diff options
| author | farah kassabri <fkassabri@habana.ai> | 2023-08-23 11:36:25 +0200 |
|---|---|---|
| committer | Oded Gabbay <ogabbay@kernel.org> | 2023-10-09 11:37:21 +0200 |
| commit | 764bfd138f359423b299b7bf3fcbabb56b981ef5 (patch) | |
| tree | 56e1e9704fad58fc8f122c91c4a4b8dc73181a05 /drivers/accel/habanalabs/gaudi2/gaudi2.c | |
| parent | accel/habanalabs/gaudi2: handle eq health heartbeat check (diff) | |
| download | linux-764bfd138f359423b299b7bf3fcbabb56b981ef5.tar.xz linux-764bfd138f359423b299b7bf3fcbabb56b981ef5.zip | |
accel/habanalabs/gaudi2: add eq health check using irq
This is the second patch for applying the eq health check mechanism
which will add support for the interrupt flow for gaudi2 asic.
More info about the interrupt mechanism:
set a dedicated msix for the eq error interrupt, and add
interrupt handler for it.
when FW detects some issue with EQ like EQ_FULL, it'll
raise that interrupt and driver should reset the device.
Driver will inform the FW which msix index to use through
the already existing handshake mechanism which will
send msix info message to fw.
Signed-off-by: farah kassabri <fkassabri@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Diffstat (limited to 'drivers/accel/habanalabs/gaudi2/gaudi2.c')
| -rw-r--r-- | drivers/accel/habanalabs/gaudi2/gaudi2.c | 16 |
1 files changed, 16 insertions, 0 deletions
diff --git a/drivers/accel/habanalabs/gaudi2/gaudi2.c b/drivers/accel/habanalabs/gaudi2/gaudi2.c index e507847bf460..b0ba62b691ec 100644 --- a/drivers/accel/habanalabs/gaudi2/gaudi2.c +++ b/drivers/accel/habanalabs/gaudi2/gaudi2.c @@ -4175,6 +4175,8 @@ static const char *gaudi2_irq_name(u16 irq_number) return "gaudi2 unexpected error"; case GAUDI2_IRQ_NUM_USER_FIRST ... GAUDI2_IRQ_NUM_USER_LAST: return "gaudi2 user completion"; + case GAUDI2_IRQ_NUM_EQ_ERROR: + return "gaudi2 eq error"; default: return "invalid"; } @@ -4317,6 +4319,15 @@ static int gaudi2_enable_msix(struct hl_device *hdev) } } + irq = pci_irq_vector(hdev->pdev, GAUDI2_IRQ_NUM_EQ_ERROR); + rc = request_threaded_irq(irq, NULL, hl_irq_eq_error_interrupt_thread_handler, + IRQF_ONESHOT, gaudi2_irq_name(GAUDI2_IRQ_NUM_EQ_ERROR), + hdev); + if (rc) { + dev_err(hdev->dev, "Failed to request IRQ %d", irq); + goto free_user_irq; + } + gaudi2->hw_cap_initialized |= HW_CAP_MSIX; return 0; @@ -4376,6 +4387,7 @@ static void gaudi2_sync_irqs(struct hl_device *hdev) } synchronize_irq(pci_irq_vector(hdev->pdev, GAUDI2_IRQ_NUM_EVENT_QUEUE)); + synchronize_irq(pci_irq_vector(hdev->pdev, GAUDI2_IRQ_NUM_EQ_ERROR)); } static void gaudi2_disable_msix(struct hl_device *hdev) @@ -4412,6 +4424,9 @@ static void gaudi2_disable_msix(struct hl_device *hdev) cq = &hdev->completion_queue[GAUDI2_RESERVED_CQ_CS_COMPLETION]; free_irq(irq, cq); + irq = pci_irq_vector(hdev->pdev, GAUDI2_IRQ_NUM_EQ_ERROR); + free_irq(irq, hdev); + pci_free_irq_vectors(hdev->pdev); gaudi2->hw_cap_initialized &= ~HW_CAP_MSIX; @@ -11345,6 +11360,7 @@ static int gaudi2_ack_mmu_page_fault_or_access_error(struct hl_device *hdev, u64 static void gaudi2_get_msi_info(__le32 *table) { table[CPUCP_EVENT_QUEUE_MSI_TYPE] = cpu_to_le32(GAUDI2_EVENT_QUEUE_MSIX_IDX); + table[CPUCP_EVENT_QUEUE_ERR_MSI_TYPE] = cpu_to_le32(GAUDI2_IRQ_NUM_EQ_ERROR); } static int gaudi2_map_pll_idx_to_fw_idx(u32 pll_idx) |