CTL_RPVF_POS (48) +/** These bits[47:44] select the Physical function number within the MAC */ +#define CN23XX_PKT_INPUT_CTL_PF_NUM_MASK (0x7) +#define CN23XX_PKT_INPUT_CTL_PF_NUM_POS (45) +/** These bits[43:32] select the function number within the PF */ +#define CN23XX_PKT_INPUT_CTL_VF_NUM_MASK (0x1FFF) +#define CN23XX_PKT_INPUT_CTL_VF_NUM_POS (32) +#define CN23XX_PKT_INPUT_CTL_MAC_NUM_MASK (0x3) +#define CN23XX_PKT_INPUT_CTL_MAC_NUM_POS (29) +#define CN23XX_PKT_IN_DONE_WMARK_MASK (0xFFFFULL) +#define CN23XX_PKT_IN_DONE_WMARK_BIT_POS (32) +#define CN23XX_PKT_IN_DONE_CNT_MASK (0x00000000FFFFFFFFULL) + +#ifdef __LITTLE_ENDIAN_BITFIELD +#define CN23XX_PKT_INPUT_CTL_MASK \ + (CN23XX_PKT_INPUT_CTL_RDSIZE | \ + CN23XX_PKT_INPUT_CTL_DATA_ES_64B_SWAP | \ + CN23XX_PKT_INPUT_CTL_USE_CSR) +#else +#define CN23XX_PKT_INPUT_CTL_MASK \ + (CN23XX_PKT_INPUT_CTL_RDSIZE | \ + CN23XX_PKT_INPUT_CTL_DATA_ES_64B_SWAP | \ + CN23XX_PKT_INPUT_CTL_USE_CSR | \ + CN23XX_PKT_INPUT_CTL_GATHER_ES_64B_SWAP) +#endif + +/** Masks for SLI_PKT_IN_DONE(0..63)_CNTS Register */ +#define CN23XX_IN_DONE_CNTS_PI_INT BIT_ULL(62) +#define CN23XX_IN_DONE_CNTS_CINT_ENB BIT_ULL(48) + +/*############################ OUTPUT QUEUE #########################*/ + +/* 64 registers for Output queue control - SLI_PKT(0..63)_OUTPUT_CONTROL */ +#define CN23XX_SLI_OQ_PKT_CONTROL_START 0x10050 + +/* 64 registers for Output queue buffer and info size - SLI_PKT0_OUT_SIZE */ +#define CN23XX_SLI_OQ0_BUFF_INFO_SIZE 0x10060 + +/* 64 registers for Output Queue Start Addr - SLI_PKT0_SLIST_BADDR */ +#define CN23XX_SLI_OQ_BASE_ADDR_START64 0x10070 + +/* 64 registers for Output Queue Packet Credits - SLI_PKT0_SLIST_BAOFF_DBELL */ +#define CN23XX_SLI_OQ_PKT_CREDITS_START 0x10080 + +/* 64 registers for Output Queue size - SLI_PKT0_SLIST_FIFO_RSIZE */ +#define CN23XX_SLI_OQ_SIZE_START 0x10090 + +/* 64 registers for Output Queue Packet Count - SLI_PKT0_CNTS */ +#define CN23XX_SLI_OQ_PKT_SENT_START 0x100B0 + +/* 64 registers for Output Queue INT Levels - SLI_PKT0_INT_LEVELS */ +#define CN23XX_SLI_OQ_PKT_INT_LEVELS_START64 0x100A0 + +/* Each Output Queue register is at a 16-byte Offset in BAR0 */ +#define CN23XX_OQ_OFFSET 0x20000 + +/* 1 (64-bit register) for Output Queue backpressure across all rings. */ +#define CN23XX_SLI_OQ_WMARK 0x29180 + +/* Global pkt control register */ +#define CN23XX_SLI_GBL_CONTROL 0x29210 + +/* Backpressure enable register for PF0 */ +#define CN23XX_SLI_OUT_BP_EN_W1S 0x29260 + +/* Backpressure enable register for PF1 */ +#define CN23XX_SLI_OUT_BP_EN2_W1S 0x29270 + +/* Backpressure disable register for PF0 */ +#define CN23XX_SLI_OUT_BP_EN_W1C 0x29280 + +/* Backpressure disable register for PF1 */ +#define CN23XX_SLI_OUT_BP_EN2_W1C 0x29290 + +/*------- Output Queue Macros ---------*/ + +#define CN23XX_SLI_OQ_PKT_CONTROL(oq) \ + (CN23XX_SLI_OQ_PKT_CONTROL_START + ((oq) * CN23XX_OQ_OFFSET)) + +#define CN23XX_SLI_OQ_BASE_ADDR64(oq) \ + (CN23XX_SLI_OQ_BASE_ADDR_START64 + ((oq) * CN23XX_OQ_OFFSET)) + +#define CN23XX_SLI_OQ_SIZE(oq) \ + (CN23XX_SLI_OQ_SIZE_START + ((oq) * CN23XX_OQ_OFFSET)) + +#define CN23XX_SLI_OQ_BUFF_INFO_SIZE(oq) \ + (CN23XX_SLI_OQ0_BUFF_INFO_SIZE + ((oq) * CN23XX_OQ_OFFSET)) + +#define CN23XX_SLI_OQ_PKTS_SENT(oq) \ + (CN23XX_SLI_OQ_PKT_SENT_START + ((oq) * CN23XX_OQ_OFFSET)) + +#define CN23XX_SLI_OQ_PKTS_CREDIT(oq) \ + (CN23XX_SLI_OQ_PKT_CREDITS_START + ((oq) * CN23XX_OQ_OFFSET)) + +#define CN23XX_SLI_OQ_PKT_INT_LEVELS(oq) \ + (CN23XX_SLI_OQ_PKT_INT_LEVELS_START64 + \ + ((oq) * CN23XX_OQ_OFFSET)) + +/*Macro's for accessing CNT and TIME separately from INT_LEVELS*/ +#define CN23XX_SLI_OQ_PKT_INT_LEVELS_CNT(oq) \ + (CN23XX_SLI_OQ_PKT_INT_LEVELS_START64 + \ + ((oq) * CN23XX_OQ_OFFSET)) + +#define CN23XX_SLI_OQ_PKT_INT_LEVELS_TIME(oq) \ + (CN23XX_SLI_OQ_PKT_INT_LEVELS_START64 + \ + ((oq) * CN23XX_OQ_OFFSET) + 4) + +/*------------------ Masks ----------------*/ +#define CN23XX_PKT_OUTPUT_CTL_TENB BIT(13) +#define CN23XX_PKT_OUTPUT_CTL_CENB BIT(12) +#define CN23XX_PKT_OUTPUT_CTL_IPTR BIT(11) +#define CN23XX_PKT_OUTPUT_CTL_ES BIT(9) +#define CN23XX_PKT_OUTPUT_CTL_NSR BIT(8) +#define CN23XX_PKT_OUTPUT_CTL_ROR BIT(7) +#define CN23XX_PKT_OUTPUT_CTL_DPTR BIT(6) +#define CN23XX_PKT_OUTPUT_CTL_BMODE BIT(5) +#define CN23XX_PKT_OUTPUT_CTL_ES_P BIT(3) +#define CN23XX_PKT_OUTPUT_CTL_NSR_P BIT(2) +#define CN23XX_PKT_OUTPUT_CTL_ROR_P BIT(1) +#define CN23XX_PKT_OUTPUT_CTL_RING_ENB BIT(0) + +/*######################### Mailbox Reg Macros ########################*/ +#define CN23XX_SLI_PKT_MBOX_INT_START 0x10210 +#define CN23XX_SLI_PKT_PF_VF_MBOX_SIG_START 0x10200 +#define CN23XX_SLI_MAC_PF_MBOX_INT_START 0x27380 + +#define CN23XX_SLI_MBOX_OFFSET 0x20000 +#define CN23XX_SLI_MBOX_SIG_IDX_OFFSET 0x8 + +#define CN23XX_SLI_PKT_MBOX_INT(q) \ + (CN23XX_SLI_PKT_MBOX_INT_START + ((q) * CN23XX_SLI_MBOX_OFFSET)) + +#define CN23XX_SLI_PKT_PF_VF_MBOX_SIG(q, idx) \ + (CN23XX_SLI_PKT_PF_VF_MBOX_SIG_START + \ + ((q) * CN23XX_SLI_MBOX_OFFSET + \ + (idx) * CN23XX_SLI_MBOX_SIG_IDX_OFFSET)) + +#define CN23XX_SLI_MAC_PF_MBOX_INT(mac, pf) \ + (CN23XX_SLI_MAC_PF_MBOX_INT_START + \ + ((mac) * CN23XX_MAC_INT_OFFSET + \ + (pf) * CN23XX_PF_INT_OFFSET)) + +/*######################### DMA Counters #########################*/ + +/* 2 registers (64-bit) - DMA Count - 1 for each DMA counter 0/1. */ +#define CN23XX_DMA_CNT_START 0x28400 + +/* 2 registers (64-bit) - DMA Timer 0/1, contains DMA timer values */ +/* SLI_DMA_0_TIM */ +#define CN23XX_DMA_TIM_START 0x28420 + +/* 2 registers (64-bit) - DMA count & Time Interrupt threshold - + * SLI_DMA_0_INT_LEVEL + */ +#define CN23XX_DMA_INT_LEVEL_START 0x283E0 + +/* Each DMA register is at a 16-byte Offset in BAR0 */ +#define CN23XX_DMA_OFFSET 0x10 + +/*---------- DMA Counter Macros ---------*/ +#define CN23XX_DMA_CNT(dq) \ + (CN23XX_DMA_CNT_START + ((dq) * CN23XX_DMA_OFFSET)) + +#define CN23XX_DMA_INT_LEVEL(dq) \ + (CN23XX_DMA_INT_LEVEL_START + ((dq) * CN23XX_DMA_OFFSET)) + +#define CN23XX_DMA_PKT_INT_LEVEL(dq) \ + (CN23XX_DMA_INT_LEVEL_START + ((dq) * CN23XX_DMA_OFFSET)) + +#define CN23XX_DMA_TIME_INT_LEVEL(dq) \ + (CN23XX_DMA_INT_LEVEL_START + 4 + ((dq) * CN23XX_DMA_OFFSET)) + +#define CN23XX_DMA_TIM(dq) \ + (CN23XX_DMA_TIM_START + ((dq) * CN23XX_DMA_OFFSET)) + +/*######################## MSIX TABLE #########################*/ + +#define CN23XX_MSIX_TABLE_ADDR_START 0x0 +#define CN23XX_MSIX_TABLE_DATA_START 0x8 + +#define CN23XX_MSIX_TABLE_SIZE 0x10 +#define CN23XX_MSIX_TABLE_ENTRIES 0x41 + +#define CN23XX_MSIX_ENTRY_VECTOR_CTL BIT_ULL(32) + +#define CN23XX_MSIX_TABLE_ADDR(idx) \ + (CN23XX_MSIX_TABLE_ADDR_START + ((idx) * CN23XX_MSIX_TABLE_SIZE)) + +#define CN23XX_MSIX_TABLE_DATA(idx) \ + (CN23XX_MSIX_TABLE_DATA_START + ((idx) * CN23XX_MSIX_TABLE_SIZE)) + +/*######################## INTERRUPTS #########################*/ +#define CN23XX_MAC_INT_OFFSET 0x20 +#define CN23XX_PF_INT_OFFSET 0x10 + +/* 1 register (64-bit) for Interrupt Summary */ +#define CN23XX_SLI_INT_SUM64 0x27000 + +/* 4 registers (64-bit) for Interrupt Enable for each Port */ +#define CN23XX_SLI_INT_ENB64 0x27080 + +#define CN23XX_SLI_MAC_PF_INT_SUM64(mac, pf) \ + (CN23XX_SLI_INT_SUM64 + \ + ((mac) * CN23XX_MAC_INT_OFFSET) + \ + ((pf) * CN23XX_PF_INT_OFFSET)) + +#define CN23XX_SLI_MAC_PF_INT_ENB64(mac, pf) \ + (CN23XX_SLI_INT_ENB64 + \ + ((mac) * CN23XX_MAC_INT_OFFSET) + \ + ((pf) * CN23XX_PF_INT_OFFSET)) + +/* 1 register (64-bit) to indicate which Output Queue reached pkt threshold */ +#define CN23XX_SLI_PKT_CNT_INT 0x29130 + +/* 1 register (64-bit) to indicate which Output Queue reached time threshold */ +#define CN23XX_SLI_PKT_TIME_INT 0x29140 + +/*------------------ Interrupt Masks ----------------*/ + +#define CN23XX_INTR_PO_INT BIT_ULL(63) +#define CN23XX_INTR_PI_INT BIT_ULL(62) +#define CN23XX_INTR_MBOX_INT BIT_ULL(61) +#define CN23XX_INTR_RESEND BIT_ULL(60) + +#define CN23XX_INTR_CINT_ENB BIT_ULL(48) +#define CN23XX_INTR_MBOX_ENB BIT(0) + +#define CN23XX_INTR_RML_TIMEOUT_ERR (1) + +#define CN23XX_INTR_MIO_INT BIT(1) + +#define CN23XX_INTR_RESERVED1 (3 << 2) + +#define CN23XX_INTR_PKT_COUNT BIT(4) +#define CN23XX_INTR_PKT_TIME BIT(5) + +#define CN23XX_INTR_RESERVED2 (3 << 6) + +#define CN23XX_INTR_M0UPB0_ERR BIT(8) +#define CN23XX_INTR_M0UPWI_ERR BIT(9) +#define CN23XX_INTR_M0UNB0_ERR BIT(10) +#define CN23XX_INTR_M0UNWI_ERR BIT(11) + +#define CN23XX_INTR_RESERVED3 (0xFFFFFULL << 12) + +#define CN23XX_INTR_DMA0_FORCE BIT_ULL(32) +#define CN23XX_INTR_DMA1_FORCE BIT_ULL(33) + +#define CN23XX_INTR_DMA0_COUNT BIT_ULL(34) +#define CN23XX_INTR_DMA1_COUNT BIT_ULL(35) + +#define CN23XX_INTR_DMA0_TIME BIT_ULL(36) +#define CN23XX_INTR_DMA1_TIME BIT_ULL(37) + +#define CN23XX_INTR_RESERVED4 (0x7FFFFULL << 38) + +#define CN23XX_INTR_VF_MBOX BIT_ULL(57) +#define CN23XX_INTR_DMAVF_ERR BIT_ULL(58) +#define CN23XX_INTR_DMAPF_ERR BIT_ULL(59) + +#define CN23XX_INTR_PKTVF_ERR BIT_ULL(60) +#define CN23XX_INTR_PKTPF_ERR BIT_ULL(61) +#define CN23XX_INTR_PPVF_ERR BIT_ULL(62) +#define CN23XX_INTR_PPPF_ERR BIT_ULL(63) + +#define CN23XX_INTR_DMA0_DATA (CN23XX_INTR_DMA0_TIME) +#define CN23XX_INTR_DMA1_DATA (CN23XX_INTR_DMA1_TIME) + +#define CN23XX_INTR_DMA_DATA \ + (CN23XX_INTR_DMA0_DATA | CN23XX_INTR_DMA1_DATA) + +/* By fault only TIME based */ +#define CN23XX_INTR_PKT_DATA (CN23XX_INTR_PKT_TIME) +/* For both COUNT and TIME based */ +/* #define CN23XX_INTR_PKT_DATA \ + * (CN23XX_INTR_PKT_COUNT | CN23XX_INTR_PKT_TIME) + */ + +/* Sum of interrupts for all PCI-Express Data Interrupts */ +#define CN23XX_INTR_PCIE_DATA \ + (CN23XX_INTR_DMA_DATA | CN23XX_INTR_PKT_DAT) + +/* Sum of interrupts for error events */ +#define CN23XX_INTR_ERR \ + (CN23XX_INTR_M0UPB0_ERR | \ + CN23XX_INTR_M0UPWI_ERR | \ + CN23XX_INTR_M0UNB0_ERR | \ + CN23XX_INTR_M0UNWI_ERR | \ + CN23XX_INTR_DMAVF_ERR | \ + CN23XX_INTR_DMAPF_ERR | \ + CN23XX_INTR_PKTPF_ERR | \ + CN23XX_INTR_PPPF_ERR | \ + CN23XX_INTR_PPVF_ERR) + +/* Programmed Mask for Interrupt Sum */ +#define CN23XX_INTR_MASK \ + (CN23XX_INTR_DMA_DATA | \ + CN23XX_INTR_DMA0_FORCE | \ + CN23XX_INTR_DMA1_FORCE | \ + CN23XX_INTR_MIO_INT | \ + CN23XX_INTR_ERR) + +/* 4 Registers (64 - bit) */ +#define CN23XX_SLI_S2M_PORT_CTL_START 0x23D80 +#define CN23XX_SLI_S2M_PORTX_CTL(port) \ + (CN23XX_SLI_S2M_PORT_CTL_START + (port * 0x10)) + +#define CN23XX_SLI_MAC_NUMBER 0x20050 + +/** PEM(0..3)_BAR1_INDEX(0..15)address is defined as + * addr = (0x00011800C0000100 |port <<24 |idx <<3 ) + * Here, port is PEM(0..3) & idx is INDEX(0..15) + */ +#define CN23XX_PEM_BAR1_INDEX_START 0x00011800C0000100ULL +#define CN23XX_PEM_OFFSET 24 +#define CN23XX_BAR1_INDEX_OFFSET 3 + +#define CN23XX_PEM_BAR1_INDEX_REG(port, idx) \ + (CN23XX_PEM_BAR1_INDEX_START + ((port) << CN23XX_PEM_OFFSET) + \ + ((idx) << CN23XX_BAR1_INDEX_OFFSET)) + +/*############################ DPI #########################*/ + +/* 1 register (64-bit) - provides DMA Enable */ +#define CN23XX_DPI_CTL 0x0001df0000000040ULL + +/* 1 register (64-bit) - Controls the DMA IO Operation */ +#define CN23XX_DPI_DMA_CONTROL 0x0001df0000000048ULL + +/* 1 register (64-bit) - Provides DMA Instr'n Queue Enable */ +#define CN23XX_DPI_REQ_GBL_ENB 0x0001df0000000050ULL + +/* 1 register (64-bit) - DPI_REQ_ERR_RSP + * Indicates which Instr'n Queue received error response from the IO sub-system + */ +#define CN23XX_DPI_REQ_ERR_RSP 0x0001df0000000058ULL + +/* 1 register (64-bit) - DPI_REQ_ERR_RST + * Indicates which Instr'n Queue dropped an Instr'n + */ +#define CN23XX_DPI_REQ_ERR_RST 0x0001df0000000060ULL + +/* 6 register (64-bit) - DPI_DMA_ENG(0..5)_EN + * Provides DMA Engine Queue Enable + */ +#define CN23XX_DPI_DMA_ENG0_ENB 0x0001df0000000080ULL +#define CN23XX_DPI_DMA_ENG_ENB(eng) (CN23XX_DPI_DMA_ENG0_ENB + (eng * 8)) + +/* 8 register (64-bit) - DPI_DMA(0..7)_REQQ_CTL + * Provides control bits for transaction on 8 Queues + */ +#define CN23XX_DPI_DMA_REQQ0_CTL 0x0001df0000000180ULL +#define CN23XX_DPI_DMA_REQQ_CTL(q_no) \ + (CN23XX_DPI_DMA_REQQ0_CTL + (q_no * 8)) + +/* 6 register (64-bit) - DPI_ENG(0..5)_BUF + * Provides DMA Engine FIFO (Queue) Size + */ +#define CN23XX_DPI_DMA_ENG0_BUF 0x0001df0000000880ULL +#define CN23XX_DPI_DMA_ENG_BUF(eng) \ + (CN23XX_DPI_DMA_ENG0_BUF + (eng * 8)) + +/* 4 Registers (64-bit) */ +#define CN23XX_DPI_SLI_PRT_CFG_START 0x0001df0000000900ULL +#define CN23XX_DPI_SLI_PRTX_CFG(port) \ + (CN23XX_DPI_SLI_PRT_CFG_START + (port * 0x8)) + +/* Masks for DPI_DMA_CONTROL Register */ +#define CN23XX_DPI_DMA_COMMIT_MODE BIT_ULL(58) +#define CN23XX_DPI_DMA_PKT_EN BIT_ULL(56) +#define CN23XX_DPI_DMA_ENB (0x0FULL << 48) +/* Set the DMA Control, to update packet count not byte count sent by DMA, + * when we use Interrupt Coalescing (CA mode) + */ +#define CN23XX_DPI_DMA_O_ADD1 BIT(19) +/*selecting 64-bit Byte Swap Mode */ +#define CN23XX_DPI_DMA_O_ES BIT(15) +#define CN23XX_DPI_DMA_O_MODE BIT(14) + +#define CN23XX_DPI_DMA_CTL_MASK \ + (CN23XX_DPI_DMA_COMMIT_MODE | \ + CN23XX_DPI_DMA_PKT_EN | \ + CN23XX_DPI_DMA_O_ES | \ + CN23XX_DPI_DMA_O_MODE) + +/*############################ RST #########################*/ + +#define CN23XX_RST_BOOT 0x0001180006001600ULL +#define CN23XX_RST_SOFT_RST 0x0001180006001680ULL + +#define CN23XX_LMC0_RESET_CTL 0x0001180088000180ULL +#define CN23XX_LMC0_RESET_CTL_DDR3RST_MASK 0x0000000000000001ULL + +#endif diff --git a/drivers/net/ethernet/cavium/liquidio/cn66xx_device.c b/drivers/net/ethernet/cavium/liquidio/cn66xx_device.c index c03d37016a48..e779af88621b 100644 --- a/drivers/net/ethernet/cavium/liquidio/cn66xx_device.c +++ b/drivers/net/ethernet/cavium/liquidio/cn66xx_device.c @@ -338,7 +338,7 @@ void lio_cn6xxx_setup_oq_regs(struct octeon_device *oct, u32 oq_no) octeon_write_csr(oct, CN6XXX_SLI_PKT_CNT_INT_ENB, intr); } -void lio_cn6xxx_enable_io_queues(struct octeon_device *oct) +int lio_cn6xxx_enable_io_queues(struct octeon_device *oct) { u32 mask; @@ -353,6 +353,8 @@ void lio_cn6xxx_enable_io_queues(struct octeon_device *oct) mask = octeon_read_csr(oct, CN6XXX_SLI_PKT_OUT_ENB); mask |= oct->io_qmask.oq; octeon_write_csr(oct, CN6XXX_SLI_PKT_OUT_ENB, mask); + + return 0; } void lio_cn6xxx_disable_io_queues(struct octeon_device *oct) @@ -418,36 +420,6 @@ void lio_cn6xxx_disable_io_queues(struct octeon_device *oct) octeon_write_csr(oct, CN6XXX_SLI_PKT_TIME_INT, d32); } -void lio_cn6xxx_reinit_regs(struct octeon_device *oct) -{ - int i; - - for (i = 0; i < MAX_OCTEON_INSTR_QUEUES(oct); i++) { - if (!(oct->io_qmask.iq & (1ULL << i))) - continue; - oct->fn_list.setup_iq_regs(oct, i); - } - - for (i = 0; i < MAX_OCTEON_OUTPUT_QUEUES(oct); i++) { - if (!(oct->io_qmask.oq & (1ULL << i))) - continue; - oct->fn_list.setup_oq_regs(oct, i); - } - - oct->fn_list.setup_device_regs(oct); - - oct->fn_list.enable_interrupt(oct->chip); - - oct->fn_list.enable_io_queues(oct); - - /* for (i = 0; i < oct->num_oqs; i++) { */ - for (i = 0; i < MAX_OCTEON_OUTPUT_QUEUES(oct); i++) { - if (!(oct->io_qmask.oq & (1ULL << i))) - continue; - writel(oct->droq[i]->max_count, oct->droq[i]->pkts_credit_reg); - } -} - void lio_cn6xxx_bar1_idx_setup(struct octeon_device *oct, u64 core_addr, @@ -507,18 +479,20 @@ lio_cn6xxx_update_read_index(struct octeon_instr_queue *iq) return new_idx; } -void lio_cn6xxx_enable_interrupt(void *chip) +void lio_cn6xxx_enable_interrupt(struct octeon_device *oct, + u8 unused __attribute__((unused))) { - struct octeon_cn6xxx *cn6xxx = (struct octeon_cn6xxx *)chip; + struct octeon_cn6xxx *cn6xxx = (struct octeon_cn6xxx *)oct->chip; u64 mask = cn6xxx->intr_mask64 | CN6XXX_INTR_DMA0_FORCE; /* Enable Interrupt */ writeq(mask, cn6xxx->intr_enb_reg64); } -void lio_cn6xxx_disable_interrupt(void *chip) +void lio_cn6xxx_disable_interrupt(struct octeon_device *oct, + u8 unused __attribute__((unused))) { - struct octeon_cn6xxx *cn6xxx = (struct octeon_cn6xxx *)chip; + struct octeon_cn6xxx *cn6xxx = (struct octeon_cn6xxx *)oct->chip; /* Disable Interrupts */ writeq(0, cn6xxx->intr_enb_reg64); @@ -714,7 +688,6 @@ int lio_setup_cn66xx_octeon_device(struct octeon_device *oct) oct->fn_list.soft_reset = lio_cn6xxx_soft_reset; oct->fn_list.setup_device_regs = lio_cn6xxx_setup_device_regs; - oct->fn_list.reinit_regs = lio_cn6xxx_reinit_regs; oct->fn_list.update_iq_read_idx = lio_cn6xxx_update_read_index; oct->fn_list.bar1_idx_setup = lio_cn6xxx_bar1_idx_setup; diff --git a/drivers/net/ethernet/cavium/liquidio/cn66xx_device.h b/drivers/net/ethernet/cavium/liquidio/cn66xx_device.h index 28c47224221a..a40a91394079 100644 --- a/drivers/net/ethernet/cavium/liquidio/cn66xx_device.h +++ b/drivers/net/ethernet/cavium/liquidio/cn66xx_device.h @@ -80,18 +80,17 @@ void lio_cn6xxx_setup_global_input_regs(struct octeon_device *oct); void lio_cn6xxx_setup_global_output_regs(struct octeon_device *oct); void lio_cn6xxx_setup_iq_regs(struct octeon_device *oct, u32 iq_no); void lio_cn6xxx_setup_oq_regs(struct octeon_device *oct, u32 oq_no); -void lio_cn6xxx_enable_io_queues(struct octeon_device *oct); +int lio_cn6xxx_enable_io_queues(struct octeon_device *oct); void lio_cn6xxx_disable_io_queues(struct octeon_device *oct); irqreturn_t lio_cn6xxx_process_interrupt_regs(void *dev); -void lio_cn6xxx_reinit_regs(struct octeon_device *oct); void lio_cn6xxx_bar1_idx_setup(struct octeon_device *oct, u64 core_addr, u32 idx, int valid); void lio_cn6xxx_bar1_idx_write(struct octeon_device *oct, u32 idx, u32 mask); u32 lio_cn6xxx_bar1_idx_read(struct octeon_device *oct, u32 idx); u32 lio_cn6xxx_update_read_index(struct octeon_instr_queue *iq); -void lio_cn6xxx_enable_interrupt(void *chip); -void lio_cn6xxx_disable_interrupt(void *chip); +void lio_cn6xxx_enable_interrupt(struct octeon_device *oct, u8 unused); +void lio_cn6xxx_disable_interrupt(struct octeon_device *oct, u8 unused); void cn6xxx_get_pcie_qlmport(struct octeon_device *oct); void lio_cn6xxx_setup_reg_address(struct octeon_device *oct, void *chip, struct octeon_reg_list *reg_list); diff --git a/drivers/net/ethernet/cavium/liquidio/cn68xx_device.c b/drivers/net/ethernet/cavium/liquidio/cn68xx_device.c index 29755bc68f12..dbf3566ead53 100644 --- a/drivers/net/ethernet/cavium/liquidio/cn68xx_device.c +++ b/drivers/net/ethernet/cavium/liquidio/cn68xx_device.c @@ -148,7 +148,6 @@ int lio_setup_cn68xx_octeon_device(struct octeon_device *oct) oct->fn_list.process_interrupt_regs = lio_cn6xxx_process_interrupt_regs; oct->fn_list.soft_reset = lio_cn68xx_soft_reset; oct->fn_list.setup_device_regs = lio_cn68xx_setup_device_regs; - oct->fn_list.reinit_regs = lio_cn6xxx_reinit_regs; oct->fn_list.update_iq_read_idx = lio_cn6xxx_update_read_index; oct->fn_list.bar1_idx_setup = lio_cn6xxx_bar1_idx_setup; diff --git a/drivers/net/ethernet/cavium/liquidio/lio_core.c b/drivers/net/ethernet/cavium/liquidio/lio_core.c new file mode 100644 index 000000000000..201eddb3013a --- /dev/null +++ b/drivers/net/ethernet/cavium/liquidio/lio_core.c @@ -0,0 +1,266 @@ +/********************************************************************** +* Author: Cavium, Inc. +* +* Contact: support@cavium.com +* Please include "LiquidIO" in the subject. +* +* Copyright (c) 2003-2015 Cavium, Inc. +* +* This file is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License, Version 2, as +* published by the Free Software Foundation. +* +* This file is distributed in the hope that it will be useful, but +* AS-IS and WITHOUT ANY WARRANTY; without even the implied warranty +* of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, TITLE, or +* NONINFRINGEMENT. See the GNU General Public License for more +* details. +* +* This file may also be available under a different license from Cavium. +* Contact Cavium, Inc. for more information +**********************************************************************/ +#include <linux/pci.h> +#include <linux/if_vlan.h> +#include "liquidio_common.h" +#include "octeon_droq.h" +#include "octeon_iq.h" +#include "response_manager.h" +#include "octeon_device.h" +#include "octeon_nic.h" +#include "octeon_main.h" +#include "octeon_network.h" + +int liquidio_set_feature(struct net_device *netdev, int cmd, u16 param1) +{ + struct lio *lio = GET_LIO(netdev); + struct octeon_device *oct = lio->oct_dev; + struct octnic_ctrl_pkt nctrl; + int ret = 0; + + memset(&nctrl, 0, sizeof(struct octnic_ctrl_pkt)); + + nctrl.ncmd.u64 = 0; + nctrl.ncmd.s.cmd = cmd; + nctrl.ncmd.s.param1 = param1; + nctrl.iq_no = lio->linfo.txpciq[0].s.q_no; + nctrl.wait_time = 100; + nctrl.netpndev = (u64)netdev; + nctrl.cb_fn = liquidio_link_ctrl_cmd_completion; + + ret = octnet_send_nic_ctrl_pkt(lio->oct_dev, &nctrl); + if (ret < 0) { + dev_err(&oct->pci_dev->dev, "Feature change failed in core (ret: 0x%x)\n", + ret); + } + return ret; +} + +void octeon_report_tx_completion_to_bql(void *txq, unsigned int pkts_compl, + unsigned int bytes_compl) +{ + struct netdev_queue *netdev_queue = txq; + + netdev_tx_completed_queue(netdev_queue, pkts_compl, bytes_compl); +} + +void octeon_update_tx_completion_counters(void *buf, int reqtype, + unsigned int *pkts_compl, + unsigned int *bytes_compl) +{ + struct octnet_buf_free_info *finfo; + struct sk_buff *skb = NULL; + struct octeon_soft_command *sc; + + switch (reqtype) { + case REQTYPE_NORESP_NET: + case REQTYPE_NORESP_NET_SG: + finfo = buf; + skb = finfo->skb; + break; + + case REQTYPE_RESP_NET_SG: + case REQTYPE_RESP_NET: + sc = buf; + skb = sc->callback_arg; + break; + + default: + return; + } + + (*pkts_compl)++; +/*TODO, Use some other pound define to suggest + * the fact that iqs are not tied to netdevs + * and can take traffic from different netdevs + * hence bql reporting is done per packet + * than in bulk. Usage of NO_NAPI in txq completion is + * a little confusing + */ + *bytes_compl += skb->len; +} + +void octeon_report_sent_bytes_to_bql(void *buf, int reqtype) +{ + struct octnet_buf_free_info *finfo; + struct sk_buff *skb; + struct octeon_soft_command *sc; + struct netdev_queue *txq; + + switch (reqtype) { + case REQTYPE_NORESP_NET: + case REQTYPE_NORESP_NET_SG: + finfo = buf; + skb = finfo->skb; + break; + + case REQTYPE_RESP_NET_SG: + case REQTYPE_RESP_NET: + sc = buf; + skb = sc->callback_arg; + break; + + default: + return; + } + + txq = netdev_get_tx_queue(skb->dev, skb_get_queue_mapping(skb)); + netdev_tx_sent_queue(txq, skb->len); +} + +void liquidio_link_ctrl_cmd_completion(void *nctrl_ptr) +{ + struct octnic_ctrl_pkt *nctrl = (struct octnic_ctrl_pkt *)nctrl_ptr; + struct net_device *netdev = (struct net_device *)nctrl->netpndev; + struct lio *lio = GET_LIO(netdev); + struct octeon_device *oct = lio->oct_dev; + u8 *mac; + + switch (nctrl->ncmd.s.cmd) { + case OCTNET_CMD_CHANGE_DEVFLAGS: + case OCTNET_CMD_SET_MULTI_LIST: + break; + + case OCTNET_CMD_CHANGE_MACADDR: + mac = ((u8 *)&nctrl->udd[0]) + 2; + netif_info(lio, probe, lio->netdev, + "MACAddr changed to %2.2x:%2.2x:%2.2x:%2.2x:%2.2x:%2.2x\n", + mac[0], mac[1], + mac[2], mac[3], + mac[4], mac[5]); + break; + + case OCTNET_CMD_CHANGE_MTU: + /* If command is successful, change the MTU. */ + netif_info(lio, probe, lio->netdev, "MTU Changed from %d to %d\n", + netdev->mtu, nctrl->ncmd.s.param1); + dev_info(&oct->pci_dev->dev, "%s MTU Changed from %d to %d\n", + netdev->name, netdev->mtu, + nctrl->ncmd.s.param1); + netdev->mtu = nctrl->ncmd.s.param1; + queue_delayed_work(lio->link_status_wq.wq, + &lio->link_status_wq.wk.work, 0); + break; + + case OCTNET_CMD_GPIO_ACCESS: + netif_info(lio, probe, lio->netdev, "LED Flashing visual identification\n"); + + break; + + case OCTNET_CMD_ID_ACTIVE: + netif_info(lio, probe, lio->netdev, "LED Flashing visual identification\n"); + + break; + + case OCTNET_CMD_LRO_ENABLE: + dev_info(&oct->pci_dev->dev, "%s LRO Enabled\n", netdev->name); + break; + + case OCTNET_CMD_LRO_DISABLE: + dev_info(&oct->pci_dev->dev, "%s LRO Disabled\n", + netdev->name); + break; + + case OCTNET_CMD_VERBOSE_ENABLE: + dev_info(&oct->pci_dev->dev, "%s Firmware debug enabled\n", + netdev->name); + break; + + case OCTNET_CMD_VERBOSE_DISABLE: + dev_info(&oct->pci_dev->dev, "%s Firmware debug disabled\n", + netdev->name); + break; + + case OCTNET_CMD_ENABLE_VLAN_FILTER: + dev_info(&oct->pci_dev->dev, "%s VLAN filter enabled\n", + netdev->name); + break; + + case OCTNET_CMD_ADD_VLAN_FILTER: + dev_info(&oct->pci_dev->dev, "%s VLAN filter %d added\n", + netdev->name, nctrl->ncmd.s.param1); + break; + + case OCTNET_CMD_DEL_VLAN_FILTER: + dev_info(&oct->pci_dev->dev, "%s VLAN filter %d removed\n", + netdev->name, nctrl->ncmd.s.param1); + break; + + case OCTNET_CMD_SET_SETTINGS: + dev_info(&oct->pci_dev->dev, "%s settings changed\n", + netdev->name); + + break; + + /* Case to handle "OCTNET_CMD_TNL_RX_CSUM_CTL" + * Command passed by NIC driver + */ + case OCTNET_CMD_TNL_RX_CSUM_CTL: + if (nctrl->ncmd.s.param1 == OCTNET_CMD_RXCSUM_ENABLE) { + netif_info(lio, probe, lio->netdev, + "RX Checksum Offload Enabled\n"); + } else if (nctrl->ncmd.s.param1 == + OCTNET_CMD_RXCSUM_DISABLE) { + netif_info(lio, probe, lio->netdev, + "RX Checksum Offload Disabled\n"); + } + break; + + /* Case to handle "OCTNET_CMD_TNL_TX_CSUM_CTL" + * Command passed by NIC driver + */ + case OCTNET_CMD_TNL_TX_CSUM_CTL: + if (nctrl->ncmd.s.param1 == OCTNET_CMD_TXCSUM_ENABLE) { + netif_info(lio, probe, lio->netdev, + "TX Checksum Offload Enabled\n"); + } else if (nctrl->ncmd.s.param1 == + OCTNET_CMD_TXCSUM_DISABLE) { + netif_info(lio, probe, lio->netdev, + "TX Checksum Offload Disabled\n"); + } + break; + + /* Case to handle "OCTNET_CMD_VXLAN_PORT_CONFIG" + * Command passed by NIC driver + */ + case OCTNET_CMD_VXLAN_PORT_CONFIG: + if (nctrl->ncmd.s.more == OCTNET_CMD_VXLAN_PORT_ADD) { + netif_info(lio, probe, lio->netdev, + "VxLAN Destination UDP PORT:%d ADDED\n", + nctrl->ncmd.s.param1); + } else if (nctrl->ncmd.s.more == + OCTNET_CMD_VXLAN_PORT_DEL) { + netif_info(lio, probe, lio->netdev, + "VxLAN Destination UDP PORT:%d DELETED\n", + nctrl->ncmd.s.param1); + } + break; + + case OCTNET_CMD_SET_FLOW_CTL: + netif_info(lio, probe, lio->netdev, "Set RX/TX flow control parameters\n"); + break; + + default: + dev_err(&oct->pci_dev->dev, "%s Unknown cmd %d\n", __func__, + nctrl->ncmd.s.cmd); + } +} diff --git a/drivers/net/ethernet/cavium/liquidio/lio_ethtool.c b/drivers/net/ethernet/cavium/liquidio/lio_ethtool.c index 289eb8907922..f163e0abbeb2 100644 --- a/drivers/net/ethernet/cavium/liquidio/lio_ethtool.c +++ b/drivers/net/ethernet/cavium/liquidio/lio_ethtool.c @@ -32,6 +32,7 @@ #include "octeon_network.h" #include "cn66xx_regs.h" #include "cn66xx_device.h" +#include "cn23xx_pf_device.h" static int octnet_get_link_stats(struct net_device *netdev); @@ -75,6 +76,7 @@ enum { #define ARRAY_LENGTH(a) (sizeof(a) / sizeof((a)[0])) #define OCT_ETHTOOL_REGDUMP_LEN 4096 +#define OCT_ETHTOOL_REGDUMP_LEN_23XX (4096 * 11) #define OCT_ETHTOOL_REGSVER 1 /* statistics of PF */ @@ -188,6 +190,10 @@ static const char oct_droq_stats_strings[][ETH_GSTRING_LEN] = { "buffer_alloc_failure", }; +/* LiquidIO driver private flags */ +static const char oct_priv_flags_strings[][ETH_GSTRING_LEN] = { +}; + #define OCTNIC_NCMD_AUTONEG_ON 0x1 #define OCTNIC_NCMD_PHY_ON 0x2 @@ -259,6 +265,13 @@ lio_ethtool_get_channels(struct net_device *dev, max_tx = CFG_GET_IQ_MAX_Q(conf6x); rx_count = CFG_GET_NUM_RXQS_NIC_IF(conf6x, lio->ifidx); tx_count = CFG_GET_NUM_TXQS_NIC_IF(conf6x, lio->ifidx); + } else if (OCTEON_CN23XX_PF(oct)) { + struct octeon_config *conf23 = CHIP_FIELD(oct, cn23xx_pf, conf); + + max_rx = CFG_GET_OQ_MAX_Q(conf23); + max_tx = CFG_GET_IQ_MAX_Q(conf23); + rx_count = CFG_GET_NUM_RXQS_NIC_IF(conf23, lio->ifidx); + tx_count = CFG_GET_NUM_TXQS_NIC_IF(conf23, lio->ifidx); } channel->max_rx = max_rx; @@ -290,18 +303,16 @@ lio_get_eeprom(struct net_device *netdev, struct ethtool_eeprom *eeprom, struct lio *lio = GET_LIO(netdev); struct octeon_device *oct_dev = lio->oct_dev; struct octeon_board_info *board_info; - int len; - if (eeprom->offset != 0) + if (eeprom->offset) return -EINVAL; eeprom->magic = oct_dev->pci_dev->vendor; board_info = (struct octeon_board_info *)(&oct_dev->boardinfo); - len = - sprintf((char *)bytes, - "boardname:%s serialnum:%s maj:%lld min:%lld\n", - board_info->name, board_info->serial_number, - board_info->major, board_info->minor); + sprintf((char *)bytes, + "boardname:%s serialnum:%s maj:%lld min:%lld\n", + board_info->name, board_info->serial_number, + board_info->major, board_info->minor); return 0; } @@ -333,6 +344,32 @@ static int octnet_gpio_access(struct net_device *netdev, int addr, int val) return 0; } +static int octnet_id_active(struct net_device *netdev, int val) +{ + struct lio *lio = GET_LIO(netdev); + struct octeon_device *oct = lio->oct_dev; + struct octnic_ctrl_pkt nctrl; + int ret = 0; + + memset(&nctrl, 0, sizeof(struct octnic_ctrl_pkt)); + + nctrl.ncmd.u64 = 0; + nctrl.ncmd.s.cmd = OCTNET_CMD_ID_ACTIVE; + nctrl.ncmd.s.param1 = val; + nctrl.iq_no = lio->linfo.txpciq[0].s.q_no; + nctrl.wait_time = 100; + nctrl.netpndev = (u64)netdev; + nctrl.cb_fn = liquidio_link_ctrl_cmd_completion; + + ret = octnet_send_nic_ctrl_pkt(lio->oct_dev, &nctrl); + if (ret < 0) { + dev_err(&oct->pci_dev->dev, "Failed to configure gpio value\n"); + return -EINVAL; + } + + return 0; +} + /* Callback for when mdio command response arrives */ static void octnet_mdio_resp_callback(struct octeon_device *oct, @@ -406,7 +443,7 @@ octnet_mdio45_access(struct lio *lio, int op, int loc, int *value) dev_err(&oct_dev->pci_dev->dev, "octnet_mdio45_access instruction failed status: %x\n", retval); - retval = -EBUSY; + retval = -EBUSY; } else { /* Sleep on a wait queue till the cond flag indicates that the * response arrived @@ -476,6 +513,11 @@ static int lio_set_phys_id(struct net_device *netdev, &value); if (ret) return ret; + } else if (oct->chip_id == OCTEON_CN23XX_PF_VID) { + octnet_id_active(netdev, LED_IDENTIFICATION_ON); + + /* returns 0 since updates are asynchronous */ + return 0; } else { return -EINVAL; } @@ -521,7 +563,10 @@ static int lio_set_phys_id(struct net_device *netdev, &lio->phy_beacon_val); if (ret) return ret; + } else if (oct->chip_id == OCTEON_CN23XX_PF_VID) { + octnet_id_active(netdev, LED_IDENTIFICATION_OFF); + return 0; } else { return -EINVAL; } @@ -550,6 +595,13 @@ lio_ethtool_get_ringparam(struct net_device *netdev, rx_max_pending = CN6XXX_MAX_OQ_DESCRIPTORS; rx_pending = CFG_GET_NUM_RX_DESCS_NIC_IF(conf6x, lio->ifidx); tx_pending = CFG_GET_NUM_TX_DESCS_NIC_IF(conf6x, lio->ifidx); + } else if (OCTEON_CN23XX_PF(oct)) { + struct octeon_config *conf23 = CHIP_FIELD(oct, cn23xx_pf, conf); + + tx_max_pending = CN23XX_MAX_IQ_DESCRIPTORS; + rx_max_pending = CN23XX_MAX_OQ_DESCRIPTORS; + rx_pending = CFG_GET_NUM_RX_DESCS_NIC_IF(conf23, lio->ifidx); + tx_pending = CFG_GET_NUM_TX_DESCS_NIC_IF(conf23, lio->ifidx); } if (lio->mtu > OCTNET_DEFAULT_FRM_SIZE - OCTNET_FRM_HEADER_SIZE) { @@ -610,6 +662,69 @@ lio_get_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause) pause->rx_pause = oct->rx_pause; } +static int +lio_set_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause) +{ + /* Notes: Not supporting any auto negotiation in these + * drivers. + */ + struct lio *lio = GET_LIO(netdev); + struct octeon_device *oct = lio->oct_dev; + struct octnic_ctrl_pkt nctrl; + struct oct_link_info *linfo = &lio->linfo; + + int ret = 0; + + if (oct->chip_id != OCTEON_CN23XX_PF_VID) + return -EINVAL; + + if (linfo->link.s.duplex == 0) { + /*no flow control for half duplex*/ + if (pause->rx_pause || pause->tx_pause) + return -EINVAL; + } + + /*do not support autoneg of link flow control*/ + if (pause->autoneg == AUTONEG_ENABLE) + return -EINVAL; + + memset(&nctrl, 0, sizeof(struct octnic_ctrl_pkt)); + + nctrl.ncmd.u64 = 0; + nctrl.ncmd.s.cmd = OCTNET_CMD_SET_FLOW_CTL; + nctrl.iq_no = lio->linfo.txpciq[0].s.q_no; + nctrl.wait_time = 100; + nctrl.netpndev = (u64)netdev; + nctrl.cb_fn = liquidio_link_ctrl_cmd_completion; + + if (pause->rx_pause) { + /*enable rx pause*/ + nctrl.ncmd.s.param1 = 1; + } else { + /*disable rx pause*/ + nctrl.ncmd.s.param1 = 0; + } + + if (pause->tx_pause) { + /*enable tx pause*/ + nctrl.ncmd.s.param2 = 1; + } else { + /*disable tx pause*/ + nctrl.ncmd.s.param2 = 0; + } + + ret = octnet_send_nic_ctrl_pkt(lio->oct_dev, &nctrl); + if (ret < 0) { + dev_err(&oct->pci_dev->dev, "Failed to set pause parameter\n"); + return -EINVAL; + } + + oct->rx_pause = pause->rx_pause; + oct->tx_pause = pause->tx_pause; + + return 0; +} + static void lio_get_ethtool_stats(struct net_device *netdev, struct ethtool_stats *stats __attribute__((unused)), @@ -877,6 +992,27 @@ lio_get_ethtool_stats(struct net_device *netdev, } } +static void lio_get_priv_flags_strings(struct lio *lio, u8 *data) +{ + struct octeon_device *oct_dev = lio->oct_dev; + int i; + + switch (oct_dev->chip_id) { + case OCTEON_CN23XX_PF_VID: + for (i = 0; i < ARRAY_SIZE(oct_priv_flags_strings); i++) { + sprintf(data, "%s", oct_priv_flags_strings[i]); + data += ETH_GSTRING_LEN; + } + break; + case OCTEON_CN68XX: + case OCTEON_CN66XX: + break; + default: + netif_info(lio, drv, lio->netdev, "Unknown Chip !!\n"); + break; + } +} + static void lio_get_strings(struct net_device *netdev, u32 stringset, u8 *data) { struct lio *lio = GET_LIO(netdev); @@ -916,12 +1052,31 @@ static void lio_get_strings(struct net_device *netdev, u32 stringset, u8 *data) } break; + case ETH_SS_PRIV_FLAGS: + lio_get_priv_flags_strings(lio, data); + break; default: netif_info(lio, drv, lio->netdev, "Unknown Stringset !!\n"); break; } } +static int lio_get_priv_flags_ss_count(struct lio *lio) +{ + struct octeon_device *oct_dev = lio->oct_dev; + + switch (oct_dev->chip_id) { + case OCTEON_CN23XX_PF_VID: + return ARRAY_SIZE(oct_priv_flags_strings); + case OCTEON_CN68XX: + case OCTEON_CN66XX: + return -EOPNOTSUPP; + default: + netif_info(lio, drv, lio->netdev, "Unknown Chip !!\n"); + return -EOPNOTSUPP; + } +} + static int lio_get_sset_count(struct net_device *netdev, int sset) { struct lio *lio = GET_LIO(netdev); @@ -932,6 +1087,8 @@ static int lio_get_sset_count(struct net_device *netdev, int sset) return (ARRAY_SIZE(oct_stats_strings) + ARRAY_SIZE(oct_iq_stats_strings) * oct_dev->num_iqs + ARRAY_SIZE(oct_droq_stats_strings) * oct_dev->num_oqs); + case ETH_SS_PRIV_FLAGS: + return lio_get_priv_flags_ss_count(lio); default: return -EOPNOTSUPP; } @@ -948,6 +1105,16 @@ static int lio_get_intr_coalesce(struct net_device *netdev, intrmod_cfg = &oct->intrmod; switch (oct->chip_id) { + case OCTEON_CN23XX_PF_VID: + if (!intrmod_cfg->rx_enable) { + intr_coal->rx_coalesce_usecs = intrmod_cfg->rx_usecs; + intr_coal->rx_max_coalesced_frames = + intrmod_cfg->rx_frames; + } + if (!intrmod_cfg->tx_enable) + intr_coal->tx_max_coalesced_frames = + intrmod_cfg->tx_frames; + break; case OCTEON_CN68XX: case OCTEON_CN66XX: { struct octeon_cn6xxx *cn6xxx = @@ -983,7 +1150,15 @@ static int lio_get_intr_coalesce(struct net_device *netdev, intr_coal->rx_coalesce_usecs_low = intrmod_cfg->rx_mintmr_trigger; intr_coal->rx_max_coalesced_frames_low = - intrmod_cfg->rx_mincnt_trigger; + intrmod_cfg->rx_mincnt_trigger; + } + if (OCTEON_CN23XX_PF(oct) && + (intrmod_cfg->tx_enable)) { + intr_coal->use_adaptive_tx_coalesce = intrmod_cfg->tx_enable; + intr_coal->tx_max_coalesced_frames_high = + intrmod_cfg->tx_maxcnt_trigger; + intr_coal->tx_max_coalesced_frames_low = + intrmod_cfg->tx_mincnt_trigger; } return 0; } @@ -1060,11 +1235,11 @@ static void octnet_nic_stats_callback(struct octeon_device *oct_dev, u32 status, void *ptr) { - struct octeon_soft_command *sc = (struct octeon_soft_command *)ptr; - struct oct_nic_stats_resp *resp = (struct oct_nic_stats_resp *) - sc->virtrptr; - struct oct_nic_stats_ctrl *ctrl = (struct oct_nic_stats_ctrl *) - sc->ctxptr; + struct octeon_soft_command *sc = (struct octeon_soft_command *)ptr; + struct oct_nic_stats_resp *resp = + (struct oct_nic_stats_resp *)sc->virtrptr; + struct oct_nic_stats_ctrl *ctrl = + (struct oct_nic_stats_ctrl *)sc->ctxptr; struct nic_rx_stats *rsp_rstats = &resp->stats.fromwire; struct nic_tx_stats *rsp_tstats = &resp->stats.fromhost; @@ -1314,14 +1489,35 @@ oct_cfg_rx_intrcnt(struct lio *lio, struct ethtool_coalesce *intr_coal) CFG_SET_OQ_INTR_PKT(cn6xxx->conf, rx_max_coalesced_frames); break; } + case OCTEON_CN23XX_PF_VID: { + int q_no; + + if (!intr_coal->rx_max_coalesced_frames) + rx_max_coalesced_frames = oct->intrmod.rx_frames; + else + rx_max_coalesced_frames = + intr_coal->rx_max_coalesced_frames; + for (q_no = 0; q_no < oct->num_oqs; q_no++) { + q_no += oct->sriov_info.pf_srn; + octeon_write_csr64( + oct, CN23XX_SLI_OQ_PKT_INT_LEVELS(q_no), + (octeon_read_csr64( + oct, CN23XX_SLI_OQ_PKT_INT_LEVELS(q_no)) & + (0x3fffff00000000UL)) | + rx_max_coalesced_frames); + /*consider setting resend bit*/ + } + oct->intrmod.rx_frames = rx_max_coalesced_frames; + break; + } default: return -EINVAL; } return 0; } -static int oct_cfg_rx_intrtime(struct lio *lio, struct ethtool_coalesce - *intr_coal) +static int oct_cfg_rx_intrtime(struct lio *lio, + struct ethtool_coalesce *intr_coal) { struct octeon_device *oct = lio->oct_dev; u32 time_threshold, rx_coalesce_usecs; @@ -1346,6 +1542,27 @@ static int oct_cfg_rx_intrtime(struct lio *lio, struct ethtool_coalesce CFG_SET_OQ_INTR_TIME(cn6xxx->conf, rx_coalesce_usecs); break; } + case OCTEON_CN23XX_PF_VID: { + u64 time_threshold; + int q_no; + + if (!intr_coal->rx_coalesce_usecs) + rx_coalesce_usecs = oct->intrmod.rx_usecs; + else + rx_coalesce_usecs = intr_coal->rx_coalesce_usecs; + time_threshold = + cn23xx_pf_get_oq_ticks(oct, (u32)rx_coalesce_usecs); + for (q_no = 0; q_no < oct->num_oqs; q_no++) { + q_no += oct->sriov_info.pf_srn; + octeon_write_csr64(oct, + CN23XX_SLI_OQ_PKT_INT_LEVELS(q_no), + (oct->intrmod.rx_frames | + (time_threshold << 32))); + /*consider writing to resend bit here*/ + } + oct->intrmod.rx_usecs = rx_coalesce_usecs; + break; + } default: return -EINVAL; } @@ -1358,12 +1575,37 @@ oct_cfg_tx_intrcnt(struct lio *lio, struct ethtool_coalesce *intr_coal __attribute__((unused))) { struct octeon_device *oct = lio->oct_dev; + u32 iq_intr_pkt; + void __iomem *inst_cnt_reg; + u64 val; /* Config Cnt based interrupt values */ switch (oct->chip_id) { case OCTEON_CN68XX: case OCTEON_CN66XX: break; + case OCTEON_CN23XX_PF_VID: { + int q_no; + + if (!intr_coal->tx_max_coalesced_frames) + iq_intr_pkt = CN23XX_DEF_IQ_INTR_THRESHOLD & + CN23XX_PKT_IN_DONE_WMARK_MASK; + else + iq_intr_pkt = intr_coal->tx_max_coalesced_frames & + CN23XX_PKT_IN_DONE_WMARK_MASK; + for (q_no = 0; q_no < oct->num_iqs; q_no++) { + inst_cnt_reg = (oct->instr_queue[q_no])->inst_cnt_reg; + val = readq(inst_cnt_reg); + /*clear wmark and count.dont want to write count back*/ + val = (val & 0xFFFF000000000000ULL) | + ((u64)iq_intr_pkt + << CN23XX_PKT_IN_DONE_WMARK_BIT_POS); + writeq(val, inst_cnt_reg); + /*consider setting resend bit*/ + } + oct->intrmod.tx_frames = iq_intr_pkt; + break; + } default: return -EINVAL; } @@ -1399,6 +1641,8 @@ static int lio_set_intr_coalesce(struct net_device *netdev, return -EINVAL; } break; + case OCTEON_CN23XX_PF_VID: + break; default: return -EINVAL; } @@ -1541,9 +1785,237 @@ static int lio_nway_reset(struct net_device *netdev) } /* Return register dump len. */ -static int lio_get_regs_len(struct net_device *dev __attribute__((unused))) +static int lio_get_regs_len(struct net_device *dev) { - return OCT_ETHTOOL_REGDUMP_LEN; + struct lio *lio = GET_LIO(dev); + struct octeon_device *oct = lio->oct_dev; + + switch (oct->chip_id) { + case OCTEON_CN23XX_PF_VID: + return OCT_ETHTOOL_REGDUMP_LEN_23XX; + default: + return OCT_ETHTOOL_REGDUMP_LEN; + } +} + +static int cn23xx_read_csr_reg(char *s, struct octeon_device *oct) +{ + u32 reg; + u8 pf_num = oct->pf_num; + int len = 0; + int i; + + /* PCI Window Registers */ + + len += sprintf(s + len, "\n\t Octeon CSR Registers\n\n"); + + /*0x29030 or 0x29040*/ + reg = CN23XX_SLI_PKT_MAC_RINFO64(oct->pcie_port, oct->pf_num); + len += sprintf(s + len, + "\n[%08x] (SLI_PKT_MAC%d_PF%d_RINFO): %016llx\n", + reg, oct->pcie_port, oct->pf_num, + (u64)octeon_read_csr64(oct, reg)); + + /*0x27080 or 0x27090*/ + reg = CN23XX_SLI_MAC_PF_INT_ENB64(oct->pcie_port, oct->pf_num); + len += + sprintf(s + len, "\n[%08x] (SLI_MAC%d_PF%d_INT_ENB): %016llx\n", + reg, oct->pcie_port, oct->pf_num, + (u64)octeon_read_csr64(oct, reg)); + + /*0x27000 or 0x27010*/ + reg = CN23XX_SLI_MAC_PF_INT_SUM64(oct->pcie_port, oct->pf_num); + len += + sprintf(s + len, "\n[%08x] (SLI_MAC%d_PF%d_INT_SUM): %016llx\n", + reg, oct->pcie_port, oct->pf_num, + (u64)octeon_read_csr64(oct, reg)); + + /*0x29120*/ + reg = 0x29120; + len += sprintf(s + len, "\n[%08x] (SLI_PKT_MEM_CTL): %016llx\n", reg, + (u64)octeon_read_csr64(oct, reg)); + + /*0x27300*/ + reg = 0x27300 + oct->pcie_port * CN23XX_MAC_INT_OFFSET + + (oct->pf_num) * CN23XX_PF_INT_OFFSET; + len += sprintf( + s + len, "\n[%08x] (SLI_MAC%d_PF%d_PKT_VF_INT): %016llx\n", reg, + oct->pcie_port, oct->pf_num, (u64)octeon_read_csr64(oct, reg)); + + /*0x27200*/ + reg = 0x27200 + oct->pcie_port * CN23XX_MAC_INT_OFFSET + + (oct->pf_num) * CN23XX_PF_INT_OFFSET; + len += sprintf(s + len, + "\n[%08x] (SLI_MAC%d_PF%d_PP_VF_INT): %016llx\n", + reg, oct->pcie_port, oct->pf_num, + (u64)octeon_read_csr64(oct, reg)); + + /*29130*/ + reg = CN23XX_SLI_PKT_CNT_INT; + len += sprintf(s + len, "\n[%08x] (SLI_PKT_CNT_INT): %016llx\n", reg, + (u64)octeon_read_csr64(oct, reg)); + + /*0x29140*/ + reg = CN23XX_SLI_PKT_TIME_INT; + len += sprintf(s + len, "\n[%08x] (SLI_PKT_TIME_INT): %016llx\n", reg, + (u64)octeon_read_csr64(oct, reg)); + + /*0x29160*/ + reg = 0x29160; + len += sprintf(s + len, "\n[%08x] (SLI_PKT_INT): %016llx\n", reg, + (u64)octeon_read_csr64(oct, reg)); + + /*0x29180*/ + reg = CN23XX_SLI_OQ_WMARK; + len += sprintf(s + len, "\n[%08x] (SLI_PKT_OUTPUT_WMARK): %016llx\n", + reg, (u64)octeon_read_csr64(oct, reg)); + + /*0x291E0*/ + reg = CN23XX_SLI_PKT_IOQ_RING_RST; + len += sprintf(s + len, "\n[%08x] (SLI_PKT_RING_RST): %016llx\n", reg, + (u64)octeon_read_csr64(oct, reg)); + + /*0x29210*/ + reg = CN23XX_SLI_GBL_CONTROL; + len += sprintf(s + len, + "\n[%08x] (SLI_PKT_GBL_CONTROL): %016llx\n", reg, + (u64)octeon_read_csr64(oct, reg)); + + /*0x29220*/ + reg = 0x29220; + len += sprintf(s + len, "\n[%08x] (SLI_PKT_BIST_STATUS): %016llx\n", + reg, (u64)octeon_read_csr64(oct, reg)); + + /*PF only*/ + if (pf_num == 0) { + /*0x29260*/ + reg = CN23XX_SLI_OUT_BP_EN_W1S; + len += sprintf(s + len, + "\n[%08x] (SLI_PKT_OUT_BP_EN_W1S): %016llx\n", + reg, (u64)octeon_read_csr64(oct, reg)); + } else if (pf_num == 1) { + /*0x29270*/ + reg = CN23XX_SLI_OUT_BP_EN2_W1S; + len += sprintf(s + len, + "\n[%08x] (SLI_PKT_OUT_BP_EN2_W1S): %016llx\n", + reg, (u64)octeon_read_csr64(oct, reg)); + } + + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + reg = CN23XX_SLI_OQ_BUFF_INFO_SIZE(i); + len += + sprintf(s + len, "\n[%08x] (SLI_PKT%d_OUT_SIZE): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10040*/ + for (i = 0; i < CN23XX_MAX_INPUT_QUEUES; i++) { + reg = CN23XX_SLI_IQ_INSTR_COUNT64(i); + len += sprintf(s + len, + "\n[%08x] (SLI_PKT_IN_DONE%d_CNTS): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10080*/ + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + reg = CN23XX_SLI_OQ_PKTS_CREDIT(i); + len += sprintf(s + len, + "\n[%08x] (SLI_PKT%d_SLIST_BAOFF_DBELL): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10090*/ + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + reg = CN23XX_SLI_OQ_SIZE(i); + len += sprintf( + s + len, "\n[%08x] (SLI_PKT%d_SLIST_FIFO_RSIZE): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10050*/ + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + reg = CN23XX_SLI_OQ_PKT_CONTROL(i); + len += sprintf( + s + len, + "\n[%08x] (SLI_PKT%d__OUTPUT_CONTROL): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10070*/ + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + reg = CN23XX_SLI_OQ_BASE_ADDR64(i); + len += sprintf(s + len, + "\n[%08x] (SLI_PKT%d_SLIST_BADDR): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x100a0*/ + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + reg = CN23XX_SLI_OQ_PKT_INT_LEVELS(i); + len += sprintf(s + len, + "\n[%08x] (SLI_PKT%d_INT_LEVELS): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x100b0*/ + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + reg = CN23XX_SLI_OQ_PKTS_SENT(i); + len += sprintf(s + len, "\n[%08x] (SLI_PKT%d_CNTS): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x100c0*/ + for (i = 0; i < CN23XX_MAX_OUTPUT_QUEUES; i++) { + reg = 0x100c0 + i * CN23XX_OQ_OFFSET; + len += sprintf(s + len, + "\n[%08x] (SLI_PKT%d_ERROR_INFO): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + + /*0x10000*/ + for (i = 0; i < CN23XX_MAX_INPUT_QUEUES; i++) { + reg = CN23XX_SLI_IQ_PKT_CONTROL64(i); + len += sprintf( + s + len, + "\n[%08x] (SLI_PKT%d_INPUT_CONTROL): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10010*/ + for (i = 0; i < CN23XX_MAX_INPUT_QUEUES; i++) { + reg = CN23XX_SLI_IQ_BASE_ADDR64(i); + len += sprintf( + s + len, + "\n[%08x] (SLI_PKT%d_INSTR_BADDR): %016llx\n", reg, + i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10020*/ + for (i = 0; i < CN23XX_MAX_INPUT_QUEUES; i++) { + reg = CN23XX_SLI_IQ_DOORBELL(i); + len += sprintf( + s + len, + "\n[%08x] (SLI_PKT%d_INSTR_BAOFF_DBELL): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10030*/ + for (i = 0; i < CN23XX_MAX_INPUT_QUEUES; i++) { + reg = CN23XX_SLI_IQ_SIZE(i); + len += sprintf( + s + len, + "\n[%08x] (SLI_PKT%d_INSTR_FIFO_RSIZE): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + /*0x10040*/ + for (i = 0; i < CN23XX_MAX_INPUT_QUEUES; i++) + reg = CN23XX_SLI_IQ_INSTR_COUNT64(i); + len += sprintf(s + len, + "\n[%08x] (SLI_PKT_IN_DONE%d_CNTS): %016llx\n", + reg, i, (u64)octeon_read_csr64(oct, reg)); + } + + return len; } static int cn6xxx_read_csr_reg(char *s, struct octeon_device *oct) @@ -1688,6 +2160,10 @@ static void lio_get_regs(struct net_device *dev, regs->version = OCT_ETHTOOL_REGSVER; switch (oct->chip_id) { + case OCTEON_CN23XX_PF_VID: + memset(regbuf, 0, OCT_ETHTOOL_REGDUMP_LEN_23XX); + len += cn23xx_read_csr_reg(regbuf + len, oct); + break; case OCTEON_CN68XX: case OCTEON_CN66XX: memset(regbuf, 0, OCT_ETHTOOL_REGDUMP_LEN); @@ -1729,6 +2205,7 @@ static const struct ethtool_ops lio_ethtool_ops = { .get_strings = lio_get_strings, .get_ethtool_stats = lio_get_ethtool_stats, .get_pauseparam = lio_get_pauseparam, + .set_pauseparam = lio_set_pauseparam, .get_regs_len = lio_get_regs_len, .get_regs = lio_get_regs, .get_msglevel = lio_get_msglevel, diff --git a/drivers/net/ethernet/cavium/liquidio/lio_main.c b/drivers/net/ethernet/cavium/liquidio/lio_main.c index 20d6942edf40..afc6f9dc8119 100644 --- a/drivers/net/ethernet/cavium/liquidio/lio_main.c +++ b/drivers/net/ethernet/cavium/liquidio/lio_main.c @@ -21,11 +21,10 @@ **********************************************************************/ #include <linux/version.h> #include <linux/pci.h> -#include <linux/net_tstamp.h> -#include <linux/if_vlan.h> #include <linux/firmware.h> #include <linux/ptp_clock_kernel.h> #include <net/vxlan.h> +#include <linux/kthread.h> #include "liquidio_common.h" #include "octeon_droq.h" #include "octeon_iq.h" @@ -37,6 +36,7 @@ #include "cn66xx_regs.h" #include "cn66xx_device.h" #include "cn68xx_device.h" +#include "cn23xx_pf_device.h" #include "liquidio_image.h" MODULE_AUTHOR("Cavium Networks, <support@cavium.com>"); @@ -52,11 +52,6 @@ module_param(ddr_timeout, int, 0644); MODULE_PARM_DESC(ddr_timeout, "Number of milliseconds to wait for DDR initialization. 0 waits for ddr_timeout to be set to non-zero value before starting to check"); -static u32 console_bitmask; -module_param(console_bitmask, int, 0644); -MODULE_PARM_DESC(console_bitmask, - "Bitmask indicating which consoles have debug output redirected to syslog."); - #define DEFAULT_MSG_ENABLE (NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_LINK) #define INCR_INSTRQUEUE_PKT_COUNT(octeon_dev_ptr, iq_no, field, count) \ @@ -102,6 +97,14 @@ struct liquidio_if_cfg_resp { u64 status; }; +struct liquidio_rx_ctl_context { + int octeon_id; + + wait_queue_head_t wc; + + int cond; +}; + struct oct_link_status_resp { u64 rh; struct oct_link_info link_info; @@ -139,7 +142,8 @@ union tx_info { #define OCTNIC_MAX_SG (MAX_SKB_FRAGS) #define OCTNIC_GSO_MAX_HEADER_SIZE 128 -#define OCTNIC_GSO_MAX_SIZE (GSO_MAX_SIZE - OCTNIC_GSO_MAX_HEADER_SIZE) +#define OCTNIC_GSO_MAX_SIZE \ + (CN23XX_DEFAULT_INPUT_JABBER - OCTNIC_GSO_MAX_HEADER_SIZE) /** Structure of a node in list of gather components maintained by * NIC driver for each network device. @@ -162,27 +166,6 @@ struct octnic_gather { u64 sg_dma_ptr; }; -/** This structure is used by NIC driver to store information required - * to free the sk_buff when the packet has been fetched by Octeon. - * Bytes offset below assume worst-case of a 64-bit system. - */ -struct octnet_buf_free_info { - /** Bytes 1-8. Pointer to network device private structure. */ - struct lio *lio; - - /** Bytes 9-16. Pointer to sk_buff. */ - struct sk_buff *skb; - - /** Bytes 17-24. Pointer to gather list. */ - struct octnic_gather *g; - - /** Bytes 25-32. Physical address of skb->data or gather list. */ - u64 dptr; - - /** Bytes 33-47. Piggybacked soft command, if any */ - struct octeon_soft_command *sc; -}; - struct handshake { struct completion init; struct completion started; @@ -198,6 +181,7 @@ struct octeon_device_priv { }; static int octeon_device_init(struct octeon_device *); +static int liquidio_stop(struct net_device *netdev); static void liquidio_remove(struct pci_dev *pdev); static int liquidio_probe(struct pci_dev *pdev, const struct pci_device_id *ent); @@ -219,6 +203,20 @@ static void octeon_droq_bh(unsigned long pdev) continue; reschedule |= octeon_droq_process_packets(oct, oct->droq[q_no], MAX_PACKET_BUDGET); + lio_enable_irq(oct->droq[q_no], NULL); + + if (OCTEON_CN23XX_PF(oct) && oct->msix_on) { + /* set time and cnt interrupt thresholds for this DROQ + * for NAPI + */ + int adjusted_q_no = q_no + oct->sriov_info.pf_srn; + + octeon_write_csr64( + oct, CN23XX_SLI_OQ_PKT_INT_LEVELS(adjusted_q_no), + 0x5700000040ULL); + octeon_write_csr64( + oct, CN23XX_SLI_OQ_PKTS_SENT(adjusted_q_no), 0); + } } if (reschedule) @@ -252,76 +250,6 @@ static int lio_wait_for_oq_pkts(struct octeon_device *oct) return pkt_cnt; } -void octeon_report_tx_completion_to_bql(void *txq, unsigned int pkts_compl, - unsigned int bytes_compl) -{ - struct netdev_queue *netdev_queue = txq; - - netdev_tx_completed_queue(netdev_queue, pkts_compl, bytes_compl); -} - -void octeon_update_tx_completion_counters(void *buf, int reqtype, - unsigned int *pkts_compl, - unsigned int *bytes_compl) -{ - struct octnet_buf_free_info *finfo; - struct sk_buff *skb = NULL; - struct octeon_soft_command *sc; - - switch (reqtype) { - case REQTYPE_NORESP_NET: - case REQTYPE_NORESP_NET_SG: - finfo = buf; - skb = finfo->skb; - break; - - case REQTYPE_RESP_NET_SG: - case REQTYPE_RESP_NET: - sc = buf; - skb = sc->callback_arg; - break; - - default: - return; - } - - (*pkts_compl)++; - *bytes_compl += skb->len; -} - -void octeon_report_sent_bytes_to_bql(void *buf, int reqtype) -{ - struct octnet_buf_free_info *finfo; - struct sk_buff *skb; - struct octeon_soft_command *sc; - struct netdev_queue *txq; - - switch (reqtype) { - case REQTYPE_NORESP_NET: - case REQTYPE_NORESP_NET_SG: - finfo = buf; - skb = finfo->skb; - break; - - case REQTYPE_RESP_NET_SG: - case REQTYPE_RESP_NET: - sc = buf; - skb = sc->callback_arg; - break; - - default: - return; - } - - txq = netdev_get_tx_queue(skb->dev, skb_get_queue_mapping(skb)); - netdev_tx_sent_queue(txq, skb->len); -} - -int octeon_console_debug_enabled(u32 console) -{ - return (console_bitmask >> (console)) & 0x1; -} - /** * \brief Forces all IO queues off on a given device * @param oct Pointer to Octeon device @@ -441,7 +369,7 @@ static void stop_pci_io(struct octeon_device *oct) pci_disable_device(oct->pci_dev); /* Disable interrupts */ - oct->fn_list.disable_interrupt(oct->chip); + oct->fn_list.disable_interrupt(oct, OCTEON_ALL_INTR); pcierror_quiesce_device(oct); @@ -570,6 +498,9 @@ static const struct pci_device_id liquidio_pci_tbl[] = { { /* 66xx */ PCI_VENDOR_ID_CAVIUM, 0x92, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0 }, + { /* 23xx pf */ + PCI_VENDOR_ID_CAVIUM, 0x9702, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0 + }, { 0, 0, 0, 0, 0, 0, 0 } @@ -587,7 +518,6 @@ static struct pci_driver liquidio_pci_driver = { .suspend = liquidio_suspend, .resume = liquidio_resume, #endif - }; /** @@ -936,6 +866,52 @@ static void print_link_info(struct net_device *netdev) } /** + * \brief Routine to notify MTU change + * @param work work_struct data structure + */ +static void octnet_link_status_change(struct work_struct *work) +{ + struct cavium_wk *wk = (struct cavium_wk *)work; + struct lio *lio = (struct lio *)wk->ctxptr; + + rtnl_lock(); + call_netdevice_notifiers(NETDEV_CHANGEMTU, lio->netdev); + rtnl_unlock(); +} + +/** + * \brief Sets up the mtu status change work + * @param netdev network device + */ +static inline int setup_link_status_change_wq(struct net_device *netdev) +{ + struct lio *lio = GET_LIO(netdev); + struct octeon_device *oct = lio->oct_dev; + + lio->link_status_wq.wq = alloc_workqueue("link-status", + WQ_MEM_RECLAIM, 0); + if (!lio->link_status_wq.wq) { + dev_err(&oct->pci_dev->dev, "unable to create cavium link status wq\n"); + return -1; + } + INIT_DELAYED_WORK(&lio->link_status_wq.wk.work, + octnet_link_status_change); + lio->link_status_wq.wk.ctxptr = lio; + + return 0; +} + +static inline void cleanup_link_status_change_wq(struct net_device *netdev) +{ + struct lio *lio = GET_LIO(netdev); + + if (lio->link_status_wq.wq) { + cancel_delayed_work_sync(&lio->link_status_wq.wk.work); + destroy_workqueue(lio->link_status_wq.wq); + } +} + +/** * \brief Update link status * @param netdev network device * @param ls link status structure @@ -973,8 +949,6 @@ static void update_txq_status(struct octeon_device *oct, int iq_num) struct lio *lio; struct octeon_instr_queue *iq = oct->instr_queue[iq_num]; - /*octeon_update_iq_read_idx(oct, iq);*/ - netdev = oct->props[iq->ifidx].netdev; /* This is needed because the first IQ does not have @@ -1002,12 +976,32 @@ static void update_txq_status(struct octeon_device *oct, int iq_num) } } +static +int liquidio_schedule_msix_droq_pkt_handler(struct octeon_droq *droq, u64 ret) +{ + struct octeon_device *oct = droq->oct_dev; + struct octeon_device_priv *oct_priv = + (struct octeon_device_priv *)oct->priv; + + if (droq->ops.poll_mode) { + droq->ops.napi_fn(droq); + } else { + if (ret & MSIX_PO_INT) { + tasklet_schedule(&oct_priv->droq_tasklet); + return 1; + } + /* this will be flushed periodically by check iq db */ + if (ret & MSIX_PI_INT) + return 0; + } + return 0; +} + /** * \brief Droq packet processor sceduler * @param oct octeon device */ -static -void liquidio_schedule_droq_pkt_handlers(struct octeon_device *oct) +static void liquidio_schedule_droq_pkt_handlers(struct octeon_device *oct) { struct octeon_device_priv *oct_priv = (struct octeon_device_priv *)oct->priv; @@ -1032,19 +1026,36 @@ void liquidio_schedule_droq_pkt_handlers(struct octeon_device *oct) } } +static irqreturn_t +liquidio_msix_intr_handler(int irq __attribute__((unused)), void *dev) +{ + u64 ret; + struct octeon_ioq_vector *ioq_vector = (struct octeon_ioq_vector *)dev; + struct octeon_device *oct = ioq_vector->oct_dev; + struct octeon_droq *droq = oct->droq[ioq_vector->droq_index]; + + ret = oct->fn_list.msix_interrupt_handler(ioq_vector); + + if ((ret & MSIX_PO_INT) || (ret & MSIX_PI_INT)) + liquidio_schedule_msix_droq_pkt_handler(droq, ret); + + return IRQ_HANDLED; +} + /** * \brief Interrupt handler for octeon * @param irq unused * @param dev octeon device */ static -irqreturn_t liquidio_intr_handler(int irq __attribute__((unused)), void *dev) +irqreturn_t liquidio_legacy_intr_handler(int irq __attribute__((unused)), + void *dev) { struct octeon_device *oct = (struct octeon_device *)dev; irqreturn_t ret; /* Disable our interrupts for the duration of ISR */ - oct->fn_list.disable_interrupt(oct->chip); + oct->fn_list.disable_interrupt(oct, OCTEON_ALL_INTR); ret = oct->fn_list.process_interrupt_regs(oct); @@ -1053,7 +1064,7 @@ irqreturn_t liquidio_intr_handler(int irq __attribute__((unused)), void *dev) /* Re-enable our interrupts */ if (!(atomic_read(&oct->status) == OCT_DEV_IN_RESET)) - oct->fn_list.enable_interrupt(oct->chip); + oct->fn_list.enable_interrupt(oct, OCTEON_ALL_INTR); return ret; } @@ -1067,22 +1078,204 @@ irqreturn_t liquidio_intr_handler(int irq __attribute__((unused)), void *dev) static int octeon_setup_interrupt(struct octeon_device *oct) { int irqret, err; + struct msix_entry *msix_entries; + int i; + int num_ioq_vectors; + int num_alloc_ioq_vectors; - err = pci_enable_msi(oct->pci_dev); - if (err) - dev_warn(&oct->pci_dev->dev, "Reverting to legacy interrupts. Error: %d\n", - err); - else - oct->flags |= LIO_FLAG_MSI_ENABLED; - - irqret = request_irq(oct->pci_dev->irq, liquidio_intr_handler, - IRQF_SHARED, "octeon", oct); - if (irqret) { - if (oct->flags & LIO_FLAG_MSI_ENABLED) - pci_disable_msi(oct->pci_dev); - dev_err(&oct->pci_dev->dev, "Request IRQ failed with code: %d\n", - irqret); - return 1; + if (OCTEON_CN23XX_PF(oct) && oct->msix_on) { + oct->num_msix_irqs = oct->sriov_info.num_pf_rings; + /* one non ioq interrupt for handling sli_mac_pf_int_sum */ + oct->num_msix_irqs += 1; + + oct->msix_entries = kcalloc( + oct->num_msix_irqs, sizeof(struct msix_entry), GFP_KERNEL); + if (!oct->msix_entries) + return 1; + + msix_entries = (struct msix_entry *)oct->msix_entries; + /*Assumption is that pf msix vectors start from pf srn to pf to + * trs and not from 0. if not change this code + */ + for (i = 0; i < oct->num_msix_irqs - 1; i++) + msix_entries[i].entry = oct->sriov_info.pf_srn + i; + msix_entries[oct->num_msix_irqs - 1].entry = + oct->sriov_info.trs; + num_alloc_ioq_vectors = pci_enable_msix_range( + oct->pci_dev, msix_entries, + oct->num_msix_irqs, + oct->num_msix_irqs); + if (num_alloc_ioq_vectors < 0) { + dev_err(&oct->pci_dev->dev, "unable to Allocate MSI-X interrupts\n"); + kfree(oct->msix_entries); + oct->msix_entries = NULL; + return 1; + } + dev_dbg(&oct->pci_dev->dev, "OCTEON: Enough MSI-X interrupts are allocated...\n"); + + num_ioq_vectors = oct->num_msix_irqs; + + /** For PF, there is one non-ioq interrupt handler */ + num_ioq_vectors -= 1; + irqret = request_irq(msix_entries[num_ioq_vectors].vector, + liquidio_legacy_intr_handler, 0, "octeon", + oct); + if (irqret) { + dev_err(&oct->pci_dev->dev, + "OCTEON: Request_irq failed for MSIX interrupt Error: %d\n", + irqret); + pci_disable_msix(oct->pci_dev); + kfree(oct->msix_entries); + oct->msix_entries = NULL; + return 1; + } + + for (i = 0; i < num_ioq_vectors; i++) { + irqret = request_irq(msix_entries[i].vector, + liquidio_msix_intr_handler, 0, + "octeon", &oct->ioq_vector[i]); + if (irqret) { + dev_err(&oct->pci_dev->dev, + "OCTEON: Request_irq failed for MSIX interrupt Error: %d\n", + irqret); + /** Freeing the non-ioq irq vector here . */ + free_irq(msix_entries[num_ioq_vectors].vector, + oct); + + while (i) { + i--; + /** clearing affinity mask. */ + irq_set_affinity_hint( + msix_entries[i].vector, NULL); + free_irq(msix_entries[i].vector, + &oct->ioq_vector[i]); + } + pci_disable_msix(oct->pci_dev); + kfree(oct->msix_entries); + oct->msix_entries = NULL; + return 1; + } + oct->ioq_vector[i].vector = msix_entries[i].vector; + /* assign the cpu mask for this msix interrupt vector */ + irq_set_affinity_hint( + msix_entries[i].vector, + (&oct->ioq_vector[i].affinity_mask)); + } + dev_dbg(&oct->pci_dev->dev, "OCTEON[%d]: MSI-X enabled\n", + oct->octeon_id); + } else { + err = pci_enable_msi(oct->pci_dev); + if (err) + dev_warn(&oct->pci_dev->dev, "Reverting to legacy interrupts. Error: %d\n", + err); + else + oct->flags |= LIO_FLAG_MSI_ENABLED; + + irqret = request_irq(oct->pci_dev->irq, + liquidio_legacy_intr_handler, IRQF_SHARED, + "octeon", oct); + if (irqret) { + if (oct->flags & LIO_FLAG_MSI_ENABLED) + pci_disable_msi(oct->pci_dev); + dev_err(&oct->pci_dev->dev, "Request IRQ failed with code: %d\n", + irqret); + return 1; + } + } + return 0; +} + +static int liquidio_watchdog(void *param) +{ + u64 wdog; + u16 mask_of_stuck_cores = 0; + u16 mask_of_crashed_cores = 0; + int core_num; + u8 core_is_stuck[LIO_MAX_CORES]; + u8 core_crashed[LIO_MAX_CORES]; + struct octeon_device *oct = param; + + memset(core_is_stuck, 0, sizeof(core_is_stuck)); + memset(core_crashed, 0, sizeof(core_crashed)); + + while (!kthread_should_stop()) { + mask_of_crashed_cores = + (u16)octeon_read_csr64(oct, CN23XX_SLI_SCRATCH2); + + for (core_num = 0; core_num < LIO_MAX_CORES; core_num++) { + if (!core_is_stuck[core_num]) { + wdog = lio_pci_readq(oct, CIU3_WDOG(core_num)); + + /* look at watchdog state field */ + wdog &= CIU3_WDOG_MASK; + if (wdog) { + /* this watchdog timer has expired */ + core_is_stuck[core_num] = + LIO_MONITOR_WDOG_EXPIRE; + mask_of_stuck_cores |= (1 << core_num); + } + } + + if (!core_crashed[core_num]) + core_crashed[core_num] = + (mask_of_crashed_cores >> core_num) & 1; + } + + if (mask_of_stuck_cores) { + for (core_num = 0; core_num < LIO_MAX_CORES; + core_num++) { + if (core_is_stuck[core_num] == 1) { + dev_err(&oct->pci_dev->dev, + "ERROR: Octeon core %d is stuck!\n", + core_num); + /* 2 means we have printk'd an error + * so no need to repeat the same printk + */ + core_is_stuck[core_num] = + LIO_MONITOR_CORE_STUCK_MSGD; + } + } + } + + if (mask_of_crashed_cores) { + for (core_num = 0; core_num < LIO_MAX_CORES; + core_num++) { + if (core_crashed[core_num] == 1) { + dev_err(&oct->pci_dev->dev, + "ERROR: Octeon core %d crashed! See oct-fwdump for details.\n", + core_num); + /* 2 means we have printk'd an error + * so no need to repeat the same printk + */ + core_crashed[core_num] = + LIO_MONITOR_CORE_STUCK_MSGD; + } + } + } +#ifdef CONFIG_MODULE_UNLOAD + if (mask_of_stuck_cores || mask_of_crashed_cores) { + /* make module refcount=0 so that rmmod will work */ + long refcount; + + refcount = module_refcount(THIS_MODULE); + + while (refcount > 0) { + module_put(THIS_MODULE); + refcount = module_refcount(THIS_MODULE); + } + + /* compensate for and withstand an unlikely (but still + * possible) race condition + */ + while (refcount < 0) { + try_module_get(THIS_MODULE); + refcount = module_refcount(THIS_MODULE); + } + } +#endif + /* sleep for two seconds */ + set_current_state(TASK_INTERRUPTIBLE); + schedule_timeout(2 * HZ); } return 0; @@ -1107,6 +1300,9 @@ liquidio_probe(struct pci_dev *pdev, return -ENOMEM; } + if (pdev->device == OCTEON_CN23XX_PF_VID) + oct_dev->msix_on = LIO_FLAG_MSIX_ENABLED; + dev_info(&pdev->dev, "Initializing device %x:%x.\n", (u32)pdev->vendor, (u32)pdev->device); @@ -1130,6 +1326,30 @@ liquidio_probe(struct pci_dev *pdev, return -ENOMEM; } + if (OCTEON_CN23XX_PF(oct_dev)) { + u64 scratch1; + u8 bus, device, function; + + scratch1 = octeon_read_csr64(oct_dev, CN23XX_SLI_SCRATCH1); + if (!(scratch1 & 4ULL)) { + /* Bit 2 of SLI_SCRATCH_1 is a flag that indicates that + * the lio watchdog kernel thread is running for this + * NIC. Each NIC gets one watchdog kernel thread. + */ + scratch1 |= 4ULL; + octeon_write_csr64(oct_dev, CN23XX_SLI_SCRATCH1, + scratch1); + + bus = pdev->bus->number; + device = PCI_SLOT(pdev->devfn); + function = PCI_FUNC(pdev->devfn); + oct_dev->watchdog_task = kthread_create( + liquidio_watchdog, oct_dev, + "liowd/%02hhx:%02hhx.%hhx", bus, device, function); + wake_up_process(oct_dev->watchdog_task); + } + } + oct_dev->rx_pause = 1; oct_dev->tx_pause = 1; @@ -1146,6 +1366,7 @@ liquidio_probe(struct pci_dev *pdev, static void octeon_destroy_resources(struct octeon_device *oct) { int i; + struct msix_entry *msix_entries; struct octeon_device_priv *oct_priv = (struct octeon_device_priv *)oct->priv; @@ -1190,21 +1411,40 @@ static void octeon_destroy_resources(struct octeon_device *oct) dev_err(&oct->pci_dev->dev, "OQ had pending packets\n"); /* Disable interrupts */ - oct->fn_list.disable_interrupt(oct->chip); + oct->fn_list.disable_interrupt(oct, OCTEON_ALL_INTR); + + if (oct->msix_on) { + msix_entries = (struct msix_entry *)oct->msix_entries; + for (i = 0; i < oct->num_msix_irqs - 1; i++) { + /* clear the affinity_cpumask */ + irq_set_affinity_hint(msix_entries[i].vector, + NULL); + free_irq(msix_entries[i].vector, + &oct->ioq_vector[i]); + } + /* non-iov vector's argument is oct struct */ + free_irq(msix_entries[i].vector, oct); - /* Release the interrupt line */ - free_irq(oct->pci_dev->irq, oct); + pci_disable_msix(oct->pci_dev); + kfree(oct->msix_entries); + oct->msix_entries = NULL; + } else { + /* Release the interrupt line */ + free_irq(oct->pci_dev->irq, oct); - if (oct->flags & LIO_FLAG_MSI_ENABLED) - pci_disable_msi(oct->pci_dev); + if (oct->flags & LIO_FLAG_MSI_ENABLED) + pci_disable_msi(oct->pci_dev); + } - /* fallthrough */ + if (OCTEON_CN23XX_PF(oct)) + octeon_free_ioq_vector(oct); + /* fallthrough */ case OCT_DEV_IN_RESET: case OCT_DEV_DROQ_INIT_DONE: /*atomic_set(&oct->status, OCT_DEV_DROQ_INIT_DONE);*/ mdelay(100); for (i = 0; i < MAX_OCTEON_OUTPUT_QUEUES(oct); i++) { - if (!(oct->io_qmask.oq & (1ULL << i))) + if (!(oct->io_qmask.oq & BIT_ULL(i))) continue; octeon_delete_droq(oct, i); } @@ -1226,16 +1466,15 @@ static void octeon_destroy_resources(struct octeon_device *oct) octeon_delete_response_list(oct); /* fallthrough */ - case OCT_DEV_SC_BUFF_POOL_INIT_DONE: - octeon_free_sc_buffer_pool(oct); - - /* fallthrough */ case OCT_DEV_INSTR_QUEUE_INIT_DONE: for (i = 0; i < MAX_OCTEON_INSTR_QUEUES(oct); i++) { - if (!(oct->io_qmask.iq & (1ULL << i))) + if (!(oct->io_qmask.iq & BIT_ULL(i))) continue; octeon_delete_instr_queue(oct, i); } + /* fallthrough */ + case OCT_DEV_SC_BUFF_POOL_INIT_DONE: + octeon_free_sc_buffer_pool(oct); /* fallthrough */ case OCT_DEV_DISPATCH_INIT_DONE: @@ -1244,9 +1483,9 @@ static void octeon_destroy_resources(struct octeon_device *oct) /* fallthrough */ case OCT_DEV_PCI_MAP_DONE: - /* Soft reset the octeon device before exiting */ - oct->fn_list.soft_reset(oct); + if ((!OCTEON_CN23XX_PF(oct)) || !oct->octeon_id) + oct->fn_list.soft_reset(oct); octeon_unmap_pci_barx(oct, 0); octeon_unmap_pci_barx(oct, 1); @@ -1264,23 +1503,89 @@ static void octeon_destroy_resources(struct octeon_device *oct) } /** + * \brief Callback for rx ctrl + * @param status status of request + * @param buf pointer to resp structure + */ +static void rx_ctl_callback(struct octeon_device *oct, + u32 status, + void *buf) +{ + struct octeon_soft_command *sc = (struct octeon_soft_command *)buf; + struct liquidio_rx_ctl_context *ctx; + + ctx = (struct liquidio_rx_ctl_context *)sc->ctxptr; + + oct = lio_get_device(ctx->octeon_id); + if (status) + dev_err(&oct->pci_dev->dev, "rx ctl instruction failed. Status: %llx\n", + CVM_CAST64(status)); + WRITE_ONCE(ctx->cond, 1); + + /* This barrier is required to be sure that the response has been + * written fully before waking up the handler + */ + wmb(); + + wake_up_interruptible(&ctx->wc); +} + +/** * \brief Send Rx control command * @param lio per-network private data * @param start_stop whether to start or stop */ static void send_rx_ctrl_cmd(struct lio *lio, int start_stop) { - struct octnic_ctrl_pkt nctrl; + struct octeon_soft_command *sc; + struct liquidio_rx_ctl_context *ctx; + union octnet_cmd *ncmd; + int ctx_size = sizeof(struct liquidio_rx_ctl_context); + struct octeon_device *oct = (struct octeon_device *)lio->oct_dev; + int retval; - memset(&nctrl, 0, sizeof(struct octnic_ctrl_pkt)); + if (oct->props[lio->ifidx].rx_on == start_stop) + return; - nctrl.ncmd.s.cmd = OCTNET_CMD_RX_CTL; - nctrl.ncmd.s.param1 = start_stop; - nctrl.iq_no = lio->linfo.txpciq[0].s.q_no; - nctrl.netpndev = (u64)lio->netdev; + sc = (struct octeon_soft_command *) + octeon_alloc_soft_command(oct, OCTNET_CMD_SIZE, + 16, ctx_size); - if (octnet_send_nic_ctrl_pkt(lio->oct_dev, &nctrl) < 0) + ncmd = (union octnet_cmd *)sc->virtdptr; + ctx = (struct liquidio_rx_ctl_context *)sc->ctxptr; + + WRITE_ONCE(ctx->cond, 0); + ctx->octeon_id = lio_get_device_id(oct); + init_waitqueue_head(&ctx->wc); + + ncmd->u64 = 0; + ncmd->s.cmd = OCTNET_CMD_RX_CTL; + ncmd->s.param1 = start_stop; + + octeon_swap_8B_data((u64 *)ncmd, (OCTNET_CMD_SIZE >> 3)); + + sc->iq_no = lio->linfo.txpciq[0].s.q_no; + + octeon_prepare_soft_command(oct, sc, OPCODE_NIC, + OPCODE_NIC_CMD, 0, 0, 0); + + sc->callback = rx_ctl_callback; + sc->callback_arg = sc; + sc->wait_time = 5000; + + retval = octeon_send_soft_command(oct, sc); + if (retval == IQ_SEND_FAILED) { netif_info(lio, rx_err, lio->netdev, "Failed to send RX Control message\n"); + } else { + /* Sleep on a wait queue till the cond flag indicates that the + * response arrived or timed-out. + */ + if (sleep_cond(&ctx->wc, &ctx->cond) == -EINTR) + return; + oct->props[lio->ifidx].rx_on = start_stop; + } + + octeon_free_soft_command(oct, sc); } /** @@ -1307,21 +1612,24 @@ static void liquidio_destroy_nic_device(struct octeon_device *oct, int ifidx) dev_dbg(&oct->pci_dev->dev, "NIC device cleanup\n"); - send_rx_ctrl_cmd(lio, 0); - if (atomic_read(&lio->ifstate) & LIO_IFSTATE_RUNNING) - txqs_stop(netdev); + liquidio_stop(netdev); if (oct->props[lio->ifidx].napi_enabled == 1) { list_for_each_entry_safe(napi, n, &netdev->napi_list, dev_list) napi_disable(napi); oct->props[lio->ifidx].napi_enabled = 0; + + if (OCTEON_CN23XX_PF(oct)) + oct->droq[0]->ops.poll_mode = 0; } if (atomic_read(&lio->ifstate) & LIO_IFSTATE_REGISTERED) unregister_netdev(netdev); + cleanup_link_status_change_wq(netdev); + delete_glists(lio); free_netdev(netdev); @@ -1374,6 +1682,9 @@ static void liquidio_remove(struct pci_dev *pdev) dev_dbg(&oct_dev->pci_dev->dev, "Stopping device\n"); + if (oct_dev->watchdog_task) + kthread_stop(oct_dev->watchdog_task); + if (oct_dev->app_mode && (oct_dev->app_mode == CVM_DRV_NIC_APP)) liquidio_stop_nic_module(oct_dev); @@ -1417,6 +1728,12 @@ static int octeon_chip_specific_setup(struct octeon_device *oct) s = "CN66XX"; break; + case OCTEON_CN23XX_PCIID_PF: + oct->chip_id = OCTEON_CN23XX_PF_VID; + ret = setup_cn23xx_octeon_pf_device(oct); + s = "CN23XX"; + break; + default: s = "?"; dev_err(&oct->pci_dev->dev, "Unknown device found (dev_id: %x)\n", @@ -1867,7 +2184,7 @@ static void if_cfg_callback(struct octeon_device *oct, struct liquidio_if_cfg_context *ctx; resp = (struct liquidio_if_cfg_resp *)sc->virtrptr; - ctx = (struct liquidio_if_cfg_context *)sc->ctxptr; + ctx = (struct liquidio_if_cfg_context *)sc->ctxptr; oct = lio_get_device(ctx->octeon_id); if (resp->status) @@ -2060,11 +2377,14 @@ static void napi_schedule_wrapper(void *param) */ static void liquidio_napi_drv_callback(void *arg) { + struct octeon_device *oct; struct octeon_droq *droq = arg; int this_cpu = smp_processor_id(); - if (droq->cpu_id == this_cpu) { - napi_schedule(&droq->napi); + oct = droq->oct_dev; + + if (OCTEON_CN23XX_PF(oct) || droq->cpu_id == this_cpu) { + napi_schedule_irqoff(&droq->napi); } else { struct call_single_data *csd = &droq->csd; @@ -2173,17 +2493,15 @@ static inline int setup_io_queues(struct octeon_device *octeon_dev, lio->ifidx), NULL); if (retval) { dev_err(&octeon_dev->pci_dev->dev, - " %s : Runtime DROQ(RxQ) creation failed.\n", + "%s : Runtime DROQ(RxQ) creation failed.\n", __func__); return 1; } droq = octeon_dev->droq[q_no]; napi = &droq->napi; - dev_dbg(&octeon_dev->pci_dev->dev, - "netif_napi_add netdev:%llx oct:%llx\n", - (u64)netdev, - (u64)octeon_dev); + dev_dbg(&octeon_dev->pci_dev->dev, "netif_napi_add netdev:%llx oct:%llx pf_num:%d\n", + (u64)netdev, (u64)octeon_dev, octeon_dev->pf_num); netif_napi_add(netdev, napi, liquidio_napi_poll, 64); /* designate a CPU for this droq */ @@ -2195,6 +2513,14 @@ static inline int setup_io_queues(struct octeon_device *octeon_dev, octeon_register_droq_ops(octeon_dev, q_no, &droq_ops); } + if (OCTEON_CN23XX_PF(octeon_dev)) { + /* 23XX PF can receive control messages (via the first PF-owned + * droq) from the firmware even if the ethX interface is down, + * so that's why poll_mode must be off for the first droq. + */ + octeon_dev->droq[0]->ops.poll_mode = 0; + } + /* set up IQs. */ for (q = 0; q < lio->linfo.num_txpciq; q++) { num_tx_descs = CFG_GET_NUM_TX_DESCS_NIC_IF(octeon_get_conf @@ -2235,7 +2561,7 @@ static void octnet_poll_check_txq_status(struct work_struct *work) * \brief Sets up the txq poll check * @param netdev network device */ -static inline void setup_tx_poll_fn(struct net_device *netdev) +static inline int setup_tx_poll_fn(struct net_device *netdev) { struct lio *lio = GET_LIO(netdev); struct octeon_device *oct = lio->oct_dev; @@ -2244,21 +2570,24 @@ static inline void setup_tx_poll_fn(struct net_device *netdev) WQ_MEM_RECLAIM, 0); if (!lio->txq_status_wq.wq) { dev_err(&oct->pci_dev->dev, "unable to create cavium txq status wq\n"); - return; + return -1; } INIT_DELAYED_WORK(&lio->txq_status_wq.wk.work, octnet_poll_check_txq_status); lio->txq_status_wq.wk.ctxptr = lio; queue_delayed_work(lio->txq_status_wq.wq, &lio->txq_status_wq.wk.work, msecs_to_jiffies(1)); + return 0; } static inline void cleanup_tx_poll_fn(struct net_device *netdev) { struct lio *lio = GET_LIO(netdev); - cancel_delayed_work_sync(&lio->txq_status_wq.wk.work); - destroy_workqueue(lio->txq_status_wq.wq); + if (lio->txq_status_wq.wq) { + cancel_delayed_work_sync(&lio->txq_status_wq.wk.work); + destroy_workqueue(lio->txq_status_wq.wq); + } } /** @@ -2276,24 +2605,34 @@ static int liquidio_open(struct net_device *netdev) napi_enable(napi); oct->props[lio->ifidx].napi_enabled = 1; + + if (OCTEON_CN23XX_PF(oct)) + oct->droq[0]->ops.poll_mode = 1; } oct_ptp_open(netdev); ifstate_set(lio, LIO_IFSTATE_RUNNING); - setup_tx_poll_fn(netdev); - - start_txq(netdev); + /* Ready for link status updates */ + lio->intf_open = 1; netif_info(lio, ifup, lio->netdev, "Interface Open, ready for traffic\n"); + if (OCTEON_CN23XX_PF(oct)) { + if (!oct->msix_on) + if (setup_tx_poll_fn(netdev)) + return -1; + } else { + if (setup_tx_poll_fn(netdev)) + return -1; + } + + start_txq(netdev); + /* tell Octeon to start forwarding packets to host */ send_rx_ctrl_cmd(lio, 1); - /* Ready for link status updates */ - lio->intf_open = 1; - dev_info(&oct->pci_dev->dev, "%s interface is opened\n", netdev->name); @@ -2328,7 +2667,12 @@ static int liquidio_stop(struct net_device *netdev) /* Now it should be safe to tell Octeon that nic interface is down. */ send_rx_ctrl_cmd(lio, 0); - cleanup_tx_poll_fn(netdev); + if (OCTEON_CN23XX_PF(oct)) { + if (!oct->msix_on) + cleanup_tx_poll_fn(netdev); + } else { + cleanup_tx_poll_fn(netdev); + } if (lio->ptp_clock) { ptp_clock_unregister(lio->ptp_clock); @@ -2340,143 +2684,6 @@ static int liquidio_stop(struct net_device *netdev) return 0; } -void liquidio_link_ctrl_cmd_completion(void *nctrl_ptr) -{ - struct octnic_ctrl_pkt *nctrl = (struct octnic_ctrl_pkt *)nctrl_ptr; - struct net_device *netdev = (struct net_device *)nctrl->netpndev; - struct lio *lio = GET_LIO(netdev); - struct octeon_device *oct = lio->oct_dev; - u8 *mac; - - switch (nctrl->ncmd.s.cmd) { - case OCTNET_CMD_CHANGE_DEVFLAGS: - case OCTNET_CMD_SET_MULTI_LIST: - break; - - case OCTNET_CMD_CHANGE_MACADDR: - mac = ((u8 *)&nctrl->udd[0]) + 2; - netif_info(lio, probe, lio->netdev, - "%s %2.2x:%2.2x:%2.2x:%2.2x:%2.2x:%2.2x\n", - "MACAddr changed to", mac[0], mac[1], - mac[2], mac[3], mac[4], mac[5]); - break; - - case OCTNET_CMD_CHANGE_MTU: - /* If command is successful, change the MTU. */ - netif_info(lio, probe, lio->netdev, " MTU Changed from %d to %d\n", - netdev->mtu, nctrl->ncmd.s.param1); - dev_info(&oct->pci_dev->dev, "%s MTU Changed from %d to %d\n", - netdev->name, netdev->mtu, - nctrl->ncmd.s.param1); - rtnl_lock(); - netdev->mtu = nctrl->ncmd.s.param1; - call_netdevice_notifiers(NETDEV_CHANGEMTU, netdev); - rtnl_unlock(); - break; - - case OCTNET_CMD_GPIO_ACCESS: - netif_info(lio, probe, lio->netdev, "LED Flashing visual identification\n"); - - break; - - case OCTNET_CMD_LRO_ENABLE: - dev_info(&oct->pci_dev->dev, "%s LRO Enabled\n", netdev->name); - break; - - case OCTNET_CMD_LRO_DISABLE: - dev_info(&oct->pci_dev->dev, "%s LRO Disabled\n", - netdev->name); - break; - - case OCTNET_CMD_VERBOSE_ENABLE: - dev_info(&oct->pci_dev->dev, "%s LRO Enabled\n", netdev->name); - break; - - case OCTNET_CMD_VERBOSE_DISABLE: - dev_info(&oct->pci_dev->dev, "%s LRO Disabled\n", - netdev->name); - break; - - case OCTNET_CMD_ENABLE_VLAN_FILTER: - dev_info(&oct->pci_dev->dev, "%s VLAN filter enabled\n", - netdev->name); - break; - - case OCTNET_CMD_ADD_VLAN_FILTER: - dev_info(&oct->pci_dev->dev, "%s VLAN filter %d added\n", - netdev->name, nctrl->ncmd.s.param1); - break; - - case OCTNET_CMD_DEL_VLAN_FILTER: - dev_info(&oct->pci_dev->dev, "%s VLAN filter %d removed\n", - netdev->name, nctrl->ncmd.s.param1); - break; - - case OCTNET_CMD_SET_SETTINGS: - dev_info(&oct->pci_dev->dev, "%s settings changed\n", - netdev->name); - - break; - /* Case to handle "OCTNET_CMD_TNL_RX_CSUM_CTL" - * Command passed by NIC driver - */ - case OCTNET_CMD_TNL_RX_CSUM_CTL: - if (nctrl->ncmd.s.param1 == OCTNET_CMD_RXCSUM_ENABLE) { - netif_info(lio, probe, lio->netdev, - "%s RX Checksum Offload Enabled\n", - netdev->name); - } else if (nctrl->ncmd.s.param1 == - OCTNET_CMD_RXCSUM_DISABLE) { - netif_info(lio, probe, lio->netdev, - "%s RX Checksum Offload Disabled\n", - netdev->name); - } - break; - - /* Case to handle "OCTNET_CMD_TNL_TX_CSUM_CTL" - * Command passed by NIC driver - */ - case OCTNET_CMD_TNL_TX_CSUM_CTL: - if (nctrl->ncmd.s.param1 == OCTNET_CMD_TXCSUM_ENABLE) { - netif_info(lio, probe, lio->netdev, - "%s TX Checksum Offload Enabled\n", - netdev->name); - } else if (nctrl->ncmd.s.param1 == - OCTNET_CMD_TXCSUM_DISABLE) { - netif_info(lio, probe, lio->netdev, - "%s TX Checksum Offload Disabled\n", - netdev->name); - } - break; - - /* Case to handle "OCTNET_CMD_VXLAN_PORT_CONFIG" - * Command passed by NIC driver - */ - case OCTNET_CMD_VXLAN_PORT_CONFIG: - if (nctrl->ncmd.s.more == OCTNET_CMD_VXLAN_PORT_ADD) { - netif_info(lio, probe, lio->netdev, - "%s VxLAN Destination UDP PORT:%d ADDED\n", - netdev->name, - nctrl->ncmd.s.param1); - } else if (nctrl->ncmd.s.more == - OCTNET_CMD_VXLAN_PORT_DEL) { - netif_info(lio, probe, lio->netdev, - "%s VxLAN Destination UDP PORT:%d DELETED\n", - netdev->name, - nctrl->ncmd.s.param1); - } - break; - - case OCTNET_CMD_SET_FLOW_CTL: - netif_info(lio, probe, lio->netdev, "Set RX/TX flow control parameters\n"); - break; - - default: - dev_err(&oct->pci_dev->dev, "%s Unknown cmd %d\n", __func__, - nctrl->ncmd.s.cmd); - } -} - /** * \brief Converts a mask based on net device flags * @param netdev network device @@ -2817,8 +3024,7 @@ static void handle_timestamp(struct octeon_device *oct, */ static inline int send_nic_timestamp_pkt(struct octeon_device *oct, struct octnic_data_pkt *ndata, - struct octnet_buf_free_info *finfo, - int xmit_more) + struct octnet_buf_free_info *finfo) { int retval; struct octeon_soft_command *sc; @@ -2846,9 +3052,15 @@ static inline int send_nic_timestamp_pkt(struct octeon_device *oct, sc->callback_arg = finfo->skb; sc->iq_no = ndata->q_no; - len = (u32)((struct octeon_instr_ih2 *)(&sc->cmd.cmd2.ih2))->dlengsz; + if (OCTEON_CN23XX_PF(oct)) + len = (u32)((struct octeon_instr_ih3 *) + (&sc->cmd.cmd3.ih3))->dlengsz; + else + len = (u32)((struct octeon_instr_ih2 *) + (&sc->cmd.cmd2.ih2))->dlengsz; + + ring_doorbell = 1; - ring_doorbell = !xmit_more; retval = octeon_send_command(oct, sc->iq_no, ring_doorbell, &sc->cmd, sc, len, ndata->reqtype); @@ -2881,7 +3093,7 @@ static int liquidio_xmit(struct sk_buff *skb, struct net_device *netdev) union tx_info *tx_info; int status = 0; int q_idx = 0, iq_no = 0; - int xmit_more, j; + int j; u64 dptr = 0; u32 tag = 0; @@ -2980,7 +3192,10 @@ static int liquidio_xmit(struct sk_buff *skb, struct net_device *netdev) return NETDEV_TX_BUSY; } - ndata.cmd.cmd2.dptr = dptr; + if (OCTEON_CN23XX_PF(oct)) + ndata.cmd.cmd3.dptr = dptr; + else + ndata.cmd.cmd2.dptr = dptr; finfo->dptr = dptr; ndata.reqtype = REQTYPE_NORESP_NET; @@ -3055,15 +3270,23 @@ static int liquidio_xmit(struct sk_buff *skb, struct net_device *netdev) g->sg_size, DMA_TO_DEVICE); dptr = g->sg_dma_ptr; - ndata.cmd.cmd2.dptr = dptr; + if (OCTEON_CN23XX_PF(oct)) + ndata.cmd.cmd3.dptr = dptr; + else + ndata.cmd.cmd2.dptr = dptr; finfo->dptr = dptr; finfo->g = g; ndata.reqtype = REQTYPE_NORESP_NET_SG; } - irh = (struct octeon_instr_irh *)&ndata.cmd.cmd2.irh; - tx_info = (union tx_info *)&ndata.cmd.cmd2.ossp[0]; + if (OCTEON_CN23XX_PF(oct)) { + irh = (struct octeon_instr_irh *)&ndata.cmd.cmd3.irh; + tx_info = (union tx_info *)&ndata.cmd.cmd3.ossp[0]; + } else { + irh = (struct octeon_instr_irh *)&ndata.cmd.cmd2.irh; + tx_info = (union tx_info *)&ndata.cmd.cmd2.ossp[0]; + } if (skb_shinfo(skb)->gso_size) { tx_info->s.gso_size = skb_shinfo(skb)->gso_size; @@ -3077,12 +3300,10 @@ static int liquidio_xmit(struct sk_buff *skb, struct net_device *netdev) irh->vlan = skb_vlan_tag_get(skb) & 0xfff; } - xmit_more = skb->xmit_more; - if (unlikely(cmdsetup.s.timestamp)) - status = send_nic_timestamp_pkt(oct, &ndata, finfo, xmit_more); + status = send_nic_timestamp_pkt(oct, &ndata, finfo); else - status = octnet_send_nic_data_pkt(oct, &ndata, xmit_more); + status = octnet_send_nic_data_pkt(oct, &ndata); if (status == IQ_SEND_FAILED) goto lio_xmit_failed; @@ -3190,8 +3411,8 @@ static int liquidio_vlan_rx_kill_vid(struct net_device *netdev, * OCTNET_CMD_RXCSUM_DISABLE * @returns SUCCESS or FAILURE */ -int liquidio_set_rxcsum_command(struct net_device *netdev, int command, - u8 rx_cmd) +static int liquidio_set_rxcsum_command(struct net_device *netdev, int command, + u8 rx_cmd) { struct lio *lio = GET_LIO(netdev); struct octeon_device *oct = lio->oct_dev; @@ -3249,31 +3470,6 @@ static int liquidio_vxlan_port_command(struct net_device *netdev, int command, return ret; } -int liquidio_set_feature(struct net_device *netdev, int cmd, u16 param1) -{ - struct lio *lio = GET_LIO(netdev); - struct octeon_device *oct = lio->oct_dev; - struct octnic_ctrl_pkt nctrl; - int ret = 0; - - memset(&nctrl, 0, sizeof(struct octnic_ctrl_pkt)); - - nctrl.ncmd.u64 = 0; - nctrl.ncmd.s.cmd = cmd; - nctrl.ncmd.s.param1 = param1; - nctrl.iq_no = lio->linfo.txpciq[0].s.q_no; - nctrl.wait_time = 100; - nctrl.netpndev = (u64)netdev; - nctrl.cb_fn = liquidio_link_ctrl_cmd_completion; - - ret = octnet_send_nic_ctrl_pkt(lio->oct_dev, &nctrl); - if (ret < 0) { - dev_err(&oct->pci_dev->dev, "Feature change failed in core (ret: 0x%x)\n", - ret); - } - return ret; -} - /** \brief Net device fix features * @param netdev pointer to network device * @param request features requested @@ -3492,8 +3688,9 @@ static int setup_nic_devices(struct octeon_device *octeon_dev) union oct_nic_if_cfg if_cfg; unsigned int base_queue; unsigned int gmx_port_id; - u32 resp_size, ctx_size; + u32 resp_size, ctx_size, data_size; u32 ifidx_or_pfnum; + struct lio_version *vdata; /* This is to handle link status changes */ octeon_register_dispatch_fn(octeon_dev, OPCODE_NIC, @@ -3515,21 +3712,37 @@ static int setup_nic_devices(struct octeon_device *octeon_dev) for (i = 0; i < octeon_dev->ifcount; i++) { resp_size = sizeof(struct liquidio_if_cfg_resp); ctx_size = sizeof(struct liquidio_if_cfg_context); + data_size = sizeof(struct lio_version); sc = (struct octeon_soft_command *) - octeon_alloc_soft_command(octeon_dev, 0, + octeon_alloc_soft_command(octeon_dev, data_size, resp_size, ctx_size); resp = (struct liquidio_if_cfg_resp *)sc->virtrptr; ctx = (struct liquidio_if_cfg_context *)sc->ctxptr; + vdata = (struct lio_version *)sc->virtdptr; + + *((u64 *)vdata) = 0; + vdata->major = cpu_to_be16(LIQUIDIO_BASE_MAJOR_VERSION); + vdata->minor = cpu_to_be16(LIQUIDIO_BASE_MINOR_VERSION); + vdata->micro = cpu_to_be16(LIQUIDIO_BASE_MICRO_VERSION); - num_iqueues = - CFG_GET_NUM_TXQS_NIC_IF(octeon_get_conf(octeon_dev), i); - num_oqueues = - CFG_GET_NUM_RXQS_NIC_IF(octeon_get_conf(octeon_dev), i); - base_queue = - CFG_GET_BASE_QUE_NIC_IF(octeon_get_conf(octeon_dev), i); - gmx_port_id = - CFG_GET_GMXID_NIC_IF(octeon_get_conf(octeon_dev), i); - ifidx_or_pfnum = i; + if (OCTEON_CN23XX_PF(octeon_dev)) { + num_iqueues = octeon_dev->sriov_info.num_pf_rings; + num_oqueues = octeon_dev->sriov_info.num_pf_rings; + base_queue = octeon_dev->sriov_info.pf_srn; + + gmx_port_id = octeon_dev->pf_num; + ifidx_or_pfnum = octeon_dev->pf_num; + } else { + num_iqueues = CFG_GET_NUM_TXQS_NIC_IF( + octeon_get_conf(octeon_dev), i); + num_oqueues = CFG_GET_NUM_RXQS_NIC_IF( + octeon_get_conf(octeon_dev), i); + base_queue = CFG_GET_BASE_QUE_NIC_IF( + octeon_get_conf(octeon_dev), i); + gmx_port_id = CFG_GET_GMXID_NIC_IF( + octeon_get_conf(octeon_dev), i); + ifidx_or_pfnum = i; + } dev_dbg(&octeon_dev->pci_dev->dev, "requesting config for interface %d, iqs %d, oqs %d\n", @@ -3566,7 +3779,11 @@ static int setup_nic_devices(struct octeon_device *octeon_dev) /* Sleep on a wait queue till the cond flag indicates that the * response arrived or timed-out. */ - sleep_cond(&ctx->wc, &ctx->cond); + if (sleep_cond(&ctx->wc, &ctx->cond) == -EINTR) { + dev_err(&octeon_dev->pci_dev->dev, "Wait interrupted\n"); + goto setup_nic_wait_intr; + } + retval = resp->status; if (retval) { dev_err(&octeon_dev->pci_dev->dev, "iq/oq config failed\n"); @@ -3633,12 +3850,16 @@ static int setup_nic_devices(struct octeon_device *octeon_dev) lio->msg_enable = netif_msg_init(debug, DEFAULT_MSG_ENABLE); - lio->dev_capability = NETIF_F_HIGHDMA - | NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM - | NETIF_F_SG | NETIF_F_RXCSUM - | NETIF_F_GRO - | NETIF_F_TSO | NETIF_F_TSO6 - | NETIF_F_LRO; + if (OCTEON_CN23XX_PF(octeon_dev) || + OCTEON_CN6XXX(octeon_dev)) { + lio->dev_capability = NETIF_F_HIGHDMA + | NETIF_F_IP_CSUM + | NETIF_F_IPV6_CSUM + | NETIF_F_SG | NETIF_F_RXCSUM + | NETIF_F_GRO + | NETIF_F_TSO | NETIF_F_TSO6 + | NETIF_F_LRO; + } netif_set_gso_max_size(netdev, OCTNIC_GSO_MAX_SIZE); /* Copy of transmit encapsulation capabilities: @@ -3713,7 +3934,10 @@ static int setup_nic_devices(struct octeon_device *octeon_dev) /* Register ethtool support */ liquidio_set_ethtool_ops(netdev); - octeon_dev->priv_flags = 0x0; + if (lio->oct_dev->chip_id == OCTEON_CN23XX_PF_VID) + octeon_dev->priv_flags = OCT_PRIV_FLAG_DEFAULT; + else + octeon_dev->priv_flags = 0x0; if (netdev->features & NETIF_F_LRO) liquidio_set_feature(netdev, OCTNET_CMD_LRO_ENABLE, @@ -3725,6 +3949,9 @@ static int setup_nic_devices(struct octeon_device *octeon_dev) liquidio_set_feature(netdev, OCTNET_CMD_VERBOSE_ENABLE, 0); + if (setup_link_status_change_wq(netdev)) + goto setup_nic_dev_fail; + /* Register the network device with the OS */ if (register_netdev(netdev)) { dev_err(&octeon_dev->pci_dev->dev, "Device registration failed\n"); @@ -3760,6 +3987,8 @@ setup_nic_dev_fail: octeon_free_soft_command(octeon_dev, sc); +setup_nic_wait_intr: + while (i--) { dev_err(&octeon_dev->pci_dev->dev, "NIC ifidx:%d Setup failed\n", i); @@ -3789,8 +4018,7 @@ static int liquidio_init_nic_module(struct octeon_device *oct) /* run port_config command for each port */ oct->ifcount = num_nic_ports; - memset(oct->props, 0, - sizeof(struct octdev_props) * num_nic_ports); + memset(oct->props, 0, sizeof(struct octdev_props) * num_nic_ports); for (i = 0; i < MAX_OCTEON_LINKS; i++) oct->props[i].gmxport = -1; @@ -3806,7 +4034,7 @@ static int liquidio_init_nic_module(struct octeon_device *oct) /* Initialize interrupt moderation params */ intrmod_cfg = &((struct octeon_device *)oct)->intrmod; intrmod_cfg->rx_enable = 1; - intrmod_cfg->check_intrvl = LIO_INTRMOD_CHECK_INTERVAL; + intrmod_cfg->check_intrvl = LIO_INTRMOD_CHECK_INTERVAL; intrmod_cfg->maxpkt_ratethr = LIO_INTRMOD_MAXPKT_RATETHR; intrmod_cfg->minpkt_ratethr = LIO_INTRMOD_MINPKT_RATETHR; intrmod_cfg->rx_maxcnt_trigger = LIO_INTRMOD_RXMAXCNT_TRIGGER; @@ -3818,6 +4046,7 @@ static int liquidio_init_nic_module(struct octeon_device *oct) intrmod_cfg->tx_mincnt_trigger = LIO_INTRMOD_TXMINCNT_TRIGGER; intrmod_cfg->rx_frames = CFG_GET_OQ_INTR_PKT(octeon_get_conf(oct)); intrmod_cfg->rx_usecs = CFG_GET_OQ_INTR_TIME(octeon_get_conf(oct)); + intrmod_cfg->tx_frames = CFG_GET_IQ_INTR_PKT(octeon_get_conf(oct)); dev_dbg(&oct->pci_dev->dev, "Network interfaces ready\n"); return retval; @@ -3880,6 +4109,7 @@ static void nic_starter(struct work_struct *work) static int octeon_device_init(struct octeon_device *octeon_dev) { int j, ret; + int fw_loaded = 0; char bootcmd[] = "\n"; struct octeon_device_priv *oct_priv = (struct octeon_device_priv *)octeon_dev->priv; @@ -3901,9 +4131,23 @@ static int octeon_device_init(struct octeon_device *octeon_dev) octeon_dev->app_mode = CVM_DRV_INVALID_APP; - /* Do a soft reset of the Octeon device. */ - if (octeon_dev->fn_list.soft_reset(octeon_dev)) + if (OCTEON_CN23XX_PF(octeon_dev)) { + if (!cn23xx_fw_loaded(octeon_dev)) { + fw_loaded = 0; + /* Do a soft reset of the Octeon device. */ + if (octeon_dev->fn_list.soft_reset(octeon_dev)) + return 1; + /* things might have changed */ + if (!cn23xx_fw_loaded(octeon_dev)) + fw_loaded = 0; + else + fw_loaded = 1; + } else { + fw_loaded = 1; + } + } else if (octeon_dev->fn_list.soft_reset(octeon_dev)) { return 1; + } /* Initialize the dispatch mechanism used to push packets arriving on * Octeon Output queues. @@ -3925,6 +4169,22 @@ static int octeon_device_init(struct octeon_device *octeon_dev) octeon_set_io_queues_off(octeon_dev); + if (OCTEON_CN23XX_PF(octeon_dev)) { + ret = octeon_dev->fn_list.setup_device_regs(octeon_dev); + if (ret) { + dev_err(&octeon_dev->pci_dev->dev, "OCTEON: Failed to configure device registers\n"); + return ret; + } + } + + /* Initialize soft command buffer pool + */ + if (octeon_setup_sc_buffer_pool(octeon_dev)) { + dev_err(&octeon_dev->pci_dev->dev, "sc buffer pool allocation failed\n"); + return 1; + } + atomic_set(&octeon_dev->status, OCT_DEV_SC_BUFF_POOL_INIT_DONE); + /* Setup the data structures that manage this Octeon's Input queues. */ if (octeon_setup_instr_queues(octeon_dev)) { dev_err(&octeon_dev->pci_dev->dev, @@ -3936,14 +4196,6 @@ static int octeon_device_init(struct octeon_device *octeon_dev) } atomic_set(&octeon_dev->status, OCT_DEV_INSTR_QUEUE_INIT_DONE); - /* Initialize soft command buffer pool - */ - if (octeon_setup_sc_buffer_pool(octeon_dev)) { - dev_err(&octeon_dev->pci_dev->dev, "sc buffer pool allocation failed\n"); - return 1; - } - atomic_set(&octeon_dev->status, OCT_DEV_SC_BUFF_POOL_INIT_DONE); - /* Initialize lists to manage the requests of different types that * arrive from user & kernel applications for this octeon device. */ @@ -3963,15 +4215,23 @@ static int octeon_device_init(struct octeon_device *octeon_dev) atomic_set(&octeon_dev->status, OCT_DEV_DROQ_INIT_DONE); - /* The input and output queue registers were setup earlier (the queues - * were not enabled). Any additional registers that need to be - * programmed should be done now. - */ - ret = octeon_dev->fn_list.setup_device_regs(octeon_dev); - if (ret) { - dev_err(&octeon_dev->pci_dev->dev, - "Failed to configure device registers\n"); - return ret; + if (OCTEON_CN23XX_PF(octeon_dev)) { + if (octeon_allocate_ioq_vector(octeon_dev)) { + dev_err(&octeon_dev->pci_dev->dev, "OCTEON: ioq vector allocation failed\n"); + return 1; + } + + } else { + /* The input and output queue registers were setup earlier (the + * queues were not enabled). Any additional registers + * that need to be programmed should be done now. + */ + ret = octeon_dev->fn_list.setup_device_regs(octeon_dev); + if (ret) { + dev_err(&octeon_dev->pci_dev->dev, + "Failed to configure device registers\n"); + return ret; + } } /* Initialize the tasklet that handles output queue packet processing.*/ @@ -3985,63 +4245,76 @@ static int octeon_device_init(struct octeon_device *octeon_dev) return 1; /* Enable Octeon device interrupts */ - octeon_dev->fn_list.enable_interrupt(octeon_dev->chip); + octeon_dev->fn_list.enable_interrupt(octeon_dev, OCTEON_ALL_INTR); /* Enable the input and output queues for this Octeon device */ - octeon_dev->fn_list.enable_io_queues(octeon_dev); + ret = octeon_dev->fn_list.enable_io_queues(octeon_dev); + if (ret) { + dev_err(&octeon_dev->pci_dev->dev, "Failed to enable input/output queues"); + return ret; + } atomic_set(&octeon_dev->status, OCT_DEV_IO_QUEUES_DONE); - dev_dbg(&octeon_dev->pci_dev->dev, "Waiting for DDR initialization...\n"); - - if (ddr_timeout == 0) - dev_info(&octeon_dev->pci_dev->dev, "WAITING. Set ddr_timeout to non-zero value to proceed with initialization.\n"); + if ((!OCTEON_CN23XX_PF(octeon_dev)) || !fw_loaded) { + dev_dbg(&octeon_dev->pci_dev->dev, "Waiting for DDR initialization...\n"); + if (!ddr_timeout) { + dev_info(&octeon_dev->pci_dev->dev, + "WAITING. Set ddr_timeout to non-zero value to proceed with initialization.\n"); + } - schedule_timeout_uninterruptible(HZ * LIO_RESET_SECS); + schedule_timeout_uninterruptible(HZ * LIO_RESET_SECS); - /* Wait for the octeon to initialize DDR after the soft-reset. */ - while (ddr_timeout == 0) { - set_current_state(TASK_INTERRUPTIBLE); - if (schedule_timeout(HZ / 10)) { - /* user probably pressed Control-C */ + /* Wait for the octeon to initialize DDR after the soft-reset.*/ + while (!ddr_timeout) { + set_current_state(TASK_INTERRUPTIBLE); + if (schedule_timeout(HZ / 10)) { + /* user probably pressed Control-C */ + return 1; + } + } + ret = octeon_wait_for_ddr_init(octeon_dev, &ddr_timeout); + if (ret) { + dev_err(&octeon_dev->pci_dev->dev, + "DDR not initialized. Please confirm that board is configured to boot from Flash, ret: %d\n", + ret); return 1; } - } - ret = octeon_wait_for_ddr_init(octeon_dev, &ddr_timeout); - if (ret) { - dev_err(&octeon_dev->pci_dev->dev, - "DDR not initialized. Please confirm that board is configured to boot from Flash, ret: %d\n", - ret); - return 1; - } - if (octeon_wait_for_bootloader(octeon_dev, 1000) != 0) { - dev_err(&octeon_dev->pci_dev->dev, "Board not responding\n"); - return 1; - } + if (octeon_wait_for_bootloader(octeon_dev, 1000)) { + dev_err(&octeon_dev->pci_dev->dev, "Board not responding\n"); + return 1; + } - /* Divert uboot to take commands from host instead. */ - ret = octeon_console_send_cmd(octeon_dev, bootcmd, 50); + /* Divert uboot to take commands from host instead. */ + ret = octeon_console_send_cmd(octeon_dev, bootcmd, 50); - dev_dbg(&octeon_dev->pci_dev->dev, "Initializing consoles\n"); - ret = octeon_init_consoles(octeon_dev); - if (ret) { - dev_err(&octeon_dev->pci_dev->dev, "Could not access board consoles\n"); - return 1; - } - ret = octeon_add_console(octeon_dev, 0); - if (ret) { - dev_err(&octeon_dev->pci_dev->dev, "Could not access board console\n"); - return 1; - } + dev_dbg(&octeon_dev->pci_dev->dev, "Initializing consoles\n"); + ret = octeon_init_consoles(octeon_dev); + if (ret) { + dev_err(&octeon_dev->pci_dev->dev, "Could not access board consoles\n"); + return 1; + } + ret = octeon_add_console(octeon_dev, 0); + if (ret) { + dev_err(&octeon_dev->pci_dev->dev, "Could not access board console\n"); + return 1; + } - atomic_set(&octeon_dev->status, OCT_DEV_CONSOLE_INIT_DONE); + atomic_set(&octeon_dev->status, OCT_DEV_CONSOLE_INIT_DONE); - dev_dbg(&octeon_dev->pci_dev->dev, "Loading firmware\n"); - ret = load_firmware(octeon_dev); - if (ret) { - dev_err(&octeon_dev->pci_dev->dev, "Could not load firmware to board\n"); - return 1; + dev_dbg(&octeon_dev->pci_dev->dev, "Loading firmware\n"); + ret = load_firmware(octeon_dev); + if (ret) { + dev_err(&octeon_dev->pci_dev->dev, "Could not load firmware to board\n"); + return 1; + } + /* set bit 1 of SLI_SCRATCH_1 to indicate that firmware is + * loaded + */ + if (OCTEON_CN23XX_PF(octeon_dev)) + octeon_write_csr64(octeon_dev, CN23XX_SLI_SCRATCH1, + 2ULL); } handshake[octeon_dev->octeon_id].init_ok = 1; @@ -4057,7 +4330,6 @@ static int octeon_device_init(struct octeon_device *octeon_dev) octeon_dev->droq[j]->pkts_credit_reg); /* Packets can start arriving on the output queues from this point. */ - return 0; } diff --git a/drivers/net/ethernet/cavium/liquidio/liquidio_common.h b/drivers/net/ethernet/cavium/liquidio/liquidio_common.h index 199a8b9c7dc5..0d990accb65e 100644 --- a/drivers/net/ethernet/cavium/liquidio/liquidio_common.h +++ b/drivers/net/ethernet/cavium/liquidio/liquidio_common.h @@ -30,10 +30,24 @@ #include "octeon_config.h" -#define LIQUIDIO_BASE_VERSION "1.4" -#define LIQUIDIO_MICRO_VERSION ".1" #define LIQUIDIO_PACKAGE "" -#define LIQUIDIO_VERSION "1.4.1" +#define LIQUIDIO_BASE_MAJOR_VERSION 1 +#define LIQUIDIO_BASE_MINOR_VERSION 4 +#define LIQUIDIO_BASE_MICRO_VERSION 1 +#define LIQUIDIO_BASE_VERSION __stringify(LIQUIDIO_BASE_MAJOR_VERSION) "." \ + __stringify(LIQUIDIO_BASE_MINOR_VERSION) +#define LIQUIDIO_MICRO_VERSION "." __stringify(LIQUIDIO_BASE_MICRO_VERSION) +#define LIQUIDIO_VERSION LIQUIDIO_PACKAGE \ + __stringify(LIQUIDIO_BASE_MAJOR_VERSION) "." \ + __stringify(LIQUIDIO_BASE_MINOR_VERSION) \ + "." __stringify(LIQUIDIO_BASE_MICRO_VERSION) + +struct lio_version { + u16 major; + u16 minor; + u16 micro; + u16 reserved; +}; #define CONTROL_IQ 0 /** Tag types used by Octeon cores in its work. */ @@ -218,6 +232,9 @@ static inline void add_sg_size(struct octeon_sg_entry *sg_entry, #define OCTNET_CMD_ADD_VLAN_FILTER 0x17 #define OCTNET_CMD_DEL_VLAN_FILTER 0x18 #define OCTNET_CMD_VXLAN_PORT_CONFIG 0x19 + +#define OCTNET_CMD_ID_ACTIVE 0x1a + #define OCTNET_CMD_VXLAN_PORT_ADD 0x0 #define OCTNET_CMD_VXLAN_PORT_DEL 0x1 #define OCTNET_CMD_RXCSUM_ENABLE 0x0 @@ -296,6 +313,13 @@ union octnet_cmd { #define OCTNET_CMD_SIZE (sizeof(union octnet_cmd)) +/*pkiih3 + irh + ossp[0] + ossp[1] + rdp + rptr = 40 bytes */ +#define LIO_SOFTCMDRESP_IH2 40 +#define LIO_SOFTCMDRESP_IH3 (40 + 8) + +#define LIO_PCICMD_O2 24 +#define LIO_PCICMD_O3 (24 + 8) + /* Instruction Header(DPI) - for OCTEON-III models */ struct octeon_instr_ih3 { #ifdef __BIG_ENDIAN_BITFIELD @@ -814,6 +838,8 @@ struct oct_link_stats { #define VITESSE_PHY_GPIO_DRIVEOFF 0x4 #define VITESSE_PHY_GPIO_HIGH 0x2 #define VITESSE_PHY_GPIO_LOW 0x3 +#define LED_IDENTIFICATION_ON 0x1 +#define LED_IDENTIFICATION_OFF 0x0 struct oct_mdio_cmd { u64 op; @@ -832,7 +858,7 @@ struct oct_mdio_cmd { /* intrmod: max. packets to trigger interrupt */ #define LIO_INTRMOD_RXMAXCNT_TRIGGER 384 /* intrmod: min. packets to trigger interrupt */ -#define LIO_INTRMOD_RXMINCNT_TRIGGER 1 +#define LIO_INTRMOD_RXMINCNT_TRIGGER 0 /* intrmod: max. time to trigger interrupt */ #define LIO_INTRMOD_RXMAXTMR_TRIGGER 128 /* 66xx:intrmod: min. time to trigger interrupt diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_config.h b/drivers/net/ethernet/cavium/liquidio/octeon_config.h index b3396e3a8bab..c76556809ed1 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_config.h +++ b/drivers/net/ethernet/cavium/liquidio/octeon_config.h @@ -64,6 +64,34 @@ #define DEFAULT_NUM_NIC_PORTS_68XX 4 #define DEFAULT_NUM_NIC_PORTS_68XX_210NV 2 +/* CN23xx IQ configuration macros */ +#define CN23XX_MAX_RINGS_PER_PF_PASS_1_0 12 +#define CN23XX_MAX_RINGS_PER_PF_PASS_1_1 32 +#define CN23XX_MAX_RINGS_PER_PF 64 + +#define CN23XX_MAX_INPUT_QUEUES CN23XX_MAX_RINGS_PER_PF +#define CN23XX_MAX_IQ_DESCRIPTORS 2048 +#define CN23XX_DB_MIN 1 +#define CN23XX_DB_MAX 8 +#define CN23XX_DB_TIMEOUT 1 + +#define CN23XX_MAX_OUTPUT_QUEUES CN23XX_MAX_RINGS_PER_PF +#define CN23XX_MAX_OQ_DESCRIPTORS 2048 +#define CN23XX_OQ_BUF_SIZE 1536 +#define CN23XX_OQ_PKTSPER_INTR 128 +/*#define CAVIUM_ONLY_CN23XX_RX_PERF*/ +#define CN23XX_OQ_REFIL_THRESHOLD 128 + +#define CN23XX_OQ_INTR_PKT 64 +#define CN23XX_OQ_INTR_TIME 100 +#define DEFAULT_NUM_NIC_PORTS_23XX 1 + +#define CN23XX_CFG_IO_QUEUES CN23XX_MAX_RINGS_PER_PF +/* PEMs count */ +#define CN23XX_MAX_MACS 4 + +#define CN23XX_DEF_IQ_INTR_THRESHOLD 32 +#define CN23XX_DEF_IQ_INTR_BYTE_THRESHOLD (64 * 1024) /* common OCTEON configuration macros */ #define CN6XXX_CFG_IO_QUEUES 32 #define OCTEON_32BYTE_INSTR 32 @@ -92,6 +120,9 @@ #define CFG_GET_IQ_DB_MIN(cfg) ((cfg)->iq.db_min) #define CFG_GET_IQ_DB_TIMEOUT(cfg) ((cfg)->iq.db_timeout) +#define CFG_GET_IQ_INTR_PKT(cfg) ((cfg)->iq.iq_intr_pkt) +#define CFG_SET_IQ_INTR_PKT(cfg, val) (cfg)->iq.iq_intr_pkt = val + #define CFG_GET_OQ_MAX_Q(cfg) ((cfg)->oq.max_oqs) #define CFG_GET_OQ_INFO_PTR(cfg) ((cfg)->oq.info_ptr) #define CFG_GET_OQ_PKTS_PER_INTR(cfg) ((cfg)->oq.pkts_per_intr) @@ -140,19 +171,24 @@ enum lio_card_type { LIO_210SV = 0, /* Two port, 66xx */ LIO_210NV, /* Two port, 68xx */ - LIO_410NV /* Four port, 68xx */ + LIO_410NV, /* Four port, 68xx */ + LIO_23XX /* 23xx */ }; #define LIO_210SV_NAME "210sv" #define LIO_210NV_NAME "210nv" #define LIO_410NV_NAME "410nv" +#define LIO_23XX_NAME "23xx" /** Structure to define the configuration attributes for each Input queue. * Applicable to all Octeon processors **/ struct octeon_iq_config { #ifdef __BIG_ENDIAN_BITFIELD - u64 reserved:32; + u64 reserved:16; + + /** Tx interrupt packets. Applicable to 23xx only */ + u64 iq_intr_pkt:16; /** Minimum ticks to wait before checking for pending instructions. */ u64 db_timeout:16; @@ -192,7 +228,10 @@ struct octeon_iq_config { /** Minimum ticks to wait before checking for pending instructions. */ u64 db_timeout:16; - u64 reserved:32; + /** Tx interrupt packets. Applicable to 23xx only */ + u64 iq_intr_pkt:16; + + u64 reserved:16; #endif }; @@ -416,11 +455,15 @@ struct octeon_config { #define DISPATCH_LIST_SIZE BIT(OPCODE_MASK_BITS) /* Maximum number of Octeon Instruction (command) queues */ -#define MAX_OCTEON_INSTR_QUEUES(oct) CN6XXX_MAX_INPUT_QUEUES -/* Maximum number of Octeon Output queues */ -#define MAX_OCTEON_OUTPUT_QUEUES(oct) CN6XXX_MAX_OUTPUT_QUEUES +#define MAX_OCTEON_INSTR_QUEUES(oct) \ + (OCTEON_CN23XX_PF(oct) ? CN23XX_MAX_INPUT_QUEUES : \ + CN6XXX_MAX_INPUT_QUEUES) -#define MAX_POSSIBLE_OCTEON_INSTR_QUEUES CN6XXX_MAX_INPUT_QUEUES -#define MAX_POSSIBLE_OCTEON_OUTPUT_QUEUES CN6XXX_MAX_OUTPUT_QUEUES +/* Maximum number of Octeon Instruction (command) queues */ +#define MAX_OCTEON_OUTPUT_QUEUES(oct) \ + (OCTEON_CN23XX_PF(oct) ? CN23XX_MAX_OUTPUT_QUEUES : \ + CN6XXX_MAX_OUTPUT_QUEUES) +#define MAX_POSSIBLE_OCTEON_INSTR_QUEUES CN23XX_MAX_INPUT_QUEUES +#define MAX_POSSIBLE_OCTEON_OUTPUT_QUEUES CN23XX_MAX_OUTPUT_QUEUES #endif /* __OCTEON_CONFIG_H__ */ diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_console.c b/drivers/net/ethernet/cavium/liquidio/octeon_console.c index bbb50ea66f16..01a50f3b0c8e 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_console.c +++ b/drivers/net/ethernet/cavium/liquidio/octeon_console.c @@ -25,12 +25,13 @@ */ #include <linux/pci.h> #include <linux/netdevice.h> +#include <linux/crc32.h> #include "liquidio_common.h" #include "octeon_droq.h" #include "octeon_iq.h" #include "response_manager.h" #include "octeon_device.h" -#include "octeon_main.h" +#include "liquidio_image.h" #include "octeon_mem_ops.h" static void octeon_remote_lock(void); @@ -40,6 +41,10 @@ static u64 cvmx_bootmem_phy_named_block_find(struct octeon_device *oct, u32 flags); static int octeon_console_read(struct octeon_device *oct, u32 console_num, char *buffer, u32 buf_size); +static u32 console_bitmask; +module_param(console_bitmask, int, 0644); +MODULE_PARM_DESC(console_bitmask, + "Bitmask indicating which consoles have debug output redirected to syslog."); #define MIN(a, b) min((a), (b)) #define CAST_ULL(v) ((u64)(v)) @@ -177,6 +182,15 @@ struct octeon_pci_console_desc { __cvmx_bootmem_desc_get(oct, addr, \ offsetof(struct cvmx_bootmem_named_block_desc, field), \ SIZEOF_FIELD(struct cvmx_bootmem_named_block_desc, field)) +/** + * \brief determines if a given console has debug enabled. + * @param console console to check + * @returns 1 = enabled. 0 otherwise + */ +static int octeon_console_debug_enabled(u32 console) +{ + return (console_bitmask >> (console)) & 0x1; +} /** * This function is the implementation of the get macros defined @@ -709,3 +723,104 @@ static int octeon_console_read(struct octeon_device *oct, u32 console_num, return bytes_to_read; } + +#define FBUF_SIZE (4 * 1024 * 1024) +u8 fbuf[FBUF_SIZE]; + +int octeon_download_firmware(struct octeon_device *oct, const u8 *data, + size_t size) +{ + int ret = 0; + u8 *p = fbuf; + u32 crc32_result; + u64 load_addr; + u32 image_len; + struct octeon_firmware_file_header *h; + u32 i, rem; + + if (size < sizeof(struct octeon_firmware_file_header)) { + dev_err(&oct->pci_dev->dev, "Firmware file too small (%d < %d).\n", + (u32)size, + (u32)sizeof(struct octeon_firmware_file_header)); + return -EINVAL; + } + + h = (struct octeon_firmware_file_header *)data; + + if (be32_to_cpu(h->magic) != LIO_NIC_MAGIC) { + dev_err(&oct->pci_dev->dev, "Unrecognized firmware file.\n"); + return -EINVAL; + } + + crc32_result = crc32((unsigned int)~0, data, + sizeof(struct octeon_firmware_file_header) - + sizeof(u32)) ^ ~0U; + if (crc32_result != be32_to_cpu(h->crc32)) { + dev_err(&oct->pci_dev->dev, "Firmware CRC mismatch (0x%08x != 0x%08x).\n", + crc32_result, be32_to_cpu(h->crc32)); + return -EINVAL; + } + + if (strncmp(LIQUIDIO_PACKAGE, h->version, strlen(LIQUIDIO_PACKAGE))) { + dev_err(&oct->pci_dev->dev, "Unmatched firmware package type. Expected %s, got %s.\n", + LIQUIDIO_PACKAGE, h->version); + return -EINVAL; + } + + if (memcmp(LIQUIDIO_BASE_VERSION, h->version + strlen(LIQUIDIO_PACKAGE), + strlen(LIQUIDIO_BASE_VERSION))) { + dev_err(&oct->pci_dev->dev, "Unmatched firmware version. Expected %s.x, got %s.\n", + LIQUIDIO_BASE_VERSION, + h->version + strlen(LIQUIDIO_PACKAGE)); + return -EINVAL; + } + + if (be32_to_cpu(h->num_images) > LIO_MAX_IMAGES) { + dev_err(&oct->pci_dev->dev, "Too many images in firmware file (%d).\n", + be32_to_cpu(h->num_images)); + return -EINVAL; + } + + dev_info(&oct->pci_dev->dev, "Firmware version: %s\n", h->version); + snprintf(oct->fw_info.liquidio_firmware_version, 32, "LIQUIDIO: %s", + h->version); + + data += sizeof(struct octeon_firmware_file_header); + + dev_info(&oct->pci_dev->dev, "%s: Loading %d images\n", __func__, + be32_to_cpu(h->num_images)); + /* load all images */ + for (i = 0; i < be32_to_cpu(h->num_images); i++) { + load_addr = be64_to_cpu(h->desc[i].addr); + image_len = be32_to_cpu(h->desc[i].len); + + dev_info(&oct->pci_dev->dev, "Loading firmware %d at %llx\n", + image_len, load_addr); + + /* Write in 4MB chunks*/ + rem = image_len; + + while (rem) { + if (rem < FBUF_SIZE) + size = rem; + else + size = FBUF_SIZE; + + memcpy(p, data, size); + + /* download the image */ + octeon_pci_write_core_mem(oct, load_addr, p, (u32)size); + + data += size; + rem -= (u32)size; + load_addr += size; + } + } + dev_info(&oct->pci_dev->dev, "Writing boot command: %s\n", + h->bootcmd); + + /* Invoke the bootcmd */ + ret = octeon_console_send_cmd(oct, h->bootcmd, 50); + + return 0; +} diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_device.c b/drivers/net/ethernet/cavium/liquidio/octeon_device.c index 0eb504a4379a..586b68899b06 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_device.c +++ b/drivers/net/ethernet/cavium/liquidio/octeon_device.c @@ -20,7 +20,6 @@ * Contact Cavium, Inc. for more information **********************************************************************/ #include <linux/pci.h> -#include <linux/crc32.h> #include <linux/netdevice.h> #include <linux/vmalloc.h> #include "liquidio_common.h" @@ -32,8 +31,7 @@ #include "octeon_network.h" #include "cn66xx_regs.h" #include "cn66xx_device.h" -#include "liquidio_image.h" -#include "octeon_mem_ops.h" +#include "cn23xx_pf_device.h" /** Default configuration * for CN66XX OCTEON Models. @@ -420,6 +418,108 @@ static struct octeon_config default_cn68xx_210nv_conf = { , }; +static struct octeon_config default_cn23xx_conf = { + .card_type = LIO_23XX, + .card_name = LIO_23XX_NAME, + /** IQ attributes */ + .iq = { + .max_iqs = CN23XX_CFG_IO_QUEUES, + .pending_list_size = (CN23XX_MAX_IQ_DESCRIPTORS * + CN23XX_CFG_IO_QUEUES), + .instr_type = OCTEON_64BYTE_INSTR, + .db_min = CN23XX_DB_MIN, + .db_timeout = CN23XX_DB_TIMEOUT, + .iq_intr_pkt = CN23XX_DEF_IQ_INTR_THRESHOLD, + }, + + /** OQ attributes */ + .oq = { + .max_oqs = CN23XX_CFG_IO_QUEUES, + .info_ptr = OCTEON_OQ_INFOPTR_MODE, + .pkts_per_intr = CN23XX_OQ_PKTSPER_INTR, + .refill_threshold = CN23XX_OQ_REFIL_THRESHOLD, + .oq_intr_pkt = CN23XX_OQ_INTR_PKT, + .oq_intr_time = CN23XX_OQ_INTR_TIME, + }, + + .num_nic_ports = DEFAULT_NUM_NIC_PORTS_23XX, + .num_def_rx_descs = CN23XX_MAX_OQ_DESCRIPTORS, + .num_def_tx_descs = CN23XX_MAX_IQ_DESCRIPTORS, + .def_rx_buf_size = CN23XX_OQ_BUF_SIZE, + + /* For ethernet interface 0: Port cfg Attributes */ + .nic_if_cfg[0] = { + /* Max Txqs: Half for each of the two ports :max_iq/2 */ + .max_txqs = MAX_TXQS_PER_INTF, + + /* Actual configured value. Range could be: 1...max_txqs */ + .num_txqs = DEF_TXQS_PER_INTF, + + /* Max Rxqs: Half for each of the two ports :max_oq/2 */ + .max_rxqs = MAX_RXQS_PER_INTF, + + /* Actual configured value. Range could be: 1...max_rxqs */ + .num_rxqs = DEF_RXQS_PER_INTF, + + /* Num of desc for rx rings */ + .num_rx_descs = CN23XX_MAX_OQ_DESCRIPTORS, + + /* Num of desc for tx rings */ + .num_tx_descs = CN23XX_MAX_IQ_DESCRIPTORS, + + /* SKB size, We need not change buf size even for Jumbo frames. + * Octeon can send jumbo frames in 4 consecutive descriptors, + */ + .rx_buf_size = CN23XX_OQ_BUF_SIZE, + + .base_queue = BASE_QUEUE_NOT_REQUESTED, + + .gmx_port_id = 0, + }, + + .nic_if_cfg[1] = { + /* Max Txqs: Half for each of the two ports :max_iq/2 */ + .max_txqs = MAX_TXQS_PER_INTF, + + /* Actual configured value. Range could be: 1...max_txqs */ + .num_txqs = DEF_TXQS_PER_INTF, + + /* Max Rxqs: Half for each of the two ports :max_oq/2 */ + .max_rxqs = MAX_RXQS_PER_INTF, + + /* Actual configured value. Range could be: 1...max_rxqs */ + .num_rxqs = DEF_RXQS_PER_INTF, + + /* Num of desc for rx rings */ + .num_rx_descs = CN23XX_MAX_OQ_DESCRIPTORS, + + /* Num of desc for tx rings */ + .num_tx_descs = CN23XX_MAX_IQ_DESCRIPTORS, + + /* SKB size, We need not change buf size even for Jumbo frames. + * Octeon can send jumbo frames in 4 consecutive descriptors, + */ + .rx_buf_size = CN23XX_OQ_BUF_SIZE, + + .base_queue = BASE_QUEUE_NOT_REQUESTED, + + .gmx_port_id = 1, + }, + + .misc = { + /* Host driver link query interval */ + .oct_link_query_interval = 100, + + /* Octeon link query interval */ + .host_link_query_interval = 500, + + .enable_sli_oq_bp = 0, + + /* Control queue group */ + .ctrlq_grp = 1, + } +}; + enum { OCTEON_CONFIG_TYPE_DEFAULT = 0, NUM_OCTEON_CONFS, @@ -487,6 +587,8 @@ static void *__retrieve_octeon_config_info(struct octeon_device *oct, } else if ((oct->chip_id == OCTEON_CN68XX) && (card_type == LIO_410NV)) { ret = (void *)&default_cn68xx_conf; + } else if (oct->chip_id == OCTEON_CN23XX_PF_VID) { + ret = (void *)&default_cn23xx_conf; } break; default: @@ -501,7 +603,8 @@ static int __verify_octeon_config_info(struct octeon_device *oct, void *conf) case OCTEON_CN66XX: case OCTEON_CN68XX: return lio_validate_cn6xxx_config_info(oct, conf); - + case OCTEON_CN23XX_PF_VID: + return 0; default: break; } @@ -541,107 +644,6 @@ static char *get_oct_app_string(u32 app_mode) return oct_dev_app_str[CVM_DRV_INVALID_APP - CVM_DRV_APP_START]; } -u8 fbuf[4 * 1024 * 1024]; - -int octeon_download_firmware(struct octeon_device *oct, const u8 *data, - size_t size) -{ - int ret = 0; - u8 *p = fbuf; - u32 crc32_result; - u64 load_addr; - u32 image_len; - struct octeon_firmware_file_header *h; - u32 i, rem, base_len = strlen(LIQUIDIO_BASE_VERSION); - char *base; - - if (size < sizeof(struct octeon_firmware_file_header)) { - dev_err(&oct->pci_dev->dev, "Firmware file too small (%d < %d).\n", - (u32)size, - (u32)sizeof(struct octeon_firmware_file_header)); - return -EINVAL; - } - - h = (struct octeon_firmware_file_header *)data; - - if (be32_to_cpu(h->magic) != LIO_NIC_MAGIC) { - dev_err(&oct->pci_dev->dev, "Unrecognized firmware file.\n"); - return -EINVAL; - } - - crc32_result = crc32((unsigned int)~0, data, - sizeof(struct octeon_firmware_file_header) - - sizeof(u32)) ^ ~0U; - if (crc32_result != be32_to_cpu(h->crc32)) { - dev_err(&oct->pci_dev->dev, "Firmware CRC mismatch (0x%08x != 0x%08x).\n", - crc32_result, be32_to_cpu(h->crc32)); - return -EINVAL; - } - - if (strncmp(LIQUIDIO_PACKAGE, h->version, strlen(LIQUIDIO_PACKAGE))) { - dev_err(&oct->pci_dev->dev, "Unmatched firmware package type. Expected %s, got %s.\n", - LIQUIDIO_PACKAGE, h->version); - return -EINVAL; - } - - base = h->version + strlen(LIQUIDIO_PACKAGE); - ret = memcmp(LIQUIDIO_BASE_VERSION, base, base_len); - if (ret) { - dev_err(&oct->pci_dev->dev, "Unmatched firmware version. Expected %s.x, got %s.\n", - LIQUIDIO_BASE_VERSION, base); - return -EINVAL; - } - - if (be32_to_cpu(h->num_images) > LIO_MAX_IMAGES) { - dev_err(&oct->pci_dev->dev, "Too many images in firmware file (%d).\n", - be32_to_cpu(h->num_images)); - return -EINVAL; - } - - dev_info(&oct->pci_dev->dev, "Firmware version: %s\n", h->version); - snprintf(oct->fw_info.liquidio_firmware_version, 32, "LIQUIDIO: %s", - h->version); - - data += sizeof(struct octeon_firmware_file_header); - - dev_info(&oct->pci_dev->dev, "%s: Loading %d images\n", __func__, - be32_to_cpu(h->num_images)); - /* load all images */ - for (i = 0; i < be32_to_cpu(h->num_images); i++) { - load_addr = be64_to_cpu(h->desc[i].addr); - image_len = be32_to_cpu(h->desc[i].len); - - dev_info(&oct->pci_dev->dev, "Loading firmware %d at %llx\n", - image_len, load_addr); - - /* Write in 4MB chunks*/ - rem = image_len; - - while (rem) { - if (rem < (4 * 1024 * 1024)) - size = rem; - else - size = 4 * 1024 * 1024; - - memcpy(p, data, size); - - /* download the image */ - octeon_pci_write_core_mem(oct, load_addr, p, (u32)size); - - data += size; - rem -= (u32)size; - load_addr += size; - } - } - dev_info(&oct->pci_dev->dev, "Writing boot command: %s\n", - h->bootcmd); - - /* Invoke the bootcmd */ - ret = octeon_console_send_cmd(oct, h->bootcmd, 50); - - return 0; -} - void octeon_free_device_mem(struct octeon_device *oct) { int i; @@ -676,6 +678,9 @@ static struct octeon_device *octeon_allocate_device_mem(u32 pci_id, configsize = sizeof(struct octeon_cn6xxx); break; + case OCTEON_CN23XX_PF_VID: + configsize = sizeof(struct octeon_cn23xx_pf); + break; default: pr_err("%s: Unknown PCI Device: 0x%x\n", __func__, @@ -741,6 +746,45 @@ struct octeon_device *octeon_allocate_device(u32 pci_id, return oct; } +int +octeon_allocate_ioq_vector(struct octeon_device *oct) +{ + int i, num_ioqs = 0; + struct octeon_ioq_vector *ioq_vector; + int cpu_num; + int size; + + if (OCTEON_CN23XX_PF(oct)) + num_ioqs = oct->sriov_info.num_pf_rings; + size = sizeof(struct octeon_ioq_vector) * num_ioqs; + + oct->ioq_vector = vmalloc(size); + if (!oct->ioq_vector) + return 1; + memset(oct->ioq_vector, 0, size); + for (i = 0; i < num_ioqs; i++) { + ioq_vector = &oct->ioq_vector[i]; + ioq_vector->oct_dev = oct; + ioq_vector->iq_index = i; + ioq_vector->droq_index = i; + + cpu_num = i % num_online_cpus(); + cpumask_set_cpu(cpu_num, &ioq_vector->affinity_mask); + + if (oct->chip_id == OCTEON_CN23XX_PF_VID) + ioq_vector->ioq_num = i + oct->sriov_info.pf_srn; + else + ioq_vector->ioq_num = i; + } + return 0; +} + +void +octeon_free_ioq_vector(struct octeon_device *oct) +{ + vfree(oct->ioq_vector); +} + /* this function is only for setting up the first queue */ int octeon_setup_instr_queues(struct octeon_device *oct) { @@ -749,10 +793,12 @@ int octeon_setup_instr_queues(struct octeon_device *oct) union oct_txpciq txpciq; int numa_node = cpu_to_node(iq_no % num_online_cpus()); - /* this causes queue 0 to be default queue */ if (OCTEON_CN6XXX(oct)) num_descs = CFG_GET_NUM_DEF_TX_DESCS(CHIP_FIELD(oct, cn6xxx, conf)); + else if (OCTEON_CN23XX_PF(oct)) + num_descs = CFG_GET_NUM_DEF_TX_DESCS(CHIP_FIELD(oct, cn23xx_pf, + conf)); oct->num_iqs = 0; @@ -769,6 +815,7 @@ int octeon_setup_instr_queues(struct octeon_device *oct) oct->instr_queue[0]->ifidx = 0; txpciq.u64 = 0; txpciq.s.q_no = iq_no; + txpciq.s.pkind = oct->pfvf_hsword.pkind; txpciq.s.use_qpg = 0; txpciq.s.qpg = 0; if (octeon_init_instr_queue(oct, txpciq, num_descs)) { @@ -788,14 +835,17 @@ int octeon_setup_output_queues(struct octeon_device *oct) u32 oq_no = 0; int numa_node = cpu_to_node(oq_no % num_online_cpus()); - /* this causes queue 0 to be default queue */ if (OCTEON_CN6XXX(oct)) { num_descs = CFG_GET_NUM_DEF_RX_DESCS(CHIP_FIELD(oct, cn6xxx, conf)); desc_size = CFG_GET_DEF_RX_BUF_SIZE(CHIP_FIELD(oct, cn6xxx, conf)); + } else if (OCTEON_CN23XX_PF(oct)) { + num_descs = CFG_GET_NUM_DEF_RX_DESCS(CHIP_FIELD(oct, cn23xx_pf, + conf)); + desc_size = CFG_GET_DEF_RX_BUF_SIZE(CHIP_FIELD(oct, cn23xx_pf, + conf)); } - oct->num_oqs = 0; oct->droq[0] = vmalloc_node(sizeof(*oct->droq[0]), numa_node); if (!oct->droq[0]) @@ -812,10 +862,10 @@ int octeon_setup_output_queues(struct octeon_device *oct) void octeon_set_io_queues_off(struct octeon_device *oct) { - /* Disable the i/p and o/p queues for this Octeon. */ - - octeon_write_csr(oct, CN6XXX_SLI_PKT_INSTR_ENB, 0); - octeon_write_csr(oct, CN6XXX_SLI_PKT_OUT_ENB, 0); + if (OCTEON_CN6XXX(oct)) { + octeon_write_csr(oct, CN6XXX_SLI_PKT_INSTR_ENB, 0); + octeon_write_csr(oct, CN6XXX_SLI_PKT_OUT_ENB, 0); + } } void octeon_set_droq_pkt_op(struct octeon_device *oct, @@ -825,14 +875,16 @@ void octeon_set_droq_pkt_op(struct octeon_device *oct, u32 reg_val = 0; /* Disable the i/p and o/p queues for this Octeon. */ - reg_val = octeon_read_csr(oct, CN6XXX_SLI_PKT_OUT_ENB); + if (OCTEON_CN6XXX(oct)) { + reg_val = octeon_read_csr(oct, CN6XXX_SLI_PKT_OUT_ENB); - if (enable) - reg_val = reg_val | (1 << q_no); - else - reg_val = reg_val & (~(1 << q_no)); + if (enable) + reg_val = reg_val | (1 << q_no); + else + reg_val = reg_val & (~(1 << q_no)); - octeon_write_csr(oct, CN6XXX_SLI_PKT_OUT_ENB, reg_val); + octeon_write_csr(oct, CN6XXX_SLI_PKT_OUT_ENB, reg_val); + } } int octeon_init_dispatch_list(struct octeon_device *oct) @@ -1019,6 +1071,9 @@ int octeon_core_drv_init(struct octeon_recv_info *recv_info, void *buf) if (OCTEON_CN6XXX(oct)) num_nic_ports = CFG_GET_NUM_NIC_PORTS(CHIP_FIELD(oct, cn6xxx, conf)); + else if (OCTEON_CN23XX_PF(oct)) + num_nic_ports = + CFG_GET_NUM_NIC_PORTS(CHIP_FIELD(oct, cn23xx_pf, conf)); if (atomic_read(&oct->status) >= OCT_DEV_RUNNING) { dev_err(&oct->pci_dev->dev, "Received CORE OK when device state is 0x%x\n", @@ -1046,6 +1101,12 @@ int octeon_core_drv_init(struct octeon_recv_info *recv_info, void *buf) } oct->fw_info.app_cap_flags = recv_pkt->rh.r_core_drv_init.app_cap_flags; oct->fw_info.app_mode = (u32)recv_pkt->rh.r_core_drv_init.app_mode; + oct->pfvf_hsword.app_mode = (u32)recv_pkt->rh.r_core_drv_init.app_mode; + + oct->pfvf_hsword.pkind = recv_pkt->rh.r_core_drv_init.pkind; + + for (i = 0; i < oct->num_iqs; i++) + oct->instr_queue[i]->txpciq.s.pkind = oct->pfvf_hsword.pkind; atomic_set(&oct->status, OCT_DEV_CORE_OK); @@ -1108,8 +1169,10 @@ struct octeon_config *octeon_get_conf(struct octeon_device *oct) if (OCTEON_CN6XXX(oct)) { default_oct_conf = (struct octeon_config *)(CHIP_FIELD(oct, cn6xxx, conf)); + } else if (OCTEON_CN23XX_PF(oct)) { + default_oct_conf = (struct octeon_config *) + (CHIP_FIELD(oct, cn23xx_pf, conf)); } - return default_oct_conf; } @@ -1141,7 +1204,9 @@ u64 lio_pci_readq(struct octeon_device *oct, u64 addr) * So write MSB first */ addrhi = (addr >> 32); - if ((oct->chip_id == OCTEON_CN66XX) || (oct->chip_id == OCTEON_CN68XX)) + if ((oct->chip_id == OCTEON_CN66XX) || + (oct->chip_id == OCTEON_CN68XX) || + (oct->chip_id == OCTEON_CN23XX_PF_VID)) addrhi |= 0x00060000; writel(addrhi, oct->reg_list.pci_win_rd_addr_hi); @@ -1185,8 +1250,15 @@ int octeon_mem_access_ok(struct octeon_device *oct) u64 lmc0_reset_ctl; /* Check to make sure a DDR interface is enabled */ - lmc0_reset_ctl = lio_pci_readq(oct, CN6XXX_LMC0_RESET_CTL); - access_okay = (lmc0_reset_ctl & CN6XXX_LMC0_RESET_CTL_DDR3RST_MASK); + if (OCTEON_CN23XX_PF(oct)) { + lmc0_reset_ctl = lio_pci_readq(oct, CN23XX_LMC0_RESET_CTL); + access_okay = + (lmc0_reset_ctl & CN23XX_LMC0_RESET_CTL_DDR3RST_MASK); + } else { + lmc0_reset_ctl = lio_pci_readq(oct, CN6XXX_LMC0_RESET_CTL); + access_okay = + (lmc0_reset_ctl & CN6XXX_LMC0_RESET_CTL_DDR3RST_MASK); + } return access_okay ? 0 : 1; } @@ -1226,3 +1298,39 @@ int lio_get_device_id(void *dev) return octeon_dev->octeon_id; return -1; } + +void lio_enable_irq(struct octeon_droq *droq, struct octeon_instr_queue *iq) +{ + u64 instr_cnt; + struct octeon_device *oct = NULL; + + /* the whole thing needs to be atomic, ideally */ + if (droq) { + spin_lock_bh(&droq->lock); + writel(droq->pkt_count, droq->pkts_sent_reg); + droq->pkt_count = 0; + spin_unlock_bh(&droq->lock); + oct = droq->oct_dev; + } + if (iq) { + spin_lock_bh(&iq->lock); + writel(iq->pkt_in_done, iq->inst_cnt_reg); + iq->pkt_in_done = 0; + spin_unlock_bh(&iq->lock); + oct = iq->oct_dev; + } + /*write resend. Writing RESEND in SLI_PKTX_CNTS should be enough + *to trigger tx interrupts as well, if they are pending. + */ + if (oct && OCTEON_CN23XX_PF(oct)) { + if (droq) + writeq(CN23XX_INTR_RESEND, droq->pkts_sent_reg); + /*we race with firmrware here. read and write the IN_DONE_CNTS*/ + else if (iq) { + instr_cnt = readq(iq->inst_cnt_reg); + writeq(((instr_cnt & 0xFFFFFFFF00000000ULL) | + CN23XX_INTR_RESEND), + iq->inst_cnt_reg); + } + } +} diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_device.h b/drivers/net/ethernet/cavium/liquidio/octeon_device.h index 01edfb404346..da15c2ae9330 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_device.h +++ b/drivers/net/ethernet/cavium/liquidio/octeon_device.h @@ -30,13 +30,19 @@ /** PCI VendorId Device Id */ #define OCTEON_CN68XX_PCIID 0x91177d #define OCTEON_CN66XX_PCIID 0x92177d - +#define OCTEON_CN23XX_PCIID_PF 0x9702177d /** Driver identifies chips by these Ids, created by clubbing together * DeviceId+RevisionId; Where Revision Id is not used to distinguish * between chips, a value of 0 is used for revision id. */ #define OCTEON_CN68XX 0x0091 #define OCTEON_CN66XX 0x0092 +#define OCTEON_CN23XX_PF_VID 0x9702 + +/**RevisionId for the chips */ +#define OCTEON_CN23XX_REV_1_0 0x00 +#define OCTEON_CN23XX_REV_1_1 0x01 +#define OCTEON_CN23XX_REV_2_0 0x80 /** Endian-swap modes supported by Octeon. */ enum octeon_pci_swap_mode { @@ -46,6 +52,9 @@ enum octeon_pci_swap_mode { OCTEON_PCI_32BIT_LW_SWAP = 3 }; +#define OCTEON_OUTPUT_INTR (2) +#define OCTEON_ALL_INTR 0xff + /*--------------- PCI BAR1 index registers -------------*/ /* BAR1 Mask */ @@ -198,9 +207,9 @@ struct octeon_fn_list { void (*setup_oq_regs)(struct octeon_device *, u32); irqreturn_t (*process_interrupt_regs)(void *); + u64 (*msix_interrupt_handler)(void *); int (*soft_reset)(struct octeon_device *); int (*setup_device_regs)(struct octeon_device *); - void (*reinit_regs)(struct octeon_device *); void (*bar1_idx_setup)(struct octeon_device *, u64, u32, int); void (*bar1_idx_write)(struct octeon_device *, u32, u32); u32 (*bar1_idx_read)(struct octeon_device *, u32); @@ -209,10 +218,10 @@ struct octeon_fn_list { void (*enable_oq_pkt_time_intr)(struct octeon_device *, u32); void (*disable_oq_pkt_time_intr)(struct octeon_device *, u32); - void (*enable_interrupt)(void *); - void (*disable_interrupt)(void *); + void (*enable_interrupt)(struct octeon_device *, u8); + void (*disable_interrupt)(struct octeon_device *, u8); - void (*enable_io_queues)(struct octeon_device *); + int (*enable_io_queues)(struct octeon_device *); void (*disable_io_queues)(struct octeon_device *); }; @@ -266,11 +275,72 @@ struct octdev_props { /* Each interface in the Octeon device has a network * device pointer (used for OS specific calls). */ + int rx_on; int napi_enabled; int gmxport; struct net_device *netdev; }; +#define LIO_FLAG_MSIX_ENABLED 0x1 +#define MSIX_PO_INT 0x1 +#define MSIX_PI_INT 0x2 + +struct octeon_pf_vf_hs_word { +#ifdef __LITTLE_ENDIAN_BITFIELD + /** PKIND value assigned for the DPI interface */ + u64 pkind : 8; + + /** OCTEON core clock multiplier */ + u64 core_tics_per_us : 16; + + /** OCTEON coprocessor clock multiplier */ + u64 coproc_tics_per_us : 16; + + /** app that currently running on OCTEON */ + u64 app_mode : 8; + + /** RESERVED */ + u64 reserved : 16; + +#else + + /** RESERVED */ + u64 reserved : 16; + + /** app that currently running on OCTEON */ + u64 app_mode : 8; + + /** OCTEON coprocessor clock multiplier */ + u64 coproc_tics_per_us : 16; + + /** OCTEON core clock multiplier */ + u64 core_tics_per_us : 16; + + /** PKIND value assigned for the DPI interface */ + u64 pkind : 8; +#endif +}; + +struct octeon_sriov_info { + /* Actual rings left for PF device */ + u32 num_pf_rings; + + /* SRN of PF usable IO queues */ + u32 pf_srn; + /* total pf rings */ + u32 trs; + +}; + +struct octeon_ioq_vector { + struct octeon_device *oct_dev; + int iq_index; + int droq_index; + int vector; + struct cpumask affinity_mask; + u32 ioq_num; +}; + /** The Octeon device. * Each Octeon device has this structure to represent all its * components. @@ -296,7 +366,7 @@ struct octeon_device { /** Octeon Chip type. */ u16 chip_id; u16 rev_id; - + u16 pf_num; /** This device's id - set by the driver. */ u32 octeon_id; @@ -305,7 +375,6 @@ struct octeon_device { u16 flags; #define LIO_FLAG_MSI_ENABLED (u32)(1 << 1) -#define LIO_FLAG_MSIX_ENABLED (u32)(1 << 2) /** The state of this device */ atomic_t status; @@ -395,6 +464,19 @@ struct octeon_device { void *priv; + int num_msix_irqs; + + void *msix_entries; + + struct octeon_sriov_info sriov_info; + + struct octeon_pf_vf_hs_word pfvf_hsword; + + int msix_on; + + /** IOq information of it's corresponding MSI-X interrupt. */ + struct octeon_ioq_vector *ioq_vector; + int rx_pause; int tx_pause; @@ -402,12 +484,15 @@ struct octeon_device { /* private flags to control driver-specific features through ethtool */ u32 priv_flags; + + void *watchdog_task; }; #define OCT_DRV_ONLINE 1 #define OCT_DRV_OFFLINE 2 #define OCTEON_CN6XXX(oct) ((oct->chip_id == OCTEON_CN66XX) || \ (oct->chip_id == OCTEON_CN68XX)) +#define OCTEON_CN23XX_PF(oct) (oct->chip_id == OCTEON_CN23XX_PF_VID) #define CHIP_FIELD(oct, TYPE, field) \ (((struct octeon_ ## TYPE *)(oct->chip))->field) @@ -661,13 +746,24 @@ void *oct_get_config_info(struct octeon_device *oct, u16 card_type); */ struct octeon_config *octeon_get_conf(struct octeon_device *oct); +void octeon_free_ioq_vector(struct octeon_device *oct); +int octeon_allocate_ioq_vector(struct octeon_device *oct); +void lio_enable_irq(struct octeon_droq *droq, struct octeon_instr_queue *iq); + /* LiquidIO driver pivate flags */ enum { OCT_PRIV_FLAG_TX_BYTES = 0, /* Tx interrupts by pending byte count */ }; -static inline void lio_set_priv_flag(struct octeon_device *octdev, u32 flag, - u32 val) +#define OCT_PRIV_FLAG_DEFAULT 0x0 + +static inline u32 lio_get_priv_flag(struct octeon_device *octdev, u32 flag) +{ + return !!(octdev->priv_flags & (0x1 << flag)); +} + +static inline void lio_set_priv_flag(struct octeon_device *octdev, + u32 flag, u32 val) { if (val) octdev->priv_flags |= (0x1 << flag); diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_droq.c b/drivers/net/ethernet/cavium/liquidio/octeon_droq.c index e0afe4c1fd01..f60e5320daf4 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_droq.c +++ b/drivers/net/ethernet/cavium/liquidio/octeon_droq.c @@ -31,6 +31,7 @@ #include "octeon_network.h" #include "cn66xx_regs.h" #include "cn66xx_device.h" +#include "cn23xx_pf_device.h" #define CVM_MIN(d1, d2) (((d1) < (d2)) ? (d1) : (d2)) #define CVM_MAX(d1, d2) (((d1) > (d2)) ? (d1) : (d2)) @@ -92,22 +93,25 @@ static inline void *octeon_get_dispatch_arg(struct octeon_device *octeon_dev, return fn_arg; } -/** Check for packets on Droq. This function should be called with - * lock held. +/** Check for packets on Droq. This function should be called with lock held. * @param droq - Droq on which count is checked. * @return Returns packet count. */ u32 octeon_droq_check_hw_for_pkts(struct octeon_droq *droq) { u32 pkt_count = 0; + u32 last_count; pkt_count = readl(droq->pkts_sent_reg); - if (pkt_count) { - atomic_add(pkt_count, &droq->pkts_pending); - writel(pkt_count, droq->pkts_sent_reg); - } - return pkt_count; + last_count = pkt_count - droq->pkt_count; + droq->pkt_count = pkt_count; + + /* we shall write to cnts at napi irq enable or end of droq tasklet */ + if (last_count) + atomic_add(last_count, &droq->pkts_pending); + + return last_count; } static void octeon_droq_compute_max_packet_bufs(struct octeon_droq *droq) @@ -259,6 +263,11 @@ int octeon_init_droq(struct octeon_device *oct, c_pkts_per_intr = (u32)CFG_GET_OQ_PKTS_PER_INTR(conf6x); c_refill_threshold = (u32)CFG_GET_OQ_REFILL_THRESHOLD(conf6x); + } else if (OCTEON_CN23XX_PF(oct)) { + struct octeon_config *conf23 = CHIP_FIELD(oct, cn23xx_pf, conf); + + c_pkts_per_intr = (u32)CFG_GET_OQ_PKTS_PER_INTR(conf23); + c_refill_threshold = (u32)CFG_GET_OQ_REFILL_THRESHOLD(conf23); } else { return 1; } @@ -564,7 +573,7 @@ octeon_droq_dispatch_pkt(struct octeon_device *oct, (unsigned int)rh->r.opcode, (unsigned int)rh->r.subcode); droq->stats.dropped_nodispatch++; - } /* else (dispatch_fn ... */ + } return cnt; } @@ -735,16 +744,20 @@ octeon_droq_process_packets(struct octeon_device *oct, u32 pkt_count = 0, pkts_processed = 0; struct list_head *tmp, *tmp2; + /* Grab the droq lock */ + spin_lock(&droq->lock); + + octeon_droq_check_hw_for_pkts(droq); pkt_count = atomic_read(&droq->pkts_pending); - if (!pkt_count) + + if (!pkt_count) { + spin_unlock(&droq->lock); return 0; + } if (pkt_count > budget) pkt_count = budget; - /* Grab the droq lock */ - spin_lock(&droq->lock); - pkts_processed = octeon_droq_fast_process_packets(oct, droq, pkt_count); atomic_sub(pkts_processed, &droq->pkts_pending); @@ -789,6 +802,8 @@ octeon_droq_process_poll_pkts(struct octeon_device *oct, spin_lock(&droq->lock); while (total_pkts_processed < budget) { + octeon_droq_check_hw_for_pkts(droq); + pkts_available = CVM_MIN((budget - total_pkts_processed), (u32)(atomic_read(&droq->pkts_pending))); @@ -803,8 +818,6 @@ octeon_droq_process_poll_pkts(struct octeon_device *oct, atomic_sub(pkts_processed, &droq->pkts_pending); total_pkts_processed += pkts_processed; - - octeon_droq_check_hw_for_pkts(droq); } spin_unlock(&droq->lock); @@ -874,8 +887,11 @@ octeon_process_droq_poll_cmd(struct octeon_device *oct, u32 q_no, int cmd, return 0; } break; + case OCTEON_CN23XX_PF_VID: { + lio_enable_irq(oct->droq[q_no], oct->instr_queue[q_no]); + } + break; } - return 0; } diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_droq.h b/drivers/net/ethernet/cavium/liquidio/octeon_droq.h index 5a6fb9113bbd..5be002d5dba4 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_droq.h +++ b/drivers/net/ethernet/cavium/liquidio/octeon_droq.h @@ -261,6 +261,8 @@ struct octeon_droq { u32 q_no; + u32 pkt_count; + struct octeon_droq_ops ops; struct octeon_device *oct_dev; diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_iq.h b/drivers/net/ethernet/cavium/liquidio/octeon_iq.h index ff4b1d6f007b..e4d426ba18dc 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_iq.h +++ b/drivers/net/ethernet/cavium/liquidio/octeon_iq.h @@ -88,6 +88,8 @@ struct octeon_instr_queue { /** A spinlock to protect while posting on the ring. */ spinlock_t post_lock; + u32 pkt_in_done; + /** A spinlock to protect access to the input ring.*/ spinlock_t iq_flush_running_lock; diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_main.h b/drivers/net/ethernet/cavium/liquidio/octeon_main.h index bc14e4c27332..366298f7bcb2 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_main.h +++ b/drivers/net/ethernet/cavium/liquidio/octeon_main.h @@ -38,12 +38,26 @@ #define DRV_NAME "LiquidIO" -/** - * \brief determines if a given console has debug enabled. - * @param console console to check - * @returns 1 = enabled. 0 otherwise +/** This structure is used by NIC driver to store information required + * to free the sk_buff when the packet has been fetched by Octeon. + * Bytes offset below assume worst-case of a 64-bit system. */ -int octeon_console_debug_enabled(u32 console); +struct octnet_buf_free_info { + /** Bytes 1-8. Pointer to network device private structure. */ + struct lio *lio; + + /** Bytes 9-16. Pointer to sk_buff. */ + struct sk_buff *skb; + + /** Bytes 17-24. Pointer to gather list. */ + struct octnic_gather *g; + + /** Bytes 25-32. Physical address of skb->data or gather list. */ + u64 dptr; + + /** Bytes 33-47. Piggybacked soft command, if any */ + struct octeon_soft_command *sc; +}; /* BQL-related functions */ void octeon_report_sent_bytes_to_bql(void *buf, int reqtype); @@ -167,22 +181,26 @@ cnnic_numa_alloc_aligned_dma(u32 size, #define cnnic_free_aligned_dma(pci_dev, ptr, size, orig_ptr, dma_addr) \ free_pages(orig_ptr, get_order(size)) -static inline void +static inline int sleep_cond(wait_queue_head_t *wait_queue, int *condition) { + int errno = 0; wait_queue_t we; init_waitqueue_entry(&we, current); add_wait_queue(wait_queue, &we); while (!(READ_ONCE(*condition))) { set_current_state(TASK_INTERRUPTIBLE); - if (signal_pending(current)) + if (signal_pending(current)) { + errno = -EINTR; goto out; + } schedule(); } out: set_current_state(TASK_RUNNING); remove_wait_queue(wait_queue, &we); + return errno; } static inline void diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_mem_ops.c b/drivers/net/ethernet/cavium/liquidio/octeon_mem_ops.c index 95a4bbedf557..0dc081a99b30 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_mem_ops.c +++ b/drivers/net/ethernet/cavium/liquidio/octeon_mem_ops.c @@ -19,7 +19,6 @@ * This file may also be available under a different license from Cavium. * Contact Cavium, Inc. for more information **********************************************************************/ -#include <linux/pci.h> #include <linux/netdevice.h> #include "liquidio_common.h" #include "octeon_droq.h" diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_network.h b/drivers/net/ethernet/cavium/liquidio/octeon_network.h index fb820dc7fcb7..e5d1debd05ad 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_network.h +++ b/drivers/net/ethernet/cavium/liquidio/octeon_network.h @@ -26,8 +26,6 @@ #ifndef __OCTEON_NETWORK_H__ #define __OCTEON_NETWORK_H__ -#include <linux/version.h> -#include <linux/dma-mapping.h> #include <linux/ptp_clock_kernel.h> #define LIO_MAX_MTU_SIZE (OCTNET_MAX_FRM_SIZE - OCTNET_FRM_HEADER_SIZE) @@ -124,11 +122,21 @@ struct lio { /* work queue for txq status */ struct cavium_wq txq_status_wq; + + /* work queue for link status */ + struct cavium_wq link_status_wq; + }; #define LIO_SIZE (sizeof(struct lio)) #define GET_LIO(netdev) ((struct lio *)netdev_priv(netdev)) +#define CIU3_WDOG(c) (0x1010000020000ULL + (c << 3)) +#define CIU3_WDOG_MASK 12ULL +#define LIO_MONITOR_WDOG_EXPIRE 1 +#define LIO_MONITOR_CORE_STUCK_MSGD 2 +#define LIO_MAX_CORES 12 + /** * \brief Enable or disable feature * @param netdev pointer to network device diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_nic.c b/drivers/net/ethernet/cavium/liquidio/octeon_nic.c index 166727be928f..40ac1fe88956 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_nic.c +++ b/drivers/net/ethernet/cavium/liquidio/octeon_nic.c @@ -19,7 +19,6 @@ * This file may also be available under a different license from Cavium. * Contact Cavium, Inc. for more information **********************************************************************/ -#include <linux/interrupt.h> #include <linux/pci.h> #include <linux/netdevice.h> #include "liquidio_common.h" @@ -36,6 +35,7 @@ octeon_alloc_soft_command_resp(struct octeon_device *oct, u32 rdatasize) { struct octeon_soft_command *sc; + struct octeon_instr_ih3 *ih3; struct octeon_instr_ih2 *ih2; struct octeon_instr_irh *irh; struct octeon_instr_rdp *rdp; @@ -52,10 +52,19 @@ octeon_alloc_soft_command_resp(struct octeon_device *oct, /* Add in the response related fields. Opcode and Param are already * there. */ - ih2 = (struct octeon_instr_ih2 *)&sc->cmd.cmd2.ih2; - rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd2.rdp; - irh = (struct octeon_instr_irh *)&sc->cmd.cmd2.irh; - ih2->fsz = 40; /* irh + ossp[0] + ossp[1] + rdp + rptr = 40 bytes */ + if (OCTEON_CN23XX_PF(oct)) { + ih3 = (struct octeon_instr_ih3 *)&sc->cmd.cmd3.ih3; + rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd3.rdp; + irh = (struct octeon_instr_irh *)&sc->cmd.cmd3.irh; + /*pkiih3 + irh + ossp[0] + ossp[1] + rdp + rptr = 40 bytes */ + ih3->fsz = LIO_SOFTCMDRESP_IH3; + } else { + ih2 = (struct octeon_instr_ih2 *)&sc->cmd.cmd2.ih2; + rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd2.rdp; + irh = (struct octeon_instr_irh *)&sc->cmd.cmd2.irh; + /* irh + ossp[0] + ossp[1] + rdp + rptr = 40 bytes */ + ih2->fsz = LIO_SOFTCMDRESP_IH2; + } irh->rflag = 1; /* a response is required */ @@ -64,7 +73,10 @@ octeon_alloc_soft_command_resp(struct octeon_device *oct, *sc->status_word = COMPLETION_WORD_INIT; - sc->cmd.cmd2.rptr = sc->dmarptr; + if (OCTEON_CN23XX_PF(oct)) + sc->cmd.cmd3.rptr = sc->dmarptr; + else + sc->cmd.cmd2.rptr = sc->dmarptr; sc->wait_time = 1000; sc->timeout = jiffies + sc->wait_time; @@ -73,12 +85,9 @@ octeon_alloc_soft_command_resp(struct octeon_device *oct, } int octnet_send_nic_data_pkt(struct octeon_device *oct, - struct octnic_data_pkt *ndata, - u32 xmit_more) + struct octnic_data_pkt *ndata) { - int ring_doorbell; - - ring_doorbell = !xmit_more; + int ring_doorbell = 1; return octeon_send_command(oct, ndata->q_no, ring_doorbell, &ndata->cmd, ndata->buf, ndata->datasize, @@ -183,8 +192,8 @@ octnet_send_nic_ctrl_pkt(struct octeon_device *oct, retval = octeon_send_soft_command(oct, sc); if (retval == IQ_SEND_FAILED) { octeon_free_soft_command(oct, sc); - dev_err(&oct->pci_dev->dev, "%s soft command:%d send failed status: %x\n", - __func__, nctrl->ncmd.s.cmd, retval); + dev_err(&oct->pci_dev->dev, "%s pf_num:%d soft command:%d send failed status: %x\n", + __func__, oct->pf_num, nctrl->ncmd.s.cmd, retval); spin_unlock_bh(&oct->cmd_resp_wqlock); return -1; } diff --git a/drivers/net/ethernet/cavium/liquidio/octeon_nic.h b/drivers/net/ethernet/cavium/liquidio/octeon_nic.h index b71a2bbe4bee..4b8da67b995f 100644 --- a/drivers/net/ethernet/cavium/liquidio/octeon_nic.h +++ b/drivers/net/ethernet/cavium/liquidio/octeon_nic.h @@ -138,7 +138,7 @@ octnet_prepare_pci_cmd_o2(struct octeon_device *oct, /* assume that rflag is cleared so therefore front data will only have * irh and ossp[0], ossp[1] for a total of 32 bytes */ - ih2->fsz = 24; + ih2->fsz = LIO_PCICMD_O2; ih2->tagtype = ORDERED_TAG; ih2->grp = DEFAULT_POW_GRP; @@ -196,7 +196,7 @@ octnet_prepare_pci_cmd_o3(struct octeon_device *oct, */ ih3->pkind = oct->instr_queue[setup->s.iq_no]->txpciq.s.pkind; /*PKI IH*/ - ih3->fsz = 24 + 8; + ih3->fsz = LIO_PCICMD_O3; if (!setup->s.gather) { ih3->dlengsz = setup->s.u.datasize; @@ -278,7 +278,7 @@ octeon_alloc_soft_command_resp(struct octeon_device *oct, * queue should be stopped, and IQ_SEND_OK if it sent okay. */ int octnet_send_nic_data_pkt(struct octeon_device *oct, - struct octnic_data_pkt *ndata, u32 xmit_more); + struct octnic_data_pkt *ndata); /** Send a NIC control packet to the device * @param oct - octeon device pointer diff --git a/drivers/net/ethernet/cavium/liquidio/request_manager.c b/drivers/net/ethernet/cavium/liquidio/request_manager.c index d32492f185ff..90866bb50033 100644 --- a/drivers/net/ethernet/cavium/liquidio/request_manager.c +++ b/drivers/net/ethernet/cavium/liquidio/request_manager.c @@ -30,6 +30,7 @@ #include "octeon_main.h" #include "octeon_network.h" #include "cn66xx_device.h" +#include "cn23xx_pf_device.h" #define INCR_INSTRQUEUE_PKT_COUNT(octeon_dev_ptr, iq_no, field, count) \ (octeon_dev_ptr->instr_queue[iq_no]->stats.field += count) @@ -71,7 +72,8 @@ int octeon_init_instr_queue(struct octeon_device *oct, if (OCTEON_CN6XXX(oct)) conf = &(CFG_GET_IQ_CFG(CHIP_FIELD(oct, cn6xxx, conf))); - + else if (OCTEON_CN23XX_PF(oct)) + conf = &(CFG_GET_IQ_CFG(CHIP_FIELD(oct, cn23xx_pf, conf))); if (!conf) { dev_err(&oct->pci_dev->dev, "Unsupported Chip %x\n", oct->chip_id); @@ -88,6 +90,7 @@ int octeon_init_instr_queue(struct octeon_device *oct, q_size = (u32)conf->instr_type * num_descs; iq = oct->instr_queue[iq_no]; + iq->oct_dev = oct; set_dev_node(&oct->pci_dev->dev, numa_node); @@ -181,6 +184,9 @@ int octeon_delete_instr_queue(struct octeon_device *oct, u32 iq_no) if (OCTEON_CN6XXX(oct)) desc_size = CFG_GET_IQ_INSTR_TYPE(CHIP_FIELD(oct, cn6xxx, conf)); + else if (OCTEON_CN23XX_PF(oct)) + desc_size = + CFG_GET_IQ_INSTR_TYPE(CHIP_FIELD(oct, cn23xx_pf, conf)); vfree(iq->request_list); @@ -383,7 +389,12 @@ lio_process_iq_request_list(struct octeon_device *oct, case REQTYPE_SOFT_COMMAND: sc = buf; - irh = (struct octeon_instr_irh *)&sc->cmd.cmd2.irh; + if (OCTEON_CN23XX_PF(oct)) + irh = (struct octeon_instr_irh *) + &sc->cmd.cmd3.irh; + else + irh = (struct octeon_instr_irh *) + &sc->cmd.cmd2.irh; if (irh->rflag) { /* We're expecting a response from Octeon. * It's up to lio_process_ordered_list() to @@ -499,6 +510,7 @@ static void __check_db_timeout(struct octeon_device *oct, u64 iq_no) if (!oct) return; + iq = oct->instr_queue[iq_no]; if (!iq) return; @@ -514,6 +526,8 @@ static void __check_db_timeout(struct octeon_device *oct, u64 iq_no) /* Flush the instruction queue */ octeon_flush_iq(oct, iq, 1, 0); + + lio_enable_irq(NULL, iq); } /* Called by the Poll thread at regular intervals to check the instruction @@ -580,6 +594,8 @@ octeon_prepare_soft_command(struct octeon_device *oct, { struct octeon_config *oct_cfg; struct octeon_instr_ih2 *ih2; + struct octeon_instr_ih3 *ih3; + struct octeon_instr_pki_ih3 *pki_ih3; struct octeon_instr_irh *irh; struct octeon_instr_rdp *rdp; @@ -588,36 +604,88 @@ octeon_prepare_soft_command(struct octeon_device *oct, oct_cfg = octeon_get_conf(oct); - ih2 = (struct octeon_instr_ih2 *)&sc->cmd.cmd2.ih2; - ih2->tagtype = ATOMIC_TAG; - ih2->tag = LIO_CONTROL; - ih2->raw = 1; - ih2->grp = CFG_GET_CTRL_Q_GRP(oct_cfg); - - if (sc->datasize) { - ih2->dlengsz = sc->datasize; - ih2->rs = 1; - } - - irh = (struct octeon_instr_irh *)&sc->cmd.cmd2.irh; - irh->opcode = opcode; - irh->subcode = subcode; - - /* opcode/subcode specific parameters (ossp) */ - irh->ossp = irh_ossp; - sc->cmd.cmd2.ossp[0] = ossp0; - sc->cmd.cmd2.ossp[1] = ossp1; - - if (sc->rdatasize) { - rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd2.rdp; - rdp->pcie_port = oct->pcie_port; - rdp->rlen = sc->rdatasize; + if (OCTEON_CN23XX_PF(oct)) { + ih3 = (struct octeon_instr_ih3 *)&sc->cmd.cmd3.ih3; + + ih3->pkind = oct->instr_queue[sc->iq_no]->txpciq.s.pkind; + + pki_ih3 = (struct octeon_instr_pki_ih3 *)&sc->cmd.cmd3.pki_ih3; + + pki_ih3->w = 1; + pki_ih3->raw = 1; + pki_ih3->utag = 1; + pki_ih3->uqpg = + oct->instr_queue[sc->iq_no]->txpciq.s.use_qpg; + pki_ih3->utt = 1; + pki_ih3->tag = LIO_CONTROL; + pki_ih3->tagtype = ATOMIC_TAG; + pki_ih3->qpg = + oct->instr_queue[sc->iq_no]->txpciq.s.qpg; + pki_ih3->pm = 0x7; + pki_ih3->sl = 8; + + if (sc->datasize) + ih3->dlengsz = sc->datasize; + + irh = (struct octeon_instr_irh *)&sc->cmd.cmd3.irh; + irh->opcode = opcode; + irh->subcode = subcode; + + /* opcode/subcode specific parameters (ossp) */ + irh->ossp = irh_ossp; + sc->cmd.cmd3.ossp[0] = ossp0; + sc->cmd.cmd3.ossp[1] = ossp1; + + if (sc->rdatasize) { + rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd3.rdp; + rdp->pcie_port = oct->pcie_port; + rdp->rlen = sc->rdatasize; + + irh->rflag = 1; + /*PKI IH3*/ + /* pki_ih3 irh+ossp[0]+ossp[1]+rdp+rptr = 48 bytes */ + ih3->fsz = LIO_SOFTCMDRESP_IH3; + } else { + irh->rflag = 0; + /*PKI IH3*/ + /* pki_h3 + irh + ossp[0] + ossp[1] = 32 bytes */ + ih3->fsz = LIO_PCICMD_O3; + } - irh->rflag = 1; - ih2->fsz = 40; /* irh+ossp[0]+ossp[1]+rdp+rptr = 40 bytes */ } else { - irh->rflag = 0; - ih2->fsz = 24; /* irh + ossp[0] + ossp[1] = 24 bytes */ + ih2 = (struct octeon_instr_ih2 *)&sc->cmd.cmd2.ih2; + ih2->tagtype = ATOMIC_TAG; + ih2->tag = LIO_CONTROL; + ih2->raw = 1; + ih2->grp = CFG_GET_CTRL_Q_GRP(oct_cfg); + + if (sc->datasize) { + ih2->dlengsz = sc->datasize; + ih2->rs = 1; + } + + irh = (struct octeon_instr_irh *)&sc->cmd.cmd2.irh; + irh->opcode = opcode; + irh->subcode = subcode; + + /* opcode/subcode specific parameters (ossp) */ + irh->ossp = irh_ossp; + sc->cmd.cmd2.ossp[0] = ossp0; + sc->cmd.cmd2.ossp[1] = ossp1; + + if (sc->rdatasize) { + rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd2.rdp; + rdp->pcie_port = oct->pcie_port; + rdp->rlen = sc->rdatasize; + + irh->rflag = 1; + /* irh+ossp[0]+ossp[1]+rdp+rptr = 40 bytes */ + ih2->fsz = LIO_SOFTCMDRESP_IH2; + } else { + irh->rflag = 0; + /* irh + ossp[0] + ossp[1] = 24 bytes */ + ih2->fsz = LIO_PCICMD_O2; + } } } @@ -625,23 +693,39 @@ int octeon_send_soft_command(struct octeon_device *oct, struct octeon_soft_command *sc) { struct octeon_instr_ih2 *ih2; + struct octeon_instr_ih3 *ih3; struct octeon_instr_irh *irh; u32 len; - ih2 = (struct octeon_instr_ih2 *)&sc->cmd.cmd2.ih2; - if (ih2->dlengsz) { - WARN_ON(!sc->dmadptr); - sc->cmd.cmd2.dptr = sc->dmadptr; - } - irh = (struct octeon_instr_irh *)&sc->cmd.cmd2.irh; - if (irh->rflag) { - WARN_ON(!sc->dmarptr); - WARN_ON(!sc->status_word); - *sc->status_word = COMPLETION_WORD_INIT; - - sc->cmd.cmd2.rptr = sc->dmarptr; + if (OCTEON_CN23XX_PF(oct)) { + ih3 = (struct octeon_instr_ih3 *)&sc->cmd.cmd3.ih3; + if (ih3->dlengsz) { + WARN_ON(!sc->dmadptr); + sc->cmd.cmd3.dptr = sc->dmadptr; + } + irh = (struct octeon_instr_irh *)&sc->cmd.cmd3.irh; + if (irh->rflag) { + WARN_ON(!sc->dmarptr); + WARN_ON(!sc->status_word); + *sc->status_word = COMPLETION_WORD_INIT; + sc->cmd.cmd3.rptr = sc->dmarptr; + } + len = (u32)ih3->dlengsz; + } else { + ih2 = (struct octeon_instr_ih2 *)&sc->cmd.cmd2.ih2; + if (ih2->dlengsz) { + WARN_ON(!sc->dmadptr); + sc->cmd.cmd2.dptr = sc->dmadptr; + } + irh = (struct octeon_instr_irh *)&sc->cmd.cmd2.irh; + if (irh->rflag) { + WARN_ON(!sc->dmarptr); + WARN_ON(!sc->status_word); + *sc->status_word = COMPLETION_WORD_INIT; + sc->cmd.cmd2.rptr = sc->dmarptr; + } + len = (u32)ih2->dlengsz; } - len = (u32)ih2->dlengsz; if (sc->wait_time) sc->timeout = jiffies + sc->wait_time; diff --git a/drivers/net/ethernet/cavium/liquidio/response_manager.c b/drivers/net/ethernet/cavium/liquidio/response_manager.c index 709049e36627..be52178d8cb6 100644 --- a/drivers/net/ethernet/cavium/liquidio/response_manager.c +++ b/drivers/net/ethernet/cavium/liquidio/response_manager.c @@ -91,8 +91,13 @@ int lio_process_ordered_list(struct octeon_device *octeon_dev, sc = (struct octeon_soft_command *)ordered_sc_list-> head.next; - rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd2.rdp; - rptr = sc->cmd.cmd2.rptr; + if (OCTEON_CN23XX_PF(octeon_dev)) { + rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd3.rdp; + rptr = sc->cmd.cmd3.rptr; + } else { + rdp = (struct octeon_instr_rdp *)&sc->cmd.cmd2.rdp; + rptr = sc->cmd.cmd2.rptr; + } status = OCTEON_REQUEST_PENDING; diff --git a/drivers/net/ethernet/cavium/thunder/Makefile b/drivers/net/ethernet/cavium/thunder/Makefile index 5c4615ccaa14..6b4d4add7353 100644 --- a/drivers/net/ethernet/cavium/thunder/Makefile +++ b/drivers/net/ethernet/cavium/thunder/Makefile @@ -2,6 +2,7 @@ # Makefile for Cavium's Thunder ethernet device # +obj-$(CONFIG_THUNDER_NIC_RGX) += thunder_xcv.o obj-$(CONFIG_THUNDER_NIC_BGX) += thunder_bgx.o obj-$(CONFIG_THUNDER_NIC_PF) += nicpf.o obj-$(CONFIG_THUNDER_NIC_VF) += nicvf.o diff --git a/drivers/net/ethernet/cavium/thunder/nic.h b/drivers/net/ethernet/cavium/thunder/nic.h index 83025bb4737c..30426109711c 100644 --- a/drivers/net/ethernet/cavium/thunder/nic.h +++ b/drivers/net/ethernet/cavium/thunder/nic.h @@ -20,6 +20,17 @@ #define PCI_DEVICE_ID_THUNDER_NIC_VF 0xA034 #define PCI_DEVICE_ID_THUNDER_BGX 0xA026 +/* Subsystem device IDs */ +#define PCI_SUBSYS_DEVID_88XX_NIC_PF 0xA11E +#define PCI_SUBSYS_DEVID_81XX_NIC_PF 0xA21E +#define PCI_SUBSYS_DEVID_83XX_NIC_PF 0xA31E + +#define PCI_SUBSYS_DEVID_88XX_PASS1_NIC_VF 0xA11E +#define PCI_SUBSYS_DEVID_88XX_NIC_VF 0xA134 +#define PCI_SUBSYS_DEVID_81XX_NIC_VF 0xA234 +#define PCI_SUBSYS_DEVID_83XX_NIC_VF 0xA334 + + /* PCI BAR nos */ #define PCI_CFG_REG_BAR_NUM 0 #define PCI_MSIX_REG_BAR_NUM 4 @@ -41,40 +52,8 @@ /* Max pkinds */ #define NIC_MAX_PKIND 16 -/* Rx Channels */ -/* Receive channel configuration in TNS bypass mode - * Below is configuration in TNS bypass mode - * BGX0-LMAC0-CHAN0 - VNIC CHAN0 - * BGX0-LMAC1-CHAN0 - VNIC CHAN16 - * ... - * BGX1-LMAC0-CHAN0 - VNIC CHAN128 - * ... - * BGX1-LMAC3-CHAN0 - VNIC CHAN174 - */ -#define NIC_INTF_COUNT 2 /* Interfaces btw VNIC and TNS/BGX */ -#define NIC_CHANS_PER_INF 128 -#define NIC_MAX_CHANS (NIC_INTF_COUNT * NIC_CHANS_PER_INF) -#define NIC_CPI_COUNT 2048 /* No of channel parse indices */ - -/* TNS bypass mode: 1-1 mapping between VNIC and BGX:LMAC */ -#define NIC_MAX_BGX MAX_BGX_PER_CN88XX -#define NIC_CPI_PER_BGX (NIC_CPI_COUNT / NIC_MAX_BGX) -#define NIC_MAX_CPI_PER_LMAC 64 /* Max when CPI_ALG is IP diffserv */ -#define NIC_RSSI_PER_BGX (NIC_RSSI_COUNT / NIC_MAX_BGX) - -/* Tx scheduling */ -#define NIC_MAX_TL4 1024 -#define NIC_MAX_TL4_SHAPERS 256 /* 1 shaper for 4 TL4s */ -#define NIC_MAX_TL3 256 -#define NIC_MAX_TL3_SHAPERS 64 /* 1 shaper for 4 TL3s */ -#define NIC_MAX_TL2 64 -#define NIC_MAX_TL2_SHAPERS 2 /* 1 shaper for 32 TL2s */ -#define NIC_MAX_TL1 2 - -/* TNS bypass mode */ -#define NIC_TL2_PER_BGX 32 -#define NIC_TL4_PER_BGX (NIC_MAX_TL4 / NIC_MAX_BGX) -#define NIC_TL4_PER_LMAC (NIC_MAX_TL4 / NIC_CHANS_PER_INF) +/* Max when CPI_ALG is IP diffserv */ +#define NIC_MAX_CPI_PER_LMAC 64 /* NIC VF Interrupts */ #define NICVF_INTR_CQ 0 @@ -148,7 +127,6 @@ struct nicvf_cq_poll { struct napi_struct napi; }; -#define NIC_RSSI_COUNT 4096 /* Total no of RSS indices */ #define NIC_MAX_RSS_HASH_BITS 8 #define NIC_MAX_RSS_IDR_TBL_SIZE (1 << NIC_MAX_RSS_HASH_BITS) #define RSS_HASH_KEY_SIZE 5 /* 320 bit key */ @@ -273,12 +251,14 @@ struct nicvf { struct net_device *netdev; struct pci_dev *pdev; void __iomem *reg_base; +#define MAX_QUEUES_PER_QSET 8 struct queue_set *qs; struct nicvf_cq_poll *napi[8]; u8 vf_id; u8 sqs_id; bool sqs_mode; bool hw_tso; + bool t88; /* Receive buffer alloc */ u32 rb_page_offset; @@ -325,7 +305,7 @@ struct nicvf { bool msix_enabled; u8 num_vec; struct msix_entry msix_entries[NIC_VF_MSIX_VECTORS]; - char irq_name[NIC_VF_MSIX_VECTORS][20]; + char irq_name[NIC_VF_MSIX_VECTORS][IFNAMSIZ + 15]; bool irq_allocated[NIC_VF_MSIX_VECTORS]; cpumask_var_t affinity_mask[NIC_VF_MSIX_VECTORS]; @@ -368,6 +348,7 @@ struct nicvf { #define NIC_MBOX_MSG_PNICVF_PTR 0x14 /* Get primary qset nicvf ptr */ #define NIC_MBOX_MSG_SNICVF_PTR 0x15 /* Send sqet nicvf ptr to PVF */ #define NIC_MBOX_MSG_LOOPBACK 0x16 /* Set interface in loopback */ +#define NIC_MBOX_MSG_RESET_STAT_COUNTER 0x17 /* Reset statistics counters */ #define NIC_MBOX_MSG_CFG_DONE 0xF0 /* VF configuration done */ #define NIC_MBOX_MSG_SHUTDOWN 0xF1 /* VF is being shutdown */ @@ -484,6 +465,31 @@ struct set_loopback { bool enable; }; +/* Reset statistics counters */ +struct reset_stat_cfg { + u8 msg; + /* Bitmap to select NIC_PF_VNIC(vf_id)_RX_STAT(0..13) */ + u16 rx_stat_mask; + /* Bitmap to select NIC_PF_VNIC(vf_id)_TX_STAT(0..4) */ + u8 tx_stat_mask; + /* Bitmap to select NIC_PF_QS(0..127)_RQ(0..7)_STAT(0..1) + * bit14, bit15 NIC_PF_QS(vf_id)_RQ7_STAT(0..1) + * bit12, bit13 NIC_PF_QS(vf_id)_RQ6_STAT(0..1) + * .. + * bit2, bit3 NIC_PF_QS(vf_id)_RQ1_STAT(0..1) + * bit0, bit1 NIC_PF_QS(vf_id)_RQ0_STAT(0..1) + */ + u16 rq_stat_mask; + /* Bitmap to select NIC_PF_QS(0..127)_SQ(0..7)_STAT(0..1) + * bit14, bit15 NIC_PF_QS(vf_id)_SQ7_STAT(0..1) + * bit12, bit13 NIC_PF_QS(vf_id)_SQ6_STAT(0..1) + * .. + * bit2, bit3 NIC_PF_QS(vf_id)_SQ1_STAT(0..1) + * bit0, bit1 NIC_PF_QS(vf_id)_SQ0_STAT(0..1) + */ + u16 sq_stat_mask; +}; + /* 128 bit shared memory between PF and each VF */ union nic_mbx { struct { u8 msg; } msg; @@ -501,6 +507,7 @@ union nic_mbx { struct sqs_alloc sqs_alloc; struct nicvf_ptr nicvf; struct set_loopback lbk; + struct reset_stat_cfg reset_stat; }; #define NIC_NODE_ID_MASK 0x03 @@ -514,7 +521,14 @@ static inline int nic_get_node_id(struct pci_dev *pdev) static inline bool pass1_silicon(struct pci_dev *pdev) { - return pdev->revision < 8; + return (pdev->revision < 8) && + (pdev->subsystem_device == PCI_SUBSYS_DEVID_88XX_NIC_PF); +} + +static inline bool pass2_silicon(struct pci_dev *pdev) +{ + return (pdev->revision >= 8) && + (pdev->subsystem_device == PCI_SUBSYS_DEVID_88XX_NIC_PF); } int nicvf_set_real_num_queues(struct net_device *netdev, diff --git a/drivers/net/ethernet/cavium/thunder/nic_main.c b/drivers/net/ethernet/cavium/thunder/nic_main.c index 16ed20357c5c..2bbf4cbf08b2 100644 --- a/drivers/net/ethernet/cavium/thunder/nic_main.c +++ b/drivers/net/ethernet/cavium/thunder/nic_main.c @@ -20,8 +20,25 @@ #define DRV_NAME "thunder-nic" #define DRV_VERSION "1.0" +struct hw_info { + u8 bgx_cnt; + u8 chans_per_lmac; + u8 chans_per_bgx; /* Rx/Tx chans */ + u8 chans_per_rgx; + u8 chans_per_lbk; + u16 cpi_cnt; + u16 rssi_cnt; + u16 rss_ind_tbl_size; + u16 tl4_cnt; + u16 tl3_cnt; + u8 tl2_cnt; + u8 tl1_cnt; + bool tl1_per_bgx; /* TL1 per BGX or per LMAC */ +}; + struct nicpf { struct pci_dev *pdev; + struct hw_info *hw; u8 node; unsigned int flags; u8 num_vf_en; /* No of VF enabled */ @@ -36,22 +53,22 @@ struct nicpf { #define NIC_SET_VF_LMAC_MAP(bgx, lmac) (((bgx & 0xF) << 4) | (lmac & 0xF)) #define NIC_GET_BGX_FROM_VF_LMAC_MAP(map) ((map >> 4) & 0xF) #define NIC_GET_LMAC_FROM_VF_LMAC_MAP(map) (map & 0xF) - u8 vf_lmac_map[MAX_LMAC]; + u8 *vf_lmac_map; struct delayed_work dwork; struct workqueue_struct *check_link; - u8 link[MAX_LMAC]; - u8 duplex[MAX_LMAC]; - u32 speed[MAX_LMAC]; + u8 *link; + u8 *duplex; + u32 *speed; u16 cpi_base[MAX_NUM_VFS_SUPPORTED]; u16 rssi_base[MAX_NUM_VFS_SUPPORTED]; - u16 rss_ind_tbl_size; bool mbx_lock[MAX_NUM_VFS_SUPPORTED]; /* MSI-X */ bool msix_enabled; u8 num_vec; - struct msix_entry msix_entries[NIC_PF_MSIX_VECTORS]; + struct msix_entry *msix_entries; bool irq_allocated[NIC_PF_MSIX_VECTORS]; + char irq_name[NIC_PF_MSIX_VECTORS][20]; }; /* Supported devices */ @@ -89,9 +106,22 @@ static u64 nic_reg_read(struct nicpf *nic, u64 offset) /* PF -> VF mailbox communication APIs */ static void nic_enable_mbx_intr(struct nicpf *nic) { - /* Enable mailbox interrupt for all 128 VFs */ - nic_reg_write(nic, NIC_PF_MAILBOX_ENA_W1S, ~0ull); - nic_reg_write(nic, NIC_PF_MAILBOX_ENA_W1S + sizeof(u64), ~0ull); + int vf_cnt = pci_sriov_get_totalvfs(nic->pdev); + +#define INTR_MASK(vfs) ((vfs < 64) ? (BIT_ULL(vfs) - 1) : (~0ull)) + + /* Clear it, to avoid spurious interrupts (if any) */ + nic_reg_write(nic, NIC_PF_MAILBOX_INT, INTR_MASK(vf_cnt)); + + /* Enable mailbox interrupt for all VFs */ + nic_reg_write(nic, NIC_PF_MAILBOX_ENA_W1S, INTR_MASK(vf_cnt)); + /* One mailbox intr enable reg per 64 VFs */ + if (vf_cnt > 64) { + nic_reg_write(nic, NIC_PF_MAILBOX_INT + sizeof(u64), + INTR_MASK(vf_cnt - 64)); + nic_reg_write(nic, NIC_PF_MAILBOX_ENA_W1S + sizeof(u64), + INTR_MASK(vf_cnt - 64)); + } } static void nic_clear_mbx_intr(struct nicpf *nic, int vf, int mbx_reg) @@ -144,7 +174,7 @@ static void nic_mbx_send_ready(struct nicpf *nic, int vf) mbx.nic_cfg.tns_mode = NIC_TNS_BYPASS_MODE; - if (vf < MAX_LMAC) { + if (vf < nic->num_vf_en) { bgx_idx = NIC_GET_BGX_FROM_VF_LMAC_MAP(nic->vf_lmac_map[vf]); lmac = NIC_GET_LMAC_FROM_VF_LMAC_MAP(nic->vf_lmac_map[vf]); @@ -155,7 +185,7 @@ static void nic_mbx_send_ready(struct nicpf *nic, int vf) mbx.nic_cfg.sqs_mode = (vf >= nic->num_vf_en) ? true : false; mbx.nic_cfg.node_id = nic->node; - mbx.nic_cfg.loopback_supported = vf < MAX_LMAC; + mbx.nic_cfg.loopback_supported = vf < nic->num_vf_en; nic_send_msg_to_vf(nic, vf, &mbx); } @@ -248,14 +278,27 @@ static int nic_update_hw_frs(struct nicpf *nic, int new_frs, int vf) /* Set minimum transmit packet size */ static void nic_set_tx_pkt_pad(struct nicpf *nic, int size) { - int lmac; + int lmac, max_lmac; + u16 sdevid; u64 lmac_cfg; - /* Max value that can be set is 60 */ - if (size > 60) - size = 60; + /* There is a issue in HW where-in while sending GSO sized + * pkts as part of TSO, if pkt len falls below this size + * NIC will zero PAD packet and also updates IP total length. + * Hence set this value to lessthan min pkt size of MAC+IP+TCP + * headers, BGX will do the padding to transmit 64 byte pkt. + */ + if (size > 52) + size = 52; + + pci_read_config_word(nic->pdev, PCI_SUBSYSTEM_ID, &sdevid); + /* 81xx's RGX has only one LMAC */ + if (sdevid == PCI_SUBSYS_DEVID_81XX_NIC_PF) + max_lmac = ((nic->hw->bgx_cnt - 1) * MAX_LMAC_PER_BGX) + 1; + else + max_lmac = nic->hw->bgx_cnt * MAX_LMAC_PER_BGX; - for (lmac = 0; lmac < (MAX_BGX_PER_CN88XX * MAX_LMAC_PER_BGX); lmac++) { + for (lmac = 0; lmac < max_lmac; lmac++) { lmac_cfg = nic_reg_read(nic, NIC_PF_LMAC_0_7_CFG | (lmac << 3)); lmac_cfg &= ~(0xF << 2); lmac_cfg |= ((size / 4) << 2); @@ -275,7 +318,7 @@ static void nic_set_lmac_vf_mapping(struct nicpf *nic) nic->num_vf_en = 0; - for (bgx = 0; bgx < NIC_MAX_BGX; bgx++) { + for (bgx = 0; bgx < nic->hw->bgx_cnt; bgx++) { if (!(bgx_map & (1 << bgx))) continue; lmac_cnt = bgx_get_lmac_count(nic->node, bgx); @@ -295,28 +338,125 @@ static void nic_set_lmac_vf_mapping(struct nicpf *nic) nic_reg_write(nic, NIC_PF_LMAC_0_7_CREDIT + (lmac * 8), lmac_credit); + + /* On CN81XX there are only 8 VFs but max possible no of + * interfaces are 9. + */ + if (nic->num_vf_en >= pci_sriov_get_totalvfs(nic->pdev)) { + nic->num_vf_en = pci_sriov_get_totalvfs(nic->pdev); + break; + } } } +static void nic_free_lmacmem(struct nicpf *nic) +{ + kfree(nic->vf_lmac_map); + kfree(nic->link); + kfree(nic->duplex); + kfree(nic->speed); +} + +static int nic_get_hw_info(struct nicpf *nic) +{ + u8 max_lmac; + u16 sdevid; + struct hw_info *hw = nic->hw; + + pci_read_config_word(nic->pdev, PCI_SUBSYSTEM_ID, &sdevid); + + switch (sdevid) { + case PCI_SUBSYS_DEVID_88XX_NIC_PF: + hw->bgx_cnt = MAX_BGX_PER_CN88XX; + hw->chans_per_lmac = 16; + hw->chans_per_bgx = 128; + hw->cpi_cnt = 2048; + hw->rssi_cnt = 4096; + hw->rss_ind_tbl_size = NIC_MAX_RSS_IDR_TBL_SIZE; + hw->tl3_cnt = 256; + hw->tl2_cnt = 64; + hw->tl1_cnt = 2; + hw->tl1_per_bgx = true; + break; + case PCI_SUBSYS_DEVID_81XX_NIC_PF: + hw->bgx_cnt = MAX_BGX_PER_CN81XX; + hw->chans_per_lmac = 8; + hw->chans_per_bgx = 32; + hw->chans_per_rgx = 8; + hw->chans_per_lbk = 24; + hw->cpi_cnt = 512; + hw->rssi_cnt = 256; + hw->rss_ind_tbl_size = 32; /* Max RSSI / Max interfaces */ + hw->tl3_cnt = 64; + hw->tl2_cnt = 16; + hw->tl1_cnt = 10; + hw->tl1_per_bgx = false; + break; + case PCI_SUBSYS_DEVID_83XX_NIC_PF: + hw->bgx_cnt = MAX_BGX_PER_CN83XX; + hw->chans_per_lmac = 8; + hw->chans_per_bgx = 32; + hw->chans_per_lbk = 64; + hw->cpi_cnt = 2048; + hw->rssi_cnt = 1024; + hw->rss_ind_tbl_size = 64; /* Max RSSI / Max interfaces */ + hw->tl3_cnt = 256; + hw->tl2_cnt = 64; + hw->tl1_cnt = 18; + hw->tl1_per_bgx = false; + break; + } + hw->tl4_cnt = MAX_QUEUES_PER_QSET * pci_sriov_get_totalvfs(nic->pdev); + + /* Allocate memory for LMAC tracking elements */ + max_lmac = hw->bgx_cnt * MAX_LMAC_PER_BGX; + nic->vf_lmac_map = kmalloc_array(max_lmac, sizeof(u8), GFP_KERNEL); + if (!nic->vf_lmac_map) + goto error; + nic->link = kmalloc_array(max_lmac, sizeof(u8), GFP_KERNEL); + if (!nic->link) + goto error; + nic->duplex = kmalloc_array(max_lmac, sizeof(u8), GFP_KERNEL); + if (!nic->duplex) + goto error; + nic->speed = kmalloc_array(max_lmac, sizeof(u32), GFP_KERNEL); + if (!nic->speed) + goto error; + return 0; + +error: + nic_free_lmacmem(nic); + return -ENOMEM; +} + #define BGX0_BLOCK 8 #define BGX1_BLOCK 9 -static void nic_init_hw(struct nicpf *nic) +static int nic_init_hw(struct nicpf *nic) { - int i; + int i, err; u64 cqm_cfg; + /* Get HW capability info */ + err = nic_get_hw_info(nic); + if (err) + return err; + /* Enable NIC HW block */ nic_reg_write(nic, NIC_PF_CFG, 0x3); /* Enable backpressure */ nic_reg_write(nic, NIC_PF_BP_CFG, (1ULL << 6) | 0x03); - /* Disable TNS mode on both interfaces */ - nic_reg_write(nic, NIC_PF_INTF_0_1_SEND_CFG, - (NIC_TNS_BYPASS_MODE << 7) | BGX0_BLOCK); - nic_reg_write(nic, NIC_PF_INTF_0_1_SEND_CFG | (1 << 8), - (NIC_TNS_BYPASS_MODE << 7) | BGX1_BLOCK); + /* TNS and TNS bypass modes are present only on 88xx */ + if (nic->pdev->subsystem_device == PCI_SUBSYS_DEVID_88XX_NIC_PF) { + /* Disable TNS mode on both interfaces */ + nic_reg_write(nic, NIC_PF_INTF_0_1_SEND_CFG, + (NIC_TNS_BYPASS_MODE << 7) | BGX0_BLOCK); + nic_reg_write(nic, NIC_PF_INTF_0_1_SEND_CFG | (1 << 8), + (NIC_TNS_BYPASS_MODE << 7) | BGX1_BLOCK); + } + nic_reg_write(nic, NIC_PF_INTF_0_1_BP_CFG, (1ULL << 63) | BGX0_BLOCK); nic_reg_write(nic, NIC_PF_INTF_0_1_BP_CFG + (1 << 8), @@ -346,11 +486,14 @@ static void nic_init_hw(struct nicpf *nic) cqm_cfg = nic_reg_read(nic, NIC_PF_CQM_CFG); if (cqm_cfg < NICPF_CQM_MIN_DROP_LEVEL) nic_reg_write(nic, NIC_PF_CQM_CFG, NICPF_CQM_MIN_DROP_LEVEL); + + return 0; } /* Channel parse index configuration */ static void nic_config_cpi(struct nicpf *nic, struct cpi_cfg_msg *cfg) { + struct hw_info *hw = nic->hw; u32 vnic, bgx, lmac, chan; u32 padd, cpi_count = 0; u64 cpi_base, cpi, rssi_base, rssi; @@ -360,9 +503,9 @@ static void nic_config_cpi(struct nicpf *nic, struct cpi_cfg_msg *cfg) bgx = NIC_GET_BGX_FROM_VF_LMAC_MAP(nic->vf_lmac_map[vnic]); lmac = NIC_GET_LMAC_FROM_VF_LMAC_MAP(nic->vf_lmac_map[vnic]); - chan = (lmac * MAX_BGX_CHANS_PER_LMAC) + (bgx * NIC_CHANS_PER_INF); - cpi_base = (lmac * NIC_MAX_CPI_PER_LMAC) + (bgx * NIC_CPI_PER_BGX); - rssi_base = (lmac * nic->rss_ind_tbl_size) + (bgx * NIC_RSSI_PER_BGX); + chan = (lmac * hw->chans_per_lmac) + (bgx * hw->chans_per_bgx); + cpi_base = vnic * NIC_MAX_CPI_PER_LMAC; + rssi_base = vnic * hw->rss_ind_tbl_size; /* Rx channel configuration */ nic_reg_write(nic, NIC_PF_CHAN_0_255_RX_BP_CFG | (chan << 3), @@ -434,7 +577,7 @@ static void nic_send_rss_size(struct nicpf *nic, int vf) msg = (u64 *)&mbx; mbx.rss_size.msg = NIC_MBOX_MSG_RSS_SIZE; - mbx.rss_size.ind_tbl_size = nic->rss_ind_tbl_size; + mbx.rss_size.ind_tbl_size = nic->hw->rss_ind_tbl_size; nic_send_msg_to_vf(nic, vf, &mbx); } @@ -481,7 +624,7 @@ static void nic_config_rss(struct nicpf *nic, struct rss_cfg_msg *cfg) /* 4 level transmit side scheduler configutation * for TNS bypass mode * - * Sample configuration for SQ0 + * Sample configuration for SQ0 on 88xx * VNIC0-SQ0 -> TL4(0) -> TL3[0] -> TL2[0] -> TL1[0] -> BGX0 * VNIC1-SQ0 -> TL4(8) -> TL3[2] -> TL2[0] -> TL1[0] -> BGX0 * VNIC2-SQ0 -> TL4(16) -> TL3[4] -> TL2[1] -> TL1[0] -> BGX0 @@ -494,6 +637,7 @@ static void nic_config_rss(struct nicpf *nic, struct rss_cfg_msg *cfg) static void nic_tx_channel_cfg(struct nicpf *nic, u8 vnic, struct sq_cfg_msg *sq) { + struct hw_info *hw = nic->hw; u32 bgx, lmac, chan; u32 tl2, tl3, tl4; u32 rr_quantum; @@ -512,21 +656,28 @@ static void nic_tx_channel_cfg(struct nicpf *nic, u8 vnic, /* 24 bytes for FCS, IPG and preamble */ rr_quantum = ((NIC_HW_MAX_FRS + 24) / 4); - if (!sq->sqs_mode) { - tl4 = (lmac * NIC_TL4_PER_LMAC) + (bgx * NIC_TL4_PER_BGX); - } else { - for (svf = 0; svf < MAX_SQS_PER_VF; svf++) { - if (nic->vf_sqs[pqs_vnic][svf] == vnic) - break; + /* For 88xx 0-511 TL4 transmits via BGX0 and + * 512-1023 TL4s transmit via BGX1. + */ + if (hw->tl1_per_bgx) { + tl4 = bgx * (hw->tl4_cnt / hw->bgx_cnt); + if (!sq->sqs_mode) { + tl4 += (lmac * MAX_QUEUES_PER_QSET); + } else { + for (svf = 0; svf < MAX_SQS_PER_VF; svf++) { + if (nic->vf_sqs[pqs_vnic][svf] == vnic) + break; + } + tl4 += (MAX_LMAC_PER_BGX * MAX_QUEUES_PER_QSET); + tl4 += (lmac * MAX_QUEUES_PER_QSET * MAX_SQS_PER_VF); + tl4 += (svf * MAX_QUEUES_PER_QSET); } - tl4 = (MAX_LMAC_PER_BGX * NIC_TL4_PER_LMAC); - tl4 += (lmac * NIC_TL4_PER_LMAC * MAX_SQS_PER_VF); - tl4 += (svf * NIC_TL4_PER_LMAC); - tl4 += (bgx * NIC_TL4_PER_BGX); + } else { + tl4 = (vnic * MAX_QUEUES_PER_QSET); } tl4 += sq_idx; - tl3 = tl4 / (NIC_MAX_TL4 / NIC_MAX_TL3); + tl3 = tl4 / (hw->tl4_cnt / hw->tl3_cnt); nic_reg_write(nic, NIC_PF_QSET_0_127_SQ_0_7_CFG2 | ((u64)vnic << NIC_QS_ID_SHIFT) | ((u32)sq_idx << NIC_Q_NUM_SHIFT), tl4); @@ -534,8 +685,19 @@ static void nic_tx_channel_cfg(struct nicpf *nic, u8 vnic, ((u64)vnic << 27) | ((u32)sq_idx << 24) | rr_quantum); nic_reg_write(nic, NIC_PF_TL3_0_255_CFG | (tl3 << 3), rr_quantum); - chan = (lmac * MAX_BGX_CHANS_PER_LMAC) + (bgx * NIC_CHANS_PER_INF); - nic_reg_write(nic, NIC_PF_TL3_0_255_CHAN | (tl3 << 3), chan); + + /* On 88xx 0-127 channels are for BGX0 and + * 127-255 channels for BGX1. + * + * On 81xx/83xx TL3_CHAN reg should be configured with channel + * within LMAC i.e 0-7 and not the actual channel number like on 88xx + */ + chan = (lmac * hw->chans_per_lmac) + (bgx * hw->chans_per_bgx); + if (hw->tl1_per_bgx) + nic_reg_write(nic, NIC_PF_TL3_0_255_CHAN | (tl3 << 3), chan); + else + nic_reg_write(nic, NIC_PF_TL3_0_255_CHAN | (tl3 << 3), 0); + /* Enable backpressure on the channel */ nic_reg_write(nic, NIC_PF_CHAN_0_255_TX_CFG | (chan << 3), 1); @@ -544,6 +706,16 @@ static void nic_tx_channel_cfg(struct nicpf *nic, u8 vnic, nic_reg_write(nic, NIC_PF_TL2_0_63_CFG | (tl2 << 3), rr_quantum); /* No priorities as of now */ nic_reg_write(nic, NIC_PF_TL2_0_63_PRI | (tl2 << 3), 0x00); + + /* Unlike 88xx where TL2s 0-31 transmits to TL1 '0' and rest to TL1 '1' + * on 81xx/83xx TL2 needs to be configured to transmit to one of the + * possible LMACs. + * + * This register doesn't exist on 88xx. + */ + if (!hw->tl1_per_bgx) + nic_reg_write(nic, NIC_PF_TL2_LMAC | (tl2 << 3), + lmac + (bgx * MAX_LMAC_PER_BGX)); } /* Send primary nicvf pointer to secondary QS's VF */ @@ -615,7 +787,7 @@ static int nic_config_loopback(struct nicpf *nic, struct set_loopback *lbk) { int bgx_idx, lmac_idx; - if (lbk->vf_id > MAX_LMAC) + if (lbk->vf_id >= nic->num_vf_en) return -1; bgx_idx = NIC_GET_BGX_FROM_VF_LMAC_MAP(nic->vf_lmac_map[lbk->vf_id]); @@ -626,6 +798,67 @@ static int nic_config_loopback(struct nicpf *nic, struct set_loopback *lbk) return 0; } +/* Reset statistics counters */ +static int nic_reset_stat_counters(struct nicpf *nic, + int vf, struct reset_stat_cfg *cfg) +{ + int i, stat, qnum; + u64 reg_addr; + + for (i = 0; i < RX_STATS_ENUM_LAST; i++) { + if (cfg->rx_stat_mask & BIT(i)) { + reg_addr = NIC_PF_VNIC_0_127_RX_STAT_0_13 | + (vf << NIC_QS_ID_SHIFT) | + (i << 3); + nic_reg_write(nic, reg_addr, 0); + } + } + + for (i = 0; i < TX_STATS_ENUM_LAST; i++) { + if (cfg->tx_stat_mask & BIT(i)) { + reg_addr = NIC_PF_VNIC_0_127_TX_STAT_0_4 | + (vf << NIC_QS_ID_SHIFT) | + (i << 3); + nic_reg_write(nic, reg_addr, 0); + } + } + + for (i = 0; i <= 15; i++) { + qnum = i >> 1; + stat = i & 1 ? 1 : 0; + reg_addr = (vf << NIC_QS_ID_SHIFT) | + (qnum << NIC_Q_NUM_SHIFT) | (stat << 3); + if (cfg->rq_stat_mask & BIT(i)) { + reg_addr |= NIC_PF_QSET_0_127_RQ_0_7_STAT_0_1; + nic_reg_write(nic, reg_addr, 0); + } + if (cfg->sq_stat_mask & BIT(i)) { + reg_addr |= NIC_PF_QSET_0_127_SQ_0_7_STAT_0_1; + nic_reg_write(nic, reg_addr, 0); + } + } + return 0; +} + +static void nic_enable_tunnel_parsing(struct nicpf *nic, int vf) +{ + u64 prot_def = (IPV6_PROT << 32) | (IPV4_PROT << 16) | ET_PROT; + u64 vxlan_prot_def = (IPV6_PROT_DEF << 32) | + (IPV4_PROT_DEF) << 16 | ET_PROT_DEF; + + /* Configure tunnel parsing parameters */ + nic_reg_write(nic, NIC_PF_RX_GENEVE_DEF, + (1ULL << 63 | UDP_GENEVE_PORT_NUM)); + nic_reg_write(nic, NIC_PF_RX_GENEVE_PROT_DEF, + ((7ULL << 61) | prot_def)); + nic_reg_write(nic, NIC_PF_RX_NVGRE_PROT_DEF, + ((7ULL << 61) | prot_def)); + nic_reg_write(nic, NIC_PF_RX_VXLAN_DEF_0_1, + ((1ULL << 63) | UDP_VXLAN_PORT_NUM)); + nic_reg_write(nic, NIC_PF_RX_VXLAN_PROT_DEF, + ((0xfULL << 60) | vxlan_prot_def)); +} + static void nic_enable_vf(struct nicpf *nic, int vf, bool enable) { int bgx, lmac; @@ -664,18 +897,17 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf) mbx_addr += sizeof(u64); } - dev_dbg(&nic->pdev->dev, "%s: Mailbox msg %d from VF%d\n", + dev_dbg(&nic->pdev->dev, "%s: Mailbox msg 0x%02x from VF%d\n", __func__, mbx.msg.msg, vf); switch (mbx.msg.msg) { case NIC_MBOX_MSG_READY: nic_mbx_send_ready(nic, vf); - if (vf < MAX_LMAC) { + if (vf < nic->num_vf_en) { nic->link[vf] = 0; nic->duplex[vf] = 0; nic->speed[vf] = 0; } - ret = 1; - break; + goto unlock; case NIC_MBOX_MSG_QS_CFG: reg_addr = NIC_PF_QSET_0_127_CFG | (mbx.qs.num << NIC_QS_ID_SHIFT); @@ -693,6 +925,15 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf) (mbx.rq.qs_num << NIC_QS_ID_SHIFT) | (mbx.rq.rq_num << NIC_Q_NUM_SHIFT); nic_reg_write(nic, reg_addr, mbx.rq.cfg); + /* Enable CQE_RX2_S extension in CQE_RX descriptor. + * This gets appended by default on 81xx/83xx chips, + * for consistency enabling the same on 88xx pass2 + * where this is introduced. + */ + if (pass2_silicon(nic->pdev)) + nic_reg_write(nic, NIC_PF_RX_CFG, 0x01); + if (!pass1_silicon(nic->pdev)) + nic_enable_tunnel_parsing(nic, vf); break; case NIC_MBOX_MSG_RQ_BP_CFG: reg_addr = NIC_PF_QSET_0_127_RQ_0_7_BP_CFG | @@ -717,8 +958,10 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf) nic_tx_channel_cfg(nic, mbx.qs.num, &mbx.sq); break; case NIC_MBOX_MSG_SET_MAC: - if (vf >= nic->num_vf_en) + if (vf >= nic->num_vf_en) { + ret = -1; /* NACK */ break; + } lmac = mbx.mac.vf_id; bgx = NIC_GET_BGX_FROM_VF_LMAC_MAP(nic->vf_lmac_map[lmac]); lmac = NIC_GET_LMAC_FROM_VF_LMAC_MAP(nic->vf_lmac_map[lmac]); @@ -767,25 +1010,38 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf) case NIC_MBOX_MSG_LOOPBACK: ret = nic_config_loopback(nic, &mbx.lbk); break; + case NIC_MBOX_MSG_RESET_STAT_COUNTER: + ret = nic_reset_stat_counters(nic, vf, &mbx.reset_stat); + break; default: dev_err(&nic->pdev->dev, "Invalid msg from VF%d, msg 0x%x\n", vf, mbx.msg.msg); break; } - if (!ret) + if (!ret) { nic_mbx_send_ack(nic, vf); - else if (mbx.msg.msg != NIC_MBOX_MSG_READY) + } else if (mbx.msg.msg != NIC_MBOX_MSG_READY) { + dev_err(&nic->pdev->dev, "NACK for MBOX 0x%02x from VF %d\n", + mbx.msg.msg, vf); nic_mbx_send_nack(nic, vf); + } unlock: nic->mbx_lock[vf] = false; } -static void nic_mbx_intr_handler (struct nicpf *nic, int mbx) +static irqreturn_t nic_mbx_intr_handler(int irq, void *nic_irq) { + struct nicpf *nic = (struct nicpf *)nic_irq; + int mbx; u64 intr; u8 vf, vf_per_mbx_reg = 64; + if (irq == nic->msix_entries[NIC_PF_INTR_ID_MBOX0].vector) + mbx = 0; + else + mbx = 1; + intr = nic_reg_read(nic, NIC_PF_MAILBOX_INT + (mbx << 3)); dev_dbg(&nic->pdev->dev, "PF interrupt Mbox%d 0x%llx\n", mbx, intr); for (vf = 0; vf < vf_per_mbx_reg; vf++) { @@ -797,23 +1053,6 @@ static void nic_mbx_intr_handler (struct nicpf *nic, int mbx) nic_clear_mbx_intr(nic, vf, mbx); } } -} - -static irqreturn_t nic_mbx0_intr_handler (int irq, void *nic_irq) -{ - struct nicpf *nic = (struct nicpf *)nic_irq; - - nic_mbx_intr_handler(nic, 0); - - return IRQ_HANDLED; -} - -static irqreturn_t nic_mbx1_intr_handler (int irq, void *nic_irq) -{ - struct nicpf *nic = (struct nicpf *)nic_irq; - - nic_mbx_intr_handler(nic, 1); - return IRQ_HANDLED; } @@ -821,7 +1060,13 @@ static int nic_enable_msix(struct nicpf *nic) { int i, ret; - nic->num_vec = NIC_PF_MSIX_VECTORS; + nic->num_vec = pci_msix_vec_count(nic->pdev); + + nic->msix_entries = kmalloc_array(nic->num_vec, + sizeof(struct msix_entry), + GFP_KERNEL); + if (!nic->msix_entries) + return -ENOMEM; for (i = 0; i < nic->num_vec; i++) nic->msix_entries[i].entry = i; @@ -829,8 +1074,9 @@ static int nic_enable_msix(struct nicpf *nic) ret = pci_enable_msix(nic->pdev, nic->msix_entries, nic->num_vec); if (ret) { dev_err(&nic->pdev->dev, - "Request for #%d msix vectors failed\n", - nic->num_vec); + "Request for #%d msix vectors failed, returned %d\n", + nic->num_vec, ret); + kfree(nic->msix_entries); return ret; } @@ -842,6 +1088,7 @@ static void nic_disable_msix(struct nicpf *nic) { if (nic->msix_enabled) { pci_disable_msix(nic->pdev); + kfree(nic->msix_entries); nic->msix_enabled = 0; nic->num_vec = 0; } @@ -860,27 +1107,26 @@ static void nic_free_all_interrupts(struct nicpf *nic) static int nic_register_interrupts(struct nicpf *nic) { - int ret; + int i, ret; /* Enable MSI-X */ ret = nic_enable_msix(nic); if (ret) return ret; - /* Register mailbox interrupt handlers */ - ret = request_irq(nic->msix_entries[NIC_PF_INTR_ID_MBOX0].vector, - nic_mbx0_intr_handler, 0, "NIC Mbox0", nic); - if (ret) - goto fail; - - nic->irq_allocated[NIC_PF_INTR_ID_MBOX0] = true; + /* Register mailbox interrupt handler */ + for (i = NIC_PF_INTR_ID_MBOX0; i < nic->num_vec; i++) { + sprintf(nic->irq_name[i], + "NICPF Mbox%d", (i - NIC_PF_INTR_ID_MBOX0)); - ret = request_irq(nic->msix_entries[NIC_PF_INTR_ID_MBOX1].vector, - nic_mbx1_intr_handler, 0, "NIC Mbox1", nic); - if (ret) - goto fail; + ret = request_irq(nic->msix_entries[i].vector, + nic_mbx_intr_handler, 0, + nic->irq_name[i], nic); + if (ret) + goto fail; - nic->irq_allocated[NIC_PF_INTR_ID_MBOX1] = true; + nic->irq_allocated[i] = true; + } /* Enable mailbox interrupt */ nic_enable_mbx_intr(nic); @@ -889,6 +1135,7 @@ static int nic_register_interrupts(struct nicpf *nic) fail: dev_err(&nic->pdev->dev, "Request irq failed\n"); nic_free_all_interrupts(nic); + nic_disable_msix(nic); return ret; } @@ -903,6 +1150,12 @@ static int nic_num_sqs_en(struct nicpf *nic, int vf_en) int pos, sqs_per_vf = MAX_SQS_PER_VF_SINGLE_NODE; u16 total_vf; + /* Secondary Qsets are needed only if CPU count is + * morethan MAX_QUEUES_PER_QSET. + */ + if (num_online_cpus() <= MAX_QUEUES_PER_QSET) + return 0; + /* Check if its a multi-node environment */ if (nr_node_ids > 1) sqs_per_vf = MAX_SQS_PER_VF; @@ -1008,6 +1261,12 @@ static int nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (!nic) return -ENOMEM; + nic->hw = devm_kzalloc(dev, sizeof(struct hw_info), GFP_KERNEL); + if (!nic->hw) { + devm_kfree(dev, nic); + return -ENOMEM; + } + pci_set_drvdata(pdev, nic); nic->pdev = pdev; @@ -1047,13 +1306,12 @@ static int nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent) nic->node = nic_get_node_id(pdev); - nic_set_lmac_vf_mapping(nic); - /* Initialize hardware */ - nic_init_hw(nic); + err = nic_init_hw(nic); + if (err) + goto err_release_regions; - /* Set RSS TBL size for each VF */ - nic->rss_ind_tbl_size = NIC_MAX_RSS_IDR_TBL_SIZE; + nic_set_lmac_vf_mapping(nic); /* Register interrupts */ err = nic_register_interrupts(nic); @@ -1086,6 +1344,9 @@ err_unregister_interrupts: err_release_regions: pci_release_regions(pdev); err_disable_device: + nic_free_lmacmem(nic); + devm_kfree(dev, nic->hw); + devm_kfree(dev, nic); pci_disable_device(pdev); pci_set_drvdata(pdev, NULL); return err; @@ -1106,6 +1367,11 @@ static void nic_remove(struct pci_dev *pdev) nic_unregister_interrupts(nic); pci_release_regions(pdev); + + nic_free_lmacmem(nic); + devm_kfree(&pdev->dev, nic->hw); + devm_kfree(&pdev->dev, nic); + pci_disable_device(pdev); pci_set_drvdata(pdev, NULL); } diff --git a/drivers/net/ethernet/cavium/thunder/nic_reg.h b/drivers/net/ethernet/cavium/thunder/nic_reg.h index fab35a593898..edf779f5a227 100644 --- a/drivers/net/ethernet/cavium/thunder/nic_reg.h +++ b/drivers/net/ethernet/cavium/thunder/nic_reg.h @@ -36,6 +36,20 @@ #define NIC_PF_MAILBOX_ENA_W1C (0x0450) #define NIC_PF_MAILBOX_ENA_W1S (0x0470) #define NIC_PF_RX_ETYPE_0_7 (0x0500) +#define NIC_PF_RX_GENEVE_DEF (0x0580) +#define UDP_GENEVE_PORT_NUM 0x17C1ULL +#define NIC_PF_RX_GENEVE_PROT_DEF (0x0588) +#define IPV6_PROT 0x86DDULL +#define IPV4_PROT 0x800ULL +#define ET_PROT 0x6558ULL +#define NIC_PF_RX_NVGRE_PROT_DEF (0x0598) +#define NIC_PF_RX_VXLAN_DEF_0_1 (0x05A0) +#define UDP_VXLAN_PORT_NUM 0x12B5 +#define NIC_PF_RX_VXLAN_PROT_DEF (0x05B0) +#define IPV6_PROT_DEF 0x2ULL +#define IPV4_PROT_DEF 0x1ULL +#define ET_PROT_DEF 0x3ULL +#define NIC_PF_RX_CFG (0x05D0) #define NIC_PF_PKIND_0_15_CFG (0x0600) #define NIC_PF_ECC0_FLIP0 (0x1000) #define NIC_PF_ECC1_FLIP0 (0x1008) @@ -103,6 +117,7 @@ #define NIC_PF_SW_SYNC_RX_DONE (0x490008) #define NIC_PF_TL2_0_63_CFG (0x500000) #define NIC_PF_TL2_0_63_PRI (0x520000) +#define NIC_PF_TL2_LMAC (0x540000) #define NIC_PF_TL2_0_63_SH_STATUS (0x580000) #define NIC_PF_TL3A_0_63_CFG (0x5F0000) #define NIC_PF_TL3_0_255_CFG (0x600000) diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_main.c b/drivers/net/ethernet/cavium/thunder/nicvf_main.c index a19e73f11d73..45a13f718863 100644 --- a/drivers/net/ethernet/cavium/thunder/nicvf_main.c +++ b/drivers/net/ethernet/cavium/thunder/nicvf_main.c @@ -29,10 +29,20 @@ static const struct pci_device_id nicvf_id_table[] = { { PCI_DEVICE_SUB(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_NIC_VF, - PCI_VENDOR_ID_CAVIUM, 0xA134) }, + PCI_VENDOR_ID_CAVIUM, + PCI_SUBSYS_DEVID_88XX_NIC_VF) }, { PCI_DEVICE_SUB(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_PASS1_NIC_VF, - PCI_VENDOR_ID_CAVIUM, 0xA11E) }, + PCI_VENDOR_ID_CAVIUM, + PCI_SUBSYS_DEVID_88XX_PASS1_NIC_VF) }, + { PCI_DEVICE_SUB(PCI_VENDOR_ID_CAVIUM, + PCI_DEVICE_ID_THUNDER_NIC_VF, + PCI_VENDOR_ID_CAVIUM, + PCI_SUBSYS_DEVID_81XX_NIC_VF) }, + { PCI_DEVICE_SUB(PCI_VENDOR_ID_CAVIUM, + PCI_DEVICE_ID_THUNDER_NIC_VF, + PCI_VENDOR_ID_CAVIUM, + PCI_SUBSYS_DEVID_83XX_NIC_VF) }, { 0, } /* end of table */ }; @@ -134,15 +144,19 @@ int nicvf_send_msg_to_pf(struct nicvf *nic, union nic_mbx *mbx) /* Wait for previous message to be acked, timeout 2sec */ while (!nic->pf_acked) { - if (nic->pf_nacked) + if (nic->pf_nacked) { + netdev_err(nic->netdev, + "PF NACK to mbox msg 0x%02x from VF%d\n", + (mbx->msg.msg & 0xFF), nic->vf_id); return -EINVAL; + } msleep(sleep); if (nic->pf_acked) break; timeout -= sleep; if (!timeout) { netdev_err(nic->netdev, - "PF didn't ack to mbox msg %d from VF%d\n", + "PF didn't ACK to mbox msg 0x%02x from VF%d\n", (mbx->msg.msg & 0xFF), nic->vf_id); return -EBUSY; } @@ -352,13 +366,7 @@ static int nicvf_rss_init(struct nicvf *nic) rss->enable = true; - /* Using the HW reset value for now */ - rss->key[0] = 0xFEED0BADFEED0BADULL; - rss->key[1] = 0xFEED0BADFEED0BADULL; - rss->key[2] = 0xFEED0BADFEED0BADULL; - rss->key[3] = 0xFEED0BADFEED0BADULL; - rss->key[4] = 0xFEED0BADFEED0BADULL; - + netdev_rss_key_fill(rss->key, RSS_HASH_KEY_SIZE * sizeof(u64)); nicvf_set_rss_key(nic); rss->cfg = RSS_IP_HASH_ENA | RSS_TCP_HASH_ENA | RSS_UDP_HASH_ENA; @@ -507,12 +515,15 @@ static int nicvf_init_resources(struct nicvf *nic) static void nicvf_snd_pkt_handler(struct net_device *netdev, struct cmp_queue *cq, - struct cqe_send_t *cqe_tx, int cqe_type) + struct cqe_send_t *cqe_tx, + int cqe_type, int budget, + unsigned int *tx_pkts, unsigned int *tx_bytes) { struct sk_buff *skb = NULL; struct nicvf *nic = netdev_priv(netdev); struct snd_queue *sq; struct sq_hdr_subdesc *hdr; + struct sq_hdr_subdesc *tso_sqe; sq = &nic->qs->sq[cqe_tx->sq_idx]; @@ -527,17 +538,23 @@ static void nicvf_snd_pkt_handler(struct net_device *netdev, nicvf_check_cqe_tx_errs(nic, cq, cqe_tx); skb = (struct sk_buff *)sq->skbuff[cqe_tx->sqe_ptr]; - /* For TSO offloaded packets only one SQE will have a valid SKB */ if (skb) { + /* Check for dummy descriptor used for HW TSO offload on 88xx */ + if (hdr->dont_send) { + /* Get actual TSO descriptors and free them */ + tso_sqe = + (struct sq_hdr_subdesc *)GET_SQ_DESC(sq, hdr->rsvd2); + nicvf_put_sq_desc(sq, tso_sqe->subdesc_cnt + 1); + } nicvf_put_sq_desc(sq, hdr->subdesc_cnt + 1); prefetch(skb); - dev_consume_skb_any(skb); + (*tx_pkts)++; + *tx_bytes += skb->len; + napi_consume_skb(skb, budget); sq->skbuff[cqe_tx->sqe_ptr] = (u64)NULL; } else { - /* In case of HW TSO, HW sends a CQE for each segment of a TSO - * packet instead of a single CQE for the whole TSO packet - * transmitted. Each of this CQE points to the same SQE, so - * avoid freeing same SQE multiple times. + /* In case of SW TSO on 88xx, only last segment will have + * a SKB attached, so just free SQEs here. */ if (!nic->hw_tso) nicvf_put_sq_desc(sq, hdr->subdesc_cnt + 1); @@ -648,6 +665,7 @@ static int nicvf_cq_intr_handler(struct net_device *netdev, u8 cq_idx, struct cmp_queue *cq = &qs->cq[cq_idx]; struct cqe_rx_t *cq_desc; struct netdev_queue *txq; + unsigned int tx_pkts = 0, tx_bytes = 0; spin_lock_bh(&cq->lock); loop: @@ -686,7 +704,8 @@ loop: break; case CQE_TYPE_SEND: nicvf_snd_pkt_handler(netdev, cq, - (void *)cq_desc, CQE_TYPE_SEND); + (void *)cq_desc, CQE_TYPE_SEND, + budget, &tx_pkts, &tx_bytes); tx_done++; break; case CQE_TYPE_INVALID: @@ -715,6 +734,9 @@ done: netdev = nic->pnicvf->netdev; txq = netdev_get_tx_queue(netdev, nicvf_netdev_qidx(nic, cq_idx)); + if (tx_pkts) + netdev_tx_completed_queue(txq, tx_pkts, tx_bytes); + nic = nic->pnicvf; if (netif_tx_queue_stopped(txq) && netif_carrier_ok(netdev)) { netif_tx_start_queue(txq); @@ -928,16 +950,19 @@ static int nicvf_register_interrupts(struct nicvf *nic) int vector; for_each_cq_irq(irq) - sprintf(nic->irq_name[irq], "NICVF%d CQ%d", - nic->vf_id, irq); + sprintf(nic->irq_name[irq], "%s-rxtx-%d", + nic->pnicvf->netdev->name, + nicvf_netdev_qidx(nic, irq)); for_each_sq_irq(irq) - sprintf(nic->irq_name[irq], "NICVF%d SQ%d", - nic->vf_id, irq - NICVF_INTR_ID_SQ); + sprintf(nic->irq_name[irq], "%s-sq-%d", + nic->pnicvf->netdev->name, + nicvf_netdev_qidx(nic, irq - NICVF_INTR_ID_SQ)); for_each_rbdr_irq(irq) - sprintf(nic->irq_name[irq], "NICVF%d RBDR%d", - nic->vf_id, irq - NICVF_INTR_ID_RBDR); + sprintf(nic->irq_name[irq], "%s-rbdr-%d", + nic->pnicvf->netdev->name, + nic->sqs_mode ? (nic->sqs_id + 1) : 0); /* Register CQ interrupts */ for (irq = 0; irq < nic->qs->cq_cnt; irq++) { @@ -961,8 +986,9 @@ static int nicvf_register_interrupts(struct nicvf *nic) } /* Register QS error interrupt */ - sprintf(nic->irq_name[NICVF_INTR_ID_QS_ERR], - "NICVF%d Qset error", nic->vf_id); + sprintf(nic->irq_name[NICVF_INTR_ID_QS_ERR], "%s-qset-err-%d", + nic->pnicvf->netdev->name, + nic->sqs_mode ? (nic->sqs_id + 1) : 0); irq = NICVF_INTR_ID_QS_ERR; ret = request_irq(nic->msix_entries[irq].vector, nicvf_qs_err_intr_handler, @@ -1141,6 +1167,9 @@ int nicvf_stop(struct net_device *netdev) netif_tx_disable(netdev); + for (qidx = 0; qidx < netdev->num_tx_queues; qidx++) + netdev_tx_reset_queue(netdev_get_tx_queue(netdev, qidx)); + /* Free resources */ nicvf_config_data_transfer(nic, false); @@ -1191,7 +1220,7 @@ int nicvf_open(struct net_device *netdev) } /* Check if we got MAC address from PF or else generate a radom MAC */ - if (is_zero_ether_addr(netdev->dev_addr)) { + if (!nic->sqs_mode && is_zero_ether_addr(netdev->dev_addr)) { eth_hw_addr_random(netdev); nicvf_hw_set_mac_addr(nic, netdev); } @@ -1502,6 +1531,7 @@ static int nicvf_probe(struct pci_dev *pdev, const struct pci_device_id *ent) struct net_device *netdev; struct nicvf *nic; int err, qcount; + u16 sdevid; err = pci_enable_device(pdev); if (err) { @@ -1527,14 +1557,13 @@ static int nicvf_probe(struct pci_dev *pdev, const struct pci_device_id *ent) goto err_release_regions; } - qcount = MAX_CMP_QUEUES_PER_QS; + qcount = netif_get_num_default_rss_queues(); /* Restrict multiqset support only for host bound VFs */ if (pdev->is_virtfn) { /* Set max number of queues per VF */ - qcount = roundup(num_online_cpus(), MAX_CMP_QUEUES_PER_QS); - qcount = min(qcount, - (MAX_SQS_PER_VF + 1) * MAX_CMP_QUEUES_PER_QS); + qcount = min_t(int, num_online_cpus(), + (MAX_SQS_PER_VF + 1) * MAX_CMP_QUEUES_PER_QS); } netdev = alloc_etherdev_mqs(sizeof(struct nicvf), qcount, qcount); @@ -1575,6 +1604,10 @@ static int nicvf_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (!pass1_silicon(nic->pdev)) nic->hw_tso = true; + pci_read_config_word(nic->pdev, PCI_SUBSYSTEM_ID, &sdevid); + if (sdevid == 0xA134) + nic->t88 = true; + /* Check if this VF is in QS only mode */ if (nic->sqs_mode) return 0; diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_queues.c b/drivers/net/ethernet/cavium/thunder/nicvf_queues.c index 0ff8e60deccb..a4fc50155881 100644 --- a/drivers/net/ethernet/cavium/thunder/nicvf_queues.c +++ b/drivers/net/ethernet/cavium/thunder/nicvf_queues.c @@ -479,6 +479,16 @@ void nicvf_config_vlan_stripping(struct nicvf *nic, netdev_features_t features) NIC_QSET_RQ_GEN_CFG, 0, rq_cfg); } +static void nicvf_reset_rcv_queue_stats(struct nicvf *nic) +{ + union nic_mbx mbx = {}; + + /* Reset all RXQ's stats */ + mbx.reset_stat.msg = NIC_MBOX_MSG_RESET_STAT_COUNTER; + mbx.reset_stat.rq_stat_mask = 0xFFFF; + nicvf_send_msg_to_pf(nic, &mbx); +} + /* Configures receive queue */ static void nicvf_rcv_queue_config(struct nicvf *nic, struct queue_set *qs, int qidx, bool enable) @@ -762,10 +772,10 @@ int nicvf_set_qset_resources(struct nicvf *nic) nic->qs = qs; /* Set count of each queue */ - qs->rbdr_cnt = RBDR_CNT; - qs->rq_cnt = RCV_QUEUE_CNT; - qs->sq_cnt = SND_QUEUE_CNT; - qs->cq_cnt = CMP_QUEUE_CNT; + qs->rbdr_cnt = DEFAULT_RBDR_CNT; + qs->rq_cnt = min_t(u8, MAX_RCV_QUEUES_PER_QS, num_online_cpus()); + qs->sq_cnt = min_t(u8, MAX_SND_QUEUES_PER_QS, num_online_cpus()); + qs->cq_cnt = max_t(u8, qs->rq_cnt, qs->sq_cnt); /* Set queue lengths */ qs->rbdr_len = RCV_BUF_COUNT; @@ -812,6 +822,11 @@ int nicvf_config_data_transfer(struct nicvf *nic, bool enable) nicvf_free_resources(nic); } + /* Reset RXQ's stats. + * SQ's stats will get reset automatically once SQ is reset. + */ + nicvf_reset_rcv_queue_stats(nic); + return 0; } @@ -938,6 +953,8 @@ static int nicvf_tso_count_subdescs(struct sk_buff *skb) return num_edescs + sh->gso_segs; } +#define POST_CQE_DESC_COUNT 2 + /* Get the number of SQ descriptors needed to xmit this skb */ static int nicvf_sq_subdesc_required(struct nicvf *nic, struct sk_buff *skb) { @@ -948,6 +965,10 @@ static int nicvf_sq_subdesc_required(struct nicvf *nic, struct sk_buff *skb) return subdesc_cnt; } + /* Dummy descriptors to get TSO pkt completion notification */ + if (nic->t88 && nic->hw_tso && skb_shinfo(skb)->gso_size) + subdesc_cnt += POST_CQE_DESC_COUNT; + if (skb_shinfo(skb)->nr_frags) subdesc_cnt += skb_shinfo(skb)->nr_frags; @@ -965,14 +986,21 @@ nicvf_sq_add_hdr_subdesc(struct nicvf *nic, struct snd_queue *sq, int qentry, struct sq_hdr_subdesc *hdr; hdr = (struct sq_hdr_subdesc *)GET_SQ_DESC(sq, qentry); - sq->skbuff[qentry] = (u64)skb; - memset(hdr, 0, SND_QUEUE_DESC_SIZE); hdr->subdesc_type = SQ_DESC_TYPE_HEADER; - /* Enable notification via CQE after processing SQE */ - hdr->post_cqe = 1; - /* No of subdescriptors following this */ - hdr->subdesc_cnt = subdesc_cnt; + + if (nic->t88 && nic->hw_tso && skb_shinfo(skb)->gso_size) { + /* post_cqe = 0, to avoid HW posting a CQE for every TSO + * segment transmitted on 88xx. + */ + hdr->subdesc_cnt = subdesc_cnt - POST_CQE_DESC_COUNT; + } else { + sq->skbuff[qentry] = (u64)skb; + /* Enable notification via CQE after processing SQE */ + hdr->post_cqe = 1; + /* No of subdescriptors following this */ + hdr->subdesc_cnt = subdesc_cnt; + } hdr->tot_len = len; /* Offload checksum calculation to HW */ @@ -1023,6 +1051,55 @@ static inline void nicvf_sq_add_gather_subdesc(struct snd_queue *sq, int qentry, gather->addr = data; } +/* Add HDR + IMMEDIATE subdescriptors right after descriptors of a TSO + * packet so that a CQE is posted as a notifation for transmission of + * TSO packet. + */ +static inline void nicvf_sq_add_cqe_subdesc(struct snd_queue *sq, int qentry, + int tso_sqe, struct sk_buff *skb) +{ + struct sq_imm_subdesc *imm; + struct sq_hdr_subdesc *hdr; + + sq->skbuff[qentry] = (u64)skb; + + hdr = (struct sq_hdr_subdesc *)GET_SQ_DESC(sq, qentry); + memset(hdr, 0, SND_QUEUE_DESC_SIZE); + hdr->subdesc_type = SQ_DESC_TYPE_HEADER; + /* Enable notification via CQE after processing SQE */ + hdr->post_cqe = 1; + /* There is no packet to transmit here */ + hdr->dont_send = 1; + hdr->subdesc_cnt = POST_CQE_DESC_COUNT - 1; + hdr->tot_len = 1; + /* Actual TSO header SQE index, needed for cleanup */ + hdr->rsvd2 = tso_sqe; + + qentry = nicvf_get_nxt_sqentry(sq, qentry); + imm = (struct sq_imm_subdesc *)GET_SQ_DESC(sq, qentry); + memset(imm, 0, SND_QUEUE_DESC_SIZE); + imm->subdesc_type = SQ_DESC_TYPE_IMMEDIATE; + imm->len = 1; +} + +static inline void nicvf_sq_doorbell(struct nicvf *nic, struct sk_buff *skb, + int sq_num, int desc_cnt) +{ + struct netdev_queue *txq; + + txq = netdev_get_tx_queue(nic->pnicvf->netdev, + skb_get_queue_mapping(skb)); + + netdev_tx_sent_queue(txq, skb->len); + + /* make sure all memory stores are done before ringing doorbell */ + smp_wmb(); + + /* Inform HW to xmit all TSO segments */ + nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_DOOR, + sq_num, desc_cnt); +} + /* Segment a TSO packet into 'gso_size' segments and append * them to SQ for transfer */ @@ -1082,12 +1159,8 @@ static int nicvf_sq_append_tso(struct nicvf *nic, struct snd_queue *sq, /* Save SKB in the last segment for freeing */ sq->skbuff[hdr_qentry] = (u64)skb; - /* make sure all memory stores are done before ringing doorbell */ - smp_wmb(); + nicvf_sq_doorbell(nic, skb, sq_num, desc_cnt); - /* Inform HW to xmit all TSO segments */ - nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_DOOR, - sq_num, desc_cnt); nic->drv_stats.tx_tso++; return 1; } @@ -1096,7 +1169,7 @@ static int nicvf_sq_append_tso(struct nicvf *nic, struct snd_queue *sq, int nicvf_sq_append_skb(struct nicvf *nic, struct sk_buff *skb) { int i, size; - int subdesc_cnt; + int subdesc_cnt, tso_sqe = 0; int sq_num, qentry; struct queue_set *qs; struct snd_queue *sq; @@ -1131,6 +1204,7 @@ int nicvf_sq_append_skb(struct nicvf *nic, struct sk_buff *skb) /* Add SQ header subdesc */ nicvf_sq_add_hdr_subdesc(nic, sq, qentry, subdesc_cnt - 1, skb, skb->len); + tso_sqe = qentry; /* Add SQ gather subdescs */ qentry = nicvf_get_nxt_sqentry(sq, qentry); @@ -1154,12 +1228,13 @@ int nicvf_sq_append_skb(struct nicvf *nic, struct sk_buff *skb) } doorbell: - /* make sure all memory stores are done before ringing doorbell */ - smp_wmb(); + if (nic->t88 && skb_shinfo(skb)->gso_size) { + qentry = nicvf_get_nxt_sqentry(sq, qentry); + nicvf_sq_add_cqe_subdesc(sq, qentry, tso_sqe, skb); + } + + nicvf_sq_doorbell(nic, skb, sq_num, subdesc_cnt); - /* Inform HW to xmit new packet */ - nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_DOOR, - sq_num, subdesc_cnt); return 1; append_fail: @@ -1184,13 +1259,23 @@ struct sk_buff *nicvf_get_rcv_skb(struct nicvf *nic, struct cqe_rx_t *cqe_rx) int frag; int payload_len = 0; struct sk_buff *skb = NULL; - struct sk_buff *skb_frag = NULL; - struct sk_buff *prev_frag = NULL; + struct page *page; + int offset; u16 *rb_lens = NULL; u64 *rb_ptrs = NULL; rb_lens = (void *)cqe_rx + (3 * sizeof(u64)); - rb_ptrs = (void *)cqe_rx + (6 * sizeof(u64)); + /* Except 88xx pass1 on all other chips CQE_RX2_S is added to + * CQE_RX at word6, hence buffer pointers move by word + * + * Use existing 'hw_tso' flag which will be set for all chips + * except 88xx pass1 instead of a additional cache line + * access (or miss) by using pci dev's revision. + */ + if (!nic->hw_tso) + rb_ptrs = (void *)cqe_rx + (6 * sizeof(u64)); + else + rb_ptrs = (void *)cqe_rx + (7 * sizeof(u64)); netdev_dbg(nic->netdev, "%s rb_cnt %d rb0_ptr %llx rb0_sz %d\n", __func__, cqe_rx->rb_cnt, cqe_rx->rb0_ptr, cqe_rx->rb0_sz); @@ -1208,22 +1293,10 @@ struct sk_buff *nicvf_get_rcv_skb(struct nicvf *nic, struct cqe_rx_t *cqe_rx) skb_put(skb, payload_len); } else { /* Add fragments */ - skb_frag = nicvf_rb_ptr_to_skb(nic, *rb_ptrs, - payload_len); - if (!skb_frag) { - dev_kfree_skb(skb); - return NULL; - } - - if (!skb_shinfo(skb)->frag_list) - skb_shinfo(skb)->frag_list = skb_frag; - else - prev_frag->next = skb_frag; - - prev_frag = skb_frag; - skb->len += payload_len; - skb->data_len += payload_len; - skb_frag->len = payload_len; + page = virt_to_page(phys_to_virt(*rb_ptrs)); + offset = phys_to_virt(*rb_ptrs) - page_address(page); + skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, page, + offset, payload_len, RCV_FRAG_LEN); } /* Next buffer pointer */ rb_ptrs++; diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_queues.h b/drivers/net/ethernet/cavium/thunder/nicvf_queues.h index 6673e1133523..869f3386028b 100644 --- a/drivers/net/ethernet/cavium/thunder/nicvf_queues.h +++ b/drivers/net/ethernet/cavium/thunder/nicvf_queues.h @@ -57,10 +57,7 @@ #define CMP_QUEUE_SIZE6 6ULL /* 64K entries */ /* Default queue count per QS, its lengths and threshold values */ -#define RBDR_CNT 1 -#define RCV_QUEUE_CNT 8 -#define SND_QUEUE_CNT 8 -#define CMP_QUEUE_CNT 8 /* Max of RCV and SND qcount */ +#define DEFAULT_RBDR_CNT 1 #define SND_QSIZE SND_QUEUE_SIZE2 #define SND_QUEUE_LEN (1ULL << (SND_QSIZE + 10)) diff --git a/drivers/net/ethernet/cavium/thunder/thunder_bgx.c b/drivers/net/ethernet/cavium/thunder/thunder_bgx.c index 63a39ac97d53..8bbaedbb7b94 100644 --- a/drivers/net/ethernet/cavium/thunder/thunder_bgx.c +++ b/drivers/net/ethernet/cavium/thunder/thunder_bgx.c @@ -28,6 +28,9 @@ struct lmac { struct bgx *bgx; int dmac; u8 mac[ETH_ALEN]; + u8 lmac_type; + u8 lane_to_sds; + bool use_training; bool link_up; int lmacid; /* ID within BGX */ int lmacid_bd; /* ID on board */ @@ -43,14 +46,13 @@ struct lmac { struct bgx { u8 bgx_id; - u8 qlm_mode; struct lmac lmac[MAX_LMAC_PER_BGX]; int lmac_count; - int lmac_type; - int lane_to_sds; - int use_training; + u8 max_lmac; void __iomem *reg_base; struct pci_dev *pdev; + bool is_dlm; + bool is_rgx; }; static struct bgx *bgx_vnic[MAX_BGX_THUNDER]; @@ -61,6 +63,7 @@ static int bgx_xaui_check_link(struct lmac *lmac); /* Supported devices */ static const struct pci_device_id bgx_id_table[] = { { PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_BGX) }, + { PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_RGX) }, { 0, } /* end of table */ }; @@ -124,8 +127,8 @@ unsigned bgx_get_map(int node) int i; unsigned map = 0; - for (i = 0; i < MAX_BGX_PER_CN88XX; i++) { - if (bgx_vnic[(node * MAX_BGX_PER_CN88XX) + i]) + for (i = 0; i < MAX_BGX_PER_NODE; i++) { + if (bgx_vnic[(node * MAX_BGX_PER_NODE) + i]) map |= (1 << i); } @@ -138,7 +141,7 @@ int bgx_get_lmac_count(int node, int bgx_idx) { struct bgx *bgx; - bgx = bgx_vnic[(node * MAX_BGX_PER_CN88XX) + bgx_idx]; + bgx = bgx_vnic[(node * MAX_BGX_PER_NODE) + bgx_idx]; if (bgx) return bgx->lmac_count; @@ -153,7 +156,7 @@ void bgx_get_lmac_link_state(int node, int bgx_idx, int lmacid, void *status) struct bgx *bgx; struct lmac *lmac; - bgx = bgx_vnic[(node * MAX_BGX_PER_CN88XX) + bgx_idx]; + bgx = bgx_vnic[(node * MAX_BGX_PER_NODE) + bgx_idx]; if (!bgx) return; @@ -166,7 +169,7 @@ EXPORT_SYMBOL(bgx_get_lmac_link_state); const u8 *bgx_get_lmac_mac(int node, int bgx_idx, int lmacid) { - struct bgx *bgx = bgx_vnic[(node * MAX_BGX_PER_CN88XX) + bgx_idx]; + struct bgx *bgx = bgx_vnic[(node * MAX_BGX_PER_NODE) + bgx_idx]; if (bgx) return bgx->lmac[lmacid].mac; @@ -177,7 +180,7 @@ EXPORT_SYMBOL(bgx_get_lmac_mac); void bgx_set_lmac_mac(int node, int bgx_idx, int lmacid, const u8 *mac) { - struct bgx *bgx = bgx_vnic[(node * MAX_BGX_PER_CN88XX) + bgx_idx]; + struct bgx *bgx = bgx_vnic[(node * MAX_BGX_PER_NODE) + bgx_idx]; if (!bgx) return; @@ -188,11 +191,13 @@ EXPORT_SYMBOL(bgx_set_lmac_mac); void bgx_lmac_rx_tx_enable(int node, int bgx_idx, int lmacid, bool enable) { - struct bgx *bgx = bgx_vnic[(node * MAX_BGX_PER_CN88XX) + bgx_idx]; + struct bgx *bgx = bgx_vnic[(node * MAX_BGX_PER_NODE) + bgx_idx]; + struct lmac *lmac; u64 cfg; if (!bgx) return; + lmac = &bgx->lmac[lmacid]; cfg = bgx_reg_read(bgx, lmacid, BGX_CMRX_CFG); if (enable) @@ -200,6 +205,9 @@ void bgx_lmac_rx_tx_enable(int node, int bgx_idx, int lmacid, bool enable) else cfg &= ~(CMR_PKT_RX_EN | CMR_PKT_TX_EN); bgx_reg_write(bgx, lmacid, BGX_CMRX_CFG, cfg); + + if (bgx->is_rgx) + xcv_setup_link(enable ? lmac->link_up : 0, lmac->last_speed); } EXPORT_SYMBOL(bgx_lmac_rx_tx_enable); @@ -266,9 +274,12 @@ static void bgx_sgmii_change_link_state(struct lmac *lmac) port_cfg = bgx_reg_read(bgx, lmac->lmacid, BGX_GMP_GMI_PRTX_CFG); - /* renable lmac */ + /* Re-enable lmac */ cmr_cfg |= CMR_EN; bgx_reg_write(bgx, lmac->lmacid, BGX_CMRX_CFG, cmr_cfg); + + if (bgx->is_rgx && (cmr_cfg & (CMR_PKT_RX_EN | CMR_PKT_TX_EN))) + xcv_setup_link(lmac->link_up, lmac->last_speed); } static void bgx_lmac_handler(struct net_device *netdev) @@ -314,7 +325,7 @@ u64 bgx_get_rx_stats(int node, int bgx_idx, int lmac, int idx) { struct bgx *bgx; - bgx = bgx_vnic[(node * MAX_BGX_PER_CN88XX) + bgx_idx]; + bgx = bgx_vnic[(node * MAX_BGX_PER_NODE) + bgx_idx]; if (!bgx) return 0; @@ -328,7 +339,7 @@ u64 bgx_get_tx_stats(int node, int bgx_idx, int lmac, int idx) { struct bgx *bgx; - bgx = bgx_vnic[(node * MAX_BGX_PER_CN88XX) + bgx_idx]; + bgx = bgx_vnic[(node * MAX_BGX_PER_NODE) + bgx_idx]; if (!bgx) return 0; @@ -356,7 +367,7 @@ void bgx_lmac_internal_loopback(int node, int bgx_idx, struct lmac *lmac; u64 cfg; - bgx = bgx_vnic[(node * MAX_BGX_PER_CN88XX) + bgx_idx]; + bgx = bgx_vnic[(node * MAX_BGX_PER_NODE) + bgx_idx]; if (!bgx) return; @@ -379,8 +390,9 @@ void bgx_lmac_internal_loopback(int node, int bgx_idx, } EXPORT_SYMBOL(bgx_lmac_internal_loopback); -static int bgx_lmac_sgmii_init(struct bgx *bgx, int lmacid) +static int bgx_lmac_sgmii_init(struct bgx *bgx, struct lmac *lmac) { + int lmacid = lmac->lmacid; u64 cfg; bgx_reg_modify(bgx, lmacid, BGX_GMP_GMI_TXX_THRESH, 0x30); @@ -409,18 +421,29 @@ static int bgx_lmac_sgmii_init(struct bgx *bgx, int lmacid) cfg |= (PCS_MRX_CTL_RST_AN | PCS_MRX_CTL_AN_EN); bgx_reg_write(bgx, lmacid, BGX_GMP_PCS_MRX_CTL, cfg); - if (bgx_poll_reg(bgx, lmacid, BGX_GMP_PCS_MRX_STATUS, - PCS_MRX_STATUS_AN_CPT, false)) { - dev_err(&bgx->pdev->dev, "BGX AN_CPT not completed\n"); - return -1; + if (lmac->lmac_type == BGX_MODE_QSGMII) { + /* Disable disparity check for QSGMII */ + cfg = bgx_reg_read(bgx, lmacid, BGX_GMP_PCS_MISCX_CTL); + cfg &= ~PCS_MISC_CTL_DISP_EN; + bgx_reg_write(bgx, lmacid, BGX_GMP_PCS_MISCX_CTL, cfg); + return 0; + } + + if (lmac->lmac_type == BGX_MODE_SGMII) { + if (bgx_poll_reg(bgx, lmacid, BGX_GMP_PCS_MRX_STATUS, + PCS_MRX_STATUS_AN_CPT, false)) { + dev_err(&bgx->pdev->dev, "BGX AN_CPT not completed\n"); + return -1; + } } return 0; } -static int bgx_lmac_xaui_init(struct bgx *bgx, int lmacid, int lmac_type) +static int bgx_lmac_xaui_init(struct bgx *bgx, struct lmac *lmac) { u64 cfg; + int lmacid = lmac->lmacid; /* Reset SPU */ bgx_reg_modify(bgx, lmacid, BGX_SPUX_CONTROL1, SPU_CTL_RESET); @@ -436,12 +459,14 @@ static int bgx_lmac_xaui_init(struct bgx *bgx, int lmacid, int lmac_type) bgx_reg_modify(bgx, lmacid, BGX_SPUX_CONTROL1, SPU_CTL_LOW_POWER); /* Set interleaved running disparity for RXAUI */ - if (bgx->lmac_type != BGX_MODE_RXAUI) - bgx_reg_modify(bgx, lmacid, - BGX_SPUX_MISC_CONTROL, SPU_MISC_CTL_RX_DIS); - else + if (lmac->lmac_type == BGX_MODE_RXAUI) bgx_reg_modify(bgx, lmacid, BGX_SPUX_MISC_CONTROL, - SPU_MISC_CTL_RX_DIS | SPU_MISC_CTL_INTLV_RDISP); + SPU_MISC_CTL_INTLV_RDISP); + + /* Clear receive packet disable */ + cfg = bgx_reg_read(bgx, lmacid, BGX_SPUX_MISC_CONTROL); + cfg &= ~SPU_MISC_CTL_RX_DIS; + bgx_reg_write(bgx, lmacid, BGX_SPUX_MISC_CONTROL, cfg); /* clear all interrupts */ cfg = bgx_reg_read(bgx, lmacid, BGX_SMUX_RX_INT); @@ -451,7 +476,7 @@ static int bgx_lmac_xaui_init(struct bgx *bgx, int lmacid, int lmac_type) cfg = bgx_reg_read(bgx, lmacid, BGX_SPUX_INT); bgx_reg_write(bgx, lmacid, BGX_SPUX_INT, cfg); - if (bgx->use_training) { + if (lmac->use_training) { bgx_reg_write(bgx, lmacid, BGX_SPUX_BR_PMD_LP_CUP, 0x00); bgx_reg_write(bgx, lmacid, BGX_SPUX_BR_PMD_LD_CUP, 0x00); bgx_reg_write(bgx, lmacid, BGX_SPUX_BR_PMD_LD_REP, 0x00); @@ -474,9 +499,9 @@ static int bgx_lmac_xaui_init(struct bgx *bgx, int lmacid, int lmac_type) bgx_reg_write(bgx, lmacid, BGX_SPUX_AN_CONTROL, cfg); cfg = bgx_reg_read(bgx, lmacid, BGX_SPUX_AN_ADV); - if (bgx->lmac_type == BGX_MODE_10G_KR) + if (lmac->lmac_type == BGX_MODE_10G_KR) cfg |= (1 << 23); - else if (bgx->lmac_type == BGX_MODE_40G_KR) + else if (lmac->lmac_type == BGX_MODE_40G_KR) cfg |= (1 << 24); else cfg &= ~((1 << 23) | (1 << 24)); @@ -511,11 +536,10 @@ static int bgx_xaui_check_link(struct lmac *lmac) { struct bgx *bgx = lmac->bgx; int lmacid = lmac->lmacid; - int lmac_type = bgx->lmac_type; + int lmac_type = lmac->lmac_type; u64 cfg; - bgx_reg_modify(bgx, lmacid, BGX_SPUX_MISC_CONTROL, SPU_MISC_CTL_RX_DIS); - if (bgx->use_training) { + if (lmac->use_training) { cfg = bgx_reg_read(bgx, lmacid, BGX_SPUX_INT); if (!(cfg & (1ull << 13))) { cfg = (1ull << 13) | (1ull << 14); @@ -556,7 +580,7 @@ static int bgx_xaui_check_link(struct lmac *lmac) BGX_SPUX_STATUS2, SPU_STATUS2_RCVFLT); if (bgx_reg_read(bgx, lmacid, BGX_SPUX_STATUS2) & SPU_STATUS2_RCVFLT) { dev_err(&bgx->pdev->dev, "Receive fault, retry training\n"); - if (bgx->use_training) { + if (lmac->use_training) { cfg = bgx_reg_read(bgx, lmacid, BGX_SPUX_INT); if (!(cfg & (1ull << 13))) { cfg = (1ull << 13) | (1ull << 14); @@ -584,11 +608,6 @@ static int bgx_xaui_check_link(struct lmac *lmac) return -1; } - /* Clear receive packet disable */ - cfg = bgx_reg_read(bgx, lmacid, BGX_SPUX_MISC_CONTROL); - cfg &= ~SPU_MISC_CTL_RX_DIS; - bgx_reg_write(bgx, lmacid, BGX_SPUX_MISC_CONTROL, cfg); - /* Check for MAC RX faults */ cfg = bgx_reg_read(bgx, lmacid, BGX_SMUX_RX_CTL); /* 0 - Link is okay, 1 - Local fault, 2 - Remote fault */ @@ -599,7 +618,7 @@ static int bgx_xaui_check_link(struct lmac *lmac) /* Rx local/remote fault seen. * Do lmac reinit to see if condition recovers */ - bgx_lmac_xaui_init(bgx, lmacid, bgx->lmac_type); + bgx_lmac_xaui_init(bgx, lmac); return -1; } @@ -623,7 +642,7 @@ static void bgx_poll_for_link(struct work_struct *work) if ((spu_link & SPU_STATUS1_RCV_LNK) && !(smu_link & SMU_RX_CTL_STATUS)) { lmac->link_up = 1; - if (lmac->bgx->lmac_type == BGX_MODE_XLAUI) + if (lmac->lmac_type == BGX_MODE_XLAUI) lmac->last_speed = 40000; else lmac->last_speed = 10000; @@ -649,6 +668,16 @@ static void bgx_poll_for_link(struct work_struct *work) queue_delayed_work(lmac->check_link, &lmac->dwork, HZ * 2); } +static int phy_interface_mode(u8 lmac_type) +{ + if (lmac_type == BGX_MODE_QSGMII) + return PHY_INTERFACE_MODE_QSGMII; + if (lmac_type == BGX_MODE_RGMII) + return PHY_INTERFACE_MODE_RGMII; + + return PHY_INTERFACE_MODE_SGMII; +} + static int bgx_lmac_enable(struct bgx *bgx, u8 lmacid) { struct lmac *lmac; @@ -657,13 +686,15 @@ static int bgx_lmac_enable(struct bgx *bgx, u8 lmacid) lmac = &bgx->lmac[lmacid]; lmac->bgx = bgx; - if (bgx->lmac_type == BGX_MODE_SGMII) { + if ((lmac->lmac_type == BGX_MODE_SGMII) || + (lmac->lmac_type == BGX_MODE_QSGMII) || + (lmac->lmac_type == BGX_MODE_RGMII)) { lmac->is_sgmii = 1; - if (bgx_lmac_sgmii_init(bgx, lmacid)) + if (bgx_lmac_sgmii_init(bgx, lmac)) return -1; } else { lmac->is_sgmii = 0; - if (bgx_lmac_xaui_init(bgx, lmacid, bgx->lmac_type)) + if (bgx_lmac_xaui_init(bgx, lmac)) return -1; } @@ -685,10 +716,10 @@ static int bgx_lmac_enable(struct bgx *bgx, u8 lmacid) /* Restore default cfg, incase low level firmware changed it */ bgx_reg_write(bgx, lmacid, BGX_CMRX_RX_DMAC_CTL, 0x03); - if ((bgx->lmac_type != BGX_MODE_XFI) && - (bgx->lmac_type != BGX_MODE_XLAUI) && - (bgx->lmac_type != BGX_MODE_40G_KR) && - (bgx->lmac_type != BGX_MODE_10G_KR)) { + if ((lmac->lmac_type != BGX_MODE_XFI) && + (lmac->lmac_type != BGX_MODE_XLAUI) && + (lmac->lmac_type != BGX_MODE_40G_KR) && + (lmac->lmac_type != BGX_MODE_10G_KR)) { if (!lmac->phydev) return -ENODEV; @@ -696,7 +727,7 @@ static int bgx_lmac_enable(struct bgx *bgx, u8 lmacid) if (phy_connect_direct(&lmac->netdev, lmac->phydev, bgx_lmac_handler, - PHY_INTERFACE_MODE_SGMII)) + phy_interface_mode(lmac->lmac_type))) return -ENODEV; phy_start_aneg(lmac->phydev); @@ -753,76 +784,19 @@ static void bgx_lmac_disable(struct bgx *bgx, u8 lmacid) bgx_flush_dmac_addrs(bgx, lmacid); - if ((bgx->lmac_type != BGX_MODE_XFI) && - (bgx->lmac_type != BGX_MODE_XLAUI) && - (bgx->lmac_type != BGX_MODE_40G_KR) && - (bgx->lmac_type != BGX_MODE_10G_KR) && lmac->phydev) + if ((lmac->lmac_type != BGX_MODE_XFI) && + (lmac->lmac_type != BGX_MODE_XLAUI) && + (lmac->lmac_type != BGX_MODE_40G_KR) && + (lmac->lmac_type != BGX_MODE_10G_KR) && lmac->phydev) phy_disconnect(lmac->phydev); lmac->phydev = NULL; } -static void bgx_set_num_ports(struct bgx *bgx) -{ - u64 lmac_count; - - switch (bgx->qlm_mode) { - case QLM_MODE_SGMII: - bgx->lmac_count = 4; - bgx->lmac_type = BGX_MODE_SGMII; - bgx->lane_to_sds = 0; - break; - case QLM_MODE_XAUI_1X4: - bgx->lmac_count = 1; - bgx->lmac_type = BGX_MODE_XAUI; - bgx->lane_to_sds = 0xE4; - break; - case QLM_MODE_RXAUI_2X2: - bgx->lmac_count = 2; - bgx->lmac_type = BGX_MODE_RXAUI; - bgx->lane_to_sds = 0xE4; - break; - case QLM_MODE_XFI_4X1: - bgx->lmac_count = 4; - bgx->lmac_type = BGX_MODE_XFI; - bgx->lane_to_sds = 0; - break; - case QLM_MODE_XLAUI_1X4: - bgx->lmac_count = 1; - bgx->lmac_type = BGX_MODE_XLAUI; - bgx->lane_to_sds = 0xE4; - break; - case QLM_MODE_10G_KR_4X1: - bgx->lmac_count = 4; - bgx->lmac_type = BGX_MODE_10G_KR; - bgx->lane_to_sds = 0; - bgx->use_training = 1; - break; - case QLM_MODE_40G_KR4_1X4: - bgx->lmac_count = 1; - bgx->lmac_type = BGX_MODE_40G_KR; - bgx->lane_to_sds = 0xE4; - bgx->use_training = 1; - break; - default: - bgx->lmac_count = 0; - break; - } - - /* Check if low level firmware has programmed LMAC count - * based on board type, if yes consider that otherwise - * the default static values - */ - lmac_count = bgx_reg_read(bgx, 0, BGX_CMR_RX_LMACS) & 0x7; - if (lmac_count != 4) - bgx->lmac_count = lmac_count; -} - static void bgx_init_hw(struct bgx *bgx) { int i; - - bgx_set_num_ports(bgx); + struct lmac *lmac; bgx_reg_modify(bgx, 0, BGX_CMR_GLOBAL_CFG, CMR_GLOBAL_CFG_FCS_STRIP); if (bgx_reg_read(bgx, 0, BGX_CMR_BIST_STATUS)) @@ -830,17 +804,9 @@ static void bgx_init_hw(struct bgx *bgx) /* Set lmac type and lane2serdes mapping */ for (i = 0; i < bgx->lmac_count; i++) { - if (bgx->lmac_type == BGX_MODE_RXAUI) { - if (i) - bgx->lane_to_sds = 0x0e; - else - bgx->lane_to_sds = 0x04; - bgx_reg_write(bgx, i, BGX_CMRX_CFG, - (bgx->lmac_type << 8) | bgx->lane_to_sds); - continue; - } + lmac = &bgx->lmac[i]; bgx_reg_write(bgx, i, BGX_CMRX_CFG, - (bgx->lmac_type << 8) | (bgx->lane_to_sds + i)); + (lmac->lmac_type << 8) | lmac->lane_to_sds); bgx->lmac[i].lmacid_bd = lmac_count; lmac_count++; } @@ -863,55 +829,212 @@ static void bgx_init_hw(struct bgx *bgx) bgx_reg_write(bgx, 0, BGX_CMR_RX_STREERING + (i * 8), 0x00); } -static void bgx_get_qlm_mode(struct bgx *bgx) +static u8 bgx_get_lane2sds_cfg(struct bgx *bgx, struct lmac *lmac) +{ + return (u8)(bgx_reg_read(bgx, lmac->lmacid, BGX_CMRX_CFG) & 0xFF); +} + +static void bgx_print_qlm_mode(struct bgx *bgx, u8 lmacid) { struct device *dev = &bgx->pdev->dev; - int lmac_type; - int train_en; + struct lmac *lmac; + char str[20]; + u8 dlm; - /* Read LMAC0 type to figure out QLM mode - * This is configured by low level firmware - */ - lmac_type = bgx_reg_read(bgx, 0, BGX_CMRX_CFG); - lmac_type = (lmac_type >> 8) & 0x07; + if (lmacid > bgx->max_lmac) + return; - train_en = bgx_reg_read(bgx, 0, BGX_SPUX_BR_PMD_CRTL) & - SPU_PMD_CRTL_TRAIN_EN; + lmac = &bgx->lmac[lmacid]; + dlm = (lmacid / 2) + (bgx->bgx_id * 2); + if (!bgx->is_dlm) + sprintf(str, "BGX%d QLM mode", bgx->bgx_id); + else + sprintf(str, "BGX%d DLM%d mode", bgx->bgx_id, dlm); - switch (lmac_type) { + switch (lmac->lmac_type) { case BGX_MODE_SGMII: - bgx->qlm_mode = QLM_MODE_SGMII; - dev_info(dev, "BGX%d QLM mode: SGMII\n", bgx->bgx_id); + dev_info(dev, "%s: SGMII\n", (char *)str); break; case BGX_MODE_XAUI: - bgx->qlm_mode = QLM_MODE_XAUI_1X4; - dev_info(dev, "BGX%d QLM mode: XAUI\n", bgx->bgx_id); + dev_info(dev, "%s: XAUI\n", (char *)str); break; case BGX_MODE_RXAUI: - bgx->qlm_mode = QLM_MODE_RXAUI_2X2; - dev_info(dev, "BGX%d QLM mode: RXAUI\n", bgx->bgx_id); + dev_info(dev, "%s: RXAUI\n", (char *)str); break; case BGX_MODE_XFI: - if (!train_en) { - bgx->qlm_mode = QLM_MODE_XFI_4X1; - dev_info(dev, "BGX%d QLM mode: XFI\n", bgx->bgx_id); - } else { - bgx->qlm_mode = QLM_MODE_10G_KR_4X1; - dev_info(dev, "BGX%d QLM mode: 10G_KR\n", bgx->bgx_id); - } + if (!lmac->use_training) + dev_info(dev, "%s: XFI\n", (char *)str); + else + dev_info(dev, "%s: 10G_KR\n", (char *)str); break; case BGX_MODE_XLAUI: - if (!train_en) { - bgx->qlm_mode = QLM_MODE_XLAUI_1X4; - dev_info(dev, "BGX%d QLM mode: XLAUI\n", bgx->bgx_id); - } else { - bgx->qlm_mode = QLM_MODE_40G_KR4_1X4; - dev_info(dev, "BGX%d QLM mode: 40G_KR4\n", bgx->bgx_id); - } + if (!lmac->use_training) + dev_info(dev, "%s: XLAUI\n", (char *)str); + else + dev_info(dev, "%s: 40G_KR4\n", (char *)str); + break; + case BGX_MODE_QSGMII: + if ((lmacid == 0) && + (bgx_get_lane2sds_cfg(bgx, lmac) != lmacid)) + return; + if ((lmacid == 2) && + (bgx_get_lane2sds_cfg(bgx, lmac) == lmacid)) + return; + dev_info(dev, "%s: QSGMII\n", (char *)str); + break; + case BGX_MODE_RGMII: + dev_info(dev, "%s: RGMII\n", (char *)str); + break; + case BGX_MODE_INVALID: + /* Nothing to do */ + break; + } +} + +static void lmac_set_lane2sds(struct bgx *bgx, struct lmac *lmac) +{ + switch (lmac->lmac_type) { + case BGX_MODE_SGMII: + case BGX_MODE_XFI: + lmac->lane_to_sds = lmac->lmacid; + break; + case BGX_MODE_XAUI: + case BGX_MODE_XLAUI: + case BGX_MODE_RGMII: + lmac->lane_to_sds = 0xE4; + break; + case BGX_MODE_RXAUI: + lmac->lane_to_sds = (lmac->lmacid) ? 0xE : 0x4; + break; + case BGX_MODE_QSGMII: + /* There is no way to determine if DLM0/2 is QSGMII or + * DLM1/3 is configured to QSGMII as bootloader will + * configure all LMACs, so take whatever is configured + * by low level firmware. + */ + lmac->lane_to_sds = bgx_get_lane2sds_cfg(bgx, lmac); break; default: - bgx->qlm_mode = QLM_MODE_SGMII; - dev_info(dev, "BGX%d QLM default mode: SGMII\n", bgx->bgx_id); + lmac->lane_to_sds = 0; + break; + } +} + +static void lmac_set_training(struct bgx *bgx, struct lmac *lmac, int lmacid) +{ + if ((lmac->lmac_type != BGX_MODE_10G_KR) && + (lmac->lmac_type != BGX_MODE_40G_KR)) { + lmac->use_training = 0; + return; + } + + lmac->use_training = bgx_reg_read(bgx, lmacid, BGX_SPUX_BR_PMD_CRTL) & + SPU_PMD_CRTL_TRAIN_EN; +} + +static void bgx_set_lmac_config(struct bgx *bgx, u8 idx) +{ + struct lmac *lmac; + struct lmac *olmac; + u64 cmr_cfg; + u8 lmac_type; + u8 lane_to_sds; + + lmac = &bgx->lmac[idx]; + + if (!bgx->is_dlm || bgx->is_rgx) { + /* Read LMAC0 type to figure out QLM mode + * This is configured by low level firmware + */ + cmr_cfg = bgx_reg_read(bgx, 0, BGX_CMRX_CFG); + lmac->lmac_type = (cmr_cfg >> 8) & 0x07; + if (bgx->is_rgx) + lmac->lmac_type = BGX_MODE_RGMII; + lmac_set_training(bgx, lmac, 0); + lmac_set_lane2sds(bgx, lmac); + return; + } + + /* On 81xx BGX can be split across 2 DLMs + * firmware programs lmac_type of LMAC0 and LMAC2 + */ + if ((idx == 0) || (idx == 2)) { + cmr_cfg = bgx_reg_read(bgx, idx, BGX_CMRX_CFG); + lmac_type = (u8)((cmr_cfg >> 8) & 0x07); + lane_to_sds = (u8)(cmr_cfg & 0xFF); + /* Check if config is not reset value */ + if ((lmac_type == 0) && (lane_to_sds == 0xE4)) + lmac->lmac_type = BGX_MODE_INVALID; + else + lmac->lmac_type = lmac_type; + lmac_set_training(bgx, lmac, lmac->lmacid); + lmac_set_lane2sds(bgx, lmac); + + /* Set LMAC type of other lmac on same DLM i.e LMAC 1/3 */ + olmac = &bgx->lmac[idx + 1]; + olmac->lmac_type = lmac->lmac_type; + lmac_set_training(bgx, olmac, olmac->lmacid); + lmac_set_lane2sds(bgx, olmac); + } +} + +static bool is_dlm0_in_bgx_mode(struct bgx *bgx) +{ + struct lmac *lmac; + + if (!bgx->is_dlm) + return true; + + lmac = &bgx->lmac[0]; + if (lmac->lmac_type == BGX_MODE_INVALID) + return false; + + return true; +} + +static void bgx_get_qlm_mode(struct bgx *bgx) +{ + struct lmac *lmac; + struct lmac *lmac01; + struct lmac *lmac23; + u8 idx; + + /* Init all LMAC's type to invalid */ + for (idx = 0; idx < bgx->max_lmac; idx++) { + lmac = &bgx->lmac[idx]; + lmac->lmacid = idx; + lmac->lmac_type = BGX_MODE_INVALID; + lmac->use_training = false; + } + + /* It is assumed that low level firmware sets this value */ + bgx->lmac_count = bgx_reg_read(bgx, 0, BGX_CMR_RX_LMACS) & 0x7; + if (bgx->lmac_count > bgx->max_lmac) + bgx->lmac_count = bgx->max_lmac; + + for (idx = 0; idx < bgx->max_lmac; idx++) + bgx_set_lmac_config(bgx, idx); + + if (!bgx->is_dlm || bgx->is_rgx) { + bgx_print_qlm_mode(bgx, 0); + return; + } + + if (bgx->lmac_count) { + bgx_print_qlm_mode(bgx, 0); + bgx_print_qlm_mode(bgx, 2); + } + + /* If DLM0 is not in BGX mode then LMAC0/1 have + * to be configured with serdes lanes of DLM1 + */ + if (is_dlm0_in_bgx_mode(bgx) || (bgx->lmac_count > 2)) + return; + for (idx = 0; idx < bgx->lmac_count; idx++) { + lmac01 = &bgx->lmac[idx]; + lmac23 = &bgx->lmac[idx + 2]; + lmac01->lmac_type = lmac23->lmac_type; + lmac01->lane_to_sds = lmac23->lane_to_sds; } } @@ -1042,7 +1165,7 @@ static int bgx_init_of_phy(struct bgx *bgx) } lmac++; - if (lmac == MAX_LMAC_PER_BGX) { + if (lmac == bgx->max_lmac) { of_node_put(node); break; } @@ -1087,6 +1210,7 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent) struct device *dev = &pdev->dev; struct bgx *bgx = NULL; u8 lmac; + u16 sdevid; bgx = devm_kzalloc(dev, sizeof(*bgx), GFP_KERNEL); if (!bgx) @@ -1115,10 +1239,30 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent) err = -ENOMEM; goto err_release_regions; } - bgx->bgx_id = (pci_resource_start(pdev, PCI_CFG_REG_BAR_NUM) >> 24) & 1; - bgx->bgx_id += nic_get_node_id(pdev) * MAX_BGX_PER_CN88XX; - bgx_vnic[bgx->bgx_id] = bgx; + pci_read_config_word(pdev, PCI_DEVICE_ID, &sdevid); + if (sdevid != PCI_DEVICE_ID_THUNDER_RGX) { + bgx->bgx_id = + (pci_resource_start(pdev, PCI_CFG_REG_BAR_NUM) >> 24) & 1; + bgx->bgx_id += nic_get_node_id(pdev) * MAX_BGX_PER_NODE; + bgx->max_lmac = MAX_LMAC_PER_BGX; + bgx_vnic[bgx->bgx_id] = bgx; + } else { + bgx->is_rgx = true; + bgx->max_lmac = 1; + bgx->bgx_id = MAX_BGX_PER_CN81XX - 1; + bgx_vnic[bgx->bgx_id] = bgx; + xcv_init_hw(); + } + + /* On 81xx all are DLMs and on 83xx there are 3 BGX QLMs and one + * BGX i.e BGX2 can be split across 2 DLMs. + */ + pci_read_config_word(pdev, PCI_SUBSYSTEM_ID, &sdevid); + if ((sdevid == PCI_SUBSYS_DEVID_81XX_BGX) || + ((sdevid == PCI_SUBSYS_DEVID_83XX_BGX) && (bgx->bgx_id == 2))) + bgx->is_dlm = true; + bgx_get_qlm_mode(bgx); err = bgx_init_phy(bgx); @@ -1133,6 +1277,8 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (err) { dev_err(dev, "BGX%d failed to enable lmac%d\n", bgx->bgx_id, lmac); + while (lmac) + bgx_lmac_disable(bgx, --lmac); goto err_enable; } } diff --git a/drivers/net/ethernet/cavium/thunder/thunder_bgx.h b/drivers/net/ethernet/cavium/thunder/thunder_bgx.h index 42010d2e5ddf..d59c71e4a000 100644 --- a/drivers/net/ethernet/cavium/thunder/thunder_bgx.h +++ b/drivers/net/ethernet/cavium/thunder/thunder_bgx.h @@ -9,8 +9,20 @@ #ifndef THUNDER_BGX_H #define THUNDER_BGX_H -#define MAX_BGX_THUNDER 8 /* Max 4 nodes, 2 per node */ +/* PCI device ID */ +#define PCI_DEVICE_ID_THUNDER_BGX 0xA026 +#define PCI_DEVICE_ID_THUNDER_RGX 0xA054 + +/* Subsystem device IDs */ +#define PCI_SUBSYS_DEVID_88XX_BGX 0xA126 +#define PCI_SUBSYS_DEVID_81XX_BGX 0xA226 +#define PCI_SUBSYS_DEVID_83XX_BGX 0xA326 + +#define MAX_BGX_THUNDER 8 /* Max 2 nodes, 4 per node */ #define MAX_BGX_PER_CN88XX 2 +#define MAX_BGX_PER_CN81XX 3 /* 2 BGXs + 1 RGX */ +#define MAX_BGX_PER_CN83XX 4 +#define MAX_BGX_PER_NODE 4 #define MAX_LMAC_PER_BGX 4 #define MAX_BGX_CHANS_PER_LMAC 16 #define MAX_DMAC_PER_LMAC 8 @@ -18,8 +30,6 @@ #define MAX_DMAC_PER_LMAC_TNS_BYPASS_MODE 2 -#define MAX_LMAC (MAX_BGX_PER_CN88XX * MAX_LMAC_PER_BGX) - /* Registers */ #define BGX_CMRX_CFG 0x00 #define CMR_PKT_TX_EN BIT_ULL(13) @@ -136,6 +146,7 @@ #define BGX_GMP_PCS_ANX_AN_RESULTS 0x30020 #define BGX_GMP_PCS_SGM_AN_ADV 0x30068 #define BGX_GMP_PCS_MISCX_CTL 0x30078 +#define PCS_MISC_CTL_DISP_EN BIT_ULL(13) #define PCS_MISC_CTL_GMX_ENO BIT_ULL(11) #define PCS_MISC_CTL_SAMP_PT_MASK 0x7Full #define BGX_GMP_GMI_PRTX_CFG 0x38020 @@ -194,6 +205,9 @@ void bgx_set_lmac_mac(int node, int bgx_idx, int lmacid, const u8 *mac); void bgx_get_lmac_link_state(int node, int bgx_idx, int lmacid, void *status); void bgx_lmac_internal_loopback(int node, int bgx_idx, int lmac_idx, bool enable); +void xcv_init_hw(void); +void xcv_setup_link(bool link_up, int link_speed); + u64 bgx_get_rx_stats(int node, int bgx_idx, int lmac, int idx); u64 bgx_get_tx_stats(int node, int bgx_idx, int lmac, int idx); #define BGX_RX_STATS_COUNT 11 @@ -213,16 +227,9 @@ enum LMAC_TYPE { BGX_MODE_XLAUI = 4, /* 4 lanes, 10.3125 Gbaud */ BGX_MODE_10G_KR = 3,/* 1 lane, 10.3125 Gbaud */ BGX_MODE_40G_KR = 4,/* 4 lanes, 10.3125 Gbaud */ -}; - -enum qlm_mode { - QLM_MODE_SGMII, /* SGMII, each lane independent */ - QLM_MODE_XAUI_1X4, /* 1 XAUI or DXAUI, 4 lanes */ - QLM_MODE_RXAUI_2X2, /* 2 RXAUI, 2 lanes each */ - QLM_MODE_XFI_4X1, /* 4 XFI, 1 lane each */ - QLM_MODE_XLAUI_1X4, /* 1 XLAUI, 4 lanes each */ - QLM_MODE_10G_KR_4X1, /* 4 10GBASE-KR, 1 lane each */ - QLM_MODE_40G_KR4_1X4, /* 1 40GBASE-KR4, 4 lanes each */ + BGX_MODE_RGMII = 5, + BGX_MODE_QSGMII = 6, + BGX_MODE_INVALID = 7, }; #endif /* THUNDER_BGX_H */ diff --git a/drivers/net/ethernet/cavium/thunder/thunder_xcv.c b/drivers/net/ethernet/cavium/thunder/thunder_xcv.c new file mode 100644 index 000000000000..67befedef709 --- /dev/null +++ b/drivers/net/ethernet/cavium/thunder/thunder_xcv.c @@ -0,0 +1,235 @@ +/* + * Copyright (C) 2016 Cavium, Inc. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of version 2 of the GNU General Public License + * as published by the Free Software Foundation. + */ + +#include <linux/acpi.h> +#include <linux/module.h> +#include <linux/interrupt.h> +#include <linux/pci.h> +#include <linux/netdevice.h> +#include <linux/etherdevice.h> +#include <linux/phy.h> +#include <linux/of.h> +#include <linux/of_mdio.h> +#include <linux/of_net.h> + +#include "nic.h" +#include "thunder_bgx.h" + +#define DRV_NAME "thunder-xcv" +#define DRV_VERSION "1.0" + +/* Register offsets */ +#define XCV_RESET 0x00 +#define PORT_EN BIT_ULL(63) +#define CLK_RESET BIT_ULL(15) +#define DLL_RESET BIT_ULL(11) +#define COMP_EN BIT_ULL(7) +#define TX_PKT_RESET BIT_ULL(3) +#define TX_DATA_RESET BIT_ULL(2) +#define RX_PKT_RESET BIT_ULL(1) +#define RX_DATA_RESET BIT_ULL(0) +#define XCV_DLL_CTL 0x10 +#define CLKRX_BYP BIT_ULL(23) +#define CLKTX_BYP BIT_ULL(15) +#define XCV_COMP_CTL 0x20 +#define DRV_BYP BIT_ULL(63) +#define XCV_CTL 0x30 +#define XCV_INT 0x40 +#define XCV_INT_W1S 0x48 +#define XCV_INT_ENA_W1C 0x50 +#define XCV_INT_ENA_W1S 0x58 +#define XCV_INBND_STATUS 0x80 +#define XCV_BATCH_CRD_RET 0x100 + +struct xcv { + void __iomem *reg_base; + struct pci_dev *pdev; +}; + +static struct xcv *xcv; + +/* Supported devices */ +static const struct pci_device_id xcv_id_table[] = { + { PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, 0xA056) }, + { 0, } /* end of table */ +}; + +MODULE_AUTHOR("Cavium Inc"); +MODULE_DESCRIPTION("Cavium Thunder RGX/XCV Driver"); +MODULE_LICENSE("GPL v2"); +MODULE_VERSION(DRV_VERSION); +MODULE_DEVICE_TABLE(pci, xcv_id_table); + +void xcv_init_hw(void) +{ + u64 cfg; + + /* Take DLL out of reset */ + cfg = readq_relaxed(xcv->reg_base + XCV_RESET); + cfg &= ~DLL_RESET; + writeq_relaxed(cfg, xcv->reg_base + XCV_RESET); + + /* Take clock tree out of reset */ + cfg = readq_relaxed(xcv->reg_base + XCV_RESET); + cfg &= ~CLK_RESET; + writeq_relaxed(cfg, xcv->reg_base + XCV_RESET); + /* Wait for DLL to lock */ + msleep(1); + + /* Configure DLL - enable or bypass + * TX no bypass, RX bypass + */ + cfg = readq_relaxed(xcv->reg_base + XCV_DLL_CTL); + cfg &= ~0xFF03; + cfg |= CLKRX_BYP; + writeq_relaxed(cfg, xcv->reg_base + XCV_DLL_CTL); + + /* Enable compensation controller and force the + * write to be visible to HW by readig back. + */ + cfg = readq_relaxed(xcv->reg_base + XCV_RESET); + cfg |= COMP_EN; + writeq_relaxed(cfg, xcv->reg_base + XCV_RESET); + readq_relaxed(xcv->reg_base + XCV_RESET); + /* Wait for compensation state machine to lock */ + msleep(10); + + /* enable the XCV block */ + cfg = readq_relaxed(xcv->reg_base + XCV_RESET); + cfg |= PORT_EN; + writeq_relaxed(cfg, xcv->reg_base + XCV_RESET); + + cfg = readq_relaxed(xcv->reg_base + XCV_RESET); + cfg |= CLK_RESET; + writeq_relaxed(cfg, xcv->reg_base + XCV_RESET); +} +EXPORT_SYMBOL(xcv_init_hw); + +void xcv_setup_link(bool link_up, int link_speed) +{ + u64 cfg; + int speed = 2; + + if (!xcv) { + dev_err(&xcv->pdev->dev, + "XCV init not done, probe may have failed\n"); + return; + } + + if (link_speed == 100) + speed = 1; + else if (link_speed == 10) + speed = 0; + + if (link_up) { + /* set operating speed */ + cfg = readq_relaxed(xcv->reg_base + XCV_CTL); + cfg &= ~0x03; + cfg |= speed; + writeq_relaxed(cfg, xcv->reg_base + XCV_CTL); + + /* Reset datapaths */ + cfg = readq_relaxed(xcv->reg_base + XCV_RESET); + cfg |= TX_DATA_RESET | RX_DATA_RESET; + writeq_relaxed(cfg, xcv->reg_base + XCV_RESET); + + /* Enable the packet flow */ + cfg = readq_relaxed(xcv->reg_base + XCV_RESET); + cfg |= TX_PKT_RESET | RX_PKT_RESET; + writeq_relaxed(cfg, xcv->reg_base + XCV_RESET); + + /* Return credits to RGX */ + writeq_relaxed(0x01, xcv->reg_base + XCV_BATCH_CRD_RET); + } else { + /* Disable packet flow */ + cfg = readq_relaxed(xcv->reg_base + XCV_RESET); + cfg &= ~(TX_PKT_RESET | RX_PKT_RESET); + writeq_relaxed(cfg, xcv->reg_base + XCV_RESET); + readq_relaxed(xcv->reg_base + XCV_RESET); + } +} +EXPORT_SYMBOL(xcv_setup_link); + +static int xcv_probe(struct pci_dev *pdev, const struct pci_device_id *ent) +{ + int err; + struct device *dev = &pdev->dev; + + xcv = devm_kzalloc(dev, sizeof(struct xcv), GFP_KERNEL); + if (!xcv) + return -ENOMEM; + xcv->pdev = pdev; + + pci_set_drvdata(pdev, xcv); + + err = pci_enable_device(pdev); + if (err) { + dev_err(dev, "Failed to enable PCI device\n"); + goto err_kfree; + } + + err = pci_request_regions(pdev, DRV_NAME); + if (err) { + dev_err(dev, "PCI request regions failed 0x%x\n", err); + goto err_disable_device; + } + + /* MAP configuration registers */ + xcv->reg_base = pcim_iomap(pdev, PCI_CFG_REG_BAR_NUM, 0); + if (!xcv->reg_base) { + dev_err(dev, "XCV: Cannot map CSR memory space, aborting\n"); + err = -ENOMEM; + goto err_release_regions; + } + + return 0; + +err_release_regions: + pci_release_regions(pdev); +err_disable_device: + pci_disable_device(pdev); +err_kfree: + devm_kfree(dev, xcv); + xcv = NULL; + return err; +} + +static void xcv_remove(struct pci_dev *pdev) +{ + struct device *dev = &pdev->dev; + + if (xcv) { + devm_kfree(dev, xcv); + xcv = NULL; + } + + pci_release_regions(pdev); + pci_disable_device(pdev); +} + +static struct pci_driver xcv_driver = { + .name = DRV_NAME, + .id_table = xcv_id_table, + .probe = xcv_probe, + .remove = xcv_remove, +}; + +static int __init xcv_init_module(void) +{ + pr_info("%s, ver %s\n", DRV_NAME, DRV_VERSION); + + return pci_register_driver(&xcv_driver); +} + +static void __exit xcv_cleanup_module(void) +{ + pci_unregister_driver(&xcv_driver); +} + +module_init(xcv_init_module); +module_exit(xcv_cleanup_module); |