From 975ab7b32b90c97046ddbdd53798391b7d8a6a1e Mon Sep 17 00:00:00 2001
From: Oded Gabbay <oded.gabbay@gmail.com>
Date: Tue, 1 Sep 2020 11:22:05 +0300
Subject: [PATCH] habanalabs: count dropped CS because max CS in-flight
There is a case where the user reaches the maximum number of CS in-flight.
In that case, the driver rejects the new CS of the user with EAGAIN. Count
that event so the user can query the driver later to see if it happened.
Reviewed-by: Tomer Tayar <ttayar@habana.ai>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
---
drivers/misc/habanalabs/common/command_submission.c | 5 ++++-
include/uapi/misc/habanalabs.h | 2 ++
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/misc/habanalabs/common/command_submission.c b/drivers/misc/habanalabs/common/command_submission.c
index a811a9fdf13b7..470bffbe9bdc7 100644
--- a/drivers/misc/habanalabs/common/command_submission.c
+++ b/drivers/misc/habanalabs/common/command_submission.c
@@ -252,6 +252,8 @@ static void cs_counters_aggregate(struct hl_device *hdev, struct hl_ctx *ctx)
ctx->cs_counters.parsing_drop_cnt;
hdev->aggregated_cs_counters.queue_full_drop_cnt +=
ctx->cs_counters.queue_full_drop_cnt;
+ hdev->aggregated_cs_counters.max_cs_in_flight_drop_cnt +=
+ ctx->cs_counters.max_cs_in_flight_drop_cnt;
}
static void cs_do_release(struct kref *ref)
@@ -431,8 +433,9 @@ static int allocate_cs(struct hl_device *hdev, struct hl_ctx *ctx,
(hdev->asic_prop.max_pending_cs - 1)];
if (other && !completion_done(&other->completion)) {
- dev_dbg(hdev->dev,
+ dev_dbg_ratelimited(hdev->dev,
"Rejecting CS because of too many in-flights CS\n");
+ ctx->cs_counters.max_cs_in_flight_drop_cnt++;
rc = -EAGAIN;
goto free_fence;
}
diff --git a/include/uapi/misc/habanalabs.h b/include/uapi/misc/habanalabs.h
index a2dcad29340fe..69fb44d352922 100644
--- a/include/uapi/misc/habanalabs.h
+++ b/include/uapi/misc/habanalabs.h
@@ -401,12 +401,14 @@ struct hl_info_sync_manager {
* @parsing_drop_cnt: dropped due to error in packet parsing
* @queue_full_drop_cnt: dropped due to queue full
* @device_in_reset_drop_cnt: dropped due to device in reset
+ * @max_cs_in_flight_drop_cnt: dropped due to maximum CS in-flight
*/
struct hl_cs_counters {
__u64 out_of_mem_drop_cnt;
__u64 parsing_drop_cnt;
__u64 queue_full_drop_cnt;
__u64 device_in_reset_drop_cnt;
+ __u64 max_cs_in_flight_drop_cnt;
};
struct hl_info_cs_counters {
--
GitLab