habanalabs: flush EQ workers in hard reset

During hard-reset, there can be multiple events received from the H/W. For
each event, the driver opens a worker thread to handle it. For some of the
events, the driver will read/write registers in the code that handles the
event.

In case of hard-reset, we must prevent reads/writes to the registers during
the reset operation because the device might get stuck if that happens.

Therefore, flush the EQ workers before resetting the device (in hard-reset
only). Additional events won't arrive as we synced and disabled the
interrupts.

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Tomer Tayar <ttayar@habana.ai>
This commit is contained in:
Oded Gabbay 2019-11-17 17:41:57 +02:00
parent 1af69d30c4
commit 55f6d68097

View File

@ -887,13 +887,19 @@ int hl_device_reset(struct hl_device *hdev, bool hard_reset,
/* Go over all the queues, release all CS and their jobs */ /* Go over all the queues, release all CS and their jobs */
hl_cs_rollback_all(hdev); hl_cs_rollback_all(hdev);
/* Kill processes here after CS rollback. This is because the process if (hard_reset) {
* can't really exit until all its CSs are done, which is what we /* Kill processes here after CS rollback. This is because the
* do in cs rollback * process can't really exit until all its CSs are done, which
*/ * is what we do in cs rollback
if (hard_reset) */
device_kill_open_processes(hdev); device_kill_open_processes(hdev);
/* Flush the Event queue workers to make sure no other thread is
* reading or writing to registers during the reset
*/
flush_workqueue(hdev->eq_wq);
}
/* Release kernel context */ /* Release kernel context */
if ((hard_reset) && (hl_ctx_put(hdev->kernel_ctx) == 1)) if ((hard_reset) && (hl_ctx_put(hdev->kernel_ctx) == 1))
hdev->kernel_ctx = NULL; hdev->kernel_ctx = NULL;