[Skiboot] [PATCH v8 21/24] MPIPL: Invalidate dump

[Vasant Hegde]

On 06/28/2019 07:17 AM, Nicholas Piggin wrote:
> Vasant Hegde's on June 17, 2019 3:10 am:
>> Post dump process, kernel sends FREE_PRESERVE_MEMEORY notification
>> to OPAL. OPAL will clear metadata section and tags.
>>
>> Signed-off-by: Vasant Hegde <hegdevasant at linux.vnet.ibm.com>
>> ---
>>   core/opal-dump.c | 10 ++++++++++
>>   1 file changed, 10 insertions(+)
>>
>> diff --git a/core/opal-dump.c b/core/opal-dump.c
>> index e53991784..555908114 100644
>> --- a/core/opal-dump.c
>> +++ b/core/opal-dump.c
>> @@ -309,6 +309,16 @@ static int64_t opal_mpipl_update(enum mpipl_ops ops,
>>   			prlog(PR_NOTICE, "Payload unregistered for MPIPL\n");
>>   		break;
>>   	case OPAL_MPIPL_FREE_PRESERVED_MEMORY:
>> +		/* Clear tags */
>> +		memset(&mpipl_tags, 0, (sizeof(u64) * MAX_MPIPL_TAGS));
>> +		max_tags = 0;
>> +		/* Release memory */
>> +		free(mpipl_opal_data);
>> +		mpipl_opal_data = NULL;
>> +		/* Clear MDRT table */
>> +		memset((void *)MDRT_TABLE_BASE, 0, MDRT_TABLE_SIZE);
>> +		/* Set MDRT count to max allocated count */
>> +		ntuple_mdrt->act_cnt = MDRT_TABLE_SIZE / sizeof(struct mdrt_table);
> 
> Any particular reason you add this here in this patch rather than where
> the call type was defined?

I have added patch based on dump workflow instead of adding everything in one patch.
(Prep work -> OPAL dump registration -> kernel API for registration -> handle 
assert part ->
   trigger dump -> Post processing -> Invalidate dump).

I thought this will make it easy for others to understand.

-Vasant