[Skiboot] [PATCH 5/7] npu2-opencapi: ODL should be in reset when enabled

[Frederic Barrat]

Le 06/03/2019 à 01:38, Andrew Donnellan a écrit :
> On 2/3/19 12:52 am, Frederic Barrat wrote:
>> We haven't hit any problem so far, but from the ODL designer, the ODL
>> should be in reset when it is enabled.
>>
>> Signed-off-by: Frederic Barrat <fbarrat at linux.ibm.com>
> 
> So to be clear - we don't need to deassert reset ourselves, that happens 
> in hardware?

No, we need to deassert ourselves. With this patch, we put the ODL in 
reset very early, before it's enabled. The ODL stays in reset until we 
call fundamental reset and start link training. At boot, the 
assert_odl_assert() we have as part of the link training procedure will 
therefore be a no-op, since the ODL is already in reset. It's only a 
scom write and it's going to be required when triggering a reset from 
the OS, so I'm keeping it. The ODL is taken out of reset when we call 
deassert_odl_reset() in our link training procedure, as usual.
I'll add a better description in the commit message.

   Fred


> Assuming that:
> 
> Reviewed-by: Andrew Donnellan <andrew.donnellan at au1.ibm.com>
> 
>> ---
>>   hw/npu2-opencapi.c | 6 ++++++
>>   1 file changed, 6 insertions(+)
>>
>> diff --git a/hw/npu2-opencapi.c b/hw/npu2-opencapi.c
>> index 1cdae100..a0c1dba1 100644
>> --- a/hw/npu2-opencapi.c
>> +++ b/hw/npu2-opencapi.c
>> @@ -308,6 +308,12 @@ static void enable_odl_phy_mux(uint32_t gcid, int 
>> index)
>>           assert(false);
>>       }
>> +    /*
>> +     * ODL must be in reset when enabling.
>> +     * It stays in reset until the link is trained
>> +     */
>> +    assert_odl_reset(gcid, index);
>> +
>>       /* PowerBus OLL PHY Training Config Register */
>>       xscom_read(gcid, phy_config_scom, &reg);
>>
>