blob: 57b241417e8cd255d8582ed3c1d1fbe34fb5d7f1 [file] [log] [blame]
// SPDX-License-Identifier: GPL-2.0
/* Copyright (c) 2012-2018, The Linux Foundation. All rights reserved.
* Copyright (C) 2018-2023 Linaro Ltd.
*/
#include <linux/types.h>
#include <linux/atomic.h>
#include <linux/bitfield.h>
#include <linux/bug.h>
#include <linux/io.h>
#include <linux/firmware.h>
#include <linux/module.h>
#include <linux/of.h>
#include <linux/of_address.h>
#include <linux/platform_device.h>
#include <linux/pm_runtime.h>
#include <linux/firmware/qcom/qcom_scm.h>
#include <linux/soc/qcom/mdt_loader.h>
#include "ipa.h"
#include "ipa_power.h"
#include "ipa_data.h"
#include "ipa_endpoint.h"
#include "ipa_resource.h"
#include "ipa_cmd.h"
#include "ipa_reg.h"
#include "ipa_mem.h"
#include "ipa_table.h"
#include "ipa_smp2p.h"
#include "ipa_modem.h"
#include "ipa_uc.h"
#include "ipa_interrupt.h"
#include "gsi_trans.h"
#include "ipa_sysfs.h"
/**
* DOC: The IP Accelerator
*
* This driver supports the Qualcomm IP Accelerator (IPA), which is a
* networking component found in many Qualcomm SoCs. The IPA is connected
* to the application processor (AP), but is also connected (and partially
* controlled by) other "execution environments" (EEs), such as a modem.
*
* The IPA is the conduit between the AP and the modem that carries network
* traffic. This driver presents a network interface representing the
* connection of the modem to external (e.g. LTE) networks.
*
* The IPA provides protocol checksum calculation, offloading this work
* from the AP. The IPA offers additional functionality, including routing,
* filtering, and NAT support, but that more advanced functionality is not
* currently supported. Despite that, some resources--including routing
* tables and filter tables--are defined in this driver because they must
* be initialized even when the advanced hardware features are not used.
*
* There are two distinct layers that implement the IPA hardware, and this
* is reflected in the organization of the driver. The generic software
* interface (GSI) is an integral component of the IPA, providing a
* well-defined communication layer between the AP subsystem and the IPA
* core. The GSI implements a set of "channels" used for communication
* between the AP and the IPA.
*
* The IPA layer uses GSI channels to implement its "endpoints". And while
* a GSI channel carries data between the AP and the IPA, a pair of IPA
* endpoints is used to carry traffic between two EEs. Specifically, the main
* modem network interface is implemented by two pairs of endpoints: a TX
* endpoint on the AP coupled with an RX endpoint on the modem; and another
* RX endpoint on the AP receiving data from a TX endpoint on the modem.
*/
/* The name of the GSI firmware file relative to /lib/firmware */
#define IPA_FW_PATH_DEFAULT "ipa_fws.mdt"
#define IPA_PAS_ID 15
/* Shift of 19.2 MHz timestamp to achieve lower resolution timestamps */
/* IPA v5.5+ does not specify Qtime timestamp config for DPL */
#define DPL_TIMESTAMP_SHIFT 14 /* ~1.172 kHz, ~853 usec per tick */
#define TAG_TIMESTAMP_SHIFT 14
#define NAT_TIMESTAMP_SHIFT 24 /* ~1.144 Hz, ~874 msec per tick */
/* Divider for 19.2 MHz crystal oscillator clock to get common timer clock */
#define IPA_XO_CLOCK_DIVIDER 192 /* 1 is subtracted where used */
/**
* enum ipa_firmware_loader: How GSI firmware gets loaded
*
* @IPA_LOADER_DEFER: System not ready; try again later
* @IPA_LOADER_SELF: AP loads GSI firmware
* @IPA_LOADER_MODEM: Modem loads GSI firmware, signals when done
* @IPA_LOADER_SKIP: Neither AP nor modem need to load GSI firmware
* @IPA_LOADER_INVALID: GSI firmware loader specification is invalid
*/
enum ipa_firmware_loader {
IPA_LOADER_DEFER,
IPA_LOADER_SELF,
IPA_LOADER_MODEM,
IPA_LOADER_SKIP,
IPA_LOADER_INVALID,
};
/**
* ipa_setup() - Set up IPA hardware
* @ipa: IPA pointer
*
* Perform initialization that requires issuing immediate commands on
* the command TX endpoint. If the modem is doing GSI firmware load
* and initialization, this function will be called when an SMP2P
* interrupt has been signaled by the modem. Otherwise it will be
* called from ipa_probe() after GSI firmware has been successfully
* loaded, authenticated, and started by Trust Zone.
*/
int ipa_setup(struct ipa *ipa)
{
struct ipa_endpoint *exception_endpoint;
struct ipa_endpoint *command_endpoint;
struct device *dev = ipa->dev;
int ret;
ret = gsi_setup(&ipa->gsi);
if (ret)
return ret;
ret = ipa_power_setup(ipa);
if (ret)
goto err_gsi_teardown;
ipa_endpoint_setup(ipa);
/* We need to use the AP command TX endpoint to perform other
* initialization, so we enable first.
*/
command_endpoint = ipa->name_map[IPA_ENDPOINT_AP_COMMAND_TX];
ret = ipa_endpoint_enable_one(command_endpoint);
if (ret)
goto err_endpoint_teardown;
ret = ipa_mem_setup(ipa); /* No matching teardown required */
if (ret)
goto err_command_disable;
ret = ipa_table_setup(ipa); /* No matching teardown required */
if (ret)
goto err_command_disable;
/* Enable the exception handling endpoint, and tell the hardware
* to use it by default.
*/
exception_endpoint = ipa->name_map[IPA_ENDPOINT_AP_LAN_RX];
ret = ipa_endpoint_enable_one(exception_endpoint);
if (ret)
goto err_command_disable;
ipa_endpoint_default_route_set(ipa, exception_endpoint->endpoint_id);
/* We're all set. Now prepare for communication with the modem */
ret = ipa_qmi_setup(ipa);
if (ret)
goto err_default_route_clear;
ipa->setup_complete = true;
dev_info(dev, "IPA driver setup completed successfully\n");
return 0;
err_default_route_clear:
ipa_endpoint_default_route_clear(ipa);
ipa_endpoint_disable_one(exception_endpoint);
err_command_disable:
ipa_endpoint_disable_one(command_endpoint);
err_endpoint_teardown:
ipa_endpoint_teardown(ipa);
ipa_power_teardown(ipa);
err_gsi_teardown:
gsi_teardown(&ipa->gsi);
return ret;
}
/**
* ipa_teardown() - Inverse of ipa_setup()
* @ipa: IPA pointer
*/
static void ipa_teardown(struct ipa *ipa)
{
struct ipa_endpoint *exception_endpoint;
struct ipa_endpoint *command_endpoint;
/* We're going to tear everything down, as if setup never completed */
ipa->setup_complete = false;
ipa_qmi_teardown(ipa);
ipa_endpoint_default_route_clear(ipa);
exception_endpoint = ipa->name_map[IPA_ENDPOINT_AP_LAN_RX];
ipa_endpoint_disable_one(exception_endpoint);
command_endpoint = ipa->name_map[IPA_ENDPOINT_AP_COMMAND_TX];
ipa_endpoint_disable_one(command_endpoint);
ipa_endpoint_teardown(ipa);
ipa_power_teardown(ipa);
gsi_teardown(&ipa->gsi);
}
static void
ipa_hardware_config_bcr(struct ipa *ipa, const struct ipa_data *data)
{
const struct reg *reg;
u32 val;
/* IPA v4.5+ has no backward compatibility register */
if (ipa->version >= IPA_VERSION_4_5)
return;
reg = ipa_reg(ipa, IPA_BCR);
val = data->backward_compat;
iowrite32(val, ipa->reg_virt + reg_offset(reg));
}
static void ipa_hardware_config_tx(struct ipa *ipa)
{
enum ipa_version version = ipa->version;
const struct reg *reg;
u32 offset;
u32 val;
if (version <= IPA_VERSION_4_0 || version >= IPA_VERSION_4_5)
return;
/* Disable PA mask to allow HOLB drop */
reg = ipa_reg(ipa, IPA_TX_CFG);
offset = reg_offset(reg);
val = ioread32(ipa->reg_virt + offset);
val &= ~reg_bit(reg, PA_MASK_EN);
iowrite32(val, ipa->reg_virt + offset);
}
static void ipa_hardware_config_clkon(struct ipa *ipa)
{
enum ipa_version version = ipa->version;
const struct reg *reg;
u32 val;
if (version >= IPA_VERSION_4_5)
return;
if (version < IPA_VERSION_4_0 && version != IPA_VERSION_3_1)
return;
/* Implement some hardware workarounds */
reg = ipa_reg(ipa, CLKON_CFG);
if (version == IPA_VERSION_3_1) {
/* Disable MISC clock gating */
val = reg_bit(reg, CLKON_MISC);
} else { /* IPA v4.0+ */
/* Enable open global clocks in the CLKON configuration */
val = reg_bit(reg, CLKON_GLOBAL);
val |= reg_bit(reg, GLOBAL_2X_CLK);
}
iowrite32(val, ipa->reg_virt + reg_offset(reg));
}
/* Configure bus access behavior for IPA components */
static void ipa_hardware_config_comp(struct ipa *ipa)
{
const struct reg *reg;
u32 offset;
u32 val;
/* Nothing to configure prior to IPA v4.0 */
if (ipa->version < IPA_VERSION_4_0)
return;
reg = ipa_reg(ipa, COMP_CFG);
offset = reg_offset(reg);
val = ioread32(ipa->reg_virt + offset);
if (ipa->version == IPA_VERSION_4_0) {
val &= ~reg_bit(reg, IPA_QMB_SELECT_CONS_EN);
val &= ~reg_bit(reg, IPA_QMB_SELECT_PROD_EN);
val &= ~reg_bit(reg, IPA_QMB_SELECT_GLOBAL_EN);
} else if (ipa->version < IPA_VERSION_4_5) {
val |= reg_bit(reg, GSI_MULTI_AXI_MASTERS_DIS);
} else {
/* For IPA v4.5+ FULL_FLUSH_WAIT_RS_CLOSURE_EN is 0 */
}
val |= reg_bit(reg, GSI_MULTI_INORDER_RD_DIS);
val |= reg_bit(reg, GSI_MULTI_INORDER_WR_DIS);
iowrite32(val, ipa->reg_virt + offset);
}
/* Configure DDR and (possibly) PCIe max read/write QSB values */
static void
ipa_hardware_config_qsb(struct ipa *ipa, const struct ipa_data *data)
{
const struct ipa_qsb_data *data0;
const struct ipa_qsb_data *data1;
const struct reg *reg;
u32 val;
/* QMB 0 represents DDR; QMB 1 (if present) represents PCIe */
data0 = &data->qsb_data[IPA_QSB_MASTER_DDR];
if (data->qsb_count > 1)
data1 = &data->qsb_data[IPA_QSB_MASTER_PCIE];
/* Max outstanding write accesses for QSB masters */
reg = ipa_reg(ipa, QSB_MAX_WRITES);
val = reg_encode(reg, GEN_QMB_0_MAX_WRITES, data0->max_writes);
if (data->qsb_count > 1)
val |= reg_encode(reg, GEN_QMB_1_MAX_WRITES, data1->max_writes);
iowrite32(val, ipa->reg_virt + reg_offset(reg));
/* Max outstanding read accesses for QSB masters */
reg = ipa_reg(ipa, QSB_MAX_READS);
val = reg_encode(reg, GEN_QMB_0_MAX_READS, data0->max_reads);
if (ipa->version >= IPA_VERSION_4_0)
val |= reg_encode(reg, GEN_QMB_0_MAX_READS_BEATS,
data0->max_reads_beats);
if (data->qsb_count > 1) {
val = reg_encode(reg, GEN_QMB_1_MAX_READS, data1->max_reads);
if (ipa->version >= IPA_VERSION_4_0)
val |= reg_encode(reg, GEN_QMB_1_MAX_READS_BEATS,
data1->max_reads_beats);
}
iowrite32(val, ipa->reg_virt + reg_offset(reg));
}
/* The internal inactivity timer clock is used for the aggregation timer */
#define TIMER_FREQUENCY 32000 /* 32 KHz inactivity timer clock */
/* Compute the value to use in the COUNTER_CFG register AGGR_GRANULARITY
* field to represent the given number of microseconds. The value is one
* less than the number of timer ticks in the requested period. 0 is not
* a valid granularity value (so for example @usec must be at least 16 for
* a TIMER_FREQUENCY of 32000).
*/
static __always_inline u32 ipa_aggr_granularity_val(u32 usec)
{
return DIV_ROUND_CLOSEST(usec * TIMER_FREQUENCY, USEC_PER_SEC) - 1;
}
/* IPA uses unified Qtime starting at IPA v4.5, implementing various
* timestamps and timers independent of the IPA core clock rate. The
* Qtimer is based on a 56-bit timestamp incremented at each tick of
* a 19.2 MHz SoC crystal oscillator (XO clock).
*
* For IPA timestamps (tag, NAT, data path logging) a lower resolution
* timestamp is achieved by shifting the Qtimer timestamp value right
* some number of bits to produce the low-order bits of the coarser
* granularity timestamp.
*
* For timers, a common timer clock is derived from the XO clock using
* a divider (we use 192, to produce a 100kHz timer clock). From
* this common clock, three "pulse generators" are used to produce
* timer ticks at a configurable frequency. IPA timers (such as
* those used for aggregation or head-of-line block handling) now
* define their period based on one of these pulse generators.
*/
static void ipa_qtime_config(struct ipa *ipa)
{
const struct reg *reg;
u32 offset;
u32 val;
/* Timer clock divider must be disabled when we change the rate */
reg = ipa_reg(ipa, TIMERS_XO_CLK_DIV_CFG);
iowrite32(0, ipa->reg_virt + reg_offset(reg));
reg = ipa_reg(ipa, QTIME_TIMESTAMP_CFG);
if (ipa->version < IPA_VERSION_5_5) {
/* Set DPL time stamp resolution to use Qtime (not 1 msec) */
val = reg_encode(reg, DPL_TIMESTAMP_LSB, DPL_TIMESTAMP_SHIFT);
val |= reg_bit(reg, DPL_TIMESTAMP_SEL);
}
/* Configure tag and NAT Qtime timestamp resolution as well */
val = reg_encode(reg, TAG_TIMESTAMP_LSB, TAG_TIMESTAMP_SHIFT);
val = reg_encode(reg, NAT_TIMESTAMP_LSB, NAT_TIMESTAMP_SHIFT);
iowrite32(val, ipa->reg_virt + reg_offset(reg));
/* Set granularity of pulse generators used for other timers */
reg = ipa_reg(ipa, TIMERS_PULSE_GRAN_CFG);
val = reg_encode(reg, PULSE_GRAN_0, IPA_GRAN_100_US);
val |= reg_encode(reg, PULSE_GRAN_1, IPA_GRAN_1_MS);
if (ipa->version >= IPA_VERSION_5_0) {
val |= reg_encode(reg, PULSE_GRAN_2, IPA_GRAN_10_MS);
val |= reg_encode(reg, PULSE_GRAN_3, IPA_GRAN_10_MS);
} else {
val |= reg_encode(reg, PULSE_GRAN_2, IPA_GRAN_1_MS);
}
iowrite32(val, ipa->reg_virt + reg_offset(reg));
/* Actual divider is 1 more than value supplied here */
reg = ipa_reg(ipa, TIMERS_XO_CLK_DIV_CFG);
offset = reg_offset(reg);
val = reg_encode(reg, DIV_VALUE, IPA_XO_CLOCK_DIVIDER - 1);
iowrite32(val, ipa->reg_virt + offset);
/* Divider value is set; re-enable the common timer clock divider */
val |= reg_bit(reg, DIV_ENABLE);
iowrite32(val, ipa->reg_virt + offset);
}
/* Before IPA v4.5 timing is controlled by a counter register */
static void ipa_hardware_config_counter(struct ipa *ipa)
{
u32 granularity = ipa_aggr_granularity_val(IPA_AGGR_GRANULARITY);
const struct reg *reg;
u32 val;
reg = ipa_reg(ipa, COUNTER_CFG);
/* If defined, EOT_COAL_GRANULARITY is 0 */
val = reg_encode(reg, AGGR_GRANULARITY, granularity);
iowrite32(val, ipa->reg_virt + reg_offset(reg));
}
static void ipa_hardware_config_timing(struct ipa *ipa)
{
if (ipa->version < IPA_VERSION_4_5)
ipa_hardware_config_counter(ipa);
else
ipa_qtime_config(ipa);
}
static void ipa_hardware_config_hashing(struct ipa *ipa)
{
const struct reg *reg;
/* Other than IPA v4.2, all versions enable "hashing". Starting
* with IPA v5.0, the filter and router tables are implemented
* differently, but the default configuration enables this feature
* (now referred to as "cacheing"), so there's nothing to do here.
*/
if (ipa->version != IPA_VERSION_4_2)
return;
/* IPA v4.2 does not support hashed tables, so disable them */
reg = ipa_reg(ipa, FILT_ROUT_HASH_EN);
/* IPV6_ROUTER_HASH, IPV6_FILTER_HASH, IPV4_ROUTER_HASH,
* IPV4_FILTER_HASH are all zero.
*/
iowrite32(0, ipa->reg_virt + reg_offset(reg));
}
static void ipa_idle_indication_cfg(struct ipa *ipa,
u32 enter_idle_debounce_thresh,
bool const_non_idle_enable)
{
const struct reg *reg;
u32 val;
if (ipa->version < IPA_VERSION_3_5_1)
return;
reg = ipa_reg(ipa, IDLE_INDICATION_CFG);
val = reg_encode(reg, ENTER_IDLE_DEBOUNCE_THRESH,
enter_idle_debounce_thresh);
if (const_non_idle_enable)
val |= reg_bit(reg, CONST_NON_IDLE_ENABLE);
iowrite32(val, ipa->reg_virt + reg_offset(reg));
}
/**
* ipa_hardware_dcd_config() - Enable dynamic clock division on IPA
* @ipa: IPA pointer
*
* Configures when the IPA signals it is idle to the global clock
* controller, which can respond by scaling down the clock to save
* power.
*/
static void ipa_hardware_dcd_config(struct ipa *ipa)
{
/* Recommended values for IPA 3.5 and later according to IPA HPG */
ipa_idle_indication_cfg(ipa, 256, false);
}
static void ipa_hardware_dcd_deconfig(struct ipa *ipa)
{
/* Power-on reset values */
ipa_idle_indication_cfg(ipa, 0, true);
}
/**
* ipa_hardware_config() - Primitive hardware initialization
* @ipa: IPA pointer
* @data: IPA configuration data
*/
static void ipa_hardware_config(struct ipa *ipa, const struct ipa_data *data)
{
ipa_hardware_config_bcr(ipa, data);
ipa_hardware_config_tx(ipa);
ipa_hardware_config_clkon(ipa);
ipa_hardware_config_comp(ipa);
ipa_hardware_config_qsb(ipa, data);
ipa_hardware_config_timing(ipa);
ipa_hardware_config_hashing(ipa);
ipa_hardware_dcd_config(ipa);
}
/**
* ipa_hardware_deconfig() - Inverse of ipa_hardware_config()
* @ipa: IPA pointer
*
* This restores the power-on reset values (even if they aren't different)
*/
static void ipa_hardware_deconfig(struct ipa *ipa)
{
/* Mostly we just leave things as we set them. */
ipa_hardware_dcd_deconfig(ipa);
}
/**
* ipa_config() - Configure IPA hardware
* @ipa: IPA pointer
* @data: IPA configuration data
*
* Perform initialization requiring IPA power to be enabled.
*/
static int ipa_config(struct ipa *ipa, const struct ipa_data *data)
{
int ret;
ipa_hardware_config(ipa, data);
ret = ipa_mem_config(ipa);
if (ret)
goto err_hardware_deconfig;
ret = ipa_interrupt_config(ipa);
if (ret)
goto err_mem_deconfig;
ipa_uc_config(ipa);
ret = ipa_endpoint_config(ipa);
if (ret)
goto err_uc_deconfig;
ipa_table_config(ipa); /* No deconfig required */
/* Assign resource limitation to each group; no deconfig required */
ret = ipa_resource_config(ipa, data->resource_data);
if (ret)
goto err_endpoint_deconfig;
ret = ipa_modem_config(ipa);
if (ret)
goto err_endpoint_deconfig;
return 0;
err_endpoint_deconfig:
ipa_endpoint_deconfig(ipa);
err_uc_deconfig:
ipa_uc_deconfig(ipa);
ipa_interrupt_deconfig(ipa);
err_mem_deconfig:
ipa_mem_deconfig(ipa);
err_hardware_deconfig:
ipa_hardware_deconfig(ipa);
return ret;
}
/**
* ipa_deconfig() - Inverse of ipa_config()
* @ipa: IPA pointer
*/
static void ipa_deconfig(struct ipa *ipa)
{
ipa_modem_deconfig(ipa);
ipa_endpoint_deconfig(ipa);
ipa_uc_deconfig(ipa);
ipa_interrupt_deconfig(ipa);
ipa_mem_deconfig(ipa);
ipa_hardware_deconfig(ipa);
}
static int ipa_firmware_load(struct device *dev)
{
const struct firmware *fw;
struct device_node *node;
struct resource res;
phys_addr_t phys;
const char *path;
ssize_t size;
void *virt;
int ret;
node = of_parse_phandle(dev->of_node, "memory-region", 0);
if (!node) {
dev_err(dev, "DT error getting \"memory-region\" property\n");
return -EINVAL;
}
ret = of_address_to_resource(node, 0, &res);
of_node_put(node);
if (ret) {
dev_err(dev, "error %d getting \"memory-region\" resource\n",
ret);
return ret;
}
/* Use name from DTB if specified; use default for *any* error */
ret = of_property_read_string(dev->of_node, "firmware-name", &path);
if (ret) {
dev_dbg(dev, "error %d getting \"firmware-name\" resource\n",
ret);
path = IPA_FW_PATH_DEFAULT;
}
ret = request_firmware(&fw, path, dev);
if (ret) {
dev_err(dev, "error %d requesting \"%s\"\n", ret, path);
return ret;
}
phys = res.start;
size = (size_t)resource_size(&res);
virt = memremap(phys, size, MEMREMAP_WC);
if (!virt) {
dev_err(dev, "unable to remap firmware memory\n");
ret = -ENOMEM;
goto out_release_firmware;
}
ret = qcom_mdt_load(dev, fw, path, IPA_PAS_ID, virt, phys, size, NULL);
if (ret)
dev_err(dev, "error %d loading \"%s\"\n", ret, path);
else if ((ret = qcom_scm_pas_auth_and_reset(IPA_PAS_ID)))
dev_err(dev, "error %d authenticating \"%s\"\n", ret, path);
memunmap(virt);
out_release_firmware:
release_firmware(fw);
return ret;
}
static const struct of_device_id ipa_match[] = {
{
.compatible = "qcom,msm8998-ipa",
.data = &ipa_data_v3_1,
},
{
.compatible = "qcom,sdm845-ipa",
.data = &ipa_data_v3_5_1,
},
{
.compatible = "qcom,sc7180-ipa",
.data = &ipa_data_v4_2,
},
{
.compatible = "qcom,sdx55-ipa",
.data = &ipa_data_v4_5,
},
{
.compatible = "qcom,sm6350-ipa",
.data = &ipa_data_v4_7,
},
{
.compatible = "qcom,sm8350-ipa",
.data = &ipa_data_v4_9,
},
{
.compatible = "qcom,sc7280-ipa",
.data = &ipa_data_v4_11,
},
{
.compatible = "qcom,sdx65-ipa",
.data = &ipa_data_v5_0,
},
{
.compatible = "qcom,sm8550-ipa",
.data = &ipa_data_v5_5,
},
{ },
};
MODULE_DEVICE_TABLE(of, ipa_match);
/* Check things that can be validated at build time. This just
* groups these things BUILD_BUG_ON() calls don't clutter the rest
* of the code.
* */
static void ipa_validate_build(void)
{
/* At one time we assumed a 64-bit build, allowing some do_div()
* calls to be replaced by simple division or modulo operations.
* We currently only perform divide and modulo operations on u32,
* u16, or size_t objects, and of those only size_t has any chance
* of being a 64-bit value. (It should be guaranteed 32 bits wide
* on a 32-bit build, but there is no harm in verifying that.)
*/
BUILD_BUG_ON(!IS_ENABLED(CONFIG_64BIT) && sizeof(size_t) != 4);
/* Code assumes the EE ID for the AP is 0 (zeroed structure field) */
BUILD_BUG_ON(GSI_EE_AP != 0);
/* There's no point if we have no channels or event rings */
BUILD_BUG_ON(!GSI_CHANNEL_COUNT_MAX);
BUILD_BUG_ON(!GSI_EVT_RING_COUNT_MAX);
/* GSI hardware design limits */
BUILD_BUG_ON(GSI_CHANNEL_COUNT_MAX > 32);
BUILD_BUG_ON(GSI_EVT_RING_COUNT_MAX > 31);
/* The number of TREs in a transaction is limited by the channel's
* TLV FIFO size. A transaction structure uses 8-bit fields
* to represents the number of TREs it has allocated and used.
*/
BUILD_BUG_ON(GSI_TLV_MAX > U8_MAX);
/* This is used as a divisor */
BUILD_BUG_ON(!IPA_AGGR_GRANULARITY);
/* Aggregation granularity value can't be 0, and must fit */
BUILD_BUG_ON(!ipa_aggr_granularity_val(IPA_AGGR_GRANULARITY));
}
static enum ipa_firmware_loader ipa_firmware_loader(struct device *dev)
{
bool modem_init;
const char *str;
int ret;
/* Look up the old and new properties by name */
modem_init = of_property_read_bool(dev->of_node, "modem-init");
ret = of_property_read_string(dev->of_node, "qcom,gsi-loader", &str);
/* If the new property doesn't exist, it's legacy behavior */
if (ret == -EINVAL) {
if (modem_init)
return IPA_LOADER_MODEM;
goto out_self;
}
/* Any other error on the new property means it's poorly defined */
if (ret)
return IPA_LOADER_INVALID;
/* New property value exists; if old one does too, that's invalid */
if (modem_init)
return IPA_LOADER_INVALID;
/* Modem loads GSI firmware for "modem" */
if (!strcmp(str, "modem"))
return IPA_LOADER_MODEM;
/* No GSI firmware load is needed for "skip" */
if (!strcmp(str, "skip"))
return IPA_LOADER_SKIP;
/* Any value other than "self" is an error */
if (strcmp(str, "self"))
return IPA_LOADER_INVALID;
out_self:
/* We need Trust Zone to load firmware; make sure it's available */
if (qcom_scm_is_available())
return IPA_LOADER_SELF;
return IPA_LOADER_DEFER;
}
/**
* ipa_probe() - IPA platform driver probe function
* @pdev: Platform device pointer
*
* Return: 0 if successful, or a negative error code (possibly
* EPROBE_DEFER)
*
* This is the main entry point for the IPA driver. Initialization proceeds
* in several stages:
* - The "init" stage involves activities that can be initialized without
* access to the IPA hardware.
* - The "config" stage requires IPA power to be active so IPA registers
* can be accessed, but does not require the use of IPA immediate commands.
* - The "setup" stage uses IPA immediate commands, and so requires the GSI
* layer to be initialized.
*
* A Boolean Device Tree "modem-init" property determines whether GSI
* initialization will be performed by the AP (Trust Zone) or the modem.
* If the AP does GSI initialization, the setup phase is entered after
* this has completed successfully. Otherwise the modem initializes
* the GSI layer and signals it has finished by sending an SMP2P interrupt
* to the AP; this triggers the start if IPA setup.
*/
static int ipa_probe(struct platform_device *pdev)
{
struct device *dev = &pdev->dev;
struct ipa_interrupt *interrupt;
enum ipa_firmware_loader loader;
const struct ipa_data *data;
struct ipa_power *power;
struct ipa *ipa;
int ret;
ipa_validate_build();
/* Get configuration data early; needed for power initialization */
data = of_device_get_match_data(dev);
if (!data) {
dev_err(dev, "matched hardware not supported\n");
return -ENODEV;
}
if (!ipa_version_supported(data->version)) {
dev_err(dev, "unsupported IPA version %u\n", data->version);
return -EINVAL;
}
if (!data->modem_route_count) {
dev_err(dev, "modem_route_count cannot be zero\n");
return -EINVAL;
}
loader = ipa_firmware_loader(dev);
if (loader == IPA_LOADER_INVALID)
return -EINVAL;
if (loader == IPA_LOADER_DEFER)
return -EPROBE_DEFER;
/* The IPA interrupt might not be ready when we're probed, so this
* might return -EPROBE_DEFER.
*/
interrupt = ipa_interrupt_init(pdev);
if (IS_ERR(interrupt))
return PTR_ERR(interrupt);
/* The clock and interconnects might not be ready when we're probed,
* so this might return -EPROBE_DEFER.
*/
power = ipa_power_init(dev, data->power_data);
if (IS_ERR(power)) {
ret = PTR_ERR(power);
goto err_interrupt_exit;
}
/* No more EPROBE_DEFER. Allocate and initialize the IPA structure */
ipa = kzalloc(sizeof(*ipa), GFP_KERNEL);
if (!ipa) {
ret = -ENOMEM;
goto err_power_exit;
}
ipa->dev = dev;
dev_set_drvdata(dev, ipa);
ipa->interrupt = interrupt;
ipa->power = power;
ipa->version = data->version;
ipa->modem_route_count = data->modem_route_count;
init_completion(&ipa->completion);
ret = ipa_reg_init(ipa, pdev);
if (ret)
goto err_kfree_ipa;
ret = ipa_mem_init(ipa, pdev, data->mem_data);
if (ret)
goto err_reg_exit;
ret = gsi_init(&ipa->gsi, pdev, ipa->version, data->endpoint_count,
data->endpoint_data);
if (ret)
goto err_mem_exit;
/* Result is a non-zero mask of endpoints that support filtering */
ret = ipa_endpoint_init(ipa, data->endpoint_count, data->endpoint_data);
if (ret)
goto err_gsi_exit;
ret = ipa_table_init(ipa);
if (ret)
goto err_endpoint_exit;
ret = ipa_smp2p_init(ipa, pdev, loader == IPA_LOADER_MODEM);
if (ret)
goto err_table_exit;
/* Power needs to be active for config and setup */
ret = pm_runtime_get_sync(dev);
if (WARN_ON(ret < 0))
goto err_power_put;
ret = ipa_config(ipa, data);
if (ret)
goto err_power_put;
dev_info(dev, "IPA driver initialized");
/* If the modem is loading GSI firmware, it will trigger a call to
* ipa_setup() when it has finished. In that case we're done here.
*/
if (loader == IPA_LOADER_MODEM)
goto done;
if (loader == IPA_LOADER_SELF) {
/* The AP is loading GSI firmware; do so now */
ret = ipa_firmware_load(dev);
if (ret)
goto err_deconfig;
} /* Otherwise loader == IPA_LOADER_SKIP */
/* GSI firmware is loaded; proceed to setup */
ret = ipa_setup(ipa);
if (ret)
goto err_deconfig;
done:
pm_runtime_mark_last_busy(dev);
(void)pm_runtime_put_autosuspend(dev);
return 0;
err_deconfig:
ipa_deconfig(ipa);
err_power_put:
pm_runtime_put_noidle(dev);
ipa_smp2p_exit(ipa);
err_table_exit:
ipa_table_exit(ipa);
err_endpoint_exit:
ipa_endpoint_exit(ipa);
err_gsi_exit:
gsi_exit(&ipa->gsi);
err_mem_exit:
ipa_mem_exit(ipa);
err_reg_exit:
ipa_reg_exit(ipa);
err_kfree_ipa:
kfree(ipa);
err_power_exit:
ipa_power_exit(power);
err_interrupt_exit:
ipa_interrupt_exit(interrupt);
return ret;
}
static void ipa_remove(struct platform_device *pdev)
{
struct ipa_interrupt *interrupt;
struct ipa_power *power;
struct device *dev;
struct ipa *ipa;
int ret;
ipa = dev_get_drvdata(&pdev->dev);
dev = ipa->dev;
WARN_ON(dev != &pdev->dev);
power = ipa->power;
interrupt = ipa->interrupt;
/* Prevent the modem from triggering a call to ipa_setup(). This
* also ensures a modem-initiated setup that's underway completes.
*/
ipa_smp2p_irq_disable_setup(ipa);
ret = pm_runtime_get_sync(dev);
if (WARN_ON(ret < 0))
goto out_power_put;
if (ipa->setup_complete) {
ret = ipa_modem_stop(ipa);
/* If starting or stopping is in progress, try once more */
if (ret == -EBUSY) {
usleep_range(USEC_PER_MSEC, 2 * USEC_PER_MSEC);
ret = ipa_modem_stop(ipa);
}
if (ret) {
/*
* Not cleaning up here properly might also yield a
* crash later on. As the device is still unregistered
* in this case, this might even yield a crash later on.
*/
dev_err(dev, "Failed to stop modem (%pe), leaking resources\n",
ERR_PTR(ret));
return;
}
ipa_teardown(ipa);
}
ipa_deconfig(ipa);
out_power_put:
pm_runtime_put_noidle(dev);
ipa_smp2p_exit(ipa);
ipa_table_exit(ipa);
ipa_endpoint_exit(ipa);
gsi_exit(&ipa->gsi);
ipa_mem_exit(ipa);
ipa_reg_exit(ipa);
kfree(ipa);
ipa_power_exit(power);
ipa_interrupt_exit(interrupt);
dev_info(dev, "IPA driver removed");
}
static const struct attribute_group *ipa_attribute_groups[] = {
&ipa_attribute_group,
&ipa_feature_attribute_group,
&ipa_endpoint_id_attribute_group,
&ipa_modem_attribute_group,
NULL,
};
static struct platform_driver ipa_driver = {
.probe = ipa_probe,
.remove_new = ipa_remove,
.shutdown = ipa_remove,
.driver = {
.name = "ipa",
.pm = &ipa_pm_ops,
.of_match_table = ipa_match,
.dev_groups = ipa_attribute_groups,
},
};
module_platform_driver(ipa_driver);
MODULE_LICENSE("GPL v2");
MODULE_DESCRIPTION("Qualcomm IP Accelerator device driver");