[PATCH net-next v7 0/8] Add Auxiliary driver support

[PATCH net-next v7 0/8] Add Auxiliary driver support

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 3081 bytes --]

Add auxiliary device driver for Broadcom devices.
The bnxt_en driver will register and initialize an aux device
if RDMA is enabled in the underlying device.
The bnxt_re driver will then probe and initialize the
RoCE interfaces with the infiniband stack.

We got rid of the bnxt_en_ops which the bnxt_re driver used to
communicate with bnxt_en.
Similarly  We have tried to clean up most of the bnxt_ulp_ops.
In most of the cases we used the functions and entry points provided
by the auxiliary bus driver framework.
And now these are the minimal functions needed to support the functionality.

We will try to work on getting rid of the remaining if we find any
other viable option in future.

v1->v2:
- Incorporated review comments including usage of ulp_id &
  complex function indirections.
- Used function calls provided by the auxiliary bus interface
  instead of proprietary calls.
- Refactor code to remove ROCE driver's access to bnxt structure.

v2->v3:
- Addressed review comments including cleanup of some unnecessary wrappers
- Fixed warnings seen during cross compilation

v3->v4:
- Cleaned up bnxt_ulp.c and bnxt_ulp.h further
- Removed some more dead code
- Sending the patchset as a standalone series

v4->v5:
- Removed the SRIOV config callback which bnxt_en driver was calling into
  bnxt_re driver.
- Removed excessive checks for rdev and other pointers.

v5->v6:
- Removed excessive checks for dev and other pointers
- Remove runtime interrupt vector allocation. bnxt_en preallocates
interrupt vectors for bnxt_re to use.

v6->v7:
- Removed incorrect usage of inline
- Updated Kconfig to select AUXILIARY BUS support
- Addressed various comments including removal of unnecessary forward
  declaration, using static functions where possible, unnecessary jump,
  cleanup logic, etc..
- Added Leon's Reviewed-by, to the commit log in the patches, from
  previous version.

Please apply. Thanks.

The following are changes since commit 55b98837e37da723c8b73ec0b48fe68c682b57d7
  Merge branch 'vsock-update-tools-and-error-handling'
and are available in the git repository at:
  https://github.com/ajitkhaparde1/net-next/tree/aux-bus-v7


Ajit Khaparde (7):
  bnxt_en: Add auxiliary driver support
  RDMA/bnxt_re: Use auxiliary driver interface
  bnxt_en: Remove usage of ulp_id
  bnxt_en: Use direct API instead of indirection
  bnxt_en: Use auxiliary bus calls over proprietary calls
  RDMA/bnxt_re: Remove the sriov config callback
  bnxt_en: Remove runtime interrupt vector allocation

Hongguang Gao (1):
  bnxt_en: Remove struct bnxt access from RoCE driver

 drivers/infiniband/hw/bnxt_re/bnxt_re.h       |  10 +-
 drivers/infiniband/hw/bnxt_re/main.c          | 635 +++++++-----------
 drivers/net/ethernet/broadcom/bnxt/bnxt.c     |  11 +-
 drivers/net/ethernet/broadcom/bnxt/bnxt.h     |   8 +
 .../net/ethernet/broadcom/bnxt/bnxt_sriov.c   |   7 +-
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 489 ++++++--------
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h |  57 +-
 7 files changed, 494 insertions(+), 723 deletions(-)

-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

[PATCH net-next v7 1/8] bnxt_en: Add auxiliary driver support

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 11242 bytes --]

Add auxiliary driver support.
An auxiliary device will be created if the hardware indicates
support for RDMA.
The bnxt_ulp_probe() function has been removed and a new
bnxt_rdma_aux_device_add() function has been added.
The bnxt_free_msix_vecs() and bnxt_req_msix_vecs() will now hold
the RTNL lock when they call the bnxt_close_nic()and bnxt_open_nic()
since the device close and open need to be protected under RTNL lock.
The operations between the bnxt_en and bnxt_re will be protected
using the en_ops_lock.
This will be used by the bnxt_re driver in a follow-on patch
to create ROCE interfaces.

Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Selvin Xavier <selvin.xavier@broadcom.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/net/ethernet/broadcom/Kconfig         |   1 +
 drivers/net/ethernet/broadcom/bnxt/bnxt.c     |   8 +-
 drivers/net/ethernet/broadcom/bnxt/bnxt.h     |   8 +
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 164 +++++++++++++++---
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h |   4 +-
 5 files changed, 158 insertions(+), 27 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/Kconfig b/drivers/net/ethernet/broadcom/Kconfig
index f4ca0c6c0f51..948586bf1b5b 100644
--- a/drivers/net/ethernet/broadcom/Kconfig
+++ b/drivers/net/ethernet/broadcom/Kconfig
@@ -213,6 +213,7 @@ config BNXT
 	select NET_DEVLINK
 	select PAGE_POOL
 	select DIMLIB
+	select AUXILIARY_BUS
 	help
 	  This driver supports Broadcom NetXtreme-C/E 10/25/40/50 gigabit
 	  Ethernet cards.  To compile this driver as a module, choose M here:
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 16ce7a90610c..c5cbf2307e07 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -13178,6 +13178,9 @@ static void bnxt_remove_one(struct pci_dev *pdev)
 	struct net_device *dev = pci_get_drvdata(pdev);
 	struct bnxt *bp = netdev_priv(dev);
 
+	bnxt_rdma_aux_device_uninit(bp);
+	bnxt_aux_dev_free(bp);
+
 	if (BNXT_PF(bp))
 		bnxt_sriov_disable(bp);
 
@@ -13776,11 +13779,13 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	bnxt_dl_fw_reporters_create(bp);
 
+	bnxt_rdma_aux_device_init(bp);
+
 	bnxt_print_device_info(bp);
 
 	pci_save_state(pdev);
-	return 0;
 
+	return 0;
 init_err_cleanup:
 	bnxt_dl_unregister(bp);
 init_err_dl:
@@ -13824,7 +13829,6 @@ static void bnxt_shutdown(struct pci_dev *pdev)
 	if (netif_running(dev))
 		dev_close(dev);
 
-	bnxt_ulp_shutdown(bp);
 	bnxt_clear_int_mode(bp);
 	pci_disable_device(pdev);
 
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 5163ef4a49ea..4df2da81708e 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -24,6 +24,7 @@
 #include <linux/interrupt.h>
 #include <linux/rhashtable.h>
 #include <linux/crash_dump.h>
+#include <linux/auxiliary_bus.h>
 #include <net/devlink.h>
 #include <net/dst_metadata.h>
 #include <net/xdp.h>
@@ -1631,6 +1632,12 @@ struct bnxt_fw_health {
 #define BNXT_FW_IF_RETRY		10
 #define BNXT_FW_SLOT_RESET_RETRY	4
 
+struct bnxt_aux_dev {
+	struct auxiliary_device aux_dev;
+	struct bnxt_en_dev *edev;
+	int id;
+};
+
 enum board_idx {
 	BCM57301,
 	BCM57302,
@@ -1852,6 +1859,7 @@ struct bnxt {
 #define BNXT_CHIP_P4_PLUS(bp)			\
 	(BNXT_CHIP_P4(bp) || BNXT_CHIP_P5(bp))
 
+	struct bnxt_aux_dev	*aux_dev;
 	struct bnxt_en_dev	*edev;
 
 	struct bnxt_napi	**bnapi;
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index 2e54bf4fc7a7..899d9adfd35f 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -25,6 +25,8 @@
 #include "bnxt_hwrm.h"
 #include "bnxt_ulp.h"
 
+static DEFINE_IDA(bnxt_aux_dev_ids);
+
 static int bnxt_register_dev(struct bnxt_en_dev *edev, unsigned int ulp_id,
 			     struct bnxt_ulp_ops *ulp_ops, void *handle)
 {
@@ -32,7 +34,6 @@ static int bnxt_register_dev(struct bnxt_en_dev *edev, unsigned int ulp_id,
 	struct bnxt *bp = netdev_priv(dev);
 	struct bnxt_ulp *ulp;
 
-	ASSERT_RTNL();
 	if (ulp_id >= BNXT_MAX_ULP)
 		return -EINVAL;
 
@@ -50,7 +51,7 @@ static int bnxt_register_dev(struct bnxt_en_dev *edev, unsigned int ulp_id,
 			return -ENOMEM;
 	}
 
-	atomic_set(&ulp->ref_count, 0);
+	atomic_set(&ulp->ref_count, 1);
 	ulp->handle = handle;
 	rcu_assign_pointer(ulp->ulp_ops, ulp_ops);
 
@@ -69,10 +70,11 @@ static int bnxt_unregister_dev(struct bnxt_en_dev *edev, unsigned int ulp_id)
 	struct bnxt_ulp *ulp;
 	int i = 0;
 
-	ASSERT_RTNL();
 	if (ulp_id >= BNXT_MAX_ULP)
 		return -EINVAL;
 
+	edev->flags |= BNXT_EN_FLAG_ULP_STOPPED;
+
 	ulp = &edev->ulp_tbl[ulp_id];
 	if (!rcu_access_pointer(ulp->ulp_ops)) {
 		netdev_err(bp->dev, "ulp id %d not registered\n", ulp_id);
@@ -126,7 +128,6 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id,
 	int total_vecs;
 	int rc = 0;
 
-	ASSERT_RTNL();
 	if (ulp_id != BNXT_ROCE_ULP)
 		return -EINVAL;
 
@@ -149,10 +150,12 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id,
 		max_idx = min_t(int, bp->total_irqs, max_cp_rings);
 		idx = max_idx - avail_msix;
 	}
+
 	edev->ulp_tbl[ulp_id].msix_base = idx;
 	edev->ulp_tbl[ulp_id].msix_requested = avail_msix;
 	hw_resc = &bp->hw_resc;
 	total_vecs = idx + avail_msix;
+	rtnl_lock();
 	if (bp->total_irqs < total_vecs ||
 	    (BNXT_NEW_RM(bp) && hw_resc->resv_irqs < total_vecs)) {
 		if (netif_running(dev)) {
@@ -162,6 +165,7 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id,
 			rc = bnxt_reserve_rings(bp, true);
 		}
 	}
+	rtnl_unlock();
 	if (rc) {
 		edev->ulp_tbl[ulp_id].msix_requested = 0;
 		return -EAGAIN;
@@ -184,7 +188,6 @@ static int bnxt_free_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id)
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
 
-	ASSERT_RTNL();
 	if (ulp_id != BNXT_ROCE_ULP)
 		return -EINVAL;
 
@@ -193,10 +196,13 @@ static int bnxt_free_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id)
 
 	edev->ulp_tbl[ulp_id].msix_requested = 0;
 	edev->flags &= ~BNXT_EN_FLAG_MSIX_REQUESTED;
+	rtnl_lock();
 	if (netif_running(dev) && !(edev->flags & BNXT_EN_FLAG_ULP_STOPPED)) {
 		bnxt_close_nic(bp, true, false);
 		bnxt_open_nic(bp, true, false);
 	}
+	rtnl_unlock();
+
 	return 0;
 }
 
@@ -347,25 +353,6 @@ void bnxt_ulp_sriov_cfg(struct bnxt *bp, int num_vfs)
 	}
 }
 
-void bnxt_ulp_shutdown(struct bnxt *bp)
-{
-	struct bnxt_en_dev *edev = bp->edev;
-	struct bnxt_ulp_ops *ops;
-	int i;
-
-	if (!edev)
-		return;
-
-	for (i = 0; i < BNXT_MAX_ULP; i++) {
-		struct bnxt_ulp *ulp = &edev->ulp_tbl[i];
-
-		ops = rtnl_dereference(ulp->ulp_ops);
-		if (!ops || !ops->ulp_shutdown)
-			continue;
-		ops->ulp_shutdown(ulp->handle);
-	}
-}
-
 void bnxt_ulp_irq_stop(struct bnxt *bp)
 {
 	struct bnxt_en_dev *edev = bp->edev;
@@ -475,6 +462,135 @@ static const struct bnxt_en_ops bnxt_en_ops_tbl = {
 	.bnxt_register_fw_async_events	= bnxt_register_async_events,
 };
 
+void bnxt_aux_dev_free(struct bnxt *bp)
+{
+	kfree(bp->aux_dev);
+	bp->aux_dev = NULL;
+}
+
+static struct bnxt_aux_dev *bnxt_aux_dev_alloc(struct bnxt *bp)
+{
+	return kzalloc(sizeof(struct bnxt_aux_dev), GFP_KERNEL);
+}
+
+void bnxt_rdma_aux_device_uninit(struct bnxt *bp)
+{
+	struct bnxt_aux_dev *bnxt_adev;
+	struct auxiliary_device *adev;
+
+	/* Skip if no auxiliary device init was done. */
+	if (!(bp->flags & BNXT_FLAG_ROCE_CAP))
+		return;
+
+	bnxt_adev = bp->aux_dev;
+	adev = &bnxt_adev->aux_dev;
+	auxiliary_device_delete(adev);
+	auxiliary_device_uninit(adev);
+	if (bnxt_adev->id >= 0)
+		ida_free(&bnxt_aux_dev_ids, bnxt_adev->id);
+}
+
+static void bnxt_aux_dev_release(struct device *dev)
+{
+	struct bnxt_aux_dev *bnxt_adev =
+		container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
+	struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
+
+	bnxt_adev->edev->en_ops = NULL;
+	kfree(bnxt_adev->edev);
+	bnxt_adev->edev = NULL;
+	bp->edev = NULL;
+}
+
+static void bnxt_set_edev_info(struct bnxt_en_dev *edev, struct bnxt *bp)
+{
+	edev->en_ops = &bnxt_en_ops_tbl;
+	edev->net = bp->dev;
+	edev->pdev = bp->pdev;
+	edev->l2_db_size = bp->db_size;
+	edev->l2_db_size_nc = bp->db_size;
+
+	if (bp->flags & BNXT_FLAG_ROCEV1_CAP)
+		edev->flags |= BNXT_EN_FLAG_ROCEV1_CAP;
+	if (bp->flags & BNXT_FLAG_ROCEV2_CAP)
+		edev->flags |= BNXT_EN_FLAG_ROCEV2_CAP;
+}
+
+static int bnxt_rdma_aux_device_add(struct bnxt *bp)
+{
+	struct bnxt_aux_dev *bnxt_adev = bp->aux_dev;
+	struct bnxt_en_dev *edev = bnxt_adev->edev;
+	struct auxiliary_device *aux_dev;
+	int ret;
+
+	edev = kzalloc(sizeof(*edev), GFP_KERNEL);
+	if (!edev)
+		return -ENOMEM;
+
+	aux_dev = &bnxt_adev->aux_dev;
+	aux_dev->id = bnxt_adev->id;
+	aux_dev->name = "rdma";
+	aux_dev->dev.parent = &bp->pdev->dev;
+	aux_dev->dev.release = bnxt_aux_dev_release;
+
+	bnxt_adev->edev = edev;
+	bp->edev = edev;
+	bnxt_set_edev_info(edev, bp);
+
+	ret = auxiliary_device_init(aux_dev);
+	if (ret)
+		goto free_edev;
+
+	ret = auxiliary_device_add(aux_dev);
+	if (ret)
+		goto aux_dev_uninit;
+
+	return 0;
+aux_dev_uninit:
+	auxiliary_device_uninit(aux_dev);
+free_edev:
+	kfree(edev);
+	bp->edev = NULL;
+
+	return ret;
+}
+
+void bnxt_rdma_aux_device_init(struct bnxt *bp)
+{
+	int rc;
+
+	if (!(bp->flags & BNXT_FLAG_ROCE_CAP))
+		return;
+
+	bp->aux_dev = bnxt_aux_dev_alloc(bp);
+	if (!bp->aux_dev)
+		goto skip_ida_init;
+
+	bp->aux_dev->id = ida_alloc(&bnxt_aux_dev_ids, GFP_KERNEL);
+	if (bp->aux_dev->id < 0) {
+		netdev_warn(bp->dev,
+			    "ida alloc failed for ROCE auxiliary device\n");
+		goto skip_aux_init;
+	}
+
+	/* If aux bus init fails, continue with netdev init. */
+	rc = bnxt_rdma_aux_device_add(bp);
+	if (rc) {
+		netdev_warn(bp->dev,
+			    "Failed to add auxiliary device for ROCE\n");
+		goto aux_add_failed;
+	}
+	return;
+
+aux_add_failed:
+	ida_free(&bnxt_aux_dev_ids, bp->aux_dev->id);
+	bp->aux_dev->id = -1;
+skip_aux_init:
+	bnxt_aux_dev_free(bp);
+skip_ida_init:
+	bp->flags &= ~BNXT_FLAG_ROCE_CAP;
+}
+
 struct bnxt_en_dev *bnxt_ulp_probe(struct net_device *dev)
 {
 	struct bnxt *bp = netdev_priv(dev);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
index 42b50abc3e91..47c7131e5549 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
@@ -102,10 +102,12 @@ int bnxt_get_ulp_stat_ctxs(struct bnxt *bp);
 void bnxt_ulp_stop(struct bnxt *bp);
 void bnxt_ulp_start(struct bnxt *bp, int err);
 void bnxt_ulp_sriov_cfg(struct bnxt *bp, int num_vfs);
-void bnxt_ulp_shutdown(struct bnxt *bp);
 void bnxt_ulp_irq_stop(struct bnxt *bp);
 void bnxt_ulp_irq_restart(struct bnxt *bp, int err);
 void bnxt_ulp_async_events(struct bnxt *bp, struct hwrm_async_event_cmpl *cmpl);
+void bnxt_rdma_aux_device_uninit(struct bnxt *bp);
+void bnxt_rdma_aux_device_init(struct bnxt *bp);
+void bnxt_aux_dev_free(struct bnxt *bp);
 struct bnxt_en_dev *bnxt_ulp_probe(struct net_device *dev);
 
 #endif
-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

[PATCH net-next v7 2/8] RDMA/bnxt_re: Use auxiliary driver interface

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 20171 bytes --]

Use auxiliary driver interface for driver load, unload ROCE driver.
The driver does not need to register the interface using the netdev
notifier anymore. Removed the bnxt_re_dev_list which is not needed.
Currently probe, remove and shutdown ops have been implemented for
the auxiliary device.
Also remove exccessve validation checks for rdev.

Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Selvin Xavier <selvin.xavier@broadcom.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/bnxt_re/bnxt_re.h       |   9 +-
 drivers/infiniband/hw/bnxt_re/main.c          | 395 ++++++------------
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c |  29 +-
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h |   3 -
 4 files changed, 132 insertions(+), 304 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/bnxt_re.h b/drivers/infiniband/hw/bnxt_re/bnxt_re.h
index 785c37cae3c0..b0465c8d229a 100644
--- a/drivers/infiniband/hw/bnxt_re/bnxt_re.h
+++ b/drivers/infiniband/hw/bnxt_re/bnxt_re.h
@@ -89,13 +89,6 @@ struct bnxt_re_ring_attr {
 	u8		mode;
 };
 
-struct bnxt_re_work {
-	struct work_struct	work;
-	unsigned long		event;
-	struct bnxt_re_dev      *rdev;
-	struct net_device	*vlan_dev;
-};
-
 struct bnxt_re_sqp_entries {
 	struct bnxt_qplib_sge sge;
 	u64 wrid;
@@ -132,6 +125,7 @@ struct bnxt_re_dev {
 #define BNXT_RE_FLAG_ERR_DEVICE_DETACHED       17
 #define BNXT_RE_FLAG_ISSUE_ROCE_STATS          29
 	struct net_device		*netdev;
+	struct notifier_block		nb;
 	unsigned int			version, major, minor;
 	struct bnxt_qplib_chip_ctx	*chip_ctx;
 	struct bnxt_en_dev		*en_dev;
@@ -194,5 +188,4 @@ static inline struct device *rdev_to_dev(struct bnxt_re_dev *rdev)
 		return  &rdev->ibdev.dev;
 	return NULL;
 }
-
 #endif
diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index 8c0c80a8d338..165ae11b927d 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -48,6 +48,7 @@
 #include <net/ipv6.h>
 #include <net/addrconf.h>
 #include <linux/if_ether.h>
+#include <linux/auxiliary_bus.h>
 
 #include <rdma/ib_verbs.h>
 #include <rdma/ib_user_verbs.h>
@@ -74,14 +75,14 @@ MODULE_DESCRIPTION(BNXT_RE_DESC " Driver");
 MODULE_LICENSE("Dual BSD/GPL");
 
 /* globals */
-static struct list_head bnxt_re_dev_list = LIST_HEAD_INIT(bnxt_re_dev_list);
-/* Mutex to protect the list of bnxt_re devices added */
-static DEFINE_MUTEX(bnxt_re_dev_lock);
-static struct workqueue_struct *bnxt_re_wq;
-static void bnxt_re_remove_device(struct bnxt_re_dev *rdev);
-static void bnxt_re_dealloc_driver(struct ib_device *ib_dev);
+static DEFINE_MUTEX(bnxt_re_mutex);
+
 static void bnxt_re_stop_irq(void *handle);
 static void bnxt_re_dev_stop(struct bnxt_re_dev *rdev);
+static int bnxt_re_netdev_event(struct notifier_block *notifier,
+				unsigned long event, void *ptr);
+static struct bnxt_re_dev *bnxt_re_from_netdev(struct net_device *netdev);
+static void bnxt_re_dev_uninit(struct bnxt_re_dev *rdev);
 
 static void bnxt_re_set_drv_mode(struct bnxt_re_dev *rdev, u8 mode)
 {
@@ -233,7 +234,6 @@ static void bnxt_re_stop(void *p)
 
 	if (!rdev)
 		return;
-	ASSERT_RTNL();
 
 	/* L2 driver invokes this callback during device error/crash or device
 	 * reset. Current RoCE driver doesn't recover the device in case of
@@ -269,9 +269,6 @@ static void bnxt_re_sriov_config(void *p, int num_vfs)
 {
 	struct bnxt_re_dev *rdev = p;
 
-	if (!rdev)
-		return;
-
 	if (test_bit(BNXT_RE_FLAG_ERR_DEVICE_DETACHED, &rdev->flags))
 		return;
 	rdev->num_vfs = num_vfs;
@@ -282,16 +279,14 @@ static void bnxt_re_sriov_config(void *p, int num_vfs)
 	}
 }
 
-static void bnxt_re_shutdown(void *p)
+static void bnxt_re_shutdown(struct auxiliary_device *adev)
 {
-	struct bnxt_re_dev *rdev = p;
+	struct bnxt_re_dev *rdev = auxiliary_get_drvdata(adev);
 
 	if (!rdev)
 		return;
-	ASSERT_RTNL();
-	/* Release the MSIx vectors before queuing unregister */
-	bnxt_re_stop_irq(rdev);
-	ib_unregister_device_queued(&rdev->ibdev);
+	ib_unregister_device(&rdev->ibdev);
+	bnxt_re_dev_uninit(rdev);
 }
 
 static void bnxt_re_stop_irq(void *handle)
@@ -346,11 +341,9 @@ static void bnxt_re_start_irq(void *handle, struct bnxt_msix_entry *ent)
 }
 
 static struct bnxt_ulp_ops bnxt_re_ulp_ops = {
-	.ulp_async_notifier = NULL,
 	.ulp_stop = bnxt_re_stop,
 	.ulp_start = bnxt_re_start,
 	.ulp_sriov_config = bnxt_re_sriov_config,
-	.ulp_shutdown = bnxt_re_shutdown,
 	.ulp_irq_stop = bnxt_re_stop_irq,
 	.ulp_irq_restart = bnxt_re_start_irq
 };
@@ -380,9 +373,6 @@ static int bnxt_re_register_netdev(struct bnxt_re_dev *rdev)
 	struct bnxt_en_dev *en_dev;
 	int rc = 0;
 
-	if (!rdev)
-		return -EINVAL;
-
 	en_dev = rdev->en_dev;
 
 	rc = en_dev->en_ops->bnxt_register_device(en_dev, BNXT_ROCE_ULP,
@@ -401,7 +391,6 @@ static int bnxt_re_free_msix(struct bnxt_re_dev *rdev)
 
 	en_dev = rdev->en_dev;
 
-
 	rc = en_dev->en_ops->bnxt_free_msix(rdev->en_dev, BNXT_ROCE_ULP);
 
 	return rc;
@@ -412,9 +401,6 @@ static int bnxt_re_request_msix(struct bnxt_re_dev *rdev)
 	int rc = 0, num_msix_want = BNXT_RE_MAX_MSIX, num_msix_got;
 	struct bnxt_en_dev *en_dev;
 
-	if (!rdev)
-		return -EINVAL;
-
 	en_dev = rdev->en_dev;
 
 	num_msix_want = min_t(u32, BNXT_RE_MAX_MSIX, num_online_cpus());
@@ -458,12 +444,17 @@ static void bnxt_re_fill_fw_msg(struct bnxt_fw_msg *fw_msg, void *msg,
 static int bnxt_re_net_ring_free(struct bnxt_re_dev *rdev,
 				 u16 fw_ring_id, int type)
 {
-	struct bnxt_en_dev *en_dev = rdev->en_dev;
+	struct bnxt_en_dev *en_dev;
 	struct hwrm_ring_free_input req = {0};
 	struct hwrm_ring_free_output resp;
 	struct bnxt_fw_msg fw_msg;
 	int rc = -EINVAL;
 
+	if (!rdev)
+		return rc;
+
+	en_dev = rdev->en_dev;
+
 	if (!en_dev)
 		return rc;
 
@@ -584,21 +575,6 @@ static int bnxt_re_net_stats_ctx_alloc(struct bnxt_re_dev *rdev,
 
 /* Device */
 
-static bool is_bnxt_re_dev(struct net_device *netdev)
-{
-	struct ethtool_drvinfo drvinfo;
-
-	if (netdev->ethtool_ops && netdev->ethtool_ops->get_drvinfo) {
-		memset(&drvinfo, 0, sizeof(drvinfo));
-		netdev->ethtool_ops->get_drvinfo(netdev, &drvinfo);
-
-		if (strcmp(drvinfo.driver, "bnxt_en"))
-			return false;
-		return true;
-	}
-	return false;
-}
-
 static struct bnxt_re_dev *bnxt_re_from_netdev(struct net_device *netdev)
 {
 	struct ib_device *ibdev =
@@ -609,31 +585,6 @@ static struct bnxt_re_dev *bnxt_re_from_netdev(struct net_device *netdev)
 	return container_of(ibdev, struct bnxt_re_dev, ibdev);
 }
 
-static struct bnxt_en_dev *bnxt_re_dev_probe(struct net_device *netdev)
-{
-	struct bnxt_en_dev *en_dev;
-	struct pci_dev *pdev;
-
-	en_dev = bnxt_ulp_probe(netdev);
-	if (IS_ERR(en_dev))
-		return en_dev;
-
-	pdev = en_dev->pdev;
-	if (!pdev)
-		return ERR_PTR(-EINVAL);
-
-	if (!(en_dev->flags & BNXT_EN_FLAG_ROCE_CAP)) {
-		dev_info(&pdev->dev,
-			"%s: probe error: RoCE is not supported on this device",
-			ROCE_DRV_MODULE_NAME);
-		return ERR_PTR(-ENODEV);
-	}
-
-	dev_hold(netdev);
-
-	return en_dev;
-}
-
 static ssize_t hw_rev_show(struct device *device, struct device_attribute *attr,
 			   char *buf)
 {
@@ -679,7 +630,6 @@ static const struct ib_device_ops bnxt_re_dev_ops = {
 	.create_qp = bnxt_re_create_qp,
 	.create_srq = bnxt_re_create_srq,
 	.create_user_ah = bnxt_re_create_ah,
-	.dealloc_driver = bnxt_re_dealloc_driver,
 	.dealloc_pd = bnxt_re_dealloc_pd,
 	.dealloc_ucontext = bnxt_re_dealloc_ucontext,
 	.del_gid = bnxt_re_del_gid,
@@ -744,18 +694,7 @@ static int bnxt_re_register_ib(struct bnxt_re_dev *rdev)
 	return ib_register_device(ibdev, "bnxt_re%d", &rdev->en_dev->pdev->dev);
 }
 
-static void bnxt_re_dev_remove(struct bnxt_re_dev *rdev)
-{
-	dev_put(rdev->netdev);
-	rdev->netdev = NULL;
-	mutex_lock(&bnxt_re_dev_lock);
-	list_del_rcu(&rdev->list);
-	mutex_unlock(&bnxt_re_dev_lock);
-
-	synchronize_rcu();
-}
-
-static struct bnxt_re_dev *bnxt_re_dev_add(struct net_device *netdev,
+static struct bnxt_re_dev *bnxt_re_dev_add(struct bnxt_aux_dev *aux_dev,
 					   struct bnxt_en_dev *en_dev)
 {
 	struct bnxt_re_dev *rdev;
@@ -768,8 +707,8 @@ static struct bnxt_re_dev *bnxt_re_dev_add(struct net_device *netdev,
 		return NULL;
 	}
 	/* Default values */
-	rdev->netdev = netdev;
-	dev_hold(rdev->netdev);
+	rdev->nb.notifier_call = NULL;
+	rdev->netdev = en_dev->net;
 	rdev->en_dev = en_dev;
 	rdev->id = rdev->en_dev->pdev->devfn;
 	INIT_LIST_HEAD(&rdev->qp_list);
@@ -784,9 +723,6 @@ static struct bnxt_re_dev *bnxt_re_dev_add(struct net_device *netdev,
 	rdev->cosq[0] = 0xFFFF;
 	rdev->cosq[1] = 0xFFFF;
 
-	mutex_lock(&bnxt_re_dev_lock);
-	list_add_tail_rcu(&rdev->list, &bnxt_re_dev_list);
-	mutex_unlock(&bnxt_re_dev_lock);
 	return rdev;
 }
 
@@ -1323,7 +1259,7 @@ static int bnxt_re_ib_init(struct bnxt_re_dev *rdev)
 		pr_err("Failed to register with IB: %#x\n", rc);
 		return rc;
 	}
-	dev_info(rdev_to_dev(rdev), "Device registered successfully");
+	dev_info(rdev_to_dev(rdev), "Device registered with IB successfully");
 	ib_get_eth_speed(&rdev->ibdev, 1, &rdev->active_speed,
 			 &rdev->active_width);
 	set_bit(BNXT_RE_FLAG_ISSUE_ROCE_STATS, &rdev->flags);
@@ -1541,135 +1477,43 @@ static int bnxt_re_dev_init(struct bnxt_re_dev *rdev, u8 wqe_mode)
 	return rc;
 }
 
-static void bnxt_re_dev_unreg(struct bnxt_re_dev *rdev)
-{
-	struct net_device *netdev = rdev->netdev;
-
-	bnxt_re_dev_remove(rdev);
-
-	if (netdev)
-		dev_put(netdev);
-}
-
-static int bnxt_re_dev_reg(struct bnxt_re_dev **rdev, struct net_device *netdev)
+static int bnxt_re_add_device(struct auxiliary_device *adev, u8 wqe_mode)
 {
+	struct bnxt_aux_dev *aux_dev =
+		container_of(adev, struct bnxt_aux_dev, aux_dev);
 	struct bnxt_en_dev *en_dev;
+	struct bnxt_re_dev *rdev;
 	int rc = 0;
 
-	if (!is_bnxt_re_dev(netdev))
-		return -ENODEV;
+	/* en_dev should never be NULL as long as adev and aux_dev are valid. */
+	en_dev = aux_dev->edev;
 
-	en_dev = bnxt_re_dev_probe(netdev);
-	if (IS_ERR(en_dev)) {
-		if (en_dev != ERR_PTR(-ENODEV))
-			ibdev_err(&(*rdev)->ibdev, "%s: Failed to probe\n",
-				  ROCE_DRV_MODULE_NAME);
-		rc = PTR_ERR(en_dev);
-		goto exit;
-	}
-	*rdev = bnxt_re_dev_add(netdev, en_dev);
-	if (!*rdev) {
+	rdev = bnxt_re_dev_add(aux_dev, en_dev);
+	if (!rdev || !rdev_to_dev(rdev)) {
 		rc = -ENOMEM;
-		dev_put(netdev);
 		goto exit;
 	}
-exit:
-	return rc;
-}
-
-static void bnxt_re_remove_device(struct bnxt_re_dev *rdev)
-{
-	bnxt_re_dev_uninit(rdev);
-	pci_dev_put(rdev->en_dev->pdev);
-	bnxt_re_dev_unreg(rdev);
-}
-
-static int bnxt_re_add_device(struct bnxt_re_dev **rdev,
-			      struct net_device *netdev, u8 wqe_mode)
-{
-	int rc;
 
-	rc = bnxt_re_dev_reg(rdev, netdev);
-	if (rc == -ENODEV)
-		return rc;
-	if (rc) {
-		pr_err("Failed to register with the device %s: %#x\n",
-		       netdev->name, rc);
-		return rc;
-	}
+	rc = bnxt_re_dev_init(rdev, wqe_mode);
+	if (rc)
+		goto re_dev_dealloc;
 
-	pci_dev_get((*rdev)->en_dev->pdev);
-	rc = bnxt_re_dev_init(*rdev, wqe_mode);
+	rc = bnxt_re_ib_init(rdev);
 	if (rc) {
-		pci_dev_put((*rdev)->en_dev->pdev);
-		bnxt_re_dev_unreg(*rdev);
+		pr_err("Failed to register with IB: %s",
+			aux_dev->aux_dev.name);
+		goto re_dev_uninit;
 	}
+	auxiliary_set_drvdata(adev, rdev);
 
-	return rc;
-}
-
-static void bnxt_re_dealloc_driver(struct ib_device *ib_dev)
-{
-	struct bnxt_re_dev *rdev =
-		container_of(ib_dev, struct bnxt_re_dev, ibdev);
-
-	dev_info(rdev_to_dev(rdev), "Unregistering Device");
-
-	rtnl_lock();
-	bnxt_re_remove_device(rdev);
-	rtnl_unlock();
-}
-
-/* Handle all deferred netevents tasks */
-static void bnxt_re_task(struct work_struct *work)
-{
-	struct bnxt_re_work *re_work;
-	struct bnxt_re_dev *rdev;
-	int rc = 0;
-
-	re_work = container_of(work, struct bnxt_re_work, work);
-	rdev = re_work->rdev;
-
-	if (re_work->event == NETDEV_REGISTER) {
-		rc = bnxt_re_ib_init(rdev);
-		if (rc) {
-			ibdev_err(&rdev->ibdev,
-				  "Failed to register with IB: %#x", rc);
-			rtnl_lock();
-			bnxt_re_remove_device(rdev);
-			rtnl_unlock();
-			goto exit;
-		}
-		goto exit;
-	}
-
-	if (!ib_device_try_get(&rdev->ibdev))
-		goto exit;
+	return 0;
 
-	switch (re_work->event) {
-	case NETDEV_UP:
-		bnxt_re_dispatch_event(&rdev->ibdev, NULL, 1,
-				       IB_EVENT_PORT_ACTIVE);
-		break;
-	case NETDEV_DOWN:
-		bnxt_re_dev_stop(rdev);
-		break;
-	case NETDEV_CHANGE:
-		if (!netif_carrier_ok(rdev->netdev))
-			bnxt_re_dev_stop(rdev);
-		else if (netif_carrier_ok(rdev->netdev))
-			bnxt_re_dispatch_event(&rdev->ibdev, NULL, 1,
-					       IB_EVENT_PORT_ACTIVE);
-		ib_get_eth_speed(&rdev->ibdev, 1, &rdev->active_speed,
-				 &rdev->active_width);
-		break;
-	default:
-		break;
-	}
-	ib_device_put(&rdev->ibdev);
+re_dev_uninit:
+	bnxt_re_dev_uninit(rdev);
+re_dev_dealloc:
+	ib_dealloc_device(&rdev->ibdev);
 exit:
-	put_device(&rdev->ibdev.dev);
-	kfree(re_work);
+	return rc;
 }
 
 /*
@@ -1690,109 +1534,130 @@ static int bnxt_re_netdev_event(struct notifier_block *notifier,
 				unsigned long event, void *ptr)
 {
 	struct net_device *real_dev, *netdev = netdev_notifier_info_to_dev(ptr);
-	struct bnxt_re_work *re_work;
 	struct bnxt_re_dev *rdev;
-	int rc = 0;
-	bool sch_work = false;
-	bool release = true;
 
 	real_dev = rdma_vlan_dev_real_dev(netdev);
 	if (!real_dev)
 		real_dev = netdev;
 
-	rdev = bnxt_re_from_netdev(real_dev);
-	if (!rdev && event != NETDEV_REGISTER)
-		return NOTIFY_OK;
-
 	if (real_dev != netdev)
 		goto exit;
 
-	switch (event) {
-	case NETDEV_REGISTER:
-		if (rdev)
-			break;
-		rc = bnxt_re_add_device(&rdev, real_dev,
-					BNXT_QPLIB_WQE_MODE_STATIC);
-		if (!rc)
-			sch_work = true;
-		release = false;
-		break;
+	rdev = bnxt_re_from_netdev(real_dev);
+	if (!rdev)
+		return NOTIFY_DONE;
 
-	case NETDEV_UNREGISTER:
-		ib_unregister_device_queued(&rdev->ibdev);
-		break;
 
+	switch (event) {
+	case NETDEV_UP:
+	case NETDEV_DOWN:
+	case NETDEV_CHANGE:
+		bnxt_re_dispatch_event(&rdev->ibdev, NULL, 1,
+					netif_carrier_ok(real_dev) ?
+					IB_EVENT_PORT_ACTIVE :
+					IB_EVENT_PORT_ERR);
+		break;
 	default:
-		sch_work = true;
 		break;
 	}
-	if (sch_work) {
-		/* Allocate for the deferred task */
-		re_work = kzalloc(sizeof(*re_work), GFP_KERNEL);
-		if (re_work) {
-			get_device(&rdev->ibdev.dev);
-			re_work->rdev = rdev;
-			re_work->event = event;
-			re_work->vlan_dev = (real_dev == netdev ?
-					     NULL : netdev);
-			INIT_WORK(&re_work->work, bnxt_re_task);
-			queue_work(bnxt_re_wq, &re_work->work);
-		}
-	}
-
+	ib_device_put(&rdev->ibdev);
 exit:
-	if (rdev && release)
-		ib_device_put(&rdev->ibdev);
 	return NOTIFY_DONE;
 }
 
-static struct notifier_block bnxt_re_netdev_notifier = {
-	.notifier_call = bnxt_re_netdev_event
-};
+#define BNXT_ADEV_NAME "bnxt_en"
 
-static int __init bnxt_re_mod_init(void)
+static void bnxt_re_remove(struct auxiliary_device *adev)
 {
-	int rc = 0;
+	struct bnxt_re_dev *rdev = auxiliary_get_drvdata(adev);
 
-	pr_info("%s: %s", ROCE_DRV_MODULE_NAME, version);
+	if (!rdev)
+		return;
 
-	bnxt_re_wq = create_singlethread_workqueue("bnxt_re");
-	if (!bnxt_re_wq)
-		return -ENOMEM;
+	mutex_lock(&bnxt_re_mutex);
+	if (rdev->nb.notifier_call) {
+		unregister_netdevice_notifier(&rdev->nb);
+		rdev->nb.notifier_call = NULL;
+	} else {
+		/* If notifier is null, we should have already done a
+		 * clean up before coming here.
+		 */
+		goto skip_remove;
+	}
+
+	ib_unregister_device(&rdev->ibdev);
+	ib_dealloc_device(&rdev->ibdev);
+	bnxt_re_dev_uninit(rdev);
+skip_remove:
+	mutex_unlock(&bnxt_re_mutex);
+}
+
+static int bnxt_re_probe(struct auxiliary_device *adev,
+			 const struct auxiliary_device_id *id)
+{
+	struct bnxt_re_dev *rdev;
+	int rc;
+
+	mutex_lock(&bnxt_re_mutex);
+	rc = bnxt_re_add_device(adev, BNXT_QPLIB_WQE_MODE_STATIC);
+	if (rc) {
+		mutex_unlock(&bnxt_re_mutex);
+		return rc;
+	}
 
-	INIT_LIST_HEAD(&bnxt_re_dev_list);
+	rdev = auxiliary_get_drvdata(adev);
 
-	rc = register_netdevice_notifier(&bnxt_re_netdev_notifier);
+	rdev->nb.notifier_call = bnxt_re_netdev_event;
+	rc = register_netdevice_notifier(&rdev->nb);
 	if (rc) {
+		rdev->nb.notifier_call = NULL;
 		pr_err("%s: Cannot register to netdevice_notifier",
 		       ROCE_DRV_MODULE_NAME);
-		goto err_netdev;
+		goto err;
 	}
+
+	mutex_unlock(&bnxt_re_mutex);
 	return 0;
 
-err_netdev:
-	destroy_workqueue(bnxt_re_wq);
+err:
+	mutex_unlock(&bnxt_re_mutex);
+	bnxt_re_remove(adev);
 
 	return rc;
 }
 
-static void __exit bnxt_re_mod_exit(void)
+static const struct auxiliary_device_id bnxt_re_id_table[] = {
+	{ .name = BNXT_ADEV_NAME ".rdma", },
+	{},
+};
+
+MODULE_DEVICE_TABLE(auxiliary, bnxt_re_id_table);
+
+static struct auxiliary_driver bnxt_re_driver = {
+	.name = "rdma",
+	.probe = bnxt_re_probe,
+	.remove = bnxt_re_remove,
+	.shutdown = bnxt_re_shutdown,
+	.id_table = bnxt_re_id_table,
+};
+
+static int __init bnxt_re_mod_init(void)
 {
-	struct bnxt_re_dev *rdev;
+	int rc = 0;
 
-	unregister_netdevice_notifier(&bnxt_re_netdev_notifier);
-	if (bnxt_re_wq)
-		destroy_workqueue(bnxt_re_wq);
-	list_for_each_entry(rdev, &bnxt_re_dev_list, list) {
-		/* VF device removal should be called before the removal
-		 * of PF device. Queue VFs unregister first, so that VFs
-		 * shall be removed before the PF during the call of
-		 * ib_unregister_driver.
-		 */
-		if (rdev->is_virtfn)
-			ib_unregister_device(&rdev->ibdev);
+	pr_info("%s: %s", ROCE_DRV_MODULE_NAME, version);
+	rc = auxiliary_driver_register(&bnxt_re_driver);
+	if (rc) {
+		pr_err("%s: Failed to register auxiliary driver\n",
+			ROCE_DRV_MODULE_NAME);
+		return rc;
 	}
-	ib_unregister_driver(RDMA_DRIVER_BNXT_RE);
+	return 0;
+}
+
+static void __exit bnxt_re_mod_exit(void)
+{
+	auxiliary_driver_unregister(&bnxt_re_driver);
 }
 
 module_init(bnxt_re_mod_init);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index 899d9adfd35f..bc1f8c189251 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -19,6 +19,7 @@
 #include <linux/irq.h>
 #include <asm/byteorder.h>
 #include <linux/bitmap.h>
+#include <linux/auxiliary_bus.h>
 
 #include "bnxt_hsi.h"
 #include "bnxt.h"
@@ -73,8 +74,6 @@ static int bnxt_unregister_dev(struct bnxt_en_dev *edev, unsigned int ulp_id)
 	if (ulp_id >= BNXT_MAX_ULP)
 		return -EINVAL;
 
-	edev->flags |= BNXT_EN_FLAG_ULP_STOPPED;
-
 	ulp = &edev->ulp_tbl[ulp_id];
 	if (!rcu_access_pointer(ulp->ulp_ops)) {
 		netdev_err(bp->dev, "ulp id %d not registered\n", ulp_id);
@@ -590,29 +589,3 @@ void bnxt_rdma_aux_device_init(struct bnxt *bp)
 skip_ida_init:
 	bp->flags &= ~BNXT_FLAG_ROCE_CAP;
 }
-
-struct bnxt_en_dev *bnxt_ulp_probe(struct net_device *dev)
-{
-	struct bnxt *bp = netdev_priv(dev);
-	struct bnxt_en_dev *edev;
-
-	edev = bp->edev;
-	if (!edev) {
-		edev = kzalloc(sizeof(*edev), GFP_KERNEL);
-		if (!edev)
-			return ERR_PTR(-ENOMEM);
-		edev->en_ops = &bnxt_en_ops_tbl;
-		edev->net = dev;
-		edev->pdev = bp->pdev;
-		edev->l2_db_size = bp->db_size;
-		edev->l2_db_size_nc = bp->db_size;
-		bp->edev = edev;
-	}
-	edev->flags &= ~BNXT_EN_FLAG_ROCE_CAP;
-	if (bp->flags & BNXT_FLAG_ROCEV1_CAP)
-		edev->flags |= BNXT_EN_FLAG_ROCEV1_CAP;
-	if (bp->flags & BNXT_FLAG_ROCEV2_CAP)
-		edev->flags |= BNXT_EN_FLAG_ROCEV2_CAP;
-	return bp->edev;
-}
-EXPORT_SYMBOL(bnxt_ulp_probe);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
index 47c7131e5549..d90f8beb6d0e 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
@@ -31,7 +31,6 @@ struct bnxt_ulp_ops {
 	void (*ulp_stop)(void *);
 	void (*ulp_start)(void *);
 	void (*ulp_sriov_config)(void *, int);
-	void (*ulp_shutdown)(void *);
 	void (*ulp_irq_stop)(void *);
 	void (*ulp_irq_restart)(void *, struct bnxt_msix_entry *);
 };
@@ -108,6 +107,4 @@ void bnxt_ulp_async_events(struct bnxt *bp, struct hwrm_async_event_cmpl *cmpl);
 void bnxt_rdma_aux_device_uninit(struct bnxt *bp);
 void bnxt_rdma_aux_device_init(struct bnxt *bp);
 void bnxt_aux_dev_free(struct bnxt *bp);
-struct bnxt_en_dev *bnxt_ulp_probe(struct net_device *dev);
-
 #endif
-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

[PATCH net-next v7 3/8] bnxt_en: Remove usage of ulp_id

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 19437 bytes --]

Since the driver continues to use the single ULP model,
the extra complexity and indirection is unnecessary.
Remove the usage of ulp_id from the code.

Suggested-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Selvin Xavier <selvin.xavier@broadcom.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/bnxt_re/main.c          |  24 +-
 drivers/net/ethernet/broadcom/bnxt/bnxt.c     |   2 +-
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 208 ++++++++----------
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h |  26 +--
 4 files changed, 114 insertions(+), 146 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index 165ae11b927d..814f1d2e31b5 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -363,8 +363,7 @@ static int bnxt_re_unregister_netdev(struct bnxt_re_dev *rdev)
 
 	en_dev = rdev->en_dev;
 
-	rc = en_dev->en_ops->bnxt_unregister_device(rdev->en_dev,
-						    BNXT_ROCE_ULP);
+	rc = en_dev->en_ops->bnxt_unregister_device(rdev->en_dev);
 	return rc;
 }
 
@@ -375,7 +374,7 @@ static int bnxt_re_register_netdev(struct bnxt_re_dev *rdev)
 
 	en_dev = rdev->en_dev;
 
-	rc = en_dev->en_ops->bnxt_register_device(en_dev, BNXT_ROCE_ULP,
+	rc = en_dev->en_ops->bnxt_register_device(en_dev,
 						  &bnxt_re_ulp_ops, rdev);
 	rdev->qplib_res.pdev = rdev->en_dev->pdev;
 	return rc;
@@ -384,16 +383,15 @@ static int bnxt_re_register_netdev(struct bnxt_re_dev *rdev)
 static int bnxt_re_free_msix(struct bnxt_re_dev *rdev)
 {
 	struct bnxt_en_dev *en_dev;
-	int rc;
 
 	if (!rdev)
 		return -EINVAL;
 
 	en_dev = rdev->en_dev;
 
-	rc = en_dev->en_ops->bnxt_free_msix(rdev->en_dev, BNXT_ROCE_ULP);
+	en_dev->en_ops->bnxt_free_msix(rdev->en_dev);
 
-	return rc;
+	return 0;
 }
 
 static int bnxt_re_request_msix(struct bnxt_re_dev *rdev)
@@ -405,7 +403,7 @@ static int bnxt_re_request_msix(struct bnxt_re_dev *rdev)
 
 	num_msix_want = min_t(u32, BNXT_RE_MAX_MSIX, num_online_cpus());
 
-	num_msix_got = en_dev->en_ops->bnxt_request_msix(en_dev, BNXT_ROCE_ULP,
+	num_msix_got = en_dev->en_ops->bnxt_request_msix(en_dev,
 							 rdev->msix_entries,
 							 num_msix_want);
 	if (num_msix_got < BNXT_RE_MIN_MSIX) {
@@ -468,7 +466,7 @@ static int bnxt_re_net_ring_free(struct bnxt_re_dev *rdev,
 	req.ring_id = cpu_to_le16(fw_ring_id);
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, BNXT_ROCE_ULP, &fw_msg);
+	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
 	if (rc)
 		ibdev_err(&rdev->ibdev, "Failed to free HW ring:%d :%#x",
 			  req.ring_id, rc);
@@ -505,7 +503,7 @@ static int bnxt_re_net_ring_alloc(struct bnxt_re_dev *rdev,
 	req.int_mode = ring_attr->mode;
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, BNXT_ROCE_ULP, &fw_msg);
+	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
 	if (!rc)
 		*fw_ring_id = le16_to_cpu(resp.ring_id);
 
@@ -533,7 +531,7 @@ static int bnxt_re_net_stats_ctx_free(struct bnxt_re_dev *rdev,
 	req.stat_ctx_id = cpu_to_le32(fw_stats_ctx_id);
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, BNXT_ROCE_ULP, &fw_msg);
+	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
 	if (rc)
 		ibdev_err(&rdev->ibdev, "Failed to free HW stats context %#x",
 			  rc);
@@ -566,7 +564,7 @@ static int bnxt_re_net_stats_ctx_alloc(struct bnxt_re_dev *rdev,
 	req.stat_ctx_flags = STAT_CTX_ALLOC_REQ_STAT_CTX_FLAGS_ROCE;
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, BNXT_ROCE_ULP, &fw_msg);
+	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
 	if (!rc)
 		*fw_stats_ctx_id = le32_to_cpu(resp.stat_ctx_id);
 
@@ -1052,7 +1050,7 @@ static int bnxt_re_query_hwrm_pri2cos(struct bnxt_re_dev *rdev, u8 dir,
 
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, BNXT_ROCE_ULP, &fw_msg);
+	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
 	if (rc)
 		return rc;
 
@@ -1235,7 +1233,7 @@ static void bnxt_re_query_hwrm_intf_version(struct bnxt_re_dev *rdev)
 	req.hwrm_intf_upd = HWRM_VERSION_UPDATE;
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, BNXT_ROCE_ULP, &fw_msg);
+	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
 	if (rc) {
 		ibdev_err(&rdev->ibdev, "Failed to query HW version, rc = 0x%x",
 			  rc);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index c5cbf2307e07..fc0054c284c5 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -5538,7 +5538,7 @@ int bnxt_hwrm_vnic_cfg(struct bnxt *bp, u16 vnic_id)
 #endif
 	if ((bp->flags & BNXT_FLAG_STRIP_VLAN) || def_vlan)
 		req->flags |= cpu_to_le32(VNIC_CFG_REQ_FLAGS_VLAN_STRIP_MODE);
-	if (!vnic_id && bnxt_ulp_registered(bp->edev, BNXT_ROCE_ULP))
+	if (!vnic_id && bnxt_ulp_registered(bp->edev))
 		req->flags |= cpu_to_le32(bnxt_get_roce_vnic_mode(bp));
 
 	return hwrm_req_send(bp, req);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index bc1f8c189251..988b09a9e99e 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -28,59 +28,44 @@
 
 static DEFINE_IDA(bnxt_aux_dev_ids);
 
-static int bnxt_register_dev(struct bnxt_en_dev *edev, unsigned int ulp_id,
-			     struct bnxt_ulp_ops *ulp_ops, void *handle)
+static int bnxt_register_dev(struct bnxt_en_dev *edev,
+			     struct bnxt_ulp_ops *ulp_ops,
+			     void *handle)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
+	unsigned int max_stat_ctxs;
 	struct bnxt_ulp *ulp;
 
-	if (ulp_id >= BNXT_MAX_ULP)
-		return -EINVAL;
-
-	ulp = &edev->ulp_tbl[ulp_id];
-	if (rcu_access_pointer(ulp->ulp_ops)) {
-		netdev_err(bp->dev, "ulp id %d already registered\n", ulp_id);
-		return -EBUSY;
-	}
-	if (ulp_id == BNXT_ROCE_ULP) {
-		unsigned int max_stat_ctxs;
+	max_stat_ctxs = bnxt_get_max_func_stat_ctxs(bp);
+	if (max_stat_ctxs <= BNXT_MIN_ROCE_STAT_CTXS ||
+	    bp->cp_nr_rings == max_stat_ctxs)
+		return -ENOMEM;
 
-		max_stat_ctxs = bnxt_get_max_func_stat_ctxs(bp);
-		if (max_stat_ctxs <= BNXT_MIN_ROCE_STAT_CTXS ||
-		    bp->cp_nr_rings == max_stat_ctxs)
-			return -ENOMEM;
-	}
+	ulp = kzalloc(sizeof(*ulp), GFP_KERNEL);
+	if (!ulp)
+		return -ENOMEM;
 
-	atomic_set(&ulp->ref_count, 1);
+	edev->ulp_tbl = ulp;
 	ulp->handle = handle;
 	rcu_assign_pointer(ulp->ulp_ops, ulp_ops);
 
-	if (ulp_id == BNXT_ROCE_ULP) {
-		if (test_bit(BNXT_STATE_OPEN, &bp->state))
-			bnxt_hwrm_vnic_cfg(bp, 0);
-	}
+	if (test_bit(BNXT_STATE_OPEN, &bp->state))
+		bnxt_hwrm_vnic_cfg(bp, 0);
 
 	return 0;
 }
 
-static int bnxt_unregister_dev(struct bnxt_en_dev *edev, unsigned int ulp_id)
+static int bnxt_unregister_dev(struct bnxt_en_dev *edev)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
 	struct bnxt_ulp *ulp;
 	int i = 0;
 
-	if (ulp_id >= BNXT_MAX_ULP)
-		return -EINVAL;
-
-	ulp = &edev->ulp_tbl[ulp_id];
-	if (!rcu_access_pointer(ulp->ulp_ops)) {
-		netdev_err(bp->dev, "ulp id %d not registered\n", ulp_id);
-		return -EINVAL;
-	}
-	if (ulp_id == BNXT_ROCE_ULP && ulp->msix_requested)
-		edev->en_ops->bnxt_free_msix(edev, ulp_id);
+	ulp = edev->ulp_tbl;
+	if (ulp->msix_requested)
+		edev->en_ops->bnxt_free_msix(edev);
 
 	if (ulp->max_async_event_id)
 		bnxt_hwrm_func_drv_rgtr(bp, NULL, 0, true);
@@ -93,6 +78,8 @@ static int bnxt_unregister_dev(struct bnxt_en_dev *edev, unsigned int ulp_id)
 		msleep(100);
 		i++;
 	}
+	kfree(ulp);
+	edev->ulp_tbl = NULL;
 	return 0;
 }
 
@@ -101,8 +88,8 @@ static void bnxt_fill_msix_vecs(struct bnxt *bp, struct bnxt_msix_entry *ent)
 	struct bnxt_en_dev *edev = bp->edev;
 	int num_msix, idx, i;
 
-	num_msix = edev->ulp_tbl[BNXT_ROCE_ULP].msix_requested;
-	idx = edev->ulp_tbl[BNXT_ROCE_ULP].msix_base;
+	num_msix = edev->ulp_tbl->msix_requested;
+	idx = edev->ulp_tbl->msix_base;
 	for (i = 0; i < num_msix; i++) {
 		ent[i].vector = bp->irq_tbl[idx + i].vector;
 		ent[i].ring_idx = idx + i;
@@ -116,8 +103,9 @@ static void bnxt_fill_msix_vecs(struct bnxt *bp, struct bnxt_msix_entry *ent)
 	}
 }
 
-static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id,
-			      struct bnxt_msix_entry *ent, int num_msix)
+static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev,
+			      struct bnxt_msix_entry *ent,
+			      int num_msix)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
@@ -127,13 +115,10 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id,
 	int total_vecs;
 	int rc = 0;
 
-	if (ulp_id != BNXT_ROCE_ULP)
-		return -EINVAL;
-
 	if (!(bp->flags & BNXT_FLAG_USING_MSIX))
 		return -ENODEV;
 
-	if (edev->ulp_tbl[ulp_id].msix_requested)
+	if (edev->ulp_tbl->msix_requested)
 		return -EAGAIN;
 
 	max_cp_rings = bnxt_get_max_func_cp_rings(bp);
@@ -150,8 +135,8 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id,
 		idx = max_idx - avail_msix;
 	}
 
-	edev->ulp_tbl[ulp_id].msix_base = idx;
-	edev->ulp_tbl[ulp_id].msix_requested = avail_msix;
+	edev->ulp_tbl->msix_base = idx;
+	edev->ulp_tbl->msix_requested = avail_msix;
 	hw_resc = &bp->hw_resc;
 	total_vecs = idx + avail_msix;
 	rtnl_lock();
@@ -166,7 +151,7 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id,
 	}
 	rtnl_unlock();
 	if (rc) {
-		edev->ulp_tbl[ulp_id].msix_requested = 0;
+		edev->ulp_tbl->msix_requested = 0;
 		return -EAGAIN;
 	}
 
@@ -175,25 +160,22 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id,
 
 		resv_msix = hw_resc->resv_irqs - bp->cp_nr_rings;
 		avail_msix = min_t(int, resv_msix, avail_msix);
-		edev->ulp_tbl[ulp_id].msix_requested = avail_msix;
+		edev->ulp_tbl->msix_requested = avail_msix;
 	}
 	bnxt_fill_msix_vecs(bp, ent);
 	edev->flags |= BNXT_EN_FLAG_MSIX_REQUESTED;
 	return avail_msix;
 }
 
-static int bnxt_free_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id)
+static void bnxt_free_msix_vecs(struct bnxt_en_dev *edev)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
 
-	if (ulp_id != BNXT_ROCE_ULP)
-		return -EINVAL;
-
 	if (!(edev->flags & BNXT_EN_FLAG_MSIX_REQUESTED))
-		return 0;
+		return;
 
-	edev->ulp_tbl[ulp_id].msix_requested = 0;
+	edev->ulp_tbl->msix_requested = 0;
 	edev->flags &= ~BNXT_EN_FLAG_MSIX_REQUESTED;
 	rtnl_lock();
 	if (netif_running(dev) && !(edev->flags & BNXT_EN_FLAG_ULP_STOPPED)) {
@@ -202,43 +184,43 @@ static int bnxt_free_msix_vecs(struct bnxt_en_dev *edev, unsigned int ulp_id)
 	}
 	rtnl_unlock();
 
-	return 0;
+	return;
 }
 
 int bnxt_get_ulp_msix_num(struct bnxt *bp)
 {
-	if (bnxt_ulp_registered(bp->edev, BNXT_ROCE_ULP)) {
+	if (bnxt_ulp_registered(bp->edev)) {
 		struct bnxt_en_dev *edev = bp->edev;
 
-		return edev->ulp_tbl[BNXT_ROCE_ULP].msix_requested;
+		return edev->ulp_tbl->msix_requested;
 	}
 	return 0;
 }
 
 int bnxt_get_ulp_msix_base(struct bnxt *bp)
 {
-	if (bnxt_ulp_registered(bp->edev, BNXT_ROCE_ULP)) {
+	if (bnxt_ulp_registered(bp->edev)) {
 		struct bnxt_en_dev *edev = bp->edev;
 
-		if (edev->ulp_tbl[BNXT_ROCE_ULP].msix_requested)
-			return edev->ulp_tbl[BNXT_ROCE_ULP].msix_base;
+		if (edev->ulp_tbl->msix_requested)
+			return edev->ulp_tbl->msix_base;
 	}
 	return 0;
 }
 
 int bnxt_get_ulp_stat_ctxs(struct bnxt *bp)
 {
-	if (bnxt_ulp_registered(bp->edev, BNXT_ROCE_ULP)) {
+	if (bnxt_ulp_registered(bp->edev)) {
 		struct bnxt_en_dev *edev = bp->edev;
 
-		if (edev->ulp_tbl[BNXT_ROCE_ULP].msix_requested)
+		if (edev->ulp_tbl->msix_requested)
 			return BNXT_MIN_ROCE_STAT_CTXS;
 	}
 
 	return 0;
 }
 
-static int bnxt_send_msg(struct bnxt_en_dev *edev, unsigned int ulp_id,
+static int bnxt_send_msg(struct bnxt_en_dev *edev,
 			 struct bnxt_fw_msg *fw_msg)
 {
 	struct net_device *dev = edev->net;
@@ -248,7 +230,7 @@ static int bnxt_send_msg(struct bnxt_en_dev *edev, unsigned int ulp_id,
 	u32 resp_len;
 	int rc;
 
-	if (ulp_id != BNXT_ROCE_ULP && bp->fw_reset_state)
+	if (bp->fw_reset_state)
 		return -EBUSY;
 
 	rc = hwrm_req_init(bp, req, 0 /* don't care */);
@@ -287,27 +269,24 @@ void bnxt_ulp_stop(struct bnxt *bp)
 {
 	struct bnxt_en_dev *edev = bp->edev;
 	struct bnxt_ulp_ops *ops;
-	int i;
+	struct bnxt_ulp *ulp;
 
 	if (!edev)
 		return;
 
 	edev->flags |= BNXT_EN_FLAG_ULP_STOPPED;
-	for (i = 0; i < BNXT_MAX_ULP; i++) {
-		struct bnxt_ulp *ulp = &edev->ulp_tbl[i];
-
-		ops = rtnl_dereference(ulp->ulp_ops);
-		if (!ops || !ops->ulp_stop)
-			continue;
-		ops->ulp_stop(ulp->handle);
-	}
+	ulp = edev->ulp_tbl;
+	ops = rtnl_dereference(ulp->ulp_ops);
+	if (!ops || !ops->ulp_stop)
+		return;
+	ops->ulp_stop(ulp->handle);
 }
 
 void bnxt_ulp_start(struct bnxt *bp, int err)
 {
 	struct bnxt_en_dev *edev = bp->edev;
 	struct bnxt_ulp_ops *ops;
-	int i;
+	struct bnxt_ulp *ulp;
 
 	if (!edev)
 		return;
@@ -317,39 +296,33 @@ void bnxt_ulp_start(struct bnxt *bp, int err)
 	if (err)
 		return;
 
-	for (i = 0; i < BNXT_MAX_ULP; i++) {
-		struct bnxt_ulp *ulp = &edev->ulp_tbl[i];
-
-		ops = rtnl_dereference(ulp->ulp_ops);
-		if (!ops || !ops->ulp_start)
-			continue;
-		ops->ulp_start(ulp->handle);
-	}
+	ulp = edev->ulp_tbl;
+	ops = rtnl_dereference(ulp->ulp_ops);
+	if (!ops || !ops->ulp_start)
+		return;
+	ops->ulp_start(ulp->handle);
 }
 
 void bnxt_ulp_sriov_cfg(struct bnxt *bp, int num_vfs)
 {
 	struct bnxt_en_dev *edev = bp->edev;
 	struct bnxt_ulp_ops *ops;
-	int i;
+	struct bnxt_ulp *ulp;
 
 	if (!edev)
 		return;
+	ulp = edev->ulp_tbl;
 
-	for (i = 0; i < BNXT_MAX_ULP; i++) {
-		struct bnxt_ulp *ulp = &edev->ulp_tbl[i];
-
-		rcu_read_lock();
-		ops = rcu_dereference(ulp->ulp_ops);
-		if (!ops || !ops->ulp_sriov_config) {
-			rcu_read_unlock();
-			continue;
-		}
-		bnxt_ulp_get(ulp);
+	rcu_read_lock();
+	ops = rcu_dereference(ulp->ulp_ops);
+	if (!ops || !ops->ulp_sriov_config) {
 		rcu_read_unlock();
-		ops->ulp_sriov_config(ulp->handle, num_vfs);
-		bnxt_ulp_put(ulp);
+		return;
 	}
+	bnxt_ulp_get(ulp);
+	rcu_read_unlock();
+	ops->ulp_sriov_config(ulp->handle, num_vfs);
+	bnxt_ulp_put(ulp);
 }
 
 void bnxt_ulp_irq_stop(struct bnxt *bp)
@@ -360,8 +333,8 @@ void bnxt_ulp_irq_stop(struct bnxt *bp)
 	if (!edev || !(edev->flags & BNXT_EN_FLAG_MSIX_REQUESTED))
 		return;
 
-	if (bnxt_ulp_registered(bp->edev, BNXT_ROCE_ULP)) {
-		struct bnxt_ulp *ulp = &edev->ulp_tbl[BNXT_ROCE_ULP];
+	if (bnxt_ulp_registered(bp->edev)) {
+		struct bnxt_ulp *ulp = edev->ulp_tbl;
 
 		if (!ulp->msix_requested)
 			return;
@@ -381,8 +354,8 @@ void bnxt_ulp_irq_restart(struct bnxt *bp, int err)
 	if (!edev || !(edev->flags & BNXT_EN_FLAG_MSIX_REQUESTED))
 		return;
 
-	if (bnxt_ulp_registered(bp->edev, BNXT_ROCE_ULP)) {
-		struct bnxt_ulp *ulp = &edev->ulp_tbl[BNXT_ROCE_ULP];
+	if (bnxt_ulp_registered(bp->edev)) {
+		struct bnxt_ulp *ulp = edev->ulp_tbl;
 		struct bnxt_msix_entry *ent = NULL;
 
 		if (!ulp->msix_requested)
@@ -409,41 +382,38 @@ void bnxt_ulp_async_events(struct bnxt *bp, struct hwrm_async_event_cmpl *cmpl)
 	u16 event_id = le16_to_cpu(cmpl->event_id);
 	struct bnxt_en_dev *edev = bp->edev;
 	struct bnxt_ulp_ops *ops;
-	int i;
+	struct bnxt_ulp *ulp;
 
-	if (!edev)
+	if (!bnxt_ulp_registered(edev))
 		return;
 
+	ulp = edev->ulp_tbl;
+
 	rcu_read_lock();
-	for (i = 0; i < BNXT_MAX_ULP; i++) {
-		struct bnxt_ulp *ulp = &edev->ulp_tbl[i];
-
-		ops = rcu_dereference(ulp->ulp_ops);
-		if (!ops || !ops->ulp_async_notifier)
-			continue;
-		if (!ulp->async_events_bmap ||
-		    event_id > ulp->max_async_event_id)
-			continue;
-
-		/* Read max_async_event_id first before testing the bitmap. */
-		smp_rmb();
-		if (test_bit(event_id, ulp->async_events_bmap))
-			ops->ulp_async_notifier(ulp->handle, cmpl);
-	}
+
+	ops = rcu_dereference(ulp->ulp_ops);
+	if (!ops || !ops->ulp_async_notifier)
+		goto exit;
+	if (!ulp->async_events_bmap || event_id > ulp->max_async_event_id)
+		goto exit;
+
+	/* Read max_async_event_id first before testing the bitmap. */
+	smp_rmb();
+	if (test_bit(event_id, ulp->async_events_bmap))
+		ops->ulp_async_notifier(ulp->handle, cmpl);
+exit:
 	rcu_read_unlock();
 }
 
-static int bnxt_register_async_events(struct bnxt_en_dev *edev, unsigned int ulp_id,
-				      unsigned long *events_bmap, u16 max_id)
+static int bnxt_register_async_events(struct bnxt_en_dev *edev,
+				      unsigned long *events_bmap,
+				      u16 max_id)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
 	struct bnxt_ulp *ulp;
 
-	if (ulp_id >= BNXT_MAX_ULP)
-		return -EINVAL;
-
-	ulp = &edev->ulp_tbl[ulp_id];
+	ulp = edev->ulp_tbl;
 	ulp->async_events_bmap = events_bmap;
 	/* Make sure bnxt_ulp_async_events() sees this order */
 	smp_wmb();
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
index d90f8beb6d0e..5433f3e17746 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
@@ -64,7 +64,7 @@ struct bnxt_en_dev {
 	#define BNXT_EN_FLAG_MSIX_REQUESTED	0x4
 	#define BNXT_EN_FLAG_ULP_STOPPED	0x8
 	const struct bnxt_en_ops	*en_ops;
-	struct bnxt_ulp			ulp_tbl[BNXT_MAX_ULP];
+	struct bnxt_ulp			*ulp_tbl;
 	int				l2_db_size;	/* Doorbell BAR size in
 							 * bytes mapped by L2
 							 * driver.
@@ -76,21 +76,21 @@ struct bnxt_en_dev {
 };
 
 struct bnxt_en_ops {
-	int (*bnxt_register_device)(struct bnxt_en_dev *, unsigned int,
-				    struct bnxt_ulp_ops *, void *);
-	int (*bnxt_unregister_device)(struct bnxt_en_dev *, unsigned int);
-	int (*bnxt_request_msix)(struct bnxt_en_dev *, unsigned int,
-				 struct bnxt_msix_entry *, int);
-	int (*bnxt_free_msix)(struct bnxt_en_dev *, unsigned int);
-	int (*bnxt_send_fw_msg)(struct bnxt_en_dev *, unsigned int,
-				struct bnxt_fw_msg *);
-	int (*bnxt_register_fw_async_events)(struct bnxt_en_dev *, unsigned int,
-					     unsigned long *, u16);
+	int (*bnxt_register_device)(struct bnxt_en_dev *edev,
+				    struct bnxt_ulp_ops *ulp_ops, void *handle);
+	int (*bnxt_unregister_device)(struct bnxt_en_dev *edev);
+	int (*bnxt_request_msix)(struct bnxt_en_dev *edev,
+				 struct bnxt_msix_entry *ent, int num_msix);
+	void (*bnxt_free_msix)(struct bnxt_en_dev *edev);
+	int (*bnxt_send_fw_msg)(struct bnxt_en_dev *edev,
+				struct bnxt_fw_msg *fw_msg);
+	int (*bnxt_register_fw_async_events)(struct bnxt_en_dev *edev,
+					     unsigned long *events_bmap, u16 max_id);
 };
 
-static inline bool bnxt_ulp_registered(struct bnxt_en_dev *edev, int ulp_id)
+static inline bool bnxt_ulp_registered(struct bnxt_en_dev *edev)
 {
-	if (edev && rcu_access_pointer(edev->ulp_tbl[ulp_id].ulp_ops))
+	if (edev && edev->ulp_tbl)
 		return true;
 	return false;
 }
-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

[PATCH net-next v7 4/8] bnxt_en: Use direct API instead of indirection

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 14346 bytes --]

For a single ULP user there is no need for complicating function
indirection calls. Remove all this complexity in favour of direct
function calls exported by the bnxt_en driver. This allows to
simplify the code greatly. Also remove unused ulp_async_notifier.

Suggested-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Selvin Xavier <selvin.xavier@broadcom.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/bnxt_re/main.c          | 71 ++++-------------
 drivers/net/ethernet/broadcom/bnxt/bnxt.c     |  1 -
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 76 +++++--------------
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h | 25 +++---
 4 files changed, 45 insertions(+), 128 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index 814f1d2e31b5..e1777c8e6d5e 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -350,23 +350,6 @@ static struct bnxt_ulp_ops bnxt_re_ulp_ops = {
 
 /* RoCE -> Net driver */
 
-/* Driver registration routines used to let the networking driver (bnxt_en)
- * to know that the RoCE driver is now installed
- */
-static int bnxt_re_unregister_netdev(struct bnxt_re_dev *rdev)
-{
-	struct bnxt_en_dev *en_dev;
-	int rc;
-
-	if (!rdev)
-		return -EINVAL;
-
-	en_dev = rdev->en_dev;
-
-	rc = en_dev->en_ops->bnxt_unregister_device(rdev->en_dev);
-	return rc;
-}
-
 static int bnxt_re_register_netdev(struct bnxt_re_dev *rdev)
 {
 	struct bnxt_en_dev *en_dev;
@@ -374,26 +357,12 @@ static int bnxt_re_register_netdev(struct bnxt_re_dev *rdev)
 
 	en_dev = rdev->en_dev;
 
-	rc = en_dev->en_ops->bnxt_register_device(en_dev,
-						  &bnxt_re_ulp_ops, rdev);
-	rdev->qplib_res.pdev = rdev->en_dev->pdev;
+	rc = bnxt_register_dev(en_dev, &bnxt_re_ulp_ops, rdev);
+	if (!rc)
+		rdev->qplib_res.pdev = rdev->en_dev->pdev;
 	return rc;
 }
 
-static int bnxt_re_free_msix(struct bnxt_re_dev *rdev)
-{
-	struct bnxt_en_dev *en_dev;
-
-	if (!rdev)
-		return -EINVAL;
-
-	en_dev = rdev->en_dev;
-
-	en_dev->en_ops->bnxt_free_msix(rdev->en_dev);
-
-	return 0;
-}
-
 static int bnxt_re_request_msix(struct bnxt_re_dev *rdev)
 {
 	int rc = 0, num_msix_want = BNXT_RE_MAX_MSIX, num_msix_got;
@@ -403,9 +372,9 @@ static int bnxt_re_request_msix(struct bnxt_re_dev *rdev)
 
 	num_msix_want = min_t(u32, BNXT_RE_MAX_MSIX, num_online_cpus());
 
-	num_msix_got = en_dev->en_ops->bnxt_request_msix(en_dev,
-							 rdev->msix_entries,
-							 num_msix_want);
+	num_msix_got = bnxt_req_msix_vecs(en_dev,
+					  rdev->msix_entries,
+					  num_msix_want);
 	if (num_msix_got < BNXT_RE_MIN_MSIX) {
 		rc = -EINVAL;
 		goto done;
@@ -466,7 +435,7 @@ static int bnxt_re_net_ring_free(struct bnxt_re_dev *rdev,
 	req.ring_id = cpu_to_le16(fw_ring_id);
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
+	rc = bnxt_send_msg(en_dev, &fw_msg);
 	if (rc)
 		ibdev_err(&rdev->ibdev, "Failed to free HW ring:%d :%#x",
 			  req.ring_id, rc);
@@ -503,7 +472,7 @@ static int bnxt_re_net_ring_alloc(struct bnxt_re_dev *rdev,
 	req.int_mode = ring_attr->mode;
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
+	rc = bnxt_send_msg(en_dev, &fw_msg);
 	if (!rc)
 		*fw_ring_id = le16_to_cpu(resp.ring_id);
 
@@ -531,7 +500,7 @@ static int bnxt_re_net_stats_ctx_free(struct bnxt_re_dev *rdev,
 	req.stat_ctx_id = cpu_to_le32(fw_stats_ctx_id);
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
+	rc = bnxt_send_msg(en_dev, &fw_msg);
 	if (rc)
 		ibdev_err(&rdev->ibdev, "Failed to free HW stats context %#x",
 			  rc);
@@ -564,7 +533,7 @@ static int bnxt_re_net_stats_ctx_alloc(struct bnxt_re_dev *rdev,
 	req.stat_ctx_flags = STAT_CTX_ALLOC_REQ_STAT_CTX_FLAGS_ROCE;
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
+	rc = bnxt_send_msg(en_dev, &fw_msg);
 	if (!rc)
 		*fw_stats_ctx_id = le32_to_cpu(resp.stat_ctx_id);
 
@@ -1050,7 +1019,7 @@ static int bnxt_re_query_hwrm_pri2cos(struct bnxt_re_dev *rdev, u8 dir,
 
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
+	rc = bnxt_send_msg(en_dev, &fw_msg);
 	if (rc)
 		return rc;
 
@@ -1233,7 +1202,7 @@ static void bnxt_re_query_hwrm_intf_version(struct bnxt_re_dev *rdev)
 	req.hwrm_intf_upd = HWRM_VERSION_UPDATE;
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-	rc = en_dev->en_ops->bnxt_send_fw_msg(en_dev, &fw_msg);
+	rc = bnxt_send_msg(en_dev, &fw_msg);
 	if (rc) {
 		ibdev_err(&rdev->ibdev, "Failed to query HW version, rc = 0x%x",
 			  rc);
@@ -1296,20 +1265,12 @@ static void bnxt_re_dev_uninit(struct bnxt_re_dev *rdev)
 		bnxt_re_net_ring_free(rdev, rdev->rcfw.creq.ring_id, type);
 		bnxt_qplib_free_rcfw_channel(&rdev->rcfw);
 	}
-	if (test_and_clear_bit(BNXT_RE_FLAG_GOT_MSIX, &rdev->flags)) {
-		rc = bnxt_re_free_msix(rdev);
-		if (rc)
-			ibdev_warn(&rdev->ibdev,
-				   "Failed to free MSI-X vectors: %#x", rc);
-	}
+	if (test_and_clear_bit(BNXT_RE_FLAG_GOT_MSIX, &rdev->flags))
+		bnxt_free_msix_vecs(rdev->en_dev);
 
 	bnxt_re_destroy_chip_ctx(rdev);
-	if (test_and_clear_bit(BNXT_RE_FLAG_NETDEV_REGISTERED, &rdev->flags)) {
-		rc = bnxt_re_unregister_netdev(rdev);
-		if (rc)
-			ibdev_warn(&rdev->ibdev,
-				   "Failed to unregister with netdev: %#x", rc);
-	}
+	if (test_and_clear_bit(BNXT_RE_FLAG_NETDEV_REGISTERED, &rdev->flags))
+		bnxt_unregister_dev(rdev->en_dev);
 }
 
 /* worker thread for polling periodic events. Now used for QoS programming*/
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index fc0054c284c5..f24e08bbeaf4 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -2414,7 +2414,6 @@ static int bnxt_async_event_process(struct bnxt *bp,
 	}
 	bnxt_queue_sp_work(bp);
 async_event_process_exit:
-	bnxt_ulp_async_events(bp, cmpl);
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index 988b09a9e99e..e7539946759d 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -28,9 +28,9 @@
 
 static DEFINE_IDA(bnxt_aux_dev_ids);
 
-static int bnxt_register_dev(struct bnxt_en_dev *edev,
-			     struct bnxt_ulp_ops *ulp_ops,
-			     void *handle)
+int bnxt_register_dev(struct bnxt_en_dev *edev,
+		      struct bnxt_ulp_ops *ulp_ops,
+		      void *handle)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
@@ -55,8 +55,9 @@ static int bnxt_register_dev(struct bnxt_en_dev *edev,
 
 	return 0;
 }
+EXPORT_SYMBOL(bnxt_register_dev);
 
-static int bnxt_unregister_dev(struct bnxt_en_dev *edev)
+void bnxt_unregister_dev(struct bnxt_en_dev *edev)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
@@ -65,7 +66,7 @@ static int bnxt_unregister_dev(struct bnxt_en_dev *edev)
 
 	ulp = edev->ulp_tbl;
 	if (ulp->msix_requested)
-		edev->en_ops->bnxt_free_msix(edev);
+		bnxt_free_msix_vecs(edev);
 
 	if (ulp->max_async_event_id)
 		bnxt_hwrm_func_drv_rgtr(bp, NULL, 0, true);
@@ -80,8 +81,9 @@ static int bnxt_unregister_dev(struct bnxt_en_dev *edev)
 	}
 	kfree(ulp);
 	edev->ulp_tbl = NULL;
-	return 0;
+	return;
 }
+EXPORT_SYMBOL(bnxt_unregister_dev);
 
 static void bnxt_fill_msix_vecs(struct bnxt *bp, struct bnxt_msix_entry *ent)
 {
@@ -103,7 +105,7 @@ static void bnxt_fill_msix_vecs(struct bnxt *bp, struct bnxt_msix_entry *ent)
 	}
 }
 
-static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev,
+int bnxt_req_msix_vecs(struct bnxt_en_dev *edev,
 			      struct bnxt_msix_entry *ent,
 			      int num_msix)
 {
@@ -166,8 +168,9 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev,
 	edev->flags |= BNXT_EN_FLAG_MSIX_REQUESTED;
 	return avail_msix;
 }
+EXPORT_SYMBOL(bnxt_req_msix_vecs);
 
-static void bnxt_free_msix_vecs(struct bnxt_en_dev *edev)
+void bnxt_free_msix_vecs(struct bnxt_en_dev *edev)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
@@ -186,6 +189,7 @@ static void bnxt_free_msix_vecs(struct bnxt_en_dev *edev)
 
 	return;
 }
+EXPORT_SYMBOL(bnxt_free_msix_vecs);
 
 int bnxt_get_ulp_msix_num(struct bnxt *bp)
 {
@@ -220,7 +224,7 @@ int bnxt_get_ulp_stat_ctxs(struct bnxt *bp)
 	return 0;
 }
 
-static int bnxt_send_msg(struct bnxt_en_dev *edev,
+int bnxt_send_msg(struct bnxt_en_dev *edev,
 			 struct bnxt_fw_msg *fw_msg)
 {
 	struct net_device *dev = edev->net;
@@ -254,6 +258,7 @@ static int bnxt_send_msg(struct bnxt_en_dev *edev,
 	hwrm_req_drop(bp, req);
 	return rc;
 }
+EXPORT_SYMBOL(bnxt_send_msg);
 
 static void bnxt_ulp_get(struct bnxt_ulp *ulp)
 {
@@ -313,14 +318,11 @@ void bnxt_ulp_sriov_cfg(struct bnxt *bp, int num_vfs)
 		return;
 	ulp = edev->ulp_tbl;
 
-	rcu_read_lock();
 	ops = rcu_dereference(ulp->ulp_ops);
-	if (!ops || !ops->ulp_sriov_config) {
-		rcu_read_unlock();
+	if (!ops || !ops->ulp_sriov_config)
 		return;
-	}
+
 	bnxt_ulp_get(ulp);
-	rcu_read_unlock();
 	ops->ulp_sriov_config(ulp->handle, num_vfs);
 	bnxt_ulp_put(ulp);
 }
@@ -377,37 +379,9 @@ void bnxt_ulp_irq_restart(struct bnxt *bp, int err)
 	}
 }
 
-void bnxt_ulp_async_events(struct bnxt *bp, struct hwrm_async_event_cmpl *cmpl)
-{
-	u16 event_id = le16_to_cpu(cmpl->event_id);
-	struct bnxt_en_dev *edev = bp->edev;
-	struct bnxt_ulp_ops *ops;
-	struct bnxt_ulp *ulp;
-
-	if (!bnxt_ulp_registered(edev))
-		return;
-
-	ulp = edev->ulp_tbl;
-
-	rcu_read_lock();
-
-	ops = rcu_dereference(ulp->ulp_ops);
-	if (!ops || !ops->ulp_async_notifier)
-		goto exit;
-	if (!ulp->async_events_bmap || event_id > ulp->max_async_event_id)
-		goto exit;
-
-	/* Read max_async_event_id first before testing the bitmap. */
-	smp_rmb();
-	if (test_bit(event_id, ulp->async_events_bmap))
-		ops->ulp_async_notifier(ulp->handle, cmpl);
-exit:
-	rcu_read_unlock();
-}
-
-static int bnxt_register_async_events(struct bnxt_en_dev *edev,
-				      unsigned long *events_bmap,
-				      u16 max_id)
+int bnxt_register_async_events(struct bnxt_en_dev *edev,
+			       unsigned long *events_bmap,
+			       u16 max_id)
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
@@ -421,15 +395,7 @@ static int bnxt_register_async_events(struct bnxt_en_dev *edev,
 	bnxt_hwrm_func_drv_rgtr(bp, events_bmap, max_id + 1, true);
 	return 0;
 }
-
-static const struct bnxt_en_ops bnxt_en_ops_tbl = {
-	.bnxt_register_device	= bnxt_register_dev,
-	.bnxt_unregister_device	= bnxt_unregister_dev,
-	.bnxt_request_msix	= bnxt_req_msix_vecs,
-	.bnxt_free_msix		= bnxt_free_msix_vecs,
-	.bnxt_send_fw_msg	= bnxt_send_msg,
-	.bnxt_register_fw_async_events	= bnxt_register_async_events,
-};
+EXPORT_SYMBOL(bnxt_register_async_events);
 
 void bnxt_aux_dev_free(struct bnxt *bp)
 {
@@ -465,7 +431,6 @@ static void bnxt_aux_dev_release(struct device *dev)
 		container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
 	struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
 
-	bnxt_adev->edev->en_ops = NULL;
 	kfree(bnxt_adev->edev);
 	bnxt_adev->edev = NULL;
 	bp->edev = NULL;
@@ -473,7 +438,6 @@ static void bnxt_aux_dev_release(struct device *dev)
 
 static void bnxt_set_edev_info(struct bnxt_en_dev *edev, struct bnxt *bp)
 {
-	edev->en_ops = &bnxt_en_ops_tbl;
 	edev->net = bp->dev;
 	edev->pdev = bp->pdev;
 	edev->l2_db_size = bp->db_size;
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
index 5433f3e17746..3bda864db1be 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
@@ -26,8 +26,6 @@ struct bnxt_msix_entry {
 };
 
 struct bnxt_ulp_ops {
-	/* async_notifier() cannot sleep (in BH context) */
-	void (*ulp_async_notifier)(void *, struct hwrm_async_event_cmpl *);
 	void (*ulp_stop)(void *);
 	void (*ulp_start)(void *);
 	void (*ulp_sriov_config)(void *, int);
@@ -63,7 +61,6 @@ struct bnxt_en_dev {
 						 BNXT_EN_FLAG_ROCEV2_CAP)
 	#define BNXT_EN_FLAG_MSIX_REQUESTED	0x4
 	#define BNXT_EN_FLAG_ULP_STOPPED	0x8
-	const struct bnxt_en_ops	*en_ops;
 	struct bnxt_ulp			*ulp_tbl;
 	int				l2_db_size;	/* Doorbell BAR size in
 							 * bytes mapped by L2
@@ -75,19 +72,6 @@ struct bnxt_en_dev {
 							 */
 };
 
-struct bnxt_en_ops {
-	int (*bnxt_register_device)(struct bnxt_en_dev *edev,
-				    struct bnxt_ulp_ops *ulp_ops, void *handle);
-	int (*bnxt_unregister_device)(struct bnxt_en_dev *edev);
-	int (*bnxt_request_msix)(struct bnxt_en_dev *edev,
-				 struct bnxt_msix_entry *ent, int num_msix);
-	void (*bnxt_free_msix)(struct bnxt_en_dev *edev);
-	int (*bnxt_send_fw_msg)(struct bnxt_en_dev *edev,
-				struct bnxt_fw_msg *fw_msg);
-	int (*bnxt_register_fw_async_events)(struct bnxt_en_dev *edev,
-					     unsigned long *events_bmap, u16 max_id);
-};
-
 static inline bool bnxt_ulp_registered(struct bnxt_en_dev *edev)
 {
 	if (edev && edev->ulp_tbl)
@@ -107,4 +91,13 @@ void bnxt_ulp_async_events(struct bnxt *bp, struct hwrm_async_event_cmpl *cmpl);
 void bnxt_rdma_aux_device_uninit(struct bnxt *bp);
 void bnxt_rdma_aux_device_init(struct bnxt *bp);
 void bnxt_aux_dev_free(struct bnxt *bp);
+int bnxt_register_dev(struct bnxt_en_dev *edev, struct bnxt_ulp_ops *ulp_ops,
+		      void *handle);
+void bnxt_unregister_dev(struct bnxt_en_dev *edev);
+int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, struct bnxt_msix_entry *ent,
+		       int num_msix);
+void bnxt_free_msix_vecs(struct bnxt_en_dev *edev);
+int bnxt_send_msg(struct bnxt_en_dev *edev, struct bnxt_fw_msg *fw_msg);
+int bnxt_register_async_events(struct bnxt_en_dev *edev,
+			       unsigned long *events_bmap, u16 max_id);
 #endif
-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

[PATCH net-next v7 5/8] bnxt_en: Use auxiliary bus calls over proprietary calls

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 7351 bytes --]

Wherever possible use the function ops provided by auxiliary bus
instead of using proprietary ops.

Defined bnxt_re_suspend and bnxt_re_resume calls which can be
invoked by the bnxt_en driver instead of the ULP stop/start calls.

Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Selvin Xavier <selvin.xavier@broadcom.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/bnxt_re/main.c          | 102 +++++++++++-------
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c |  40 ++++---
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h |   2 -
 3 files changed, 87 insertions(+), 57 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index e1777c8e6d5e..ae04198c33b7 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -226,45 +226,6 @@ static void bnxt_re_set_resource_limits(struct bnxt_re_dev *rdev)
 		bnxt_re_limit_vf_res(&rdev->qplib_ctx, num_vfs);
 }
 
-/* for handling bnxt_en callbacks later */
-static void bnxt_re_stop(void *p)
-{
-	struct bnxt_re_dev *rdev = p;
-	struct bnxt *bp;
-
-	if (!rdev)
-		return;
-
-	/* L2 driver invokes this callback during device error/crash or device
-	 * reset. Current RoCE driver doesn't recover the device in case of
-	 * error. Handle the error by dispatching fatal events to all qps
-	 * ie. by calling bnxt_re_dev_stop and release the MSIx vectors as
-	 * L2 driver want to modify the MSIx table.
-	 */
-	bp = netdev_priv(rdev->netdev);
-
-	ibdev_info(&rdev->ibdev, "Handle device stop call from L2 driver");
-	/* Check the current device state from L2 structure and move the
-	 * device to detached state if FW_FATAL_COND is set.
-	 * This prevents more commands to HW during clean-up,
-	 * in case the device is already in error.
-	 */
-	if (test_bit(BNXT_STATE_FW_FATAL_COND, &bp->state))
-		set_bit(ERR_DEVICE_DETACHED, &rdev->rcfw.cmdq.flags);
-
-	bnxt_re_dev_stop(rdev);
-	bnxt_re_stop_irq(rdev);
-	/* Move the device states to detached and  avoid sending any more
-	 * commands to HW
-	 */
-	set_bit(BNXT_RE_FLAG_ERR_DEVICE_DETACHED, &rdev->flags);
-	set_bit(ERR_DEVICE_DETACHED, &rdev->rcfw.cmdq.flags);
-}
-
-static void bnxt_re_start(void *p)
-{
-}
-
 static void bnxt_re_sriov_config(void *p, int num_vfs)
 {
 	struct bnxt_re_dev *rdev = p;
@@ -341,8 +302,6 @@ static void bnxt_re_start_irq(void *handle, struct bnxt_msix_entry *ent)
 }
 
 static struct bnxt_ulp_ops bnxt_re_ulp_ops = {
-	.ulp_stop = bnxt_re_stop,
-	.ulp_start = bnxt_re_start,
 	.ulp_sriov_config = bnxt_re_sriov_config,
 	.ulp_irq_stop = bnxt_re_stop_irq,
 	.ulp_irq_restart = bnxt_re_start_irq
@@ -1585,6 +1544,65 @@ static int bnxt_re_probe(struct auxiliary_device *adev,
 	return rc;
 }
 
+static int bnxt_re_suspend(struct auxiliary_device *adev, pm_message_t state)
+{
+	struct bnxt_re_dev *rdev = auxiliary_get_drvdata(adev);
+	struct bnxt *bp;
+
+	if (!rdev)
+		return 0;
+
+	mutex_lock(&bnxt_re_mutex);
+	/* L2 driver may invoke this callback during device error/crash or device
+	 * reset. Current RoCE driver doesn't recover the device in case of
+	 * error. Handle the error by dispatching fatal events to all qps
+	 * ie. by calling bnxt_re_dev_stop and release the MSIx vectors as
+	 * L2 driver want to modify the MSIx table.
+	 */
+	bp = netdev_priv(rdev->netdev);
+
+	ibdev_info(&rdev->ibdev, "Handle device suspend call");
+	/* Check the current device state from L2 structure and move the
+	 * device to detached state if FW_FATAL_COND is set.
+	 * This prevents more commands to HW during clean-up,
+	 * in case the device is already in error.
+	 */
+	if (test_bit(BNXT_STATE_FW_FATAL_COND, &bp->state))
+		set_bit(ERR_DEVICE_DETACHED, &rdev->rcfw.cmdq.flags);
+
+	bnxt_re_dev_stop(rdev);
+	bnxt_re_stop_irq(rdev);
+	/* Move the device states to detached and  avoid sending any more
+	 * commands to HW
+	 */
+	set_bit(BNXT_RE_FLAG_ERR_DEVICE_DETACHED, &rdev->flags);
+	set_bit(ERR_DEVICE_DETACHED, &rdev->rcfw.cmdq.flags);
+	mutex_unlock(&bnxt_re_mutex);
+
+	return 0;
+}
+
+static int bnxt_re_resume(struct auxiliary_device *adev)
+{
+	struct bnxt_re_dev *rdev = auxiliary_get_drvdata(adev);
+
+	if (!rdev)
+		return 0;
+
+	mutex_lock(&bnxt_re_mutex);
+	/* L2 driver may invoke this callback during device recovery, resume.
+	 * reset. Current RoCE driver doesn't recover the device in case of
+	 * error. Handle the error by dispatching fatal events to all qps
+	 * ie. by calling bnxt_re_dev_stop and release the MSIx vectors as
+	 * L2 driver want to modify the MSIx table.
+	 */
+
+	ibdev_info(&rdev->ibdev, "Handle device resume call");
+	mutex_unlock(&bnxt_re_mutex);
+
+	return 0;
+}
+
 static const struct auxiliary_device_id bnxt_re_id_table[] = {
 	{ .name = BNXT_ADEV_NAME ".rdma", },
 	{},
@@ -1597,6 +1615,8 @@ static struct auxiliary_driver bnxt_re_driver = {
 	.probe = bnxt_re_probe,
 	.remove = bnxt_re_remove,
 	.shutdown = bnxt_re_shutdown,
+	.suspend = bnxt_re_suspend,
+	.resume = bnxt_re_resume,
 	.id_table = bnxt_re_id_table,
 };
 
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index e7539946759d..e08b51ee8886 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -272,26 +272,31 @@ static void bnxt_ulp_put(struct bnxt_ulp *ulp)
 
 void bnxt_ulp_stop(struct bnxt *bp)
 {
+	struct bnxt_aux_dev *bnxt_aux = bp->aux_dev;
 	struct bnxt_en_dev *edev = bp->edev;
-	struct bnxt_ulp_ops *ops;
-	struct bnxt_ulp *ulp;
 
 	if (!edev)
 		return;
 
 	edev->flags |= BNXT_EN_FLAG_ULP_STOPPED;
-	ulp = edev->ulp_tbl;
-	ops = rtnl_dereference(ulp->ulp_ops);
-	if (!ops || !ops->ulp_stop)
-		return;
-	ops->ulp_stop(ulp->handle);
+	if (bnxt_aux) {
+		struct auxiliary_device *adev;
+
+		adev = &bnxt_aux->aux_dev;
+		if (adev->dev.driver) {
+			struct auxiliary_driver *adrv;
+			pm_message_t pm = {};
+
+			adrv = to_auxiliary_drv(adev->dev.driver);
+			adrv->suspend(adev, pm);
+		}
+	}
 }
 
 void bnxt_ulp_start(struct bnxt *bp, int err)
 {
+	struct bnxt_aux_dev *bnxt_aux = bp->aux_dev;
 	struct bnxt_en_dev *edev = bp->edev;
-	struct bnxt_ulp_ops *ops;
-	struct bnxt_ulp *ulp;
 
 	if (!edev)
 		return;
@@ -301,11 +306,18 @@ void bnxt_ulp_start(struct bnxt *bp, int err)
 	if (err)
 		return;
 
-	ulp = edev->ulp_tbl;
-	ops = rtnl_dereference(ulp->ulp_ops);
-	if (!ops || !ops->ulp_start)
-		return;
-	ops->ulp_start(ulp->handle);
+	if (bnxt_aux) {
+		struct auxiliary_device *adev;
+
+		adev = &bnxt_aux->aux_dev;
+		if (adev->dev.driver) {
+			struct auxiliary_driver *adrv;
+
+			adrv = to_auxiliary_drv(adev->dev.driver);
+			adrv->resume(adev);
+		}
+	}
+
 }
 
 void bnxt_ulp_sriov_cfg(struct bnxt *bp, int num_vfs)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
index 3bda864db1be..6525aad3b642 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
@@ -26,8 +26,6 @@ struct bnxt_msix_entry {
 };
 
 struct bnxt_ulp_ops {
-	void (*ulp_stop)(void *);
-	void (*ulp_start)(void *);
 	void (*ulp_sriov_config)(void *, int);
 	void (*ulp_irq_stop)(void *);
 	void (*ulp_irq_restart)(void *, struct bnxt_msix_entry *);
-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

[PATCH net-next v7 6/8] bnxt_en: Remove struct bnxt access from RoCE driver

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 6317 bytes --]

From: Hongguang Gao <hongguang.gao@broadcom.com>

Decouple RoCE driver from directly accessing L2's private bnxt
structure. Move the fields needed by RoCE driver into bnxt_en_dev.
They'll be passed to RoCE driver by bnxt_rdma_aux_device_add()
function.

Signed-off-by: Hongguang Gao <hongguang.gao@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Selvin Xavier <selvin.xavier@broadcom.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/bnxt_re/main.c          | 22 ++++++-------------
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c |  9 ++++++++
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h | 11 ++++++++++
 3 files changed, 27 insertions(+), 15 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index ae04198c33b7..3251078e9fe3 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -112,16 +112,14 @@ static int bnxt_re_setup_chip_ctx(struct bnxt_re_dev *rdev, u8 wqe_mode)
 {
 	struct bnxt_qplib_chip_ctx *chip_ctx;
 	struct bnxt_en_dev *en_dev;
-	struct bnxt *bp;
 
 	en_dev = rdev->en_dev;
-	bp = netdev_priv(en_dev->net);
 
 	chip_ctx = kzalloc(sizeof(*chip_ctx), GFP_KERNEL);
 	if (!chip_ctx)
 		return -ENOMEM;
-	chip_ctx->chip_num = bp->chip_num;
-	chip_ctx->hw_stats_size = bp->hw_ring_stats_size;
+	chip_ctx->chip_num = en_dev->chip_num;
+	chip_ctx->hw_stats_size = en_dev->hw_ring_stats_size;
 
 	rdev->chip_ctx = chip_ctx;
 	/* rest members to follow eventually */
@@ -129,7 +127,7 @@ static int bnxt_re_setup_chip_ctx(struct bnxt_re_dev *rdev, u8 wqe_mode)
 	rdev->qplib_res.cctx = rdev->chip_ctx;
 	rdev->rcfw.res = &rdev->qplib_res;
 	rdev->qplib_res.dattr = &rdev->dev_attr;
-	rdev->qplib_res.is_vf = BNXT_VF(bp);
+	rdev->qplib_res.is_vf = BNXT_EN_VF(en_dev);
 
 	bnxt_re_set_drv_mode(rdev, wqe_mode);
 	if (bnxt_qplib_determine_atomics(en_dev->pdev))
@@ -142,10 +140,7 @@ static int bnxt_re_setup_chip_ctx(struct bnxt_re_dev *rdev, u8 wqe_mode)
 
 static void bnxt_re_get_sriov_func_type(struct bnxt_re_dev *rdev)
 {
-	struct bnxt *bp;
-
-	bp = netdev_priv(rdev->en_dev->net);
-	if (BNXT_VF(bp))
+	if (BNXT_EN_VF(rdev->en_dev))
 		rdev->is_virtfn = 1;
 }
 
@@ -957,7 +952,6 @@ static int bnxt_re_query_hwrm_pri2cos(struct bnxt_re_dev *rdev, u8 dir,
 				      u64 *cid_map)
 {
 	struct hwrm_queue_pri2cos_qcfg_input req = {0};
-	struct bnxt *bp = netdev_priv(rdev->netdev);
 	struct hwrm_queue_pri2cos_qcfg_output resp;
 	struct bnxt_en_dev *en_dev = rdev->en_dev;
 	struct bnxt_fw_msg fw_msg;
@@ -974,7 +968,7 @@ static int bnxt_re_query_hwrm_pri2cos(struct bnxt_re_dev *rdev, u8 dir,
 	flags |= (dir & 0x01);
 	flags |= HWRM_QUEUE_PRI2COS_QCFG_INPUT_FLAGS_IVLAN;
 	req.flags = cpu_to_le32(flags);
-	req.port_id = bp->pf.port_id;
+	req.port_id = en_dev->pf_port_id;
 
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
@@ -1547,7 +1541,6 @@ static int bnxt_re_probe(struct auxiliary_device *adev,
 static int bnxt_re_suspend(struct auxiliary_device *adev, pm_message_t state)
 {
 	struct bnxt_re_dev *rdev = auxiliary_get_drvdata(adev);
-	struct bnxt *bp;
 
 	if (!rdev)
 		return 0;
@@ -1559,15 +1552,14 @@ static int bnxt_re_suspend(struct auxiliary_device *adev, pm_message_t state)
 	 * ie. by calling bnxt_re_dev_stop and release the MSIx vectors as
 	 * L2 driver want to modify the MSIx table.
 	 */
-	bp = netdev_priv(rdev->netdev);
 
 	ibdev_info(&rdev->ibdev, "Handle device suspend call");
-	/* Check the current device state from L2 structure and move the
+	/* Check the current device state from bnxt_en_dev and move the
 	 * device to detached state if FW_FATAL_COND is set.
 	 * This prevents more commands to HW during clean-up,
 	 * in case the device is already in error.
 	 */
-	if (test_bit(BNXT_STATE_FW_FATAL_COND, &bp->state))
+	if (test_bit(BNXT_STATE_FW_FATAL_COND, &rdev->en_dev->en_state))
 		set_bit(ERR_DEVICE_DETACHED, &rdev->rcfw.cmdq.flags);
 
 	bnxt_re_dev_stop(rdev);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index e08b51ee8886..8bc78d5525ad 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -288,6 +288,7 @@ void bnxt_ulp_stop(struct bnxt *bp)
 			pm_message_t pm = {};
 
 			adrv = to_auxiliary_drv(adev->dev.driver);
+			edev->en_state = bp->state;
 			adrv->suspend(adev, pm);
 		}
 	}
@@ -314,6 +315,7 @@ void bnxt_ulp_start(struct bnxt *bp, int err)
 			struct auxiliary_driver *adrv;
 
 			adrv = to_auxiliary_drv(adev->dev.driver);
+			edev->en_state = bp->state;
 			adrv->resume(adev);
 		}
 	}
@@ -459,6 +461,13 @@ static void bnxt_set_edev_info(struct bnxt_en_dev *edev, struct bnxt *bp)
 		edev->flags |= BNXT_EN_FLAG_ROCEV1_CAP;
 	if (bp->flags & BNXT_FLAG_ROCEV2_CAP)
 		edev->flags |= BNXT_EN_FLAG_ROCEV2_CAP;
+	if (bp->flags & BNXT_FLAG_VF)
+		edev->flags |= BNXT_EN_FLAG_VF;
+
+	edev->chip_num = bp->chip_num;
+	edev->hw_ring_stats_size = bp->hw_ring_stats_size;
+	edev->pf_port_id = bp->pf.port_id;
+	edev->en_state = bp->state;
 }
 
 static int bnxt_rdma_aux_device_add(struct bnxt *bp)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
index 6525aad3b642..ce939df24574 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
@@ -59,6 +59,9 @@ struct bnxt_en_dev {
 						 BNXT_EN_FLAG_ROCEV2_CAP)
 	#define BNXT_EN_FLAG_MSIX_REQUESTED	0x4
 	#define BNXT_EN_FLAG_ULP_STOPPED	0x8
+	#define BNXT_EN_FLAG_VF			0x10
+#define BNXT_EN_VF(edev)	((edev)->flags & BNXT_EN_FLAG_VF)
+
 	struct bnxt_ulp			*ulp_tbl;
 	int				l2_db_size;	/* Doorbell BAR size in
 							 * bytes mapped by L2
@@ -68,6 +71,14 @@ struct bnxt_en_dev {
 							 * bytes mapped as non-
 							 * cacheable.
 							 */
+	u16				chip_num;
+	u16				hw_ring_stats_size;
+	u16				pf_port_id;
+	unsigned long			en_state;	/* Could be checked in
+							 * RoCE driver suspend
+							 * mode only. Will be
+							 * updated in resume.
+							 */
 };
 
 static inline bool bnxt_ulp_registered(struct bnxt_en_dev *edev)
-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

[PATCH net-next v7 7/8] RDMA/bnxt_re: Remove the sriov config callback

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 4837 bytes --]

Remove the SRIOV config callback which the bnxt_en was calling
to reconfigure the chip resources for a PF device when VFs are
created. The code is now modified to provision the VF resources
based on the total VF count instead of the actual VF count.
This allows the SRIOV config callback to be removed from the
list of ulp_ops.

Suggested-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/bnxt_re/main.c          | 11 ++++---
 .../net/ethernet/broadcom/bnxt/bnxt_sriov.c   |  7 +----
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 29 -------------------
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h |  1 -
 4 files changed, 8 insertions(+), 40 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index 3251078e9fe3..b287c3ba9352 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -221,13 +221,12 @@ static void bnxt_re_set_resource_limits(struct bnxt_re_dev *rdev)
 		bnxt_re_limit_vf_res(&rdev->qplib_ctx, num_vfs);
 }
 
-static void bnxt_re_sriov_config(void *p, int num_vfs)
+static void bnxt_re_vf_res_config(struct bnxt_re_dev *rdev)
 {
-	struct bnxt_re_dev *rdev = p;
 
 	if (test_bit(BNXT_RE_FLAG_ERR_DEVICE_DETACHED, &rdev->flags))
 		return;
-	rdev->num_vfs = num_vfs;
+	rdev->num_vfs = pci_sriov_get_totalvfs(rdev->en_dev->pdev);
 	if (!bnxt_qplib_is_chip_gen_p5(rdev->chip_ctx)) {
 		bnxt_re_set_resource_limits(rdev);
 		bnxt_qplib_set_func_resources(&rdev->qplib_res, &rdev->rcfw,
@@ -297,7 +296,6 @@ static void bnxt_re_start_irq(void *handle, struct bnxt_msix_entry *ent)
 }
 
 static struct bnxt_ulp_ops bnxt_re_ulp_ops = {
-	.ulp_sriov_config = bnxt_re_sriov_config,
 	.ulp_irq_stop = bnxt_re_stop_irq,
 	.ulp_irq_restart = bnxt_re_start_irq
 };
@@ -1369,6 +1367,11 @@ static int bnxt_re_dev_init(struct bnxt_re_dev *rdev, u8 wqe_mode)
 		INIT_DELAYED_WORK(&rdev->worker, bnxt_re_worker);
 		set_bit(BNXT_RE_FLAG_QOS_WORK_REG, &rdev->flags);
 		schedule_delayed_work(&rdev->worker, msecs_to_jiffies(30000));
+		/*
+		 * Use the total VF count since the actual VF count may not be
+		 * available at this point.
+		 */
+		bnxt_re_vf_res_config(rdev);
 	}
 
 	return 0;
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
index a4cba7cb2783..3ed3a2b3b3a9 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
@@ -749,7 +749,6 @@ int bnxt_cfg_hw_sriov(struct bnxt *bp, int *num_vfs, bool reset)
 		*num_vfs = rc;
 	}
 
-	bnxt_ulp_sriov_cfg(bp, *num_vfs);
 	return 0;
 }
 
@@ -823,10 +822,8 @@ static int bnxt_sriov_enable(struct bnxt *bp, int *num_vfs)
 		goto err_out2;
 
 	rc = pci_enable_sriov(bp->pdev, *num_vfs);
-	if (rc) {
-		bnxt_ulp_sriov_cfg(bp, 0);
+	if (rc)
 		goto err_out2;
-	}
 
 	return 0;
 
@@ -872,8 +869,6 @@ void bnxt_sriov_disable(struct bnxt *bp)
 	rtnl_lock();
 	bnxt_restore_pf_fw_resources(bp);
 	rtnl_unlock();
-
-	bnxt_ulp_sriov_cfg(bp, 0);
 }
 
 int bnxt_sriov_configure(struct pci_dev *pdev, int num_vfs)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index 8bc78d5525ad..2164b4419263 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -260,16 +260,6 @@ int bnxt_send_msg(struct bnxt_en_dev *edev,
 }
 EXPORT_SYMBOL(bnxt_send_msg);
 
-static void bnxt_ulp_get(struct bnxt_ulp *ulp)
-{
-	atomic_inc(&ulp->ref_count);
-}
-
-static void bnxt_ulp_put(struct bnxt_ulp *ulp)
-{
-	atomic_dec(&ulp->ref_count);
-}
-
 void bnxt_ulp_stop(struct bnxt *bp)
 {
 	struct bnxt_aux_dev *bnxt_aux = bp->aux_dev;
@@ -322,25 +312,6 @@ void bnxt_ulp_start(struct bnxt *bp, int err)
 
 }
 
-void bnxt_ulp_sriov_cfg(struct bnxt *bp, int num_vfs)
-{
-	struct bnxt_en_dev *edev = bp->edev;
-	struct bnxt_ulp_ops *ops;
-	struct bnxt_ulp *ulp;
-
-	if (!edev)
-		return;
-	ulp = edev->ulp_tbl;
-
-	ops = rcu_dereference(ulp->ulp_ops);
-	if (!ops || !ops->ulp_sriov_config)
-		return;
-
-	bnxt_ulp_get(ulp);
-	ops->ulp_sriov_config(ulp->handle, num_vfs);
-	bnxt_ulp_put(ulp);
-}
-
 void bnxt_ulp_irq_stop(struct bnxt *bp)
 {
 	struct bnxt_en_dev *edev = bp->edev;
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
index ce939df24574..0e15b35c36dd 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
@@ -26,7 +26,6 @@ struct bnxt_msix_entry {
 };
 
 struct bnxt_ulp_ops {
-	void (*ulp_sriov_config)(void *, int);
 	void (*ulp_irq_stop)(void *);
 	void (*ulp_irq_restart)(void *, struct bnxt_msix_entry *);
 };
-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

[PATCH net-next v7 8/8] bnxt_en: Remove runtime interrupt vector allocation

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 14258 bytes --]

Modified the bnxt_en code to create and pre-configure RDMA devices
with the right MSI-X vector count for the ROCE driver to use.
This is to align the ROCE driver to the auxiliary device model which
will simply bind the driver without getting into PCI-related handling.
All PCI-related logic will now be in the bnxt_en driver.

Suggested-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
---
 drivers/infiniband/hw/bnxt_re/bnxt_re.h       |   1 -
 drivers/infiniband/hw/bnxt_re/main.c          |  48 ++---
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 166 ++++++------------
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h |   6 +-
 4 files changed, 64 insertions(+), 157 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/bnxt_re.h b/drivers/infiniband/hw/bnxt_re/bnxt_re.h
index b0465c8d229a..5a2baf49ecaa 100644
--- a/drivers/infiniband/hw/bnxt_re/bnxt_re.h
+++ b/drivers/infiniband/hw/bnxt_re/bnxt_re.h
@@ -129,7 +129,6 @@ struct bnxt_re_dev {
 	unsigned int			version, major, minor;
 	struct bnxt_qplib_chip_ctx	*chip_ctx;
 	struct bnxt_en_dev		*en_dev;
-	struct bnxt_msix_entry		msix_entries[BNXT_RE_MAX_MSIX];
 	int				num_msix;
 
 	int				id;
diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index b287c3ba9352..8d991fa184ae 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -262,7 +262,7 @@ static void bnxt_re_stop_irq(void *handle)
 static void bnxt_re_start_irq(void *handle, struct bnxt_msix_entry *ent)
 {
 	struct bnxt_re_dev *rdev = (struct bnxt_re_dev *)handle;
-	struct bnxt_msix_entry *msix_ent = rdev->msix_entries;
+	struct bnxt_msix_entry *msix_ent = rdev->en_dev->msix_entries;
 	struct bnxt_qplib_rcfw *rcfw = &rdev->rcfw;
 	struct bnxt_qplib_nq *nq;
 	int indx, rc;
@@ -281,7 +281,7 @@ static void bnxt_re_start_irq(void *handle, struct bnxt_msix_entry *ent)
 	 * in device sctructure.
 	 */
 	for (indx = 0; indx < rdev->num_msix; indx++)
-		rdev->msix_entries[indx].vector = ent[indx].vector;
+		rdev->en_dev->msix_entries[indx].vector = ent[indx].vector;
 
 	bnxt_qplib_rcfw_start_irq(rcfw, msix_ent[BNXT_RE_AEQ_IDX].vector,
 				  false);
@@ -315,32 +315,6 @@ static int bnxt_re_register_netdev(struct bnxt_re_dev *rdev)
 	return rc;
 }
 
-static int bnxt_re_request_msix(struct bnxt_re_dev *rdev)
-{
-	int rc = 0, num_msix_want = BNXT_RE_MAX_MSIX, num_msix_got;
-	struct bnxt_en_dev *en_dev;
-
-	en_dev = rdev->en_dev;
-
-	num_msix_want = min_t(u32, BNXT_RE_MAX_MSIX, num_online_cpus());
-
-	num_msix_got = bnxt_req_msix_vecs(en_dev,
-					  rdev->msix_entries,
-					  num_msix_want);
-	if (num_msix_got < BNXT_RE_MIN_MSIX) {
-		rc = -EINVAL;
-		goto done;
-	}
-	if (num_msix_got != num_msix_want) {
-		ibdev_warn(&rdev->ibdev,
-			   "Requested %d MSI-X vectors, got %d\n",
-			   num_msix_want, num_msix_got);
-	}
-	rdev->num_msix = num_msix_got;
-done:
-	return rc;
-}
-
 static void bnxt_re_init_hwrm_hdr(struct bnxt_re_dev *rdev, struct input *hdr,
 				  u16 opcd, u16 crid, u16 trid)
 {
@@ -785,7 +759,7 @@ static u32 bnxt_re_get_nqdb_offset(struct bnxt_re_dev *rdev, u16 indx)
 	return bnxt_qplib_is_chip_gen_p5(rdev->chip_ctx) ?
 		(rdev->is_virtfn ? BNXT_RE_GEN_P5_VF_NQ_DB :
 				   BNXT_RE_GEN_P5_PF_NQ_DB) :
-				   rdev->msix_entries[indx].db_offset;
+				   rdev->en_dev->msix_entries[indx].db_offset;
 }
 
 static void bnxt_re_cleanup_res(struct bnxt_re_dev *rdev)
@@ -810,7 +784,7 @@ static int bnxt_re_init_res(struct bnxt_re_dev *rdev)
 	for (i = 1; i < rdev->num_msix ; i++) {
 		db_offt = bnxt_re_get_nqdb_offset(rdev, i);
 		rc = bnxt_qplib_enable_nq(rdev->en_dev->pdev, &rdev->nq[i - 1],
-					  i - 1, rdev->msix_entries[i].vector,
+					  i - 1, rdev->en_dev->msix_entries[i].vector,
 					  db_offt, &bnxt_re_cqn_handler,
 					  &bnxt_re_srqn_handler);
 		if (rc) {
@@ -897,7 +871,7 @@ static int bnxt_re_alloc_res(struct bnxt_re_dev *rdev)
 		rattr.type = type;
 		rattr.mode = RING_ALLOC_REQ_INT_MODE_MSIX;
 		rattr.depth = BNXT_QPLIB_NQE_MAX_CNT - 1;
-		rattr.lrid = rdev->msix_entries[i + 1].ring_idx;
+		rattr.lrid = rdev->en_dev->msix_entries[i + 1].ring_idx;
 		rc = bnxt_re_net_ring_alloc(rdev, &rattr, &nq->ring_id);
 		if (rc) {
 			ibdev_err(&rdev->ibdev,
@@ -1217,7 +1191,7 @@ static void bnxt_re_dev_uninit(struct bnxt_re_dev *rdev)
 		bnxt_qplib_free_rcfw_channel(&rdev->rcfw);
 	}
 	if (test_and_clear_bit(BNXT_RE_FLAG_GOT_MSIX, &rdev->flags))
-		bnxt_free_msix_vecs(rdev->en_dev);
+		rdev->num_msix = 0;
 
 	bnxt_re_destroy_chip_ctx(rdev);
 	if (test_and_clear_bit(BNXT_RE_FLAG_NETDEV_REGISTERED, &rdev->flags))
@@ -1262,13 +1236,15 @@ static int bnxt_re_dev_init(struct bnxt_re_dev *rdev, u8 wqe_mode)
 	/* Check whether VF or PF */
 	bnxt_re_get_sriov_func_type(rdev);
 
-	rc = bnxt_re_request_msix(rdev);
-	if (rc) {
+	if (!rdev->en_dev->ulp_tbl->msix_requested) {
 		ibdev_err(&rdev->ibdev,
 			  "Failed to get MSI-X vectors: %#x\n", rc);
 		rc = -EINVAL;
 		goto fail;
 	}
+	ibdev_dbg(&rdev->ibdev, "Got %d MSI-X vectors\n",
+		  rdev->en_dev->ulp_tbl->msix_requested);
+	rdev->num_msix = rdev->en_dev->ulp_tbl->msix_requested;
 	set_bit(BNXT_RE_FLAG_GOT_MSIX, &rdev->flags);
 
 	bnxt_re_query_hwrm_intf_version(rdev);
@@ -1292,14 +1268,14 @@ static int bnxt_re_dev_init(struct bnxt_re_dev *rdev, u8 wqe_mode)
 	rattr.type = type;
 	rattr.mode = RING_ALLOC_REQ_INT_MODE_MSIX;
 	rattr.depth = BNXT_QPLIB_CREQE_MAX_CNT - 1;
-	rattr.lrid = rdev->msix_entries[BNXT_RE_AEQ_IDX].ring_idx;
+	rattr.lrid = rdev->en_dev->msix_entries[BNXT_RE_AEQ_IDX].ring_idx;
 	rc = bnxt_re_net_ring_alloc(rdev, &rattr, &creq->ring_id);
 	if (rc) {
 		ibdev_err(&rdev->ibdev, "Failed to allocate CREQ: %#x\n", rc);
 		goto free_rcfw;
 	}
 	db_offt = bnxt_re_get_nqdb_offset(rdev, BNXT_RE_AEQ_IDX);
-	vid = rdev->msix_entries[BNXT_RE_AEQ_IDX].vector;
+	vid = rdev->en_dev->msix_entries[BNXT_RE_AEQ_IDX].vector;
 	rc = bnxt_qplib_enable_rcfw_channel(&rdev->rcfw,
 					    vid, db_offt, rdev->is_virtfn,
 					    &bnxt_re_aeq_handler);
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index 2164b4419263..75ae0f499793 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -28,6 +28,30 @@
 
 static DEFINE_IDA(bnxt_aux_dev_ids);
 
+static void bnxt_fill_msix_vecs(struct bnxt *bp, struct bnxt_msix_entry *ent)
+{
+	struct bnxt_en_dev *edev = bp->edev;
+	int num_msix, idx, i;
+
+	if (!edev->ulp_tbl->msix_requested) {
+		netdev_warn(bp->dev, "Requested MSI-X vectors insufficient\n");
+		return;
+	}
+	num_msix = edev->ulp_tbl->msix_requested;
+	idx = edev->ulp_tbl->msix_base;
+	for (i = 0; i < num_msix; i++) {
+		ent[i].vector = bp->irq_tbl[idx + i].vector;
+		ent[i].ring_idx = idx + i;
+		if (bp->flags & BNXT_FLAG_CHIP_P5) {
+			ent[i].db_offset = DB_PF_OFFSET_P5;
+			if (BNXT_VF(bp))
+				ent[i].db_offset = DB_VF_OFFSET_P5;
+		} else {
+			ent[i].db_offset = (idx + i) * 0x80;
+		}
+	}
+}
+
 int bnxt_register_dev(struct bnxt_en_dev *edev,
 		      struct bnxt_ulp_ops *ulp_ops,
 		      void *handle)
@@ -42,17 +66,18 @@ int bnxt_register_dev(struct bnxt_en_dev *edev,
 	    bp->cp_nr_rings == max_stat_ctxs)
 		return -ENOMEM;
 
-	ulp = kzalloc(sizeof(*ulp), GFP_KERNEL);
+	ulp = edev->ulp_tbl;
 	if (!ulp)
 		return -ENOMEM;
 
-	edev->ulp_tbl = ulp;
 	ulp->handle = handle;
 	rcu_assign_pointer(ulp->ulp_ops, ulp_ops);
 
 	if (test_bit(BNXT_STATE_OPEN, &bp->state))
 		bnxt_hwrm_vnic_cfg(bp, 0);
 
+	bnxt_fill_msix_vecs(bp, bp->edev->msix_entries);
+	edev->flags |= BNXT_EN_FLAG_MSIX_REQUESTED;
 	return 0;
 }
 EXPORT_SYMBOL(bnxt_register_dev);
@@ -66,7 +91,7 @@ void bnxt_unregister_dev(struct bnxt_en_dev *edev)
 
 	ulp = edev->ulp_tbl;
 	if (ulp->msix_requested)
-		bnxt_free_msix_vecs(edev);
+		edev->flags &= ~BNXT_EN_FLAG_MSIX_REQUESTED;
 
 	if (ulp->max_async_event_id)
 		bnxt_hwrm_func_drv_rgtr(bp, NULL, 0, true);
@@ -79,126 +104,17 @@ void bnxt_unregister_dev(struct bnxt_en_dev *edev)
 		msleep(100);
 		i++;
 	}
-	kfree(ulp);
-	edev->ulp_tbl = NULL;
 	return;
 }
 EXPORT_SYMBOL(bnxt_unregister_dev);
 
-static void bnxt_fill_msix_vecs(struct bnxt *bp, struct bnxt_msix_entry *ent)
-{
-	struct bnxt_en_dev *edev = bp->edev;
-	int num_msix, idx, i;
-
-	num_msix = edev->ulp_tbl->msix_requested;
-	idx = edev->ulp_tbl->msix_base;
-	for (i = 0; i < num_msix; i++) {
-		ent[i].vector = bp->irq_tbl[idx + i].vector;
-		ent[i].ring_idx = idx + i;
-		if (bp->flags & BNXT_FLAG_CHIP_P5) {
-			ent[i].db_offset = DB_PF_OFFSET_P5;
-			if (BNXT_VF(bp))
-				ent[i].db_offset = DB_VF_OFFSET_P5;
-		} else {
-			ent[i].db_offset = (idx + i) * 0x80;
-		}
-	}
-}
-
-int bnxt_req_msix_vecs(struct bnxt_en_dev *edev,
-			      struct bnxt_msix_entry *ent,
-			      int num_msix)
-{
-	struct net_device *dev = edev->net;
-	struct bnxt *bp = netdev_priv(dev);
-	struct bnxt_hw_resc *hw_resc;
-	int max_idx, max_cp_rings;
-	int avail_msix, idx;
-	int total_vecs;
-	int rc = 0;
-
-	if (!(bp->flags & BNXT_FLAG_USING_MSIX))
-		return -ENODEV;
-
-	if (edev->ulp_tbl->msix_requested)
-		return -EAGAIN;
-
-	max_cp_rings = bnxt_get_max_func_cp_rings(bp);
-	avail_msix = bnxt_get_avail_msix(bp, num_msix);
-	if (!avail_msix)
-		return -ENOMEM;
-	if (avail_msix > num_msix)
-		avail_msix = num_msix;
-
-	if (BNXT_NEW_RM(bp)) {
-		idx = bp->cp_nr_rings;
-	} else {
-		max_idx = min_t(int, bp->total_irqs, max_cp_rings);
-		idx = max_idx - avail_msix;
-	}
-
-	edev->ulp_tbl->msix_base = idx;
-	edev->ulp_tbl->msix_requested = avail_msix;
-	hw_resc = &bp->hw_resc;
-	total_vecs = idx + avail_msix;
-	rtnl_lock();
-	if (bp->total_irqs < total_vecs ||
-	    (BNXT_NEW_RM(bp) && hw_resc->resv_irqs < total_vecs)) {
-		if (netif_running(dev)) {
-			bnxt_close_nic(bp, true, false);
-			rc = bnxt_open_nic(bp, true, false);
-		} else {
-			rc = bnxt_reserve_rings(bp, true);
-		}
-	}
-	rtnl_unlock();
-	if (rc) {
-		edev->ulp_tbl->msix_requested = 0;
-		return -EAGAIN;
-	}
-
-	if (BNXT_NEW_RM(bp)) {
-		int resv_msix;
-
-		resv_msix = hw_resc->resv_irqs - bp->cp_nr_rings;
-		avail_msix = min_t(int, resv_msix, avail_msix);
-		edev->ulp_tbl->msix_requested = avail_msix;
-	}
-	bnxt_fill_msix_vecs(bp, ent);
-	edev->flags |= BNXT_EN_FLAG_MSIX_REQUESTED;
-	return avail_msix;
-}
-EXPORT_SYMBOL(bnxt_req_msix_vecs);
-
-void bnxt_free_msix_vecs(struct bnxt_en_dev *edev)
-{
-	struct net_device *dev = edev->net;
-	struct bnxt *bp = netdev_priv(dev);
-
-	if (!(edev->flags & BNXT_EN_FLAG_MSIX_REQUESTED))
-		return;
-
-	edev->ulp_tbl->msix_requested = 0;
-	edev->flags &= ~BNXT_EN_FLAG_MSIX_REQUESTED;
-	rtnl_lock();
-	if (netif_running(dev) && !(edev->flags & BNXT_EN_FLAG_ULP_STOPPED)) {
-		bnxt_close_nic(bp, true, false);
-		bnxt_open_nic(bp, true, false);
-	}
-	rtnl_unlock();
-
-	return;
-}
-EXPORT_SYMBOL(bnxt_free_msix_vecs);
-
 int bnxt_get_ulp_msix_num(struct bnxt *bp)
 {
-	if (bnxt_ulp_registered(bp->edev)) {
-		struct bnxt_en_dev *edev = bp->edev;
+	u32 roce_msix = BNXT_VF(bp) ?
+			BNXT_MAX_VF_ROCE_MSIX : BNXT_MAX_ROCE_MSIX;
 
-		return edev->ulp_tbl->msix_requested;
-	}
-	return 0;
+	return ((bp->flags & BNXT_FLAG_ROCE_CAP) && bp->aux_dev ?
+		min_t(u32, roce_msix, num_online_cpus()) : 0);
 }
 
 int bnxt_get_ulp_msix_base(struct bnxt *bp)
@@ -416,6 +332,8 @@ static void bnxt_aux_dev_release(struct device *dev)
 		container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
 	struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
 
+	kfree(bnxt_adev->edev->ulp_tbl);
+	bnxt_adev->edev->ulp_tbl = NULL;
 	kfree(bnxt_adev->edev);
 	bnxt_adev->edev = NULL;
 	bp->edev = NULL;
@@ -439,6 +357,8 @@ static void bnxt_set_edev_info(struct bnxt_en_dev *edev, struct bnxt *bp)
 	edev->hw_ring_stats_size = bp->hw_ring_stats_size;
 	edev->pf_port_id = bp->pf.port_id;
 	edev->en_state = bp->state;
+
+	edev->ulp_tbl->msix_requested = bnxt_get_ulp_msix_num(bp);
 }
 
 static int bnxt_rdma_aux_device_add(struct bnxt *bp)
@@ -446,12 +366,21 @@ static int bnxt_rdma_aux_device_add(struct bnxt *bp)
 	struct bnxt_aux_dev *bnxt_adev = bp->aux_dev;
 	struct bnxt_en_dev *edev = bnxt_adev->edev;
 	struct auxiliary_device *aux_dev;
+	struct bnxt_ulp *ulp;
 	int ret;
 
 	edev = kzalloc(sizeof(*edev), GFP_KERNEL);
 	if (!edev)
 		return -ENOMEM;
 
+	ulp = kzalloc(sizeof(*ulp), GFP_KERNEL);
+	if (!ulp) {
+		ret = -ENOMEM;
+		goto free_edev;
+	}
+
+	edev->ulp_tbl = ulp;
+
 	aux_dev = &bnxt_adev->aux_dev;
 	aux_dev->id = bnxt_adev->id;
 	aux_dev->name = "rdma";
@@ -464,7 +393,7 @@ static int bnxt_rdma_aux_device_add(struct bnxt *bp)
 
 	ret = auxiliary_device_init(aux_dev);
 	if (ret)
-		goto free_edev;
+		goto free_ulp;
 
 	ret = auxiliary_device_add(aux_dev);
 	if (ret)
@@ -473,6 +402,9 @@ static int bnxt_rdma_aux_device_add(struct bnxt *bp)
 	return 0;
 aux_dev_uninit:
 	auxiliary_device_uninit(aux_dev);
+free_ulp:
+	kfree(ulp);
+	bp->edev->ulp_tbl = NULL;
 free_edev:
 	kfree(edev);
 	bp->edev = NULL;
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
index 0e15b35c36dd..01a9dc336312 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.h
@@ -15,6 +15,8 @@
 
 #define BNXT_MIN_ROCE_CP_RINGS	2
 #define BNXT_MIN_ROCE_STAT_CTXS	1
+#define BNXT_MAX_ROCE_MSIX	9
+#define BNXT_MAX_VF_ROCE_MSIX	2
 
 struct hwrm_async_event_cmpl;
 struct bnxt;
@@ -51,6 +53,7 @@ struct bnxt_ulp {
 struct bnxt_en_dev {
 	struct net_device *net;
 	struct pci_dev *pdev;
+	struct bnxt_msix_entry			msix_entries[BNXT_MAX_ROCE_MSIX];
 	u32 flags;
 	#define BNXT_EN_FLAG_ROCEV1_CAP		0x1
 	#define BNXT_EN_FLAG_ROCEV2_CAP		0x2
@@ -102,9 +105,6 @@ void bnxt_aux_dev_free(struct bnxt *bp);
 int bnxt_register_dev(struct bnxt_en_dev *edev, struct bnxt_ulp_ops *ulp_ops,
 		      void *handle);
 void bnxt_unregister_dev(struct bnxt_en_dev *edev);
-int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, struct bnxt_msix_entry *ent,
-		       int num_msix);
-void bnxt_free_msix_vecs(struct bnxt_en_dev *edev);
 int bnxt_send_msg(struct bnxt_en_dev *edev, struct bnxt_fw_msg *fw_msg);
 int bnxt_register_async_events(struct bnxt_en_dev *edev,
 			       unsigned long *events_bmap, u16 max_id);
-- 
2.37.1 (Apple Git-137.1)


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

Re: [PATCH net-next v7 1/8] bnxt_en: Add auxiliary driver support

[Jakub Kicinski]

On Thu, 12 Jan 2023 12:29:32 -0800 Ajit Khaparde wrote:
> Add auxiliary driver support.
> An auxiliary device will be created if the hardware indicates
> support for RDMA.
> The bnxt_ulp_probe() function has been removed and a new
> bnxt_rdma_aux_device_add() function has been added.
> The bnxt_free_msix_vecs() and bnxt_req_msix_vecs() will now hold
> the RTNL lock when they call the bnxt_close_nic()and bnxt_open_nic()
> since the device close and open need to be protected under RTNL lock.
> The operations between the bnxt_en and bnxt_re will be protected
> using the en_ops_lock.
> This will be used by the bnxt_re driver in a follow-on patch
> to create ROCE interfaces.

> --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
> +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
> @@ -13178,6 +13178,9 @@ static void bnxt_remove_one(struct pci_dev *pdev)
>  	struct net_device *dev = pci_get_drvdata(pdev);
>  	struct bnxt *bp = netdev_priv(dev);
>  
> +	bnxt_rdma_aux_device_uninit(bp);
> +	bnxt_aux_dev_free(bp);

You still free bp->aux_dev synchronously..

> +void bnxt_aux_dev_free(struct bnxt *bp)
> +{
> +	kfree(bp->aux_dev);

.. here. Which is called on .remove of the PCI device.

> +	bp->aux_dev = NULL;
> +}
> +
> +static struct bnxt_aux_dev *bnxt_aux_dev_alloc(struct bnxt *bp)
> +{
> +	return kzalloc(sizeof(struct bnxt_aux_dev), GFP_KERNEL);
> +}
> +
> +void bnxt_rdma_aux_device_uninit(struct bnxt *bp)
> +{
> +	struct bnxt_aux_dev *bnxt_adev;
> +	struct auxiliary_device *adev;
> +
> +	/* Skip if no auxiliary device init was done. */
> +	if (!(bp->flags & BNXT_FLAG_ROCE_CAP))
> +		return;
> +
> +	bnxt_adev = bp->aux_dev;
> +	adev = &bnxt_adev->aux_dev;
> +	auxiliary_device_delete(adev);
> +	auxiliary_device_uninit(adev);
> +	if (bnxt_adev->id >= 0)
> +		ida_free(&bnxt_aux_dev_ids, bnxt_adev->id);
> +}
> +
> +static void bnxt_aux_dev_release(struct device *dev)
> +{
> +	struct bnxt_aux_dev *bnxt_adev =
> +		container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
> +	struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
> +
> +	bnxt_adev->edev->en_ops = NULL;
> +	kfree(bnxt_adev->edev);

And yet the reference counted "release" function accesses the bp->adev
like it must exist.

This seems odd to me - why do we need refcounting on devices at all 
if we can free them synchronously? To be clear - I'm not sure this is
wrong, just seems odd.

> +	bnxt_adev->edev = NULL;
> +	bp->edev = NULL;
> +}

Re: [PATCH net-next v7 1/8] bnxt_en: Add auxiliary driver support

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 2818 bytes --]

On Fri, Jan 13, 2023 at 10:10 PM Jakub Kicinski <kuba@kernel.org> wrote:
>
> On Thu, 12 Jan 2023 12:29:32 -0800 Ajit Khaparde wrote:
> > Add auxiliary driver support.
> > An auxiliary device will be created if the hardware indicates
> > support for RDMA.
> > The bnxt_ulp_probe() function has been removed and a new
> > bnxt_rdma_aux_device_add() function has been added.
> > The bnxt_free_msix_vecs() and bnxt_req_msix_vecs() will now hold
> > the RTNL lock when they call the bnxt_close_nic()and bnxt_open_nic()
> > since the device close and open need to be protected under RTNL lock.
> > The operations between the bnxt_en and bnxt_re will be protected
> > using the en_ops_lock.
> > This will be used by the bnxt_re driver in a follow-on patch
> > to create ROCE interfaces.
>
> > --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
> > +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
> > @@ -13178,6 +13178,9 @@ static void bnxt_remove_one(struct pci_dev *pdev)
> >       struct net_device *dev = pci_get_drvdata(pdev);
> >       struct bnxt *bp = netdev_priv(dev);
> >
> > +     bnxt_rdma_aux_device_uninit(bp);
> > +     bnxt_aux_dev_free(bp);
>
> You still free bp->aux_dev synchronously..
>
> > +void bnxt_aux_dev_free(struct bnxt *bp)
> > +{
> > +     kfree(bp->aux_dev);
>
> .. here. Which is called on .remove of the PCI device.
>
> > +     bp->aux_dev = NULL;
> > +}
> > +
> > +static struct bnxt_aux_dev *bnxt_aux_dev_alloc(struct bnxt *bp)
> > +{
> > +     return kzalloc(sizeof(struct bnxt_aux_dev), GFP_KERNEL);
> > +}
> > +
> > +void bnxt_rdma_aux_device_uninit(struct bnxt *bp)
> > +{
> > +     struct bnxt_aux_dev *bnxt_adev;
> > +     struct auxiliary_device *adev;
> > +
> > +     /* Skip if no auxiliary device init was done. */
> > +     if (!(bp->flags & BNXT_FLAG_ROCE_CAP))
> > +             return;
> > +
> > +     bnxt_adev = bp->aux_dev;
> > +     adev = &bnxt_adev->aux_dev;
> > +     auxiliary_device_delete(adev);
> > +     auxiliary_device_uninit(adev);
> > +     if (bnxt_adev->id >= 0)
> > +             ida_free(&bnxt_aux_dev_ids, bnxt_adev->id);
> > +}
> > +
> > +static void bnxt_aux_dev_release(struct device *dev)
> > +{
> > +     struct bnxt_aux_dev *bnxt_adev =
> > +             container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
> > +     struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
> > +
> > +     bnxt_adev->edev->en_ops = NULL;
> > +     kfree(bnxt_adev->edev);
>
> And yet the reference counted "release" function accesses the bp->adev
> like it must exist.
>
> This seems odd to me - why do we need refcounting on devices at all
> if we can free them synchronously? To be clear - I'm not sure this is
> wrong, just seems odd.
I followed the existing implementations in that regard. Thanks

>
> > +     bnxt_adev->edev = NULL;
> > +     bp->edev = NULL;
> > +}

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

Re: [PATCH net-next v7 1/8] bnxt_en: Add auxiliary driver support

[Jakub Kicinski]

On Sat, 14 Jan 2023 12:39:09 -0800 Ajit Khaparde wrote:
> > > +static void bnxt_aux_dev_release(struct device *dev)
> > > +{
> > > +     struct bnxt_aux_dev *bnxt_adev =
> > > +             container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
> > > +     struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
> > > +
> > > +     bnxt_adev->edev->en_ops = NULL;
> > > +     kfree(bnxt_adev->edev);  
> >
> > And yet the reference counted "release" function accesses the bp->adev
> > like it must exist.
> >
> > This seems odd to me - why do we need refcounting on devices at all
> > if we can free them synchronously? To be clear - I'm not sure this is
> > wrong, just seems odd.  
> I followed the existing implementations in that regard. Thanks

Leon, could you take a look? Is there no problem in assuming bnxt_adev
is still around in the release function?

Re: [PATCH net-next v7 1/8] bnxt_en: Add auxiliary driver support

[Leon Romanovsky]

On Mon, Jan 16, 2023 at 08:56:25PM -0800, Jakub Kicinski wrote:
> On Sat, 14 Jan 2023 12:39:09 -0800 Ajit Khaparde wrote:
> > > > +static void bnxt_aux_dev_release(struct device *dev)
> > > > +{
> > > > +     struct bnxt_aux_dev *bnxt_adev =
> > > > +             container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
> > > > +     struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
> > > > +
> > > > +     bnxt_adev->edev->en_ops = NULL;
> > > > +     kfree(bnxt_adev->edev);  
> > >
> > > And yet the reference counted "release" function accesses the bp->adev
> > > like it must exist.
> > >
> > > This seems odd to me - why do we need refcounting on devices at all
> > > if we can free them synchronously? To be clear - I'm not sure this is
> > > wrong, just seems odd.  
> > I followed the existing implementations in that regard. Thanks
> 
> Leon, could you take a look? Is there no problem in assuming bnxt_adev
> is still around in the release function?

You caught a real bug. The auxdev idea is very simple - it needs to
behave like driver core, but in the driver itself.

As such, bnxt_aux_dev_free() shouldn't be called after bnxt_rdma_aux_device_uninit().
Device will be released through auxiliary_device_uninit();

BTW, line 325 from below shouldn't exist too.

  312 void bnxt_rdma_aux_device_uninit(struct bnxt *bp)
  313 {
...
  325         if (bnxt_adev->id >= 0)
  326                 ida_free(&bnxt_aux_dev_ids, bnxt_adev->id);

And one line bnxt_aux_dev_alloc() needs to be deleted too.

Thanks

Re: [PATCH net-next v7 1/8] bnxt_en: Add auxiliary driver support

[Leon Romanovsky]

On Tue, Jan 17, 2023 at 02:31:01PM +0200, Leon Romanovsky wrote:
> On Mon, Jan 16, 2023 at 08:56:25PM -0800, Jakub Kicinski wrote:
> > On Sat, 14 Jan 2023 12:39:09 -0800 Ajit Khaparde wrote:
> > > > > +static void bnxt_aux_dev_release(struct device *dev)
> > > > > +{
> > > > > +     struct bnxt_aux_dev *bnxt_adev =
> > > > > +             container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
> > > > > +     struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
> > > > > +
> > > > > +     bnxt_adev->edev->en_ops = NULL;
> > > > > +     kfree(bnxt_adev->edev);  
> > > >
> > > > And yet the reference counted "release" function accesses the bp->adev
> > > > like it must exist.
> > > >
> > > > This seems odd to me - why do we need refcounting on devices at all
> > > > if we can free them synchronously? To be clear - I'm not sure this is
> > > > wrong, just seems odd.  
> > > I followed the existing implementations in that regard. Thanks
> > 
> > Leon, could you take a look? Is there no problem in assuming bnxt_adev
> > is still around in the release function?
> 
> You caught a real bug. The auxdev idea is very simple - it needs to
> behave like driver core, but in the driver itself.

BTW, this can be classic example why assigning NULL pointers after
release is bad practice. It hides this class of errors.

+void bnxt_aux_dev_free(struct bnxt *bp)
+{
+       kfree(bp->aux_dev);
+       bp->aux_dev = NULL;
+}

Thanks

Re: [PATCH net-next v7 1/8] bnxt_en: Add auxiliary driver support

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 1728 bytes --]

On Tue, Jan 17, 2023 at 4:31 AM Leon Romanovsky <leonro@nvidia.com> wrote:
>
> On Mon, Jan 16, 2023 at 08:56:25PM -0800, Jakub Kicinski wrote:
> > On Sat, 14 Jan 2023 12:39:09 -0800 Ajit Khaparde wrote:
> > > > > +static void bnxt_aux_dev_release(struct device *dev)
> > > > > +{
> > > > > +     struct bnxt_aux_dev *bnxt_adev =
> > > > > +             container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
> > > > > +     struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
> > > > > +
> > > > > +     bnxt_adev->edev->en_ops = NULL;
> > > > > +     kfree(bnxt_adev->edev);
> > > >
> > > > And yet the reference counted "release" function accesses the bp->adev
> > > > like it must exist.
> > > >
> > > > This seems odd to me - why do we need refcounting on devices at all
> > > > if we can free them synchronously? To be clear - I'm not sure this is
> > > > wrong, just seems odd.
> > > I followed the existing implementations in that regard. Thanks
> >
> > Leon, could you take a look? Is there no problem in assuming bnxt_adev
> > is still around in the release function?
>
> You caught a real bug. The auxdev idea is very simple - it needs to
> behave like driver core, but in the driver itself.
>
> As such, bnxt_aux_dev_free() shouldn't be called after bnxt_rdma_aux_device_uninit().
> Device will be released through auxiliary_device_uninit();
>
> BTW, line 325 from below shouldn't exist too.
>
>   312 void bnxt_rdma_aux_device_uninit(struct bnxt *bp)
>   313 {
> ...
>   325         if (bnxt_adev->id >= 0)
>   326                 ida_free(&bnxt_aux_dev_ids, bnxt_adev->id);
>
> And one line bnxt_aux_dev_alloc() needs to be deleted too.
>
> Thanks
Thanks.
We are reviewing the comments and will have an update soon.

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]

Re: [PATCH net-next v7 1/8] bnxt_en: Add auxiliary driver support

[Ajit Khaparde]

[-- Attachment #1: Type: text/plain, Size: 2088 bytes --]

On Tue, Jan 17, 2023 at 4:31 AM Leon Romanovsky <leonro@nvidia.com> wrote:
>
> On Mon, Jan 16, 2023 at 08:56:25PM -0800, Jakub Kicinski wrote:
> > On Sat, 14 Jan 2023 12:39:09 -0800 Ajit Khaparde wrote:
> > > > > +static void bnxt_aux_dev_release(struct device *dev)
> > > > > +{
> > > > > +     struct bnxt_aux_dev *bnxt_adev =
> > > > > +             container_of(dev, struct bnxt_aux_dev, aux_dev.dev);
> > > > > +     struct bnxt *bp = netdev_priv(bnxt_adev->edev->net);
> > > > > +
> > > > > +     bnxt_adev->edev->en_ops = NULL;
> > > > > +     kfree(bnxt_adev->edev);
> > > >
> > > > And yet the reference counted "release" function accesses the bp->adev
> > > > like it must exist.
> > > >
> > > > This seems odd to me - why do we need refcounting on devices at all
> > > > if we can free them synchronously? To be clear - I'm not sure this is
> > > > wrong, just seems odd.
> > > I followed the existing implementations in that regard. Thanks
> >
> > Leon, could you take a look? Is there no problem in assuming bnxt_adev
> > is still around in the release function?
>
> You caught a real bug. The auxdev idea is very simple - it needs to
> behave like driver core, but in the driver itself.
>
> As such, bnxt_aux_dev_free() shouldn't be called after bnxt_rdma_aux_device_uninit().
> Device will be released through auxiliary_device_uninit();
Ok. But..
bnxt_aux_dev_free() is actually freeing up the private memory allocated
for holding the pointer returned by my_aux_dev_alloc(xxx);
The aux device is freed via the auxiliary_device_uninit only.

>
> BTW, line 325 from below shouldn't exist too.
ACK

>
>   312 void bnxt_rdma_aux_device_uninit(struct bnxt *bp)
>   313 {
> ...
>   325         if (bnxt_adev->id >= 0)
>   326                 ida_free(&bnxt_aux_dev_ids, bnxt_adev->id);
>
> And one line bnxt_aux_dev_alloc() needs to be deleted too.
To avoid confusion, I will refactor and rename the code handling
auxiliary_device alloc, cleanup and the alloc, cleanup of priv
pointers used for bookkeeping.

I hope the new patchset will address the concerns raised.

>
> Thanks

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]