[PATCH 5/8] accel/habanalabs/gaudi2: assume hard-reset by FW upon MC SEI severe error

From: Ofir Bitton
Date: Mon May 27 2024 - 08:13:21 EST


From: Tomer Tayar <ttayar@xxxxxxxxx>

FW initiates a hard reset upon an MC SEI severe error.
Align the driver to expect this reset and avoid accessing the device
until the reset is done.

Signed-off-by: Tomer Tayar <ttayar@xxxxxxxxx>
Reviewed-by: Ofir Bitton <obitton@xxxxxxxxx>
---
drivers/accel/habanalabs/gaudi2/gaudi2.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/accel/habanalabs/gaudi2/gaudi2.c b/drivers/accel/habanalabs/gaudi2/gaudi2.c
index 18cc7b773650..4791582d157c 100644
--- a/drivers/accel/habanalabs/gaudi2/gaudi2.c
+++ b/drivers/accel/habanalabs/gaudi2/gaudi2.c
@@ -10004,6 +10004,7 @@ static void gaudi2_handle_eqe(struct hl_device *hdev, struct hl_eq_entry *eq_ent
if (gaudi2_handle_hbm_mc_sei_err(hdev, event_type, &eq_entry->sei_data)) {
reset_flags |= HL_DRV_RESET_FW_FATAL_ERR;
reset_required = true;
+ is_critical = eq_entry->sei_data.hdr.is_critical;
}
error_count++;
break;
@@ -10235,8 +10236,7 @@ static void gaudi2_handle_eqe(struct hl_device *hdev, struct hl_eq_entry *eq_ent
gaudi2_print_event(hdev, event_type, true,
"No error cause for H/W event %u", event_type);

- if ((gaudi2_irq_map_table[event_type].reset != EVENT_RESET_TYPE_NONE) ||
- reset_required) {
+ if ((gaudi2_irq_map_table[event_type].reset != EVENT_RESET_TYPE_NONE) || reset_required) {
if (reset_required ||
(gaudi2_irq_map_table[event_type].reset == EVENT_RESET_TYPE_HARD))
reset_flags |= HL_DRV_RESET_HARD;
--
2.34.1