@@ -15,6 +15,7 @@
* SOFTWARE.
*
*/
+#include <linux/bug.h>
#include <linux/errno.h>
#include <linux/module.h>
#include <linux/spinlock.h>
@@ -27,6 +28,8 @@
#include "usnic_ib_sysfs.h"
#include "usnic_transport.h"
+#define DFLT_RQ_IDX 0
+
const char *usnic_ib_qp_grp_state_to_string(enum ib_qp_state state)
{
switch (state) {
@@ -58,83 +61,31 @@ int usnic_ib_qp_grp_dump_hdr(char *buf, int buf_sz)
int usnic_ib_qp_grp_dump_rows(void *obj, char *buf, int buf_sz)
{
struct usnic_ib_qp_grp *qp_grp = obj;
- struct usnic_fwd_filter_hndl *default_filter_hndl;
+ struct usnic_ib_qp_grp_flow *default_flow;
if (obj) {
- default_filter_hndl = list_first_entry(&qp_grp->filter_hndls,
- struct usnic_fwd_filter_hndl, link);
+ default_flow = list_first_entry(&qp_grp->flows_lst,
+ struct usnic_ib_qp_grp_flow, link);
return scnprintf(buf, buf_sz, "|%d\t|%s\t|%d\t|%hu\t|%d",
qp_grp->ibqp.qp_num,
usnic_ib_qp_grp_state_to_string(
qp_grp->state),
qp_grp->owner_pid,
usnic_vnic_get_index(qp_grp->vf->vnic),
- default_filter_hndl->id);
+ default_flow->flow->flow_id);
} else {
return scnprintf(buf, buf_sz, "|N/A\t|N/A\t|N/A\t|N/A\t|N/A");
}
}
-static int add_fwd_filter(struct usnic_ib_qp_grp *qp_grp,
- struct usnic_fwd_filter *fwd_filter)
-{
- struct usnic_fwd_filter_hndl *filter_hndl;
- int status;
- struct usnic_vnic_res_chunk *chunk;
- int rq_idx;
-
- lockdep_assert_held(&qp_grp->lock);
-
- chunk = usnic_ib_qp_grp_get_chunk(qp_grp, USNIC_VNIC_RES_TYPE_RQ);
- if (IS_ERR_OR_NULL(chunk) || chunk->cnt < 1) {
- usnic_err("Failed to get RQ info for qp_grp %u\n",
- qp_grp->grp_id);
- return -EFAULT;
- }
-
- rq_idx = chunk->res[0]->vnic_idx;
-
- switch (qp_grp->transport) {
- case USNIC_TRANSPORT_ROCE_CUSTOM:
- status = usnic_fwd_add_usnic_filter(qp_grp->ufdev,
- usnic_vnic_get_index(qp_grp->vf->vnic),
- rq_idx,
- fwd_filter,
- &filter_hndl);
- break;
- default:
- usnic_err("Unable to install filter for qp_grp %u for transport %d",
- qp_grp->grp_id, qp_grp->transport);
- status = -EINVAL;
- }
-
- if (status)
- return status;
-
- list_add_tail(&filter_hndl->link, &qp_grp->filter_hndls);
- return 0;
-}
-
-static int del_all_filters(struct usnic_ib_qp_grp *qp_grp)
+static struct usnic_vnic_res_chunk *
+get_qp_res_chunk(struct usnic_ib_qp_grp *qp_grp)
{
- int err, status;
- struct usnic_fwd_filter_hndl *filter_hndl, *tmp;
-
lockdep_assert_held(&qp_grp->lock);
-
- status = 0;
-
- list_for_each_entry_safe(filter_hndl, tmp,
- &qp_grp->filter_hndls, link) {
- list_del(&filter_hndl->link);
- err = usnic_fwd_del_filter(filter_hndl);
- if (err) {
- usnic_err("Failed to delete filter %u of qp_grp %d\n",
- filter_hndl->id, qp_grp->grp_id);
- }
- status |= err;
- }
-
- return status;
+ /*
+ * The QP res chunk, used to derive qp indices,
+ * are just indices of the RQs
+ */
+ return usnic_ib_qp_grp_get_chunk(qp_grp, USNIC_VNIC_RES_TYPE_RQ);
}
static int enable_qp_grp(struct usnic_ib_qp_grp *qp_grp)
@@ -149,22 +100,20 @@ static int enable_qp_grp(struct usnic_ib_qp_grp *qp_grp)
vnic_idx = usnic_vnic_get_index(qp_grp->vf->vnic);
- res_chunk = usnic_ib_qp_grp_get_chunk(qp_grp, USNIC_VNIC_RES_TYPE_RQ);
+ res_chunk = get_qp_res_chunk(qp_grp);
if (IS_ERR_OR_NULL(res_chunk)) {
- usnic_err("Unable to get %s with err %ld\n",
- usnic_vnic_res_type_to_str(USNIC_VNIC_RES_TYPE_RQ),
- PTR_ERR(res_chunk));
+ usnic_err("Unable to get qp res with err %ld\n",
+ PTR_ERR(res_chunk));
return res_chunk ? PTR_ERR(res_chunk) : -ENOMEM;
}
for (i = 0; i < res_chunk->cnt; i++) {
res = res_chunk->res[i];
- status = usnic_fwd_enable_rq(qp_grp->ufdev, vnic_idx,
+ status = usnic_fwd_enable_qp(qp_grp->ufdev, vnic_idx,
res->vnic_idx);
if (status) {
- usnic_err("Failed to enable rq %d of %s:%d\n with err %d\n",
- res->vnic_idx,
- netdev_name(qp_grp->ufdev->netdev),
+ usnic_err("Failed to enable qp %d of %s:%d\n with err %d\n",
+ res->vnic_idx, qp_grp->ufdev->name,
vnic_idx, status);
goto out_err;
}
@@ -175,7 +124,7 @@ static int enable_qp_grp(struct usnic_ib_qp_grp *qp_grp)
out_err:
for (i--; i >= 0; i--) {
res = res_chunk->res[i];
- usnic_fwd_disable_rq(qp_grp->ufdev, vnic_idx,
+ usnic_fwd_disable_qp(qp_grp->ufdev, vnic_idx,
res->vnic_idx);
}
@@ -192,22 +141,21 @@ static int disable_qp_grp(struct usnic_ib_qp_grp *qp_grp)
lockdep_assert_held(&qp_grp->lock);
vnic_idx = usnic_vnic_get_index(qp_grp->vf->vnic);
- res_chunk = usnic_ib_qp_grp_get_chunk(qp_grp, USNIC_VNIC_RES_TYPE_RQ);
+ res_chunk = get_qp_res_chunk(qp_grp);
if (IS_ERR_OR_NULL(res_chunk)) {
- usnic_err("Unable to get %s with err %ld\n",
- usnic_vnic_res_type_to_str(USNIC_VNIC_RES_TYPE_RQ),
+ usnic_err("Unable to get qp res with err %ld\n",
PTR_ERR(res_chunk));
return res_chunk ? PTR_ERR(res_chunk) : -ENOMEM;
}
for (i = 0; i < res_chunk->cnt; i++) {
res = res_chunk->res[i];
- status = usnic_fwd_disable_rq(qp_grp->ufdev, vnic_idx,
+ status = usnic_fwd_disable_qp(qp_grp->ufdev, vnic_idx,
res->vnic_idx);
if (status) {
usnic_err("Failed to disable rq %d of %s:%d\n with err %d\n",
res->vnic_idx,
- netdev_name(qp_grp->ufdev->netdev),
+ qp_grp->ufdev->name,
vnic_idx, status);
}
}
@@ -216,17 +164,148 @@ static int disable_qp_grp(struct usnic_ib_qp_grp *qp_grp)
}
+static int init_filter_action(struct usnic_ib_qp_grp *qp_grp,
+ struct usnic_filter_action *uaction)
+{
+ struct usnic_vnic_res_chunk *res_chunk;
+
+ res_chunk = usnic_ib_qp_grp_get_chunk(qp_grp, USNIC_VNIC_RES_TYPE_RQ);
+ if (IS_ERR_OR_NULL(res_chunk)) {
+ usnic_err("Unable to get %s with err %ld\n",
+ usnic_vnic_res_type_to_str(USNIC_VNIC_RES_TYPE_RQ),
+ PTR_ERR(res_chunk));
+ return res_chunk ? PTR_ERR(res_chunk) : -ENOMEM;
+ }
+
+ uaction->vnic_idx = usnic_vnic_get_index(qp_grp->vf->vnic);
+ uaction->action.type = FILTER_ACTION_RQ_STEERING;
+ uaction->action.u.rq_idx = res_chunk->res[DFLT_RQ_IDX]->vnic_idx;
+
+ return 0;
+}
+
+static struct usnic_ib_qp_grp_flow*
+create_roce_custom_flow(struct usnic_ib_qp_grp *qp_grp,
+ struct usnic_transport_spec *trans_spec)
+{
+ uint16_t port_num;
+ int err;
+ struct filter filter;
+ struct usnic_filter_action uaction;
+ struct usnic_ib_qp_grp_flow *qp_flow;
+ struct usnic_fwd_flow *flow;
+ enum usnic_transport_type trans_type;
+
+ trans_type = trans_spec->trans_type;
+ port_num = trans_spec->usnic_roce.port_num;
+
+ /* Reserve Port */
+ port_num = usnic_transport_rsrv_port(trans_type, port_num);
+ if (port_num == 0)
+ return ERR_PTR(-EINVAL);
+
+ /* Create Flow */
+ usnic_fwd_init_usnic_filter(&filter, port_num);
+ err = init_filter_action(qp_grp, &uaction);
+ if (err)
+ goto out_unreserve_port;
+
+ flow = usnic_fwd_alloc_flow(qp_grp->ufdev, &filter, &uaction);
+ if (IS_ERR_OR_NULL(flow)) {
+ usnic_err("Unable to alloc flow failed with err %ld\n",
+ PTR_ERR(flow));
+ err = (flow) ? PTR_ERR(flow) : -EFAULT;
+ goto out_unreserve_port;
+ }
+
+ /* Create Flow Handle */
+ qp_flow = kzalloc(sizeof(*qp_flow), GFP_ATOMIC);
+ if (IS_ERR_OR_NULL(qp_flow)) {
+ err = (qp_flow) ? PTR_ERR(qp_flow) : -ENOMEM;
+ goto out_dealloc_flow;
+ }
+ qp_flow->flow = flow;
+ qp_flow->trans_type = trans_type;
+ qp_flow->usnic_roce.port_num = port_num;
+ qp_flow->qp_grp = qp_grp;
+ return qp_flow;
+
+out_dealloc_flow:
+ usnic_fwd_dealloc_flow(flow);
+out_unreserve_port:
+ usnic_transport_unrsrv_port(trans_type, port_num);
+ return ERR_PTR(err);
+}
+
+static void release_roce_custom_flow(struct usnic_ib_qp_grp_flow *qp_flow)
+{
+ usnic_fwd_dealloc_flow(qp_flow->flow);
+ usnic_transport_unrsrv_port(qp_flow->trans_type,
+ qp_flow->usnic_roce.port_num);
+ kfree(qp_flow);
+}
+
+static struct usnic_ib_qp_grp_flow*
+create_and_add_flow(struct usnic_ib_qp_grp *qp_grp,
+ struct usnic_transport_spec *trans_spec)
+{
+ struct usnic_ib_qp_grp_flow *qp_flow;
+ enum usnic_transport_type trans_type;
+
+ trans_type = trans_spec->trans_type;
+ switch (trans_type) {
+ case USNIC_TRANSPORT_ROCE_CUSTOM:
+ qp_flow = create_roce_custom_flow(qp_grp, trans_spec);
+ break;
+ default:
+ usnic_err("Unsupported transport %u\n",
+ trans_spec->trans_type);
+ return ERR_PTR(-EINVAL);
+ }
+
+ if (!IS_ERR_OR_NULL(qp_flow))
+ list_add_tail(&qp_flow->link, &qp_grp->flows_lst);
+
+
+ return qp_flow;
+}
+
+static void release_and_remove_flow(struct usnic_ib_qp_grp_flow *qp_flow)
+{
+ list_del(&qp_flow->link);
+
+ switch (qp_flow->trans_type) {
+ case USNIC_TRANSPORT_ROCE_CUSTOM:
+ release_roce_custom_flow(qp_flow);
+ break;
+ default:
+ WARN(1, "Unsupported transport %u\n",
+ qp_flow->trans_type);
+ break;
+ }
+}
+
+static void release_and_remove_all_flows(struct usnic_ib_qp_grp *qp_grp)
+{
+ struct usnic_ib_qp_grp_flow *qp_flow, *tmp;
+ list_for_each_entry_safe(qp_flow, tmp, &qp_grp->flows_lst, link)
+ release_and_remove_flow(qp_flow);
+}
+
int usnic_ib_qp_grp_modify(struct usnic_ib_qp_grp *qp_grp,
enum ib_qp_state new_state,
- struct usnic_fwd_filter *fwd_filter)
+ void *data)
{
int status = 0;
int vnic_idx;
struct ib_event ib_event;
enum ib_qp_state old_state;
+ struct usnic_transport_spec *trans_spec;
+ struct usnic_ib_qp_grp_flow *qp_flow;
old_state = qp_grp->state;
vnic_idx = usnic_vnic_get_index(qp_grp->vf->vnic);
+ trans_spec = (struct usnic_transport_spec *) data;
spin_lock(&qp_grp->lock);
switch (new_state) {
@@ -236,13 +315,14 @@ int usnic_ib_qp_grp_modify(struct usnic_ib_qp_grp *qp_grp,
/* NO-OP */
break;
case IB_QPS_INIT:
- status = del_all_filters(qp_grp);
+ release_and_remove_all_flows(qp_grp);
+ status = 0;
break;
case IB_QPS_RTR:
case IB_QPS_RTS:
case IB_QPS_ERR:
status = disable_qp_grp(qp_grp);
- status &= del_all_filters(qp_grp);
+ release_and_remove_all_flows(qp_grp);
break;
default:
status = -EINVAL;
@@ -251,10 +331,35 @@ int usnic_ib_qp_grp_modify(struct usnic_ib_qp_grp *qp_grp,
case IB_QPS_INIT:
switch (old_state) {
case IB_QPS_RESET:
- status = add_fwd_filter(qp_grp, fwd_filter);
+ if (trans_spec) {
+ qp_flow = create_and_add_flow(qp_grp,
+ trans_spec);
+ if (IS_ERR_OR_NULL(qp_flow)) {
+ status = (qp_flow) ? PTR_ERR(qp_flow) : -EFAULT;
+ break;
+ }
+ } else {
+ /*
+ * Optional to specify filters.
+ */
+ status = 0;
+ }
break;
case IB_QPS_INIT:
- status = add_fwd_filter(qp_grp, fwd_filter);
+ if (trans_spec) {
+ qp_flow = create_and_add_flow(qp_grp,
+ trans_spec);
+ if (IS_ERR_OR_NULL(qp_flow)) {
+ status = (qp_flow) ? PTR_ERR(qp_flow) : -EFAULT;
+ break;
+ }
+ } else {
+ /*
+ * Doesn't make sense to go into INIT state
+ * from INIT state w/o adding filters.
+ */
+ status = -EINVAL;
+ }
break;
case IB_QPS_RTR:
status = disable_qp_grp(qp_grp);
@@ -295,14 +400,14 @@ int usnic_ib_qp_grp_modify(struct usnic_ib_qp_grp *qp_grp,
qp_grp->ibqp.qp_context);
break;
case IB_QPS_INIT:
- status = del_all_filters(qp_grp);
+ release_and_remove_all_flows(qp_grp);
qp_grp->ibqp.event_handler(&ib_event,
qp_grp->ibqp.qp_context);
break;
case IB_QPS_RTR:
case IB_QPS_RTS:
status = disable_qp_grp(qp_grp);
- status &= del_all_filters(qp_grp);
+ release_and_remove_all_flows(qp_grp);
qp_grp->ibqp.event_handler(&ib_event,
qp_grp->ibqp.qp_context);
break;
@@ -435,16 +540,33 @@ static void log_spec(struct usnic_vnic_res_spec *res_spec)
usnic_dbg("%s\n", buf);
}
+static int qp_grp_id_from_flow(struct usnic_ib_qp_grp_flow *qp_flow,
+ uint32_t *id)
+{
+ enum usnic_transport_type trans_type = qp_flow->trans_type;
+
+ switch (trans_type) {
+ case USNIC_TRANSPORT_ROCE_CUSTOM:
+ *id = qp_flow->usnic_roce.port_num;
+ break;
+ default:
+ usnic_err("Unsupported transport %u\n", trans_type);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
struct usnic_ib_qp_grp *
-usnic_ib_qp_grp_create(struct usnic_fwd_dev *ufdev,
- struct usnic_ib_vf *vf,
+usnic_ib_qp_grp_create(struct usnic_fwd_dev *ufdev, struct usnic_ib_vf *vf,
struct usnic_ib_pd *pd,
struct usnic_vnic_res_spec *res_spec,
- enum usnic_transport_type transport)
+ struct usnic_transport_spec *transport_spec)
{
struct usnic_ib_qp_grp *qp_grp;
- u16 port_num;
int err;
+ enum usnic_transport_type transport = transport_spec->trans_type;
+ struct usnic_ib_qp_grp_flow *qp_flow;
lockdep_assert_held(&vf->lock);
@@ -457,13 +579,6 @@ usnic_ib_qp_grp_create(struct usnic_fwd_dev *ufdev,
return ERR_PTR(err);
}
- port_num = usnic_transport_rsrv_port(transport, 0);
- if (!port_num) {
- usnic_err("Unable to allocate port for %s\n",
- netdev_name(ufdev->netdev));
- return ERR_PTR(-EINVAL);
- }
-
qp_grp = kzalloc(sizeof(*qp_grp), GFP_ATOMIC);
if (!qp_grp) {
usnic_err("Unable to alloc qp_grp - Out of memory\n");
@@ -477,53 +592,59 @@ usnic_ib_qp_grp_create(struct usnic_fwd_dev *ufdev,
PTR_ERR(qp_grp->res_chunk_list) : -ENOMEM;
usnic_err("Unable to alloc res for %d with err %d\n",
qp_grp->grp_id, err);
- goto out_free_port;
+ goto out_free_qp_grp;
}
- INIT_LIST_HEAD(&qp_grp->filter_hndls);
+ err = qp_grp_and_vf_bind(vf, pd, qp_grp);
+ if (err)
+ goto out_free_res;
+
+ INIT_LIST_HEAD(&qp_grp->flows_lst);
spin_lock_init(&qp_grp->lock);
qp_grp->ufdev = ufdev;
- qp_grp->transport = transport;
- qp_grp->filters[DFLT_FILTER_IDX].transport = transport;
- qp_grp->filters[DFLT_FILTER_IDX].port_num = port_num;
qp_grp->state = IB_QPS_RESET;
qp_grp->owner_pid = current->pid;
- /* qp_num is same as default filter port_num */
- qp_grp->ibqp.qp_num = qp_grp->filters[DFLT_FILTER_IDX].port_num;
- qp_grp->grp_id = qp_grp->ibqp.qp_num;
+ qp_flow = create_and_add_flow(qp_grp, transport_spec);
+ if (IS_ERR_OR_NULL(qp_flow)) {
+ usnic_err("Unable to create and add flow with err %ld\n",
+ PTR_ERR(qp_flow));
+ err = (qp_flow) ? PTR_ERR(qp_flow) : -EFAULT;
+ goto out_qp_grp_vf_unbind;
+ }
- err = qp_grp_and_vf_bind(vf, pd, qp_grp);
+ err = qp_grp_id_from_flow(qp_flow, &qp_grp->grp_id);
if (err)
- goto out_free_port;
+ goto out_release_flow;
+ qp_grp->ibqp.qp_num = qp_grp->grp_id;
usnic_ib_sysfs_qpn_add(qp_grp);
return qp_grp;
-out_free_port:
+out_release_flow:
+ release_and_remove_flow(qp_flow);
+out_qp_grp_vf_unbind:
+ qp_grp_and_vf_unbind(qp_grp);
+out_free_res:
+ free_qp_grp_res(qp_grp->res_chunk_list);
+out_free_qp_grp:
kfree(qp_grp);
- usnic_transport_unrsrv_port(transport, port_num);
return ERR_PTR(err);
}
void usnic_ib_qp_grp_destroy(struct usnic_ib_qp_grp *qp_grp)
{
- u16 default_port_num;
- enum usnic_transport_type transport;
WARN_ON(qp_grp->state != IB_QPS_RESET);
lockdep_assert_held(&qp_grp->vf->lock);
- transport = qp_grp->filters[DFLT_FILTER_IDX].transport;
- default_port_num = qp_grp->filters[DFLT_FILTER_IDX].port_num;
-
usnic_ib_sysfs_qpn_remove(qp_grp);
qp_grp_and_vf_unbind(qp_grp);
+ release_and_remove_all_flows(qp_grp);
free_qp_grp_res(qp_grp->res_chunk_list);
kfree(qp_grp);
- usnic_transport_unrsrv_port(transport, default_port_num);
}
struct usnic_vnic_res_chunk*
@@ -26,9 +26,6 @@
#include "usnic_fwd.h"
#include "usnic_vnic.h"
-#define MAX_QP_GRP_FILTERS 10
-#define DFLT_FILTER_IDX 0
-
/*
* The qp group struct represents all the hw resources needed to present a ib_qp
*/
@@ -38,11 +35,8 @@ struct usnic_ib_qp_grp {
int grp_id;
struct usnic_fwd_dev *ufdev;
- short unsigned filter_cnt;
- struct usnic_fwd_filter filters[MAX_QP_GRP_FILTERS];
- struct list_head filter_hndls;
- enum usnic_transport_type transport;
struct usnic_ib_ucontext *ctx;
+ struct list_head flows_lst;
struct usnic_vnic_res_chunk **res_chunk_list;
@@ -55,6 +49,18 @@ struct usnic_ib_qp_grp {
struct kobject kobj;
};
+struct usnic_ib_qp_grp_flow {
+ struct usnic_fwd_flow *flow;
+ enum usnic_transport_type trans_type;
+ union {
+ struct {
+ uint16_t port_num;
+ } usnic_roce;
+ };
+ struct usnic_ib_qp_grp *qp_grp;
+ struct list_head link;
+};
+
static const struct
usnic_vnic_res_spec min_transport_spec[USNIC_TRANSPORT_MAX] = {
{ /*USNIC_TRANSPORT_UNKNOWN*/
@@ -79,11 +85,11 @@ struct usnic_ib_qp_grp *
usnic_ib_qp_grp_create(struct usnic_fwd_dev *ufdev, struct usnic_ib_vf *vf,
struct usnic_ib_pd *pd,
struct usnic_vnic_res_spec *res_spec,
- enum usnic_transport_type transport);
+ struct usnic_transport_spec *trans_spec);
void usnic_ib_qp_grp_destroy(struct usnic_ib_qp_grp *qp_grp);
int usnic_ib_qp_grp_modify(struct usnic_ib_qp_grp *qp_grp,
enum ib_qp_state new_state,
- struct usnic_fwd_filter *fwd_filter);
+ void *data);
struct usnic_vnic_res_chunk
*usnic_ib_qp_grp_get_chunk(struct usnic_ib_qp_grp *qp_grp,
enum usnic_vnic_res_type type);
This patch ports usnic_ib_qp_grp.[hc] to the new interface of usnic_fwd.h. Signed-off-by: Upinder Malhi <umalhi@cisco.com> --- drivers/infiniband/hw/usnic/usnic_ib_qp_grp.c | 347 +++++++++++++++++--------- drivers/infiniband/hw/usnic/usnic_ib_qp_grp.h | 24 +- 2 files changed, 249 insertions(+), 122 deletions(-)