* [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update
@ 2023-10-11 18:31 Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 1/3] RDMA/bnxt_re: Update HW interface headers Selvin Xavier
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: Selvin Xavier @ 2023-10-11 18:31 UTC (permalink / raw
To: jgg, leon; +Cc: linux-rdma, andrew.gospodarek, Selvin Xavier
[-- Attachment #1: Type: text/plain, Size: 575 bytes --]
Reports async error events received from the HW.
Please review and apply
Regards,
Selvin
v1 -> v2:
- Remove couple of unused defines in patch 1
- Avoid dev_err and use ibdev_err/ibdev_dbg
Chandramohan Akula (3):
RDMA/bnxt_re: Update HW interface headers
RDMA/bnxt_re: Report async events and errors
RDMA/bnxt_re: Do not report SRQ error in srq notification
drivers/infiniband/hw/bnxt_re/main.c | 172 ++++++++++++++++++++++++++++---
drivers/infiniband/hw/bnxt_re/roce_hsi.h | 58 +++++++++++
2 files changed, 216 insertions(+), 14 deletions(-)
--
2.5.5
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4224 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH for-next v2 1/3] RDMA/bnxt_re: Update HW interface headers
2023-10-11 18:31 [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update Selvin Xavier
@ 2023-10-11 18:31 ` Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 2/3] RDMA/bnxt_re: Report async events and errors Selvin Xavier
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Selvin Xavier @ 2023-10-11 18:31 UTC (permalink / raw
To: jgg, leon
Cc: linux-rdma, andrew.gospodarek, Chandramohan Akula, Selvin Xavier
[-- Attachment #1: Type: text/plain, Size: 6514 bytes --]
From: Chandramohan Akula <chandramohan.akula@broadcom.com>
Updating the HW structures for the affiliated event and error
reporting. Newly added interface structures will be used in the
followup patch.
Signed-off-by: Chandramohan Akula <chandramohan.akula@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
---
drivers/infiniband/hw/bnxt_re/roce_hsi.h | 58 ++++++++++++++++++++++++++++++++
1 file changed, 58 insertions(+)
diff --git a/drivers/infiniband/hw/bnxt_re/roce_hsi.h b/drivers/infiniband/hw/bnxt_re/roce_hsi.h
index 4a10303..84b5acd 100644
--- a/drivers/infiniband/hw/bnxt_re/roce_hsi.h
+++ b/drivers/infiniband/hw/bnxt_re/roce_hsi.h
@@ -2919,6 +2919,35 @@ struct creq_qp_error_notification {
u8 status;
u8 req_slow_path_state;
u8 req_err_state_reason;
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_NO_ERROR 0X0UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_OPCODE_ERROR 0X1UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_TIMEOUT_RETRY_LIMIT 0X2UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RNR_TIMEOUT_RETRY_LIMIT 0X3UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_NAK_ARRIVAL_1 0X4UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_NAK_ARRIVAL_2 0X5UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_NAK_ARRIVAL_3 0X6UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_NAK_ARRIVAL_4 0X7UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RX_MEMORY_ERROR 0X8UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_TX_MEMORY_ERROR 0X9UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_READ_RESP_LENGTH 0XAUL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_INVALID_READ_RESP 0XBUL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_ILLEGAL_BIND 0XCUL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_ILLEGAL_FAST_REG 0XDUL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_ILLEGAL_INVALIDATE 0XEUL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_CMP_ERROR 0XFUL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RETRAN_LOCAL_ERROR 0X10UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_WQE_FORMAT_ERROR 0X11UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_ORRQ_FORMAT_ERROR 0X12UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_INVALID_AVID_ERROR 0X13UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_AV_DOMAIN_ERROR 0X14UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_CQ_LOAD_ERROR 0X15UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_SERV_TYPE_ERROR 0X16UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_INVALID_OP_ERROR 0X17UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_TX_PCI_ERROR 0X18UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RX_PCI_ERROR 0X19UL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_PROD_WQE_MSMTCH_ERROR 0X1AUL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_PSN_RANGE_CHECK_ERROR 0X1BUL
+ #define CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RETX_SETUP_ERROR 0X1CUL
__le32 xid;
u8 v;
#define CREQ_QP_ERROR_NOTIFICATION_V 0x1UL
@@ -2928,6 +2957,35 @@ struct creq_qp_error_notification {
CREQ_QP_ERROR_NOTIFICATION_EVENT_QP_ERROR_NOTIFICATION
u8 res_slow_path_state;
u8 res_err_state_reason;
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_NO_ERROR 0x0UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_EXCEED_MAX 0x1UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_PAYLOAD_LENGTH_MISMATCH 0x2UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_EXCEEDS_WQE 0x3UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_OPCODE_ERROR 0x4UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_PSN_SEQ_ERROR_RETRY_LIMIT 0x5UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_INVALID_R_KEY 0x6UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_DOMAIN_ERROR 0x7UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_NO_PERMISSION 0x8UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_RANGE_ERROR 0x9UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_INVALID_R_KEY 0xaUL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_DOMAIN_ERROR 0xbUL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_NO_PERMISSION 0xcUL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_RANGE_ERROR 0xdUL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_IRRQ_OFLOW 0xeUL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_UNSUPPORTED_OPCODE 0xfUL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_UNALIGN_ATOMIC 0x10UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_REM_INVALIDATE 0x11UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_MEMORY_ERROR 0x12UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_SRQ_ERROR 0x13UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_CMP_ERROR 0x14UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_INVALID_DUP_RKEY 0x15UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_WQE_FORMAT_ERROR 0x16UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_IRRQ_FORMAT_ERROR 0x17UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_CQ_LOAD_ERROR 0x18UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_SRQ_LOAD_ERROR 0x19UL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_PCI_ERROR 0x1bUL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_PCI_ERROR 0x1cUL
+ #define CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_PSN_NOT_FOUND 0x1dUL
__le16 sq_cons_idx;
__le16 rq_cons_idx;
};
--
2.5.5
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4224 bytes --]
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH for-next v2 2/3] RDMA/bnxt_re: Report async events and errors
2023-10-11 18:31 [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 1/3] RDMA/bnxt_re: Update HW interface headers Selvin Xavier
@ 2023-10-11 18:31 ` Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 3/3] RDMA/bnxt_re: Do not report SRQ error in srq notification Selvin Xavier
2023-10-15 8:52 ` [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update Leon Romanovsky
3 siblings, 0 replies; 5+ messages in thread
From: Selvin Xavier @ 2023-10-11 18:31 UTC (permalink / raw
To: jgg, leon
Cc: linux-rdma, andrew.gospodarek, Chandramohan Akula, Selvin Xavier
[-- Attachment #1: Type: text/plain, Size: 9249 bytes --]
From: Chandramohan Akula <chandramohan.akula@broadcom.com>
Report QP, SRQ and CQ async events and errors.
Signed-off-by: Chandramohan Akula <chandramohan.akula@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
---
drivers/infiniband/hw/bnxt_re/main.c | 165 +++++++++++++++++++++++++++++++++--
1 file changed, 156 insertions(+), 9 deletions(-)
diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index c9066aa..ae72207 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -970,6 +970,9 @@ static int bnxt_re_handle_unaffi_async_event(struct creq_func_event
static int bnxt_re_handle_qp_async_event(struct creq_qp_event *qp_event,
struct bnxt_re_qp *qp)
{
+ struct bnxt_re_srq *srq = container_of(qp->qplib_qp.srq, struct bnxt_re_srq,
+ qplib_srq);
+ struct creq_qp_error_notification *err_event;
struct ib_event event = {};
unsigned int flags;
@@ -980,14 +983,146 @@ static int bnxt_re_handle_qp_async_event(struct creq_qp_event *qp_event,
bnxt_re_unlock_cqs(qp, flags);
}
- if (qp->qplib_qp.srq) {
- event.device = &qp->rdev->ibdev;
- event.element.qp = &qp->ib_qp;
- event.event = IB_EVENT_QP_LAST_WQE_REACHED;
+ event.device = &qp->rdev->ibdev;
+ event.element.qp = &qp->ib_qp;
+ event.event = IB_EVENT_QP_FATAL;
+
+ err_event = (struct creq_qp_error_notification *)qp_event;
+
+ switch (err_event->req_err_state_reason) {
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_OPCODE_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_TIMEOUT_RETRY_LIMIT:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RNR_TIMEOUT_RETRY_LIMIT:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_NAK_ARRIVAL_2:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_NAK_ARRIVAL_3:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_INVALID_READ_RESP:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_ILLEGAL_BIND:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_ILLEGAL_FAST_REG:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_ILLEGAL_INVALIDATE:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RETRAN_LOCAL_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_AV_DOMAIN_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_PROD_WQE_MSMTCH_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_PSN_RANGE_CHECK_ERROR:
+ event.event = IB_EVENT_QP_ACCESS_ERR;
+ break;
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_NAK_ARRIVAL_1:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_NAK_ARRIVAL_4:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_READ_RESP_LENGTH:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_WQE_FORMAT_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_ORRQ_FORMAT_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_INVALID_AVID_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_SERV_TYPE_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_INVALID_OP_ERROR:
+ event.event = IB_EVENT_QP_REQ_ERR;
+ break;
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RX_MEMORY_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_TX_MEMORY_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_CMP_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_CQ_LOAD_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_TX_PCI_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RX_PCI_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_REQ_ERR_STATE_REASON_REQ_RETX_SETUP_ERROR:
+ event.event = IB_EVENT_QP_FATAL;
+ break;
+
+ default:
+ break;
}
- if (event.device && qp->ib_qp.event_handler)
+ switch (err_event->res_err_state_reason) {
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_EXCEED_MAX:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_PAYLOAD_LENGTH_MISMATCH:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_PSN_SEQ_ERROR_RETRY_LIMIT:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_INVALID_R_KEY:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_DOMAIN_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_NO_PERMISSION:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_RANGE_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_INVALID_R_KEY:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_DOMAIN_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_NO_PERMISSION:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_RANGE_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_UNALIGN_ATOMIC:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_PSN_NOT_FOUND:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_INVALID_DUP_RKEY:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_IRRQ_FORMAT_ERROR:
+ event.event = IB_EVENT_QP_ACCESS_ERR;
+ break;
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_EXCEEDS_WQE:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_WQE_FORMAT_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_UNSUPPORTED_OPCODE:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_REM_INVALIDATE:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_OPCODE_ERROR:
+ event.event = IB_EVENT_QP_REQ_ERR;
+ break;
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_IRRQ_OFLOW:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_CMP_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_CQ_LOAD_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_TX_PCI_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_RX_PCI_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_MEMORY_ERROR:
+ event.event = IB_EVENT_QP_FATAL;
+ break;
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_SRQ_LOAD_ERROR:
+ case CREQ_QP_ERROR_NOTIFICATION_RES_ERR_STATE_REASON_RES_SRQ_ERROR:
+ if (srq)
+ event.event = IB_EVENT_SRQ_ERR;
+ break;
+ default:
+ break;
+ }
+
+ if (err_event->res_err_state_reason || err_event->req_err_state_reason) {
+ ibdev_dbg(&qp->rdev->ibdev,
+ "%s %s qp_id: %d cons (%d %d) req (%d %d) res (%d %d)\n",
+ __func__, rdma_is_kernel_res(&qp->ib_qp.res) ? "kernel" : "user",
+ qp->qplib_qp.id,
+ err_event->sq_cons_idx,
+ err_event->rq_cons_idx,
+ err_event->req_slow_path_state,
+ err_event->req_err_state_reason,
+ err_event->res_slow_path_state,
+ err_event->res_err_state_reason);
+ } else {
+ if (srq)
+ event.event = IB_EVENT_QP_LAST_WQE_REACHED;
+ }
+
+ if (event.event == IB_EVENT_SRQ_ERR && srq->ib_srq.event_handler) {
+ (*srq->ib_srq.event_handler)(&event,
+ srq->ib_srq.srq_context);
+ } else if (event.device && qp->ib_qp.event_handler) {
qp->ib_qp.event_handler(&event, qp->ib_qp.qp_context);
+ }
+
+ return 0;
+}
+
+static int bnxt_re_handle_cq_async_error(void *event, struct bnxt_re_cq *cq)
+{
+ struct creq_cq_error_notification *cqerr;
+ struct ib_event ibevent = {};
+
+ cqerr = event;
+ switch (cqerr->cq_err_reason) {
+ case CREQ_CQ_ERROR_NOTIFICATION_CQ_ERR_REASON_REQ_CQ_INVALID_ERROR:
+ case CREQ_CQ_ERROR_NOTIFICATION_CQ_ERR_REASON_REQ_CQ_OVERFLOW_ERROR:
+ case CREQ_CQ_ERROR_NOTIFICATION_CQ_ERR_REASON_REQ_CQ_LOAD_ERROR:
+ case CREQ_CQ_ERROR_NOTIFICATION_CQ_ERR_REASON_RES_CQ_INVALID_ERROR:
+ case CREQ_CQ_ERROR_NOTIFICATION_CQ_ERR_REASON_RES_CQ_OVERFLOW_ERROR:
+ case CREQ_CQ_ERROR_NOTIFICATION_CQ_ERR_REASON_RES_CQ_LOAD_ERROR:
+ ibevent.event = IB_EVENT_CQ_ERR;
+ default:
+ break;
+ }
+
+ if (ibevent.event == IB_EVENT_CQ_ERR && cq->ib_cq.event_handler) {
+ ibevent.element.cq = &cq->ib_cq;
+ ibevent.device = &cq->rdev->ibdev;
+
+ ibdev_dbg(&cq->rdev->ibdev,
+ "%s err reason %d\n", __func__, cqerr->cq_err_reason);
+ cq->ib_cq.event_handler(&ibevent, cq->ib_cq.cq_context);
+ }
return 0;
}
@@ -995,6 +1130,10 @@ static int bnxt_re_handle_qp_async_event(struct creq_qp_event *qp_event,
static int bnxt_re_handle_affi_async_event(struct creq_qp_event *affi_async,
void *obj)
{
+ struct bnxt_qplib_qp *lib_qp;
+ struct bnxt_qplib_cq *lib_cq;
+ struct bnxt_re_qp *qp;
+ struct bnxt_re_cq *cq;
int rc = 0;
u8 event;
@@ -1002,11 +1141,19 @@ static int bnxt_re_handle_affi_async_event(struct creq_qp_event *affi_async,
return rc; /* QP was already dead, still return success */
event = affi_async->event;
- if (event == CREQ_QP_EVENT_EVENT_QP_ERROR_NOTIFICATION) {
- struct bnxt_qplib_qp *lib_qp = obj;
- struct bnxt_re_qp *qp = container_of(lib_qp, struct bnxt_re_qp,
- qplib_qp);
+ switch (event) {
+ case CREQ_QP_EVENT_EVENT_QP_ERROR_NOTIFICATION:
+ lib_qp = obj;
+ qp = container_of(lib_qp, struct bnxt_re_qp, qplib_qp);
rc = bnxt_re_handle_qp_async_event(affi_async, qp);
+ break;
+ case CREQ_QP_EVENT_EVENT_CQ_ERROR_NOTIFICATION:
+ lib_cq = obj;
+ cq = container_of(lib_cq, struct bnxt_re_cq, qplib_cq);
+ rc = bnxt_re_handle_cq_async_error(affi_async, cq);
+ break;
+ default:
+ rc = -EINVAL;
}
return rc;
}
--
2.5.5
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4224 bytes --]
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH for-next v2 3/3] RDMA/bnxt_re: Do not report SRQ error in srq notification
2023-10-11 18:31 [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 1/3] RDMA/bnxt_re: Update HW interface headers Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 2/3] RDMA/bnxt_re: Report async events and errors Selvin Xavier
@ 2023-10-11 18:31 ` Selvin Xavier
2023-10-15 8:52 ` [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update Leon Romanovsky
3 siblings, 0 replies; 5+ messages in thread
From: Selvin Xavier @ 2023-10-11 18:31 UTC (permalink / raw
To: jgg, leon
Cc: linux-rdma, andrew.gospodarek, Chandramohan Akula, Selvin Xavier
[-- Attachment #1: Type: text/plain, Size: 1185 bytes --]
From: Chandramohan Akula <chandramohan.akula@broadcom.com>
In the SRQ notification handler, do not report the SRQ_ERROR
in the default event case, as there was no error.
Signed-off-by: Chandramohan Akula <chandramohan.akula@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
---
drivers/infiniband/hw/bnxt_re/main.c | 7 ++-----
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index ae72207..bd3deb2 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -1187,13 +1187,10 @@ static int bnxt_re_srqn_handler(struct bnxt_qplib_nq *nq,
ib_event.device = &srq->rdev->ibdev;
ib_event.element.srq = &srq->ib_srq;
- if (event == NQ_SRQ_EVENT_EVENT_SRQ_THRESHOLD_EVENT)
- ib_event.event = IB_EVENT_SRQ_LIMIT_REACHED;
- else
- ib_event.event = IB_EVENT_SRQ_ERR;
if (srq->ib_srq.event_handler) {
- /* Lock event_handler? */
+ if (event == NQ_SRQ_EVENT_EVENT_SRQ_THRESHOLD_EVENT)
+ ib_event.event = IB_EVENT_SRQ_LIMIT_REACHED;
(*srq->ib_srq.event_handler)(&ib_event,
srq->ib_srq.srq_context);
}
--
2.5.5
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4224 bytes --]
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update
2023-10-11 18:31 [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update Selvin Xavier
` (2 preceding siblings ...)
2023-10-11 18:31 ` [PATCH for-next v2 3/3] RDMA/bnxt_re: Do not report SRQ error in srq notification Selvin Xavier
@ 2023-10-15 8:52 ` Leon Romanovsky
3 siblings, 0 replies; 5+ messages in thread
From: Leon Romanovsky @ 2023-10-15 8:52 UTC (permalink / raw
To: jgg, Selvin Xavier; +Cc: linux-rdma, andrew.gospodarek
On Wed, 11 Oct 2023 11:31:34 -0700, Selvin Xavier wrote:
> Reports async error events received from the HW.
> Please review and apply
>
> Regards,
> Selvin
>
> v1 -> v2:
> - Remove couple of unused defines in patch 1
> - Avoid dev_err and use ibdev_err/ibdev_dbg
>
> [...]
Applied, thanks!
[1/3] RDMA/bnxt_re: Update HW interface headers
https://git.kernel.org/rdma/rdma/c/d60a779673defc
[2/3] RDMA/bnxt_re: Report async events and errors
https://git.kernel.org/rdma/rdma/c/b02fd3f79ec326
[3/3] RDMA/bnxt_re: Do not report SRQ error in srq notification
https://git.kernel.org/rdma/rdma/c/45cfa8864cd3ae
Best regards,
--
Leon Romanovsky <leon@kernel.org>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2023-10-15 8:52 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-10-11 18:31 [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 1/3] RDMA/bnxt_re: Update HW interface headers Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 2/3] RDMA/bnxt_re: Report async events and errors Selvin Xavier
2023-10-11 18:31 ` [PATCH for-next v2 3/3] RDMA/bnxt_re: Do not report SRQ error in srq notification Selvin Xavier
2023-10-15 8:52 ` [PATCH for-next v2 0/3] RDMA/bnxt_re: Async events update Leon Romanovsky
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.