On Tue, May 19, 2015 at 9:45 PM, Jason Gunthorpe
<[email protected]> wrote:
> On Sun, May 17, 2015 at 04:36:17PM +0300, Or Gerlitz wrote:
>> From: Matan Barak <[email protected]>
>>
>> Add a new ib_cq_init_attr structure which contains the
>> previous cqe (minimum number of CQ entries) and comp_vector
>> (completion vector) in addition to a new flags field.
>> All vendors' create_cq callbacks are changed in order
>> to work with the new API.
>>
>> This commit does not change any functionality.
>
> This seems reasonable to me.
>
>> @@ -1341,6 +1341,7 @@ ssize_t ib_uverbs_create_cq(struct ib_uverbs_file
>> *file,
>> struct ib_uverbs_event_file *ev_file = NULL;
>> struct ib_cq *cq;
>> int ret;
>> + struct ib_cq_init_attr attr = {.cqe = 0};
>
> This doesn't seem necessary, it is unconditionally set below:
>
Almost :) It also zeros (default value) all other fields. I could
replace it with a memset if it's clearer.
>> if (out_len < sizeof resp)
>> return -ENOSPC;
>> @@ -1376,8 +1377,9 @@ ssize_t ib_uverbs_create_cq(struct ib_uverbs_file
>> *file,
>> INIT_LIST_HEAD(&obj->comp_list);
>> INIT_LIST_HEAD(&obj->async_list);
>>
>> - cq = file->device->ib_dev->create_cq(file->device->ib_dev, cmd.cqe,
>> - cmd.comp_vector,
>> + attr.cqe = cmd.cqe;
>> + attr.comp_vector = cmd.comp_vector;
>> + cq = file->device->ib_dev->create_cq(file->device->ib_dev, &attr,
>> file->ucontext, &udata);
>
>> +++ b/drivers/infiniband/core/verbs.c
>> @@ -1013,8 +1013,9 @@ struct ib_cq *ib_create_cq(struct ib_device *device,
>> void *cq_context, int cqe, int comp_vector)
>> {
>> struct ib_cq *cq;
>> + struct ib_cq_init_attr attr = {.cqe = cqe, .comp_vector = comp_vector};
>>
>> - cq = device->create_cq(device, cqe, comp_vector, NULL, NULL);
>> + cq = device->create_cq(device, &attr, NULL, NULL);
>
> Hum, I guess it makes sense to stop flowing ib_cq_init_attr at this
> point, for this patch, but it does seem a bit weird from an API design
> perspective.
>
I guess you suggest that ib_create_cq will take ib_cq_init_attr * instead of
separate parameters for cqe and comp_vector. Ok, I'll change that for next
patch set.
>> diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
>> index 8d59479..ad0e2ea 100644
>> +++ b/include/rdma/ib_verbs.h
>> @@ -173,6 +173,12 @@ struct ib_odp_caps {
>> } per_transport_caps;
>> };
>>
>> +struct ib_cq_init_attr {
>> + int cqe;
>
> unsigned int cqe
>
> Can't be negative..
>
Correct, but that's the current API as well. I'll change that.
>> + struct ib_cq * (*create_cq)(struct ib_device *device,
>> + struct ib_cq_init_attr *attr,
>
> const struct ib_cq_init_attr *attr,
>
> And related changes that will cause.
>
I'll change that, thanks!
Matan
> Jason
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
