* [net-next 8/9] ixgbevf: Remove unneeded and obsolete comment
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352990387-3872-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 7 -------
1 file changed, 7 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 57ae5cd..a52b14e 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1681,13 +1681,6 @@ void ixgbevf_reinit_locked(struct ixgbevf_adapter *adapter)
while (test_and_set_bit(__IXGBEVF_RESETTING, &adapter->state))
msleep(1);
- /*
- * Check if PF is up before re-init. If not then skip until
- * later when the PF is up and ready to service requests from
- * the VF via mailbox. If the VF is up and running then the
- * watchdog task will continue to schedule reset tasks until
- * the PF is up and running.
- */
ixgbevf_down(adapter);
ixgbevf_up(adapter);
--
1.7.11.7
^ permalink raw reply related
* [net-next 6/9] ixgbevf: Remove mailbox spinlock from the reset function
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352990387-3872-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
The spinlocks are not required during reset. There won't be any
contention for the mailbox resource.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 4 ----
1 file changed, 4 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 8b8a685..592fe99 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1702,15 +1702,11 @@ void ixgbevf_reset(struct ixgbevf_adapter *adapter)
struct ixgbe_hw *hw = &adapter->hw;
struct net_device *netdev = adapter->netdev;
- spin_lock_bh(&adapter->mbx_lock);
-
if (hw->mac.ops.reset_hw(hw))
hw_dbg(hw, "PF still resetting\n");
else
hw->mac.ops.init_hw(hw);
- spin_unlock_bh(&adapter->mbx_lock);
-
if (is_valid_ether_addr(adapter->hw.mac.addr)) {
memcpy(netdev->dev_addr, adapter->hw.mac.addr,
netdev->addr_len);
--
1.7.11.7
^ permalink raw reply related
* [net-next 7/9] ixgbevf: White space and comments clean up
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352990387-3872-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 19 ++++++-------------
1 file changed, 6 insertions(+), 13 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 592fe99..57ae5cd 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -121,7 +121,6 @@ static inline void ixgbevf_release_rx_desc(struct ixgbe_hw *hw,
* @direction: 0 for Rx, 1 for Tx, -1 for other causes
* @queue: queue to map the corresponding interrupt to
* @msix_vector: the vector to map to the corresponding queue
- *
*/
static void ixgbevf_set_ivar(struct ixgbevf_adapter *adapter, s8 direction,
u8 queue, u8 msix_vector)
@@ -380,7 +379,6 @@ static void ixgbevf_alloc_rx_buffers(struct ixgbevf_adapter *adapter,
no_buffers:
if (rx_ring->next_to_use != i) {
rx_ring->next_to_use = i;
-
ixgbevf_release_rx_desc(&adapter->hw, rx_ring, i);
}
}
@@ -765,7 +763,6 @@ static irqreturn_t ixgbevf_msix_other(int irq, void *data)
return IRQ_HANDLED;
}
-
/**
* ixgbevf_msix_clean_rings - single unshared vector rx clean (all queues)
* @irq: unused
@@ -1224,12 +1221,13 @@ static int ixgbevf_write_uc_addr_list(struct net_device *netdev)
}
/**
- * ixgbevf_set_rx_mode - Multicast set
+ * ixgbevf_set_rx_mode - Multicast and unicast set
* @netdev: network interface device structure
*
* The set_rx_method entry point is called whenever the multicast address
- * list or the network interface flags are updated. This routine is
- * responsible for configuring the hardware for proper multicast mode.
+ * list, unicast address list or the network interface flags are updated.
+ * This routine is responsible for configuring the hardware for proper
+ * multicast mode and configuring requested unicast filters.
**/
static void ixgbevf_set_rx_mode(struct net_device *netdev)
{
@@ -1588,7 +1586,6 @@ static void ixgbevf_clean_tx_ring(struct ixgbevf_adapter *adapter,
return;
/* Free all the Tx ring sk_buffs */
-
for (i = 0; i < tx_ring->count; i++) {
tx_buffer_info = &tx_ring->tx_buffer_info[i];
ixgbevf_unmap_and_free_tx_resource(tx_ring, tx_buffer_info);
@@ -1757,6 +1754,7 @@ static int ixgbevf_acquire_msix_vectors(struct ixgbevf_adapter *adapter,
*/
adapter->num_msix_vectors = vectors;
}
+
return err;
}
@@ -2053,7 +2051,7 @@ static int __devinit ixgbevf_sw_init(struct ixgbevf_adapter *adapter)
goto out;
}
memcpy(adapter->netdev->dev_addr, adapter->hw.mac.addr,
- adapter->netdev->addr_len);
+ adapter->netdev->addr_len);
}
/* lock to protect mailbox accesses */
@@ -2198,7 +2196,6 @@ static void ixgbevf_watchdog_task(struct work_struct *work)
* Always check the link on the watchdog because we have
* no LSC interrupt
*/
-
spin_lock_bh(&adapter->mbx_lock);
need_reset = hw->mac.ops.check_link(hw, &link_speed, &link_up, false);
@@ -2704,9 +2701,6 @@ static int ixgbevf_tso(struct ixgbevf_ring *tx_ring,
static bool ixgbevf_tx_csum(struct ixgbevf_ring *tx_ring,
struct sk_buff *skb, u32 tx_flags)
{
-
-
-
u32 vlan_macip_lens = 0;
u32 mss_l4len_idx = 0;
u32 type_tucmd = 0;
@@ -2896,7 +2890,6 @@ static void ixgbevf_tx_queue(struct ixgbevf_ring *tx_ring, int tx_flags,
olinfo_status |= (1 << IXGBE_ADVTXD_IDX_SHIFT);
if (tx_flags & IXGBE_TX_FLAGS_IPV4)
olinfo_status |= IXGBE_ADVTXD_POPTS_IXSM;
-
}
/*
--
1.7.11.7
^ permalink raw reply related
* [net-next 4/9] ixgbevf: Remove the ring adapter pointer value
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352990387-3872-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
It is unused - remove it.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index 1211fa0..dbdf39b 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -58,7 +58,6 @@ struct ixgbevf_ring {
struct ixgbevf_ring *next;
struct net_device *netdev;
struct device *dev;
- struct ixgbevf_adapter *adapter; /* backlink */
void *desc; /* descriptor ring memory */
dma_addr_t dma; /* phys. address of descriptor ring */
unsigned int size; /* length in bytes */
--
1.7.11.7
^ permalink raw reply related
* [net-next 5/9] ixgbevf: Remove checking for mac.ops function pointers
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352990387-3872-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
The function pointers will always be set - there is no good reason to
check them. Also just remove get_bus_info() call as the VF has no bus
info to report.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 56 ++++++++---------------
1 file changed, 18 insertions(+), 38 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index a001684..8b8a685 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1150,9 +1150,6 @@ static int ixgbevf_vlan_rx_add_vid(struct net_device *netdev, u16 vid)
struct ixgbe_hw *hw = &adapter->hw;
int err;
- if (!hw->mac.ops.set_vfta)
- return -EOPNOTSUPP;
-
spin_lock_bh(&adapter->mbx_lock);
/* add VID to filter table */
@@ -1181,8 +1178,7 @@ static int ixgbevf_vlan_rx_kill_vid(struct net_device *netdev, u16 vid)
spin_lock_bh(&adapter->mbx_lock);
/* remove VID from filter table */
- if (hw->mac.ops.set_vfta)
- err = hw->mac.ops.set_vfta(hw, vid, 0, false);
+ err = hw->mac.ops.set_vfta(hw, vid, 0, false);
spin_unlock_bh(&adapter->mbx_lock);
@@ -1243,8 +1239,7 @@ static void ixgbevf_set_rx_mode(struct net_device *netdev)
spin_lock_bh(&adapter->mbx_lock);
/* reprogram multicast list */
- if (hw->mac.ops.update_mc_addr_list)
- hw->mac.ops.update_mc_addr_list(hw, netdev);
+ hw->mac.ops.update_mc_addr_list(hw, netdev);
ixgbevf_write_uc_addr_list(netdev);
@@ -1414,12 +1409,10 @@ static void ixgbevf_up_complete(struct ixgbevf_adapter *adapter)
spin_lock_bh(&adapter->mbx_lock);
- if (hw->mac.ops.set_rar) {
- if (is_valid_ether_addr(hw->mac.addr))
- hw->mac.ops.set_rar(hw, 0, hw->mac.addr, 0);
- else
- hw->mac.ops.set_rar(hw, 0, hw->mac.perm_addr, 0);
- }
+ if (is_valid_ether_addr(hw->mac.addr))
+ hw->mac.ops.set_rar(hw, 0, hw->mac.addr, 0);
+ else
+ hw->mac.ops.set_rar(hw, 0, hw->mac.perm_addr, 0);
spin_unlock_bh(&adapter->mbx_lock);
@@ -2201,6 +2194,7 @@ static void ixgbevf_watchdog_task(struct work_struct *work)
struct ixgbe_hw *hw = &adapter->hw;
u32 link_speed = adapter->link_speed;
bool link_up = adapter->link_up;
+ s32 need_reset;
adapter->flags |= IXGBE_FLAG_IN_WATCHDOG_TASK;
@@ -2208,29 +2202,20 @@ static void ixgbevf_watchdog_task(struct work_struct *work)
* Always check the link on the watchdog because we have
* no LSC interrupt
*/
- if (hw->mac.ops.check_link) {
- s32 need_reset;
- spin_lock_bh(&adapter->mbx_lock);
+ spin_lock_bh(&adapter->mbx_lock);
- need_reset = hw->mac.ops.check_link(hw, &link_speed,
- &link_up, false);
+ need_reset = hw->mac.ops.check_link(hw, &link_speed, &link_up, false);
- spin_unlock_bh(&adapter->mbx_lock);
+ spin_unlock_bh(&adapter->mbx_lock);
- if (need_reset) {
- adapter->link_up = link_up;
- adapter->link_speed = link_speed;
- netif_carrier_off(netdev);
- netif_tx_stop_all_queues(netdev);
- schedule_work(&adapter->reset_task);
- goto pf_has_reset;
- }
- } else {
- /* always assume link is up, if no check link
- * function */
- link_speed = IXGBE_LINK_SPEED_10GB_FULL;
- link_up = true;
+ if (need_reset) {
+ adapter->link_up = link_up;
+ adapter->link_speed = link_speed;
+ netif_carrier_off(netdev);
+ netif_tx_stop_all_queues(netdev);
+ schedule_work(&adapter->reset_task);
+ goto pf_has_reset;
}
adapter->link_up = link_up;
adapter->link_speed = link_speed;
@@ -3070,8 +3055,7 @@ static int ixgbevf_set_mac(struct net_device *netdev, void *p)
spin_lock_bh(&adapter->mbx_lock);
- if (hw->mac.ops.set_rar)
- hw->mac.ops.set_rar(hw, 0, hw->mac.addr, 0);
+ hw->mac.ops.set_rar(hw, 0, hw->mac.addr, 0);
spin_unlock_bh(&adapter->mbx_lock);
@@ -3396,10 +3380,6 @@ static int __devinit ixgbevf_probe(struct pci_dev *pdev,
if (err)
goto err_sw_init;
- /* pick up the PCI bus settings for reporting later */
- if (hw->mac.ops.get_bus_info)
- hw->mac.ops.get_bus_info(hw);
-
strcpy(netdev->name, "eth%d");
err = register_netdev(netdev);
--
1.7.11.7
^ permalink raw reply related
* [net-next 2/9] ixgbevf: Streamline the rx buffer allocation
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352990387-3872-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
Moves allocation of local variable to section where it is needed and
removes unnecessary if statement.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 9d88153..b46dff8 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -341,15 +341,16 @@ static void ixgbevf_alloc_rx_buffers(struct ixgbevf_adapter *adapter,
struct pci_dev *pdev = adapter->pdev;
union ixgbe_adv_rx_desc *rx_desc;
struct ixgbevf_rx_buffer *bi;
- struct sk_buff *skb;
unsigned int i = rx_ring->next_to_use;
bi = &rx_ring->rx_buffer_info[i];
while (cleaned_count--) {
rx_desc = IXGBEVF_RX_DESC(rx_ring, i);
- skb = bi->skb;
- if (!skb) {
+
+ if (!bi->skb) {
+ struct sk_buff *skb;
+
skb = netdev_alloc_skb_ip_align(rx_ring->netdev,
rx_ring->rx_buf_len);
if (!skb) {
@@ -357,8 +358,7 @@ static void ixgbevf_alloc_rx_buffers(struct ixgbevf_adapter *adapter,
goto no_buffers;
}
bi->skb = skb;
- }
- if (!bi->dma) {
+
bi->dma = dma_map_single(&pdev->dev, skb->data,
rx_ring->rx_buf_len,
DMA_FROM_DEVICE);
--
1.7.11.7
^ permalink raw reply related
* [net-next 3/9] ixgbevf: Fix unnecessary dereference where local var is available.
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352990387-3872-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
Remove dereference of hw pointer from adapter structure since a pointer
to the hw structure has already been allocated off the stack. Also clean
up useless parenthesis.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index b46dff8..a001684 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1312,8 +1312,8 @@ static inline void ixgbevf_rx_desc_queue_enable(struct ixgbevf_adapter *adapter,
"not set within the polling period\n", rxr);
}
- ixgbevf_release_rx_desc(&adapter->hw, &adapter->rx_ring[rxr],
- (adapter->rx_ring[rxr].count - 1));
+ ixgbevf_release_rx_desc(hw, &adapter->rx_ring[rxr],
+ adapter->rx_ring[rxr].count - 1);
}
static void ixgbevf_save_reset_stats(struct ixgbevf_adapter *adapter)
--
1.7.11.7
^ permalink raw reply related
* [net-next 1/9] ioat: Do not enable DCA if tag map is invalid
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Alexander Duyck, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352990387-3872-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Alexander Duyck <alexander.h.duyck@intel.com>
I have encountered several systems that have an invalid tag map. This
invalid tag map results in only two tags being generated 0x1F which is
usually applied to the first core in a Hyper-threaded pair and 0x00 which
is applied to the second core in a Hyper-threaded pair. The net result of
all this is that DCA causes significant cache thrash because the 0x1F tag
will send traffic to the second socket, which the 0x00 tag will not DCA tag
the frame resulting in no benefit.
For now the best solution from the driver's perspective is to just disable
DCA if the tag map is invalid. The correct solution for this would be to
have the BIOS on affected systems updated so that the correct tags are
generated for a given APIC ID.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/dma/ioat/dca.c | 23 +++++++++++++++++++++++
1 file changed, 23 insertions(+)
diff --git a/drivers/dma/ioat/dca.c b/drivers/dma/ioat/dca.c
index abd9038..d666807 100644
--- a/drivers/dma/ioat/dca.c
+++ b/drivers/dma/ioat/dca.c
@@ -604,6 +604,23 @@ static int ioat3_dca_count_dca_slots(void *iobase, u16 dca_offset)
return slots;
}
+static inline int dca3_tag_map_invalid(u8 *tag_map)
+{
+ /*
+ * If the tag map is not programmed by the BIOS the default is:
+ * 0x80 0x80 0x80 0x80 0x80 0x00 0x00 0x00
+ *
+ * This an invalid map and will result in only 2 possible tags
+ * 0x1F and 0x00. 0x00 is an invalid DCA tag so we know that
+ * this entire definition is invalid.
+ */
+ return ((tag_map[0] == DCA_TAG_MAP_VALID) &&
+ (tag_map[1] == DCA_TAG_MAP_VALID) &&
+ (tag_map[2] == DCA_TAG_MAP_VALID) &&
+ (tag_map[3] == DCA_TAG_MAP_VALID) &&
+ (tag_map[4] == DCA_TAG_MAP_VALID));
+}
+
struct dca_provider * __devinit
ioat3_dca_init(struct pci_dev *pdev, void __iomem *iobase)
{
@@ -674,6 +691,12 @@ ioat3_dca_init(struct pci_dev *pdev, void __iomem *iobase)
ioatdca->tag_map[i] = bit & DCA_TAG_MAP_MASK;
}
+ if (dca3_tag_map_invalid(ioatdca->tag_map)) {
+ dev_err(&pdev->dev, "APICID_TAG_MAP set incorrectly by BIOS, disabling DCA\n");
+ free_dca_provider(dca);
+ return NULL;
+ }
+
err = register_dca_provider(dca, &pdev->dev);
if (err) {
free_dca_provider(dca);
--
1.7.11.7
^ permalink raw reply related
* [net-next 0/9][pull request] Intel Wired LAN Driver Updates
From: Jeff Kirsher @ 2012-11-15 14:39 UTC (permalink / raw)
To: davem; +Cc: Jeff Kirsher, netdev, gospo, sassmann
This series contains updates to ioat (DCA) and ixgbevf.
The following are changes since commit 702ed3c1a9dfe4dfe112f13542d0c9d689f5008b:
Merge tag 'batman-adv-for-davem' of git://git.open-mesh.org/linux-merge
and are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
Alexander Duyck (1):
ioat: Do not enable DCA if tag map is invalid
Greg Rose (8):
ixgbevf: Streamline the rx buffer allocation
ixgbevf: Fix unnecessary dereference where local var is available.
ixgbevf: Remove the ring adapter pointer value
ixgbevf: Remove checking for mac.ops function pointers
ixgbevf: Remove mailbox spinlock from the reset function
ixgbevf: White space and comments clean up
ixgbevf: Remove unneeded and obsolete comment
ixgbevf: Add checksum statistics counters to rings
drivers/dma/ioat/dca.c | 23 ++++
drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 3 +-
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 123 +++++++++-------------
3 files changed, 72 insertions(+), 77 deletions(-)
--
1.7.11.7
^ permalink raw reply
* Re: [PATCH 0/9 v4] use efficient this_cpu_* helper
From: Christoph Lameter @ 2012-11-15 14:19 UTC (permalink / raw)
To: Tejun Heo; +Cc: David Miller, NetDev, Shan Wei, Kernel-Maillist
In-Reply-To: <50A1A7A9.6050501@gmail.com>
[-- Attachment #1: Type: TEXT/PLAIN, Size: 2786 bytes --]
Tejon: Could you pick up this patchset?
On Tue, 13 Nov 2012, Shan Wei wrote:
> this_cpu_ptr/this_cpu_read is faster than per_cpu_ptr(p, smp_processor_id())
> and can reduce memory accesses.
> The latter helper needs to find the offset for current cpu,
> and needs more assembler instructions which objdump shows in following.
>
> this_cpu_ptr relocates and address. this_cpu_read() relocates the address
> and performs the fetch. If you want to operate on rda(defined as per_cpu)
> then you can only use this_cpu_ptr. this_cpu_read() saves you more instructions
> since it can do the relocation and the fetch in one instruction.
>
> per_cpu_ptr(p, smp_processor_id())：
> 1e: 65 8b 04 25 00 00 00 00 mov %gs:0x0,%eax
> 26: 48 98 cltq
> 28: 31 f6 xor %esi,%esi
> 2a: 48 c7 c7 00 00 00 00 mov $0x0,%rdi
> 31: 48 8b 04 c5 00 00 00 00 mov 0x0(,%rax,8),%rax
> 39: c7 44 10 04 14 00 00 00 movl $0x14,0x4(%rax,%rdx,1)
>
> this_cpu_ptr(p)
> 1e: 65 48 03 14 25 00 00 00 00 add %gs:0x0,%rdx
> 27: 31 f6 xor %esi,%esi
> 29: c7 42 04 14 00 00 00 movl $0x14,0x4(%rdx)
> 30: 48 c7 c7 00 00 00 00 mov $0x0,%rdi
>
>
>
> Changelog V4:
> 1. [read|write]ing fields of struct rds_ib_cache_head using __this_cpu_* operation for rds subsystem.
> see patch2
> 2. fix bug in xfrm to read pointer. see patch3.
> 3. avoid type cast in patch7.
>
> Changelog V3:
> 1. use this_cpu_read directly read member of per-cpu variable,
> so that droping the this_cpu_ptr operation.
> 2. for preemption off and bottom halves off case,
> use __this_cpu_read instead of this_cpu_read.
>
> Changelog V2:
> 1. Use this_cpu_read directly instead of ref to field of per-cpu variable.
> 2. Patch5 about ftrace is dropped from this series.
> 3. Add new patch9 to replace get_cpu;per_cpu_ptr;put_cpu with this_cpu_add opt.
> 4. For preemption disable case, use __this_cpu_read instead.
>
>
> $ git diff --stat d4185bbf62a5d8d777ee445db1581beb17882a07
> drivers/clocksource/arm_generic.c | 2 +-
> kernel/padata.c | 5 ++---
> kernel/rcutree.c | 2 +-
> kernel/trace/blktrace.c | 2 +-
> kernel/trace/trace.c | 5 +----
> net/batman-adv/main.h | 4 +---
> net/core/flow.c | 4 +---
> net/openvswitch/datapath.c | 4 ++--
> net/openvswitch/vport.c | 5 ++---
> net/rds/ib.h | 2 +-
> net/rds/ib_recv.c | 24 +++++++++++++-----------
> net/xfrm/xfrm_ipcomp.c | 8 +++-----
> 12 files changed, 29 insertions(+), 38 deletions(-)
>
^ permalink raw reply
* Re: [PATCH] tcp: fix retransmission in repair mode
From: Pavel Emelyanov @ 2012-11-15 14:10 UTC (permalink / raw)
To: Andrey Vagin, David S. Miller
Cc: netdev, criu, linux-kernel, Alexey Kuznetsov, James Morris,
Hideaki YOSHIFUJI, Patrick McHardy
In-Reply-To: <1352988197-14414-1-git-send-email-avagin@openvz.org>
On 11/15/2012 06:03 PM, Andrey Vagin wrote:
> From: Andrew Vagin <avagin@openvz.org>
>
> Currently if a socket was repaired with a few packet in a write queue,
> a kernel bug may be triggered:
>
> kernel BUG at net/ipv4/tcp_output.c:2330!
> RIP: 0010:[<ffffffff8155784f>] tcp_retransmit_skb+0x5ff/0x610
>
> According to the initial realization v3.4-rc2-963-gc0e88ff,
> all skb-s should look like already posted. This patch fixes code
> according with this sentence.
>
> Here are three points, which were not done in the initial patch:
> 1. A tcp send head should not be changed
> 2. Initialize TSO state of a skb
> 3. Reset the retransmission time
>
> This patch moves logic from tcp_sendmsg to tcp_write_xmit. A packet
> passes the ussual way, but isn't sent to network. This patch solves
> all described problems and handles tcp_sendpages.
>
> Cc: Pavel Emelyanov <xemul@parallels.com>
> Cc: "David S. Miller" <davem@davemloft.net>
> Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
> Cc: James Morris <jmorris@namei.org>
> Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
> Cc: Patrick McHardy <kaber@trash.net>
> Signed-off-by: Andrey Vagin <avagin@openvz.org>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
^ permalink raw reply
* [PATCH net-next v2 1/3] ipip: fix sparse warnings in ipip_netlink_parms()
From: Nicolas Dichtel @ 2012-11-15 14:06 UTC (permalink / raw)
To: eric.dumazet; +Cc: netdev, davem, fengguang.wu, Nicolas Dichtel
In-Reply-To: <1352985602.4497.31.camel@edumazet-glaptop>
This change fixes two sparse warnings triggered by casting the ip addresses
from netlink messages in an u32 instead of be32. This change corrects that
in order to resolve the sparse warnings.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
v2: enhance commit log and add "Reported-by" line
net/ipv4/ipip.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c
index 64686e1..c26c171 100644
--- a/net/ipv4/ipip.c
+++ b/net/ipv4/ipip.c
@@ -864,10 +864,10 @@ static void ipip_netlink_parms(struct nlattr *data[],
parms->link = nla_get_u32(data[IFLA_IPTUN_LINK]);
if (data[IFLA_IPTUN_LOCAL])
- parms->iph.saddr = nla_get_u32(data[IFLA_IPTUN_LOCAL]);
+ parms->iph.saddr = nla_get_be32(data[IFLA_IPTUN_LOCAL]);
if (data[IFLA_IPTUN_REMOTE])
- parms->iph.daddr = nla_get_u32(data[IFLA_IPTUN_REMOTE]);
+ parms->iph.daddr = nla_get_be32(data[IFLA_IPTUN_REMOTE]);
if (data[IFLA_IPTUN_TTL]) {
parms->iph.ttl = nla_get_u8(data[IFLA_IPTUN_TTL]);
--
1.7.12
^ permalink raw reply related
* [PATCH net-next v2 2/3] sit: fix sparse warnings
From: Nicolas Dichtel @ 2012-11-15 14:06 UTC (permalink / raw)
To: eric.dumazet; +Cc: netdev, davem, fengguang.wu, Nicolas Dichtel
In-Reply-To: <1352988402-16950-1-git-send-email-nicolas.dichtel@6wind.com>
This change fixes several sparse warnings about endianness problem. The wrong
nla_*() functions were used.
It also fix a sparse warning about a flag test (field i_flags). This field is
used in this file like a local flag only, so it is more an u16 (gre uses it as a
be16). This sparse warning was already there before the patch that add netlink
management, the code has just been moved.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
v2: enhance commit log and add "Reported-by" line
net/ipv6/sit.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c
index 7bd2a06..ca6c2c8 100644
--- a/net/ipv6/sit.c
+++ b/net/ipv6/sit.c
@@ -228,7 +228,7 @@ static int ipip6_tunnel_create(struct net_device *dev)
goto out;
ipip6_tunnel_clone_6rd(dev, sitn);
- if (t->parms.i_flags & SIT_ISATAP)
+ if ((__force u16)t->parms.i_flags & SIT_ISATAP)
dev->priv_flags |= IFF_ISATAP;
err = register_netdevice(dev);
@@ -1240,10 +1240,10 @@ static void ipip6_netlink_parms(struct nlattr *data[],
parms->link = nla_get_u32(data[IFLA_IPTUN_LINK]);
if (data[IFLA_IPTUN_LOCAL])
- parms->iph.saddr = nla_get_u32(data[IFLA_IPTUN_LOCAL]);
+ parms->iph.saddr = nla_get_be32(data[IFLA_IPTUN_LOCAL]);
if (data[IFLA_IPTUN_REMOTE])
- parms->iph.daddr = nla_get_u32(data[IFLA_IPTUN_REMOTE]);
+ parms->iph.daddr = nla_get_be32(data[IFLA_IPTUN_REMOTE]);
if (data[IFLA_IPTUN_TTL]) {
parms->iph.ttl = nla_get_u8(data[IFLA_IPTUN_TTL]);
@@ -1258,7 +1258,7 @@ static void ipip6_netlink_parms(struct nlattr *data[],
parms->iph.frag_off = htons(IP_DF);
if (data[IFLA_IPTUN_FLAGS])
- parms->i_flags = nla_get_u16(data[IFLA_IPTUN_FLAGS]);
+ parms->i_flags = nla_get_be16(data[IFLA_IPTUN_FLAGS]);
}
static int ipip6_newlink(struct net *src_net, struct net_device *dev,
@@ -1337,7 +1337,7 @@ static int ipip6_fill_info(struct sk_buff *skb, const struct net_device *dev)
nla_put_u8(skb, IFLA_IPTUN_TOS, parm->iph.tos) ||
nla_put_u8(skb, IFLA_IPTUN_PMTUDISC,
!!(parm->iph.frag_off & htons(IP_DF))) ||
- nla_put_u16(skb, IFLA_IPTUN_FLAGS, parm->i_flags))
+ nla_put_be16(skb, IFLA_IPTUN_FLAGS, parm->i_flags))
goto nla_put_failure;
return 0;
--
1.7.12
^ permalink raw reply related
* [PATCH net-next v2 3/3] ip6tnl: fix sparse warnings in ip6_tnl_netlink_parms()
From: Nicolas Dichtel @ 2012-11-15 14:06 UTC (permalink / raw)
To: eric.dumazet; +Cc: netdev, davem, fengguang.wu, Nicolas Dichtel
In-Reply-To: <1352988402-16950-1-git-send-email-nicolas.dichtel@6wind.com>
This change fixes a sparse warning triggered by casting the flowinfo from
netlink messages in an u32 instead of be32. This change corrects that in order
to resolve the sparse warning.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
v2: enhance commit log and add "Reported-by" line
net/ipv6/ip6_tunnel.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index ab4d056..bf3a549 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -1568,7 +1568,7 @@ static void ip6_tnl_netlink_parms(struct nlattr *data[],
parms->encap_limit = nla_get_u8(data[IFLA_IPTUN_ENCAP_LIMIT]);
if (data[IFLA_IPTUN_FLOWINFO])
- parms->flowinfo = nla_get_u32(data[IFLA_IPTUN_FLOWINFO]);
+ parms->flowinfo = nla_get_be32(data[IFLA_IPTUN_FLOWINFO]);
if (data[IFLA_IPTUN_FLAGS])
parms->flags = nla_get_u32(data[IFLA_IPTUN_FLAGS]);
--
1.7.12
^ permalink raw reply related
* [PATCH] tcp: fix retransmission in repair mode
From: Andrey Vagin @ 2012-11-15 14:03 UTC (permalink / raw)
To: netdev
Cc: criu, linux-kernel, Andrew Vagin, Pavel Emelyanov,
David S. Miller, Alexey Kuznetsov, James Morris,
Hideaki YOSHIFUJI, Patrick McHardy
From: Andrew Vagin <avagin@openvz.org>
Currently if a socket was repaired with a few packet in a write queue,
a kernel bug may be triggered:
kernel BUG at net/ipv4/tcp_output.c:2330!
RIP: 0010:[<ffffffff8155784f>] tcp_retransmit_skb+0x5ff/0x610
According to the initial realization v3.4-rc2-963-gc0e88ff,
all skb-s should look like already posted. This patch fixes code
according with this sentence.
Here are three points, which were not done in the initial patch:
1. A tcp send head should not be changed
2. Initialize TSO state of a skb
3. Reset the retransmission time
This patch moves logic from tcp_sendmsg to tcp_write_xmit. A packet
passes the ussual way, but isn't sent to network. This patch solves
all described problems and handles tcp_sendpages.
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: James Morris <jmorris@namei.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: Andrey Vagin <avagin@openvz.org>
---
net/ipv4/tcp.c | 4 ++--
net/ipv4/tcp_output.c | 4 ++++
2 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 197c000..083092e 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1212,7 +1212,7 @@ new_segment:
wait_for_sndbuf:
set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
wait_for_memory:
- if (copied && likely(!tp->repair))
+ if (copied)
tcp_push(sk, flags & ~MSG_MORE, mss_now, TCP_NAGLE_PUSH);
if ((err = sk_stream_wait_memory(sk, &timeo)) != 0)
@@ -1223,7 +1223,7 @@ wait_for_memory:
}
out:
- if (copied && likely(!tp->repair))
+ if (copied)
tcp_push(sk, flags, mss_now, tp->nonagle);
release_sock(sk);
return copied + copied_syn;
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index cfe6ffe..2798706 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1986,6 +1986,9 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
tso_segs = tcp_init_tso_segs(sk, skb, mss_now);
BUG_ON(!tso_segs);
+ if (unlikely(tp->repair) && tp->repair_queue == TCP_SEND_QUEUE)
+ goto repair; /* Skip network transmission */
+
cwnd_quota = tcp_cwnd_test(tp, skb);
if (!cwnd_quota)
break;
@@ -2026,6 +2029,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
if (unlikely(tcp_transmit_skb(sk, skb, 1, gfp)))
break;
+repair:
/* Advance the send_head. This one is sent out.
* This call will increment packets_out.
*/
--
1.7.11.7
^ permalink raw reply related
* [PATCHv2 net-next] add DOVE extensions for VXLAN
From: David L Stevens @ 2012-11-15 13:46 UTC (permalink / raw)
To: David Miller, Stephen Hemminger; +Cc: netdev
This patch provides extensions to VXLAN for supporting Distributed
Overlay Virtual Ethernet (DOVE) networks. The patch includes:
+ a dove flag per VXLAN device to enable DOVE extensions
+ ARP reduction, whereby a bridge-connected VXLAN tunnel endpoint
answers ARP requests from the local bridge on behalf of
remote DOVE clients
+ route short-circuiting (aka L3 switching). Known destination IP
addresses use the corresponding destination MAC address for
switching rather than going to a (possibly remote) router first.
+ netlink notification messages for forwarding table and L3 switching
misses
Changes since v1:
- split "dove" flag into separate feature flags:
- "proxy" for ARP reduction
- "rsc" for route short circuiting
- "l2miss" for L2 switch miss notifications
- "l3miss" for L3 switch miss notifications
Signed-off-by: David L Stevens <dlstevens@us.ibm.com>
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 8aca888..6503091 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -29,6 +29,8 @@
#include <linux/etherdevice.h>
#include <linux/if_ether.h>
#include <linux/hash.h>
+#include <net/arp.h>
+#include <net/ndisc.h>
#include <net/ip.h>
#include <net/icmp.h>
#include <net/udp.h>
@@ -111,6 +113,10 @@ struct vxlan_dev {
__u8 tos; /* TOS override */
__u8 ttl;
bool learn;
+ bool proxy; /* ARP reduction */
+ bool rsc; /* route short-circuit */
+ bool l2miss; /* L2 miss notifications */
+ bool l3miss; /* L3 miss notifications */
unsigned long age_interval;
struct timer_list age_timer;
@@ -155,6 +161,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
struct nda_cacheinfo ci;
struct nlmsghdr *nlh;
struct ndmsg *ndm;
+ bool send_ip, send_eth;
nlh = nlmsg_put(skb, portid, seq, type, sizeof(*ndm), flags);
if (nlh == NULL)
@@ -162,16 +169,28 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
ndm = nlmsg_data(nlh);
memset(ndm, 0, sizeof(*ndm));
- ndm->ndm_family = AF_BRIDGE;
+
+ send_eth = send_ip = true;
+
+ if (type == RTM_GETNEIGH) {
+ int i;
+
+ ndm->ndm_family = AF_INET;
+ send_ip = fdb->remote_ip != 0;
+ send_eth = 0;
+ for (i=0; i<ETH_ALEN; ++i)
+ send_eth |= !!fdb->eth_addr[i];
+ } else
+ ndm->ndm_family = AF_BRIDGE;
ndm->ndm_state = fdb->state;
ndm->ndm_ifindex = vxlan->dev->ifindex;
ndm->ndm_flags = NTF_SELF;
ndm->ndm_type = NDA_DST;
- if (nla_put(skb, NDA_LLADDR, ETH_ALEN, &fdb->eth_addr))
+ if (send_eth && nla_put(skb, NDA_LLADDR, ETH_ALEN, &fdb->eth_addr))
goto nla_put_failure;
- if (nla_put_be32(skb, NDA_DST, fdb->remote_ip))
+ if (send_ip && nla_put_be32(skb, NDA_DST, fdb->remote_ip))
goto nla_put_failure;
ci.ndm_used = jiffies_to_clock_t(now - fdb->used);
@@ -223,6 +242,29 @@ errout:
rtnl_set_sk_err(net, RTNLGRP_NEIGH, err);
}
+static void vxlan_ip_miss(struct net_device *dev, __be32 ipa)
+{
+ struct vxlan_dev *vxlan = netdev_priv(dev);
+ struct vxlan_fdb f;
+
+ memset(&f, 0, sizeof f);
+ f.state = NUD_STALE;
+ f.remote_ip = ipa; /* goes to NDA_DST */
+
+ vxlan_fdb_notify(vxlan, &f, RTM_GETNEIGH);
+}
+
+static void vxlan_fdb_miss(struct vxlan_dev *vxlan, const u8 eth_addr[ETH_ALEN])
+{
+ struct vxlan_fdb f;
+
+ memset(&f, 0, sizeof f);
+ f.state = NUD_STALE;
+ memcpy(f.eth_addr, eth_addr, ETH_ALEN);
+
+ vxlan_fdb_notify(vxlan, &f, RTM_GETNEIGH);
+}
+
/* Hash Ethernet address */
static u32 eth_hash(const unsigned char *addr)
{
@@ -552,6 +594,8 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
goto drop;
}
+ skb_reset_mac_header(skb);
+
/* Re-examine inner Ethernet packet */
oip = ip_hdr(skb);
skb->protocol = eth_type_trans(skb, vxlan->dev);
@@ -600,6 +644,117 @@ drop:
return 0;
}
+static int arp_reduce(struct net_device *dev, struct sk_buff *skb)
+{
+ struct vxlan_dev *vxlan = netdev_priv(dev);
+ struct arphdr *parp;
+ u8 *arpptr, *sha;
+ __be32 sip, tip;
+ struct neighbour *n;
+
+ if (dev->flags & IFF_NOARP)
+ goto out;
+
+ if (!pskb_may_pull(skb, arp_hdr_len(dev))) {
+ dev->stats.tx_dropped++;
+ goto out;
+ }
+ parp = arp_hdr(skb);
+
+ if ((parp->ar_hrd != htons(ARPHRD_ETHER) &&
+ parp->ar_hrd != htons(ARPHRD_IEEE802)) ||
+ parp->ar_pro != htons(ETH_P_IP) ||
+ parp->ar_op != htons(ARPOP_REQUEST) ||
+ parp->ar_hln != dev->addr_len ||
+ parp->ar_pln != 4)
+ goto out;
+ arpptr = (u8 *)parp + sizeof(struct arphdr);
+ sha = arpptr;
+ arpptr += dev->addr_len; /* sha */
+ memcpy(&sip, arpptr, sizeof(sip));
+ arpptr += sizeof(sip);
+ arpptr += dev->addr_len; /* tha */
+ memcpy(&tip, arpptr, sizeof(tip));
+
+ if (ipv4_is_loopback(tip) ||
+ ipv4_is_multicast(tip))
+ goto out;
+
+ n = neigh_lookup(&arp_tbl, &tip, dev);
+
+ if (n) {
+ struct vxlan_dev *vxlan = netdev_priv(dev);
+ struct vxlan_fdb *f;
+ struct sk_buff *reply;
+
+ if (!(n->nud_state & NUD_CONNECTED)) {
+ neigh_release(n);
+ goto out;
+ }
+
+ f = vxlan_find_mac(vxlan, n->ha);
+ if (f && f->remote_ip == 0) {
+ /* bridge-local neighbor */
+ neigh_release(n);
+ goto out;
+ }
+
+ reply = arp_create(ARPOP_REPLY, ETH_P_ARP, sip, dev, tip, sha,
+ n->ha, sha);
+
+ neigh_release(n);
+
+ skb_reset_mac_header(reply);
+ __skb_pull(reply, skb_network_offset(reply));
+ reply->ip_summed = CHECKSUM_UNNECESSARY;
+ reply->pkt_type = PACKET_HOST;
+
+ if (netif_rx_ni(reply) == NET_RX_DROP)
+ dev->stats.rx_dropped++;
+ } else if (vxlan->l3miss)
+ vxlan_ip_miss(dev, tip);
+out:
+ consume_skb(skb);
+ return NETDEV_TX_OK;
+}
+
+static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
+{
+ struct vxlan_dev *vxlan = netdev_priv(dev);
+ struct neighbour *n;
+ struct iphdr *pip;
+
+ if (is_multicast_ether_addr(eth_hdr(skb)->h_dest))
+ return false;
+
+ n = NULL;
+ switch (ntohs(eth_hdr(skb)->h_proto)) {
+ case ETH_P_IP:
+ if (!pskb_may_pull(skb, sizeof(struct iphdr)))
+ return false;
+ pip = ip_hdr(skb);
+ n = neigh_lookup(&arp_tbl, &pip->daddr, dev);
+ break;
+ default:
+ return false;
+ }
+
+ if (n) {
+ bool diff;
+
+ diff = compare_ether_addr(eth_hdr(skb)->h_dest, n->ha) != 0;
+ if (diff) {
+ memcpy(eth_hdr(skb)->h_source, eth_hdr(skb)->h_dest,
+ dev->addr_len);
+ memcpy(eth_hdr(skb)->h_dest, n->ha, dev->addr_len);
+ }
+ neigh_release(n);
+ return diff;
+ } else if (vxlan->l3miss)
+ vxlan_ip_miss(dev, pip->daddr);
+ return false;
+}
+
/* Extract dsfield from inner protocol */
static inline u8 vxlan_get_dsfield(const struct iphdr *iph,
const struct sk_buff *skb)
@@ -622,22 +777,6 @@ static inline u8 vxlan_ecn_encap(u8 tos,
return INET_ECN_encapsulate(tos, inner);
}
-static __be32 vxlan_find_dst(struct vxlan_dev *vxlan, struct sk_buff *skb)
-{
- const struct ethhdr *eth = (struct ethhdr *) skb->data;
- const struct vxlan_fdb *f;
-
- if (is_multicast_ether_addr(eth->h_dest))
- return vxlan->gaddr;
-
- f = vxlan_find_mac(vxlan, eth->h_dest);
- if (f)
- return f->remote_ip;
- else
- return vxlan->gaddr;
-
-}
-
static void vxlan_sock_free(struct sk_buff *skb)
{
sock_put(skb->sk);
@@ -684,6 +823,7 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
struct vxlan_dev *vxlan = netdev_priv(dev);
struct rtable *rt;
const struct iphdr *old_iph;
+ struct ethhdr *eth;
struct iphdr *iph;
struct vxlanhdr *vxh;
struct udphdr *uh;
@@ -694,10 +834,50 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
__be16 df = 0;
__u8 tos, ttl;
int err;
+ bool rsc = false;
+ const struct vxlan_fdb *f;
+
+ skb_reset_mac_header(skb);
+ eth = eth_hdr(skb);
+
+ if (vxlan->proxy && ntohs(eth->h_proto) == ETH_P_ARP)
+ return arp_reduce(dev, skb);
+ else if (vxlan->rsc && ntohs(eth->h_proto) == ETH_P_IP)
+ rsc = route_shortcircuit(dev, skb);
- dst = vxlan_find_dst(vxlan, skb);
- if (!dst)
+ f = vxlan_find_mac(vxlan, eth->h_dest);
+ if (f == NULL) {
+ rsc = false;
+ dst = vxlan->gaddr;
+ if (!dst && vxlan->l2miss &&
+ !is_multicast_ether_addr(eth->h_dest))
+ vxlan_fdb_miss(vxlan, eth->h_dest);
+ } else
+ dst = f->remote_ip;
+
+ if (!dst) {
+ if (rsc) {
+ __skb_pull(skb, skb_network_offset(skb));
+ skb->ip_summed = CHECKSUM_NONE;
+ skb->pkt_type = PACKET_HOST;
+
+ /* short-circuited back to local bridge */
+ if (netif_rx(skb) == NET_RX_SUCCESS) {
+ struct vxlan_stats *stats =
+ this_cpu_ptr(vxlan->stats);
+
+ u64_stats_update_begin(&stats->syncp);
+ stats->tx_packets++;
+ stats->tx_bytes += pkt_len;
+ u64_stats_update_end(&stats->syncp);
+ } else {
+ dev->stats.tx_errors++;
+ dev->stats.tx_aborted_errors++;
+ }
+ return NETDEV_TX_OK;
+ }
goto drop;
+ }
/* Need space for new headers (invalidates iph ptr) */
if (skb_cow_head(skb, VXLAN_HEADROOM))
@@ -1020,6 +1200,10 @@ static const struct nla_policy vxlan_policy[IFLA_VXLAN_MAX + 1] = {
[IFLA_VXLAN_AGEING] = { .type = NLA_U32 },
[IFLA_VXLAN_LIMIT] = { .type = NLA_U32 },
[IFLA_VXLAN_PORT_RANGE] = { .len = sizeof(struct ifla_vxlan_port_range) },
+ [IFLA_VXLAN_PROXY] = { .type = NLA_U8 },
+ [IFLA_VXLAN_RSC] = { .type = NLA_U8 },
+ [IFLA_VXLAN_L2MISS] = { .type = NLA_U8 },
+ [IFLA_VXLAN_L3MISS] = { .type = NLA_U8 },
};
static int vxlan_validate(struct nlattr *tb[], struct nlattr *data[])
@@ -1118,6 +1302,18 @@ static int vxlan_newlink(struct net *net, struct net_device *dev,
else
vxlan->age_interval = FDB_AGE_DEFAULT;
+ if (data[IFLA_VXLAN_PROXY] && nla_get_u8(data[IFLA_VXLAN_PROXY]))
+ vxlan->proxy = true;
+
+ if (data[IFLA_VXLAN_RSC] && nla_get_u8(data[IFLA_VXLAN_RSC]))
+ vxlan->rsc = true;
+
+ if (data[IFLA_VXLAN_L2MISS] && nla_get_u8(data[IFLA_VXLAN_L2MISS]))
+ vxlan->l2miss = true;
+
+ if (data[IFLA_VXLAN_L3MISS] && nla_get_u8(data[IFLA_VXLAN_L3MISS]))
+ vxlan->l3miss = true;
+
if (data[IFLA_VXLAN_LIMIT])
vxlan->addrmax = nla_get_u32(data[IFLA_VXLAN_LIMIT]);
@@ -1154,6 +1350,7 @@ static size_t vxlan_get_size(const struct net_device *dev)
nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_TTL */
nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_TOS */
nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_LEARNING */
+ nla_total_size(sizeof(__u32)) + /* IFLA_VXLAN_DOVE */
nla_total_size(sizeof(__u32)) + /* IFLA_VXLAN_AGEING */
nla_total_size(sizeof(__u32)) + /* IFLA_VXLAN_LIMIT */
nla_total_size(sizeof(struct ifla_vxlan_port_range)) +
@@ -1183,6 +1380,10 @@ static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)
if (nla_put_u8(skb, IFLA_VXLAN_TTL, vxlan->ttl) ||
nla_put_u8(skb, IFLA_VXLAN_TOS, vxlan->tos) ||
nla_put_u8(skb, IFLA_VXLAN_LEARNING, vxlan->learn) ||
+ nla_put_u8(skb, IFLA_VXLAN_PROXY, vxlan->proxy) ||
+ nla_put_u8(skb, IFLA_VXLAN_RSC, vxlan->rsc) ||
+ nla_put_u8(skb, IFLA_VXLAN_L2MISS, vxlan->l2miss) ||
+ nla_put_u8(skb, IFLA_VXLAN_L3MISS, vxlan->l3miss) ||
nla_put_u32(skb, IFLA_VXLAN_AGEING, vxlan->age_interval) ||
nla_put_u32(skb, IFLA_VXLAN_LIMIT, vxlan->addrmax))
goto nla_put_failure;
diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index 5c80cb1..89695ff 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -285,6 +285,10 @@ enum {
IFLA_VXLAN_AGEING,
IFLA_VXLAN_LIMIT,
IFLA_VXLAN_PORT_RANGE,
+ IFLA_VXLAN_PROXY,
+ IFLA_VXLAN_RSC,
+ IFLA_VXLAN_L2MISS,
+ IFLA_VXLAN_L3MISS,
__IFLA_VXLAN_MAX
};
#define IFLA_VXLAN_MAX (__IFLA_VXLAN_MAX - 1)
^ permalink raw reply related
* [PATCHv2 iproute2] add DOVE extensions for iproute2
From: David L Stevens @ 2012-11-15 13:49 UTC (permalink / raw)
To: David Miller, Stephen Hemminger; +Cc: netdev
This patch adds a new flag to iproute2 for vxlan devices to enable
DOVE features. It also adds support for L2 and L3 switch lookup miss
netlink messages to "ip monitor".
Changes since v1:
- split "dove" flag into separate feature flags:
- "proxy" for ARP reduction
- "rsc" for route short circuiting
- "l2miss" for L2 switch miss notifications
- "l3miss" for L3 switch miss notifications
Signed-off-by: David L Stevens <dlstevens@us.ibm.com>
diff --git a/include/linux/if_link.h b/include/linux/if_link.h
index 012d95a..a163702 100644
--- a/include/linux/if_link.h
+++ b/include/linux/if_link.h
@@ -283,6 +283,10 @@ enum {
IFLA_VXLAN_AGEING,
IFLA_VXLAN_LIMIT,
IFLA_VXLAN_PORT_RANGE,
+ IFLA_VXLAN_PROXY,
+ IFLA_VXLAN_RSC,
+ IFLA_VXLAN_L2MISS,
+ IFLA_VXLAN_L3MISS,
__IFLA_VXLAN_MAX
};
#define IFLA_VXLAN_MAX (__IFLA_VXLAN_MAX - 1)
diff --git a/ip/iplink_vxlan.c b/ip/iplink_vxlan.c
index ba5c4ab..f2e6bef 100644
--- a/ip/iplink_vxlan.c
+++ b/ip/iplink_vxlan.c
@@ -26,6 +26,8 @@ static void explain(void)
fprintf(stderr, "Usage: ... vxlan id VNI [ group ADDR ] [ local ADDR ]\n");
fprintf(stderr, " [ ttl TTL ] [ tos TOS ] [ dev PHYS_DEV ]\n");
fprintf(stderr, " [ port MIN MAX ] [ [no]learning ]\n");
+ fprintf(stderr, " [ [no]proxy ] [ [no]rsc ]\n");
+ fprintf(stderr, " [ [no]l2miss ] [ [no]l3miss ]\n");
fprintf(stderr, "\n");
fprintf(stderr, "Where: VNI := 0-16777215\n");
fprintf(stderr, " ADDR := { IP_ADDRESS | any }\n");
@@ -44,6 +46,10 @@ static int vxlan_parse_opt(struct link_util *lu, int argc, char **argv,
__u8 tos = 0;
__u8 ttl = 0;
__u8 learning = 1;
+ __u8 proxy = 0;
+ __u8 rsc = 0;
+ __u8 l2miss = 0;
+ __u8 l3miss = 0;
__u8 noage = 0;
__u32 age = 0;
__u32 maxaddr = 0;
@@ -123,6 +129,22 @@ static int vxlan_parse_opt(struct link_util *lu, int argc, char **argv,
learning = 0;
} else if (!matches(*argv, "learning")) {
learning = 1;
+ } else if (!matches(*argv, "noproxy")) {
+ proxy = 0;
+ } else if (!matches(*argv, "proxy")) {
+ proxy = 1;
+ } else if (!matches(*argv, "norsc")) {
+ rsc = 0;
+ } else if (!matches(*argv, "rsc")) {
+ rsc = 1;
+ } else if (!matches(*argv, "nol2miss")) {
+ l2miss = 0;
+ } else if (!matches(*argv, "l2miss")) {
+ l2miss = 1;
+ } else if (!matches(*argv, "nol3miss")) {
+ l3miss = 0;
+ } else if (!matches(*argv, "l3miss")) {
+ l3miss = 1;
} else if (matches(*argv, "help") == 0) {
explain();
return -1;
@@ -148,6 +170,10 @@ static int vxlan_parse_opt(struct link_util *lu, int argc, char **argv,
addattr8(n, 1024, IFLA_VXLAN_TTL, ttl);
addattr8(n, 1024, IFLA_VXLAN_TOS, tos);
addattr8(n, 1024, IFLA_VXLAN_LEARNING, learning);
+ addattr8(n, 1024, IFLA_VXLAN_PROXY, proxy);
+ addattr8(n, 1024, IFLA_VXLAN_RSC, rsc);
+ addattr8(n, 1024, IFLA_VXLAN_L2MISS, l2miss);
+ addattr8(n, 1024, IFLA_VXLAN_L3MISS, l3miss);
if (noage)
addattr32(n, 1024, IFLA_VXLAN_AGEING, 0);
else if (age)
@@ -213,6 +239,18 @@ static void vxlan_print_opt(struct link_util *lu, FILE *f, struct rtattr *tb[])
if (tb[IFLA_VXLAN_LEARNING] &&
!rta_getattr_u8(tb[IFLA_VXLAN_LEARNING]))
fputs("nolearning ", f);
+
+ if (tb[IFLA_VXLAN_PROXY] && rta_getattr_u8(tb[IFLA_VXLAN_PROXY]))
+ fputs("proxy ", f);
+
+ if (tb[IFLA_VXLAN_RSC] && rta_getattr_u8(tb[IFLA_VXLAN_RSC]))
+ fputs("rsc ", f);
+
+ if (tb[IFLA_VXLAN_L2MISS] && rta_getattr_u8(tb[IFLA_VXLAN_L2MISS]))
+ fputs("l2miss ", f);
+
+ if (tb[IFLA_VXLAN_L3MISS] && rta_getattr_u8(tb[IFLA_VXLAN_L3MISS]))
+ fputs("l3miss ", f);
if (tb[IFLA_VXLAN_TOS] &&
(tos = rta_getattr_u8(tb[IFLA_VXLAN_TOS]))) {
diff --git a/ip/ipmonitor.c b/ip/ipmonitor.c
index 4b1d469..7a7cc88 100644
--- a/ip/ipmonitor.c
+++ b/ip/ipmonitor.c
@@ -67,7 +67,8 @@ int accept_msg(const struct sockaddr_nl *who,
print_addrlabel(who, n, arg);
return 0;
}
- if (n->nlmsg_type == RTM_NEWNEIGH || n->nlmsg_type == RTM_DELNEIGH) {
+ if (n->nlmsg_type == RTM_NEWNEIGH || n->nlmsg_type == RTM_DELNEIGH ||
+ n->nlmsg_type == RTM_GETNEIGH) {
if (prefix_banner)
fprintf(fp, "[NEIGH]");
print_neigh(who, n, arg);
diff --git a/ip/ipneigh.c b/ip/ipneigh.c
index 56e56b2..1b7600b 100644
--- a/ip/ipneigh.c
+++ b/ip/ipneigh.c
@@ -189,7 +189,8 @@ int print_neigh(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
struct rtattr * tb[NDA_MAX+1];
char abuf[256];
- if (n->nlmsg_type != RTM_NEWNEIGH && n->nlmsg_type != RTM_DELNEIGH) {
+ if (n->nlmsg_type != RTM_NEWNEIGH && n->nlmsg_type != RTM_DELNEIGH &&
+ n->nlmsg_type != RTM_GETNEIGH) {
fprintf(stderr, "Not RTM_NEWNEIGH: %08x %08x %08x\n",
n->nlmsg_len, n->nlmsg_type, n->nlmsg_flags);
@@ -251,6 +252,8 @@ int print_neigh(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
if (n->nlmsg_type == RTM_DELNEIGH)
fprintf(fp, "delete ");
+ else if (n->nlmsg_type == RTM_GETNEIGH)
+ fprintf(fp, "miss ");
if (tb[NDA_DST]) {
fprintf(fp, "%s ",
format_host(r->ndm_family,
^ permalink raw reply related
* [PATCHv2] sctp: fix /proc/net/sctp/ memory leak
From: Tommi Rantala @ 2012-11-15 13:49 UTC (permalink / raw)
To: netdev
Cc: Neil Horman, Vlad Yasevich, Sridhar Samudrala, David S. Miller,
linux-sctp, Dave Jones, Tommi Rantala, Eric W. Biederman
Commit 13d782f ("sctp: Make the proc files per network namespace.")
changed the /proc/net/sctp/ struct file_operations opener functions to
use single_open_net() and seq_open_net().
Avoid leaking memory by using single_release_net() and seq_release_net()
as the release functions.
Discovered with Trinity (the syscall fuzzer).
Signed-off-by: Tommi Rantala <tt.rantala@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
---
net/sctp/proc.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/net/sctp/proc.c b/net/sctp/proc.c
index c3bea26..9966e7b 100644
--- a/net/sctp/proc.c
+++ b/net/sctp/proc.c
@@ -102,7 +102,7 @@ static const struct file_operations sctp_snmp_seq_fops = {
.open = sctp_snmp_seq_open,
.read = seq_read,
.llseek = seq_lseek,
- .release = single_release,
+ .release = single_release_net,
};
/* Set up the proc fs entry for 'snmp' object. */
@@ -251,7 +251,7 @@ static const struct file_operations sctp_eps_seq_fops = {
.open = sctp_eps_seq_open,
.read = seq_read,
.llseek = seq_lseek,
- .release = seq_release,
+ .release = seq_release_net,
};
/* Set up the proc fs entry for 'eps' object. */
@@ -372,7 +372,7 @@ static const struct file_operations sctp_assocs_seq_fops = {
.open = sctp_assocs_seq_open,
.read = seq_read,
.llseek = seq_lseek,
- .release = seq_release,
+ .release = seq_release_net,
};
/* Set up the proc fs entry for 'assocs' object. */
@@ -517,7 +517,7 @@ static const struct file_operations sctp_remaddr_seq_fops = {
.open = sctp_remaddr_seq_open,
.read = seq_read,
.llseek = seq_lseek,
- .release = seq_release,
+ .release = seq_release_net,
};
int __net_init sctp_remaddr_proc_init(struct net *net)
--
1.7.9.5
^ permalink raw reply related
* Re: [RFC] tcp: use order-3 pages in tcp_sendmsg()
From: Eric Dumazet @ 2012-11-15 13:47 UTC (permalink / raw)
To: Yan, Zheng; +Cc: netdev
In-Reply-To: <CAAM7YAm+XJKTxJStLMoo4RRV85oogN5wCHfXorkEiUHKqNKHDQ@mail.gmail.com>
On Thu, 2012-11-15 at 15:52 +0800, Yan, Zheng wrote:
> LLC misses happen on the receiver size. It means most pages allocated by the
> senders are cache hot. But when using order-3 pages, 2048 * 32k = 64M, 64M
> is much larger than LLC size.
By the way, this 2048*32k is wrong, as the receiver only uses fragments
in pages, and its not related to the "tcp: use order-3 pages in
tcp_sendmsg()" commit
Many drivers use order-0 pages to hold ethernet frames, regardless of
what was used by the sender ;)
^ permalink raw reply
* Re: [PATCH net-next 1/3] ipip: fix sparse warnings in ipip_netlink_parms()
From: Eric Dumazet @ 2012-11-15 13:20 UTC (permalink / raw)
To: nicolas.dichtel; +Cc: netdev, davem
In-Reply-To: <50A4B8B0.1090409@6wind.com>
On Thu, 2012-11-15 at 10:41 +0100, Nicolas Dichtel wrote:
> Le 15/11/2012 09:53, Nicolas Dichtel a écrit :
> > Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
> I forget "Reported-by: Fengguang Wu <fengguang.wu@intel.com>"
>
> Should I resend the serie?
Your patches generally lack changelogs and proper attribution.
Having a good changelog can avoid to actually look at the patch content,
its incredibly useful in the long term.
^ permalink raw reply
* Re: [PATCH net-next 1/3] ipip: fix sparse warnings in ipip_netlink_parms()
From: Eric Dumazet @ 2012-11-15 13:15 UTC (permalink / raw)
To: Nicolas Dichtel; +Cc: netdev, davem
In-Reply-To: <1352969616-30476-1-git-send-email-nicolas.dichtel@6wind.com>
On Thu, 2012-11-15 at 09:53 +0100, Nicolas Dichtel wrote:
> Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
> ---
> net/ipv4/ipip.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c
> index 64686e1..c26c171 100644
> --- a/net/ipv4/ipip.c
> +++ b/net/ipv4/ipip.c
> @@ -864,10 +864,10 @@ static void ipip_netlink_parms(struct nlattr *data[],
> parms->link = nla_get_u32(data[IFLA_IPTUN_LINK]);
>
> if (data[IFLA_IPTUN_LOCAL])
> - parms->iph.saddr = nla_get_u32(data[IFLA_IPTUN_LOCAL]);
> + parms->iph.saddr = nla_get_be32(data[IFLA_IPTUN_LOCAL]);
>
> if (data[IFLA_IPTUN_REMOTE])
> - parms->iph.daddr = nla_get_u32(data[IFLA_IPTUN_REMOTE]);
> + parms->iph.daddr = nla_get_be32(data[IFLA_IPTUN_REMOTE]);
>
> if (data[IFLA_IPTUN_TTL]) {
> parms->iph.ttl = nla_get_u8(data[IFLA_IPTUN_TTL]);
No credit for who reported this bug ?
^ permalink raw reply
* Re: [PATCH] sctp: fix /proc/net/sctp/ memory leak
From: Eric Dumazet @ 2012-11-15 13:13 UTC (permalink / raw)
To: Tommi Rantala
Cc: netdev, Neil Horman, Vlad Yasevich, Sridhar Samudrala,
David S. Miller, linux-sctp, Dave Jones
In-Reply-To: <1352967802-2335-1-git-send-email-tt.rantala@gmail.com>
On Thu, 2012-11-15 at 10:23 +0200, Tommi Rantala wrote:
> We are using single_open_net() and seq_open_net() in the opener
> functions, so avoid leaking memory by using single_release_net() and
> seq_release_net() as the struct file_operations release functions.
>
> Discovered with Trinity (the syscall fuzzer).
>
> Signed-off-by: Tommi Rantala <tt.rantala@gmail.com>
> Acked-by: Neil Horman <nhorman@tuxdriver.com>
> ---
> net/sctp/proc.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/net/sctp/proc.c b/net/sctp/proc.c
> index c3bea26..9966e7b 100644
> --- a/net/sctp/proc.c
> +++ b/net/sctp/proc.c
> @@ -102,7 +102,7 @@ static const struct file_operations sctp_snmp_seq_fops = {
> .open = sctp_snmp_seq_open,
> .read = seq_read,
> .llseek = seq_lseek,
> - .release = single_release,
> + .release = single_release_net,
> };
>
It seems you forgot to do do some work to pinpoint which commit added
this bug. Please add this in the changelog to ease maintainer and stable
teams work, and CC the author of said commit so that he has a chance to
Ack your patch.
Thanks !
^ permalink raw reply
* Re: [RFC] tcp: use order-3 pages in tcp_sendmsg()
From: Eric Dumazet @ 2012-11-15 13:06 UTC (permalink / raw)
To: Yan, Zheng; +Cc: netdev
In-Reply-To: <CAAM7YAm+XJKTxJStLMoo4RRV85oogN5wCHfXorkEiUHKqNKHDQ@mail.gmail.com>
On Thu, 2012-11-15 at 15:52 +0800, Yan, Zheng wrote:
> On Mon, Sep 17, 2012 at 3:49 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> > We currently use per socket page reserve for tcp_sendmsg() operations.
> >
> > Its done to raise the probability of coalescing small write() in to
> > single segments in the skbs.
> >
> > But it wastes a lot of memory for applications handling a lot of mostly
> > idle sockets, since each socket holds one page in sk->sk_sndmsg_page
> >
> > I did a small experiment to use order-3 pages and it gave me a 10% boost
> > of performance, because each TSO skb can use only two frags of 32KB,
> > instead of 16 frags of 4KB, so we spend less time in ndo_start_xmit() to
> > setup the tx descriptor and TX completion path to unmap the frags and
> > free them.
> >
> > We also spend less time in tcp_sendmsg(), because we call page allocator
> > 8x less often.
> >
> > Now back to the per socket page, what about trying to factorize it ?
> >
> > Since we can sleep (or/and do a cpu migration) in tcp_sendmsg(), we cant
> > really use a percpu page reserve as we do in __netdev_alloc_frag()
> >
> > We could instead use a per thread reserve, at the cost of adding a test
> > in task exit handler.
> >
> > Recap :
> >
> > 1) Use a per thread page reserve instead of a per socket one
> > 2) Use order-3 pages (or order-0 pages if page size is >= 32768)
> >
> >
>
> Hi,
>
> This commit makes one of our test case on core 2 machine drop in performance
> by about 60%. The test case runs 2048 instances of netperf 64k stream test at
> the same time. Analysis showed using order-3 pages causes more LLC misses,
> most new LLC misses happen when the senders copy data to the socket buffer.
> If revert to use single page, the sender side only trigger a few LLC
> misses, most
> LLC misses happen on the receiver size. It means most pages allocated by the
> senders are cache hot. But when using order-3 pages, 2048 * 32k = 64M, 64M
> is much larger than LLC size. Should this regression be worried? or
> our test case
> is too unpractical?
Hi Yan
You forgot to give some basic information with this mail, like the
hardware configuration, NIC driver, ...
Increasing performance can sometime change the balance you had on a
prior workload.
Number of in flight bytes do not depend on the order of the pages, but
sizes of TCP buffers (receiver, sender)
TCP Small queue was an attempt to reduce the number of in-flight bytes,
you should try to change either SO_SNDBUF or SO_RCVBUF settings (instead
of letting the system autotune them) if you really need 2048 concurrent
flows.
Otherwise, each flow can consume up to 6 MB of memory, so obviously your
cpu caches wont hold 2048*6MB of memory...
If the sender is faster (because of this commit), but receiver is slow
to drain the receive queues, then you can have a situation where the
consumed memory on receiver is higher and the receiver might be actually
slower.
Thanks
^ permalink raw reply
* [v6 PATCH 8/8] csiostor: Chelsio FCoE offload driver submission (sources part 5).
From: Naresh Kumar Inna @ 2012-11-15 17:11 UTC (permalink / raw)
To: JBottomley, linux-scsi, dm, leedom; +Cc: netdev, naresh, chethan
In-Reply-To: <1352999484-17812-1-git-send-email-naresh@chelsio.com>
This patch contains the hardware interfacing functionality (chip/firmware
initialization/setup) and error handling. It also has slow path event
handling functionality, the Makefile and Kconfig changes.
Signed-off-by: Naresh Kumar Inna <naresh@chelsio.com>
---
V3: Added const qualifier to adapter descriptor table.
V5: Re-wrote error paths to reduce lock/unlock ambiguity.
drivers/scsi/Kconfig | 1 +
drivers/scsi/Makefile | 1 +
drivers/scsi/csiostor/Kconfig | 19 +
drivers/scsi/csiostor/Makefile | 11 +
drivers/scsi/csiostor/csio_hw.c | 4395 +++++++++++++++++++++++++++++++++++++++
5 files changed, 4427 insertions(+), 0 deletions(-)
create mode 100644 drivers/scsi/csiostor/Kconfig
create mode 100644 drivers/scsi/csiostor/Makefile
create mode 100644 drivers/scsi/csiostor/csio_hw.c
diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
index 74bf1aa..af7a3e7 100644
--- a/drivers/scsi/Kconfig
+++ b/drivers/scsi/Kconfig
@@ -1812,6 +1812,7 @@ config SCSI_VIRTIO
This is the virtual HBA driver for virtio. If the kernel will
be used in a virtual machine, say Y or M.
+source "drivers/scsi/csiostor/Kconfig"
endif # SCSI_LOWLEVEL
diff --git a/drivers/scsi/Makefile b/drivers/scsi/Makefile
index 888f73a..8739aa7 100644
--- a/drivers/scsi/Makefile
+++ b/drivers/scsi/Makefile
@@ -90,6 +90,7 @@ obj-$(CONFIG_SCSI_QLA_FC) += qla2xxx/
obj-$(CONFIG_SCSI_QLA_ISCSI) += libiscsi.o qla4xxx/
obj-$(CONFIG_SCSI_LPFC) += lpfc/
obj-$(CONFIG_SCSI_BFA_FC) += bfa/
+obj-$(CONFIG_SCSI_CHELSIO_FCOE) += csiostor/
obj-$(CONFIG_SCSI_PAS16) += pas16.o
obj-$(CONFIG_SCSI_T128) += t128.o
obj-$(CONFIG_SCSI_DMX3191D) += dmx3191d.o
diff --git a/drivers/scsi/csiostor/Kconfig b/drivers/scsi/csiostor/Kconfig
new file mode 100644
index 0000000..4d03b03
--- /dev/null
+++ b/drivers/scsi/csiostor/Kconfig
@@ -0,0 +1,19 @@
+config SCSI_CHELSIO_FCOE
+ tristate "Chelsio Communications FCoE support"
+ depends on PCI && SCSI
+ select SCSI_FC_ATTRS
+ select FW_LOADER
+ help
+ This driver supports FCoE Offload functionality over
+ Chelsio T4-based 10Gb Converged Network Adapters.
+
+ For general information about Chelsio and our products, visit
+ our website at <http://www.chelsio.com>.
+
+ For customer support, please visit our customer support page at
+ <http://www.chelsio.com/support.html>.
+
+ Please send feedback to <linux-bugs@chelsio.com>.
+
+ To compile this driver as a module choose M here; the module
+ will be called csiostor.
diff --git a/drivers/scsi/csiostor/Makefile b/drivers/scsi/csiostor/Makefile
new file mode 100644
index 0000000..b581966
--- /dev/null
+++ b/drivers/scsi/csiostor/Makefile
@@ -0,0 +1,11 @@
+#
+## Chelsio FCoE driver
+#
+##
+
+ccflags-y += -I$(srctree)/drivers/net/ethernet/chelsio/cxgb4
+
+obj-$(CONFIG_SCSI_CHELSIO_FCOE) += csiostor.o
+
+csiostor-objs := csio_attr.o csio_init.o csio_lnode.o csio_scsi.o \
+ csio_hw.o csio_isr.o csio_mb.o csio_rnode.o csio_wr.o
diff --git a/drivers/scsi/csiostor/csio_hw.c b/drivers/scsi/csiostor/csio_hw.c
new file mode 100644
index 0000000..963c6c1
--- /dev/null
+++ b/drivers/scsi/csiostor/csio_hw.c
@@ -0,0 +1,4395 @@
+/*
+ * This file is part of the Chelsio FCoE driver for Linux.
+ *
+ * Copyright (c) 2008-2012 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/pci.h>
+#include <linux/pci_regs.h>
+#include <linux/firmware.h>
+#include <linux/stddef.h>
+#include <linux/delay.h>
+#include <linux/string.h>
+#include <linux/compiler.h>
+#include <linux/jiffies.h>
+#include <linux/kernel.h>
+#include <linux/log2.h>
+
+#include "csio_hw.h"
+#include "csio_lnode.h"
+#include "csio_rnode.h"
+
+int csio_force_master;
+int csio_dbg_level = 0xFEFF;
+unsigned int csio_port_mask = 0xf;
+
+/* Default FW event queue entries. */
+static uint32_t csio_evtq_sz = CSIO_EVTQ_SIZE;
+
+/* Default MSI param level */
+int csio_msi = 2;
+
+/* FCoE function instances */
+static int dev_num;
+
+/* FCoE Adapter types & its description */
+static const struct csio_adap_desc csio_fcoe_adapters[] = {
+ {"T440-Dbg 10G", "Chelsio T440-Dbg 10G [FCoE]"},
+ {"T420-CR 10G", "Chelsio T420-CR 10G [FCoE]"},
+ {"T422-CR 10G/1G", "Chelsio T422-CR 10G/1G [FCoE]"},
+ {"T440-CR 10G", "Chelsio T440-CR 10G [FCoE]"},
+ {"T420-BCH 10G", "Chelsio T420-BCH 10G [FCoE]"},
+ {"T440-BCH 10G", "Chelsio T440-BCH 10G [FCoE]"},
+ {"T440-CH 10G", "Chelsio T440-CH 10G [FCoE]"},
+ {"T420-SO 10G", "Chelsio T420-SO 10G [FCoE]"},
+ {"T420-CX4 10G", "Chelsio T420-CX4 10G [FCoE]"},
+ {"T420-BT 10G", "Chelsio T420-BT 10G [FCoE]"},
+ {"T404-BT 1G", "Chelsio T404-BT 1G [FCoE]"},
+ {"B420-SR 10G", "Chelsio B420-SR 10G [FCoE]"},
+ {"B404-BT 1G", "Chelsio B404-BT 1G [FCoE]"},
+ {"T480-CR 10G", "Chelsio T480-CR 10G [FCoE]"},
+ {"T440-LP-CR 10G", "Chelsio T440-LP-CR 10G [FCoE]"},
+ {"T4 FPGA", "Chelsio T4 FPGA [FCoE]"}
+};
+
+static void csio_mgmtm_cleanup(struct csio_mgmtm *);
+static void csio_hw_mbm_cleanup(struct csio_hw *);
+
+/* State machine forward declarations */
+static void csio_hws_uninit(struct csio_hw *, enum csio_hw_ev);
+static void csio_hws_configuring(struct csio_hw *, enum csio_hw_ev);
+static void csio_hws_initializing(struct csio_hw *, enum csio_hw_ev);
+static void csio_hws_ready(struct csio_hw *, enum csio_hw_ev);
+static void csio_hws_quiescing(struct csio_hw *, enum csio_hw_ev);
+static void csio_hws_quiesced(struct csio_hw *, enum csio_hw_ev);
+static void csio_hws_resetting(struct csio_hw *, enum csio_hw_ev);
+static void csio_hws_removing(struct csio_hw *, enum csio_hw_ev);
+static void csio_hws_pcierr(struct csio_hw *, enum csio_hw_ev);
+
+static void csio_hw_initialize(struct csio_hw *hw);
+static void csio_evtq_stop(struct csio_hw *hw);
+static void csio_evtq_start(struct csio_hw *hw);
+
+int csio_is_hw_ready(struct csio_hw *hw)
+{
+ return csio_match_state(hw, csio_hws_ready);
+}
+
+int csio_is_hw_removing(struct csio_hw *hw)
+{
+ return csio_match_state(hw, csio_hws_removing);
+}
+
+
+/*
+ * csio_hw_wait_op_done_val - wait until an operation is completed
+ * @hw: the HW module
+ * @reg: the register to check for completion
+ * @mask: a single-bit field within @reg that indicates completion
+ * @polarity: the value of the field when the operation is completed
+ * @attempts: number of check iterations
+ * @delay: delay in usecs between iterations
+ * @valp: where to store the value of the register at completion time
+ *
+ * Wait until an operation is completed by checking a bit in a register
+ * up to @attempts times. If @valp is not NULL the value of the register
+ * at the time it indicated completion is stored there. Returns 0 if the
+ * operation completes and -EAGAIN otherwise.
+ */
+static int
+csio_hw_wait_op_done_val(struct csio_hw *hw, int reg, uint32_t mask,
+ int polarity, int attempts, int delay, uint32_t *valp)
+{
+ uint32_t val;
+ while (1) {
+ val = csio_rd_reg32(hw, reg);
+
+ if (!!(val & mask) == polarity) {
+ if (valp)
+ *valp = val;
+ return 0;
+ }
+
+ if (--attempts == 0)
+ return -EAGAIN;
+ if (delay)
+ udelay(delay);
+ }
+}
+
+void
+csio_set_reg_field(struct csio_hw *hw, uint32_t reg, uint32_t mask,
+ uint32_t value)
+{
+ uint32_t val = csio_rd_reg32(hw, reg) & ~mask;
+
+ csio_wr_reg32(hw, val | value, reg);
+ /* Flush */
+ csio_rd_reg32(hw, reg);
+
+}
+
+/*
+ * csio_hw_mc_read - read from MC through backdoor accesses
+ * @hw: the hw module
+ * @addr: address of first byte requested
+ * @data: 64 bytes of data containing the requested address
+ * @ecc: where to store the corresponding 64-bit ECC word
+ *
+ * Read 64 bytes of data from MC starting at a 64-byte-aligned address
+ * that covers the requested address @addr. If @parity is not %NULL it
+ * is assigned the 64-bit ECC word for the read data.
+ */
+int
+csio_hw_mc_read(struct csio_hw *hw, uint32_t addr, uint32_t *data,
+ uint64_t *ecc)
+{
+ int i;
+
+ if (csio_rd_reg32(hw, MC_BIST_CMD) & START_BIST)
+ return -EBUSY;
+ csio_wr_reg32(hw, addr & ~0x3fU, MC_BIST_CMD_ADDR);
+ csio_wr_reg32(hw, 64, MC_BIST_CMD_LEN);
+ csio_wr_reg32(hw, 0xc, MC_BIST_DATA_PATTERN);
+ csio_wr_reg32(hw, BIST_OPCODE(1) | START_BIST | BIST_CMD_GAP(1),
+ MC_BIST_CMD);
+ i = csio_hw_wait_op_done_val(hw, MC_BIST_CMD, START_BIST,
+ 0, 10, 1, NULL);
+ if (i)
+ return i;
+
+#define MC_DATA(i) MC_BIST_STATUS_REG(MC_BIST_STATUS_RDATA, i)
+
+ for (i = 15; i >= 0; i--)
+ *data++ = htonl(csio_rd_reg32(hw, MC_DATA(i)));
+ if (ecc)
+ *ecc = csio_rd_reg64(hw, MC_DATA(16));
+#undef MC_DATA
+ return 0;
+}
+
+/*
+ * csio_hw_edc_read - read from EDC through backdoor accesses
+ * @hw: the hw module
+ * @idx: which EDC to access
+ * @addr: address of first byte requested
+ * @data: 64 bytes of data containing the requested address
+ * @ecc: where to store the corresponding 64-bit ECC word
+ *
+ * Read 64 bytes of data from EDC starting at a 64-byte-aligned address
+ * that covers the requested address @addr. If @parity is not %NULL it
+ * is assigned the 64-bit ECC word for the read data.
+ */
+int
+csio_hw_edc_read(struct csio_hw *hw, int idx, uint32_t addr, uint32_t *data,
+ uint64_t *ecc)
+{
+ int i;
+
+ idx *= EDC_STRIDE;
+ if (csio_rd_reg32(hw, EDC_BIST_CMD + idx) & START_BIST)
+ return -EBUSY;
+ csio_wr_reg32(hw, addr & ~0x3fU, EDC_BIST_CMD_ADDR + idx);
+ csio_wr_reg32(hw, 64, EDC_BIST_CMD_LEN + idx);
+ csio_wr_reg32(hw, 0xc, EDC_BIST_DATA_PATTERN + idx);
+ csio_wr_reg32(hw, BIST_OPCODE(1) | BIST_CMD_GAP(1) | START_BIST,
+ EDC_BIST_CMD + idx);
+ i = csio_hw_wait_op_done_val(hw, EDC_BIST_CMD + idx, START_BIST,
+ 0, 10, 1, NULL);
+ if (i)
+ return i;
+
+#define EDC_DATA(i) (EDC_BIST_STATUS_REG(EDC_BIST_STATUS_RDATA, i) + idx)
+
+ for (i = 15; i >= 0; i--)
+ *data++ = htonl(csio_rd_reg32(hw, EDC_DATA(i)));
+ if (ecc)
+ *ecc = csio_rd_reg64(hw, EDC_DATA(16));
+#undef EDC_DATA
+ return 0;
+}
+
+/*
+ * csio_mem_win_rw - read/write memory through PCIE memory window
+ * @hw: the adapter
+ * @addr: address of first byte requested
+ * @data: MEMWIN0_APERTURE bytes of data containing the requested address
+ * @dir: direction of transfer 1 => read, 0 => write
+ *
+ * Read/write MEMWIN0_APERTURE bytes of data from MC starting at a
+ * MEMWIN0_APERTURE-byte-aligned address that covers the requested
+ * address @addr.
+ */
+static int
+csio_mem_win_rw(struct csio_hw *hw, u32 addr, __be32 *data, int dir)
+{
+ int i;
+
+ /*
+ * Setup offset into PCIE memory window. Address must be a
+ * MEMWIN0_APERTURE-byte-aligned address. (Read back MA register to
+ * ensure that changes propagate before we attempt to use the new
+ * values.)
+ */
+ csio_wr_reg32(hw, addr & ~(MEMWIN0_APERTURE - 1),
+ PCIE_MEM_ACCESS_OFFSET);
+ csio_rd_reg32(hw, PCIE_MEM_ACCESS_OFFSET);
+
+ /* Collecting data 4 bytes at a time upto MEMWIN0_APERTURE */
+ for (i = 0; i < MEMWIN0_APERTURE; i = i + sizeof(__be32)) {
+ if (dir)
+ *data++ = csio_rd_reg32(hw, (MEMWIN0_BASE + i));
+ else
+ csio_wr_reg32(hw, *data++, (MEMWIN0_BASE + i));
+ }
+
+ return 0;
+}
+
+/*
+ * csio_memory_rw - read/write EDC 0, EDC 1 or MC via PCIE memory window
+ * @hw: the csio_hw
+ * @mtype: memory type: MEM_EDC0, MEM_EDC1 or MEM_MC
+ * @addr: address within indicated memory type
+ * @len: amount of memory to transfer
+ * @buf: host memory buffer
+ * @dir: direction of transfer 1 => read, 0 => write
+ *
+ * Reads/writes an [almost] arbitrary memory region in the firmware: the
+ * firmware memory address, length and host buffer must be aligned on
+ * 32-bit boudaries. The memory is transferred as a raw byte sequence
+ * from/to the firmware's memory. If this memory contains data
+ * structures which contain multi-byte integers, it's the callers
+ * responsibility to perform appropriate byte order conversions.
+ */
+static int
+csio_memory_rw(struct csio_hw *hw, int mtype, u32 addr, u32 len,
+ uint32_t *buf, int dir)
+{
+ uint32_t pos, start, end, offset, memoffset;
+ int ret;
+ __be32 *data;
+
+ /*
+ * Argument sanity checks ...
+ */
+ if ((addr & 0x3) || (len & 0x3))
+ return -EINVAL;
+
+ data = kzalloc(MEMWIN0_APERTURE, GFP_KERNEL);
+ if (!data)
+ return -ENOMEM;
+
+ /* Offset into the region of memory which is being accessed
+ * MEM_EDC0 = 0
+ * MEM_EDC1 = 1
+ * MEM_MC = 2
+ */
+ memoffset = (mtype * (5 * 1024 * 1024));
+
+ /* Determine the PCIE_MEM_ACCESS_OFFSET */
+ addr = addr + memoffset;
+
+ /*
+ * The underlaying EDC/MC read routines read MEMWIN0_APERTURE bytes
+ * at a time so we need to round down the start and round up the end.
+ * We'll start copying out of the first line at (addr - start) a word
+ * at a time.
+ */
+ start = addr & ~(MEMWIN0_APERTURE-1);
+ end = (addr + len + MEMWIN0_APERTURE-1) & ~(MEMWIN0_APERTURE-1);
+ offset = (addr - start)/sizeof(__be32);
+
+ for (pos = start; pos < end; pos += MEMWIN0_APERTURE, offset = 0) {
+ /*
+ * If we're writing, copy the data from the caller's memory
+ * buffer
+ */
+ if (!dir) {
+ /*
+ * If we're doing a partial write, then we need to do
+ * a read-modify-write ...
+ */
+ if (offset || len < MEMWIN0_APERTURE) {
+ ret = csio_mem_win_rw(hw, pos, data, 1);
+ if (ret) {
+ kfree(data);
+ return ret;
+ }
+ }
+ while (offset < (MEMWIN0_APERTURE/sizeof(__be32)) &&
+ len > 0) {
+ data[offset++] = *buf++;
+ len -= sizeof(__be32);
+ }
+ }
+
+ /*
+ * Transfer a block of memory and bail if there's an error.
+ */
+ ret = csio_mem_win_rw(hw, pos, data, dir);
+ if (ret) {
+ kfree(data);
+ return ret;
+ }
+
+ /*
+ * If we're reading, copy the data into the caller's memory
+ * buffer.
+ */
+ if (dir)
+ while (offset < (MEMWIN0_APERTURE/sizeof(__be32)) &&
+ len > 0) {
+ *buf++ = data[offset++];
+ len -= sizeof(__be32);
+ }
+ }
+
+ kfree(data);
+
+ return 0;
+}
+
+static int
+csio_memory_write(struct csio_hw *hw, int mtype, u32 addr, u32 len, __be32 *buf)
+{
+ return csio_memory_rw(hw, mtype, addr, len, buf, 0);
+}
+
+/*
+ * EEPROM reads take a few tens of us while writes can take a bit over 5 ms.
+ */
+#define EEPROM_MAX_RD_POLL 40
+#define EEPROM_MAX_WR_POLL 6
+#define EEPROM_STAT_ADDR 0x7bfc
+#define VPD_BASE 0x400
+#define VPD_BASE_OLD 0
+#define VPD_LEN 512
+#define VPD_INFO_FLD_HDR_SIZE 3
+
+/*
+ * csio_hw_seeprom_read - read a serial EEPROM location
+ * @hw: hw to read
+ * @addr: EEPROM virtual address
+ * @data: where to store the read data
+ *
+ * Read a 32-bit word from a location in serial EEPROM using the card's PCI
+ * VPD capability. Note that this function must be called with a virtual
+ * address.
+ */
+static int
+csio_hw_seeprom_read(struct csio_hw *hw, uint32_t addr, uint32_t *data)
+{
+ uint16_t val = 0;
+ int attempts = EEPROM_MAX_RD_POLL;
+ uint32_t base = hw->params.pci.vpd_cap_addr;
+
+ if (addr >= EEPROMVSIZE || (addr & 3))
+ return -EINVAL;
+
+ pci_write_config_word(hw->pdev, base + PCI_VPD_ADDR, (uint16_t)addr);
+
+ do {
+ udelay(10);
+ pci_read_config_word(hw->pdev, base + PCI_VPD_ADDR, &val);
+ } while (!(val & PCI_VPD_ADDR_F) && --attempts);
+
+ if (!(val & PCI_VPD_ADDR_F)) {
+ csio_err(hw, "reading EEPROM address 0x%x failed\n", addr);
+ return -EINVAL;
+ }
+
+ pci_read_config_dword(hw->pdev, base + PCI_VPD_DATA, data);
+ *data = le32_to_cpu(*data);
+ return 0;
+}
+
+/*
+ * Partial EEPROM Vital Product Data structure. Includes only the ID and
+ * VPD-R sections.
+ */
+struct t4_vpd_hdr {
+ u8 id_tag;
+ u8 id_len[2];
+ u8 id_data[ID_LEN];
+ u8 vpdr_tag;
+ u8 vpdr_len[2];
+};
+
+/*
+ * csio_hw_get_vpd_keyword_val - Locates an information field keyword in
+ * the VPD
+ * @v: Pointer to buffered vpd data structure
+ * @kw: The keyword to search for
+ *
+ * Returns the value of the information field keyword or
+ * -EINVAL otherwise.
+ */
+static int
+csio_hw_get_vpd_keyword_val(const struct t4_vpd_hdr *v, const char *kw)
+{
+ int32_t i;
+ int32_t offset , len;
+ const uint8_t *buf = &v->id_tag;
+ const uint8_t *vpdr_len = &v->vpdr_tag;
+ offset = sizeof(struct t4_vpd_hdr);
+ len = (uint16_t)vpdr_len[1] + ((uint16_t)vpdr_len[2] << 8);
+
+ if (len + sizeof(struct t4_vpd_hdr) > VPD_LEN)
+ return -EINVAL;
+
+ for (i = offset; (i + VPD_INFO_FLD_HDR_SIZE) <= (offset + len);) {
+ if (memcmp(buf + i , kw, 2) == 0) {
+ i += VPD_INFO_FLD_HDR_SIZE;
+ return i;
+ }
+
+ i += VPD_INFO_FLD_HDR_SIZE + buf[i+2];
+ }
+
+ return -EINVAL;
+}
+
+static int
+csio_pci_capability(struct pci_dev *pdev, int cap, int *pos)
+{
+ *pos = pci_find_capability(pdev, cap);
+ if (*pos)
+ return 0;
+
+ return -1;
+}
+
+/*
+ * csio_hw_get_vpd_params - read VPD parameters from VPD EEPROM
+ * @hw: HW module
+ * @p: where to store the parameters
+ *
+ * Reads card parameters stored in VPD EEPROM.
+ */
+static int
+csio_hw_get_vpd_params(struct csio_hw *hw, struct csio_vpd *p)
+{
+ int i, ret, ec, sn, addr;
+ uint8_t *vpd, csum;
+ const struct t4_vpd_hdr *v;
+ /* To get around compilation warning from strstrip */
+ char *s;
+
+ if (csio_is_valid_vpd(hw))
+ return 0;
+
+ ret = csio_pci_capability(hw->pdev, PCI_CAP_ID_VPD,
+ &hw->params.pci.vpd_cap_addr);
+ if (ret)
+ return -EINVAL;
+
+ vpd = kzalloc(VPD_LEN, GFP_ATOMIC);
+ if (vpd == NULL)
+ return -ENOMEM;
+
+ /*
+ * Card information normally starts at VPD_BASE but early cards had
+ * it at 0.
+ */
+ ret = csio_hw_seeprom_read(hw, VPD_BASE, (uint32_t *)(vpd));
+ addr = *vpd == 0x82 ? VPD_BASE : VPD_BASE_OLD;
+
+ for (i = 0; i < VPD_LEN; i += 4) {
+ ret = csio_hw_seeprom_read(hw, addr + i, (uint32_t *)(vpd + i));
+ if (ret) {
+ kfree(vpd);
+ return ret;
+ }
+ }
+
+ /* Reset the VPD flag! */
+ hw->flags &= (~CSIO_HWF_VPD_VALID);
+
+ v = (const struct t4_vpd_hdr *)vpd;
+
+#define FIND_VPD_KW(var, name) do { \
+ var = csio_hw_get_vpd_keyword_val(v, name); \
+ if (var < 0) { \
+ csio_err(hw, "missing VPD keyword " name "\n"); \
+ kfree(vpd); \
+ return -EINVAL; \
+ } \
+} while (0)
+
+ FIND_VPD_KW(i, "RV");
+ for (csum = 0; i >= 0; i--)
+ csum += vpd[i];
+
+ if (csum) {
+ csio_err(hw, "corrupted VPD EEPROM, actual csum %u\n", csum);
+ kfree(vpd);
+ return -EINVAL;
+ }
+ FIND_VPD_KW(ec, "EC");
+ FIND_VPD_KW(sn, "SN");
+#undef FIND_VPD_KW
+
+ memcpy(p->id, v->id_data, ID_LEN);
+ s = strstrip(p->id);
+ memcpy(p->ec, vpd + ec, EC_LEN);
+ s = strstrip(p->ec);
+ i = vpd[sn - VPD_INFO_FLD_HDR_SIZE + 2];
+ memcpy(p->sn, vpd + sn, min(i, SERNUM_LEN));
+ s = strstrip(p->sn);
+
+ csio_valid_vpd_copied(hw);
+
+ kfree(vpd);
+ return 0;
+}
+
+/*
+ * csio_hw_sf1_read - read data from the serial flash
+ * @hw: the HW module
+ * @byte_cnt: number of bytes to read
+ * @cont: whether another operation will be chained
+ * @lock: whether to lock SF for PL access only
+ * @valp: where to store the read data
+ *
+ * Reads up to 4 bytes of data from the serial flash. The location of
+ * the read needs to be specified prior to calling this by issuing the
+ * appropriate commands to the serial flash.
+ */
+static int
+csio_hw_sf1_read(struct csio_hw *hw, uint32_t byte_cnt, int32_t cont,
+ int32_t lock, uint32_t *valp)
+{
+ int ret;
+
+ if (!byte_cnt || byte_cnt > 4)
+ return -EINVAL;
+ if (csio_rd_reg32(hw, SF_OP) & SF_BUSY)
+ return -EBUSY;
+
+ cont = cont ? SF_CONT : 0;
+ lock = lock ? SF_LOCK : 0;
+
+ csio_wr_reg32(hw, lock | cont | BYTECNT(byte_cnt - 1), SF_OP);
+ ret = csio_hw_wait_op_done_val(hw, SF_OP, SF_BUSY, 0, SF_ATTEMPTS,
+ 10, NULL);
+ if (!ret)
+ *valp = csio_rd_reg32(hw, SF_DATA);
+ return ret;
+}
+
+/*
+ * csio_hw_sf1_write - write data to the serial flash
+ * @hw: the HW module
+ * @byte_cnt: number of bytes to write
+ * @cont: whether another operation will be chained
+ * @lock: whether to lock SF for PL access only
+ * @val: value to write
+ *
+ * Writes up to 4 bytes of data to the serial flash. The location of
+ * the write needs to be specified prior to calling this by issuing the
+ * appropriate commands to the serial flash.
+ */
+static int
+csio_hw_sf1_write(struct csio_hw *hw, uint32_t byte_cnt, uint32_t cont,
+ int32_t lock, uint32_t val)
+{
+ if (!byte_cnt || byte_cnt > 4)
+ return -EINVAL;
+ if (csio_rd_reg32(hw, SF_OP) & SF_BUSY)
+ return -EBUSY;
+
+ cont = cont ? SF_CONT : 0;
+ lock = lock ? SF_LOCK : 0;
+
+ csio_wr_reg32(hw, val, SF_DATA);
+ csio_wr_reg32(hw, cont | BYTECNT(byte_cnt - 1) | OP_WR | lock, SF_OP);
+
+ return csio_hw_wait_op_done_val(hw, SF_OP, SF_BUSY, 0, SF_ATTEMPTS,
+ 10, NULL);
+}
+
+/*
+ * csio_hw_flash_wait_op - wait for a flash operation to complete
+ * @hw: the HW module
+ * @attempts: max number of polls of the status register
+ * @delay: delay between polls in ms
+ *
+ * Wait for a flash operation to complete by polling the status register.
+ */
+static int
+csio_hw_flash_wait_op(struct csio_hw *hw, int32_t attempts, int32_t delay)
+{
+ int ret;
+ uint32_t status;
+
+ while (1) {
+ ret = csio_hw_sf1_write(hw, 1, 1, 1, SF_RD_STATUS);
+ if (ret != 0)
+ return ret;
+
+ ret = csio_hw_sf1_read(hw, 1, 0, 1, &status);
+ if (ret != 0)
+ return ret;
+
+ if (!(status & 1))
+ return 0;
+ if (--attempts == 0)
+ return -EAGAIN;
+ if (delay)
+ msleep(delay);
+ }
+}
+
+/*
+ * csio_hw_read_flash - read words from serial flash
+ * @hw: the HW module
+ * @addr: the start address for the read
+ * @nwords: how many 32-bit words to read
+ * @data: where to store the read data
+ * @byte_oriented: whether to store data as bytes or as words
+ *
+ * Read the specified number of 32-bit words from the serial flash.
+ * If @byte_oriented is set the read data is stored as a byte array
+ * (i.e., big-endian), otherwise as 32-bit words in the platform's
+ * natural endianess.
+ */
+static int
+csio_hw_read_flash(struct csio_hw *hw, uint32_t addr, uint32_t nwords,
+ uint32_t *data, int32_t byte_oriented)
+{
+ int ret;
+
+ if (addr + nwords * sizeof(uint32_t) > hw->params.sf_size || (addr & 3))
+ return -EINVAL;
+
+ addr = swab32(addr) | SF_RD_DATA_FAST;
+
+ ret = csio_hw_sf1_write(hw, 4, 1, 0, addr);
+ if (ret != 0)
+ return ret;
+
+ ret = csio_hw_sf1_read(hw, 1, 1, 0, data);
+ if (ret != 0)
+ return ret;
+
+ for ( ; nwords; nwords--, data++) {
+ ret = csio_hw_sf1_read(hw, 4, nwords > 1, nwords == 1, data);
+ if (nwords == 1)
+ csio_wr_reg32(hw, 0, SF_OP); /* unlock SF */
+ if (ret)
+ return ret;
+ if (byte_oriented)
+ *data = htonl(*data);
+ }
+ return 0;
+}
+
+/*
+ * csio_hw_write_flash - write up to a page of data to the serial flash
+ * @hw: the hw
+ * @addr: the start address to write
+ * @n: length of data to write in bytes
+ * @data: the data to write
+ *
+ * Writes up to a page of data (256 bytes) to the serial flash starting
+ * at the given address. All the data must be written to the same page.
+ */
+static int
+csio_hw_write_flash(struct csio_hw *hw, uint32_t addr,
+ uint32_t n, const uint8_t *data)
+{
+ int ret = -EINVAL;
+ uint32_t buf[64];
+ uint32_t i, c, left, val, offset = addr & 0xff;
+
+ if (addr >= hw->params.sf_size || offset + n > SF_PAGE_SIZE)
+ return -EINVAL;
+
+ val = swab32(addr) | SF_PROG_PAGE;
+
+ ret = csio_hw_sf1_write(hw, 1, 0, 1, SF_WR_ENABLE);
+ if (ret != 0)
+ goto unlock;
+
+ ret = csio_hw_sf1_write(hw, 4, 1, 1, val);
+ if (ret != 0)
+ goto unlock;
+
+ for (left = n; left; left -= c) {
+ c = min(left, 4U);
+ for (val = 0, i = 0; i < c; ++i)
+ val = (val << 8) + *data++;
+
+ ret = csio_hw_sf1_write(hw, c, c != left, 1, val);
+ if (ret)
+ goto unlock;
+ }
+ ret = csio_hw_flash_wait_op(hw, 8, 1);
+ if (ret)
+ goto unlock;
+
+ csio_wr_reg32(hw, 0, SF_OP); /* unlock SF */
+
+ /* Read the page to verify the write succeeded */
+ ret = csio_hw_read_flash(hw, addr & ~0xff, ARRAY_SIZE(buf), buf, 1);
+ if (ret)
+ return ret;
+
+ if (memcmp(data - n, (uint8_t *)buf + offset, n)) {
+ csio_err(hw,
+ "failed to correctly write the flash page at %#x\n",
+ addr);
+ return -EINVAL;
+ }
+
+ return 0;
+
+unlock:
+ csio_wr_reg32(hw, 0, SF_OP); /* unlock SF */
+ return ret;
+}
+
+/*
+ * csio_hw_flash_erase_sectors - erase a range of flash sectors
+ * @hw: the HW module
+ * @start: the first sector to erase
+ * @end: the last sector to erase
+ *
+ * Erases the sectors in the given inclusive range.
+ */
+static int
+csio_hw_flash_erase_sectors(struct csio_hw *hw, int32_t start, int32_t end)
+{
+ int ret = 0;
+
+ while (start <= end) {
+
+ ret = csio_hw_sf1_write(hw, 1, 0, 1, SF_WR_ENABLE);
+ if (ret != 0)
+ goto out;
+
+ ret = csio_hw_sf1_write(hw, 4, 0, 1,
+ SF_ERASE_SECTOR | (start << 8));
+ if (ret != 0)
+ goto out;
+
+ ret = csio_hw_flash_wait_op(hw, 14, 500);
+ if (ret != 0)
+ goto out;
+
+ start++;
+ }
+out:
+ if (ret)
+ csio_err(hw, "erase of flash sector %d failed, error %d\n",
+ start, ret);
+ csio_wr_reg32(hw, 0, SF_OP); /* unlock SF */
+ return 0;
+}
+
+/*
+ * csio_hw_flash_cfg_addr - return the address of the flash
+ * configuration file
+ * @hw: the HW module
+ *
+ * Return the address within the flash where the Firmware Configuration
+ * File is stored.
+ */
+static unsigned int
+csio_hw_flash_cfg_addr(struct csio_hw *hw)
+{
+ if (hw->params.sf_size == 0x100000)
+ return FPGA_FLASH_CFG_OFFSET;
+ else
+ return FLASH_CFG_OFFSET;
+}
+
+static void
+csio_hw_print_fw_version(struct csio_hw *hw, char *str)
+{
+ csio_info(hw, "%s: %u.%u.%u.%u\n", str,
+ FW_HDR_FW_VER_MAJOR_GET(hw->fwrev),
+ FW_HDR_FW_VER_MINOR_GET(hw->fwrev),
+ FW_HDR_FW_VER_MICRO_GET(hw->fwrev),
+ FW_HDR_FW_VER_BUILD_GET(hw->fwrev));
+}
+
+/*
+ * csio_hw_get_fw_version - read the firmware version
+ * @hw: HW module
+ * @vers: where to place the version
+ *
+ * Reads the FW version from flash.
+ */
+static int
+csio_hw_get_fw_version(struct csio_hw *hw, uint32_t *vers)
+{
+ return csio_hw_read_flash(hw, FW_IMG_START +
+ offsetof(struct fw_hdr, fw_ver), 1,
+ vers, 0);
+}
+
+/*
+ * csio_hw_get_tp_version - read the TP microcode version
+ * @hw: HW module
+ * @vers: where to place the version
+ *
+ * Reads the TP microcode version from flash.
+ */
+static int
+csio_hw_get_tp_version(struct csio_hw *hw, u32 *vers)
+{
+ return csio_hw_read_flash(hw, FLASH_FW_START +
+ offsetof(struct fw_hdr, tp_microcode_ver), 1,
+ vers, 0);
+}
+
+/*
+ * csio_hw_check_fw_version - check if the FW is compatible with
+ * this driver
+ * @hw: HW module
+ *
+ * Checks if an adapter's FW is compatible with the driver. Returns 0
+ * if there's exact match, a negative error if the version could not be
+ * read or there's a major/minor version mismatch/minor.
+ */
+static int
+csio_hw_check_fw_version(struct csio_hw *hw)
+{
+ int ret, major, minor, micro;
+
+ ret = csio_hw_get_fw_version(hw, &hw->fwrev);
+ if (!ret)
+ ret = csio_hw_get_tp_version(hw, &hw->tp_vers);
+ if (ret)
+ return ret;
+
+ major = FW_HDR_FW_VER_MAJOR_GET(hw->fwrev);
+ minor = FW_HDR_FW_VER_MINOR_GET(hw->fwrev);
+ micro = FW_HDR_FW_VER_MICRO_GET(hw->fwrev);
+
+ if (major != FW_VERSION_MAJOR) { /* major mismatch - fail */
+ csio_err(hw, "card FW has major version %u, driver wants %u\n",
+ major, FW_VERSION_MAJOR);
+ return -EINVAL;
+ }
+
+ if (minor == FW_VERSION_MINOR && micro == FW_VERSION_MICRO)
+ return 0; /* perfect match */
+
+ /* Minor/micro version mismatch */
+ return -EINVAL;
+}
+
+/*
+ * csio_hw_fw_dload - download firmware.
+ * @hw: HW module
+ * @fw_data: firmware image to write.
+ * @size: image size
+ *
+ * Write the supplied firmware image to the card's serial flash.
+ */
+static int
+csio_hw_fw_dload(struct csio_hw *hw, uint8_t *fw_data, uint32_t size)
+{
+ uint32_t csum;
+ int32_t addr;
+ int ret;
+ uint32_t i;
+ uint8_t first_page[SF_PAGE_SIZE];
+ const uint32_t *p = (const uint32_t *)fw_data;
+ struct fw_hdr *hdr = (struct fw_hdr *)fw_data;
+ uint32_t sf_sec_size;
+
+ if ((!hw->params.sf_size) || (!hw->params.sf_nsec)) {
+ csio_err(hw, "Serial Flash data invalid\n");
+ return -EINVAL;
+ }
+
+ if (!size) {
+ csio_err(hw, "FW image has no data\n");
+ return -EINVAL;
+ }
+
+ if (size & 511) {
+ csio_err(hw, "FW image size not multiple of 512 bytes\n");
+ return -EINVAL;
+ }
+
+ if (ntohs(hdr->len512) * 512 != size) {
+ csio_err(hw, "FW image size differs from size in FW header\n");
+ return -EINVAL;
+ }
+
+ if (size > FW_MAX_SIZE) {
+ csio_err(hw, "FW image too large, max is %u bytes\n",
+ FW_MAX_SIZE);
+ return -EINVAL;
+ }
+
+ for (csum = 0, i = 0; i < size / sizeof(csum); i++)
+ csum += ntohl(p[i]);
+
+ if (csum != 0xffffffff) {
+ csio_err(hw, "corrupted firmware image, checksum %#x\n", csum);
+ return -EINVAL;
+ }
+
+ sf_sec_size = hw->params.sf_size / hw->params.sf_nsec;
+ i = DIV_ROUND_UP(size, sf_sec_size); /* # of sectors spanned */
+
+ csio_dbg(hw, "Erasing sectors... start:%d end:%d\n",
+ FW_START_SEC, FW_START_SEC + i - 1);
+
+ ret = csio_hw_flash_erase_sectors(hw, FW_START_SEC,
+ FW_START_SEC + i - 1);
+ if (ret) {
+ csio_err(hw, "Flash Erase failed\n");
+ goto out;
+ }
+
+ /*
+ * We write the correct version at the end so the driver can see a bad
+ * version if the FW write fails. Start by writing a copy of the
+ * first page with a bad version.
+ */
+ memcpy(first_page, fw_data, SF_PAGE_SIZE);
+ ((struct fw_hdr *)first_page)->fw_ver = htonl(0xffffffff);
+ ret = csio_hw_write_flash(hw, FW_IMG_START, SF_PAGE_SIZE, first_page);
+ if (ret)
+ goto out;
+
+ csio_dbg(hw, "Writing Flash .. start:%d end:%d\n",
+ FW_IMG_START, FW_IMG_START + size);
+
+ addr = FW_IMG_START;
+ for (size -= SF_PAGE_SIZE; size; size -= SF_PAGE_SIZE) {
+ addr += SF_PAGE_SIZE;
+ fw_data += SF_PAGE_SIZE;
+ ret = csio_hw_write_flash(hw, addr, SF_PAGE_SIZE, fw_data);
+ if (ret)
+ goto out;
+ }
+
+ ret = csio_hw_write_flash(hw,
+ FW_IMG_START +
+ offsetof(struct fw_hdr, fw_ver),
+ sizeof(hdr->fw_ver),
+ (const uint8_t *)&hdr->fw_ver);
+
+out:
+ if (ret)
+ csio_err(hw, "firmware download failed, error %d\n", ret);
+ return ret;
+}
+
+static int
+csio_hw_get_flash_params(struct csio_hw *hw)
+{
+ int ret;
+ uint32_t info = 0;
+
+ ret = csio_hw_sf1_write(hw, 1, 1, 0, SF_RD_ID);
+ if (!ret)
+ ret = csio_hw_sf1_read(hw, 3, 0, 1, &info);
+ csio_wr_reg32(hw, 0, SF_OP); /* unlock SF */
+ if (ret != 0)
+ return ret;
+
+ if ((info & 0xff) != 0x20) /* not a Numonix flash */
+ return -EINVAL;
+ info >>= 16; /* log2 of size */
+ if (info >= 0x14 && info < 0x18)
+ hw->params.sf_nsec = 1 << (info - 16);
+ else if (info == 0x18)
+ hw->params.sf_nsec = 64;
+ else
+ return -EINVAL;
+ hw->params.sf_size = 1 << info;
+
+ return 0;
+}
+
+static void
+csio_set_pcie_completion_timeout(struct csio_hw *hw, u8 range)
+{
+ uint16_t val;
+ uint32_t pcie_cap;
+
+ if (!csio_pci_capability(hw->pdev, PCI_CAP_ID_EXP, &pcie_cap)) {
+ pci_read_config_word(hw->pdev,
+ pcie_cap + PCI_EXP_DEVCTL2, &val);
+ val &= 0xfff0;
+ val |= range ;
+ pci_write_config_word(hw->pdev,
+ pcie_cap + PCI_EXP_DEVCTL2, val);
+ }
+}
+
+
+/*
+ * Return the specified PCI-E Configuration Space register from our Physical
+ * Function. We try first via a Firmware LDST Command since we prefer to let
+ * the firmware own all of these registers, but if that fails we go for it
+ * directly ourselves.
+ */
+static uint32_t
+csio_read_pcie_cfg4(struct csio_hw *hw, int reg)
+{
+ u32 val = 0;
+ struct csio_mb *mbp;
+ int rv;
+ struct fw_ldst_cmd *ldst_cmd;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ pci_read_config_dword(hw->pdev, reg, &val);
+ return val;
+ }
+
+ csio_mb_ldst(hw, mbp, CSIO_MB_DEFAULT_TMO, reg);
+
+ rv = csio_mb_issue(hw, mbp);
+
+ /*
+ * If the LDST Command suucceeded, exctract the returned register
+ * value. Otherwise read it directly ourself.
+ */
+ if (rv == 0) {
+ ldst_cmd = (struct fw_ldst_cmd *)(mbp->mb);
+ val = ntohl(ldst_cmd->u.pcie.data[0]);
+ } else
+ pci_read_config_dword(hw->pdev, reg, &val);
+
+ mempool_free(mbp, hw->mb_mempool);
+
+ return val;
+} /* csio_read_pcie_cfg4 */
+
+static int
+csio_hw_set_mem_win(struct csio_hw *hw)
+{
+ u32 bar0;
+
+ /*
+ * Truncation intentional: we only read the bottom 32-bits of the
+ * 64-bit BAR0/BAR1 ... We use the hardware backdoor mechanism to
+ * read BAR0 instead of using pci_resource_start() because we could be
+ * operating from within a Virtual Machine which is trapping our
+ * accesses to our Configuration Space and we need to set up the PCI-E
+ * Memory Window decoders with the actual addresses which will be
+ * coming across the PCI-E link.
+ */
+ bar0 = csio_read_pcie_cfg4(hw, PCI_BASE_ADDRESS_0);
+ bar0 &= PCI_BASE_ADDRESS_MEM_MASK;
+
+ /*
+ * Set up memory window for accessing adapter memory ranges. (Read
+ * back MA register to ensure that changes propagate before we attempt
+ * to use the new values.)
+ */
+ csio_wr_reg32(hw, (bar0 + MEMWIN0_BASE) | BIR(0) |
+ WINDOW(ilog2(MEMWIN0_APERTURE) - 10),
+ PCIE_MEM_ACCESS_REG(PCIE_MEM_ACCESS_BASE_WIN, 0));
+ csio_wr_reg32(hw, (bar0 + MEMWIN1_BASE) | BIR(0) |
+ WINDOW(ilog2(MEMWIN1_APERTURE) - 10),
+ PCIE_MEM_ACCESS_REG(PCIE_MEM_ACCESS_BASE_WIN, 1));
+ csio_wr_reg32(hw, (bar0 + MEMWIN2_BASE) | BIR(0) |
+ WINDOW(ilog2(MEMWIN2_APERTURE) - 10),
+ PCIE_MEM_ACCESS_REG(PCIE_MEM_ACCESS_BASE_WIN, 2));
+ csio_rd_reg32(hw, PCIE_MEM_ACCESS_REG(PCIE_MEM_ACCESS_BASE_WIN, 2));
+ return 0;
+} /* csio_hw_set_mem_win */
+
+
+
+/*****************************************************************************/
+/* HW State machine assists */
+/*****************************************************************************/
+
+static int
+csio_hw_dev_ready(struct csio_hw *hw)
+{
+ uint32_t reg;
+ int cnt = 6;
+
+ while (((reg = csio_rd_reg32(hw, PL_WHOAMI)) == 0xFFFFFFFF) &&
+ (--cnt != 0))
+ mdelay(100);
+
+ if ((cnt == 0) && (((int32_t)(SOURCEPF_GET(reg)) < 0) ||
+ (SOURCEPF_GET(reg) >= CSIO_MAX_PFN))) {
+ csio_err(hw, "PL_WHOAMI returned 0x%x, cnt:%d\n", reg, cnt);
+ return -EIO;
+ }
+
+ hw->pfn = SOURCEPF_GET(reg);
+
+ return 0;
+}
+
+/*
+ * csio_do_hello - Perform the HELLO FW Mailbox command and process response.
+ * @hw: HW module
+ * @state: Device state
+ *
+ * FW_HELLO_CMD has to be polled for completion.
+ */
+static int
+csio_do_hello(struct csio_hw *hw, enum csio_dev_state *state)
+{
+ struct csio_mb *mbp;
+ int rv = 0;
+ enum csio_dev_master master;
+ enum fw_retval retval;
+ uint8_t mpfn;
+ char state_str[16];
+ int retries = FW_CMD_HELLO_RETRIES;
+
+ memset(state_str, 0, sizeof(state_str));
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ rv = -ENOMEM;
+ CSIO_INC_STATS(hw, n_err_nomem);
+ goto out;
+ }
+
+ master = csio_force_master ? CSIO_MASTER_MUST : CSIO_MASTER_MAY;
+
+retry:
+ csio_mb_hello(hw, mbp, CSIO_MB_DEFAULT_TMO, hw->pfn,
+ hw->pfn, master, NULL);
+
+ rv = csio_mb_issue(hw, mbp);
+ if (rv) {
+ csio_err(hw, "failed to issue HELLO cmd. ret:%d.\n", rv);
+ goto out_free_mb;
+ }
+
+ csio_mb_process_hello_rsp(hw, mbp, &retval, state, &mpfn);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "HELLO cmd failed with ret: %d\n", retval);
+ rv = -EINVAL;
+ goto out_free_mb;
+ }
+
+ /* Firmware has designated us to be master */
+ if (hw->pfn == mpfn) {
+ hw->flags |= CSIO_HWF_MASTER;
+ } else if (*state == CSIO_DEV_STATE_UNINIT) {
+ /*
+ * If we're not the Master PF then we need to wait around for
+ * the Master PF Driver to finish setting up the adapter.
+ *
+ * Note that we also do this wait if we're a non-Master-capable
+ * PF and there is no current Master PF; a Master PF may show up
+ * momentarily and we wouldn't want to fail pointlessly. (This
+ * can happen when an OS loads lots of different drivers rapidly
+ * at the same time). In this case, the Master PF returned by
+ * the firmware will be PCIE_FW_MASTER_MASK so the test below
+ * will work ...
+ */
+
+ int waiting = FW_CMD_HELLO_TIMEOUT;
+
+ /*
+ * Wait for the firmware to either indicate an error or
+ * initialized state. If we see either of these we bail out
+ * and report the issue to the caller. If we exhaust the
+ * "hello timeout" and we haven't exhausted our retries, try
+ * again. Otherwise bail with a timeout error.
+ */
+ for (;;) {
+ uint32_t pcie_fw;
+
+ msleep(50);
+ waiting -= 50;
+
+ /*
+ * If neither Error nor Initialialized are indicated
+ * by the firmware keep waiting till we exaust our
+ * timeout ... and then retry if we haven't exhausted
+ * our retries ...
+ */
+ pcie_fw = csio_rd_reg32(hw, PCIE_FW);
+ if (!(pcie_fw & (PCIE_FW_ERR|PCIE_FW_INIT))) {
+ if (waiting <= 0) {
+ if (retries-- > 0)
+ goto retry;
+
+ rv = -ETIMEDOUT;
+ break;
+ }
+ continue;
+ }
+
+ /*
+ * We either have an Error or Initialized condition
+ * report errors preferentially.
+ */
+ if (state) {
+ if (pcie_fw & PCIE_FW_ERR) {
+ *state = CSIO_DEV_STATE_ERR;
+ rv = -ETIMEDOUT;
+ } else if (pcie_fw & PCIE_FW_INIT)
+ *state = CSIO_DEV_STATE_INIT;
+ }
+
+ /*
+ * If we arrived before a Master PF was selected and
+ * there's not a valid Master PF, grab its identity
+ * for our caller.
+ */
+ if (mpfn == PCIE_FW_MASTER_MASK &&
+ (pcie_fw & PCIE_FW_MASTER_VLD))
+ mpfn = PCIE_FW_MASTER_GET(pcie_fw);
+ break;
+ }
+ hw->flags &= ~CSIO_HWF_MASTER;
+ }
+
+ switch (*state) {
+ case CSIO_DEV_STATE_UNINIT:
+ strcpy(state_str, "Initializing");
+ break;
+ case CSIO_DEV_STATE_INIT:
+ strcpy(state_str, "Initialized");
+ break;
+ case CSIO_DEV_STATE_ERR:
+ strcpy(state_str, "Error");
+ break;
+ default:
+ strcpy(state_str, "Unknown");
+ break;
+ }
+
+ if (hw->pfn == mpfn)
+ csio_info(hw, "PF: %d, Coming up as MASTER, HW state: %s\n",
+ hw->pfn, state_str);
+ else
+ csio_info(hw,
+ "PF: %d, Coming up as SLAVE, Master PF: %d, HW state: %s\n",
+ hw->pfn, mpfn, state_str);
+
+out_free_mb:
+ mempool_free(mbp, hw->mb_mempool);
+out:
+ return rv;
+}
+
+/*
+ * csio_do_bye - Perform the BYE FW Mailbox command and process response.
+ * @hw: HW module
+ *
+ */
+static int
+csio_do_bye(struct csio_hw *hw)
+{
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ csio_mb_bye(hw, mbp, CSIO_MB_DEFAULT_TMO, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of BYE command failed\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ mempool_free(mbp, hw->mb_mempool);
+
+ return 0;
+}
+
+/*
+ * csio_do_reset- Perform the device reset.
+ * @hw: HW module
+ * @fw_rst: FW reset
+ *
+ * If fw_rst is set, issues FW reset mbox cmd otherwise
+ * does PIO reset.
+ * Performs reset of the function.
+ */
+static int
+csio_do_reset(struct csio_hw *hw, bool fw_rst)
+{
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+
+ if (!fw_rst) {
+ /* PIO reset */
+ csio_wr_reg32(hw, PIORSTMODE | PIORST, PL_RST);
+ mdelay(2000);
+ return 0;
+ }
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ csio_mb_reset(hw, mbp, CSIO_MB_DEFAULT_TMO,
+ PIORSTMODE | PIORST, 0, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of RESET command failed.n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "RESET cmd failed with ret:0x%x.\n", retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ mempool_free(mbp, hw->mb_mempool);
+
+ return 0;
+}
+
+static int
+csio_hw_validate_caps(struct csio_hw *hw, struct csio_mb *mbp)
+{
+ struct fw_caps_config_cmd *rsp = (struct fw_caps_config_cmd *)mbp->mb;
+ uint16_t caps;
+
+ caps = ntohs(rsp->fcoecaps);
+
+ if (!(caps & FW_CAPS_CONFIG_FCOE_INITIATOR)) {
+ csio_err(hw, "No FCoE Initiator capability in the firmware.\n");
+ return -EINVAL;
+ }
+
+ if (!(caps & FW_CAPS_CONFIG_FCOE_CTRL_OFLD)) {
+ csio_err(hw, "No FCoE Control Offload capability\n");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/*
+ * csio_hw_fw_halt - issue a reset/halt to FW and put uP into RESET
+ * @hw: the HW module
+ * @mbox: mailbox to use for the FW RESET command (if desired)
+ * @force: force uP into RESET even if FW RESET command fails
+ *
+ * Issues a RESET command to firmware (if desired) with a HALT indication
+ * and then puts the microprocessor into RESET state. The RESET command
+ * will only be issued if a legitimate mailbox is provided (mbox <=
+ * PCIE_FW_MASTER_MASK).
+ *
+ * This is generally used in order for the host to safely manipulate the
+ * adapter without fear of conflicting with whatever the firmware might
+ * be doing. The only way out of this state is to RESTART the firmware
+ * ...
+ */
+static int
+csio_hw_fw_halt(struct csio_hw *hw, uint32_t mbox, int32_t force)
+{
+ enum fw_retval retval = 0;
+
+ /*
+ * If a legitimate mailbox is provided, issue a RESET command
+ * with a HALT indication.
+ */
+ if (mbox <= PCIE_FW_MASTER_MASK) {
+ struct csio_mb *mbp;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ csio_mb_reset(hw, mbp, CSIO_MB_DEFAULT_TMO,
+ PIORSTMODE | PIORST, FW_RESET_CMD_HALT(1),
+ NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of RESET command failed!\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ mempool_free(mbp, hw->mb_mempool);
+ }
+
+ /*
+ * Normally we won't complete the operation if the firmware RESET
+ * command fails but if our caller insists we'll go ahead and put the
+ * uP into RESET. This can be useful if the firmware is hung or even
+ * missing ... We'll have to take the risk of putting the uP into
+ * RESET without the cooperation of firmware in that case.
+ *
+ * We also force the firmware's HALT flag to be on in case we bypassed
+ * the firmware RESET command above or we're dealing with old firmware
+ * which doesn't have the HALT capability. This will serve as a flag
+ * for the incoming firmware to know that it's coming out of a HALT
+ * rather than a RESET ... if it's new enough to understand that ...
+ */
+ if (retval == 0 || force) {
+ csio_set_reg_field(hw, CIM_BOOT_CFG, UPCRST, UPCRST);
+ csio_set_reg_field(hw, PCIE_FW, PCIE_FW_HALT, PCIE_FW_HALT);
+ }
+
+ /*
+ * And we always return the result of the firmware RESET command
+ * even when we force the uP into RESET ...
+ */
+ return retval ? -EINVAL : 0;
+}
+
+/*
+ * csio_hw_fw_restart - restart the firmware by taking the uP out of RESET
+ * @hw: the HW module
+ * @reset: if we want to do a RESET to restart things
+ *
+ * Restart firmware previously halted by csio_hw_fw_halt(). On successful
+ * return the previous PF Master remains as the new PF Master and there
+ * is no need to issue a new HELLO command, etc.
+ *
+ * We do this in two ways:
+ *
+ * 1. If we're dealing with newer firmware we'll simply want to take
+ * the chip's microprocessor out of RESET. This will cause the
+ * firmware to start up from its start vector. And then we'll loop
+ * until the firmware indicates it's started again (PCIE_FW.HALT
+ * reset to 0) or we timeout.
+ *
+ * 2. If we're dealing with older firmware then we'll need to RESET
+ * the chip since older firmware won't recognize the PCIE_FW.HALT
+ * flag and automatically RESET itself on startup.
+ */
+static int
+csio_hw_fw_restart(struct csio_hw *hw, uint32_t mbox, int32_t reset)
+{
+ if (reset) {
+ /*
+ * Since we're directing the RESET instead of the firmware
+ * doing it automatically, we need to clear the PCIE_FW.HALT
+ * bit.
+ */
+ csio_set_reg_field(hw, PCIE_FW, PCIE_FW_HALT, 0);
+
+ /*
+ * If we've been given a valid mailbox, first try to get the
+ * firmware to do the RESET. If that works, great and we can
+ * return success. Otherwise, if we haven't been given a
+ * valid mailbox or the RESET command failed, fall back to
+ * hitting the chip with a hammer.
+ */
+ if (mbox <= PCIE_FW_MASTER_MASK) {
+ csio_set_reg_field(hw, CIM_BOOT_CFG, UPCRST, 0);
+ msleep(100);
+ if (csio_do_reset(hw, true) == 0)
+ return 0;
+ }
+
+ csio_wr_reg32(hw, PIORSTMODE | PIORST, PL_RST);
+ msleep(2000);
+ } else {
+ int ms;
+
+ csio_set_reg_field(hw, CIM_BOOT_CFG, UPCRST, 0);
+ for (ms = 0; ms < FW_CMD_MAX_TIMEOUT; ) {
+ if (!(csio_rd_reg32(hw, PCIE_FW) & PCIE_FW_HALT))
+ return 0;
+ msleep(100);
+ ms += 100;
+ }
+ return -ETIMEDOUT;
+ }
+ return 0;
+}
+
+/*
+ * csio_hw_fw_upgrade - perform all of the steps necessary to upgrade FW
+ * @hw: the HW module
+ * @mbox: mailbox to use for the FW RESET command (if desired)
+ * @fw_data: the firmware image to write
+ * @size: image size
+ * @force: force upgrade even if firmware doesn't cooperate
+ *
+ * Perform all of the steps necessary for upgrading an adapter's
+ * firmware image. Normally this requires the cooperation of the
+ * existing firmware in order to halt all existing activities
+ * but if an invalid mailbox token is passed in we skip that step
+ * (though we'll still put the adapter microprocessor into RESET in
+ * that case).
+ *
+ * On successful return the new firmware will have been loaded and
+ * the adapter will have been fully RESET losing all previous setup
+ * state. On unsuccessful return the adapter may be completely hosed ...
+ * positive errno indicates that the adapter is ~probably~ intact, a
+ * negative errno indicates that things are looking bad ...
+ */
+static int
+csio_hw_fw_upgrade(struct csio_hw *hw, uint32_t mbox,
+ const u8 *fw_data, uint32_t size, int32_t force)
+{
+ const struct fw_hdr *fw_hdr = (const struct fw_hdr *)fw_data;
+ int reset, ret;
+
+ ret = csio_hw_fw_halt(hw, mbox, force);
+ if (ret != 0 && !force)
+ return ret;
+
+ ret = csio_hw_fw_dload(hw, (uint8_t *) fw_data, size);
+ if (ret != 0)
+ return ret;
+
+ /*
+ * Older versions of the firmware don't understand the new
+ * PCIE_FW.HALT flag and so won't know to perform a RESET when they
+ * restart. So for newly loaded older firmware we'll have to do the
+ * RESET for it so it starts up on a clean slate. We can tell if
+ * the newly loaded firmware will handle this right by checking
+ * its header flags to see if it advertises the capability.
+ */
+ reset = ((ntohl(fw_hdr->flags) & FW_HDR_FLAGS_RESET_HALT) == 0);
+ return csio_hw_fw_restart(hw, mbox, reset);
+}
+
+
+/*
+ * csio_hw_fw_config_file - setup an adapter via a Configuration File
+ * @hw: the HW module
+ * @mbox: mailbox to use for the FW command
+ * @mtype: the memory type where the Configuration File is located
+ * @maddr: the memory address where the Configuration File is located
+ * @finiver: return value for CF [fini] version
+ * @finicsum: return value for CF [fini] checksum
+ * @cfcsum: return value for CF computed checksum
+ *
+ * Issue a command to get the firmware to process the Configuration
+ * File located at the specified mtype/maddress. If the Configuration
+ * File is processed successfully and return value pointers are
+ * provided, the Configuration File "[fini] section version and
+ * checksum values will be returned along with the computed checksum.
+ * It's up to the caller to decide how it wants to respond to the
+ * checksums not matching but it recommended that a prominant warning
+ * be emitted in order to help people rapidly identify changed or
+ * corrupted Configuration Files.
+ *
+ * Also note that it's possible to modify things like "niccaps",
+ * "toecaps",etc. between processing the Configuration File and telling
+ * the firmware to use the new configuration. Callers which want to
+ * do this will need to "hand-roll" their own CAPS_CONFIGS commands for
+ * Configuration Files if they want to do this.
+ */
+static int
+csio_hw_fw_config_file(struct csio_hw *hw,
+ unsigned int mtype, unsigned int maddr,
+ uint32_t *finiver, uint32_t *finicsum, uint32_t *cfcsum)
+{
+ struct csio_mb *mbp;
+ struct fw_caps_config_cmd *caps_cmd;
+ int rv = -EINVAL;
+ enum fw_retval ret;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+ /*
+ * Tell the firmware to process the indicated Configuration File.
+ * If there are no errors and the caller has provided return value
+ * pointers for the [fini] section version, checksum and computed
+ * checksum, pass those back to the caller.
+ */
+ caps_cmd = (struct fw_caps_config_cmd *)(mbp->mb);
+ CSIO_INIT_MBP(mbp, caps_cmd, CSIO_MB_DEFAULT_TMO, hw, NULL, 1);
+ caps_cmd->op_to_write =
+ htonl(FW_CMD_OP(FW_CAPS_CONFIG_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_READ);
+ caps_cmd->cfvalid_to_len16 =
+ htonl(FW_CAPS_CONFIG_CMD_CFVALID |
+ FW_CAPS_CONFIG_CMD_MEMTYPE_CF(mtype) |
+ FW_CAPS_CONFIG_CMD_MEMADDR64K_CF(maddr >> 16) |
+ FW_LEN16(*caps_cmd));
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_CAPS_CONFIG_CMD failed!\n");
+ goto out;
+ }
+
+ ret = csio_mb_fw_retval(mbp);
+ if (ret != FW_SUCCESS) {
+ csio_dbg(hw, "FW_CAPS_CONFIG_CMD returned %d!\n", rv);
+ goto out;
+ }
+
+ if (finiver)
+ *finiver = ntohl(caps_cmd->finiver);
+ if (finicsum)
+ *finicsum = ntohl(caps_cmd->finicsum);
+ if (cfcsum)
+ *cfcsum = ntohl(caps_cmd->cfcsum);
+
+ /* Validate device capabilities */
+ if (csio_hw_validate_caps(hw, mbp)) {
+ rv = -ENOENT;
+ goto out;
+ }
+
+ /*
+ * And now tell the firmware to use the configuration we just loaded.
+ */
+ caps_cmd->op_to_write =
+ htonl(FW_CMD_OP(FW_CAPS_CONFIG_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_WRITE);
+ caps_cmd->cfvalid_to_len16 = htonl(FW_LEN16(*caps_cmd));
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_CAPS_CONFIG_CMD failed!\n");
+ goto out;
+ }
+
+ ret = csio_mb_fw_retval(mbp);
+ if (ret != FW_SUCCESS) {
+ csio_dbg(hw, "FW_CAPS_CONFIG_CMD returned %d!\n", rv);
+ goto out;
+ }
+
+ rv = 0;
+out:
+ mempool_free(mbp, hw->mb_mempool);
+ return rv;
+}
+
+/*
+ * csio_get_device_params - Get device parameters.
+ * @hw: HW module
+ *
+ */
+static int
+csio_get_device_params(struct csio_hw *hw)
+{
+ struct csio_wrm *wrm = csio_hw_to_wrm(hw);
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+ u32 param[6];
+ int i, j = 0;
+
+ /* Initialize portids to -1 */
+ for (i = 0; i < CSIO_MAX_PPORTS; i++)
+ hw->pport[i].portid = -1;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ /* Get port vec information. */
+ param[0] = FW_PARAM_DEV(PORTVEC);
+
+ /* Get Core clock. */
+ param[1] = FW_PARAM_DEV(CCLK);
+
+ /* Get EQ id start and end. */
+ param[2] = FW_PARAM_PFVF(EQ_START);
+ param[3] = FW_PARAM_PFVF(EQ_END);
+
+ /* Get IQ id start and end. */
+ param[4] = FW_PARAM_PFVF(IQFLINT_START);
+ param[5] = FW_PARAM_PFVF(IQFLINT_END);
+
+ csio_mb_params(hw, mbp, CSIO_MB_DEFAULT_TMO, hw->pfn, 0,
+ ARRAY_SIZE(param), param, NULL, false, NULL);
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_PARAMS_CMD(read) failed!\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ csio_mb_process_read_params_rsp(hw, mbp, &retval,
+ ARRAY_SIZE(param), param);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_PARAMS_CMD(read) failed with ret:0x%x!\n",
+ retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ /* cache the information. */
+ hw->port_vec = param[0];
+ hw->vpd.cclk = param[1];
+ wrm->fw_eq_start = param[2];
+ wrm->fw_iq_start = param[4];
+
+ /* Using FW configured max iqs & eqs */
+ if ((hw->flags & CSIO_HWF_USING_SOFT_PARAMS) ||
+ !csio_is_hw_master(hw)) {
+ hw->cfg_niq = param[5] - param[4] + 1;
+ hw->cfg_neq = param[3] - param[2] + 1;
+ csio_dbg(hw, "Using fwconfig max niqs %d neqs %d\n",
+ hw->cfg_niq, hw->cfg_neq);
+ }
+
+ hw->port_vec &= csio_port_mask;
+
+ hw->num_pports = hweight32(hw->port_vec);
+
+ csio_dbg(hw, "Port vector: 0x%x, #ports: %d\n",
+ hw->port_vec, hw->num_pports);
+
+ for (i = 0; i < hw->num_pports; i++) {
+ while ((hw->port_vec & (1 << j)) == 0)
+ j++;
+ hw->pport[i].portid = j++;
+ csio_dbg(hw, "Found Port:%d\n", hw->pport[i].portid);
+ }
+ mempool_free(mbp, hw->mb_mempool);
+
+ return 0;
+}
+
+
+/*
+ * csio_config_device_caps - Get and set device capabilities.
+ * @hw: HW module
+ *
+ */
+static int
+csio_config_device_caps(struct csio_hw *hw)
+{
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+ int rv = -EINVAL;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ /* Get device capabilities */
+ csio_mb_caps_config(hw, mbp, CSIO_MB_DEFAULT_TMO, 0, 0, 0, 0, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_CAPS_CONFIG_CMD(r) failed!\n");
+ goto out;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_CAPS_CONFIG_CMD(r) returned %d!\n", retval);
+ goto out;
+ }
+
+ /* Validate device capabilities */
+ if (csio_hw_validate_caps(hw, mbp))
+ goto out;
+
+ /* Don't config device capabilities if already configured */
+ if (hw->fw_state == CSIO_DEV_STATE_INIT) {
+ rv = 0;
+ goto out;
+ }
+
+ /* Write back desired device capabilities */
+ csio_mb_caps_config(hw, mbp, CSIO_MB_DEFAULT_TMO, true, true,
+ false, true, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_CAPS_CONFIG_CMD(w) failed!\n");
+ goto out;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_CAPS_CONFIG_CMD(w) returned %d!\n", retval);
+ goto out;
+ }
+
+ rv = 0;
+out:
+ mempool_free(mbp, hw->mb_mempool);
+ return rv;
+}
+
+static int
+csio_config_global_rss(struct csio_hw *hw)
+{
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ csio_rss_glb_config(hw, mbp, CSIO_MB_DEFAULT_TMO,
+ FW_RSS_GLB_CONFIG_CMD_MODE_BASICVIRTUAL,
+ FW_RSS_GLB_CONFIG_CMD_TNLMAPEN |
+ FW_RSS_GLB_CONFIG_CMD_HASHTOEPLITZ |
+ FW_RSS_GLB_CONFIG_CMD_TNLALLLKP,
+ NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_RSS_GLB_CONFIG_CMD failed!\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_RSS_GLB_CONFIG_CMD returned 0x%x!\n", retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ mempool_free(mbp, hw->mb_mempool);
+
+ return 0;
+}
+
+/*
+ * csio_config_pfvf - Configure Physical/Virtual functions settings.
+ * @hw: HW module
+ *
+ */
+static int
+csio_config_pfvf(struct csio_hw *hw)
+{
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ /*
+ * For now, allow all PFs to access to all ports using a pmask
+ * value of 0xF (M_FW_PFVF_CMD_PMASK). Once we have VFs, we will
+ * need to provide access based on some rule.
+ */
+ csio_mb_pfvf(hw, mbp, CSIO_MB_DEFAULT_TMO, hw->pfn, 0, CSIO_NEQ,
+ CSIO_NETH_CTRL, CSIO_NIQ_FLINT, 0, 0, CSIO_NVI, CSIO_CMASK,
+ CSIO_PMASK, CSIO_NEXACTF, CSIO_R_CAPS, CSIO_WX_CAPS, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_PFVF_CMD failed!\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_PFVF_CMD returned 0x%x!\n", retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ mempool_free(mbp, hw->mb_mempool);
+
+ return 0;
+}
+
+/*
+ * csio_enable_ports - Bring up all available ports.
+ * @hw: HW module.
+ *
+ */
+static int
+csio_enable_ports(struct csio_hw *hw)
+{
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+ uint8_t portid;
+ int i;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ for (i = 0; i < hw->num_pports; i++) {
+ portid = hw->pport[i].portid;
+
+ /* Read PORT information */
+ csio_mb_port(hw, mbp, CSIO_MB_DEFAULT_TMO, portid,
+ false, 0, 0, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "failed to issue FW_PORT_CMD(r) port:%d\n",
+ portid);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ csio_mb_process_read_port_rsp(hw, mbp, &retval,
+ &hw->pport[i].pcap);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_PORT_CMD(r) port:%d failed: 0x%x\n",
+ portid, retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ /* Write back PORT information */
+ csio_mb_port(hw, mbp, CSIO_MB_DEFAULT_TMO, portid, true,
+ (PAUSE_RX | PAUSE_TX), hw->pport[i].pcap, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "failed to issue FW_PORT_CMD(w) port:%d\n",
+ portid);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_PORT_CMD(w) port:%d failed :0x%x\n",
+ portid, retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ } /* For all ports */
+
+ mempool_free(mbp, hw->mb_mempool);
+
+ return 0;
+}
+
+/*
+ * csio_get_fcoe_resinfo - Read fcoe fw resource info.
+ * @hw: HW module
+ * Issued with lock held.
+ */
+static int
+csio_get_fcoe_resinfo(struct csio_hw *hw)
+{
+ struct csio_fcoe_res_info *res_info = &hw->fres_info;
+ struct fw_fcoe_res_info_cmd *rsp;
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ /* Get FCoE FW resource information */
+ csio_fcoe_read_res_info_init_mb(hw, mbp, CSIO_MB_DEFAULT_TMO, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "failed to issue FW_FCOE_RES_INFO_CMD\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ rsp = (struct fw_fcoe_res_info_cmd *)(mbp->mb);
+ retval = FW_CMD_RETVAL_GET(ntohl(rsp->retval_len16));
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_FCOE_RES_INFO_CMD failed with ret x%x\n",
+ retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ res_info->e_d_tov = ntohs(rsp->e_d_tov);
+ res_info->r_a_tov_seq = ntohs(rsp->r_a_tov_seq);
+ res_info->r_a_tov_els = ntohs(rsp->r_a_tov_els);
+ res_info->r_r_tov = ntohs(rsp->r_r_tov);
+ res_info->max_xchgs = ntohl(rsp->max_xchgs);
+ res_info->max_ssns = ntohl(rsp->max_ssns);
+ res_info->used_xchgs = ntohl(rsp->used_xchgs);
+ res_info->used_ssns = ntohl(rsp->used_ssns);
+ res_info->max_fcfs = ntohl(rsp->max_fcfs);
+ res_info->max_vnps = ntohl(rsp->max_vnps);
+ res_info->used_fcfs = ntohl(rsp->used_fcfs);
+ res_info->used_vnps = ntohl(rsp->used_vnps);
+
+ csio_dbg(hw, "max ssns:%d max xchgs:%d\n", res_info->max_ssns,
+ res_info->max_xchgs);
+ mempool_free(mbp, hw->mb_mempool);
+
+ return 0;
+}
+
+static int
+csio_hw_check_fwconfig(struct csio_hw *hw, u32 *param)
+{
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+ u32 _param[1];
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ /*
+ * Find out whether we're dealing with a version of
+ * the firmware which has configuration file support.
+ */
+ _param[0] = (FW_PARAMS_MNEM(FW_PARAMS_MNEM_DEV) |
+ FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_DEV_CF));
+
+ csio_mb_params(hw, mbp, CSIO_MB_DEFAULT_TMO, hw->pfn, 0,
+ ARRAY_SIZE(_param), _param, NULL, false, NULL);
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_PARAMS_CMD(read) failed!\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ csio_mb_process_read_params_rsp(hw, mbp, &retval,
+ ARRAY_SIZE(_param), _param);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_PARAMS_CMD(read) failed with ret:0x%x!\n",
+ retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ mempool_free(mbp, hw->mb_mempool);
+ *param = _param[0];
+
+ return 0;
+}
+
+static int
+csio_hw_flash_config(struct csio_hw *hw, u32 *fw_cfg_param, char *path)
+{
+ int ret = 0;
+ const struct firmware *cf;
+ struct pci_dev *pci_dev = hw->pdev;
+ struct device *dev = &pci_dev->dev;
+
+ unsigned int mtype = 0, maddr = 0;
+ uint32_t *cfg_data;
+ int value_to_add = 0;
+
+ if (request_firmware(&cf, CSIO_CF_FNAME, dev) < 0) {
+ csio_err(hw, "could not find config file " CSIO_CF_FNAME
+ ",err: %d\n", ret);
+ return -ENOENT;
+ }
+
+ if (cf->size%4 != 0)
+ value_to_add = 4 - (cf->size % 4);
+
+ cfg_data = kzalloc(cf->size+value_to_add, GFP_KERNEL);
+ if (cfg_data == NULL)
+ return -ENOMEM;
+
+ memcpy((void *)cfg_data, (const void *)cf->data, cf->size);
+
+ if (csio_hw_check_fwconfig(hw, fw_cfg_param) != 0)
+ return -EINVAL;
+
+ mtype = FW_PARAMS_PARAM_Y_GET(*fw_cfg_param);
+ maddr = FW_PARAMS_PARAM_Z_GET(*fw_cfg_param) << 16;
+
+ ret = csio_memory_write(hw, mtype, maddr,
+ cf->size + value_to_add, cfg_data);
+ if (ret == 0) {
+ csio_info(hw, "config file upgraded to " CSIO_CF_FNAME "\n");
+ strncpy(path, "/lib/firmware/" CSIO_CF_FNAME, 64);
+ }
+
+ kfree(cfg_data);
+ release_firmware(cf);
+
+ return ret;
+}
+
+/*
+ * HW initialization: contact FW, obtain config, perform basic init.
+ *
+ * If the firmware we're dealing with has Configuration File support, then
+ * we use that to perform all configuration -- either using the configuration
+ * file stored in flash on the adapter or using a filesystem-local file
+ * if available.
+ *
+ * If we don't have configuration file support in the firmware, then we'll
+ * have to set things up the old fashioned way with hard-coded register
+ * writes and firmware commands ...
+ */
+
+/*
+ * Attempt to initialize the HW via a Firmware Configuration File.
+ */
+static int
+csio_hw_use_fwconfig(struct csio_hw *hw, int reset, u32 *fw_cfg_param)
+{
+ unsigned int mtype, maddr;
+ int rv;
+ uint32_t finiver, finicsum, cfcsum;
+ int using_flash;
+ char path[64];
+
+ /*
+ * Reset device if necessary
+ */
+ if (reset) {
+ rv = csio_do_reset(hw, true);
+ if (rv != 0)
+ goto bye;
+ }
+
+ /*
+ * If we have a configuration file in host ,
+ * then use that. Otherwise, use the configuration file stored
+ * in the HW flash ...
+ */
+ spin_unlock_irq(&hw->lock);
+ rv = csio_hw_flash_config(hw, fw_cfg_param, path);
+ spin_lock_irq(&hw->lock);
+ if (rv != 0) {
+ if (rv == -ENOENT) {
+ /*
+ * config file was not found. Use default
+ * config file from flash.
+ */
+ mtype = FW_MEMTYPE_CF_FLASH;
+ maddr = csio_hw_flash_cfg_addr(hw);
+ using_flash = 1;
+ } else {
+ /*
+ * we revert back to the hardwired config if
+ * flashing failed.
+ */
+ goto bye;
+ }
+ } else {
+ mtype = FW_PARAMS_PARAM_Y_GET(*fw_cfg_param);
+ maddr = FW_PARAMS_PARAM_Z_GET(*fw_cfg_param) << 16;
+ using_flash = 0;
+ }
+
+ hw->cfg_store = (uint8_t)mtype;
+
+ /*
+ * Issue a Capability Configuration command to the firmware to get it
+ * to parse the Configuration File.
+ */
+ rv = csio_hw_fw_config_file(hw, mtype, maddr, &finiver,
+ &finicsum, &cfcsum);
+ if (rv != 0)
+ goto bye;
+
+ hw->cfg_finiver = finiver;
+ hw->cfg_finicsum = finicsum;
+ hw->cfg_cfcsum = cfcsum;
+ hw->cfg_csum_status = true;
+
+ if (finicsum != cfcsum) {
+ csio_warn(hw,
+ "Config File checksum mismatch: csum=%#x, computed=%#x\n",
+ finicsum, cfcsum);
+
+ hw->cfg_csum_status = false;
+ }
+
+ /*
+ * Note that we're operating with parameters
+ * not supplied by the driver, rather than from hard-wired
+ * initialization constants buried in the driver.
+ */
+ hw->flags |= CSIO_HWF_USING_SOFT_PARAMS;
+
+ /* device parameters */
+ rv = csio_get_device_params(hw);
+ if (rv != 0)
+ goto bye;
+
+ /* Configure SGE */
+ csio_wr_sge_init(hw);
+
+ /*
+ * And finally tell the firmware to initialize itself using the
+ * parameters from the Configuration File.
+ */
+ /* Post event to notify completion of configuration */
+ csio_post_event(&hw->sm, CSIO_HWE_INIT);
+
+ csio_info(hw,
+ "Firmware Configuration File %s, version %#x, computed checksum %#x\n",
+ (using_flash ? "in device FLASH" : path), finiver, cfcsum);
+
+ return 0;
+
+ /*
+ * Something bad happened. Return the error ...
+ */
+bye:
+ hw->flags &= ~CSIO_HWF_USING_SOFT_PARAMS;
+ csio_dbg(hw, "Configuration file error %d\n", rv);
+ return rv;
+}
+
+/*
+ * Attempt to initialize the adapter via hard-coded, driver supplied
+ * parameters ...
+ */
+static int
+csio_hw_no_fwconfig(struct csio_hw *hw, int reset)
+{
+ int rv;
+ /*
+ * Reset device if necessary
+ */
+ if (reset) {
+ rv = csio_do_reset(hw, true);
+ if (rv != 0)
+ goto out;
+ }
+
+ /* Get and set device capabilities */
+ rv = csio_config_device_caps(hw);
+ if (rv != 0)
+ goto out;
+
+ /* Config Global RSS command */
+ rv = csio_config_global_rss(hw);
+ if (rv != 0)
+ goto out;
+
+ /* Configure PF/VF capabilities of device */
+ rv = csio_config_pfvf(hw);
+ if (rv != 0)
+ goto out;
+
+ /* device parameters */
+ rv = csio_get_device_params(hw);
+ if (rv != 0)
+ goto out;
+
+ /* Configure SGE */
+ csio_wr_sge_init(hw);
+
+ /* Post event to notify completion of configuration */
+ csio_post_event(&hw->sm, CSIO_HWE_INIT);
+
+out:
+ return rv;
+}
+
+/*
+ * Returns -EINVAL if attempts to flash the firmware failed
+ * else returns 0,
+ * if flashing was not attempted because the card had the
+ * latest firmware ECANCELED is returned
+ */
+static int
+csio_hw_flash_fw(struct csio_hw *hw)
+{
+ int ret = -ECANCELED;
+ const struct firmware *fw;
+ const struct fw_hdr *hdr;
+ u32 fw_ver;
+ struct pci_dev *pci_dev = hw->pdev;
+ struct device *dev = &pci_dev->dev ;
+
+ if (request_firmware(&fw, CSIO_FW_FNAME, dev) < 0) {
+ csio_err(hw, "could not find firmware image " CSIO_FW_FNAME
+ ",err: %d\n", ret);
+ return -EINVAL;
+ }
+
+ hdr = (const struct fw_hdr *)fw->data;
+ fw_ver = ntohl(hdr->fw_ver);
+ if (FW_HDR_FW_VER_MAJOR_GET(fw_ver) != FW_VERSION_MAJOR)
+ return -EINVAL; /* wrong major version, won't do */
+
+ /*
+ * If the flash FW is unusable or we found something newer, load it.
+ */
+ if (FW_HDR_FW_VER_MAJOR_GET(hw->fwrev) != FW_VERSION_MAJOR ||
+ fw_ver > hw->fwrev) {
+ ret = csio_hw_fw_upgrade(hw, hw->pfn, fw->data, fw->size,
+ /*force=*/false);
+ if (!ret)
+ csio_info(hw, "firmware upgraded to version %pI4 from "
+ CSIO_FW_FNAME "\n", &hdr->fw_ver);
+ else
+ csio_err(hw, "firmware upgrade failed! err=%d\n", ret);
+ }
+
+ release_firmware(fw);
+
+ return ret;
+}
+
+
+/*
+ * csio_hw_configure - Configure HW
+ * @hw - HW module
+ *
+ */
+static void
+csio_hw_configure(struct csio_hw *hw)
+{
+ int reset = 1;
+ int rv;
+ u32 param[1];
+
+ rv = csio_hw_dev_ready(hw);
+ if (rv != 0) {
+ CSIO_INC_STATS(hw, n_err_fatal);
+ csio_post_event(&hw->sm, CSIO_HWE_FATAL);
+ goto out;
+ }
+
+ /* HW version */
+ hw->chip_ver = (char)csio_rd_reg32(hw, PL_REV);
+
+ /* Needed for FW download */
+ rv = csio_hw_get_flash_params(hw);
+ if (rv != 0) {
+ csio_err(hw, "Failed to get serial flash params rv:%d\n", rv);
+ csio_post_event(&hw->sm, CSIO_HWE_FATAL);
+ goto out;
+ }
+
+ /* Set pci completion timeout value to 4 seconds. */
+ csio_set_pcie_completion_timeout(hw, 0xd);
+
+ csio_hw_set_mem_win(hw);
+
+ rv = csio_hw_get_fw_version(hw, &hw->fwrev);
+ if (rv != 0)
+ goto out;
+
+ csio_hw_print_fw_version(hw, "Firmware revision");
+
+ rv = csio_do_hello(hw, &hw->fw_state);
+ if (rv != 0) {
+ CSIO_INC_STATS(hw, n_err_fatal);
+ csio_post_event(&hw->sm, CSIO_HWE_FATAL);
+ goto out;
+ }
+
+ /* Read vpd */
+ rv = csio_hw_get_vpd_params(hw, &hw->vpd);
+ if (rv != 0)
+ goto out;
+
+ if (csio_is_hw_master(hw) && hw->fw_state != CSIO_DEV_STATE_INIT) {
+ rv = csio_hw_check_fw_version(hw);
+ if (rv == -EINVAL) {
+
+ /* Do firmware update */
+ spin_unlock_irq(&hw->lock);
+ rv = csio_hw_flash_fw(hw);
+ spin_lock_irq(&hw->lock);
+
+ if (rv == 0) {
+ reset = 0;
+ /*
+ * Note that the chip was reset as part of the
+ * firmware upgrade so we don't reset it again
+ * below and grab the new firmware version.
+ */
+ rv = csio_hw_check_fw_version(hw);
+ }
+ }
+ /*
+ * If the firmware doesn't support Configuration
+ * Files, use the old Driver-based, hard-wired
+ * initialization. Otherwise, try using the
+ * Configuration File support and fall back to the
+ * Driver-based initialization if there's no
+ * Configuration File found.
+ */
+ if (csio_hw_check_fwconfig(hw, param) == 0) {
+ rv = csio_hw_use_fwconfig(hw, reset, param);
+ if (rv == -ENOENT)
+ goto out;
+ if (rv != 0) {
+ csio_info(hw,
+ "No Configuration File present "
+ "on adapter. Using hard-wired "
+ "configuration parameters.\n");
+ rv = csio_hw_no_fwconfig(hw, reset);
+ }
+ } else {
+ rv = csio_hw_no_fwconfig(hw, reset);
+ }
+
+ if (rv != 0)
+ goto out;
+
+ } else {
+ if (hw->fw_state == CSIO_DEV_STATE_INIT) {
+
+ /* device parameters */
+ rv = csio_get_device_params(hw);
+ if (rv != 0)
+ goto out;
+
+ /* Get device capabilities */
+ rv = csio_config_device_caps(hw);
+ if (rv != 0)
+ goto out;
+
+ /* Configure SGE */
+ csio_wr_sge_init(hw);
+
+ /* Post event to notify completion of configuration */
+ csio_post_event(&hw->sm, CSIO_HWE_INIT);
+ goto out;
+ }
+ } /* if not master */
+
+out:
+ return;
+}
+
+/*
+ * csio_hw_initialize - Initialize HW
+ * @hw - HW module
+ *
+ */
+static void
+csio_hw_initialize(struct csio_hw *hw)
+{
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+ int rv;
+ int i;
+
+ if (csio_is_hw_master(hw) && hw->fw_state != CSIO_DEV_STATE_INIT) {
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp)
+ goto out;
+
+ csio_mb_initialize(hw, mbp, CSIO_MB_DEFAULT_TMO, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FW_INITIALIZE_CMD failed!\n");
+ goto free_and_out;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FW_INITIALIZE_CMD returned 0x%x!\n",
+ retval);
+ goto free_and_out;
+ }
+
+ mempool_free(mbp, hw->mb_mempool);
+ }
+
+ rv = csio_get_fcoe_resinfo(hw);
+ if (rv != 0) {
+ csio_err(hw, "Failed to read fcoe resource info: %d\n", rv);
+ goto out;
+ }
+
+ spin_unlock_irq(&hw->lock);
+ rv = csio_config_queues(hw);
+ spin_lock_irq(&hw->lock);
+
+ if (rv != 0) {
+ csio_err(hw, "Config of queues failed!: %d\n", rv);
+ goto out;
+ }
+
+ for (i = 0; i < hw->num_pports; i++)
+ hw->pport[i].mod_type = FW_PORT_MOD_TYPE_NA;
+
+ if (csio_is_hw_master(hw) && hw->fw_state != CSIO_DEV_STATE_INIT) {
+ rv = csio_enable_ports(hw);
+ if (rv != 0) {
+ csio_err(hw, "Failed to enable ports: %d\n", rv);
+ goto out;
+ }
+ }
+
+ csio_post_event(&hw->sm, CSIO_HWE_INIT_DONE);
+ return;
+
+free_and_out:
+ mempool_free(mbp, hw->mb_mempool);
+out:
+ return;
+}
+
+#define PF_INTR_MASK (PFSW | PFCIM)
+
+/*
+ * csio_hw_intr_enable - Enable HW interrupts
+ * @hw: Pointer to HW module.
+ *
+ * Enable interrupts in HW registers.
+ */
+static void
+csio_hw_intr_enable(struct csio_hw *hw)
+{
+ uint16_t vec = (uint16_t)csio_get_mb_intr_idx(csio_hw_to_mbm(hw));
+ uint32_t pf = SOURCEPF_GET(csio_rd_reg32(hw, PL_WHOAMI));
+ uint32_t pl = csio_rd_reg32(hw, PL_INT_ENABLE);
+
+ /*
+ * Set aivec for MSI/MSIX. PCIE_PF_CFG.INTXType is set up
+ * by FW, so do nothing for INTX.
+ */
+ if (hw->intr_mode == CSIO_IM_MSIX)
+ csio_set_reg_field(hw, MYPF_REG(PCIE_PF_CFG),
+ AIVEC(AIVEC_MASK), vec);
+ else if (hw->intr_mode == CSIO_IM_MSI)
+ csio_set_reg_field(hw, MYPF_REG(PCIE_PF_CFG),
+ AIVEC(AIVEC_MASK), 0);
+
+ csio_wr_reg32(hw, PF_INTR_MASK, MYPF_REG(PL_PF_INT_ENABLE));
+
+ /* Turn on MB interrupts - this will internally flush PIO as well */
+ csio_mb_intr_enable(hw);
+
+ /* These are common registers - only a master can modify them */
+ if (csio_is_hw_master(hw)) {
+ /*
+ * Disable the Serial FLASH interrupt, if enabled!
+ */
+ pl &= (~SF);
+ csio_wr_reg32(hw, pl, PL_INT_ENABLE);
+
+ csio_wr_reg32(hw, ERR_CPL_EXCEED_IQE_SIZE |
+ EGRESS_SIZE_ERR | ERR_INVALID_CIDX_INC |
+ ERR_CPL_OPCODE_0 | ERR_DROPPED_DB |
+ ERR_DATA_CPL_ON_HIGH_QID1 |
+ ERR_DATA_CPL_ON_HIGH_QID0 | ERR_BAD_DB_PIDX3 |
+ ERR_BAD_DB_PIDX2 | ERR_BAD_DB_PIDX1 |
+ ERR_BAD_DB_PIDX0 | ERR_ING_CTXT_PRIO |
+ ERR_EGR_CTXT_PRIO | INGRESS_SIZE_ERR,
+ SGE_INT_ENABLE3);
+ csio_set_reg_field(hw, PL_INT_MAP0, 0, 1 << pf);
+ }
+
+ hw->flags |= CSIO_HWF_HW_INTR_ENABLED;
+
+}
+
+/*
+ * csio_hw_intr_disable - Disable HW interrupts
+ * @hw: Pointer to HW module.
+ *
+ * Turn off Mailbox and PCI_PF_CFG interrupts.
+ */
+void
+csio_hw_intr_disable(struct csio_hw *hw)
+{
+ uint32_t pf = SOURCEPF_GET(csio_rd_reg32(hw, PL_WHOAMI));
+
+ if (!(hw->flags & CSIO_HWF_HW_INTR_ENABLED))
+ return;
+
+ hw->flags &= ~CSIO_HWF_HW_INTR_ENABLED;
+
+ csio_wr_reg32(hw, 0, MYPF_REG(PL_PF_INT_ENABLE));
+ if (csio_is_hw_master(hw))
+ csio_set_reg_field(hw, PL_INT_MAP0, 1 << pf, 0);
+
+ /* Turn off MB interrupts */
+ csio_mb_intr_disable(hw);
+
+}
+
+static void
+csio_hw_fatal_err(struct csio_hw *hw)
+{
+ csio_set_reg_field(hw, SGE_CONTROL, GLOBALENABLE, 0);
+ csio_hw_intr_disable(hw);
+
+ /* Do not reset HW, we may need FW state for debugging */
+ csio_fatal(hw, "HW Fatal error encountered!\n");
+}
+
+/*****************************************************************************/
+/* START: HW SM */
+/*****************************************************************************/
+/*
+ * csio_hws_uninit - Uninit state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_uninit(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_CFG:
+ csio_set_state(&hw->sm, csio_hws_configuring);
+ csio_hw_configure(hw);
+ break;
+
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_hws_configuring - Configuring state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_configuring(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_INIT:
+ csio_set_state(&hw->sm, csio_hws_initializing);
+ csio_hw_initialize(hw);
+ break;
+
+ case CSIO_HWE_INIT_DONE:
+ csio_set_state(&hw->sm, csio_hws_ready);
+ /* Fan out event to all lnode SMs */
+ csio_notify_lnodes(hw, CSIO_LN_NOTIFY_HWREADY);
+ break;
+
+ case CSIO_HWE_FATAL:
+ csio_set_state(&hw->sm, csio_hws_uninit);
+ break;
+
+ case CSIO_HWE_PCI_REMOVE:
+ csio_do_bye(hw);
+ break;
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_hws_initializing - Initialiazing state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_initializing(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_INIT_DONE:
+ csio_set_state(&hw->sm, csio_hws_ready);
+
+ /* Fan out event to all lnode SMs */
+ csio_notify_lnodes(hw, CSIO_LN_NOTIFY_HWREADY);
+
+ /* Enable interrupts */
+ csio_hw_intr_enable(hw);
+ break;
+
+ case CSIO_HWE_FATAL:
+ csio_set_state(&hw->sm, csio_hws_uninit);
+ break;
+
+ case CSIO_HWE_PCI_REMOVE:
+ csio_do_bye(hw);
+ break;
+
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_hws_ready - Ready state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_ready(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ /* Remember the event */
+ hw->evtflag = evt;
+
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_HBA_RESET:
+ case CSIO_HWE_FW_DLOAD:
+ case CSIO_HWE_SUSPEND:
+ case CSIO_HWE_PCI_REMOVE:
+ case CSIO_HWE_PCIERR_DETECTED:
+ csio_set_state(&hw->sm, csio_hws_quiescing);
+ /* cleanup all outstanding cmds */
+ if (evt == CSIO_HWE_HBA_RESET ||
+ evt == CSIO_HWE_PCIERR_DETECTED)
+ csio_scsim_cleanup_io(csio_hw_to_scsim(hw), false);
+ else
+ csio_scsim_cleanup_io(csio_hw_to_scsim(hw), true);
+
+ csio_hw_intr_disable(hw);
+ csio_hw_mbm_cleanup(hw);
+ csio_evtq_stop(hw);
+ csio_notify_lnodes(hw, CSIO_LN_NOTIFY_HWSTOP);
+ csio_evtq_flush(hw);
+ csio_mgmtm_cleanup(csio_hw_to_mgmtm(hw));
+ csio_post_event(&hw->sm, CSIO_HWE_QUIESCED);
+ break;
+
+ case CSIO_HWE_FATAL:
+ csio_set_state(&hw->sm, csio_hws_uninit);
+ break;
+
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_hws_quiescing - Quiescing state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_quiescing(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_QUIESCED:
+ switch (hw->evtflag) {
+ case CSIO_HWE_FW_DLOAD:
+ csio_set_state(&hw->sm, csio_hws_resetting);
+ /* Download firmware */
+ /* Fall through */
+
+ case CSIO_HWE_HBA_RESET:
+ csio_set_state(&hw->sm, csio_hws_resetting);
+ /* Start reset of the HBA */
+ csio_notify_lnodes(hw, CSIO_LN_NOTIFY_HWRESET);
+ csio_wr_destroy_queues(hw, false);
+ csio_do_reset(hw, false);
+ csio_post_event(&hw->sm, CSIO_HWE_HBA_RESET_DONE);
+ break;
+
+ case CSIO_HWE_PCI_REMOVE:
+ csio_set_state(&hw->sm, csio_hws_removing);
+ csio_notify_lnodes(hw, CSIO_LN_NOTIFY_HWREMOVE);
+ csio_wr_destroy_queues(hw, true);
+ /* Now send the bye command */
+ csio_do_bye(hw);
+ break;
+
+ case CSIO_HWE_SUSPEND:
+ csio_set_state(&hw->sm, csio_hws_quiesced);
+ break;
+
+ case CSIO_HWE_PCIERR_DETECTED:
+ csio_set_state(&hw->sm, csio_hws_pcierr);
+ csio_wr_destroy_queues(hw, false);
+ break;
+
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+
+ }
+ break;
+
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_hws_quiesced - Quiesced state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_quiesced(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_RESUME:
+ csio_set_state(&hw->sm, csio_hws_configuring);
+ csio_hw_configure(hw);
+ break;
+
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_hws_resetting - HW Resetting state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_resetting(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_HBA_RESET_DONE:
+ csio_evtq_start(hw);
+ csio_set_state(&hw->sm, csio_hws_configuring);
+ csio_hw_configure(hw);
+ break;
+
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_hws_removing - PCI Hotplug removing state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_removing(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_HBA_RESET:
+ if (!csio_is_hw_master(hw))
+ break;
+ /*
+ * The BYE should have alerady been issued, so we cant
+ * use the mailbox interface. Hence we use the PL_RST
+ * register directly.
+ */
+ csio_err(hw, "Resetting HW and waiting 2 seconds...\n");
+ csio_wr_reg32(hw, PIORSTMODE | PIORST, PL_RST);
+ mdelay(2000);
+ break;
+
+ /* Should never receive any new events */
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+
+ }
+}
+
+/*
+ * csio_hws_pcierr - PCI Error state
+ * @hw - HW module
+ * @evt - Event
+ *
+ */
+static void
+csio_hws_pcierr(struct csio_hw *hw, enum csio_hw_ev evt)
+{
+ hw->prev_evt = hw->cur_evt;
+ hw->cur_evt = evt;
+ CSIO_INC_STATS(hw, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_HWE_PCIERR_SLOT_RESET:
+ csio_evtq_start(hw);
+ csio_set_state(&hw->sm, csio_hws_configuring);
+ csio_hw_configure(hw);
+ break;
+
+ default:
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+}
+
+/*****************************************************************************/
+/* END: HW SM */
+/*****************************************************************************/
+
+/* Slow path handlers */
+struct intr_info {
+ unsigned int mask; /* bits to check in interrupt status */
+ const char *msg; /* message to print or NULL */
+ short stat_idx; /* stat counter to increment or -1 */
+ unsigned short fatal; /* whether the condition reported is fatal */
+};
+
+/*
+ * csio_handle_intr_status - table driven interrupt handler
+ * @hw: HW instance
+ * @reg: the interrupt status register to process
+ * @acts: table of interrupt actions
+ *
+ * A table driven interrupt handler that applies a set of masks to an
+ * interrupt status word and performs the corresponding actions if the
+ * interrupts described by the mask have occured. The actions include
+ * optionally emitting a warning or alert message. The table is terminated
+ * by an entry specifying mask 0. Returns the number of fatal interrupt
+ * conditions.
+ */
+static int
+csio_handle_intr_status(struct csio_hw *hw, unsigned int reg,
+ const struct intr_info *acts)
+{
+ int fatal = 0;
+ unsigned int mask = 0;
+ unsigned int status = csio_rd_reg32(hw, reg);
+
+ for ( ; acts->mask; ++acts) {
+ if (!(status & acts->mask))
+ continue;
+ if (acts->fatal) {
+ fatal++;
+ csio_fatal(hw, "Fatal %s (0x%x)\n",
+ acts->msg, status & acts->mask);
+ } else if (acts->msg)
+ csio_info(hw, "%s (0x%x)\n",
+ acts->msg, status & acts->mask);
+ mask |= acts->mask;
+ }
+ status &= mask;
+ if (status) /* clear processed interrupts */
+ csio_wr_reg32(hw, status, reg);
+ return fatal;
+}
+
+/*
+ * Interrupt handler for the PCIE module.
+ */
+static void
+csio_pcie_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info sysbus_intr_info[] = {
+ { RNPP, "RXNP array parity error", -1, 1 },
+ { RPCP, "RXPC array parity error", -1, 1 },
+ { RCIP, "RXCIF array parity error", -1, 1 },
+ { RCCP, "Rx completions control array parity error", -1, 1 },
+ { RFTP, "RXFT array parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info pcie_port_intr_info[] = {
+ { TPCP, "TXPC array parity error", -1, 1 },
+ { TNPP, "TXNP array parity error", -1, 1 },
+ { TFTP, "TXFT array parity error", -1, 1 },
+ { TCAP, "TXCA array parity error", -1, 1 },
+ { TCIP, "TXCIF array parity error", -1, 1 },
+ { RCAP, "RXCA array parity error", -1, 1 },
+ { OTDD, "outbound request TLP discarded", -1, 1 },
+ { RDPE, "Rx data parity error", -1, 1 },
+ { TDUE, "Tx uncorrectable data error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info pcie_intr_info[] = {
+ { MSIADDRLPERR, "MSI AddrL parity error", -1, 1 },
+ { MSIADDRHPERR, "MSI AddrH parity error", -1, 1 },
+ { MSIDATAPERR, "MSI data parity error", -1, 1 },
+ { MSIXADDRLPERR, "MSI-X AddrL parity error", -1, 1 },
+ { MSIXADDRHPERR, "MSI-X AddrH parity error", -1, 1 },
+ { MSIXDATAPERR, "MSI-X data parity error", -1, 1 },
+ { MSIXDIPERR, "MSI-X DI parity error", -1, 1 },
+ { PIOCPLPERR, "PCI PIO completion FIFO parity error", -1, 1 },
+ { PIOREQPERR, "PCI PIO request FIFO parity error", -1, 1 },
+ { TARTAGPERR, "PCI PCI target tag FIFO parity error", -1, 1 },
+ { CCNTPERR, "PCI CMD channel count parity error", -1, 1 },
+ { CREQPERR, "PCI CMD channel request parity error", -1, 1 },
+ { CRSPPERR, "PCI CMD channel response parity error", -1, 1 },
+ { DCNTPERR, "PCI DMA channel count parity error", -1, 1 },
+ { DREQPERR, "PCI DMA channel request parity error", -1, 1 },
+ { DRSPPERR, "PCI DMA channel response parity error", -1, 1 },
+ { HCNTPERR, "PCI HMA channel count parity error", -1, 1 },
+ { HREQPERR, "PCI HMA channel request parity error", -1, 1 },
+ { HRSPPERR, "PCI HMA channel response parity error", -1, 1 },
+ { CFGSNPPERR, "PCI config snoop FIFO parity error", -1, 1 },
+ { FIDPERR, "PCI FID parity error", -1, 1 },
+ { INTXCLRPERR, "PCI INTx clear parity error", -1, 1 },
+ { MATAGPERR, "PCI MA tag parity error", -1, 1 },
+ { PIOTAGPERR, "PCI PIO tag parity error", -1, 1 },
+ { RXCPLPERR, "PCI Rx completion parity error", -1, 1 },
+ { RXWRPERR, "PCI Rx write parity error", -1, 1 },
+ { RPLPERR, "PCI replay buffer parity error", -1, 1 },
+ { PCIESINT, "PCI core secondary fault", -1, 1 },
+ { PCIEPINT, "PCI core primary fault", -1, 1 },
+ { UNXSPLCPLERR, "PCI unexpected split completion error", -1,
+ 0 },
+ { 0, NULL, 0, 0 }
+ };
+
+ int fat;
+
+ fat = csio_handle_intr_status(hw,
+ PCIE_CORE_UTL_SYSTEM_BUS_AGENT_STATUS,
+ sysbus_intr_info) +
+ csio_handle_intr_status(hw,
+ PCIE_CORE_UTL_PCI_EXPRESS_PORT_STATUS,
+ pcie_port_intr_info) +
+ csio_handle_intr_status(hw, PCIE_INT_CAUSE, pcie_intr_info);
+ if (fat)
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * TP interrupt handler.
+ */
+static void csio_tp_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info tp_intr_info[] = {
+ { 0x3fffffff, "TP parity error", -1, 1 },
+ { FLMTXFLSTEMPTY, "TP out of Tx pages", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, TP_INT_CAUSE, tp_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * SGE interrupt handler.
+ */
+static void csio_sge_intr_handler(struct csio_hw *hw)
+{
+ uint64_t v;
+
+ static struct intr_info sge_intr_info[] = {
+ { ERR_CPL_EXCEED_IQE_SIZE,
+ "SGE received CPL exceeding IQE size", -1, 1 },
+ { ERR_INVALID_CIDX_INC,
+ "SGE GTS CIDX increment too large", -1, 0 },
+ { ERR_CPL_OPCODE_0, "SGE received 0-length CPL", -1, 0 },
+ { ERR_DROPPED_DB, "SGE doorbell dropped", -1, 0 },
+ { ERR_DATA_CPL_ON_HIGH_QID1 | ERR_DATA_CPL_ON_HIGH_QID0,
+ "SGE IQID > 1023 received CPL for FL", -1, 0 },
+ { ERR_BAD_DB_PIDX3, "SGE DBP 3 pidx increment too large", -1,
+ 0 },
+ { ERR_BAD_DB_PIDX2, "SGE DBP 2 pidx increment too large", -1,
+ 0 },
+ { ERR_BAD_DB_PIDX1, "SGE DBP 1 pidx increment too large", -1,
+ 0 },
+ { ERR_BAD_DB_PIDX0, "SGE DBP 0 pidx increment too large", -1,
+ 0 },
+ { ERR_ING_CTXT_PRIO,
+ "SGE too many priority ingress contexts", -1, 0 },
+ { ERR_EGR_CTXT_PRIO,
+ "SGE too many priority egress contexts", -1, 0 },
+ { INGRESS_SIZE_ERR, "SGE illegal ingress QID", -1, 0 },
+ { EGRESS_SIZE_ERR, "SGE illegal egress QID", -1, 0 },
+ { 0, NULL, 0, 0 }
+ };
+
+ v = (uint64_t)csio_rd_reg32(hw, SGE_INT_CAUSE1) |
+ ((uint64_t)csio_rd_reg32(hw, SGE_INT_CAUSE2) << 32);
+ if (v) {
+ csio_fatal(hw, "SGE parity error (%#llx)\n",
+ (unsigned long long)v);
+ csio_wr_reg32(hw, (uint32_t)(v & 0xFFFFFFFF),
+ SGE_INT_CAUSE1);
+ csio_wr_reg32(hw, (uint32_t)(v >> 32), SGE_INT_CAUSE2);
+ }
+
+ v |= csio_handle_intr_status(hw, SGE_INT_CAUSE3, sge_intr_info);
+
+ if (csio_handle_intr_status(hw, SGE_INT_CAUSE3, sge_intr_info) ||
+ v != 0)
+ csio_hw_fatal_err(hw);
+}
+
+#define CIM_OBQ_INTR (OBQULP0PARERR | OBQULP1PARERR | OBQULP2PARERR |\
+ OBQULP3PARERR | OBQSGEPARERR | OBQNCSIPARERR)
+#define CIM_IBQ_INTR (IBQTP0PARERR | IBQTP1PARERR | IBQULPPARERR |\
+ IBQSGEHIPARERR | IBQSGELOPARERR | IBQNCSIPARERR)
+
+/*
+ * CIM interrupt handler.
+ */
+static void csio_cim_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info cim_intr_info[] = {
+ { PREFDROPINT, "CIM control register prefetch drop", -1, 1 },
+ { CIM_OBQ_INTR, "CIM OBQ parity error", -1, 1 },
+ { CIM_IBQ_INTR, "CIM IBQ parity error", -1, 1 },
+ { MBUPPARERR, "CIM mailbox uP parity error", -1, 1 },
+ { MBHOSTPARERR, "CIM mailbox host parity error", -1, 1 },
+ { TIEQINPARERRINT, "CIM TIEQ outgoing parity error", -1, 1 },
+ { TIEQOUTPARERRINT, "CIM TIEQ incoming parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info cim_upintr_info[] = {
+ { RSVDSPACEINT, "CIM reserved space access", -1, 1 },
+ { ILLTRANSINT, "CIM illegal transaction", -1, 1 },
+ { ILLWRINT, "CIM illegal write", -1, 1 },
+ { ILLRDINT, "CIM illegal read", -1, 1 },
+ { ILLRDBEINT, "CIM illegal read BE", -1, 1 },
+ { ILLWRBEINT, "CIM illegal write BE", -1, 1 },
+ { SGLRDBOOTINT, "CIM single read from boot space", -1, 1 },
+ { SGLWRBOOTINT, "CIM single write to boot space", -1, 1 },
+ { BLKWRBOOTINT, "CIM block write to boot space", -1, 1 },
+ { SGLRDFLASHINT, "CIM single read from flash space", -1, 1 },
+ { SGLWRFLASHINT, "CIM single write to flash space", -1, 1 },
+ { BLKWRFLASHINT, "CIM block write to flash space", -1, 1 },
+ { SGLRDEEPROMINT, "CIM single EEPROM read", -1, 1 },
+ { SGLWREEPROMINT, "CIM single EEPROM write", -1, 1 },
+ { BLKRDEEPROMINT, "CIM block EEPROM read", -1, 1 },
+ { BLKWREEPROMINT, "CIM block EEPROM write", -1, 1 },
+ { SGLRDCTLINT , "CIM single read from CTL space", -1, 1 },
+ { SGLWRCTLINT , "CIM single write to CTL space", -1, 1 },
+ { BLKRDCTLINT , "CIM block read from CTL space", -1, 1 },
+ { BLKWRCTLINT , "CIM block write to CTL space", -1, 1 },
+ { SGLRDPLINT , "CIM single read from PL space", -1, 1 },
+ { SGLWRPLINT , "CIM single write to PL space", -1, 1 },
+ { BLKRDPLINT , "CIM block read from PL space", -1, 1 },
+ { BLKWRPLINT , "CIM block write to PL space", -1, 1 },
+ { REQOVRLOOKUPINT , "CIM request FIFO overwrite", -1, 1 },
+ { RSPOVRLOOKUPINT , "CIM response FIFO overwrite", -1, 1 },
+ { TIMEOUTINT , "CIM PIF timeout", -1, 1 },
+ { TIMEOUTMAINT , "CIM PIF MA timeout", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ int fat;
+
+ fat = csio_handle_intr_status(hw, CIM_HOST_INT_CAUSE,
+ cim_intr_info) +
+ csio_handle_intr_status(hw, CIM_HOST_UPACC_INT_CAUSE,
+ cim_upintr_info);
+ if (fat)
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * ULP RX interrupt handler.
+ */
+static void csio_ulprx_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info ulprx_intr_info[] = {
+ { 0x1800000, "ULPRX context error", -1, 1 },
+ { 0x7fffff, "ULPRX parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, ULP_RX_INT_CAUSE, ulprx_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * ULP TX interrupt handler.
+ */
+static void csio_ulptx_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info ulptx_intr_info[] = {
+ { PBL_BOUND_ERR_CH3, "ULPTX channel 3 PBL out of bounds", -1,
+ 0 },
+ { PBL_BOUND_ERR_CH2, "ULPTX channel 2 PBL out of bounds", -1,
+ 0 },
+ { PBL_BOUND_ERR_CH1, "ULPTX channel 1 PBL out of bounds", -1,
+ 0 },
+ { PBL_BOUND_ERR_CH0, "ULPTX channel 0 PBL out of bounds", -1,
+ 0 },
+ { 0xfffffff, "ULPTX parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, ULP_TX_INT_CAUSE, ulptx_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * PM TX interrupt handler.
+ */
+static void csio_pmtx_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info pmtx_intr_info[] = {
+ { PCMD_LEN_OVFL0, "PMTX channel 0 pcmd too large", -1, 1 },
+ { PCMD_LEN_OVFL1, "PMTX channel 1 pcmd too large", -1, 1 },
+ { PCMD_LEN_OVFL2, "PMTX channel 2 pcmd too large", -1, 1 },
+ { ZERO_C_CMD_ERROR, "PMTX 0-length pcmd", -1, 1 },
+ { 0xffffff0, "PMTX framing error", -1, 1 },
+ { OESPI_PAR_ERROR, "PMTX oespi parity error", -1, 1 },
+ { DB_OPTIONS_PAR_ERROR, "PMTX db_options parity error", -1,
+ 1 },
+ { ICSPI_PAR_ERROR, "PMTX icspi parity error", -1, 1 },
+ { C_PCMD_PAR_ERROR, "PMTX c_pcmd parity error", -1, 1},
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, PM_TX_INT_CAUSE, pmtx_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * PM RX interrupt handler.
+ */
+static void csio_pmrx_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info pmrx_intr_info[] = {
+ { ZERO_E_CMD_ERROR, "PMRX 0-length pcmd", -1, 1 },
+ { 0x3ffff0, "PMRX framing error", -1, 1 },
+ { OCSPI_PAR_ERROR, "PMRX ocspi parity error", -1, 1 },
+ { DB_OPTIONS_PAR_ERROR, "PMRX db_options parity error", -1,
+ 1 },
+ { IESPI_PAR_ERROR, "PMRX iespi parity error", -1, 1 },
+ { E_PCMD_PAR_ERROR, "PMRX e_pcmd parity error", -1, 1},
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, PM_RX_INT_CAUSE, pmrx_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * CPL switch interrupt handler.
+ */
+static void csio_cplsw_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info cplsw_intr_info[] = {
+ { CIM_OP_MAP_PERR, "CPLSW CIM op_map parity error", -1, 1 },
+ { CIM_OVFL_ERROR, "CPLSW CIM overflow", -1, 1 },
+ { TP_FRAMING_ERROR, "CPLSW TP framing error", -1, 1 },
+ { SGE_FRAMING_ERROR, "CPLSW SGE framing error", -1, 1 },
+ { CIM_FRAMING_ERROR, "CPLSW CIM framing error", -1, 1 },
+ { ZERO_SWITCH_ERROR, "CPLSW no-switch error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, CPL_INTR_CAUSE, cplsw_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * LE interrupt handler.
+ */
+static void csio_le_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info le_intr_info[] = {
+ { LIPMISS, "LE LIP miss", -1, 0 },
+ { LIP0, "LE 0 LIP error", -1, 0 },
+ { PARITYERR, "LE parity error", -1, 1 },
+ { UNKNOWNCMD, "LE unknown command", -1, 1 },
+ { REQQPARERR, "LE request queue parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, LE_DB_INT_CAUSE, le_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * MPS interrupt handler.
+ */
+static void csio_mps_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info mps_rx_intr_info[] = {
+ { 0xffffff, "MPS Rx parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info mps_tx_intr_info[] = {
+ { TPFIFO, "MPS Tx TP FIFO parity error", -1, 1 },
+ { NCSIFIFO, "MPS Tx NC-SI FIFO parity error", -1, 1 },
+ { TXDATAFIFO, "MPS Tx data FIFO parity error", -1, 1 },
+ { TXDESCFIFO, "MPS Tx desc FIFO parity error", -1, 1 },
+ { BUBBLE, "MPS Tx underflow", -1, 1 },
+ { SECNTERR, "MPS Tx SOP/EOP error", -1, 1 },
+ { FRMERR, "MPS Tx framing error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info mps_trc_intr_info[] = {
+ { FILTMEM, "MPS TRC filter parity error", -1, 1 },
+ { PKTFIFO, "MPS TRC packet FIFO parity error", -1, 1 },
+ { MISCPERR, "MPS TRC misc parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info mps_stat_sram_intr_info[] = {
+ { 0x1fffff, "MPS statistics SRAM parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info mps_stat_tx_intr_info[] = {
+ { 0xfffff, "MPS statistics Tx FIFO parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info mps_stat_rx_intr_info[] = {
+ { 0xffffff, "MPS statistics Rx FIFO parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+ static struct intr_info mps_cls_intr_info[] = {
+ { MATCHSRAM, "MPS match SRAM parity error", -1, 1 },
+ { MATCHTCAM, "MPS match TCAM parity error", -1, 1 },
+ { HASHSRAM, "MPS hash SRAM parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ int fat;
+
+ fat = csio_handle_intr_status(hw, MPS_RX_PERR_INT_CAUSE,
+ mps_rx_intr_info) +
+ csio_handle_intr_status(hw, MPS_TX_INT_CAUSE,
+ mps_tx_intr_info) +
+ csio_handle_intr_status(hw, MPS_TRC_INT_CAUSE,
+ mps_trc_intr_info) +
+ csio_handle_intr_status(hw, MPS_STAT_PERR_INT_CAUSE_SRAM,
+ mps_stat_sram_intr_info) +
+ csio_handle_intr_status(hw, MPS_STAT_PERR_INT_CAUSE_TX_FIFO,
+ mps_stat_tx_intr_info) +
+ csio_handle_intr_status(hw, MPS_STAT_PERR_INT_CAUSE_RX_FIFO,
+ mps_stat_rx_intr_info) +
+ csio_handle_intr_status(hw, MPS_CLS_INT_CAUSE,
+ mps_cls_intr_info);
+
+ csio_wr_reg32(hw, 0, MPS_INT_CAUSE);
+ csio_rd_reg32(hw, MPS_INT_CAUSE); /* flush */
+ if (fat)
+ csio_hw_fatal_err(hw);
+}
+
+#define MEM_INT_MASK (PERR_INT_CAUSE | ECC_CE_INT_CAUSE | ECC_UE_INT_CAUSE)
+
+/*
+ * EDC/MC interrupt handler.
+ */
+static void csio_mem_intr_handler(struct csio_hw *hw, int idx)
+{
+ static const char name[3][5] = { "EDC0", "EDC1", "MC" };
+
+ unsigned int addr, cnt_addr, v;
+
+ if (idx <= MEM_EDC1) {
+ addr = EDC_REG(EDC_INT_CAUSE, idx);
+ cnt_addr = EDC_REG(EDC_ECC_STATUS, idx);
+ } else {
+ addr = MC_INT_CAUSE;
+ cnt_addr = MC_ECC_STATUS;
+ }
+
+ v = csio_rd_reg32(hw, addr) & MEM_INT_MASK;
+ if (v & PERR_INT_CAUSE)
+ csio_fatal(hw, "%s FIFO parity error\n", name[idx]);
+ if (v & ECC_CE_INT_CAUSE) {
+ uint32_t cnt = ECC_CECNT_GET(csio_rd_reg32(hw, cnt_addr));
+
+ csio_wr_reg32(hw, ECC_CECNT_MASK, cnt_addr);
+ csio_warn(hw, "%u %s correctable ECC data error%s\n",
+ cnt, name[idx], cnt > 1 ? "s" : "");
+ }
+ if (v & ECC_UE_INT_CAUSE)
+ csio_fatal(hw, "%s uncorrectable ECC data error\n", name[idx]);
+
+ csio_wr_reg32(hw, v, addr);
+ if (v & (PERR_INT_CAUSE | ECC_UE_INT_CAUSE))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * MA interrupt handler.
+ */
+static void csio_ma_intr_handler(struct csio_hw *hw)
+{
+ uint32_t v, status = csio_rd_reg32(hw, MA_INT_CAUSE);
+
+ if (status & MEM_PERR_INT_CAUSE)
+ csio_fatal(hw, "MA parity error, parity status %#x\n",
+ csio_rd_reg32(hw, MA_PARITY_ERROR_STATUS));
+ if (status & MEM_WRAP_INT_CAUSE) {
+ v = csio_rd_reg32(hw, MA_INT_WRAP_STATUS);
+ csio_fatal(hw,
+ "MA address wrap-around error by client %u to address %#x\n",
+ MEM_WRAP_CLIENT_NUM_GET(v), MEM_WRAP_ADDRESS_GET(v) << 4);
+ }
+ csio_wr_reg32(hw, status, MA_INT_CAUSE);
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * SMB interrupt handler.
+ */
+static void csio_smb_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info smb_intr_info[] = {
+ { MSTTXFIFOPARINT, "SMB master Tx FIFO parity error", -1, 1 },
+ { MSTRXFIFOPARINT, "SMB master Rx FIFO parity error", -1, 1 },
+ { SLVFIFOPARINT, "SMB slave FIFO parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, SMB_INT_CAUSE, smb_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * NC-SI interrupt handler.
+ */
+static void csio_ncsi_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info ncsi_intr_info[] = {
+ { CIM_DM_PRTY_ERR, "NC-SI CIM parity error", -1, 1 },
+ { MPS_DM_PRTY_ERR, "NC-SI MPS parity error", -1, 1 },
+ { TXFIFO_PRTY_ERR, "NC-SI Tx FIFO parity error", -1, 1 },
+ { RXFIFO_PRTY_ERR, "NC-SI Rx FIFO parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, NCSI_INT_CAUSE, ncsi_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * XGMAC interrupt handler.
+ */
+static void csio_xgmac_intr_handler(struct csio_hw *hw, int port)
+{
+ uint32_t v = csio_rd_reg32(hw, PORT_REG(port, XGMAC_PORT_INT_CAUSE));
+
+ v &= TXFIFO_PRTY_ERR | RXFIFO_PRTY_ERR;
+ if (!v)
+ return;
+
+ if (v & TXFIFO_PRTY_ERR)
+ csio_fatal(hw, "XGMAC %d Tx FIFO parity error\n", port);
+ if (v & RXFIFO_PRTY_ERR)
+ csio_fatal(hw, "XGMAC %d Rx FIFO parity error\n", port);
+ csio_wr_reg32(hw, v, PORT_REG(port, XGMAC_PORT_INT_CAUSE));
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * PL interrupt handler.
+ */
+static void csio_pl_intr_handler(struct csio_hw *hw)
+{
+ static struct intr_info pl_intr_info[] = {
+ { FATALPERR, "T4 fatal parity error", -1, 1 },
+ { PERRVFID, "PL VFID_MAP parity error", -1, 1 },
+ { 0, NULL, 0, 0 }
+ };
+
+ if (csio_handle_intr_status(hw, PL_PL_INT_CAUSE, pl_intr_info))
+ csio_hw_fatal_err(hw);
+}
+
+/*
+ * csio_hw_slow_intr_handler - control path interrupt handler
+ * @hw: HW module
+ *
+ * Interrupt handler for non-data global interrupt events, e.g., errors.
+ * The designation 'slow' is because it involves register reads, while
+ * data interrupts typically don't involve any MMIOs.
+ */
+int
+csio_hw_slow_intr_handler(struct csio_hw *hw)
+{
+ uint32_t cause = csio_rd_reg32(hw, PL_INT_CAUSE);
+
+ if (!(cause & CSIO_GLBL_INTR_MASK)) {
+ CSIO_INC_STATS(hw, n_plint_unexp);
+ return 0;
+ }
+
+ csio_dbg(hw, "Slow interrupt! cause: 0x%x\n", cause);
+
+ CSIO_INC_STATS(hw, n_plint_cnt);
+
+ if (cause & CIM)
+ csio_cim_intr_handler(hw);
+
+ if (cause & MPS)
+ csio_mps_intr_handler(hw);
+
+ if (cause & NCSI)
+ csio_ncsi_intr_handler(hw);
+
+ if (cause & PL)
+ csio_pl_intr_handler(hw);
+
+ if (cause & SMB)
+ csio_smb_intr_handler(hw);
+
+ if (cause & XGMAC0)
+ csio_xgmac_intr_handler(hw, 0);
+
+ if (cause & XGMAC1)
+ csio_xgmac_intr_handler(hw, 1);
+
+ if (cause & XGMAC_KR0)
+ csio_xgmac_intr_handler(hw, 2);
+
+ if (cause & XGMAC_KR1)
+ csio_xgmac_intr_handler(hw, 3);
+
+ if (cause & PCIE)
+ csio_pcie_intr_handler(hw);
+
+ if (cause & MC)
+ csio_mem_intr_handler(hw, MEM_MC);
+
+ if (cause & EDC0)
+ csio_mem_intr_handler(hw, MEM_EDC0);
+
+ if (cause & EDC1)
+ csio_mem_intr_handler(hw, MEM_EDC1);
+
+ if (cause & LE)
+ csio_le_intr_handler(hw);
+
+ if (cause & TP)
+ csio_tp_intr_handler(hw);
+
+ if (cause & MA)
+ csio_ma_intr_handler(hw);
+
+ if (cause & PM_TX)
+ csio_pmtx_intr_handler(hw);
+
+ if (cause & PM_RX)
+ csio_pmrx_intr_handler(hw);
+
+ if (cause & ULP_RX)
+ csio_ulprx_intr_handler(hw);
+
+ if (cause & CPL_SWITCH)
+ csio_cplsw_intr_handler(hw);
+
+ if (cause & SGE)
+ csio_sge_intr_handler(hw);
+
+ if (cause & ULP_TX)
+ csio_ulptx_intr_handler(hw);
+
+ /* Clear the interrupts just processed for which we are the master. */
+ csio_wr_reg32(hw, cause & CSIO_GLBL_INTR_MASK, PL_INT_CAUSE);
+ csio_rd_reg32(hw, PL_INT_CAUSE); /* flush */
+
+ return 1;
+}
+
+/*****************************************************************************
+ * HW <--> mailbox interfacing routines.
+ ****************************************************************************/
+/*
+ * csio_mberr_worker - Worker thread (dpc) for mailbox/error completions
+ *
+ * @data: Private data pointer.
+ *
+ * Called from worker thread context.
+ */
+static void
+csio_mberr_worker(void *data)
+{
+ struct csio_hw *hw = (struct csio_hw *)data;
+ struct csio_mbm *mbm = &hw->mbm;
+ LIST_HEAD(cbfn_q);
+ struct csio_mb *mbp_next;
+ int rv;
+
+ del_timer_sync(&mbm->timer);
+
+ spin_lock_irq(&hw->lock);
+ if (list_empty(&mbm->cbfn_q)) {
+ spin_unlock_irq(&hw->lock);
+ return;
+ }
+
+ list_splice_tail_init(&mbm->cbfn_q, &cbfn_q);
+ mbm->stats.n_cbfnq = 0;
+
+ /* Try to start waiting mailboxes */
+ if (!list_empty(&mbm->req_q)) {
+ mbp_next = list_first_entry(&mbm->req_q, struct csio_mb, list);
+ list_del_init(&mbp_next->list);
+
+ rv = csio_mb_issue(hw, mbp_next);
+ if (rv != 0)
+ list_add_tail(&mbp_next->list, &mbm->req_q);
+ else
+ CSIO_DEC_STATS(mbm, n_activeq);
+ }
+ spin_unlock_irq(&hw->lock);
+
+ /* Now callback completions */
+ csio_mb_completions(hw, &cbfn_q);
+}
+
+/*
+ * csio_hw_mb_timer - Top-level Mailbox timeout handler.
+ *
+ * @data: private data pointer
+ *
+ **/
+static void
+csio_hw_mb_timer(uintptr_t data)
+{
+ struct csio_hw *hw = (struct csio_hw *)data;
+ struct csio_mb *mbp = NULL;
+
+ spin_lock_irq(&hw->lock);
+ mbp = csio_mb_tmo_handler(hw);
+ spin_unlock_irq(&hw->lock);
+
+ /* Call back the function for the timed-out Mailbox */
+ if (mbp)
+ mbp->mb_cbfn(hw, mbp);
+
+}
+
+/*
+ * csio_hw_mbm_cleanup - Cleanup Mailbox module.
+ * @hw: HW module
+ *
+ * Called with lock held, should exit with lock held.
+ * Cancels outstanding mailboxes (waiting, in-flight) and gathers them
+ * into a local queue. Drops lock and calls the completions. Holds
+ * lock and returns.
+ */
+static void
+csio_hw_mbm_cleanup(struct csio_hw *hw)
+{
+ LIST_HEAD(cbfn_q);
+
+ csio_mb_cancel_all(hw, &cbfn_q);
+
+ spin_unlock_irq(&hw->lock);
+ csio_mb_completions(hw, &cbfn_q);
+ spin_lock_irq(&hw->lock);
+}
+
+/*****************************************************************************
+ * Event handling
+ ****************************************************************************/
+int
+csio_enqueue_evt(struct csio_hw *hw, enum csio_evt type, void *evt_msg,
+ uint16_t len)
+{
+ struct csio_evt_msg *evt_entry = NULL;
+
+ if (type >= CSIO_EVT_MAX)
+ return -EINVAL;
+
+ if (len > CSIO_EVT_MSG_SIZE)
+ return -EINVAL;
+
+ if (hw->flags & CSIO_HWF_FWEVT_STOP)
+ return -EINVAL;
+
+ if (list_empty(&hw->evt_free_q)) {
+ csio_err(hw, "Failed to alloc evt entry, msg type %d len %d\n",
+ type, len);
+ return -ENOMEM;
+ }
+
+ evt_entry = list_first_entry(&hw->evt_free_q,
+ struct csio_evt_msg, list);
+ list_del_init(&evt_entry->list);
+
+ /* copy event msg and queue the event */
+ evt_entry->type = type;
+ memcpy((void *)evt_entry->data, evt_msg, len);
+ list_add_tail(&evt_entry->list, &hw->evt_active_q);
+
+ CSIO_DEC_STATS(hw, n_evt_freeq);
+ CSIO_INC_STATS(hw, n_evt_activeq);
+
+ return 0;
+}
+
+static int
+csio_enqueue_evt_lock(struct csio_hw *hw, enum csio_evt type, void *evt_msg,
+ uint16_t len, bool msg_sg)
+{
+ struct csio_evt_msg *evt_entry = NULL;
+ struct csio_fl_dma_buf *fl_sg;
+ uint32_t off = 0;
+ unsigned long flags;
+ int n, ret = 0;
+
+ if (type >= CSIO_EVT_MAX)
+ return -EINVAL;
+
+ if (len > CSIO_EVT_MSG_SIZE)
+ return -EINVAL;
+
+ spin_lock_irqsave(&hw->lock, flags);
+ if (hw->flags & CSIO_HWF_FWEVT_STOP) {
+ ret = -EINVAL;
+ goto out;
+ }
+
+ if (list_empty(&hw->evt_free_q)) {
+ csio_err(hw, "Failed to alloc evt entry, msg type %d len %d\n",
+ type, len);
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ evt_entry = list_first_entry(&hw->evt_free_q,
+ struct csio_evt_msg, list);
+ list_del_init(&evt_entry->list);
+
+ /* copy event msg and queue the event */
+ evt_entry->type = type;
+
+ /* If Payload in SG list*/
+ if (msg_sg) {
+ fl_sg = (struct csio_fl_dma_buf *) evt_msg;
+ for (n = 0; (n < CSIO_MAX_FLBUF_PER_IQWR && off < len); n++) {
+ memcpy((void *)((uintptr_t)evt_entry->data + off),
+ fl_sg->flbufs[n].vaddr,
+ fl_sg->flbufs[n].len);
+ off += fl_sg->flbufs[n].len;
+ }
+ } else
+ memcpy((void *)evt_entry->data, evt_msg, len);
+
+ list_add_tail(&evt_entry->list, &hw->evt_active_q);
+ CSIO_DEC_STATS(hw, n_evt_freeq);
+ CSIO_INC_STATS(hw, n_evt_activeq);
+out:
+ spin_unlock_irqrestore(&hw->lock, flags);
+ return ret;
+}
+
+static void
+csio_free_evt(struct csio_hw *hw, struct csio_evt_msg *evt_entry)
+{
+ if (evt_entry) {
+ spin_lock_irq(&hw->lock);
+ list_del_init(&evt_entry->list);
+ list_add_tail(&evt_entry->list, &hw->evt_free_q);
+ CSIO_DEC_STATS(hw, n_evt_activeq);
+ CSIO_INC_STATS(hw, n_evt_freeq);
+ spin_unlock_irq(&hw->lock);
+ }
+}
+
+void
+csio_evtq_flush(struct csio_hw *hw)
+{
+ uint32_t count;
+ count = 30;
+ while (hw->flags & CSIO_HWF_FWEVT_PENDING && count--) {
+ spin_unlock_irq(&hw->lock);
+ msleep(2000);
+ spin_lock_irq(&hw->lock);
+ }
+
+ CSIO_DB_ASSERT(!(hw->flags & CSIO_HWF_FWEVT_PENDING));
+}
+
+static void
+csio_evtq_stop(struct csio_hw *hw)
+{
+ hw->flags |= CSIO_HWF_FWEVT_STOP;
+}
+
+static void
+csio_evtq_start(struct csio_hw *hw)
+{
+ hw->flags &= ~CSIO_HWF_FWEVT_STOP;
+}
+
+static void
+csio_evtq_cleanup(struct csio_hw *hw)
+{
+ struct list_head *evt_entry, *next_entry;
+
+ /* Release outstanding events from activeq to freeq*/
+ if (!list_empty(&hw->evt_active_q))
+ list_splice_tail_init(&hw->evt_active_q, &hw->evt_free_q);
+
+ hw->stats.n_evt_activeq = 0;
+ hw->flags &= ~CSIO_HWF_FWEVT_PENDING;
+
+ /* Freeup event entry */
+ list_for_each_safe(evt_entry, next_entry, &hw->evt_free_q) {
+ kfree(evt_entry);
+ CSIO_DEC_STATS(hw, n_evt_freeq);
+ }
+
+ hw->stats.n_evt_freeq = 0;
+}
+
+
+static void
+csio_process_fwevtq_entry(struct csio_hw *hw, void *wr, uint32_t len,
+ struct csio_fl_dma_buf *flb, void *priv)
+{
+ __u8 op;
+ __be64 *data;
+ void *msg = NULL;
+ uint32_t msg_len = 0;
+ bool msg_sg = 0;
+
+ op = ((struct rss_header *) wr)->opcode;
+ if (op == CPL_FW6_PLD) {
+ CSIO_INC_STATS(hw, n_cpl_fw6_pld);
+ if (!flb || !flb->totlen) {
+ CSIO_INC_STATS(hw, n_cpl_unexp);
+ return;
+ }
+
+ msg = (void *) flb;
+ msg_len = flb->totlen;
+ msg_sg = 1;
+
+ data = (__be64 *) msg;
+ } else if (op == CPL_FW6_MSG || op == CPL_FW4_MSG) {
+
+ CSIO_INC_STATS(hw, n_cpl_fw6_msg);
+ /* skip RSS header */
+ msg = (void *)((uintptr_t)wr + sizeof(__be64));
+ msg_len = (op == CPL_FW6_MSG) ? sizeof(struct cpl_fw6_msg) :
+ sizeof(struct cpl_fw4_msg);
+
+ data = (__be64 *) msg;
+ } else {
+ csio_warn(hw, "unexpected CPL %#x on FW event queue\n", op);
+ CSIO_INC_STATS(hw, n_cpl_unexp);
+ return;
+ }
+
+ /*
+ * Enqueue event to EventQ. Events processing happens
+ * in Event worker thread context
+ */
+ if (csio_enqueue_evt_lock(hw, CSIO_EVT_FW, msg,
+ (uint16_t)msg_len, msg_sg))
+ CSIO_INC_STATS(hw, n_evt_drop);
+}
+
+void
+csio_evtq_worker(struct work_struct *work)
+{
+ struct csio_hw *hw = container_of(work, struct csio_hw, evtq_work);
+ struct list_head *evt_entry, *next_entry;
+ LIST_HEAD(evt_q);
+ struct csio_evt_msg *evt_msg;
+ struct cpl_fw6_msg *msg;
+ struct csio_rnode *rn;
+ int rv = 0;
+ uint8_t evtq_stop = 0;
+
+ csio_dbg(hw, "event worker thread active evts#%d\n",
+ hw->stats.n_evt_activeq);
+
+ spin_lock_irq(&hw->lock);
+ while (!list_empty(&hw->evt_active_q)) {
+ list_splice_tail_init(&hw->evt_active_q, &evt_q);
+ spin_unlock_irq(&hw->lock);
+
+ list_for_each_safe(evt_entry, next_entry, &evt_q) {
+ evt_msg = (struct csio_evt_msg *) evt_entry;
+
+ /* Drop events if queue is STOPPED */
+ spin_lock_irq(&hw->lock);
+ if (hw->flags & CSIO_HWF_FWEVT_STOP)
+ evtq_stop = 1;
+ spin_unlock_irq(&hw->lock);
+ if (evtq_stop) {
+ CSIO_INC_STATS(hw, n_evt_drop);
+ goto free_evt;
+ }
+
+ switch (evt_msg->type) {
+ case CSIO_EVT_FW:
+ msg = (struct cpl_fw6_msg *)(evt_msg->data);
+
+ if ((msg->opcode == CPL_FW6_MSG ||
+ msg->opcode == CPL_FW4_MSG) &&
+ !msg->type) {
+ rv = csio_mb_fwevt_handler(hw,
+ msg->data);
+ if (!rv)
+ break;
+ /* Handle any remaining fw events */
+ csio_fcoe_fwevt_handler(hw,
+ msg->opcode, msg->data);
+ } else if (msg->opcode == CPL_FW6_PLD) {
+
+ csio_fcoe_fwevt_handler(hw,
+ msg->opcode, msg->data);
+ } else {
+ csio_warn(hw,
+ "Unhandled FW msg op %x type %x\n",
+ msg->opcode, msg->type);
+ CSIO_INC_STATS(hw, n_evt_drop);
+ }
+ break;
+
+ case CSIO_EVT_MBX:
+ csio_mberr_worker(hw);
+ break;
+
+ case CSIO_EVT_DEV_LOSS:
+ memcpy(&rn, evt_msg->data, sizeof(rn));
+ csio_rnode_devloss_handler(rn);
+ break;
+
+ default:
+ csio_warn(hw, "Unhandled event %x on evtq\n",
+ evt_msg->type);
+ CSIO_INC_STATS(hw, n_evt_unexp);
+ break;
+ }
+free_evt:
+ csio_free_evt(hw, evt_msg);
+ }
+
+ spin_lock_irq(&hw->lock);
+ }
+ hw->flags &= ~CSIO_HWF_FWEVT_PENDING;
+ spin_unlock_irq(&hw->lock);
+}
+
+int
+csio_fwevtq_handler(struct csio_hw *hw)
+{
+ int rv;
+
+ if (csio_q_iqid(hw, hw->fwevt_iq_idx) == CSIO_MAX_QID) {
+ CSIO_INC_STATS(hw, n_int_stray);
+ return -EINVAL;
+ }
+
+ rv = csio_wr_process_iq_idx(hw, hw->fwevt_iq_idx,
+ csio_process_fwevtq_entry, NULL);
+ return rv;
+}
+
+/****************************************************************************
+ * Entry points
+ ****************************************************************************/
+
+/* Management module */
+/*
+ * csio_mgmt_req_lookup - Lookup the given IO req exist in Active Q.
+ * mgmt - mgmt module
+ * @io_req - io request
+ *
+ * Return - 0:if given IO Req exists in active Q.
+ * -EINVAL :if lookup fails.
+ */
+int
+csio_mgmt_req_lookup(struct csio_mgmtm *mgmtm, struct csio_ioreq *io_req)
+{
+ struct list_head *tmp;
+
+ /* Lookup ioreq in the ACTIVEQ */
+ list_for_each(tmp, &mgmtm->active_q) {
+ if (io_req == (struct csio_ioreq *)tmp)
+ return 0;
+ }
+ return -EINVAL;
+}
+
+#define ECM_MIN_TMO 1000 /* Minimum timeout value for req */
+
+/*
+ * csio_mgmts_tmo_handler - MGMT IO Timeout handler.
+ * @data - Event data.
+ *
+ * Return - none.
+ */
+static void
+csio_mgmt_tmo_handler(uintptr_t data)
+{
+ struct csio_mgmtm *mgmtm = (struct csio_mgmtm *) data;
+ struct list_head *tmp;
+ struct csio_ioreq *io_req;
+
+ csio_dbg(mgmtm->hw, "Mgmt timer invoked!\n");
+
+ spin_lock_irq(&mgmtm->hw->lock);
+
+ list_for_each(tmp, &mgmtm->active_q) {
+ io_req = (struct csio_ioreq *) tmp;
+ io_req->tmo -= min_t(uint32_t, io_req->tmo, ECM_MIN_TMO);
+
+ if (!io_req->tmo) {
+ /* Dequeue the request from retry Q. */
+ tmp = csio_list_prev(tmp);
+ list_del_init(&io_req->sm.sm_list);
+ if (io_req->io_cbfn) {
+ /* io_req will be freed by completion handler */
+ io_req->wr_status = -ETIMEDOUT;
+ io_req->io_cbfn(mgmtm->hw, io_req);
+ } else {
+ CSIO_DB_ASSERT(0);
+ }
+ }
+ }
+
+ /* If retry queue is not empty, re-arm timer */
+ if (!list_empty(&mgmtm->active_q))
+ mod_timer(&mgmtm->mgmt_timer,
+ jiffies + msecs_to_jiffies(ECM_MIN_TMO));
+ spin_unlock_irq(&mgmtm->hw->lock);
+}
+
+static void
+csio_mgmtm_cleanup(struct csio_mgmtm *mgmtm)
+{
+ struct csio_hw *hw = mgmtm->hw;
+ struct csio_ioreq *io_req;
+ struct list_head *tmp;
+ uint32_t count;
+
+ count = 30;
+ /* Wait for all outstanding req to complete gracefully */
+ while ((!list_empty(&mgmtm->active_q)) && count--) {
+ spin_unlock_irq(&hw->lock);
+ msleep(2000);
+ spin_lock_irq(&hw->lock);
+ }
+
+ /* release outstanding req from ACTIVEQ */
+ list_for_each(tmp, &mgmtm->active_q) {
+ io_req = (struct csio_ioreq *) tmp;
+ tmp = csio_list_prev(tmp);
+ list_del_init(&io_req->sm.sm_list);
+ mgmtm->stats.n_active--;
+ if (io_req->io_cbfn) {
+ /* io_req will be freed by completion handler */
+ io_req->wr_status = -ETIMEDOUT;
+ io_req->io_cbfn(mgmtm->hw, io_req);
+ }
+ }
+}
+
+/*
+ * csio_mgmt_init - Mgmt module init entry point
+ * @mgmtsm - mgmt module
+ * @hw - HW module
+ *
+ * Initialize mgmt timer, resource wait queue, active queue,
+ * completion q. Allocate Egress and Ingress
+ * WR queues and save off the queue index returned by the WR
+ * module for future use. Allocate and save off mgmt reqs in the
+ * mgmt_req_freelist for future use. Make sure their SM is initialized
+ * to uninit state.
+ * Returns: 0 - on success
+ * -ENOMEM - on error.
+ */
+static int
+csio_mgmtm_init(struct csio_mgmtm *mgmtm, struct csio_hw *hw)
+{
+ struct timer_list *timer = &mgmtm->mgmt_timer;
+
+ init_timer(timer);
+ timer->function = csio_mgmt_tmo_handler;
+ timer->data = (unsigned long)mgmtm;
+
+ INIT_LIST_HEAD(&mgmtm->active_q);
+ INIT_LIST_HEAD(&mgmtm->cbfn_q);
+
+ mgmtm->hw = hw;
+ /*mgmtm->iq_idx = hw->fwevt_iq_idx;*/
+
+ return 0;
+}
+
+/*
+ * csio_mgmtm_exit - MGMT module exit entry point
+ * @mgmtsm - mgmt module
+ *
+ * This function called during MGMT module uninit.
+ * Stop timers, free ioreqs allocated.
+ * Returns: None
+ *
+ */
+static void
+csio_mgmtm_exit(struct csio_mgmtm *mgmtm)
+{
+ del_timer_sync(&mgmtm->mgmt_timer);
+}
+
+
+/**
+ * csio_hw_start - Kicks off the HW State machine
+ * @hw: Pointer to HW module.
+ *
+ * It is assumed that the initialization is a synchronous operation.
+ * So when we return afer posting the event, the HW SM should be in
+ * the ready state, if there were no errors during init.
+ */
+int
+csio_hw_start(struct csio_hw *hw)
+{
+ spin_lock_irq(&hw->lock);
+ csio_post_event(&hw->sm, CSIO_HWE_CFG);
+ spin_unlock_irq(&hw->lock);
+
+ if (csio_is_hw_ready(hw))
+ return 0;
+ else
+ return -EINVAL;
+}
+
+int
+csio_hw_stop(struct csio_hw *hw)
+{
+ csio_post_event(&hw->sm, CSIO_HWE_PCI_REMOVE);
+
+ if (csio_is_hw_removing(hw))
+ return 0;
+ else
+ return -EINVAL;
+}
+
+/* Max reset retries */
+#define CSIO_MAX_RESET_RETRIES 3
+
+/**
+ * csio_hw_reset - Reset the hardware
+ * @hw: HW module.
+ *
+ * Caller should hold lock across this function.
+ */
+int
+csio_hw_reset(struct csio_hw *hw)
+{
+ if (!csio_is_hw_master(hw))
+ return -EPERM;
+
+ if (hw->rst_retries >= CSIO_MAX_RESET_RETRIES) {
+ csio_dbg(hw, "Max hw reset attempts reached..");
+ return -EINVAL;
+ }
+
+ hw->rst_retries++;
+ csio_post_event(&hw->sm, CSIO_HWE_HBA_RESET);
+
+ if (csio_is_hw_ready(hw)) {
+ hw->rst_retries = 0;
+ hw->stats.n_reset_start = jiffies_to_msecs(jiffies);
+ return 0;
+ } else
+ return -EINVAL;
+}
+
+/*
+ * csio_hw_get_device_id - Caches the Adapter's vendor & device id.
+ * @hw: HW module.
+ */
+static void
+csio_hw_get_device_id(struct csio_hw *hw)
+{
+ /* Is the adapter device id cached already ?*/
+ if (csio_is_dev_id_cached(hw))
+ return;
+
+ /* Get the PCI vendor & device id */
+ pci_read_config_word(hw->pdev, PCI_VENDOR_ID,
+ &hw->params.pci.vendor_id);
+ pci_read_config_word(hw->pdev, PCI_DEVICE_ID,
+ &hw->params.pci.device_id);
+
+ csio_dev_id_cached(hw);
+
+} /* csio_hw_get_device_id */
+
+/*
+ * csio_hw_set_description - Set the model, description of the hw.
+ * @hw: HW module.
+ * @ven_id: PCI Vendor ID
+ * @dev_id: PCI Device ID
+ */
+static void
+csio_hw_set_description(struct csio_hw *hw, uint16_t ven_id, uint16_t dev_id)
+{
+ uint32_t adap_type, prot_type;
+
+ if (ven_id == CSIO_VENDOR_ID) {
+ prot_type = (dev_id & CSIO_ASIC_DEVID_PROTO_MASK);
+ adap_type = (dev_id & CSIO_ASIC_DEVID_TYPE_MASK);
+
+ if (prot_type == CSIO_FPGA) {
+ memcpy(hw->model_desc,
+ csio_fcoe_adapters[13].description, 32);
+ } else if (prot_type == CSIO_T4_FCOE_ASIC) {
+ memcpy(hw->hw_ver,
+ csio_fcoe_adapters[adap_type].model_no, 16);
+ memcpy(hw->model_desc,
+ csio_fcoe_adapters[adap_type].description, 32);
+ } else {
+ char tempName[32] = "Chelsio FCoE Controller";
+ memcpy(hw->model_desc, tempName, 32);
+
+ CSIO_DB_ASSERT(0);
+ }
+ }
+} /* csio_hw_set_description */
+
+/**
+ * csio_hw_init - Initialize HW module.
+ * @hw: Pointer to HW module.
+ *
+ * Initialize the members of the HW module.
+ */
+int
+csio_hw_init(struct csio_hw *hw)
+{
+ int rv = -EINVAL;
+ uint32_t i;
+ uint16_t ven_id, dev_id;
+ struct csio_evt_msg *evt_entry;
+
+ INIT_LIST_HEAD(&hw->sm.sm_list);
+ csio_init_state(&hw->sm, csio_hws_uninit);
+ spin_lock_init(&hw->lock);
+ INIT_LIST_HEAD(&hw->sln_head);
+
+ /* Get the PCI vendor & device id */
+ csio_hw_get_device_id(hw);
+
+ strcpy(hw->name, CSIO_HW_NAME);
+
+ /* Set the model & its description */
+
+ ven_id = hw->params.pci.vendor_id;
+ dev_id = hw->params.pci.device_id;
+
+ csio_hw_set_description(hw, ven_id, dev_id);
+
+ /* Initialize default log level */
+ hw->params.log_level = (uint32_t) csio_dbg_level;
+
+ csio_set_fwevt_intr_idx(hw, -1);
+ csio_set_nondata_intr_idx(hw, -1);
+
+ /* Init all the modules: Mailbox, WorkRequest and Transport */
+ if (csio_mbm_init(csio_hw_to_mbm(hw), hw, csio_hw_mb_timer))
+ goto err;
+
+ rv = csio_wrm_init(csio_hw_to_wrm(hw), hw);
+ if (rv)
+ goto err_mbm_exit;
+
+ rv = csio_scsim_init(csio_hw_to_scsim(hw), hw);
+ if (rv)
+ goto err_wrm_exit;
+
+ rv = csio_mgmtm_init(csio_hw_to_mgmtm(hw), hw);
+ if (rv)
+ goto err_scsim_exit;
+ /* Pre-allocate evtq and initialize them */
+ INIT_LIST_HEAD(&hw->evt_active_q);
+ INIT_LIST_HEAD(&hw->evt_free_q);
+ for (i = 0; i < csio_evtq_sz; i++) {
+
+ evt_entry = kzalloc(sizeof(struct csio_evt_msg), GFP_KERNEL);
+ if (!evt_entry) {
+ csio_err(hw, "Failed to initialize eventq");
+ goto err_evtq_cleanup;
+ }
+
+ list_add_tail(&evt_entry->list, &hw->evt_free_q);
+ CSIO_INC_STATS(hw, n_evt_freeq);
+ }
+
+ hw->dev_num = dev_num;
+ dev_num++;
+
+ return 0;
+
+err_evtq_cleanup:
+ csio_evtq_cleanup(hw);
+ csio_mgmtm_exit(csio_hw_to_mgmtm(hw));
+err_scsim_exit:
+ csio_scsim_exit(csio_hw_to_scsim(hw));
+err_wrm_exit:
+ csio_wrm_exit(csio_hw_to_wrm(hw), hw);
+err_mbm_exit:
+ csio_mbm_exit(csio_hw_to_mbm(hw));
+err:
+ return rv;
+}
+
+/**
+ * csio_hw_exit - Un-initialize HW module.
+ * @hw: Pointer to HW module.
+ *
+ */
+void
+csio_hw_exit(struct csio_hw *hw)
+{
+ csio_evtq_cleanup(hw);
+ csio_mgmtm_exit(csio_hw_to_mgmtm(hw));
+ csio_scsim_exit(csio_hw_to_scsim(hw));
+ csio_wrm_exit(csio_hw_to_wrm(hw), hw);
+ csio_mbm_exit(csio_hw_to_mbm(hw));
+}
--
1.7.1
^ permalink raw reply related
* [v6 PATCH 6/8] csiostor: Chelsio FCoE offload driver submission (sources part 3).
From: Naresh Kumar Inna @ 2012-11-15 17:11 UTC (permalink / raw)
To: JBottomley, linux-scsi, dm, leedom; +Cc: netdev, naresh, chethan
In-Reply-To: <1352999484-17812-1-git-send-email-naresh@chelsio.com>
This patch contains code to implement the local and remote node port
functionality. It includes tracking the firmware events for changes to
the states of these ports.
Signed-off-by: Naresh Kumar Inna <naresh@chelsio.com>
---
V5: Re-wrote error paths to reduce lock/unlock ambiguity.
drivers/scsi/csiostor/csio_lnode.c | 2133 ++++++++++++++++++++++++++++++++++++
drivers/scsi/csiostor/csio_rnode.c | 912 +++++++++++++++
2 files changed, 3045 insertions(+), 0 deletions(-)
create mode 100644 drivers/scsi/csiostor/csio_lnode.c
create mode 100644 drivers/scsi/csiostor/csio_rnode.c
diff --git a/drivers/scsi/csiostor/csio_lnode.c b/drivers/scsi/csiostor/csio_lnode.c
new file mode 100644
index 0000000..551959e
--- /dev/null
+++ b/drivers/scsi/csiostor/csio_lnode.c
@@ -0,0 +1,2133 @@
+/*
+ * This file is part of the Chelsio FCoE driver for Linux.
+ *
+ * Copyright (c) 2008-2012 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/utsname.h>
+#include <scsi/scsi_device.h>
+#include <scsi/scsi_transport_fc.h>
+#include <asm/unaligned.h>
+#include <scsi/fc/fc_els.h>
+#include <scsi/fc/fc_fs.h>
+#include <scsi/fc/fc_gs.h>
+#include <scsi/fc/fc_ms.h>
+
+#include "csio_hw.h"
+#include "csio_mb.h"
+#include "csio_lnode.h"
+#include "csio_rnode.h"
+
+int csio_fcoe_rnodes = 1024;
+int csio_fdmi_enable = 1;
+
+#define PORT_ID_PTR(_x) ((uint8_t *)(&_x) + 1)
+
+/* Lnode SM declarations */
+static void csio_lns_uninit(struct csio_lnode *, enum csio_ln_ev);
+static void csio_lns_online(struct csio_lnode *, enum csio_ln_ev);
+static void csio_lns_ready(struct csio_lnode *, enum csio_ln_ev);
+static void csio_lns_offline(struct csio_lnode *, enum csio_ln_ev);
+
+static int csio_ln_mgmt_submit_req(struct csio_ioreq *,
+ void (*io_cbfn) (struct csio_hw *, struct csio_ioreq *),
+ enum fcoe_cmn_type, struct csio_dma_buf *, uint32_t);
+
+/* LN event mapping */
+static enum csio_ln_ev fwevt_to_lnevt[] = {
+ CSIO_LNE_NONE, /* None */
+ CSIO_LNE_NONE, /* PLOGI_ACC_RCVD */
+ CSIO_LNE_NONE, /* PLOGI_RJT_RCVD */
+ CSIO_LNE_NONE, /* PLOGI_RCVD */
+ CSIO_LNE_NONE, /* PLOGO_RCVD */
+ CSIO_LNE_NONE, /* PRLI_ACC_RCVD */
+ CSIO_LNE_NONE, /* PRLI_RJT_RCVD */
+ CSIO_LNE_NONE, /* PRLI_RCVD */
+ CSIO_LNE_NONE, /* PRLO_RCVD */
+ CSIO_LNE_NONE, /* NPORT_ID_CHGD */
+ CSIO_LNE_LOGO, /* FLOGO_RCVD */
+ CSIO_LNE_LOGO, /* CLR_VIRT_LNK_RCVD */
+ CSIO_LNE_FAB_INIT_DONE,/* FLOGI_ACC_RCVD */
+ CSIO_LNE_NONE, /* FLOGI_RJT_RCVD */
+ CSIO_LNE_FAB_INIT_DONE,/* FDISC_ACC_RCVD */
+ CSIO_LNE_NONE, /* FDISC_RJT_RCVD */
+ CSIO_LNE_NONE, /* FLOGI_TMO_MAX_RETRY */
+ CSIO_LNE_NONE, /* IMPL_LOGO_ADISC_ACC */
+ CSIO_LNE_NONE, /* IMPL_LOGO_ADISC_RJT */
+ CSIO_LNE_NONE, /* IMPL_LOGO_ADISC_CNFLT */
+ CSIO_LNE_NONE, /* PRLI_TMO */
+ CSIO_LNE_NONE, /* ADISC_TMO */
+ CSIO_LNE_NONE, /* RSCN_DEV_LOST */
+ CSIO_LNE_NONE, /* SCR_ACC_RCVD */
+ CSIO_LNE_NONE, /* ADISC_RJT_RCVD */
+ CSIO_LNE_NONE, /* LOGO_SNT */
+ CSIO_LNE_NONE, /* PROTO_ERR_IMPL_LOGO */
+};
+
+#define CSIO_FWE_TO_LNE(_evt) ((_evt > PROTO_ERR_IMPL_LOGO) ? \
+ CSIO_LNE_NONE : \
+ fwevt_to_lnevt[_evt])
+
+#define csio_ct_rsp(cp) (((struct fc_ct_hdr *)cp)->ct_cmd)
+#define csio_ct_reason(cp) (((struct fc_ct_hdr *)cp)->ct_reason)
+#define csio_ct_expl(cp) (((struct fc_ct_hdr *)cp)->ct_explan)
+#define csio_ct_get_pld(cp) ((void *)(((uint8_t *)cp) + FC_CT_HDR_LEN))
+
+/*
+ * csio_ln_match_by_portid - lookup lnode using given portid.
+ * @hw: HW module
+ * @portid: port-id.
+ *
+ * If found, returns lnode matching given portid otherwise returns NULL.
+ */
+static struct csio_lnode *
+csio_ln_lookup_by_portid(struct csio_hw *hw, uint8_t portid)
+{
+ struct csio_lnode *ln = hw->rln;
+ struct list_head *tmp;
+
+ /* Match siblings lnode with portid */
+ list_for_each(tmp, &hw->sln_head) {
+ ln = (struct csio_lnode *) tmp;
+ if (ln->portid == portid)
+ return ln;
+ }
+
+ return NULL;
+}
+
+/*
+ * csio_ln_lookup_by_vnpi - Lookup lnode using given vnp id.
+ * @hw - HW module
+ * @vnpi - vnp index.
+ * Returns - If found, returns lnode matching given vnp id
+ * otherwise returns NULL.
+ */
+static struct csio_lnode *
+csio_ln_lookup_by_vnpi(struct csio_hw *hw, uint32_t vnp_id)
+{
+ struct list_head *tmp1, *tmp2;
+ struct csio_lnode *sln = NULL, *cln = NULL;
+
+ if (list_empty(&hw->sln_head)) {
+ CSIO_INC_STATS(hw, n_lnlkup_miss);
+ return NULL;
+ }
+ /* Traverse sibling lnodes */
+ list_for_each(tmp1, &hw->sln_head) {
+ sln = (struct csio_lnode *) tmp1;
+
+ /* Match sibling lnode */
+ if (sln->vnp_flowid == vnp_id)
+ return sln;
+
+ if (list_empty(&sln->cln_head))
+ continue;
+
+ /* Traverse children lnodes */
+ list_for_each(tmp2, &sln->cln_head) {
+ cln = (struct csio_lnode *) tmp2;
+
+ if (cln->vnp_flowid == vnp_id)
+ return cln;
+ }
+ }
+ CSIO_INC_STATS(hw, n_lnlkup_miss);
+ return NULL;
+}
+
+/**
+ * csio_lnode_lookup_by_wwpn - Lookup lnode using given wwpn.
+ * @hw: HW module.
+ * @wwpn: WWPN.
+ *
+ * If found, returns lnode matching given wwpn, returns NULL otherwise.
+ */
+struct csio_lnode *
+csio_lnode_lookup_by_wwpn(struct csio_hw *hw, uint8_t *wwpn)
+{
+ struct list_head *tmp1, *tmp2;
+ struct csio_lnode *sln = NULL, *cln = NULL;
+
+ if (list_empty(&hw->sln_head)) {
+ CSIO_INC_STATS(hw, n_lnlkup_miss);
+ return NULL;
+ }
+ /* Traverse sibling lnodes */
+ list_for_each(tmp1, &hw->sln_head) {
+ sln = (struct csio_lnode *) tmp1;
+
+ /* Match sibling lnode */
+ if (!memcmp(csio_ln_wwpn(sln), wwpn, 8))
+ return sln;
+
+ if (list_empty(&sln->cln_head))
+ continue;
+
+ /* Traverse children lnodes */
+ list_for_each(tmp2, &sln->cln_head) {
+ cln = (struct csio_lnode *) tmp2;
+
+ if (!memcmp(csio_ln_wwpn(cln), wwpn, 8))
+ return cln;
+ }
+ }
+ return NULL;
+}
+
+/* FDMI */
+static void
+csio_fill_ct_iu(void *buf, uint8_t type, uint8_t sub_type, uint16_t op)
+{
+ struct fc_ct_hdr *cmd = (struct fc_ct_hdr *)buf;
+ cmd->ct_rev = FC_CT_REV;
+ cmd->ct_fs_type = type;
+ cmd->ct_fs_subtype = sub_type;
+ cmd->ct_cmd = op;
+}
+
+static int
+csio_hostname(uint8_t *buf, size_t buf_len)
+{
+ if (snprintf(buf, buf_len, "%s", init_utsname()->nodename) > 0)
+ return 0;
+ return -1;
+}
+
+static int
+csio_osname(uint8_t *buf, size_t buf_len)
+{
+ if (snprintf(buf, buf_len, "%s %s %s",
+ init_utsname()->sysname,
+ init_utsname()->release,
+ init_utsname()->version) > 0)
+ return 0;
+
+ return -1;
+}
+
+static inline void
+csio_append_attrib(uint8_t **ptr, uint16_t type, uint8_t *val, uint16_t len)
+{
+ struct fc_fdmi_attr_entry *ae = (struct fc_fdmi_attr_entry *)*ptr;
+ ae->type = htons(type);
+ len += 4; /* includes attribute type and length */
+ len = (len + 3) & ~3; /* should be multiple of 4 bytes */
+ ae->len = htons(len);
+ memset(ae->value, 0, len - 4);
+ memcpy(ae->value, val, len);
+ *ptr += len;
+}
+
+/*
+ * csio_ln_fdmi_done - FDMI registeration completion
+ * @hw: HW context
+ * @fdmi_req: fdmi request
+ */
+static void
+csio_ln_fdmi_done(struct csio_hw *hw, struct csio_ioreq *fdmi_req)
+{
+ void *cmd;
+ struct csio_lnode *ln = fdmi_req->lnode;
+
+ if (fdmi_req->wr_status != FW_SUCCESS) {
+ csio_ln_dbg(ln, "WR error:%x in processing fdmi rpa cmd\n",
+ fdmi_req->wr_status);
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ }
+
+ cmd = fdmi_req->dma_buf.vaddr;
+ if (ntohs(csio_ct_rsp(cmd)) != FC_FS_ACC) {
+ csio_ln_dbg(ln, "fdmi rpa cmd rejected reason %x expl %x\n",
+ csio_ct_reason(cmd), csio_ct_expl(cmd));
+ }
+}
+
+/*
+ * csio_ln_fdmi_rhba_cbfn - RHBA completion
+ * @hw: HW context
+ * @fdmi_req: fdmi request
+ */
+static void
+csio_ln_fdmi_rhba_cbfn(struct csio_hw *hw, struct csio_ioreq *fdmi_req)
+{
+ void *cmd;
+ uint8_t *pld;
+ uint32_t len = 0;
+ struct csio_lnode *ln = fdmi_req->lnode;
+ struct fs_fdmi_attrs *attrib_blk;
+ struct fc_fdmi_port_name *port_name;
+ uint8_t buf[64];
+ uint32_t val;
+ uint8_t *fc4_type;
+
+ if (fdmi_req->wr_status != FW_SUCCESS) {
+ csio_ln_dbg(ln, "WR error:%x in processing fdmi rhba cmd\n",
+ fdmi_req->wr_status);
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ }
+
+ cmd = fdmi_req->dma_buf.vaddr;
+ if (ntohs(csio_ct_rsp(cmd)) != FC_FS_ACC) {
+ csio_ln_dbg(ln, "fdmi rhba cmd rejected reason %x expl %x\n",
+ csio_ct_reason(cmd), csio_ct_expl(cmd));
+ }
+
+ if (!csio_is_rnode_ready(fdmi_req->rnode)) {
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ return;
+ }
+
+ /* Prepare CT hdr for RPA cmd */
+ memset(cmd, 0, FC_CT_HDR_LEN);
+ csio_fill_ct_iu(cmd, FC_FST_MGMT, FC_FDMI_SUBTYPE, htons(FC_FDMI_RPA));
+
+ /* Prepare RPA payload */
+ pld = (uint8_t *)csio_ct_get_pld(cmd);
+ port_name = (struct fc_fdmi_port_name *)pld;
+ memcpy(&port_name->portname, csio_ln_wwpn(ln), 8);
+ pld += sizeof(*port_name);
+
+ /* Start appending Port attributes */
+ attrib_blk = (struct fs_fdmi_attrs *)pld;
+ attrib_blk->numattrs = 0;
+ len += sizeof(attrib_blk->numattrs);
+ pld += sizeof(attrib_blk->numattrs);
+
+ fc4_type = &buf[0];
+ memset(fc4_type, 0, FC_FDMI_PORT_ATTR_FC4TYPES_LEN);
+ fc4_type[2] = 1;
+ fc4_type[7] = 1;
+ csio_append_attrib(&pld, FC_FDMI_PORT_ATTR_FC4TYPES,
+ fc4_type, FC_FDMI_PORT_ATTR_FC4TYPES_LEN);
+ attrib_blk->numattrs++;
+ val = htonl(FC_PORTSPEED_1GBIT | FC_PORTSPEED_10GBIT);
+ csio_append_attrib(&pld, FC_FDMI_PORT_ATTR_SUPPORTEDSPEED,
+ (uint8_t *)&val,
+ FC_FDMI_PORT_ATTR_SUPPORTEDSPEED_LEN);
+ attrib_blk->numattrs++;
+
+ if (hw->pport[ln->portid].link_speed == FW_PORT_CAP_SPEED_1G)
+ val = htonl(FC_PORTSPEED_1GBIT);
+ else if (hw->pport[ln->portid].link_speed == FW_PORT_CAP_SPEED_10G)
+ val = htonl(FC_PORTSPEED_10GBIT);
+ else
+ val = htonl(CSIO_HBA_PORTSPEED_UNKNOWN);
+ csio_append_attrib(&pld, FC_FDMI_PORT_ATTR_CURRENTPORTSPEED,
+ (uint8_t *)&val,
+ FC_FDMI_PORT_ATTR_CURRENTPORTSPEED_LEN);
+ attrib_blk->numattrs++;
+
+ val = htonl(ln->ln_sparm.csp.sp_bb_data);
+ csio_append_attrib(&pld, FC_FDMI_PORT_ATTR_MAXFRAMESIZE,
+ (uint8_t *)&val, FC_FDMI_PORT_ATTR_MAXFRAMESIZE_LEN);
+ attrib_blk->numattrs++;
+
+ strcpy(buf, "csiostor");
+ csio_append_attrib(&pld, FC_FDMI_PORT_ATTR_OSDEVICENAME, buf,
+ (uint16_t)strlen(buf));
+ attrib_blk->numattrs++;
+
+ if (!csio_hostname(buf, sizeof(buf))) {
+ csio_append_attrib(&pld, FC_FDMI_PORT_ATTR_HOSTNAME,
+ buf, (uint16_t)strlen(buf));
+ attrib_blk->numattrs++;
+ }
+ attrib_blk->numattrs = ntohl(attrib_blk->numattrs);
+ len = (uint32_t)(pld - (uint8_t *)cmd);
+
+ /* Submit FDMI RPA request */
+ spin_lock_irq(&hw->lock);
+ if (csio_ln_mgmt_submit_req(fdmi_req, csio_ln_fdmi_done,
+ FCOE_CT, &fdmi_req->dma_buf, len)) {
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ csio_ln_dbg(ln, "Failed to issue fdmi rpa req\n");
+ }
+ spin_unlock_irq(&hw->lock);
+}
+
+/*
+ * csio_ln_fdmi_dprt_cbfn - DPRT completion
+ * @hw: HW context
+ * @fdmi_req: fdmi request
+ */
+static void
+csio_ln_fdmi_dprt_cbfn(struct csio_hw *hw, struct csio_ioreq *fdmi_req)
+{
+ void *cmd;
+ uint8_t *pld;
+ uint32_t len = 0;
+ uint32_t maxpayload = htonl(65536);
+ struct fc_fdmi_hba_identifier *hbaid;
+ struct csio_lnode *ln = fdmi_req->lnode;
+ struct fc_fdmi_rpl *reg_pl;
+ struct fs_fdmi_attrs *attrib_blk;
+ uint8_t buf[64];
+
+ if (fdmi_req->wr_status != FW_SUCCESS) {
+ csio_ln_dbg(ln, "WR error:%x in processing fdmi dprt cmd\n",
+ fdmi_req->wr_status);
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ }
+
+ if (!csio_is_rnode_ready(fdmi_req->rnode)) {
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ return;
+ }
+ cmd = fdmi_req->dma_buf.vaddr;
+ if (ntohs(csio_ct_rsp(cmd)) != FC_FS_ACC) {
+ csio_ln_dbg(ln, "fdmi dprt cmd rejected reason %x expl %x\n",
+ csio_ct_reason(cmd), csio_ct_expl(cmd));
+ }
+
+ /* Prepare CT hdr for RHBA cmd */
+ memset(cmd, 0, FC_CT_HDR_LEN);
+ csio_fill_ct_iu(cmd, FC_FST_MGMT, FC_FDMI_SUBTYPE, htons(FC_FDMI_RHBA));
+ len = FC_CT_HDR_LEN;
+
+ /* Prepare RHBA payload */
+ pld = (uint8_t *)csio_ct_get_pld(cmd);
+ hbaid = (struct fc_fdmi_hba_identifier *)pld;
+ memcpy(&hbaid->id, csio_ln_wwpn(ln), 8); /* HBA identifer */
+ pld += sizeof(*hbaid);
+
+ /* Register one port per hba */
+ reg_pl = (struct fc_fdmi_rpl *)pld;
+ reg_pl->numport = ntohl(1);
+ memcpy(®_pl->port[0].portname, csio_ln_wwpn(ln), 8);
+ pld += sizeof(*reg_pl);
+
+ /* Start appending HBA attributes hba */
+ attrib_blk = (struct fs_fdmi_attrs *)pld;
+ attrib_blk->numattrs = 0;
+ len += sizeof(attrib_blk->numattrs);
+ pld += sizeof(attrib_blk->numattrs);
+
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_NODENAME, csio_ln_wwnn(ln),
+ FC_FDMI_HBA_ATTR_NODENAME_LEN);
+ attrib_blk->numattrs++;
+
+ memset(buf, 0, sizeof(buf));
+
+ strcpy(buf, "Chelsio Communications");
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_MANUFACTURER, buf,
+ (uint16_t)strlen(buf));
+ attrib_blk->numattrs++;
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_SERIALNUMBER,
+ hw->vpd.sn, (uint16_t)sizeof(hw->vpd.sn));
+ attrib_blk->numattrs++;
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_MODEL, hw->vpd.id,
+ (uint16_t)sizeof(hw->vpd.id));
+ attrib_blk->numattrs++;
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_MODELDESCRIPTION,
+ hw->model_desc, (uint16_t)strlen(hw->model_desc));
+ attrib_blk->numattrs++;
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_HARDWAREVERSION,
+ hw->hw_ver, (uint16_t)sizeof(hw->hw_ver));
+ attrib_blk->numattrs++;
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_FIRMWAREVERSION,
+ hw->fwrev_str, (uint16_t)strlen(hw->fwrev_str));
+ attrib_blk->numattrs++;
+
+ if (!csio_osname(buf, sizeof(buf))) {
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_OSNAMEVERSION,
+ buf, (uint16_t)strlen(buf));
+ attrib_blk->numattrs++;
+ }
+
+ csio_append_attrib(&pld, FC_FDMI_HBA_ATTR_MAXCTPAYLOAD,
+ (uint8_t *)&maxpayload,
+ FC_FDMI_HBA_ATTR_MAXCTPAYLOAD_LEN);
+ len = (uint32_t)(pld - (uint8_t *)cmd);
+ attrib_blk->numattrs++;
+ attrib_blk->numattrs = ntohl(attrib_blk->numattrs);
+
+ /* Submit FDMI RHBA request */
+ spin_lock_irq(&hw->lock);
+ if (csio_ln_mgmt_submit_req(fdmi_req, csio_ln_fdmi_rhba_cbfn,
+ FCOE_CT, &fdmi_req->dma_buf, len)) {
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ csio_ln_dbg(ln, "Failed to issue fdmi rhba req\n");
+ }
+ spin_unlock_irq(&hw->lock);
+}
+
+/*
+ * csio_ln_fdmi_dhba_cbfn - DHBA completion
+ * @hw: HW context
+ * @fdmi_req: fdmi request
+ */
+static void
+csio_ln_fdmi_dhba_cbfn(struct csio_hw *hw, struct csio_ioreq *fdmi_req)
+{
+ struct csio_lnode *ln = fdmi_req->lnode;
+ void *cmd;
+ struct fc_fdmi_port_name *port_name;
+ uint32_t len;
+
+ if (fdmi_req->wr_status != FW_SUCCESS) {
+ csio_ln_dbg(ln, "WR error:%x in processing fdmi dhba cmd\n",
+ fdmi_req->wr_status);
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ }
+
+ if (!csio_is_rnode_ready(fdmi_req->rnode)) {
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ return;
+ }
+ cmd = fdmi_req->dma_buf.vaddr;
+ if (ntohs(csio_ct_rsp(cmd)) != FC_FS_ACC) {
+ csio_ln_dbg(ln, "fdmi dhba cmd rejected reason %x expl %x\n",
+ csio_ct_reason(cmd), csio_ct_expl(cmd));
+ }
+
+ /* Send FDMI cmd to de-register any Port attributes if registered
+ * before
+ */
+
+ /* Prepare FDMI DPRT cmd */
+ memset(cmd, 0, FC_CT_HDR_LEN);
+ csio_fill_ct_iu(cmd, FC_FST_MGMT, FC_FDMI_SUBTYPE, htons(FC_FDMI_DPRT));
+ len = FC_CT_HDR_LEN;
+ port_name = (struct fc_fdmi_port_name *)csio_ct_get_pld(cmd);
+ memcpy(&port_name->portname, csio_ln_wwpn(ln), 8);
+ len += sizeof(*port_name);
+
+ /* Submit FDMI request */
+ spin_lock_irq(&hw->lock);
+ if (csio_ln_mgmt_submit_req(fdmi_req, csio_ln_fdmi_dprt_cbfn,
+ FCOE_CT, &fdmi_req->dma_buf, len)) {
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ csio_ln_dbg(ln, "Failed to issue fdmi dprt req\n");
+ }
+ spin_unlock_irq(&hw->lock);
+}
+
+/**
+ * csio_ln_fdmi_start - Start an FDMI request.
+ * @ln: lnode
+ * @context: session context
+ *
+ * Issued with lock held.
+ */
+int
+csio_ln_fdmi_start(struct csio_lnode *ln, void *context)
+{
+ struct csio_ioreq *fdmi_req;
+ struct csio_rnode *fdmi_rn = (struct csio_rnode *)context;
+ void *cmd;
+ struct fc_fdmi_hba_identifier *hbaid;
+ uint32_t len;
+
+ if (!(ln->flags & CSIO_LNF_FDMI_ENABLE))
+ return -EPROTONOSUPPORT;
+
+ if (!csio_is_rnode_ready(fdmi_rn))
+ CSIO_INC_STATS(ln, n_fdmi_err);
+
+ /* Send FDMI cmd to de-register any HBA attributes if registered
+ * before
+ */
+
+ fdmi_req = ln->mgmt_req;
+ fdmi_req->lnode = ln;
+ fdmi_req->rnode = fdmi_rn;
+
+ /* Prepare FDMI DHBA cmd */
+ cmd = fdmi_req->dma_buf.vaddr;
+ memset(cmd, 0, FC_CT_HDR_LEN);
+ csio_fill_ct_iu(cmd, FC_FST_MGMT, FC_FDMI_SUBTYPE, htons(FC_FDMI_DHBA));
+ len = FC_CT_HDR_LEN;
+
+ hbaid = (struct fc_fdmi_hba_identifier *)csio_ct_get_pld(cmd);
+ memcpy(&hbaid->id, csio_ln_wwpn(ln), 8);
+ len += sizeof(*hbaid);
+
+ /* Submit FDMI request */
+ if (csio_ln_mgmt_submit_req(fdmi_req, csio_ln_fdmi_dhba_cbfn,
+ FCOE_CT, &fdmi_req->dma_buf, len)) {
+ CSIO_INC_STATS(ln, n_fdmi_err);
+ csio_ln_dbg(ln, "Failed to issue fdmi dhba req\n");
+ }
+
+ return 0;
+}
+
+/*
+ * csio_ln_vnp_read_cbfn - vnp read completion handler.
+ * @hw: HW lnode
+ * @cbfn: Completion handler.
+ *
+ * Reads vnp response and updates ln parameters.
+ */
+static void
+csio_ln_vnp_read_cbfn(struct csio_hw *hw, struct csio_mb *mbp)
+{
+ struct csio_lnode *ln = ((struct csio_lnode *)mbp->priv);
+ struct fw_fcoe_vnp_cmd *rsp = (struct fw_fcoe_vnp_cmd *)(mbp->mb);
+ struct fc_els_csp *csp;
+ struct fc_els_cssp *clsp;
+ enum fw_retval retval;
+
+ retval = FW_CMD_RETVAL_GET(ntohl(rsp->alloc_to_len16));
+ if (retval != FW_SUCCESS) {
+ csio_err(hw, "FCOE VNP read cmd returned error:0x%x\n", retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return;
+ }
+
+ spin_lock_irq(&hw->lock);
+
+ memcpy(ln->mac, rsp->vnport_mac, sizeof(ln->mac));
+ memcpy(&ln->nport_id, &rsp->vnport_mac[3],
+ sizeof(uint8_t)*3);
+ ln->nport_id = ntohl(ln->nport_id);
+ ln->nport_id = ln->nport_id>>8;
+
+ /* Update WWNs */
+ /*
+ * This may look like a duplication of what csio_fcoe_enable_link()
+ * does, but is absolutely necessary if the vnpi changes between
+ * a FCOE LINK UP and FCOE LINK DOWN.
+ */
+ memcpy(csio_ln_wwnn(ln), rsp->vnport_wwnn, 8);
+ memcpy(csio_ln_wwpn(ln), rsp->vnport_wwpn, 8);
+
+ /* Copy common sparam */
+ csp = (struct fc_els_csp *)rsp->cmn_srv_parms;
+ ln->ln_sparm.csp.sp_hi_ver = csp->sp_hi_ver;
+ ln->ln_sparm.csp.sp_lo_ver = csp->sp_lo_ver;
+ ln->ln_sparm.csp.sp_bb_cred = ntohs(csp->sp_bb_cred);
+ ln->ln_sparm.csp.sp_features = ntohs(csp->sp_features);
+ ln->ln_sparm.csp.sp_bb_data = ntohs(csp->sp_bb_data);
+ ln->ln_sparm.csp.sp_r_a_tov = ntohl(csp->sp_r_a_tov);
+ ln->ln_sparm.csp.sp_e_d_tov = ntohl(csp->sp_e_d_tov);
+
+ /* Copy word 0 & word 1 of class sparam */
+ clsp = (struct fc_els_cssp *)rsp->clsp_word_0_1;
+ ln->ln_sparm.clsp[2].cp_class = ntohs(clsp->cp_class);
+ ln->ln_sparm.clsp[2].cp_init = ntohs(clsp->cp_init);
+ ln->ln_sparm.clsp[2].cp_recip = ntohs(clsp->cp_recip);
+ ln->ln_sparm.clsp[2].cp_rdfs = ntohs(clsp->cp_rdfs);
+
+ spin_unlock_irq(&hw->lock);
+
+ mempool_free(mbp, hw->mb_mempool);
+
+ /* Send an event to update local attribs */
+ csio_lnode_async_event(ln, CSIO_LN_FC_ATTRIB_UPDATE);
+}
+
+/*
+ * csio_ln_vnp_read - Read vnp params.
+ * @ln: lnode
+ * @cbfn: Completion handler.
+ *
+ * Issued with lock held.
+ */
+static int
+csio_ln_vnp_read(struct csio_lnode *ln,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct csio_hw *hw = ln->hwp;
+ struct csio_mb *mbp;
+
+ /* Allocate Mbox request */
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ /* Prepare VNP Command */
+ csio_fcoe_vnp_read_init_mb(ln, mbp,
+ CSIO_MB_DEFAULT_TMO,
+ ln->fcf_flowid,
+ ln->vnp_flowid,
+ cbfn);
+
+ /* Issue MBOX cmd */
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Failed to issue mbox FCoE VNP command\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/*
+ * csio_fcoe_enable_link - Enable fcoe link.
+ * @ln: lnode
+ * @enable: enable/disable
+ * Issued with lock held.
+ * Issues mbox cmd to bring up FCOE link on port associated with given ln.
+ */
+static int
+csio_fcoe_enable_link(struct csio_lnode *ln, bool enable)
+{
+ struct csio_hw *hw = ln->hwp;
+ struct csio_mb *mbp;
+ enum fw_retval retval;
+ uint8_t portid;
+ uint8_t sub_op;
+ struct fw_fcoe_link_cmd *lcmd;
+ int i;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ portid = ln->portid;
+ sub_op = enable ? FCOE_LINK_UP : FCOE_LINK_DOWN;
+
+ csio_dbg(hw, "bringing FCOE LINK %s on Port:%d\n",
+ sub_op ? "UP" : "DOWN", portid);
+
+ csio_write_fcoe_link_cond_init_mb(ln, mbp, CSIO_MB_DEFAULT_TMO,
+ portid, sub_op, 0, 0, 0, NULL);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "failed to issue FCOE LINK cmd on port[%d]\n",
+ portid);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ retval = csio_mb_fw_retval(mbp);
+ if (retval != FW_SUCCESS) {
+ csio_err(hw,
+ "FCOE LINK %s cmd on port[%d] failed with "
+ "ret:x%x\n", sub_op ? "UP" : "DOWN", portid, retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ if (!enable)
+ goto out;
+
+ lcmd = (struct fw_fcoe_link_cmd *)mbp->mb;
+
+ memcpy(csio_ln_wwnn(ln), lcmd->vnport_wwnn, 8);
+ memcpy(csio_ln_wwpn(ln), lcmd->vnport_wwpn, 8);
+
+ for (i = 0; i < CSIO_MAX_PPORTS; i++)
+ if (hw->pport[i].portid == portid)
+ memcpy(hw->pport[i].mac, lcmd->phy_mac, 6);
+
+out:
+ mempool_free(mbp, hw->mb_mempool);
+ return 0;
+}
+
+/*
+ * csio_ln_read_fcf_cbfn - Read fcf parameters
+ * @ln: lnode
+ *
+ * read fcf response and Update ln fcf information.
+ */
+static void
+csio_ln_read_fcf_cbfn(struct csio_hw *hw, struct csio_mb *mbp)
+{
+ struct csio_lnode *ln = (struct csio_lnode *)mbp->priv;
+ struct csio_fcf_info *fcf_info;
+ struct fw_fcoe_fcf_cmd *rsp =
+ (struct fw_fcoe_fcf_cmd *)(mbp->mb);
+ enum fw_retval retval;
+
+ retval = FW_CMD_RETVAL_GET(ntohl(rsp->retval_len16));
+ if (retval != FW_SUCCESS) {
+ csio_ln_err(ln, "FCOE FCF cmd failed with ret x%x\n",
+ retval);
+ mempool_free(mbp, hw->mb_mempool);
+ return;
+ }
+
+ spin_lock_irq(&hw->lock);
+ fcf_info = ln->fcfinfo;
+ fcf_info->priority = FW_FCOE_FCF_CMD_PRIORITY_GET(
+ ntohs(rsp->priority_pkd));
+ fcf_info->vf_id = ntohs(rsp->vf_id);
+ fcf_info->vlan_id = rsp->vlan_id;
+ fcf_info->max_fcoe_size = ntohs(rsp->max_fcoe_size);
+ fcf_info->fka_adv = be32_to_cpu(rsp->fka_adv);
+ fcf_info->fcfi = FW_FCOE_FCF_CMD_FCFI_GET(ntohl(rsp->op_to_fcfi));
+ fcf_info->fpma = FW_FCOE_FCF_CMD_FPMA_GET(rsp->fpma_to_portid);
+ fcf_info->spma = FW_FCOE_FCF_CMD_SPMA_GET(rsp->fpma_to_portid);
+ fcf_info->login = FW_FCOE_FCF_CMD_LOGIN_GET(rsp->fpma_to_portid);
+ fcf_info->portid = FW_FCOE_FCF_CMD_PORTID_GET(rsp->fpma_to_portid);
+ memcpy(fcf_info->fc_map, rsp->fc_map, sizeof(fcf_info->fc_map));
+ memcpy(fcf_info->mac, rsp->mac, sizeof(fcf_info->mac));
+ memcpy(fcf_info->name_id, rsp->name_id, sizeof(fcf_info->name_id));
+ memcpy(fcf_info->fabric, rsp->fabric, sizeof(fcf_info->fabric));
+ memcpy(fcf_info->spma_mac, rsp->spma_mac, sizeof(fcf_info->spma_mac));
+
+ spin_unlock_irq(&hw->lock);
+
+ mempool_free(mbp, hw->mb_mempool);
+}
+
+/*
+ * csio_ln_read_fcf_entry - Read fcf entry.
+ * @ln: lnode
+ * @cbfn: Completion handler.
+ *
+ * Issued with lock held.
+ */
+static int
+csio_ln_read_fcf_entry(struct csio_lnode *ln,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct csio_hw *hw = ln->hwp;
+ struct csio_mb *mbp;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ /* Get FCoE FCF information */
+ csio_fcoe_read_fcf_init_mb(ln, mbp, CSIO_MB_DEFAULT_TMO,
+ ln->portid, ln->fcf_flowid, cbfn);
+
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "failed to issue FCOE FCF cmd\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/*
+ * csio_handle_link_up - Logical Linkup event.
+ * @hw - HW module.
+ * @portid - Physical port number
+ * @fcfi - FCF index.
+ * @vnpi - VNP index.
+ * Returns - none.
+ *
+ * This event is received from FW, when virtual link is established between
+ * Physical port[ENode] and FCF. If its new vnpi, then local node object is
+ * created on this FCF and set to [ONLINE] state.
+ * Lnode waits for FW_RDEV_CMD event to be received indicating that
+ * Fabric login is completed and lnode moves to [READY] state.
+ *
+ * This called with hw lock held
+ */
+static void
+csio_handle_link_up(struct csio_hw *hw, uint8_t portid, uint32_t fcfi,
+ uint32_t vnpi)
+{
+ struct csio_lnode *ln = NULL;
+
+ /* Lookup lnode based on vnpi */
+ ln = csio_ln_lookup_by_vnpi(hw, vnpi);
+ if (!ln) {
+ /* Pick lnode based on portid */
+ ln = csio_ln_lookup_by_portid(hw, portid);
+ if (!ln) {
+ csio_err(hw, "failed to lookup fcoe lnode on port:%d\n",
+ portid);
+ CSIO_DB_ASSERT(0);
+ return;
+ }
+
+ /* Check if lnode has valid vnp flowid */
+ if (ln->vnp_flowid != CSIO_INVALID_IDX) {
+ /* New VN-Port */
+ spin_unlock_irq(&hw->lock);
+ csio_lnode_alloc(hw);
+ spin_lock_irq(&hw->lock);
+ if (!ln) {
+ csio_err(hw,
+ "failed to allocate fcoe lnode"
+ "for port:%d vnpi:x%x\n",
+ portid, vnpi);
+ CSIO_DB_ASSERT(0);
+ return;
+ }
+ ln->portid = portid;
+ }
+ ln->vnp_flowid = vnpi;
+ ln->dev_num &= ~0xFFFF;
+ ln->dev_num |= vnpi;
+ }
+
+ /*Initialize fcfi */
+ ln->fcf_flowid = fcfi;
+
+ csio_info(hw, "Port:%d - FCOE LINK UP\n", portid);
+
+ CSIO_INC_STATS(ln, n_link_up);
+
+ /* Send LINKUP event to SM */
+ csio_post_event(&ln->sm, CSIO_LNE_LINKUP);
+}
+
+/*
+ * csio_post_event_rns
+ * @ln - FCOE lnode
+ * @evt - Given rnode event
+ * Returns - none
+ *
+ * Posts given rnode event to all FCOE rnodes connected with given Lnode.
+ * This routine is invoked when lnode receives LINK_DOWN/DOWN_LINK/CLOSE
+ * event.
+ *
+ * This called with hw lock held
+ */
+static void
+csio_post_event_rns(struct csio_lnode *ln, enum csio_rn_ev evt)
+{
+ struct csio_rnode *rnhead = (struct csio_rnode *) &ln->rnhead;
+ struct list_head *tmp, *next;
+ struct csio_rnode *rn;
+
+ list_for_each_safe(tmp, next, &rnhead->sm.sm_list) {
+ rn = (struct csio_rnode *) tmp;
+ csio_post_event(&rn->sm, evt);
+ }
+}
+
+/*
+ * csio_cleanup_rns
+ * @ln - FCOE lnode
+ * Returns - none
+ *
+ * Frees all FCOE rnodes connected with given Lnode.
+ *
+ * This called with hw lock held
+ */
+static void
+csio_cleanup_rns(struct csio_lnode *ln)
+{
+ struct csio_rnode *rnhead = (struct csio_rnode *) &ln->rnhead;
+ struct list_head *tmp, *next_rn;
+ struct csio_rnode *rn;
+
+ list_for_each_safe(tmp, next_rn, &rnhead->sm.sm_list) {
+ rn = (struct csio_rnode *) tmp;
+ csio_put_rnode(ln, rn);
+ }
+
+}
+
+/*
+ * csio_post_event_lns
+ * @ln - FCOE lnode
+ * @evt - Given lnode event
+ * Returns - none
+ *
+ * Posts given lnode event to all FCOE lnodes connected with given Lnode.
+ * This routine is invoked when lnode receives LINK_DOWN/DOWN_LINK/CLOSE
+ * event.
+ *
+ * This called with hw lock held
+ */
+static void
+csio_post_event_lns(struct csio_lnode *ln, enum csio_ln_ev evt)
+{
+ struct list_head *tmp;
+ struct csio_lnode *cln, *sln;
+
+ /* If NPIV lnode, send evt only to that and return */
+ if (csio_is_npiv_ln(ln)) {
+ csio_post_event(&ln->sm, evt);
+ return;
+ }
+
+ sln = ln;
+ /* Traverse children lnodes list and send evt */
+ list_for_each(tmp, &sln->cln_head) {
+ cln = (struct csio_lnode *) tmp;
+ csio_post_event(&cln->sm, evt);
+ }
+
+ /* Send evt to parent lnode */
+ csio_post_event(&ln->sm, evt);
+}
+
+/*
+ * csio_ln_down - Lcoal nport is down
+ * @ln - FCOE Lnode
+ * Returns - none
+ *
+ * Sends LINK_DOWN events to Lnode and its associated NPIVs lnodes.
+ *
+ * This called with hw lock held
+ */
+static void
+csio_ln_down(struct csio_lnode *ln)
+{
+ csio_post_event_lns(ln, CSIO_LNE_LINK_DOWN);
+}
+
+/*
+ * csio_handle_link_down - Logical Linkdown event.
+ * @hw - HW module.
+ * @portid - Physical port number
+ * @fcfi - FCF index.
+ * @vnpi - VNP index.
+ * Returns - none
+ *
+ * This event is received from FW, when virtual link goes down between
+ * Physical port[ENode] and FCF. Lnode and its associated NPIVs lnode hosted on
+ * this vnpi[VN-Port] will be de-instantiated.
+ *
+ * This called with hw lock held
+ */
+static void
+csio_handle_link_down(struct csio_hw *hw, uint8_t portid, uint32_t fcfi,
+ uint32_t vnpi)
+{
+ struct csio_fcf_info *fp;
+ struct csio_lnode *ln;
+
+ /* Lookup lnode based on vnpi */
+ ln = csio_ln_lookup_by_vnpi(hw, vnpi);
+ if (ln) {
+ fp = ln->fcfinfo;
+ CSIO_INC_STATS(ln, n_link_down);
+
+ /*Warn if linkdown received if lnode is not in ready state */
+ if (!csio_is_lnode_ready(ln)) {
+ csio_ln_warn(ln,
+ "warn: FCOE link is already in offline "
+ "Ignoring Fcoe linkdown event on portid %d\n",
+ portid);
+ CSIO_INC_STATS(ln, n_evt_drop);
+ return;
+ }
+
+ /* Verify portid */
+ if (fp->portid != portid) {
+ csio_ln_warn(ln,
+ "warn: FCOE linkdown recv with "
+ "invalid port %d\n", portid);
+ CSIO_INC_STATS(ln, n_evt_drop);
+ return;
+ }
+
+ /* verify fcfi */
+ if (ln->fcf_flowid != fcfi) {
+ csio_ln_warn(ln,
+ "warn: FCOE linkdown recv with "
+ "invalid fcfi x%x\n", fcfi);
+ CSIO_INC_STATS(ln, n_evt_drop);
+ return;
+ }
+
+ csio_info(hw, "Port:%d - FCOE LINK DOWN\n", portid);
+
+ /* Send LINK_DOWN event to lnode s/m */
+ csio_ln_down(ln);
+
+ return;
+ } else {
+ csio_warn(hw,
+ "warn: FCOE linkdown recv with invalid vnpi x%x\n",
+ vnpi);
+ CSIO_INC_STATS(hw, n_evt_drop);
+ }
+}
+
+/*
+ * csio_is_lnode_ready - Checks FCOE lnode is in ready state.
+ * @ln: Lnode module
+ *
+ * Returns True if FCOE lnode is in ready state.
+ */
+int
+csio_is_lnode_ready(struct csio_lnode *ln)
+{
+ return (csio_get_state(ln) == ((csio_sm_state_t)csio_lns_ready));
+}
+
+/*****************************************************************************/
+/* START: Lnode SM */
+/*****************************************************************************/
+/*
+ * csio_lns_uninit - The request in uninit state.
+ * @ln - FCOE lnode.
+ * @evt - Event to be processed.
+ *
+ * Process the given lnode event which is currently in "uninit" state.
+ * Invoked with HW lock held.
+ * Return - none.
+ */
+static void
+csio_lns_uninit(struct csio_lnode *ln, enum csio_ln_ev evt)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+ struct csio_lnode *rln = hw->rln;
+ int rv;
+
+ CSIO_INC_STATS(ln, n_evt_sm[evt]);
+ switch (evt) {
+ case CSIO_LNE_LINKUP:
+ csio_set_state(&ln->sm, csio_lns_online);
+ /* Read FCF only for physical lnode */
+ if (csio_is_phys_ln(ln)) {
+ rv = csio_ln_read_fcf_entry(ln,
+ csio_ln_read_fcf_cbfn);
+ if (rv != 0) {
+ /* TODO: Send HW RESET event */
+ CSIO_INC_STATS(ln, n_err);
+ break;
+ }
+
+ /* Add FCF record */
+ list_add_tail(&ln->fcfinfo->list, &rln->fcf_lsthead);
+ }
+
+ rv = csio_ln_vnp_read(ln, csio_ln_vnp_read_cbfn);
+ if (rv != 0) {
+ /* TODO: Send HW RESET event */
+ CSIO_INC_STATS(ln, n_err);
+ }
+ break;
+
+ case CSIO_LNE_DOWN_LINK:
+ break;
+
+ default:
+ csio_ln_dbg(ln,
+ "unexp ln event %d recv from did:x%x in "
+ "ln state[uninit].\n", evt, ln->nport_id);
+ CSIO_INC_STATS(ln, n_evt_unexp);
+ break;
+ } /* switch event */
+}
+
+/*
+ * csio_lns_online - The request in online state.
+ * @ln - FCOE lnode.
+ * @evt - Event to be processed.
+ *
+ * Process the given lnode event which is currently in "online" state.
+ * Invoked with HW lock held.
+ * Return - none.
+ */
+static void
+csio_lns_online(struct csio_lnode *ln, enum csio_ln_ev evt)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ CSIO_INC_STATS(ln, n_evt_sm[evt]);
+ switch (evt) {
+ case CSIO_LNE_LINKUP:
+ csio_ln_warn(ln,
+ "warn: FCOE link is up already "
+ "Ignoring linkup on port:%d\n", ln->portid);
+ CSIO_INC_STATS(ln, n_evt_drop);
+ break;
+
+ case CSIO_LNE_FAB_INIT_DONE:
+ csio_set_state(&ln->sm, csio_lns_ready);
+
+ spin_unlock_irq(&hw->lock);
+ csio_lnode_async_event(ln, CSIO_LN_FC_LINKUP);
+ spin_lock_irq(&hw->lock);
+
+ break;
+
+ case CSIO_LNE_LINK_DOWN:
+ /* Fall through */
+ case CSIO_LNE_DOWN_LINK:
+ csio_set_state(&ln->sm, csio_lns_uninit);
+ if (csio_is_phys_ln(ln)) {
+ /* Remove FCF entry */
+ list_del_init(&ln->fcfinfo->list);
+ }
+ break;
+
+ default:
+ csio_ln_dbg(ln,
+ "unexp ln event %d recv from did:x%x in "
+ "ln state[uninit].\n", evt, ln->nport_id);
+ CSIO_INC_STATS(ln, n_evt_unexp);
+
+ break;
+ } /* switch event */
+}
+
+/*
+ * csio_lns_ready - The request in ready state.
+ * @ln - FCOE lnode.
+ * @evt - Event to be processed.
+ *
+ * Process the given lnode event which is currently in "ready" state.
+ * Invoked with HW lock held.
+ * Return - none.
+ */
+static void
+csio_lns_ready(struct csio_lnode *ln, enum csio_ln_ev evt)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ CSIO_INC_STATS(ln, n_evt_sm[evt]);
+ switch (evt) {
+ case CSIO_LNE_FAB_INIT_DONE:
+ csio_ln_dbg(ln,
+ "ignoring event %d recv from did x%x"
+ "in ln state[ready].\n", evt, ln->nport_id);
+ CSIO_INC_STATS(ln, n_evt_drop);
+ break;
+
+ case CSIO_LNE_LINK_DOWN:
+ csio_set_state(&ln->sm, csio_lns_offline);
+ csio_post_event_rns(ln, CSIO_RNFE_DOWN);
+
+ spin_unlock_irq(&hw->lock);
+ csio_lnode_async_event(ln, CSIO_LN_FC_LINKDOWN);
+ spin_lock_irq(&hw->lock);
+
+ if (csio_is_phys_ln(ln)) {
+ /* Remove FCF entry */
+ list_del_init(&ln->fcfinfo->list);
+ }
+ break;
+
+ case CSIO_LNE_DOWN_LINK:
+ csio_set_state(&ln->sm, csio_lns_offline);
+ csio_post_event_rns(ln, CSIO_RNFE_DOWN);
+
+ /* Host need to issue aborts in case if FW has not returned
+ * WRs with status "ABORTED"
+ */
+ spin_unlock_irq(&hw->lock);
+ csio_lnode_async_event(ln, CSIO_LN_FC_LINKDOWN);
+ spin_lock_irq(&hw->lock);
+
+ if (csio_is_phys_ln(ln)) {
+ /* Remove FCF entry */
+ list_del_init(&ln->fcfinfo->list);
+ }
+ break;
+
+ case CSIO_LNE_CLOSE:
+ csio_set_state(&ln->sm, csio_lns_uninit);
+ csio_post_event_rns(ln, CSIO_RNFE_CLOSE);
+ break;
+
+ case CSIO_LNE_LOGO:
+ csio_set_state(&ln->sm, csio_lns_offline);
+ csio_post_event_rns(ln, CSIO_RNFE_DOWN);
+ break;
+
+ default:
+ csio_ln_dbg(ln,
+ "unexp ln event %d recv from did:x%x in "
+ "ln state[uninit].\n", evt, ln->nport_id);
+ CSIO_INC_STATS(ln, n_evt_unexp);
+ CSIO_DB_ASSERT(0);
+ break;
+ } /* switch event */
+}
+
+/*
+ * csio_lns_offline - The request in offline state.
+ * @ln - FCOE lnode.
+ * @evt - Event to be processed.
+ *
+ * Process the given lnode event which is currently in "offline" state.
+ * Invoked with HW lock held.
+ * Return - none.
+ */
+static void
+csio_lns_offline(struct csio_lnode *ln, enum csio_ln_ev evt)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+ struct csio_lnode *rln = hw->rln;
+ int rv;
+
+ CSIO_INC_STATS(ln, n_evt_sm[evt]);
+ switch (evt) {
+ case CSIO_LNE_LINKUP:
+ csio_set_state(&ln->sm, csio_lns_online);
+ /* Read FCF only for physical lnode */
+ if (csio_is_phys_ln(ln)) {
+ rv = csio_ln_read_fcf_entry(ln,
+ csio_ln_read_fcf_cbfn);
+ if (rv != 0) {
+ /* TODO: Send HW RESET event */
+ CSIO_INC_STATS(ln, n_err);
+ break;
+ }
+
+ /* Add FCF record */
+ list_add_tail(&ln->fcfinfo->list, &rln->fcf_lsthead);
+ }
+
+ rv = csio_ln_vnp_read(ln, csio_ln_vnp_read_cbfn);
+ if (rv != 0) {
+ /* TODO: Send HW RESET event */
+ CSIO_INC_STATS(ln, n_err);
+ }
+ break;
+
+ case CSIO_LNE_LINK_DOWN:
+ case CSIO_LNE_DOWN_LINK:
+ case CSIO_LNE_LOGO:
+ csio_ln_dbg(ln,
+ "ignoring event %d recv from did x%x"
+ "in ln state[offline].\n", evt, ln->nport_id);
+ CSIO_INC_STATS(ln, n_evt_drop);
+ break;
+
+ case CSIO_LNE_CLOSE:
+ csio_set_state(&ln->sm, csio_lns_uninit);
+ csio_post_event_rns(ln, CSIO_RNFE_CLOSE);
+ break;
+
+ default:
+ csio_ln_dbg(ln,
+ "unexp ln event %d recv from did:x%x in "
+ "ln state[offline]\n", evt, ln->nport_id);
+ CSIO_INC_STATS(ln, n_evt_unexp);
+ CSIO_DB_ASSERT(0);
+ break;
+ } /* switch event */
+}
+
+/*****************************************************************************/
+/* END: Lnode SM */
+/*****************************************************************************/
+
+static void
+csio_free_fcfinfo(struct kref *kref)
+{
+ struct csio_fcf_info *fcfinfo = container_of(kref,
+ struct csio_fcf_info, kref);
+ kfree(fcfinfo);
+}
+
+/* Helper routines for attributes */
+/*
+ * csio_lnode_state_to_str - Get current state of FCOE lnode.
+ * @ln - lnode
+ * @str - state of lnode.
+ *
+ */
+void
+csio_lnode_state_to_str(struct csio_lnode *ln, int8_t *str)
+{
+ if (csio_get_state(ln) == ((csio_sm_state_t)csio_lns_uninit)) {
+ strcpy(str, "UNINIT");
+ return;
+ }
+ if (csio_get_state(ln) == ((csio_sm_state_t)csio_lns_ready)) {
+ strcpy(str, "READY");
+ return;
+ }
+ if (csio_get_state(ln) == ((csio_sm_state_t)csio_lns_offline)) {
+ strcpy(str, "OFFLINE");
+ return;
+ }
+ strcpy(str, "UNKNOWN");
+} /* csio_lnode_state_to_str */
+
+
+int
+csio_get_phy_port_stats(struct csio_hw *hw, uint8_t portid,
+ struct fw_fcoe_port_stats *port_stats)
+{
+ struct csio_mb *mbp;
+ struct fw_fcoe_port_cmd_params portparams;
+ enum fw_retval retval;
+ int idx;
+
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ csio_err(hw, "FCoE FCF PARAMS command out of memory!\n");
+ return -EINVAL;
+ }
+ portparams.portid = portid;
+
+ for (idx = 1; idx <= 3; idx++) {
+ portparams.idx = (idx-1)*6 + 1;
+ portparams.nstats = 6;
+ if (idx == 3)
+ portparams.nstats = 4;
+ csio_fcoe_read_portparams_init_mb(hw, mbp, CSIO_MB_DEFAULT_TMO,
+ &portparams, NULL);
+ if (csio_mb_issue(hw, mbp)) {
+ csio_err(hw, "Issue of FCoE port params failed!\n");
+ mempool_free(mbp, hw->mb_mempool);
+ return -EINVAL;
+ }
+ csio_mb_process_portparams_rsp(hw, mbp, &retval,
+ &portparams, port_stats);
+ }
+
+ mempool_free(mbp, hw->mb_mempool);
+ return 0;
+}
+
+/*
+ * csio_ln_mgmt_wr_handler -Mgmt Work Request handler.
+ * @wr - WR.
+ * @len - WR len.
+ * This handler is invoked when an outstanding mgmt WR is completed.
+ * Its invoked in the context of FW event worker thread for every
+ * mgmt event received.
+ * Return - none.
+ */
+
+static void
+csio_ln_mgmt_wr_handler(struct csio_hw *hw, void *wr, uint32_t len)
+{
+ struct csio_mgmtm *mgmtm = csio_hw_to_mgmtm(hw);
+ struct csio_ioreq *io_req = NULL;
+ struct fw_fcoe_els_ct_wr *wr_cmd;
+
+
+ wr_cmd = (struct fw_fcoe_els_ct_wr *) wr;
+
+ if (len < sizeof(struct fw_fcoe_els_ct_wr)) {
+ csio_err(mgmtm->hw,
+ "Invalid ELS CT WR length recvd, len:%x\n", len);
+ mgmtm->stats.n_err++;
+ return;
+ }
+
+ io_req = (struct csio_ioreq *) ((uintptr_t) wr_cmd->cookie);
+ io_req->wr_status = csio_wr_status(wr_cmd);
+
+ /* lookup ioreq exists in our active Q */
+ spin_lock_irq(&hw->lock);
+ if (csio_mgmt_req_lookup(mgmtm, io_req) != 0) {
+ csio_err(mgmtm->hw,
+ "Error- Invalid IO handle recv in WR. handle: %p\n",
+ io_req);
+ mgmtm->stats.n_err++;
+ spin_unlock_irq(&hw->lock);
+ return;
+ }
+
+ mgmtm = csio_hw_to_mgmtm(hw);
+
+ /* Dequeue from active queue */
+ list_del_init(&io_req->sm.sm_list);
+ mgmtm->stats.n_active--;
+ spin_unlock_irq(&hw->lock);
+
+ /* io_req will be freed by completion handler */
+ if (io_req->io_cbfn)
+ io_req->io_cbfn(hw, io_req);
+}
+
+/**
+ * csio_fcoe_fwevt_handler - Event handler for Firmware FCoE events.
+ * @hw: HW module
+ * @cpl_op: CPL opcode
+ * @cmd: FW cmd/WR.
+ *
+ * Process received FCoE cmd/WR event from FW.
+ */
+void
+csio_fcoe_fwevt_handler(struct csio_hw *hw, __u8 cpl_op, __be64 *cmd)
+{
+ struct csio_lnode *ln;
+ struct csio_rnode *rn;
+ uint8_t portid, opcode = *(uint8_t *)cmd;
+ struct fw_fcoe_link_cmd *lcmd;
+ struct fw_wr_hdr *wr;
+ struct fw_rdev_wr *rdev_wr;
+ enum fw_fcoe_link_status lstatus;
+ uint32_t fcfi, rdev_flowid, vnpi;
+ enum csio_ln_ev evt;
+
+ if (cpl_op == CPL_FW6_MSG && opcode == FW_FCOE_LINK_CMD) {
+
+ lcmd = (struct fw_fcoe_link_cmd *)cmd;
+ lstatus = lcmd->lstatus;
+ portid = FW_FCOE_LINK_CMD_PORTID_GET(
+ ntohl(lcmd->op_to_portid));
+ fcfi = FW_FCOE_LINK_CMD_FCFI_GET(ntohl(lcmd->sub_opcode_fcfi));
+ vnpi = FW_FCOE_LINK_CMD_VNPI_GET(ntohl(lcmd->vnpi_pkd));
+
+ if (lstatus == FCOE_LINKUP) {
+
+ /* HW lock here */
+ spin_lock_irq(&hw->lock);
+ csio_handle_link_up(hw, portid, fcfi, vnpi);
+ spin_unlock_irq(&hw->lock);
+ /* HW un lock here */
+
+ } else if (lstatus == FCOE_LINKDOWN) {
+
+ /* HW lock here */
+ spin_lock_irq(&hw->lock);
+ csio_handle_link_down(hw, portid, fcfi, vnpi);
+ spin_unlock_irq(&hw->lock);
+ /* HW un lock here */
+ } else {
+ csio_warn(hw, "Unexpected FCOE LINK status:0x%x\n",
+ ntohl(lcmd->lstatus));
+ CSIO_INC_STATS(hw, n_cpl_unexp);
+ }
+ } else if (cpl_op == CPL_FW6_PLD) {
+ wr = (struct fw_wr_hdr *) (cmd + 4);
+ if (FW_WR_OP_GET(be32_to_cpu(wr->hi))
+ == FW_RDEV_WR) {
+
+ rdev_wr = (struct fw_rdev_wr *) (cmd + 4);
+
+ rdev_flowid = FW_RDEV_WR_FLOWID_GET(
+ ntohl(rdev_wr->alloc_to_len16));
+ vnpi = FW_RDEV_WR_ASSOC_FLOWID_GET(
+ ntohl(rdev_wr->flags_to_assoc_flowid));
+
+ csio_dbg(hw,
+ "FW_RDEV_WR: flowid:x%x ev_cause:x%x "
+ "vnpi:0x%x\n", rdev_flowid,
+ rdev_wr->event_cause, vnpi);
+
+ if (rdev_wr->protocol != PROT_FCOE) {
+ csio_err(hw,
+ "FW_RDEV_WR: invalid proto:x%x "
+ "received with flowid:x%x\n",
+ rdev_wr->protocol,
+ rdev_flowid);
+ CSIO_INC_STATS(hw, n_evt_drop);
+ return;
+ }
+
+ /* HW lock here */
+ spin_lock_irq(&hw->lock);
+ ln = csio_ln_lookup_by_vnpi(hw, vnpi);
+ if (!ln) {
+ csio_err(hw,
+ "FW_DEV_WR: invalid vnpi:x%x received "
+ "with flowid:x%x\n", vnpi, rdev_flowid);
+ CSIO_INC_STATS(hw, n_evt_drop);
+ goto out_pld;
+ }
+
+ rn = csio_confirm_rnode(ln, rdev_flowid,
+ &rdev_wr->u.fcoe_rdev);
+ if (!rn) {
+ csio_ln_dbg(ln,
+ "Failed to confirm rnode "
+ "for flowid:x%x\n", rdev_flowid);
+ CSIO_INC_STATS(hw, n_evt_drop);
+ goto out_pld;
+ }
+
+ /* save previous event for debugging */
+ ln->prev_evt = ln->cur_evt;
+ ln->cur_evt = rdev_wr->event_cause;
+ CSIO_INC_STATS(ln, n_evt_fw[rdev_wr->event_cause]);
+
+ /* Translate all the fabric events to lnode SM events */
+ evt = CSIO_FWE_TO_LNE(rdev_wr->event_cause);
+ if (evt) {
+ csio_ln_dbg(ln,
+ "Posting event to lnode event:%d "
+ "cause:%d flowid:x%x\n", evt,
+ rdev_wr->event_cause, rdev_flowid);
+ csio_post_event(&ln->sm, evt);
+ }
+
+ /* Handover event to rn SM here. */
+ csio_rnode_fwevt_handler(rn, rdev_wr->event_cause);
+out_pld:
+ spin_unlock_irq(&hw->lock);
+ return;
+ } else {
+ csio_warn(hw, "unexpected WR op(0x%x) recv\n",
+ FW_WR_OP_GET(be32_to_cpu((wr->hi))));
+ CSIO_INC_STATS(hw, n_cpl_unexp);
+ }
+ } else if (cpl_op == CPL_FW6_MSG) {
+ wr = (struct fw_wr_hdr *) (cmd);
+ if (FW_WR_OP_GET(be32_to_cpu(wr->hi)) == FW_FCOE_ELS_CT_WR) {
+ csio_ln_mgmt_wr_handler(hw, wr,
+ sizeof(struct fw_fcoe_els_ct_wr));
+ } else {
+ csio_warn(hw, "unexpected WR op(0x%x) recv\n",
+ FW_WR_OP_GET(be32_to_cpu((wr->hi))));
+ CSIO_INC_STATS(hw, n_cpl_unexp);
+ }
+ } else {
+ csio_warn(hw, "unexpected CPL op(0x%x) recv\n", opcode);
+ CSIO_INC_STATS(hw, n_cpl_unexp);
+ }
+}
+
+/**
+ * csio_lnode_start - Kickstart lnode discovery.
+ * @ln: lnode
+ *
+ * This routine kickstarts the discovery by issuing an FCOE_LINK (up) command.
+ */
+int
+csio_lnode_start(struct csio_lnode *ln)
+{
+ int rv = 0;
+ if (csio_is_phys_ln(ln) && !(ln->flags & CSIO_LNF_LINK_ENABLE)) {
+ rv = csio_fcoe_enable_link(ln, 1);
+ ln->flags |= CSIO_LNF_LINK_ENABLE;
+ }
+
+ return rv;
+}
+
+/**
+ * csio_lnode_stop - Stop the lnode.
+ * @ln: lnode
+ *
+ * This routine is invoked by HW module to stop lnode and its associated NPIV
+ * lnodes.
+ */
+void
+csio_lnode_stop(struct csio_lnode *ln)
+{
+ csio_post_event_lns(ln, CSIO_LNE_DOWN_LINK);
+ if (csio_is_phys_ln(ln) && (ln->flags & CSIO_LNF_LINK_ENABLE)) {
+ csio_fcoe_enable_link(ln, 0);
+ ln->flags &= ~CSIO_LNF_LINK_ENABLE;
+ }
+ csio_ln_dbg(ln, "stopping ln :%p\n", ln);
+}
+
+/**
+ * csio_lnode_close - Close an lnode.
+ * @ln: lnode
+ *
+ * This routine is invoked by HW module to close an lnode and its
+ * associated NPIV lnodes. Lnode and its associated NPIV lnodes are
+ * set to uninitialized state.
+ */
+void
+csio_lnode_close(struct csio_lnode *ln)
+{
+ csio_post_event_lns(ln, CSIO_LNE_CLOSE);
+ if (csio_is_phys_ln(ln))
+ ln->vnp_flowid = CSIO_INVALID_IDX;
+
+ csio_ln_dbg(ln, "closed ln :%p\n", ln);
+}
+
+/*
+ * csio_ln_prep_ecwr - Prepare ELS/CT WR.
+ * @io_req - IO request.
+ * @wr_len - WR len
+ * @immd_len - WR immediate data
+ * @sub_op - Sub opcode
+ * @sid - source portid.
+ * @did - destination portid
+ * @flow_id - flowid
+ * @fw_wr - ELS/CT WR to be prepared.
+ * Returns: 0 - on success
+ */
+static int
+csio_ln_prep_ecwr(struct csio_ioreq *io_req, uint32_t wr_len,
+ uint32_t immd_len, uint8_t sub_op, uint32_t sid,
+ uint32_t did, uint32_t flow_id, uint8_t *fw_wr)
+{
+ struct fw_fcoe_els_ct_wr *wr;
+ uint32_t port_id;
+
+ wr = (struct fw_fcoe_els_ct_wr *)fw_wr;
+ wr->op_immdlen = cpu_to_be32(FW_WR_OP(FW_FCOE_ELS_CT_WR) |
+ FW_FCOE_ELS_CT_WR_IMMDLEN(immd_len));
+
+ wr_len = DIV_ROUND_UP(wr_len, 16);
+ wr->flowid_len16 = cpu_to_be32(FW_WR_FLOWID(flow_id) |
+ FW_WR_LEN16(wr_len));
+ wr->els_ct_type = sub_op;
+ wr->ctl_pri = 0;
+ wr->cp_en_class = 0;
+ wr->cookie = io_req->fw_handle;
+ wr->iqid = (uint16_t)cpu_to_be16(csio_q_physiqid(
+ io_req->lnode->hwp, io_req->iq_idx));
+ wr->fl_to_sp = FW_FCOE_ELS_CT_WR_SP(1);
+ wr->tmo_val = (uint8_t) io_req->tmo;
+ port_id = htonl(sid);
+ memcpy(wr->l_id, PORT_ID_PTR(port_id), 3);
+ port_id = htonl(did);
+ memcpy(wr->r_id, PORT_ID_PTR(port_id), 3);
+
+ /* Prepare RSP SGL */
+ wr->rsp_dmalen = cpu_to_be32(io_req->dma_buf.len);
+ wr->rsp_dmaaddr = cpu_to_be64(io_req->dma_buf.paddr);
+ return 0;
+}
+
+/*
+ * csio_ln_mgmt_submit_wr - Post elsct work request.
+ * @mgmtm - mgmtm
+ * @io_req - io request.
+ * @sub_op - ELS or CT request type
+ * @pld - Dma Payload buffer
+ * @pld_len - Payload len
+ * Prepares ELSCT Work request and sents it to FW.
+ * Returns: 0 - on success
+ */
+static int
+csio_ln_mgmt_submit_wr(struct csio_mgmtm *mgmtm, struct csio_ioreq *io_req,
+ uint8_t sub_op, struct csio_dma_buf *pld,
+ uint32_t pld_len)
+{
+ struct csio_wr_pair wrp;
+ struct csio_lnode *ln = io_req->lnode;
+ struct csio_rnode *rn = io_req->rnode;
+ struct csio_hw *hw = mgmtm->hw;
+ uint8_t fw_wr[64];
+ struct ulptx_sgl dsgl;
+ uint32_t wr_size = 0;
+ uint8_t im_len = 0;
+ uint32_t wr_off = 0;
+
+ int ret = 0;
+
+ /* Calculate WR Size for this ELS REQ */
+ wr_size = sizeof(struct fw_fcoe_els_ct_wr);
+
+ /* Send as immediate data if pld < 256 */
+ if (pld_len < 256) {
+ wr_size += ALIGN(pld_len, 8);
+ im_len = (uint8_t)pld_len;
+ } else
+ wr_size += sizeof(struct ulptx_sgl);
+
+ /* Roundup WR size in units of 16 bytes */
+ wr_size = ALIGN(wr_size, 16);
+
+ /* Get WR to send ELS REQ */
+ ret = csio_wr_get(hw, mgmtm->eq_idx, wr_size, &wrp);
+ if (ret != 0) {
+ csio_err(hw, "Failed to get WR for ec_req %p ret:%d\n",
+ io_req, ret);
+ return ret;
+ }
+
+ /* Prepare Generic WR used by all ELS/CT cmd */
+ csio_ln_prep_ecwr(io_req, wr_size, im_len, sub_op,
+ ln->nport_id, rn->nport_id,
+ csio_rn_flowid(rn),
+ &fw_wr[0]);
+
+ /* Copy ELS/CT WR CMD */
+ csio_wr_copy_to_wrp(&fw_wr[0], &wrp, wr_off,
+ sizeof(struct fw_fcoe_els_ct_wr));
+ wr_off += sizeof(struct fw_fcoe_els_ct_wr);
+
+ /* Copy payload to Immediate section of WR */
+ if (im_len)
+ csio_wr_copy_to_wrp(pld->vaddr, &wrp, wr_off, im_len);
+ else {
+ /* Program DSGL to dma payload */
+ dsgl.cmd_nsge = htonl(ULPTX_CMD(ULP_TX_SC_DSGL) |
+ ULPTX_MORE | ULPTX_NSGE(1));
+ dsgl.len0 = cpu_to_be32(pld_len);
+ dsgl.addr0 = cpu_to_be64(pld->paddr);
+ csio_wr_copy_to_wrp(&dsgl, &wrp, ALIGN(wr_off, 8),
+ sizeof(struct ulptx_sgl));
+ }
+
+ /* Issue work request to xmit ELS/CT req to FW */
+ csio_wr_issue(mgmtm->hw, mgmtm->eq_idx, false);
+ return ret;
+}
+
+/*
+ * csio_ln_mgmt_submit_req - Submit FCOE Mgmt request.
+ * @io_req - IO Request
+ * @io_cbfn - Completion handler.
+ * @req_type - ELS or CT request type
+ * @pld - Dma Payload buffer
+ * @pld_len - Payload len
+ *
+ *
+ * This API used submit managment ELS/CT request.
+ * This called with hw lock held
+ * Returns: 0 - on success
+ * -ENOMEM - on error.
+ */
+static int
+csio_ln_mgmt_submit_req(struct csio_ioreq *io_req,
+ void (*io_cbfn) (struct csio_hw *, struct csio_ioreq *),
+ enum fcoe_cmn_type req_type, struct csio_dma_buf *pld,
+ uint32_t pld_len)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(io_req->lnode);
+ struct csio_mgmtm *mgmtm = csio_hw_to_mgmtm(hw);
+ int rv;
+
+ io_req->io_cbfn = io_cbfn; /* Upper layer callback handler */
+ io_req->fw_handle = (uintptr_t) (io_req);
+ io_req->eq_idx = mgmtm->eq_idx;
+ io_req->iq_idx = mgmtm->iq_idx;
+
+ rv = csio_ln_mgmt_submit_wr(mgmtm, io_req, req_type, pld, pld_len);
+ if (rv == 0) {
+ list_add_tail(&io_req->sm.sm_list, &mgmtm->active_q);
+ mgmtm->stats.n_active++;
+ }
+ return rv;
+}
+
+/*
+ * csio_ln_fdmi_init - FDMI Init entry point.
+ * @ln: lnode
+ */
+static int
+csio_ln_fdmi_init(struct csio_lnode *ln)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+ struct csio_dma_buf *dma_buf;
+
+ /* Allocate MGMT request required for FDMI */
+ ln->mgmt_req = kzalloc(sizeof(struct csio_ioreq), GFP_KERNEL);
+ if (!ln->mgmt_req) {
+ csio_ln_err(ln, "Failed to alloc ioreq for FDMI\n");
+ CSIO_INC_STATS(hw, n_err_nomem);
+ return -ENOMEM;
+ }
+
+ /* Allocate Dma buffers for FDMI response Payload */
+ dma_buf = &ln->mgmt_req->dma_buf;
+ dma_buf->len = 2048;
+ dma_buf->vaddr = pci_alloc_consistent(hw->pdev, dma_buf->len,
+ &dma_buf->paddr);
+ if (!dma_buf->vaddr) {
+ csio_err(hw, "Failed to alloc DMA buffer for FDMI!\n");
+ kfree(ln->mgmt_req);
+ ln->mgmt_req = NULL;
+ return -ENOMEM;
+ }
+
+ ln->flags |= CSIO_LNF_FDMI_ENABLE;
+ return 0;
+}
+
+/*
+ * csio_ln_fdmi_exit - FDMI exit entry point.
+ * @ln: lnode
+ */
+static int
+csio_ln_fdmi_exit(struct csio_lnode *ln)
+{
+ struct csio_dma_buf *dma_buf;
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ if (!ln->mgmt_req)
+ return 0;
+
+ dma_buf = &ln->mgmt_req->dma_buf;
+ if (dma_buf->vaddr)
+ pci_free_consistent(hw->pdev, dma_buf->len, dma_buf->vaddr,
+ dma_buf->paddr);
+
+ kfree(ln->mgmt_req);
+ return 0;
+}
+
+int
+csio_scan_done(struct csio_lnode *ln, unsigned long ticks,
+ unsigned long time, unsigned long max_scan_ticks,
+ unsigned long delta_scan_ticks)
+{
+ int rv = 0;
+
+ if (time >= max_scan_ticks)
+ return 1;
+
+ if (!ln->tgt_scan_tick)
+ ln->tgt_scan_tick = ticks;
+
+ if (((ticks - ln->tgt_scan_tick) >= delta_scan_ticks)) {
+ if (!ln->last_scan_ntgts)
+ ln->last_scan_ntgts = ln->n_scsi_tgts;
+ else {
+ if (ln->last_scan_ntgts == ln->n_scsi_tgts)
+ return 1;
+
+ ln->last_scan_ntgts = ln->n_scsi_tgts;
+ }
+ ln->tgt_scan_tick = ticks;
+ }
+ return rv;
+}
+
+/*
+ * csio_notify_lnodes:
+ * @hw: HW module
+ * @note: Notification
+ *
+ * Called from the HW SM to fan out notifications to the
+ * Lnode SM. Since the HW SM is entered with lock held,
+ * there is no need to hold locks here.
+ *
+ */
+void
+csio_notify_lnodes(struct csio_hw *hw, enum csio_ln_notify note)
+{
+ struct list_head *tmp;
+ struct csio_lnode *ln;
+
+ csio_dbg(hw, "Notifying all nodes of event %d\n", note);
+
+ /* Traverse children lnodes list and send evt */
+ list_for_each(tmp, &hw->sln_head) {
+ ln = (struct csio_lnode *) tmp;
+
+ switch (note) {
+ case CSIO_LN_NOTIFY_HWREADY:
+ csio_lnode_start(ln);
+ break;
+
+ case CSIO_LN_NOTIFY_HWRESET:
+ case CSIO_LN_NOTIFY_HWREMOVE:
+ csio_lnode_close(ln);
+ break;
+
+ case CSIO_LN_NOTIFY_HWSTOP:
+ csio_lnode_stop(ln);
+ break;
+
+ default:
+ break;
+
+ }
+ }
+}
+
+/*
+ * csio_disable_lnodes:
+ * @hw: HW module
+ * @portid:port id
+ * @disable: disable/enable flag.
+ * If disable=1, disables all lnode hosted on given physical port.
+ * otherwise enables all the lnodes on given phsysical port.
+ * This routine need to called with hw lock held.
+ */
+void
+csio_disable_lnodes(struct csio_hw *hw, uint8_t portid, bool disable)
+{
+ struct list_head *tmp;
+ struct csio_lnode *ln;
+
+ csio_dbg(hw, "Notifying event to all nodes of port:%d\n", portid);
+
+ /* Traverse sibling lnodes list and send evt */
+ list_for_each(tmp, &hw->sln_head) {
+ ln = (struct csio_lnode *) tmp;
+ if (ln->portid != portid)
+ continue;
+
+ if (disable)
+ csio_lnode_stop(ln);
+ else
+ csio_lnode_start(ln);
+ }
+}
+
+/*
+ * csio_ln_init - Initialize an lnode.
+ * @ln: lnode
+ *
+ */
+static int
+csio_ln_init(struct csio_lnode *ln)
+{
+ int rv = -EINVAL;
+ struct csio_lnode *rln, *pln;
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ csio_init_state(&ln->sm, csio_lns_uninit);
+ ln->vnp_flowid = CSIO_INVALID_IDX;
+ ln->fcf_flowid = CSIO_INVALID_IDX;
+
+ if (csio_is_root_ln(ln)) {
+
+ /* This is the lnode used during initialization */
+
+ ln->fcfinfo = kzalloc(sizeof(struct csio_fcf_info), GFP_KERNEL);
+ if (!ln->fcfinfo) {
+ csio_ln_err(ln, "Failed to alloc FCF record\n");
+ CSIO_INC_STATS(hw, n_err_nomem);
+ goto err;
+ }
+
+ INIT_LIST_HEAD(&ln->fcf_lsthead);
+ kref_init(&ln->fcfinfo->kref);
+
+ if (csio_fdmi_enable && csio_ln_fdmi_init(ln))
+ goto err;
+
+ } else { /* Either a non-root physical or a virtual lnode */
+
+ /*
+ * THe rest is common for non-root physical and NPIV lnodes.
+ * Just get references to all other modules
+ */
+ rln = csio_root_lnode(ln);
+
+ if (csio_is_npiv_ln(ln)) {
+ /* NPIV */
+ pln = csio_parent_lnode(ln);
+ kref_get(&pln->fcfinfo->kref);
+ ln->fcfinfo = pln->fcfinfo;
+ } else {
+ /* Another non-root physical lnode (FCF) */
+ ln->fcfinfo = kzalloc(sizeof(struct csio_fcf_info),
+ GFP_KERNEL);
+ if (!ln->fcfinfo) {
+ csio_ln_err(ln, "Failed to alloc FCF info\n");
+ CSIO_INC_STATS(hw, n_err_nomem);
+ goto err;
+ }
+
+ kref_init(&ln->fcfinfo->kref);
+
+ if (csio_fdmi_enable && csio_ln_fdmi_init(ln))
+ goto err;
+ }
+
+ } /* if (!csio_is_root_ln(ln)) */
+
+ return 0;
+err:
+ return rv;
+}
+
+static void
+csio_ln_exit(struct csio_lnode *ln)
+{
+ struct csio_lnode *pln;
+
+ csio_cleanup_rns(ln);
+ if (csio_is_npiv_ln(ln)) {
+ pln = csio_parent_lnode(ln);
+ kref_put(&pln->fcfinfo->kref, csio_free_fcfinfo);
+ } else {
+ kref_put(&ln->fcfinfo->kref, csio_free_fcfinfo);
+ if (csio_fdmi_enable)
+ csio_ln_fdmi_exit(ln);
+ }
+ ln->fcfinfo = NULL;
+}
+
+/**
+ * csio_lnode_init - Initialize the members of an lnode.
+ * @ln: lnode
+ *
+ */
+int
+csio_lnode_init(struct csio_lnode *ln, struct csio_hw *hw,
+ struct csio_lnode *pln)
+{
+ int rv = -EINVAL;
+
+ /* Link this lnode to hw */
+ csio_lnode_to_hw(ln) = hw;
+
+ /* Link child to parent if child lnode */
+ if (pln)
+ ln->pln = pln;
+ else
+ ln->pln = NULL;
+
+ /* Initialize scsi_tgt and timers to zero */
+ ln->n_scsi_tgts = 0;
+ ln->last_scan_ntgts = 0;
+ ln->tgt_scan_tick = 0;
+
+ /* Initialize rnode list */
+ INIT_LIST_HEAD(&ln->rnhead);
+ INIT_LIST_HEAD(&ln->cln_head);
+
+ /* Initialize log level for debug */
+ ln->params.log_level = hw->params.log_level;
+
+ if (csio_ln_init(ln))
+ goto err;
+
+ /* Add lnode to list of sibling or children lnodes */
+ spin_lock_irq(&hw->lock);
+ list_add_tail(&ln->sm.sm_list, pln ? &pln->cln_head : &hw->sln_head);
+ if (pln)
+ pln->num_vports++;
+ spin_unlock_irq(&hw->lock);
+
+ hw->num_lns++;
+
+ return 0;
+err:
+ csio_lnode_to_hw(ln) = NULL;
+ return rv;
+}
+
+/**
+ * csio_lnode_exit - De-instantiate an lnode.
+ * @ln: lnode
+ *
+ */
+void
+csio_lnode_exit(struct csio_lnode *ln)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ csio_ln_exit(ln);
+
+ /* Remove this lnode from hw->sln_head */
+ spin_lock_irq(&hw->lock);
+
+ list_del_init(&ln->sm.sm_list);
+
+ /* If it is children lnode, decrement the
+ * counter in its parent lnode
+ */
+ if (ln->pln)
+ ln->pln->num_vports--;
+
+ /* Update root lnode pointer */
+ if (list_empty(&hw->sln_head))
+ hw->rln = NULL;
+ else
+ hw->rln = (struct csio_lnode *)csio_list_next(&hw->sln_head);
+
+ spin_unlock_irq(&hw->lock);
+
+ csio_lnode_to_hw(ln) = NULL;
+ hw->num_lns--;
+}
diff --git a/drivers/scsi/csiostor/csio_rnode.c b/drivers/scsi/csiostor/csio_rnode.c
new file mode 100644
index 0000000..b0ae430
--- /dev/null
+++ b/drivers/scsi/csiostor/csio_rnode.c
@@ -0,0 +1,912 @@
+/*
+ * This file is part of the Chelsio FCoE driver for Linux.
+ *
+ * Copyright (c) 2008-2012 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/string.h>
+#include <scsi/scsi_device.h>
+#include <scsi/scsi_transport_fc.h>
+#include <scsi/fc/fc_els.h>
+#include <scsi/fc/fc_fs.h>
+
+#include "csio_hw.h"
+#include "csio_lnode.h"
+#include "csio_rnode.h"
+
+static int csio_rnode_init(struct csio_rnode *, struct csio_lnode *);
+static void csio_rnode_exit(struct csio_rnode *);
+
+/* Static machine forward declarations */
+static void csio_rns_uninit(struct csio_rnode *, enum csio_rn_ev);
+static void csio_rns_ready(struct csio_rnode *, enum csio_rn_ev);
+static void csio_rns_offline(struct csio_rnode *, enum csio_rn_ev);
+static void csio_rns_disappeared(struct csio_rnode *, enum csio_rn_ev);
+
+/* RNF event mapping */
+static enum csio_rn_ev fwevt_to_rnevt[] = {
+ CSIO_RNFE_NONE, /* None */
+ CSIO_RNFE_LOGGED_IN, /* PLOGI_ACC_RCVD */
+ CSIO_RNFE_NONE, /* PLOGI_RJT_RCVD */
+ CSIO_RNFE_PLOGI_RECV, /* PLOGI_RCVD */
+ CSIO_RNFE_LOGO_RECV, /* PLOGO_RCVD */
+ CSIO_RNFE_PRLI_DONE, /* PRLI_ACC_RCVD */
+ CSIO_RNFE_NONE, /* PRLI_RJT_RCVD */
+ CSIO_RNFE_PRLI_RECV, /* PRLI_RCVD */
+ CSIO_RNFE_PRLO_RECV, /* PRLO_RCVD */
+ CSIO_RNFE_NONE, /* NPORT_ID_CHGD */
+ CSIO_RNFE_LOGO_RECV, /* FLOGO_RCVD */
+ CSIO_RNFE_NONE, /* CLR_VIRT_LNK_RCVD */
+ CSIO_RNFE_LOGGED_IN, /* FLOGI_ACC_RCVD */
+ CSIO_RNFE_NONE, /* FLOGI_RJT_RCVD */
+ CSIO_RNFE_LOGGED_IN, /* FDISC_ACC_RCVD */
+ CSIO_RNFE_NONE, /* FDISC_RJT_RCVD */
+ CSIO_RNFE_NONE, /* FLOGI_TMO_MAX_RETRY */
+ CSIO_RNFE_NONE, /* IMPL_LOGO_ADISC_ACC */
+ CSIO_RNFE_NONE, /* IMPL_LOGO_ADISC_RJT */
+ CSIO_RNFE_NONE, /* IMPL_LOGO_ADISC_CNFLT */
+ CSIO_RNFE_NONE, /* PRLI_TMO */
+ CSIO_RNFE_NONE, /* ADISC_TMO */
+ CSIO_RNFE_NAME_MISSING, /* RSCN_DEV_LOST */
+ CSIO_RNFE_NONE, /* SCR_ACC_RCVD */
+ CSIO_RNFE_NONE, /* ADISC_RJT_RCVD */
+ CSIO_RNFE_NONE, /* LOGO_SNT */
+ CSIO_RNFE_LOGO_RECV, /* PROTO_ERR_IMPL_LOGO */
+};
+
+#define CSIO_FWE_TO_RNFE(_evt) ((_evt > PROTO_ERR_IMPL_LOGO) ? \
+ CSIO_RNFE_NONE : \
+ fwevt_to_rnevt[_evt])
+int
+csio_is_rnode_ready(struct csio_rnode *rn)
+{
+ return csio_match_state(rn, csio_rns_ready);
+}
+
+static int
+csio_is_rnode_uninit(struct csio_rnode *rn)
+{
+ return csio_match_state(rn, csio_rns_uninit);
+}
+
+static int
+csio_is_rnode_wka(uint8_t rport_type)
+{
+ if ((rport_type == FLOGI_VFPORT) ||
+ (rport_type == FDISC_VFPORT) ||
+ (rport_type == NS_VNPORT) ||
+ (rport_type == FDMI_VNPORT))
+ return 1;
+
+ return 0;
+}
+
+/*
+ * csio_rn_lookup - Finds the rnode with the given flowid
+ * @ln - lnode
+ * @flowid - flowid.
+ *
+ * Does the rnode lookup on the given lnode and flowid.If no matching entry
+ * found, NULL is returned.
+ */
+static struct csio_rnode *
+csio_rn_lookup(struct csio_lnode *ln, uint32_t flowid)
+{
+ struct csio_rnode *rnhead = (struct csio_rnode *) &ln->rnhead;
+ struct list_head *tmp;
+ struct csio_rnode *rn;
+
+ list_for_each(tmp, &rnhead->sm.sm_list) {
+ rn = (struct csio_rnode *) tmp;
+ if (rn->flowid == flowid)
+ return rn;
+ }
+
+ return NULL;
+}
+
+/*
+ * csio_rn_lookup_wwpn - Finds the rnode with the given wwpn
+ * @ln: lnode
+ * @wwpn: wwpn
+ *
+ * Does the rnode lookup on the given lnode and wwpn. If no matching entry
+ * found, NULL is returned.
+ */
+static struct csio_rnode *
+csio_rn_lookup_wwpn(struct csio_lnode *ln, uint8_t *wwpn)
+{
+ struct csio_rnode *rnhead = (struct csio_rnode *) &ln->rnhead;
+ struct list_head *tmp;
+ struct csio_rnode *rn;
+
+ list_for_each(tmp, &rnhead->sm.sm_list) {
+ rn = (struct csio_rnode *) tmp;
+ if (!memcmp(csio_rn_wwpn(rn), wwpn, 8))
+ return rn;
+ }
+
+ return NULL;
+}
+
+/**
+ * csio_rnode_lookup_portid - Finds the rnode with the given portid
+ * @ln: lnode
+ * @portid: port id
+ *
+ * Lookup the rnode list for a given portid. If no matching entry
+ * found, NULL is returned.
+ */
+struct csio_rnode *
+csio_rnode_lookup_portid(struct csio_lnode *ln, uint32_t portid)
+{
+ struct csio_rnode *rnhead = (struct csio_rnode *) &ln->rnhead;
+ struct list_head *tmp;
+ struct csio_rnode *rn;
+
+ list_for_each(tmp, &rnhead->sm.sm_list) {
+ rn = (struct csio_rnode *) tmp;
+ if (rn->nport_id == portid)
+ return rn;
+ }
+
+ return NULL;
+}
+
+static int
+csio_rn_dup_flowid(struct csio_lnode *ln, uint32_t rdev_flowid,
+ uint32_t *vnp_flowid)
+{
+ struct csio_rnode *rnhead;
+ struct list_head *tmp, *tmp1;
+ struct csio_rnode *rn;
+ struct csio_lnode *ln_tmp;
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ list_for_each(tmp1, &hw->sln_head) {
+ ln_tmp = (struct csio_lnode *) tmp1;
+ if (ln_tmp == ln)
+ continue;
+
+ rnhead = (struct csio_rnode *)&ln_tmp->rnhead;
+ list_for_each(tmp, &rnhead->sm.sm_list) {
+
+ rn = (struct csio_rnode *) tmp;
+ if (csio_is_rnode_ready(rn)) {
+ if (rn->flowid == rdev_flowid) {
+ *vnp_flowid = csio_ln_flowid(ln_tmp);
+ return 1;
+ }
+ }
+ }
+ }
+
+ return 0;
+}
+
+static struct csio_rnode *
+csio_alloc_rnode(struct csio_lnode *ln)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ struct csio_rnode *rn = mempool_alloc(hw->rnode_mempool, GFP_ATOMIC);
+ if (!rn)
+ goto err;
+
+ memset(rn, 0, sizeof(struct csio_rnode));
+ if (csio_rnode_init(rn, ln))
+ goto err_free;
+
+ CSIO_INC_STATS(ln, n_rnode_alloc);
+
+ return rn;
+
+err_free:
+ mempool_free(rn, hw->rnode_mempool);
+err:
+ CSIO_INC_STATS(ln, n_rnode_nomem);
+ return NULL;
+}
+
+static void
+csio_free_rnode(struct csio_rnode *rn)
+{
+ struct csio_hw *hw = csio_lnode_to_hw(csio_rnode_to_lnode(rn));
+
+ csio_rnode_exit(rn);
+ CSIO_INC_STATS(rn->lnp, n_rnode_free);
+ mempool_free(rn, hw->rnode_mempool);
+}
+
+/*
+ * csio_get_rnode - Gets rnode with the given flowid
+ * @ln - lnode
+ * @flowid - flow id.
+ *
+ * Does the rnode lookup on the given lnode and flowid. If no matching
+ * rnode found, then new rnode with given npid is allocated and returned.
+ */
+static struct csio_rnode *
+csio_get_rnode(struct csio_lnode *ln, uint32_t flowid)
+{
+ struct csio_rnode *rn;
+
+ rn = csio_rn_lookup(ln, flowid);
+ if (!rn) {
+ rn = csio_alloc_rnode(ln);
+ if (!rn)
+ return NULL;
+
+ rn->flowid = flowid;
+ }
+
+ return rn;
+}
+
+/*
+ * csio_put_rnode - Frees the given rnode
+ * @ln - lnode
+ * @flowid - flow id.
+ *
+ * Does the rnode lookup on the given lnode and flowid. If no matching
+ * rnode found, then new rnode with given npid is allocated and returned.
+ */
+void
+csio_put_rnode(struct csio_lnode *ln, struct csio_rnode *rn)
+{
+ CSIO_DB_ASSERT(csio_is_rnode_uninit(rn) != 0);
+ csio_free_rnode(rn);
+}
+
+/*
+ * csio_confirm_rnode - confirms rnode based on wwpn.
+ * @ln: lnode
+ * @rdev_flowid: remote device flowid
+ * @rdevp: remote device params
+ * This routines searches other rnode in list having same wwpn of new rnode.
+ * If there is a match, then matched rnode is returned and otherwise new rnode
+ * is returned.
+ * returns rnode.
+ */
+struct csio_rnode *
+csio_confirm_rnode(struct csio_lnode *ln, uint32_t rdev_flowid,
+ struct fcoe_rdev_entry *rdevp)
+{
+ uint8_t rport_type;
+ struct csio_rnode *rn, *match_rn;
+ uint32_t vnp_flowid;
+ uint32_t *port_id;
+
+ port_id = (uint32_t *)&rdevp->r_id[0];
+ rport_type =
+ FW_RDEV_WR_RPORT_TYPE_GET(rdevp->rd_xfer_rdy_to_rport_type);
+
+ /* Drop rdev event for cntrl port */
+ if (rport_type == FAB_CTLR_VNPORT) {
+ csio_ln_dbg(ln,
+ "Unhandled rport_type:%d recv in rdev evt "
+ "ssni:x%x\n", rport_type, rdev_flowid);
+ return NULL;
+ }
+
+ /* Lookup on flowid */
+ rn = csio_rn_lookup(ln, rdev_flowid);
+ if (!rn) {
+
+ /* Drop events with duplicate flowid */
+ if (csio_rn_dup_flowid(ln, rdev_flowid, &vnp_flowid)) {
+ csio_ln_warn(ln,
+ "ssni:%x already active on vnpi:%x",
+ rdev_flowid, vnp_flowid);
+ return NULL;
+ }
+
+ /* Lookup on wwpn for NPORTs */
+ rn = csio_rn_lookup_wwpn(ln, rdevp->wwpn);
+ if (!rn)
+ goto alloc_rnode;
+
+ } else {
+ /* Lookup well-known ports with nport id */
+ if (csio_is_rnode_wka(rport_type)) {
+ match_rn = csio_rnode_lookup_portid(ln,
+ ((ntohl(*port_id) >> 8) & CSIO_DID_MASK));
+ if (match_rn == NULL) {
+ csio_rn_flowid(rn) = CSIO_INVALID_IDX;
+ goto alloc_rnode;
+ }
+
+ /*
+ * Now compare the wwpn to confirm that
+ * same port relogged in. If so update the matched rn.
+ * Else, go ahead and alloc a new rnode.
+ */
+ if (!memcmp(csio_rn_wwpn(match_rn), rdevp->wwpn, 8)) {
+ if (csio_is_rnode_ready(rn)) {
+ csio_ln_warn(ln,
+ "rnode is already"
+ "active ssni:x%x\n",
+ rdev_flowid);
+ CSIO_ASSERT(0);
+ }
+ csio_rn_flowid(rn) = CSIO_INVALID_IDX;
+ rn = match_rn;
+
+ /* Update rn */
+ goto found_rnode;
+ }
+ csio_rn_flowid(rn) = CSIO_INVALID_IDX;
+ goto alloc_rnode;
+ }
+
+ /* wwpn match */
+ if (!memcmp(csio_rn_wwpn(rn), rdevp->wwpn, 8))
+ goto found_rnode;
+
+ /* Search for rnode that have same wwpn */
+ match_rn = csio_rn_lookup_wwpn(ln, rdevp->wwpn);
+ if (match_rn != NULL) {
+ csio_ln_dbg(ln,
+ "ssni:x%x changed for rport name(wwpn):%llx "
+ "did:x%x\n", rdev_flowid,
+ wwn_to_u64(rdevp->wwpn),
+ match_rn->nport_id);
+ csio_rn_flowid(rn) = CSIO_INVALID_IDX;
+ rn = match_rn;
+ } else {
+ csio_ln_dbg(ln,
+ "rnode wwpn mismatch found ssni:x%x "
+ "name(wwpn):%llx\n",
+ rdev_flowid,
+ wwn_to_u64(csio_rn_wwpn(rn)));
+ if (csio_is_rnode_ready(rn)) {
+ csio_ln_warn(ln,
+ "rnode is already active "
+ "wwpn:%llx ssni:x%x\n",
+ wwn_to_u64(csio_rn_wwpn(rn)),
+ rdev_flowid);
+ CSIO_ASSERT(0);
+ }
+ csio_rn_flowid(rn) = CSIO_INVALID_IDX;
+ goto alloc_rnode;
+ }
+ }
+
+found_rnode:
+ csio_ln_dbg(ln, "found rnode:%p ssni:x%x name(wwpn):%llx\n",
+ rn, rdev_flowid, wwn_to_u64(rdevp->wwpn));
+
+ /* Update flowid */
+ csio_rn_flowid(rn) = rdev_flowid;
+
+ /* update rdev entry */
+ rn->rdev_entry = rdevp;
+ CSIO_INC_STATS(ln, n_rnode_match);
+ return rn;
+
+alloc_rnode:
+ rn = csio_get_rnode(ln, rdev_flowid);
+ if (!rn)
+ return NULL;
+
+ csio_ln_dbg(ln, "alloc rnode:%p ssni:x%x name(wwpn):%llx\n",
+ rn, rdev_flowid, wwn_to_u64(rdevp->wwpn));
+
+ /* update rdev entry */
+ rn->rdev_entry = rdevp;
+ return rn;
+}
+
+/*
+ * csio_rn_verify_rparams - verify rparams.
+ * @ln: lnode
+ * @rn: rnode
+ * @rdevp: remote device params
+ * returns success if rparams are verified.
+ */
+static int
+csio_rn_verify_rparams(struct csio_lnode *ln, struct csio_rnode *rn,
+ struct fcoe_rdev_entry *rdevp)
+{
+ uint8_t null[8];
+ uint8_t rport_type;
+ uint8_t fc_class;
+ uint32_t *did;
+
+ did = (uint32_t *) &rdevp->r_id[0];
+ rport_type =
+ FW_RDEV_WR_RPORT_TYPE_GET(rdevp->rd_xfer_rdy_to_rport_type);
+ switch (rport_type) {
+ case FLOGI_VFPORT:
+ rn->role = CSIO_RNFR_FABRIC;
+ if (((ntohl(*did) >> 8) & CSIO_DID_MASK) != FC_FID_FLOGI) {
+ csio_ln_err(ln, "ssni:x%x invalid fabric portid\n",
+ csio_rn_flowid(rn));
+ return -EINVAL;
+ }
+ /* NPIV support */
+ if (FW_RDEV_WR_NPIV_GET(rdevp->vft_to_qos))
+ ln->flags |= CSIO_LNF_NPIVSUPP;
+
+ break;
+
+ case NS_VNPORT:
+ rn->role = CSIO_RNFR_NS;
+ if (((ntohl(*did) >> 8) & CSIO_DID_MASK) != FC_FID_DIR_SERV) {
+ csio_ln_err(ln, "ssni:x%x invalid fabric portid\n",
+ csio_rn_flowid(rn));
+ return -EINVAL;
+ }
+ break;
+
+ case REG_FC4_VNPORT:
+ case REG_VNPORT:
+ rn->role = CSIO_RNFR_NPORT;
+ if (rdevp->event_cause == PRLI_ACC_RCVD ||
+ rdevp->event_cause == PRLI_RCVD) {
+ if (FW_RDEV_WR_TASK_RETRY_ID_GET(
+ rdevp->enh_disc_to_tgt))
+ rn->fcp_flags |= FCP_SPPF_OVLY_ALLOW;
+
+ if (FW_RDEV_WR_RETRY_GET(rdevp->enh_disc_to_tgt))
+ rn->fcp_flags |= FCP_SPPF_RETRY;
+
+ if (FW_RDEV_WR_CONF_CMPL_GET(rdevp->enh_disc_to_tgt))
+ rn->fcp_flags |= FCP_SPPF_CONF_COMPL;
+
+ if (FW_RDEV_WR_TGT_GET(rdevp->enh_disc_to_tgt))
+ rn->role |= CSIO_RNFR_TARGET;
+
+ if (FW_RDEV_WR_INI_GET(rdevp->enh_disc_to_tgt))
+ rn->role |= CSIO_RNFR_INITIATOR;
+ }
+
+ break;
+
+ case FDMI_VNPORT:
+ case FAB_CTLR_VNPORT:
+ rn->role = 0;
+ break;
+
+ default:
+ csio_ln_err(ln, "ssni:x%x invalid rport type recv x%x\n",
+ csio_rn_flowid(rn), rport_type);
+ return -EINVAL;
+ }
+
+ /* validate wwpn/wwnn for Name server/remote port */
+ if (rport_type == REG_VNPORT || rport_type == NS_VNPORT) {
+ memset(null, 0, 8);
+ if (!memcmp(rdevp->wwnn, null, 8)) {
+ csio_ln_err(ln,
+ "ssni:x%x invalid wwnn received from"
+ " rport did:x%x\n",
+ csio_rn_flowid(rn),
+ (ntohl(*did) & CSIO_DID_MASK));
+ return -EINVAL;
+ }
+
+ if (!memcmp(rdevp->wwpn, null, 8)) {
+ csio_ln_err(ln,
+ "ssni:x%x invalid wwpn received from"
+ " rport did:x%x\n",
+ csio_rn_flowid(rn),
+ (ntohl(*did) & CSIO_DID_MASK));
+ return -EINVAL;
+ }
+
+ }
+
+ /* Copy wwnn, wwpn and nport id */
+ rn->nport_id = (ntohl(*did) >> 8) & CSIO_DID_MASK;
+ memcpy(csio_rn_wwnn(rn), rdevp->wwnn, 8);
+ memcpy(csio_rn_wwpn(rn), rdevp->wwpn, 8);
+ rn->rn_sparm.csp.sp_bb_data = ntohs(rdevp->rcv_fr_sz);
+ fc_class = FW_RDEV_WR_CLASS_GET(rdevp->vft_to_qos);
+ rn->rn_sparm.clsp[fc_class - 1].cp_class = htons(FC_CPC_VALID);
+ return 0;
+}
+
+static void
+__csio_reg_rnode(struct csio_rnode *rn)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ spin_unlock_irq(&hw->lock);
+ csio_reg_rnode(rn);
+ spin_lock_irq(&hw->lock);
+
+ if (rn->role & CSIO_RNFR_TARGET)
+ ln->n_scsi_tgts++;
+
+ if (rn->nport_id == FC_FID_MGMT_SERV)
+ csio_ln_fdmi_start(ln, (void *) rn);
+}
+
+static void
+__csio_unreg_rnode(struct csio_rnode *rn)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+ LIST_HEAD(tmp_q);
+ int cmpl = 0;
+
+ if (!list_empty(&rn->host_cmpl_q)) {
+ csio_dbg(hw, "Returning completion queue I/Os\n");
+ list_splice_tail_init(&rn->host_cmpl_q, &tmp_q);
+ cmpl = 1;
+ }
+
+ if (rn->role & CSIO_RNFR_TARGET) {
+ ln->n_scsi_tgts--;
+ ln->last_scan_ntgts--;
+ }
+
+ spin_unlock_irq(&hw->lock);
+ csio_unreg_rnode(rn);
+ spin_lock_irq(&hw->lock);
+
+ /* Cleanup I/Os that were waiting for rnode to unregister */
+ if (cmpl)
+ csio_scsi_cleanup_io_q(csio_hw_to_scsim(hw), &tmp_q);
+
+}
+
+/*****************************************************************************/
+/* START: Rnode SM */
+/*****************************************************************************/
+
+/*
+ * csio_rns_uninit -
+ * @rn - rnode
+ * @evt - SM event.
+ *
+ */
+static void
+csio_rns_uninit(struct csio_rnode *rn, enum csio_rn_ev evt)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ int ret = 0;
+
+ CSIO_INC_STATS(rn, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_RNFE_LOGGED_IN:
+ case CSIO_RNFE_PLOGI_RECV:
+ ret = csio_rn_verify_rparams(ln, rn, rn->rdev_entry);
+ if (!ret) {
+ csio_set_state(&rn->sm, csio_rns_ready);
+ __csio_reg_rnode(rn);
+ } else {
+ CSIO_INC_STATS(rn, n_err_inval);
+ }
+ break;
+ case CSIO_RNFE_LOGO_RECV:
+ csio_ln_dbg(ln,
+ "ssni:x%x Ignoring event %d recv "
+ "in rn state[uninit]\n", csio_rn_flowid(rn), evt);
+ CSIO_INC_STATS(rn, n_evt_drop);
+ break;
+ default:
+ csio_ln_dbg(ln,
+ "ssni:x%x unexp event %d recv "
+ "in rn state[uninit]\n", csio_rn_flowid(rn), evt);
+ CSIO_INC_STATS(rn, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_rns_ready -
+ * @rn - rnode
+ * @evt - SM event.
+ *
+ */
+static void
+csio_rns_ready(struct csio_rnode *rn, enum csio_rn_ev evt)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ int ret = 0;
+
+ CSIO_INC_STATS(rn, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_RNFE_LOGGED_IN:
+ case CSIO_RNFE_PLOGI_RECV:
+ csio_ln_dbg(ln,
+ "ssni:x%x Ignoring event %d recv from did:x%x "
+ "in rn state[ready]\n", csio_rn_flowid(rn), evt,
+ rn->nport_id);
+ CSIO_INC_STATS(rn, n_evt_drop);
+ break;
+
+ case CSIO_RNFE_PRLI_DONE:
+ case CSIO_RNFE_PRLI_RECV:
+ ret = csio_rn_verify_rparams(ln, rn, rn->rdev_entry);
+ if (!ret)
+ __csio_reg_rnode(rn);
+ else
+ CSIO_INC_STATS(rn, n_err_inval);
+
+ break;
+ case CSIO_RNFE_DOWN:
+ csio_set_state(&rn->sm, csio_rns_offline);
+ __csio_unreg_rnode(rn);
+
+ /* FW expected to internally aborted outstanding SCSI WRs
+ * and return all SCSI WRs to host with status "ABORTED".
+ */
+ break;
+
+ case CSIO_RNFE_LOGO_RECV:
+ csio_set_state(&rn->sm, csio_rns_offline);
+
+ __csio_unreg_rnode(rn);
+
+ /* FW expected to internally aborted outstanding SCSI WRs
+ * and return all SCSI WRs to host with status "ABORTED".
+ */
+ break;
+
+ case CSIO_RNFE_CLOSE:
+ /*
+ * Each rnode receives CLOSE event when driver is removed or
+ * device is reset
+ * Note: All outstanding IOs on remote port need to returned
+ * to uppper layer with appropriate error before sending
+ * CLOSE event
+ */
+ csio_set_state(&rn->sm, csio_rns_uninit);
+ __csio_unreg_rnode(rn);
+ break;
+
+ case CSIO_RNFE_NAME_MISSING:
+ csio_set_state(&rn->sm, csio_rns_disappeared);
+ __csio_unreg_rnode(rn);
+
+ /*
+ * FW expected to internally aborted outstanding SCSI WRs
+ * and return all SCSI WRs to host with status "ABORTED".
+ */
+
+ break;
+
+ default:
+ csio_ln_dbg(ln,
+ "ssni:x%x unexp event %d recv from did:x%x "
+ "in rn state[uninit]\n", csio_rn_flowid(rn), evt,
+ rn->nport_id);
+ CSIO_INC_STATS(rn, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_rns_offline -
+ * @rn - rnode
+ * @evt - SM event.
+ *
+ */
+static void
+csio_rns_offline(struct csio_rnode *rn, enum csio_rn_ev evt)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ int ret = 0;
+
+ CSIO_INC_STATS(rn, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_RNFE_LOGGED_IN:
+ case CSIO_RNFE_PLOGI_RECV:
+ ret = csio_rn_verify_rparams(ln, rn, rn->rdev_entry);
+ if (!ret) {
+ csio_set_state(&rn->sm, csio_rns_ready);
+ __csio_reg_rnode(rn);
+ } else {
+ CSIO_INC_STATS(rn, n_err_inval);
+ csio_post_event(&rn->sm, CSIO_RNFE_CLOSE);
+ }
+ break;
+
+ case CSIO_RNFE_DOWN:
+ csio_ln_dbg(ln,
+ "ssni:x%x Ignoring event %d recv from did:x%x "
+ "in rn state[offline]\n", csio_rn_flowid(rn), evt,
+ rn->nport_id);
+ CSIO_INC_STATS(rn, n_evt_drop);
+ break;
+
+ case CSIO_RNFE_CLOSE:
+ /* Each rnode receives CLOSE event when driver is removed or
+ * device is reset
+ * Note: All outstanding IOs on remote port need to returned
+ * to uppper layer with appropriate error before sending
+ * CLOSE event
+ */
+ csio_set_state(&rn->sm, csio_rns_uninit);
+ break;
+
+ case CSIO_RNFE_NAME_MISSING:
+ csio_set_state(&rn->sm, csio_rns_disappeared);
+ break;
+
+ default:
+ csio_ln_dbg(ln,
+ "ssni:x%x unexp event %d recv from did:x%x "
+ "in rn state[offline]\n", csio_rn_flowid(rn), evt,
+ rn->nport_id);
+ CSIO_INC_STATS(rn, n_evt_unexp);
+ break;
+ }
+}
+
+/*
+ * csio_rns_disappeared -
+ * @rn - rnode
+ * @evt - SM event.
+ *
+ */
+static void
+csio_rns_disappeared(struct csio_rnode *rn, enum csio_rn_ev evt)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ int ret = 0;
+
+ CSIO_INC_STATS(rn, n_evt_sm[evt]);
+
+ switch (evt) {
+ case CSIO_RNFE_LOGGED_IN:
+ case CSIO_RNFE_PLOGI_RECV:
+ ret = csio_rn_verify_rparams(ln, rn, rn->rdev_entry);
+ if (!ret) {
+ csio_set_state(&rn->sm, csio_rns_ready);
+ __csio_reg_rnode(rn);
+ } else {
+ CSIO_INC_STATS(rn, n_err_inval);
+ csio_post_event(&rn->sm, CSIO_RNFE_CLOSE);
+ }
+ break;
+
+ case CSIO_RNFE_CLOSE:
+ /* Each rnode receives CLOSE event when driver is removed or
+ * device is reset.
+ * Note: All outstanding IOs on remote port need to returned
+ * to uppper layer with appropriate error before sending
+ * CLOSE event
+ */
+ csio_set_state(&rn->sm, csio_rns_uninit);
+ break;
+
+ case CSIO_RNFE_DOWN:
+ case CSIO_RNFE_NAME_MISSING:
+ csio_ln_dbg(ln,
+ "ssni:x%x Ignoring event %d recv from did x%x"
+ "in rn state[disappeared]\n", csio_rn_flowid(rn),
+ evt, rn->nport_id);
+ break;
+
+ default:
+ csio_ln_dbg(ln,
+ "ssni:x%x unexp event %d recv from did x%x"
+ "in rn state[disappeared]\n", csio_rn_flowid(rn),
+ evt, rn->nport_id);
+ CSIO_INC_STATS(rn, n_evt_unexp);
+ break;
+ }
+}
+
+/*****************************************************************************/
+/* END: Rnode SM */
+/*****************************************************************************/
+
+/*
+ * csio_rnode_devloss_handler - Device loss event handler
+ * @rn: rnode
+ *
+ * Post event to close rnode SM and free rnode.
+ */
+void
+csio_rnode_devloss_handler(struct csio_rnode *rn)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+
+ /* ignore if same rnode came back as online */
+ if (csio_is_rnode_ready(rn))
+ return;
+
+ csio_post_event(&rn->sm, CSIO_RNFE_CLOSE);
+
+ /* Free rn if in uninit state */
+ if (csio_is_rnode_uninit(rn))
+ csio_put_rnode(ln, rn);
+}
+
+/**
+ * csio_rnode_fwevt_handler - Event handler for firmware rnode events.
+ * @rn: rnode
+ *
+ */
+void
+csio_rnode_fwevt_handler(struct csio_rnode *rn, uint8_t fwevt)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ enum csio_rn_ev evt;
+
+ evt = CSIO_FWE_TO_RNFE(fwevt);
+ if (!evt) {
+ csio_ln_err(ln, "ssni:x%x Unhandled FW Rdev event: %d\n",
+ csio_rn_flowid(rn), fwevt);
+ CSIO_INC_STATS(rn, n_evt_unexp);
+ return;
+ }
+ CSIO_INC_STATS(rn, n_evt_fw[fwevt]);
+
+ /* Track previous & current events for debugging */
+ rn->prev_evt = rn->cur_evt;
+ rn->cur_evt = fwevt;
+
+ /* Post event to rnode SM */
+ csio_post_event(&rn->sm, evt);
+
+ /* Free rn if in uninit state */
+ if (csio_is_rnode_uninit(rn))
+ csio_put_rnode(ln, rn);
+}
+
+/*
+ * csio_rnode_init - Initialize rnode.
+ * @rn: RNode
+ * @ln: Associated lnode
+ *
+ * Caller is responsible for holding the lock. The lock is required
+ * to be held for inserting the rnode in ln->rnhead list.
+ */
+static int
+csio_rnode_init(struct csio_rnode *rn, struct csio_lnode *ln)
+{
+ csio_rnode_to_lnode(rn) = ln;
+ csio_init_state(&rn->sm, csio_rns_uninit);
+ INIT_LIST_HEAD(&rn->host_cmpl_q);
+ csio_rn_flowid(rn) = CSIO_INVALID_IDX;
+
+ /* Add rnode to list of lnodes->rnhead */
+ list_add_tail(&rn->sm.sm_list, &ln->rnhead);
+
+ return 0;
+}
+
+static void
+csio_rnode_exit(struct csio_rnode *rn)
+{
+ list_del_init(&rn->sm.sm_list);
+ CSIO_DB_ASSERT(list_empty(&rn->host_cmpl_q));
+}
--
1.7.1
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox