> On 6 May 2021, at 09:29, Leon Romanovsky <leon@xxxxxxxxxx> wrote:
>
> On Tue, May 04, 2021 at 01:46:24PM +0200, Håkon Bugge wrote:
>> Both the PKEY and GID tables in an HCA can hold in the order of
>> hundreds entries. Reading them are expensive. Partly because the API
>> for retrieving them only returns a single entry at a time. Further, on
>> certain implementations, e.g., CX-3, the VFs are paravirtualized in
>> this respect and have to rely on the PF driver to perform the
>> read. This again demands VF to PF communication.
>>
>> IB Core's cache is refreshed on all events. Hence, filter the refresh
>> of the PKEY and GID caches based on the event received being
>> IB_EVENT_PKEY_CHANGE and IB_EVENT_GID_CHANGE respectively.
>>
>> Signed-off-by: Håkon Bugge <haakon.bugge@xxxxxxxxxx>
>> ---
>> drivers/infiniband/core/cache.c | 25 ++++++++++++++++---------
>> 1 file changed, 16 insertions(+), 9 deletions(-)
>>
>> diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c
>> index 5c9fac7..531ae6b 100644
>> --- a/drivers/infiniband/core/cache.c
>> +++ b/drivers/infiniband/core/cache.c
>> @@ -1472,10 +1472,14 @@ static int config_non_roce_gid_cache(struct ib_device *device,
>> }
>>
>> static int
>> -ib_cache_update(struct ib_device *device, u8 port, bool enforce_security)
>> +ib_cache_update(struct ib_device *device, u8 port, enum ib_event_type event,
>> + bool reg_dev, bool enforce_security)
>> {
>> struct ib_port_attr *tprops = NULL;
>> - struct ib_pkey_cache *pkey_cache = NULL, *old_pkey_cache;
>> + struct ib_pkey_cache *pkey_cache = NULL;
>> + struct ib_pkey_cache *old_pkey_cache = NULL;
>> + bool update_pkey_cache = (reg_dev || event == IB_EVENT_PKEY_CHANGE);
>> + bool update_gid_cache = (reg_dev || event == IB_EVENT_GID_CHANGE);
>
> I'm not super excited to see "events" in this function and would be more
> happy to see it is handled in ib_cache_event_task() while the function
> signature will be:
> ib_cache_update(struct ib_device *device, u8 port, bool all_gids, bool
> all_pkeys, bool enforce_security)
I was thinking the other way around; if a new entity, FOO, gets cached, to be updated on IB_EVENT_FOO_CHANGE, the change to support this would be more contained; you only need to change ib_cache_update().
But by all means, will send a v2 based on you recommendation.
Thxs, Håkon
>
> Thanks
