mbox series

[i2c-next,v9,0/5] i2c: aspeed: Add bus idle waiting logic for multi-master use cases

Message ID 20181030210917.32711-1-jae.hyun.yoo@linux.intel.com (mailing list archive)
Headers show
Series i2c: aspeed: Add bus idle waiting logic for multi-master use cases | expand

Message

Jae Hyun Yoo Oct. 30, 2018, 9:09 p.m. UTC
In multi-master environment, this driver's master cannot know
exactly when a 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 bus timeout duration before starting
a master_xfer operation.

To support this change, it introduces changes on i2c-core-base to
make that able to read the bus timeout and master transfer retries
count values from device tree properties.

Please review this patch set.

Thanks,

-Jae

Changes since v8:
- Fixed a build break when CONFIG_I2C_SLAVE is not set.

Changes since v7:
- Simplified the bus idle waiting logic using jiffy based timer APIs.

Changes since v6:
- Changed the 'timeout-ms' property name to 'bus-timeout-ms'.

Changes since v5:
- Changed using of property reading API to device_property_read_u32.

Changes since v4:
- Moved the property reading code into i2c-base-core and changed the
  property name to 'timeout-ms'. Also, added '#retries' property reading
  code.
- Changed bus busy checking logic to make that check slave_state instead
  of 'Transfer Mode State Machine' reg value.

Changes since v3:
- Changed the property name to 'timeout' and made it use the
  default setting in i2c-core when not specified.

Changes since v2:
- Changed the property name to 'aspeed,timeout' and made it to
  update the adapter's timeout configuration.

Changes since v1:
- Changed define names of timeout related.

Jae Hyun Yoo (5):
  dt-bindings: i2c: Add 'bus-timeout-ms' and '#retries' properties as
    common optional
  i2c: core: Add support reading of 'bus-timeout-ms' and '#retries'
    properties
  dt-bindings: i2c: aspeed: Add 'bus-timeout-ms' property as an optional
    property
  i2c: aspeed: Remove hard-coded bus timeout value setting
  i2c: aspeed: Add bus idle waiting logic for multi-master use cases

 .../devicetree/bindings/i2c/i2c-aspeed.txt    |  3 +
 Documentation/devicetree/bindings/i2c/i2c.txt |  6 ++
 drivers/i2c/busses/i2c-aspeed.c               | 56 +++++++++++++------
 drivers/i2c/i2c-core-base.c                   | 12 +++-
 4 files changed, 59 insertions(+), 18 deletions(-)

Comments

Wolfram Sang Jan. 15, 2019, 10:47 p.m. UTC | #1
On Tue, Oct 30, 2018 at 02:09:11PM -0700, Jae Hyun Yoo wrote:
> In multi-master environment, this driver's master cannot know
> exactly when a 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 bus timeout duration before starting
> a master_xfer operation.
> 
> To support this change, it introduces changes on i2c-core-base to
> make that able to read the bus timeout and master transfer retries
> count values from device tree properties.
> 
> Please review this patch set.

Marking this as "Changes requested". I think if the mutex approach
works, these patches become obsolete then.
Jae Hyun Yoo Jan. 15, 2019, 11:32 p.m. UTC | #2
On 1/15/2019 2:47 PM, Wolfram Sang wrote:
> On Tue, Oct 30, 2018 at 02:09:11PM -0700, Jae Hyun Yoo wrote:
>> In multi-master environment, this driver's master cannot know
>> exactly when a 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 bus timeout duration before starting
>> a master_xfer operation.
>>
>> To support this change, it introduces changes on i2c-core-base to
>> make that able to read the bus timeout and master transfer retries
>> count values from device tree properties.
>>
>> Please review this patch set.
> 
> Marking this as "Changes requested". I think if the mutex approach
> works, these patches become obsolete then.
> 

Hi Wolfram,

Thanks for your touching base. I made and tested a new patch in my
local. Actually, I couldn't use a mutex because it should provide
locking between driver context and interrupt context. Instead, after
taking lots of experiments, I checked that idle waiting can be supported
by this H/W so I didn't need to use any locking method in the driver
code for multi-master handling. I'll submit a new patch tomorrow using a
different thread. Please drop this patch set.

Thanks,
Jae