On Thu, Oct 3, 2019 at 6:22 PM Tomer Tayar <tta...@habana.ai> wrote:
>
> Jobs on some queues must be provided with a handle to a driver command
> buffer object, while for other queues, jobs must be provided with an
> address to a command buffer.
> Currently the distinction is done based on the queue type, which is less
> flexible if the same type of queue behaves differently on different
> types of ASIC.
> This patch adds a new queue property for this target, which is
> configured per queue type per ASIC type.
>
> Signed-off-by: Tomer Tayar <tta...@habana.ai>
> ---
> drivers/misc/habanalabs/command_submission.c | 4 +++-
> drivers/misc/habanalabs/goya/goya.c | 3 +++
> drivers/misc/habanalabs/habanalabs.h | 3 +++
> 3 files changed, 9 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/misc/habanalabs/command_submission.c
> b/drivers/misc/habanalabs/command_submission.c
> index a9ac045dcfde..f44205540520 100644
> --- a/drivers/misc/habanalabs/command_submission.c
> +++ b/drivers/misc/habanalabs/command_submission.c
> @@ -414,7 +414,9 @@ static struct hl_cb *validate_queue_index(struct
> hl_device *hdev,
> "Queue index %d is restricted for the kernel
> driver\n",
> chunk->queue_index);
> return NULL;
> - } else if (hw_queue_prop->type == QUEUE_TYPE_INT) {
> + }
> +
> + if (!hw_queue_prop->requires_kernel_cb) {
> *ext_queue = false;
> return (struct hl_cb *) (uintptr_t) chunk->cb_handle;
> }
> diff --git a/drivers/misc/habanalabs/goya/goya.c
> b/drivers/misc/habanalabs/goya/goya.c
> index 09caef7642fd..71693fcffb16 100644
> --- a/drivers/misc/habanalabs/goya/goya.c
> +++ b/drivers/misc/habanalabs/goya/goya.c
> @@ -337,17 +337,20 @@ void goya_get_fixed_properties(struct hl_device *hdev)
> for (i = 0 ; i < NUMBER_OF_EXT_HW_QUEUES ; i++) {
> prop->hw_queues_props[i].type = QUEUE_TYPE_EXT;
> prop->hw_queues_props[i].driver_only = 0;
> + prop->hw_queues_props[i].requires_kernel_cb = 1;
> }
>
> for (; i < NUMBER_OF_EXT_HW_QUEUES + NUMBER_OF_CPU_HW_QUEUES ; i++) {
> prop->hw_queues_props[i].type = QUEUE_TYPE_CPU;
> prop->hw_queues_props[i].driver_only = 1;
> + prop->hw_queues_props[i].requires_kernel_cb = 0;
> }
>
> for (; i < NUMBER_OF_EXT_HW_QUEUES + NUMBER_OF_CPU_HW_QUEUES +
> NUMBER_OF_INT_HW_QUEUES; i++) {
> prop->hw_queues_props[i].type = QUEUE_TYPE_INT;
> prop->hw_queues_props[i].driver_only = 0;
> + prop->hw_queues_props[i].requires_kernel_cb = 0;
> }
>
> for (; i < HL_MAX_QUEUES; i++)
> diff --git a/drivers/misc/habanalabs/habanalabs.h
> b/drivers/misc/habanalabs/habanalabs.h
> index c3d24ffad9fa..f47f4b22cb6b 100644
> --- a/drivers/misc/habanalabs/habanalabs.h
> +++ b/drivers/misc/habanalabs/habanalabs.h
> @@ -98,10 +98,13 @@ enum hl_queue_type {
> * @type: queue type.
> * @driver_only: true if only the driver is allowed to send a job to this
> queue,
> * false otherwise.
> + * @requires_kernel_cb: true if a CB handle must be provided for jobs on this
> + * queue, false otherwise (a CB address must be
> provided).
> */
> struct hw_queue_properties {
> enum hl_queue_type type;
> u8 driver_only;
> + u8 requires_kernel_cb;
> };
>
> /**
> --
> 2.17.1
>
Patches 1-4 are:
Reviewed-by: Oded Gabbay <oded.gab...@gmail.com>
However, I'm only applying patches 1 & 2 to -next, because 3 & 4
mostly add code which is not relevant to GOYA, without much changing
existing code, as opposed to patches 1&2.
Oded
