From 673b975f1fbad5382f76afdb987e7513c5f4b71b Mon Sep 17 00:00:00 2001 From: Dean Luick Date: Wed, 31 Aug 2016 07:24:33 -0700 Subject: IB/hfi1: Add QSFP sanity pre-check Sometimes a QSFP device does not respond in the expected time after a power-on. Add a read pre-check/retry when starting the link on driver load. Reviewed-by: Easwar Hariharan Signed-off-by: Dean Luick Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c index b32638d..ec3635a 100644 --- a/drivers/infiniband/hw/hfi1/chip.c +++ b/drivers/infiniband/hw/hfi1/chip.c @@ -9490,6 +9490,78 @@ static void init_lcb(struct hfi1_devdata *dd) write_csr(dd, DC_LCB_CFG_TX_FIFOS_RESET, 0x00); } +/* + * Perform a test read on the QSFP. Return 0 on success, -ERRNO + * on error. + */ +static int test_qsfp_read(struct hfi1_pportdata *ppd) +{ + int ret; + u8 status; + + /* report success if not a QSFP */ + if (ppd->port_type != PORT_TYPE_QSFP) + return 0; + + /* read byte 2, the status byte */ + ret = one_qsfp_read(ppd, ppd->dd->hfi1_id, 2, &status, 1); + if (ret < 0) + return ret; + if (ret != 1) + return -EIO; + + return 0; /* success */ +} + +/* + * Values for QSFP retry. + * + * Give up after 10s (20 x 500ms). The overall timeout was empirically + * arrived at from experience on a large cluster. + */ +#define MAX_QSFP_RETRIES 20 +#define QSFP_RETRY_WAIT 500 /* msec */ + +/* + * Try a QSFP read. If it fails, schedule a retry for later. + * Called on first link activation after driver load. + */ +static void try_start_link(struct hfi1_pportdata *ppd) +{ + if (test_qsfp_read(ppd)) { + /* read failed */ + if (ppd->qsfp_retry_count >= MAX_QSFP_RETRIES) { + dd_dev_err(ppd->dd, "QSFP not responding, giving up\n"); + return; + } + dd_dev_info(ppd->dd, + "QSFP not responding, waiting and retrying %d\n", + (int)ppd->qsfp_retry_count); + ppd->qsfp_retry_count++; + queue_delayed_work(ppd->hfi1_wq, &ppd->start_link_work, + msecs_to_jiffies(QSFP_RETRY_WAIT)); + return; + } + ppd->qsfp_retry_count = 0; + + /* + * Tune the SerDes to a ballpark setting for optimal signal and bit + * error rate. Needs to be done before starting the link. + */ + tune_serdes(ppd); + start_link(ppd); +} + +/* + * Workqueue function to start the link after a delay. + */ +void handle_start_link(struct work_struct *work) +{ + struct hfi1_pportdata *ppd = container_of(work, struct hfi1_pportdata, + start_link_work.work); + try_start_link(ppd); +} + int bringup_serdes(struct hfi1_pportdata *ppd) { struct hfi1_devdata *dd = ppd->dd; @@ -9525,14 +9597,8 @@ int bringup_serdes(struct hfi1_pportdata *ppd) set_qsfp_int_n(ppd, 1); } - /* - * Tune the SerDes to a ballpark setting for - * optimal signal and bit error rate - * Needs to be done before starting the link - */ - tune_serdes(ppd); - - return start_link(ppd); + try_start_link(ppd); + return 0; } void hfi1_quiet_serdes(struct hfi1_pportdata *ppd) @@ -9549,6 +9615,10 @@ void hfi1_quiet_serdes(struct hfi1_pportdata *ppd) ppd->driver_link_ready = 0; ppd->link_enabled = 0; + ppd->qsfp_retry_count = MAX_QSFP_RETRIES; /* prevent more retries */ + flush_delayed_work(&ppd->start_link_work); + cancel_delayed_work_sync(&ppd->start_link_work); + ppd->offline_disabled_reason = HFI1_ODR_MASK(OPA_LINKDOWN_REASON_SMA_DISABLED); set_link_down_reason(ppd, OPA_LINKDOWN_REASON_SMA_DISABLED, 0, diff --git a/drivers/infiniband/hw/hfi1/chip.h b/drivers/infiniband/hw/hfi1/chip.h index ed11107..e295737 100644 --- a/drivers/infiniband/hw/hfi1/chip.h +++ b/drivers/infiniband/hw/hfi1/chip.h @@ -706,6 +706,7 @@ void handle_link_up(struct work_struct *work); void handle_link_down(struct work_struct *work); void handle_link_downgrade(struct work_struct *work); void handle_link_bounce(struct work_struct *work); +void handle_start_link(struct work_struct *work); void handle_sma_message(struct work_struct *work); void reset_qsfp(struct hfi1_pportdata *ppd); void qsfp_event(struct work_struct *work); diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h index a021e66..28b9128 100644 --- a/drivers/infiniband/hw/hfi1/hfi.h +++ b/drivers/infiniband/hw/hfi1/hfi.h @@ -605,6 +605,7 @@ struct hfi1_pportdata { struct work_struct freeze_work; struct work_struct link_downgrade_work; struct work_struct link_bounce_work; + struct delayed_work start_link_work; /* host link state variables */ struct mutex hls_lock; u32 host_link_state; @@ -659,6 +660,7 @@ struct hfi1_pportdata { u8 linkinit_reason; u8 local_tx_rate; /* rate given to 8051 firmware */ u8 last_pstate; /* info only */ + u8 qsfp_retry_count; /* placeholders for IB MAD packet settings */ u8 overrun_threshold; diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c index b793545..000dc07 100644 --- a/drivers/infiniband/hw/hfi1/init.c +++ b/drivers/infiniband/hw/hfi1/init.c @@ -500,6 +500,7 @@ void hfi1_init_pportdata(struct pci_dev *pdev, struct hfi1_pportdata *ppd, INIT_WORK(&ppd->link_downgrade_work, handle_link_downgrade); INIT_WORK(&ppd->sma_message_work, handle_sma_message); INIT_WORK(&ppd->link_bounce_work, handle_link_bounce); + INIT_DELAYED_WORK(&ppd->start_link_work, handle_start_link); INIT_WORK(&ppd->linkstate_active_work, receive_interrupt_work); INIT_WORK(&ppd->qsfp_info.qsfp_work, qsfp_event); -- cgit v0.10.2