[PATCH v5 2/4] scsi:ufs-qcom: add specific support for Qualcomm platforms
From: Yaniv Gardi
Date: Wed Jan 07 2015 - 10:44:34 EST
This change adds support for Qualcomm platforms that use UFS driver.
It adds for example:
- PM specific operations during hibern8, suspend, resume, clock setup
- qcom-ufs generic phy driver initialization, calibration,
power-on/off sequence, etc.
- UFS Controller specific configuration
- Rate, Gear, Mode negotiation between device and controller
Signed-off-by: Yaniv Gardi <ygardi@xxxxxxxxxxxxxx>
---
drivers/scsi/ufs/Kconfig | 12 +
drivers/scsi/ufs/Makefile | 1 +
drivers/scsi/ufs/ufs-qcom.c | 1004 +++++++++++++++++++++++++++++++++++++++++++
drivers/scsi/ufs/ufs-qcom.h | 170 ++++++++
4 files changed, 1187 insertions(+)
create mode 100644 drivers/scsi/ufs/ufs-qcom.c
create mode 100644 drivers/scsi/ufs/ufs-qcom.h
diff --git a/drivers/scsi/ufs/Kconfig b/drivers/scsi/ufs/Kconfig
index 6e07b2a..d7d56ad 100644
--- a/drivers/scsi/ufs/Kconfig
+++ b/drivers/scsi/ufs/Kconfig
@@ -70,3 +70,15 @@ config SCSI_UFSHCD_PLATFORM
If you have a controller with this interface, say Y or M here.
If unsure, say N.
+
+config SCSI_UFS_QCOM
+ bool "QCOM specific hooks to UFS controller platform driver"
+ depends on SCSI_UFSHCD_PLATFORM && ARCH_MSM
+ help
+ This selects the QCOM specific additions to UFSHCD platform driver.
+ UFS host on QCOM needs some vendor specific configuration before
+ accessing the hardware which includes PHY configuration and vendor
+ specific registers.
+
+ Select this if you have UFS controller on QCOM chipset.
+ If unsure, say N.
diff --git a/drivers/scsi/ufs/Makefile b/drivers/scsi/ufs/Makefile
index 1e5bd48..8303bcc 100644
--- a/drivers/scsi/ufs/Makefile
+++ b/drivers/scsi/ufs/Makefile
@@ -1,4 +1,5 @@
# UFSHCD makefile
+obj-$(CONFIG_SCSI_UFS_QCOM) += ufs-qcom.o
obj-$(CONFIG_SCSI_UFSHCD) += ufshcd.o
obj-$(CONFIG_SCSI_UFSHCD_PCI) += ufshcd-pci.o
obj-$(CONFIG_SCSI_UFSHCD_PLATFORM) += ufshcd-pltfrm.o
diff --git a/drivers/scsi/ufs/ufs-qcom.c b/drivers/scsi/ufs/ufs-qcom.c
new file mode 100644
index 0000000..9217af9
--- /dev/null
+++ b/drivers/scsi/ufs/ufs-qcom.c
@@ -0,0 +1,1004 @@
+/*
+ * Copyright (c) 2013-2015, Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/time.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/phy/phy.h>
+
+#include <linux/phy/phy-qcom-ufs.h>
+#include "ufshcd.h"
+#include "unipro.h"
+#include "ufs-qcom.h"
+#include "ufshci.h"
+
+static struct ufs_qcom_host *ufs_qcom_hosts[MAX_UFS_QCOM_HOSTS];
+
+static void ufs_qcom_get_speed_mode(struct ufs_pa_layer_attr *p, char *result);
+static int ufs_qcom_get_bus_vote(struct ufs_qcom_host *host,
+ const char *speed_mode);
+static int ufs_qcom_set_bus_vote(struct ufs_qcom_host *host, int vote);
+
+static int ufs_qcom_get_connected_tx_lanes(struct ufs_hba *hba, u32 *tx_lanes)
+{
+ int err = 0;
+
+ err = ufshcd_dme_get(hba,
+ UIC_ARG_MIB(PA_CONNECTEDTXDATALANES), tx_lanes);
+ if (err)
+ dev_err(hba->dev, "%s: couldn't read PA_CONNECTEDTXDATALANES %d\n",
+ __func__, err);
+
+ return err;
+}
+
+static int ufs_qcom_host_clk_get(struct device *dev,
+ const char *name, struct clk **clk_out)
+{
+ struct clk *clk;
+ int err = 0;
+
+ clk = devm_clk_get(dev, name);
+ if (IS_ERR(clk)) {
+ err = PTR_ERR(clk);
+ dev_err(dev, "%s: failed to get %s err %d",
+ __func__, name, err);
+ } else {
+ *clk_out = clk;
+ }
+
+ return err;
+}
+
+static int ufs_qcom_host_clk_enable(struct device *dev,
+ const char *name, struct clk *clk)
+{
+ int err = 0;
+
+ err = clk_prepare_enable(clk);
+ if (err)
+ dev_err(dev, "%s: %s enable failed %d\n", __func__, name, err);
+
+ return err;
+}
+
+static void ufs_qcom_disable_lane_clks(struct ufs_qcom_host *host)
+{
+ if (!host->is_lane_clks_enabled)
+ return;
+
+ clk_disable_unprepare(host->tx_l1_sync_clk);
+ clk_disable_unprepare(host->tx_l0_sync_clk);
+ clk_disable_unprepare(host->rx_l1_sync_clk);
+ clk_disable_unprepare(host->rx_l0_sync_clk);
+
+ host->is_lane_clks_enabled = false;
+}
+
+static int ufs_qcom_enable_lane_clks(struct ufs_qcom_host *host)
+{
+ int err = 0;
+ struct device *dev = host->hba->dev;
+
+ if (host->is_lane_clks_enabled)
+ return 0;
+
+ err = ufs_qcom_host_clk_enable(dev, "rx_lane0_sync_clk",
+ host->rx_l0_sync_clk);
+ if (err)
+ goto out;
+
+ err = ufs_qcom_host_clk_enable(dev, "tx_lane0_sync_clk",
+ host->tx_l0_sync_clk);
+ if (err)
+ goto disable_rx_l0;
+
+ err = ufs_qcom_host_clk_enable(dev, "rx_lane1_sync_clk",
+ host->rx_l1_sync_clk);
+ if (err)
+ goto disable_tx_l0;
+
+ err = ufs_qcom_host_clk_enable(dev, "tx_lane1_sync_clk",
+ host->tx_l1_sync_clk);
+ if (err)
+ goto disable_rx_l1;
+
+ host->is_lane_clks_enabled = true;
+ goto out;
+
+disable_rx_l1:
+ clk_disable_unprepare(host->rx_l1_sync_clk);
+disable_tx_l0:
+ clk_disable_unprepare(host->tx_l0_sync_clk);
+disable_rx_l0:
+ clk_disable_unprepare(host->rx_l0_sync_clk);
+out:
+ return err;
+}
+
+static int ufs_qcom_init_lane_clks(struct ufs_qcom_host *host)
+{
+ int err = 0;
+ struct device *dev = host->hba->dev;
+
+ err = ufs_qcom_host_clk_get(dev,
+ "rx_lane0_sync_clk", &host->rx_l0_sync_clk);
+ if (err)
+ goto out;
+
+ err = ufs_qcom_host_clk_get(dev,
+ "tx_lane0_sync_clk", &host->tx_l0_sync_clk);
+ if (err)
+ goto out;
+
+ err = ufs_qcom_host_clk_get(dev, "rx_lane1_sync_clk",
+ &host->rx_l1_sync_clk);
+ if (err)
+ goto out;
+
+ err = ufs_qcom_host_clk_get(dev, "tx_lane1_sync_clk",
+ &host->tx_l1_sync_clk);
+out:
+ return err;
+}
+
+static int ufs_qcom_link_startup_post_change(struct ufs_hba *hba)
+{
+ struct ufs_qcom_host *host = hba->priv;
+ struct phy *phy = host->generic_phy;
+ u32 tx_lanes;
+ int err = 0;
+
+ err = ufs_qcom_get_connected_tx_lanes(hba, &tx_lanes);
+ if (err)
+ goto out;
+
+ err = ufs_qcom_phy_set_tx_lane_enable(phy, tx_lanes);
+ if (err)
+ dev_err(hba->dev, "%s: ufs_qcom_phy_set_tx_lane_enable failed\n",
+ __func__);
+
+out:
+ return err;
+}
+
+static int ufs_qcom_check_hibern8(struct ufs_hba *hba)
+{
+ int err;
+ u32 tx_fsm_val = 0;
+ unsigned long timeout = jiffies + msecs_to_jiffies(HBRN8_POLL_TOUT_MS);
+
+ do {
+ err = ufshcd_dme_get(hba,
+ UIC_ARG_MIB(MPHY_TX_FSM_STATE), &tx_fsm_val);
+ if (err || tx_fsm_val == TX_FSM_HIBERN8)
+ break;
+
+ /* sleep for max. 200us */
+ usleep_range(100, 200);
+ } while (time_before(jiffies, timeout));
+
+ /*
+ * we might have scheduled out for long during polling so
+ * check the state again.
+ */
+ if (time_after(jiffies, timeout))
+ err = ufshcd_dme_get(hba,
+ UIC_ARG_MIB(MPHY_TX_FSM_STATE), &tx_fsm_val);
+
+ if (err) {
+ dev_err(hba->dev, "%s: unable to get TX_FSM_STATE, err %d\n",
+ __func__, err);
+ } else if (tx_fsm_val != TX_FSM_HIBERN8) {
+ err = tx_fsm_val;
+ dev_err(hba->dev, "%s: invalid TX_FSM_STATE = %d\n",
+ __func__, err);
+ }
+
+ return err;
+}
+
+static int ufs_qcom_power_up_sequence(struct ufs_hba *hba)
+{
+ struct ufs_qcom_host *host = hba->priv;
+ struct phy *phy = host->generic_phy;
+ int ret = 0;
+ u8 major;
+ u16 minor, step;
+ bool is_rate_B = (UFS_QCOM_LIMIT_HS_RATE == PA_HS_MODE_B)
+ ? true : false;
+
+ /* Assert PHY reset and apply PHY calibration values */
+ ufs_qcom_assert_reset(hba);
+ /* provide 1ms delay to let the reset pulse propagate */
+ usleep_range(1000, 1100);
+
+ ufs_qcom_get_controller_revision(hba, &major, &minor, &step);
+ ufs_qcom_phy_save_controller_version(phy, major, minor, step);
+ ret = ufs_qcom_phy_calibrate_phy(phy, is_rate_B);
+ if (ret) {
+ dev_err(hba->dev, "%s: ufs_qcom_phy_calibrate_phy() failed, ret = %d\n",
+ __func__, ret);
+ goto out;
+ }
+
+ /* De-assert PHY reset and start serdes */
+ ufs_qcom_deassert_reset(hba);
+
+ /*
+ * after reset deassertion, phy will need all ref clocks,
+ * voltage, current to settle down before starting serdes.
+ */
+ usleep_range(1000, 1100);
+ ret = ufs_qcom_phy_start_serdes(phy);
+ if (ret) {
+ dev_err(hba->dev, "%s: ufs_qcom_phy_start_serdes() failed, ret = %d\n",
+ __func__, ret);
+ goto out;
+ }
+
+ ret = ufs_qcom_phy_is_pcs_ready(phy);
+ if (ret)
+ dev_err(hba->dev, "%s: is_physical_coding_sublayer_ready() failed, ret = %d\n",
+ __func__, ret);
+
+out:
+ return ret;
+}
+
+/*
+ * The UTP controller has a number of internal clock gating cells (CGCs).
+ * Internal hardware sub-modules within the UTP controller control the CGCs.
+ * Hardware CGCs disable the clock to inactivate UTP sub-modules not involved
+ * in a specific operation, UTP controller CGCs are by default disabled and
+ * this function enables them (after every UFS link startup) to save some power
+ * leakage.
+ */
+static void ufs_qcom_enable_hw_clk_gating(struct ufs_hba *hba)
+{
+ ufshcd_writel(hba,
+ ufshcd_readl(hba, REG_UFS_CFG2) | REG_UFS_CFG2_CGC_EN_ALL,
+ REG_UFS_CFG2);
+
+ /* Ensure that HW clock gating is enabled before next operations */
+ mb();
+}
+
+static int ufs_qcom_hce_enable_notify(struct ufs_hba *hba, bool status)
+{
+ struct ufs_qcom_host *host = hba->priv;
+ int err = 0;
+
+ switch (status) {
+ case PRE_CHANGE:
+ ufs_qcom_power_up_sequence(hba);
+ /*
+ * The PHY PLL output is the source of tx/rx lane symbol
+ * clocks, hence, enable the lane clocks only after PHY
+ * is initialized.
+ */
+ err = ufs_qcom_enable_lane_clks(host);
+ break;
+ case POST_CHANGE:
+ /* check if UFS PHY moved from DISABLED to HIBERN8 */
+ err = ufs_qcom_check_hibern8(hba);
+ ufs_qcom_enable_hw_clk_gating(hba);
+
+ break;
+ default:
+ dev_err(hba->dev, "%s: invalid status %d\n", __func__, status);
+ err = -EINVAL;
+ break;
+ }
+ return err;
+}
+
+/**
+ * Returns non-zero for success (which rate of core_clk) and 0
+ * in case of a failure
+ */
+static unsigned long
+ufs_qcom_cfg_timers(struct ufs_hba *hba, u32 gear, u32 hs, u32 rate)
+{
+ struct ufs_clk_info *clki;
+ u32 core_clk_period_in_ns;
+ u32 tx_clk_cycles_per_us = 0;
+ unsigned long core_clk_rate = 0;
+ u32 core_clk_cycles_per_us = 0;
+
+ static u32 pwm_fr_table[][2] = {
+ {UFS_PWM_G1, 0x1},
+ {UFS_PWM_G2, 0x1},
+ {UFS_PWM_G3, 0x1},
+ {UFS_PWM_G4, 0x1},
+ };
+
+ static u32 hs_fr_table_rA[][2] = {
+ {UFS_HS_G1, 0x1F},
+ {UFS_HS_G2, 0x3e},
+ };
+
+ static u32 hs_fr_table_rB[][2] = {
+ {UFS_HS_G1, 0x24},
+ {UFS_HS_G2, 0x49},
+ };
+
+ if (gear == 0) {
+ dev_err(hba->dev, "%s: invalid gear = %d\n", __func__, gear);
+ goto out_error;
+ }
+
+ list_for_each_entry(clki, &hba->clk_list_head, list) {
+ if (!strcmp(clki->name, "core_clk"))
+ core_clk_rate = clk_get_rate(clki->clk);
+ }
+
+ /* If frequency is smaller than 1MHz, set to 1MHz */
+ if (core_clk_rate < DEFAULT_CLK_RATE_HZ)
+ core_clk_rate = DEFAULT_CLK_RATE_HZ;
+
+ core_clk_cycles_per_us = core_clk_rate / USEC_PER_SEC;
+ ufshcd_writel(hba, core_clk_cycles_per_us, REG_UFS_SYS1CLK_1US);
+
+ core_clk_period_in_ns = NSEC_PER_SEC / core_clk_rate;
+ core_clk_period_in_ns <<= OFFSET_CLK_NS_REG;
+ core_clk_period_in_ns &= MASK_CLK_NS_REG;
+
+ switch (hs) {
+ case FASTAUTO_MODE:
+ case FAST_MODE:
+ if (rate == PA_HS_MODE_A) {
+ if (gear > ARRAY_SIZE(hs_fr_table_rA)) {
+ dev_err(hba->dev,
+ "%s: index %d exceeds table size %zu\n",
+ __func__, gear,
+ ARRAY_SIZE(hs_fr_table_rA));
+ goto out_error;
+ }
+ tx_clk_cycles_per_us = hs_fr_table_rA[gear-1][1];
+ } else if (rate == PA_HS_MODE_B) {
+ if (gear > ARRAY_SIZE(hs_fr_table_rB)) {
+ dev_err(hba->dev,
+ "%s: index %d exceeds table size %zu\n",
+ __func__, gear,
+ ARRAY_SIZE(hs_fr_table_rB));
+ goto out_error;
+ }
+ tx_clk_cycles_per_us = hs_fr_table_rB[gear-1][1];
+ } else {
+ dev_err(hba->dev, "%s: invalid rate = %d\n",
+ __func__, rate);
+ goto out_error;
+ }
+ break;
+ case SLOWAUTO_MODE:
+ case SLOW_MODE:
+ if (gear > ARRAY_SIZE(pwm_fr_table)) {
+ dev_err(hba->dev,
+ "%s: index %d exceeds table size %zu\n",
+ __func__, gear,
+ ARRAY_SIZE(pwm_fr_table));
+ goto out_error;
+ }
+ tx_clk_cycles_per_us = pwm_fr_table[gear-1][1];
+ break;
+ case UNCHANGED:
+ default:
+ dev_err(hba->dev, "%s: invalid mode = %d\n", __func__, hs);
+ goto out_error;
+ }
+
+ /* this register 2 fields shall be written at once */
+ ufshcd_writel(hba, core_clk_period_in_ns | tx_clk_cycles_per_us,
+ REG_UFS_TX_SYMBOL_CLK_NS_US);
+ goto out;
+
+out_error:
+ core_clk_rate = 0;
+out:
+ return core_clk_rate;
+}
+
+static int ufs_qcom_link_startup_notify(struct ufs_hba *hba, bool status)
+{
+ unsigned long core_clk_rate = 0;
+ u32 core_clk_cycles_per_100ms;
+
+ switch (status) {
+ case PRE_CHANGE:
+ core_clk_rate = ufs_qcom_cfg_timers(hba, UFS_PWM_G1,
+ SLOWAUTO_MODE, 0);
+ if (!core_clk_rate) {
+ dev_err(hba->dev, "%s: ufs_qcom_cfg_timers() failed\n",
+ __func__);
+ return -EINVAL;
+ }
+ core_clk_cycles_per_100ms =
+ (core_clk_rate / MSEC_PER_SEC) * 100;
+ ufshcd_writel(hba, core_clk_cycles_per_100ms,
+ REG_UFS_PA_LINK_STARTUP_TIMER);
+ break;
+ case POST_CHANGE:
+ ufs_qcom_link_startup_post_change(hba);
+ break;
+ default:
+ break;
+ }
+
+ return 0;
+}
+
+static int ufs_qcom_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
+{
+ struct ufs_qcom_host *host = hba->priv;
+ struct phy *phy = host->generic_phy;
+ int ret = 0;
+
+ if (ufs_qcom_is_link_off(hba)) {
+ /*
+ * Disable the tx/rx lane symbol clocks before PHY is
+ * powered down as the PLL source should be disabled
+ * after downstream clocks are disabled.
+ */
+ ufs_qcom_disable_lane_clks(host);
+ phy_power_off(phy);
+
+ /* Assert PHY soft reset */
+ ufs_qcom_assert_reset(hba);
+ goto out;
+ }
+
+ /*
+ * If UniPro link is not active, PHY ref_clk, main PHY analog power
+ * rail and low noise analog power rail for PLL can be switched off.
+ */
+ if (!ufs_qcom_is_link_active(hba))
+ phy_power_off(phy);
+
+out:
+ return ret;
+}
+
+static int ufs_qcom_resume(struct ufs_hba *hba, enum ufs_pm_op pm_op)
+{
+ struct ufs_qcom_host *host = hba->priv;
+ struct phy *phy = host->generic_phy;
+ int err;
+
+ err = phy_power_on(phy);
+ if (err) {
+ dev_err(hba->dev, "%s: failed enabling regs, err = %d\n",
+ __func__, err);
+ goto out;
+ }
+
+ hba->is_sys_suspended = false;
+
+out:
+ return err;
+}
+
+struct ufs_qcom_dev_params {
+ u32 pwm_rx_gear; /* pwm rx gear to work in */
+ u32 pwm_tx_gear; /* pwm tx gear to work in */
+ u32 hs_rx_gear; /* hs rx gear to work in */
+ u32 hs_tx_gear; /* hs tx gear to work in */
+ u32 rx_lanes; /* number of rx lanes */
+ u32 tx_lanes; /* number of tx lanes */
+ u32 rx_pwr_pwm; /* rx pwm working pwr */
+ u32 tx_pwr_pwm; /* tx pwm working pwr */
+ u32 rx_pwr_hs; /* rx hs working pwr */
+ u32 tx_pwr_hs; /* tx hs working pwr */
+ u32 hs_rate; /* rate A/B to work in HS */
+ u32 desired_working_mode;
+};
+
+static int ufs_qcom_get_pwr_dev_param(struct ufs_qcom_dev_params *qcom_param,
+ struct ufs_pa_layer_attr *dev_max,
+ struct ufs_pa_layer_attr *agreed_pwr)
+{
+ int min_qcom_gear;
+ int min_dev_gear;
+ bool is_dev_sup_hs = false;
+ bool is_qcom_max_hs = false;
+
+ if (dev_max->pwr_rx == FAST_MODE)
+ is_dev_sup_hs = true;
+
+ if (qcom_param->desired_working_mode == FAST) {
+ is_qcom_max_hs = true;
+ min_qcom_gear = min_t(u32, qcom_param->hs_rx_gear,
+ qcom_param->hs_tx_gear);
+ } else {
+ min_qcom_gear = min_t(u32, qcom_param->pwm_rx_gear,
+ qcom_param->pwm_tx_gear);
+ }
+
+ /*
+ * device doesn't support HS but qcom_param->desired_working_mode is
+ * HS, thus device and qcom_param don't agree
+ */
+ if (!is_dev_sup_hs && is_qcom_max_hs) {
+ pr_err("%s: failed to agree on power mode (device doesn't support HS but requested power is HS)\n",
+ __func__);
+ return -ENOTSUPP;
+ } else if (is_dev_sup_hs && is_qcom_max_hs) {
+ /*
+ * since device supports HS, it supports FAST_MODE.
+ * since qcom_param->desired_working_mode is also HS
+ * then final decision (FAST/FASTAUTO) is done according
+ * to qcom_params as it is the restricting factor
+ */
+ agreed_pwr->pwr_rx = agreed_pwr->pwr_tx =
+ qcom_param->rx_pwr_hs;
+ } else {
+ /*
+ * here qcom_param->desired_working_mode is PWM.
+ * it doesn't matter whether device supports HS or PWM,
+ * in both cases qcom_param->desired_working_mode will
+ * determine the mode
+ */
+ agreed_pwr->pwr_rx = agreed_pwr->pwr_tx =
+ qcom_param->rx_pwr_pwm;
+ }
+
+ /*
+ * we would like tx to work in the minimum number of lanes
+ * between device capability and vendor preferences.
+ * the same decision will be made for rx
+ */
+ agreed_pwr->lane_tx = min_t(u32, dev_max->lane_tx,
+ qcom_param->tx_lanes);
+ agreed_pwr->lane_rx = min_t(u32, dev_max->lane_rx,
+ qcom_param->rx_lanes);
+
+ /* device maximum gear is the minimum between device rx and tx gears */
+ min_dev_gear = min_t(u32, dev_max->gear_rx, dev_max->gear_tx);
+
+ /*
+ * if both device capabilities and vendor pre-defined preferences are
+ * both HS or both PWM then set the minimum gear to be the chosen
+ * working gear.
+ * if one is PWM and one is HS then the one that is PWM get to decide
+ * what is the gear, as it is the one that also decided previously what
+ * pwr the device will be configured to.
+ */
+ if ((is_dev_sup_hs && is_qcom_max_hs) ||
+ (!is_dev_sup_hs && !is_qcom_max_hs))
+ agreed_pwr->gear_rx = agreed_pwr->gear_tx =
+ min_t(u32, min_dev_gear, min_qcom_gear);
+ else if (!is_dev_sup_hs)
+ agreed_pwr->gear_rx = agreed_pwr->gear_tx = min_dev_gear;
+ else
+ agreed_pwr->gear_rx = agreed_pwr->gear_tx = min_qcom_gear;
+
+ agreed_pwr->hs_rate = qcom_param->hs_rate;
+ return 0;
+}
+
+static int ufs_qcom_update_bus_bw_vote(struct ufs_qcom_host *host)
+{
+ int vote;
+ int err = 0;
+ char mode[BUS_VECTOR_NAME_LEN];
+
+ ufs_qcom_get_speed_mode(&host->dev_req_params, mode);
+
+ vote = ufs_qcom_get_bus_vote(host, mode);
+ if (vote >= 0)
+ err = ufs_qcom_set_bus_vote(host, vote);
+ else
+ err = vote;
+
+ if (err)
+ dev_err(host->hba->dev, "%s: failed %d\n", __func__, err);
+ else
+ host->bus_vote.saved_vote = vote;
+ return err;
+}
+
+static int ufs_qcom_pwr_change_notify(struct ufs_hba *hba,
+ bool status,
+ struct ufs_pa_layer_attr *dev_max_params,
+ struct ufs_pa_layer_attr *dev_req_params)
+{
+ u32 val;
+ struct ufs_qcom_host *host = hba->priv;
+ struct phy *phy = host->generic_phy;
+ struct ufs_qcom_dev_params ufs_qcom_cap;
+ int ret = 0;
+ int res = 0;
+
+ if (!dev_req_params) {
+ pr_err("%s: incoming dev_req_params is NULL\n", __func__);
+ ret = -EINVAL;
+ goto out;
+ }
+
+ switch (status) {
+ case PRE_CHANGE:
+ ufs_qcom_cap.tx_lanes = UFS_QCOM_LIMIT_NUM_LANES_TX;
+ ufs_qcom_cap.rx_lanes = UFS_QCOM_LIMIT_NUM_LANES_RX;
+ ufs_qcom_cap.hs_rx_gear = UFS_QCOM_LIMIT_HSGEAR_RX;
+ ufs_qcom_cap.hs_tx_gear = UFS_QCOM_LIMIT_HSGEAR_TX;
+ ufs_qcom_cap.pwm_rx_gear = UFS_QCOM_LIMIT_PWMGEAR_RX;
+ ufs_qcom_cap.pwm_tx_gear = UFS_QCOM_LIMIT_PWMGEAR_TX;
+ ufs_qcom_cap.rx_pwr_pwm = UFS_QCOM_LIMIT_RX_PWR_PWM;
+ ufs_qcom_cap.tx_pwr_pwm = UFS_QCOM_LIMIT_TX_PWR_PWM;
+ ufs_qcom_cap.rx_pwr_hs = UFS_QCOM_LIMIT_RX_PWR_HS;
+ ufs_qcom_cap.tx_pwr_hs = UFS_QCOM_LIMIT_TX_PWR_HS;
+ ufs_qcom_cap.hs_rate = UFS_QCOM_LIMIT_HS_RATE;
+ ufs_qcom_cap.desired_working_mode =
+ UFS_QCOM_LIMIT_DESIRED_MODE;
+
+ ret = ufs_qcom_get_pwr_dev_param(&ufs_qcom_cap,
+ dev_max_params,
+ dev_req_params);
+ if (ret) {
+ pr_err("%s: failed to determine capabilities\n",
+ __func__);
+ goto out;
+ }
+
+ break;
+ case POST_CHANGE:
+ if (!ufs_qcom_cfg_timers(hba, dev_req_params->gear_rx,
+ dev_req_params->pwr_rx,
+ dev_req_params->hs_rate)) {
+ dev_err(hba->dev, "%s: ufs_qcom_cfg_timers() failed\n",
+ __func__);
+ /*
+ * we return error code at the end of the routine,
+ * but continue to configure UFS_PHY_TX_LANE_ENABLE
+ * and bus voting as usual
+ */
+ ret = -EINVAL;
+ }
+
+ val = ~(MAX_U32 << dev_req_params->lane_tx);
+ res = ufs_qcom_phy_set_tx_lane_enable(phy, val);
+ if (res) {
+ dev_err(hba->dev, "%s: ufs_qcom_phy_set_tx_lane_enable() failed res = %d\n",
+ __func__, res);
+ ret = res;
+ }
+
+ /* cache the power mode parameters to use internally */
+ memcpy(&host->dev_req_params,
+ dev_req_params, sizeof(*dev_req_params));
+ ufs_qcom_update_bus_bw_vote(host);
+ break;
+ default:
+ ret = -EINVAL;
+ break;
+ }
+out:
+ return ret;
+}
+
+/**
+ * ufs_qcom_advertise_quirks - advertise the known QCOM UFS controller quirks
+ * @hba: host controller instance
+ *
+ * QCOM UFS host controller might have some non standard behaviours (quirks)
+ * than what is specified by UFSHCI specification. Advertise all such
+ * quirks to standard UFS host controller driver so standard takes them into
+ * account.
+ */
+static void ufs_qcom_advertise_quirks(struct ufs_hba *hba)
+{
+ u8 major;
+ u16 minor, step;
+
+ ufs_qcom_get_controller_revision(hba, &major, &minor, &step);
+
+ /*
+ * TBD
+ * here we should be advertising controller quirks according to
+ * controller version.
+ */
+}
+
+static int ufs_qcom_get_bus_vote(struct ufs_qcom_host *host,
+ const char *speed_mode)
+{
+ struct device *dev = host->hba->dev;
+ struct device_node *np = dev->of_node;
+ int err;
+ const char *key = "qcom,bus-vector-names";
+
+ if (!speed_mode) {
+ err = -EINVAL;
+ goto out;
+ }
+
+ if (host->bus_vote.is_max_bw_needed && !!strcmp(speed_mode, "MIN"))
+ err = of_property_match_string(np, key, "MAX");
+ else
+ err = of_property_match_string(np, key, speed_mode);
+
+out:
+ if (err < 0)
+ dev_err(dev, "%s: Invalid %s mode %d\n",
+ __func__, speed_mode, err);
+ return err;
+}
+
+static int ufs_qcom_set_bus_vote(struct ufs_qcom_host *host, int vote)
+{
+ int err = 0;
+
+ if (vote != host->bus_vote.curr_vote)
+ host->bus_vote.curr_vote = vote;
+
+ return err;
+}
+
+static void ufs_qcom_get_speed_mode(struct ufs_pa_layer_attr *p, char *result)
+{
+ int gear = max_t(u32, p->gear_rx, p->gear_tx);
+ int lanes = max_t(u32, p->lane_rx, p->lane_tx);
+ int pwr;
+
+ /* default to PWM Gear 1, Lane 1 if power mode is not initialized */
+ if (!gear)
+ gear = 1;
+
+ if (!lanes)
+ lanes = 1;
+
+ if (!p->pwr_rx && !p->pwr_tx) {
+ pwr = SLOWAUTO_MODE;
+ snprintf(result, BUS_VECTOR_NAME_LEN, "MIN");
+ } else if (p->pwr_rx == FAST_MODE || p->pwr_rx == FASTAUTO_MODE ||
+ p->pwr_tx == FAST_MODE || p->pwr_tx == FASTAUTO_MODE) {
+ pwr = FAST_MODE;
+ snprintf(result, BUS_VECTOR_NAME_LEN, "%s_R%s_G%d_L%d", "HS",
+ p->hs_rate == PA_HS_MODE_B ? "B" : "A", gear, lanes);
+ } else {
+ pwr = SLOW_MODE;
+ snprintf(result, BUS_VECTOR_NAME_LEN, "%s_G%d_L%d",
+ "PWM", gear, lanes);
+ }
+}
+
+static int ufs_qcom_setup_clocks(struct ufs_hba *hba, bool on)
+{
+ struct ufs_qcom_host *host = hba->priv;
+ int err = 0;
+ int vote = 0;
+
+ /*
+ * In case ufs_qcom_init() is not yet done, simply ignore.
+ * This ufs_qcom_setup_clocks() shall be called from
+ * ufs_qcom_init() after init is done.
+ */
+ if (!host)
+ return 0;
+
+ if (on) {
+ err = ufs_qcom_phy_enable_iface_clk(host->generic_phy);
+ if (err)
+ goto out;
+
+ err = ufs_qcom_phy_enable_ref_clk(host->generic_phy);
+ if (err) {
+ dev_err(hba->dev, "%s enable phy ref clock failed, err=%d\n",
+ __func__, err);
+ ufs_qcom_phy_disable_iface_clk(host->generic_phy);
+ goto out;
+ }
+ /* enable the device ref clock */
+ ufs_qcom_phy_enable_dev_ref_clk(host->generic_phy);
+ vote = host->bus_vote.saved_vote;
+ if (vote == host->bus_vote.min_bw_vote)
+ ufs_qcom_update_bus_bw_vote(host);
+ } else {
+ /* M-PHY RMMI interface clocks can be turned off */
+ ufs_qcom_phy_disable_iface_clk(host->generic_phy);
+ if (!ufs_qcom_is_link_active(hba)) {
+ /* turn off UFS local PHY ref_clk */
+ ufs_qcom_phy_disable_ref_clk(host->generic_phy);
+ /* disable device ref_clk */
+ ufs_qcom_phy_disable_dev_ref_clk(host->generic_phy);
+ }
+ vote = host->bus_vote.min_bw_vote;
+ }
+
+ err = ufs_qcom_set_bus_vote(host, vote);
+ if (err)
+ dev_err(hba->dev, "%s: set bus vote failed %d\n",
+ __func__, err);
+
+out:
+ return err;
+}
+
+static ssize_t
+show_ufs_to_mem_max_bus_bw(struct device *dev, struct device_attribute *attr,
+ char *buf)
+{
+ struct ufs_hba *hba = dev_get_drvdata(dev);
+ struct ufs_qcom_host *host = hba->priv;
+
+ return snprintf(buf, PAGE_SIZE, "%u\n",
+ host->bus_vote.is_max_bw_needed);
+}
+
+static ssize_t
+store_ufs_to_mem_max_bus_bw(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct ufs_hba *hba = dev_get_drvdata(dev);
+ struct ufs_qcom_host *host = hba->priv;
+ uint32_t value;
+
+ if (!kstrtou32(buf, 0, &value)) {
+ host->bus_vote.is_max_bw_needed = !!value;
+ ufs_qcom_update_bus_bw_vote(host);
+ }
+
+ return count;
+}
+
+static int ufs_qcom_bus_register(struct ufs_qcom_host *host)
+{
+ int err;
+ struct device *dev = host->hba->dev;
+ struct device_node *np = dev->of_node;
+
+ err = of_property_count_strings(np, "qcom,bus-vector-names");
+ if (err < 0 ) {
+ dev_err(dev, "%s: qcom,bus-vector-names not specified correctly %d\n",
+ __func__, err);
+ goto out;
+ }
+
+ /* cache the vote index for minimum and maximum bandwidth */
+ host->bus_vote.min_bw_vote = ufs_qcom_get_bus_vote(host, "MIN");
+ host->bus_vote.max_bw_vote = ufs_qcom_get_bus_vote(host, "MAX");
+
+ host->bus_vote.max_bus_bw.show = show_ufs_to_mem_max_bus_bw;
+ host->bus_vote.max_bus_bw.store = store_ufs_to_mem_max_bus_bw;
+ sysfs_attr_init(&host->bus_vote.max_bus_bw.attr);
+ host->bus_vote.max_bus_bw.attr.name = "max_bus_bw";
+ host->bus_vote.max_bus_bw.attr.mode = S_IRUGO | S_IWUSR;
+ err = device_create_file(dev, &host->bus_vote.max_bus_bw);
+out:
+ return err;
+}
+
+#define ANDROID_BOOT_DEV_MAX 30
+static char android_boot_dev[ANDROID_BOOT_DEV_MAX];
+static int get_android_boot_dev(char *str)
+{
+ strlcpy(android_boot_dev, str, ANDROID_BOOT_DEV_MAX);
+ return 1;
+}
+__setup("androidboot.bootdevice=", get_android_boot_dev);
+
+/**
+ * ufs_qcom_init - bind phy with controller
+ * @hba: host controller instance
+ *
+ * Binds PHY with controller and powers up PHY enabling clocks
+ * and regulators.
+ *
+ * Returns -EPROBE_DEFER if binding fails, returns negative error
+ * on phy power up failure and returns zero on success.
+ */
+static int ufs_qcom_init(struct ufs_hba *hba)
+{
+ int err;
+ struct device *dev = hba->dev;
+ struct ufs_qcom_host *host;
+
+ if (strlen(android_boot_dev) && strcmp(android_boot_dev, dev_name(dev)))
+ return -ENODEV;
+
+ host = devm_kzalloc(dev, sizeof(*host), GFP_KERNEL);
+ if (!host) {
+ err = -ENOMEM;
+ dev_err(dev, "%s: no memory for qcom ufs host\n", __func__);
+ goto out;
+ }
+
+ host->hba = hba;
+ hba->priv = (void *)host;
+
+ host->generic_phy = devm_phy_get(dev, "ufsphy");
+
+ if (IS_ERR(host->generic_phy)) {
+ err = PTR_ERR(host->generic_phy);
+ dev_err(dev, "%s: PHY get failed %d\n", __func__, err);
+ goto out;
+ }
+
+ err = ufs_qcom_bus_register(host);
+ if (err)
+ goto out_host_free;
+
+ phy_init(host->generic_phy);
+ err = phy_power_on(host->generic_phy);
+ if (err)
+ goto out_unregister_bus;
+
+ err = ufs_qcom_init_lane_clks(host);
+ if (err)
+ goto out_disable_phy;
+
+ ufs_qcom_advertise_quirks(hba);
+
+ hba->caps |= UFSHCD_CAP_CLK_GATING | UFSHCD_CAP_CLK_SCALING;
+ hba->caps |= UFSHCD_CAP_AUTO_BKOPS_SUSPEND;
+
+ ufs_qcom_setup_clocks(hba, true);
+
+ if (hba->dev->id < MAX_UFS_QCOM_HOSTS)
+ ufs_qcom_hosts[hba->dev->id] = host;
+
+ goto out;
+
+out_disable_phy:
+ phy_power_off(host->generic_phy);
+out_unregister_bus:
+ phy_exit(host->generic_phy);
+out_host_free:
+ devm_kfree(dev, host);
+ hba->priv = NULL;
+out:
+ return err;
+}
+
+static void ufs_qcom_exit(struct ufs_hba *hba)
+{
+ struct ufs_qcom_host *host = hba->priv;
+
+ ufs_qcom_disable_lane_clks(host);
+ phy_power_off(host->generic_phy);
+}
+
+static
+void ufs_qcom_clk_scale_notify(struct ufs_hba *hba)
+{
+ struct ufs_qcom_host *host = hba->priv;
+ struct ufs_pa_layer_attr *dev_req_params = &host->dev_req_params;
+
+ if (!dev_req_params)
+ return;
+
+ ufs_qcom_cfg_timers(hba, dev_req_params->gear_rx,
+ dev_req_params->pwr_rx,
+ dev_req_params->hs_rate);
+}
+
+/**
+ * struct ufs_hba_qcom_vops - UFS QCOM specific variant operations
+ *
+ * The variant operations configure the necessary controller and PHY
+ * handshake during initialization.
+ */
+static const struct ufs_hba_variant_ops ufs_hba_qcom_vops = {
+ .name = "qcom",
+ .init = ufs_qcom_init,
+ .exit = ufs_qcom_exit,
+ .clk_scale_notify = ufs_qcom_clk_scale_notify,
+ .setup_clocks = ufs_qcom_setup_clocks,
+ .hce_enable_notify = ufs_qcom_hce_enable_notify,
+ .link_startup_notify = ufs_qcom_link_startup_notify,
+ .pwr_change_notify = ufs_qcom_pwr_change_notify,
+ .suspend = ufs_qcom_suspend,
+ .resume = ufs_qcom_resume,
+};
+EXPORT_SYMBOL(ufs_hba_qcom_vops);
diff --git a/drivers/scsi/ufs/ufs-qcom.h b/drivers/scsi/ufs/ufs-qcom.h
new file mode 100644
index 0000000..9a6febd
--- /dev/null
+++ b/drivers/scsi/ufs/ufs-qcom.h
@@ -0,0 +1,170 @@
+/* Copyright (c) 2013-2015, The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#ifndef UFS_QCOM_H_
+#define UFS_QCOM_H_
+
+#define MAX_UFS_QCOM_HOSTS 1
+#define MAX_U32 (~(u32)0)
+#define MPHY_TX_FSM_STATE 0x41
+#define TX_FSM_HIBERN8 0x1
+#define HBRN8_POLL_TOUT_MS 100
+#define DEFAULT_CLK_RATE_HZ 1000000
+#define BUS_VECTOR_NAME_LEN 32
+
+#define UFS_HW_VER_MAJOR_SHFT (28)
+#define UFS_HW_VER_MAJOR_MASK (0x000F << UFS_HW_VER_MAJOR_SHFT)
+#define UFS_HW_VER_MINOR_SHFT (16)
+#define UFS_HW_VER_MINOR_MASK (0x0FFF << UFS_HW_VER_MINOR_SHFT)
+#define UFS_HW_VER_STEP_SHFT (0)
+#define UFS_HW_VER_STEP_MASK (0xFFFF << UFS_HW_VER_STEP_SHFT)
+
+/* vendor specific pre-defined parameters */
+#define SLOW 1
+#define FAST 2
+
+#define UFS_QCOM_LIMIT_NUM_LANES_RX 2
+#define UFS_QCOM_LIMIT_NUM_LANES_TX 2
+#define UFS_QCOM_LIMIT_HSGEAR_RX UFS_HS_G2
+#define UFS_QCOM_LIMIT_HSGEAR_TX UFS_HS_G2
+#define UFS_QCOM_LIMIT_PWMGEAR_RX UFS_PWM_G4
+#define UFS_QCOM_LIMIT_PWMGEAR_TX UFS_PWM_G4
+#define UFS_QCOM_LIMIT_RX_PWR_PWM SLOW_MODE
+#define UFS_QCOM_LIMIT_TX_PWR_PWM SLOW_MODE
+#define UFS_QCOM_LIMIT_RX_PWR_HS FAST_MODE
+#define UFS_QCOM_LIMIT_TX_PWR_HS FAST_MODE
+#define UFS_QCOM_LIMIT_HS_RATE PA_HS_MODE_B
+#define UFS_QCOM_LIMIT_DESIRED_MODE FAST
+
+/* QCOM UFS host controller vendor specific registers */
+enum {
+ REG_UFS_SYS1CLK_1US = 0xC0,
+ REG_UFS_TX_SYMBOL_CLK_NS_US = 0xC4,
+ REG_UFS_LOCAL_PORT_ID_REG = 0xC8,
+ REG_UFS_PA_ERR_CODE = 0xCC,
+ REG_UFS_RETRY_TIMER_REG = 0xD0,
+ REG_UFS_PA_LINK_STARTUP_TIMER = 0xD8,
+ REG_UFS_CFG1 = 0xDC,
+ REG_UFS_CFG2 = 0xE0,
+ REG_UFS_HW_VERSION = 0xE4,
+
+ UFS_DBG_RD_REG_UAWM = 0x100,
+ UFS_DBG_RD_REG_UARM = 0x200,
+ UFS_DBG_RD_REG_TXUC = 0x300,
+ UFS_DBG_RD_REG_RXUC = 0x400,
+ UFS_DBG_RD_REG_DFC = 0x500,
+ UFS_DBG_RD_REG_TRLUT = 0x600,
+ UFS_DBG_RD_REG_TMRLUT = 0x700,
+ UFS_UFS_DBG_RD_REG_OCSC = 0x800,
+
+ UFS_UFS_DBG_RD_DESC_RAM = 0x1500,
+ UFS_UFS_DBG_RD_PRDT_RAM = 0x1700,
+ UFS_UFS_DBG_RD_RESP_RAM = 0x1800,
+ UFS_UFS_DBG_RD_EDTL_RAM = 0x1900,
+};
+
+/* bit definitions for REG_UFS_CFG2 register */
+#define UAWM_HW_CGC_EN (1 << 0)
+#define UARM_HW_CGC_EN (1 << 1)
+#define TXUC_HW_CGC_EN (1 << 2)
+#define RXUC_HW_CGC_EN (1 << 3)
+#define DFC_HW_CGC_EN (1 << 4)
+#define TRLUT_HW_CGC_EN (1 << 5)
+#define TMRLUT_HW_CGC_EN (1 << 6)
+#define OCSC_HW_CGC_EN (1 << 7)
+
+#define REG_UFS_CFG2_CGC_EN_ALL (UAWM_HW_CGC_EN | UARM_HW_CGC_EN |\
+ TXUC_HW_CGC_EN | RXUC_HW_CGC_EN |\
+ DFC_HW_CGC_EN | TRLUT_HW_CGC_EN |\
+ TMRLUT_HW_CGC_EN | OCSC_HW_CGC_EN)
+
+/* bit offset */
+enum {
+ OFFSET_UFS_PHY_SOFT_RESET = 1,
+ OFFSET_CLK_NS_REG = 10,
+};
+
+/* bit masks */
+enum {
+ MASK_UFS_PHY_SOFT_RESET = 0x2,
+ MASK_TX_SYMBOL_CLK_1US_REG = 0x3FF,
+ MASK_CLK_NS_REG = 0xFFFC00,
+};
+
+enum ufs_qcom_phy_init_type {
+ UFS_PHY_INIT_FULL,
+ UFS_PHY_INIT_CFG_RESTORE,
+};
+
+static inline void
+ufs_qcom_get_controller_revision(struct ufs_hba *hba,
+ u8 *major, u16 *minor, u16 *step)
+{
+ u32 ver = ufshcd_readl(hba, REG_UFS_HW_VERSION);
+
+ *major = (ver & UFS_HW_VER_MAJOR_MASK) >> UFS_HW_VER_MAJOR_SHFT;
+ *minor = (ver & UFS_HW_VER_MINOR_MASK) >> UFS_HW_VER_MINOR_SHFT;
+ *step = (ver & UFS_HW_VER_STEP_MASK) >> UFS_HW_VER_STEP_SHFT;
+};
+
+static inline void ufs_qcom_assert_reset(struct ufs_hba *hba)
+{
+ ufshcd_rmwl(hba, MASK_UFS_PHY_SOFT_RESET,
+ 1 << OFFSET_UFS_PHY_SOFT_RESET, REG_UFS_CFG1);
+
+ /*
+ * Make sure assertion of ufs phy reset is written to
+ * register before returning
+ */
+ mb();
+}
+
+static inline void ufs_qcom_deassert_reset(struct ufs_hba *hba)
+{
+ ufshcd_rmwl(hba, MASK_UFS_PHY_SOFT_RESET,
+ 0 << OFFSET_UFS_PHY_SOFT_RESET, REG_UFS_CFG1);
+
+ /*
+ * Make sure de-assertion of ufs phy reset is written to
+ * register before returning
+ */
+ mb();
+}
+
+struct ufs_qcom_bus_vote {
+ uint32_t client_handle;
+ uint32_t curr_vote;
+ int min_bw_vote;
+ int max_bw_vote;
+ int saved_vote;
+ bool is_max_bw_needed;
+ struct device_attribute max_bus_bw;
+};
+
+struct ufs_qcom_host {
+ struct phy *generic_phy;
+ struct ufs_hba *hba;
+ struct ufs_qcom_bus_vote bus_vote;
+ struct ufs_pa_layer_attr dev_req_params;
+ struct clk *rx_l0_sync_clk;
+ struct clk *tx_l0_sync_clk;
+ struct clk *rx_l1_sync_clk;
+ struct clk *tx_l1_sync_clk;
+ bool is_lane_clks_enabled;
+};
+
+#define ufs_qcom_is_link_off(hba) ufshcd_is_link_off(hba)
+#define ufs_qcom_is_link_active(hba) ufshcd_is_link_active(hba)
+#define ufs_qcom_is_link_hibern8(hba) ufshcd_is_link_hibern8(hba)
+
+#endif /* UFS_QCOM_H_ */
--
1.8.2.1
--
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/