From 051868d93cfd342b4ff8e6297b93a6a43dbe81b3 Mon Sep 17 00:00:00 2001 From: farah kassabri Date: Sun, 27 Aug 2023 19:01:20 +0300 Subject: accel/habanalabs: prevent sending heartbeat before events are enabled After the heartbeat mechanism is now expanded to be used also for EQ health check, we shouldn't send heartbeat messages to FW before driver allow events to be received from FW. Because if the driver will send two heartbeats before it enables events to be received from FW, then the EQ health check will fail and reset the device. Signed-off-by: farah kassabri Reviewed-by: Oded Gabbay Signed-off-by: Oded Gabbay --- drivers/accel/habanalabs/common/device.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) (limited to 'drivers/accel') diff --git a/drivers/accel/habanalabs/common/device.c b/drivers/accel/habanalabs/common/device.c index 1d1ccd8d5c75..0c9ba09c1b75 100644 --- a/drivers/accel/habanalabs/common/device.c +++ b/drivers/accel/habanalabs/common/device.c @@ -994,12 +994,7 @@ static void hl_device_eq_heartbeat(struct hl_device *hdev) u64 event_mask = HL_NOTIFIER_EVENT_DEVICE_RESET | HL_NOTIFIER_EVENT_DEVICE_UNAVAILABLE; struct asic_fixed_properties *prop = &hdev->asic_prop; - /* - * This feature supported in FW version 1.12.0 45.2.0 and above, - * only on those FW versions eq_health_check_supported will be set. - * Start checking eq health only after driver has enabled events from FW. - */ - if (!prop->cpucp_info.eq_health_check_supported || !hdev->init_done) + if (!prop->cpucp_info.eq_health_check_supported) return; if (hdev->eq_heartbeat_received) @@ -1015,7 +1010,8 @@ static void hl_device_heartbeat(struct work_struct *work) struct hl_info_fw_err_info info = {0}; u64 event_mask = HL_NOTIFIER_EVENT_DEVICE_RESET | HL_NOTIFIER_EVENT_DEVICE_UNAVAILABLE; - if (!hl_device_operational(hdev, NULL)) + /* Start heartbeat checks only after driver has enabled events from FW */ + if (!hl_device_operational(hdev, NULL) || !hdev->init_done) goto reschedule; /* -- cgit v1.2.3