Hello,
I've done further investigations on this issue, since I wasn't
understanding why we were loosing interrupts where we shouldn't. And I
think I found the root cause.
The irq handlers top and bottom are written in a way considering that
every interrupt will be handled by the top part in the hard irq handler,
even if the irq thread is still running in the bottom part. That's why
we are using a FIFO for storing multiple timestamps.
But by using triggered-buffer setup, we are using the ONESHOT irq mode
that is disabling irq until the thread processing is done. In this way,
we are never reliably catching data interrupts when the thread is
running, and there is no need to have a FIFO for storing timestamps.
The correct way would be to have the top part in the hard irq handler
without using ONESHOT mode, and use only the thread for reading FIFO
data in the triggered-buffer interrupt.
I have done a patch which is working correctly on my side. But since I
am not able to trigger the issue in the first time, I cannot guarantee
this is fixing the problem.
Martin,
is it possible for you to test the following patch and tell me if it is
working better like that?
For sure, there is a conception issue here, because in the current setup
having a FIFO for storing timestamps is completely useless.
JB
Subject: [PATCH] iio: imu: inv_mpu6050: use top irq handler in the hard irq
handler
Signed-off-by: Jean-Baptiste Maneyrol <jmaneyrol@xxxxxxxxxxxxxx>
---
drivers/iio/imu/inv_mpu6050/inv_mpu_core.c | 2 +-
drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c | 6 +++---
drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c | 2 +-
3 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c
b/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c
index 7d64be3..ed883ef 100644
--- a/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c
+++ b/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c
@@ -941,7 +941,7 @@ int inv_mpu_core_probe(struct regmap *regmap, int
irq, const char *name,
indio_dev->modes = INDIO_BUFFER_TRIGGERED;
result = iio_triggered_buffer_setup(indio_dev,
- inv_mpu6050_irq_handler,
+ NULL,
inv_mpu6050_read_fifo,
NULL);
if (result) {
diff --git a/drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c
b/drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c
index ff81c6a..6935815 100644
--- a/drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c
+++ b/drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c
@@ -102,8 +102,8 @@ int inv_reset_fifo(struct iio_dev *indio_dev)
*/
irqreturn_t inv_mpu6050_irq_handler(int irq, void *p)
{
- struct iio_poll_func *pf = p;
- struct iio_dev *indio_dev = pf->indio_dev;
+ struct iio_trigger *trig = p;
+ struct iio_dev *indio_dev = iio_trigger_get_drvdata(trig);
struct inv_mpu6050_state *st = iio_priv(indio_dev);
s64 timestamp;
@@ -111,7 +111,7 @@ irqreturn_t inv_mpu6050_irq_handler(int irq, void *p)
kfifo_in_spinlocked(&st->timestamps, ×tamp, 1,
&st->time_stamp_lock);
- return IRQ_WAKE_THREAD;
+ return iio_trigger_generic_data_rdy_poll(irq, trig);
}
/**
diff --git a/drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c
b/drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c
index f963f9f..1cb9ed7 100644
--- a/drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c
+++ b/drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c
@@ -130,7 +130,7 @@ int inv_mpu6050_probe_trigger(struct iio_dev *indio_dev)
return -ENOMEM;
ret = devm_request_irq(&indio_dev->dev, st->irq,
- &iio_trigger_generic_data_rdy_poll,
+ &inv_mpu6050_irq_handler,
IRQF_TRIGGER_RISING,
"inv_mpu",
st->trig);
--
2.7.4
On 11/04/2018 18:42, Martin Kelly wrote:
On 04/11/2018 12:01 AM, Jean-Baptiste Maneyrol wrote:
This is OK for me.
Jonathan will tell us about EBUSY error code for sure if it is not
correct.
JB
Sounds good; once we hear from Jonathan, I will submit the next revision.
On 10/04/2018 20:08, Martin Kelly wrote:
Thanks, replies inline. I made all the changes you mentioned except
the one about -EBUSY. Let me know what you think regarding
-EBUSY/-ENODEV and then I'll send a new revision with all the changes
included.
On 04/10/2018 02:06 AM, Jean-Baptiste Maneyrol wrote:
Hello,
some more concerns after a deeper look.
Thanks.
JB
On 09/04/2018 22:21, Martin Kelly wrote:
Currently, we support only rising edge interrupts, and in fact we
assume
that the interrupt we're given is rising edge (and things won't
work if
it's not). However, the device supports rising edge, falling edge,
level
low, and level high interrupts.
Empirically, on my system, switching to level interrupts has fixed a
problem I had with significant (~40%) interrupt loss with edge
interrupts. This issue is likely related to the SoC I'm using
(Allwinner
H3), but being able to switch the interrupt type is still a very
useful
workaround.
I tested this with each interrupt type and verified correct
behavior in
a logic analyzer.
Add support for these interrupt types while also eliminating the error
case of the device tree and driver using different interrupt types.
Signed-off-by: Martin Kelly <mkelly@xxxxxxxx>
---
drivers/iio/imu/inv_mpu6050/inv_mpu_core.c | 33
++++++++++++++++++++++++++-
drivers/iio/imu/inv_mpu6050/inv_mpu_i2c.c | 5 ++--
drivers/iio/imu/inv_mpu6050/inv_mpu_iio.h | 14 ++++++++++--
drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c | 13 +++++++++++
drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c | 4 ++--
5 files changed, 61 insertions(+), 8 deletions(-)
v2:
- Changed to ACK level interrupts at the start of the bottom half
thread instead
of at the end of it. Without this, the sample timestamps get
distorted because
the time to handle the bottom half thread delays future
interrupts. With this
change, the timestamps appear evenly distributed at the right
frequency.
v3:
- Sent version 2 too quickly. Now that the ACK is moved to the top
of the
function, the "goto out" logic is unnecessary, so we can clean
it up.
v4:
- Moved the ACK inside the mutex.
v5:
- Check interrupt status in all cases rather than only in the level
triggered
case, and warn if we get spurious interrupts.
- Rename INV_MPU6050_LEVEL_TRIGGERED to INV_MPU6050_LATCH_INT_EN to
match
datasheet naming.
- Write st->irq_mask prior to device power off to make sure it is
fully set.
- Write st->irq_mask instead of 0 in inv_mpu6050_deselect_bypass.
- Make irq_type local instead of part of the driver state, as we
use it only at
probe time and never again.
- Remove the comment about bus lockups, as I have determined them
to be
unrelated.
- Add missing documentation for irq_type and irq_mask.
diff --git a/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c
b/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c
index 7d64be353403..b711e6260d9a 100644
--- a/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c
+++ b/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c
@@ -24,6 +24,7 @@
#include <linux/spinlock.h>
#include <linux/iio/iio.h>
#include <linux/acpi.h>
+#include <linux/platform_device.h>
#include "inv_mpu_iio.h"
/*
@@ -52,6 +53,7 @@ static const struct inv_mpu6050_reg_map
reg_set_6500 = {
.raw_accl = INV_MPU6050_REG_RAW_ACCEL,
.temperature = INV_MPU6050_REG_TEMPERATURE,
.int_enable = INV_MPU6050_REG_INT_ENABLE,
+ .int_status = INV_MPU6050_REG_INT_STATUS,
.pwr_mgmt_1 = INV_MPU6050_REG_PWR_MGMT_1,
.pwr_mgmt_2 = INV_MPU6050_REG_PWR_MGMT_2,
.int_pin_cfg = INV_MPU6050_REG_INT_PIN_CFG,
@@ -278,6 +280,10 @@ static int inv_mpu6050_init_config(struct
iio_dev *indio_dev)
if (result)
return result;
+ result = regmap_write(st->map, st->reg->int_pin_cfg,
st->irq_mask);
+ if (result)
+ return result;
+
memcpy(&st->chip_config, hw_info[st->chip_type].config,
sizeof(struct inv_mpu6050_chip_config));
result = inv_mpu6050_set_power_itg(st, false);
@@ -882,6 +888,8 @@ int inv_mpu_core_probe(struct regmap *regmap,
int irq, const char *name,
struct inv_mpu6050_platform_data *pdata;
struct device *dev = regmap_get_device(regmap);
int result;
+ struct irq_data *desc;
+ int irq_type;
indio_dev = devm_iio_device_alloc(dev, sizeof(*st));
if (!indio_dev)
@@ -913,6 +921,29 @@ int inv_mpu_core_probe(struct regmap *regmap,
int irq, const char *name,
st->plat_data = *pdata;
}
+ desc = irq_get_irq_data(irq);
+ if (!desc) {
+ dev_err(dev, "Could not find IRQ %d\n", irq);
+ return -EBUSY;
I would prefer -ENODEV instead. There is nothing busy as far as I
understand.
I picked -EBUSY based on guidance from this thread:
https://lists.gt.net/linux/kernel/1010603
akpm seemed happy with this treatment of -EBUSY and -ENODEV, but I
can't see whether or not this guidance was formalized.
Please let me know which error code is more appropriate, and I'll
change it.
+ }
+
+ irq_type = irqd_get_trigger_type(desc);
+ if (irq_type == IRQF_TRIGGER_RISING)
+ st->irq_mask = INV_MPU6050_ACTIVE_HIGH;
+ else if (irq_type == IRQF_TRIGGER_FALLING)
+ st->irq_mask = INV_MPU6050_ACTIVE_LOW;
+ else if (irq_type == IRQF_TRIGGER_HIGH)
+ st->irq_mask = INV_MPU6050_ACTIVE_HIGH |
+ INV_MPU6050_LATCH_INT_EN;
+ else if (irq_type == IRQF_TRIGGER_LOW)
+ st->irq_mask = INV_MPU6050_ACTIVE_LOW |
+ INV_MPU6050_LATCH_INT_EN;
+ else {
+ dev_err(dev, "Invalid interrupt type 0x%x specified\n",
+ irq_type);
+ return -EBUSY;
Same here, -ENODEV makes more sense if I'm understanding well.
+ }
+
/* power is turned on inside check chip type*/
result = inv_check_and_setup_chip(st);
if (result)
@@ -948,7 +979,7 @@ int inv_mpu_core_probe(struct regmap *regmap,
int irq, const char *name,
dev_err(dev, "configure buffer fail %d\n", result);
return result;
}
- result = inv_mpu6050_probe_trigger(indio_dev);
+ result = inv_mpu6050_probe_trigger(indio_dev, irq_type);
if (result) {
dev_err(dev, "trigger probe fail %d\n", result);
goto out_unreg_ring;
diff --git a/drivers/iio/imu/inv_mpu6050/inv_mpu_i2c.c
b/drivers/iio/imu/inv_mpu6050/inv_mpu_i2c.c
index fcd7a92b6cf8..1b02d2b69174 100644
--- a/drivers/iio/imu/inv_mpu6050/inv_mpu_i2c.c
+++ b/drivers/iio/imu/inv_mpu6050/inv_mpu_i2c.c
@@ -44,8 +44,7 @@ static int inv_mpu6050_select_bypass(struct
i2c_mux_core *muxc, u32 chan_id)
if (!ret) {
st->powerup_count++;
ret = regmap_write(st->map, st->reg->int_pin_cfg,
- INV_MPU6050_INT_PIN_CFG |
- INV_MPU6050_BIT_BYPASS_EN);
+ st->irq_mask | INV_MPU6050_BIT_BYPASS_EN);
}
write_error:
mutex_unlock(&st->lock);
@@ -60,7 +59,7 @@ static int inv_mpu6050_deselect_bypass(struct
i2c_mux_core *muxc, u32 chan_id)
mutex_lock(&st->lock);
/* It doesn't really mattter, if any of the calls fails */
- regmap_write(st->map, st->reg->int_pin_cfg,
INV_MPU6050_INT_PIN_CFG);
+ regmap_write(st->map, st->reg->int_pin_cfg, st->irq_mask);
st->powerup_count--;
if (!st->powerup_count)
regmap_write(st->map, st->reg->pwr_mgmt_1,
diff --git a/drivers/iio/imu/inv_mpu6050/inv_mpu_iio.h
b/drivers/iio/imu/inv_mpu6050/inv_mpu_iio.h
index 065794162d65..064e3b28fdb0 100644
--- a/drivers/iio/imu/inv_mpu6050/inv_mpu_iio.h
+++ b/drivers/iio/imu/inv_mpu6050/inv_mpu_iio.h
@@ -40,6 +40,7 @@
* @raw_accl: Address of first accel register.
* @temperature: temperature register
* @int_enable: Interrupt enable register.
+ * @int_status: Interrupt status register.
* @pwr_mgmt_1: Controls chip's power state and clock source.
* @pwr_mgmt_2: Controls power state of individual sensors.
* @int_pin_cfg; Controls interrupt pin configuration.
@@ -60,6 +61,7 @@ struct inv_mpu6050_reg_map {
u8 raw_accl;
u8 temperature;
u8 int_enable;
+ u8 int_status;
u8 pwr_mgmt_1;
u8 pwr_mgmt_2;
u8 int_pin_cfg;
@@ -125,6 +127,7 @@ struct inv_mpu6050_hw {
* @timestamps: kfifo queue to store time stamp.
* @map regmap pointer.
* @irq interrupt number.
+ * @irq_mask the int_pin_cfg mask to configure interrupt type
*/
struct inv_mpu6050_state {
#define TIMESTAMP_FIFO_SIZE 16
@@ -143,6 +146,7 @@ struct inv_mpu6050_state {
DECLARE_KFIFO(timestamps, long long, TIMESTAMP_FIFO_SIZE);
struct regmap *map;
int irq;
+ u16 irq_mask;
It is data for a 8 bits register. u8 should be sufficient.
Good catch.
};
/*register and associated bit definition*/
@@ -159,6 +163,8 @@ struct inv_mpu6050_state {
#define INV_MPU6050_BITS_GYRO_OUT 0x70
#define INV_MPU6050_REG_INT_ENABLE 0x38
+#define INV_MPU6050_REG_INT_STATUS 0x3a
+#define INV_MPU6050_BIT_RAW_DATA_RDY_INT 0x01
#define INV_MPU6050_BIT_DATA_RDY_EN 0x01
#define INV_MPU6050_BIT_DMP_INT_EN 0x02
Beware you are mixing register and bit defines here. Better put the
new definition below and let the interrupt enable BIT defines just
below the corresponding REG define.
@@ -190,6 +196,11 @@ struct inv_mpu6050_state {
#define INV_MPU6050_FIFO_COUNT_BYTE 2
#define INV_MPU6050_FIFO_THRESHOLD 500
+#define INV_MPU6050_ACTIVE_HIGH 0x00
+#define INV_MPU6050_ACTIVE_LOW 0x80
+/* enable level triggering */
+#define INV_MPU6050_LATCH_INT_EN 0x20
Would be better to have this amoung the bit defines of register
INT_PIN_CFG. I would prefer just to add these defines below the
REG_INT_PIN_CFG:
#define INV_MPU6050_BIT_ACTIVE_LOW 0x80
#define INV_MPU6050_BIT_LATCH_INT_EN 0x20
+
/* mpu6500 registers */
#define INV_MPU6500_REG_ACCEL_CONFIG_2 0x1D
#define INV_MPU6500_REG_ACCEL_OFFSET 0x77
@@ -216,7 +227,6 @@ struct inv_mpu6050_state {
#define INV_MPU6050_REG_INT_PIN_CFG 0x37
#define INV_MPU6050_BIT_BYPASS_EN 0x2
-#define INV_MPU6050_INT_PIN_CFG 0
/* init parameters */
#define INV_MPU6050_INIT_FIFO_RATE 50
@@ -287,7 +297,7 @@ enum inv_mpu6050_clock_sel_e {
irqreturn_t inv_mpu6050_irq_handler(int irq, void *p);
irqreturn_t inv_mpu6050_read_fifo(int irq, void *p);
-int inv_mpu6050_probe_trigger(struct iio_dev *indio_dev);
+int inv_mpu6050_probe_trigger(struct iio_dev *indio_dev, int
irq_type);
void inv_mpu6050_remove_trigger(struct inv_mpu6050_state *st);
int inv_reset_fifo(struct iio_dev *indio_dev);
int inv_mpu6050_switch_engine(struct inv_mpu6050_state *st, bool
en, u32 mask);
diff --git a/drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c
b/drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c
index ff81c6aa009d..8f1f637fb972 100644
--- a/drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c
+++ b/drivers/iio/imu/inv_mpu6050/inv_mpu_ring.c
@@ -127,8 +127,21 @@ irqreturn_t inv_mpu6050_read_fifo(int irq,
void *p)
u8 data[INV_MPU6050_OUTPUT_DATA_SIZE];
u16 fifo_count;
s64 timestamp;
+ int int_status;
mutex_lock(&st->lock);
+
+ /* ack interrupt and check status */
+ result = regmap_read(st->map, st->reg->int_status, &int_status);
+ if (result)
+ dev_err(regmap_get_device(st->map),
If we cannot read int status, I would exit with error using
flush_fifo error path. This would ensure the interrupt would be
resetted.
I will do that. Depending on the error cause, resetting the FIFO may
not work either (for instance, if you get bus lockup, all I2C
transactions will fail). But it's at least worth a shot as it will
solve some error cases.
+ "failed to ack interrupt\n");
+ if (!(int_status & INV_MPU6050_BIT_RAW_DATA_RDY_INT)) {
+ dev_warn(regmap_get_device(st->map),
+ "spurious interrupt with status 0x%x\n", int_status);
+ goto end_session;
+ }
+
if (!(st->chip_config.accl_fifo_enable |
st->chip_config.gyro_fifo_enable))
goto end_session;
diff --git a/drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c
b/drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c
index f963f9fc98c0..b8c5584e4252 100644
--- a/drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c
+++ b/drivers/iio/imu/inv_mpu6050/inv_mpu_trigger.c
@@ -117,7 +117,7 @@ static const struct iio_trigger_ops
inv_mpu_trigger_ops = {
.set_trigger_state = &inv_mpu_data_rdy_trigger_set_state,
};
-int inv_mpu6050_probe_trigger(struct iio_dev *indio_dev)
+int inv_mpu6050_probe_trigger(struct iio_dev *indio_dev, int
irq_type)
{
int ret;
struct inv_mpu6050_state *st = iio_priv(indio_dev);
@@ -131,7 +131,7 @@ int inv_mpu6050_probe_trigger(struct iio_dev
*indio_dev)
ret = devm_request_irq(&indio_dev->dev, st->irq,
&iio_trigger_generic_data_rdy_poll,
- IRQF_TRIGGER_RISING,
+ irq_type,
"inv_mpu",
st->trig);
if (ret)
--
2.11.0
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
