Messages in this thread Patch in this message | | | From | Oded Gabbay <> | Subject | [PATCH 3/5] habanalabs: fail reset if device is not idle | Date | Tue, 23 Feb 2021 00:08:04 +0200 |
| |
After any reset (soft or hard) the device (the engines/QMANs) should be idle. If they are not idle, fail the reset. If it is soft-reset, the driver will try to do hard-reset automatically. If it is hard-reset, the driver will make the device non-operational.
Signed-off-by: Oded Gabbay <ogabbay@kernel.org> --- drivers/misc/habanalabs/common/device.c | 25 +++++++++++-------------- 1 file changed, 11 insertions(+), 14 deletions(-)
diff --git a/drivers/misc/habanalabs/common/device.c b/drivers/misc/habanalabs/common/device.c index 8cc3264ae378..48d301a03d62 100644 --- a/drivers/misc/habanalabs/common/device.c +++ b/drivers/misc/habanalabs/common/device.c @@ -71,21 +71,8 @@ static void hpriv_release(struct kref *ref) kfree(hpriv); - if (hdev->reset_upon_device_release) { - u64 idle_mask[HL_BUSY_ENGINES_MASK_EXT_SIZE] = {0}; - - /* We try soft reset first */ + if (hdev->reset_upon_device_release) hl_device_reset(hdev, false, false); - - /* If device is not idle perform hard reset */ - if (!hdev->asic_funcs->is_device_idle(hdev, idle_mask, - HL_BUSY_ENGINES_MASK_EXT_SIZE, NULL)) { - dev_info(hdev->dev, - "device is not idle (mask %#llx %#llx) after soft reset, performing hard reset", - idle_mask[0], idle_mask[1]); - hl_device_reset(hdev, true, false); - } - } } void hl_hpriv_get(struct hl_fpriv *hpriv) @@ -921,6 +908,7 @@ static int device_kill_open_processes(struct hl_device *hdev, u32 timeout) int hl_device_reset(struct hl_device *hdev, bool hard_reset, bool from_hard_reset_thread) { + u64 idle_mask[HL_BUSY_ENGINES_MASK_EXT_SIZE] = {0}; int i, rc; if (!hdev->init_done) { @@ -1140,6 +1128,15 @@ int hl_device_reset(struct hl_device *hdev, bool hard_reset, goto out_err; } + /* If device is not idle fail the reset process */ + if (!hdev->asic_funcs->is_device_idle(hdev, idle_mask, + HL_BUSY_ENGINES_MASK_EXT_SIZE, NULL)) { + dev_err(hdev->dev, + "device is not idle (mask %#llx %#llx) after reset\n", + idle_mask[0], idle_mask[1]); + goto out_err; + } + /* Check that the communication with the device is working */ rc = hdev->asic_funcs->test_queues(hdev); if (rc) { -- 2.25.1
| |