* [PATCH i2c-next v3 1/3] dt-bindings: i2c: aspeed: Add 'timeout' property as an optional property
2018-09-26 21:58 [PATCH i2c-next v3 0/3] i2c: aspeed: Add bus idle waiting logic for multi-master use cases Jae Hyun Yoo
@ 2018-09-26 21:58 ` Jae Hyun Yoo
2018-09-27 20:56 ` Rob Herring
2018-09-26 21:58 ` [PATCH i2c-next v3 2/3] i2c: aspeed: Add 'aspeed,timeout' DT property reading code Jae Hyun Yoo
2018-09-26 21:58 ` [PATCH i2c-next v3 3/3] i2c: aspeed: Add bus idle waiting logic for multi-master use cases Jae Hyun Yoo
2 siblings, 1 reply; 8+ messages in thread
From: Jae Hyun Yoo @ 2018-09-26 21:58 UTC (permalink / raw)
To: Brendan Higgins, Wolfram Sang, Benjamin Herrenschmidt,
Joel Stanley, Rob Herring, Mark Rutland, Andrew Jeffery,
linux-i2c, openbmc, devicetree, linux-arm-kernel, linux-aspeed,
linux-kernel
Cc: Jarkko Nikula, James Feist, Vernon Mauery, Jae Hyun Yoo
This commit adds 'aspeed,timeout' property as an optional property
which can be used for setting 'timeout' value of
'struct i2c_adapter'. With this patch, the timeout value can be
set through an I2C_TIMEOUT ioctl on cdev, or through this optional
DT property.
Signed-off-by: Jae Hyun Yoo <jae.hyun.yoo@linux.intel.com>
---
Documentation/devicetree/bindings/i2c/i2c-aspeed.txt | 3 +++
1 file changed, 3 insertions(+)
diff --git a/Documentation/devicetree/bindings/i2c/i2c-aspeed.txt b/Documentation/devicetree/bindings/i2c/i2c-aspeed.txt
index 8fbd8633a387..d6965b360fbc 100644
--- a/Documentation/devicetree/bindings/i2c/i2c-aspeed.txt
+++ b/Documentation/devicetree/bindings/i2c/i2c-aspeed.txt
@@ -17,6 +17,9 @@ Optional Properties:
specified
- multi-master : states that there is another master active on this bus.
+- aspeed,timeout : I2C bus timeout in microseconds defaults to 5 seconds when
+ not specified.
+
Example:
i2c {
--
2.19.0
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH i2c-next v3 1/3] dt-bindings: i2c: aspeed: Add 'timeout' property as an optional property
2018-09-26 21:58 ` [PATCH i2c-next v3 1/3] dt-bindings: i2c: aspeed: Add 'timeout' property as an optional property Jae Hyun Yoo
@ 2018-09-27 20:56 ` Rob Herring
2018-09-27 21:35 ` Jae Hyun Yoo
0 siblings, 1 reply; 8+ messages in thread
From: Rob Herring @ 2018-09-27 20:56 UTC (permalink / raw)
To: Jae Hyun Yoo
Cc: Brendan Higgins, Wolfram Sang, Benjamin Herrenschmidt,
Joel Stanley, Mark Rutland, Andrew Jeffery, linux-i2c, openbmc,
devicetree, linux-arm-kernel, linux-aspeed, linux-kernel,
Jarkko Nikula, James Feist, Vernon Mauery
On Wed, Sep 26, 2018 at 02:58:40PM -0700, Jae Hyun Yoo wrote:
> This commit adds 'aspeed,timeout' property as an optional property
> which can be used for setting 'timeout' value of
> 'struct i2c_adapter'. With this patch, the timeout value can be
> set through an I2C_TIMEOUT ioctl on cdev, or through this optional
> DT property.
Isn't controlling this from userspace or relying on a default
sufficient? I can't see this needing to be highly tuned for each
platform.
However, if we do have a property, it should be common.
>
> Signed-off-by: Jae Hyun Yoo <jae.hyun.yoo@linux.intel.com>
> ---
> Documentation/devicetree/bindings/i2c/i2c-aspeed.txt | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/Documentation/devicetree/bindings/i2c/i2c-aspeed.txt b/Documentation/devicetree/bindings/i2c/i2c-aspeed.txt
> index 8fbd8633a387..d6965b360fbc 100644
> --- a/Documentation/devicetree/bindings/i2c/i2c-aspeed.txt
> +++ b/Documentation/devicetree/bindings/i2c/i2c-aspeed.txt
> @@ -17,6 +17,9 @@ Optional Properties:
> specified
> - multi-master : states that there is another master active on this bus.
>
> +- aspeed,timeout : I2C bus timeout in microseconds defaults to 5 seconds when
> + not specified.
> +
> Example:
>
> i2c {
> --
> 2.19.0
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH i2c-next v3 1/3] dt-bindings: i2c: aspeed: Add 'timeout' property as an optional property
2018-09-27 20:56 ` Rob Herring
@ 2018-09-27 21:35 ` Jae Hyun Yoo
0 siblings, 0 replies; 8+ messages in thread
From: Jae Hyun Yoo @ 2018-09-27 21:35 UTC (permalink / raw)
To: Rob Herring
Cc: Brendan Higgins, Wolfram Sang, Benjamin Herrenschmidt,
Joel Stanley, Mark Rutland, Andrew Jeffery, linux-i2c, openbmc,
devicetree, linux-arm-kernel, linux-aspeed, linux-kernel,
Jarkko Nikula, James Feist, Vernon Mauery
Hi Rob,
On 9/27/2018 1:56 PM, Rob Herring wrote:
> On Wed, Sep 26, 2018 at 02:58:40PM -0700, Jae Hyun Yoo wrote:
>> This commit adds 'aspeed,timeout' property as an optional property
>> which can be used for setting 'timeout' value of
>> 'struct i2c_adapter'. With this patch, the timeout value can be
>> set through an I2C_TIMEOUT ioctl on cdev, or through this optional
>> DT property.
>
> Isn't controlling this from userspace or relying on a default
> sufficient? I can't see this needing to be highly tuned for each
> platform.
>
It can be controlled using an ioctl command on an I2C cdev from
userspace if CONFIG_I2C_CHARDEV is enabled. A couple of I2C drivers use
their own specific default value for it but in general the common
default value (1 second) which is set by i2c core is sufficient.
But it still needs to be tuned for specific cases based on attached
devices' characteristic, on packet length, on bus speed and on etc.
Specifically in Aspeed I2C driver for BMC, it should be tuned
to support multi-master use cases properly, and it needs a device tree
property to apply this timeout value from the probing time of the
module.
> However, if we do have a property, it should be common.
>
Okay, I'll change it to 'timeout'.
Thanks a lot,
Jae
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH i2c-next v3 2/3] i2c: aspeed: Add 'aspeed,timeout' DT property reading code
2018-09-26 21:58 [PATCH i2c-next v3 0/3] i2c: aspeed: Add bus idle waiting logic for multi-master use cases Jae Hyun Yoo
2018-09-26 21:58 ` [PATCH i2c-next v3 1/3] dt-bindings: i2c: aspeed: Add 'timeout' property as an optional property Jae Hyun Yoo
@ 2018-09-26 21:58 ` Jae Hyun Yoo
2018-09-27 3:11 ` Joel Stanley
2018-09-26 21:58 ` [PATCH i2c-next v3 3/3] i2c: aspeed: Add bus idle waiting logic for multi-master use cases Jae Hyun Yoo
2 siblings, 1 reply; 8+ messages in thread
From: Jae Hyun Yoo @ 2018-09-26 21:58 UTC (permalink / raw)
To: Brendan Higgins, Wolfram Sang, Benjamin Herrenschmidt,
Joel Stanley, Rob Herring, Mark Rutland, Andrew Jeffery,
linux-i2c, openbmc, devicetree, linux-arm-kernel, linux-aspeed,
linux-kernel
Cc: Jarkko Nikula, James Feist, Vernon Mauery, Jae Hyun Yoo
This commit adds reading code of the 'aspeed,timeout' DT property
to set 'timeout' value in adapter configuration. This value still
case be configured through an I2C_TIMEOUT ioctl on cdev too.
Signed-off-by: Jae Hyun Yoo <jae.hyun.yoo@linux.intel.com>
---
drivers/i2c/busses/i2c-aspeed.c | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/drivers/i2c/busses/i2c-aspeed.c b/drivers/i2c/busses/i2c-aspeed.c
index 8dc9161ced38..0d934ce0c028 100644
--- a/drivers/i2c/busses/i2c-aspeed.c
+++ b/drivers/i2c/busses/i2c-aspeed.c
@@ -115,6 +115,9 @@
/* 0x18 : I2CD Slave Device Address Register */
#define ASPEED_I2CD_DEV_ADDR_MASK GENMASK(6, 0)
+/* Timeout */
+#define ASPEED_I2C_BUS_TIMEOUT_US_DEFAULT (5 * 1000 * 1000)
+
enum aspeed_i2c_master_state {
ASPEED_I2C_MASTER_INACTIVE,
ASPEED_I2C_MASTER_START,
@@ -885,6 +888,7 @@ static int aspeed_i2c_probe_bus(struct platform_device *pdev)
struct clk *parent_clk;
struct resource *res;
int irq, ret;
+ u32 timeout_us;
bus = devm_kzalloc(&pdev->dev, sizeof(*bus), GFP_KERNEL);
if (!bus)
@@ -918,6 +922,11 @@ static int aspeed_i2c_probe_bus(struct platform_device *pdev)
bus->bus_frequency = 100000;
}
+ ret = of_property_read_u32(pdev->dev.of_node, "aspeed,timeout",
+ &timeout_us);
+ if (ret)
+ timeout_us = ASPEED_I2C_BUS_TIMEOUT_US_DEFAULT;
+
match = of_match_node(aspeed_i2c_bus_of_table, pdev->dev.of_node);
if (!match)
bus->get_clk_reg_val = aspeed_i2c_24xx_get_clk_reg_val;
@@ -930,7 +939,7 @@ static int aspeed_i2c_probe_bus(struct platform_device *pdev)
init_completion(&bus->cmd_complete);
bus->adap.owner = THIS_MODULE;
bus->adap.retries = 0;
- bus->adap.timeout = 5 * HZ;
+ bus->adap.timeout = usecs_to_jiffies(timeout_us);
bus->adap.algo = &aspeed_i2c_algo;
bus->adap.dev.parent = &pdev->dev;
bus->adap.dev.of_node = pdev->dev.of_node;
--
2.19.0
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH i2c-next v3 2/3] i2c: aspeed: Add 'aspeed,timeout' DT property reading code
2018-09-26 21:58 ` [PATCH i2c-next v3 2/3] i2c: aspeed: Add 'aspeed,timeout' DT property reading code Jae Hyun Yoo
@ 2018-09-27 3:11 ` Joel Stanley
2018-09-27 17:41 ` Jae Hyun Yoo
0 siblings, 1 reply; 8+ messages in thread
From: Joel Stanley @ 2018-09-27 3:11 UTC (permalink / raw)
To: Jae Hyun Yoo
Cc: Brendan Higgins, Wolfram Sang, Benjamin Herrenschmidt,
Rob Herring, Mark Rutland, Andrew Jeffery, linux-i2c,
OpenBMC Maillist, devicetree, Linux ARM, linux-aspeed,
Linux Kernel Mailing List, jarkko.nikula, James Feist,
Vernon Mauery
On Thu, 27 Sep 2018 at 01:58, Jae Hyun Yoo <jae.hyun.yoo@linux.intel.com> wrote:
>
> This commit adds reading code of the 'aspeed,timeout' DT property
> to set 'timeout' value in adapter configuration. This value still
> case be configured through an I2C_TIMEOUT ioctl on cdev too.
>
> Signed-off-by: Jae Hyun Yoo <jae.hyun.yoo@linux.intel.com>
> ---
> drivers/i2c/busses/i2c-aspeed.c | 11 ++++++++++-
> 1 file changed, 10 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/i2c/busses/i2c-aspeed.c b/drivers/i2c/busses/i2c-aspeed.c
> index 8dc9161ced38..0d934ce0c028 100644
> --- a/drivers/i2c/busses/i2c-aspeed.c
> +++ b/drivers/i2c/busses/i2c-aspeed.c
> @@ -115,6 +115,9 @@
> /* 0x18 : I2CD Slave Device Address Register */
> #define ASPEED_I2CD_DEV_ADDR_MASK GENMASK(6, 0)
>
> +/* Timeout */
> +#define ASPEED_I2C_BUS_TIMEOUT_US_DEFAULT (5 * 1000 * 1000)
The 5 seconds time out is way too long. On a system that doesn't have
functional i2c, this holds up boot for a long time as most i2c client
drivers try to initialise their device and fail. I realise you're not
changing the value, but we should pick a better default. 1 second?
Half a second?
> +
> enum aspeed_i2c_master_state {
> ASPEED_I2C_MASTER_INACTIVE,
> ASPEED_I2C_MASTER_START,
> @@ -885,6 +888,7 @@ static int aspeed_i2c_probe_bus(struct platform_device *pdev)
> struct clk *parent_clk;
> struct resource *res;
> int irq, ret;
> + u32 timeout_us;
>
> bus = devm_kzalloc(&pdev->dev, sizeof(*bus), GFP_KERNEL);
> if (!bus)
> @@ -918,6 +922,11 @@ static int aspeed_i2c_probe_bus(struct platform_device *pdev)
> bus->bus_frequency = 100000;
> }
>
> + ret = of_property_read_u32(pdev->dev.of_node, "aspeed,timeout",
> + &timeout_us);
Can we make this binding generic? It's not specific to aspeed's
hardware. Getting the value could even part of the i2c core.
I read the previous thread with Wolfram. I think this would still fit
with what Wolfram suggested, but please forgive my jetlagged brain if
I've missed something.
Cheers,
Joel
> + if (ret)
> + timeout_us = ASPEED_I2C_BUS_TIMEOUT_US_DEFAULT;
> +
> match = of_match_node(aspeed_i2c_bus_of_table, pdev->dev.of_node);
> if (!match)
> bus->get_clk_reg_val = aspeed_i2c_24xx_get_clk_reg_val;
> @@ -930,7 +939,7 @@ static int aspeed_i2c_probe_bus(struct platform_device *pdev)
> init_completion(&bus->cmd_complete);
> bus->adap.owner = THIS_MODULE;
> bus->adap.retries = 0;
> - bus->adap.timeout = 5 * HZ;
> + bus->adap.timeout = usecs_to_jiffies(timeout_us);
> bus->adap.algo = &aspeed_i2c_algo;
> bus->adap.dev.parent = &pdev->dev;
> bus->adap.dev.of_node = pdev->dev.of_node;
> --
> 2.19.0
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH i2c-next v3 2/3] i2c: aspeed: Add 'aspeed,timeout' DT property reading code
2018-09-27 3:11 ` Joel Stanley
@ 2018-09-27 17:41 ` Jae Hyun Yoo
0 siblings, 0 replies; 8+ messages in thread
From: Jae Hyun Yoo @ 2018-09-27 17:41 UTC (permalink / raw)
To: Joel Stanley
Cc: Brendan Higgins, Wolfram Sang, Benjamin Herrenschmidt,
Rob Herring, Mark Rutland, Andrew Jeffery, linux-i2c,
OpenBMC Maillist, devicetree, Linux ARM, linux-aspeed,
Linux Kernel Mailing List, jarkko.nikula, James Feist,
Vernon Mauery
Hi Joel,
On 9/26/2018 8:11 PM, Joel Stanley wrote:
> On Thu, 27 Sep 2018 at 01:58, Jae Hyun Yoo <jae.hyun.yoo@linux.intel.com> wrote:
>>
>> +/* Timeout */
>> +#define ASPEED_I2C_BUS_TIMEOUT_US_DEFAULT (5 * 1000 * 1000)
>
> The 5 seconds time out is way too long. On a system that doesn't have
> functional i2c, this holds up boot for a long time as most i2c client
> drivers try to initialise their device and fail. I realise you're not
> changing the value, but we should pick a better default. 1 second?
> Half a second?
>
I agree with you. We could probably use 1 second as default which can
cover the most of general cases. If so, we don't need to make the
default setting in this driver because i2c-core-base will default
adap->timeout to 1 second if the value is 0 when a driver registers an
adapter. Will fix this code.
>>
>> + ret = of_property_read_u32(pdev->dev.of_node, "aspeed,timeout",
>> + &timeout_us);
>
> Can we make this binding generic? It's not specific to aspeed's
> hardware. Getting the value could even part of the i2c core.
>
> I read the previous thread with Wolfram. I think this would still fit
> with what Wolfram suggested, but please forgive my jetlagged brain if
> I've missed something.
>
It followed the way of the existing i2c-mpc driver uses 'fsl,timeout'
for the same purpose. Though, I also want to make it as a generic as you
suggested like 'timeout' in milliseconds unit, not in microseconds unit.
If making it a generic property is acceptable, I'll fix it too.
Wolfram, can you please share your thought on it?
Thanks,
Jae
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH i2c-next v3 3/3] i2c: aspeed: Add bus idle waiting logic for multi-master use cases
2018-09-26 21:58 [PATCH i2c-next v3 0/3] i2c: aspeed: Add bus idle waiting logic for multi-master use cases Jae Hyun Yoo
2018-09-26 21:58 ` [PATCH i2c-next v3 1/3] dt-bindings: i2c: aspeed: Add 'timeout' property as an optional property Jae Hyun Yoo
2018-09-26 21:58 ` [PATCH i2c-next v3 2/3] i2c: aspeed: Add 'aspeed,timeout' DT property reading code Jae Hyun Yoo
@ 2018-09-26 21:58 ` Jae Hyun Yoo
2 siblings, 0 replies; 8+ messages in thread
From: Jae Hyun Yoo @ 2018-09-26 21:58 UTC (permalink / raw)
To: Brendan Higgins, Wolfram Sang, Benjamin Herrenschmidt,
Joel Stanley, Rob Herring, Mark Rutland, Andrew Jeffery,
linux-i2c, openbmc, devicetree, linux-arm-kernel, linux-aspeed,
linux-kernel
Cc: Jarkko Nikula, James Feist, Vernon Mauery, Jae Hyun Yoo
In multi-master environment, this driver's master cannot know
exactly when peer master sends data to this driver's slave so a
case can be happened that this master tries to send data through
the master_xfer function but slave data from peer master is still
being processed by this driver.
To prevent state corruption in the case, this patch adds checking
code if any slave operation is ongoing and it waits up to the
timeout duration before starting a master_xfer operation.
Signed-off-by: Jae Hyun Yoo <jae.hyun.yoo@linux.intel.com>
---
drivers/i2c/busses/i2c-aspeed.c | 62 ++++++++++++++++++++++++---------
1 file changed, 46 insertions(+), 16 deletions(-)
diff --git a/drivers/i2c/busses/i2c-aspeed.c b/drivers/i2c/busses/i2c-aspeed.c
index 0d934ce0c028..957649009b53 100644
--- a/drivers/i2c/busses/i2c-aspeed.c
+++ b/drivers/i2c/busses/i2c-aspeed.c
@@ -12,6 +12,7 @@
#include <linux/clk.h>
#include <linux/completion.h>
+#include <linux/delay.h>
#include <linux/err.h>
#include <linux/errno.h>
#include <linux/i2c.h>
@@ -99,6 +100,7 @@
ASPEED_I2CD_INTR_TX_ACK)
/* 0x14 : I2CD Command/Status Register */
+#define ASPEED_I2CD_XFER_MODE_STS_MASK GENMASK(22, 19)
#define ASPEED_I2CD_SCL_LINE_STS BIT(18)
#define ASPEED_I2CD_SDA_LINE_STS BIT(17)
#define ASPEED_I2CD_BUS_BUSY_STS BIT(16)
@@ -115,8 +117,9 @@
/* 0x18 : I2CD Slave Device Address Register */
#define ASPEED_I2CD_DEV_ADDR_MASK GENMASK(6, 0)
-/* Timeout */
+/* Timeout & Busy checking */
#define ASPEED_I2C_BUS_TIMEOUT_US_DEFAULT (5 * 1000 * 1000)
+#define ASPEED_I2C_BUS_BUSY_CHECK_INTERVAL_US (10 * 1000)
enum aspeed_i2c_master_state {
ASPEED_I2C_MASTER_INACTIVE,
@@ -159,6 +162,8 @@ struct aspeed_i2c_bus {
int cmd_err;
/* Protected only by i2c_lock_bus */
int master_xfer_result;
+ /* Multi-master */
+ bool multi_master;
#if IS_ENABLED(CONFIG_I2C_SLAVE)
struct i2c_client *slave;
enum aspeed_i2c_slave_state slave_state;
@@ -599,27 +604,50 @@ static irqreturn_t aspeed_i2c_bus_irq(int irq, void *dev_id)
return irq_remaining ? IRQ_NONE : IRQ_HANDLED;
}
+static int aspeed_i2c_check_bus_busy(struct aspeed_i2c_bus *bus)
+{
+ u32 status_check_mask = ASPEED_I2CD_BUS_BUSY_STS;
+ ktime_t timeout;
+
+ if (bus->multi_master) {
+ might_sleep();
+ timeout = ktime_add_us(ktime_get(),
+ jiffies_to_usecs(bus->adap.timeout));
+ /*
+ * ASPEED_I2CD_XFER_MODE_STS_MASK is marked as
+ * 'for debugging purpose only' in datasheet but ASPEED
+ * confirmed that this reflects real information and good to be
+ * used in practical code. It will be used only in multi-master
+ * use cases.
+ */
+ status_check_mask |= ASPEED_I2CD_XFER_MODE_STS_MASK;
+ }
+
+ for (;;) {
+ if (!(readl(bus->base + ASPEED_I2C_CMD_REG) &
+ status_check_mask))
+ return 0;
+ if (!bus->multi_master)
+ break;
+ if (ktime_compare(ktime_get(), timeout) > 0)
+ break;
+ usleep_range((ASPEED_I2C_BUS_BUSY_CHECK_INTERVAL_US >> 2) + 1,
+ ASPEED_I2C_BUS_BUSY_CHECK_INTERVAL_US);
+ }
+
+ return aspeed_i2c_recover_bus(bus);
+}
+
static int aspeed_i2c_master_xfer(struct i2c_adapter *adap,
struct i2c_msg *msgs, int num)
{
struct aspeed_i2c_bus *bus = i2c_get_adapdata(adap);
unsigned long time_left, flags;
- int ret = 0;
-
- spin_lock_irqsave(&bus->lock, flags);
- bus->cmd_err = 0;
- /* If bus is busy, attempt recovery. We assume a single master
- * environment.
- */
- if (readl(bus->base + ASPEED_I2C_CMD_REG) & ASPEED_I2CD_BUS_BUSY_STS) {
- spin_unlock_irqrestore(&bus->lock, flags);
- ret = aspeed_i2c_recover_bus(bus);
- if (ret)
- return ret;
- spin_lock_irqsave(&bus->lock, flags);
- }
+ if (aspeed_i2c_check_bus_busy(bus))
+ return -EAGAIN;
+ spin_lock_irqsave(&bus->lock, flags);
bus->cmd_err = 0;
bus->msgs = msgs;
bus->msgs_index = 0;
@@ -830,7 +858,9 @@ static int aspeed_i2c_init(struct aspeed_i2c_bus *bus,
if (ret < 0)
return ret;
- if (!of_property_read_bool(pdev->dev.of_node, "multi-master"))
+ if (of_property_read_bool(pdev->dev.of_node, "multi-master"))
+ bus->multi_master = true;
+ else
fun_ctrl_reg |= ASPEED_I2CD_MULTI_MASTER_DIS;
/* Enable Master Mode */
--
2.19.0
^ permalink raw reply related [flat|nested] 8+ messages in thread