* [PATCH v1 51/63] input: Atmel: limit the max bytes transferred while reading T5 messages
2019-08-16 8:37 [PATCH v1 50/63] Input: Atmel: handle ReportID "0x00" while processing T5 messages Jiada Wang
@ 2019-08-16 8:37 ` Jiada Wang
2019-08-16 8:37 ` [PATCH v1 52/63] Input: Atmel: use T44 object to process " Jiada Wang
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Jiada Wang @ 2019-08-16 8:37 UTC (permalink / raw)
To: nick, dmitry.torokhov; +Cc: linux-input, linux-kernel, jiada_wang, george_davis
From: Deepak Das <deepak_das@mentor.com>
"a008387 Input: atmel_mxt_ts: Limit the max bytes transferred in an i2c
transaction" limited the Max bytes transferred in i2c transaction while
reading T5 message in mxt_process_messages_until_invalid().
mxt_process_messages_t44() reads the T44 message which contains the
pending T5 message count. If the number of pending T5 messages returned
by T44 message is too high then there is a risk of i2c transaction
timeout while reading T5 messages in mxt_process_messages_t44().
To avoid i2c transaction timeout, this commit limits the max bytes
transferred in a single transaction while reading T5 messages in
mxt_process_messages_t44() depending on the property 'atmel,mtu'.
This improvement is created based on the commit "a008387 Input:
atmel_mxt_ts: Limit the max bytes transferred in an i2c
transaction"
Signed-off-by: Deepak Das <deepak_das@mentor.com>
Signed-off-by: Sanjeev Chugh <sanjeev_chugh@mentor.com>
Signed-off-by: George G. Davis <george_davis@mentor.com>
Signed-off-by: Jiada Wang <jiada_wang@mentor.com>
---
drivers/input/touchscreen/atmel_mxt_ts.c | 73 ++++++++++++++----------
1 file changed, 44 insertions(+), 29 deletions(-)
diff --git a/drivers/input/touchscreen/atmel_mxt_ts.c b/drivers/input/touchscreen/atmel_mxt_ts.c
index 2041a82a4551..47c1e72152de 100644
--- a/drivers/input/touchscreen/atmel_mxt_ts.c
+++ b/drivers/input/touchscreen/atmel_mxt_ts.c
@@ -1472,11 +1472,32 @@ static int mxt_read_and_process_messages(struct mxt_data *data, u8 count)
return num_valid;
}
+static u8 mxt_max_msg_read_count(struct mxt_data *data, u8 max_T5_msg_count)
+{
+ u8 T5_msg_count_limit = data->mtu / data->T5_msg_size;
+
+ if (!data->mtu)
+ return max_T5_msg_count;
+
+ if (data->mtu < data->T5_msg_size) {
+ WARN(1, "mtu set is lesser than the T5 message size\n");
+ /* Return count of 1, as fallback */
+ return 1;
+ }
+ /*
+ * Return maximum number of T5 messages in single i2c transaction
+ * based on "atmel,mtu" property.
+ */
+ return min(T5_msg_count_limit, max_T5_msg_count);
+}
+
static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
{
struct device *dev = &data->client->dev;
int ret;
- u8 count, num_left;
+ u8 T5_msg_count, total_pending;
+ u8 total_processed = 0;
+ u8 processed_valid = 0;
/* Read T44 and T5 together */
ret = __mxt_read_reg(data->client, data->T44_address,
@@ -1486,18 +1507,19 @@ static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
return IRQ_NONE;
}
- count = data->msg_buf[0];
+ T5_msg_count = data->msg_buf[0];
/*
* This condition may be caused by the CHG line being configured in
* Mode 0. It results in unnecessary I2C operations but it is benign.
*/
- if (count == 0)
+ if (!T5_msg_count)
return IRQ_NONE;
- if (count > data->max_reportid) {
- dev_warn(dev, "T44 count %d exceeded max report id\n", count);
- count = data->max_reportid;
+ if (T5_msg_count > data->max_reportid) {
+ dev_warn(dev, "T44 count %d exceeded max report id\n",
+ T5_msg_count);
+ T5_msg_count = data->max_reportid;
}
/* Process first message */
@@ -1507,16 +1529,25 @@ static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
return IRQ_NONE;
}
- num_left = count - 1;
+ total_pending = T5_msg_count - 1;
+ if (!total_pending)
+ goto end;
/* Process remaining messages if necessary */
- if (num_left) {
- ret = mxt_read_and_process_messages(data, num_left);
+ T5_msg_count = mxt_max_msg_read_count(data, total_pending);
+
+ do {
+ if ((total_pending - total_processed) < T5_msg_count)
+ T5_msg_count = total_pending - total_processed;
+ ret = mxt_read_and_process_messages(data, T5_msg_count);
if (ret < 0)
goto end;
- else if (ret != num_left)
- dev_warn(dev, "Unexpected invalid message\n");
- }
+ total_processed += T5_msg_count;
+ processed_valid += ret;
+ } while (total_processed < total_pending);
+
+ if (processed_valid != total_pending)
+ dev_warn(dev, "Unexpected invalid message\n");
end:
if (data->update_input) {
@@ -1527,29 +1558,13 @@ static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
return IRQ_HANDLED;
}
-static u8 mxt_max_msg_read_count(struct mxt_data *data)
-{
- u8 count_limit = data->mtu / data->T5_msg_size;
-
- if (!data->mtu)
- return data->max_reportid;
-
- if (data->mtu < data->T5_msg_size) {
- WARN(1, "mtu set is lesser than the T5 message size\n");
- /* Return count of 1, as fallback */
- return 1;
- }
-
- return min(count_limit, data->max_reportid);
-}
-
static int mxt_process_messages_until_invalid(struct mxt_data *data)
{
struct device *dev = &data->client->dev;
int count, read;
int tries;
- count = mxt_max_msg_read_count(data);
+ count = mxt_max_msg_read_count(data, data->max_reportid);
tries = (data->max_reportid / count) + 1;
/* Read messages until we force an invalid */
--
2.19.2
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PATCH v1 52/63] Input: Atmel: use T44 object to process T5 messages
2019-08-16 8:37 [PATCH v1 50/63] Input: Atmel: handle ReportID "0x00" while processing T5 messages Jiada Wang
2019-08-16 8:37 ` [PATCH v1 51/63] input: Atmel: limit the max bytes transferred while reading " Jiada Wang
@ 2019-08-16 8:37 ` Jiada Wang
2019-08-16 8:37 ` [PATCH v1 53/63] Input: atmel_mxt_ts: use gpiod_set_value_cansleep for reset pin Jiada Wang
2019-08-16 8:37 ` [PATCH v1 54/63] Input: atmel_mxt_ts: Avoid race condition in freeing of input device Jiada Wang
3 siblings, 0 replies; 5+ messages in thread
From: Jiada Wang @ 2019-08-16 8:37 UTC (permalink / raw)
To: nick, dmitry.torokhov; +Cc: linux-input, linux-kernel, jiada_wang, george_davis
From: Deepak Das <deepak_das@mentor.com>
T44 object returns the count of valid T5 messages in the buffer. According
to atmel, this count should be the main criteria to read the number of T5
messages.
Following is the statement from atmel confirming the same :-
"For the readout of messages we recommend to stop after the last message
is read out from the buffer. One way to identify the amount of new messages
is to read T44. The other way is to monitor the /CHG line which indicates
independent of mode 0 or mode 1 if there are still data in the buffer.
0xFF indicates that there is no message pending anymore, but it is not
recommended to use this as the main criteria to control the
data transfer."
This commit modifies the logic to readout the T5 messages on the basis
of T44 object.
Signed-off-by: Deepak Das <deepak_das@mentor.com>
Signed-off-by: Sanjeev Chugh <sanjeev_chugh@mentor.com>
Signed-off-by: George G. Davis <george_davis@mentor.com>
Signed-off-by: Jiada Wang <jiada_wang@mentor.com>
---
drivers/input/touchscreen/atmel_mxt_ts.c | 55 +++++++++++++++---------
1 file changed, 35 insertions(+), 20 deletions(-)
diff --git a/drivers/input/touchscreen/atmel_mxt_ts.c b/drivers/input/touchscreen/atmel_mxt_ts.c
index 47c1e72152de..5c980e74e6b0 100644
--- a/drivers/input/touchscreen/atmel_mxt_ts.c
+++ b/drivers/input/touchscreen/atmel_mxt_ts.c
@@ -1491,7 +1491,7 @@ static u8 mxt_max_msg_read_count(struct mxt_data *data, u8 max_T5_msg_count)
return min(T5_msg_count_limit, max_T5_msg_count);
}
-static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
+static int mxt_process_messages_t44(struct mxt_data *data)
{
struct device *dev = &data->client->dev;
int ret;
@@ -1504,7 +1504,7 @@ static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
data->T5_msg_size + 1, data->msg_buf);
if (ret) {
dev_err(dev, "Failed to read T44 and T5 (%d)\n", ret);
- return IRQ_NONE;
+ return ret;
}
T5_msg_count = data->msg_buf[0];
@@ -1514,7 +1514,7 @@ static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
* Mode 0. It results in unnecessary I2C operations but it is benign.
*/
if (!T5_msg_count)
- return IRQ_NONE;
+ return processed_valid;
if (T5_msg_count > data->max_reportid) {
dev_warn(dev, "T44 count %d exceeded max report id\n",
@@ -1526,12 +1526,14 @@ static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
ret = mxt_proc_message(data, data->msg_buf + 1);
if (ret < 0) {
dev_warn(dev, "Unexpected invalid message\n");
- return IRQ_NONE;
+ return ret;
}
total_pending = T5_msg_count - 1;
- if (!total_pending)
+ if (!total_pending) {
+ processed_valid = 1;
goto end;
+ }
/* Process remaining messages if necessary */
T5_msg_count = mxt_max_msg_read_count(data, total_pending);
@@ -1555,7 +1557,7 @@ static irqreturn_t mxt_process_messages_t44(struct mxt_data *data)
data->update_input = false;
}
- return IRQ_HANDLED;
+ return processed_valid;
}
static int mxt_process_messages_until_invalid(struct mxt_data *data)
@@ -1585,7 +1587,7 @@ static int mxt_process_messages_until_invalid(struct mxt_data *data)
return -EBUSY;
}
-static irqreturn_t mxt_process_messages(struct mxt_data *data)
+static int mxt_process_messages(struct mxt_data *data)
{
int total_handled, num_handled;
u8 count = data->last_message_count;
@@ -1596,7 +1598,7 @@ static irqreturn_t mxt_process_messages(struct mxt_data *data)
/* include final invalid message */
total_handled = mxt_read_and_process_messages(data, count + 1);
if (total_handled < 0)
- return IRQ_NONE;
+ return total_handled;
/* if there were invalid messages, then we are done */
else if (total_handled <= count)
goto update_count;
@@ -1605,7 +1607,7 @@ static irqreturn_t mxt_process_messages(struct mxt_data *data)
do {
num_handled = mxt_read_and_process_messages(data, 2);
if (num_handled < 0)
- return IRQ_NONE;
+ return num_handled;
total_handled += num_handled;
@@ -1621,12 +1623,13 @@ static irqreturn_t mxt_process_messages(struct mxt_data *data)
data->update_input = false;
}
- return IRQ_HANDLED;
+ return total_handled;
}
static irqreturn_t mxt_interrupt(int irq, void *dev_id)
{
struct mxt_data *data = dev_id;
+ int ret;
complete(&data->chg_completion);
@@ -1634,17 +1637,22 @@ static irqreturn_t mxt_interrupt(int irq, void *dev_id)
if (data->flash && &data->flash->work)
cancel_delayed_work_sync(&data->flash->work);
- return IRQ_RETVAL(mxt_check_bootloader(data));
+ ret = mxt_check_bootloader(data);
+ return IRQ_RETVAL(ret);
}
if (!data->object_table)
return IRQ_HANDLED;
- if (data->T44_address) {
- return mxt_process_messages_t44(data);
- } else {
- return mxt_process_messages(data);
- }
+ if (data->T44_address)
+ ret = mxt_process_messages_t44(data);
+ else
+ ret = mxt_process_messages(data);
+
+ if (ret <= 0)
+ return IRQ_NONE;
+ else
+ return IRQ_HANDLED;
}
static int mxt_t6_command(struct mxt_data *data, u16 cmd_offset,
@@ -1779,8 +1787,11 @@ static int mxt_acquire_irq(struct mxt_data *data)
}
if (data->object_table && data->use_retrigen_workaround) {
- error = mxt_process_messages_until_invalid(data);
- if (error)
+ if (data->T44_address)
+ error = mxt_process_messages_t44(data);
+ else
+ error = mxt_process_messages_until_invalid(data);
+ if (error < 0)
return error;
}
@@ -4116,8 +4127,12 @@ static int mxt_start(struct mxt_data *data)
* Discard any touch messages still in message buffer
* from before chip went to sleep
*/
- ret = mxt_process_messages_until_invalid(data);
- if (ret)
+
+ if (data->T44_address)
+ ret = mxt_process_messages_t44(data);
+ else
+ ret = mxt_process_messages_until_invalid(data);
+ if (ret < 0)
break;
ret = mxt_set_t7_power_cfg(data, MXT_POWER_CFG_RUN);
--
2.19.2
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PATCH v1 53/63] Input: atmel_mxt_ts: use gpiod_set_value_cansleep for reset pin
2019-08-16 8:37 [PATCH v1 50/63] Input: Atmel: handle ReportID "0x00" while processing T5 messages Jiada Wang
2019-08-16 8:37 ` [PATCH v1 51/63] input: Atmel: limit the max bytes transferred while reading " Jiada Wang
2019-08-16 8:37 ` [PATCH v1 52/63] Input: Atmel: use T44 object to process " Jiada Wang
@ 2019-08-16 8:37 ` Jiada Wang
2019-08-16 8:37 ` [PATCH v1 54/63] Input: atmel_mxt_ts: Avoid race condition in freeing of input device Jiada Wang
3 siblings, 0 replies; 5+ messages in thread
From: Jiada Wang @ 2019-08-16 8:37 UTC (permalink / raw)
To: nick, dmitry.torokhov; +Cc: linux-input, linux-kernel, jiada_wang, george_davis
From: Balasubramani Vivekanandan <balasubramani_vivekanandan@mentor.com>
In case of remote display, touch controller will be also remote.
In such cases, the reset pin of the touch controller will be
controlled through bridging ICs like Deserilizer and Serializer.
Therefore accessing the gpio pins require transactions with the
external IC. Using the function gpiod_set_value will print a
warning like below
WARNING: CPU: 0 PID: 576 at drivers/gpio/gpiolib.c:1441 gpiod_set_value+0x34/0x60()
CPU: 0 PID: 576 Comm: modprobe Not tainted 3.14.79-08377-g84ea22f-dirty #4
Backtrace:
[<80011c58>] (dump_backtrace) from [<80011e60>] (show_stack+0x18/0x1c)
[<80011e48>] (show_stack) from [<8052d7ac>] (dump_stack+0x7c/0x9c)
[<8052d730>] (dump_stack) from [<800241bc>] (warn_slowpath_common+0x74/0x9c)
[<80024148>] (warn_slowpath_common) from [<80024288>] (warn_slowpath_null+0x24/0x2c)
[<80024264>] (warn_slowpath_null) from [<8029e070>] (gpiod_set_value+0x34/0x60)
[<8029e03c>] (gpiod_set_value) from [<7f492e98>] (mxt_probe+0x1e0/0x718 [atmel_mxt_ts])
[<7f492cb8>] (mxt_probe [atmel_mxt_ts]) from [<803c4d34>] (i2c_device_probe+0xcc/0xec)
[<803c4c68>] (i2c_device_probe) from [<803252a0>] (driver_probe_device+0xc0/0x200)
Signed-off-by: Balasubramani Vivekanandan <balasubramani_vivekanandan@mentor.com>
Signed-off-by: Vladimir Zapolskiy <vladimir_zapolskiy@mentor.com>
Signed-off-by: Sanjeev Chugh <sanjeev_chugh@mentor.com>
Signed-off-by: George G. Davis <george_davis@mentor.com>
Signed-off-by: Jiada Wang <jiada_wang@mentor.com>
---
drivers/input/touchscreen/atmel_mxt_ts.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/input/touchscreen/atmel_mxt_ts.c b/drivers/input/touchscreen/atmel_mxt_ts.c
index 5c980e74e6b0..79fc6561f6ad 100644
--- a/drivers/input/touchscreen/atmel_mxt_ts.c
+++ b/drivers/input/touchscreen/atmel_mxt_ts.c
@@ -2492,7 +2492,7 @@ static void mxt_regulator_enable(struct mxt_data *data)
if (!data->reg_vdd || !data->reg_avdd)
return;
- gpiod_set_value(data->reset_gpio, 0);
+ gpiod_set_value_cansleep(data->reset_gpio, 0);
error = regulator_enable(data->reg_vdd);
if (error)
@@ -2510,7 +2510,7 @@ static void mxt_regulator_enable(struct mxt_data *data)
* voltage
*/
msleep(MXT_REGULATOR_DELAY);
- gpiod_set_value(data->reset_gpio, 1);
+ gpiod_set_value_cansleep(data->reset_gpio, 1);
msleep(MXT_CHG_DELAY);
retry_wait:
@@ -4501,7 +4501,7 @@ static int mxt_probe(struct i2c_client *client, const struct i2c_device_id *id)
disable_irq(data->irq);
} else if (data->reset_gpio) {
msleep(MXT_RESET_GPIO_TIME);
- gpiod_set_value(data->reset_gpio, 1);
+ gpiod_set_value_cansleep(data->reset_gpio, 1);
msleep(MXT_RESET_INVALID_CHG);
} else {
dev_dbg(&client->dev,
--
2.19.2
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PATCH v1 54/63] Input: atmel_mxt_ts: Avoid race condition in freeing of input device
2019-08-16 8:37 [PATCH v1 50/63] Input: Atmel: handle ReportID "0x00" while processing T5 messages Jiada Wang
` (2 preceding siblings ...)
2019-08-16 8:37 ` [PATCH v1 53/63] Input: atmel_mxt_ts: use gpiod_set_value_cansleep for reset pin Jiada Wang
@ 2019-08-16 8:37 ` Jiada Wang
3 siblings, 0 replies; 5+ messages in thread
From: Jiada Wang @ 2019-08-16 8:37 UTC (permalink / raw)
To: nick, dmitry.torokhov; +Cc: linux-input, linux-kernel, jiada_wang, george_davis
From: Bhuvanesh Surachari <bhuvanesh_surachari@mentor.com>
>From the static analysis of the code it seems that there could be a
race condition leading to crash while freeing input device in
mxt_free_input_device(). The backtrace of crash is as shown below:
Unable to handle kernel NULL pointer dereference at virtual address 0000003c
Internal error: Oops: 17 [#1] PREEMPT SMP ARM
CPU: 1 PID: 229 Comm: load_firmware.s Not tainted 3.14.79-00978-g58395f0ebac4 #1
PC is at kernfs_find_ns+0x14/0xf0
LR is at kernfs_find_and_get_ns+0x34/0x50
Backtrace:
[<801745c0>] (kernfs_find_ns) from [<801746e4>] (kernfs_find_and_get_ns+0x34/0x50)
[<801746b0>] (kernfs_find_and_get_ns) from [<801730c8>] (sysfs_unmerge_group+0x20/0x60)
[<801730a8>] (sysfs_unmerge_group) from [<80328490>] (pm_qos_sysfs_remove_latency+0x18/0x20)
[<80328478>] (pm_qos_sysfs_remove_latency) from [<80329844>] (dev_pm_qos_constraints_destroy+0x20/0x128)
[<80329824>] (dev_pm_qos_constraints_destroy) from [<80328510>] (dpm_sysfs_remove+0x18/0x44)
[<803284f8>] (dpm_sysfs_remove) from [<80320514>] (device_del+0x3c/0x178)
[<803204d8>] (device_del) from [<803b5ae8>] (__input_unregister_device+0x120/0x134)
[<803b59c8>] (__input_unregister_device) from [<803b5b68>] (input_unregister_device+0x54/0x74)
[<803b5b14>] (input_unregister_device) from [<7f1288bc>] (mxt_debug_enable_store+0x1a8/0x2c4 [atmel_mxt_ts])
[<7f12888c>] (mxt_debug_enable_store [atmel_mxt_ts]) from [<7f12b534>] (mxt_update_cfg_store+0xc4/0x154 [atmel_mxt_ts])
[<7f12b470>] (mxt_update_cfg_store [atmel_mxt_ts]) from [<8031f4e8>] (dev_attr_store+0x20/0x2c)
[<8031f4c8>] (dev_attr_store) from [<80172558>] (sysfs_kf_write+0x40/0x4c)
[<80172518>] (sysfs_kf_write) from [<801756fc>] (kernfs_fop_write+0xf8/0x140)
[<80175604>] (kernfs_fop_write) from [<801136e0>] (vfs_write+0xd8/0x16c)
[<80113608>] (vfs_write) from [<80113c34>] (SyS_write+0x50/0x90)
[<80113be4>] (SyS_write) from [<8000e0a0>] (ret_fast_syscall+0x0/0x38)
Note: mxt_free_input_device() is misrepresented as mxt_debug_enable_store()
in the crash backtrace. From the disassembly of the atmel_mxt_ts.ko, address
pointed by (mxt_debug_enable_store+0x1a8/0x2c4 [atmel_mxt_ts]) refers to
mxt_free_input_device()
There is speculation that this race condition may occur while
configuration (firmware) is loading and driver is being unloaded parallely.
The solution is to take a local pointer to data->input_dev with the lock
held and then to set data->input_dev to NULL to prevent any parallel
thread from executing input_unregister_device() for a second time overall
so avoiding the crash. The lock is released and the local pointer is
safely used by input_unregister_device() so this function only runs a
single time overall.
Signed-off-by: Bhuvanesh Surachari <bhuvanesh_surachari@mentor.com>
Signed-off-by: Sanjeev Chugh <sanjeev_chugh@mentor.com>
Signed-off-by: George G. Davis <george_davis@mentor.com>
Signed-off-by: Jiada Wang <jiada_wang@mentor.com>
---
drivers/input/touchscreen/atmel_mxt_ts.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/input/touchscreen/atmel_mxt_ts.c b/drivers/input/touchscreen/atmel_mxt_ts.c
index 79fc6561f6ad..35d92751e49f 100644
--- a/drivers/input/touchscreen/atmel_mxt_ts.c
+++ b/drivers/input/touchscreen/atmel_mxt_ts.c
@@ -2221,8 +2221,10 @@ static int mxt_update_cfg(struct mxt_data *data, const struct firmware *fw)
static void mxt_free_input_device(struct mxt_data *data)
{
if (data->input_dev) {
- input_unregister_device(data->input_dev);
+ struct input_dev *dev = data->input_dev;
+
data->input_dev = NULL;
+ input_unregister_device(dev);
}
}
--
2.19.2
^ permalink raw reply related [flat|nested] 5+ messages in thread