Re: [Qemu-devel] [PATCH 1/3] aspeed/i2c: interrupts should be cleared by software only

[Peter Maydell]

On 25 September 2018 at 10:12, Cédric Le Goater <address@hidden> wrote:
> On 09/25/2018 11:02 AM, Peter Maydell wrote:
>> On 14 September 2018 at 07:35, Cédric Le Goater <address@hidden> wrote:
>>> and the bus interrupt should be lowered when all interrupts have been
>>> cleared. Also, the model does not implement correctly the RX_DONE bit
>>> behavior which should be cleared to allow more data to be received.
>>> Yet to be fixed.
>>>
>>> Signed-off-by: Cédric Le Goater <address@hidden>
>>> ---
>>>  hw/i2c/aspeed_i2c.c | 21 +++++++++++++++++----
>>>  1 file changed, 17 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/hw/i2c/aspeed_i2c.c b/hw/i2c/aspeed_i2c.c
>>> index c762c7366ad9..de6b08378675 100644
>>> --- a/hw/i2c/aspeed_i2c.c
>>> +++ b/hw/i2c/aspeed_i2c.c
>>> @@ -52,6 +52,13 @@
>>>  #define I2CD_AC_TIMING_REG2     0x08       /* Clock and AC Timing Control 
>>> #1 */
>>>  #define I2CD_INTR_CTRL_REG      0x0c       /* I2CD Interrupt Control */
>>>  #define I2CD_INTR_STS_REG       0x10       /* I2CD Interrupt Status */
>>> +
>>> +#define   I2CD_INTR_SLAVE_ADDR_MATCH       (0x1 << 31) /* 0: addr1 1: 
>>> addr2 */
>>> +#define   I2CD_INTR_SLAVE_ADDR_RX_PENDING  (0x1 << 30)
>>> +/* bits[19-16] Reserved */
>>> +
>>> +/* All bits below are cleared by writing 1 */
>>> +#define   I2CD_INTR_SLAVE_INACTIVE_TIMEOUT (0x1 << 15)
>>>  #define   I2CD_INTR_SDA_DL_TIMEOUT         (0x1 << 14)
>>>  #define   I2CD_INTR_BUS_RECOVER_DONE       (0x1 << 13)
>>>  #define   I2CD_INTR_SMBUS_ALERT            (0x1 << 12) /* Bus [0-3] only */
>>> @@ -59,11 +66,16 @@
>>>  #define   I2CD_INTR_SMBUS_DEV_ALERT_ADDR   (0x1 << 10) /* Removed */
>>>  #define   I2CD_INTR_SMBUS_DEF_ADDR         (0x1 << 9)  /* Removed */
>>>  #define   I2CD_INTR_GCALL_ADDR             (0x1 << 8)  /* Removed */
>>> -#define   I2CD_INTR_SLAVE_MATCH            (0x1 << 7)  /* use RX_DONE */
>>> +#define   I2CD_INTR_SLAVE_ADDR_RX_MATCH    (0x1 << 7)  /* use RX_DONE */
>>>  #define   I2CD_INTR_SCL_TIMEOUT            (0x1 << 6)
>>>  #define   I2CD_INTR_ABNORMAL               (0x1 << 5)
>>>  #define   I2CD_INTR_NORMAL_STOP            (0x1 << 4)
>>>  #define   I2CD_INTR_ARBIT_LOSS             (0x1 << 3)
>>> +
>>> +/*
>>> + * TODO: handle correctly I2CD_INTR_RX_DONE which needs to be cleared
>>> + * to allow next data to be received.
>>> + */
>>>  #define   I2CD_INTR_RX_DONE                (0x1 << 2)
>>
>> Is this TODO something different from the behaviour involving
>> this RX_DONE bit which is fixed in patch 3 in this set ?
>
> no. It's the same but the TODO was not removed. I intended to do
> so after a possible merge.

OK. I'll fix up the commit message and drop the TODO message from
this patch when I put the series into target-arm.next, if that's ok.

thanks
-- PMM