* [PATCH net-net] net: loopback: set default mtu to 64K
From: Eric Dumazet @ 2012-09-24 8:28 UTC (permalink / raw)
To: David Miller; +Cc: netdev
From: Eric Dumazet <edumazet@google.com>
loopback current mtu of 16436 bytes allows no more than 3 MSS TCP
segments per frame, or 48 Kbytes. Changing mtu to 64K allows TCP
stack to build large frames and significantly reduces stack overhead.
Performance boost on bulk TCP transferts can be up to 30 %, partly
because we now have one ACK message for two 64KB segments, and a lower
probability of hitting /proc/sys/net/ipv4/tcp_reordering default limit.
Signed-off-by: Eric Dumazet <edumazet@google.com>
---
drivers/net/loopback.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/loopback.c b/drivers/net/loopback.c
index 4a075ba..81f8f9e 100644
--- a/drivers/net/loopback.c
+++ b/drivers/net/loopback.c
@@ -157,7 +157,7 @@ static const struct net_device_ops loopback_ops = {
*/
static void loopback_setup(struct net_device *dev)
{
- dev->mtu = (16 * 1024) + 20 + 20 + 12;
+ dev->mtu = 64 * 1024;
dev->hard_header_len = ETH_HLEN; /* 14 */
dev->addr_len = ETH_ALEN; /* 6 */
dev->tx_queue_len = 0;
^ permalink raw reply related
* [PATCH net-next v1] gianfar: Change default HW Tx queue scheduling mode
From: Claudiu Manoil @ 2012-09-24 8:39 UTC (permalink / raw)
To: netdev; +Cc: David S. Miller, Paul Gortmaker, Claudiu Manoil
This is primarily to address transmission timeout occurrences, when
multiple H/W Tx queues are being used concurrently. Because in
the priority scheduling mode the controller does not service the
Tx queues equally (but in ascending index order), Tx timeouts are
being triggered rightaway for a basic test with multiple simultaneous
connections like:
iperf -c <server_ip> -n 100M -P 8
resulting in kernel trace:
NETDEV WATCHDOG: eth1 (fsl-gianfar): transmit queue <X> timed out
------------[ cut here ]------------
WARNING: at net/sched/sch_generic.c:255
...
and controller reset during intense traffic, and possibly further
complications.
This patch changes the default H/W Tx scheduling setting (TXSCHED)
for multi-queue devices, from priority scheduling mode to a weighted
round robin mode with equal weights for all H/W Tx queues, and
addresses the issue above.
Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
---
v1: removed runtime config option of tx scheduling mode (i.e. prio_sched_en)
(to be provided in a distinct patch)
drivers/net/ethernet/freescale/gianfar.c | 11 ++++++++++-
drivers/net/ethernet/freescale/gianfar.h | 11 ++++++++++-
2 files changed, 20 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/freescale/gianfar.c b/drivers/net/ethernet/freescale/gianfar.c
index 4d5b58c..a1b52ec 100644
--- a/drivers/net/ethernet/freescale/gianfar.c
+++ b/drivers/net/ethernet/freescale/gianfar.c
@@ -394,7 +394,13 @@ static void gfar_init_mac(struct net_device *ndev)
if (ndev->features & NETIF_F_IP_CSUM)
tctrl |= TCTRL_INIT_CSUM;
- tctrl |= TCTRL_TXSCHED_PRIO;
+ if (priv->prio_sched_en)
+ tctrl |= TCTRL_TXSCHED_PRIO;
+ else {
+ tctrl |= TCTRL_TXSCHED_WRRS;
+ gfar_write(®s->tr03wt, DEFAULT_WRRS_WEIGHT);
+ gfar_write(®s->tr47wt, DEFAULT_WRRS_WEIGHT);
+ }
gfar_write(®s->tctrl, tctrl);
@@ -1160,6 +1166,9 @@ static int gfar_probe(struct platform_device *ofdev)
priv->rx_filer_enable = 1;
/* Enable most messages by default */
priv->msg_enable = (NETIF_MSG_IFUP << 1 ) - 1;
+ /* use pritority h/w tx queue scheduling for single queue devices */
+ if (priv->num_tx_queues == 1)
+ priv->prio_sched_en = 1;
/* Carrier starts down, phylib will bring it up */
netif_carrier_off(dev);
diff --git a/drivers/net/ethernet/freescale/gianfar.h b/drivers/net/ethernet/freescale/gianfar.h
index 2136c7f..4141ef2 100644
--- a/drivers/net/ethernet/freescale/gianfar.h
+++ b/drivers/net/ethernet/freescale/gianfar.h
@@ -301,8 +301,16 @@ extern const char gfar_driver_version[];
#define TCTRL_TFCPAUSE 0x00000008
#define TCTRL_TXSCHED_MASK 0x00000006
#define TCTRL_TXSCHED_INIT 0x00000000
+/* priority scheduling */
#define TCTRL_TXSCHED_PRIO 0x00000002
+/* weighted round-robin scheduling (WRRS) */
#define TCTRL_TXSCHED_WRRS 0x00000004
+/* default WRRS weight and policy setting,
+ * tailored to the tr03wt and tr47wt registers:
+ * equal weight for all Tx Qs, measured in 64byte units
+ */
+#define DEFAULT_WRRS_WEIGHT 0x18181818
+
#define TCTRL_INIT_CSUM (TCTRL_TUCSEN | TCTRL_IPCSEN)
#define IEVENT_INIT_CLEAR 0xffffffff
@@ -1098,7 +1106,8 @@ struct gfar_private {
extended_hash:1,
bd_stash_en:1,
rx_filer_enable:1,
- wol_en:1; /* Wake-on-LAN enabled */
+ wol_en:1, /* Wake-on-LAN enabled */
+ prio_sched_en:1; /* Enable priorty based Tx scheduling in Hw */
unsigned short padding;
/* PHY stuff */
--
1.7.6.5
^ permalink raw reply related
* Re: [PATCH] inet_diag: make config INET_DIAG bool
From: Gao feng @ 2012-09-24 8:48 UTC (permalink / raw)
To: Jan Engelhardt; +Cc: Stephen Hemminger, netdev, davem, kuznet
In-Reply-To: <alpine.LNX.2.01.1209231540180.15692@nerf07.vanv.qr>
于 2012年09月23日 21:40, Jan Engelhardt 写道:
> On Sunday 2012-09-23 06:36, Stephen Hemminger wrote:
>
>>> when inet_diag being compiled as module, inet_diag_handler_dump
>>> set netlink_dump_control.dump to inet_diag_dump,so if module
>>> inet_diag is unloaded,netlink will still try to call this function
>>> in netlink_dump. this will cause kernel panic.
>>
>> Another way to handle this to just get rid of inet_diag_exit
>> so that module can be loaded but not unloaded.
>
> Why don't we unset netlink_dump_control.dump on exit?
Though I like this idea,but it may cause dead lock.
netlink_dimp [mutex_lock(netlink_sock->cb_mutex) here]
|->inet_diag_dump
|->__inet_diag_dump
|->inet_diag_lock_handler [may try to load tcp_diag here,
so we need module_mutex lock]
And on module exit,we already get module_mutex lock,
if we unset netlink_sock->cb,we need to get mutex lock of netlink_sock->cb_mutex.
I think this will cause dead lock.
I don't know if I should change this patch as Stephen said,because I don't know
witch one is better.
Any comments?
^ permalink raw reply
* [net-next 0/6][pull request] Intel Wired LAN Driver Updates
From: Jeff Kirsher @ 2012-09-24 9:00 UTC (permalink / raw)
To: davem; +Cc: Jeff Kirsher, netdev, gospo, sassmann
This series contains updates to ixgbe and ixgbevf (mainly ixgbevf).
The following are changes since commit c9d2ea96ca3bbc85264803ff6bd66eb3bbefdb77:
netlink: Rearrange netlink_kernel_cfg to save space on 64-bit.
and are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
Alexander Duyck (3):
ixgbevf: Fix code for handling timeout
ixgbevf: Add fix to VF to handle multi-descriptor buffers
ixgbevf: Return error on failure to enable VLAN
Greg Rose (2):
ixgbe: Do not read the spoofed packets counter when not in IOV mode
ixgbevf: Fix AIM (Adaptive Interrupt Moderation)
Narendra K (1):
ixgbevf - Remove unused parameter in ixgbevf_receive_skb
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 5 +-
drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 5 ++
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 89 +++++++++++------------
drivers/net/ethernet/intel/ixgbevf/mbx.c | 15 +++-
drivers/net/ethernet/intel/ixgbevf/vf.c | 69 ++++++++++++++----
5 files changed, 117 insertions(+), 66 deletions(-)
--
1.7.11.4
^ permalink raw reply
* [net-next 2/6] ixgbe: Do not read the spoofed packets counter when not in IOV mode
From: Jeff Kirsher @ 2012-09-24 9:00 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1348477207-27030-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
The counter is not valid unless the controller is running in IOV mode.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 70d27a3..29465be 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -5560,8 +5560,9 @@ static void ixgbe_spoof_check(struct ixgbe_adapter *adapter)
{
u32 ssvpc;
- /* Do not perform spoof check for 82598 */
- if (adapter->hw.mac.type == ixgbe_mac_82598EB)
+ /* Do not perform spoof check for 82598 or if not in IOV mode */
+ if (adapter->hw.mac.type == ixgbe_mac_82598EB ||
+ adapter->num_vfs == 0)
return;
ssvpc = IXGBE_READ_REG(&adapter->hw, IXGBE_SSVPC);
--
1.7.11.4
^ permalink raw reply related
* [net-next 1/6] ixgbevf: Fix code for handling timeout
From: Jeff Kirsher @ 2012-09-24 9:00 UTC (permalink / raw)
To: davem; +Cc: Alexander Duyck, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1348477207-27030-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Alexander Duyck <alexander.h.duyck@intel.com>
The VF driver was not designed to correctly handle a message timeout. As
a result it is possible for one bad message to invalidate all messages
following it until the part is reset. Instead we should copy the example
in igbvf of how to handle a mailbox event and message timeout.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 46 +++++----------------
drivers/net/ethernet/intel/ixgbevf/mbx.c | 15 +++++--
drivers/net/ethernet/intel/ixgbevf/vf.c | 49 ++++++++++++++++++-----
3 files changed, 58 insertions(+), 52 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index c5ffe1d..7aa31ee 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -716,40 +716,15 @@ static void ixgbevf_set_itr(struct ixgbevf_q_vector *q_vector)
}
}
-static irqreturn_t ixgbevf_msix_mbx(int irq, void *data)
+static irqreturn_t ixgbevf_msix_other(int irq, void *data)
{
struct ixgbevf_adapter *adapter = data;
struct ixgbe_hw *hw = &adapter->hw;
- u32 msg;
- bool got_ack = false;
- if (!hw->mbx.ops.check_for_ack(hw))
- got_ack = true;
+ hw->mac.get_link_status = 1;
- if (!hw->mbx.ops.check_for_msg(hw)) {
- hw->mbx.ops.read(hw, &msg, 1);
-
- if ((msg & IXGBE_MBVFICR_VFREQ_MASK) == IXGBE_PF_CONTROL_MSG)
- mod_timer(&adapter->watchdog_timer,
- round_jiffies(jiffies + 1));
-
- if (msg & IXGBE_VT_MSGTYPE_NACK)
- pr_warn("Last Request of type %2.2x to PF Nacked\n",
- msg & 0xFF);
- /*
- * Restore the PFSTS bit in case someone is polling for a
- * return message from the PF
- */
- hw->mbx.v2p_mailbox |= IXGBE_VFMAILBOX_PFSTS;
- }
-
- /*
- * checking for the ack clears the PFACK bit. Place
- * it back in the v2p_mailbox cache so that anyone
- * polling for an ack will not miss it
- */
- if (got_ack)
- hw->mbx.v2p_mailbox |= IXGBE_VFMAILBOX_PFACK;
+ if (!test_bit(__IXGBEVF_DOWN, &adapter->state))
+ mod_timer(&adapter->watchdog_timer, jiffies);
IXGBE_WRITE_REG(hw, IXGBE_VTEIMS, adapter->eims_other);
@@ -899,10 +874,10 @@ static int ixgbevf_request_msix_irqs(struct ixgbevf_adapter *adapter)
}
err = request_irq(adapter->msix_entries[vector].vector,
- &ixgbevf_msix_mbx, 0, netdev->name, adapter);
+ &ixgbevf_msix_other, 0, netdev->name, adapter);
if (err) {
hw_dbg(&adapter->hw,
- "request_irq for msix_mbx failed: %d\n", err);
+ "request_irq for msix_other failed: %d\n", err);
goto free_queue_irqs;
}
@@ -1411,6 +1386,7 @@ static void ixgbevf_up_complete(struct ixgbevf_adapter *adapter)
ixgbevf_save_reset_stats(adapter);
ixgbevf_init_last_counter_stats(adapter);
+ hw->mac.get_link_status = 1;
mod_timer(&adapter->watchdog_timer, jiffies);
}
@@ -1589,8 +1565,6 @@ void ixgbevf_down(struct ixgbevf_adapter *adapter)
void ixgbevf_reinit_locked(struct ixgbevf_adapter *adapter)
{
- struct ixgbe_hw *hw = &adapter->hw;
-
WARN_ON(in_interrupt());
while (test_and_set_bit(__IXGBEVF_RESETTING, &adapter->state))
@@ -1603,10 +1577,8 @@ void ixgbevf_reinit_locked(struct ixgbevf_adapter *adapter)
* watchdog task will continue to schedule reset tasks until
* the PF is up and running.
*/
- if (!hw->mac.ops.reset_hw(hw)) {
- ixgbevf_down(adapter);
- ixgbevf_up(adapter);
- }
+ ixgbevf_down(adapter);
+ ixgbevf_up(adapter);
clear_bit(__IXGBEVF_RESETTING, &adapter->state);
}
diff --git a/drivers/net/ethernet/intel/ixgbevf/mbx.c b/drivers/net/ethernet/intel/ixgbevf/mbx.c
index 9c95590..d5028dd 100644
--- a/drivers/net/ethernet/intel/ixgbevf/mbx.c
+++ b/drivers/net/ethernet/intel/ixgbevf/mbx.c
@@ -86,14 +86,17 @@ static s32 ixgbevf_poll_for_ack(struct ixgbe_hw *hw)
static s32 ixgbevf_read_posted_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size)
{
struct ixgbe_mbx_info *mbx = &hw->mbx;
- s32 ret_val = IXGBE_ERR_MBX;
+ s32 ret_val = -IXGBE_ERR_MBX;
+
+ if (!mbx->ops.read)
+ goto out;
ret_val = ixgbevf_poll_for_msg(hw);
/* if ack received read message, otherwise we timed out */
if (!ret_val)
ret_val = mbx->ops.read(hw, msg, size);
-
+out:
return ret_val;
}
@@ -109,7 +112,11 @@ static s32 ixgbevf_read_posted_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size)
static s32 ixgbevf_write_posted_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size)
{
struct ixgbe_mbx_info *mbx = &hw->mbx;
- s32 ret_val;
+ s32 ret_val = -IXGBE_ERR_MBX;
+
+ /* exit if either we can't write or there isn't a defined timeout */
+ if (!mbx->ops.write || !mbx->timeout)
+ goto out;
/* send msg */
ret_val = mbx->ops.write(hw, msg, size);
@@ -117,7 +124,7 @@ static s32 ixgbevf_write_posted_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size)
/* if msg sent wait until we receive an ack */
if (!ret_val)
ret_val = ixgbevf_poll_for_ack(hw);
-
+out:
return ret_val;
}
diff --git a/drivers/net/ethernet/intel/ixgbevf/vf.c b/drivers/net/ethernet/intel/ixgbevf/vf.c
index 87b3f3b..a5e66c7 100644
--- a/drivers/net/ethernet/intel/ixgbevf/vf.c
+++ b/drivers/net/ethernet/intel/ixgbevf/vf.c
@@ -392,20 +392,23 @@ static s32 ixgbevf_check_mac_link_vf(struct ixgbe_hw *hw,
bool *link_up,
bool autoneg_wait_to_complete)
{
+ struct ixgbe_mbx_info *mbx = &hw->mbx;
+ struct ixgbe_mac_info *mac = &hw->mac;
+ s32 ret_val = 0;
u32 links_reg;
+ u32 in_msg = 0;
- if (!(hw->mbx.ops.check_for_rst(hw))) {
- *link_up = false;
- *speed = 0;
- return -1;
- }
+ /* If we were hit with a reset drop the link */
+ if (!mbx->ops.check_for_rst(hw) || !mbx->timeout)
+ mac->get_link_status = true;
- links_reg = IXGBE_READ_REG(hw, IXGBE_VFLINKS);
+ if (!mac->get_link_status)
+ goto out;
- if (links_reg & IXGBE_LINKS_UP)
- *link_up = true;
- else
- *link_up = false;
+ /* if link status is down no point in checking to see if pf is up */
+ links_reg = IXGBE_READ_REG(hw, IXGBE_VFLINKS);
+ if (!(links_reg & IXGBE_LINKS_UP))
+ goto out;
switch (links_reg & IXGBE_LINKS_SPEED_82599) {
case IXGBE_LINKS_SPEED_10G_82599:
@@ -419,7 +422,31 @@ static s32 ixgbevf_check_mac_link_vf(struct ixgbe_hw *hw,
break;
}
- return 0;
+ /* if the read failed it could just be a mailbox collision, best wait
+ * until we are called again and don't report an error */
+ if (mbx->ops.read(hw, &in_msg, 1))
+ goto out;
+
+ if (!(in_msg & IXGBE_VT_MSGTYPE_CTS)) {
+ /* msg is not CTS and is NACK we must have lost CTS status */
+ if (in_msg & IXGBE_VT_MSGTYPE_NACK)
+ ret_val = -1;
+ goto out;
+ }
+
+ /* the pf is talking, if we timed out in the past we reinit */
+ if (!mbx->timeout) {
+ ret_val = -1;
+ goto out;
+ }
+
+ /* if we passed all the tests above then the link is up and we no
+ * longer need to check for link */
+ mac->get_link_status = false;
+
+out:
+ *link_up = !mac->get_link_status;
+ return ret_val;
}
/**
--
1.7.11.4
^ permalink raw reply related
* [net-next 3/6] ixgbevf - Remove unused parameter in ixgbevf_receive_skb
From: Jeff Kirsher @ 2012-09-24 9:00 UTC (permalink / raw)
To: davem; +Cc: Narendra K, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1348477207-27030-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Narendra K <narendra_k@dell.com>
Remove 'rx_ring' parameter as it is not used in ixgbevf_receive_skb
Signed-off-by: Narendra K <narendra_k@dell.com>
Acked-by: Greg Rose <gregory.v.rose@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 7aa31ee..cb215c6 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -272,12 +272,10 @@ cont_loop:
* @q_vector: structure containing interrupt and ring information
* @skb: packet to send up
* @status: hardware indication of status of receive
- * @rx_ring: rx descriptor ring (for a specific queue) to setup
* @rx_desc: rx descriptor
**/
static void ixgbevf_receive_skb(struct ixgbevf_q_vector *q_vector,
struct sk_buff *skb, u8 status,
- struct ixgbevf_ring *ring,
union ixgbe_adv_rx_desc *rx_desc)
{
struct ixgbevf_adapter *adapter = q_vector->adapter;
@@ -461,7 +459,7 @@ static bool ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
}
skb->protocol = eth_type_trans(skb, rx_ring->netdev);
- ixgbevf_receive_skb(q_vector, skb, staterr, rx_ring, rx_desc);
+ ixgbevf_receive_skb(q_vector, skb, staterr, rx_desc);
next_desc:
rx_desc->wb.upper.status_error = 0;
--
1.7.11.4
^ permalink raw reply related
* [net-next 4/6] ixgbevf: Fix AIM (Adaptive Interrupt Moderation)
From: Jeff Kirsher @ 2012-09-24 9:00 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1348477207-27030-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
While fixing up a patch from Alex Duyck to use q_vectors in ring containers
to update the ITR I bungled it and missed actually updating the counters
in the ring container q_vectors. This patch fixes my mistake and makes
interrupt moderation actually work.
Signed-off-by: Greg Rose <gregory.v.rose@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 insertions(+)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index cb215c6..37bbd86 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -263,6 +263,8 @@ cont_loop:
tx_ring->total_bytes += total_bytes;
tx_ring->total_packets += total_packets;
u64_stats_update_end(&tx_ring->syncp);
+ q_vector->tx.total_bytes += total_bytes;
+ q_vector->tx.total_packets += total_packets;
return count < tx_ring->count;
}
@@ -488,6 +490,8 @@ next_desc:
rx_ring->total_packets += total_rx_packets;
rx_ring->total_bytes += total_rx_bytes;
u64_stats_update_end(&rx_ring->syncp);
+ q_vector->rx.total_packets += total_rx_packets;
+ q_vector->rx.total_bytes += total_rx_bytes;
return !!budget;
}
--
1.7.11.4
^ permalink raw reply related
* [net-next 5/6] ixgbevf: Add fix to VF to handle multi-descriptor buffers
From: Jeff Kirsher @ 2012-09-24 9:00 UTC (permalink / raw)
To: davem; +Cc: Alexander Duyck, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1348477207-27030-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Alexander Duyck <alexander.h.duyck@intel.com>
This change fixes the ixgbevf driver so that it can correctly drop a frame
should it receive a jumbo frame.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 5 +++++
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 14 ++++++++++++--
2 files changed, 17 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index eb26fda..383b4e1 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -261,6 +261,11 @@ enum ixbgevf_state_t {
__IXGBEVF_DOWN
};
+struct ixgbevf_cb {
+ struct sk_buff *prev;
+};
+#define IXGBE_CB(skb) ((struct ixgbevf_cb *)(skb)->cb)
+
enum ixgbevf_boards {
board_82599_vf,
board_X540_vf,
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 37bbd86..2ba15ae 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -433,11 +433,21 @@ static bool ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
if (!(staterr & IXGBE_RXD_STAT_EOP)) {
skb->next = next_buffer->skb;
- skb->next->prev = skb;
+ IXGBE_CB(skb->next)->prev = skb;
adapter->non_eop_descs++;
goto next_desc;
}
+ /* we should not be chaining buffers, if we did drop the skb */
+ if (IXGBE_CB(skb)->prev) {
+ do {
+ struct sk_buff *this = skb;
+ skb = IXGBE_CB(skb)->prev;
+ dev_kfree_skb(this);
+ } while (skb);
+ goto next_desc;
+ }
+
/* ERR_MASK will only have valid bits if EOP set */
if (unlikely(staterr & IXGBE_RXDADV_ERR_FRAME_ERR_MASK)) {
dev_kfree_skb_irq(skb);
@@ -1439,7 +1449,7 @@ static void ixgbevf_clean_rx_ring(struct ixgbevf_adapter *adapter,
rx_buffer_info->skb = NULL;
do {
struct sk_buff *this = skb;
- skb = skb->prev;
+ skb = IXGBE_CB(skb)->prev;
dev_kfree_skb(this);
} while (skb);
}
--
1.7.11.4
^ permalink raw reply related
* [net-next 6/6] ixgbevf: Return error on failure to enable VLAN
From: Jeff Kirsher @ 2012-09-24 9:00 UTC (permalink / raw)
To: davem; +Cc: Alexander Duyck, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1348477207-27030-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Alexander Duyck <alexander.h.duyck@intel.com>
With recent kernel changes we can now return errors on a failure to setup a
VLAN filter. This patch takes advantage of that opportunity so that we can
return either an EIO error in the case of a mailbox failure, or an EACCESS
error in the case of being denied access to the VLAN filter table by the
PF.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Robert Garrett <robertx.e.garrett@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 21 ++++++++++++++++-----
drivers/net/ethernet/intel/ixgbevf/vf.c | 20 ++++++++++++++++++--
2 files changed, 34 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 2ba15ae..cf372ee 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1126,36 +1126,47 @@ static int ixgbevf_vlan_rx_add_vid(struct net_device *netdev, u16 vid)
{
struct ixgbevf_adapter *adapter = netdev_priv(netdev);
struct ixgbe_hw *hw = &adapter->hw;
+ int err;
+
+ if (!hw->mac.ops.set_vfta)
+ return -EOPNOTSUPP;
spin_lock(&adapter->mbx_lock);
/* add VID to filter table */
- if (hw->mac.ops.set_vfta)
- hw->mac.ops.set_vfta(hw, vid, 0, true);
+ err = hw->mac.ops.set_vfta(hw, vid, 0, true);
spin_unlock(&adapter->mbx_lock);
+ /* translate error return types so error makes sense */
+ if (err == IXGBE_ERR_MBX)
+ return -EIO;
+
+ if (err == IXGBE_ERR_INVALID_ARGUMENT)
+ return -EACCES;
+
set_bit(vid, adapter->active_vlans);
- return 0;
+ return err;
}
static int ixgbevf_vlan_rx_kill_vid(struct net_device *netdev, u16 vid)
{
struct ixgbevf_adapter *adapter = netdev_priv(netdev);
struct ixgbe_hw *hw = &adapter->hw;
+ int err = -EOPNOTSUPP;
spin_lock(&adapter->mbx_lock);
/* remove VID from filter table */
if (hw->mac.ops.set_vfta)
- hw->mac.ops.set_vfta(hw, vid, 0, false);
+ err = hw->mac.ops.set_vfta(hw, vid, 0, false);
spin_unlock(&adapter->mbx_lock);
clear_bit(vid, adapter->active_vlans);
- return 0;
+ return err;
}
static void ixgbevf_restore_vlan(struct ixgbevf_adapter *adapter)
diff --git a/drivers/net/ethernet/intel/ixgbevf/vf.c b/drivers/net/ethernet/intel/ixgbevf/vf.c
index a5e66c7..0c7447e 100644
--- a/drivers/net/ethernet/intel/ixgbevf/vf.c
+++ b/drivers/net/ethernet/intel/ixgbevf/vf.c
@@ -349,16 +349,32 @@ static s32 ixgbevf_update_mc_addr_list_vf(struct ixgbe_hw *hw,
static s32 ixgbevf_set_vfta_vf(struct ixgbe_hw *hw, u32 vlan, u32 vind,
bool vlan_on)
{
+ struct ixgbe_mbx_info *mbx = &hw->mbx;
u32 msgbuf[2];
+ s32 err;
msgbuf[0] = IXGBE_VF_SET_VLAN;
msgbuf[1] = vlan;
/* Setting the 8 bit field MSG INFO to TRUE indicates "add" */
msgbuf[0] |= vlan_on << IXGBE_VT_MSGINFO_SHIFT;
- ixgbevf_write_msg_read_ack(hw, msgbuf, 2);
+ err = mbx->ops.write_posted(hw, msgbuf, 2);
+ if (err)
+ goto mbx_err;
- return 0;
+ err = mbx->ops.read_posted(hw, msgbuf, 2);
+ if (err)
+ goto mbx_err;
+
+ /* remove extra bits from the message */
+ msgbuf[0] &= ~IXGBE_VT_MSGTYPE_CTS;
+ msgbuf[0] &= ~(0xFF << IXGBE_VT_MSGINFO_SHIFT);
+
+ if (msgbuf[0] != (IXGBE_VF_SET_VLAN | IXGBE_VT_MSGTYPE_ACK))
+ err = IXGBE_ERR_INVALID_ARGUMENT;
+
+mbx_err:
+ return err;
}
/**
--
1.7.11.4
^ permalink raw reply related
* [PATCH net-next v3] net: use a per task frag allocator
From: Eric Dumazet @ 2012-09-24 9:04 UTC (permalink / raw)
To: David Miller; +Cc: Vijay Subramanian, netdev, Ben Hutchings, Alexander Duyck
In-Reply-To: <1348261871.2669.962.camel@edumazet-glaptop>
From: Eric Dumazet <edumazet@google.com>
We currently use a per socket order-0 page cache for tcp_sendmsg()
operations.
This page is used to build fragments for skbs.
Its done to increase probability of coalescing small write() into
single segments in skbs still in write queue (not yet sent)
But it wastes a lot of memory for applications handling many mostly
idle sockets, since each socket holds one page in sk->sk_sndmsg_page
Its also quite inefficient to build TSO 64KB packets, because we need
about 16 pages per skb on arches where PAGE_SIZE = 4096, so we hit
page allocator more than wanted.
This patch adds a per task frag allocator and uses bigger pages,
if available. An automatic fallback is done in case of memory pressure.
(up to 32768 bytes per frag, thats order-3 pages on x86)
This increases TCP stream performance by 20% on loopback device,
but also benefits on other network devices, since 8x less frags are
mapped on transmit and unmapped on tx completion. Alexander Duyck
mentioned a probable performance win on systems with IOMMU enabled.
Its possible some SG enabled hardware cant cope with bigger fragments,
but their ndo_start_xmit() should already handle this, splitting a
fragment in sub fragments, since some arches have PAGE_SIZE=65536
Successfully tested on various ethernet devices.
(ixgbe, igb, bnx2x, tg3, mellanox mlx4)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Vijay Subramanian <subramanian.vijay@gmail.com>
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
---
v3: net/sched/em_meta.c bit pointed out by Vijay
v2: uses existing page_frag structure to hold page/offset/size
convert linear_to_page()/__ip_append_data()/ip6_append_data()
remove @page and @off fields from struct inet_cork
move the destructor from tcp_v4_destroy_sock() to sk_common_release
include/linux/sched.h | 3 +
include/net/inet_sock.h | 4 -
include/net/sock.h | 27 +++++++-----
kernel/exit.c | 3 +
kernel/fork.c | 1
net/core/skbuff.c | 37 ++++-------------
net/core/sock.c | 49 ++++++++++++++++++++++-
net/ipv4/ip_output.c | 70 +++++++++++++--------------------
net/ipv4/raw.c | 19 +++++----
net/ipv4/tcp.c | 79 ++++++++++----------------------------
net/ipv4/tcp_ipv4.c | 8 ---
net/ipv6/ip6_output.c | 65 ++++++++++++-------------------
net/sched/em_meta.c | 2
13 files changed, 167 insertions(+), 200 deletions(-)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index b8c8664..a8e2413 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1530,6 +1530,9 @@ struct task_struct {
* cache last used pipe for splice
*/
struct pipe_inode_info *splice_pipe;
+
+ struct page_frag task_frag;
+
#ifdef CONFIG_TASK_DELAY_ACCT
struct task_delay_info *delays;
#endif
diff --git a/include/net/inet_sock.h b/include/net/inet_sock.h
index 613cfa4..256c1ed 100644
--- a/include/net/inet_sock.h
+++ b/include/net/inet_sock.h
@@ -101,10 +101,8 @@ struct inet_cork {
__be32 addr;
struct ip_options *opt;
unsigned int fragsize;
- struct dst_entry *dst;
int length; /* Total length of all frames */
- struct page *page;
- u32 off;
+ struct dst_entry *dst;
u8 tx_flags;
};
diff --git a/include/net/sock.h b/include/net/sock.h
index 181b711..42053759 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -247,8 +247,7 @@ struct cg_proto;
* @sk_stamp: time stamp of last packet received
* @sk_socket: Identd and reporting IO signals
* @sk_user_data: RPC layer private data
- * @sk_sndmsg_page: cached page for sendmsg
- * @sk_sndmsg_off: cached offset for sendmsg
+ * @sk_frag: cached page frag
* @sk_peek_off: current peek_offset value
* @sk_send_head: front of stuff to transmit
* @sk_security: used by security modules
@@ -362,9 +361,8 @@ struct sock {
ktime_t sk_stamp;
struct socket *sk_socket;
void *sk_user_data;
- struct page *sk_sndmsg_page;
+ struct page_frag sk_frag;
struct sk_buff *sk_send_head;
- __u32 sk_sndmsg_off;
__s32 sk_peek_off;
int sk_write_pending;
#ifdef CONFIG_SECURITY
@@ -2034,18 +2032,23 @@ static inline void sk_stream_moderate_sndbuf(struct sock *sk)
struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp);
-static inline struct page *sk_stream_alloc_page(struct sock *sk)
+/**
+ * sk_page_frag - return an appropriate page_frag
+ * @sk: socket
+ *
+ * If socket allocation mode allows current thread to sleep, it means its
+ * safe to use the per task page_frag instead of the per socket one.
+ */
+static inline struct page_frag *sk_page_frag(struct sock *sk)
{
- struct page *page = NULL;
+ if (sk->sk_allocation & __GFP_WAIT)
+ return ¤t->task_frag;
- page = alloc_pages(sk->sk_allocation, 0);
- if (!page) {
- sk_enter_memory_pressure(sk);
- sk_stream_moderate_sndbuf(sk);
- }
- return page;
+ return &sk->sk_frag;
}
+extern bool sk_page_frag_refill(struct sock *sk, struct page_frag *pfrag);
+
/*
* Default write policy as shown to user space via poll/select/SIGIO
*/
diff --git a/kernel/exit.c b/kernel/exit.c
index f65345f..42f2595 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -1046,6 +1046,9 @@ void do_exit(long code)
if (tsk->splice_pipe)
__free_pipe_info(tsk->splice_pipe);
+ if (tsk->task_frag.page)
+ put_page(tsk->task_frag.page);
+
validate_creds_for_do_exit(tsk);
preempt_disable();
diff --git a/kernel/fork.c b/kernel/fork.c
index 2c8857e..01565b9 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -330,6 +330,7 @@ static struct task_struct *dup_task_struct(struct task_struct *orig)
tsk->btrace_seq = 0;
#endif
tsk->splice_pipe = NULL;
+ tsk->task_frag.page = NULL;
account_kernel_stack(ti, 1);
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index fe00d12..2ede3cf 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1655,38 +1655,19 @@ static struct page *linear_to_page(struct page *page, unsigned int *len,
unsigned int *offset,
struct sk_buff *skb, struct sock *sk)
{
- struct page *p = sk->sk_sndmsg_page;
- unsigned int off;
+ struct page_frag *pfrag = sk_page_frag(sk);
- if (!p) {
-new_page:
- p = sk->sk_sndmsg_page = alloc_pages(sk->sk_allocation, 0);
- if (!p)
- return NULL;
-
- off = sk->sk_sndmsg_off = 0;
- /* hold one ref to this page until it's full */
- } else {
- unsigned int mlen;
-
- /* If we are the only user of the page, we can reset offset */
- if (page_count(p) == 1)
- sk->sk_sndmsg_off = 0;
- off = sk->sk_sndmsg_off;
- mlen = PAGE_SIZE - off;
- if (mlen < 64 && mlen < *len) {
- put_page(p);
- goto new_page;
- }
+ if (!sk_page_frag_refill(sk, pfrag))
+ return NULL;
- *len = min_t(unsigned int, *len, mlen);
- }
+ *len = min_t(unsigned int, *len, pfrag->size - pfrag->offset);
- memcpy(page_address(p) + off, page_address(page) + *offset, *len);
- sk->sk_sndmsg_off += *len;
- *offset = off;
+ memcpy(page_address(pfrag->page) + pfrag->offset,
+ page_address(page) + *offset, *len);
+ *offset = pfrag->offset;
+ pfrag->offset += *len;
- return p;
+ return pfrag->page;
}
static bool spd_can_coalesce(const struct splice_pipe_desc *spd,
diff --git a/net/core/sock.c b/net/core/sock.c
index 2693f76..727114c 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1744,6 +1744,45 @@ struct sk_buff *sock_alloc_send_skb(struct sock *sk, unsigned long size,
}
EXPORT_SYMBOL(sock_alloc_send_skb);
+/* On 32bit arches, an skb frag is limited to 2^15 */
+#define SKB_FRAG_PAGE_ORDER get_order(32768)
+
+bool sk_page_frag_refill(struct sock *sk, struct page_frag *pfrag)
+{
+ int order;
+
+ if (pfrag->page) {
+ if (atomic_read(&pfrag->page->_count) == 1) {
+ pfrag->offset = 0;
+ return true;
+ }
+ if (pfrag->offset < pfrag->size)
+ return true;
+ put_page(pfrag->page);
+ }
+
+ /* We restrict high order allocations to users that can afford to wait */
+ order = (sk->sk_allocation & __GFP_WAIT) ? SKB_FRAG_PAGE_ORDER : 0;
+
+ do {
+ gfp_t gfp = sk->sk_allocation;
+
+ if (order)
+ gfp |= __GFP_COMP | __GFP_NOWARN;
+ pfrag->page = alloc_pages(gfp, order);
+ if (likely(pfrag->page)) {
+ pfrag->offset = 0;
+ pfrag->size = PAGE_SIZE << order;
+ return true;
+ }
+ } while (--order >= 0);
+
+ sk_enter_memory_pressure(sk);
+ sk_stream_moderate_sndbuf(sk);
+ return false;
+}
+EXPORT_SYMBOL(sk_page_frag_refill);
+
static void __lock_sock(struct sock *sk)
__releases(&sk->sk_lock.slock)
__acquires(&sk->sk_lock.slock)
@@ -2173,8 +2212,8 @@ void sock_init_data(struct socket *sock, struct sock *sk)
sk->sk_error_report = sock_def_error_report;
sk->sk_destruct = sock_def_destruct;
- sk->sk_sndmsg_page = NULL;
- sk->sk_sndmsg_off = 0;
+ sk->sk_frag.page = NULL;
+ sk->sk_frag.offset = 0;
sk->sk_peek_off = -1;
sk->sk_peer_pid = NULL;
@@ -2417,6 +2456,12 @@ void sk_common_release(struct sock *sk)
xfrm_sk_free_policy(sk);
sk_refcnt_debug_release(sk);
+
+ if (sk->sk_frag.page) {
+ put_page(sk->sk_frag.page);
+ sk->sk_frag.page = NULL;
+ }
+
sock_put(sk);
}
EXPORT_SYMBOL(sk_common_release);
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index a5beab1..24a29a3 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -793,6 +793,7 @@ static int __ip_append_data(struct sock *sk,
struct flowi4 *fl4,
struct sk_buff_head *queue,
struct inet_cork *cork,
+ struct page_frag *pfrag,
int getfrag(void *from, char *to, int offset,
int len, int odd, struct sk_buff *skb),
void *from, int length, int transhdrlen,
@@ -987,47 +988,30 @@ alloc_new_skb:
}
} else {
int i = skb_shinfo(skb)->nr_frags;
- skb_frag_t *frag = &skb_shinfo(skb)->frags[i-1];
- struct page *page = cork->page;
- int off = cork->off;
- unsigned int left;
-
- if (page && (left = PAGE_SIZE - off) > 0) {
- if (copy >= left)
- copy = left;
- if (page != skb_frag_page(frag)) {
- if (i == MAX_SKB_FRAGS) {
- err = -EMSGSIZE;
- goto error;
- }
- skb_fill_page_desc(skb, i, page, off, 0);
- skb_frag_ref(skb, i);
- frag = &skb_shinfo(skb)->frags[i];
- }
- } else if (i < MAX_SKB_FRAGS) {
- if (copy > PAGE_SIZE)
- copy = PAGE_SIZE;
- page = alloc_pages(sk->sk_allocation, 0);
- if (page == NULL) {
- err = -ENOMEM;
- goto error;
- }
- cork->page = page;
- cork->off = 0;
- skb_fill_page_desc(skb, i, page, 0, 0);
- frag = &skb_shinfo(skb)->frags[i];
- } else {
- err = -EMSGSIZE;
- goto error;
- }
- if (getfrag(from, skb_frag_address(frag)+skb_frag_size(frag),
- offset, copy, skb->len, skb) < 0) {
- err = -EFAULT;
+ err = -ENOMEM;
+ if (!sk_page_frag_refill(sk, pfrag))
goto error;
+
+ if (!skb_can_coalesce(skb, i, pfrag->page,
+ pfrag->offset)) {
+ err = -EMSGSIZE;
+ if (i == MAX_SKB_FRAGS)
+ goto error;
+
+ __skb_fill_page_desc(skb, i, pfrag->page,
+ pfrag->offset, 0);
+ skb_shinfo(skb)->nr_frags = ++i;
+ get_page(pfrag->page);
}
- cork->off += copy;
- skb_frag_size_add(frag, copy);
+ copy = min_t(int, copy, pfrag->size - pfrag->offset);
+ if (getfrag(from,
+ page_address(pfrag->page) + pfrag->offset,
+ offset, copy, skb->len, skb) < 0)
+ goto error_efault;
+
+ pfrag->offset += copy;
+ skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
skb->len += copy;
skb->data_len += copy;
skb->truesize += copy;
@@ -1039,6 +1023,8 @@ alloc_new_skb:
return 0;
+error_efault:
+ err = -EFAULT;
error:
cork->length -= length;
IP_INC_STATS(sock_net(sk), IPSTATS_MIB_OUTDISCARDS);
@@ -1079,8 +1065,6 @@ static int ip_setup_cork(struct sock *sk, struct inet_cork *cork,
cork->dst = &rt->dst;
cork->length = 0;
cork->tx_flags = ipc->tx_flags;
- cork->page = NULL;
- cork->off = 0;
return 0;
}
@@ -1117,7 +1101,8 @@ int ip_append_data(struct sock *sk, struct flowi4 *fl4,
transhdrlen = 0;
}
- return __ip_append_data(sk, fl4, &sk->sk_write_queue, &inet->cork.base, getfrag,
+ return __ip_append_data(sk, fl4, &sk->sk_write_queue, &inet->cork.base,
+ sk_page_frag(sk), getfrag,
from, length, transhdrlen, flags);
}
@@ -1439,7 +1424,8 @@ struct sk_buff *ip_make_skb(struct sock *sk,
if (err)
return ERR_PTR(err);
- err = __ip_append_data(sk, fl4, &queue, &cork, getfrag,
+ err = __ip_append_data(sk, fl4, &queue, &cork,
+ ¤t->task_frag, getfrag,
from, length, transhdrlen, flags);
if (err) {
__ip_flush_pending_frames(sk, &queue, &cork);
diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c
index f242578..3a35c81 100644
--- a/net/ipv4/raw.c
+++ b/net/ipv4/raw.c
@@ -131,18 +131,23 @@ found:
* 0 - deliver
* 1 - block
*/
-static __inline__ int icmp_filter(struct sock *sk, struct sk_buff *skb)
+static int icmp_filter(const struct sock *sk, const struct sk_buff *skb)
{
- int type;
-
- if (!pskb_may_pull(skb, sizeof(struct icmphdr)))
+ struct icmphdr _hdr;
+ const struct icmphdr *hdr;
+
+ pr_err("icmp_filter skb_transport_offset %d data-head %ld len %d/%d\n",
+ skb_transport_offset(skb), skb->data - skb->head, skb->len, skb->data_len);
+ hdr = skb_header_pointer(skb, skb_transport_offset(skb),
+ sizeof(_hdr), &_hdr);
+ pr_err("head %p data %p hdr %p type %d\n", skb->head, skb->data, hdr, hdr ? hdr->type : -1);
+ if (!hdr)
return 1;
- type = icmp_hdr(skb)->type;
- if (type < 32) {
+ if (hdr->type < 32) {
__u32 data = raw_sk(sk)->filter.data;
- return ((1 << type) & data) != 0;
+ return ((1U << hdr->type) & data) != 0;
}
/* Do not block unknown ICMP types */
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 7b1e940..72ea475 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1150,78 +1150,43 @@ new_segment:
if (err)
goto do_fault;
} else {
- bool merge = false;
+ bool merge = true;
int i = skb_shinfo(skb)->nr_frags;
- struct page *page = sk->sk_sndmsg_page;
- int off;
-
- if (page && page_count(page) == 1)
- sk->sk_sndmsg_off = 0;
-
- off = sk->sk_sndmsg_off;
-
- if (skb_can_coalesce(skb, i, page, off) &&
- off != PAGE_SIZE) {
- /* We can extend the last page
- * fragment. */
- merge = true;
- } else if (i == MAX_SKB_FRAGS || !sg) {
- /* Need to add new fragment and cannot
- * do this because interface is non-SG,
- * or because all the page slots are
- * busy. */
- tcp_mark_push(tp, skb);
- goto new_segment;
- } else if (page) {
- if (off == PAGE_SIZE) {
- put_page(page);
- sk->sk_sndmsg_page = page = NULL;
- off = 0;
+ struct page_frag *pfrag = sk_page_frag(sk);
+
+ if (!sk_page_frag_refill(sk, pfrag))
+ goto wait_for_memory;
+
+ if (!skb_can_coalesce(skb, i, pfrag->page,
+ pfrag->offset)) {
+ if (i == MAX_SKB_FRAGS || !sg) {
+ tcp_mark_push(tp, skb);
+ goto new_segment;
}
- } else
- off = 0;
+ merge = false;
+ }
- if (copy > PAGE_SIZE - off)
- copy = PAGE_SIZE - off;
+ copy = min_t(int, copy, pfrag->size - pfrag->offset);
if (!sk_wmem_schedule(sk, copy))
goto wait_for_memory;
- if (!page) {
- /* Allocate new cache page. */
- if (!(page = sk_stream_alloc_page(sk)))
- goto wait_for_memory;
- }
-
- /* Time to copy data. We are close to
- * the end! */
err = skb_copy_to_page_nocache(sk, from, skb,
- page, off, copy);
- if (err) {
- /* If this page was new, give it to the
- * socket so it does not get leaked.
- */
- if (!sk->sk_sndmsg_page) {
- sk->sk_sndmsg_page = page;
- sk->sk_sndmsg_off = 0;
- }
+ pfrag->page,
+ pfrag->offset,
+ copy);
+ if (err)
goto do_error;
- }
/* Update the skb. */
if (merge) {
skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
} else {
- skb_fill_page_desc(skb, i, page, off, copy);
- if (sk->sk_sndmsg_page) {
- get_page(page);
- } else if (off + copy < PAGE_SIZE) {
- get_page(page);
- sk->sk_sndmsg_page = page;
- }
+ skb_fill_page_desc(skb, i, pfrag->page,
+ pfrag->offset, copy);
+ get_page(pfrag->page);
}
-
- sk->sk_sndmsg_off = off + copy;
+ pfrag->offset += copy;
}
if (!copied)
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 0a7e020..93406c5 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2200,14 +2200,6 @@ void tcp_v4_destroy_sock(struct sock *sk)
if (inet_csk(sk)->icsk_bind_hash)
inet_put_port(sk);
- /*
- * If sendmsg cached page exists, toss it.
- */
- if (sk->sk_sndmsg_page) {
- __free_page(sk->sk_sndmsg_page);
- sk->sk_sndmsg_page = NULL;
- }
-
/* TCP Cookie Transactions */
if (tp->cookie_values != NULL) {
kref_put(&tp->cookie_values->kref,
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 3dd4a37..aece3e7 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1279,8 +1279,6 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
if (dst_allfrag(rt->dst.path))
cork->flags |= IPCORK_ALLFRAG;
cork->length = 0;
- sk->sk_sndmsg_page = NULL;
- sk->sk_sndmsg_off = 0;
exthdrlen = (opt ? opt->opt_flen : 0) - rt->rt6i_nfheader_len;
length += exthdrlen;
transhdrlen += exthdrlen;
@@ -1504,48 +1502,31 @@ alloc_new_skb:
}
} else {
int i = skb_shinfo(skb)->nr_frags;
- skb_frag_t *frag = &skb_shinfo(skb)->frags[i-1];
- struct page *page = sk->sk_sndmsg_page;
- int off = sk->sk_sndmsg_off;
- unsigned int left;
-
- if (page && (left = PAGE_SIZE - off) > 0) {
- if (copy >= left)
- copy = left;
- if (page != skb_frag_page(frag)) {
- if (i == MAX_SKB_FRAGS) {
- err = -EMSGSIZE;
- goto error;
- }
- skb_fill_page_desc(skb, i, page, sk->sk_sndmsg_off, 0);
- skb_frag_ref(skb, i);
- frag = &skb_shinfo(skb)->frags[i];
- }
- } else if(i < MAX_SKB_FRAGS) {
- if (copy > PAGE_SIZE)
- copy = PAGE_SIZE;
- page = alloc_pages(sk->sk_allocation, 0);
- if (page == NULL) {
- err = -ENOMEM;
- goto error;
- }
- sk->sk_sndmsg_page = page;
- sk->sk_sndmsg_off = 0;
+ struct page_frag *pfrag = sk_page_frag(sk);
- skb_fill_page_desc(skb, i, page, 0, 0);
- frag = &skb_shinfo(skb)->frags[i];
- } else {
- err = -EMSGSIZE;
+ err = -ENOMEM;
+ if (!sk_page_frag_refill(sk, pfrag))
goto error;
+
+ if (!skb_can_coalesce(skb, i, pfrag->page,
+ pfrag->offset)) {
+ err = -EMSGSIZE;
+ if (i == MAX_SKB_FRAGS)
+ goto error;
+
+ __skb_fill_page_desc(skb, i, pfrag->page,
+ pfrag->offset, 0);
+ skb_shinfo(skb)->nr_frags = ++i;
+ get_page(pfrag->page);
}
+ copy = min_t(int, copy, pfrag->size - pfrag->offset);
if (getfrag(from,
- skb_frag_address(frag) + skb_frag_size(frag),
- offset, copy, skb->len, skb) < 0) {
- err = -EFAULT;
- goto error;
- }
- sk->sk_sndmsg_off += copy;
- skb_frag_size_add(frag, copy);
+ page_address(pfrag->page) + pfrag->offset,
+ offset, copy, skb->len, skb) < 0)
+ goto error_efault;
+
+ pfrag->offset += copy;
+ skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
skb->len += copy;
skb->data_len += copy;
skb->truesize += copy;
@@ -1554,7 +1535,11 @@ alloc_new_skb:
offset += copy;
length -= copy;
}
+
return 0;
+
+error_efault:
+ err = -EFAULT;
error:
cork->length -= length;
IP6_INC_STATS(sock_net(sk), rt->rt6i_idev, IPSTATS_MIB_OUTDISCARDS);
diff --git a/net/sched/em_meta.c b/net/sched/em_meta.c
index 4ab6e33..7c3de6f 100644
--- a/net/sched/em_meta.c
+++ b/net/sched/em_meta.c
@@ -461,7 +461,7 @@ META_COLLECTOR(int_sk_sndtimeo)
META_COLLECTOR(int_sk_sendmsg_off)
{
SKIP_NONLOCAL(skb);
- dst->value = skb->sk->sk_sndmsg_off;
+ dst->value = skb->sk->sk_frag.offset;
}
META_COLLECTOR(int_sk_write_pend)
^ permalink raw reply related
* Re: [PATCH] inet_diag: make config INET_DIAG bool
From: Eric Dumazet @ 2012-09-24 9:42 UTC (permalink / raw)
To: Gao feng; +Cc: Jan Engelhardt, Stephen Hemminger, netdev, davem, kuznet
In-Reply-To: <50601E6C.3070106@cn.fujitsu.com>
On Mon, 2012-09-24 at 16:48 +0800, Gao feng wrote:
> 于 2012年09月23日 21:40, Jan Engelhardt 写道:
> > On Sunday 2012-09-23 06:36, Stephen Hemminger wrote:
> >
> >>> when inet_diag being compiled as module, inet_diag_handler_dump
> >>> set netlink_dump_control.dump to inet_diag_dump,so if module
> >>> inet_diag is unloaded,netlink will still try to call this function
> >>> in netlink_dump. this will cause kernel panic.
> >>
> >> Another way to handle this to just get rid of inet_diag_exit
> >> so that module can be loaded but not unloaded.
> >
> > Why don't we unset netlink_dump_control.dump on exit?
>
>
> Though I like this idea,but it may cause dead lock.
> netlink_dimp [mutex_lock(netlink_sock->cb_mutex) here]
> |->inet_diag_dump
> |->__inet_diag_dump
> |->inet_diag_lock_handler [may try to load tcp_diag here,
> so we need module_mutex lock]
>
> And on module exit,we already get module_mutex lock,
> if we unset netlink_sock->cb,we need to get mutex lock of netlink_sock->cb_mutex.
>
> I think this will cause dead lock.
>
> I don't know if I should change this patch as Stephen said,because I don't know
> witch one is better.
>
> Any comments?
In fact I didnt fully understand the problem you try to address.
If you want to prevent module being unloaded, you need to add proper
module_get()/module_put()
So I would add a "struct module *module;" in struct sock_diag_handler
and use it appropriately.
But then, I might have totally misunderstood the problem.
Care to explain how you trigger the bug ?
^ permalink raw reply
* Re: [PATCH] inet_diag: make config INET_DIAG bool
From: Gao feng @ 2012-09-24 10:17 UTC (permalink / raw)
To: Eric Dumazet; +Cc: Jan Engelhardt, Stephen Hemminger, netdev, davem, kuznet
In-Reply-To: <1348479725.26828.297.camel@edumazet-glaptop>
于 2012年09月24日 17:42, Eric Dumazet 写道:
> In fact I didnt fully understand the problem you try to address.
>
> If you want to prevent module being unloaded, you need to add proper
> module_get()/module_put()
>
> So I would add a "struct module *module;" in struct sock_diag_handler
> and use it appropriately.
Yes, I try to add reference of the module,but I can't find a proper
location to call module_get and module_put.
module_get should be called when userspace program use netlink to
send dump request to the kernel,and module_put should be called when
the dump is completed. I am right?
BUT the userspace program may only call netlink_sendmsg without call
netlink_recvmsg.so the reference of the module will be incorrect.
>
> But then, I might have totally misunderstood the problem.
>
> Care to explain how you trigger the bug ?
It's very easy to trigger this bug,
you can exec "while :; do ss -a ; done" in one terminal
and exec "while :; do rmmod tcp_diag && rmmod inet_diag; done"
in another terminal.
^ permalink raw reply
* Re: [PATCH net-next] gianfar: Change default HW Tx queue scheduling mode
From: Claudiu Manoil @ 2012-09-24 10:46 UTC (permalink / raw)
To: David Miller; +Cc: netdev, paul.gortmaker
In-Reply-To: <20120921.133528.2219558580262606729.davem@davemloft.net>
On 09/21/2012 08:35 PM, David Miller wrote:
> From: Claudiu Manoil <claudiu.manoil@freescale.com>
> Date: Thu, 20 Sep 2012 18:57:54 +0300
>
>> The TXSCHED setting may be changed at runtime, via sysfs, for devices
>> using multiple H/W Tx queues. For single queue devices this config
>> option is disabled, as the TXSCHED setting is superfluous in those cases.
>
> Please, no special driver specific sysfs knobs.
>
> Extend ethtool as necessary for your needs, so that any driver can
> provide this kind of configurability in a consistent way for users..
>
Ok.
Resent the patch addressing the watchdog timeout issue, without the
runtime configuration part.
(see http://patchwork.ozlabs.org/patch/186315/).
This config option seems to be better off in a separate patch,
especially if it would imply changes to ethtool.
What would you think of such an ethtool extension? Beside allowing the
user to choose b/w letting the H/W to service its Tx queues equally or
according to a built-in priority scheme, it could also provide api to
configure scheduling weights per Tx queue for a weighted round-robin
queue selection scheme, or, to configure only a single weight for all
the queues (for a balanced scheme). (Are there currently other drivers/
devices that would use such an option?)
Thanks,
Claudiu
^ permalink raw reply
* linux-next: manual merge of the userns tree with the tree
From: Stephen Rothwell @ 2012-09-24 11:31 UTC (permalink / raw)
To: Eric W. Biederman; +Cc: linux-next, linux-kernel, David Miller, netdev
[-- Attachment #1: Type: text/plain, Size: 2326 bytes --]
Hi Eric,
Today's linux-next merge of the userns tree got a conflict in
kernel/audit.c between commit 15e473046cb6 ("netlink: Rename pid to
portid to avoid confusion") from the net-next tree and commits
8aa14b64981e ("audit: Simply AUDIT_TTY_SET and AUDIT_TTY_GET") and
017143fecb33 ("audit: Remove the unused uid parameter from
audit_receive_filter") from the userns tree.
I fixed it up (see below) and can carry the fix as necessary (no action
is required).
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
diff --cc kernel/audit.c
index e0cf64a,511488a..0000000
--- a/kernel/audit.c
+++ b/kernel/audit.c
@@@ -782,8 -769,8 +769,8 @@@ static int audit_receive_msg(struct sk_
}
/* fallthrough */
case AUDIT_LIST:
- err = audit_receive_filter(msg_type, NETLINK_CB(skb).pid,
+ err = audit_receive_filter(msg_type, NETLINK_CB(skb).portid,
- uid, seq, data, nlmsg_len(nlh),
+ seq, data, nlmsg_len(nlh),
loginuid, sessionid, sid);
break;
case AUDIT_ADD_RULE:
@@@ -801,8 -788,8 +788,8 @@@
}
/* fallthrough */
case AUDIT_LIST_RULES:
- err = audit_receive_filter(msg_type, NETLINK_CB(skb).pid,
+ err = audit_receive_filter(msg_type, NETLINK_CB(skb).portid,
- uid, seq, data, nlmsg_len(nlh),
+ seq, data, nlmsg_len(nlh),
loginuid, sessionid, sid);
break;
case AUDIT_TRIM:
@@@ -878,21 -865,14 +865,14 @@@
break;
case AUDIT_TTY_GET: {
struct audit_tty_status s;
- struct task_struct *tsk;
- unsigned long flags;
-
- rcu_read_lock();
- tsk = find_task_by_vpid(pid);
- if (tsk && lock_task_sighand(tsk, &flags)) {
- s.enabled = tsk->signal->audit_tty != 0;
- unlock_task_sighand(tsk, &flags);
- } else
- err = -ESRCH;
- rcu_read_unlock();
-
- if (!err)
- audit_send_reply(NETLINK_CB(skb).portid, seq,
- AUDIT_TTY_GET, 0, 0, &s, sizeof(s));
+ struct task_struct *tsk = current;
+
+ spin_lock_irq(&tsk->sighand->siglock);
+ s.enabled = tsk->signal->audit_tty != 0;
+ spin_unlock_irq(&tsk->sighand->siglock);
+
- audit_send_reply(NETLINK_CB(skb).pid, seq,
++ audit_send_reply(NETLINK_CB(skb).portid, seq,
+ AUDIT_TTY_GET, 0, 0, &s, sizeof(s));
break;
}
case AUDIT_TTY_SET: {
[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply
* Re: [PATCH] inet_diag: make config INET_DIAG bool
From: Eric Dumazet @ 2012-09-24 11:32 UTC (permalink / raw)
To: Gao feng; +Cc: Jan Engelhardt, Stephen Hemminger, netdev, davem, kuznet
In-Reply-To: <50603355.4010001@cn.fujitsu.com>
On Mon, 2012-09-24 at 18:17 +0800, Gao feng wrote:
> 于 2012年09月24日 17:42, Eric Dumazet 写道:
> > In fact I didnt fully understand the problem you try to address.
> >
> > If you want to prevent module being unloaded, you need to add proper
> > module_get()/module_put()
> >
> > So I would add a "struct module *module;" in struct sock_diag_handler
> > and use it appropriately.
>
> Yes, I try to add reference of the module,but I can't find a proper
> location to call module_get and module_put.
>
> module_get should be called when userspace program use netlink to
> send dump request to the kernel,and module_put should be called when
> the dump is completed. I am right?
>
> BUT the userspace program may only call netlink_sendmsg without call
> netlink_recvmsg.so the reference of the module will be incorrect.
check ->dump() and ->done() methods
^ permalink raw reply
* [V5 PATCH 5/8] csiostor: Chelsio FCoE offload driver submission (sources part 2).
From: Naresh Kumar Inna @ 2012-09-24 17:17 UTC (permalink / raw)
To: JBottomley, linux-scsi, dm, leedom; +Cc: netdev, naresh, chethan
In-Reply-To: <1348507072-25625-1-git-send-email-naresh@chelsio.com>
This patch contains code for the FC transport template callbacks and the
Mailbox module functionality. The FC transport callbacks include Virtual
Node ports creation and deletion, FC session registration, unregistration and
teardown. The Mailbox module provides services to issue/track/cancel
mailbox commands and wrappers for them.
Signed-off-by: Naresh Kumar Inna <naresh@chelsio.com>
---
V5: Re-wrote error paths to reduce lock/unlock ambiguity.
drivers/scsi/csiostor/csio_attr.c | 804 +++++++++++++++++
drivers/scsi/csiostor/csio_mb.c | 1769 +++++++++++++++++++++++++++++++++++++
2 files changed, 2573 insertions(+), 0 deletions(-)
create mode 100644 drivers/scsi/csiostor/csio_attr.c
create mode 100644 drivers/scsi/csiostor/csio_mb.c
diff --git a/drivers/scsi/csiostor/csio_attr.c b/drivers/scsi/csiostor/csio_attr.c
new file mode 100644
index 0000000..106f9db
--- /dev/null
+++ b/drivers/scsi/csiostor/csio_attr.c
@@ -0,0 +1,804 @@
+/*
+ * 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/string.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/pci.h>
+#include <linux/mm.h>
+#include <linux/jiffies.h>
+#include <scsi/fc/fc_fs.h>
+
+#include "csio_init.h"
+
+static void
+csio_vport_set_state(struct csio_lnode *ln);
+
+/*
+ * csio_reg_rnode - Register a remote port with FC transport.
+ * @rn: Rnode representing remote port.
+ *
+ * Call fc_remote_port_add() to register this remote port with FC transport.
+ * If remote port is Initiator OR Target OR both, change the role appropriately.
+ *
+ */
+void
+csio_reg_rnode(struct csio_rnode *rn)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ struct Scsi_Host *shost = csio_ln_to_shost(ln);
+ struct fc_rport_identifiers ids;
+ struct fc_rport *rport;
+ struct csio_service_parms *sp;
+
+ ids.node_name = wwn_to_u64(csio_rn_wwnn(rn));
+ ids.port_name = wwn_to_u64(csio_rn_wwpn(rn));
+ ids.port_id = rn->nport_id;
+ ids.roles = FC_RPORT_ROLE_UNKNOWN;
+
+ if (rn->role & CSIO_RNFR_INITIATOR || rn->role & CSIO_RNFR_TARGET) {
+ rport = rn->rport;
+ CSIO_ASSERT(rport != NULL);
+ goto update_role;
+ }
+
+ rn->rport = fc_remote_port_add(shost, 0, &ids);
+ if (!rn->rport) {
+ csio_ln_err(ln, "Failed to register rport = 0x%x.\n",
+ rn->nport_id);
+ return;
+ }
+
+ ln->num_reg_rnodes++;
+ rport = rn->rport;
+ spin_lock_irq(shost->host_lock);
+ *((struct csio_rnode **)rport->dd_data) = rn;
+ spin_unlock_irq(shost->host_lock);
+
+ sp = &rn->rn_sparm;
+ rport->maxframe_size = sp->csp.sp_bb_data;
+ if (ntohs(sp->clsp[2].cp_class) & FC_CPC_VALID)
+ rport->supported_classes = FC_COS_CLASS3;
+ else
+ rport->supported_classes = FC_COS_UNSPECIFIED;
+update_role:
+ if (rn->role & CSIO_RNFR_INITIATOR)
+ ids.roles |= FC_RPORT_ROLE_FCP_INITIATOR;
+ if (rn->role & CSIO_RNFR_TARGET) {
+ ids.roles |= FC_RPORT_ROLE_FCP_TARGET;
+ ln->n_scsi_tgts++;
+ }
+
+ if (ids.roles != FC_RPORT_ROLE_UNKNOWN)
+ fc_remote_port_rolechg(rport, ids.roles);
+
+ rn->scsi_id = rport->scsi_target_id;
+
+ csio_ln_dbg(ln, "Remote port x%x role 0x%x registered\n",
+ rn->nport_id, ids.roles);
+}
+
+/*
+ * csio_unreg_rnode - Unregister a remote port with FC transport.
+ * @rn: Rnode representing remote port.
+ *
+ * Call fc_remote_port_delete() to unregister this remote port with FC
+ * transport.
+ *
+ */
+void
+csio_unreg_rnode(struct csio_rnode *rn)
+{
+ struct csio_lnode *ln = csio_rnode_to_lnode(rn);
+ struct fc_rport *rport = rn->rport;
+
+ rn->role &= ~(CSIO_RNFR_INITIATOR | CSIO_RNFR_TARGET);
+ fc_remote_port_delete(rport);
+ ln->num_reg_rnodes--;
+
+ if (ln->n_scsi_tgts)
+ ln->n_scsi_tgts--;
+
+ if (ln->last_scan_ntgts)
+ ln->last_scan_ntgts--;
+
+ csio_ln_dbg(ln, "Remote port x%x un-registered\n", rn->nport_id);
+}
+
+/*
+ * csio_lnode_async_event - Async events from local port.
+ * @ln: lnode representing local port.
+ *
+ * Async events from local node that FC transport/SCSI ML
+ * should be made aware of (Eg: RSCN).
+ */
+void
+csio_lnode_async_event(struct csio_lnode *ln, enum csio_ln_fc_evt fc_evt)
+{
+ switch (fc_evt) {
+ case CSIO_LN_FC_RSCN:
+ /* Get payload of rscn from ln */
+ /* For each RSCN entry */
+ /*
+ * fc_host_post_event(shost,
+ * fc_get_event_number(),
+ * FCH_EVT_RSCN,
+ * rscn_entry);
+ */
+ break;
+ case CSIO_LN_FC_LINKUP:
+ /* send fc_host_post_event */
+ /* set vport state */
+ if (csio_is_npiv_ln(ln))
+ csio_vport_set_state(ln);
+
+ break;
+ case CSIO_LN_FC_LINKDOWN:
+ /* send fc_host_post_event */
+ /* set vport state */
+ if (csio_is_npiv_ln(ln))
+ csio_vport_set_state(ln);
+
+ break;
+ case CSIO_LN_FC_ATTRIB_UPDATE:
+ csio_fchost_attr_init(ln);
+ break;
+ default:
+ break;
+ }
+}
+
+/*
+ * csio_fchost_attr_init - Initialize FC transport attributes
+ * @ln: Lnode.
+ *
+ */
+void
+csio_fchost_attr_init(struct csio_lnode *ln)
+{
+ struct Scsi_Host *shost = csio_ln_to_shost(ln);
+
+ fc_host_node_name(shost) = wwn_to_u64(csio_ln_wwnn(ln));
+ fc_host_port_name(shost) = wwn_to_u64(csio_ln_wwpn(ln));
+
+ fc_host_supported_classes(shost) = FC_COS_CLASS3;
+ fc_host_max_npiv_vports(shost) =
+ (csio_lnode_to_hw(ln))->fres_info.max_vnps;
+ fc_host_supported_speeds(shost) = FC_PORTSPEED_10GBIT |
+ FC_PORTSPEED_1GBIT;
+
+ fc_host_maxframe_size(shost) = ln->ln_sparm.csp.sp_bb_data;
+ memset(fc_host_supported_fc4s(shost), 0,
+ sizeof(fc_host_supported_fc4s(shost)));
+ fc_host_supported_fc4s(shost)[7] = 1;
+
+ memset(fc_host_active_fc4s(shost), 0,
+ sizeof(fc_host_active_fc4s(shost)));
+ fc_host_active_fc4s(shost)[7] = 1;
+}
+
+/*
+ * csio_get_host_port_id - sysfs entries for nport_id is
+ * populated/cached from this function
+ */
+static void
+csio_get_host_port_id(struct Scsi_Host *shost)
+{
+ struct csio_lnode *ln = shost_priv(shost);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ spin_lock_irq(&hw->lock);
+ fc_host_port_id(shost) = ln->nport_id;
+ spin_unlock_irq(&hw->lock);
+}
+
+/*
+ * csio_get_port_type - Return FC local port type.
+ * @shost: scsi host.
+ *
+ */
+static void
+csio_get_host_port_type(struct Scsi_Host *shost)
+{
+ struct csio_lnode *ln = shost_priv(shost);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ spin_lock_irq(&hw->lock);
+ if (csio_is_npiv_ln(ln))
+ fc_host_port_type(shost) = FC_PORTTYPE_NPIV;
+ else
+ fc_host_port_type(shost) = FC_PORTTYPE_NPORT;
+ spin_unlock_irq(&hw->lock);
+}
+
+/*
+ * csio_get_port_state - Return FC local port state.
+ * @shost: scsi host.
+ *
+ */
+static void
+csio_get_host_port_state(struct Scsi_Host *shost)
+{
+ struct csio_lnode *ln = shost_priv(shost);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+ char state[16];
+
+ spin_lock_irq(&hw->lock);
+
+ csio_lnode_state_to_str(ln, state);
+ if (!strcmp(state, "READY"))
+ fc_host_port_state(shost) = FC_PORTSTATE_ONLINE;
+ else if (!strcmp(state, "OFFLINE"))
+ fc_host_port_state(shost) = FC_PORTSTATE_LINKDOWN;
+ else
+ fc_host_port_state(shost) = FC_PORTSTATE_UNKNOWN;
+
+ spin_unlock_irq(&hw->lock);
+}
+
+/*
+ * csio_get_host_speed - Return link speed to FC transport.
+ * @shost: scsi host.
+ *
+ */
+static void
+csio_get_host_speed(struct Scsi_Host *shost)
+{
+ struct csio_lnode *ln = shost_priv(shost);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ spin_lock_irq(&hw->lock);
+ switch (hw->pport[ln->portid].link_speed) {
+ case FW_PORT_CAP_SPEED_1G:
+ fc_host_speed(shost) = FC_PORTSPEED_1GBIT;
+ break;
+ case FW_PORT_CAP_SPEED_10G:
+ fc_host_speed(shost) = FC_PORTSPEED_10GBIT;
+ break;
+ default:
+ fc_host_speed(shost) = FC_PORTSPEED_UNKNOWN;
+ break;
+ }
+ spin_unlock_irq(&hw->lock);
+}
+
+/*
+ * csio_get_host_fabric_name - Return fabric name
+ * @shost: scsi host.
+ *
+ */
+static void
+csio_get_host_fabric_name(struct Scsi_Host *shost)
+{
+ struct csio_lnode *ln = shost_priv(shost);
+ struct csio_rnode *rn = NULL;
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ spin_lock_irq(&hw->lock);
+ rn = csio_rnode_lookup_portid(ln, FC_FID_FLOGI);
+ if (rn)
+ fc_host_fabric_name(shost) = wwn_to_u64(csio_rn_wwnn(rn));
+ else
+ fc_host_fabric_name(shost) = 0;
+ spin_unlock_irq(&hw->lock);
+}
+
+/*
+ * csio_get_host_speed - Return FC transport statistics.
+ * @ln: Lnode.
+ *
+ */
+static struct fc_host_statistics *
+csio_get_stats(struct Scsi_Host *shost)
+{
+ struct csio_lnode *ln = shost_priv(shost);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+ struct fc_host_statistics *fhs = &ln->fch_stats;
+ struct fw_fcoe_port_stats fcoe_port_stats;
+ uint64_t seconds;
+
+ memset(&fcoe_port_stats, 0, sizeof(struct fw_fcoe_port_stats));
+ csio_get_phy_port_stats(hw, ln->portid, &fcoe_port_stats);
+
+ fhs->tx_frames += (fcoe_port_stats.tx_bcast_frames +
+ fcoe_port_stats.tx_mcast_frames +
+ fcoe_port_stats.tx_ucast_frames +
+ fcoe_port_stats.tx_offload_frames);
+ fhs->tx_words += (fcoe_port_stats.tx_bcast_bytes +
+ fcoe_port_stats.tx_mcast_bytes +
+ fcoe_port_stats.tx_ucast_bytes +
+ fcoe_port_stats.tx_offload_bytes) /
+ CSIO_WORD_TO_BYTE;
+ fhs->rx_frames += (fcoe_port_stats.rx_bcast_frames +
+ fcoe_port_stats.rx_mcast_frames +
+ fcoe_port_stats.rx_ucast_frames);
+ fhs->rx_words += (fcoe_port_stats.rx_bcast_bytes +
+ fcoe_port_stats.rx_mcast_bytes +
+ fcoe_port_stats.rx_ucast_bytes) /
+ CSIO_WORD_TO_BYTE;
+ fhs->error_frames += fcoe_port_stats.rx_err_frames;
+ fhs->fcp_input_requests += ln->stats.n_input_requests;
+ fhs->fcp_output_requests += ln->stats.n_output_requests;
+ fhs->fcp_control_requests += ln->stats.n_control_requests;
+ fhs->fcp_input_megabytes += ln->stats.n_input_bytes >> 20;
+ fhs->fcp_output_megabytes += ln->stats.n_output_bytes >> 20;
+ fhs->link_failure_count = ln->stats.n_link_down;
+ /* Reset stats for the device */
+ seconds = jiffies_to_msecs(jiffies) - hw->stats.n_reset_start;
+ do_div(seconds, 1000);
+ fhs->seconds_since_last_reset = seconds;
+
+ return fhs;
+}
+
+/*
+ * csio_set_rport_loss_tmo - Set the rport dev loss timeout
+ * @rport: fc rport.
+ * @timeout: new value for dev loss tmo.
+ *
+ * If timeout is non zero set the dev_loss_tmo to timeout, else set
+ * dev_loss_tmo to one.
+ */
+static void
+csio_set_rport_loss_tmo(struct fc_rport *rport, uint32_t timeout)
+{
+ if (timeout)
+ rport->dev_loss_tmo = timeout;
+ else
+ rport->dev_loss_tmo = 1;
+}
+
+static void
+csio_vport_set_state(struct csio_lnode *ln)
+{
+ struct fc_vport *fc_vport = ln->fc_vport;
+ struct csio_lnode *pln = ln->pln;
+ char state[16];
+
+ /* Set fc vport state based on phyiscal lnode */
+ csio_lnode_state_to_str(pln, state);
+ if (strcmp(state, "READY")) {
+ fc_vport_set_state(fc_vport, FC_VPORT_LINKDOWN);
+ return;
+ }
+
+ if (!(pln->flags & CSIO_LNF_NPIVSUPP)) {
+ fc_vport_set_state(fc_vport, FC_VPORT_NO_FABRIC_SUPP);
+ return;
+ }
+
+ /* Set fc vport state based on virtual lnode */
+ csio_lnode_state_to_str(ln, state);
+ if (strcmp(state, "READY")) {
+ fc_vport_set_state(fc_vport, FC_VPORT_LINKDOWN);
+ return;
+ }
+ fc_vport_set_state(fc_vport, FC_VPORT_ACTIVE);
+}
+
+static int
+csio_fcoe_alloc_vnp(struct csio_hw *hw, struct csio_lnode *ln)
+{
+ struct csio_lnode *pln;
+ struct csio_mb *mbp;
+ struct fw_fcoe_vnp_cmd *rsp;
+ int ret = 0;
+ int retry = 0;
+
+ /* Issue VNP cmd to alloc vport */
+ /* Allocate Mbox request */
+ spin_lock_irq(&hw->lock);
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ pln = ln->pln;
+ ln->fcf_flowid = pln->fcf_flowid;
+ ln->portid = pln->portid;
+
+ csio_fcoe_vnp_alloc_init_mb(ln, mbp, CSIO_MB_DEFAULT_TMO,
+ pln->fcf_flowid, pln->vnp_flowid, 0,
+ csio_ln_wwnn(ln), csio_ln_wwpn(ln), NULL);
+
+ for (retry = 0; retry < 3; retry++) {
+ /* FW is expected to complete vnp cmd in immediate mode
+ * without much delay.
+ * Otherwise, there will be increase in IO latency since HW
+ * lock is held till completion of vnp mbox cmd.
+ */
+ ret = csio_mb_issue(hw, mbp);
+ if (ret != -EBUSY)
+ break;
+
+ /* Retry if mbox returns busy */
+ spin_unlock_irq(&hw->lock);
+ msleep(2000);
+ spin_lock_irq(&hw->lock);
+ }
+
+ if (ret) {
+ csio_ln_err(ln, "Failed to issue mbox FCoE VNP command\n");
+ goto out_free;
+ }
+
+ /* Process Mbox response of VNP command */
+ rsp = (struct fw_fcoe_vnp_cmd *)(mbp->mb);
+ if (FW_CMD_RETVAL_GET(ntohl(rsp->alloc_to_len16)) != FW_SUCCESS) {
+ csio_ln_err(ln, "FCOE VNP ALLOC cmd returned 0x%x!\n",
+ FW_CMD_RETVAL_GET(ntohl(rsp->alloc_to_len16)));
+ ret = -EINVAL;
+ goto out_free;
+ }
+
+ ln->vnp_flowid = FW_FCOE_VNP_CMD_VNPI_GET(
+ ntohl(rsp->gen_wwn_to_vnpi));
+ memcpy(csio_ln_wwnn(ln), rsp->vnport_wwnn, 8);
+ memcpy(csio_ln_wwpn(ln), rsp->vnport_wwpn, 8);
+
+ csio_ln_dbg(ln, "FCOE VNPI: 0x%x\n", ln->vnp_flowid);
+ csio_ln_dbg(ln, "\tWWNN: %x%x%x%x%x%x%x%x\n",
+ ln->ln_sparm.wwnn[0], ln->ln_sparm.wwnn[1],
+ ln->ln_sparm.wwnn[2], ln->ln_sparm.wwnn[3],
+ ln->ln_sparm.wwnn[4], ln->ln_sparm.wwnn[5],
+ ln->ln_sparm.wwnn[6], ln->ln_sparm.wwnn[7]);
+ csio_ln_dbg(ln, "\tWWPN: %x%x%x%x%x%x%x%x\n",
+ ln->ln_sparm.wwpn[0], ln->ln_sparm.wwpn[1],
+ ln->ln_sparm.wwpn[2], ln->ln_sparm.wwpn[3],
+ ln->ln_sparm.wwpn[4], ln->ln_sparm.wwpn[5],
+ ln->ln_sparm.wwpn[6], ln->ln_sparm.wwpn[7]);
+
+out_free:
+ mempool_free(mbp, hw->mb_mempool);
+out:
+ spin_unlock_irq(&hw->lock);
+ return ret;
+}
+
+static int
+csio_fcoe_free_vnp(struct csio_hw *hw, struct csio_lnode *ln)
+{
+ struct csio_lnode *pln;
+ struct csio_mb *mbp;
+ struct fw_fcoe_vnp_cmd *rsp;
+ int ret = 0;
+ int retry = 0;
+
+ /* Issue VNP cmd to free vport */
+ /* Allocate Mbox request */
+
+ spin_lock_irq(&hw->lock);
+ mbp = mempool_alloc(hw->mb_mempool, GFP_ATOMIC);
+ if (!mbp) {
+ CSIO_INC_STATS(hw, n_err_nomem);
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ pln = ln->pln;
+
+ csio_fcoe_vnp_free_init_mb(ln, mbp, CSIO_MB_DEFAULT_TMO,
+ ln->fcf_flowid, ln->vnp_flowid,
+ NULL);
+
+ for (retry = 0; retry < 3; retry++) {
+ ret = csio_mb_issue(hw, mbp);
+ if (ret != -EBUSY)
+ break;
+
+ /* Retry if mbox returns busy */
+ spin_unlock_irq(&hw->lock);
+ msleep(2000);
+ spin_lock_irq(&hw->lock);
+ }
+
+ if (ret) {
+ csio_ln_err(ln, "Failed to issue mbox FCoE VNP command\n");
+ goto out_free;
+ }
+
+ /* Process Mbox response of VNP command */
+ rsp = (struct fw_fcoe_vnp_cmd *)(mbp->mb);
+ if (FW_CMD_RETVAL_GET(ntohl(rsp->alloc_to_len16)) != FW_SUCCESS) {
+ csio_ln_err(ln, "FCOE VNP FREE cmd returned 0x%x!\n",
+ FW_CMD_RETVAL_GET(ntohl(rsp->alloc_to_len16)));
+ ret = -EINVAL;
+ }
+
+out_free:
+ mempool_free(mbp, hw->mb_mempool);
+out:
+ spin_unlock_irq(&hw->lock);
+ return ret;
+}
+
+static int
+csio_vport_create(struct fc_vport *fc_vport, bool disable)
+{
+ struct Scsi_Host *shost = fc_vport->shost;
+ struct csio_lnode *pln = shost_priv(shost);
+ struct csio_lnode *ln = NULL;
+ struct csio_hw *hw = csio_lnode_to_hw(pln);
+ uint8_t wwn[8];
+ int ret = -1;
+
+ ln = csio_shost_init(hw, &fc_vport->dev, false, pln);
+ if (!ln)
+ goto error;
+
+ if (fc_vport->node_name != 0) {
+ u64_to_wwn(fc_vport->node_name, wwn);
+
+ if (!CSIO_VALID_WWN(wwn)) {
+ csio_ln_err(ln,
+ "vport create failed. Invalid wwnn\n");
+ goto error;
+ }
+ memcpy(csio_ln_wwnn(ln), wwn, 8);
+ }
+
+ if (fc_vport->port_name != 0) {
+ u64_to_wwn(fc_vport->port_name, wwn);
+
+ if (!CSIO_VALID_WWN(wwn)) {
+ csio_ln_err(ln,
+ "vport create failed. Invalid wwpn\n");
+ goto error;
+ }
+
+ if (csio_lnode_lookup_by_wwpn(hw, wwn)) {
+ csio_ln_err(ln,
+ "vport create failed. wwpn already exists\n");
+ goto error;
+ }
+ memcpy(csio_ln_wwpn(ln), wwn, 8);
+ }
+
+ fc_vport_set_state(fc_vport, FC_VPORT_INITIALIZING);
+
+ if (csio_fcoe_alloc_vnp(hw, ln))
+ goto error;
+
+ *(struct csio_lnode **)fc_vport->dd_data = ln;
+ ln->fc_vport = fc_vport;
+ if (!fc_vport->node_name)
+ fc_vport->node_name = wwn_to_u64(csio_ln_wwnn(ln));
+ if (!fc_vport->port_name)
+ fc_vport->port_name = wwn_to_u64(csio_ln_wwpn(ln));
+ csio_fchost_attr_init(ln);
+ return 0;
+error:
+ if (ln)
+ csio_shost_exit(ln);
+
+ return ret;
+}
+
+static int
+csio_vport_delete(struct fc_vport *fc_vport)
+{
+ struct csio_lnode *ln = *(struct csio_lnode **)fc_vport->dd_data;
+ struct Scsi_Host *shost = csio_ln_to_shost(ln);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+ int rmv;
+
+ spin_lock_irq(&hw->lock);
+ rmv = csio_is_hw_removing(hw);
+ spin_unlock_irq(&hw->lock);
+
+ if (rmv) {
+ csio_shost_exit(ln);
+ return 0;
+ }
+
+ /* Quiesce ios and send remove event to lnode */
+ scsi_block_requests(shost);
+ spin_lock_irq(&hw->lock);
+ csio_scsim_cleanup_io_lnode(csio_hw_to_scsim(hw), ln);
+ csio_lnode_close(ln);
+ spin_unlock_irq(&hw->lock);
+ scsi_unblock_requests(shost);
+
+ /* Free vnp */
+ if (fc_vport->vport_state != FC_VPORT_DISABLED)
+ csio_fcoe_free_vnp(hw, ln);
+
+ csio_shost_exit(ln);
+ return 0;
+}
+
+static int
+csio_vport_disable(struct fc_vport *fc_vport, bool disable)
+{
+ struct csio_lnode *ln = *(struct csio_lnode **)fc_vport->dd_data;
+ struct Scsi_Host *shost = csio_ln_to_shost(ln);
+ struct csio_hw *hw = csio_lnode_to_hw(ln);
+
+ /* disable vport */
+ if (disable) {
+ /* Quiesce ios and send stop event to lnode */
+ scsi_block_requests(shost);
+ spin_lock_irq(&hw->lock);
+ csio_scsim_cleanup_io_lnode(csio_hw_to_scsim(hw), ln);
+ csio_lnode_stop(ln);
+ spin_unlock_irq(&hw->lock);
+ scsi_unblock_requests(shost);
+
+ /* Free vnp */
+ csio_fcoe_free_vnp(hw, ln);
+ fc_vport_set_state(fc_vport, FC_VPORT_DISABLED);
+ csio_ln_err(ln, "vport disabled\n");
+ return 0;
+ } else {
+ /* enable vport */
+ fc_vport_set_state(fc_vport, FC_VPORT_INITIALIZING);
+ if (csio_fcoe_alloc_vnp(hw, ln)) {
+ csio_ln_err(ln, "vport enabled failed.\n");
+ return -1;
+ }
+ csio_ln_err(ln, "vport enabled\n");
+ return 0;
+ }
+}
+
+static void
+csio_dev_loss_tmo_callbk(struct fc_rport *rport)
+{
+ struct csio_rnode *rn;
+ struct csio_hw *hw;
+ struct csio_lnode *ln;
+
+ rn = *((struct csio_rnode **)rport->dd_data);
+ ln = csio_rnode_to_lnode(rn);
+ hw = csio_lnode_to_hw(ln);
+
+ spin_lock_irq(&hw->lock);
+
+ /* return if driver is being removed or same rnode comes back online */
+ if (csio_is_hw_removing(hw) || csio_is_rnode_ready(rn))
+ goto out;
+
+ csio_ln_dbg(ln, "devloss timeout on rnode:%p portid:x%x flowid:x%x\n",
+ rn, rn->nport_id, csio_rn_flowid(rn));
+
+ CSIO_INC_STATS(ln, n_dev_loss_tmo);
+
+ /*
+ * enqueue devloss event to event worker thread to serialize all
+ * rnode events.
+ */
+ if (csio_enqueue_evt(hw, CSIO_EVT_DEV_LOSS, &rn, sizeof(rn))) {
+ CSIO_INC_STATS(hw, n_evt_drop);
+ goto out;
+ }
+
+ if (!(hw->flags & CSIO_HWF_FWEVT_PENDING)) {
+ hw->flags |= CSIO_HWF_FWEVT_PENDING;
+ spin_unlock_irq(&hw->lock);
+ schedule_work(&hw->evtq_work);
+ return;
+ }
+
+out:
+ spin_unlock_irq(&hw->lock);
+}
+
+/* FC transport functions template - Physical port */
+struct fc_function_template csio_fc_transport_funcs = {
+ .show_host_node_name = 1,
+ .show_host_port_name = 1,
+ .show_host_supported_classes = 1,
+ .show_host_supported_fc4s = 1,
+ .show_host_maxframe_size = 1,
+
+ .get_host_port_id = csio_get_host_port_id,
+ .show_host_port_id = 1,
+
+ .get_host_port_type = csio_get_host_port_type,
+ .show_host_port_type = 1,
+
+ .get_host_port_state = csio_get_host_port_state,
+ .show_host_port_state = 1,
+
+ .show_host_active_fc4s = 1,
+ .get_host_speed = csio_get_host_speed,
+ .show_host_speed = 1,
+ .get_host_fabric_name = csio_get_host_fabric_name,
+ .show_host_fabric_name = 1,
+
+ .get_fc_host_stats = csio_get_stats,
+
+ .dd_fcrport_size = sizeof(struct csio_rnode *),
+ .show_rport_maxframe_size = 1,
+ .show_rport_supported_classes = 1,
+
+ .set_rport_dev_loss_tmo = csio_set_rport_loss_tmo,
+ .show_rport_dev_loss_tmo = 1,
+
+ .show_starget_port_id = 1,
+ .show_starget_node_name = 1,
+ .show_starget_port_name = 1,
+
+ .dev_loss_tmo_callbk = csio_dev_loss_tmo_callbk,
+ .dd_fcvport_size = sizeof(struct csio_lnode *),
+
+ .vport_create = csio_vport_create,
+ .vport_disable = csio_vport_disable,
+ .vport_delete = csio_vport_delete,
+};
+
+/* FC transport functions template - Virtual port */
+struct fc_function_template csio_fc_transport_vport_funcs = {
+ .show_host_node_name = 1,
+ .show_host_port_name = 1,
+ .show_host_supported_classes = 1,
+ .show_host_supported_fc4s = 1,
+ .show_host_maxframe_size = 1,
+
+ .get_host_port_id = csio_get_host_port_id,
+ .show_host_port_id = 1,
+
+ .get_host_port_type = csio_get_host_port_type,
+ .show_host_port_type = 1,
+
+ .get_host_port_state = csio_get_host_port_state,
+ .show_host_port_state = 1,
+ .show_host_active_fc4s = 1,
+
+ .get_host_speed = csio_get_host_speed,
+ .show_host_speed = 1,
+
+ .get_host_fabric_name = csio_get_host_fabric_name,
+ .show_host_fabric_name = 1,
+
+ .get_fc_host_stats = csio_get_stats,
+
+ .dd_fcrport_size = sizeof(struct csio_rnode *),
+ .show_rport_maxframe_size = 1,
+ .show_rport_supported_classes = 1,
+
+ .set_rport_dev_loss_tmo = csio_set_rport_loss_tmo,
+ .show_rport_dev_loss_tmo = 1,
+
+ .show_starget_port_id = 1,
+ .show_starget_node_name = 1,
+ .show_starget_port_name = 1,
+
+ .dev_loss_tmo_callbk = csio_dev_loss_tmo_callbk,
+
+};
diff --git a/drivers/scsi/csiostor/csio_mb.c b/drivers/scsi/csiostor/csio_mb.c
new file mode 100644
index 0000000..100afd1
--- /dev/null
+++ b/drivers/scsi/csiostor/csio_mb.c
@@ -0,0 +1,1769 @@
+/*
+ * 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/delay.h>
+#include <linux/jiffies.h>
+#include <linux/string.h>
+#include <scsi/scsi_transport_fc.h>
+
+#include "csio_hw.h"
+#include "csio_lnode.h"
+#include "csio_rnode.h"
+#include "csio_mb.h"
+#include "csio_wr.h"
+
+#define csio_mb_is_host_owner(__owner) ((__owner) == CSIO_MBOWNER_PL)
+
+/* MB Command/Response Helpers */
+/*
+ * csio_mb_fw_retval - FW return value from a mailbox response.
+ * @mbp: Mailbox structure
+ *
+ */
+enum fw_retval
+csio_mb_fw_retval(struct csio_mb *mbp)
+{
+ struct fw_cmd_hdr *hdr;
+
+ hdr = (struct fw_cmd_hdr *)(mbp->mb);
+
+ return FW_CMD_RETVAL_GET(ntohl(hdr->lo));
+}
+
+/*
+ * csio_mb_hello - FW HELLO command helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @m_mbox: Master mailbox number, if any.
+ * @a_mbox: Mailbox number for asycn notifications.
+ * @master: Device mastership.
+ * @cbfn: Callback, if any.
+ *
+ */
+void
+csio_mb_hello(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo,
+ uint32_t m_mbox, uint32_t a_mbox, enum csio_dev_master master,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_hello_cmd *cmdp = (struct fw_hello_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, 1);
+
+ cmdp->op_to_write = htonl(FW_CMD_OP(FW_HELLO_CMD) |
+ FW_CMD_REQUEST | FW_CMD_WRITE);
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ cmdp->err_to_clearinit = htonl(
+ FW_HELLO_CMD_MASTERDIS(master == CSIO_MASTER_CANT) |
+ FW_HELLO_CMD_MASTERFORCE(master == CSIO_MASTER_MUST) |
+ FW_HELLO_CMD_MBMASTER(master == CSIO_MASTER_MUST ?
+ m_mbox : FW_HELLO_CMD_MBMASTER_MASK) |
+ FW_HELLO_CMD_MBASYNCNOT(a_mbox) |
+ FW_HELLO_CMD_STAGE(FW_HELLO_CMD_STAGE_OS) |
+ FW_HELLO_CMD_CLEARINIT);
+
+}
+
+/*
+ * csio_mb_process_hello_rsp - FW HELLO response processing helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @retval: Mailbox return value from Firmware
+ * @state: State that the function is in.
+ * @mpfn: Master pfn
+ *
+ */
+void
+csio_mb_process_hello_rsp(struct csio_hw *hw, struct csio_mb *mbp,
+ enum fw_retval *retval, enum csio_dev_state *state,
+ uint8_t *mpfn)
+{
+ struct fw_hello_cmd *rsp = (struct fw_hello_cmd *)(mbp->mb);
+ uint32_t value;
+
+ *retval = FW_CMD_RETVAL_GET(ntohl(rsp->retval_len16));
+
+ if (*retval == FW_SUCCESS) {
+ hw->fwrev = ntohl(rsp->fwrev);
+
+ value = ntohl(rsp->err_to_clearinit);
+ *mpfn = FW_HELLO_CMD_MBMASTER_GET(value);
+
+ if (value & FW_HELLO_CMD_INIT)
+ *state = CSIO_DEV_STATE_INIT;
+ else if (value & FW_HELLO_CMD_ERR)
+ *state = CSIO_DEV_STATE_ERR;
+ else
+ *state = CSIO_DEV_STATE_UNINIT;
+ }
+}
+
+/*
+ * csio_mb_bye - FW BYE command helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @cbfn: Callback, if any.
+ *
+ */
+void
+csio_mb_bye(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_bye_cmd *cmdp = (struct fw_bye_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, 1);
+
+ cmdp->op_to_write = htonl(FW_CMD_OP(FW_BYE_CMD) |
+ FW_CMD_REQUEST | FW_CMD_WRITE);
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+}
+
+/*
+ * csio_mb_reset - FW RESET command helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @reset: Type of reset.
+ * @cbfn: Callback, if any.
+ *
+ */
+void
+csio_mb_reset(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo,
+ int reset, int halt,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_reset_cmd *cmdp = (struct fw_reset_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, 1);
+
+ cmdp->op_to_write = htonl(FW_CMD_OP(FW_RESET_CMD) |
+ FW_CMD_REQUEST | FW_CMD_WRITE);
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ cmdp->val = htonl(reset);
+ cmdp->halt_pkd = htonl(halt);
+
+}
+
+/*
+ * csio_mb_params - FW PARAMS command helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @tmo: Command timeout.
+ * @pf: PF number.
+ * @vf: VF number.
+ * @nparams: Number of paramters
+ * @params: Parameter mnemonic array.
+ * @val: Parameter value array.
+ * @wr: Write/Read PARAMS.
+ * @cbfn: Callback, if any.
+ *
+ */
+void
+csio_mb_params(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo,
+ unsigned int pf, unsigned int vf, unsigned int nparams,
+ const u32 *params, u32 *val, bool wr,
+ void (*cbfn)(struct csio_hw *, struct csio_mb *))
+{
+ uint32_t i;
+ uint32_t temp_params = 0, temp_val = 0;
+ struct fw_params_cmd *cmdp = (struct fw_params_cmd *)(mbp->mb);
+ __be32 *p = &cmdp->param[0].mnem;
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, 1);
+
+ cmdp->op_to_vfn = htonl(FW_CMD_OP(FW_PARAMS_CMD) |
+ FW_CMD_REQUEST |
+ (wr ? FW_CMD_WRITE : FW_CMD_READ) |
+ FW_PARAMS_CMD_PFN(pf) |
+ FW_PARAMS_CMD_VFN(vf));
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+ /* Write Params */
+ if (wr) {
+ while (nparams--) {
+ temp_params = *params++;
+ temp_val = *val++;
+
+ *p++ = htonl(temp_params);
+ *p++ = htonl(temp_val);
+ }
+ } else {
+ for (i = 0; i < nparams; i++, p += 2) {
+ temp_params = *params++;
+ *p = htonl(temp_params);
+ }
+ }
+
+}
+
+/*
+ * csio_mb_process_read_params_rsp - FW PARAMS response processing helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @retval: Mailbox return value from Firmware
+ * @nparams: Number of parameters
+ * @val: Parameter value array.
+ *
+ */
+void
+csio_mb_process_read_params_rsp(struct csio_hw *hw, struct csio_mb *mbp,
+ enum fw_retval *retval, unsigned int nparams,
+ u32 *val)
+{
+ struct fw_params_cmd *rsp = (struct fw_params_cmd *)(mbp->mb);
+ uint32_t i;
+ __be32 *p = &rsp->param[0].val;
+
+ *retval = FW_CMD_RETVAL_GET(ntohl(rsp->retval_len16));
+
+ if (*retval == FW_SUCCESS)
+ for (i = 0; i < nparams; i++, p += 2)
+ *val++ = ntohl(*p);
+}
+
+/*
+ * csio_mb_ldst - FW LDST command
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @tmo: timeout
+ * @reg: register
+ *
+ */
+void
+csio_mb_ldst(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo, int reg)
+{
+ struct fw_ldst_cmd *ldst_cmd = (struct fw_ldst_cmd *)(mbp->mb);
+ CSIO_INIT_MBP(mbp, ldst_cmd, tmo, hw, NULL, 1);
+
+ /*
+ * Construct and send the Firmware LDST Command to retrieve the
+ * specified PCI-E Configuration Space register.
+ */
+ ldst_cmd->op_to_addrspace =
+ htonl(FW_CMD_OP(FW_LDST_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_READ |
+ FW_LDST_CMD_ADDRSPACE(FW_LDST_ADDRSPC_FUNC_PCIE));
+ ldst_cmd->cycles_to_len16 = htonl(FW_LEN16(struct fw_ldst_cmd));
+ ldst_cmd->u.pcie.select_naccess = FW_LDST_CMD_NACCESS(1);
+ ldst_cmd->u.pcie.ctrl_to_fn =
+ (FW_LDST_CMD_LC | FW_LDST_CMD_FN(hw->pfn));
+ ldst_cmd->u.pcie.r = (uint8_t)reg;
+}
+
+/*
+ *
+ * csio_mb_caps_config - FW Read/Write Capabilities command helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @wr: Write if 1, Read if 0
+ * @init: Turn on initiator mode.
+ * @tgt: Turn on target mode.
+ * @cofld: If 1, Control Offload for FCoE
+ * @cbfn: Callback, if any.
+ *
+ * This helper assumes that cmdp has MB payload from a previous CAPS
+ * read command.
+ */
+void
+csio_mb_caps_config(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo,
+ bool wr, bool init, bool tgt, bool cofld,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_caps_config_cmd *cmdp =
+ (struct fw_caps_config_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, wr ? 0 : 1);
+
+ cmdp->op_to_write = htonl(FW_CMD_OP(FW_CAPS_CONFIG_CMD) |
+ FW_CMD_REQUEST |
+ (wr ? FW_CMD_WRITE : FW_CMD_READ));
+ cmdp->cfvalid_to_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+ /* Read config */
+ if (!wr)
+ return;
+
+ /* Write config */
+ cmdp->fcoecaps = 0;
+
+ if (cofld)
+ cmdp->fcoecaps |= htons(FW_CAPS_CONFIG_FCOE_CTRL_OFLD);
+ if (init)
+ cmdp->fcoecaps |= htons(FW_CAPS_CONFIG_FCOE_INITIATOR);
+ if (tgt)
+ cmdp->fcoecaps |= htons(FW_CAPS_CONFIG_FCOE_TARGET);
+}
+
+void
+csio_rss_glb_config(struct csio_hw *hw, struct csio_mb *mbp,
+ uint32_t tmo, uint8_t mode, unsigned int flags,
+ void (*cbfn)(struct csio_hw *, struct csio_mb *))
+{
+ struct fw_rss_glb_config_cmd *cmdp =
+ (struct fw_rss_glb_config_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, 1);
+
+ cmdp->op_to_write = htonl(FW_CMD_OP(FW_RSS_GLB_CONFIG_CMD) |
+ FW_CMD_REQUEST | FW_CMD_WRITE);
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+ if (mode == FW_RSS_GLB_CONFIG_CMD_MODE_MANUAL) {
+ cmdp->u.manual.mode_pkd =
+ htonl(FW_RSS_GLB_CONFIG_CMD_MODE(mode));
+ } else if (mode == FW_RSS_GLB_CONFIG_CMD_MODE_BASICVIRTUAL) {
+ cmdp->u.basicvirtual.mode_pkd =
+ htonl(FW_RSS_GLB_CONFIG_CMD_MODE(mode));
+ cmdp->u.basicvirtual.synmapen_to_hashtoeplitz = htonl(flags);
+ }
+}
+
+
+/*
+ * csio_mb_pfvf - FW Write PF/VF capabilities command helper.
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @pf:
+ * @vf:
+ * @txq:
+ * @txq_eht_ctrl:
+ * @rxqi:
+ * @rxq:
+ * @tc:
+ * @vi:
+ * @pmask:
+ * @rcaps:
+ * @wxcaps:
+ * @cbfn: Callback, if any.
+ *
+ */
+void
+csio_mb_pfvf(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo,
+ unsigned int pf, unsigned int vf, unsigned int txq,
+ unsigned int txq_eth_ctrl, unsigned int rxqi,
+ unsigned int rxq, unsigned int tc, unsigned int vi,
+ unsigned int cmask, unsigned int pmask, unsigned int nexactf,
+ unsigned int rcaps, unsigned int wxcaps,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_pfvf_cmd *cmdp = (struct fw_pfvf_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, 1);
+
+ cmdp->op_to_vfn = htonl(FW_CMD_OP(FW_PFVF_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_WRITE |
+ FW_PFVF_CMD_PFN(pf) |
+ FW_PFVF_CMD_VFN(vf));
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ cmdp->niqflint_niq = htonl(FW_PFVF_CMD_NIQFLINT(rxqi) |
+ FW_PFVF_CMD_NIQ(rxq));
+
+ cmdp->type_to_neq = htonl(FW_PFVF_CMD_TYPE |
+ FW_PFVF_CMD_CMASK(cmask) |
+ FW_PFVF_CMD_PMASK(pmask) |
+ FW_PFVF_CMD_NEQ(txq));
+ cmdp->tc_to_nexactf = htonl(FW_PFVF_CMD_TC(tc) |
+ FW_PFVF_CMD_NVI(vi) |
+ FW_PFVF_CMD_NEXACTF(nexactf));
+ cmdp->r_caps_to_nethctrl = htonl(FW_PFVF_CMD_R_CAPS(rcaps) |
+ FW_PFVF_CMD_WX_CAPS(wxcaps) |
+ FW_PFVF_CMD_NETHCTRL(txq_eth_ctrl));
+}
+
+#define CSIO_ADVERT_MASK (FW_PORT_CAP_SPEED_100M | FW_PORT_CAP_SPEED_1G |\
+ FW_PORT_CAP_SPEED_10G | FW_PORT_CAP_ANEG)
+
+/*
+ * csio_mb_port- FW PORT command helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @tmo: COmmand timeout
+ * @portid: Port ID to get/set info
+ * @wr: Write/Read PORT information.
+ * @fc: Flow control
+ * @caps: Port capabilites to set.
+ * @cbfn: Callback, if any.
+ *
+ */
+void
+csio_mb_port(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo,
+ uint8_t portid, bool wr, uint32_t fc, uint16_t caps,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_port_cmd *cmdp = (struct fw_port_cmd *)(mbp->mb);
+ unsigned int lfc = 0, mdi = FW_PORT_MDI(FW_PORT_MDI_AUTO);
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, 1);
+
+ cmdp->op_to_portid = htonl(FW_CMD_OP(FW_PORT_CMD) |
+ FW_CMD_REQUEST |
+ (wr ? FW_CMD_EXEC : FW_CMD_READ) |
+ FW_PORT_CMD_PORTID(portid));
+ if (!wr) {
+ cmdp->action_to_len16 = htonl(
+ FW_PORT_CMD_ACTION(FW_PORT_ACTION_GET_PORT_INFO) |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ return;
+ }
+
+ /* Set port */
+ cmdp->action_to_len16 = htonl(
+ FW_PORT_CMD_ACTION(FW_PORT_ACTION_L1_CFG) |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+ if (fc & PAUSE_RX)
+ lfc |= FW_PORT_CAP_FC_RX;
+ if (fc & PAUSE_TX)
+ lfc |= FW_PORT_CAP_FC_TX;
+
+ if (!(caps & FW_PORT_CAP_ANEG))
+ cmdp->u.l1cfg.rcap = htonl((caps & CSIO_ADVERT_MASK) | lfc);
+ else
+ cmdp->u.l1cfg.rcap = htonl((caps & CSIO_ADVERT_MASK) |
+ lfc | mdi);
+}
+
+/*
+ * csio_mb_process_read_port_rsp - FW PORT command response processing helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @retval: Mailbox return value from Firmware
+ * @caps: port capabilities
+ *
+ */
+void
+csio_mb_process_read_port_rsp(struct csio_hw *hw, struct csio_mb *mbp,
+ enum fw_retval *retval, uint16_t *caps)
+{
+ struct fw_port_cmd *rsp = (struct fw_port_cmd *)(mbp->mb);
+
+ *retval = FW_CMD_RETVAL_GET(ntohl(rsp->action_to_len16));
+
+ if (*retval == FW_SUCCESS)
+ *caps = ntohs(rsp->u.info.pcap);
+}
+
+/*
+ * csio_mb_initialize - FW INITIALIZE command helper
+ * @hw: The HW structure
+ * @mbp: Mailbox structure
+ * @tmo: COmmand timeout
+ * @cbfn: Callback, if any.
+ *
+ */
+void
+csio_mb_initialize(struct csio_hw *hw, struct csio_mb *mbp, uint32_t tmo,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_initialize_cmd *cmdp = (struct fw_initialize_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, tmo, hw, cbfn, 1);
+
+ cmdp->op_to_write = htonl(FW_CMD_OP(FW_INITIALIZE_CMD) |
+ FW_CMD_REQUEST | FW_CMD_WRITE);
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+}
+
+/*
+ * csio_mb_iq_alloc - Initializes the mailbox to allocate an
+ * Ingress DMA queue in the firmware.
+ *
+ * @hw: The hw structure
+ * @mbp: Mailbox structure to initialize
+ * @priv: Private object
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @iq_params: Ingress queue params needed for allocation.
+ * @cbfn: The call-back function
+ *
+ *
+ */
+static void
+csio_mb_iq_alloc(struct csio_hw *hw, struct csio_mb *mbp, void *priv,
+ uint32_t mb_tmo, struct csio_iq_params *iq_params,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_iq_cmd *cmdp = (struct fw_iq_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, priv, cbfn, 1);
+
+ cmdp->op_to_vfn = htonl(FW_CMD_OP(FW_IQ_CMD) |
+ FW_CMD_REQUEST | FW_CMD_EXEC |
+ FW_IQ_CMD_PFN(iq_params->pfn) |
+ FW_IQ_CMD_VFN(iq_params->vfn));
+
+ cmdp->alloc_to_len16 = htonl(FW_IQ_CMD_ALLOC |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+ cmdp->type_to_iqandstindex = htonl(
+ FW_IQ_CMD_VIID(iq_params->viid) |
+ FW_IQ_CMD_TYPE(iq_params->type) |
+ FW_IQ_CMD_IQASYNCH(iq_params->iqasynch));
+
+ cmdp->fl0size = htons(iq_params->fl0size);
+ cmdp->fl0size = htons(iq_params->fl1size);
+
+} /* csio_mb_iq_alloc */
+
+/*
+ * csio_mb_iq_write - Initializes the mailbox for writing into an
+ * Ingress DMA Queue.
+ *
+ * @hw: The HW structure
+ * @mbp: Mailbox structure to initialize
+ * @priv: Private object
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @cascaded_req: TRUE - if this request is cascased with iq-alloc request.
+ * @iq_params: Ingress queue params needed for writing.
+ * @cbfn: The call-back function
+ *
+ * NOTE: We OR relevant bits with cmdp->XXX, instead of just equating,
+ * because this IQ write request can be cascaded with a previous
+ * IQ alloc request, and we dont want to over-write the bits set by
+ * that request. This logic will work even in a non-cascaded case, since the
+ * cmdp structure is zeroed out by CSIO_INIT_MBP.
+ */
+static void
+csio_mb_iq_write(struct csio_hw *hw, struct csio_mb *mbp, void *priv,
+ uint32_t mb_tmo, bool cascaded_req,
+ struct csio_iq_params *iq_params,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_iq_cmd *cmdp = (struct fw_iq_cmd *)(mbp->mb);
+
+ uint32_t iq_start_stop = (iq_params->iq_start) ?
+ FW_IQ_CMD_IQSTART(1) :
+ FW_IQ_CMD_IQSTOP(1);
+
+ /*
+ * If this IQ write is cascaded with IQ alloc request, do not
+ * re-initialize with 0's.
+ *
+ */
+ if (!cascaded_req)
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, priv, cbfn, 1);
+
+ cmdp->op_to_vfn |= htonl(FW_CMD_OP(FW_IQ_CMD) |
+ FW_CMD_REQUEST | FW_CMD_WRITE |
+ FW_IQ_CMD_PFN(iq_params->pfn) |
+ FW_IQ_CMD_VFN(iq_params->vfn));
+ cmdp->alloc_to_len16 |= htonl(iq_start_stop |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ cmdp->iqid |= htons(iq_params->iqid);
+ cmdp->fl0id |= htons(iq_params->fl0id);
+ cmdp->fl1id |= htons(iq_params->fl1id);
+ cmdp->type_to_iqandstindex |= htonl(
+ FW_IQ_CMD_IQANDST(iq_params->iqandst) |
+ FW_IQ_CMD_IQANUS(iq_params->iqanus) |
+ FW_IQ_CMD_IQANUD(iq_params->iqanud) |
+ FW_IQ_CMD_IQANDSTINDEX(iq_params->iqandstindex));
+ cmdp->iqdroprss_to_iqesize |= htons(
+ FW_IQ_CMD_IQPCIECH(iq_params->iqpciech) |
+ FW_IQ_CMD_IQDCAEN(iq_params->iqdcaen) |
+ FW_IQ_CMD_IQDCACPU(iq_params->iqdcacpu) |
+ FW_IQ_CMD_IQINTCNTTHRESH(iq_params->iqintcntthresh) |
+ FW_IQ_CMD_IQCPRIO(iq_params->iqcprio) |
+ FW_IQ_CMD_IQESIZE(iq_params->iqesize));
+
+ cmdp->iqsize |= htons(iq_params->iqsize);
+ cmdp->iqaddr |= cpu_to_be64(iq_params->iqaddr);
+
+ if (iq_params->type == 0) {
+ cmdp->iqns_to_fl0congen |= htonl(
+ FW_IQ_CMD_IQFLINTIQHSEN(iq_params->iqflintiqhsen)|
+ FW_IQ_CMD_IQFLINTCONGEN(iq_params->iqflintcongen));
+ }
+
+ if (iq_params->fl0size && iq_params->fl0addr &&
+ (iq_params->fl0id != 0xFFFF)) {
+
+ cmdp->iqns_to_fl0congen |= htonl(
+ FW_IQ_CMD_FL0HOSTFCMODE(iq_params->fl0hostfcmode)|
+ FW_IQ_CMD_FL0CPRIO(iq_params->fl0cprio) |
+ FW_IQ_CMD_FL0PADEN(iq_params->fl0paden) |
+ FW_IQ_CMD_FL0PACKEN(iq_params->fl0packen));
+ cmdp->fl0dcaen_to_fl0cidxfthresh |= htons(
+ FW_IQ_CMD_FL0DCAEN(iq_params->fl0dcaen) |
+ FW_IQ_CMD_FL0DCACPU(iq_params->fl0dcacpu) |
+ FW_IQ_CMD_FL0FBMIN(iq_params->fl0fbmin) |
+ FW_IQ_CMD_FL0FBMAX(iq_params->fl0fbmax) |
+ FW_IQ_CMD_FL0CIDXFTHRESH(iq_params->fl0cidxfthresh));
+ cmdp->fl0size |= htons(iq_params->fl0size);
+ cmdp->fl0addr |= cpu_to_be64(iq_params->fl0addr);
+ }
+} /* csio_mb_iq_write */
+
+/*
+ * csio_mb_iq_alloc_write - Initializes the mailbox for allocating an
+ * Ingress DMA Queue.
+ *
+ * @hw: The HW structure
+ * @mbp: Mailbox structure to initialize
+ * @priv: Private data.
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @iq_params: Ingress queue params needed for allocation & writing.
+ * @cbfn: The call-back function
+ *
+ *
+ */
+void
+csio_mb_iq_alloc_write(struct csio_hw *hw, struct csio_mb *mbp, void *priv,
+ uint32_t mb_tmo, struct csio_iq_params *iq_params,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ csio_mb_iq_alloc(hw, mbp, priv, mb_tmo, iq_params, cbfn);
+ csio_mb_iq_write(hw, mbp, priv, mb_tmo, true, iq_params, cbfn);
+} /* csio_mb_iq_alloc_write */
+
+/*
+ * csio_mb_iq_alloc_write_rsp - Process the allocation & writing
+ * of ingress DMA queue mailbox's response.
+ *
+ * @hw: The HW structure.
+ * @mbp: Mailbox structure to initialize.
+ * @retval: Firmware return value.
+ * @iq_params: Ingress queue parameters, after allocation and write.
+ *
+ */
+void
+csio_mb_iq_alloc_write_rsp(struct csio_hw *hw, struct csio_mb *mbp,
+ enum fw_retval *ret_val,
+ struct csio_iq_params *iq_params)
+{
+ struct fw_iq_cmd *rsp = (struct fw_iq_cmd *)(mbp->mb);
+
+ *ret_val = FW_CMD_RETVAL_GET(ntohl(rsp->alloc_to_len16));
+ if (*ret_val == FW_SUCCESS) {
+ iq_params->physiqid = ntohs(rsp->physiqid);
+ iq_params->iqid = ntohs(rsp->iqid);
+ iq_params->fl0id = ntohs(rsp->fl0id);
+ iq_params->fl1id = ntohs(rsp->fl1id);
+ } else {
+ iq_params->physiqid = iq_params->iqid =
+ iq_params->fl0id = iq_params->fl1id = 0;
+ }
+} /* csio_mb_iq_alloc_write_rsp */
+
+/*
+ * csio_mb_iq_free - Initializes the mailbox for freeing a
+ * specified Ingress DMA Queue.
+ *
+ * @hw: The HW structure
+ * @mbp: Mailbox structure to initialize
+ * @priv: Private data
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @iq_params: Parameters of ingress queue, that is to be freed.
+ * @cbfn: The call-back function
+ *
+ *
+ */
+void
+csio_mb_iq_free(struct csio_hw *hw, struct csio_mb *mbp, void *priv,
+ uint32_t mb_tmo, struct csio_iq_params *iq_params,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_iq_cmd *cmdp = (struct fw_iq_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, priv, cbfn, 1);
+
+ cmdp->op_to_vfn = htonl(FW_CMD_OP(FW_IQ_CMD) |
+ FW_CMD_REQUEST | FW_CMD_EXEC |
+ FW_IQ_CMD_PFN(iq_params->pfn) |
+ FW_IQ_CMD_VFN(iq_params->vfn));
+ cmdp->alloc_to_len16 = htonl(FW_IQ_CMD_FREE |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ cmdp->type_to_iqandstindex = htonl(FW_IQ_CMD_TYPE(iq_params->type));
+
+ cmdp->iqid = htons(iq_params->iqid);
+ cmdp->fl0id = htons(iq_params->fl0id);
+ cmdp->fl1id = htons(iq_params->fl1id);
+
+} /* csio_mb_iq_free */
+
+/*
+ * csio_mb_eq_ofld_alloc - Initializes the mailbox for allocating
+ * an offload-egress queue.
+ *
+ * @hw: The HW structure
+ * @mbp: Mailbox structure to initialize
+ * @priv: Private data
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @eq_ofld_params: (Offload) Egress queue paramters.
+ * @cbfn: The call-back function
+ *
+ *
+ */
+static void
+csio_mb_eq_ofld_alloc(struct csio_hw *hw, struct csio_mb *mbp, void *priv,
+ uint32_t mb_tmo, struct csio_eq_params *eq_ofld_params,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_eq_ofld_cmd *cmdp = (struct fw_eq_ofld_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, priv, cbfn, 1);
+ cmdp->op_to_vfn = htonl(FW_CMD_OP(FW_EQ_OFLD_CMD) |
+ FW_CMD_REQUEST | FW_CMD_EXEC |
+ FW_EQ_OFLD_CMD_PFN(eq_ofld_params->pfn) |
+ FW_EQ_OFLD_CMD_VFN(eq_ofld_params->vfn));
+ cmdp->alloc_to_len16 = htonl(FW_EQ_OFLD_CMD_ALLOC |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+} /* csio_mb_eq_ofld_alloc */
+
+/*
+ * csio_mb_eq_ofld_write - Initializes the mailbox for writing
+ * an alloacted offload-egress queue.
+ *
+ * @hw: The HW structure
+ * @mbp: Mailbox structure to initialize
+ * @priv: Private data
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @cascaded_req: TRUE - if this request is cascased with Eq-alloc request.
+ * @eq_ofld_params: (Offload) Egress queue paramters.
+ * @cbfn: The call-back function
+ *
+ *
+ * NOTE: We OR relevant bits with cmdp->XXX, instead of just equating,
+ * because this EQ write request can be cascaded with a previous
+ * EQ alloc request, and we dont want to over-write the bits set by
+ * that request. This logic will work even in a non-cascaded case, since the
+ * cmdp structure is zeroed out by CSIO_INIT_MBP.
+ */
+static void
+csio_mb_eq_ofld_write(struct csio_hw *hw, struct csio_mb *mbp, void *priv,
+ uint32_t mb_tmo, bool cascaded_req,
+ struct csio_eq_params *eq_ofld_params,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_eq_ofld_cmd *cmdp = (struct fw_eq_ofld_cmd *)(mbp->mb);
+
+ uint32_t eq_start_stop = (eq_ofld_params->eqstart) ?
+ FW_EQ_OFLD_CMD_EQSTART : FW_EQ_OFLD_CMD_EQSTOP;
+
+ /*
+ * If this EQ write is cascaded with EQ alloc request, do not
+ * re-initialize with 0's.
+ *
+ */
+ if (!cascaded_req)
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, priv, cbfn, 1);
+
+ cmdp->op_to_vfn |= htonl(FW_CMD_OP(FW_EQ_OFLD_CMD) |
+ FW_CMD_REQUEST | FW_CMD_WRITE |
+ FW_EQ_OFLD_CMD_PFN(eq_ofld_params->pfn) |
+ FW_EQ_OFLD_CMD_VFN(eq_ofld_params->vfn));
+ cmdp->alloc_to_len16 |= htonl(eq_start_stop |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+ cmdp->eqid_pkd |= htonl(FW_EQ_OFLD_CMD_EQID(eq_ofld_params->eqid));
+
+ cmdp->fetchszm_to_iqid |= htonl(
+ FW_EQ_OFLD_CMD_HOSTFCMODE(eq_ofld_params->hostfcmode) |
+ FW_EQ_OFLD_CMD_CPRIO(eq_ofld_params->cprio) |
+ FW_EQ_OFLD_CMD_PCIECHN(eq_ofld_params->pciechn) |
+ FW_EQ_OFLD_CMD_IQID(eq_ofld_params->iqid));
+
+ cmdp->dcaen_to_eqsize |= htonl(
+ FW_EQ_OFLD_CMD_DCAEN(eq_ofld_params->dcaen) |
+ FW_EQ_OFLD_CMD_DCACPU(eq_ofld_params->dcacpu) |
+ FW_EQ_OFLD_CMD_FBMIN(eq_ofld_params->fbmin) |
+ FW_EQ_OFLD_CMD_FBMAX(eq_ofld_params->fbmax) |
+ FW_EQ_OFLD_CMD_CIDXFTHRESHO(eq_ofld_params->cidxfthresho) |
+ FW_EQ_OFLD_CMD_CIDXFTHRESH(eq_ofld_params->cidxfthresh) |
+ FW_EQ_OFLD_CMD_EQSIZE(eq_ofld_params->eqsize));
+
+ cmdp->eqaddr |= cpu_to_be64(eq_ofld_params->eqaddr);
+
+} /* csio_mb_eq_ofld_write */
+
+/*
+ * csio_mb_eq_ofld_alloc_write - Initializes the mailbox for allocation
+ * writing into an Engress DMA Queue.
+ *
+ * @hw: The HW structure
+ * @mbp: Mailbox structure to initialize
+ * @priv: Private data.
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @eq_ofld_params: (Offload) Egress queue paramters.
+ * @cbfn: The call-back function
+ *
+ *
+ */
+void
+csio_mb_eq_ofld_alloc_write(struct csio_hw *hw, struct csio_mb *mbp,
+ void *priv, uint32_t mb_tmo,
+ struct csio_eq_params *eq_ofld_params,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ csio_mb_eq_ofld_alloc(hw, mbp, priv, mb_tmo, eq_ofld_params, cbfn);
+ csio_mb_eq_ofld_write(hw, mbp, priv, mb_tmo, true,
+ eq_ofld_params, cbfn);
+} /* csio_mb_eq_ofld_alloc_write */
+
+/*
+ * csio_mb_eq_ofld_alloc_write_rsp - Process the allocation
+ * & write egress DMA queue mailbox's response.
+ *
+ * @hw: The HW structure.
+ * @mbp: Mailbox structure to initialize.
+ * @retval: Firmware return value.
+ * @eq_ofld_params: (Offload) Egress queue paramters.
+ *
+ */
+void
+csio_mb_eq_ofld_alloc_write_rsp(struct csio_hw *hw,
+ struct csio_mb *mbp, enum fw_retval *ret_val,
+ struct csio_eq_params *eq_ofld_params)
+{
+ struct fw_eq_ofld_cmd *rsp = (struct fw_eq_ofld_cmd *)(mbp->mb);
+
+ *ret_val = FW_CMD_RETVAL_GET(ntohl(rsp->alloc_to_len16));
+
+ if (*ret_val == FW_SUCCESS) {
+ eq_ofld_params->eqid = FW_EQ_OFLD_CMD_EQID_GET(
+ ntohl(rsp->eqid_pkd));
+ eq_ofld_params->physeqid = FW_EQ_OFLD_CMD_PHYSEQID_GET(
+ ntohl(rsp->physeqid_pkd));
+ } else
+ eq_ofld_params->eqid = 0;
+
+} /* csio_mb_eq_ofld_alloc_write_rsp */
+
+/*
+ * csio_mb_eq_ofld_free - Initializes the mailbox for freeing a
+ * specified Engress DMA Queue.
+ *
+ * @hw: The HW structure
+ * @mbp: Mailbox structure to initialize
+ * @priv: Private data area.
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @eq_ofld_params: (Offload) Egress queue paramters, that is to be freed.
+ * @cbfn: The call-back function
+ *
+ *
+ */
+void
+csio_mb_eq_ofld_free(struct csio_hw *hw, struct csio_mb *mbp, void *priv,
+ uint32_t mb_tmo, struct csio_eq_params *eq_ofld_params,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_eq_ofld_cmd *cmdp = (struct fw_eq_ofld_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, priv, cbfn, 1);
+
+ cmdp->op_to_vfn = htonl(FW_CMD_OP(FW_EQ_OFLD_CMD) |
+ FW_CMD_REQUEST | FW_CMD_EXEC |
+ FW_EQ_OFLD_CMD_PFN(eq_ofld_params->pfn) |
+ FW_EQ_OFLD_CMD_VFN(eq_ofld_params->vfn));
+ cmdp->alloc_to_len16 = htonl(FW_EQ_OFLD_CMD_FREE |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ cmdp->eqid_pkd = htonl(FW_EQ_OFLD_CMD_EQID(eq_ofld_params->eqid));
+
+} /* csio_mb_eq_ofld_free */
+
+/*
+ * csio_write_fcoe_link_cond_init_mb - Initialize Mailbox to write FCoE link
+ * condition.
+ *
+ * @ln: The Lnode structure
+ * @mbp: Mailbox structure to initialize
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @cbfn: The call back function.
+ *
+ *
+ */
+void
+csio_write_fcoe_link_cond_init_mb(struct csio_lnode *ln, struct csio_mb *mbp,
+ uint32_t mb_tmo, uint8_t port_id, uint32_t sub_opcode,
+ uint8_t cos, bool link_status, uint32_t fcfi,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_fcoe_link_cmd *cmdp =
+ (struct fw_fcoe_link_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, ln, cbfn, 1);
+
+ cmdp->op_to_portid = htonl((
+ FW_CMD_OP(FW_FCOE_LINK_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_WRITE |
+ FW_FCOE_LINK_CMD_PORTID(port_id)));
+ cmdp->sub_opcode_fcfi = htonl(
+ FW_FCOE_LINK_CMD_SUB_OPCODE(sub_opcode) |
+ FW_FCOE_LINK_CMD_FCFI(fcfi));
+ cmdp->lstatus = link_status;
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+} /* csio_write_fcoe_link_cond_init_mb */
+
+/*
+ * csio_fcoe_read_res_info_init_mb - Initializes the mailbox for reading FCoE
+ * resource information(FW_GET_RES_INFO_CMD).
+ *
+ * @hw: The HW structure
+ * @mbp: Mailbox structure to initialize
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @cbfn: The call-back function
+ *
+ *
+ */
+void
+csio_fcoe_read_res_info_init_mb(struct csio_hw *hw, struct csio_mb *mbp,
+ uint32_t mb_tmo,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_fcoe_res_info_cmd *cmdp =
+ (struct fw_fcoe_res_info_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, hw, cbfn, 1);
+
+ cmdp->op_to_read = htonl((FW_CMD_OP(FW_FCOE_RES_INFO_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_READ));
+
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+} /* csio_fcoe_read_res_info_init_mb */
+
+/*
+ * csio_fcoe_vnp_alloc_init_mb - Initializes the mailbox for allocating VNP
+ * in the firmware (FW_FCOE_VNP_CMD).
+ *
+ * @ln: The Lnode structure.
+ * @mbp: Mailbox structure to initialize.
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @fcfi: FCF Index.
+ * @vnpi: vnpi
+ * @iqid: iqid
+ * @vnport_wwnn: vnport WWNN
+ * @vnport_wwpn: vnport WWPN
+ * @cbfn: The call-back function.
+ *
+ *
+ */
+void
+csio_fcoe_vnp_alloc_init_mb(struct csio_lnode *ln, struct csio_mb *mbp,
+ uint32_t mb_tmo, uint32_t fcfi, uint32_t vnpi, uint16_t iqid,
+ uint8_t vnport_wwnn[8], uint8_t vnport_wwpn[8],
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_fcoe_vnp_cmd *cmdp =
+ (struct fw_fcoe_vnp_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, ln, cbfn, 1);
+
+ cmdp->op_to_fcfi = htonl((FW_CMD_OP(FW_FCOE_VNP_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_EXEC |
+ FW_FCOE_VNP_CMD_FCFI(fcfi)));
+
+ cmdp->alloc_to_len16 = htonl(FW_FCOE_VNP_CMD_ALLOC |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+ cmdp->gen_wwn_to_vnpi = htonl(FW_FCOE_VNP_CMD_VNPI(vnpi));
+
+ cmdp->iqid = htons(iqid);
+
+ if (!wwn_to_u64(vnport_wwnn) && !wwn_to_u64(vnport_wwpn))
+ cmdp->gen_wwn_to_vnpi |= htonl(FW_FCOE_VNP_CMD_GEN_WWN);
+
+ if (vnport_wwnn)
+ memcpy(cmdp->vnport_wwnn, vnport_wwnn, 8);
+ if (vnport_wwpn)
+ memcpy(cmdp->vnport_wwpn, vnport_wwpn, 8);
+
+} /* csio_fcoe_vnp_alloc_init_mb */
+
+/*
+ * csio_fcoe_vnp_read_init_mb - Prepares VNP read cmd.
+ * @ln: The Lnode structure.
+ * @mbp: Mailbox structure to initialize.
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @fcfi: FCF Index.
+ * @vnpi: vnpi
+ * @cbfn: The call-back handler.
+ */
+void
+csio_fcoe_vnp_read_init_mb(struct csio_lnode *ln, struct csio_mb *mbp,
+ uint32_t mb_tmo, uint32_t fcfi, uint32_t vnpi,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_fcoe_vnp_cmd *cmdp =
+ (struct fw_fcoe_vnp_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, ln, cbfn, 1);
+ cmdp->op_to_fcfi = htonl(FW_CMD_OP(FW_FCOE_VNP_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_READ |
+ FW_FCOE_VNP_CMD_FCFI(fcfi));
+ cmdp->alloc_to_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ cmdp->gen_wwn_to_vnpi = htonl(FW_FCOE_VNP_CMD_VNPI(vnpi));
+}
+
+/*
+ * csio_fcoe_vnp_free_init_mb - Initializes the mailbox for freeing an
+ * alloacted VNP in the firmware (FW_FCOE_VNP_CMD).
+ *
+ * @ln: The Lnode structure.
+ * @mbp: Mailbox structure to initialize.
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @fcfi: FCF flow id
+ * @vnpi: VNP flow id
+ * @cbfn: The call-back function.
+ * Return: None
+ */
+void
+csio_fcoe_vnp_free_init_mb(struct csio_lnode *ln, struct csio_mb *mbp,
+ uint32_t mb_tmo, uint32_t fcfi, uint32_t vnpi,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_fcoe_vnp_cmd *cmdp =
+ (struct fw_fcoe_vnp_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, ln, cbfn, 1);
+
+ cmdp->op_to_fcfi = htonl(FW_CMD_OP(FW_FCOE_VNP_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_EXEC |
+ FW_FCOE_VNP_CMD_FCFI(fcfi));
+ cmdp->alloc_to_len16 = htonl(FW_FCOE_VNP_CMD_FREE |
+ FW_CMD_LEN16(sizeof(*cmdp) / 16));
+ cmdp->gen_wwn_to_vnpi = htonl(FW_FCOE_VNP_CMD_VNPI(vnpi));
+}
+
+/*
+ * csio_fcoe_read_fcf_init_mb - Initializes the mailbox to read the
+ * FCF records.
+ *
+ * @ln: The Lnode structure
+ * @mbp: Mailbox structure to initialize
+ * @mb_tmo: Mailbox time-out period (in ms).
+ * @fcf_params: FC-Forwarder parameters.
+ * @cbfn: The call-back function
+ *
+ *
+ */
+void
+csio_fcoe_read_fcf_init_mb(struct csio_lnode *ln, struct csio_mb *mbp,
+ uint32_t mb_tmo, uint32_t portid, uint32_t fcfi,
+ void (*cbfn) (struct csio_hw *, struct csio_mb *))
+{
+ struct fw_fcoe_fcf_cmd *cmdp =
+ (struct fw_fcoe_fcf_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, ln, cbfn, 1);
+
+ cmdp->op_to_fcfi = htonl(FW_CMD_OP(FW_FCOE_FCF_CMD) |
+ FW_CMD_REQUEST |
+ FW_CMD_READ |
+ FW_FCOE_FCF_CMD_FCFI(fcfi));
+ cmdp->retval_len16 = htonl(FW_CMD_LEN16(sizeof(*cmdp) / 16));
+
+} /* csio_fcoe_read_fcf_init_mb */
+
+void
+csio_fcoe_read_portparams_init_mb(struct csio_hw *hw, struct csio_mb *mbp,
+ uint32_t mb_tmo,
+ struct fw_fcoe_port_cmd_params *portparams,
+ void (*cbfn)(struct csio_hw *,
+ struct csio_mb *))
+{
+ struct fw_fcoe_stats_cmd *cmdp = (struct fw_fcoe_stats_cmd *)(mbp->mb);
+
+ CSIO_INIT_MBP(mbp, cmdp, mb_tmo, hw, cbfn, 1);
+ mbp->mb_size = 64;
+
+ cmdp->op_to_flowid = htonl(FW_CMD_OP(FW_FCOE_STATS_CMD) |
+ FW_CMD_REQUEST | FW_CMD_READ);
+ cmdp->free_to_len16 = htonl(FW_CMD_LEN16(CSIO_MAX_MB_SIZE/16));
+
+ cmdp->u.ctl.nstats_port = FW_FCOE_STATS_CMD_NSTATS(portparams->nstats) |
+ FW_FCOE_STATS_CMD_PORT(portparams->portid);
+
+ cmdp->u.ctl.port_valid_ix = FW_FCOE_STATS_CMD_IX(portparams->idx) |
+ FW_FCOE_STATS_CMD_PORT_VALID;
+
+} /* csio_fcoe_read_portparams_init_mb */
+
+void
+csio_mb_process_portparams_rsp(
+ struct csio_hw *hw,
+ struct csio_mb *mbp,
+ enum fw_retval *retval,
+ struct fw_fcoe_port_cmd_params *portparams,
+ struct fw_fcoe_port_stats *portstats
+ )
+{
+ struct fw_fcoe_stats_cmd *rsp = (struct fw_fcoe_stats_cmd *)(mbp->mb);
+ struct fw_fcoe_port_stats stats;
+ uint8_t *src;
+ uint8_t *dst;
+
+ *retval = FW_CMD_RETVAL_GET(ntohl(rsp->free_to_len16));
+
+ memset(&stats, 0, sizeof(struct fw_fcoe_port_stats));
+
+ if (*retval == FW_SUCCESS) {
+ dst = (uint8_t *)(&stats) + ((portparams->idx - 1) * 8);
+ src = (uint8_t *)rsp + (CSIO_STATS_OFFSET * 8);
+ memcpy(dst, src, (portparams->nstats * 8));
+ if (portparams->idx == 1) {
+ /* Get the first 6 flits from the Mailbox */
+ portstats->tx_bcast_bytes =
+ be64_to_cpu(stats.tx_bcast_bytes);
+ portstats->tx_bcast_frames =
+ be64_to_cpu(stats.tx_bcast_frames);
+ portstats->tx_mcast_bytes =
+ be64_to_cpu(stats.tx_mcast_bytes);
+ portstats->tx_mcast_frames =
+ be64_to_cpu(stats.tx_mcast_frames);
+ portstats->tx_ucast_bytes =
+ be64_to_cpu(stats.tx_ucast_bytes);
+ portstats->tx_ucast_frames =
+ be64_to_cpu(stats.tx_ucast_frames);
+ }
+ if (portparams->idx == 7) {
+ /* Get the second 6 flits from the Mailbox */
+ portstats->tx_drop_frames =
+ be64_to_cpu(stats.tx_drop_frames);
+ portstats->tx_offload_bytes =
+ be64_to_cpu(stats.tx_offload_bytes);
+ portstats->tx_offload_frames =
+ be64_to_cpu(stats.tx_offload_frames);
+#if 0
+ portstats->rx_pf_bytes =
+ be64_to_cpu(stats.rx_pf_bytes);
+ portstats->rx_pf_frames =
+ be64_to_cpu(stats.rx_pf_frames);
+#endif
+ portstats->rx_bcast_bytes =
+ be64_to_cpu(stats.rx_bcast_bytes);
+ portstats->rx_bcast_frames =
+ be64_to_cpu(stats.rx_bcast_frames);
+ portstats->rx_mcast_bytes =
+ be64_to_cpu(stats.rx_mcast_bytes);
+ }
+ if (portparams->idx == 13) {
+ /* Get the last 4 flits from the Mailbox */
+ portstats->rx_mcast_frames =
+ be64_to_cpu(stats.rx_mcast_frames);
+ portstats->rx_ucast_bytes =
+ be64_to_cpu(stats.rx_ucast_bytes);
+ portstats->rx_ucast_frames =
+ be64_to_cpu(stats.rx_ucast_frames);
+ portstats->rx_err_frames =
+ be64_to_cpu(stats.rx_err_frames);
+ }
+ }
+}
+
+/* Entry points/APIs for MB module */
+/*
+ * csio_mb_intr_enable - Enable Interrupts from mailboxes.
+ * @hw: The HW structure
+ *
+ * Enables CIM interrupt bit in appropriate INT_ENABLE registers.
+ */
+void
+csio_mb_intr_enable(struct csio_hw *hw)
+{
+ csio_wr_reg32(hw, MBMSGRDYINTEN(1), MYPF_REG(CIM_PF_HOST_INT_ENABLE));
+ csio_rd_reg32(hw, MYPF_REG(CIM_PF_HOST_INT_ENABLE));
+}
+
+/*
+ * csio_mb_intr_disable - Disable Interrupts from mailboxes.
+ * @hw: The HW structure
+ *
+ * Disable bit in HostInterruptEnable CIM register.
+ */
+void
+csio_mb_intr_disable(struct csio_hw *hw)
+{
+ csio_wr_reg32(hw, MBMSGRDYINTEN(0), MYPF_REG(CIM_PF_HOST_INT_ENABLE));
+ csio_rd_reg32(hw, MYPF_REG(CIM_PF_HOST_INT_ENABLE));
+}
+
+static void
+csio_mb_dump_fw_dbg(struct csio_hw *hw, __be64 *cmd)
+{
+ struct fw_debug_cmd *dbg = (struct fw_debug_cmd *)cmd;
+
+ if ((FW_DEBUG_CMD_TYPE_GET(ntohl(dbg->op_type))) == 1) {
+ csio_info(hw, "FW print message:\n");
+ csio_info(hw, "\tdebug->dprtstridx = %d\n",
+ ntohs(dbg->u.prt.dprtstridx));
+ csio_info(hw, "\tdebug->dprtstrparam0 = 0x%x\n",
+ ntohl(dbg->u.prt.dprtstrparam0));
+ csio_info(hw, "\tdebug->dprtstrparam1 = 0x%x\n",
+ ntohl(dbg->u.prt.dprtstrparam1));
+ csio_info(hw, "\tdebug->dprtstrparam2 = 0x%x\n",
+ ntohl(dbg->u.prt.dprtstrparam2));
+ csio_info(hw, "\tdebug->dprtstrparam3 = 0x%x\n",
+ ntohl(dbg->u.prt.dprtstrparam3));
+ } else {
+ /* This is a FW assertion */
+ csio_fatal(hw, "FW assertion at %.16s:%u, val0 %#x, val1 %#x\n",
+ dbg->u.assert.filename_0_7,
+ ntohl(dbg->u.assert.line),
+ ntohl(dbg->u.assert.x),
+ ntohl(dbg->u.assert.y));
+ }
+}
+
+static void
+csio_mb_debug_cmd_handler(struct csio_hw *hw)
+{
+ int i;
+ __be64 cmd[CSIO_MB_MAX_REGS];
+ uint32_t ctl_reg = PF_REG(hw->pfn, CIM_PF_MAILBOX_CTRL);
+ uint32_t data_reg = PF_REG(hw->pfn, CIM_PF_MAILBOX_DATA);
+ int size = sizeof(struct fw_debug_cmd);
+
+ /* Copy mailbox data */
+ for (i = 0; i < size; i += 8)
+ cmd[i / 8] = cpu_to_be64(csio_rd_reg64(hw, data_reg + i));
+
+ csio_mb_dump_fw_dbg(hw, cmd);
+
+ /* Notify FW of mailbox by setting owner as UP */
+ csio_wr_reg32(hw, MBMSGVALID | MBINTREQ | MBOWNER(CSIO_MBOWNER_FW),
+ ctl_reg);
+
+ csio_rd_reg32(hw, ctl_reg);
+ wmb();
+}
+
+/*
+ * csio_mb_issue - generic routine for issuing Mailbox commands.
+ * @hw: The HW structure
+ * @mbp: Mailbox command to issue
+ *
+ * Caller should hold hw lock across this call.
+ */
+int
+csio_mb_issue(struct csio_hw *hw, struct csio_mb *mbp)
+{
+ uint32_t owner, ctl;
+ int i;
+ uint32_t ii;
+ __be64 *cmd = mbp->mb;
+ __be64 hdr;
+ struct csio_mbm *mbm = &hw->mbm;
+ uint32_t ctl_reg = PF_REG(hw->pfn, CIM_PF_MAILBOX_CTRL);
+ uint32_t data_reg = PF_REG(hw->pfn, CIM_PF_MAILBOX_DATA);
+ int size = mbp->mb_size;
+ int rv = -EINVAL;
+ struct fw_cmd_hdr *fw_hdr;
+
+ /* Determine mode */
+ if (mbp->mb_cbfn == NULL) {
+ /* Need to issue/get results in the same context */
+ if (mbp->tmo < CSIO_MB_POLL_FREQ) {
+ csio_err(hw, "Invalid tmo: 0x%x\n", mbp->tmo);
+ goto error_out;
+ }
+ } else if (!csio_is_host_intr_enabled(hw) ||
+ !csio_is_hw_intr_enabled(hw)) {
+ csio_err(hw, "Cannot issue mailbox in interrupt mode 0x%x\n",
+ *((uint8_t *)mbp->mb));
+ goto error_out;
+ }
+
+ if (mbm->mcurrent != NULL) {
+ /* Queue mbox cmd, if another mbox cmd is active */
+ if (mbp->mb_cbfn == NULL) {
+ rv = -EBUSY;
+ csio_dbg(hw, "Couldnt own Mailbox %x op:0x%x\n",
+ hw->pfn, *((uint8_t *)mbp->mb));
+
+ goto error_out;
+ } else {
+ list_add_tail(&mbp->list, &mbm->req_q);
+ CSIO_INC_STATS(mbm, n_activeq);
+
+ return 0;
+ }
+ }
+
+ /* Now get ownership of mailbox */
+ owner = MBOWNER_GET(csio_rd_reg32(hw, ctl_reg));
+
+ if (!csio_mb_is_host_owner(owner)) {
+
+ for (i = 0; (owner == CSIO_MBOWNER_NONE) && (i < 3); i++)
+ owner = MBOWNER_GET(csio_rd_reg32(hw, ctl_reg));
+ /*
+ * Mailbox unavailable. In immediate mode, fail the command.
+ * In other modes, enqueue the request.
+ */
+ if (!csio_mb_is_host_owner(owner)) {
+ if (mbp->mb_cbfn == NULL) {
+ rv = owner ? -EBUSY : -ETIMEDOUT;
+
+ csio_dbg(hw,
+ "Couldnt own Mailbox %x op:0x%x "
+ "owner:%x\n",
+ hw->pfn, *((uint8_t *)mbp->mb), owner);
+ goto error_out;
+ } else {
+ if (mbm->mcurrent == NULL) {
+ csio_err(hw,
+ "Couldnt own Mailbox %x "
+ "op:0x%x owner:%x\n",
+ hw->pfn, *((uint8_t *)mbp->mb),
+ owner);
+ csio_err(hw,
+ "No outstanding driver"
+ " mailbox as well\n");
+ goto error_out;
+ }
+ }
+ }
+ }
+
+ /* Mailbox is available, copy mailbox data into it */
+ for (i = 0; i < size; i += 8) {
+ csio_wr_reg64(hw, be64_to_cpu(*cmd), data_reg + i);
+ cmd++;
+ }
+
+ CSIO_DUMP_MB(hw, hw->pfn, data_reg);
+
+ /* Start completion timers in non-immediate modes and notify FW */
+ if (mbp->mb_cbfn != NULL) {
+ mbm->mcurrent = mbp;
+ mod_timer(&mbm->timer, jiffies + msecs_to_jiffies(mbp->tmo));
+ csio_wr_reg32(hw, MBMSGVALID | MBINTREQ |
+ MBOWNER(CSIO_MBOWNER_FW), ctl_reg);
+ } else
+ csio_wr_reg32(hw, MBMSGVALID | MBOWNER(CSIO_MBOWNER_FW),
+ ctl_reg);
+
+ /* Flush posted writes */
+ csio_rd_reg32(hw, ctl_reg);
+ wmb();
+
+ CSIO_INC_STATS(mbm, n_req);
+
+ if (mbp->mb_cbfn)
+ return 0;
+
+ /* Poll for completion in immediate mode */
+ cmd = mbp->mb;
+
+ for (ii = 0; ii < mbp->tmo; ii += CSIO_MB_POLL_FREQ) {
+ mdelay(CSIO_MB_POLL_FREQ);
+
+ /* Check for response */
+ ctl = csio_rd_reg32(hw, ctl_reg);
+ if (csio_mb_is_host_owner(MBOWNER_GET(ctl))) {
+
+ if (!(ctl & MBMSGVALID)) {
+ csio_wr_reg32(hw, 0, ctl_reg);
+ continue;
+ }
+
+ CSIO_DUMP_MB(hw, hw->pfn, data_reg);
+
+ hdr = cpu_to_be64(csio_rd_reg64(hw, data_reg));
+ fw_hdr = (struct fw_cmd_hdr *)&hdr;
+
+ switch (FW_CMD_OP_GET(ntohl(fw_hdr->hi))) {
+ case FW_DEBUG_CMD:
+ csio_mb_debug_cmd_handler(hw);
+ continue;
+ }
+
+ /* Copy response */
+ for (i = 0; i < size; i += 8)
+ *cmd++ = cpu_to_be64(csio_rd_reg64
+ (hw, data_reg + i));
+ csio_wr_reg32(hw, 0, ctl_reg);
+
+ if (FW_CMD_RETVAL_GET(*(mbp->mb)))
+ CSIO_INC_STATS(mbm, n_err);
+
+ CSIO_INC_STATS(mbm, n_rsp);
+ return 0;
+ }
+ }
+
+ CSIO_INC_STATS(mbm, n_tmo);
+
+ csio_err(hw, "Mailbox %x op:0x%x timed out!\n",
+ hw->pfn, *((uint8_t *)cmd));
+
+ return -ETIMEDOUT;
+
+error_out:
+ CSIO_INC_STATS(mbm, n_err);
+ return rv;
+}
+
+/*
+ * csio_mb_completions - Completion handler for Mailbox commands
+ * @hw: The HW structure
+ * @cbfn_q: Completion queue.
+ *
+ */
+void
+csio_mb_completions(struct csio_hw *hw, struct list_head *cbfn_q)
+{
+ struct csio_mb *mbp;
+ struct csio_mbm *mbm = &hw->mbm;
+ enum fw_retval rv;
+
+ while (!list_empty(cbfn_q)) {
+ mbp = list_first_entry(cbfn_q, struct csio_mb, list);
+ list_del_init(&mbp->list);
+
+ rv = csio_mb_fw_retval(mbp);
+ if ((rv != FW_SUCCESS) && (rv != FW_HOSTERROR))
+ CSIO_INC_STATS(mbm, n_err);
+ else if (rv != FW_HOSTERROR)
+ CSIO_INC_STATS(mbm, n_rsp);
+
+ if (mbp->mb_cbfn)
+ mbp->mb_cbfn(hw, mbp);
+ }
+}
+
+static void
+csio_mb_portmod_changed(struct csio_hw *hw, uint8_t port_id)
+{
+ static char *mod_str[] = {
+ NULL, "LR", "SR", "ER", "TWINAX", "active TWINAX", "LRM"
+ };
+
+ struct csio_pport *port = &hw->pport[port_id];
+
+ if (port->mod_type == FW_PORT_MOD_TYPE_NONE)
+ csio_info(hw, "Port:%d - port module unplugged\n", port_id);
+ else if (port->mod_type < ARRAY_SIZE(mod_str))
+ csio_info(hw, "Port:%d - %s port module inserted\n", port_id,
+ mod_str[port->mod_type]);
+ else if (port->mod_type == FW_PORT_MOD_TYPE_NOTSUPPORTED)
+ csio_info(hw,
+ "Port:%d - unsupported optical port module "
+ "inserted\n", port_id);
+ else if (port->mod_type == FW_PORT_MOD_TYPE_UNKNOWN)
+ csio_info(hw,
+ "Port:%d - unknown port module inserted, forcing "
+ "TWINAX\n", port_id);
+ else if (port->mod_type == FW_PORT_MOD_TYPE_ERROR)
+ csio_info(hw, "Port:%d - transceiver module error\n", port_id);
+ else
+ csio_info(hw, "Port:%d - unknown module type %d inserted\n",
+ port_id, port->mod_type);
+}
+
+int
+csio_mb_fwevt_handler(struct csio_hw *hw, __be64 *cmd)
+{
+ uint8_t opcode = *(uint8_t *)cmd;
+ struct fw_port_cmd *pcmd;
+ uint8_t port_id;
+ uint32_t link_status;
+ uint16_t action;
+ uint8_t mod_type;
+
+ if (opcode == FW_PORT_CMD) {
+ pcmd = (struct fw_port_cmd *)cmd;
+ port_id = FW_PORT_CMD_PORTID_GET(
+ ntohl(pcmd->op_to_portid));
+ action = FW_PORT_CMD_ACTION_GET(
+ ntohl(pcmd->action_to_len16));
+ if (action != FW_PORT_ACTION_GET_PORT_INFO) {
+ csio_err(hw, "Unhandled FW_PORT_CMD action: %u\n",
+ action);
+ return -EINVAL;
+ }
+
+ link_status = ntohl(pcmd->u.info.lstatus_to_modtype);
+ mod_type = FW_PORT_CMD_MODTYPE_GET(link_status);
+
+ hw->pport[port_id].link_status =
+ FW_PORT_CMD_LSTATUS_GET(link_status);
+ hw->pport[port_id].link_speed =
+ FW_PORT_CMD_LSPEED_GET(link_status);
+
+ csio_info(hw, "Port:%x - LINK %s\n", port_id,
+ FW_PORT_CMD_LSTATUS_GET(link_status) ? "UP" : "DOWN");
+
+ if (mod_type != hw->pport[port_id].mod_type) {
+ hw->pport[port_id].mod_type = mod_type;
+ csio_mb_portmod_changed(hw, port_id);
+ }
+ } else if (opcode == FW_DEBUG_CMD) {
+ csio_mb_dump_fw_dbg(hw, cmd);
+ } else {
+ csio_dbg(hw, "Gen MB can't handle op:0x%x on evtq.\n", opcode);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/*
+ * csio_mb_isr_handler - Handle mailboxes related interrupts.
+ * @hw: The HW structure
+ *
+ * Called from the ISR to handle Mailbox related interrupts.
+ * HW Lock should be held across this call.
+ */
+int
+csio_mb_isr_handler(struct csio_hw *hw)
+{
+ struct csio_mbm *mbm = &hw->mbm;
+ struct csio_mb *mbp = mbm->mcurrent;
+ __be64 *cmd;
+ uint32_t ctl, cim_cause, pl_cause;
+ int i;
+ uint32_t ctl_reg = PF_REG(hw->pfn, CIM_PF_MAILBOX_CTRL);
+ uint32_t data_reg = PF_REG(hw->pfn, CIM_PF_MAILBOX_DATA);
+ int size;
+ __be64 hdr;
+ struct fw_cmd_hdr *fw_hdr;
+
+ pl_cause = csio_rd_reg32(hw, MYPF_REG(PL_PF_INT_CAUSE));
+ cim_cause = csio_rd_reg32(hw, MYPF_REG(CIM_PF_HOST_INT_CAUSE));
+
+ if (!(pl_cause & PFCIM) || !(cim_cause & MBMSGRDYINT)) {
+ CSIO_INC_STATS(hw, n_mbint_unexp);
+ return -EINVAL;
+ }
+
+ /*
+ * The cause registers below HAVE to be cleared in the SAME
+ * order as below: The low level cause register followed by
+ * the upper level cause register. In other words, CIM-cause
+ * first followed by PL-Cause next.
+ */
+ csio_wr_reg32(hw, MBMSGRDYINT, MYPF_REG(CIM_PF_HOST_INT_CAUSE));
+ csio_wr_reg32(hw, PFCIM, MYPF_REG(PL_PF_INT_CAUSE));
+
+ ctl = csio_rd_reg32(hw, ctl_reg);
+
+ if (csio_mb_is_host_owner(MBOWNER_GET(ctl))) {
+
+ CSIO_DUMP_MB(hw, hw->pfn, data_reg);
+
+ if (!(ctl & MBMSGVALID)) {
+ csio_warn(hw,
+ "Stray mailbox interrupt recvd,"
+ " mailbox data not valid\n");
+ csio_wr_reg32(hw, 0, ctl_reg);
+ /* Flush */
+ csio_rd_reg32(hw, ctl_reg);
+ return -EINVAL;
+ }
+
+ hdr = cpu_to_be64(csio_rd_reg64(hw, data_reg));
+ fw_hdr = (struct fw_cmd_hdr *)&hdr;
+
+ switch (FW_CMD_OP_GET(ntohl(fw_hdr->hi))) {
+ case FW_DEBUG_CMD:
+ csio_mb_debug_cmd_handler(hw);
+ return -EINVAL;
+#if 0
+ case FW_ERROR_CMD:
+ case FW_INITIALIZE_CMD: /* When we are not master */
+#endif
+ }
+
+ CSIO_ASSERT(mbp != NULL);
+
+ cmd = mbp->mb;
+ size = mbp->mb_size;
+ /* Get response */
+ for (i = 0; i < size; i += 8)
+ *cmd++ = cpu_to_be64(csio_rd_reg64
+ (hw, data_reg + i));
+
+ csio_wr_reg32(hw, 0, ctl_reg);
+ /* Flush */
+ csio_rd_reg32(hw, ctl_reg);
+
+ mbm->mcurrent = NULL;
+
+ /* Add completion to tail of cbfn queue */
+ list_add_tail(&mbp->list, &mbm->cbfn_q);
+ CSIO_INC_STATS(mbm, n_cbfnq);
+
+ /*
+ * Enqueue event to EventQ. Events processing happens
+ * in Event worker thread context
+ */
+ if (csio_enqueue_evt(hw, CSIO_EVT_MBX, mbp, sizeof(mbp)))
+ CSIO_INC_STATS(hw, n_evt_drop);
+
+ return 0;
+
+ } else {
+ /*
+ * We can get here if mailbox MSIX vector is shared,
+ * or in INTx case. Or a stray interrupt.
+ */
+ csio_dbg(hw, "Host not owner, no mailbox interrupt\n");
+ CSIO_INC_STATS(hw, n_int_stray);
+ return -EINVAL;
+ }
+}
+
+/*
+ * csio_mb_tmo_handler - Timeout handler
+ * @hw: The HW structure
+ *
+ */
+struct csio_mb *
+csio_mb_tmo_handler(struct csio_hw *hw)
+{
+ struct csio_mbm *mbm = &hw->mbm;
+ struct csio_mb *mbp = mbm->mcurrent;
+ struct fw_cmd_hdr *fw_hdr;
+
+ /*
+ * Could be a race b/w the completion handler and the timer
+ * and the completion handler won that race.
+ */
+ if (mbp == NULL) {
+ CSIO_DB_ASSERT(0);
+ return NULL;
+ }
+
+ fw_hdr = (struct fw_cmd_hdr *)(mbp->mb);
+
+ csio_dbg(hw, "Mailbox num:%x op:0x%x timed out\n", hw->pfn,
+ FW_CMD_OP_GET(ntohl(fw_hdr->hi)));
+
+ mbm->mcurrent = NULL;
+ CSIO_INC_STATS(mbm, n_tmo);
+ fw_hdr->lo = htonl(FW_CMD_RETVAL(FW_ETIMEDOUT));
+
+ return mbp;
+}
+
+/*
+ * csio_mb_cancel_all - Cancel all waiting commands.
+ * @hw: The HW structure
+ * @cbfn_q: The callback queue.
+ *
+ * Caller should hold hw lock across this call.
+ */
+void
+csio_mb_cancel_all(struct csio_hw *hw, struct list_head *cbfn_q)
+{
+ struct csio_mb *mbp;
+ struct csio_mbm *mbm = &hw->mbm;
+ struct fw_cmd_hdr *hdr;
+ struct list_head *tmp;
+
+ if (mbm->mcurrent) {
+ mbp = mbm->mcurrent;
+
+ /* Stop mailbox completion timer */
+ del_timer_sync(&mbm->timer);
+
+ /* Add completion to tail of cbfn queue */
+ list_add_tail(&mbp->list, cbfn_q);
+ mbm->mcurrent = NULL;
+ }
+
+ if (!list_empty(&mbm->req_q)) {
+ list_splice_tail_init(&mbm->req_q, cbfn_q);
+ mbm->stats.n_activeq = 0;
+ }
+
+ if (!list_empty(&mbm->cbfn_q)) {
+ list_splice_tail_init(&mbm->cbfn_q, cbfn_q);
+ mbm->stats.n_cbfnq = 0;
+ }
+
+ if (list_empty(cbfn_q))
+ return;
+
+ list_for_each(tmp, cbfn_q) {
+ mbp = (struct csio_mb *)tmp;
+ hdr = (struct fw_cmd_hdr *)(mbp->mb);
+
+ csio_dbg(hw, "Cancelling pending mailbox num %x op:%x\n",
+ hw->pfn, FW_CMD_OP_GET(ntohl(hdr->hi)));
+
+ CSIO_INC_STATS(mbm, n_cancel);
+ hdr->lo = htonl(FW_CMD_RETVAL(FW_HOSTERROR));
+ }
+}
+
+/*
+ * csio_mbm_init - Initialize Mailbox module
+ * @mbm: Mailbox module
+ * @hw: The HW structure
+ * @timer: Timing function for interrupting mailboxes
+ *
+ * Initialize timer and the request/response queues.
+ */
+int
+csio_mbm_init(struct csio_mbm *mbm, struct csio_hw *hw,
+ void (*timer_fn)(uintptr_t))
+{
+ struct timer_list *timer = &mbm->timer;
+
+ init_timer(timer);
+ timer->function = timer_fn;
+ timer->data = (unsigned long)hw;
+
+ INIT_LIST_HEAD(&mbm->req_q);
+ INIT_LIST_HEAD(&mbm->cbfn_q);
+ csio_set_mb_intr_idx(mbm, -1);
+
+ return 0;
+}
+
+/*
+ * csio_mbm_exit - Uninitialize mailbox module
+ * @mbm: Mailbox module
+ *
+ * Stop timer.
+ */
+void
+csio_mbm_exit(struct csio_mbm *mbm)
+{
+ del_timer_sync(&mbm->timer);
+
+ CSIO_DB_ASSERT(mbm->mcurrent == NULL);
+ CSIO_DB_ASSERT(list_empty(&mbm->req_q));
+ CSIO_DB_ASSERT(list_empty(&mbm->cbfn_q));
+}
--
1.7.1
^ permalink raw reply related
* [V5 PATCH 8/8] csiostor: Chelsio FCoE offload driver submission (sources part 5).
From: Naresh Kumar Inna @ 2012-09-24 17:17 UTC (permalink / raw)
To: JBottomley, linux-scsi, dm, leedom; +Cc: netdev, naresh, chethan
In-Reply-To: <1348507072-25625-1-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..83dfc80
--- /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,
+ 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
* RE: removing the timer from cdc-ncm
From: Alexey ORISHKO @ 2012-09-24 12:13 UTC (permalink / raw)
To: Oliver Neukum
Cc: bjorn@mork.no, netdev@vger.kernel.org, linux-usb@vger.kernel.org
In-Reply-To: <6020948.EvS2esPpc0@linux-lqwf.site>
> -----Original Message-----
> From: Oliver Neukum [mailto:oneukum@suse.de]
> Sent: Friday, September 21, 2012 10:35 AM
> To: Alexey ORISHKO; bjorn@mork.no; netdev@vger.kernel.org; linux-
> usb@vger.kernel.org
> Subject: removing the timer from cdc-ncm
>
> Hi,
>
> here is the patch that does everything I consider theoretically
> necessary to have bundling of frames in usbnet and adapting cdc-ncm to
> it.
>
> I'd appreciate any review in case I am doing something stupid.
Hi Oliver,
I've tried to apply it, but always got a message that "patch does not apply".
What is the base for this patch?
Regards,
Alexey
^ permalink raw reply
* Re: removing the timer from cdc-ncm
From: Oliver Neukum @ 2012-09-24 12:15 UTC (permalink / raw)
To: Alexey ORISHKO
Cc: bjorn@mork.no, netdev@vger.kernel.org, linux-usb@vger.kernel.org
In-Reply-To: <2AC7D4AD8BA1C640B4C60C61C8E520154A8BED545F@EXDCVYMBSTM006.EQ1STM.local>
On Monday 24 September 2012 14:13:50 Alexey ORISHKO wrote:
> I've tried to apply it, but always got a message that "patch does not apply".
> What is the base for this patch?
Greg KH's tree, specifically commit 6dab7ede9390d4d937cb89feca932e4fd575d2da
HTH
Oliver
^ permalink raw reply
* linux-next: build failure after merge of the userns tree
From: Stephen Rothwell @ 2012-09-24 12:18 UTC (permalink / raw)
To: Eric W. Biederman; +Cc: linux-next, linux-kernel, David Miller, netdev
[-- Attachment #1: Type: text/plain, Size: 1436 bytes --]
Hi Eric,
After merging the userns tree, today's linux-next build (x86_64
allmodconfig) failed like this:
kernel/audit.c: In function 'audit_receive_msg':
kernel/audit.c:753:37: error: 'struct netlink_skb_parms' has no member named 'pid'
I missed the conflict between commit 35ce9888ad2a ("audit: Properly set
the origin port id of audit messages") from the userns tree and commit
15e473046cb6 ("netlink: Rename pid to portid to avoid confusion") from
the net-next tree. I have added the following merge fix patch and can
carry it as necessary.
From 55a9ee70146f37c559333b1f35154618c708203f Mon Sep 17 00:00:00 2001
From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 24 Sep 2012 22:05:38 +1000
Subject: [PATCH] audit: fix up for pid to portid change
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
kernel/audit.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/audit.c b/kernel/audit.c
index ea9b67a..4d0ceed 100644
--- a/kernel/audit.c
+++ b/kernel/audit.c
@@ -750,7 +750,7 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
size--;
audit_log_n_untrustedstring(ab, data, size);
}
- audit_set_pid(ab, NETLINK_CB(skb).pid);
+ audit_set_pid(ab, NETLINK_CB(skb).portid);
audit_log_end(ab);
}
break;
--
1.7.10.280.gaa39
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply related
* [PATCH net-next] filter: add XOR instruction for use with X/K
From: Daniel Borkmann @ 2012-09-24 12:23 UTC (permalink / raw)
To: davem; +Cc: netdev, eric.dumazet
BPF_S_ANC_ALU_XOR_X has been added a while ago, but as an 'ancillary'
operation that is invoked through a negative offset in K within BPF
load operations. Since BPF_MOD has recently been added, BPF_XOR should
also be part of the common ALU operations. Removing BPF_S_ANC_ALU_XOR_X
might not be an option since this is exposed to user space.
Signed-off-by: Daniel Borkmann <daniel.borkmann@tik.ee.ethz.ch>
---
include/linux/filter.h | 3 +++
net/core/filter.c | 12 +++++++++---
2 files changed, 12 insertions(+), 3 deletions(-)
diff --git a/include/linux/filter.h b/include/linux/filter.h
index 3cf5fd5..2ded090 100644
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -75,6 +75,7 @@ struct sock_fprog { /* Required for SO_ATTACH_FILTER. */
#define BPF_RSH 0x70
#define BPF_NEG 0x80
#define BPF_MOD 0x90
+#define BPF_XOR 0xa0
#define BPF_JA 0x00
#define BPF_JEQ 0x10
@@ -204,6 +205,8 @@ enum {
BPF_S_ALU_AND_X,
BPF_S_ALU_OR_K,
BPF_S_ALU_OR_X,
+ BPF_S_ALU_XOR_K,
+ BPF_S_ALU_XOR_X,
BPF_S_ALU_LSH_K,
BPF_S_ALU_LSH_X,
BPF_S_ALU_RSH_K,
diff --git a/net/core/filter.c b/net/core/filter.c
index fbe3a8d..3d92ebb 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -187,6 +187,13 @@ unsigned int sk_run_filter(const struct sk_buff *skb,
case BPF_S_ALU_OR_K:
A |= K;
continue;
+ case BPF_S_ANC_ALU_XOR_X:
+ case BPF_S_ALU_XOR_X:
+ A ^= X;
+ continue;
+ case BPF_S_ALU_XOR_K:
+ A ^= K;
+ continue;
case BPF_S_ALU_LSH_X:
A <<= X;
continue;
@@ -334,9 +341,6 @@ load_b:
case BPF_S_ANC_CPU:
A = raw_smp_processor_id();
continue;
- case BPF_S_ANC_ALU_XOR_X:
- A ^= X;
- continue;
case BPF_S_ANC_NLATTR: {
struct nlattr *nla;
@@ -483,6 +487,8 @@ int sk_chk_filter(struct sock_filter *filter, unsigned int flen)
[BPF_ALU|BPF_AND|BPF_X] = BPF_S_ALU_AND_X,
[BPF_ALU|BPF_OR|BPF_K] = BPF_S_ALU_OR_K,
[BPF_ALU|BPF_OR|BPF_X] = BPF_S_ALU_OR_X,
+ [BPF_ALU|BPF_XOR|BPF_K] = BPF_S_ALU_XOR_K,
+ [BPF_ALU|BPF_XOR|BPF_X] = BPF_S_ALU_XOR_X,
[BPF_ALU|BPF_LSH|BPF_K] = BPF_S_ALU_LSH_K,
[BPF_ALU|BPF_LSH|BPF_X] = BPF_S_ALU_LSH_X,
[BPF_ALU|BPF_RSH|BPF_K] = BPF_S_ALU_RSH_K,
^ permalink raw reply related
* Re: [PATCH net-next v4 0/1] Add support of ECMPv6
From: Nicolas Dichtel @ 2012-09-24 12:28 UTC (permalink / raw)
To: David Miller; +Cc: bernat, netdev, yoshfuji
In-Reply-To: <20120921.134833.1294721941180523200.davem@davemloft.net>
Le 21/09/2012 19:48, David Miller a écrit :
> From: Nicolas Dichtel <nicolas.dichtel@6wind.com>
> Date: Fri, 21 Sep 2012 11:59:04 +0200
>
>> v4: remove compilation options to choose multipath algorithm for next hop
>> selection. Now the choice can be done at run time via
>> /proc/sys/net/ipv6/route/multipath_algorithm
>
> Please specify this in the routing configuration protocol, rather than
> via some obscure procfs file.
Just to be sure to understand, the goal is to configure the algorithm when the
route is added? Thus, resurrecting RTA_MP_ALGO and having one algo per route?
^ permalink raw reply
* Re: [PATCH net-next] filter: add XOR instruction for use with X/K
From: Eric Dumazet @ 2012-09-24 13:43 UTC (permalink / raw)
To: Daniel Borkmann; +Cc: davem, netdev
In-Reply-To: <20120924122359.GB25156@thinkbox>
On Mon, 2012-09-24 at 14:23 +0200, Daniel Borkmann wrote:
> BPF_S_ANC_ALU_XOR_X has been added a while ago, but as an 'ancillary'
> operation that is invoked through a negative offset in K within BPF
> load operations. Since BPF_MOD has recently been added, BPF_XOR should
> also be part of the common ALU operations. Removing BPF_S_ANC_ALU_XOR_X
> might not be an option since this is exposed to user space.
Please note we dont expose BPF_S_ANC_ALU_XOR_X to user space.
We expose SKF_AD_ALU_XOR_X instead.
But it seems easier to leave it to keep this patch small (not touching
various JIT implementations, even if followup are welcomed)
Acked-by: Eric Dumazet <edumazet@google.com>
^ permalink raw reply
* Re: [RFC PATCHv2 bridge 7/7] bridge: Add the ability to show dump the vlan map from a bridge port
From: Vlad Yasevich @ 2012-09-24 13:48 UTC (permalink / raw)
To: Ben Hutchings; +Cc: netdev, shemminger
In-Reply-To: <1348334132.2521.100.camel@bwh-desktop.uk.solarflarecom.com>
On 09/22/2012 01:15 PM, Ben Hutchings wrote:
> On Wed, 2012-09-19 at 08:42 -0400, Vlad Yasevich wrote:
>> Using the RTM_GETLINK dump the vlan map of a given bridge port.
> [...]
>
> This enlarges the RTM_GETLINK response quite a bit. I think perhaps
> this should be optional, like IFLA_VFINFO_LIST is now.
>
> Ben.
>
Only for AF_BRIDGE and only if VLANs are set. I guess I could add
a filter option similar to VFINFO as well, if you think that's necessary.
-vlad
^ permalink raw reply
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